Please use this thread to discuss findings, that can be made public. Save that other massive thread.
Please only comment if you are a beta tester, Work for Giv, involved in the coding.
Others please read away, but leave this thread for its intended purpose.
Beta version as of 12Dec
Inverter D0-292, A0-292
BMS - 3092
.
Hi all, anyone else been added to the testing? Not heard anything back yet and I'm curious with tomorrow slated as the date of general release.
@mpartington is one I know of
@hoggy maybe another when I asked Dan
Have fed back in a cold related issue and reserve behaviour.
Normally not seen, but below 0C is having odd effects
No beta firmware for me yet I’m afraid.
I got inverter firmware Tuesday afternoon and battery firmware yesterday morning. Not much to report, values all over the place as haven't been able to cycle my battery. This has mainly been due to the low charge rate of 1kW imposed as battery was around 6 degrees last night, meaning I could only put 4kWh in, rather than usual fill. I think the limits for charging at low temperature seem very restrictive, as they are not sufficient to generate the heat needed to increase the charge rate, so you get stuck in the cold zone. Sure I read that LiFePo4 should be good for higher charge rates down to 5 C, so hope this is something that could be relaxed. Looking at the old software, even at full charge it seems to take a while to warm the cells (not BMS) from 5 C to over 10C. Honestly unsure what useful data they would get from mine in such a short space of time, unless it was more a test of remote installation.
Same firmware as TheDragon
That reduced charge rate would be a massive issue for me. I currently charge 17.7kw of battery from a single AC3 with a 5 hour window. I currently get 69% charge in, if they cut the charge speed to 1kw..... Bad. Very bad 🙁 Will find out forsure tomorrow I guess.
The current low temps are very unusual for the UK. Its not like we get them regularly, is it?
Zakalwe According to the Met Office website the weather station nearest me (Scampton) shows winter monthly average minimum temps of:
Dec: 1.12c
Jan: 1.02c
Feb: 2.26c
Zakalwe looking at my history, my garage has been hovering around 3 degrees most nights (so not extreme), and haven't seen the BMS temperature over 7.5 degrees at time of charge at 21.30, could be worse at normal later go time of 1.30. Only been an issue last night following the firmware update. I worry that even typical winter temperatures could significantly limit overnight charge rates on systems not in the house. I really hope I'm wrong and it is only a few nights a year.
I've found a wierd like the exact opposite of what should be happening
At 0 and below we all agree charging is halted.
At 0 to 8C I get full speed 3iw
8 and above it drops to 1660W
Which is like wierd.
Mentioned it as this isnt correct.
If anything 0.1C 1660w for me from 0 to 5
Then allow it upto 0.2c from 5C 3kw for me
Above 10C you can go higher still
Charging and discharging is measured as a portion of capacity in Ah
3 x 5.2 is 306Ah
0.1C is 30.6A x 52ish
0.2C 61.2A x 52 ish
TheDragon (GivEnergy) update.
Not a simple temp issue.
Did a charge to 60%. Noticed it had dropped to 1600w
Discharged down to 35%. By now were 8C and rising
Charged again, full speed 3kw till SoC hit 40% then dropped to 1600w
So some logic a miss here.
Is there any pattern once you look at the cell (or battery) voltages along with the temperature?
THALL same form me as well as I charge the batteries over a 4 hour window - don't think I'll be updating anytime soon until that's sorted, unless it's automatically updated.
Defeats the purpose of requirements, as cold weather I generally need more and hope this doesn't bork it!
TheDragon (GivEnergy)
Which battery temp are you monitoring? Mine were 6 Deg in GE app yesterday and charging at 0.1C. Started same rate tonight. Started at 1.08kW at ?? Deg, now dropped to 0.863kW (temp currently 8 Deg). I wonder if I'm near full @54V and it's slowing.
Update: (not sure how to add screenshots) the charge profile last night is either very clever, or very lucky (thank goodness the battery is also charging from solar during the day!).
21.31 : charge schedule started -1080W
22.55: drops to 864W
00:57: drops to 339W
01.23: drops to 0W- voltages peak at 56.93V and 56.6V
01.29 schedule ends
3.4kWh added in period. Voltage profile show full charge was achieved
Now did it know I had a 4 hr schedule and map this out perfectly, or is the final charge to 100% very very slow, in which case will it ever get to 100% in a 4 hr window? As part of the software is to balance and run checks whilst idle, this idle period is important. Of course this could just be part of the initial system learning and next time will be much faster as this is the first time I hit 100%.
Peak bms temp rise from 8.4 to 9.8 degrees. Reduces from 22.55 as charge rate decreases.
I'm feeding all this back though the standard support email address (as reply to original email from GE) so hopefully that's the correct route. Although will be too late for today's release 🤔
TheDragon (GivEnergy) So you have found a bug in the new firmware a day before mass release? Fingers crossed this is something simple 😄
TheDragon (GivEnergy) Can you keep quiet about this until they've released it please as my battery is nice and toasty indoors so I don't care about temperatures 😆
I monitor 3 packs, each has 5 sensors
1 BMS and 4 Cell sensors. For the Cells the 1 and 4 look to be nearer the outer edges, 2 and 3 are much slower changing, so assume in the middle of the block.
Ha Yeah, only issues so far are all temp related, which clearly couldn't be tested in the lab easily and the temps dropped severely only recently
Well no new version issued to test, post finding these odd behaviors at low temps.
Could be a good sign its an easy fix, or it will get done later in a minor release update.
Need to now see if the issue goes away on Monday when it warms back up. Or someone with indoor batteries is testing and not seeing it.
TheDragon (GivEnergy) that's a nice but of code, how do I replicate those battery monitoring graphs? The standard charts are about the limit of my HA dashboard capability 🤣.
mpartington
The standard ones are ok, these are Mini-Graph from the Lovlace HACS store
@mpartington
That is pretty much how Li batteries charge when close to 100%
ispookie666 Yeah, think this is throttling a bit too early. But could just be cold related
ispookie666 think it must be part of the new battery calibration method. It's just charged back up off solar from 95 to 100% and not taken anywhere near the length of time. Looking at the point the charge reduces to 864W , this took 90 mins yesterday to drop to zero (battery full) and only around 30 mins today . That was it's first trip to full, so think it was trying to learn the 100% point.
mpartington
I don't have surplus solar 😅 . Wont see surplus solar till March
Have you done any reaction time testing yet? How fast it now reacts to the change in load or solar?
TheDragon (GivEnergy) not yet, GivTCP was hanging for long periods, which I put down to inverter in debug mode. All seems much better now though. What sort of tests do you do? Wait for it to go idle and turn a kettle on? Anecdotally, the wife was ironing yesterday, and the smart meter display didn't seem to be on for as long as usual whilst the iron was cutting in/out.
Yes, put GivTCP in a fast refresh 10s or less.
Watch the grid load, turn the kettle on, if your lucky you'll see the spike. Previous was a good 2 or 3 refreshes
Or Press Pause. Wait for the grid to show on your iHD/GivTCP. Then set back to play. 5s or less it'll be on battery fully
What's the experience on accuracy for multi packs? I hope is still just that it's learning, but a little disillusioned at the moment after all the positive comments. 2x5.2kWh pack, battery thought it was at 98% at end of charge period (it's hit the top point 3 times now), however only managed to get a net discharge of 5.7kWh. Even considering dod, the cold etc. it's a lot less than I saw on the old firmware. To be fair, this is the first time I have 'fully' discharged with this firmware, but the battery voltages are now sat at 51.125 (20%) and 51.092 (14%). We were cooking, so they were lower at the time discharge stopped, but I thought it had compensation for that. So clearly not empty, although Inverter says 6%. Put a couple of minutes charge in and restarted the inverter, which brought it back to life for a few minutes, but now idle.
mpartington
Just a guess, but I would expect that a 'learning' system such as this, would probably need at least 10 days/cycles to get to a reasonable level of calibration, with incrementally smaller benefits coming after that (plus ongoing adjustments to account for all of the other variable factors, like usage pattern, battery aging, weather/seasonal changes etc.).
Thanks for being a beta guinea pig, by the way. I offered but never heard anything back.
ToothyChris Hope that's the case. Put another 5 mins charge in and left battery on pause. Now stabilised at 51.61V on app, inverter 3% and idle. Restart has no effect. Hopefully it is smarter than im giving it credit for, but it should be more like 20%. Really hope it doesn't think 20% is zero in the calibration, or will be fun few days. Just trying to keep if above 10 degrees now with burst charging, so have good charge rate to fill it later.
They only asked me on Wednesday, so think the chances of delivering a production version yesterday were always slim
mpartington
Yep, when I saw that a few people had been added to the beta last week, I pretty much knew that the Friday launch wasn't going to happen. A lesson that I hope GE will learn from....
mpartington
Monday Tuesday will be a good comparison as it warms up again.
I was getting g 12.6 to 13kwh into my 3x5.2 which was about acceptable.
Now full at 11.6kwh. So about 10% capacity down. Ties in with the difference between 20C and 0C capacity charts. But China are investigating this too.
Stable as rocks at the moment, all 3 tracking well. Within 4% of each other from 100 to 4 and back again.
Try leaving them idle for an hour to let OCV do its job
Just curious TheDragon (GivEnergy), do you have multiple batteries on a single inverter (AC or hybrid?), or do you have multiple inverters, each with a battery?
Just curious to see if the software is likely to help the battery SOC balance, with multiple (AC) inverters....
ToothyChris single inverter, multi battery AC3
TheDragon (GivEnergy) what's OCV? I'm probably being too impatient, but on the plus side above 12 Deg was giving me full charge rate. Thinking between 10 and 12 are 0.2C. It was drawing 100W, but seems to have stopped now The change in capacity is odd though, as you would think you would see that kind of temperature effect in the old firmware. It definitely seems to be going down faster with use now than it was, compared to the start of the week (3007 firmware).
Just noticed it's gone from 4 to 8% and started discharging againwithout intervention. That's a good sign 👍
mpartington that's the OCV (Open Circuit Voltage) calibration.
When idle it will disconnect the batteries and test the voltage.
Then adjust the SoC. For each, when complete the Inverter adjusts its value and if above cutoff, power is again available.
Then you can see, ooh I've still got power, and it springs back into life for a bit.
Only when OCV calibration agree within 4% of each other and below cutoff. Will it be the end.
All this happens automatically
TheDragon (GivEnergy) sounds clever stuff! Well I'm happy, currently charging at 3kW, with a start temp of 11 Deg. Think with the big temperature impact on capacity and charge rate, my next project will be some removable insulation for the batteries for winter. Might hold off and see what people have most success with and copy
Remember back in the summer was getting 13kwh in my batteries now 11.6.
The losses overall are still crazy though. Losing 2kwh on a full cycle.
Done 2 full cycles today and charge kWh is 4.2kwh higher than discharge kWh.
So wish I had got 2x 8.2s instead of 3 x 5.2s
80% DoD sucks
TheDragon (GivEnergy) I'm sure mine were previously operating at 90%. One of the lead installers (Ivoltz) with very close ties to GE said they had been updated with a firmware update, despite white the data plate says. Bet that's been reversed in this update. I did notice he posted more recently they were only 80%, so who knows.... feels more like 80% now
Just seen that step back in charge rate from 2880 to 2164W , despite battery being at 15 degrees. There's a chap on the Facebook group who's also seen it with battery indoors, so not looking like low temperature is the only issue. Might be reading more into it, but looks like it's backing off to maintain the voltage below 54V?!
Update:
Still can't work how to add photos ☹️, would make thinks much easier. Any ideas?
Charge rate and lowest battery temp of the 10 sensors quoted:
21.30: charge started 2880 W (10.7deg)
22.35: throttles to 2170 W (14 deg)
00.26: starts reducing linearly to 1081W by 00.42 (17 Deg)
00.26: also point of peak voltage (54.04V and 54.18)
00.57 starts to drop gradually to zero (16.5 Deg)
01.10 charge complete
01.30 my scheduled charge period ends
From perceived empty, added approx 8.3 kWh (ref GivTCP).
Clearly not cold related, but could it be mapping out the charge rates based on required charge and available time? Rather than shoving in in as fast as possible? (My bet is still on a bug 🙂)
Also battery SoC at least 5% higher this morning, no reason to think over night load was any different, so feeling more positive about the firmware. Still got a 7% difference between individual batteries though (which is more than yesterday). One only reached an indicated 90%. Hopefully that will continue to reduce with time.
mpartington
Images upload to IMDb
Select below the picture icon
Then select html code full. Paste the bit after src.
Paste it where it says https. In the square brackets give the picture a header if you wish
Internet movie database?
Step 1: Upload to postimages.org
Step 2: Click on the icon to the right of the Markdown line.
Step 3: Return to this page and hit Ctrl C.
Easy peasy.
chris_p my wine infused typing.
ImgBB
Zakalwe here goes 😉
I've just noticed this is a slightly different flavour. The examples provided previously show the charge rate dropping from 0.2C to 0.1C, whereas this is a drop from max to 0.2C.
Previously we were seeing the throttling down to 1kWh from 0.2C. Now it is throttling to 0.2C, so wonder if it is a simple case of when it hits a temperature limit where it's supposed to increase charge rate, it sets the lower limit instead. My battery was about 14 deg at the time it throttled, so interesting if any internal temperatures above 14 at begining of charge start at the 0.2C limit rather than max (might be waiting a while to test that one).
Although I would have thought that would have been obvious in the in house testing. Maybe it actually needs to move from a lower limit to a higher one, as doubt they tested below a cool room temp. Will check tonight if it throttles at the same temperature.
mpartington yep van see your throttling is less than mine. 3kw to 1.6kw.
Your looks 3kw to about 2kw.
My packs are 306Ah so 0.1C is around 30.6A 1660W
TheDragon (GivEnergy) based on the battery voltage chart you posted, is a buffer of 20% being applied at both the top and bottom limits?
Battery was resting 1.10 until 1.30 (after reaching 100%) assume it's the voltage value just before it starts discharging?
https://postimg.cc/XBnTVN2k
It's now finished for the day
https://postimg.cc/gwGrbGYG
6.8kWh discharged today (value from portal/GivTCP), which is much better than yesterday, but still low.
App says 7.26 kWh discharged, which is odd, not sure why it would be different.
Any updates peeps?
Expexting another version today. With my detailed summary items addressed. I will re check, report back to GE.
The staged charge, in the cold, i know was bugging me, likely never see cold like that again. But it was sorted.
Didnt get any new FW to try. Maybe it'll be in my Christmas stocking with the lump of coal.
BMS 3094 just landed.
Charge speed restrictions removed as requested.
A few other tweaks for batteries than have no idle period to allow the self calibrate at idle to function
Will report back findings
[unknown] Sounds like it is getting close to release with each iteration.
[unknown] Excellent thanks for the update. Cheers
Keep an eye out later today, 3094 should be going to those taking part.
Please feed back to Dan/engineers via the betatester email any issues, good or bad.
Anything you like is just as valuable input as the bits you dont.
TheDragon (GivEnergy) How's the latest version going from your experience, please?
Going fine. Testing some .minor iterations of 3094 logic. To see benefits or not.
The official response will come from GivEnergy on when and how it will roll out. Some older Hybrids may need a modification or not able to load it
TheDragon (GivEnergy) Sounds good. What kind of modification? Memory size too small?
Yay for being an early adopter!
hoggy No yay. You got no Cap either I take it?
BMS 3096 landed this am. My main niggle around reserve may finally be nailed.
It's taken a lot of work and head scratching, but be jesus we might have nailed it.
Capacity issue, may not actually be an issue after all. It seems to be linked to the Pinv and Pbatt values. Which are reporting wrong since changing from inveter 535 to Beta 292.
TheDragon (GivEnergy) is that an inverter software update to fix? So far so good with 3096 for me
3097 - addresses the calibration running too often, for those who are idle for long periods, there is no need to re-calibrate every hour. ie those using battery for EPS and not charging routinely. Battery is there for backup only.
Reserve is maintain by small bursts of trickle charge, interleaved with OCV calibration. So self discharge (Powering the inverter) doesn't drain the batters too far below your set level.
So OCV is now limited to 2x per 24hr period.
Battery upper threshold altered and some logic added to ensure all batts charge to the upper capacity, even if one gets there 1st
curious folks if gen 2 invertor firmware allows 4000w battery charge/discharge?
michaelmcgoldrick 3600 max on the datasheet
THALL
I think that question has come from a rumour the 5kW gen2 inverter might be getting a firmware upgrade to support it.
94jdh Hmm not sure GE would set that. The range on the 9.5kw battery is 3600 - 4700w. I guess they might set it and document it as "up to" 4000. But then will that be on Gen 2 flat front or curved.... does the curved have newer better guts than gen 2 flat. Would be great for a lot of people if they did allow it though!
On D0.290-A0.290 (battery 3011) since yesterday. AC3 Gen1 and 5.2Bat
Was at 84% before the firmware update and 0% after. It did get up to 100% with solar and some grid import, although seemed to get to 0% very quickly yesterday empty by 5pm (most likely not true 0!) without any big appliance use.
Overnight it put 1kWh into the battery, supposedly getting to 100% but as of 0819hrs this morning I'm at 5% battery.
Hoping that things improve after a few cycles, but a little worried so far!
Edited to add : According to the app, the battery voltage is 52.64, 19c. It's been idle for 40 minutes (dropped from 5% to 4% during that).
TX200 I'm my experience it does nothing for the first 40 mins, whilst it lets the cells rest. It should then start the process. I asked for that I inital time to be shortened, but haven't seen any changes to that piece of logic
Just to repost from the other thread, this is the result of my 3rd full cycle. Plus my 100% down to 5% capacity appears to be 11.96kw. I have 17.7kw of battery storage.
I am not on the latest BMS firmware, 3088 for me.
As I mentioned in the 'other' thread (I was still posting there as it seemed far more active than this one), I have just requested a calibration which I've been told should be done within the next hour as but they're waiting to hear back from the technical director to make sure they're happy with them doing it. Will let you all know how it goes from there shortly!!
Ra Their support is working today? I don't think I have ever had a reply on a weekend before. So thats good!
THALL Messaged them on Facebook as I'd seen other people do it in the past 😉
Ra
Once you've done a full calibration run.
Please let us know.
How the next regular cycle goes
Today is a lazy day. Just a trickle charge in all day.
Set Reserve to 100% and SoC was at 4%.
So it will just slowly charge 290W
mpartington it did a jump from 5% to 100% yesterday around 1pm. Guess it might not do another recalculation until the same time today. Who knows where it'll end up. 🤣
We volunteered to take part in the Beta testing a few days ago and have just had a reply asking us to confirm if we still want to take part.
However, the email says the following: "* please can i ask that we use this email chain for the beta testing only and any other support requests go to support@givenergy.co.uk * but the reply was sent from "support@givenergy.co.uk".
Who are others responding too, is it support@givenergy.co.uk or the original email address that they said to express interest on "betatesting@givenergy.co.uk", or a different email address for feedback?
Woodlands I've just been replying to the email that the initial beta contact was made from.
Woodlands I have been sending emails to both addresses and kept that chain going.
Mine was also sent from support@
Guessing by chain they mean the email subject? They're using freshdesk for support, so wouldn't be surprised if it's linking the emails into a ticket by sender and subject line?
TX200 I've assumed that too.
Calibration has been running for about 90 minutes now and down to 50.59V....
Just cooked lunch and got the iBoost running to try and make sure I'm not wasting any energy I could use myself 🙂
I am not on any social media so not sure my calibration email request will be seen until Monday.
THALL Mine was done via the beta testing mail chain.
judgepd Ah thanks for that. Fingers crossed mine will be seen.
same here. message received from support not betatesting@givenergy.co.uk but it is clear they want replies to the beta email to me.
Not sure if my calibration is working properly. It emptied the battery at maximum export and sat drawing a few hundred watts at 0% for several hours without changing the SOC. It's now stopped drawing at all and is just sat idle at 4% and 50.85V. I assumed it would refill as part of the calibration process.
they have said this to me.
"our support software will keep all replies in the chain in the relevant group under the same ticket number, meaning that i can keep track of everything.
if you send a new email to support that will go into the support group of the system.
I replied just before 3pm today to say that I was content to take part in the Beta testing, but from checking the System data it appears that something changed about 2:30pm, as 5 second monitoring appears to have started from then.
However, the battery was at 100% SOC, because of some sunshine today, but it has now dropped to 0% and is charging from the grid.
Is that what others have experience. No email to confirm when they take over and a full battery emptied in a few seconds.
i did get an email saying my system had been updated though perhaps because i had queried the email address?
Well, I was hoping my battery percentage might adjust this lunchtime (it did yesterday), but nope. It started charging around 1pm to get from 2% to 4% and is now at 3%
Hoping tonight's timed charge is bigger than yesterday's 1kWh. 🤞
Woodlands same here. The inverter was updated first at around 14.30. Saw the fast response time straight away.
Then around 3pm, it fell from 60% to 0.0%, currently charging at 300w and hasn't moved in an hour from 0.0%.
Guess this is considered normal?
mine updated at 3.15pm
soc was 66% at update
dropped to 60% then jumped to 100
still saying 100% an hour later despite constant approx 200w draw for almost an hour.
2x 8.2kwh 100% DOD on an ac3
Inveter bug report submitted. Relating to timed charge, discharging during the charge window.
I was able to replicate a users fault, so I now seem this a bug.
If the inverter is set to charge to 100% for example 0000 to 0600, the inverter finishes charging 0430, it remains idle till 0600.
If you set the charge to % to Amy other value not 100% the same is not true.
The inverter begins charging at the required time, once the charge level is reached, it stops as expected. Then however it begins discharging immediately to meet home demand for a few seconds maybe a min. Then stops and charges again back to the set level.
Stops, begins discharging and the cycle repeats.
Alternating charge and discharge at the set level till the charge time ends, when it returns to normal Eco/Dynamic mode.
Attached is a test I replicated from another users questions. With the time set, charge finished around 1615, and from the 2nd picture see the discharge, charge alternating.
I was able to exactly repeat the issue, hence I am now reporting it formally as a bug.
Paul.
TheDragon (GivEnergy) Is this really a problem as the discharge/recharge might keep the battery temp up in cold weather?
OK, I admit it. I am thoroughly confused and in the dark as to how the beta firmware is expected to behave. This is making any meaningful contribution to system testing difficult to say the least. Can anyone with definitive answers (not conjecture or theories) to the following questions please enlighten me.
Can table look up and SOC value correction occur any time the inverter is idle or does the battery also have to be at specific state of charge (full, 100%, 99%, 98%, empty, reserve threshold, 4%, 5%)?
Having gone into an idle inverter state is there a waiting period (and if so what is it) before table look up takes place?
Having done a table look up is there a waiting period or other criteria that have to be met before SOC update commences?
Is it expected that the SOC will get updated very slowly and in small incremental steps and if so why? (It is tempting to think that this is so the user does not see sudden jumps in the SOC value but surely this cannot be the only reason).
Having commenced the SOC update how long does it carry on for ( until the SOC eventually matches the looked up value, a certain number of iterations, a certain length of time)?
Having commenced the SOC update can it be interrupted at any time by a call for battery charge/discharge taking the inverter out of idle?
Is there a minimum time that has to elapse between one table look up and the next?
How can you tell that a table look up and/or incremental SOC update is taking place at any given moment in time?
Can you initiate a table look up and/or incremental SOC update manually?
If the SOC value increments towards the looked up value so slowly, is subject to interruption and can only restart after a prolonged period; are we all wasting our time?
I really want to help in this process but, as you can probably tell, I am finding it rather frustrating!
judgepd So it only did the empty part of the calibration and hasn't refilled. I'll be interested to see how much the battery takes when it charges tonight.
I'm afraid GE haven't managed to turn my frustrations with them around with this beta firmware and it seems far from ready for release for general release at least on my system at the moment. I'll keep giving them feedback for what it's worth.
TheDragon (GivEnergy) not sure if that's a new bug, pretty sure it's been around since day 1 Pre all this software change.
Oscillating charge? like this kind of thing?
Think it's a carry over from the old "Winter Mode"
Woodstock
1 seems to be under 50% and at full
2 30 mins. But the readme indicates 60min. Not seen 30 I'm seeing idle and the SoC starts to alter.
The rest I don't know answers to.
hoggy interesting read that old thread.
While I've got the Devs attention, it might now get fixed.
TheDragon (GivEnergy) Looking at that old post of mine, I raised a help call at the time and GE said it was expected behaviour and couldn't be changed.
They were only able to update the Inverter Software to D/AO-290 so still waiting to receive the battery software file so I can update the battery software, as they can't do it remotely.
Battery dropped from 100% to 0% at about 15:30 and has been charging from the grid at about 290watts for the last 6 hours to reach 4% and seems to be sticking there.
The 100% to 0% was in a few seconds, with no export to grid or use, so hopefully should still be in the battery, once the software recognizes it and allows it to be used.
Battery Voltage is reported as 55.21v, which I think should represent 100%.
If it doesn't jump back up soon, I may switch it all off, as don't want to risk it going too high overnight.
There are a few people now running slightly different versions of the beta firmware. I'm looking for experience on a particular issue to try and identify if it's a common issue and if so where it crept in.
The firmware has some.logic at empty (possibly below 50%) when battery is idle. If the SoC is lower than the recorded idle voltage, then it will increase the SoC to match.
On 3011, I've seen this not working correctly on my installation (2x 5.2). It seems to do the opposite i.e reduce the SoC to well below the SoC based on voltage. Tonight my batteries are reading empty, but the voltage suggests they should be at 30%. They haven't increased as expected. I'm wondering if this could be linked to some of the capacity issues people made seeing.
Has anyone seen this working correctly on earlier versions such as 3088 ?
Luckily, it has now switched back to showing 55% and discharging to house load.
All green LEDS alight on the master 8.2kWh battery, but only one green LED showing on the slave 8.2kWh battery, plus green status lights.
mpartington I see a very similar pattern on 3088. Battery hit the reserve tonight and voltage was 3.2V. After 45mins or so it jumped the SoC up a few % and then drained down again. I agree with your thought that this is likely the reason for the lower than expected capacity.
I'm interested to see the changes for those who have had a calibration run done, as I'm not sure the corrections it makes are great for handling a system that starts off well out of sync. A reset might improve matters.
My battery is now charging from reserve and so far it seems to be taking a lot more than it has the past few days. Perhaps I just needed to hit the bottom.
mpartington
Let us know how reverting to 3094 goes.
TheDragon (GivEnergy) how many Watts are we talking on average once the target SoC is achieved? On old (July 2022) software, I was seeing about 30W reported by the energy meter stats.
mpartington I'm on 3011 (1x 5.2Bat). Second overnight charge since applied. It's put < 1 kWh in overnight.
Believe it's under reading too, when the percentage is around 5% and has been idle for some time, the voltage seems to suggest around 40%
My battery was 84% full when the update was applied and 0% immediately after. With a bit of grid and solar it charged to 100% and updated it's soc. But it was apparently empty by 5pm that day and empty circa 8am the day after.
Well, I'm going to say it. For me, this firmware is a disaster. After a supposed calibration yesterday, it took 3.4kWh last night from empty to full on an 8.2kWh battery. It's already down to 75% and only discharged 1kWh which means the capacity is still massively down on what it should be. It also didn't sit at 100% until the timed charge period ended, it constantly discharged and charged the battery like it's done before at SOC lower than 100%.
An awful lot needs fixing.
judgepd agree something has gone badly wrong in this release. It has been getting better and better, but not this one.
I decided to revert to an older version 3094 (where I know the OCV works) last night, (around 10pm) the battery SoC immediately jumped to where they should have been and I added around 6.3kWh, when I'd have expected at least 8 kWh (ideally more) charging from empty to full. This confirms.the obvious that the batteries were actually around 30% full, but the system thought they were empty. The top points also corrected and now it is all looking well aligned again.
Will see how this goes for a few days and want to do a full cycle today, but have already fed this back as well.
TX200 Does seem like it is wrongly calculating the empty point. I'm wondering if each time the battery is seen as empty and idle, the OCV is kicking in as it sees the error, but instead of correcting it, it makes it worse. So every iteration, you loss more and more capacity. Definitely something that needs fixing ASAP, or people.rolling back to the old version / an earlier beta.
3011 may be working better (initially) for those of us that have been on the trial for longer, as I think Paul has said the calibration points are stored and not lost, so are reused. Whereas yours would be defined for the first time by 3011. Mine screwed up when these were changed by the OCV function running.
All just a theory though, hopefully GE can tell us the reason
mpartington I'm now at "15%" battery and will shortly be emailing some more (polite) feedback.
Question for those running Inverter Firmware Version D0.290-A0.290 on an AC3 with dual 8.2kWh batteries. Is it this Inverter Software that makes the system react quicker or a later version than 3007 of the BMS software, or a combination of both?
Woodlands
It's the inverter firmware that's making it quicker.
I dont think the new inverter firmware is compatible with the previous BMS versions. Why, I dont know, but they go together.
As Matt says 3011 seems to have gone wrong, compared to the steay improvement seen in the 309x series.
I have fed back my feelings on 3011, the shortfalls seen and 3094 seems to be a good starting point.
3088 is ok, but capacity down for me and offeres no improvement over 3094; others seen more severe capacity issues.
Waiting for the next move.
GivEnergy are seeing all feedback, Dan M is collating it and feeding up. So it's not ignored for sure. Due to the number of testers with the firmware, it's a heck of a job. So bear with him
TheDragon (GivEnergy) Thanks for that explanation. We currently have Inverter Firmware Version D0.290-A0.290, but still running 3007 BMS, as we are waiting to receive an updated BMS software file, as Dan was not able to update it remotely yesterday.
Noticed this morning that it appeared to be re-acting quicker, but just wanted to check.
Yesterday it dropped from 100% to 0% instantly and then took over 6 hours of charging at 290 watts to then suddenly jump to 55% and then about an hour later to 100%.
On 3011 myself - all working fine so far. Definitely more stable. Assume coincidental update to mobile app, but that seems much more stable than it's been in the recent weeks, so think GE gradually getting the ship back in the right direction.... now if they could just deliver me the Gen 2 inverters I ordered back in July 2022 that are now not appearing until April 2023....!
Givenergy 5kw Gen 1 (waiting for the never-appearing Gen 2 as ordered!)
2 x 9.5Kwh Batteries
Woodlands
Shame there isnt an official repository to pull them from.
I have every version since early Nov. But not 3011 (Done remote), which I don't like anyway.
3094 was a good one for me and Matt. Capacity was better, not up with 3007.
Just my update from yesterday. As previously mentioned I was seeing about 1kWh capacity from my 8.2kWh battery so I requested a calibration cycle to be carried out. This commenced at 11am Saturday and finished at 11pm that night.
It took quite a while to discharge down to 0% as I suspected it would but managed to have my iboost on and washing machine etc so this energy wasn't exported 🙂
The battery now seems to be behaving itself in terms of capacity/discharge expectations but will see how it goes as it has only been 12 hours.
So far so good.
ac3 with 2 x 100% DOD (depth of discharge) 8.2 kwh batteries
D0.290-A0.290 and 3011 BMS (battery management software)
recap- mine updated at 3.15pm on the 14th
soc was 66% at update
dropped to 60% then jumped to 100
still saying 100% an hour later despite constant approx 200w draw for almost an hour.
2x 8.2kwh 100% DOD on an ac3
I decided to try and empty the batteries so at 00.30 hrs I started emptying them into the immersions at 3kw.
at 03.15 they had dropped to 21% at which point I gave up and set them to charge till 07.25. target 100%
reached 94% at 07.25 (grid charge stopped)
92% at 08.30
98% at 10am due to some good solar power and also reported 100% for just 5 minutes then dropped 98% again
98% at 10.10 discharging 300 watts despite 1.2kw solar available.
97% at 11.00 discharging 116watts. 1.5kw solar available
currently 97% and idle/1.5kw available.
Ra Did yours refill as part of the calibration cycle? Mine just emptied and didn't recharge so is still messed up.
judgepd If this firmware is the fix then not to worry as warranty should mean a new battery for you.
So last night, soc jump of 12% 88 - 100. The odd 2% dip from 100% has gone. Discharge so far 12.1kw today and down to 22%. In short its a bit better, but at this rate of learning I think another 5 or 6 cycles for me. 10 total, 10 days. Oh and I did not get a calibration.
THALL Don't follow sorry
judgepd Sorry, I mean if this firmware is meant to fix all issues. But the result for you is only 20% of battery capacity usable then GE will have to swap it under warranty.
THALL Ah, I see.
judgepd Do you see any differences in capacity between the first charge on the beta vs the latest? I was very worried Friday/Saturday. but suddenly lasts nights charge cycle is better. 80% discharged says 12.7kw 80% should be 14.1kw. Taking out the bit of solar that went in I am missing 15% ish capacity. Much better than over 30% yesterday.
THALL Capacity still seems to be somewhere between 3 and 3.5kWh. It might have improved fractionally but will take a month at this rate.
Todays data. 100% down to 5% total discharge without solar charge 14.5kw from a possible 16.8kw assuming the 5% is being kept as reserve. Sadly the system will not fully charge tonight but this does appear to be better. If it gets to 16kw discharge i'll be happy.
THALL starting to think I'm doing something wrong.
judgepd I think where it went wrong for me is the firmware assuming my battery was at 0% after it was applied. Think I've got half a battery available but it won't go below what it thinks is 3% to get to it.
If you can, install the BMS when battery is flat.
Or ask for a calibration.
Will give you a level playing field to begin with, instead of trying to balance out any number of reasons why the SoC is different.
@judgepd have you access to GivTCP? Are all 16 cells in each pack the same?
You might have a dodgy cell connection, or worse still a bad cell.
Previous FW just counted A in and A out.
This does that aswell as periodic checks it's still in check.
If a cell is out, it will upset the readings, which before might have gone unnoticed till the cell failed. Now itll catch issues earlier.
Ask Dan Maddocks to check your cell alignment aswell; if you are unable to check for yourself.
This screenshot shows 3 packs, each has 16 cells, looks as 1 line as there all perfect. A bad cell would show itself very obvious here.
Note the timings each pack max cell voltage is seen, all pretty close to each other. Another indication of a perfectly balanced pack set.
Reading through the other thread, seems a calibration fixes those naughty batteries.
Less @judgepd who may have something else going on.
If they can widen the bounds of the voltage tracking and increase the reserve SoC charge to 250w x pack count. This will fix all my issues.
291W with 3 packs gives 1.8A trickle charge, below the threshold the OCV logic see's as a charge.
So OCV can kick in while its actually trickle charging, maintaining of getting upto reserve.
Which is a no no, as voltage is abnormally high
TheDragon (GivEnergy) Just had a look at the cell voltages and this is from during the calibration yesterday.
judgepd there lies your answer.
Cell 3 is fubar, cell 10 is out of spec.
Is this dip in 3 and 10 seen on the next discharge? That is highlighting 2 weak cells
TheDragon (GivEnergy) I've also had a look at the delta between the max and min cell voltages.
judgepd that delta should be low mv numbers.
Can you show the 1st graph from a subsequent discharge and charge cycle since calibration.
To see if it was just lazy, or needs attention
judgepd nothing seen, just a blank screen
TheDragon (GivEnergy) helps if I actually save the image upload
judgepd just seeing a grey screen with circle rotating.
TheDragon (GivEnergy) hmm, working for me and got my wife to open them on her phone too
Strange. The apps says I have 8% battery (in home mode), but I have two green lights on the battery (3 including the status one). Doesn't that mean 50% ?
TX200 25 to 49
8% on inverter could be just inside that if a little high
judgepd
Now I can see them
Cell 3 had a wobble just after 12 on the discharge. But then got back in line.
I wonder if it has a slightly bad weld.
Keep an eye on it for sure now wr know which one is misbehaving
judgepd ooh, that looks familiar! Same as I see on my system (that is being replaced tomorrow).
Well, I guess that's sort of good news if at least an underlying fault has been spotted. I've emailed Dan with more info asking if the can have a look at those cell voltages.
Many thanks for the advice.
judgepd Yes, it charged at 100% for about 30 minutes at which point the SOC stated it was 100% which obviously had me worried straight away but it then just charged at 750W for another 11'ish hours.
I'm going to email Dan this morning with my thoughts but I think (like a lot of others) that the firmware sees whatever the battery SOC is at time of application as 0% hence why my battery only thought it had about 1kWh capacity as I was around 85% SOC when it was applied :/
fluctuating charge when Target set to less than 100% replicated here.
I recall @"TheDragon" saying he was reporting this as a bug?
looks like every 20 minutes for 5 minutes once target has been reached as opposed to holding ay target previously.
BMS 3011 D0.290-A0.290
Been having a look again this morning and the overnight charges (always from empty to full) were:
6.8kWh (last one before beta firmware), 2.5, 2.7, 3.2, 3, 3.4, 3.5, so it does seem to be improving but very slowly. I also checked the cell voltages and they were not as dramatically out as previous times but still don't look great to me:
Used the report tool to check my data for yesterday after. Total usable power for my system should be around 16.5kw allowing for a bit of loss from conversion and the 4% reserve. Yesterday discharge taking away the solar is 15.2kw. So only missing 1.3kw ish. I appear to be missing data from 2am til 2:30 so not idea if I got another soc jump. But overall I am happy.
THALL I appear to be missing data from 2am til 2:30
Me too, must have been portal issues again.
I'm on 3011 and yesterday my (AC-3.0)inverter was idle for 2hrs 40min during which time solar generation was continuously exported to the grid. The displayed SOC was 96% and the battery voltage between 52.42V and 52.49V for the entire period. I would have expected the inverter to remain idle if the battery was full but as it clearly wasn't, by either measure, I don't understand what was going on! Thoughts anyone?
Woodstock How many cycles has your system done on the beta firmware? I ask because for my first 3 cycles my inverter did this, but at 98% soc. However a couple of cycles later it is now hitting 100% and staying there. I am not sure 3 - 4 cycles is enough as per GE's instruction. I think 5 or 6 to see more normal behaviour.
THALL interesting. Mines done overnight charges of 1kWh, 0.9kWh, 1kWh so far.
Would be nice if it suddenly starts finding the extra capacity from tonight!
I have asked for a recalibration (yesterday) but not heard anything back yet.
TX200 When you got the firmware how much charge was in your system?
THALL 84% (and with it reporting 0% after... Not good perhaps!!)
TX200 Not sure, might be dodgy cells as mentioned above for judge.
I actually thought I had a good chance of seeing issues with my 2 different size batteries, 1 18 months old the other brand new.
Sounded more and more likely that the new firmware needs to come with a compulsory calibration like all the others did
chris_p Some seem to have been ok with a few cycles, think it depends on initial charge state as suggested. I'd much rather see a manual calibration cycle trigger option than a compulsory one, but suspect I'll be disappointed.
It would be really great if this new learning firmware ended the need for calibration cycles though. That would be another win if that was possible.
THALL I hope mine are okay (just starting to look into GivTCP)
I've only had 4 soc crashes that I've noticed.
TX200 Yeah I need to take a look at givtcp. Had a quick look using a windows docker which was not the best idea lol. I'll grab a pi from work or something.
I've been mucking around more today and did a forced export to empty the battery and then have been charging it but at a steady 600W to see if that will bring the cell voltages closer. They certainly don't seem to like being hit hard:
max discharge and steady charge
Also, does anyone know the difference in GivTCP between 'battery SOC' and 'SOC' as mine don't agree. I also pull the battery SOC from the GE api and that doesn't agree with the other 2 either!
Woodstock I've seen this on mine (3011). On closer inspection, I saw the SoC has dropped from 100% to 96% with no battery discharge, so suspect in my case the battery was actually full. Saying that I did restart the inverter and it immediately proceeded to add an additional 200W to take it to 100% (before falling back again)
TX200 It sounds like you are in exactly the same position I was, 8.2kWh battery that was 85% charged when the firmware update was applied and afterwards I could only access approximately 1kWh of it.
I asked for a calibration on Saturday which completed that evening and it is much better now but still early days!!
Ra Very similar to me...except a calibration didn't fix things.
judgepd 🙁 The calibration done on my battery is different to the one that was done when first installed. Not sure why?....
Ra In what way? I've had them before and they always empty and then refill back to full. The one I had on the beta just emptied and then didn't refill.
THALL The battery had solar charged to a displayed 100% SOC earlier in the day and stood idle for about 35 minutes. The voltage drop curve had just started to flatten out at around 52.49V when the sun went in and the battery discharged down to the displayed 96% SOC. The sun came out again and the battery then stayed idle at 96% for the next 2hours 40 minutes (continuous solar excess export to the grid and battery voltage between 52.42V and 52.49V) until dusk when it started discharging again. It's a mystery to me.
Today the battery has solar charged to a displayed 100% SOC again and gone into an idle state. This time I have locked it in idle by setting battery charge and discharge set to 0, disabling ECO mode and disabling all timed charges/discharges/exports (anyone know why disabling ECO mode turns on timed export?). Waiting with interest to see what happens when the battery voltage has sat for a while at wherever it stabilises.
judgepd My initial calibration on installation was just drain down at maximum discharge and then charge to '100%' at maximum charge rate.
This calibration discharged like normal and then charged at maximum rate for about 30 minutes and then dropped down to about 30% charge rate for the remainder. Charge time was about 12hrs in total (11am-11pm)
Ra Maybe I didn't leave mine for long enough to complete properly.
judgepd This is the comparison between the two.
Top picture was installation, bottom picture was this Saturday
judgepd Do you know if three separate SOC values existed in pre beta releases?
Woodstock Yes sorry, these aren't new values. 2 are from GivTCP and one is from the GE api.
Was going to webchat GE, but can't seem to find it...
Thought the option was on their contact page?
TX200 Seems to have been removed, so no chat option right now. Email only? Hmm
chris_p Again, another communications debarcle. Regardless of technical issues they need better consumer support, they are too big now to have minimal staff.
My battery is stuck with only about 5% usable capacity too.
Maybe it's actually full but the f/w thinks otherwise? It takes about 0.5kWh charge overnight and then empties pretty soon after, then sits idle. PV is still being processed OK.
Gen2 hybrid. F/w D0.999-A0.999
Riscy What is your battery firmware version? Also what voltage does your battery say it is at ?
THALL 3011, and 52.5V (climbing slowly - maybe as it's warming up through the day)
Riscy voltage will rise with temp, and fall too. One reason why the SoC self OCV was capped to twice per 24hr, to stop the SoC wandering
@TheDragon (GivEnergy) have GE said anything to you about people only getting a tiny proporation of their battery?
No, but have in conversation mentioned their own battery is down. So im sure it'll get sorted
Everyone can see loss of capacity. Those with silly losses, I would say calibration should be done as a matter or course, which will either resolve, or highlight there's a underlying issue. But im not the boss
TheDragon (GivEnergy) their own battery is down
They only have 1 test rig?! Or am I reading to much in to that wording?
No sure what is happening to my system now. I had it charging all day at 500W to try to bring the cell voltages together, it got to 97% and then without me changing anything it just started exporting at max discharge. I can only assume GE have triggered another calibration but they haven't told me they are doing it. I'm sorry to keep moaning but I'm starting to get really cheesed off.
judgepd I think you should see a notification in the portal if they have started a calibration (Not the app)
Ra I think you should see a notification in the portal if they have started a calibration (Not the app)
Yes you should. But why would they kick one off without telling you, you have not long just had one, which is what exposed the weak cell in the 1st place
Riscy has at any point (since installing 3011) during the day after you've 'run out' has the battery increased the calculated SoC to match the idle voltage and started to discharge again? (The OCV function)
My observations was when OCV ran (in this version), it didn't realign the SoC correctly with the higher voltage, which maybe why it's stuck with a wrong bottom limit.
I personally had approx 30% I couldn't access. I was to impatient to see if it cleared (or got worse) over several days and was lucky enough to be able to revert back to a previous firmware.
You can see in the attached example,.I pause the batteries at 21.30 (a deliberate trial as I thought the SoC was a little high). 40 mins later the OCV function kicks in to realign the SoC with voltage, but it goes bananas. 52.3 V resting should be approx 30%, but it pushes it down to 22% from 50% and still dropping. The other battery was not affected as the SoC was above 50%, so the function did not run on it.
The next day at empty, the same happens to the other battery.
Long winded reason for my question, I was just wondering if there is any evidence of the OCV function adjusting upwards in 3011, which should recover some of this lost capacity (my suspicion is it's broken)
mpartington Before I had my calibration, my battery went flat very quickly so I forced it into idle and my SOC % went from 4% to 23% and then stopped. Ran it flat again and tried to do this again but it didn't do anything the second time. Battery voltage was 53V
Ra Nothing showing there (there was last time they did it). At the moment, all my settings indicate that it shouldn't be discharging but I'm just going to leave it in case they have triggered something.
Ra but it didn't do anything the second time.
isnt there a 24 hour only run twice rule that may have stopped it?
just found this in the inverter notifications section. amusing? new software was was at 15:13
BMS Over Voltage 2023-01-14 15:18:06 2023-01-14 15:18:27
Battery Voltage Low 2023-01-14 15:17:50 2023-01-14 15:18:06
BMS Over Voltage 2023-01-14 15:17:45 2023-01-14 15:22:10
BMS Communication Fail 2023-01-14 15:17:34 2023-01-14 15:17:50
Battery Voltage Low 2023-01-14 15:17:24 2023-01-14 15:22:10
positor1 Looks like I have a 72hour only run once rule stopping mine! 😆
positor1 yes, think that's a possibility. I've asked in feedback for that time constraint to be removed or reduced as it means I'd probably go 48 hrs without a correction if it needs one on consecutive days, but I do understand there may be other reasons for it around the reserve. However I'm sure anything is achievable.
positor1 I had all of those when my calibration was run before. I've just had them all again but I haven't been told that a calibration is running again.
Woodstock Well the battery voltage stabilised at 52.4xV (displaying 100% SOC) and stuck there for the next 2hrs 45min with the inverter at idle. Now the sun has gone the battery is discharging at between 58W and 107W despite "Discharge power" under "Battery options" in the portal being set to 0.
No sign of any learning/SOC adjustment taking place since the day of beta installation (Friday the thirteenth!) when the displayed SOC rose by 12% (at the same time as battery voltage was falling) of the roughly 60% needed to align it with the true state of the battery.
Prior to beta install my battery had an effective capacity of 7.2kWH and thought 52.8V was 100% SOC. Three days later it has an effective capacity of 6.5kWh and thinks 52.45V is 100% SOC!
mpartington This is excellent discussion... Finally feel I am getting good technical insight. The SoC hasn't recalculated; no. Nor has it started to discharge again.
Rather, I see this:
This repeats each night when it does a timed-charge, settles at 53V, then fully discharges and settles at the 51.5V.
Riscy to be honest, those voltages aren't a million miles off what I'm seeing. The firmware was loaded remotely and the SoC was around 80% at the time. They aren't far off at the lower limit what you would expect, which is probably why it's not recalculating (Although you mentioned before it climbed to 52.5v, so odd it didn't recalculate then, unless it's the 24hrs thing). Why you have such a loss in capacity very strange. Are you a GivTCP user and able to check the individual cell voltages?
mpartington
Riscy With that range of voltage, mine is 51.7 to 53.2. Thats 4 to 100
So this is typical of the days since the beta.
The efficiency is an odd one. I don't think it's purely an SOC issue because it's the same pattern every day. If it was SOC causing early discharge cut-off, then the error (more charge in than out) would surely accumulate over time and the battery would end up totally full. I.e. after a few days of that it would take no charge at all overnight - which isn't what I'm observing.
You mention GivTCP. Have you got a plot of the cell voltages, all 16 of them for each pack
Have you had a full calibration forcing down to 45v done?
I see you are operating in the upper but of the charge curve and not using the bulk
The voltage is rising way too fast at the beginning, which is what is indicating the battery is actually nearly full. When discharging its being pulled right down through either a loose wire or bad cell, so the voltage is way lower than it should be, causing premature ending.
No idea what is happening with my system now. It randomly, and without me triggering it exported all it's charge (of only a maximum of 3kWh or so) and is now sitting idle at 0% 45.93V and has been like this for several hours. I can only assume this was triggered by GE but they didn't tell me they were doing it which I find unacceptable (apologies to them if this is not the case). I have not had a response back from them since suggesting that the cell voltages for my system did not look correct but on this latest discharge they are all over the place again but not the same cells as last time.
Joined the betatesters today.
3.6 PV on microinverters, AC 3 and 8.2 100%. only use portal / 5 min updates.
DO535-AO535 / 3007 changed to DO290 - AO290 / 3011
Immediate calibration.
My battery was at 1% @ 54V. Dumped 4.5kW down to 45V then charged 8.7 kW to 56V
Took 6 hrs 45 then back into normal operations.
Will see how it goes.
Interesting, I am still on the 3088 bms
That is exactly what i would see as a normal calibration. Dump the lot to 45v and charge full. 8.7kwh a tad under, but over the actual capacity. So all is good so far.
Welcome to the gang
3088 is the same as 3011, i didn't see any differences at all.
TheDragon (GivEnergy) Superb info. I shall set up logging on all the voltages and log tonight's behaviour.
Riscy
If you go to history, the log will already be there since you installed GivTCP, unless you've changed the retention
Got this out of GivTCP (5.2 Bat) - it's been idle since 6pm - app says 7%
` Battery_Cell_1_Voltage: 3.312,
Battery_Cell_2_Voltage: 3.314,
Battery_Cell_3_Voltage: 3.312,
Battery_Cell_4_Voltage: 3.312,
Battery_Cell_5_Voltage: 3.314,
Battery_Cell_6_Voltage: 3.312,
Battery_Cell_7_Voltage: 3.312,
Battery_Cell_8_Voltage: 3.311,
Battery_Cell_9_Voltage: 3.312,
Battery_Cell_10_Voltage: 3.312,
Battery_Cell_11_Voltage: 3.314,
Battery_Cell_12_Voltage: 3.312,
Battery_Cell_13_Voltage: 3.312,
Battery_Cell_14_Voltage: 3.312,
Battery_Cell_15_Voltage: 3.312,
Battery_Cell_16_Voltage: 3.312,
Battery_Cells: 16,
Battery_Cycles: 235,
Battery_Design_Capacity: 102.0,
Battery_Firmware_Version: 3011,
Battery_Remaining_Capacity: 6.25,
Battery_SOC: 16,
Battery_Serial_Number: BE2144G153,
Battery_Temperature: 20.2,
Battery_USB_present: true,
Battery_Voltage: 53.0 ,
`
Edit, temperature was 17c - missed that off the copy and paste after re ordering the voltages.
TheDragon (GivEnergy) I've gone somewhat hardcore and rolled my own sensor entities and data processing. I use GivTCP to publish to MQTT and then do my own thing from there... I'll shortly add a change into GivTCP with an option to disable the value smoothing, for similar reasons.
<OT> I've been burned by HA integrations having breaking-changes, resulting in entity names going awry. So I process everything through Node-RED as an abstraction into my own sensor names. I can then pipe long-term history (i.e. several years) into an Influx DB, knowing that the sensor entity IDs will remain consistent in there.
Anyway. I've put the individual cell voltages in now, and it's coming into HA, so I'll see tomorrow. I also added a sensor for "cell voltage range" (i.e. max-min) as a long-term measure of battery health perhaps.
TX200 you need a full cal down to 45v. Those cells indicate pretty fully charged.
Breakthrough! Finally found a setting that is allowing the displayed SOC% to edge closer to the true value.
ECO disabled.
Timed export disabled (you have to do this because disabling ECO mode enables timed export).
All other timed charges/discharges disabled.
Battery charge power set to 3000W.
Battery discharge power set to 0W.
The displayed SOC% is now reducing while the battery voltage is static at between 52.43V and 52.44V. It has lost 5% in the last 4 hours so only another 18 hours until it gets close to where it should be!
Now for the interesting bit ...
The only other time I have seen the displayed SOC move and the battery voltage remain static was on beta installation day, following the displayed SOC drop to 0% which most people see. I had very similar settings to this morning/evening but the displayed SOC was rising rather than falling. It stopped rising when solar generation dropped to zero. Today the displayed SOC needed to drop but did not move during the day when solar was generating. This evening after solar generation fell to zero the displayed SOC started to drop. Solar generation should not make any difference with an AC inverter but I wonder if there is some code shared with the hybrid inverter that is sticking its nose in where it is not wanted.
If the displayed SOC stops falling tomorrow morning when the sun comes up I will think I am on to something.
I was initially horrified by how slowly the displayed SOC was converging with reality but in this one instance it would be quite useful if it doesn't get there before solar generation kicks in and the theory can be tested! 😀
Woodstock The other way of also proving it is to turn off the solar in the day and see if the SOC continues to drop. Perhaps for not too long though if the sun is shining brightly!
My latest update. 6 full cycles and I have a normal graph. What I can not explain is how it charged from 5% to 92% in 5 hours on a single AC3 in minus 3c temps. I charge at 2910w, so 5 hours charge is 14550w max, but the graph suggests I charged 15400w nearly 1kw more? I assume there are factors I do not understand, I would of expected 1kw less due to conversion loss etc. But look, a normal looking data plot! No spikes, happy days.
Woodstock I'm trying something similar today (any reason why you leave the discharge at 3000W?)and the SOC is very slowly dropping, around 1% every 45 minutes or so, BUT, the voltage is also still slowly dropping and just seems to never stabilise. I'm not sure today if that's because of the cold weather, it's 5 degrees in my garage where the battery is and the battery is currently around 10 degrees and still dropping from it's overnight charge high.
THALL the charge is limited to 0.33C under 10C
So only a small battery will ever really notice.
Basically 30A per 100Ah capacity per battery.
TheDragon (GivEnergy) Thats fine, the temp data says the battery was above 11c for the whole charge cycle.
My battery currently exporting lots of electric, managed to get through on the telephone line. 😀
Glad I'm working from home today. Stuck a wash load on to use some of the export.
They said it should empty to zero and be ready to charge from solar (it looks like a nice day today!).
judgepd So it looks like this might actually be working. The voltage has stabilised around 52.4V and the SOC is continuing to drop around 1% every 50 minutes. I believe the actual SOC is around 40% so it still needs to adjust all the way from 92% which at the current rate will take around 40 hours unless it starts to adjust quicker.
judgepd I set battery discharge to 0W not 3000W. Original thinking was belt and braces to ensure inverter remained idle after dark but I have tried many and various setting combinations over the past few days. The settings outlined above are just information as to what was in place when the displayed SOC% finally started falling (albeit very slowly).
4kWh emptied from the battery (circa 0.8 of that was pre calibration)
2.3kWh exported (rest used by house including dishwasher and washing machine!)
It's now charging, got to 100% and is still charging but slower (half grid, half solar). Think it put 1.1kWh in during the calibration, so it's got a way to go to get to true 100%
Fingers crossed again!
Woodstock Sorry, I mixed up charge and discharge in your original post.
TX200 Sounds like it is doing the same calibration cycle mine did on Saturday. Charged to '100%' really quickly and then just sat at a slow charge rate 700W?? until it saw the voltage about to spike and shut off.
Ra yeah, 945 W for me. Glad it's a sunny day. Solar not covering full house usage and battery, but better than nothing! 😀
Woodstock Well it's looking as though the solar generating or not was a red herring. Sun has been up for a couple of hours now and the displayed SOC% is still falling.
After midnight the battery voltage also started dropping ( having remained within a 0.02V range for the previous four hours). It has dropped from 52.42V to 52.36V since midnight but this could be temperature related.
Between 18.15 yesterday evening and 11.15 this morning the displayed SOC has fallen from 93% to 73%. For the whole period thus far the inverter has remained idle and the system has recorded no battery charge or discharge.
Just going to leave as is now until something changes. The firmware may at some point decide that the displayed SOC% has fallen far enough to match the actual battery SOC% and the downward slope of the displayed SOC%
line will flatten off. Or perhaps the displayed SOC% will just keep dropping until it get to 0%. Watch this space! 😀
Woodstock I've never been pleased to see the SOC dropping before like I am now!
judgepd I've never been so pleased to see a battery charging from actual 0% to hopefully actual 100% 🤣
Does anyone know if the betatesting@ email is being monitored? I've pinged it a few times since Friday but no response.
I could really do with someone initiating a recalibration. I logged all the cell voltages last night and they're consistent across the battery, but the capacity is definitely being judged wrongly. But 1st-line support weren't keen...
Riscy I had a couple of responses at the start but nothing in response to my last 2 or 3 emails. I basically begged for a response in the last one (which to be fair I only sent this morning).
I have had no reply to any of my feedback. So Wednesday 11th was my last contact with GE, i tried the beta email and support emails asking for calibration. Silence. Also the chat option on their main site is still not available. Last option is to phone them.
Our system is 1 month old and I'm getting to the end of my tether with it [5kW hybrid inverter and 9.5kWh battery]
I was locked into the drop to zero situation and was told by GivEnergy that I would have to wait for a solution when new firmware was released. I joined the beta test programme to jump that queue and for a few days it seemed to have worked, but it hasn't . This morning the battery was again at zero and in bright sunshine my system was reporting only 50watts of generation. I restarted the inverter and the generation figure jumped immediately to 1.6kW but the battery remained idle at 0% - actually it showed that it was feeding the grid at times. So 1kW of the energy from my panels is going straight out to the grip instead of filling up my battery for later consumption.
I've sent three emails with no response from GE. So I've now given my supplier/installer notice that I want a solution and if he doesn't come up with one I think I'm going to demand a refund.
Riscy I've not had any replies, but I guess they're probably getting quite a lot, hard to keep up (I have to deal with IT tickets in my day job, so know the feeling).
Managed to get my calibration started via a phone call to them. They checked with the beta manager whilst I was on hold.
Guess maybe try ringing again tomorrow, maybe you'll get a different support person? Fingers crossed!
It may be significant that our system is outside and we had a hard frost last night.
ed-form what temperature is (or was) the battery according to the app/portal?
TX200
After sunrise this morning the system showed -0.48degrees C
ed-form If ANY cell is 0C or lower, all charging is stopped.
TX200 I agree. I'm hugely sympathetic. I consult on software development - exactly to help people dig their teams out of this kind of development-hell.
That's why I've been very patient, and it's also why I'd rather email instead of phoning. But you're right, at some point I'll need to look after my own interests and pick up the phone instead.
ed-form Not entirely related, but my PV has peaked at 80W today. I'm assuming this is because they're covered in frost, rather than something related to the firmware... But I have had f/w issues where the inverter would go Idle and the PV power would drop to near-zero even when the voltage was >200V.
ed-form There is your answer, the BMS is in protect the cells mode. Till above 0 on all cells, it will NOT charge.
Thats down the chemistry and cannot be changed, unless you want shorted cells
Riscy While there's still some sunshine, restart your inverter and see if your generation figure improves - mine went from 35-5-W up to 1.6kW.
TheDragon (GivEnergy)
Aha! I reread the data sheet for the battery more carefully, and saw that the temperature range for safe charging is 0 to 50 C. If I had understood this, I'd have found and indoor space for the battery. I wonder if I should build it a little house?
ed-form I had the same idea. Tried a soft-restart via the Portal about an hour ago but it didn't make any difference.
Or did you physically power it down and re-start it? Cheers.
judgepd Well, the SOC is still dropping and the voltage is staying stable. It's dropped from 99% to 85% today but I think it needs to get down below 50% before it's correct. If this only happens when the system is idle then I have to leave it idle for the next 2 days. That doesn't sound correct to me and I don't see when this could ever happen 'naturally'.
I applied for the beta over a week ago and was told I'd been put on it. But I hadn't as the firmware references were the same as before. I mailed a few times, no response so I've given up. I tried to help but no good if there's no-one at the other end. I understand they are busy but then most of us are. I will continue to monitor the various threads and keep my fingers crossed.
judgepd It looks like the adjustment time is indeed consistently once every 50 minutes. All of my SOC% drops (bar one) so far have been 1%. The one anomaly was a drop of 2% where the 50 minute wait coincided with a battery temperature drop. If my calculations are correct the displayed and true SOC% on my system should coincide around 18:00 this evening.
I cannot believe that this is the way the update is supposed to operate as how many people have their inverter in idle mode for such a long period? Taking 50 minutes to adjust 1% (coupled with how often the inverter is likely to be in idle mode for 50 minutes) I would be surprised if correction could keep pace with the displayed SOC% drift.
Woodstock Absolutely identical behaviour to mine, every 50 minutes the SOC drops 1% except for once where it dropped 2% (but could have been temperature related). I guess if this had been running the whole time then it wouldn't have got so far out of step and therefore take so long to correct, but I still don't see it kicking in unless the system is forced to be idle.
My 5.2 battery is still charging, 5.9kWh put in so far today (0.8 of that was pre calibration). Guess it can't be far off full now!
judgepd It's as if the "lookup, compare, adjust" loop is including the wait for battery voltage stabilisation each round rather than just once before starting.
Woodstock You could be on to something there, unless they are deliberately trying to stop people seeing big jumps in the SOC by adjusting it more subtly.
Possible. So if SoC map is > 30% SoC measured, then do a different loop cycle than the small steps, ie measure once a min and adjust.
Maybe submit your own idea to them for how to improve the process
TX200 charged 10.25-15.25hrs, 5.3kWh put in I believe. 😀
judgepd by adjusting it more subtly.
But if its wildly out, like on initial install, then surely thats ok
TheDragon (GivEnergy) Oh, I'd be totally fine with it, I just wonder if they want to get away from the 'SOC jump' problem.
TheDragon (GivEnergy)
I had always assumed that the feedback loop would employ bigger % corrections to start with, and then get incrementally smaller with each iteration. I'm not sure why I think that, but it felt instinctively the way to do it, to home in on the ideal calibration as quickly as possible.
Just thinking aloud.
ed-form a few people have done and will apply some extra insulation in winter, shouldn't cost a lot to do
I don't really understand why it doesn't just adjust in one go. If it knows what the SOC should be, why doesn't it just set the SOC to that value. It would still need the system to be idle to read a correct voltage but then it could just adjust in 1 go. Yes, you'd get a jump but at least it would just get it over and done with. Or is that the whole problem that they are trying to fix!
TheDragon (GivEnergy) Yes something along those lines, but I don't think that you would ever want to be doing a 50 minute loop no matter how far out the SOC % was. Something more along the lines of "always do a one minute loop and if the looked up SOC is more than 30% from the displayed SOC then adjust by 5% rather instead of 1%" perhaps. There is an issue with not wanting users to see sudden jumps in displayed SOC but this has to weighed against the time needed to get the SOC back in line. If the adjustment process has to wait 50 minutes for battery voltage stabilisation before commencing then you may not get many opportunities for it to run, followed by a limited window in which to do its stuff.
Cdent
I came back to the forum just now to discuss that possibility - I could see a small lean-to enclosure with an opening roof for the warm weather and a door for access and possibly a tiny oil-filled heater bar with a frost-stat to keep the whole rig above zero.
Or did you physically power it down and re-start it? Cheers
I simply did a soft restart with the button in the My Inverter screen.
sponev
Looking at the problems with capacity we're probably best keeping on the old firmware for now!
94jdh Your choice
Its beta after all, if you dont wish to try it, perhaps offer suggestions. Then feel free to request to revert.
A positive beta result needs many people feeding back
TheDragon (GivEnergy) yep. And people coming on at different times too. Suspect 3011 would have worked for me if I'd upgraded from a different version using the beta logic.
Now I'm recalibrated (well the battery, not me!) I hope to see good results. Fingers crossed! 🤞
I may just be being dim here, but when is an inverter ever idle (unless forced)? Mine is always either charging or supplying load to the house.
Simkna If it's empty or full are the only times I can think of.
and here's my GivTCP battery details (with thanks to @hoggy for the guide!)
App says 98% full, it's been mostly idle since 1540hrs, after a long charge, just 100-200W draw since.
`
"Battery_Cell_1_Temperature": 27.1,
"Battery_Cell_2_Temperature": 26.1,
"Battery_Cell_3_Temperature": 25.5,
"Battery_Cell_4_Temperature": 27.2,
"Battery_Cell_1_Voltage": 3.311,
"Battery_Cell_2_Voltage": 3.312,
"Battery_Cell_3_Voltage": 3.312,
"Battery_Cell_4_Voltage": 3.312,
"Battery_Cell_5_Voltage": 3.312,
"Battery_Cell_6_Voltage": 3.314,
"Battery_Cell_7_Voltage": 3.312,
"Battery_Cell_8_Voltage": 3.311,
"Battery_Cell_9_Voltage": 3.312,
"Battery_Cell_10_Voltage": 3.312,
"Battery_Cell_11_Voltage": 3.314,
"Battery_Cell_12_Voltage": 3.312,
"Battery_Cell_13_Voltage": 3.314,
"Battery_Cell_14_Voltage": 3.314,
"Battery_Cell_15_Voltage": 3.312,
"Battery_Cell_16_Voltage": 3.314,
"Battery_Capacity": 102.0,
"Battery_Cells": 16,
"Battery_Cycles": 241,
"Battery_Design_Capacity": 102.0,
"Battery_Firmware_Version": 3011,
"Battery_Remaining_Capacity": 96.31,
"Battery_SOC": 94,
"Battery_Temperature": 25.2,
"Battery_USB_present": true,
"Battery_Voltage": 52.997
`
So virtually never, then?
TheDragon (GivEnergy)
Whilst I agree in principle - I'd argue that extensive company QA testing should have ironed this out before releasing a product.
We seem to accept that products (in general) can be released incomplete or unfinished - a basic example being my kids having to download updates to their games consoles for the game to run as advertised, in my time this wasn't an option and the product lived or died on it's release so had to be on point.
Of course there are benefits and improvements from having ths facility available but it shouldn't be for relying on the customer to be the guinea pig to resolve their problems although many beta testers will hopefully reduce the time period to attain a satisfactory result.
At the end of the day it should work as advertised/sold from the offset and I'd be furious if I had the severity of problems some customers are currently experiencing.
94jdh The number of systems effected is not known publicly as far as I know - people just bandy about 1000's and "everyone's" affected. Well the latter is not the case.
It was said earlier that the 0% and 100% stops are set when the system is "overcharged" and "undercharged" - so presumably systems that tend to sit between that and never reach one end or other normally drift and end up with the result seen.
During initial testing of the firmware this may not have happened - possibly because of the way the systems were operated and it only came to light once several systems had displayed the problem.
My system is not affected and it may be because as a rule during the whole of last year my battery rarely stayed full or even got full on too many days and rarely stayed above 4% for many consecutive days either.
I've been doing some tests myself for the last few days for a couple of reasons - one being to have some comparison data to work out if an E7 tariff would be worth while (I'm currently on a pretty good fixed tariff so don't want to change yet) - so I've been charging the battery overnight and it takes 9kWh from empty and assuming no charge in the day from the sun, over 8kWh is extracted.
judgepd If it's empty or full are the only times I can think of.
Reserve Hold, Paused
Woodstock guessing it's either unintended behaviour, or a massive improvement opportunity. Certainly doesn't do that at the bottom end (although this adjustment was overzealous)
Just reinstalled 3088, so will see what happens to mine. Can you guess when I installed it 🤣
mpartington Can you guess when I installed i
Sorry, cant see it for your master battery digging a hole at 0%
94jdh This is indisputable. Fortunately there is a group of folk prepared to give up their time to try to fix this, or at least make it manageable. But we can't forget why we're here which is basically that Giv released a product that is not as advertised and not fit for purpose as it currently stands. We all want the beta to be successful of course, but the background to all of this is abundantly clear.
What also worries me is that some folk who are on the beta are struggling to get feedback e.g. with requests for calibration etc. That's just ridiculous given how they've volunteered their time. Giv need to sort this asap I'd suggest as it's getting quite messy again. Folk are justifiably getting concerned again so really some official, transparent feedback is needed. What level of fix is achievable and when?
sponev Very well said.
My only remaining concern is the 3088 bms I have is not the final version, there a newer ones in testing. But for me this 3088 appears to be working as desired. So every chance changing firmware again may kick up a new issue....(s)
THALL 3088 appears to be working as desired
Including expected capacity? kWh in vs the design kWh
Woodstock If my calculations are correct the displayed and true SOC% on my system should coincide around 18:00 this evening.
Has yours finished adjusting? Mine is still dropping and has taken 22 hours to adjust from 99% to 75%.
judgepd No not finished yet. It appears that the battery voltage drop is down to more than temperature variation. The displayed SOC% adjustment is having to chase the voltage drop downwards, lengthening the whole process. I am just hoping that it is not the adjustment process itself that is dropping the voltage!
Woodstock Hmm, my voltage is still dropping too. Like you say, at this rate they'll both just chase each other down to zero eventually.
mpartington I dream of having an adjustment curve that steep. Mine is so shallow I am beginning to wonder if I am just seeing battery leakage. Would have to be pretty extreme though at full to empty in 3.5 days.
TheDragon (GivEnergy)
If i pause the system. how long would it take for the system to check itself so I can resume?
Woodstock My voltage has now started dropping faster, despite a constant battery temperature. The SOC continues to adjust down at 1% every 50 minutes.
The thing that is annoying me the most is not the battery situation, I knew it was a beta and might not work properly, but it's the lack of responses from GE, and not just to me, to the other beta testers too.
This company is a shambles. If they are looking at this forum, then I probably haven't done anything to help myself here but I haven't said anything that isn't true.
TheDragon (GivEnergy) Cycle 6 and 7 have found most of my expected usable capacity. Taking off 3% for conversion, 5% reserve (it never drops to 4% though it is set to) I am missing 5 more percentage ish. Last nights charge slowed at 90% but I think the temp dropped under 10.5c the data says 11c.
I rose this morning to -2 degrees C and a hard frost. My battery has either failed, or turned itself completely off. No lights at all on the front. What do I do now?
ed-form Sounds like a call/web chat to GE, good luck.
ed-form Have you tried turning it back on? Button top right on the battery
My charge from last night. This is 7 full cycles. I am not a fan of the slowed charge from 90% however. I have added the temp data on this chart. But I only just managed to finish the charge within my 5 hour window due to the slowed charge. But still looking good! Though this is firmware 3088 not the newer 3011.
Have you tried turning it back on? Button top right on the battery
I'm frightened to mess with it. What I have done is emailed the contractor who actually fitted the system [not the supplier]. He's very experienced and knowledgeable. He telephoned me this morning while travelling to today's job, and told me that he'd look at my system when he gets to his destination and then speak to GE.
I've also asked him to quote for moving the system indoors, which is not a big job - the wiring passes through a large cupboard where the equipment could go, so he'd only have to shorten the AC cables - a bit more difficult with the cables from the PV array which will have to be rerouted from the loft into the cupboard, but there is an available route.
I'll post again when he's come back to me.
THALL I'm not sure about the significantly slowed charge from 90% either, it always used to slow down but this seems quite a bit now. The other thing I've noticed is that as soon as it hits 100% it starts to discharge again, whereas when I used to set a charge from 0030-0430, it wouldn't discharge until 0430 when the timed period ended.
Ra
Strange as I set my keep warm in cold charge 0000 to 0600.
Its finished at 0430 and sits idle till 0600.
That's using timed charge and eco. No discharge settings set. All off.
ed-form
I pressed the on/off button and the battery lit the lowest of the status lights and whistled at me. The status light was flashing. Held the on/off button in till the battery turned off again.
Ra
Yes the batteries do now discharge when the set level or 100% has been reached whereas they used to hold at the level until the end of the timed charge period. BMS 3011
Ra I do not have that, when I get 100% it stays there now. It used to jump up frm 98 to 100 then straight down to 98 but that appears to have resolved itself. I have told the GE I am happy with this new firmware and am more than willing to try the later release. But emails appear to be going into a black hole since I joined the beta last wednesday. My remaining concern is the newer firmware or even final release may have other issues. But so far I am happy.
THALL Mine is doing the 98%-100%-98% thing but am willing to let it slide at the moment. It has been a couple of days since I fully discharged the battery and judging by the weather forecast it will be a few days more. By the time the end of the week/weekend comes I would normally expect an SOC drop so will keep an eye on it!!
Ra No cheap rate to grid charge from?
judgepd So, my mistake, the voltage isn't dropping any faster, the scale on the graph threw me off. I do wonder whether that voltage is dropping too fast considering the battery discharge is off.
I stand by my rant though.
THALL I have Octopus Go 0030-0430 so I have just set to charge from whatever it is currently at at 0030 to 100%
Bit more sun about at the moment so it has charged back up to 100% this morning and then immediately dropped to 98% and is now sat there idle.
judgepd No change overnight on my system. Displayed SOC still dropping 1% every 50 minutes and battery voltage still dropping at a steady (if fairly slow) rate.
Since setting this test going my displayed SOC has dropped from 93% to 47% and battery voltage has dropped from 52.41V to 52.3V in the space of 40 hours. Inverter has remained idle throughout and battery data is showing no power having gone in or out of the battery. I was kind of hoping that something would change when SOC dropped below 50% but no such luck. It would be good to know what happens when the displayed SOC reaches 0% but not sure I want to wait another 40 hours!
Not sure if we are seeing an adjustment cycle every 50 minutes or just a very even battery drain. It would be good to have sight of the battery voltage versus SOC% versus temperature lookup table in use by the algorithm to help clear this up.
Wrapped my battery (in attached garage, hung on house wall) in blankets last night and managed to keep temperature above 6 degrees C. 🙂
Can I get a feel for which inverter models people have, as part of this Beta?
Am I the only one with a Gen2 Hybrid?
Today's issues:
judgepd Interesting. Your SOC% curve is identical to mine but your battery voltage is dropping quite a bit faster.
AC3 for me.
Woodstock I thought that, but I think it's just the scale on the graph. This from home assistant shows that it's still steady (although I'm not convinced GivTCP reports the voltage correctly).
Riscy Check my forum image. Wonder if this would be a way of getting round the lack of signatures.
judgepd Perfect idea!
AC3 here too
I thought I would share this. Just checked my dashboard since I first posted good news this morning if you notice my battery is charging but its capacity is shrinking! Before that load appeared it dropped from 42% charge to 33% charge while apparently charging from solar according to this graph..... The drop was 42% down to 12% in 1 hour which should be a constant 2.7kw load! I don't get it :-( It looked really good this morning.
Looked closer, from 10:15am until 11:10 I dropped 30% of my battery, thats 5.31kw, in 1 hour 5 minutes from a single AC3! Can someone tell me how? What data can I show you because I am getting ZERO response from GE about this. No contact for 1 week today.
I really would like to offer some help/advice (as I would like for myself), but if GE aren't going to communicate with the beta testers then there is not a lot we can do to help them and are just making guesses about what is/is supposed to be happening.
I don't know whether to keep leaving my system idle to correct the SOC as they won't tell me if what is happening is correct/expected.
judgepd Yes, yours is not dropping fast, but still quicker than mine which only fell 0.18V in 40 hours.
We are still running on Inverter Firmware D0.290-A0.290 & Battery Firmware 3007, as waiting to receive the beta battery firmware. But, from what people have been describing with the latest version, I am not sure I want to try it until it is fixed.
However, with the current configuration (touch wood) it all seems to be working OK, after the initial shock of everything dropping to zero and taking over six hours to recover!
Have noticed that when charging overnight, it does slow down from 90% to gradually reach 100%, just like it does with our EV batteries, so probably some better protection in there to prevent overcharging the home batteries.
Also, noticed that the inverter is re-acting quicker to changes, as both import & export during the day has reduced by about 50% from what it was before.
I caved and tried phoning on my lunch break. Just under 20 minutes without answer. Might google to see if they are still trading......
I thought I'd try something different too and stop the battery idling and let it charge from solar....it won't charge or discharge or export and seems to be stuck in idle no matter what I do.
judgepd Have you tried fully turning all the GE kit off and starting it up again?
THALL I can only do it remotely via the portal as I'm not at home but I've tried several restarts, reset to default, enable and disable all settings.
judgepd Any data available? In the inverter card remote control does it say all your changes applied succesfully?
You have most likely done all this, I am purely running down all the things I can think of to assist.
THALL Thank you. I've 'read successfully' from all the remote control values and then stored them all again multiple times. I'll try to ignore until I get home (I will not succeed!)
judgepd Crazily, nothing is changing, no charge in or out yet the voltage is going up!
My only reply to all issues is WTF. So todays data, 100% state of charge, down to 19% discharge amount 8kw. Its like someone just ripped out my 8kwh battery.
AC-3.0 100% DOD 8.2kWh D0.290-A0.290 3011
Does anyone know what is significant about a battery discharge rate of 107W?
I keep getting this (and sometimes an 81W charge rate) at spurious times that bear no relation to the inverter/battery settings. For example my system is currently set for timed export (battery discharge set at 3000W) but the battery data is showing a constant discharge of 107W. Similar thing occurred a couple of days ago when in ECO mode (battery charge power 3000W, battery discharge power 0W, no solar generation). Just a constant battery discharge of 107W!
THALL It really is time for GivEnergy to comment now. Even the guys who are willing to help are getting thoroughly hacked off. Statement asap please GivEnergy.
Woodstock I've had that 107W issue on the beta, it's another thing that I reported to GE and they didn't reply to.
Woodstock Not seen that. I am on 290 and 3088.
I am now trying to get support on the phone again. Got 20 miuntes available 😄
The only person that seems to be getting anything at all out of them at the moment is @TheDragon (GivEnergy) .
Any further light you can shed on all the shenanigans since further beta rollout, @TheDragon (GivEnergy) ?
Those struggling - have you got the old firmware and instructions to roll back? Chinese New Year is in 4 days and all…
(These releases still feel a bit Alpha rather than Beta to my eye)
I've been following the various threads for a few weeks whilst waiting for my equipment to be installed.
The main thing that seems to jump out at me regarding lack of response is around resource.
How many devs do they actually have working on this? It seems @thedragon is in contact with one who I'm presuming is chinese?
So who is actually parsing all of the feedback being given?
judgepd That's interesting. I am now going to take my system out of timed export and put it into normal ECO mode (solar charging if I'm lucky) as you did. See if I can stop the 107W discharge (and if I get the voltage rise with no charge/discharge).
As the system dashboard shows battery temperature at 3.4 degrees C, I started the battery with the on/off button. It notified me of a BMS Over Voltage and has not begun to charge, although a small surplus of generation is available. The battery is also missing from the animated display in the dashboard.
Woodstock yep, seen that and reported it. Think that one is connected to the inverter firmware, rather than BMS firmware.
If you pause the battery on the app, it shows this 107W value. The grid measurements match my IHD, so I'm not sure if there is a real 107W discharge, or if it's just a spurious value. I confirmed on the portal that the app had actually set my discharge rates to zero, and not a random number.
Maybe one for @TheDragon (GivEnergy) to check with his clamp meter 🙂
Ok I spoke to a human. The issue I just reported has apparently not been seen, might be uniqiue to me. But was noted my system has not fully run to 0% since being on the beta firmware and the thought is a full calibration will be needed for this new firmware so the voltage starts from as low as possible and builds up. I have just agreed to start a full calibration right now and will look at the data again on Friday.
Currently got 20% to discharge, then a recharge of 17.7kw via single AC3...... Just over 1 hour to discharge then around 6 hours to charge. Around £7 worth. Should finish charging just before my cheap rate starts lol. Note when beta testing ask for the calibration button to be added somewhere on my dashboard 😉
I am backing to a more positive state of mind now. To be fair I think every other battery firmware I have needed a calibration.
I set a timed charge of 90%, it ended up at 100% - looks like it updated the soc.
It stayed at 100% although did power the house (60ish W) during the rest of the timed charge.
Guess that makes sense for a 90% target, use it to get back to 90% - although it only made it to 98% ten mins after the timed charge ended.
Exported a bit of solar today as a result, but not too bad.
Guess it'll get better with a few more cycles too.
THALL depends what charge rate! Mine did most of it at 945 W
Woodstock Did it in two stages. Turned off timed export and the constant 107W discharge went to 0W. Turned on ECO mode (solar exporting >1kW to the grid) and the battery is now in constant 81W charge!!!!!!!!!!!
Looking like next port of call is a total power off and on again. Has anyone done this on the beta software and had the system successfully restart (not work properly, just come back into operation)?
@Woodstock 107w, discharge while paused - known about and reported. Inverter FW issue. Hit pause while in normal Eco and the battery doesn't actually fall silent.
Hook Dev I know - isn't a Dev, he's the boss and deffo not Chinese. However the firmware for BMS is coming out of china and the feedback being sent back to China.
Simkna I have mentioned to the Beta Co-Ord it is nice to get some feedback about a submission, even if just a thanks. He is max'd out, there are a large number of people on the beta trial. With mixed results, mentioned a full calibration is a good thing, especially if you load the FW while the battery is high up the SoC %.
Ive loaded with and without a calibration and it done while flat, its minimal issues. Not sure why they don't run a fol calibration as a mater of course on a BMS load, like they used to. There must be a reason.
I have gone quiet, as I have nothing to report that I haven't already. Anything new I discover and can be public, I will.
TX200 Nooo don't say that. I am currently discharging at the highest rate I have seen from my AC3 though, 3195w from the battery 3003w says the inverter
THALL
That matches my findings. 3000 set, 3001 Pinv, Pbatt reads 3090, actual using a DC shunt 3152
98% App reads slightly low at Max.
Average error across 0 to 3000w in 500w steps 0.21% App High. (Test Pass)
Another test with BMS3007 and 535 AC3 - Average 5% App reads high. So this has been better calibrated in this Inverter release from D0-292 (This FW was reporting Pinv as Pbatt and vice ver)sa to D0-290 (Corrected)
TheDragon (GivEnergy) Yes, had that one. This time same number (107W discharge) different scenario. I was in timed export, not paused.
According to the inverter card data I currently have a battery charging at 81W from an inverter that is outputting 63W while the displayed SOC% is dropping like a stone! As the SOC% is falling at a decent rate I am going to wait to see what happens when it gets to zero.
THALL Around £7 worth. Should finish charging just before my cheap rate starts lol
Set the lower charge rate, should mitigate some of that cost plus better for the battery
Sounds like there are no Hybrid Gen 1s in this thread, anyone know if there is at least one with a modified inverter in this beta? Or are we expected to miss out on the improved ramp up/down response time and have our own special firmware?!
chris_p Nah its ok. Just want it done as quick as possible. Plus the battery could charge faster than the 3kw my inverter can give it. They can take 4000w+ so 3000w wont upset them.
chris_p I volunteered mine, was told (over a week ago) I was on the beta but on inspection I wasn't as the release numbers haven't changed. Contacted them twice to query it but no reply so I figured it was a waste of time. Not much you can do if no-one responds. Yet more grief again on Facebook today about this, lots of unhappy people 🙁
This is my favourite image from the last 7 days testing 3kw output from a battery with 0% hehe
Hey! Hi everyone!
I have recently joined beta and have noticed extremely poor MPPT tracking with my PV, anyone else having similar issues?
If I restart the inverter it immediately holds the PV power and will pull 1kw+, then randomly will drop to 0a/5w in app.
I have Tigo optimisers on each panel, but Tigo 'run a report' and told me none of the panels are in 'rapid shutdown' and their logs (they cant provide me) say that it looks like the inverter Vpower is not met (380v~)
Today I expected to generate 6.5-7kwh and generated 3 due to the tracker being down unless I repeat restart it.
Separately, the inverter reaction is so stable and fast from battery that I am noticing that my meter is billing me some reactive power usage. The EM115 says there is 0.288a but 0w active yet my SMETS2 says 65w is being drawn which I think is reactive usage. I can reduce the value by turning off inductive loads in the house.
chris_p I have a Gen 1 but later serial number
Woodstock I still can't get mine to work properly. It will force charge, discharge or export but won't discharge to meet demand, it's like eco mode won't set.
@JaseSpace
Everyone is seeing around 10% lost capacity, inc the Giv themselves in 3088 onwards. The voltage charge curve is very gentle and wasting capacity both top and bottom ends.
I asked why were the limits set to conservative and didn't get an answer.
So hopefully went back to China to set the proper usable capacity, before we all begin mutiny on deck.
judgepd So I had to physically turn the battery off and on again to get it to meet household demand again.
judgepd Just discovered that when my system gets stuck not being able to discharge at more than 107W (ECO mode) if I perform a Settings "Reset to defaults" and a "Restart Inverter" it kicks it into discharging to meet demand. Not sure why since all the settings were already at default (toggling off the "Hide same settings" showed that "Reset to defaults" just set everything to what it already was). Might be worth a try, however, if it happens again.
joshgreenuk , could you please confirm Gen1 or Gen2 hybrid inverter, and how long the MPPT tracking (with Tigo optimisers) worked correctly before joining the beta test program.
What sort of time scale elapsed following a restart before the PV power dropped off again?
TheDragon (GivEnergy) I did get down to sub 10% missing after last nights recharge. However todays random power draw means I am currently doing a calibration. But the last cycle I think I was down to 6 or 7% missing. Things were looking good! Fingers crossed this full calibration fixes todays issue. Though saying that my system has been charging for 40 minutes and only just hit 1% charging at the full 2910w. Ooo 2% now.
sponev I offered mine at roughly the same time but no response - communication doesn't seem to be their redeeming quality.....!
I seem to remember someone else saying that a reset does more than just a reset of the visible settings. There was a suggestion that it also reset some other 'invisible' values at the server end. This may explain what you are seeing.
ToothyChris
Yes, exactly what, you would probably need to load the registers before and after and compare
I have had it almost 2 months on original, beta loaded last night. Gen 1 kw.
PV Data logs show constant ramping up and then dumping to 0a, if I reboot I get a straight 5a and then about 3-5 minutes later it dumps down to 0a and stays there for a little bit.
Here you can see my PV voltage is up and down like no tomorrow and this is when the tracker is trying to load the PV and ends up backing off, previously when tracked on older firmware this was much,MUCH more smooth and really only changed with the quality of the day. Today was sunny, clear sky, all day, managed a peak of 2kw for about an hour, but to get that I had to reboot the inverter as it was generating 50w (0a) before doing so.
Purple highlights where it couldn't hold, green was where it managed to hold to at least get a snapshot but died right after.
The lines not highlighted were mid ramp up or down and wrong, it was clear sunny day with zero shading.
The tigos have worked fine since install and pre-beta.
14:12 when there was 0 PV Tigos reported all panels in sun at almost complete open circuit, again, direct sun. This is inverter related.
So, leaving my battery idle for almost 2 days achieved nothing. It still has an effective capacity of around 3kWh. The SOC has not actually adjusted at all.
judgepd have you asked for a calibration?
TX200 Yes, it's run twice.
judgepd did your calibration do both the empty and fill parts?
Sure I read somewhere that someone only had the empty to true zero part run.
Mine got to "100%" very quickly on the recharge but continued for many hours until it found the real 100%.
TX200 yeah, that was me probably, the calibration just discharges and never seems to charge. I have reported it multiple times but not heard back from GE.
judgepd think you'll need to pick up the phone. Took me 15 mins to get through at 0850 the other day. Wasn't sure what time they open (it's not on their contact page!).
Live chat might be an option from that page too, but last few times I looked, it wasn't showing.
My calibration took 11 hours! Currently discharging, discharged 52% and the dashboard says 8kwh discharged, but the dashboard is often behind. Waiting to see if I have the same issue as yesterday. The 2% jump at the end of the charge and start of discharge has returned however. But I need to wait for another cycle to see if that corrects itself like last time.
TX200 Surely email was invented so I didn't have to interact with other people! 😆 But yes, I probably will.
Just wanted to say a big thanks to everyone in the beta (and previous alpha), you are all doing a great job observing and testing.
chris_p I have a 3.6hybrid Gen 1 on FW 3088, mine is a SA2209 if that makes any difference in terms of how old it is? Was installed end of June last year.
judgepd Same here! The displayed SOC% is now under reading by a substantial amount. The Inverter went idle at a displayed 5% which stopped discharge. I am now witnessing a very slow drop of displayed SOC% (1% every 3.5 hours) and it is down to 2%. The battery voltage is also dropping very slowly but is still showing 51.78V! Of course when the sun comes up the battery will start charging again, take the displayed SOC% back above 5% and take the inverter out of idle. I foresee a day of stumbling up and down the displayed SOC% range with little prospect of any alignment taking place. ☹️
chris_p But still no status report from Giv. Pretty disturbing given the promise to fix it before Christmas, the obvious (significant) problems still remaining and pending Chinese holidays. Is everyone still confident or do we need to start warning our suppliers about taking the kit back? Surely it's time for a formal statement?
sponev Surely it's time for a formal statement?
Completely agree. I was only thanking the beta testers at the moment you'll notice. GE devs are doing their bit and so are the technical team from the snippets we hear from various non-GE people. However, the customer relations department are AWOL as usual.
Mine has some new exciting behaviour today where the battery is charging from solar, the voltage is going up but the SOC is going down!
Calibrated 2 x 9.5 overnight, woke up to 100%, restarted inverter to get PV to kick in, battery 1 red status battery 2 fully charged, both 54.9~ on giv tcp.
Rung Giv, they said they will look, they did a force charge, and between then (30-45 ago) and now, I have drained 20% drawing just under 600w.
PV issue remains, they put debug mode on, but looks like its off again now, but shows that the inverter is loading and unloading the array.
judgepd That'll be the negative reversed hysterisis loop 🙁
judgepd That sounds familiar to what I had yesterday. Not seen before according to GE. They started a full calibration to see if that will solve it. But you have already had 2 calibrations......
THALL I'm pretty sure neither of my calibrations have worked though as they didn't recharge. I think my battery is just screwed as the cell voltages vary so much whenever the battery is being charged or discharged.
Ive just come out of a calibration of 2 x 9.5 and within about an hour they are at 60% with light discharge, PV doesnt work either, this update is much worse condition than I thought, if China are doing all the updates when are they back from holiday? I may downgrade until they are back
judgepd What kind of voltage swings do you see when it's charging?
My battery is stuck at about 80% full (I guess - the system reckons it's only about 15% full) and the cell voltages swing up from 3.25V to 3.45V when charging at around 1.5kW.
joshgreenuk a wild guess is this is a new "feature" of faster reaction times - the tracker doesn't appear to cope well with on small strings (a larger array has a lot more shock absorber / cumulative voltage to play with I guess).
This is just string 2 which has only 4 panels yes? An experiment I guess would be to shift string 1 and 2 around and see if the problem follows it (doing so with suitable precautions)
hoggy They are JA Solar 545w so not small panels. I would be expecting to see 2.5kw right now on string 1, I restart it works for about 2 or 3 minutes and then falls off. Yesterday it held it for about 3 hours then fell off.
String 2 shouldn't be producing anything its not in the sun, String 1 is 7 due south panels.
judgepd Ok tomorrow phone them, option 2, ask for Dan Maddocks. I spoke with him yesterday and he started a full calibration for me. What was funny my system said it was exporting the full 3kw for nearly 2 hours when the battery said 0% and it was exporting I can see the export on my smart meter. Then it started charging. Entire process took 11 hours for my system. Current battery curve is looking ok today. But he's not in the office today, hes in tomorrow. Apparently he is going to take a look at my battery data.
Riscy These are the cell voltages covering the last charge and a bit of discharge this morning.
hoggy Can you guess where I restarted the inverter? 
judgepd I am seeing this one on of my batteries, after the 3rd reset its now holding PV at 1.5kw which means it will now track up to the expected 2.2 for this time of day. Its taking in 800w of charge but dropping SoC, voltage relatively stable as should in theory be 95% SoC
Good solar output today so I have set a timed charge going (supplementing solar with a grid power). Expecting a bounce off the high voltage alarm but at least I should get a couple of hours of idle inverter at 100% (if I turn off timed charge and the sun stays out).
Who knows, I may even see some alignment take place (although it will be less than 24hrs since a suspected attempt to align late yesterday afternoon was curtailed when the sun went in and house demand took the inverter out of a constant 107W charging state which could really have been idle). 🤣
joshgreenuk I mean... I can take a wild stab? Guess it's another bug on the list.
Could these times where the voltage is increasing but the SOC is decreasing be this beta firmware trying to align the SOC/voltage? So far today my system has added 0.5kWh from solar and taken the SOC from 22% to 27% which gives it a capacity of 10kWh (which is too high but it's closer than it's been for a long time).
hoggy I seem to the the only one reporting it from what I can tell/see.
Just cant hold the PV anymore, I guess there is a possibility it is hardware related and not firmware and that this FW is just highlighting it? Who knows, but China is on holiday for a week in 2 days and thus I suspect there wont be any updates to FW or BMS for another 2 weeks at a guess... have asked if I can downgrade as this equipment is worthless in this state.
Feels like a solution is further away now than in November, when it was promised on 16th Dec tbh
I am seeing no reason why I would buy GE kit until this is fully resolved now. I'm OK with buying something that needs a Firmware update, if and when that update is stable, plug and play, which honestly looks like it is months, not weeks, off currently. Buying just to join an alpha test or sit with faulty equipment seems ridiculous.
I am also wondering if part of the problem here is that GE have given themselves a headache in this testing by not adequately ensuring that all the batteries involved were actually functioning correctly to start with (outside of the SoV issue), leading to a lot of wasted time and head scratching by all as they have to troubleshoot faulty batteries running the updated firmware with unexpected results. If a general rollout of a firmware happens in a similar manner it will be a logistical catastrophe with the extremely limited support that seems to be be available. That's why I now see it as being months, not weeks away sadly.
Grateful for all that you testers are doing, you are going above and beyond, and really ought to be being reimbursed for your time.
joshgreenuk I suspect there wont be any updates to FW or BMS for another 2 weeks at a guess
At least that long
phoenix I am also wondering if part of the problem here is that GE have given themselves a headache in this testing by not adequately ensuring that all the batteries involved were actually functioning correctly to start with
Very good point. I said not that long ago that this firmware test will probably highlight more people with faulty hardware rather than just SoC bug issues.
judgepd Similar to mine, around 0.2V. Anyone know if that's normal? Feels OK I guess.
judgepd Spoke too soon, the current battery in per percent raised indicates a capacity of 5kWh and it's dropping every time I redo the calculation.
judgepd Where are you getting that data? From givtecp or the dashboard?
THALL Getting 'battery in' from the GE API and I also record it when my overnight charge ends so I can tell how much is added by solar.
judgepd Hmm, yeah sadly maybe coming off the beta is the best answer for now.
All this negativity. Thought I'd better post 3 x 5.2s perfectly in Sync
Capacity overall down 10% but the tracking is so close, its almost indistinguishable
Zoomed in today PV charging
TheDragon (GivEnergy) Thats great etc. But it does feel like this firmware is a long way off from being ready for general release. From comments from those testing. After my calibration yesterday it looks like I am missing 6.6% capacity today. But yesterdays double rate drop concerns me so I want another 5 or 6 cycles before I say yey or nay to this.
But i dunno, 4 outa 10 good?
I have to say I am also not experiencing anything too worrying either, though I do not have access to givetcp or the level of understanding of some like dragon.
I am noticing a reluctance to charge above 98% or if it does it then drops to 98 and doesn't take any charge. on the other hand my 2 x 8.s's have not fully cycled since the update and the lowest i got them to was 20%.
Also today I noticed the system idle at 95% soc despite plenty of solar available.
no soc drops or increases though and i have to trust the soc is correct?
52.8v it says was 98%
positor1 52.8v it says was 98%
Not far off. My 5.2s are 53.2v when they say I'm done and stop
TheDragon (GivEnergy) I'm glad there is some positivity but you can't deny that most of the new people joining the beta aren't having the same experience as you? Got to agree with @THALL that it is still some way from general release. Once we get to mid Feb most people's batteries will unlikely be getting close to 0%? I know mine didn't very often last year and won't this year with overnight charging.
chris_p If they release it and people have the same issue I am having with PV then I suspect they will get a lot more complaints than the SoC
Sun all day, only held for what seems to be about an hour or two, 4kw generated but best day since install back in September so would have expected double minimum really.
chris_p I joined the beta on Monday, having never had a SoC issue (so presumably my batteries weren't made on a Friday 🤔).
I've seen a capacity loss and also some strangeness this morning of idling and exporting with the battery at 95% and PV slightly above house load (could be related to using the app for this mornings savings event).
Without knowing the underlying detail, it's impossible to tell how far off a good fix this is. All I can say is that it's not the version I'm running today due to the capacity loss.
stevelewis You mean you downgraded?
ToothyChris I had this happen to me. I had to get GE involved to correct the problem.
That’s actually a good call, I haven’t tried to reset to default, I doubt it will fix my PV though.
joshgreenuk Maybe, or perhaps it was enlightened self-interest as I knew it would be coming out sooner or later.
stevelewis I don’t understand your cryptic messages, did you upgrade not like it and simply downgrade? Yes or no would be a good response lol
chris_p Agree with you on the batteries not getting close to 0%. Only three months of the year when mine gets anywhere near 0% and then only once or maybe twice a week. A rethink is required if successful implementation of the fix relies on completing multiple charge/discharge cycles.
Mine is certainly working better post the calibration. But probably need a few more days to a week to see if I'm happy with it.
Like when I said charge to 90% but it ended up on 100% (due to a soc recalculation) - hopefully that will be less likely after a few more cycles!
TheDragon (GivEnergy) Mine settles at 53V once it's finished charging.
Unfortunately, it stops discharging at 52V... Meaning my current SoC is about 500Wh.
Been like this for 2 weeks now.
Riscy Something is very wrong here.
The voltages suggest its going from 20 to 70% and back again in a very short time.
All the cells seem to be similar. Indicating the high impedance is common to them all
Are you able to check the connections are all tight, Inverter, MCB and battery?
TheDragon (GivEnergy) Indicating the high impedance
Ah... So you're saying that there might be an unusually high impedance, which is then causing the voltage to go:
Is that it?
If so - this would explain a lot:
I have terrible losses. Like, 4kWh energy in, but then 2.5kWh out. Maybe this power loss is from an unusually-high impedance, resulting in burning energy into heat somewhere?
(I gave the cables a gentle pull and they seemed solidly attached... If that counts as "checking connections"...)
Urm no a gentle tug wouldn't confirm if the bolts are all tight, inc the 2 cables internal to the BMS on the right side of the input terminals of the Battery.
If there is a high resistance point, on charge the voltage will abnormally rise higher than normal, and on discharge lower.
Giv Energy should be able to see this, if its the BMS that is spiking or the inverter voltage, giving a clue where the poor connection is.
Was it like this with 3007 BMS firmware? Did you previously see high spikes on charge?
Just seen you have a single 5.2kwh.
4kwh in isn't far off the 80% DoD. The charge times looks short as its is such a small battery.
I also assume this is charging at no more than 2.6kwh, so should take around 2 hours
joshgreenuk ah, I see. I thought you meant that I "downgraded" by going to the beta 😁. No, I still have the beta installed.
TheDragon (GivEnergy) I'm not bothered about the 4kWh capacity (i.e. 80%).
Rather, it's that I only get back out about 70% of that 4kWh. That's a big old loss. Unusually high impedance is what will account for that perhaps. This has been the case since day 1, regardless of f/w.
Yes, under the original f/w it would rise to 55V when charging and then settle to 53V once charged.
Edit: and since the new f/w, the 2.6kW charge rate "fills" the battery in about 15 minutes and then shuts off. That's the overnight spike in the graph I posted.
Riscy Edit: and since the new f/w, the 2.6kW charge rate "fills" the battery in about 15 minutes and then shuts off. That's the overnight spike in the graph I posted
Yeah thats not good. AS the new FW uses the voltage for a reference as well as the power in and out, if there is a loose cable, it might account for the spike being so soon and sharp.
Without removing the covers and doing tests (Not recommended for a user) no one can be certain why its thinking its full so soon, and going flat so soon.
This has all the hallmarks of a high impedance joint somewhere.
TheDragon (GivEnergy) Thanks my friend. Much appreciated.
(sighs) shall try sending a message to Support.
AC-3.0 100% DOD 8.2kWh D0.290-A0.290 3011
Day six and things are looking a little more normal.
Observations and thoughts.
Charging did not initiate an over voltage alarm, so it looks as though battery voltage (rather than the displayed SOC reaching 100%) is now used to halt charging.
Cliff edge throttling of charging power to 250W is evident as the battery approaches full.
Three spikes and you're full? Battery voltage shows three spikes before halting the charge; the same as it did last Friday when beta was installed.
Displayed SOC% shows a sudden jump from 50% to 100%. Not really what you want to see six days after installation!
Voltage stabilisation after halting battery charge took the best part of 3.5 hours. Is it practical to wait this long in inverter idle for an accurate figure to reset the SOC?
What percentage does the true SOC get reset to when the battery is FULL? 120%? The displayed SOC never goes above 100% and only starts to fall after the FULL battery has been discharging for a while. This suggests that there is a second SOC in use which covers the whole range (including the unusable bits at both ends). This is why I keep using the qualification "displayed" SOC% in my posts. It must be a nightmare having to code for all the permutations of battery capacity and DOD!
No evidence of SOC to voltage alignment taking place. Perhaps not needed in this instance if FULL is automatically considered to be full.
Slightly worrying that my stabilised FULL battery voltage is only 52.41V. Under 3008 this was 52.8V which I already thought was too low.
Much to think about. I have thus far seen no firm evidence of any Battery voltage to SOC% alignment taking place on my system. Even if this were not the case I do not think the method used to implement the fix for beta testers stands any hope of success in the real world. The initial learning process is currently too reliant on an expected battery usage pattern and does not account for the many and varied usage patterns that undoubtedly exist. I worry that the same may be true of the day to day alignment process but I am not in a position to comment as I have not knowingly seen any alignment yet!
positor1 I experienced idle too yesterday and I suspect it was due to me tinkering with the settings on the app for the morning savings event. A reset seemed to fix the problem. Could it be the same for you?
stevelewis
I think a second reset worked or was that coincidence. I can not say. all seems good since.
Thank you
Ok so charge 8 for me. 2 cycles since a full calibration in an attempt to fix my soc dropping at twice the speed my inverter can do..... and all looks good. The odd 2% issue at the top is back, it doesn't hold 100% for very long. But this cured itself after 5 cycles pre calibration so hoping it will again. Voltages are good. However capacity appears to be around 15% short again. I am hoping another 3 or 4 cycles will get me back to 6 - 7% missing. GE said they would be looking at my data today. But I'm not sure about this firmware, need another 7 - 10 cycles usage before I can say I am confident with it.
But so far if anyone else gets invited to the beta ask for a FULL calibration when it is applied. And please wait for 6 or 7 cycles for things to settle down. Not 6 or 7 days, but 6 or 7 battery cycles, 100 down to 4% and charged back to 100%.
TX200 Like when I said charge to 90% but it ended up on 100% (due to a soc recalculation)
I'm on 3007 and have this issue. I've logged it with GE who are investigating why the Limit setting keeps turning itself off. Go and read the value in the portal for 'Enable AC Charge Upper % Limit' and see what it returns. Mine keeps disabling itself and I get 100% AC charge every night.
THALL This looks better, thanks for the detailed feedback. However, how is it going to work for those of us who don't use a full pack. For example I built in around 40% excess with an eye on future plans and typically charge/discharge 25-85 ish. Do we manage it such that we force the desired cycles manually? I'd be OK with this but lots of folk won't I suspect.
sponev In theory it will just take longer. If you are using 60% daily then it should take 2 days to "full cycle". 2 full weeks?
THALL It can be done with manual intervention but to be user-friendly, should it need that? I suspect your average user won't really be disciplined enough to manage it. Given some of the questions I see on Facebook, lots of folk who have this kit don't really understand the basics. There are also some major quirks it seems during the learning process which could cause havoc?
I think what we have here is more an alpha proof of concept, but quite a bit of refinement needed before full roll-out? The more recent results have certainly seemed better though.
sponev Short of saying as part of this release people need to allow one month (though this is affected by weather, solar) for things to settle down I am not sure how much better this is going to get. Forcing effectively 6 or 7 calibrations on someone would be rather expensive. I am still hoping after this firmware calibrations will not be needed for future firmwares.
THALL as far as calibration is concerned, a full cycle needs to span the full SoC range, two cycles of 30% to 80% isn't the same.
stevelewis Is it not? I thought a battery cycle was output total, so 30% daily needs over 3 days. If it needs the full soc range then on current data I don't think this firmware is a fix, especially if deployed in the summer. 1 cycle was not enough for things to settle down on my system.
I don't really get the initial "learning" part being so random. If it's just going off a voltage table surely for starters it should just know pretty much within a few % after 1 calibration - both upper & lower boundaries are then known, everything in the middle is tabulated from there?
Fair enough from there onwards let it do it's thing but it currently sounds like it is learning completely from scratch on each system?
(or is this what the Beta is actually trying to achieve - widespread data on battery cell voltages vs SOC from all different battery ages, sizes etc to then utilise as a baseline trend table / starting point for everyone's system)
chris_p 3rd cycle since recalibration, charge target of 90% set.
It hit 89% and stayed there for some time (no wobbles). Just a little woble in the last half hour, drop to 87%, charge to 89%, drop to 87%, recalibrate to 89%
So it ended up where I wanted it to be.
My battery has not come back online in this cold weather, although it is distinctly above freezing now and was only just below overnight.
How do I restart the thing? I've pressed the button several times but all that gets me is the lowest charge range light flashing and, over the space of a couple of minutes, the status indicator light coming on green, then immediately yellow, and finally red. The battery does not appear in the dashboard - the picture of the house has an inverter, but no battery.
Do I have to get the GE people to kickstart the thing? Or can I launch the process from within the dashboard? Can someone tell me the correct way to do this.
ed-form You hold the button on the side of the battery for a few seconds to turn it off and press it once to turn it on.
judgepd You hold the button on the side of the battery for a few seconds to turn it off and press it once to turn it on.
Yes, that's what I have done, but the battery does not appear within the system and the red status light seems to indicate it is an off or error state of some kind.
hoggy this is exactly what I've been feeding back. Perform a look up from these tables on the fly, then fine tune at the top / bottom if necessary to improve the individual battery cal curve. Can't rely on batteries being idle for hours at either empty or full. If that's how it is meant to work, then there is some odd behaviour.
I am getting a HUGE jump on the start and end of a charge/discharge, looks to take in about 20kw and give me about 14.5kw
Coupled with the fact PV is broken for 50% of the day I am giving this a couple more cycles to feedback then I will request to be downgraded as right now my Hybrid is essentially broken and may as well be off.
THALL The number of full battery cycles you are talking about could take me approaching 2 years of normal operation! It is only in the depths of winter that I get anywhere near to 4% and a full drain occurs maybe 4 to 5 times a year. I don't believe that living with all the weird and wonderful implementation side effects for that length of time is a viable option. Neither do I think it reasonable to expect customers to force charge or discharge their batteries multiple times to speed the process up. It is OK for us experimenting in beta testing but a go live involving anything more than "press the firmware download button and forget" is a non starter.
All over the show SoC:
joshgreenuk Very weird. Looks like the soc issue but it recovers? Then I can see the rapid discharge like I had on Wednesday in the middle.
I am happy with mine (ish) but I don't think this beta is the fix. Too many other issues with it.
Beta has destroyed my inverter lol, PV will ramp up to expected rate then drop to 0a then repeat, so probably stressing internals out like no tomorrow dumping and applying load on the PV.
Cant be good going 0a to 5.5a 8 times within a 5 minute update.
It never displays the right SoC at all at any point since update, but cells are very balanced.
If you call and ask to speak to someone in the beta team they should be able to downgrade you.
Woodstock Just a quick additional observation and a correction on my Day 6 report.
Additional Observation
The total battery charge figure recorded for the day was 0.9kWh. Patently ridiculous if the plotted battery charge curve is to be believed.
Correction
"Under 3008 this was 52.8V" should have read "Under 3007 this was 52.8V". I have never had 3008 installed.
Yeah but they said I am 'not giving it enough time', but the PV issue is a much, much greater issue, and if its not fixed what happens if it gets released and suddenly theirs loads of people like me?
I'd rather suffer if its being fixed, but if its not then yeah, I'll go for a downgrade. Will ring Monday if no response to my feedback emails.
joshgreenuk How long have you been on the beta?
THALL 3 days
joshgreenuk Ah id say another 4 - 5 days based on my experience. Have you had a calibration done?
THALL Yes then was told it wasnt needed in this firmware and to let it just cycle 3-4 times, which it has done 2-3 times now
PV Issue much, much, much worse, sun all day today. Tigos CCA portal shows clear difference to the loading of the panels between the 16th and the 18th, 16th the panels have a relatively smooth output voltage, 18th they are constantly either loaded or at open circuit. Maybe 30% of my daily generation is being generated, more times than not I look its ramping back up to peak and then falls off, so most of my generation is during the ramping phase of the panels.
joshgreenuk your Hybrid isn’t an old one is it?
Those before SA/SD2209 need another capacitor added - they shouldn’t have put you on Beta if so!
Edit never mind, should have scrolled up a bit
TheDragon (GivEnergy)
5 days after beta install and now it is declaring 100% soc. if all is good as it seems then cycles do not have to be full cycles as the last four of mine have only been from 98 down to 70.
interesting battery values declared on the inverter card, clearly 100 encompasses a range of voltages.
13:54:49 53.03 0 10 100
13:49:47 53.03 0 10 100
13:44:45 53.04 0 10 100
13:39:43 53.05 0 10 100
13:34:42 53.06 0 10 100
13:29:40 53.07 0 10 100
13:24:38 53.06 0 10 100
13:19:36 53.09 0 10 100
13:14:34 53.1 0 10 100
13:09:32 53.12 0 10 100
13:04:31 53.23 -145 10 100
12:59:29 53.55 -698 10 98
12:54:27 53.55 -698 10 98
12:49:25 53.55 -697 10 98
12:44:23 53.62 -698 10 98
For the record its now after 2pm and I have not heard back from GE after they said they would take a look at my data today and get back to me on the issues. I have emailed them a couple of hours ago as a nudge...... But still 4 hours of work time left I guess.
I telephoned support a couple of hours ago and one of the guys there kickstarted my battery and set it charging from the grid. Just now, I restarted the inverter which was displaying nonsense numbers, and the numbers all settled at believable values. I also set another grid charge going - I'm giving it about 5 hours because it's only had 0.3kWh and it still registered 0% capacity. I'd like to push it up to 15% or more, so I've set the minimum charge limit at 15%. We'll see what happens.
THALL I've been going backwards and forwards between accepting that they are very busy and shouldn't be too bothered about not getting a reply and being absolutely furious about it. I have to assume that at the very least the feedback emails are being read, in which case, then someone read several emails from me saying that my 8kWh battery has been reduced to at best 3kWh and then felt that it didn't even deserve a response or at the very least didn't forward it to the main support area.
judgepd I got an email back today, I think they are swamped, and they are being read. I don't think there is enough staff for the testers perhaps, I have not been furious with them at all because its beta and I asked for it, but I can understand why you would be if you only went to beta to fix a non working product.
joshgreenuk Out of interest, when did you send the email you got a reply to?
joshgreenuk My battery worked pretty well before the beta to be honest and I hadn't really ever seen the SOC jump, I was getting slightly less capacity than I would have expected but it was functioning pretty well. It seemed like joining the beta would benefit all users and GE by providing them some real usage data.
All emails possible, same, I just wanted to test the faster reactivity rate which highlighted other issues meter-wise.
I would be happy to stay with scuffed PV output but China is on holiday until at least the next 7 days, so I think I will downgrade and then return back once they either have a fix OR need the problem online again so they can log.
I'm not seeing much evidence of SoC recalculation myself. Loads of solar today, so inverter has been idle most of the afternoon, yet stopped charging at 95% and never clicked up to 100.
kai
could it be that the calibrations are so small and also only allowed once in x hours that some systems will take longer to come right?
mine was also stopping early and when at 100 with spare solar; a draw on the battery does not appear to get replenished straight away. i suspect this is because when idle it started to measure the parameters for soc and "started so would finish"
just guessing BTW
positor1 I left my system idle for more than 2 days and didn't see any sign of SOC adjustment. The SOC just steadily dropped as the voltage dropped.
Woodstock How much solar do you have? Or do you have little load?
My battery is rarely full from the sun and if as I have the last week to get some comparison data it won't last until I recharge over night again. (I have a 8.2kWh 100%DOD)
I do only have 4.6kWp spilt equal south & west so this time of year only just over 2.3kWp. Last few days generating less than 6kWh
positor1 On the other days we've had good solar then I've seen the voltage when charging get to full (at similar 95% SoC), then fall back down and eventually start charging again. Charging from grid gets to 100 SoC, but isn't showing any of the slowdown in charging as it approaches 100% that others seem to have.
Another sunny day. Almost have surplus sun. Another month I rekon and no more charging from grid till October.
Little drift on middle pack, which will self adjust later to match the rest, as it always does.
https://i.ibb.co/jR2rpR1/Screenshot-2023-01-21-10-25-32-510-io-homeassistant-companion-android.jpg"
3088- Mine not quite as tight. Hitting top limit regularly, but not bottom. Neither of the batteries wanting to record 100%. Good simulation of summer usage
(Still can't work out how people.post pictures, rather than links)
Edit... There we go...
anglefire 3.52kWp due South, installed 2009. Lower than average load. Heating and cooking by gas. Battery was sized to cover two days use without any solar generation. In the summer the battery rarely drops below 50% but in the winter it does occasionally go flat.
Empty to full took 3kWh last night and it's discharged 2.3kWh today and is now empty. Doesn't seem to be learning anything.
WARNING! WARNING! WARNING!
If you have a G100 export limit set on your system, beware.
Mine should be limited to 3.52kW export but I am currently exporting 5.42kW! Looks like the beta firmware has lost this setting or is not checking for it.
Woodstock Ah ok - our usage (according to Octopus, despite a year of Solar is still 7000kWh a year) So we are often flat and I don't yet have a E7 tariff - though hopefully getting an EV in about 12months (Assuming the CEO signs it off soon and it actually gets ordered!) so will be able to then go on either IO or Go - which based on my experiment for the last 7 days will be a significant saving for normal usage but will go up obviously charging the car.
Oh and that 7000kWh hasn't changed from before solar - we actually pulled about 3600kWh off the grid last year
Woodstock interesting as that I would have thought is related to the MPPT tracker, which I am having issues with it even holding PV power....
Woodstock
Good spot.
That could cause a few complications in the summer.
Have you fed it back?
Just checked mine, it was 5000. It had reverted to 6000.
So set it back to 5000, not that I ever export much anyway, but in summer it could bite a few in the ass
TheDragon (GivEnergy) Yes I have sent a reply to my "beta firmware installed" email. This is not the only G100 export limit problem I have seen. Back on 25/3/22 I reported that the G100 export limit is ignored during a calibration run. Chased this on 23/9/22 as it was still being ignored. My bug report remains open with nothing other than my entries in it!
AC3 8.2 DO290-AO290 3011 Full Calibration 5 days ago
System now working from 52.9 V to 57 V previous FW started at 54 V so more capacity available. I have seen SOC jumps and a SOC adjustment downwards by 20%. The software seems to be learning but it needs full cycles, because of battery size / PV and house usage this takes 2 days to achieve so still at start of process.
The inverter control is improved and "leakage" has been minimal.
Battery capacity now over 5.5 Kwh and should increase with more cycles.
Optimistic
Newman Day 8 for me and I saw my first sign of an SOC% alignment. Had to do an aggressive timed export (full to empty at 3000W) to get the displayed SOC down to 4% (it usually sticks between 5% and 6%) but then the inverter went idle. After a delay the displayed SOC started to rise and got to 7% before timed export kicked in again and stopped the process. Doh! When the inverter went idle I set the battery charge and discharge power to 0W but it didn't stop the battery dumping charge at 3180W. Feedback report raised.
The G100 break is frustrating, only GE can set the export limit, providing our installers tell them.
While this is all good news, more testers seeing SOC alignment I am still not convinced this firmware is good. As better weather approach it could take weeks to fully cycle batteries enough times. For some customers that could take weeks! But there are other big issues this firmware appear to cause. Unable to use solar power, only being able to access less than 20% of their battery capacity.
I would like to see a statement on GE's expectation and next steps.
Certainly from my side, the new firmware is working seamlessly - much like @TheDragon (GivEnergy) described. Totally solved the SoC jumping, but slightly more limited battery capacity (although I have to say some of this is inevitably contaminated data with normal winter conditions, so hard to really judge how much of this is the firmware).
Not sure what you mean by 'unable to use solar power', certainly mine's been storming away in recent sunny days (18kwh in winter ain't bad!).
It'd be interesting to see what % of people are affected by the symptoms you describe - it's obviously not all customers, so there must be a differentiating factor that could be isolated (eg certain setups, versions of inverters, sizes of batteries etc etc).
Either way, they seem to be making progress, although the chinese new year is bound to slow things as many have pointed out - I also noted from one of the posts they're building a UK factory for developing solar tech, so hopefully this will bring a lot more under their control. I have a suspicion at the moment they're a little at the whim of their chinese developers/manufacturers - if they could bring more of the development work to the UK and ultimately just use China as a manufacturing/labour hub, then I think both pace and quality would improve (plus our inverters would speed up as we wouldn't need to be sending all our data to the Chinese government first..... 😆 )
THALL only GE can set the export limit, providing our installers tell them
Not true.
It's an engineer level setting. I set my own back.
TheDragon (GivEnergy) I will keep that in mind for when the EMS arrives and I have a second AC3 fitted. That way each of my 3 export devices will be capped at the 3.68kw limit.
Another day another mystery/problem. (Well two actually, but I am going to put the second in a separate post).
At 07:50 this morning my battery started charging (at 290W) for no apparent reason. The battery voltage was at 51.17V, the displayed SOC was 2% (and had been for nearly 4 hours) and the battery temperature was 12 degrees C. There were no user settings in place to initiate a charge at this time. The charging continued at 290W until an expected timed charge kicked in at 08:30.
Anyone else seen anything like this or have any idea what was going on?
It is almost as if this was some sort of "nearly empty" battery protection but given the voltage and length of time the battery had been at 2% displayed SOC this seems unlikely.
Feedback raised.
... and now for the second mystery/problem.
The expected timed charge started at the set time of 08:30 but charged at 2.8kW instead of the 1.5kW set in battery charge power.
So I set the battery charge power to 1.6kW and the battery charged at 1.6kW.
I restarted the inverter and the battery charged at 2.8kW.
I set the battery charge power to 1.5kW and the battery charged at 1.5kW.
I restarted the inverter and the battery charged at 2.8kW.
I set the battery charge power to 1.6kW and the battery charged at 1.6kW.
I restarted the inverter and the battery charged at 2.8kW.
You get the picture?
I am in the habit of always restarting the inverter after making a change in inverter settings as it is not unknown for the change to be displayed but not actually to have taken place. This new sequence of events appears to show that making the change is working but restarting the inverter is losing it!
Feedback raised.
Woodstock I'm not really surprised that a reboot resets the charge power if I'm honest. It does depend on systems, but that would require persistent values. In my working world, you can have persistent values that can be set to reset to default on a reboot - though that is a tick box in the one system I use as the default is to retain persistent values.
So either the inverter doesn't have persistent space - but given you have always done this suggests otherwise, or the persistent space is being reset on a restart.
A question on LiFePo4 cold weather capacity. Obviously this reduces in cold temperature, but does this happen because the charge energy that can be added is reduced as well as the discharge?
For example, last night during charge I added 8kWh from empty to full. If if had been summer and the battery was warmer would this have been higher -say 9kWh?
Or do we just see the reduced capacity in terms of the retentive capacity? So I add / start with 8kWh regardless of temperature, and in summer I may get 7kWh out, but in winter this will drop to say 6kWh?
mpartington not sure. But I was consistently adding 9kWh to my 8.2 battery overnight last week.
Taking out over 8 - when nothing was added.
mpartington Just trying to get a handle on how much I'm seeing is due to temperature Vs the firmware limits. Mine is operating similarly to @TheDragon (GivEnergy) 53.0 to 51.7V, which is only approx 70% real capacity.
We know the charge energy is about 7% high in the app Vs reality, so I actually added 7.5kWh. 7.5/10.4. = 72%.
However my discharge (as expected) is much lower than this, about 6.3kWh yesterday, so 60%.
My conclusion would be I've lost 10 % in excess of the 80% dod due to current implementation of limits, and maybe another 10% due to the temperature.
This is made more stark by the fact we appeared to be getting more than 80% dod in 3007.
mpartington This is made more stark by the fact we appeared to be getting more than 80% dod in 3007.
The App readings vs reality was also out in 3007 too. But 4% high for charge. Didnt get discharge data for that release from the tester who helped me
TheDragon (GivEnergy) whilst both out, the amount of variation to each other looked quite similar at max charge rate, so I think are comparable (if not necessary correct). I'm scratching my head and wondering more about the theory of LiFePo4 batteries behaviour though. Obviously the overall battery capacity reduces with cold, but is this driven by being able to put in less charge, or is it all down to a reduction in discharge capacity?
If I need 10kWh to fill my battery at 20 degrees, will it be the same at 5 degrees?
So on a cold day, do you put the same energy in as you would on a warm day, but get less out? Or do you put less in as well?
So I've cycled mine now a few times and I'm pleased to say I haven't seen any SOC drops at all.
However, what I am seeing is reduced capacity on my 8.2kWh 100% battery. It charged from 4%-100-98% overnight but the portal claims this only took 7.6kWh of juice.
I know from experience without even looking at the values/percentages that it isn't lasting as long as the discharge graph curve is steeper than what it has been the previous 6 months.
I guess they have stopped the SOC by just removing even more of my battery capacity? If this is the case I'd rather have it back to how it was and I'll just manage the SOC drops myself.
So, just to keep @THALL happy 😀 I phoned GE this morning (waited about 30 minutes to get through). They didn't think the differing cell voltages were an issue and have triggered a calibration.
We shall see.
judgepd but I thought they had calibrated twice already?
On another note I appear to be missing about 2.5kw capacity from my 17.7kwh so around 15%. But last time this did recover after cycles 5 and 6 which should be tonight/tomorrow night for me since the full calibration.
THALL Yes, they did one that they told me about, and I suspect another one that I didn't get told about. Another calibration was the only help they could offer.
judgepd Was firmware rollback not an option? You did not have this issue pre beta did you?
Sorry to invade, but I’m watching this closely and can’t determine if it’s now working or still buggy.
What’s the general consensus?
I’m pretty close to walking away from GE, so it’d be good to make an educated decision.
Ra I guess they have stopped the SOC by just removing even more of my battery capacity? If this is the case I'd rather have it back to how it was and I'll just manage the SOC drops myself.
But aren't we all losing a lot with the previous firmware anyway? For example yesterday my 8.2 100% DOD 3007fw charged up to "100%" with 3kW of solar, by the late evening I had used 57% then boom, it crashed. I "lost" a lot to export that in theory should have gone into my battery because it really wasn't at 100% when it stopped charging.
THALL I didn't, no. I'm going to see what this calibration achieves and take it from there.
chris_p When I was on the 'ordinary' firmware, if I had an SOC drop and then fully charged, it would take about 9.5kWh of juice to get to 100%. Now, without an SOC drop, it is 7.6kWh.
I don't have givtcp etc so cannot interrogate the system to any great level of detail but something isn't quite right somewhere. Will give it a few more cycles to see if it is actually learning anything at all and go from there.
kram Depends how you look at it. Yes this firmware has fixed the SOC jump issue, I think all testers would agree on that. But it has made some other issues. Personally for my system it seems fine, but does take a couple of weeks of cycles to settle down. Professionally I would not release this as a general release anytime soon. It feel's 50% ready at best, there is issues to work out. However I think this is fixable, the hardware seems fine and software can be tweaked. An outside issue will be warmer weather, I think this firmware needs 6 or 7 full cycles and warmer weather should mean less battery use and I would not want 6 or 7 calibrations to happen.
@chris_p I agree. It seems like the soc jumps are fixed but now some testers are missing capacity and I think they are different issues. For me I am missing 15% but I think this will drop to 5-6% soon. Other people are missing 80% etc. So fixed 1 missing capacity issue but uncovered/made another missing capacity issue.
judgepd I would like to know if they can downgrade remotely. I think the downgrades I have read about were done manually which I hope is not the default.
mpartington battery capacity reduces with cold, but is this driven by being able to put in less charge [...] on a cold day, do you put the same energy in as you would on a warm day, but get less out?
Really intelligent question. I'd love to know the answer.
If speculating with some physics: the batteries get hot during charging. Heat is the only energy loss from this system - because it's not doing any other work.
If we knew the Specific Heat Capacity of LiFePO 4 (and we already roughly know the battery weights and the temp increases) then we could calculate how much energy is being lost.
It would then be possible to do this measurement on a cold battery, then repeat the experiment having pre-warmed it (e.g. with a fan heater) and run it again. This would tell you if the energy losses are "fixed" regardless of initial temperature.
THALL I would not agree (just yet) that the SOC jump issue has been fixed.
I had a 50% to 100% jump six days into testing. That's not to say that it won't sort itself out (given long enough) but the time taken to achieve this is heavily dependent upon battery usage pattern. I do not consider the problem fixed if it takes (as I estimate in the case of my usage) nearly two years of normal operation to realign the displayed SOC%. The method is showing promise but the practical implementation needs more thought.
My 8.2kWh (100% DOD) battery is now showing a useable capacity of between 5.3kWh and 6.8kWh. Definitely an issue that needs addressing.
Woodstock I am going to assume GE have not commented on that soc jump?
If there is 1 report of a soc jump on the new firmware that can not be explained by GE then this firmware is not a fix.
Does anyone have information on why ramp down charging has been introduced above 90% SOC?
It occurs to me that lowering the charge voltage at higher states of charge allows for a greater disparity between displayed SOC% and true SOC% to exist before a high voltage alarm is triggered (and a sudden jump of displayed SOC% takes place). Is this the reason for the new charge throttling or just an unintended consequence it being introduced for some other purpose?
It seems to me on a 100% DOD it shouldn't be necessary either as there is both a top and bottom buffer so it could charge at maximum rate up to a displayed 100%.
Woodstock
My assumption/guess: The voltage that the system needs to get an SOC value, is when the battery is idle - so no current flowing in (i.e. artificially high) or flowing out (artificially low). By reducing the charging current in that last 10%, they may be hoping to see a voltage which is more useful to them, than if it was still charging at 3.6kW, or whatever it might be.
Or perhaps it reduces the time which is needed at idle, following a charge, for the voltage to stabilise (and hence be valuable to them)?
Reinstalled 3088 on 17th Jan. After running out the last couple of days, the system had chance to idle and readjust (points 1&2). The allingment between my batteries and soc is now looking really good.
Master battery still needs a kick at the top limit though, as refuses to register more than 92%, but realigns. Hoping it will keep edging closer.
This is reinforcing what we all know that the software can work well, but it does need multiple full cycles to do so (I haven't had a calibration).
Haven't seen any soc drops or jumps since the first few days, when it was testing the boundaries.
If the underlying capacity issue can be fixed (that's the showstopper in my eyes), and maybe a user selectable option to calibrate on firmware install if they wish to speed up process. Could add a single 'one time use' button that appears for a short period after an update. The user can choose to press (or not) at a time of their convenience to initiate, then it's definitely on the right lines.
judgepd So the calibration has just finished after kicking off first thing this morning. It discharged to empty and charged at full rate until it got to the previous incorrect 100%. It then charged at 1kW for the rest of the day and added 9.2kWh in total. I hope this sorts things.
judgepd I take it you set the 1kw charge speed? That wasn't a fault?
THALL No, I didn't. I had it set to max the whole time. It went full rate up to 100% and then 1kW for another 6 hours at 100%.
judgepd Not sure I have seen it go that way before. Though saying that when mine was done last week it got to 0% and kept exporting at full 3kw for just over an hour 😃 Magic power.
THALL mine did a very slow charge like that.
It was saying zero when it was probably more like 70% full. So it had a long way to find true 100% !
Last night my charge was a bit odd, so pre mid night battery got to 89% and stopped charging, the voltage spiked to 56.69v. 1 hour later my battery soc climbs to 93%, and I do mean climbs it is not a jump. But to get from 89 - 93% apparently it was not drawing any power. So it must have corrected itself? Though it didn't charge anymore than that. It stayed at 93% from 0055 to 0230. Screenshots available on request.
THALL Hmm, that all seems a bit odd especially it stopping at 93% and not 100%. My first overnight charge since calibration looks okay but this morning the voltage has been climbing despite not adding any charge to the battery. Also still seeing the behaviour of the SOC dropping during the timed charge period (from 100% to 96% here).
My guess is that
Thall - your charge was stopped at 89% because of the 57 V maximum protection
judgepd - your charge reached 100% but the voltage was lower than 57 V.
My charge last night stopped at 56.7 V / 100% and stayed there, no drop to 98%.
The voltage fell to 53.8 V over the next 90 mins while idle.
The top end of the program works for me but the system capacity is rubbish, battery at 21C so I think it is the bottom end voltage that is the issue.
Newman but the system capacity is rubbish
I really shouldn't speak too soon as mine was only calibrated yesterday, but so far today it's discharged 1kWh and dropped 19% which gives a rough capacity of 5.25kWh. Hopefully this will improve after some more battery cycles.
Newman Agreed. But how long before it should continue charging? There was 2 hours 35 left on my charge window. Also why did it only correct to 93%? I wish they had sent a set of release notes with this firmware.
judgepd Same here!
Just done first full discharge post calibration and from full to empty discharged 5.9kWh. Whilst this is much better than before the calibration, I don't really understand why this isn't showing the full capacity so soon after it put in 9.2kWh during calibration.
Hopefully like @THALL has seen, this will increase quickly with the next few cycles but this would need to be very clearly communicated with users otherwise it's going to lead to an awful lot of support calls.
judgepd Sadly I have not had full discharge cycles recently as not been at home to use it all. But if I trust the dashboard and look at the percentage drop vs what it says it has discharged my loss has dropped from 15% ish to 10%. Thats over 5 partial cycles. Again that is based purely on the dashboard data I have available too now.
judgepd It just sat at 5% for 30 minutes, not discharging and then the SOC adjusted up to 8% and it started discharging again and got another 0.5kWh out of the battery. It was still discharging but I needed to start charging again to be full for this evening. This is certainly new behaviour I've not seen before.
judgepd Sounds like my last nights charge. Hmm
judgepd That is the OCV adjusting upwards, the actual SoC was higher than reported, so it adjusted up, so more energy was made available, then discharge back to 5. This is normal behavior on an early cycle
TheDragon (GivEnergy) That does make sense, but what I still don't understand is if it did a calibration yesterday, why is it already having to make adjustments to the SOC/voltage relationship?
Not expecting you to have all the answers either!
Hmm my system just stopped charging at only 92% and I am exporting 1kw solar to the grid, been like it for 10 minutes...... This is not right.
Edit, voltage spiked at 90% 0.5v then another 0.5v at 91% and hit the 56v limit at 92% soc. So last night missed the top 7% soc charged up from solar today and missing the last 8% soc..... I dunno what to think. First week of beta issues that needed a calibration. Second week of beta missing 7 - 8% soc hitting the voltage limit so different behaviour to week one.
NEED A STATEMENT FROM GIVENERGY
THALL NEED A STATEMENT FROM GIVENERGY
Couldn't agree more.
What is a statement going to do? They won't be able to do anything until after the Chinese team return to work. We additionally have to hope that the Chinese team are able to work when they return back and not off sick.
I suspect it will be early Feb before we have a revision (week after holiday ends), I doubt a statement will be coming.
I am not seeing any improvement after multiple cycles, one battery jumps from 1% to 100% with minimal charge and 100% to 1% with minimal discharge which means that 4-70% is FAST and 100-30% is FAST in-between is ok with 25% loss of capacity. (2x9.5's)
On the positive: My baseload is now 3w as measured by Octopus when running completely on battery.
A statement? Who would believe one at this stage? They been issuing "it'll be fixed in two weeks" posts for months. Then the infamous "give us 4 weeks and it absolutely will be fixed on 16th December" from their tech director, no less. What's next? Pinkie swears?
Here's my totally non-technical view- they don't know how to fix this and even if they did, it wouldn't apply to many users.Theyve been at this for the best part of a year and it still isn't sorted.Ally that to rumours that older inverters need a physical modification. My system is less than 12 months old and, if the rumours are factual, will need a physical modification. I've offered my system for beta testing twice and heard nothing back, which adds credence to that.
In the meantime, GE are happy to continue selling and installing defective systems. This problem won't be sorted until new BMS systems are installed in new batteries.
I am fortunate in that my system seems to work OK. Do I have any confidence that it will continue to work perfectly for it's expected lifetime? Hell, no.
Face it guys, you've been had here.
Having been keeping an eye of this and the other thread, and offering to take part in testing, what I would like, if not having not already been posted (which means I've missed it!!) is a clear explaination of what you are expected to see or happen through the process of the recalibrations etc.
I must admit I am concerned about the fact GE are beta testing on systems that the users (you lot!!) have been playing with and have differing set ups.
Anyway, could someone enlighten me on what is supposed to happen from when GE update a system and what a tester should expect to see and experience? A step by step would be great.
Sadly, GE will never do anything as logical or prefessional as informing testers what they should expect to see happening. Testers should have clear objectives. Simple as that.
Stevef1964 Anyway, could someone enlighten me on what is supposed to happen from when GE update a system and what a tester should expect to see and experience?
I just got told the firmware would be installed and that is it. The email also said "we will work closely with you for any issues or bugs that arise you find in an email discussion“ which they blatantly haven't done.
A statement is to show GE UK are monitoring and give a poop. Some update on testing results, or at least say how its going. Issues fixed, issues found. Something official rather than us testers posting on this forum.
I want a way forward with this. A fix. Replacement kit. Full refund. Hell at this point if i got a refund I would just order a powerwall to know I would have a fully working solution regardless of cost. Id prefer hassle free at this point.
I find it quite offensive that people on here are prepared to give up their time, and in some cases their systems, to try and help GE to get on top of this, yet get nothing at all in return - beyond non-existent carrots. Yet again GE has failed to meet even a minimally acceptable level of communication.
THALL I rejected the Powerwall route specifically to try GivEnergy, UK company and all that. I was at best 50:50 on going ahead when I became aware of the issues but did so on the back of the "definite fix" statement. I covered myself by paying for some of it on credit card, however group action might be more appropriate here if no resolution is forthcoming sharpish? It's not just a refund, it's wasted time, wasted generation, uneccesary grid import etc. GE really need to step up and advise soon as patience is now thin. They don't really have a defensible position. Come on GE, talk to us!
WARNING: The following is a brain dump. There is no guarantee that any of it is true but it may be useful to someone, somewhere.
.
.
.
.
Finally the planets aligned and all the conditions were met on my system for a displayed SOC% adjustment to take place!
Between 05:25:46 and 06:26:08 this morning my displayed SOC made a dead straight line adjustment from 49% to 13%. (NOTE: JUST OVER 60 MINUTES. See translation of the 3011 release note (around post #2004) in "the other" thread for the significance of this).
All that follows relates only to the firmware currently installed on my system, which is
Inverter D0.290-A0.290 Battery 3011.
I believe that for an adjustment to take place, all four of the following conditions have to be met.
Let's take these one at a time ...
Condition 1.
This is an indication that the battery is on the relatively flat bit of the voltage versus SOC% curve. A difference in measured voltage in this region is going to have little effect on the looked up SOC%.
Condition 2.
If the charge/discharge current is low then the battery voltage is not going to be raised or depressed by an amount large enough to badly affect the lookup (see condition 1). Also this ensures that we really are on the flat bit of the curve and that the battery voltage is not being pushed below 51.5V by a rapid discharge.
Condition 3.
Because of conditions 1 and 2 we know that we are at a true low state of charge. If the displayed SOC were to start at over 50% any adjustment would present as a large and steep drop. We don't want to see that, do we!!! Presenting a less than 50% drop stretched over 60 minutes (see release note referenced above) is less likely to be noticed and/or worry customers.
Condition 4.
This stops the table look up looping continuously. It also prevents what would otherwise show as a bad case of the jitters on the lower portion of the displayed SOC% curve.
.
.
.
.
I believe that the inverter having to be idle is a red herring brought about by condition 2 (or maybe it was a requirement in an earlier beta release).
There does not appear to be a requirement for a voltage stabilisation period either (this could also relate to a previous beta release).
I can see a flaw in the above four condition method as in the 60 minute period during which the displayed SOC% undergoes its straight line adjustment, the true SOC% will have moved. This will have an impact if the displayed SOC% is subsequently used anywhere for control purposes. Also, if the movement is large, it will make the displayed SOC% less than useful for a customer scheduling their charging/discharging based upon what they think is currently in their battery. Kind of back to square one with a displayed and true SOC% that do not match!
It is largely because of condition 3 that I have not identified an adjustment taking place until day 11 of testing. The other three conditions having played a lesser part by preventing adjustment when the displayed SOC% WAS below 50%.
Condition 3 is bad news if the displayed SOC does not drop below 50% for weeks at a time (my system in the summer)!
As yet I am unsure whether SOC% adjustment can take place under other circumstances. I suspect it can/does around the upper voltage limit, but via another mechanism not yet fully understood. The conditions I have set out above are, I believe, what is being outlined under the heading "SOC calibration at low voltage" in the translated 3011 release note referred to previously.
How many man hours of testing effort (and days of elapsed time) have been wasted; all for the want of a decent English translation of the release notes dating back to the start of beta testing? The only information I have seen is a Google translation of a copy of (what I take to be) the incremental changes made for version 3011. Talk about Chinese whispers. Ouch!!!
joshgreenuk I suspect you have a fault with the rapid %swings, rather than this being a feature of the beta software. I'd be on the phone if I were you.
stevelewis fine before beta was loaded, the jump happened immediately after the latest bms was loaded.
Will see what they say when holidays over. Today I generated all day, low steady 4-500w overcast and foggy. Minimal drops rather than consistent drops.
sponev
I’ve spent considerable time trying to get a diagnosis of my failed system. I’ve looked through the logs on my app and it appears things all started to go wrong on 15 Jan 23 or thereabouts. My batteries are flat and my inverter is permanently idle.
After seeing this and a couple of other threads I’m now suspecting ropey firmware.
I don’t think I was part of any beta testing thing but it looks like these things are updated remotely without any interaction from me.
Might I possibly be on right track?
RBilton What's your Inverter model and current firmware revisions, out of interest? I had the "totally idle and flat battery" on my Gen2 HY3.6 with f/w D0.906-A0.906, and battery 3009
Cycle 14
2 hours into a 5 hour charge window. System charged to 90%. Voltage hit 56.5v. Charge stops. Data shows 2 brief charges later in the morning 90 - 92% within 5 minutes. Then a large voltage spike to 57.09v and charged to 93%. 1 hour 30 left on the charge window. Didn't get any further.
7% missing at the top. Total capacity missing is around 21% now according to dashboard data, 21% of 17.72kw.
Might try calling Dan today to see what he thinks. No replies to my email chain.
Out of interest does anyone know what options I have for getting this fixed? Or US rather than I. My original AC3 and 8.2 batt are 18 months old so no option for refund. My 9.5kw is 3 months old and my installer is aware of the issues and open to removing and refunding. A hardware replacement seems pointless if the issues really are software based. Are there any options other than to just keep testing?
THALL what's your resting voltage at the top limit? 53v seems to be the maximum on this firmware version, which we believe needs to be increased as a part of the fix to the reduced capacity. Mine isn't showing 100% at 'full'.l.either, but I think it is as full as the firmware will allow and this is just indication.
mpartington 54v for me. graph data shows lowest 53.94v up to 54.5v
THALL odd, seems like it's constantly charging with the spikes. Although mine only peaked at 54.5v, which I think is the maximum. Maybe yours was attempting a top point calibration or something and want higher. This is my soc graph I missed off the last post. Discharge starts at 1.30am.
RBilton I'd be inclined to call them asap. It's unlikely you are on the beta firmware if you didn't apply for it. That said, I did apply for it, was told I'd been put on it, but actually wasn't. They've failed to answer any emails since! Call them.
mpartington As far as I know those trofs on mine are when the system is adjusting itself. It does add 3% more during those points but it reports only 0 - 340w at anypoint, so it appears to find capacity without putting more juice in. But doesn't get over 100%. Week 1 of the beta pre calibration I could hit 100% everyday. Basically 2 weeks 2 different issues.
My charge from last night (not from empty) would indicate a total capacity of around 7kWh and the discharge so far this morning is suggesting a total of around 5.5kWh.
@mpartington Be a bit wary of the voltages from GivTCP as they don't seem to tally with the GE portal and are usually higher (admittedly less so when the battery is closer to full or empty)
5 cycles
rubbish capacity = 6.8kWh charge / 6.3 plus 0.6 kWh discharge
Last night battery went Idle @ 5% SOC and 52.3 V
SOC corrected to 16% then discharged extra 0.6 kWh to 5% SOC and 51.4 V
I hope this corretion was to span of SOC otherwise top may now be wrong.
This software is far superior to my previous system, things (on my system) are getting better. I would not be prepared to go back to what I had.
51.4 V is equivalent to 18.5% SOC, so there is some more capacity available but doubt it will reach 8.2kWh.
Newman After my 6th cycle (pre full calibration) my system appeared to drop soc at twice the speed my inverter should be able to do. The actual capacity percentage worked out to over 6kw. Full calibration then happened by Dan at GE and first 5 cycles were fine, cycle 6 and last nights 7 my system refuses to show a soc of 100%. Max voltage kicks in and stops the charge.
I started at 0230 this morning with 93% charge, I am down to 28% dashboard reports discharge of 9.04kwh. My total system capacity is 17.72kw.
Woodstock Screenshot of the planetary alignment referred to in yesterday's brain dump.
See displayed battery SOC% adjustment between 05:25 and 06:26.
Note:
Adjustment started when battery voltage dropped below 51.5V.................... Condition 1.
Battery current was less than 4A when adjustment started................................ Condition 2.
Adjustment only started after displayed SOC% had dropped below 50%... Condition 3.
It was more than 24 hrs since any previous adjustment...................................... Condition 4.
Displayed battery SOC% adjustment lasted 60 minutes.
My battery has just emptied. Total discharge 14.4kw. Started at 93%, stopped discharging at 6%. Reserve is set to 4%.
So adding up the 13% soc I can't access, total capacity 16.71kw? Missing 1.01kw so efficiency and the cold? 6% loss?
Does this sound right to anyone?
Mine has just emptied, discharged 5.8kWh. Still some way off 8.2kWh. I wonder if the SOC might move up a bit and discharge some more.
judgepd good point. I've just pulled the voltage from the portal into GivTCP as a new sensor to take a look. I guess the portal voltage is adjusted to reflect the inverter calculation, where GivTCP are the battery voltages directly calculated by the BMS.
Managed to get 7.8kWh from my 11.8kWh usable range today. Full to 7% so far. Definitely well down on what I was getting with firmware 350 and battery 3008.
judgepd The SOC did move back up to 8% and then discharged more to make a total of 6.4kWh.
judgepd that's about what I'm getting from my 2 X 5.2s, which have similar usable capacity to you - if you assume 80% dod.
It's too low, but we are consistent.
sponev I’m in communication with both GivEnergy and the installers. Looks like I had some sort of BMS detected failure on Jan 15.
They’re investigating.
Riscy I’ll have to check tomorrow but support told me my battery is on version “0”.
I think this might be a problem.
Charged from '0%' again overnight. Portal is suggesting it has only put 7.2kWh into my 8.2kWh battery.
Looking at the voltages I'd suggest its the lower end of the voltage scale where it is nowhere near. I don't think I've seen anything beginning with a 4 since the update.
RBilton My battery firmware version is OK at 3011, but the inverter firmware is D0.999-A0.999 which feels odd but I'm assured is normal for the beta (the gen2, at least.)
I'm not on the beta software (I offered three times and never heard a peep), but I have noticed lots of people who are, talking about reduced battery capacity. I am wondering if this might actually be intentional on GE's part, to reduce the chance of damage to your batteries, during this testing period? I realise that this goes against everything that I believe about GE, as it would suggest that they have put the customers first. Just a thought....
ToothyChris If that is the case then their advertising is simply wrong and would get blowback for it. They claim 100% depth of discharge. However based on my usage yesterday I am not sure I am missing much capacity depending if the soc calculation is correct vs the meter monitoring whats gone in and out.
Current issues for me
1 where has the top 7% of my battery gone, why can't I charge past 93%
2 why is my reserve of 4% being ignored and I can only get to 6% (I thought this was fixed)
3 is soc % displayed incorrectly. The first 50% drop indicated I had used a lot less than 50% capacity yet by the time I got to empty it worked out far better.
Sadly I was expecting to get a call back from Dan last Friday and didn't, then after chasing yesterday morning I was meant to get a call back and didn't. The UK GE staff sound like a really small team and as others have said while the Chinese are on holiday nothing will happen. That currently seems to be the case.
Ra I don't think I've seen anything beginning with a 4 since the update.
I hope this isn't how they think they can avoid the crashes to 0%.
ToothyChris I am wondering if this might actually be intentional on GE's part, to reduce the chance of damage to your batteries, during this testing period?
Surely that would be an easy thing to communicate and head off everyone who is reporting it? But then this is GE, kings of crap comms.
3rd cycle since calibration and added 6.8kWh from empty to full. Not much more to add really.
Start of cycle 6 since calibration.
7 kWH charge - 51.6 V @ 4% to 53.75 V @ 97%.
Previous charge was to 100% / 56.7 V Still some improvement left to make.
I would accept 51.0 to 57.0 as full range.
Query, does anyone have multiple inverters and are on the beta firmware?
When the EMS arrives, I may add another AC3 and battery. I assume GE will then have to develope new software so will not be putting any effort into multiple inverters now (they have too many software problems already).
Newman My issue is the when for the EMS. I am stuck on the givenergy path now and could really do with a second ac3. Target Q2 was the last thing I was told about it. But I was wondering if the beta works with plant inverters and if there are any improvements for plant systems with the beta.
Has someone been playing with the portal software?
Just tried to set a timed charge (which failed) and I now cannot stop all my solar production being exported to the grid!!
Tried a reset to defaults. All the displayed settings changed in remote control and the log, but still exporting everything!
Tried a restart. Still exporting everything!
If this is someone making changes to live portal software on the fly I will be very annoyed!
Woodstock No idea. But there have been a couple of threads recently about timed charges setting not staying set.
THALL Yes, thanks. That's why I thought someone may have been playing to try and fix the problem.
Think I am getting to the bottom of it now though. I was basing "all solar export going to grid" on my smart meter but I have now spotted that a continuous 250W is actually going into the battery. Seen this before and I think that it could be that a high voltage limit has kicked in and is throttling battery charge. Looks like it kicked in at the same time as I was setting the timed charge so I was already in the inverter card and not viewing the dashboard. I tried amending settings which according to the smart meter were having no effect and it was not until I went back to the dashboard that I saw what was going on. Unfortunate coincidence that throttling kicked in at the very moment I was updating an inverter setting!
Why throttling has kicked in when the displayed SOC is 72% is another question altogether. 😄
Woodstock Hmm I see the throttle at 90%, but recently I have hit high 56v and my system stops charging and over the course of the next hour it finds another 3%.
I dunno. All I can do is report to others what my dashboard shows. But chatter about this issue appears to have slowed.
THALL I didn't have to wait long for the answer. It came in the form of a sudden jump in the displayed SOC from 78% to 100%! Throttling kicked in because the battery was nearly full. Ironically I was expecting this (based on the stabilised battery voltage at a displayed 4% SOC from last night). This was why I was setting a timed charge; to ensure full would be approached at an even rate and make it easier to observe what was going on with the high voltage processing. The displayed SOC% was reading even lower than I thought and the battery beat me to it!
The premise for the beta fix seems to be that a displayed SOC% adjustment will take place (if required) ...
A. At full, if the displayed SOC% is lower than the true SOC%.
B. In the lower half of the displayed SOC% curve (displayed SOC less than 50%) if the four conditions I have already set out are met. This should handle displayed SOC% being both higher and lower than true SOC%.
The problem appears to be that if the displayed SOC% has drifted below the true SOC% then a lower half adjustment will never take place. This is because discharge stops when the displayed SOC reaches 4%; before the battery voltage has dropped below 51.5V. Thereby ensuring condition 1 is never met. (Last night my displayed battery SOC was 4% for over eight hours and the stabilised battery voltage was between 51.7V and 51.8V for the whole period).
The above suggests that if the displayed SOC% is lower than the true SOC% then an adjustment will only ever take place at full. If the battery charge goes up and down, even to empty, (but not full) over a long period then the displayed SOC% will not be adjusted. Of course this means that when the battery eventually does fill a sudden jump will occur.
This is all supposition based on observation but (in the absence of any indication from GE on how the beta firmware is supposed to be working) it's making the best of what's available! What is not supposition is the displayed SOC% sudden jump showing on my dashboard!
Can someone please explain to me why it is thought sensible to use the displayed SOC% to control the system (stop discharge at 4%, stop charge at 100% etc.)? We know the displayed SOC% does not always reflect the true SOC% and you couldn't think of a better way of losing battery capacity than stopping charging before it is really full or stopping discharging before it is really empty! I suppose it makes the dashboard look consistent (until you get a sudden jump) but that is all that I can think of.
Woodstock I was told in a help call a long time ago that the system worked within the maximum and minimum voltages (not to the SOC) and therefore I was still getting the full use of the battery capacity even if the SOC jumped....this is very clearly not the case.
Sorry but I find it funny that everyone is trying to figure out what the firmware is doing at point x, y etc. The people we are testing this for have ALL these answers! They wrote the logic! How this has been installed on lots of systems now without a full brief of what the expected behaviour is of the software is beyond me.
chris_p Sorry, but I don't see why they need to say anything other than its meant to solve the SOC errors seen by some. How that is done is probably commercially sensitive as if it turns out to be industry leading it will give them an advantage.
As beta testers, you (Not necessarily you Chris as I don't know your status) were quick enough to jump onto the program without asking the question before jumping in. Now its not seemingly going as you thought it might, you are now asking the question.
I offered my system for the testing program, but my inverter is one of the older ones so is apparently not suitable at the moment - as it happens any loss of capacity to me currently wouldn't make a blind bit of difference this time of year as solar won't come close to filling it and I don't have a cheap over night tariff. Obviously others are not in the same boat - but the drop in capacity had been noted sometime ago by the Dragon.
anglefire Nope not in the beta, my inverter too old too, although again (like you) I'm only guessing that from third hand information as GE never bothered to reply to any one of the three times I offered.
Just thought they might help themselves by communicating more than just "load this up" which seems to be about it. They don't have to give away any trade secrets by saying "the firmware attempts to do this, in scenario b you should see this happen etc". No denying it is just a continuation of this farce.
Good morning testers. Day 16, cycle 16 for me and I appear to have a soc jump this morning. 82% straight to 100%. I have made GE support aware. Hit 57.83v.
Also interesting how your battery is actually showing 100% and staying there rather than dropping immediately to 98%
Do you think it might have been the additional demand/load in the house about that time that has thrown it in to a wobbly?
Ra The 98% only appeared for the first 5 or 6 cycles then it pegged to 100% as expected. The 98% oddness came back after a full calibration was done by GE. My system is set to charge to 100% until 0230. But yes it stays at 100% until 0255 even under 1500w load.
Thought it had gone rather quiet on the other thread - you've all decamped to this one!
Just fed the info below back to GE, but disappointed to report a very noticeable SoC jump from 90% to 100% the other morning. The other issues (below) I had already fed back to them weeks ago, but never had a response. A quick scan above suggests others are seeing the stalls and jumps on the beta. I know others have also seen the fast leakage of SoC and reduction in capacity. So it is not just me and GE must have been aware of most of these issues for a while now.
I know this is a beta and won't be perfect until complete, but for me it is the complete lack of communication with the beta testers that are spending a long time watching, analysing and feeding back that really gets me. I really wanted GE to nail this (not just the functionality but the process too).
I am on beta FW 3011...
On Weds morning I witnessed a SoC jump from 90% to 100% - I was watching at the time because it was curious that the charge power was <200W when the SoC was only 90% and battery temperature at 18C (so no need to throttle). Next minute the SoC jumped to 100%. So this looks very much like the SoC jump issue is not fixed yet.
Inline image
I now have 4 issues - 3 that I have reported about in this thread, and another that I know others have seen:
1) SoC jumps are back - see above. Charge power slowed then SoC jumped +10%. As I understand it, this is not supposed to happen with the new algorithm. This is after >10 full cycles of the battery since the last beta FW update so I don't believe this can be considered as part of settling time for the new algorithm.
2) Long pauses where the battery stops taking charge even though it is not at 100% and there is charge power available. I have watched this happen at 95%, and 98%. It sends any excess to the grid when apparently it could still be charging the battery. When I notice this now, I tend to hit Restart in the portal and after a short while it springs back into life, charging. So this appears to be some issue with the FW getting into a state that it can't get out of.
3) Leakage of charge once reserve level has been reached. There is no power discharging, though you can watch the SoC gradually decrease over minutes. This used to happen over hours/days, now it is minutes/hours. So there looks like there has been a huge increase in the kit's self-consumption. Or this is a side-effect of the new algorithm?
4) Reduction in capacity of the battery. Recently I have been doing forced charges in the morning at the end of my Eco7 window. These are always from "empty" to 100%. It varies day to day but I seem to only be getting 3.6kWh into the battery during this time. This is below what I would expect from a 6 month old 5.2kWh (80% DoD). I know others have reported this issue.
The communication on this beta program has been one-way. My questions never get answered. I think this is unfair and a big mistake on GE's part.
Are the issues I reported above known to GE already, and actively being analysed? Is there new beta FW already to address some/all of these issues? If so, can I try it?
Where are GE with the beta program? Is it finished? Or is it on hold while new issues are being fixed?
Is there an ETA for known issues to be fixed?
JaseSpace just to confirm, there is only 1 uk engineer working on the beta. There is no team. All emails about it end up in his inbox.
THALL that inspires me with confidence!
94jdh Apologies, I wish I had better news. But it seems any fix is reliant on the Chinese parent company and their engineers.
THALL This sadly confirms my suspicions. It would strongly suggest that a fix is not in sight and that Giv are basically keeping their heads down because they don't know what to do or say. There are clearly lots of issues still and no way that this could be released generally. I guess it's ultimatum time for our installers - rip it out and provide kit that is reliable. How much longer do we give this?
sponev And more and more people on the Facebook groups reporting SOC issues, getting to crunch time for GE. They can't keep hoping for the sun to shine to hide this issue while they bury their heads in the sand.
sponev I expect like me a lot of us are tied to this kit now. My AC3 and 8.2 battery are 19 months old. While I could have my new 9.5kw removed and refunded I then lose out. But it is not all doom and gloom, while it appears a chunk of my capacity is not available to me digging through the data suggests the capacity is there and looking at the battery percentage on the dashboard and app are not showing correctly. Plus the faster ramp times. So there are some perks. I need to try and see the light and hope for an amazing fix. But, will I recommend the givenergy equipment to anyone, NO. Will I advise anyone who has recently ordered or thinking of ordering to cancel, YES. Based purely on my experience over the last 4 months.
THALL Interesting that you experienced no charge throttling and your battery voltage was so high. Here is my "sudden jump". My max battery voltage was 55.48V
1 thought, I have mixed size batteries. Maybe that is why I am seeing higher voltages? Really dont know.
Another good one, the sun is out woohoo generating 4.3kw, solar charging shows no big soc jump like last nights charge. Battery got to 90% and charge speed throttled, got to 98% held that for 15 minutes and has now jumped 98 - 100%. Now dumping 4kw into the grid. The DNO will love me. I am still waiting on a ct clamp to limit grid export from my newer inverter, has been out of stock since November 😆
THALL 1 uk engineer working on the beta
Source?
TheDragon (GivEnergy) GE support when I phoned to speak to Dan earlier this week. He was busy training new installers I asked to speak to anyone else dealing with the beta and was told there is only Dan. I never got a call back or email either.
So the "Team" is China then?
TheDragon (GivEnergy) probably referring to the uk 1 person collating the data to send to china (Maddocks)
We know its all in china because of the 0 updates during the China holiday period, that is kind of a dead give away really.
TheDragon (GivEnergy) I assume so.
Just checked my phone log, I called Wednesday at 0949 left a message for a call back to discuss next steps.
Guessing that is Mr Yu's team then.
Ooo another offshoot company name to add - http://www.ladon-tech.com/
LOL that Ladon tech's designs are 1:1 copies of Giv's inverters.. obviously...
to be fair page 2 just has GE stuff aswell... http://www.ladon-tech.com/ProductInfoCategory?
So in theory could the chinese side just decide to cut ties with the uk arm?
I might just shut up and stop posting and wait quietly in the corner for a "fix" . . . . .
THALL yes it could, but its highly unlikely that would happen. Like not even remotely likely.
THALL Out of curiosity, did your 57.83V trigger an over voltage alarm in inverter notifications?
I shouldn't worry too much about not having the CT clamp. Based upon my own experience, any G100 export limit you had in place on your system will likely have been lost anyway when the beta software was installed! Reported, so if and when the fix is finally released I hope it includes code to retain any G100 export limits in place. If not someone is going to be very busy resetting them all manually!
No I don’t think they’ve done anything like that, if you look at his profile it’s not a new company or anything (2019 I think that one popped up)
I just think the same basic kit goes by different names for different regions.
Woodstock Just checked, and no alarm was triggered.
I never export from my batteries, only my solar. But my cap is the default 3.68kw. My installer did say don't worry about it, he has a clamp on order but it hasn't arrived. I don't really understand how everything can be limited though, I have 2 separate PV array's East West and my battery system. Unless the clamp goes on the main feed to the house but then the inverters are in the loft..... I dunno. Not too fussed.
JaseSpace Have to agree with regard to this statement : -
3) Leakage of charge once reserve level has been reached. There is no power discharging, though you can watch the SoC gradually decrease over minutes. This used to happen over hours/days, now it is minutes/hours. So there looks like there has been a huge increase in the kit's self-consumption. Or this is a side-effect of the new algorithm?
I don't really understand why the battery has to leak charge at all - the answer generally given is that this is how it's powering itself - but why - it's grid connected - just use that and leave the battery at a static charge.
THALL Thanks. I have had three sudden jumps to 100% on the beta firmware and no over voltage alarms. None of mine took the voltage over 57.5V ( the alarm limit set in 3002 but may it have changed since) so I was interested to see if your 57.83V had. Looks like yours was not an old school bounce off the alarm then and the high voltage processing has changed. All three of my jumps occurred during charge throttling, while yours didn't. This would explain why your voltage was higher than mine but sheds no light on why no voltage throttling occurred on your system.
New issue
Battery discharged to 95%, still getting 1.4kw solar but the battery is not topping itself up. Voltage is a flat 53.5v.
THALL That sounds like my issue #2:
2) Long pauses where the battery stops taking charge even though it is not at 100% and there is charge power available. I have watched this happen at 95%, and 98%. It sends any excess to the grid when apparently it could still be charging the battery. When I notice this now, I tend to hit Restart in the portal and after a short while it springs back into life, charging. So this appears to be some issue with the FW getting into a state that it can't get out of.
Have you tried Restarting the inverter to see if it springs back into life?
JaseSpace Nope, not at home to mess with it. I am wondering if this non charge issue is related a side affect of the reduced charging speed over 90%. I am wondering if it needs to drop below 90 before charging will start again asin this is by design rather than a fault. If this was an EV battery it would make sense, but as a home storage battery will not be at 100% for long there shouldn't be the need to reduce the speed or cap charging. Unless there is something in the battery design which means hitting 100% is an issue? I dunno I am guessing.
Hi All,
I can only apologise, I had been lead to believe you were getting regular comms via email as part of the beta group. I can see that isn't the case. I'll chase up on this and we'll get that rectified.
As some of you have guessed, our firmware team are off at the moment for the Lunar New Year, they're back in from Monday.
As a result, we haven't had any major revisions on the firmware released, but all of your data is being monitored and collated ready for patching when they're back in.
Again, thank you so much for your continued input and support, We truly couldn't do it without you!
Have a lovely weekend.
THALL That's been my experience of the new firmware as well - solar charge unable to get it above 93-96%. Gets to 100% from grid charging during the Go window though.
GeorgeRawding pops his head above the parapet and then ducks back down again, double quick......
I for one feel that these repeated apologies are now rather hollow.
All people are asking for is a reasonable level of communications. This has been repeatedly asked for, but always ignored.
ToothyChris They are utterly meaningless unless the company behaviour changes.
Thanks George.
Here we all are, passing feedback to GEUK on the assumption that it is being collated and passed on to software developers in China. Would it not be simpler to give the Chinese developers a list of the systems involved in the beta test and let them monitor the impact of their efforts remotely? They at least know how the systems should be behaving and are in a far better position to spot any anomalies than we are. Nothing to get lost in translation either.
As long as it was a strict "look but don't touch" policy and any updates required were consolidated into a release (to be installed by GEUK with my knowledge and agreement) I would not have have a problem with this.
Perhaps they are already doing this. Who knows?
The invite to take part in beta testing did include the phrase "Your system will be then closely monitored...", but it didn't specify if this was going to be by GE UK staff, GE China staff or me! 😝
I’m a little confused why the MS Teams group which was setup specifically for Beta testing 2+ years ago isn’t being used for any of this.
When local control was being developed only those who had signed an NDA could get into that particular channel on there whereby the developers were free to divulge exactly how the system works and what it’s supposed to do - the GE engineer only then has to reply once and everyone else knows what’s going on.
This email / ticket black hole thing although probably sounded good on paper - doesn’t really work if the sole person looking after it is too busy to reply to anyone.
hoggy This would be a much better way to handle a beta test group. Team, chat channel, files and documentation available in 1 place. It would be more focused and constant and not so public and just generally faster.
If this beta test was issued a single support person in the uk I totally understand why comms have failed. Especially as its most likely not their only job at the moment.
There are a lot of active testers who feedback on the forum frequently. I post daily and send feedback to the email chain every time I spot something "different".
Help us help you 😄
Wonder if we can sign up an ODM contract and just put an open stack on the battery to talk to a Victron or similar inverter 😂
Think I'm running out quicker than pre firmware upgrade
8% just before timed charge. Charged to 100%. Now at 5%
According to summary at the top of the web portal 3.8kWh in, 2.5kWh out.
Plus a steep drop towards the end of the day too.
Similar yesterday.
7% to 100 to 9%
4.6 into battery, 3.7 out.
6.4kWh out today but again it stopped discharging at 5%, recovered to 8% and then discharged back to 5% again. Seems to be the same every time my battery runs out.
TX200 ... And 3.2 in overnight to get it back to 100% hmmm.
My 8.2kWh 100% DOD battery is operating as if it were a 6.3kWh battery.
TX200 The drop from about 35% to 10% looks to be a lower half displayed SOC% adjustment typical of the beta firmware. Difficult to tell from the image but it should show as having taken around 60 minutes. If you have a look at the battery voltage, I expect that the drop started when the battery voltage fell to around 51.5V.
As well as a reduced useable battery capacity, I am finding that the displayed SOC% is getting out of step with the true SOC% much faster than it used to pre beta. The day before yesterday I had a 22% adjustment and today I have just had a 32% adjustment.
Something similar 8.2 kw with AC inverter firmware 3011 taking 4.84 to charge from 4 percent to 100 emailed GE twice the silence is deafening
Like the teams idea.
Or a whatapp group.
A discord server would be more applicable to me personally... I might even create a 'Owners' one
Screenshots of today's displayed battery SOC% adjustment.
Adjustment was from 49% down to 17%.
I would have thought by this time (16 days after beta installation) that any adjustments would be of the order of 1-2% and virtually undetectable.
Hopefully Monday we will see some updates from Giv CN
6.8kWh in from empty to full then 5.8kWh out from full to empty today. Today the battery didn't recover from 5% to 8% like I've seen every other time it's run out.
Agree with joshgreenuk - Discord would be the best plan - preserves anonymity.
WhatsApp - don't fancy my phone number being public. Teams - for anyone who uses it in business (me, 8hrs a day...) it would be a right pain managing a second login for staying anonymous.
OK, I've been following all this as best I can, and this is what I think I'm hearing:
This 'fix' just really isn't one in real world varied use scenarios
It's working for a few and failing badly for many. Seems like it can cope pretty well with batteries that are constantly cycling from full to empty but the SoC drift still exists with any variance from this this fairly basic usecase. It falls down badly for batteries that seldom get below 50% discharge, possibly because this prevents a downward SoC adjusment as posited by @Woodstock . The adjustment is artificially 'smoothed' so that is does not appear as the vertical drop that it really is. That's aside from the capacity reduction issues.
It sounds like a the underlying issues have not been resolved, and that this is really duck tape, not a 'brand new all singing dancing best in class BMS' as was claimed not long ago.
Is that about the size of it?
Is a total rewrite of the BMS what is probably required here, not just a patch on a fundamentally unsound methodology? (I'm speaking from the results here, not from any knowledge of how to do such a thing)
phoenix My system fully cycles most days. It was fine with the previous BMS scheme. I've seen some niggles with the new version (inconsistent capacity could be weather-related in its first winter, sluggish to start charging/discharging at the top and bottom of the range). But these should be fixable by adjusting parameters in the code.
Capacity loss with age is normal in a Li battery and the largest drop is in the first year... some users may be seeing this and putting it down to the new firmware. Also, maybe some systems have latent defects or just aren't performing as designed. There are only so many things a BMS can measure (volts, amps, temperature) and a "total re-write" risks replacing one set of problems with another.
stevelewis maybe some systems have latent defects or just aren't performing as designed
I suspect this applies to my system.
It's never worked quite properly since install in Nov 2022, including a deep-discharge for several days. The old firmware seems to have been hiding a latent battery issue (incredibly spikey voltage) perhaps arising from that deep-discharge abuse. The beta algorithm seems to rely more on voltage, and is revealing previously undiscovered battery issues. Important not to mix cause and effect.
Are we at the stage of telling our installers to rip it out and supply a working system? It clearly isn't working correctly and GivEnergy is refusing to make a statement. Seems like there are very few options now. Shame but I didn't pay thousands of pounds to tolerate a crock. Dismal.
Battery been idle for an hour, app says 51.41 volts
Guess that's around 18% rather than 5% ?
This morning it charged to 81% then slowed right down, and 15 mins later jumped (corrected?) to 100%
stevelewis My system fully cycles most days. It was fine with the previous BMS scheme
Right, so both the previous and this version seems to work basically fine if you cycle fully. It's the oscillations that don't hit the boundaries that necessarily cause the reported SoC to 'walk away' from the actual SoC. That is not exactly a ringing endorsement.
stevelewis Capacity loss with age is normal in a Li battery and the largest drop is in the first year
Why would capacity loss between firmware versions have anything to do with this?
stevelewis Also, maybe some systems have latent defects or just aren't performing as designed
Agreed. I suggested a while back that not checking hardware pre beta to remove this variable was a mistake that is complicating things
stevelewis There are only so many things a BMS can measure (volts, amps, temperature) and a "total re-write" risks replacing one set of problems with another
Agreed re what can be measured. Disagree that the 'old' BMS calculations, upon which this firmware seemingly still relies at its core, are good enough that risks should not be taken. One customer, who was experiencing large SoC jumps was able to accurately predict them by using a Victron SmartShunt to calculate the SoC and compare it with the GE reported version. If a Victron SmartShunt can do this, why can't GE write code that can too? Or perhaps it's the hardware that is taking the measurements that is at fault, in which case, there will likely never be a solution.
Basically, we were told this new BMS was 'revolutionary', but it isn't. It's a fudge applied to bad data, that's my assertion.
Sometimes you have to tear up what you've got and start over, and 'fear' should not remove this option from the table.
phoenix largely agree with you there. My comment on battery degradation was about the time that has passed since installation and users starting to monitor performance (or lack thereof) in a way that is arguably unheard of in a commercial product. Even with EVs, most people haven't a clue how to get at the cell data.
To be fair my 2+ yo (10MWh throughput to date) 8.2 pack hovers between 6.3 and 6.5 usable from an original 6.5 so I'd be surprised if anyone else's considerably newer ones have more degradation than that.
Ok so soc jump, over 57.21v, and a soc adjustment all in 1 morning
THALL ... and it was only two days ago that you had another SOC jump. Were you seeing them that close together pre beta? My displayed SOC% is departing from true SOC% at a much quicker rate than it used to.
Riscy preserves anonymity
Why? What's the big secret?
Who cares if your called fred Bob, or Barbara.
This isnt the dark web, it's a support group.
My 100% DOD 8.2kWh battery capacity has been hovering around 6.3kWh useable throughout beta testing .... until today.
According to the dashboard it has taken just 0.7kWh to raise the SOC from 4% to 26%. It appears I have a battery with a useable 3.2kWh now!
JaseSpace I'm not one of the beta teters but I have experienced your number 2 issue.
I've not had any major jumps in SoC value but have had a couple on minor drops to zero from 4%.
I'm on D0.499-A0.499 and 3007 with only a 3.6 HY and two 2.6 ECO batteries.
Plus I do not carrry out any timed charges or discharges or have an EV
I leave the system to do it's own thing on ECO and just monitor it.
(Bet it starts playing up now I've written this!)
Woodstock Week 1 of the beta I saw none. I have only had soc jumps since my system was calibrated. Pre beta at least once a week. My jumps were the top 7, 9 and later 13%
Stevef1964 That is interesting. If you are not on beta FW then maybe this stall issue is pre-existing and I have never noticied it before? Either way it seems to be a bug and should be addressed as it is "throwing away" good generation for no reason (OK I get paid a little for that, but I would prefer it was used to top the battery up first).
I moved from 3088 to 3011, calibration yesterday and today got to 10% and it dropped to 0. Emergency charging now.
Wasn't seeing this before on either the 290/3088 beta or on the 350/3008 stable combination.
This is after replacement of the battery with the dodgy cell.
TheDragon (GivEnergy) it's a support group.
Getting more like group therapy as each day passes 😆
kai Do you know what the differences are between 3088 and 3011?
WOOOO I just noticed my firmware has changed to 3079, my timed charge has been turned off!! I have no idea when this was applied, it was not expected, no one told me it was happening.
Was told the differences were minor tweaks. My fishing for details was unsuccessful ☹️
The calibration used 9kWh. I have a 5.2 80% and 8.2 100%, so should have a usable capacity of just under 12. Seems like something odd there.
After the crash and emergency charge yesterday, my 4h go window got the battery to 88%. This is more like the SoC I would have expected pre-beta. Previous on beta it would get to 100% in 3h. So maybe hope yet.
The emergency charge now happened at 290W instead of full pelt. Not sure the beta firmware and the SoC process are playing nicely together.
Well I have no idea what is in version 3079, I think it was applied yesterday because my timed charge was turned off and I only noticed some 40 minutes into my charge window. But it also continued to charge past my set times, it was meant to finish by 0230 but continued to 100% which finished 0325. It looks like a better charge curve from 90%, but still has the odd 2% blip at the top end. But, checking my data yesterday it appears my soc display vs capacity matched.... Maybe this is it!
stevelewis My battery capacity dropped overnight when the new firmware was installed. I can guarantee you it isn't degradation as I'm all over my portal information every single day.
Though check out the voltage plot from 90%
Well I've always tried to be fair with my comments on the beta firmware, but what I've seen on my system today is the worst I have experienced.and more in line with 3007. Worst thing is that I've been running full to empty (standard cycles) quite regularly over the last few days, so no excuse.
Early this morning, my batteries went from an indicated 30% to 5% ( about 40/35% down to 16% on battery BMS), however there was only a 0.2kWh discharge. This to all intensive purposes is a soc crash of epic proportions.
The current draw was only around 100W at the time. I had paused the batteries at 17% to give me some head room to get to the start of the saver session. It should have been zero, but there is a separate inverter software bug -another story!
The voltage recovered when resting (true zero.discharge) to 52.15v, which is more associated with over 30% (BMS) SOC.
Suprise, suprise! Charging this morning slowed significantly and around 80% I've seen a large upward jump to 100 %. Effectively, this has rendered a good 15-20% of the battery unusable during this cycle.
Strong feedback already provided.
I know some people have seen the discharge restart when the battery is higher than indicated at 'empty', but I've never seen this happen on 3011. It always seems to push soc too low relative to voltage, and then recovers with a jump to 100% during charge. Just 6.3kWh added.
3011 with 2 X 5.2 kWh batteries.
THALL Though check out the voltage plot from 90%
That looks ok? Voltage drops on demand of the battery and rises when no demand is made? And the peak of the voltage when resting is dropping slightly over time. Is that not how it is supposed to happen?
I guess Giv CN are back to work ? 🙂
My system was in debug mode 9am CN to 9am CN so hopefully they are also looking at my PV issue, looking back through the debug logs, the PV issue is larger than I thought, lots more ramping and drop offs than local mode on the app shows.
THALL very different, wonder if it's an effort to get more into the cells after the voltage peak has been hit, to restore the 'missing' capacity
I have started an unofficial Giv discord server if anyone wants to join? https://discord.gg/jkCycRsG
Looks like I might have had an SoC jump during the calibration charge - surely adjustments aren't being made mid calibration? Maybe the calibration was just to get the 2 batteries to a similar starting level given one was a fresh install.
The other thing is that the behaviour seems more like before beta with the slower emergency charge to 4% rather than the rapid jump. Will see how today goes.
chris_p But the voltage is dropping while the battery is still charging. At 90% it is still charging at 2600w, the voltage should continue to climb with it shouldn't it?
THALL Mine has always done that (at least on the beta). The voltage always drops as the SOC is still going up at the end of a charge.
THALL Ah the first 90%, not the second 90%.
judgepd And still drawing power? I have seen my soc go up on its own but not drawing power. I saw this as the recalcution stage. But previously for me at least if the battery was charging, actually taking power the voltage would climb. If the soc was going up but not pulling power then voltage would drop and settle. Hmm maybe I am totally wrong or this is unique to me.
chris_p Yep, during the charge climb in the early hours.
THALL: In order for the battery to charge, the charging voltage needs to be higher than the terminal voltage of the battery. If they want to reduce the charging current/rate (as seen in your graph, after that 90% point), then the charging voltage needs to be reduced (closer to the terminal voltage), but the SOC would still rise.
I think the issue here is that our voltage graphs are representing two things:
1) When there is no current flowing in/out of the battery, we see the terminal voltage of the battery (i.e. no-load voltage). But to complicate things, this can change depending on how long it has had to 'settle'.
2) When the battery is being charged, we see the charging voltage which must, by definition, be higher than the terminal voltage.
(Then there would actually be a third, which would be the loaded voltage, which would get lower and lower as the amount of load increased - because the battery has a finite internal resistance).
I believe that it is this which makes tracking SOC so complicated/difficult, as any reading which is taken to estimate SOC, can probably only happen when the battery has been idle for a certain time (the 'settling' I mentioned above).
ToothyChris
Thanks, was just typing a much less confident version of this!
Totally agree about this being a huge complicating factor in calculating the SoC. That was my 'toilet thought' today after seeing these charts!
ToothyChris Understood, thank you. So everything is normal then.
In that case this new firmware that I know nothing about looks pretty good. Just hoping the timed charge going over by an hour was a one off. Will find out tonight I guess.
And just thinking (rambling?) some more about this....
If the internal resistance of the battery is fixed, then it should be possible to derive a relationship between the charging voltage and the SOC, but that would also assume a fixed charging current. But I suspect that the internal resistance is not fixed (and probably not linear either), so you end up with several variables which you would need to try and map. The temperature probably also comes into the equation as well.
I assume that it is this kind of matrix/map, which GE is trying to get to, using their (claimed) learning or iterative process. But based on the comments that beta testers have posted on here, it does not appear to be learning or focusing in on something which is close to that map/matrix. If they had, then only small adjustments would be needed to the reported SOC.
I'm not sure I have seen much evidence of iterative learning at all, but then I'm only seeing your second hand observations.
TheDragon (GivEnergy) Do you mean 3079? And no I didn't even know I had it applied! No documentation, no warning. But I assume it went on yesterday hense my settings changes.
THALL by bad yes 3097 was an earlier one.
WTF I'ls 3079.
Gonna have to do some digging
TheDragon (GivEnergy) Well i would like to know too. I would like to know why it was applied without giving me a heads up. Because it cleared all my settings which was not very helpful. However, the soc numbers look good. As in 10% soc drop should be 1.772kw and the data suggests 1.7kw so looking rather good! I had full capacity available yesterday, taking away the reserve etc
DO290-AO290 / 3011 Full calibration then 7 cycles (Full,empty,full) over 14 days. AC3 8.2 100%
I consider that my system is fully commissioned and functioning correctly, I cannot answer why others have issues but offer the following that may assist. Someone more knowledgable please feel free to corect the detail.
1 High voltage protection The charge rate modulates down from 90% SOC and the battery voltage falls as the charge finishes at 100%. 90% V = 55.76 100% V = 53.75
2 Low voltage protection A sequence exists to charge the battery from 2% SOC 51.5 V to 4% 52.5V. If too much power is taken and SOC falls to 0% then the 300 W charge takes ages to recover the battery back to 4% SOC.
3 SOC The software generates a value and occasionally adjusts it's value. Large jumps which were a regular feature have gone. The SOC value works for Timed Charge / Battery Options.
4 Inverter performance has improved ie less leakage.
5 Temperature protection has been improved, mine are in an attached garage so no details.
6 Capacity 8.2 kW 100% DOD 160Ah is performing like 80% DOD. Takes 7.1 kW charge and delivers 6.6 Kw
Newman It sounds like you are one of the lucky ones, for whom the new strategy seems to be working.
Did you see incremental improvements over those 14 days? I would expect larger corrections to start with, and then smaller corrections after that.
Did the 100% voltage settle at a lower value (i.e. below 53.75V), after a rest, or did your usage not allow that scenario?
I am encouraged since I also have an AC3 setup, but the apparent loss of capacity is slightly concerning.
(I was not accepted onto the beta testing)
3079 has just arrived in my inbox.
For once you guys ahead of the curve. I'll wait to load and see what its about
Calibration was the major event, each cycle saw an improvement - it may continue to after cycle 7 but it's miles better than 535 / 3007.
Voltage dropped to 93.5 V 97% SOC as first discharge occured.
Does anyone have a 8.2 kWh battery supplying 8.2 kWh ?
Newman Based on data from yesterday, I appear to have. My export vs what the soc said matched perfectly. Well if you ignore the reserve. I will check the numbers in a bit and update this post shortly.
Ok 1% of my batteries should be 0.1772kw, I discharged 86% which should be 14.88kw. Data says I discharged 14.55kw once I remove a bit of charge from solar. Pretty damn close. Not factoring in any degradation and temperature. Please someone tell me if my math is wrong is just stupid!
Newman Pre beta, yes. Last night the fact I couldn't charge to full in my 4h Go window would suggest I'm closer to 8.2 than ever before with the beta.
My 3088 firmware seems to be happy putting in7.3kWh a night into my 8.2 100% battery
Certainly not me! Mine has just taken only 4.3kWh to go from 6% to 100% displayed SOC (including a sudden jump from 79% to 100%).
I am getting either a sudden jump to 100% or a very large (20-30%) lower half SOC% downward adjustment nearly every day now. The system is flip flopping between the displayed SOC% reading significantly too high or too low.
Still on 3088 here, which to be honest is working well other than the capacity still being low at around 6-6.5kWh.
There is an old navigators adage which goes ...
"Shipwrecks are not caused by not knowing where you are. Shipwrecks are caused by believing you are somewhere other than you actually are".
Woodstock Looks like the battery was only charging for an hour an a half at 2.4kWh
judgepd Ditto!!
8.2kWh drawn in at 95% inverter efficiency (from AC) would actually put 7.8kWh in the battery. When drawn out to AC again at 95% inverter efficiency that 7.8kWh stored would become 7.4kWh all other things being equal.
Not sure what the claimed efficiency of the GE inverter is but that would explain some of the divergence of power. Temperature drops/cell balancing and other draws would take some more.
If the readings are power drawn from AC and power put back from AC view then 8.2kWh in would reasonably only give 7.4kWh out.
Just wondering if people seeing the drops are factoring this in?
For an 8.2 kWh battery I would expect 8.7 kWh in and 7.6 kWh out. 7.1 kWh in at the moment is worrying.
kram I'm monitoring the DC in and DC out, so comparing after/before the losses and the capacity reduction for me is definitely real unfortunately. I see it in the data and also the general feel for how quickly it drains compared to 3007.
Fingers crossed 3079 has widened the voltage limits.
mpartington Fingers crossed 3079 has widened the voltage limits.
Have you done a full charge on this yet? I might load it tonight and set off a full charge and see how much it takes, vs 3011
TheDragon (GivEnergy) not yet, still at 60%, doubt it will run out today now.
chris_p Yes, portal data shows total battery input as 4.3kWh. A significant drop in capacity.
TheDragon (GivEnergy) You got your hands on the 3079? Did you get any notes with it?
I have seen an soc jump now, first for a long time since it seemed the 3011 beta had stabilised.
SOC jump from 87 to 100 at 14.10, on 30th jan 23.
charging from solar battery voltage jumped from 54.9 to 56.41. 10 mins later volts was 53.87.
at 15.45 after a steady curve voltage drop it was 52.98
THALL Did you get any notes with it?
No, but I did email back for the changes between 3079 & 3011.
The only change is the charge rate a5 56v is capped at 0.1C, which is to reduce the jumping at the top end for users with high voltages.
Mine tops out at 54v, other see 55 or 56v. So the update didn't do much for me.
A slight increase in charge energy loaded up at he top end, 12.1kwh now vs 11.7kwh
The next release is working on the capacity issues, so unless you see the 56v and stop, wait out
TheDragon (GivEnergy) Interesting. Highest voltage I have seen since I noticed the firmware applied is 55.35v compared to 57+ on 3088. Though last night I capped my charge to 85% because I foolishly believed the weather forecast.....
Do you have direct contact with the Chinese on this?
THALL
No i do not
TheDragon (GivEnergy) My battery bottom voltage is 50.4V and the upper is 54.5V. I'm missing out on firmware 3088 I guess but emails just seem to go into a black hole.
Odd behaviour during the sunny spell.
We've had a spell of good sunshine over the last few days, and each time it has risen, my system has stuck at about 25watts of generation. I have to restart the inverter to get it to read properly - typically 1kW.
ed-form I think a few people have seen this. Can anyone explain what's happening in this situation, from an electrical point of view?
TheDragon (GivEnergy) Did a full charge last night. hit 57.43v on this firmware. But charge did continue until the new standard 98 - 100% jump. Lowest voltage so far 51.8v
Reporting a bug in the firmware. Unable to post an image at the moment as imgbb is down?
But my ac3 reports 200w generation from PPV2 but my battery reports it is still charging at 2900w. The sun is out solar is generating, just my AC3 says it isn't.
The system recovered itself after 15 ish minutes.
Has been logged with GE beta email thread. I will link a pic when imgbb comes back on.
BMS FW 3078 being sent out. Not loaded yet, as its not really something bothering me
To deal with the 80 to 100 jumps people see. Not me, but other do
THALL Lowest voltage so far 51.8v
Yep, I too see a slightly lower finish voltage. Maybe testing the waters for a wider limit on a stable release.
Noting the 307x are test releases
TheDragon (GivEnergy) Any idea when this is happening? Because the last update I had I was not informed of and it turned off my timed charge.
Is it just 80 - 100 jumps? No fix for this odd 98 - 100 - 98 jumps?
No idea. The people I know with it, me included (Not loaded yet) were sent via email
TheDragon (GivEnergy) Ah ha, not me then. I do not get any emails back..... My last firmware was applied unexpectedly.
So the new app is interesting. It shows both of my batteries now and appears that only my older 8.2kw battery is soc jumping.... The 9.5 has a smooth charge curve. The 8.2kw hits the 56v sooner. This will be very useful going forward with the beta
THALL Just seen this.
My master is showing as at 98%, 52.8V and 20.2C.
My slave is showing as at 99%, 52.9V ans 18.5C.
I assume the difference in temp is because my master is directly beneath the inverter with the slave under the master.
So the slave is shielded from the heat my inverter is putting out
Other than that mine seem ok.
Stevef1964 I also see a higher master temp, but not related to inverter as inverter is above the slave and quite some distance away.
Have a couple of GE staff on the Discord now 🙂.
Loaded 3078, hoping to see my SoC be somewhat correct as the master is either 1% or 100%, done it with 60% charge and both are reporting 7%.
ed-form Potentially same as the issue I have had with the inverter struggling to keep the trackers alive. Solar has been 50w~ but mostly fixed in latest beta (yesterday). Still some way to go but produced much much more today.
I made a channel on disc and threads can be started under beta-channel to track individual problems, might be easier than thread dumping - that is why I made it 🙂
THALL Have you checked your Spam folder...?
One reason that separate data, for those who want it, is useful. Only one of your batteries is misbehaving. GivTCP would have shown you, but you dont have that. This was the next best idea.
THALL You really should install HA inside a VirtualBox, they have a ready made .vdk drive to download. It takes about 15 mins to get it up and running with givTCP, so much more useful data for you when doing all this testing and possibly beyond.
chris_p deffo beyond, next thing you will have dimmers and other crap you dont need activating as you walk around the house setting off various sensors
3078 loaded up. Let's see what it brings tonight, when Agile automations is allowed to charge my battery as it sees fit.
From the summary of changes, I'm fairly sure 3078 is to fix this behaviour.
At the bottom limit OCV was 'dropping too quickly' i.e. ended up setting a much lower soc than the resting voltage. On the next charge I saw the jumps from 80% to correct.
I wouldn't be surprised if it fixes jumps of other magnitudes as well.
Really pleased to see this fix, was the second biggest issue for me behind capacity. Hopefully this is an enabler for improvements on that. Just need some heavy use /poor solar days to test it.
TheDragon (GivEnergy) I do fire up givtcp every once in a while to check the data, but I do not have it running 24/7. I did also notice on the new app that my older 8.2 has spikes dropping to 0% soc, but I have no idea if that is a app bug or a fault with the firmware, fault with the battery. I have linked the image on the discord server as imgbb is still down and I just can't be bothered to share from another source right now.
chris_p I could, but do not want to go down a do it yourself route if possible. I inform GE support of what I saw using the default tools available. I have no idea if something I reported was the trigger for me getting 3079. But I will continue to feed back, I have no problem with my kit being used for testing. However if GE told me to please install givtcp and run it 24/7 for testing I would do that. But had no such request :-)
THALL to be fair the problem with that would be them then having to give support to something they don't really want to support, could be a can of worms. This new UI is probably the answer to that 'can of worms'.
THALL that's just a data glitch, nothing to worry about. GivTCP has some data smoothing inbuilt to filter out these drops to zero, but as it's on the data being sent by the inverter, no surprise they are also appearing on the giv charts. Safe to just ignore them
mpartington Cheers
New firmware just hit my batteries
3077
Tad scary the software changes on MY batteries without any pre warning, not even a notification in the log. All in the name of progress 😀
And no, I have no idea what this one is meant to do. Hopefully @TheDragon (GivEnergy) will hear something
THALL To be fair, that is an ideal update scenario, it is remotely updated without any loss of service, but yeah a notification would be cool to add - especially with their new push notifications!
joshgreenuk 3079 turned off my timed charge. This push update appears to have left it on so maybe that is what 3077 fixes? Just guessing
I'll ask what 3077 is. I don't have it.
Itll will have been installed direct by China.
TheDragon (GivEnergy) I suggested maybe someone from Giv could just post it in disc once with the file + readme then it saves having to email it out and us asking more questions / requesting it...
It would, but who then controls its release to public?
TheDragon (GivEnergy) People could have roles that limit them to the channels to stop public from accessing. Is anyone else on 3077 or just THALL?
joshgreenuk I am special, key customer.
Not even GivUK knew of 3077. So they now gone digging
TheDragon (GivEnergy) That's a new level of confusion, even for GE!
ToothyChris Now they know what it's like for the rest of us!
Blind leading the blind? 😃
It was to check if one issue was resolved. So who ever had it, was likely a one off and China saw a bug, testing fix.
If others start getting it, it may well be the issue checked, is fixed, what ever it was.
THALL I noticed also these spikes to zero and back up to the level it was at on the latest version of the app, but as I see no cost on my smart meter I also put it down to the system data needing to be refined so nothing to worry about.
Note: my system is not part of the trial
TheDragon (GivEnergy) Really? Wow. Hmm so now I am wondering what makes my system so interesting......
Stevef1964 it looks like my spike to 0 is happening only on my 8.2 battery and is happening daily but at different times during the day so I can not see a pattern yet
The odd spikes in import after reaching a sub 100% overnight charge seem to be a feature of 3011 as it occurs every night.
Is this stopped in subsequent betas?
THALL I'm also seeing these drops to zero for my temp readings for both my batteries.
So I would guess these drops are actually either the data from the inverter being interupted or the GE servers having blips (caused by the changes GE are making?).
positor1 Hmm I had a couple of spikes, 2 a night. On the 3079 firmware.
Stevef1964 Not sure. My newer 9.5 battery is fine, no drops at all. Looking at the new screens only my 8.2 has any kind of soc jump the 9.5 appears to have always been a nice curve. I really don't know, I expect the data means something to the development team looking at it.
I'm also seeing the charge blips to zero and back on both packs. I'm not on the beta. At least I think I'm not, I was told I was but my firmware release numbers didn't change. They didn't reply to any emails thereafter.
My packs are now running around 24% different in charge level and I've seen the odd jump to 100% on the overall figure. Infrequent and happens in the low 90s. Does this reset anything at all on the old firmware or will they just plod along out of sync until the next event, progressively getting worse?
sponev Then the drop to 0 could be purely a data bug in the app. Hopefully it is something that simple.
THALL It "feels" like a comms/data bug but it's all guess work in the weird and "wonderful" world of GivEnergy. My next test is to force my overall charge a lot lower (I don't tend to go much under 30%) to see when it all crashes to zero and what happens after that. Or maybe I just plod on with them out of sync but at least working, hmmm?
Any data loss previously just kept the previous data point and joined it to the next reported data point, making some very straight lined graphs at this point. Personally I'd prefer no joining of those points and it be empty as it gives a false impression of what was going on.
chris_p Agree. Showing the straight lines makes it difficult for many users to understand that data has been lost. Then they start questioning why their total import value does not match what is on their IHD (or other daily totals not aligning....).
But the counter argument, is that it then looks like a crash to zero on some graphs, which would probably freak people out even more.
The reality is that the GE dongles are known to be unreliable, for many users, when it comes to staying connected to the GE mothership (be that a WiFi issue, or a server issue or something else). So some kind of visual indicator of 'lost data' would certainly be beneficial, beyond the two options above. How about keeping the straight lines, but changing the colour of the area between the line and the axis?
TheDragon (GivEnergy) looks like they pushed 3077 to me at 13.37. Bit late for China unless working very late, so maybe now going wider beta release via UK team?
mpartington Mine was 12:17 yesterday
Mine has also been changed to 3077. This happened some time today. There was an otherwise unexplained spike of grid pull at 13:36 so I'm guessing it was around then (not sure if there is another way of checking exactly when it happened?).
Mine has never been changed so maybe I'm the baseline for you tresters!!! 😆
Update on 3077 in behalf of Paul 🙂
It's a release to capture and send additional debug data to allow China to dig deeper into the capacity issues that we are all reporting.
Also in separate inverter update, they've fixed the PV issues some were seeing
mpartington China on your Wi-Fi network - no concerns there then 😄. But then I guess they could probably do it via my CCTV if they wanted to. Anyone seen any weather balloons overhead? Beer o'clock....
JaseSpace not without GivTCP.
Mine was 13:44, 13:46, 13:42. Done by UK. So the debug FW must be sent out wider.
No known changes, other than its spewing out binary data in HEX for debug, in relation to capacity issues
sponev chill and grab a beer.
mpartington I am seeing the capacity loss issue and I am not part of the beta testers (although I did offer my setup) My 4 month old 9.5KWh is seeing between 11% and 18% capacity loss every cycle.
I hope they are aware the capacity loss issue is not just a result of the beta FW trials.
Apparently I was updated from 3088 to 3077 at 3.50pm this afternoon.
Only noticed by looking at the energy meter, saw I was pulling from the grid :/
The update had put my battery to sleep so have now woken it up.....
CCourtney maybe there are some other underlying issues. Installing the beta firmware was like flicking a switch though. I went from putting over 9kWh in the batteries on 3007, to struggling to hit 8kWh. 15 to 20% down some days. I know @TheDragon (GivEnergy) has done back to back testing and the difference is clear. Once they've cracked that nut though, I'm sure the new firmware will massively benefit everyone.
Moved to 3077 at 16.00 today. 7.4 kWh from battery today, hope this software change is capacity related.
also now on 3077. no idea when.
I am praying for you guys it’s good enough to fix all the issues and be a formal public release .
Puzzling behaviour around lower battery limit.
I set my system to stop discharging the battery at 15% and it works - when the battery falls to 15% while discharging into the house in the evening, it stops the discharge and declares it is idle. In the morning, however, the charge reading is typically 12% - this morning 11% - and the battery begins charging from the mains until it gets back to 15%. I do not have any setting to do this, but it happens every day.
kram It looks good so far. I will collect the data over the weekened. Still getting the odd 98 - 100 - 98% jumps. As in it holds 100% for 5 - 10 minutes. But charge going in vs out looks good.
I was seeing a much reduced capacity (8.5kWh from 12.3kWh) until my system triggered a low voltage emergency charge (couple hours at 290W). Since then it has been much closer to expected and I've gone from 100-55% and 5.2kWh discharge.
With the app displaying each battery separately, I've watched a sizeable loss of capacity unfold today... a possible reason the capacity fluctuates between cycles and is generally reduced with the new firmware.
AC3 system with 2x 5.2kWh batteries. Overnight charging to 100%, 2 hours hold, followed by discharge in 3-4 hours. Repeat daily.
Both batteries started at 100%. At mid-discharge, the battery SoCs diverged by 25% in 30 minutes. Discharge ended when one battery hit the 4% limit; the other was still at 30%, tallying with input & output energy readings.
Charging to 100%, the two batteries have re-aligned during the trickle period.
I've been considering a move to Cosy Octopus with its 2x 3 hour charging windows... just enough to charge each time and bad news with FW 3011 as there won't be a trickle period to resolve any discrepancy.
Has anyone else seen similar?
stevelewis yes, exactly same set up and have seen this a few times. Yesterday it dropped from overall soc 40 % to 4% discharging just 1.3kWh. it then charged up with only 7.6kWh. I've seen this quite a lot on the different beta versions and keep reporting it.
One battery was at 12% and the other 30% when it stopped discharging. They did converge at idle
Both batteries voltage resting were 51.8v, so closer to 20% (but still a small loss)
So for me there are 2 issues, first being a general loss of capacity (I'd have put in nearer 9kWh on 3007). The second is the quick drop from 40%, seems the SoC profile Vs voltage map was wrong.
Discharge was fairly steady from 18.00
mpartington interesting to see that the battery with the higher SoC switched over mid-charge in your case. How much of your lost capacity would be accounted for by the incomplete charge/discharge of the lagging battery in your case? From a squint at the chart and assuming your batteries are both the same size, I'm guessing 4% extra charge retained at the bottom and 5% missed at the top.
stevelewis if I go off resting voltages , then I was 51.8v at 'empty' and 53v 'full'. The resting voltages of the 2 batteries were more or less identical, but different displayed soc. So I think it's probably a tracking issue, rather than necessarily pulling capacity down. Still a big bug though!
This chart below has been shared by @TheDragon (GivEnergy) before, and I'm sure he won't mind me sharing it. The theory is that the upper and lower voltage limits need to be widened as we're operating within 70%, rather than 80% of the capacity.
3077 has been issued to help china do some debugging on why were seeing this.lower capacity.
Tracking better tonight, although only 5.3kWh DC discharge and I'm down to 25% (from 2x5.2 batteries 😞)
Matt, more than happy you sharing my diagnosis of the capacity issue.
Hopefully with the data they are getting from those with 3077, they can address the issue and get the best of both worlds.
Stable SoC readings, for which mine are rock solid, and capacity nearer the actual capable.
Even on Agile, charging at random times, I'm trying my best to break the sync, it's not happening. The worst I've managed is 10% variation over 3 packs. Which is soon corrected back to within 4% at either idle or full.
TheDragon (GivEnergy) @mpartington Thanks for the additional thoughts.
The tracking issue reduces the usable capacity by preventing discharge in one battery once the other hits the bottom limit, even though it may still be 30% full. There's a similar issue at the top end too.
As my battery is being limited by the 4% SoC, rather than voltage, I don't think that widening the voltage limits would help, but you guys may have given this more thought than me!
@mpartington My resting voltages were similar to yours.
@TheDragon My SoC values were pretty stable until today and then went south fast... I've been checking daily charge/discharge and 8.1 kWh for 5.1 kWh out was exceptional (a 25% variation between two packs).
stevelewis interesting results from last night. Similar soc dip on 1 battery (albeit at a lower soc), but this time it's the other battery. No rest period this time for the OCV to kick in - that will be typical for my usage profile. Maybe I'll have to start automating an hour's pause at appropriate conditions as a maintenance action, but I don't really want to do anything during the trial that is not open to a regular user.
Discharge was much more uniform across the whole range (5.9kWh DC total).
During charge, a mini calibration is performed at top end and there is a soc jump of around 10%. This doesn't seem to have made a significant difference to the energy added (7.1 kWh from 12% to 100 %) - still low.
(3077 software, 2x 5.2 batteries, AC coupled)
mpartington interesting. Mine are back within 5% this morning, currently discharged to 37% overall with yesterday's troublesome battery showing the higher SoC too.
This is getting boring now!
Another lower half adjustment of displayed SOC (downwards from 41% to 11%) yesterday afternoon.
Followed by another sudden jump of displayed SOC (upwards from 73% to 100%) less than 24hrs later!
Flip flop, flip flop, flip flop.
Battery (100% DOD 8.2kWh) took only 4.6kWh to charge from a displayed 4% SOC to a displayed 100% SOC.
Woodstock are you on 3077 firmware?
If so then additional data is being sent and this would have been noticed by the firmware team
TheDragon (GivEnergy) No, still on 3011.
Woodstock 3011 here too. Its not so much that the reported SoC gets out of sync that concerns me (although it's not the greatest of looks), it's the fact that there's no realignment at the bottom end to empty the batteries properly.
Woodstock bugger. Would be useful for your data to get the detailed firmware debug treatment.
My inverter is spewing data out constant since.
stevelewis I have had the displayed SOC% doing its hour long adjustment downwards at the same time as the
battery is charging and the battery voltage is increasing! Looks suspiciously like the value obtained from an instantaneous table look up is being applied bit by bit over a 60 minute period to avoid showing a "sudden jump" on the graph. Of course by the end of the hour the looked up value is no longer correct as the actual SOC% has moved.
Hi all
I've been on the Beta for about 23 days now and have been tracking key data. Certainly my system (5kW Type 1 Inverter, 9.5 kWh battery) had been pretty well behaved before the Beta, though I was getting SOC adjustments like all. I've been tracking the 'notional charge' that the battery would take every day and am displaying the data below in an I-MR control chart - a way to track if any statistically significant changes have taken place with the system:
The chart shows the first few days of various adjustments settling down, then very consistent operation, then a ramp up in capacity when the last update to 3077 happened in the last couple of days. I'll obviously share this with Givenergy but from my point of view, looks potentially promising but too soon to make any firm conclusions.
How the heck do you add an inline image? I uploaded it to imgbb but don't know how to link to it to make it display?
Andy C In imgbb choose "BBCode full linked" from the drop down link menu.
Andy C
So you are seeing an increase in capacity from 3011 to 3077
TheDragon (GivEnergy) Would he see that change? I thought 3077 is purely a data collection firmware rather than any kind of fix?
THALL Thats why I asked, as on the graph he has written changed to 3077 and the capacity went up, so not sure which was loaded before 3077 on this timeline
Andy C also interested in what you mean by 'calculated full battery charge'. Are you recording full cycles or for example saying if you put 4kWh in at 50%, the full charge is 8kWh?
My first charge after 3077 appeared to show a capacity increase - 5.7kWh for a 58% soc increase. I thought fantastic, until the next cycle, where I only discharged 1.3kW from 40% to empty. So the update had messed up the SoC relationship (it seems better now after a few complete cycles).
Since then, I'm only seeing around 7.5kWh (DC ) going in on a full charge 2x5.2), so no improvement for me unfortunately
mpartington
Same here, which is why I was a little confused when the graph above appears to show a decent improvement
Guys, be careful when trying to estimate capacity from a partial charge or discharge. Firstly, we all know the indicated SoC values are dubious, at best, but there is a good chance that the process/relationship is not linear either, so saying you 'put X kW.h in, with a rise of Y % SoC, implying a capacity of Z kW.h' - may be flawed logic.
But I fully appreciate it may be the only indicator we have.
TheDragon (GivEnergy) Yes, and it was confirmed again today, seems a significant shift.
I appreciate that the SOC/voltage relationship won't be linear, but if the system is working correctly then the kWh/percent of SOC relationship should be, otherwise looking at the SOC is a bit pointless. If some 'percents' are bigger/smaller than others than it's useless to even show it.
mpartington not all of my cycles go from 4% to 100% (well, 98%) but most have a high range this time of year. The gradient on any charge, against SOC, is rock solid. So it is reasonable to use the calculation amount added/(end% - start%) I think to calculate the notional capacity. The control chart tells me my assumption is reasonable as it is staying in control with no SOC adjustments being applied.
Andy C i.e. in control from days 8 to 21 before the software upgrade.
TheDragon (GivEnergy) 3011 before this date.
Change to 3077 03/02.
Running significantly lower voltages at low SOC, enabling higher charge into battery (c. 8 Kwh today).
Down side to this is that first low voltage protection charge at 291 W was from 2% SOC 48.8V to 4% SOC 51.89 V.
Second low voltage charge was from 2% SOC 50.78 V to 4% SOC at 52.34 V. SOC adjustment took place to 11% and voltage dropped to 48.18 V. Worse position to when it started.
I have increased reserve fro 4% to 10%
Newman mine is down to 4%, recovered a bit to 7% and remains idle. BMS reports 17,17,18%
51.7v is my low on 3 packs. Only time I see 48s is during calibration.
Inverter reactivated at 8%, unfortunately kettle switched on at 11% and voltage dropped.
on 3077 I am noticing the last two days that solar takes the batteries to 100% and shortly after they start discharging despite enough solar for the house and EDDi (hot water diverter)
currently solar is at 2kw
300w exporting
battery discharging 290w and at 97% 52.51v
eddi taking 1.5 (approximate figures as changing as i type) this has been going on for over an hour now.
boiling the kettle still resulted in a 300w export
reset to defaults stopped this and now batteries charging though at a low rate 132w.
i had to reset to defaults yesterday too to stop this. the only setting i have apart from eco is a timed charge to 70% overnight on eco 7.
positor1 Not seen this behavior on my AC3
positor1 I have not seen this on my AC3 either. Solar charges to 100% then pins it there.
Daily need to restart inverter.
I continue to see the same error first thing in the morning - with the sun well up and really bright, I see readings of 10 or 20 watts, climbing slowly to perhaps 100watts. If I restart the inverter, a few minutes later the true reading of 1.8kW comes up.
I engineered a lower half SOC correction (from 43% to 11%) this morning between 05:55 and 06:55.
No prizes for guessing what is going to happen when the displayed SOC gets to around 73%!
I have emailed GE to see if they want to turn data collection on.
Woodstock Are you on the current 3077 beta firmware? I have not had time to check data but going by the graphs alone my system seems pretty much spot on. Nothing unexpected. I know the 3077 is for data collection, but everything seems better.
Feels odd saying that considering I have other electrical issues, like Octopus failing to charge my car last night...... Intelligent Octopus Beta issues.
Since I had a calibration I haven't seen any sign of SOC adjustments, it's just been working fine (aside from capacity being down about 20%).
THALL No, I am still on 3011. Thought I may get 3077 after raising the same issue on Sunday but no-one seems interested.
judgepd I can produce downward and upward adjustments to order! 😝
Almost there ...
any minute now ...
Problem with this forum, know one knows who you are, so GivEnergy are reading this thread, i do know that for sure. But they cant assist a handle name.
Who in GivEnergy are you working with? as 3077 is 4 revisions on from 3011
TheDragon (GivEnergy) It doesn't matter if this is being read or not, if you want assistance on the beta there is only 1 uk support person to speak to and he's a busy chap. I have no doubt all of this is being read by GE and all the customers and potential customers we are sending to this forum too 😁
Woodstock give the support line a call, ask to speak to who is running the beta and I expect you he will slap the newer firmware on for you. Unless there is something about your system they are still collecting data for. As in the chinese not GE UK. In which case not much you can do but wait.
Just a reminder my system got slapped with 3077 last week and I had no notification. I am not complaining btw. At the moment my system is working very well. Only visual issue is my system doesn't seem to hold 100% for longer than a few minutes., 98% instead.
THALL Only visual issue is my system doesn't seem to hold 100% for longer than a few minutes., 98% instead.
ditto. and 52.8v max then starts discharging. I have pressed pause on the app to see if i can force the inverter idle to see if it then recalibrates.
TheDragon (GivEnergy) Forgive me but I thought it was GivEnergy that wanted my assistance. That is why I signed up for the beta test.
My handle on this forum is the same as my portal name (both are unique) and it is included at the end of each email that I send to GE. My full system details (including inverter serial number) are included at the start of each email that I send. I couldn't make it much clearer. If GE can't work out who Woodstock is, then God help us all!
As for who in GivEnergy I am working with; that would appear to be no-one. GE requested that all beta communication should be by reply to the installation notification email. I keep sending emails but since 3011 was installed on 13/1/23 I have received no return communication whatsoever. It is only by posting to this thread (and hoping that someone in GE might be following it) that I am getting any sense of contributing to the testing process at all.
positor1 I am not sure how idle it has to be to adjust, because the only time mine has been idle during this beta trial is when it finishes its timed charge and then waits for the window to end. But I have seen adjusts in soc via the graph data.
interestingly while paused the batteries still showed a variable discharge of between 2 and 30 watts.]
now unpaused , still showing 98% discharging 13 to 33 watts volts at 52.81 - 2kw solar available.
version 3011 so far not the best for me.
THALL
thank you Thall,
I do not think i have seen mine go idle once since 3011 was installed. they get to 100% for 5 minutes and then start discharging even if there are hours left till the end of the set charge period.
Woodstock I have received no return communication whatsoever
I am not aware that anyone does?
positor1 Exactly, so how am I supposed to know who I am "working with"?
incase this helps anyone. tried exporting to drop the batteries down to 95% and see if they would then start charging buit that didn't work.
i have restated the inverter and that has worked. now charging at 680 watts. odd
positor1 I can confirm if the charge is over 95% they will not charge, feature of the new firmware. Drop it to 94% and it will start charging.
Apologies for the suspense; the imgbb server was down for a while earlier.
The expected jump to 100% SOC did indeed materialise.
Points to note:
This is a clear illustration of the processing that takes place as the battery approaches full and the displayed SOC% is significantly lower than the true SOC%. (Thall's recent posts are showing the other situation when the displayed SOC% is higher than or equal to the true SOC%).
When the displayed SOC% is significantly lower than the true SOC%, there is repeated processing to ensure that the battery really is full and the displayed SOC is then reset to 100%. Fair do's, this does realign things to where they should be and heads off the over voltage alarm that used to occur.
The fly in the ointment is the lower half displayed SOC% adjustment that took place only a few hours previously. This appears to be doing a table look up and getting it spectacularly wrong, to the tune of around 30%. (I believe the table look up is instantaneous but the value returned is applied to the SOC curve over the next hour). This resulted in the displayed SOC% under reading by about 30%, with the inevitable sudden jump to 100% as the battery charged.
Both the lower half adjustment failure and the sudden jump to 100% SOC are entirely reproduceable at will on my system, as evidenced by my manipulation to produce todays events. (I wanted to get both on the same day, in the same cycle, to eliminate any other cause of the SOC misalignment).
If anyone from GE would care to contact me I will be happy to repeat the process at a time of their choosing with data collection turned on. I did advise in advance by email what I was going to do today but, as always, no reply!
Update on progress, on my 3 x 5.2kwh and AC3. Been given permission to put this out after some initial testing, among 3 testers.
FW 4077 was released yesterday, to address the capacity issues, with revision on other SoC bugs others have noticed. No readme available I'm afraid.
Capacity is up 11%, upper and lower voltages look to be wider
Below is an exact comparison of 2 days, on the left is BMS 3077, on the right 4077, i drew a line at the full kwh of 3077, so you can see 4077 is higher.
Also AC and DC charge kwh, so you can see the loss in AC to DC conversion hasn't changed between the version
Beta testers can expect this to be issued out in the coming day or do, subject to Dan's schedule.
TheDragon (GivEnergy) Great news - never heard a thing since being asked to put 3088 on manually (and still on that version today).
Lowered capacity was my only real concern at this stage.
TheDragon (GivEnergy) When the fw was applied did any charge / discharge times get turned off? Anything we should check after the fw is applied? I am away soon for an unwanted mile stone birthday so wont be around for a while to check on it.
THALL
Nothing changed that was obvious. But I also have my charging scheduled via APIs
Other than a blip in SoC reported level as the firmware was applied (Locally), the battery was at Max discharge power at the time, while cooking dinner
TheDragon (GivEnergy) Hi, is the expectation around 11% across the board? Or just what you observed in your circumstances.
Curious on whether this is 'the fix' for capacity issues, or part of many tweaks to get people back inline with spec.
I'm currently seeing around 30% loss since moving to the beta. 9.0kwh typical discharge is down to 6.5 for my 9kwh battery.
I appreciate we are all a bit different, does an 11% improvement put you back where expect to be?
Jamesrich 11% is what I personally saw, I dont have eyes on the design.
I do know they have done something different in calculating capacity, I guess before it was under estimating what can go in. Seeing 11% increase with only 0.01v/cell increase top end. So still well within the LFP spec
TheDragon (GivEnergy) good news cheers. Hope it's a good release! 😀
judgepd Not sure if it counts but just had the SOC drop from 100% to 97% without discharging any power, but also it isn't now charging back up to 100% despite plenty of solar. I think I saw this as a new 'feature' but I'm not sure why it's been implemented to be honest (does it have to drop back below 95% to start charging again?).
TheDragon (GivEnergy) Look forward to receiving the update. On 3011 I am around 21% down on battery capacity so it will be a good test. 🙂
Woodstock Please someone, get Woodstock the latest version! Prime candidate to see if these issues are truly on the road to being fixed!!
Woodstock
Yes, I was around 10-15% down, so looks to be bang on the money, without going near any of the limits where voltage drops and rises sharply.
Likely due to the tapered charge profile, stops the upper voltage spiking and triggering the BMS to cease
judgepd Yes, has to drop to sub 95% before it will charge again.
Tonight I will turn off Octopus automations, set a solid block of 6hr 0000-0600. So it will for sure fill, then test if the 100 to 98% oddity is still there
TheDragon (GivEnergy) Ah thank you, must helpful.
Just checked mine out of interest, still on 3077
TheDragon (GivEnergy) sounds promising. If they need another tester for 4077, happy to oblige as my system continues to be well behaved and stable on 3077, with an average capacity of 8.8kWhr, up from 8.4kWhr on 3011. Will also email them on the beta email.
Interesting finding from today's test.
Today's test was a repeat of yesterday's (lower half adjustment initiated at 43%, provoking a sudden jump from 74% to100%) but with a twist.
Instead of letting the displayed SOC% make its way down to the looked up value unmolested; I decided to see what would happen if I kicked off a fast charge while the adjustment was taking place.
The result was that the displayed SOC% hesitated for a bit and then started to climb. This shows that it is not just a simple case of arriving at the looked up value at the end of an hour but that charge/discharge during the adjustment period is allowed for.
What the processing is saying is ...
Looked up SOC value is x.
Currently displayed SOC value is y.
"In an hour I want to be displaying what I would otherwise have been displaying (taking into account any intervening charge/discharge) but adjusted by the difference between x and y. In getting there I will try and make the SOC% curve look as credible as possible". (Immediately following the flat line you can see the displayed SOC% slope is much shallower than would be expected for the prevailing rate of charge).
The good news is this removes concerns over the actual SOC% moving during the adjustment period.
Now all we need (to avoid this particular manifestation of the sudden displayed SOC% jump) is for the looked up value x to be correct ....
Woodstock I think 3078 addresses this behaviour. It apparently has improvements to OCV and soc reporting. Worth a try if you can get it pushed.
OK, as promised, last night i turned off all my Agile automations, reverting to a Go type charge profile (Un interupted)
Below is the charge 4hr 45min to complete.
In Orange is the User reported SoC (4 to 100%), then the 3 battery packs actual values (14,17,18 - 93,96,96)
Note no more 100 to 98 drop.
Then the Pack Voltages
Resting V 51.8(4%) - 53.2(100%) peak seen before charge slowed 54.4v
TheDragon (GivEnergy) Looks solid.
Ah so we might be close to a solution now. Hopefully the rest of the testers get this firmware and find nothing. Would be great to get this resolved
THALL sounds good, but one swallow doesn't make a summer... it took over 2 weeks for my system to start throwing wobbles with 3011.
stevelewis Fair point. But the later firmware have been better and it does appear each firmware addresses each bug. It just looks promising.
TheDragon (GivEnergy) looks good, how much did you squeeze in?
Indeed, I hope this goes out soon, I am interested to see how other battery makeups deal with it.
Just because 3.x 5.2s are near perfect, with some minor wiggle room still, performs stable, others see much higher voltages than I do, so they need testing to be sure all is well.
mpartington 12.1kwh DC. so 77% DoD. Hence there is still some room for more. But its a good improvement on 11.3
TheDragon (GivEnergy) that's encouraging, let's hope they are up for a small widening of the voltage limits. The resting voltages don't seem much different on your charts. That's still between 90% and 20% (ish) with all the usual caveats on temperature etc
mpartington Looks like 4077 is likely to be the next release installed now so let's hope it incorporates the relevant bit from 3078. I've said it before and I'll say it again, I do hope GE have a good grip on their change control procedures with the multiple releases floating around!
The thing that is bugging me is the 3011 lower half displayed SOC% adjustment uses a trigger voltage of 51.5ishV which is lower than the battery resting voltage (51.8V). This suggests that if a table look up is not initiated on the way down it never will be. The battery has to be discharging at less than 4A (200ishW) for a table look up to take place, so unless there is period of low discharge between displayed 50% SOC and displayed 4%SOC then you won't even get an adjustment on the way down. Strikes me a lot of people could be in the position of very rarely, if ever, getting a bottom end alignment (never mind those, like me, whose SOC rarely drops below 50% anyway).
Anyone else think they may be getting obsessed with this testing lark? Yesterday I caught myself thinking... "No I can't have a jacket potato for lunch because the current draw from the microwave will upset the test I am running".
Had to have cheese on toast to use the gas grill instead! 🤣
Woodstock Obsessed? Us solar/battery nerds? Never happen....
Ran out of battery (supposedly) 1.5 hr ago - showing 5%
Battery voltage is now reading 51.95 - isn’t that more like 30% ?
This is what GivTCP is currently showing
`Battery_Cell_1_Voltage: 3.272,
Battery_Cell_2_Voltage: 3.272,
Battery_Cell_3_Voltage: 3.272,
Battery_Cell_4_Voltage: 3.272,
Battery_Cell_5_Voltage: 3.272,
Battery_Cell_6_Voltage: 3.272,
Battery_Cell_7_Voltage: 3.27,
Battery_Cell_8_Voltage: 3.272,
Battery_Cell_9_Voltage: 3.272,
Battery_Cell_10_Voltage: 3.274,
Battery_Cell_11_Voltage: 3.272,
Battery_Cell_12_Voltage: 3.27,
Battery_Cell_13_Voltage: 3.272,
Battery_Cell_14_Voltage: 3.272,
Battery_Cell_15_Voltage: 3.274,
Battery_Cell_16_Voltage: 3.272,
Battery_Cell_1_Temperature: 19.1,
Battery_Cell_2_Temperature: 18.4,
Battery_Cell_3_Temperature: 18.3,
Battery_Cell_4_Temperature: 18.8,
Battery_Cells: 16,
Battery_Cycles: 265,
Battery_Design_Capacity: 102.0,
Battery_Firmware_Version: 3011,
Battery_Remaining_Capacity: 19.55,
Battery_SOC: 15,
Battery_Temperature: 21.3,
Battery_USB_present: true,
Battery_Voltage: 52.361 `
App currently showing 4% battery, 51.95V, 19c
TX200 Any evidence of a steeper section of discharge on the displayed SOC% curve (lasting about an hour in the section between 50% and 4%)? If so, you are seeing the same behaviour as me and can expect a sudden jump to 100% on the next charge cycle.
Has anyone got onto 4077 yet, or know the timetable for rollout?
We are 3kwh down on usual capacity, and its starting to cost us a £1 a day.
Woodstock
Yep, steeper decline from say 30-40% - although not always. And yep, saw a jump from 82% to 100% during overnight charge or solar charge.
Hoping this gets fixed in the next beta version, heard people mention capacity, but not sure if that includes what we are seeing. Plus no reply from GE.
Might give them a call tomorrow. 😬🤣
Jamesrich
I know of 2 others at least with it, both on here
What version is the standard non beta? Just checking if I am up to date with public releases.
Software
Inverter Firmware Version: D0.535-A0.535
Last Update Time: --
Battery Firmware Version: 3008
Inverter
Inverter Model: GIV-AC3.0
Software Version: D0.535-A0.535
Thats the latest
TheDragon (GivEnergy) Thank you.
TheDragon (GivEnergy) 3008 might be the latest but I don't think the option to install has been sent out to everyone. I was still on 3007 before getting 3011. I had the impression that 3008 was new installs only.
Woodstock
It is the latest available to the field engineers
Had an SOC drop Saturday evening for the first time on the beta 3077 from about 9%, caused the battery to be emergency charged to 4%. Still seem to have an under capacity issue as well as it only took 8.2kWh to fully charge it to 100% (its an 8.2kWh battery) as when I was on the existing production firmware when it discharged to 0% it would take >9.5kWh to charge to 100%
Battery jumped from 75% full to 100% full as of 4pm. 🫤
My 9.5kWhr battery (installed Nov 2022) has firmware 3009 installed.
Does anyone know if 4077 claims to address the SOC jump issues or is it just restoring lost battery capacity?
TX200 I had a jump from 70% to 100% full at 3:35pm. Must be the weather. 😀
Woodstock This lines up - I was an October '22 install.
My set up is currently from jumping 90% to 100%.
Primary fix is the 80-100 jump.
Secondary is the capacity calcs.
Further capacity restoration is in work, as I feel there is still more room at the lower end, certainly for me anyway
So when do mere mortal get a public release and an update from the company itself?
EmmetBrown Inverter FW D0.450-A0.45 / 9.5 Battery FW 3009. These are what they had as at installation last September.
Just been put on 4077 firmware. Daughter called me to say she heard beeping from the cupboard and when I checked I could see my inverter had gone into idle mode. Checked the software and it showed 4077.
However......happy to acknowledge that when I checked my emails, there was one there from Dan saying he had just uploaded the new firmware whereas last time it was done under the cover of darkness. Dan gets 2 internet points from me for improved communication 🙂
In lieu of Valentines Day flowers and chocolates, please find our latest Statment regarding the SOC firmware below.
Ra there was one there from Dan saying he had just uploaded the new firmware
This makes me much happier than the news that we are close to public release. Improved GE comms is what is needed, especially when faced with future issues.
GeorgeRawding Will the update force a calibration?
First thing I have noticed straight away is that my battery has reached 100% charge and stayed there rather than dropping immediately to 98%
chris_p When I had the beta firmware it didn't trigger a recalibration but I also lost at least half of my battery capacity until GE ran a calibration manually. Maybe they've changed something in the later beta firmwares to fix this.
judgepd I think so. The first firmware update I got it borked my capacity right up so I had to run a calibration (I think I had something like 20%)
Further updates since then have been ok.
GeorgeRawding Thanks George #blushing.
GeorgeRawding thanks George.
Much nicest than flowers.
Ticking my snag list off one by one. Give Dan2 a hug from me, they needs a hug, been a busy lads
3077 received
positor1 you mean 4077?
TheDragon (GivEnergy) Mine has gone from 3077 to 3076 since this morning. What that means in practice is anybody's guess.
Still 3088 for me despite sending an email through on Friday asking for 4077.
GeorgeRawding Thanks for the update and I wholeheartedly agree with the gratitude offered to the beta testers. The feedback on this channel from those testers had given me some confidence that this release would be achieved and even offering some little additions that should improve overall performance. I also think that some credit should be given to the GE team who must have felt somewhat battered at times, and hopefully that hard work will pay off.
Someone is meddling with my battery, I'm charging up, 13:38 it stopped, BMS over, BMS under voltage SoC dropped to 0, back up to previous and carried on charging.
However now in Debug mode
4077 still loaded, so maybe someone forgot I had 4077 already for the past week, and re-loaded it remotely LOL.
Will see what there up to, as I have my own monitoring on my kit
GeorgeRawding
Another dangerously bullish communication from GE, which I suspect will come back to bite them on the backside.
When the (un-named) author of this communication says "We’ll update in batches of 100", is this referring to the latest beta version, or the live version?
I for one do NOT want to have this software forced upon me. I only want it once I decide that it is stable and fit-for-purpose (based on the reaction of other users).
ToothyChris I for one do NOT want to have this software forced upon me. I only want it once I decide that it is stable and fit-for-purpose
Ditto.
@GeorgeRawding Make the link available and then leave it to the user to implement. My system is working OK-ish*....I will be annoyed if a new fresh-outta-limited-beta firmware is pushed into the system.
*Apart from some jumps >90% to 100%. And the batteries discharging at wildly different rates. And the BMS stopping all charging when one hits 100% even though the other is at 85%
I'm still on 3077. I've seen recent mention of the charge stopping if the SoC is above 95% and this is expected behaviour (!). I reported this behaviour previously as a lockup issue and it is very noticeable (when using a real time power flow app) - I'm not particularly happy to throw away another 5% of free charge (when excess generation is available) on a battery that is already substantially below the headline 5.2kWh capacity (known 80% DoD, and current capacity issues combined).
Do we know if the new bestest release candidate changes this behaviour or do I have to accept another 5% capacity reduction forever?
I'd really like the SoC to show 100% when it is properly fully charged, and I'd really like the available capacity at "100%" to be what I'd expect it to be (which I think is 80% of 5,2kWh, with some degradation with age, but maybe I am being naive?). Currently on 3077 my battery is infrequently showing at 100%, and can be anywhere from 95% to 98% but not accepting any more charge.
GeorgeRawding I have one question and does the firmware include people like me who have the older inverters without the additional hardware to cope with the faster response times?
Zakalwe Ditto. My system behaviour and quirks are very similar to yours. So working OKish and hence I'd prefer to update when we see some general stability.
I did apply for the beta and was told I was on it (I wasn't) but seeing the wild ride some of the testers had, kinda glad I missed out. Credit to those who took part, looking more promising now and thanks for sticking with it.
Woodstock We just got a 5.2 system installed 2 weeks ago. Have seen two jumps 69 to 100% and 70 to 100% jumps. Do you think this is common and entirely software related. My set up is installed in a Dunster Cabin, (so not heated but well insulated), with a feed to the house. Battery temps generally around 11deg C. My wife and myself are real newbies so any info will be gratefully accepted
TheDragon (GivEnergy)
sorry. i should have said still on 3o77>
Battery Firmware Version
3077
TheDragon (GivEnergy) same on mine. Looked like a couple of attempts before and after lunch. How do you tell if debug mode is on?
mpartington fast updates on your graphs
And there it goes... Not really empty I'd guess.
TX200 what version are you on?
chris_p 3011 still.
Expecting some sun today, so set a charge target of 85%
Of course it gets to 79% charge and then declares itself full. 🤦
TX200 Oh, isn't 3011 only a couple of versions before the current (latest) beta?
Paul_johnson1 More than likely the well documented SOC% bug. Hopefully going to be fixed by firmware updates currently undergoing beta testing. Watch this space!
mpartington Portal data updated every 5 seconds instead of the usual 5 minute updates
Is there an ETA on the new firmware release date. I'm not expecting a specific date but are we looking at days/weeks or months here?
Thanks
I wouldn't hold your breath - despite the communication from GE yesterday.
I would imagine weeks or possibly months (before they get it right and stable for most customers).
ToothyChris Enough with the negativity, for heavens sake. Everyone is frustrated their system is not working the way they wanted. Persistently bashing GE and passive aggressive cynicism is exhausting, even for users like me. I can't imagine how motivating it is for GE staff to be the butt end of it all. More carrot and a lot less stick please! Do you really think everyone at GE does not want to get this sorted, properly, as soon as they can?
For my part I would like to express my appreciattion to all the beta testers who have worked so hard helping GE get to where we are. It's been a long journey and yes, at times, one step forwards and two steps back. Without you it would have taken longer. GE? Keep your spirits up. Thanks for all your hard work, in spite of the sniping and threats of uninstalls by the few shouting the loudest. I am confident I will continue to benefit from an improving platform for years to come.
Yes, I probably am more pessimistic than most, but experience has taught me that companies and organisations that behave in this way are unlikely to change, until they are absolutely forced to (or they implode). I really do want GE to get their act together, because I quite like their kit. But they have shown no signs at all of taking this (and many other issues) seriously, or of treating their customers (or indeed their installers) with respect.
As others have said many times on here, who are business owners themselves, this is no way to run a business.
The appreciation for the testers should be coming directly, and regularly from GE themselves. But it is not.
Like you, I try to help others on here, when I can. But I also need to vent about the crap that we are having to endure from GE. Venting is a healthy way to minimise the risks associated with poor wellbeing, and god knows there is already enough in this world to get frustrated about.
You asked 'Do you really think everyone at GE does not want to get this sorted, properly, as soon as they can?' Well, there probably are many GE employees who want that to happen, but are powerless to do anything. If the GE powers REALLY wanted to resolve this then they would have a UK-based team of experienced software developers, working closely with a dedicated team of GE engineers, who can fully immerse themselves in this - and communicate in a meaningful way with those of us who were invited onto the beta test.
All views are welcomed and valid, so I have no issue with your comments, even if they felt a touch spikey at my receiving end! Some of us will naturally be optimistic, and others (like me) will be more pessimistic. That is life.
I've been documenting this issue from the start of my first communication with GE back in February 2022 about my experience of this issue. Once I have a working firmware I will be sending these details in a letter to the new CEO in the hope he will ensure they understand what has happened (and not happened) and make improvements. No matter what anyone thinks of GE, it cannot be denied that there are definitely lessons to be learned.
cluelesscris I'm staying positive about the GE equipment and ecosystem as much of it is very good. However the huge SoC issue was (and still is) a major concern for many, particularly those like me who were waiting for a very expensive install just as Givenergy stuck their heads in the sand and ignored their customers. That's pretty much undeniable and inexcusable.
The reliance on beta testers to provide any meaningful feedback (massive credit to them) is also inexcusable. If the new management structure thinks this is the way to operate, irrespective of what else they've got going on, well the future doesn't bode well. I can only hope for all our sakes, that they learn from it and act appropriately in future.
On a more positive note, it does look like there is some light at the end of the tunnel and hopefully we can all benefit soon. My own guess is that it's still several weeks away, but as long as it happens and it's reliable, I'll be happy enough with that. But please don't defend GE too much on this one. Whilst there are no doubt some talented devs working their nuts off and doing their best, the overall company response to the problem has been demonstrably poor. My guess is that they're swamped on several fronts and they need to learn to walk before they can run. Hopefully that's in hand with the new management team.
Next big test will be the new products they've promised and the new factory. It will be interesting to see how it all progresses and good luck keeping it all on the rails.
We must remember we're not in normal times. There's been a whole host of issues in this decade, COVID, supply chain issues, a war, an energy crisis. Demand for solar and battery equipment has gone through the roof.
If the Devs are in an environment anything like my workplace (do more with less people etc) then I feel a bit sorry for them. I hope they're being treated well by the management!
TX200 Absolutely, but communicate really well and manage expectations. Don't ignore the problem for months and months, then try to minimise it, then promise a fix prematurely, miss the date then stop communicating completely. At the same time don't promise equipment deliveries that don't have a chance of materialising and don't leave customers with failed equipment waiting and waiting. Oh and no apologies.
It's been a perfect storm of challenges for sure but some have managed better than others. Anyway, let's hope most of that cr@p is behind us now.
sponev sounds just like my senior leadership team 🤣
Promise the world with unrealistic timescales.
cluelesscris While I do in part agree, I believe that the carrot was the £17,000 we forked over for 4 of their 8.2 batteries and 3 of the AC3s to be installed. I am glad they are finally getting on top of this issue even if we are now 2 months past the last deadline they gave out. What bothers me and I'm sure many others is that this issue didn't appear 3 months ago. Or 6 months ago. It's been reported for more than a year but it certainly felt like it was only being taken seriously in the last 4 months.
Don't get me wrong, I am hoping to get 5kw or so installed this year and with that would like to pair an All in One. Hell I'd love to put one of their energy diverters in but that also hasn't appeared more than a year after they had it on the website and so I'll be forced to look elsewhere and that naturally then has you looking at other battery systems too.
I'm looking forward to the fix finally being made public and hope that it won't be too much longer as the sunnier days are already appearing.
And the cold issues, they didn't put anything about wrapping it in a blanket just, its IP65, and yes i know batteries dont perform that well in the cold but when they have had these kind of issues and before you know it they are releasing stuff and making the stuff that doesn't work obsolete.
My electrician says no voltage coming from the board to the battery which won't take a charge. pay money so an engineer can come to a 5.2 that wont be able to last as it 80% dod. not wot your told when buying and future upgrades when i was following evm .
Respectfully, please can we all get this chat back on topic. If it's not about the beta testing activities, it belongs somewhere else.
A tweak to 3077, was tested yesterday as 3076, less the capacity changes in 4077
Today the same tweak added to 4077 to make 4076. Lower voltage looks lower by 1v. Will post some graphs later on after a cycle or 2.
The tweak was related to OCV at low end dropping too fast on some users batteries.
A small subset of testers will have 4076 soon, if nothing else is added, Im sure Dan will send it to the wider beta team
TheDragon (GivEnergy) I know your experience is clearly different - but just getting anything other than 3088 (and, yes, I've emailed) would be great.
So there have been at least 4 versions after the one I was given...how are people being chosen to get newer versions? I'm asking you because at least I stand a chance of getting an answer from you 🙂
TheDragon (GivEnergy) likewise I'm keen to move off 3011 to get the missing third to quarter of my battery back!
Not the most ideal beta test to have some of the pool of test sites not testing the latest firmware.
I don't control which testers this goes to, I can assume, the minor tests are to a small set, then if no issues and the issue addressed, sent out wider.
I will ask on timeline for 4076 to go out to the wider pool via Dan Maddocks
TX200 Ditto!
TheDragon (GivEnergy) Is that a stabilised 50.8V at 4% SOC (instead of 51.8V)? It will be interesting to see the effect on the <50% SOC adjustment processing.
On the left BMS 4077, on the right 4076, showing the lower cutoff under load, then the lower idle point
I need to do a couple of cycles, as I said before to confirm this wasnt a fluke
Past 14 days, can see the 4076 is deffo lower
TheDragon (GivEnergy) I wonder if they have taken the opportunity to also reduce the threshold voltage at which the <50% SOC adjustment kicks in. This would get the table look up on a flatter part of the SOC curve but I do not think that this alone would account for the more than 20% error that the SOC adjustment is currently giving me. Likely something more is going on. Happy to continue looking but there is only so much that can be achieved when you are five releases behind. 😀
Woodstock Am I reading correct you are forcing the SoC adjustment? Sorry if not, I am a bit lost on your reports
System keeps charging from the mains.
My system - 12-panel roof array, a 5kW inverter and 9.5kWh battery persists in charging from the mains when it reaches its lower limit. I've set a lower limit of battery discharge at 15% and the system does as it's told, dropping smoothly to the lower limit and going idle. Then the reported battery percentage drops below 15% without feeding any power into the house and the thing starts to charge itself off the mains - which I really do not want. And there doesn't seem to be a way to tell the battery it cannot charge from the mains.
What on earth is going on?
I have a few days to decide if to change to one of a few other options… this doesn’t feel like it’s resolved.
Am I missing something?
ed-form your system appears to be working as expected. The battery discharges to the 15% limit. The measured SoC value will vary slightly over time due to battery chemistry and temperature (I see from other posts that your system is outside which won't help). If the measured SoC drops, the inverter will charge the battery slightly to maintain the reserve that you've set.
Out off interest, which beta firmware version are you running?
It was a fluke.
The low reading I'll put down to changing FW while not flat. Which is my preferred loading time.
OCV however has changed. It only happend once
4077 appeared on my system this morning.
stevelewis
I wondered if that was the cause - I guess that the battery temperature falls during the cold night and that alters the system's estimate of the charge level. I didn't realise that this would cause the system to mains-charge the battery. So this is normal behaviour - typically, the charge taken from the mains is in the order of 200Wh.
Firmware is D0.104-A0.104
ed-form My 9.5 battery has been doing this for some time. I had lower limit set at 10% at which point it would correctly go idle but would then lose about 1% every hour until it hit 0% at which point it correctly charges from the grid back to 4% but would then once again drop by 1% every 3 hours. This equates to a continuous 30W (approx) of internal system drain.
Earlier this week I rebooted the battery and the hybrid 3.6 Inv and found this improved things as I am now getting between 6 and 7 hours per 1% idle drop
Sounds like a feature request for a soft user discharge lower limit and a harder user lower limit.
Essentially stop discharge at the soft limit, but don’t grid charge until the hard limit.
If anyone from GE is watching this, kram's idea of a "stop discharge at the soft limit, but don’t grid charge until the hard limit." feature looks like a practical idea.
joshgreenuk Not so much forcing as engineering that the right system parameters prevail for the adjustment to take place. By doing this I can predict when an adjustment (and subsequent sudden jump to 100%) will happen. The system comes to its own conclusion that an adjustment is required, I am just putting the system in a position where I know it will act.
judgepd I wonder why not 4076?
ed-form not sure how this could work... you either need the reserve or you don't!
stevelewis I don't have the issue that someone else reported of a fairly quick discharge after the soft limit is reached. My system drops to 15% - my soft setting - and then very slowly falls to 13% - at which stage a grid charge kicks in. I've been set like this for a week or two now and it's never gone any lower.
Essentially I need Approx “X” % reserved, but I’m happy for it to be this low before you charge up.
It’s a tolerance factor, to optimise running costs.
For example I absolutely need 15% to keep the freezer running for 3 hours, but I’d like to keep 20% in the battery to keep some lights on. However as long as I have 15% or more don’t spend money topping it up.
It would only require one extra variable and a tweak to the logic. The discharge logic would check against upper soft limit. The grid recharge logic would check against the hard limit. Code wise it’s fairly trivial, although the extra test time might mean it’s an iterative update in one of the point update of the final release of the beta.
It stops the battery topping up tiny bits regularly when you have a good degree of confidence the sun will fill it up soon.
judgepd New behaviour on 4077, the battery has stopped charging at 97% rather than 100%.
judgepd Actually, it did charge to 100% but then dropped to 97% from usage and won't start charging again until it drops below 95%....I find this new feature unwelcome.
judgepd Did the charge ramp down before it stopped or brick wall?
judgepd Isn't this how the current public version of the software works already?
So for example a scenario where the battery is charging from the sun, gets to 96%, cloud comes over, stops charging, then sun comes out and the battery doesn't continue to charge? Who would want that?!
chris_p No, it's not quite that. Once it gets to 100% if it drops it won't charge until it drops below 95% again.
judgepd So sun charges it to 100%, cloud comes over or house demand exceeds solar, battery drops to 96%, demand drops/sun comes out and it won't charge the battery? Yeah, still don't want that.
chris_p Correct and neither do I.
sponev
An engineer from GivEnergy came on Tuesday 14th Feb and sorted it, I was very impressed with service and technical expertise.
Thanks also to EcoPartners, the original supplier/ installer who got this all moving quickly.
I’m back where I need to be, lots of generation and storage.
😀
judgepd I don't want it either! This is my issue #2 from earlier in beta and it sounds like the latest and greatest beta doesn't resolve it.
Worth suggesting to the team as a fix for what most would see as an annoyance?
judgepd
Essentially an optimised charging routine to stop cell cycling.
Again easy to make optional with a tiny bit of work. Probably on by default but can be disabled through the API only. That way you only do it if you really mean it.
All. Call for help
I read that some are seeing the inverter not charge once it gets to 100 and back down to above 95%.
Can you give me examples please of this happening, example you are charging to 100. Then discharge to 96% and the sun comes out, however the power isnt put in the batteries, but left to grid.
Then once the battery gets under 95% it will charge again.
I am going to write a consolodated report to Giv on this issue, to get them to investigate if its BMS or inverter FW driven.
So will need battery type, inverter type and graphs, evidence please.
I personally havnt seen this, but am now going to try and recreate the conditions by force charging to 100. Then discharging to 96/97% and seeing if the sun charges again
How about they actually ask the beta testers themselves and act like they're actually responsible?
Rather than passing everything through yourself...
TheDragon (GivEnergy) I have provided examples via the beta email thread back to GE already. Never had a response from them. They should know about it already as I've complained about it several times. Here is a snippet of my last feedback to them about this issue (hopefully you can see it?):
Whilst I do agree with you… I also think if they are talking to a person and that route can get a fix better fixed than not.
GE communication has been terrible IMHO, but that’s my impression from on here.
FWIW I assume that last 5% isn’t just volatile tolerance based on temperature changes or allowances for them? 100% at cold and 95% at warm night just be tolerance for night/day temperatures?
JaseSpace And here is another capture from the day before:
In this one it takes about 2h15m for the charge to kick back in, presumably as it dipped below 95% due to the load profile. Not a huge issue in the middle of a sunny day where it recharges back up quickly after awaking, but imagine a point where it is stalled and the sun goes in - when it eventually wakes up there is no excess to charge back up. You can see this at the end of the previous image to some extent although it wasn't as low as 95% at that point so not missing the full potential 5% of capacity in this example. But this is why I say it is more potential loss of capacity, since we could have been using earlier excess generation to keep the battery at max capacity, instead it allows it to discharge to 95%.
BTW I am Gen1 Hy 3.6, 5.2kWh (80% DoD), with BMS 3077 and D0.104-A0.104 Inverter FW.
JaseSpace thanks mate.
For the record, giv arent channeling anything through me! It's something I heard about and was gonna submit a report on it.
@ianfretwell if you prefer I didnt bother, then that's not me.
I spot an issue I'll take it up.
JaseSpace when did you 1st notice this? Trying to think if it's an inverter FW bug or was it introduced with a later BMS?
TheDragon (GivEnergy) It's not that I don't want you to at all, it's a good job you are - it's just that for me and many others YOU are the only source of info, GE themselves won't/don't respond to emails from beta testers (again - except for you it seems) despite them asking for us in the first place. It's just disappointing.
For me, all beta testers should all be on the same firmware at the same time - so that each firmware gets maximum relevant feedback (whereas we're all over the place currently). I'm sure any feedback I give now whilst still on 3088 is completely and utterly irrelevant - so why not upgrade me too?
TheDragon (GivEnergy) And just to show good willing 🙂
In my case by the time the charge had dropped low enough to start the recharging the solar output was too low anyway. But it still shows the (highlighted) drop from 100% when solar was still sufficient to have keep it at 100%.
](https://)
From AC3 on D0.290-A0.290 / BMS 3088
ianfretwell Depending on the purpose of the test, having all beta testers on the same version may not make sense; however, I agree that more direct comms would be beneficial, especially for those of us still on 3011!
ianfretwell Are you saying this began with BMS 3088? was it present before 3088, or was that your 1st beta load?
Inverter is the same as mine
I have just done a test, Set a force charge from the App. (Battery + icon).
It charged as expected till 97% then began discharging to meet load, despite being set on force charge. So I think these 2 issues maybe related.
Something is going on at 95% and above, just dont know what, or if its BMS or Inverter related yet.
stevelewis Correct, there were 3 tweaks issues end of last week, 3 different people, to see the changes.
TheDragon (GivEnergy) It's the only time I've seen it happen - but it's also the only time solar has ever managed to charge the battery to 100% (only a 4kW system WSW facing).
The odd times I've done forced charges to 100% there's been no solar to keep it there anyway so the immediate drop was expected.
ianfretwell OK thanks.
Do you mind if i take your graph above and add to the others from Jase?
TheDragon (GivEnergy) Absolutely - feel free.
For transparancy, here is whats been submitted
Dan,
The issue here is about charging from Solar, it will charge to full, then export solar, while discharging to load, down to 95% then start charging again. Instead of holding the battery full, it is discharging.
Same with Forced charge using teh Battery + icon. Charged as expected, then once it reached 97% began discharging to load, despite the force charge still on.
If you do the force charge at night, it will charge to full and remain there, till charge removed.
As soon as there is solar present, it has a mind of its own, ignoring the force charge option, or the available sun and discharges the battteries to load.
Attached is a PPT with some examples from me and 2 other users, one from Jase Space, is very clear example of it.
As this seems to be affecting BMS 3088, 3011, 4076 and both AC3 and Hybrid, it could be something not yet seen. Now we have some sun, its appeared and affects everyone
TheDragon (GivEnergy) First reported to GE via beta email feedback on 11 Jan (I was on 3094 at the time). Here is my feedback (verbatim):
I joined the beta program on 16 Dec but was away for Xmas so didn't spot much until back in Jan. So it may have been happening since the start of my beta, or may be from 3094, can't say for sure. There wasn't much solar around at the time so wouldn't have been as easy to spot as it is now. Never noticed it at all before being on the beta program.
History of my beta updates:
Before: 3007, D0.450-A0.451
14 Dec: 3092, D0.103-A0.103
29 Dec: 3094, (D0.103-A0.103) <- issue definitely happening as early as this set of versions
about 13 Jan: 3011, D0.104-A0.104
3 Feb: 3077, (D0.104-A0.104)
Depite repeated submissions about this and questions about whether it is new/expected, I have never had any response from GE about it, and the issue remains even today on the latest beta (as I understand it, but GE haven't moved me on from 3077 yet so I'm a little behind others still).
Thanks for taking this on, BTW. It would be great to get it fixed.
JaseSpace OK have added your text to the email, I sent above
No, not the sudden SOC% jump around 15:00. (When is that going to be fixed by the way?)
Graph shows test results of the reverse situation (empty battery, how much do I have to charge before I can discharge again).
Tried four separate timed exports between 10:20 and 12:20. At 4%, 5%, and 7% displayed SOC, the inverter just went into idle and no discharge took place. At 9% displayed SOC the timed discharge took place. What the actual SOC% was at each of the four test points is anybody's guess!
TheDragon (GivEnergy) I read that some are seeing the inverter not charge once it gets to 100 and back down to above 95%.
Forgive me folks, but I like this behavior. It smacks of being a bit kinder to the battery. If the battery is always topped up (as it will be in summer, surely?) then the life is going to suffer.
...yes, yes I know there's a buffer at the top, but we all know what happened to Mr. Creosote... I like the behavior of don't force feed when it's so nearly full. Or at least make it a setting.
<running and ducking>
BobB it's a bug that needs fixing... I bought a battery so I could use it. If you don't want to charge to 100%, there's already a limit which can be set.
Woodstock
Ok one rub of the genie lamp used up.
Just spoken to Dan Maddocks, he is going to see if he can identify your inverter from your beta emails, if they have woodstock in them.
He is going to load you up 4077
He is already going through the list one by one updating everyone to 4077. But its manual this way and slow.
stevelewis If you don't want to charge to 100%, there's already a limit which can be set.
Meh - sort of Steve. I do appreciate you can limit the AC-fed upper limit. But you cannot limit solar-fed upper limit.
I've raised a feature request.
anglefire I,m in the same position and I find it a bit concerning that no one from Givenergy has responded. If ours can't be fixed , then I would expect a replacement inverter .
BobB So you want GE to modify the firmware in a G98/G99-approved and CE/UKCA-marked product, with related changes to the API, web portal and app. That's probably £100k of work. And all this to add a feature that enables less of the battery to be used... Don't hold your breath!
jqworle The inverters can be modified - so I'm not too concerned to be honest. But it would be nice to be able to have the newer firmware - not that I have any issues with SOC jumps, so for me the time frame is not that important.
I just like new and shiny 🙂
TheDragon (GivEnergy)
In my experience this happens when an overnight charge is set to less than 100%
eg two days i had it set to charge to 70%.
solar took over the next day and charged to 100
5 for just 5 mins then started discharging ( in my case an Eddi takes it so no big deal)
when the batteries dropped to 95% i did a reset to default and the batteries started to charge from solar again. they may have dropped to lower than 95 before i changed settings of course and may have started charging on their own?
As you have reported before i saw the post i wont search for the graphs unless you ask. i did report this on the beta email thread on those days thougjh.
TheDragon (GivEnergy) Big thank you to you and Dan. 4077 has landed. Interestingly there was no period of rapid data updating and I spotted it this morning from the obvious decrease in the slope of the SOC% curve. Checked in "My Inverter, Software" and there it was. 😀
Woodstock Now can you review it, see if the issues you report have gone, feed back via emails.
See if yuour jump to order is no more
Had second 8.2 kWh battery fitted, updated to 4077 and Calibration run. Impressed with new performance but it is difficult to separate benefits of 4077 from extra battery. Only using portal which sees batteries as one 320 Ah unit.
18.1 kWh charge from 0% 45.94 V to 90% 55.83 V then voltage initially declined before 100% at 57.28 V. Stayed at 100%
Subsequent PV charge to 98%, jumped to 100% then fell to 98% with charge inhibited until 95%, exporting solar.
Cannot understate benefit of extra battery, much "smoother" operation and 3 kW demand when at 8% SOC dropped voltage to 49.08 V but recovered and inverter did not idle. I was considering extra AC3 when EMS available but not any longer. If AC3 consumes 30 W then that is 0.7 kWh/day , 5 kWh/week or 250 kWh/annum.
Calibration recharge took 8 hours and inverter temp rose to 43C in a cold garage.
Glad you are happy. Please find time to submit your findings to betatesting@givenergy.co.uk
The 100/95 issue has been reported, its not designed that way, so needs fixing. So please submit examples of it hapening, so we get more ideas on the root cause and when it was introduced
TheDragon (GivEnergy)
i think it started with 3077
positor1 when did you apply 3077?
TheDragon (GivEnergy)
it was sent to mine on Jan 14th.
stayed at 100% for 30 mins today.
TheDragon (GivEnergy) Thanks.
Note that I have mentioned this in much earlier posts but forgot to mention it in my most recent posts - when in this charge-stalled state, I can wake it from its stupor by issuing an Inverter Restart from the Portal. After a delay of >1min it comes back online and starts charging up to 100% (if there is excess available), but obviously I have lost the use of the inverter for that restart time so doesn't come for free.
So I wouldn't be surprised if this bug was introduced with inverter FW changes, but I don't know enough about it to say for sure). As I said previously, I first noticed the issue some time after this update: 29 Dec: 3094, D0.103-A0.103 and have had it ever since (through 3011, 3077 and D0.104-A0.104).
Inverter Firmware Version
D0.290-A0.290
Battery Firmware Version
3077
so must be bms not inverter software as mine has not changed
I have a 3.6 Gen1 Hybrid and the battery dropping while solar is aplenty is definitely something new. I distinctly remember my battery going idle for long periods of the day in the summer last year once it had reached 100% and it stayed there until house demand was greater than solar.
TheDragon (GivEnergy) So here is mine from yesterday:
On a positive note, on my first complete empty to full charge on 4077 I added 7.4kWh which is the highest I've seen for a very long time (should still be higher though).
judgepd
Thanks. I have already submitted finings to Giv now, But anyone who has examples of this 95% not charging from solar, examples, please submit them. betatesting@givenergy.co.uk. @Danlambert (GivEnergy) will discuss with China why this is happening
They have now acknowledged my email and asked for names of people with the issues, however I don't know this info, as the forum isnt real names.
So please submit direct.
@JaseSpace I am leaning to agree its a FW issue, as a reset gets it moving along again. As does a Pause and resume.
I was also able to replicate using the forced charge option in the APp.
Once it hot 98%, it began discharging, which it should not, force charge should just sit at full, till play (Resumed). Which it does at night, but if there is solar present, it ignores the setting. Indeed an odd one.
Been away for a while so playing catch up. Is this solar charge issue hybrid related? I don't see it on my AC3 290 firmware. My battery is still on 3077 though I have requested the newer version today for the capacity improvement.
You see similar with a charged period to 100%. It'll hit 100, then drop back a bit to 97ish or so and sit there.
D0.906-A0.906 on a GEN2 5.0kW Hybrid also charges to 100% with solar and then drops away to approx. 95% before holding the SoC level even with plenty of generation available and I am not in the beta program ............
athresuk Can you provide any graphs like above, demonstrating this?
The fact its present pre Beta, is new, no one last year noticed it
athresuk That is interesting since if I go back through my historical data, to find a day before beta where I was getting to 100% and still excess of solar (need to go back to Nov for that), then I can see it behaves differently - SoC holds at 100% for hours until excess drops, then SoC starts to deplete. This is how I'd expect it to operate. Since the beta, it no longer operates like that - SoC depletes to 95% even with excess.
Here is one example:
This example is probably better as it shows obvious battery charge power during the 100% SoC period:
One word of caution - could we actually rely on the reported SoC being correct in that version of FW? Probably not, but the fact remains that according to the reported SoC, it used to keep the battery topped up when at 100%, and it doesn't any more.
JaseSpace
Same here AC3 Way before Beta, had to go back to late September to get a good example
Issue for AC3 at least is since Beta was added. Just don't know at what point, as the early days were very little sun
TheDragon (GivEnergy) Could this be purely a display bug? Is there givTcp data available to show this drop?
THALL
Good shout, I will take a look
TheDragon (GivEnergy) I have had graphical issues previously when I last tested a beta firmware. I was seeing graph spikes of 70000w which is impossible, it was purely graphical.
THALL The current issue is definitely happening - I use a different real time app to observe the power flows and can clearly see the charge-stall state. But this app has no history so I can't comment on whether the historical graphs from the Portal were accurate (but I suspect they were). I have multiple examples.
Here you go, portal image for SoC droop on a GEN2 hybrid
Battery charged up to 100% and as the charge rate tailed off the Eddi started to mop up the excess.
Once the tank was up to temp the inverter started exporting, but the battery SoC continued to fall and stabilized at 95%
Unfortunately Home Assistant won't display data that far back in time 🙁
For me this wasn't happening pre-beta (at least in the 353/3008 combination). Might have been happening earlier than that, but in Dec it looked ok. This is from 8/12/22 and it reaches and holds 100% as expected.
Worth pointing out that the GEN2 displaying the SoC droop is from the Jan 2023 batch, so has newer FW (D0.906-A0.906) that earlier units supplied last year (FW D0.904-A0.904)
THALL
The 18th area of interest zoomed, showing the force charge deactivating itself to charge solar
Force charge off and on, got it charging again briefly
judgepd My first full discharge (100% to 5%) under 4077 was 5.7kWh and the first full charge (5% to 100%) was 6.8kWh. Both figures are only marginally better than under 3011 (100% DOD 8.2kWh battery).
Going to give it a few more cycles to see if it "learns" the true battery capacity.
Are we not actually seeing a couple of different issues here:
But did your SoC jump on demand happen?
Here is the same plot for the GEN2 with the battery power highlighted so it is more obvious :
Once the battery is up to 100% the battery is then discharging (to include some as export) until 95% and then a re-charge occurs to maintain 95% and the process repeats, but the solar generation faded away.
The GEN1 hybrid previously at this site did not do this.
TheDragon (GivEnergy) Haven't tried that yet. Want to give the system the best chance of avoiding it by letting it establish where it thinks it's at first! 🙂
Do you mind me or you sending this to betatesting@givenergy.co.uk
I can see the battery is actually discharging instead of holding at 100% till the sun is lower than load
OK
I have had a response already. Inverter FW. Issue is in work, and many thanks for the various examples.
No one has ever reported this before, till this thread, so Dan is very thankful.
Watch this space for updates on a revised inverter FW, likely Beta 1st to see if its fixed, then my feeling will be the issue fixed in the Gen 2 aswell
TheDragon (GivEnergy) go for it
TheDragon (GivEnergy) Good news on one hand, but as per the evidence I shared on here a few days ago, I reported this on 11 Jan via the GE beta feedback email thread, so clearly they aren't reading their beta feedback emails. I also mentioned it earlier on this forum thread (or perhaps the previous forum thread about SoC jumps before I realised this thread existed). So they could/should have known about this issue nearly 6 weeks ago!
Anyway thanks for taking the initiative and hassling GE with this - they seem to be listening to you at least.
My system got 100% today via solar, it is pinned at 100% Currently pv 2.1kw load 200w export 1.9kw 100% pinned. It did jump from 98% to 100% like it has always done on this beta firmware. Currently on AC3 290 and BMS 3077
THALL I can see 3 times today my charge has got to 100% and instantly dropped to 99%. On my overnight charge and then twice more from solar.
judgepd I can only assume is this on select systems? Or I've just been lucky and missed this issue.
THALL Could be on firmware 4077 as I think you said you still have 3077?
Edit: found 1 occasion of solar charging to 100% on 3077 and it stayed there. Every time since 4077 it's immediately dropped to 99%.
judgepd I am on 3077. Though I have requested 407* to get the capacity back.
judgepd Yesterday I artificially held my system (4077) at 100% SOC for several hours by leaving a timed charge to 100% in place. No blip took place.
Today I let it get to 100% though solar power (ECO mode) and had a 99% to100% to 99% (and stay there) blip!
Mine didn't hit 100% overnight (stopped at 99%). Once on solar today it hit 100% then dropped to 99% straightaway and then later in the day it hit 100%, dropped to 99% immediately and then dropped to 97% in 1 go (all the time solar was outstripping demand).
Makes it all the more weird, @THALL hasn't seen it, and his is behaving exactly as expected
@Woodstock I have found artificially holding at 100% via forced charge worked at night. When I tried it during the day, it failed and discharged, as per drawing from the 18th.
Giv are fixing something, so maybe its a particular set of circumstances to align the holes in the cheese.
Wonder if it is as simple as a rounding error with the graph? 99.9% = 99%, but 100%=100%, with that kind of fraction difference it might be causing issues. But the devs know about this now so no worries...
I said earlier that i suspected having an overnight charge to less than 100% may cause theis issue (solar to 100% for 5 to 15mins then discharge) but that isnt the case.
today mine got to 100% at 1pm and stayed there for 15 mins, then started discharging at around 300watts.
i tried pausing and resetting to defaults but no change.
Just as a balance, I am not part of the beta testing and my system has never done any of this getting to 100% and then dropping whilst there is still solar in the air.
When it gets to 100% it stays there until there is not enough solar being generated to support the load.
I'm on inverter s/w 499 and batt s/w 3007. I do not play around with discharging etc, mine stays in eco mode.
I do occasionally still get the SOC drop, but it's only from 4% when the batteries have not been able to fully charge to 100%.
Today was all about establishing the baseline for battery reserve. This is the first time ever (pre-beta and beta) that I have seen active learning taking place. During the entirety of this graph no battery charging took place but the displayed SOC% corrected first from 5-15% then from 5-10% and then at 5%. The OCV for displayed 5% SOC stabilised between 51.0V and 51.02V.
As can be seen the scenario was a bit contrived but I wanted to give the system the best chance of discovering where it actually was. In the real world this would happen (with my normal usage pattern) maybe once or twice a year. It is looking very likely that the firmware roll out is going to have to include a re-calibration.
The above is the good news, now for the not so good. Why was the displayed SOC% 10% out in the first place I hear you ask. Well this was down to a lower half (displayed SOC% <50%) adjustment that took place last night at 21:40. The adjustment dropped the displayed SOC% by 10% and had I then chosen to charge to full (rather than discharge to empty) I am pretty sure I would have seen an SOC% jump. As was happening under 3011, the lower half adjustment is simply returning an incorrect SOC% value.
Anyhow, the next step is to charge to full and see how big the system now thinks my battery is!
Update: The SOC% has just dropped to 4% and the battery voltage to 50.99V.
Woodstock for what it’s worth, I’m running 3007 and due to crap sun today no charging has taken place and the SOC has stayed at 4% all day. I’m expecting a grid charge anytime now as the battery voltage is less than 46v. Battery temperature has dropped from around 22’c to 15’c so probably accounts for much of the change.
Woodstock After getting the battery down to 51.0V at 4% SOC yesterday I ran a timed charge to 100% SOC today. The battery took on 7.0kWh, which is significantly less than I would have expected. 😞
Woodstock out of interest what was the battery voltage @100% SOC?
Woodstock Better than what I'm getting on my 9.5kwh on 3077!
Anyone know the timeframe for getting the beta testers updated to 4077?
Jamesrich I've been on 3088 since Jan 11th. I've requested a newer version on three occasions via email to the beta address - all have been ignored as best as I can tell...
Jamesrich Better than I was getting on 3011 too, but around 1kWh worse than I was getting on 3007!
alfwro Around 52.6V. Was 52.8V under 3007.
Under 4077 the 100% SOC voltage sometimes stabilises and sometimes embarks on a gradual decline. Haven't yet worked out why this varies.
These low capacities that lots of people are seeing are almost worthy of a mass warranty claim for not providing advertised capacity. I reckon I've been 20% down for well over a year now.
Woodstock as expected mine did a small charge to get back up to 4% SOC - now sitting at 51v and 4% after a minor amount of charge and discharge today. - basically back to where it was after the top up charge.
chris_p Direct line arent on comparison sites.
Everyone knows that
anglefire A stabilised 51V and 4% SOC do indeed seem to be the target values for reserve capacity but I am wondering how often in the real world a stabilised 51V will be reached. It took ten hours of no charging and three discharge adjustment cycles for my system to get there.
First time my 4077 firmware has gone down to zero % last night and looking at the portal, it claims to have put 7.88kWh into my 8.2kWh so this is a little more than I've seen previously but still below what it would do when I was on 3007 firmware (9.5kWh from 0%)
Ra but still below what it would do when I was on 3007 firmware (9.5kWh from 0%)
Was that a calibration, crash to 0%, or normal full charge from 4% to 100%?
I am still on 3077, I emailed asking for 4077 to test the capacity issue but not heard anything. Annoying. Could do with 100% capacity just now.
THALL Thus far 4077 has gained me about 0.5kWh. Still need at least another 1.0kWh to get close to 100%.
chris_p Had a pretty heavy evening in the kitchen with dinner so my battery dropped from 10% to 0% about 10.30pm, charged itself to 4% and then charged from 4-100% on Octopus Go.
So even though I'm still seeing slight SOC drops, with the 3007 firmware given the discharge/charge cycles over the last week my battery had seen this would have dropped from something more like >40%% to 0% instead of 10%
I think the best case scenario on 3007 was about 10% to 0% even if I had fully cycled it the day before.
I'm prepared to accept this firmware as long as the capacity increases some more....
Ra I wonder why it ignored the 4% reserve on the way down?
Ra so my battery dropped from 10% to 0% about 10.30pm, charged itself to 4%
Huh?! Isn't this what the new firmware is supposed to address?!
@"chris_p"#p21592Maybe discharged not sudden dropped. It's not clear. Either way it should not have gone down to 0%.
anglefire
good question, how do we identify and realise if we have the new or old inverter hardware revision?
From the reading of the announcement though it did say the update will be rolled out to every customer, so that would imply new and old systems.
Clarification on this point would be good though.
It sounds to me like that sticking point has been removed on the latest announcement. Time will tell.
It was mentioned any Hybrid with a serial lower than SA/SD 2209.... may not get on well with it however that was never "officially" confirmed one way or the other unless I missed it.
Something weird happened yesterday, on the app out battery charge level suddenly shot up from about 43 per cent to 100 per cent. when it was being restricted to an input from the grid of about 1.8 kWh.
This happened at around 3pm, when the charging time I have set ends. I checked our smart meter and there was nothing to suggest enough electricity had been supplied to charge the battery to full. It's also been quite a dull day here in Chesterfield, so there certainly wouldn't have been sufficient input from our solar array to achieve that, either.
The lights on the battery itself also showed it as being fully charged, or at least they did about half an hour ago when I first noticed this.
I reported this to support and the reply, which came very promptly, stated: "This is a Battery Firmware bug, we are aware of this and our Dev team are working on a fix for this." Not sure if this at all relevant to this discussion, but thought it might be worth mentioning.
24 hours, on the battery level is still much higher than I'd expect. Photo attached, I hope!
RogerGreen same here.
So if there really was 100 per cent in the battery, how did it get there?! I suppose one possibility is that the system had been under-reading the power going into the battery over several days, and suddenly corrected itself to give a true reading. We've had the battery just under three weeks and can't wait until the solar starts kicking out more electrons and our supplier reintroduces a cheap overnight rate.
RogerGreen that's the reason.
Its fuller than it thinks.
@Woodstock the 95+ odd not charging behaviour looks fixed on Hybrid Gen1. Report from an internal beta tester, not active on here
Waiting for my AC3 version to check that is also fixed.
Need to get info if it's fixed on Gen2 too.
This is an inverter FW update
Sorry to be a naysayer but clearly when they said dec 16th, they didn’t specify which year.
This really doesn’t seem like it’s close to being ready for formal public release any time soon.
Lots of new issues or repeating of old issues.
So do we have a list of what are the outstanding issues?
TheDragon (GivEnergy) Can't myself think of any other explanation. Thanks for your consensus!
TheDragon (GivEnergy) wonder if they've fixed the 100(ish)W discharge whilst battery is paused as well
Woodstock No idea!! One difference I've noticed is that on firmware 3007 when it used to drop to 0% and then charge to 4% was that the inverter would sit there making clicking noises. I used to isolate the battery master switch and it would stop. I'd have to remember to switch it back on before going to bed though so it could charge 🙂
TheDragon (GivEnergy) Its not fixed for me running 4077, I have a Gen1 hybrid.
That was in work, they are doing something with registers for the pause function, also linked to the Home Demand discharge either not working or full export.
Ra What FW?
As far as I know the fix i mentioned above, its on 1 inverter for internal testing. Before anyone asks, its not mine.
TheDragon (GivEnergy) I'm running 4077 with D0.191-A0.104
OK, The Charge fix is D199-A104. Which may et wider Beta, of they may wait till other fixes are incorporated.
I was merely mentioning the Fix for 95+ charging has been resolved now
TheDragon (GivEnergy) Ah ok, I'll sit and wait for that to turn up then, many thanks!
Ra Look awfully like it has hit the low voltage alarm and "sudden SOC% dropped". Did you get a "BMS Under Voltage" notification?
Not had anything similar on 4077 (or 3011) but I have been deliberately keeping the discharge rate low for capacity testing. I will include a fast bottom end discharge in the next test cycle.
Woodstock No notifications in the portal.
TheDragon (GivEnergy) I haven't got around to doing it yet, will do it later today.
Inverter DO.290 - AO.290 Battery 4077
Re-ran the slow discharge test.
Lower half SOC% adjustment took place again. There was the same 10% drop in displayed SOC, taking place over 40 minutes. (Under 3011 this adjustment was a repeatable 20% drop over 60 minutes).
Reserve level SOC% adjustments this time went from 5% to 10% (previously 5% to 15 %) and then 5% to 8% (previously 5% to 10%).
Stabilised OCV at a displayed 4% SOC was 50.88V (previously 50.99V).
Still don't know why there should be a lower half adjustment of -10% SOC when this is immediately reversed (and more) by the reserve level SOC adjustments. If there had been no lower half adjustment would I have just seen a single reserve level adjustment of 5% to 8% SOC?
Does the lower OCV at 4% SOC mean that the learning process is continuing or was it just random? Either way I am hoping that this may have increased the battery capacity and will check when the battery gets back up to 100%.
Woodstock ... and the answer was the battery took on 6.9kWh to go from 4% to 100% SOC. Less than last time! 😞
Ran the fast discharge test and the system successfully stopped discharging at reserve. Well 5% not 4% but that is not unusual!
A single reserve level SOC adjustment took place from 5% up to10%. (Because of the fast discharge there was no lower half SOC% adjustment on the way down).
Battery discharged back down to 5% and the inverter then went into idle with no power showing and the battery showing no charge/discharge.
I next ran a slow (79W) timed charge for an hour which brought the displayed battery SOC up to 6%.
At the end of the timed charge the inverter went back into idle but was now showing a continuous power of 23-27W while the battery was again showing no charge/discharge.
This state continued for several hours with the displayed SOC% slowly dropping until, after it had been at 2% for an extended period, an automated 290W charge kicked in and the displayed SOC% went up.
Why did the inverter start showing power and the battery slowly discharge after the timed charge, when neither had been doing so before? I can possibly understand a discharge from 6% to 4% SOC but why did this not show as a battery discharge and why did it continue until an automatic charge kicked in?
Woodstock the 290w is a reserve maintenance charge.
If yyou set reserve at 20 and current SoC is 10. It will slow trickle charge to reserve and stop.
I've found the 20w ish always on, is the inverter own power needs. Probably too low to register on the BMS
TheDragon (GivEnergy) Not sure about the "inverter own power needs". The battery voltage curve was flat before my slow timed charge and the inverter was reading 1-2W during this time. It was only after the slow timed charge that the inverter started reading 23-27W and the battery voltage headed downwards.
Woodstock The 20w Inverter power, is now gone. Just loaded the 95% charging AC3 fix test FW.
Hybrid was done on Thursday as mentioned earlier
When Idle, inverter reads 0W now
Better, but it's still there!
After a week of testing battery capacity and low charge level behaviour I decided to see if I could engineer a sudden jump of the displayed SOC to 100% under 4077. Same as always I started by setting the conditions for a lower half displayed SOC adjustment to happen. The adjustment took place between 11:35 and 12:15 and the SOC adjusted downwards by about 15%. Next I left a timed charge to 100% going with the battery charge power at around 1000W.
At 18:35 .... Displayed SOC 91% ...... Battery Power 1032W .... Battery Voltage 55.42V
At 18:40 .... Displayed SOC 92% ...... Battery Power 696W ...... Battery Voltage 56.1V
At 18:45 .... Displayed SOC 92% ...... Battery Power 256W ...... Battery Voltage 55.03V
Sudden jump of displayed SOC%.
At 18:50 .... Displayed SOC 100% ... Battery Power 0W ........... Battery Voltage 54.19V
At 18:55 .... Displayed SOC 100% ... Battery Power 0W ........... Battery Voltage 53.55V
At 19:00 timed charge period ended and battery discharge commenced.
My take on this is as follows ...
It looks as though the system started ramping down the battery charge power based on the displayed SOC reaching 90%. This processing was then overtaken by the battery voltage hitting 56ishV which jolted it into a "stop everything I'm full" backstop state. This roughly mirrors what was happening under 3011, with the difference being that under 4077 the lower half displayed SOC% adjustment is now more accurate. As a result the displayed SOC% is not so far out and the sudden jump to 100% is smaller.
If the lower half displayed SOC% adjustment was spot on (or close enough) then the sudden jump would not occur (or be so small as to go unnoticed) but it appears that the adjustment is still setting the SOC% lower than it should.
Hopefully this can be rectified with some further tuning of the table lookup.
That is some good feedback. Have you emailed this level of detail? In that email ask for 4076. I dont know another after that
Is 4077 being pushed to the rest of the testers? I am still on 3077, so every fix accept capacity.
THALL It is what dan mentioned yes, when I asked about Woodstock. There is also 3076 and 4076, variants of 3077 & 4077
still 3077 here
Same here, Dan isn't replying about 4077. If you speak to him it would be interesting to hear what the plan is, and whether we've been forgotten about? :-)
There is very little value being a beta tester and reporting issues on 3077. We are on CE2225G411 if he fancies giving it a push.
I am down on so much capacity, I would be better off coming off the beta, if I can't be upgraded beyond 3077. Although, I think I read that my 290 inverter firmware wouldn't be compatible with 3008 anymore.
TheDragon (GivEnergy) Yes. The post on here is a close paraphrase of what I had already submitted as feedback.
Happy to test on 4076 if you think it will help.
Not sure of the reaction I would get asking Dan for 4076 when we had him working the other weekend to get me on 4077! Maybe I'll leave that one to you. 😄
Rather frustrated that I am still on 3077.
Also when will the BMS beta program complete? It has turned into a perpetual beta of us customers testing every new feature they seem to be throwing in. How do we get off the ride? Honestly I wouldn't mind that so much if GE were more reactive/engaged with feedback. Currently stuck on a combination of FW versions that puts me in a worse position than I seemed to be in before the beta.
JaseSpace I can totally relate...3088 for me.
I've requested updates four times now to the beta email - never once been even acknowledged let alone had anything done...
Well, I'm stuck on 3011 and think other than the increased responsiveness I'm worse off (due to missing the lower bit of my battery!).
Never had a reply to any of my beta emails.
I was happy with the idea of being temporarily worse off, and expecting issues all in the name of progress.
The fact that the capacity issues were largely addressed sometime ago, is all the more frustrating. I can't offer any value towards progress, if the build I'm 'testing' is no longer relevant to engineering.
I was put on the beta nearly 7 weeks ago, and other than a debug update to 3077 a month ago, I've had no updates. Annoyingly the 30% loss in capacity far outweighs any drop off in SOC I was occasionally experiencing.
The last few posts are exactly why I feel we dodged a bullet when we asked to go on the beta but ultimately weren't included. How the flying f@ck can you enlist the help of folk in this way and then completely f@sking ignore them? It's way past time that GE gave some proper, transparent update so we can make some decisions. Elsewhere the rumblings of legal action are building and understandably so. Come on GE, time to start communicating!
Has anyone considered legal action as a group? I am not on the social media sites to know what is going on.
I think legal action is the only way they will listen. Today I have taken 1st step and formally I have submitted a request for a fix (just a formality as my lawyer advised me) and a request for compensation. I have purchased a 9.5 kWh battery and currently at best I'm getting half of that capacity. I added to that distress caused, time and electricity wasted. If they refuse I will pursue this via small claims court.
Its disappointing its come to this. From what I can see engineering have made substantial progress, with 4077 created promptly after the expected delay of Chinese New Year, which has apparently gone some way to address the capacity issues.
That was 3 weeks ago now, with no clue on what the rollout schedule is bring the beta testers up to the latest build.
THALL More and more disenchanted users on various channels, some who know what the SoC issue is and have been expecting a fix, others who are naively posting "my battery just jumped" etc and are wondering why. As predicted months back, this is widespread and the tsunami is coming if GE can't actually resolve it - and quickly. I know they made another "coming soon" statement but is it really? Fingers crossed it's a sustainable fix for everyone's sake, including GE. Hopefully they really do realise the seriousness of the problem - lack of comms and slow progress doesn't fill me with confidence unfortunately. Have they even resourced it properly?
There have been a couple of minor tweaks, 4076, and today 4075 (not loaded it yet). 4075 is supposed to force an OCV at the bottom end, not sure if this means without a wait for idle or not. I will load, and see what happens.
As well as at least 4 Hybrid inverter tweaks and an AC3 one in the space of 2 weeks. I can assume getting updates this fast out is easy, Dan is but one man. Behind which is a team busy typing, testing and typing some more, to get it right.
Caveat - My Opinion.
What work for one, clearly isn't the same for all, different configurations, different battery structures/sizes and BMS hardware iterations. To get one version for all, I'm sure is a challenge. A good tweak for one, is a step back for another, so go back and try the same tweak a different way.
TheDragon (GivEnergy) Dan is but one man
For months now we've heard this through you and I feel sorry for him as an employee of GE. Why hasn't that become a team of people by now? It won't take many people to start legal action for this be seriously damaging for GE as a company. Today is the third time I've heard someone say they've consulted legal advice and taken appropriate action.
(This isn't direct at you @TheDragon (GivEnergy))
So, progress is put out, people not happy. no updates, people not happy.
Maybe I shouldn't waste my time
Ref not direct, It felt that way
TheDragon (GivEnergy) Please continue with your updates. Its the only update anyone seems to be getting.
If you do stop putting them out generally, just send to me them directly. I won't moan - promise!
TheDragon (GivEnergy) I'm pretty confident in saying that folk will be hugely appreciative of the vast amount of time and effort you've put in. Also the mystical "Dan". But it really shouldn't be just you and Dan (and a few others). GE should be all in on this, not just the firmware guys but also whoever is responsible for keeping the (increasingly unhappy) customers appraised. I had started to be hopeful that a fix was around the corner but the last couple of weeks dearth of information and appalling treatment of some of the beta testers has me really worried again.
TheDragon (GivEnergy) I think that's what I need OCV at bottom end. I'm not getting to the bottom of the battery! So to speak. 🤣
TheDragon (GivEnergy) you shouldn't have to do the updates but thank you for doing them.
Its all a bit one way for some of us beta testers. No replies to emails. Official updates sporadic. Etc.
Not ideal, but in the absence of official Comms any reliable info is better than nothing.
TheDragon (GivEnergy) Maybe I shouldn't waste my time
GE are abusing your good will. If you were to have a problem with anything you need not look any further than that. They should be embarrassed by the work, time, effort you have put in on their behalf. And they should be ashamed that they've let this situation go on for so long.
What are your symptoms, could I try and replicate?
I've tried keeping the battery over 50%, one pack gradually drifts away, always the middle of the 3, but is always brought back inline at OCV time late evening when I've used up all my power.
I can't remember what OCV is sorry?
TheDragon (GivEnergy) doesn't happen all the time, but this is what I see
All of the above leads me to believe it's not getting anywhere near empty, certainly not as much as pre beta.
It's an AC3 Gen 1 (5.2 Bat).
Charging overnight on the grid. Often to 100% but sometimes to 80/90% if solar and weather forecast looks promising.
TX200 here's one where it went from 77% to 100% full at 1555hrs. Think it was saving session day. 😀
TX200 the next day doesn't look too interesting, but the day after it seems to drain quite quick during later part of the evening.
positor1 Open Circuit Voltage
Its how the new method calibrates SoC vs voltage. Disconnects the battery from the Inverter (electronically) then samples the terminal voltage.
TX200
So larger swings in voltage, as its a single pack, I wonder if the charge and discharge rate is actually too high for the internal resistance of the cells.
A single pack is limited in power output, 60A from the Specs, which is 3kw. Maybe this is actually too generous, and should be more like the 2.6kwh, or somewhere in between.
I bet if you have a pair, to share the load, it wouldnt swing half as much with full load/charge
TheDragon (GivEnergy) wonder if GE would provide me with a second battery free of charge to mitigate the issue. 🤣
I tend to charge overnight at 1.2kw, to keep the temperature down and fit the four hour cheap slot.
But discharge, nah, not limiting that cuz of the oven, shower, coffee maker etc. Discharge is limited to 2.6 anyway on the 5.2 Bat.
2045 5% no longer powering house
2325 4%
0305 charging rate drops, nearly full
0315/0320 jumps from 84% to 100%
And yesterday's graph for reference (was charged to 90% early that morning)
TX200 I'm the opposite, I don't think I'm getting to the top of mine!!
TX200 Agree with all that
All I'd like to see is a reply to my requests for a newer beta with some explanation of why I'm not being given anything (since Jan 11th now)... Is that really so much to ask?
ianfretwell yes, it seems a little pointless to be raising issues about things that might have been fixed in a later beta.
I get that moving everyone to the latest beta immediately might be a bad idea, but try to do it as quickly as possible (test with 1, test with 10, test with 100, rollout).
For me on my system my AC3 is on the 290 firmware and my batteries on the 3077. The only issue I have is capacity missing and even then not a huge amount so no new issues to report to the beta. But I have requested 4077 to test if it does fix the missing capacity on my mixed battery setup. But no response to my requests, which might mean they need my system and others to keep running 3077 for data? No idea.
Basically nothing new to report on my current setup.
THALL I'm in the same boat, no issues to report other than the loss in capacity, and the drop at 100% straight down to 98%.
Both issues I understand have apparently been fixed, if I can see that against my system. Then it gets a pass from me!
I would be interested to see the findings of a few more people on 4077. For me it has NOT fixed the capacity issue. From 3011 to 4077 capacity went up from 6.5kWh to 6.9 kWh. Still well short of the 7.8-8kWh I was seeing under 3007.
Woodstock I'm still down about 15-20% down on 4077. I haven't seen any SOC jumps, but didn't before the beta either to be honest. I'm still seeing the issue with the charge not holding at 100% and dropping immediately to 99% and also the issue where after hitting 100% it won't charge on solar until it drops below 95% (an inverter restart will make it start charging straightaway).
ianfretwell OK, so I have tried again via the beta email thread:
---------------
Dear GivEnergy
I will again request that I am moved off of the current beta FW versions and on to a more recent (relevant) version.
The current FW versions that GivEnergy have installed on my system are not fit for purpose.
Regards
JaseSpace The issue here is you are already in the beta, and have no way of knowing if they have kept your system on 3077 for data gathering reasons. I am still on 3077 as well. But we agreed to be in the beta, effectivly giving permission for GE to test firmwares on our hardware. There is no way to know why the testers are now spread over 4 different firmwares or the schedule of testing.
However, you would be well within your rights to request to go back to the last general release version, ie come off the beta. If you ask for this they really should do it in a timely manner.
I agree. we cant know why we are staying on 3077, it could be for a good reason, it could be they are too busy. we just cant know and having agreed to be a beta tester I see no point in speculating. I am trusting all will be good in the end.
positor1 Oh I think it will be fixed. But the time involved, communicaitons issues. Maybe they could offer extended warranty to everyone affected? Post it on their website saying contact us at blah blah if you were affected and maybe get 2 years warranty extension free? Something that might never get used but covers any what if's from all this.
From Dragons update 3 weeks ago, I assumed there is no good reason other than 'Dan's schedule'
THALL Fair in theory, but seems to be in contradiction to the statement on 14/2/2023:
"We’re currently engaged in a final round of testing. This means we’ll be updating beta tester systems with the latest firmware versions for both the inverter and the battery. Then, after careful monitoring, we’ll look to roll the updates out to every customer"
If you are correct and they are still collecting data from old FW versions then this statement would be pretty misleading.
My guess is that, as ever, they are simply moving at glacial speed due to inadequate resource allocation. God only knows how they will be able to handle a general release. The fact that they are still commissioning new systems with faulty firmware means there is an ever growing list of people to update. If they DO have a team that can do this, why would they not involve them in updating and monitoring the beta testers NOW? If they DON'T have one it's gonna take a loooooong time. I was hoping to see a swift update of all beta testers to the latest firmware to instill some confidence, but obviously this has not happened.
On the bright side it does seem like the FW is there/nearly there ...
On a side note, surely, at some point the logical call has to be to commission new systems on the new FW rather than adding them at the back of the queue if it's working well enough for general release.
THALL I think comments about us agreeing to be a beta tester so we shouldn't grumble are not correct. This is what I received initially from GE after I emailed them offering my services as a beta tester:
----------
We are contacting you as you have reached out to be part of our BETA test for the new firmware logic for the battery SOC calculation issue. We would like you to take part in this and will need you to perform a software update on your system. Please provide feedback if you are OK to perform this, we will then provide instructions on how to do this.
Your system will be then closely monitored, and we will work closely with you for any issues or bugs that arise you find in an email discussion.
Please let us know if you wish to take part in this, and we thank you for opting to be an early adopter.
Kind Regards,
This effectively forms the T&Cs of me taking part (I had nothing else, whereas I think earlier adopters may have had NDAs etc). Literally none of what they said would happen, happened:
Some beta testers do appear to be getting the service indicated above. Many are not. In my opinion, the biggest issues associated with this beta program are not really the functional problems with the system, they are associated with the conduct of GE. There is lack of information/dialogue which only serves to breed the sort of negative comments we have been seeing. I am trying so hard not to be a GE "hater" but I've definitely fallen out of love.
positor1 Comments like "we can't know why" are wrong - there is most definitely a way of knowing! The original intent of the beta program was co-operation: "we will work closely with you for any issues or bugs that arise". GE have reneged on the original beta T&Cs (mine anyway).
JaseSpace
I actually agree with you. however it seems communication is not a strong point and I cant change that.
I do not want to put undue pressure which may result in a bodged fix,,, but again i do agree.
JaseSpace I agree with your points and yes as testers we be given more information than we have been. But in your case if you got a message from GE saying we need to keep your system on firmware **** because blah blah blah would you be happy to leave your system as it is?
I have only had comms with Dan when I have phoned to ask for changes.
I have been on various beta's for things in the past. Testing hardware, testing software each one had a plan, goals, deadlines. This beta is lets say questionable. My best example is my system got slapped with 3077 and GE support UK knew nothing about 3077.......
EDIT: Just to recap another point i am not currently a GE lover and have recently actively divertered potential customers away from its products and given all the reasons and evidance as to why. This may change in the future, hopefully near future. But at the moment I do not understand why anyone would order GE kit given the current issues. I know my installer will not currently install GE kit because of these issues. I expect other decent installers are doing the same.
I tried webchatting yesterday to ask if I could be moved to a newer beta to resolve my capacity issues.
Dan wasn't available. He was training people.
I asked the webchat person to ask Dan to email me. He said he would.
So far, no email received. Time will tell, but I suspect I'll have to chase it next week, perhaps by phone.
Considering how serious this issue is one would think they have a dedicated team working on this all the time.
But I guess they see it differently.
I've spoken to my installer and he is no longer installing their kit. I have also advised all people that I know considering installing similar system to stay away from GE.
The sad thing is, the people testing this are more likely to be evangelists for the company who support / promote and install the products, than end-users.
Opting not to adequately support the testers or at least spend 10 minutes to put out a statement to explain why that hasn't happened is perhaps not the best strategy. Some transparency would go a long way towards maintaining their faith at a difficult time, when people are questioning whether they will continue to endorse GivEnergy product.
Variation on a theme.
Battery firmware 4077.
It appears that the SOC sudden jump does not exclusively require the 56V limit to be hit.
If the displayed SOC% is far enough out relative to voltage (and the charging power is high enough) then the charging ramp down processing kicks in well below a displayed 90% SOC. In these circumstances the battery may consider itself full at a lower displayed SOC% and without the 56V limit being hit.
This, of course, results in a much bigger SOC% jump!
Woodstock I checked all my data on home assistant and can't find a single case where my battery voltage has ever hit 56V. Does that mean it's never truly been full?
Woodstock My battery doesn't even get to 55V, nevermind 56V 😃
Same here, I see these batteries hitting 55 or 56v. The most ive ever seen id 54.2
Lowest 51.2
Cells go between 3.14 and 3.41 as regular as clockwork.
How are these other getting so high or so low, unless the smaller the pack, the more extreme the voltage range.
judgepd I hit 55.25v @ 90% then voltage drops as charge continues. 3077 firmware. Can't see I have hit higher for a while now.
Battery firmware 4075 just arrived out of the blue. Anyone know what this one contains?
judgepd I think the 56V is a backstop to halt charging, similar to the old over voltage alarm but it doesn't generate a notification. In normal operation I would expect the battery to consider itself full and stop charging before getting this high.
Ra Neither does mine normally. Only when the displayed SOC% is well out of alignment with the battery voltage.
Woodstock Just hit my system too. Probably even GE don't know what it contains! 😄
Woodstock I have also just been updated to 4075 (and D0.199-A0.104). This was in response to my request (and whinge) this morning. No idea what it is meant to fix, but I can see straight away that the SoC is being kept at 100% so the annoying 95% hysteresis issue appears to have been resolved (but I suspect this is the inverter update not the BMS update). I'll keep an eye on my capacity over the next few days and see what has happened there.
Hmm I wonder why you guys are getting all the love.... My system was showing as offline for a bit so I wondered if I was being splat with it. But no 👎️
Weirdly, my overnight timed charge has been changed from 90% to 85% at the same time that the firmware was updated.
TheDragon (GivEnergy) Cells go between 3.14 and 3.41
I have a range of 3.03 to 3.48. The extremes are hit during charge/discharge where higher rates cause more extreme voltages:
judgepd And my charge power was reset to max too.
THALL No change here either (yet?)
4075 here too! The winging worked, inverter is still D0.290-A0.290 tho.
I am hated 😢
THALL Not sure about the others, but you have dual batteries...could be a reason maybe...who knows.
It looks like it happened at 14:35, and my SOC jumped from 71% to 74% at that point.
judgepd Yeah. Only messing. Might be because my system is awesome and have has no problems (apart from the 98% drop and missing total capacity) 😎
@Jamesrich ah it might be China have blatted that out directly. When I got hit with 3077 it was around that time.
4075 and D0.290-A0.290. Was already stuck in the 95% hysteresis loop when it was applied. Will see if that improves.
judgepd Checked my settings... They seem to be as I had them previously (assuming the portal is reporting the actual settings in the registers).
JaseSpace Used Remote Control page to read out the register settings. One anomaly - AC charge upper % limit was actually set to 94% not the 100% reported on the main settings page. So it is definitely worth doing the manual readout step as a check!
kai It hasn't fixed it for me. Battery was at 100% used a bit to drop it to 99% and solar is now exporting and not charging the battery.
Spotted 4075 coming in (although I didn't know that was what it was) when I went to change a setting and couldn't because the inverter was offline. Tried several times without success to get the inverter back online so went out to the garage to see what state it was in. Talk about Blackpool illuminations, the battery had 4 charge lights lit and a flashing status light. Simultaneously the inverter centre circle was alternating red and green and the dongle was flashing ten to the dozen. Add in a cacophony of clicks and buzzes and my first reaction was to hit the power off button! Luckily I resisted and it settled down after a couple of minutes with the dongle eventually going solid when connection was re-established.
I am guessing this is normal behaviour for a firmware update arriving and I have just never been in the vicinity of the inverter before! 😄
AC3 plus 2 off 8.19 "100%" DoD Changed from 4077 to 4075 today.
Woodstock
4075 is 4077, with some low end OCV tweaks
THALL Its not China 😉
Leave it as I mentioned it in conversation, that there were many differing version out in the wild, and it could use some one to re-align them all the same.
judgepd First thing I did was a "Set to factory defaults" (includes an inverter restart) and the battery did a ramped charge of the remaining few percent to 100. Stayed at 100% for 10 minutes then dropped to 99% where it remains. Plenty of excess solar PV throughout. Doh!
Just restarted the inverter and the SOC has gone back up to 100% with a 421W/145W ramped charge. Seems to be operating as previously but I don't know if a supposed fix for this is included in 4075!
TheDragon (GivEnergy) With the aim of increasing capacity?
4075 has turned up for me this afternoon
THALL don’t worry, I’ve not received it either.
Woodstock Stopping the odd jumps down some are seeing. Not me, its done nothing to mine, exact same. So who ever had low end drops will be the intended target.
The 5% and above not charging is fixed in Inverter DSP firmware. Both AC3 (D0-289,A0-289) and Hybrid D0-199, A0-104
My experience is if it's operating nicely, it tops out at 54.5V. If the calibration is out, then the voltage can go right up. This was today's reset (on solar) note the jump Vs a normal charge a few days ago.
The jump was because it's been reading 90% instead of 100% at full for some time. Other than that well alligned.
4077
Grid to battery - 5.99 kwh in last night, ontop of the 32% I had remaining - which sort of indicates there has been an improvement in capacity. Previously I saw around 7.5kwh 'grid to battery' a night, on my usually empty 9.5kwh battery.
One thing I noted was the immediate drop 100 > 98% is still present. Did i understand correctly this was meant to be fixed?
I'm now on 4075, but still on the 290 firmware if thats relevant.
But that’s the problem - you can’t just add 5.99kwh onto 32% because you don’t know if 32% is the correct SOC.
I was on battery firmware 3009 and reported lack of capacity which GE confirmed they could see and downgraded me to 3007.
Gen 2 inverter and 9.5kWh battery and at times I’m storing under 6 kWh.
I now have 4075. I will do a complete empty and charge and see what happens over the weekend.
TheDragon (GivEnergy) I'm still on D0.290-A0.290 so that would explain it. Thanks.
First overnight forced charge from "empty" (SoC reported at 3% at the time due to leakage down from reserve, which still seems to be happening at a higher rate than pre-beta) to "full" (SoC reported at 100%) on my new FW (4075, D0.199-A0.104) and I DO see a noticeable improvement. In the past week (all "empty" to "full") my Grid to Battery kWh figures are reported as: 3.43, 3.38, 3.65, 3.37, 2.9, 3.51, 3.43, then today (with new FW) 4.11kWh. This is a 9 month old 5.2kWh 80% DoD battery, so I think the max is approx 4.16kWh. I'm happy with 4.11kWh. The problem is that it has been jumping about all over the place, so I'll keep monitoring it.
JaseSpace First impressions of latest FW 4075, D0.199-A0.104 on my Gen1 Hy 3.6 with 5.2kWh (80%) in relation to my previous 4 issues:
1) SoC Jumps: too early to tell on this version, but not seen anything for a while on previous version (3077)
2) Charge Stalls: This appears to have been resolved as I now see the SoC stay at 100% when there is excess generation and you can see there is charge going in to account for any discharge due to load.
3) SoC Leakage: Still appears to be happening. E.g. 5% to 4% in 30mins, 4% to 3% in 20mins, with no recorded discharge in that time. May be a function of the new SoC recalculation logic as opposed to true leakage?
4) Capacity Reduction: From a sample of 1 forced charge so far, I seem to have charged the battery close to the max expected (4.11kWh). This is a noticeable improvement.
But something else I have noticed (not brand new as was seeing it before latest update). During my overnight forced charge the charge power starts at 1730W (at 3%, 14C) and slowly ramps to 1770W (at 66%, 20C) then a step change to 2400W (until 91%, 23C), then a ramp down of charge power until full. My settings have charge power set to max (2600W). The battery was not particularly cold at any point (min of 14C) so why am I not getting my requested 2600W charge power at any point? I'm going to call this my issue 5).
Another anomaly I had last night. I had my max AC charge % limit set to 94%, but it charged all the way to 100%. I think someone else reported this recently. I'm going to call this my issue 6).
I'm sure they went to the same school of promises as my installer.........🙃
Woodstock 289 and you remin at 100%.
Seems Dan and Dan have been busy this morning
TX200 Any chnage for you yet - nothing for me still 290 & 3088.... 🙁
Hmm. New behaviour on 4075. What is going on here then:
It looks like the battery is autonomously deciding to pause charge and do some periodic dumps when at 100%. Doesn't appear to correspond to an increase in load (though there are 2 examples of that later to the right). Unless I am missing something obvious?
GeorgeRawding I can only apologise, I had been lead to believe you were getting regular comms via email as part of the beta group. I can see that isn't the case. I'll chase up on this and we'll get that rectified.
Another thing that hasn't happened.
JaseSpace I am currently charging at max speed but will need another couple of hours. But at 69% I am not see those spikes on 4075. But I will look for them later.
ianfretwell nope, not yet.
@TheDragon (GivEnergy) You might not notice this as your batteries are the same size - but do you see a difference in reported SoC for the different batteries?
My 8.2 and 5.2 report differences that seem to correlate with the relative sizing. So last night my 8.2 was claiming 43%, while the 5.2 was at 77%.
I'm not seeing any capacity loss from expected across the system as a whole, so think this is purely a reporting anomaly from different sized batteries. Just curious.
JaseSpace I am not seeing spikes but rather a constant discharge of 15-19W. Does not show on the battery data but you can see it on the inverter power. The inverter never goes into an idle state even though the battery is "full".
kai 9.5kw reporting 71% charge my 8.2kw reporting 82% charge. Both at 53.9v, 14.9c - 15.6c range 4075 firmware
That makes sense. The system is using voltage to report charge state, but voltage has no information about capacity. So the individual reported SoC, based on voltage, seems to scale for capacity, and overall is just an average.
JaseSpace The spikes roughly match missing data points on the graph?
Woodstock I thought that at first, but I have a real time power flow app running too, and I can see the discharge-to-export happening at 100% SoC on that, so I don't believe it is a portal display funny. It's as if I've filled above the L-ion overflow "pipe" in my battery!
JaseSpace , it could be cell balancing in the battery as at 100% SoC.
What does the battery Voltage chart indicate for the same period of time?
Certainly BEV chargers do this at 100% SoC to equalise the Voltage on each individual battery cell, hence the common BEV recommendation to charge to 100% say once every month.
Simon_C 
The spikes on the graph are at 10:15, 10:45 and 11:10 and the battery data timestamps don't quite align with that. The table clearly shows the exports around similar times, and in each case this is just after the battery voltage is reported as >53V. So maybe it is a voltage related event?
JaseSpace Interesting that the charging power after the first highlighted spike suggests the use of pre-set ramp down values while that before doesn't. It's as if something has caused a change of mode. Does the battery voltage curve give any clues?
Woodstock Battery voltage curve? Where do I get that?
JaseSpace Portal Power Graph tile. Click on "(+3 others)" and select "Battery Voltage".
Woodstock Never spotted that before, thanks! Here you go:
JaseSpace Do you have the 95% charging bug fix firmware? Hybrid D0-199, A0-499 or AC3 D0289,A0289?
This stops odd behaviour above 95%
kai I sure did the master, was always highest, followed by 2 and then 3.
As they are the same size, I did some testing of my own, ref voltage drop across the link cables. Aswell as looking what others such as pylon tech do for multiple packs.
They wire them in a ring, DC cables fed from both ends, like your sockets at home are. 1 MCB, 2 cables around a ring, giving the power 2 ways to go, reducing the current in one cable, to half and all equal.
This format changed the Soc to 1 and 3 equal and 2 a touch lower
So I then looked deeper, and looked at impedence of each battery to the inverter, and calculated that battery 2 had a slightly higher impedence. So I got authority to try another wire style, Star.
The inverter to MCB, MCB to a DC BusBar, then from here to each battery with identical sized cables. All 3 now equal , only differences now are due to actual battery pack manufacutured differences/tolerances/cell capacity.
This method is good for 3 or more EQUAL packs. Any other make up, it has no befefit.
Some installers are interested in this design, Giv dont see an issue if done correctly.
The actual overall capacity of the packs, probably no different, but at least all 3 packs are treated equally
Latest Beta Firmware versions
AC3
D0-289, A0-289
BMS 4075
Hybrid
D0-199, A0-499
BMS 4075
TheDragon (GivEnergy) Yesterday I was moved to the latest FW (4075, D0.199-A0.104) - this solves the 95-100% hysteresis issue (charge stall) but now I have the discharge->export feature instead.
JaseSpace , I am still getting my head around import/export and charge/discharge above and below the 0 W line.
The 6 spikes appear to be with 100% SoC and solar gen > house load, and are battery discharges that are increasing the size of grid export above that explained by the solar excess during each 5 minute period.
I would have expected to see the grid export at these times, but not the battery discharge. If they were triggered by a spike in the house load, you would expect to see that captured by the green load power line.
@joshgreenuk
Do you see this?
TheDragon (GivEnergy) I have been on 199 for around a week.
joshgreenuk But do you see this pulsing export if full?
TheDragon (GivEnergy) I havent yet applied 4095, but when set to 75% charge it charged to 100% and then sat there until the charge window was over, then it discharged to the house and I had a full day of solar so exported all day with battery full.
Will look out for it, main concern is my PV issue persists, strings still not independent. Dropped to 60w production yesterday at 4pm from 1.3kw on string 1 because string 2 went under 135v. this means everyone with east west split will have this issue.
joshgreenuk Looking back I don't see any export at all except from yesterday.
joshgreenuk Yes, that is still on the list of outstanding
What is this strings issue please?
This would be complete show stopper if the strings don’t work independently for anyone with an east west.
Has anyone noticed excess import and export has got a bit spikey since upgrade to 4075?
Just like pre-beta when the reaction time was slower?
It's not all the time, just for an hour here and there. Pretty easy to spot on the graphs.
3077
4075
4075. Got to 7%, but only discharging at an apparent constant 94w draw from the battery. The rest is coming from the grid.
Anyone else seen this yet? I was trying to empty it before tonight, but it may not happen at this rate.
kram if I remember correctly the strings are wired together - i have a south and west split system on separate strings into the inverter and are totally independent. Actually the voltage does tend to be similar first thing but the power from each varies a lot.
kram , the strings issue is specific to @joshgreenuk , who has Tigo optimisers on his panels, and has only come to light since going on the beta firmware trial. As far as I know, no-one else has reported this issue.
See much further back within this thread.
https://community.givenergy.cloud/d/1342-new-givenergy-inverter-bms-firmware-beta-testing-chat/348
kai Not on beta, but following the thread.
I have an AC30, two 5.2KWH Gen 1s, 3007 FW. Ever since I became aware of the function to get 'More information' I have monitored my batteries.
The stated SoC on the main card is more closely aligned to the Primary and is never equidistant between the two.
As I have never run the stated SoC below 15% in the 8 months I have had them, I have not observed the bottom end shenanigans.
hhall Thanks! I'd seen similar before the beta but had put it down to the failing battery that I had replaced after joining. @TheDragon (GivEnergy) explained what was happening.
The beta is working well for me. Only thing I've noticed wrong was the 95-100% issue which is already addressed by a inverter firmware. It did require a calibration and then an emergency charge at 290W to get the capacity behaving right though. Since then it has been solid.
kai yes inverter FW addressed that both Hybrid and AC3
TheDragon (GivEnergy) Any idea when they are going to roll out that inverter FW update more widely?
kai I believe it will get the rest of the minor issues rolled in 1st.
But I could be wrong. I would roll it out to beta as it's available.
But I'm not running the show
Obviously I jinxed it. Battery down to 15% reported SoC, and discharge reduced to 100W. Battery voltage at 48.8V. Expecting a drop and emergency charge soon.
'Control': { 'Battery_Charge_Rate': 100,
'Battery_Discharge_Rate': 100,Not sure where these have been set from
Jamesrich looks like I'm seeing that too
4% to 100% yesterday morning: 3.85kWh into battery from grid. (Nothing from solar).
At 4% by the end of the day: 2.54 kWh out
Back to 100% this morning: 3.03kWh into battery
(Figures from the table on the app's energy graph).
Oh and 78% to 100% jump at end of charging
TX200 On yesterday's power graph do you see a steeper section in the discharge curve (lasting about 40 minutes) in the bit below 50% SOC? The battery power will have been less than 200W at the start of the steeper section.
If so, you have experienced a lower half displayed SOC% adjustment. This is what has been causing my SOC sudden jumps to 100%. The adjustment gets the table look up wrong and places the displayed SOC% lower than in should be. You will have stopped discharging sooner than you should have, which is why you took less power to get back to 100%.
joshgreenuk Will look out for it, main concern is my PV issue persists, strings still not independent. Dropped to 60w production yesterday at 4pm from 1.3kw on string 1 because string 2 went under 135v. this means everyone with east west split will have this issue
Josh is not the only one seeing this………
I am seeing a similar issue on a GEN2 5.0kW hybrid(D0.906-A0.906) the MPPT inputs appear
to oscillate at low generation (<200W) both at sunrise and sunset.
This never happened with a GEN1 at the same location (inverter swap from Gen1 to Gen2) with the exact same solar array with Tigo optimisers.
Need to log a bit more data to see if I can help figure this one out.
Woodstock 
Did you mean the battery graph? Power graph doesn't seem to show much of interest.
42% to 17% over 55 minutes. Mostly around 115-120W during that time.
1530&1600 hrs half of the house's needs were from solar half from battery. No solar in the 1630 period or onwards.
dhb666 , I was just trying to clarify that the statement by @joshgreenuk was not totally accurate.
Jjoshgreenuk
2 days agoTheDragon I havent yet applied 4095, but when set to 75% charge it charged to 100% and then sat there until the charge window was over, then it discharged to the house and I had a full day of solar so exported all day with battery full.
Will look out for it, main concern is my PV issue persists, strings still not independent. Dropped to 60w production yesterday at 4pm from 1.3kw on string 1 because string 2 went under 135v. this means everyone with east west split will have this issue.
The only 2 instances I have come across with questionable string performance both involve Tigo optimisers.
@joshgreenuk was stating that everyone with East/West split would be having the problem. It does not appear to be the case when there are no Tigo optimisers involved.
4075 / D0.191/A0.104 firmware. SOC drop from 20% and still have the 100-98% charge issue.
TX200 That's it! Exactly what I am seeing. The reason most people don't get this is their battery discharge is too great (>4A or 200ishW) so the lower half adjustment never takes place.
The lower half adjustment sets the displayed SOC% too low and this then gets corrected by either ...
A) A prolonged idle period at (supposed) full discharge when the displayed SOC% will go through a series of correcting humps and discharges.
or (if the idle period is not long enough or the battery starts charging again before 4% is reached)
B) A sudden jump to 100% as the battery reaches a true full (you will see the battery voltage reaches a higher value of 55V plus when this happens).
I believe the lower half adjustment is there to take account of people (like me) who very rarely discharge down to reserve level. Unfortunately instead of correcting any error in the displayed SOC% it is currently introducing one!
Reported this first under 3011.
Woodstock so you still getting this on 4075?
I never see the readjustment at "empty" - even if it sits doing nothing for hours, never re-adjusts. Only corrects itself when charging to "full".
I might ask to come off the beta. I was better off on the live firmware, even with the odd soc crash. Maybe if there was more two way communication...
(I'm still on 3011)
@"TX200"#p22413Was still doing it under 4077. Test has been running on 4075 since 02:00 this morning. Watch this space. 🙂
Ra D0-199 sorts the 95% above blips and odd behaviour
Woodstock 4075 was to address the low SoC adjustment.
So interestingly waiting for your results.
TX200 Correction at "empty" was only introduced in 4077 so you won't have seen this behaviour.
Also off interest, there is an aspiration to have a beta portal, where new beta will be posted, with documentation. You can then apply at will.
I stress it's an idea being discussed to better manage beta trials. As going forward there will be many more as new kit comes to fruition in coming months.
There maybe more changes coming too. But those are under my hat for now
Woodstock so I've got the bug, but no workaround or fix. 🤣
Ra I haven't seen a case of displayed SOC% over-reading (sudden drop rather than sudden rise) since being on beta. Am I seeing your battery discharge ramp down as displayed 4% SOC was approached?
I am on 4075 too (although AC not hybrid) so will look out for this. I haven't yet been down to 4% SOC on 4075.
Woodstock I haven't noticed any ramp down, although I guess if it drops from 20% I won't see any as it thinks it still has plenty of juice to go. I have NEVER seen a SOC jump since my system was installed, only ever seen drops.
I read of people talking about 55-56V readings on here but mine never sees that, even when fully charged/charging. My capacity is slighty down on what it should be and I believe that if my system charged to the higher voltages I would achieve that, but as it stands I'm still seeing the same issue.
I will add the issue is slightly reduced, but whether thats because of my charging pattern in the winter rather than the summer, I don't know.
Thought I'd just post some pictures of 4075 and AC3 289,289
Combination.
The trace shows a calibration run, and a days normal either side.
You can see how little extra charge is in the 2 extremes of voltage. Maybe 1.5kwh in those 2 tips
Ra Same for me last night. I see your discharge rate dropped to 100W as well just before the SoC reset downwards. I'd been down around 10% on earlier beta versions and not seen this behaviour.
kai The discharge rate dropping was the reason for my question. I did not know whether it was a manual battery power adjustment or a system generated ramp down. Would be interesting to know which and what the power figure was (121W?).
Bang on 100W for me. Not manual. I hadn't touched the settings and after it dropped a reread all the registers thinking some setting had changed.
Givtcp was reporting it:
'Control': { 'Battery_Charge_Rate': 100,
'Battery_Discharge_Rate': 100,
edit: these might be % and I'm just seeing the 100.
This lasted for 90mins with 100W discharge, during which my reported SoC was falling really fast. Then my cheap period kicked in at full pelt. Got to 84% overnight, about right for my batteries from reserve.
Looking at @Ra's image it looks like his system did the same just beforehand.
kai Wow some adjustment there. I wonder if you will get a sudden displayed SOC jump to 100% on charging towards the top end. I am still discharging and testing paused/slow discharge scenarios so I have not yet got near to a displayed 4%. I await with interest to see what happens. 😀
Woodstock I can second this behaviour. My overnight baseload is typically under 200W, and soon as the batteries hit 40% one (of the 2) would start a fast soc adjustment downwards, leaving around a 30% split between the packs. Next charge it would then correct with a jump. The pack it did this could be either master, or slave (but not normally both at same time). Seen this (and reported) on most of the releases, except 4075 (so far). It's a good observation it needs the low load, I suspected this, but never was able to confirm it.
4076 was interesting in the fact it was the first time I saw it on both batteries (ignore the spike at 00.00, I had a random 1 min discharge set)
Fingers crossed 4075 is the answer
mpartington It needs a low battery discharge current (<4A), the displayed SOC to be less than 50%, to be more than 24 hours since the last adjustment and (possibly) the battery voltage to be below a certain threshold level. Haven't seen it yet on 4075 but others are seeing some new weird behaviour towards the end of discharge. Hope we haven't swapped one problem for another!
kai you updated GivTCP by any chance?
Think it’s swapped % for power now.
So you’re probably now telling it to use 100W not 100%
hoggy I'm only using givtcp to read, not write. Unless it is self-updating on the docker container then nothing should have changed in terms of version.
The system is behaving normally since it charged, full discharge rate. I think I was just misreading the 100 as W rather than %.
Woodstock guess I'll need to try to use a bit more electric in the evenings 😁
Simon_C no, loads of people on Facebook with no optimisers but on gen 2 have the SAME issue and it’s not to do with the tigos now it’s the trackers dropping both strings if conditions are met to ramp down either string. They have been playing with the voltage of shut down and start up but the problem is my north east panels will go down to 130v (4 panels) and shutdown my 7 in south - south west whilst they are generating 1.5kw+ still.
Tigo or no Tigo it’s not related anymore and is a fault in the new gen 2 firmware and beta to which I believe gen 2 is alpha version beta as they have it as bad as the first beta I had.
It’s mostly masked now, but still present and will be a issue until the trackers are independent again.
I can resolve the issue by joining the strings and having just one.
Simon_C I am correct, that is happening, removing tigos would have the same effect, whenever one string decides it needs to change the other follows suit.
The issue now is the way it’s masked is you can go from 3.5-4kw to 600w to 3.5-4kw with no change in weather conditions due to the strings adjusting each other, again fixable by joining strings.
The issue is my array orientation, and I am seeing at least 10 people with the issue, loads without optimisers but on gen 2 (alpha beta)
joshgreenuk I have gone through at least 10 variations of firmware to fix this and each time giv CN have said the trackers are now independent but they are not.
The fix is to change MPPT to 120v but I still see other string or both drop if one string is above 70v but under 135v
I know of someone else on gen 1 with beta having identical issue if not worse than me who has had to join strings to stop the issue
Sorry to post so much, one point they had it stable throughout the day but evening and morning dropped each other string, but now it’s a mix of both where the day just is all over the shop but mostly ramped high. It’s not a fix, I would run 1 string but potential to be over 600vdc at -15c
When it’s warm I join the strings and I am at 540voc in the morning which is then flawless all day into the evening.
Not Tigo related, just enhancing the voltage changes.
Trackers are no longer independent and appears that there is no intention to separate them again otherwise they would have after saying so many times they have
The low power under 4A I've mentioned lots of time could be an issue for some very low users, seeing a SoC adjustment.
As there is a small load, the voltage is slightly lower than normal, so the SoC would drop downwards.
I would want this under 2A per battery
I was seeing the opposite, when using reserve, a slow charge at 294w was triggering OCV upwards on 3 or more pack users.
Hence 2A or lower would be better, to have a threshold for OCV
-
I will re iterate this point to them and that it's not just me.
This is what I submitted.
Feel free to add your own and link it to my email.
Hi Dan,
Remeber a while back we were looking at OCV kicking in while reserve charging.
It was never fixed.well 2 users are seeing the opposite, very low users, under 200w load.
The OCV once SoC is under 50% is kicking in, as there is a load on the battery, the voltage is lower than it would be at idle.
So the SoC is actually being pulled artificially low. It can be seen on their submissions.
Now the SoC is artificially low, next time they charge near to 100%, it hits the upper limit and jumps.
So the OCV is actually causing issues, not resolving in these low load, low charge cases.
I would recommend OCV criteria.
Over 24hr since last
Under 50%
BMS under 1A.
Those should fix the issue.
TheDragon (GivEnergy) cheers
Based on the drop last night, it looks like it was loosing 2% every five minutes (<200W usage)
I’m currently using circa 245W and seeing the percentage drop 2% over 20 minutes (more believable!)
TheDragon (GivEnergy) Setting the battery discharge current that low will effectively mean that an adjustment can only take place when the battery is fully discharged and idle (or has been manually paused). Neither is likely to occur where the pattern of usage does not see a low level of discharge for months at a time. I take the fact that the displayed SOC% adjustment takes into account any actual charge/discharge during the adjustment period to show that the intention was for the adjustment to take place during periods of battery activity. If this is not to be the case can the "battery full" processing alone be relied upon to keep the relationship between displayed and actual SOC percentages in line?
I thought this new algorithm was supposed to be extremely clever and working out the percentages from multiple factors...
In its current state its not exactly ideal. Although I guess if the workaround, when it gets to empty, is to realise there's still charge left and let you use it - well that's better than nothing, although it would depend on how long it needs to be idle for (e.g. you might not need the extra capacity it finds by the time it's found it!)
Woodstock
Or they add a % to the voltage to take into account its still under some load.
As the OCV expects no load, and the voltage stable.
If you disconnect the pack via MOSFET, sample and reconnect the voltage read will be low compared to an idle pack.
The difference error would be greater the smaller the pack
It’s slightly reassuring, but also troubling if they think this sort of thing will need to run for a while.
They really need to make an official release plan and date statement asap.
joshgreenuk , somewhat concerning if it is a wider issue, though from what you have stated before, the East/West strings worked ok before going on the beta firmware. There were only 2 mentions of the issue on this forum, both with Tigo optimisers, hence the deduction.
I am having a temporary non-battery inverter swapped out for a HY 5 kW Gen 2 with 9.5 kWh battery this week, and have an East - 6 x 410W and West - 10 x 410W split, so will stick with the general release firmware until I see the issue is fixed. Even with potential reduced battery capacity from the SoC issues, it will be better than reduced solar generation, and my current setup with no battery storage.
Lets see how well the startup works in the morning compared with the current inverter.
Simon_C the gen 2 are the ones having the issue along with beta users thus I think gen 2 is running early beta version as general release.
It’s when one or the other string drops / isn’t at 135v (on the latest beta) previously it was 150v which meant string 1 (east) was just under start up all day where string 2 ( south ) was in sun with no start up.
They’re fixing it, but tbh the only solution is separating the trackers again, I don’t know why they are reluctant.
People on Fb don’t post on here, when I see a post about it it’s always gen 2 as not many gen 1 are on the beta.
I only have 4 panels on 1 and 7 on the other, meaning string 1 is easy to drop under 135v.
Since my Tigo post I found a lot more info out on it and have been working with giv on the issue. At one point I saw multiple firmwares in a week. I’ve been on this version for 2 weeks or so now awaiting the next (hopefully separating them).
The other issue is gen 2 duel are cross charging, this has also been introduced to gen 1 which don’t suffer this via beta. Thus this all points to gen 2 running beta as general.
From the 1 picture of have of inside Gen 2, Unless it’s hidden somewhere I believe Gen2 doesn’t have the isolation transformer like Gen1 on the DC side. I’m not sure the consequence of this in regards to MPPT separation. I believe it was done this way to get around the busbar limitation on Gen1 holding battery charge rates back.
G2’s controller PCB is atleast 4 revisions beyond my own which may also introduce differences.
(if someone happens to be able to get hold of the IEC report for Gen 2 it’s useful to suss out how it all works, other than me going to their test provider Intertek for it I’m yet to find it unlike Gen1 which is on the ENA supporting docs)
Edit: basically this is my long way of saying if G2 has always struggled with E/W split (I’ve not got the time to look through thousands of posts) then perhaps they aren’t actually separate MPPT inputs.
Keeping the usage above 200W yesterday eve, no rapid drop off of the battery percentage.
2055 usage went down to 94W and was at 12%
Similar wattage till:
2205 usage went up to 182W and was at 9%
Similar wattage till:
2250 ran out (6%)
3.9kWh into the battery.
Woodstock Results are in. No sudden jump at top end.
Last night went down to 5% without any drop or having to have an emergency recharge. I'd suggest this is probably because it has been less than 24hrs since the last one so it hasn't got itself out of kilter.
Still have an issue with capacity though. Missing 10%
4075 - Mine didn't flatten itself to 0% last night and need an emergency charge. I also got 7.8kwh out of it, as opposed to the 6.8 I'd seen on 3077. (Single 9.5 AC3).
I put in slightly more than the day before this morning off peak, 8.55 vs 8.34, although it didn't fill past 98%.
This is all an improvement on 3077 where I was charging about 7.5 per night from 4%. I'd say about a 13% improvement all in.
Pre-beta an empty battery would put around 10kwh in, and discharge around 9kwh out. So still someway to go, may request a calibration.
kai YAY.
Capacity vs 3007/8?
My 5.2s are 12.3kwh right now, instead of the 13.4 before. Now 78% DoD, before i was nearer 90%
Jamesrich Do you have the updated Inverter FW 289-289? This resolved all my over 95% issues, it remains nailed to 100% now till its ready to discharge.
TheDragon (GivEnergy) Not yet, still on D0.290-A0.290. Glad to hear that's the issue, I did think I'd heard that it was fixed somewhere / somehow but wasn't sure.
I guess this is part of the noise from us all on different versions, reporting issues that are already fixed.
hoggy They worked fine separately prior to beta, my Gen 1 now acts like Gen 2's, thus I believe its in the software. My wild guess is that the calculations are demanding and this is a way of optimising the performance freeing up processing for other things such as the faster switching.
Gen 2's issues are identical to Gen 1's, but worse, almost like I keep saying, they are running a very early beta as their general release.
I will chase again this week to see if there is any movement forward, the gap between updates signals to me that they are either stuck on a resolution or the resolution isn't a quick change and that part is being rewritten (more likely I think).
TheDragon (GivEnergy) Capacity seems down a bit but hard to compare as when I had 3008 I had a battery with a crap cell. Energy graph reckons 9.4kWh discharge from 100->4, which is definitely low. I should be 11.8kWh or thereabouts.
Not sure if this is new or previously reported, but I'm on 290 - 4075 and have paused my battery and it's still charging on solar at 78W. I've tried manually setting the charge power to 0W and it's still charging at 78W.
Here's the upper end plot.
Hi, that's on my list already. Mine is 121w, its a function of the number of cells
@Jamesrich It is getting better, I have the ears of the right people
That nipple at 98/100 goes away with 289/289 and 4075
joshgreenuk ah so it used to be fine, that’s fair enough. It “should” be addressable although as we’ve seen it is a bit like whack-a-mole in this things brain - will likely break something else!
I do wonder if some parts of the firmware are more open to quick changes than others - on top of all this headache with SOC they are likely working inside quite a rigid box of compliance with the likes of ENA/IEC - a step too far would potentially force them into compliance retests should the operating principle be changed significantly from what was originally tested and signed off.
hoggy I've asked if there is any updates coming this week and if the Gen 2 are using a beta version as their general release as the same identical issues are now present between gen 1 and gen 2 the only difference is the charge rate now (which IMO can be raised, as calibration is at 3kw)
joshgreenuk let me know how you get on (I gain nothing from all this beyond just been generally nosey about how stuff works)
hoggy Will do, doubt they will answer it, I have asked quite direct questions that haven't been answered 🙂 I kind of know they wont be but it doesnt hurt asking lol.
IMO I think the tracker side effect is an optimisation to free up the processing power for other features like the faster switching, but I have no basis of this except being a developer and it kind of making sense, That and I have asked that a lot and I am not getting a 'no'. 🙂
At least we are still moving forward, and to be honest, probably quite quickly if you look at the broader picture of how different everything seems to be.
joshgreenuk We just need that bit of capacity back from the bottom end. 0.15v for me low idle. 3007 was 0.15v lower. Dont sound a lot, but right on the knee of the curve, can make a difference.
I want it on the downwards part, not the top
So the east/west array thing. Is that a problem without optimisers as we are east/west? at that point it makes sense to look elsewhere.
Frustrating as each step forward seems to be two back and the official communication is woeful.
Its in work, Josh is pushing on this
D0.290-A0.290 4075 AC-3.0 100% DOD 8.2kWh
Energy for free!
Anyone else seeing load power constantly in excess of grid and/or battery power by around 25W?
Also the system data is showing that I now have inverter gains rather than inverter losses!
When discharging; the inverter is apparently outputting more power than is being taken from the battery.
When charging; the inverter is taking in less power than is output to the battery. 🤔
Woodstock No, not seeing that. I see correct AC higher on charge and lower on discharge than DC.
Well using GivTCP anyway. Gonaa go check the Portal now.
Edit. Nope just you. Lucky one
Grid ACp DCp load SoC
14:58:53 101 -988 2948 3090 4037 27
14:53:51 21 -969 2943 3091 3933 30
14:48:49 8 70 823 858 761 31
14:43:47 105 6 750 781 849 31
14:38:45 252 5 482 502 729 32
kram
I've just had my Gen2 inverters retro fitted today(from gen1) - I'll let you know once commissioning has finished. My is slightly different though, I have one inverter with one string only East and one inverter for West ( all with Tigo's).
94jdh must be big strings then.
Why not one invertor?
@Woodstock
AC vs DC right now, discharge.
Inverter Firmware Version
D0.290-A0.290
Battery Firmware Version
3077
I thought soc jumps were gone with 3077 beta. 94 to 100 happened today. do i bother reporting it?
I’m looking at adding about 3.5kW east and 3.5kW west, each on their own string and unoptimised.
I can still change the inverter and battery at a cost and it’s just another thing that’s going to cause GE to be deselected.
It’s hard, because they’re working on it, but this would be a complete blocker. Looking forward to seeing if you have issues 👍
94jdh Some good news at least. I'm watching closely to see how you get on before I push for the retrofit again 🙂 This firmware issue is a right PITA. My two 9.5s are currently 26% apart on SoC and getting worse.
TheDragon (GivEnergy) benefit of a higher discharge rate when the house demands it.
kram
Similar size to me, just over 7kW in total equally spread between the two - just awaiting the commissioning again, so hopefully all good and fingers crossed one string eliminating the problem with Tigo's.
Yes, getting there - although still a few outstanding issues stil need addressing with C rated instead of B rated breakers and earth bonding etc.
Haven't posted for a while, I had 4075 applied a few days ago so I thought I would run a control chart on the notional battery capacity and wow; what a difference:
You can see the slight increase in capacity from 3011 (8.4 kWh) to 3077 (8.7 kWh) then the dramatic increase from 3077 to 4075. Only 4 days data so too early to draw conclusions but encouraging. BTW I have a 9.5 kWh battery, new mean total capacity would be 10kWh if it stays on this trend, hence 9.5kWh usable. Have noticed no 'funnies' going on but then I never did on 3077 either. Will be interesting to see if it stays in control (note you can ignore wide upper/lower control limits on 4075 as only 4 days data, need 20 days ideally and this should narrow considerably).
Solar light at the end of the tunnel?
3.67kWh into the battery today (Vs 3.85 yesterday).
Same "4%" to "100%" charge.
Tried to stay above 200W usage once the battery got to 50% again, so not sure why the difference.
Battery did appear to correct itself from 4% to 7% before the charge time kicked in.
Plus adjusted itself from 100% to 97% before the battery stated outputting at the end of the timed charge period.
TX200 I've found that there is always some variation in kWh added per %. There are a lot of variables - temp, charge rate etc. and the voltage curve is shallow for lithium batteries so more important is long term mean average of capacity and a tight variation around that.
Andy C
How are you calculating the capacity? I have data from charging overnight flat to full, then during the day solar adds some that the heating took out between full and sun up proper. So the daily charge kwh isnt just the flat to full bit.
Unless you are manually reading data each morning.
9.5kwh on 4075 - I put 8.55 in yesterday and got 8.0 out of it before it fell to 4%
Big improvement on 6.8 on 3077, and slight improvement on 7.8 the night before, but still someway to go
Today I only got 8.3 in overnight, not sure why. Maybe it the -2oC weather has affected it?
Jamesrich Temp's will affect it. Not sure how much it affects on the GE battieres. I know how much the cold affects my EV though......
TheDragon (GivEnergy) What does Portal/My Inverter/System Data show for the same period?
TheDragon (GivEnergy) I am, as you say, making sure this is just the overnight charge and logging it manually. I do not count any emergency charge from grid at low voltage.
Woodstock
Here you see on charge the AC slightly over DC
Discharge DC above AC, as expected.
Bit hard to see, blue on blue, but its visible
kram the new inverters are working well this morning - no issues with the optimisers causing any peak and troughs - currently generating over 3kW per hour and climbing.
These are are running on the standard firmware though and one string per inverter.
Andy C I do not count any emergency charge from grid at low voltage.
You are still seeing a drop that requires the auto emergency charge? Isn't that a bug to report/been reported? Thought the point of this new firmware was to prevent this occurring?
Andy C \Emergency charge??
Shouldn't be seeing this, even on the early beta's
chris_p I still see this occasionally, have reported it.
TheDragon (GivEnergy) This is with AC3 v290 & 4075 - pretty sure that's an emergency charge happening betwen 4 and 6am isn't it ? I've never NOT seen these on either 3088 or 4075.
I heard a very sharp long beep a short while ago. Think it came from my inverter or battery.
Also noticed all my solar was being exported to the grid rather than charge the battery.
I have system data for 12:00, 12:08:22, 12:08:32, 12:13, 12:18.
Really hoping that was a firmware update being applied. Fingers crossed!
ianfretwell
Ok, i will explain this.
If there is a reserve level set, if the battery is below reserve, it will do a slow trickly charge back to reserve.
EG if you set 20%, and your current SoC is 10%, the battery will do a background charge and slowly creep up to your reserve level. Instead of a WHAM full power.
TX200 Sounds like it. A long beep, would be a BMS update
TX200 Might bem I just got slapped with AC3 firmware 289. So now on 289 and 4075. Just realised, no calibration? If no calibration is needed for every future firmware that will be very good indeed!
TheDragon (GivEnergy) I totally get that - my reserve is the default 4%.
So if that wasn't what you were referring to by -
_\Emergency charge??
Shouldn't be seeing this, even on the early beta's_then what did you mean ?
ianfretwell Emergency charge, is you've hit 0% and the battery will initiate full power charge back to 4%
Reserve levelling is slow
TheDragon (GivEnergy) Gotcha - thanks.
TheDragon (GivEnergy) cheers. Just waiting for the MyInverter software tab/page to update. Believe it only updates every so often. 🤞
THALL when they put 3011 on mine they said no calibration needed, but it turned out one was required.
The system never recovered from being at 80% charged when the firmware applied and being at 0% charged after the update. It just wouldn't let me use the full battery, only 25% ish.
Maybe some of the bug fixes applied since will help. Maybe it'll be a calibration when you move to the new methodology from the old, then never again for future updates.
Thanks and good to see it works for you.
I’m basing everything on on inverter with 2 strings, one E one W.
It’s a very annoying niggle at the back of my mind saying, oh yet another potential rabbit hole.
TX200 Yeah they did change their mind on that calibration. A calibration did help fix some issues with that early beta.
Even though the forecast for today said light snow, sleet, rain I actually have blue skys so I am heading towards 100% charge much faster than planned. Looking forward to seeing 100% and not 98% with the new 289 firmware.
So for AC3 users is this 289 and 4075 the final test? Still some issues with hybrids I know, solar charging etc.
Now on 289 from 290, must have happen in the last couple of hours. Hard to see when.
Only thing to note is that battery power is now showing 25w less than load, not sure how that is possible. 🙂
Jamesrich Exactly what I am seeing on 290!
This is an emergency charge example
Jamesrich When the battery is idle the grid import is 25W less than the load too!
289 has just arrived and it it is still doing it. For me it was 4075 that started the "free energy".
Ra Need to know what the battery charge power is. Full bore or 290W?
i've had the 'free energy' since I started on the beta - i often wondered if it was affecting the SOC. I should have mentioned it, but it was only a few watts 3 - 5w, maybe?
Since 4075 + 289 its much more pronounced.
TheDragon (GivEnergy) The slow charge (290W) back to reserve is OK when the reserve is set above 4% but it causes a problem when it kicks off at a displayed SOC% of less than 4%. It starts to do-si-do with the discharged battery SOC% adjustment processing.
It looks as though the wait time after the reserve level recharge is too short and the SOC% adjustment process picks up the battery voltage on its way back down before it has stabilised. This causes the system to think that the SOC% is higher than it really is and it is adjusted upwards which in turn initiates a battery discharge. The battery discharge pushes the displayed SOC% back down again and after a while the reserve level recharge kicks off again and the cycle recommences.
Depending upon where in the cycle battery charging proper recommences, the displayed SOC% may or may not be correct.
Last 3 days on AC3 289 with BMS 4075
Woodstock Are you saying disable a reserve charge levelling under 10%
So in thory to be meaningful, a reserve can be set 10% as the Min, or Off.
2 2 registers needed for reserve EPS use. One to say use reserve, 2nd to se the level, 10% as the Min. Which is a common sense min setting if you really want to use reserve??
Thoughts. I will draft a response back in
On 289/4075 and I still seem to have a niggle with a drop away from 100% when charged on solar/grid. Seems like it's staying at 100% for longer but still drops without any battery discharge. Each time, if I restart the inverter it charges back to 100% again.
Another Harrier jump jet-like vertical take-off for our battery this afternoon. charging up nicely from the best sunshine so far this year and some grid input, the battery level suddenly shot up from 47 per cent at 14.40hrs to 100 per cent at 14.45. Second time it's done this now, again suggests it's been under reporting the battery level and then suddenly corrects it in one big leap. . . unless anyone out there has another explanation?
TheDragon (GivEnergy) That would work. However (as long as GE continue to insist on using the displayed SOC% to control their processing) a simpler change might be to just disable reserve charge levelling when the displayed SOC% is less than or equal to 4%. The "battery discharged and at idle" displayed SOC% adjustment processing (OCV to you?) should pick up the pieces and level everything out. Depends on what the conditions for initiating the OCV are (for example not "commence 30 minutes after completing a reserve charge levelling").
TheDragon (GivEnergy) well, it still shows 3011 on the MyInverter software page. Not sure how long it might take to update the info, but feeling less likely now…
TheDragon (GivEnergy) maybe my terminology is wrong, the top up you occasionally get from low% up to 5%. Is that emergency charge or just normal behaviour?
Andy C
Normal behaviour.
If the SoC gets to 0% which under this FW should no longer happen it will be obvious 2.6/3/3.6kw
If its just dropped a but to SoC level processing, mine corrected from 5 to 3% but just remained there. I think if it went to 2% the trickle charge 0.01C would bring it back to 4%. Never seen it happen, so cant really guage when.
I know if you set reserve of 10% and just left, over time, drops down and at 8% slowly goes back up to 10. Just hovvers around to as set.
If the temperature falls, so will the SoC reading, so it can top up. Temp warms, you end up at 12/13% it will discharge to house back to 10.
You get the idea?
Is this an accurate representation? I will send this in this could be upper end levelling, need to really see longer at 100% to be sure its an issue.
But will send it in as you have raised it
TheDragon (GivEnergy) Thanks. Not sure there will be enough solar tomorrow to get to 100% but I'll check it next time it happens and won't restart the inverter.
TheDragon (GivEnergy) yes, this is normal behaviour at the bottom end. I won't use the e-word again lightly!! 🙂
It looks so promising and yet sufficiently far away you have to question if you want to spend your hard earned on it whilst it still “coming soon”.
🤯
judgepd Yesterday I had the 2% jump at the top, on 289/4075. It stayed there for 20 minutes before dropping 2%.
Charged to 100% this morning on ac3 289.
Unfortunately saw multiple nipples, with discharge and charge back to 100. It didn't hold for longer than 5 mins, if that.
The battery was not supposed to be discharging. I guess this isn't what we are meant to see?
This hunting at the top, is the upper levelling. It hits 100, check is this really 100, not quite, charge some more. Check again. it may do this a couple of times, then stop.
If we had it stop at 100 then hold, you might be missing out on that tiny bit of energy.
Here is mine, i've had beta a while now, top end is a well known charge point, we hit this at night and maybe also during the say from solar if its a good day. So it has learned where full is. Charged to 100%, nearly an hours before the end of the cheap Agile period, so just sat there till 6AM
Very heavy battery use 3x5.2s
What about this from last night. Charged to 100%, immediately dropped to 99% and just sat at 99% for the rest of the timed charge period. One odd burst of charge which didn't raise the SOC back to 100%.
judgepd
Checked in with GivEnergy. Top end small amounts of top ups are normal, if the OCV hasnt run. It will be hunting for the upper limit (offline). So will stop, test, charge, stop test till happy. Kinda broad brush analogy, but the energy is tiny amounts. The other option is get to 100 and just stop. If the SoC is slightly out, ie no OCV levelling recently, you could be slowly moving away from 100.
So to prevent this, you see these small hunts.
If enough people really have an issue with it, we can look again, but the topups are tiny in comparison to the capacity of the battery and preventing any jumps in the future happening.
On the flip side of this, the tiny jumps, is the opposite, it has slightly drifted upwards (Above is a down drift) so thinks its hit 100 early. stops, checks, nope, i was right before, down, charge, stop check.
You get the idea.
It is a compromise between silly complex firmware logic and preventing drift over time
ABOVE ARE MY PERSONAL OPINION and what i understand from seeing things develop
Happy to stand against the wall, and await the incoming bullets
Cool, thanks Dragon. So its learning about what the 'top' is.
I'll keep an eye on it, I'd expect it should settle down after a few attempts.
My thoughts on this behavior.
Hit 100 and stopped, a previous OCV has happened in last 24hr, so no need for another, may well be towards the end of the 24hr counter.
at 04:28, OCV timer finished (Asumption) oh i'm not at 100, ok begin top ups. Doh, time runs out of the charge window at 04:30. Discharge begins
I add some humanaillity (is this a real word, it is now) to the logic as my way of understanding what may well be going on in the code, that's getting ever more complex. I hope there actually now focusing on the upper and lower limit relaxing to get the capacity back, instead of tiny oddities
Jamesrich Thats my assumption.
If a few days goes by and no levelling has been able to run, I hope that there will be a fail safe option added, that users can set a time to run one manually
TheDragon (GivEnergy) That behaviour makes sense but I can see lots of queries being raised if a 'full' battery sits at 99% rather than 100% even if it makes little to no difference in reality. On my previous night charge the battery did hit 100% and stay there so I guess it will just be slightly different each time it's full.
Any news on the capacity issue as for me that's the only problem I have with the beta firmware?
Correct, just like every day if you do a flat to full charge, each day will be slightly different kWh stuffed in. temperature, voltage, previous OCV run adjustments, all influence. As was mentioned by another guy, look at the average kWh over time.
eg for me, i consider ideal conditions as its mine and i hate things not just right.
11.3, 12.1, 12,3, 13,3, back to 12.3 again last night.
Going to have a look at getting the kWh loaded at 0600 each day over the past 90 days (Limit of my database) and average them, for each BMS version
TheDragon (GivEnergy)
"Top end small amounts of top ups are normal"
I think everyone would accept this if it was confirmed as the new normal.
Woodstock Full bore!!
4075 installed 02 March
Without regular full cycles, the indicated difference between my 2 packs does seem to be increasing. Voltages remain very close. As this will be typical summer usage for many, it does suggest regular 'maintenance' cycles will be needed to keep them in line as top point cal alone doesn't appear sufficient
Understood Matt.
The option to run the levelling manually at a user schedule was something previously mentioned, in case its been missed, I will jog the power at be memory, not to forget to reserve a register in the FW for this
@mpartington Can you send this in to betatesting and cc Dan, I will discuss at the next discussion point
TheDragon (GivEnergy) even that's not ideal as it seems to be happening quite quickly, so could be needed every few days. It would be better if the 'on the fly' look ups based on current, temperature, voltage etc. could be enhanced. We know that my usage pattern is probably the worst in terms of triggering soc jumps under the current production firmware, as full cycles are rarely completed.
I've sent it in to Dan and co, but interesting to share results here to see it others are seeing similar as we've all got different setups/usage patterns. Probably not an issue (or at least as observable) on the single battery systems.
Edit - forwarded to Dan L as well as beta address (per Paul's request)
Does anyone know if "on the fly" displayed SOC% adjustment is still a thing in 4075?
I have tried provoking an adjustment but have not seen any evidence of one taking place.
Could be that the lookup has been fixed and is now totally accurate but I rather fancy that the processing has been removed/disabled.
Would be good to know, to avoid wasting time trying to test something that is possibly no longer there!
Still on 3011 here, we’ll unless the info in the portal is wrong.
Noticed a change in logging frequency for a bit today, so guess something is happening, but not sure what.
Seem to be burning through the battery this eve, reckon I’ll be on the grid soon.
TX200 I'm now on v289/4075 - BUT - I had to apply both manually. GE were unable to push the files from their end because for whatever reason they always get corrupted on the dongle.
It might well be that this has also happened to you and is why you've been left behind as it were.
Probably be worth pushing them a bit at the beta address to see if that's actually the case.
I'd also like to know if there's any 'fix' for this as such or if I'll have to do all further updates manually?
ianfretwell the upgrade to 3011 worked fine remotely.
Did a web chat last week asking Dan to contact me, no contact so far.
Emailed normal support this morning, they’re checking with the beta team - no further update yet.
Might phone them up next week if I don’t hear anything back.
TX200 My initial upgrade to the inverter v290 worked remotely too but nothing else - i.e. I had to put 3088 on manually.
I just thought it might be worth mentioning.
TX200 Are you AC3, or Hybrid?
ianfretwell Remote updates can fail for a number or reasons, mine has had many sucessfull, few fail.
Internet to Inverter, if the dongle connection gets interupted, and a packet of data get corrupted, when the inverter checks the integrity, It will fail. You would rather it fail, than just load corrupted FW and brick the inverter. So this failsafe is a GOOD thing. PITA yes, but saves your inverter control board for needing replacement. It has happened once before, on an AC3 and no manner of manual loads would bring it back from the failed OTA DSP firmware load.
Just as it failed once today, tomorrow may be fine, a solid dongle connection is key here.
Woodstock Yes I still see it sometime, not every day, as im cycling fully as its cold and the ASHP is sucking the life out the batteries by 10am, down to 50%. Then sun brings it back to 100 and tea time and evening heating drains then dry by 8pm ish. Ready to stary again at midnight.
Hard life my batteries have up to 2 cycles a day, not always fully cycled
Unless I am missing something, isn’t where the beta is now substantially better off than the last firmware release?
With that in mind, would it not make sense to roll if out as an officially supported firmware v4?
kram BMS has capacity issue for some
Inverter I am tracking 7 faults, 3 due to the rewrite, the rest, similar issues in previous versions. Might aswell get them all out at once.
Beta may roll out to wider soon, subject to a few idea getting apprved
TheDragon (GivEnergy) BMS has capacity issue for some
Isn't this for everyone on the beta?
judgepd varying degrees.
I getting average 78% DoD.
Others less.
Think the 100% DoD are getting less. Need to gather more data like the guy above who got a nice stasticical separation plot.
That is very good data
TX200 If there was a DM function, I could send it over.
If I knew your inverter serial, I could get it chased.
Dam this antonymous forum
Didn't stay for long at 100% this morning, but it continued trying to charge all the way until 04:30.
I assume it was trying to find full, but with the reduced charge rate post " 100% ", i don't think the '4 hour Octopus Go charge window' is enough to find full.
I've resumed charge, and left it charging all morning to 'do its thing'. 11.8 total charged today, minus 1.34 consumed between 4:30 and 08:45. I've calculated its managed put 10.46 kwh in, which was the maximum I saw a few months back.
I'll see how much I can discharge out of it today!
I accidentally posted it in this thread the other day, can't edit the post! So may as well post it again here: CE2133G030
I'll just try not to post my API key. 😁
It's happened again today
Low usage, but faster percentage drop
38% to 13% between 1620 to 1720, circa 115W
Incorrect soc calculations/OCV thingy.
TheDragon (GivEnergy) happy to throw anyone's daily battery kWh data into Minitab if it helps!! (that might be a dangerous offer!)
A little rub of the lamp, hopefully the Genie is home and you get an upgrade
Or it fails and you get an email with a gift of 3 little files. Doubt tonight, its late, only im daft enough to still be up and about working
TheDragon (GivEnergy) cheers. 🤞
Had some 'strange' behaviour last night. Weather has been rubbish so I knew the battery would go flat at some point but looking at the picture below, the battery stopped supplying the whole house load and reduced its output to 90W and continued this until it bottomed out. Is it supposed to do this now on the new firmware (4075)? On a more positive note, it put the most in it has for a while from empty (7.9kWh), still not 100% but better.
Had interesting battery activity last night. Actually watched this happen. Check out the bits highlighted in yellow. Batteries had discharged down to 6% and stopped. Not the 4% I have set. Checked the mobile app and it said my 8.2 battery was completely empty but my 9.5 battery still had 12%. From 20:45 to 22:05 the battery soc kept increasing up to 14% and was slightly discharging all 300w to my home. It found a bit more before the grid charge kicked in.
Capacity. Yesterday battery was grid charged to 60% My total battery capacity is 17.7kw so 60% is 10.62kw(ish). Solar to battery was 3.46kw. Total discharge 17.39kw. So do I assume the new firmwares found just over 3kw? 3.31kw? Doubt these numbers are correct, I got them from the app and dashboard. But it does appear to have found 9% + how much it discharged from 20.45 to 23:30.
For those of you that understand the data better than me if I can provide any other data / screenshots so get a better understanding of what happened please let me know what you want and I will get it.
But in simple terms, it looks pretty good :-)
What I see here (Assuming without moreinfo)
The lower voltage limit was coming close, the BMS is designed to back off of low and voltage/SoC table are out of tolerance. To prevent the crash, at this low induced rate, OCV (Levelling) can take place after an amount of settling time from heavier loads (The shallow SoC drop). You then see the SOC drop faster, this is the OCV determining OMG the voltage is lower than i thought, adjusting the Soc downwards to match.
This would have previously hit a 0% CRASH. Avoided, still maintained some load, as the ABSOLUTE lower limit is set at 2.8v which ordinally never reach unless a fault (Not even calibration goes this low)
I hope this explains what is going on here with the logic.
Yours is the opposite to RA. The SoC finished, im flat.
OCV begins after the settling period, OMG voltage is higher than SoC, i've actually got some power remaining. The BMS will allow current to flow again (At a reduced rate, depending on actual voltage level).
Looking at your graph, it was actually quite a way out and had you not induced the charge, would have found the value that was correct, and the the SoC would go down again to 5% before stopping at the correct 5% for your pack make up
Again i caveat this with what I have over time, deduced from the logic, what's New readme's with most supplied FWs to me. I have a pretty good understanding now of how the BMS logic works in most if not all circumstances.
This group has dome this for me, examples, explain what you were doing etc and GRAPHS. DATA IS KEY. Thank you all.
In the coming weeks, I will have an announcement too
TheDragon (GivEnergy) That is what I thought. It's just the first time I have seen this behaviour so wanted to confirm 🙂
Just need to get the upper end sorted and I'll be happy!!
Battery Cutoff % Limit (not Battery Reserve % Limit) .
Does/should this have any effect with the new firmware? I tried setting it to 25% and the battery discharged straight through it. (Never tried this pre-beta so it may always have done this).
OOh now there's a question. In the AC3 it used the stop at the cutoff, or the reserve, which ever was higher.
Can anyone else repeat this? I will give it a go on my own system, if not a glitch, it gonna go on the minors bug list. Its there, but not a show stopper.
Ive set this ready for a test, will kick of a manual charge to get up above these levels, as Im flat already
OCV dragged my soc down incorrectly last night as per the pictures below.
The next morning it jumped from 80% to 100% at the end of the charge, thus lost 1/5 of the battery capacity due to incorrect OCV.
Interestingly it didn't recalibrate back up whilst idle yesterday evening.
Although perhaps it can't because it's within 24hrs of the last one.
It's a shame - it could have found that battery capacity and started using it!
(3011)
TheDragon (GivEnergy)
I think there was a special firmware update for the reserve (for AC3 only). I never applied it. Maybe it hasn’t been built into the beta firmware yet?
And as for your announcement, got a new job? 🤣
Ok, am charged to 17%
Reserve moved to 4%, cutoff still at 10%, however this really isnt a realistic setting, as reserve would always be equal or higher to cutoff.
But for the test, I have forced it to be totally nonsense
Will report back once the sun goes in, or the wife turns the washer on again (At work, so cant do much back home)
Early indications, cutoff is indeed ignored, If i set reserve toe same. discharge stops as expected.
Will test some more
TheDragon (GivEnergy) So do I need to let my system empty for the OCV to start? Because the goal is to not fully empty the system if I can help it :-) But since the firmwares were applied this is the first time I have seen this happen.
Annoucement, new job. £50k a year, customer relations for GivEnergy? :-) We should start a pool going.
The idea is the OCV will happen in 3 scenarios as it is now. 4th recently added
TheDragon (GivEnergy) Ok I so either option 1 or 2 this time. This is the first time I have seen it happen and as you mentioned it would of settled if my timed charge did not kick in so there is more to fine. But the fact it found either 9% or 3.3kw I am very happy about. I would try a full charge then discharge, but 17kw on a single AC3 takes a long time.
THALL I bet it does. At least now if this drift does happen, it can "Self recover" and give you back what it robbed off you. Previous its lost, and the next time you know about it, is when the upper voltages slams into the upper safe cutoff
Under 4075 289, has anyone seen the battery discharging to meet the house load, when it's supposed to be charging?
kram
So, update, since swapping to the Gen 2 inverters I'm now seeing the dreaded cross charging which didn't happen previously with the Gen1 versions.
Nothing else has been changed so I'm guessing it has to be something software related.
TheDragon (GivEnergy) I will have an announcement too
A well paid job, shares and a free new all-in-one all feel appropriate!
What happens when you set a Battery Reserve % Limit and just leave it.
Battery Reserve % Limit set at 29% SOC and system in ECO mode throughout.
Battery temperature between 8 and 9 degrees C for the entire period.
Automated battery top ups were all at 290W.
(Top up four, early the next morning, looked identical to top ups two and three).
SOC% was restored to 28% each time (not the 29% set).
Solar PV turned off (no sun anyway).
Throughout the entire period there was no recorded battery discharge.
.
.
.
It looks as though the reserve top up processing realigns the voltage/SOC% relationship. The first top up started at a much lower voltage and put far more energy into the battery than those following.
The thing that concerns me is the amount of charge being lost when the battery is at idle. In a 24hr period the system put 1.4kWh into the battery to keep it topped up (no recorded battery discharge). Allowing for the fact that some of this energy was used initially to realign the SOC% there is still around 0.65kWh a day being used purely to keep the battery topped up to reserve level. Makes me wonder if this level of discharge is going on all the time and only becomes visible at idle. Also makes me wonder if this discharge is taken into account when calculating (not measuring) the displayed SOC%. Does the BMS alone really use 0.65kWh a day?
94jdh Cross charing on hybrids? Ah I thought it was a AC3 feature. I wonder how many of us are waiting for the EMS.
TX200 now on 4075 😀
(According to the portal)
Time to see what new bugs I can find, plus see if the OCV causes more, less or the same problems at low usage.
That is a good question. One test would be, to be sure its not a drift in some disguise, is to set reserve to 100, and leave it. If it does carry on adding power for no reason, the voltage will soon start to rise above normal and hit a limit
Response from FW team.
"Cut-Off is set at 4%, adjusting it in the new FW will have no effect"
This register has no purpose in the new FW going forward.
Option 1 - bring it back into use - Of no real benefit, use Reserve
Option 2 - Remove the slider completely.
My 2 propositions.
I would expect any engineer worth his salt, if its not needed, of no benefit, it will be rescinded in a future portal upgrade.
Maybe someone will see, and the postman brings me something nice. LOL
TheDragon (GivEnergy) does this mean that the ability to set a reserve (in case of powercut) would be removed? (Sorry for the post as not in beta, but wanted to ask a question)
Ah, I think I might have got confused earlier thinking cut off and reserve were the same things. They're different aren't they. Ooops.
THALL
Yes, very strange considering it was working fine the Gen1 inverters and effectively two independent systems.
Now I'm also seeing another strange issue, each system approx 3.6kW panels, East and West. So any charging would go to its own connected battery.
I've been generating approx 3.6kWh total, 1.9 and 1.7kWh per inverter and you'd expect that to go in each battery (less any house load) which it did before.
But no, now I'm seeing 3.4kWh going into one battery, and one discharging at around 300W?
94jdh Bonkers. Can try rebooting the inverters.
I've tried that and no joy, just on with the installation team and they're contacting Givenergy as they've not see this before.
TheDragon (GivEnergy) Don't think it is adding (replacing?) energy for no reason. Looks like it is responding to a drop in measured battery voltage causing the displayed SOC% to fall. Also the top up charge ceases at the same battery voltage each time (with the exception of the first which stopped at a slightly lower level) suggesting it is working in response to measured rather than calculated values. The question is why is the battery voltage dropping so quickly when the battery is "at idle" and there is no recorded battery discharge taking place? Is this linked to the constant 15-27W that the inverter is reporting as coming from nowhere?
Smk82 Reserve is very much there.
The Hard cutoff is no longer relevant.
In the new FW
If you set reserve of 10% and your under this, it will charge slowly to it and stay there, adding a topup if it drops.
If reserve is set it will discharge as normal to house, when it gets down to reserve, it stops.
Only EPS can access the 10 down to 4% bit.
Woodstock batteries do settle.
Voltage drops with settling, or drop in temperature.
Both happen after a charge. Cells loose heat not in use, terminal voltage drops slightly.
Chemical equalization causes the settling voltage drop. Hence OCV waits post charge.
94jdh the new Gen1 firmware Like Gen2 hybrids, also looks to be monitoring AC too.
So this cross charging has appeared, it's on the list, to improve dual control.
If you Have GivTCP there are some automations there that have pretty much nailed AC3 and Hybrid G2 to each other. They look to behaving well.
EMS functionality is the official GivEnergy intermediate bit, likely doing the very same thing. Tjis is in work still, no date, sorry.
It is Monitoring both, if one steps out of line, a swift wrap on the knuckles. Back in line
TheDragon (GivEnergy) thanks for the update, I'm still running on the SIM dongles so I may have to dump these and run from WIfi for the GivTCP - and get something to run it 24/7?
So do you think this is also why my panels are now cross charging to a single battery as well even though set up independently?
TheDragon (GivEnergy) Is 0.65kWh per day a reasonable battery leakage rate? Bearing in mind that the battery was being maintained at 29% SOC (2.4kWh) losing 27% of this in 24hrs seems excessive to me. Certainly didn't go in heat as the battery temperature remained constant to within 1°C throughout.
An AIO system 😂
On a serious note, I’m ready to pay more to replace the GE parts in the new E/W system with a suitable alternative.
Do you have anything solid that would be compelling to make me change my mind?
Thanks for the info, it’s not filling me with confidence as my above post, but I hope you get it resolved.
TheDragon (GivEnergy) When I last asked about the ETA of the EMS they said Q2 targetted launch so April to June. I am really really hoping it is on track. Having a single AC3 with 17kw of batteries is a pain. They did tell me testing of the ems is done, I did offer to test a unit for them ;-)
THALL I’d emailed offering a test as well. No response. With 2x 3.0AC inc and 16.4kwh of batteries, I manage a full charge in just under 3 hours. Also hopeful for ems
Smk82 Oh I gave up on emails ages ago. Try web chat. I currently have a single AC3 and 17.7kw of batts. My ASHP draws about 3.8kw when running flat out and then 200 - 500w background usage so dual AC3's should cut my grid draw to pretty much zero. I can't wait! My installer is ready and waiting as well.
THALL might be a stupid question, but where do you find the web chat? I looked on the website and couldn’t see anything?
Smk82 on the hopepage bottom right. But the button will only appear when someone is available. Working hours only, seems to be manned 10 til 4 ish.
THALL oh that’s fine, just want to know where to look!
RogerGreen do you have a low base load say < 200W ?
If so, you might be experiencing the same thing as me and Woodstock.
When the battery is under 50% full, it may drag the percentage down too far during a "lower end OCV" (if an OCV hasn't run in the previous 24hrs)
I've been seeing my percentage drop by 10% every 25 mins when my usage has been low (say 100-115W). Should be more like 1-2% drop.
Then the next full charged I see a jump from 78 to 100% or 80 to 100% etc.
I've just moved from beta 3011 to beta 4075, so will have to see if that behaviour changes now.
4.14kWh in overnight, which is about the 3rd highest I've seen since joining the beta. And the battery wasn't even empty at the start of charging (at 11%).
What is strange is the charge rate of 1400ish was ignored. Instead it charged at 1745 for the first 30 mins, then 2675-2708 until it got to nearly full.
The little bumps after the charge I assume are it checking it really is at 100% - which could be useful!
(4075)
TX200 Haven't seen this since 4075. Yesterday I had an upwards lower half SOC% adjustment for the first time ever (13 to 18%) so looking hopeful. At first I thought they may have disabled this processing altogether.
Woodstock excellent. I look forward to not reporting any more issues with that! Fingers crossed 🤞
Woodstock I'm NOT on the beta testing programme. I have a Gen1 3.6 Inv FW D0.450-A0.451 and a 9.5kWh battery FW 3009 (installed Sept 2022) and everyday I am seeing the same thing as you. Battery at idle (4% SOC) in the early evening and then dropping to 0% by the following morning and then emergency top up taking place or getting down to !% and solar kicking in. There appears therefore to be aprrox. a constant 27W drain when the battery is idle which equates to about 0.65kWh. A battery re-calibration hasn't improved the situation.
Could it be the inverter powering itself from battery rather than mains?
Or just the percentage updating as the temperature of the device changes when at rest?
Two displayed SOC% adjustments yesterday.
One upwards "lower half" adjustment at 12:15. (13% to 18%, on the fly).
One downwards "battery discharged" adjustment at 22:15. (5% to 2%, battery idle).
I may be wrong but I believe the "only one adjustment in a 24hr period" restriction applies just to "lower half" adjustments. I don't think there is any restriction on the number of "battery full" or "battery discharged" adjustments that can take place.
CCourtney Thank you for the information. The plot thickens!
TX200 Possibly the first. Not the second as my battery temperature has been constant for hours before and while this is going on.
Doesn't the Inverter "PLC" and the BMS systems have to draw their power from the battery in order to function when the EPS is activated, otherwise both would shutdown when the grid fails.
correct, battery inverter definitely runs from battery during power cut - had one yesterday and it was the only thing working.
TX200 200W on anything other than the smallest batteries, is under the OCV load threshold.
I would like it reduded from 4A to 2A, but last time i recommended it, the FW engineers poopoo'd me
TheDragon (GivEnergy) well, if it happens again, I'll report it again. Every time. 🤣
I guess they need to tweak it slightly, to take account of the voltage whilst in use.
Although if it quickly and dynamically adjusts when it gets to 4% and gives me the rest of the battery back to use, then maybe not so bad.
TX200 Charge rate ignored? Thats a new one, I purposly tested this, when I did the DC shunt tests, charging and discharging at set levels and comparing DC shunt to App readings. It deffo complied to the settings, allbeing the actual DC values were lower than the App, App was reading 7% high on both charge and discharge.
This is linked to sample timings, it is on the list, lower down, but im tracking this one too.
Can anyone else, set a grid charge for 10 mins set the charge to 50% power, ish, see if the actual charge is correct, or ignored. AC3 is fine.
TX200 I too have seen this 21-25w drain displayed.
In FW AC3 289, it now reads 1W.
I could easily see the BMS's and inverter itself using 25w average, after all it is a small computer in there
I asked yesterday about are there any more releases coming, to address the list of bugs Im tracking, ranging from major (Still capacity issues) and Inverter home demand, dual charge slots for Cosy tariff, to minor.
Pause not fully paused, SoC over compensating down aswell as UP, UP was the one I noticed (Reserve charge topups) BMS 4A business.
Anyway they are planing to get as many as they can in the next release, so instead of multiple, we will get one more, towards the end of the month (Yes I will be hot on their heels)
I hope this adresses most of my tracked bugs, after all me and Dan spent quite a while cross checking our tracked issues.
This will go out, to the curent crop of testers. Another round of testing, if nothing else comes up, we may be in a poisition to, add more testers, wider, feedback not really needed, unless its dire.
Then if we arent rocking in the corner with stress, its all too much, then final update, will be proper verision numbers, and upload to live servers.
By this time, It will be my train set, and as you are probably well aware, I aint like normal people
TheDragon (GivEnergy) Everything is pointing to this nearly being finished finally. Do you have any idea what the plan is for pushing to new systems (and if not can you inquire!)? I'm in limbo now, seeing no point installing a system with the old FW, and waiting for the new FW installation to be part of the commissioning process. Does seem logical for GE to make this switch as soon as the FW is on general release, but obviously things don't always happen in a logical manner!
phoenix I dont think it will be an enforced release.
It will be like any other, there if you want it.
I cant and wont nail myself or GivEnergy Directors to the wall with a date.
You will if following this channel, see the light get brighter
TheDragon (GivEnergy) But if it is the inverter draining all or part of that idle battery power in order to run it's internal electronics why isn't it taking it from the grid? It's not like the system is running in EPS mode.
In my case rebooting the battery and inverter cause that idle drain to half to around 13 to 15W but the improvement only lasts a day or two before reverting to around 30W drain.
The first three months of ownership always had the 13 to 15W idle drain.
TheDragon (GivEnergy) When the battery is at idle I am seeing the inverter power in one of four ranges. Roughly 0-1W, 6-10W, 15-19W or 23-27W. Haven't yet worked out the rhyme or reason for which range it is in at any given time.
When the battery is charging or discharging the inverter power is suggesting inverter gains (rather than losses) at rates of 360ishW or less. This switches to showing inverter losses as charge/discharge rates increase above this level. I am guessing that this is when the true inverter losses start to exceed the displayed power discrepancy at idle.
Woodstock You have AC3 289 now yes?
Can you do some charging at different power levels set in the portal.
100w or closest value, as there not liner.
200W
500w
1000W
Record battery power (DC)
Record Inverter Power (AC Power)
I would like to compare to my previous results, but these started at 500w, you are seeing apaprent gains if very low.
Charge, AC will ALWAYS be higher than DC, it has to be
Discharge, AC will ALWAYS be lower, it has to be, conversion losses.
TheDragon (GivEnergy) just in case the firmware overwrote something, I adjusted the charge rate slightly earlier and saved it. Will see what it does early hours tomorrow.
TheDragon (GivEnergy) Are you saying that "lower half" (as distinct from "battery discharged") displayed SOC% adjustments are happening by accident and are not intentional? I have maybe made a false assumption but I thought they were there to cater for usage patterns where a fully discharged battery was rare.
If they are happening by accident rather than design then lowering the threshold to 2A would certainly reduce their occurrence. My overnight base load, however, is often less than 50W so it would be necessary to set an even lower value to eliminate them.
Either way, if the displayed SOC% adjustment is accurate enough does it matter if the threshold is 4A?
TheDragon (GivEnergy) Yes. Will do.
Woodstock I was assuming accidental and not intended, but those very low users, who never go flat, it'll not run at all. This is a difficult one, as the look up will need to take into account a varying load between 0 and 4A to offset the voltage. Now that is an unknown, as this will for sure be dofferent for every install, load has less impact the bigger the battery. So pack count, cell count cell make up 16S vs 32S2 makeups.
For sure this may well be the hardest one to nail, and we may have to have a compromise.
Where it could drift, but at top you get hunting, botton you get flat, nope, flat, nope till true flat.
Me and you may well need to work outthe box on this one.
Lets see the next itteration and go from there.
TheDragon (GivEnergy) That's pretty much my situation, not so much low user but I have enough battery capacity to almost never run flat (assuming no SoC drops). That was the biggest issue with the pre 535 firmware, I'd drift and get to what seemed like 30%, then crash. Made it very hard to actually use the full capacity.
TheDragon (GivEnergy) Here you go. Not sure how repeatable the individual figures would be but the trend is clear.
This was a charge from 2% SOC.
.
.
Charge ............................ Inverter ............................. Battery ...................... Gain/
Power ............................... Power ................................ Power ........................ Loss
82 ........................................ -53 ...................................... -79 .......................... 26 G
164 ................................... -136 .................................... -159 .......................... 23 G
246 ................................... -219 .................................... -239 ...........................19 G
328 ................................... -303 .................................... -319 ...........................16 G
410 ................................... -387 .................................... -399 ...........................12 G
492 ................................... -472 .................................... -480 ............................ 8 G
574 ................................... -557 .................................... -562 ............................ 5 G
656 ................................... -645 .................................... -644 ............................ 1 L
738 ................................... -729 .................................... -727 ............................ 2 L
820 ................................... -817 .................................... -809 ............................ 8 L
902 ................................... -903 .................................... -893 .......................... 10 L
984 ................................... -990 .................................... -977 .......................... 13 L
1066 .............................. -1079 .................................. -1061 .......................... 18 L
1148 .............................. -1168 .................................. -1147 .......................... 21 L
1230 .............................. -1257 .................................. -1233 .......................... 24 L
Woodstock
OK I suspect this is also linked in with the sample rate of the currents. Sampling was reduced, then so did the accurancy of App reading vs real.
I will keep this info in my pocket, as I think its linked to existing and gonna add your data to my original power test reading spreadsheet if you dont mind
Woodstock
My results with Invert FW AC3 290 18th Jan 2023
As mine were all above 500W i didnt see the very low power, AC vs DC switching sides, clearly you can never get more out than you put in. So it is a glitch indeed
The Pause and 0W set was also documented here tooo
Ok folks.
I am now allowed to break cover. My efforts on this and other forums havn't gone unnoticed.
I recieved a job offer from GivEnergy, for which I have accepted. Leaving my regular job with the world 2nd largest defence company and joining GivEnergy, as the Head of Security & System Testing.
In April i will officially begin, however as most probabaly will deduce, I have always been here, but now can speak officially. So those before voicing their concerns over some random member, offering information. It was always founded on good backing.
So here I am, coming clean, I will also keep my forum name, just because I can.
Good night and godbless GivEnergy
Excellent. Can only be good for all!
TheDragon (GivEnergy) A very solid decision for GE that should help restore confidence. Good for you!
TheDragon (GivEnergy) Congratulations and best wishes. Thanks for all your efforts, I feel more confident now with you on board.
So can I please ask…
As it stands today firmware is sounding positive, if not perfect.
What exactly is the situation with a gen2 hybrid and an east/west split array on two strings?
I get the impression it’s broken, acting as a single string so causing reduced generation. Have I got the wrong impression here?
kram That was how it was, now much better, but still not TRULY indepedant. I have 2 beta testers purely on this issue.
It will be resolved. you are also correct, alpha testing it was near perfect. Open up to Beta, as expected more scenarios, more variations on kit, if even highlighted some real faults, hidden previous.
Now tracking of more installs, we/I get a better feel for behaviour in more circumstances.
I now also now need to know, is ANYONE not
Hybrid 199-104
AC3 289
BMS 4075.
Like THALL was lft behind on 3011, which i got resolved
I do hope your new boss is paying for all the energy you're using to test the kit!
TheDragon (GivEnergy) That was how it was, now much better, but still not TRULY indepedant. I have 2 beta testers purely on this issue.
It will be resolved.
I do hope so as that's what's being installed here in a month!
dbt85 I would be interested in what firmware you are delivered.
Please get in touch once installed
TheDragon (GivEnergy) congratulations and thank you for all your help. Wish you all the best for your future venture.
Dewy Nothing will change
Just some more structure, once I get in the chair
That’s similar to me but with all the it might /might not work for an E/W split I’m debating paying more for something that definitely works as two strings.
E/W split arrays are not cheap so I need maximum generation asap as it’s only purchased to power the car and minimise home bills and summer has use.
TheDragon (GivEnergy) 
Charge rate more or less respected today. Hopefully yesterday was a one off.
Changed it from 1462 to 1514.
the majority of the charge was between 1519 and 1558 W.
TheDragon (GivEnergy)
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
3077
Congratulations TheDragon Excellent news for both GE and this forum.
Hybrid 5.0 Gen1:
D0.450 A0.451
8.5kWh Battery:
3008
And still experiencing battery high voltage of circa 58V on EVERY charge cycle. No SOC jumps to correct for drift. Just hundreds of high voltage errors and inverter going to idle. Martin Phipp has re-calibrated the battery a dozen or so times to try and regain some capacity each time.
Total nightmare, constantly monitoring and trying to catch and pause the battery before it goes high every night and day. So…. If anything you can do in your new role to help bring our system along, even if to patch with the beta software for now, would be appreciated.
Congratulations @TheDragon (GivEnergy)
When I came on board as a customer it was due to the strong community above all else.
Your technical engagement with customers here, has gone along way to restore my confidence that any reported issues will be resolved. Glad to see you've joined Givenergy, and looking forward to buying more batteries in the future.
mattpdavey Hybrid 5.0 Gen1:
D0.450 A0.451
8.5kWh Battery:
3008
What is the serial number
positor1 D0.289-A0.289
Battery Firmware Version
3077
SD2148G201 👍
OK I will re word the question.
Anyone on the beta program, and not
Hybrid 199-104 (current Gen1 Hy version)
AC3 289 (Current AC3 version)
BMS 4075.
Ie you got left behind.
mattpdavey
Your versions are pre beta, those are the latest Hybrid Gen 1 regular release versions.
Have you been acepted on the beta program? When, who by?
CE2219G122
i wasn't sure if i should post it in a public forum so left it off sorry.
assuming i wont have to re calibrate?
positor1 No. It be fine.
Unless someone wants to buy you an extended warranty, it is of no real use to anyone, only Givenergy can identify you from this
TheDragon (GivEnergy)
CE2119G103.
nwc469 Your firmware above, isnt part of the beta trial.
Have you asked and been accepted onto the trial? via betatesting@givenergy.co.uk
Accepted 16th Jan via email had to upload files via usb had some issues with capacity multiple emails sent finally got through to tech support think inverter firmware changed and recalibration twice guess i was removed from beta without me even knowing 🤣
Strange, the Inverter FW version is the non Beta, stable public
Yet the BMS was a beta from around 2 months back
Ac3 D0.535-A0.535
Bms 3011
When you got sent files, there would have been 3,
ARMStore.bin - Inverter CPU firmware
DSPStore.bin - Inverter DSP firmware
BMS_ARM.bin - BMS firmware
TheDragon (GivEnergy)
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
3007
CE2105G137
The Inverter Firmware was updated remotely, but the BMS would not update remotely, so was sent a file to update it myself. However, it didn't work. Therefore, left it as others were reporting problems with the different versions and not getting any help/feedback from GE.
Happy to try it again with the latest versions.
PS. Have only experience the drop a few times, may be because the batteries get charged to 100% most days.
However, noticed recently that there can be a 10% difference between the 2 batteries SOC.
Woodlands What was the last version you had, as the inverter is latest, I would assune BMS 4076 or 4075 (Latest)
Yes thats right however if i recall the inverter firmware was 534 534 initially at the start of the beta they remotely upgrade but the inverter was on 534 290 I think but might be wrong there hence the manual upgrades the battery has been on 3011 since january only inverter firmware changed which remotely changed to current in february after calling support
nwc469 534 290, would be ARM updated, DSP failed
So maybe helpdesk set them both to 535 535, whish is the latest AC3 public release
TheDragon (GivEnergy)
Don't know, as the bms file that was sent does not have a version number I can see., nor did the email it was sent with.
The portal reports that the batteries are still on 3007.
No to all, despite applying and emailing requesting the same on many occasions.
Still leaves me with a system that hasn’t worked correctly for the last 8mnths. If it weren’t for Martin Phipp the engagement from GE would have been zero. Frustration would be an understatement, as I have done everything in my power to reach out to GE, describe our specific issues, engage with the beta testing… etc.etc.etc.
Woodlands Thats after the load?
Remove USB from all batteries
Put in a computer, ensure Content is empty.
Copy file recieved to the USB(s) then eject them.
Place in Slave battery, it will flash flash and beep, then reboot.
Repeat with master.
Always do slave's before master, found this to be pretty fool proof.
Once done, give it 10 mins and check the portal, software page, for BMS version
Once I actually start, I will have a GivEnergy email, i dont want to use my personal one for work use. a. not secure, b, I need a life too.
OK, I feel the beta will help your system. But till I actually get in-processed to GivEnergy I cant actually access the betatesting emails. But once I do, I will be going through them
TheDragon (GivEnergy)
The BMS file I was sent, that didn't work, was back in January.
Luckily, I had a copy of the previous version, so I put that back on the USB sticks and got it working again.
As the system is working at the moment, I'll wait for the latest version, as I don't want to risk it not working again.
Understood. And that’s all I can ask….having someone as capable (and reachable) as yourself on the case would be a real plus! Thanks
TheDragon (GivEnergy) Best decision made by GE for a long time. Congratulations, well deserved!
chris_p Will you stop giving me a hard time now? LOL
TheDragon (GivEnergy) I'll think about it 😉😂
kram e/w should work, it's worked on others inverters for years so I can't see a reason why Giv can't get it working properly (or if I'm honest how they didn't have it working out the gate). Since I also have access to a 4, 6 and 8kw set of arrays facing south 100m away, it'll be interesting to compare generation throughout the year.
@TheDragon (GivEnergy) congratulations. Can you get an ETA on the EMS device now? :-)
dbt85 I have two arrays - south and west on separate strings (Hybrid) works perfectly.
94jdh no optimisers going on my install, no shading.
Am interested in the fact that the inverters are more responsive now though.
94jdh no gen 1 hybrid and no optimisers.
Congratulations TheDragon, couldn't happen to a nicer chap! Will you be working remotely or journeying to that wonderful hinterland of Chemical Lane lol? I know it well as Foster Grant were one of my old customers.
Looking forward to the official public release of the final firmware. In the meantime my own system plods along faithfully and faultlessly, I must be one of the lucky ones.
dbt85 I have 10.2kwp on an E/W even split. Currently with a 5kw gen 1 inverter. The generation per array is good and similar.
However I initially ordered a gen 2 inverter and this is being retrofitted the week after next.
Are we saying hold off on this because the new gen 2s can't handle two arrays? Eh what! That's madness and would be a completely unacceptable situation. Just how bad is the issue please? I look at this thread reasonably regularly but must have missed that 🙁
sponev have you Tigo's fitted?
That’s what I’m trying to understand.
No tigo = no problem?
94jdh Zero shading so no Tigos.
kram Have GivEnergy said anything official about this or is it something that's cropped up during the beta testing please? Lots of folk use both inverter strings so this isn't a trivial issue. I see the post which says two beta testers are working on it but surely it's a pretty fundamental problem which needs sorting asap? Is it guesswork time again 🙁
It's been a tortuous year trying to get a fully working system in place and the retrofit of the gen 2 was the final piece of the jigsaw. Surely it's not broken before it even starts?!
sponev It cropped up in testing.
No one has ever reported it as an issue with Gen2 without Tigo. Unless I messed that.
There is also new Beta for Gen2's aswell, D0.906 A0.906
TheDragon (GivEnergy)
So without tigo the two strings work as expected on gen 2 5kW hybrid and there would be no issues with an east/west?
TheDragon (GivEnergy) OK thank you. On the back of that it seems reasonable to proceed with our retrofit of the gen 2. It does seem extremely unlikely that GivEnergy would have allowed a huge retrograde step like that. Hopefully it can also be resolved for the folk with Tigos on dual arrays. It seems specific enough to be able to diagnose reasonably easily.
I’m more positive, although still on the fence as the other system can pull 5kVA from the battery alone (circa 4.6kW) and will do 5kW from both and nearly 7kW for a 10 second peak. Pretty much like the AIO.
Brilliant news and a no-brainer as far as I can see! You'll be a great addition to the company!
Still seeing the immediate drop from 100% when charging, either forced from grid or solar as well as my own feeling that I'm not getting full capacity from my battery due to it not hitting the higher end of the voltage scale. I'm on BMS 4075, is this a BMS issue or the inverter firmware as that is still D191-A104
Ra inverter.
DSP 199 fixes this.
I've consolidated a list of all the inverter we chat about here.
With serials if I pick them out the screenshots.
There will be many more, once I actually take the reigns I'll be sifting through the beta testing emails and consolidating a full list, if I can.
Anyway the list of those lagging behind was submitted to the team.
You are on it, but I couldn't find a serial for you.
So I cant put pressure on.
Since I accepted the off, I've been backtracking of what data is available to me here,
So you can be sure once I get the actual email traffic too, things will move along.
There is a consolidated bug fix for inverter coming end of the month (dont shoot me, just relay)
TheDragon (GivEnergy) My inverter serial number is SA2209G014 🙂
Many thanks for all your help 🙂
Edit - Don't worry about the pressure, just wanted to check that a potential fix was in the pipeline more than anything
Ra it's on the list.
So next round of checks, cant do too often direct to the Boss, will likely I have direct login of my own.
Regards.
Paul
TheDragon (GivEnergy) Congratulations! Well deserved.
Congratulations. I am not on beta but follow this forum daily.
Your contributions have been brilliant and you will be a real asset to GE.
TheDragon (GivEnergy) I also see this issue where it would drop immediately from 100-99%. I've just checked and on last night's charge it didn't hit 100% at all and just stopped at 99%.
Now getting a discharge spike at 01:00 every night.
Never had these before. It could be a coincidence but these started when I set battery cutoff limit to 25% as part of testing and now I can't get rid of them! I tried a reset to factory defaults yesterday but had another spike this morning.
There is a thread about this behaviour (from someone not running beta firmware) under the title ...
"Battery Charging for 5 minutes every night from the grid when it shouldn't be"
This thread contains a lot of background information and suggestions but, as yet, no definitive cause or solution. There is a "fudge" work around but it is not clear why this is having an effect.
Anyone experienced similar, or know their cause, or how to stop them?
judgepd charging from the Grid using the new GEN2 Inverters it only tops up to 97% ( previous on Gen1 100%) but with solar manages the 100%?
Woodstock Again I am NOT part of the BETA testing but I had a similar behaviour from my system in the recent past. I had previously set the reserve cutoff at 10% on my 9.5kWh Battery, which I ran for a few days, however the battery was correctly going idle at 10% but then continued to leak charge at a rate of about 1% every 3 hours.
I decided to cancel the 10% reserve which I did by going to the Inverter Card / Settings / Battery Options / setting the reserve back to 4%. This was confirmed as done and showed the reserve now as 4%. However the batteryu continued to go idle at the previous figure of 10%. A reset to defaults did not change this behaviour, neither did an inverter reboot.
By chance I was rooting around in the Inverter Card / Remote control and found that under the Percentage Limits table and discovered that one or other (can't remember now which) of the "Battery Reserve & Limit" or "Battery Cutoff & Limit" were still saying 10%.
I changed this back to 4% with the slider and clicked the arrow head that appears to the right of the box. This saved the setting and it has worked correctly since. Hope this sorts it for you too .
Woodstock Just had a thought.
The "fudge" work around includes resetting the battery cutoff limit to 4% and it works on pre beta software. Running a reset to factory defaults under 4075 shows the battery cutoff limit being reset to 4% but it does not stop the spikes.
Is there an additional flag (not visible in remote control) that gets set at the same time as battery cutoff limit? Perhaps this flag is not included a beta factory reset since (as we now know) battery cutoff % limit is no longer a thing in beta.
Have turned something on that I cannot turn off!
CCourtney Thank you for the suggestion. It would have been "Battery Reserve % LImit".
The whole business of changes not taking effect, or not showing as having taken effect, or showing as having taken effect and then reverting back to not showing as having taken effect, is just GivEnergy's attempt to "gaslight" their customers. Not related specifically to this problem but hopefully something that will get fixed. 🙂
Right up until now "Reset to Factory Defaults" has been the one thing that has always done as requested, but today I had my first doubt! Maybe the "gaslighting" is working. 😄
Today's Weather conditions.
Since the inverter reaction time was improved I have noticed that the correcting power swings continue for a much shorter period but are of larger amplitude. What used to be 50W is now 150-200W. That was until today.
A combination of particularly spikey solar generation and the sun getting stronger as spring progresses is seeing my IHD wildly oscillating between 800W import and 800W export. Not sure this is actually a problem but it doesn't look good and the swings will likely only get bigger as summer approaches.
Woodstock Cue lots of queries from people that have devices switched by import/export threshold values that do not have a corresponding minimum duration?
pacemaker Guess what made me go look at my IHD. 😀
pacemaker What is being described it that thread is why the faster inverter reaction time was introduced. Just wanting to ensure that the fix does not have unforeseen consequences.
I am on the latest beta firmwares. But the 2% jump at the top has come back...... It went away. But just seen it again. Hopefully a 1 off?
THALL Interesting that your maximum voltage is just over 1V more than I ever see. Out of interest, what sort of voltage are you seeing when your SOC is showing your batteries to be almost empty?
mine doesn't seem to ever get over 53 at 100%. perhaps because i have not had a calibration or emptied them for months.
Woodstock I thought the faster reaction time was an accident? Something that just happened after a code change. A good bug. ? 😁
Ra 50.3v
100% for my 3x5.2 is 54.2 peak, then settles down to 53 ish.
Any more updates to firmware since I was last here? Just seeing if Dan managed to get my list worked though
@positor1 your update from 3077 was top of the list, as its the oldest.
there were half a dozen on the list i wanted (Asked) as he is the boss doing
TheDragon (GivEnergy)
no change as of 07.30 14th march 23
THALL Interesting....I assume you're not seeing full capacity? Looks like our batteries are phased out to each other slightly, mine operates in a lower voltage band and yours higher.
Ra The dashboard data suggests I am missing less than 7% total capacity. Checking the last couple of days data suggests 6.6%. Though I have only seen the OCV run once. Visibly finding more battery. But my system was empty for a good while before that kicked in. Now with better weather, warmer etc my system does not empty and doesn't fill everyday.
So far so good on 4075 (since Friday I think it was). Capacity seems much better, think the OCV lower half adjustment issues are fixed 🤞
It's hitting 100% or 90% if I set that, and I see it occasionally top up towards the end of the timed charge, so hopefully no SOC crashes will happen because of this new methodology.
Haven't run out of battery yet, before the timed charge kicks in, will have to see how that is handled if/when we get some miserable days. 😁 🌧️
TX200 Wasn't aware of that. Let's hope so!
THALL Is this perhaps just a feature of top end displayed SOC% adjustment? If the displayed SOC% is slightly under reading it jumps a couple of percent?
Woodstock No idea. After the new firmware was applied it went up to 99% and later 100%. Hence I hope this recent 98% was a blip until the OCV runs again. I also couldn't get below 6% at the bottom. Not yet anyway. But actual charge discharge numbers look really good!
Variation on a theme?
I am getting the charge to 100% displayed SOC and then drop back to 99% feature, but with a twist.
ECO mode on, there is excess solar, the displayed SOC is straight lining at 99% and my battery is discharging to the grid at 29-64W. Can't recall if we have seen this one before.
Woodstock Hmm checked my 100% charge from solar data from yesterday and while pegged at 100% it shows constant charge of 145w
THALL Yes, could be a blip or may just be the flip side of the battery continuing to charge for a short while after it has reached 100% displayed SOC. One happening when the displayed SOC% is slightly under reading and other when it is slightly over reading.
THALL Over what sort of a period? 145W is suspiciously close to a ramping set point.
Woodstock 1440 to 1610
THALL Hmm, that is a bit long to be just a top end adjustment. So between us we now have a constant charge and a constant discharge at "full". 😧
Woodstock I'm not seeing any weird charges and discharges at the top end but my system has a distinct phobia of 100% and seems much more comfortable just sitting at 99% instead.
judgepd Mine is sitting at 99% too, but it is constantly discharging and throwing in a blip charge (421W) every now and again. This short duration charge takes the battery back up to 100% before it immediately drops back to 99% and resumes a constant discharge! Do I recall someone else seeing this and saying they were waiting to see what happened when the sun went in?
Top end adjustments visible on solar only here.
Then settles at 100 as it is expected to.
I'm away with my current job, hence I'm a bit quieter than usual. The current place has a phone ban.
Those know defence, some areas are no go for phones.
TheDragon (GivEnergy) Those know defence, some areas are no go for phones
and anything that goes in must not come out..its a little known newton law...or was it einstein.
One of the more interesting repairs i did was to send the parts there and talk the person through fitting them.
positor1 Yes we have placed like that too. Dell & HP to name a few, hate it, as they don't get the broke bits back. There're shredded
Yep, areas of Porton Down for one....
Woodstock I'm seeing similar strangeness around 100%. Discharging / charging in hour 3 of 4 scheduled charge, once its "full". Seems to be a new quirk No solar just charging a 9.5 overnight om Octopus Go.
Woodstock haven't left it at 100% long enough to see if it drops to 99%, but I too am seeing the small discharge at full and with excess solar. I imagine given enough time, it would drop and charge back. The larger discharge spikes are genuine demand.
First two nights on 4075, charged to 100%, saw no discharge during the timed charge. Five small top ups first night, one the second.
Next night, 90% target, two small top ups, 60-160W discharge when not charging (during timed charge) - wondered if that is either the battery powering the inverter and itself and/or the system trying to get the voltage right to get to true 90%?
This morning, 90% target, after main charge, one small top up, and around that mostly 60W discharge, and a little circa 160W, until end of timed charge period.
And graphs :
TX200 I think a bug fix is in progress on this one. If you selected a target soc that is not 100%, then the system oscillates around it. This behaviour was also present in the standard firmware, but seems worse on beta, especially if you have a heavy load at the same time. I've had it oscillating +/-3kW, this added an unnecessary 4kWh of throughout. I now set discharge to zero for the charge period via HA, which helps a bit.
Logging via HA, so 10 second resolution
Same chart on the app
mpartington yeah that looks a bit messy on yours. I'm not too fussed about a little bit here, a little bit there if I know that 90% is actually 90%, but not great for you there.
TX200 Thanks for your graphs, as Matt says, its on my list. The lower you set the % charge to, the more obvious the oscillation is.
Its been there for a while, it looks to be made worse in beta, by the fact a pause of charge or discharge leaks 112w.
This is another 2 lines on my list, as its 2 registers affected for pause charge & discharge. 2 entries in the bug list.
So these are gonna change for sure.
There are a few issues all interlinked, which is why the next revision of inverter FW is a longer piece of work, not one fix at a time, this is a series linked in with each other.
As you know, I'm on the case, info will continue with updates as i get them
Over and Out.
TheDragon (GivEnergy) Not only did I have the phone ban but also the possiblity I would get locked in for days with no way of communicating out until the "All Clear" if there was an "incident."
Not sure they’d want some of those parts back 😂😂😂
mpartington Yes, looks similar to mine.
The dip from 99% to 97% SOC in the middle was me turning off the solar PV to see if I could jolt the system out of its constant discharge state. It then behaved as you would expect it to in ECO mode (apart from the grid export spike) but the discharge resumed when I turned the PV on again and the battery had recharged to 99%.
Only just seen this. SoC in multi battery installations are drifting apart? That's a bit depressing. I was facing this issue last year (and it's listed as issue 4 here: https://community.givenergy.cloud/d/237-reported-issues-and-resolutions).
Without wading through the full thread, is this actually a problem we're still facing even in the 4xxx firmware series? Is there a known fix or workaround or is it "new" or yet to be addressed?
New bug to report. I am on 289 and 4075. I recently mentioned I have missing dat on my reports. GE Support have looked at this and beleive it is a firmware issue. My solar generation is being logged as AC charge. At request of GE support I have forwarded all my data to the beta email thread. This does make reporting unreliable. Day to day use is fine, but I am trying to check my data to work out if changing to Flux is worth it yet so this will be a bigger issue for some.
bjb Multi battery on one inverter, track well. The tolerance is 5%
Mine and other track better than this.
Grossly different sized batteries would be a nice addition to the beta to test. Anyone got a 9.5 and 2.6?
bjb the voltages allign perfectly, however on my usage cycles, the SoCs do seem to get out of line at times. Good news is the software does bring them back together eventually. I'm probably the worst case, as I rarely fully cycle my batteries (except in winter)
I have 2x 5.2 batteries
What I do find is even if the SoC are different, the inverter seems to be operating to the voltages. So at 'empty', even if I have a split in soc, the voltages of the 2 will be the same and consistent with what I'd expect for being empty (or full). So I'd say it appears more of an indication issue than functional.
Same chart for voltage
If you only looked at the system SoC this would be invisible and trends as expected. No SoC jumps/drops for me on 4075. Overall I think it's looking good. There will always be more refinements that could be done over time.
(Just to add, the SoC splits on 3007 definitely had a functional impact)
.
TheDragon (GivEnergy) 5%? My two batteries are currently at 43 and 31% respectively. Though, as mentioned, it seems cosmetic.
kai 59% and 43 % for me at the moment. Seems to be getting wider the further down the discharge cycle I get. I'm deliberately not changing my usage habits. I don't think they lock together as well on 4075 (that could simply be they're not being cycled as deeply now the sun is out), but conversely the overall soc curve seems much better (for me)
TheDragon (GivEnergy) I’ve got a 5.2 and 8.2 on an AC3 inverter that have drifted apart over the past few weeks. Not currently on the beta but would be happy to and report back.
One of the criteria for OCV is within 5%
So if and when OCV runs, one or both will adjust via lookup of not within tolerance, one or other will lookup again.
As if voltage is same, SoC reading should also be, negating battery cell tolerances, no 2 packs are exactly the same through ALL of their charge curves.
Had a calibration done yesterday.
The 9.5kwh, put in 10.5kwh. It normally puts in 8.5kwh on 4075. I tend to get around 7.8kwh out. It drew about 950w for a two hour period after it thought it was full. I assume this is my missing 1.7kwh that I never get to use.
By the time the next charge window came around it had still had 84% battery. It shot up to 100% fairly quickly in less than an hour. It consumed just over 1.5kwh to get there. I assume it should have been fairly confident where 100% was, as it had just been there a few hours before. Instead it cycled between discharging and charging for the next 3 hours when I assume it should have been idle, and ended up charging another 0.7kwh, topping itself back up "100%". Total charge was 2.1kwh
Is this searching for the top, or a fault?
i think it is in line with what others have said, corrections i think are limited to a 24 hour repeat so one day is too short to say for certain.
Jamesrich That is normal and expected behaviour.
OCV levelling operates in a few different circumstances, top end, i call it hunting. But its effectively the same.
Charge, check, charge, check.
Is has no bearing on your use of the battery, as it would have been idle till your charging finished anyway. If your situation demands power from the battery, it will provide as normal.
It will go back to hunting next time it is full. Only when in the check phase its is satisfied, OCV flag set.
However top end it will do another each time. Its not limited to 24hr like the low end, under 4A BMS, or manually initiated (coming soon)
Another drop last night. Still not hitting the higher end of the voltage scale when 'fully' charged and not seeing full capacity of battery. Will also forward this to Dan for his information.
Ra OK cheers. I'm away with my old job right so access to internet is limited.
You need the later Inverter FW to resolve the 95% shenanigans
What i see is consistent with the latest revision -1 of the FW. Dan has a list of the outstanding FW upgrades to bring everyone I know about to the latest Inverter and BMS
Ra I'm looking on a very small screen, but is bottom end voltage really hitting 43.2V? That is incredibly low. I'm lucky to see under 51V at a similar discharge at empty
TheDragon (GivEnergy)
no change here. still
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
3077
The end of "free energy" (AKA "inverter gains")?
Since 06:30 on Sunday 12/3/23 my battery and inverter power readings have shown no instances of the above. This is irrespective of the power level. The readings look believable throughout the power range from low to high. The only thing of any note that occurred before they corrected themselves was a (battery fully discharged) Displayed SOC% adjustment from 5% to 2%. The inverter still shows 5-10W when the battery is idle (0W) but that is the only remaining anomaly.
May be a coincidence but it was the first time my battery reached 100% SOC after this that I started getting the constant discharge at 99% and blip recharges to 100% issue.
Thats a shame, that could have well been an awesome marketing line, "Free Energy for low users".
Alas, no such free energy. Cant be breaking the laws of psychics can we.......
Apologies, Dan is a busy man. I have another engineer with a super user login I can ask.
Soon It will be me doing it directly, everyone will be on the same within days of a new release being announced here.
TheDragon (GivEnergy) did it ever get clarified if “mature” inverters are going to be on the same FW branch or do they work slightly differently to get around the whole capacitor thing.
(Apologies in advance, I know someone has 2 old ones on test but from my rummaging I’m sure there’s been about 4 revisions of the Controller PCB and I’ve no idea if the “old” ones on test are as early as mine!)
Side note: as you know I don’t suffer the bug so I’m fine either way if I’m to become a “deprecated” install.
TheDragon (GivEnergy) And any news on what will happen to those of us with the older Gen 1s, is there going to be a hardware upgrade?
Woodstock The remaining anomaly is more subtle than I first thought and can be described as ....
When inverter power is < -8W ................... battery power reads correctly.
When inverter power is -8W to -1W ........ battery power always reads 0.
When inverter power is 0W ......................... battery power reads correctly.
When inverter power is 1W to 8W ........... battery power always reads 0.
When inverter power is > 8W .................... battery power reads correctly.
Woodstock interesting- did you observe this before the beta FW as well?
chris_p I do indeed have some news on that, yet to be confirmed how, when, but there is a solution yes.
Inverters 2209 or older, need an engineer visit, to run a test FW looking for the affected issue present or not. If present, an engineer will rectify and load the proper faster FW.
Internal Alpha testing has produced a working option for the table.
So rest assured, just because your inverter is on the naughty list, It can, if you wish be updated to the faster switching FW release, the newer ones have.
TheDragon (GivEnergy) ah nice, so just to clarify "will rectify" does that mean make hardware changes as well as loading the firmware?
TheDragon (GivEnergy) how do we tell how “old” our inverters are wrt the 2209? Is it just first 4 digits on the serial number you see on the web portal? (Sure I asked before but can’t find the post!)
Smk82 pretty sure it's two digit year followed by two digit week number. Yes, from the inverter serial
Believe the issue is only with the older Gen1 hybrids?
I'm going to assume my Sept 2020 is well and truly in the affected pile...
hoggy Not necessarily.
But engineer needed to load a test, if it passes, it can accept the new FW. If fails, some work needed.
Can discuss on Monday's meeting as a side huddle
Talk about transparency, well today not all good news
I have just heard ref a minor internal test for hybrid Gen 1
Fixing the tracker separation, for some, unknown as yet reason, the battery discharge slider has no effect.
Hence no beta version D0.197-A0-197 (198 was also an internal) released. why waste your time, when the bug was caught internally by the alpha test team. Yes we have a real structured layers of testing, with QA at each step.
This will once correct, will trigger a Gen2 update too, as it has the same tracker issue, in its production firmware. Such as small amount of users affected, but its there.
If confirmed by 2 testers, not a glitch in the matrix - So the FW passed back for re-work.
@joshgreenuk @ Mark for passing this info on, and formally reporting as a bug back to betatesting@givenergy.co.uk
For now
Latest hybrid Gen 1 remains D0.199-A0.199
Latest Gen2 D0.999-A0.999
Latest AC3 D0.289-A0-289
Latest BMS 4075
Regards Paul.
TheDragon (GivEnergy) ill test the 2.6 alongside mine if they want to lend it?
joshgreenuk
do they lend batteries for testing?
positor1 I doubt it.
I realise he was probably joking now. sorry I seem to have an undiagnosed condition where i take things literally. man itus? (humour)
TheDragon (GivEnergy) Sorry for the delayed response, I've also been away with work so difficult to get on here.
Is 4075 not the latest FW?
Ra
4075 is the latest BMS (battery firmware)
positor1 Well thats what I'm on and I've seen about 3/4 drops since I had it loaded. If anyone wants a definitive figure I could go back through my data and find out the cadence etc
Ra drops in Soc ?
Ra
Couldn't they be valid drops where the new BMS is re calculating? apart from fluctuations at the top I am reading that the predominant SOC problem is sorted.
I haven’t had drops or spikes in Soc longer than I can remember and do not cycle the batteries often now either
best report it then with graphs in case this is something new and worrying.
Maybe not for this thread, but as Paul is involved I thought I might as well put it here.
Shoot me down if it is already out there!
I get the impression there is a lot of user interpretation of the graphs and reports on here, which I think leads to a lot of misunderstanding and therefore wasting of effort.
I am aware and have read through the wiki (knowledge base) but I am pretty sure there is nothing in there, so:
Would it be a good idea GE to look at producing an offical easy to read "Idots Guide" (if you can remember the windows books!) to the graphs and reports and how to interpret them? Me included!!
I'm sure it would help existing users and future users.
Something worth persuing?
joshgreenuk Yes, I posted the last couple on here but also send them to Dan at GivEnergy for his information
Stevef1964 Giv will just look at the graphs their side with the logs and interpret them correctly when reported, I don't think this is really that important as such but still worthwhile exploring when the bigger issues are completely resolved.
positor1 I don't believe so, dropping to 43.2V isn't great (about 20%-0% crash)
On the flip side I'm not seeing the higher end of the voltage scale either, tops out at about 54V
TheDragon (GivEnergy) I assume those are latest beta firmware versions?
Is there a plan to update (and keep updated) the https://kb.givenergy.cloud/ firmware revision documents? ( I am saying update because the firmware version I have is not listed there and I'm not in beta program)
thanks
Smk82 No. Zero was always zero and inverter always showed losses and never gains.
Unfortunately I too saw a sudden dip on 4075 Wednesday (I had only just commented on here how well it was working - sods law!). I have been operating between 50 and 100% for the previous 7 days and the system last performed an OCV bottom end cal on 10th March. I regularly hit 100% in that time. The resting voltage at empty was a more 'normal' 51.7V (compared to @Ra ) suggesting it probably was near to the intended bottom limit.
It's been reported, but staring to make me think that a regular proactive manual/timed OCV will be necessary for some usage patterns, unless the clever learning algorithms can be further improved. This isn't as bad as it sounds as it could easily be scheduled for first hour of cheap rate, or at a point where solar is covering baseload with little export. So long as there are not overly restrictive soc boundaries between which it must be performed
And at battery level (2x 5.2kWh)
joshgreenuk out of interest, do you see a SoC split opening up between your batteries when partially cycling? My 2 seem to wander apart (voltage is spot on)
Might be worth taking them down to near zero as an experiment as mine were fine until I hit 30% for the first time in a week. Before that I was only going down to 50%
mpartington Have you seen any upward bottom half displayed SOC% adjustments?
Woodstock not yet. I saw plenty of (unnecessary) downward ones on software prior to 4075 though.
I was running approx 230W baseload prior to the drop. After the drop I did have 50 mins at idle before my charge kicked in, so in this case it appears that the SoC was over reading and there was nothing to recover.
Ra, I see you have a Hybrid, and are a couple of versions behind on the Inverter.
Can we load D0.196-A0.196.
alfwro They are BETA. So not on the KB.
Once final, they will be
Long call with Dan, inbound today/overnight new revisions for Inverter.
I also see 4074 BMS is live in the admin portal on 2 batteries (in China) So also expecting a BMS update too
Really confusing when the version numbers go backwards and then jump 1000 up and now back to going backwards!
chris_p Thats fine, Im tracking them.
There is a method to their chaos
TheDragon (GivEnergy) Do you mean the testers are going to see new firmwares for inverters and batteries from late Monday? Don't suppose you have the key notes that you can share for what these are fixing?
THALL When I know I will summerise.
Multiple, time slots, different Pause method, Both need subsequent portal updates, so wont see till then.
PV Tracker seperation
To name a few. I will sumerise what I get for each inverter.
Im also getting my logins hopefully this weekend, so my 1st task will be tracking all the existing BETA, 56 of them, documenting each one on a spreadsheet and then bringing them all to the same level for each inverter and battery type.
If i get it today, I will likly ave a busy day tomorrow. Which I like
Think that's my first lower half adjustment since getting 4075. 9% to 14% 😀
So far, no issues spotted. Think it's been a week now on it.
TheDragon (GivEnergy) Oh sorry. As you mentioned them I assumed you may have already known.
I don't mind new firmware getting slapped out. But as a tester I know I can be more valuable if I knew what each version was meant to fix/change.
mpartington OK, in that case it's looking as though top end adjustment alone is insufficient to keep the SOC% in line. 😞
My normal usage pattern is similar to yours (with an even lower base load). Since getting 4075 I have been concentrating on forcing top and bottom end processing and have not just "let it run". It seems likely that if I were to do so then I would be seeing the same as you.
Try and let it run, your system is different. Could well have a different effect, therfore better data to feed back.
However new Inverter and BMS due soon
Had an adjustment up from 5% to 13% last night but I've not seen any jumps or other adjustments. Could do with something happening at the top end as my system has a real phobia of 100% and will only hit 99% on an overnight charge but that certainly isn't a deal breaker, more of a display annoyance.
TX200 Glad it's not just me! Now all we need to know is if the adjustment was intentional or accidental. 😉
judgepd Up fro 5 to 13 would be intentional. As I mentioned way way back up, Oh ive got some spare power now Ive checked, here have it. So 5 creeps up, if its goes 8 or more, it will release this and you will see a burst of free energy
TheDragon (GivEnergy) Yeah, I assumed this was a sign of the firmware working as it should and certainly wouldn't consider it an SOC 'jump'.
Woodstock yeah, I've deliberately let it do its thing, with a few tests thrown in here and there
Me too. Tried to get a mix of scenarios in over the week. E.g. horrible weather, charge to 100%, good weather, charge to a lower percentage overnight, saving session (45% overnight charge and mid-afternoon charge too) etc. 😁
Any Gen2 Inverter owners in here? We have a FW build to test, to confirm a few things
Ideally 1 or 2 south facing normal users, and 1 or 2 east west
Serial and setup, batteries, PV etc. if interested in being part of the test team
TheDragon (GivEnergy) , what is the PV tracker update supposed to achieve?
Multiple, time slots, different Pause method, Both need subsequent portal updates, so wont see till then.
PV Tracker seperation
To name a few. I will sumerise what I get for each inverter.
I have recently swapped out a Solax (non-battery) inverter for the HY5-Gen2 and with East/West strings have noticed the PV string currents do not vary so much over the day on the GE inverter. So far I cannot tell if this is affecting the overall power output as we have not had any clear blue sky days recently.
Simon_C A low input on one tracker can pull down the other. Similar seen on Gen 1 Hybrids too.
Hence i need to test them both
TheDragon (GivEnergy) , is there a specific voltage on 1 of the strings that I should be looking out for?
Presumably at the beginning/end of the day when the West/East facing panels are shaded by the house roof.
Simon_C It seems to be when under 100v, ie one tracker has gone offline.
So if the sun is in the east, morning, the east string shoud be performing, not asleep. Likewise the end of the day, vice versa
But you’re saying it is believed to be fixed in the soon to be released beta-next version of inverter firmware?
TheDragon (GivEnergy) , this is what I have collected from the Solax inverter on a blue sky day (20/1/2023).
Note - PV1 (dark blue line) is West facing 10x410 W, PV2 (pale blue line) is East facing 6x410 W.
I will see what I can dig out from the HY5-Gen2 logs.
Simon_C almost text book, West biased
TheDragon (GivEnergy) , the best I can find in the time I have had the HY5-Gen2 is this.
String1 = West - 10 x 410 W
String2 = East - 6 x 410 W
The graph is somewhat confusing as I cannot change the trace colours.
Before 09:30, the East string is the highest output, after it is the West.
Adding the total power just confuses even more with 3 shades of yellow that are barely distinguishable.
TheDragon (GivEnergy) I've two Gen2 inverters, one runs East only on one string (9 x 390W panels) and one runs West only one string and same amount of panels.
If you need the Gen1 and as you already thought our system would benefit from the Beta (due to our SOC drift related High Voltage errors), then system info as below. As you can just about make out from the overlapping shades of yellow...both our strings are pretty spot on South facers.
SD2148G201: Hybrid 5.0 Gen1 (D0.450 A0.451) 8.5kWh Battery (3008)
mattpdavey I do need Gen1 too.
But need some Gen2 1st.
My plan is 50 of each would be a good sample set
94jdh sent you a FB message
No problem, well feel free to sign us up at first opportunity. We have only circa 70% useable battery at the moment so will be tapping Martin up for yet another re-cal next week at this rate anyway. I would be surprised if the beta could be any worse than how it currently operates 🙂
TheDragon (GivEnergy) replied on Facebook
mattpdavey I’m in the same boat sadly
TheDragon (GivEnergy) I have two strings pointing directly south, but my inverter is a Gen 1. If its of any use?
alfwro Are you wanting a ride on the lifeboat?
Soon I'll have full control and am expanding the beta. Beginning with this new FW I get this weekend (Fingers crossed) from China
Stevef1964 hey why not.
I'll chase my admin access and firmware files from China.
I don't have an EV to charge and I don't do any time discharging or fiddling around. I just leave mine in ECO mode.
I do have an Eddi which is set to only kick in when the exporting to the grid exceeds 200W and a reaction time of 5 seconds.
All fairly straightforward and simple set up.
So unless you want to do anything specific it will stay in that state. Hopefully should then give you some stable data.
TheDragon (GivEnergy)
By the looks of it, my inverter firmware is not on the latest?
BMS 4075
Inverter D0.104-A0.104
I'm SA2211G234
TheDragon (GivEnergy) Hello, yes that is fine. Many thanks!!
Just brought up the power graph and my immediate reaction was WTF is going on with the SOC% curve now!
Turned out to be a dog hair on the computer screen in just the right (wrong?) place.
It's been a long day. 😄
Woodstock HaHa.
Don't do that to me.
Gen 1 hybrid user here 🙋 willing to join the beta... If @TheDragon (GivEnergy) will have me 😬
So unfortunately 4 days after the last, I've seen another SoC drop from 23%. It discharged 100Wh in that period.
Those who are now not fully cycling, I'd really recommend throwing in a deep discharge at least one a week as part of your testing to see if you get similar results.
4075
D0.289-A0.289
2x 5.2kWh batteries
mpartington Now solar is holding the battery at a displayed 100% SOC for most of the daylight hours I am noticing a steady decline in battery voltage. Yesterday the battery started at around 52.6V but was down to 52.00V when the sun finally disappeared. The displayed SOC, however, was still showing 100%. Currently waiting for the new firmware release but you can bet I will be including a deep discharge in the testing!
(Under 3007 I was seeing 52.8V at a displayed 100% SOC and there was a minimal drop throughout the day).
@TheDragon (GivEnergy) Not a direct beta question...well, kind of. I'm going to be adding another 8.2kWh battery to my existing setup. I'm guessing I need to let my installer know that I'm in the beta as it will need the beta firmware adding when they install it.
judgepd That is correct. Would better in this situation to have the BMS bin fine sent to your installer to load on install day.
Then Re-Commission
TheDragon (GivEnergy) Thanks. I don't have a firm installation date yet so I'll see what the state of play is with the beta firmware when I have more details.
@"judgepd" Just wondering, 1 I didn't realise the 8.2's were in stock anymore? Did you find one or did your installer? 2 was the 8.2 that much cheaper than a 9.5? I went with the 9.5 for a my second as no 8.2's available etc. At the time it sounded like the same price as a 8.2 (October last year)
Ra Just checking TheDragon (GivEnergy) that you'd seen I was ok with the inverter update. I know you've not started yet so no rush. Only bumping this in case my reply got lost in the thread.
Many thanks!!
THALL A company I haven't used before were advertising 8.2s installed for what I felt was a very reasonable price. Wasn't going to increase the storage but couldn't resist and to be honest I haven't compared it to the price of a 9.5. I think I got a bit giddy when the price was so much cheaper than my original installer quoted a few months ago and didn't ask any questions!
judgepd Fair 😃
We have out 1st Gen2 on the very latest (Like i don't even have it yet)
D0.998-A0.998 Gen 2 Hybrid Firmware
4074 BMS
This is a separate trial just for Gen2 - As requested over the weekend
Ra Same for me! 🙂
Ra Wont get lost, as I have a spread sheet with all the data on it, all coloured who needs what.
Still waiting on my login
2 more done
TheDragon (GivEnergy) We have out 1st Gen2 on the very latest (Like i don't even have it yet)
D0.998-A0.998 Gen 2 Hybrid Firmware
4074 BMSThis is a separate trial just for Gen2 - As requested over the weekend
nwc469 Your firmware is attempted to be loaded at the moment. I have the ear of a field engineer at a loose end
Loading
D0.289-A0.289
BMS 4074
BMS failing. Grrrrrr
Seems to be sucessful now bms 4074 DO289 AO 289 tnx
You have BMS 4074 now too
Your upgrade tried, but is failing. May need to do manually. Please email support for the file to load manually.
If using GivTCP, set the self run to 30s or longer
You on D0.289-A0-289
BMS 4074 too now
So out of all the ones done, 3 failures, @Woodlands and @ianfretwell
and @BobSmith all BMS fail.
Which is easy to do locally - contact support for BMS 4074
Thanks for the input tonight - hopefully I can add some positive news to this update.
TheDragon (GivEnergy) Brilliant, thanks Paul.
Is this the 'end of month' update that's due to address numerous things on your list?
By the way, seeing much better performance out of my 9.5kwh now. Up to 8.8kwh today before it stopped dicharging, but its just found another 3% adjusted itself to 8%, and now discharging again!
Funny it jumped 3% when I got updated to 4075.
I’m really hoping to see some results as I now have a price to change to a libbi and to a solaredge storage system and will need to decide in the next day or so.
My major concerns are:
1) random discharges
2) East/west performance
3) capacity
4) discharge rates , eg when will a g2 inverter/battery really put out 3.6kW
Cheers for upgrading.
Another little adjustment from 7% to 12% 😀
Charging rates set in MyInverter not respected again (just like when moved from 3011 to 4075).
Assuming this is a one off thing post firmware updates.
I'll not make any changes this time and see if tomorrow's charge is as set in the portal.
Jamesrich no this is a BMS update for AC3.
I'm still waiting for the inverter updates.
The Gen2 is a separate set of tests, for PV tracking east West.
Chucked in latest BMS aswell.
Gen1 waiting for the Gen2 alpha tests
My 1st priority is those on the beta, all get to the same "Current " version"
Once we're all on the same song sheet.
A test schedule will be written to test the 8 known issues I'm tracking
For now, BMS 4074, any jumps, any odd SoC movement.
Flat to full kwh would be nice to know, over a few days. Not just one. I'm looking for average capacity.
4075 I was seeing 11.3 to 13.3 quite a range for 3x5.2s
No, not running GivTCP,. However, at about 9PM last night the SOC dropped from about 80% to zero, then charged at 290 watts until 11:30pm when it started charging at max rate. Then according to the graph, jumped from about 40% to 100% at 3AM this morning.
Remote update of the BMS has never worked before, so can you @TheDragon send me the new file and I will try and update it.
TheDragon (GivEnergy)
Yes i found
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
4074
thank you
9pm would be me and martin trying to force the firmware over. It was having none of it.
If you email support and request they send you over BMS version 4074 for manual load. Add in Paul & Martin have tried to load OTA and failed 3 x
As you are aware of OTA failing, I assume you know the process, if not, support can assist and guide you
Regards. Paul
Interesting little SOC blip after 4074 firmware applied. nice to see it was measuring.
positor1 Yeah, that was me. Cant hide. Mine did it too. You gained about 3% SoC looking at the graph. as the lookups are slightly ammended
You can slap me with 4074 too if you like. Though my only issue with the 289 and 4075 bundle appears to be missing solar to battery data.
After the updates for the new firmware and BMS it seems both Inverters missed their charging schedule, one fully and one partially.
Will monitor it tonight to see if it was a one-off.
TheDragon (GivEnergy)
I half expected to see the newer inverter firmware too? (290 290?)
Thanks Paul. All in all, i got 9.42kwh out of the 9.5kwh battery yesterday, which is a huge step forwards. Even on 3008, and the calibration I had on 3077 didn't give me anything close to that, topped out at 7.3kwh-ish.
However, i only put in 9.35kwh to get back to '100% / 98%'. It feels like the hunting at the top takes too long to play out in a Octopus Go 4 hour charge window (no solar here). It now takes 3.5 hours to get to 100%, due to the new slow down towards the top. Then it only has circa 30 mins for hunting, could it hunt with a bit more power?
I know on a calibration the Grid to Battery pulls 10.5kwh in to fill it from 0%. So I guess I'd expect a maximum daily import of 10kwh, 10.5 - 4% reserve.
Will aim to fully discharge / charge over the next few days and compile some data. If you can let us know what data you are looking for, happy to pull it together.
TheDragon (GivEnergy) Thanks - I've sent them a mail.
I know we mentioned this a little while back but I still feel that the failures in my case (which is every single attempt made remotely except the very first one) are more likely to be down to an issue writing to the USB drive element of the dongle - can it be reformatted as long as I save the MSN file to put back on afterwards ?
IT (inc. networking) is my day job and whilst I pretty much loath wi-fi for it's general inability to be any use for anything I'm not seeing any drops in connection to the dongle at all which makes me think it's not a connection issue as such.
Thanks.
positor1 289 is later.
Beta counts down, production count up
Didnt have your serial number
Just to be safe, hit restore to defaults, the re-enter them.
If the FW did this, thats a new bug, never bothered mine
James at the moment, Im wanting from BMS 4074
Capacity in relation to design over a few cycles, not just 1
SoC across packs runs pretty similar across the whole plot, current I see 4075 one is 10% less than other 2, as does @mpartington
Lower power users, dont see OCV pulling down the SoC unnecessary
likewise, slow reserve topup OCV doesnt rise artificially
TheDragon (GivEnergy) Ohhh I thought you did. Here you go CE2105G149. Happy to test whatever you've got. Just let me know what test you want done.
I have never tried that, If you put the Dongle in a laptop, the only file on it should be msn.txt
on a normal day
I have emailed you the 4074 BMS to load
I've restored them to default, but both are back offline again - I'm nipping home a lunch so I'll drop out the dongles again
4075 not 4074.
Following the 52.00V at a displayed 100% SOC on Sunday there was no sun yesterday so the battery dropped into the lower half last night. Had a lower half adjustment from 46% to 17% in the early hours. I suspect that if the conditions for this to occur had not prevailed then there would have been a sudden SOC% drop later on.
Charge spike was not of my making and I am unsure if it is associated with completion of the SOC% adjustment (not seen one before) or if it is a 01:00 charge spike (have seen these) moved to 02:00. The charge spike was recorded on the 01:58:52 data point.
System had been running in "reset to defaults" mode the last couple of days without any intervention from me.
TheDragon (GivEnergy) OK, that's 4074 up and running for me.
I deleted the ARMStoreT.bin file that existed on the dongle at the same time (leaving the MSN file).
Then ran a chkdsk against the dongle - this claimed to have found errors and repaired them.
The ARMStoreT.bin file re-appeared (!).
Deleted it again and re-ran chkdsk - no errors this time and the file did not re-appear.
Maybe this has cleared the issue for next time...
Thanks again.
ianfretwell Good to know chkdsk found issues
Yeah didn't have your serial number
@TheDragon (GivEnergy) If you're collecting serial numbers, mine is CE2105G039.
TheDragon (GivEnergy) No problem. CE2122G003.
Woodstock looks like that was heading the same direction as my 2 sudden drops on 4075.
Glad in some ways I'm not an isolated case. Fingers crossed for 4074
Biggest SoC crash to date, 58% down to Zero 🙁
Whats the latest? How far away are we from a full release, also is it the case with the new firmware that you lose some capacity and if so what?
Thanks, Keith
'Free energy' is notably reduced in 4074.
Seems to be between 5 and 9w, was 21w on 4075.
Was there ever an explanation for this? Pre beta, battery discharge was always slightly more than consumption.
I thought i had always been on 289 289 , must have had that back in jan feb then?
CE2209G233 for me. Still on 4075
Jamesrich Hope so but I am sceptical. Under 4075 I was getting "free energy" in at least three different ranges. Sometimes it was 1-8W, sometimes 15-20W and sometimes 24-29W (this was in addition to the inverter/battery power dependant "free energy"). Just received 4074 so I will keep an eye on it.
thanks @TheDragon (GivEnergy) !
kai CE2209G233 for me. Still on 4075
Done
@TheDragon (GivEnergy) Are you actually working for GE now or just poking around? :-)
You can slap mine too if you're doing a load. CE2105G149
TheDragon (GivEnergy) Very early days, but on 4074 my system has found it's love for 100% SOC again. On 4075 when charging on solar it always dropped back to 99% immediately after hitting 100% but it's been at 100% for at least 45 minutes now.
I want to see 100% again please 😃
@TheDragon (GivEnergy) Please go ahead when you can and update SA2211G234 to BMS 4074 and latest Gen1 hybrid inverter firmware (I'm still on D0.104-A0.104).
I have had some occassional battery discharge to grid behaviour appear I'd like to check is fixed.
Thanks!
TheDragon (GivEnergy) Woodlands
9pm would be me and martin trying to force the firmware over. It was having none of it.
If you email support and request they send you over BMS version 4074 for manual load. Add in Paul & Martin have tried to load OTA and failed 3 xAs you are aware of OTA failing, I assume you know the process, if not, support can assist and guide you
Regards. Paul
TheDragon (GivEnergy) OK, Just emailed support to request the file. Only just seen your reply, as been out all day.
PS. I assume (and hope) that I don't need to request a calibration for this version?
Ra G014.
I'm starting hybrids 2 night
On way home now, leave one job start next.
Lol
1 more day or dual jobs
TheDragon (GivEnergy) Just been reading through the bones of this, having experienced this on my 9.5 battery on Saturday.
I was on live chat earlier to ask what had happened and was told this firmware is in the last stage of release and has "already gone through beta".
Is that and accurate description of where this is? Beta completed and now being rolled out?
Atkinstagram I wouldn't say last stage. That wouldn't be an accurate statement. Near the end yes, if no issues in this release of BMS, then no issues in the 3 inverter types then this version could well end up being production.
But I'm not making that call till the latest iteration of the inverters is out.
AC3 has at least one more version.
Gen2 could be the last. Gen 1 waiting on good Gen 2 tests
Depending on test results on a larger sample set
TheDragon (GivEnergy) i do not think hybrids are on their last unless you are referring to one last one from 196 (soon hopefully)
@THALL - 4074 loaded
joshgreenuk The notes for 197 say its address splitt MPPT.
If not I will not close that line off
paultottie
Now on D0-197-A0.197
BMS 4074
TheDragon (GivEnergy) Many thanks. Is this a side affect of applying it? See below.
BMS Over Voltage 2023-03-21 19:54:06
THALL Yes, some spike up, some down during the load
TheDragon (GivEnergy) please could you do CE2146G203. Still on D0.289-A0.289 and 3011 beta. Cheers!
The new firmware seems to have ignored my AC charging schedule again ( even after a reset) the Inverters sit in the idle position.
Charge 2 post 4074 applied, still charging at max rate rather than specified rate of 1508
I'll try changing it slightly in the portal and see if that fixes it like previously on the upgrade from 3011 to 4075.
Could you update mine
SD2227G291
Not part of the Testing but this is the first time my battery has stayed atl 100% after night charge rather then dropping to 98%
Inverter Firmware Version
D0.904-A0.904
Battery Firmware Version
3007
Cannot see any software changes.
Rodmac76 That's the dream right there!
@TheDragon (GivEnergy) I'm currently flattening my battery every day to get you some data on max charge in, max charge out.
Is it worth me extending my charge window, from 4 to 6 hours? 4 hours does not appear to be enough on the beta firmware, due to the ramp down / hunting.
Prebeta I used to do a constant 2.8kwh in, until 100%. On beta it tails off once it gets to 90%.
Is this behavoir likely to stick around post beta?
I'm only on Go at present, but likely move to IO with 6 hours offpeak in the summer, so it won't be a big issue for me going forwards. I can see some customers caught out not being able to fully charge / hunt in 4 hours.
Jamesrich Just checked a previous overnight charge to 100% for me and it took 55 minutes to get from 94% to 100% which isn't great to be honest but maybe it was the firmware trying to work out where 100% really is.
stevelewis CE2146G203
BMS update in progress
Jamesrich If you can, get 30mins idle at full, that would be for me the filled to the brim, from idle flat.
The document the kWh each day, over 7 days. Average it
then compare to design
mine is 15.6kWh design, 12.4-13.4 kWh in each day
12.9/15.6 = 82% DoD. PASS
The tapered charge at the end will stay, the hunting at 98/100 gets less with time. My charging has increased by 20 mins overall. In return for smoother charging and no big jumps
TheDragon (GivEnergy) I did not manage to fully empty this morning. But currently doing a 16 - 100% charge. Is that close enough or do you require a down to 4% test?
TheDragon (GivEnergy) thanks Paul
I never see what I would call completely idle. There is always a game of 'battery backgammon' going on, which is doesn't seem to have an end in sight.
On this day I'd hit 100% in an hour of charge, but it spent the next 3 hours bouncing around 'hunting'
Are you looking for 30 minutes after it hits first hits 100%, and should be idle, but hunting? 
More so you feel the capacity is where it should be
Ra It didn't last, back to being stuck at 99% today.
Jamesrich Idle for a while. Here are 2 examples, one overnight full and idle
Solar full and surplus
I have some space for Gen1.
Best time to load the new BMS is when flat. Saves me having to sit and babysit the calibration
TheDragon (GivEnergy) Ooooh exciting, serial is SA2111G035. It will be closer to "flat" first thing in the morning, currently 88%. I can stop the overnight timed charge, although I probably would anyway as cloud cover is looking ok for tomorrow.
TheDragon (GivEnergy) Best time for me is also in the morning as battery is low.
Serial SA2208G029
Current Inverter F/W: D0.499-A0.499
Current Battery F/W: 3007
judgepd Mine did 100% overnight but now at 98% with the sun out :/
Hopefully this will improve with a few cycles
TheDragon (GivEnergy) no worries. I will do the math once its dark
Seems to be a very slow charge from 90-100% on 4074. 500W
kai That will be the voltage is near the limit, so back off, the fuller the glass the lower the tap, till full
Previously the tap on full, glass over flows and tripped the upper limit too early
solar charge spiked voltage at 90% to 56.27V. Can understand the backing off, but it seems to be then stuck at slow charge until full. For me that's ~ 2h.
TheDragon (GivEnergy) Doesn't a 100% DOD battery have a top and bottom buffer so that it can do max charge and discharge to it's imposed software limits?
judgepd It does, however that profile is what China have decided is the best.
TheDragon (GivEnergy) Just checking as mine is an older hybrid Gen 1 that it won't go pop like has been mentioned previously with the new firmware?! 😅
On checking the serial number, it is older than the cutoff for routine upload.
Once in production version, you will require an engineer to come and test, if the inverter can be done, or a modification be carried out.
Same for you @Stevef1964
SA2208 just outside the range I'm allowed to upload, without checking with an engineer visit
Sorry for now, you'll have to watch from the side lines.
@makse if he is near you, can check and report back, a Yay or neigh
TheDragon (GivEnergy) No reply yet, with the BMS 4074 from support. Can you chase it up?
TheDragon (GivEnergy) ok thanks for letting me know 👍
if you happy to load manually. drop me an email
paul.landregan@givenergy.co.uk
Later this evening, I will send you the file to load manually.
Needs removing battery cover and removal of the USB, which you can do before hand ready
TheDragon (GivEnergy) OK, sent you an email.
TheDragon (GivEnergy) Thank Paul, file received and loaded manually onto the USB sticks in each battery.
Software now reads as follows:
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
4074
Battery SOC jumped from 92% to 100% when they restarted.
TheDragon (GivEnergy) you need a gen 1 hybrid to test?
Hook Is the serial number SA/SD 2209 or later?
I have pretty much worked my way through the requests and upgraded to latest. I will soon be writing a test schedule for testers to report against.
I will also be doing a Beta Tester email broadcast over the weekend, where I will send to EVERY known tester. Over 130 of you.
Still would like more Gen2 testers, as these are paving the way to fix Gen1 MPPT trackers
TheDragon (GivEnergy) SA2226G445
Having real problems getting the batteries to charge from AC on either inverter since the beta update.
All settings as previous, charge 00:00 to 04:29 for Octopus Go etc.
One inverter started charging battery from 6% and managed to get to 20% before idle, the other sat idle at 62%.
I've also tried force charge from the app, checked the settings via the portal but still no AC charge from the Grid?
Tried changing percentage and other options to kick start, including reset defaults but no difference.
Day 3 after 4074 applied. Charge rate respected.
So it seems you have to save the changes after firmware updates to stop it charging at full rate. The slider looked right, but it ignored it.
After 3011-4075 upgrade I changed the settings after the first charge. This time I changed the settings after the second charge - just to see if it would sort itself out (which it didn't).
TheDragon (GivEnergy) I have a Gen2 5kwh Inv with single 9.5kwh battery and 15*420w panels all facing due south on 2 strings, 5 and 10 on each. Would this be of use for Beta?
If so then what would I need to do, just report any issues as I see them or will you expect me to follow set scripts and provide specific feedback.
Only had system a couple of weeks and still learning, so not sure I will be knowledgable enough to be of use?
THALL Ok on the 4074 firmware. My total battery capacity is 17.72. Between reserve and charge the maths suggest I have 17.01 usable. So missing 0.71kw. Recharging again now.
The battery charge rate setting in the app is giving a misleading reflection of what it is actually set to.
Prior to 4074 update grid/solar charge power was set to roughly 120w max and charge window 00:30 to 04:30.
After firmware update I changed the window to 02:30 to 04:30 and noticed the charge power was showing as 2600W so left that alone.
However this morning I can see although the change to the charge period worked it was still using the previously set 120w charge power setting.
I have since used the "Restore to default" option in the System Modes screen - apparently performed successfully - so will see what happens again tonight.
All of the above done in the app.
I have a Gen1 AC3.0 inverter, if you have spave on the beta I'd be interested in helping. S/N CE2224G001
beautiful 100% flatline for almost 2 hours . 😃
"Free energy" 4074 update.
Not seeing any improvement. The inverter continues to show power when the battery is idle at 100% SOC and the
inverter/battery power dependant "free energy" is still evident.
Woodstock OK a few are seeing this, so will put on the to be discussed list next week
2 things about 4074
1, charged from 18 - 100% today, maths suggests missing capacity is only 0.19kwh. Happy days
2, I am still seeing the 2% jump at the top.
is this the same thing? I cant be sure it was idle .
TheDragon (GivEnergy) 'Still would like more Gen2 testers, as these are paving the way to fix Gen1 MPPT trackers' - this will give a false illusion the problem is fixed and will waste more time going back to the 'drawing board' of how to separate the trackers
joshgreenuk Lend me a GEN2 and I'll test alongside a GEN1 🙂
makse i think we are being ignored on that suggestion
Givenergy upgrade program sounds like a good idea too
TheDragon (GivEnergy) Lend me a GEN2 and I'll become a GEN1 & GEN2 tester
makse seconded
pros: problem gets fixed
cons: none
After just 15 days from initial installation I had my first noticeable SoC drop this morning, a big one at that from 58% to 2%.
With a HY5 Gen2 and East/West strings I am thinking of joining in with the beta firmware testing, but wondered what state it is best to have the system when the firmware is installed and which way does the battery calibration go first. Is it full discharge then charge, so new firmware in the evening could align somewhat with a cheap overnight charge?
Also is there an optimal charge and discharge rate to configure before the calibration rather than just flat out in both directions?
Can @TheDragon (GivEnergy) advise?
TheDragon (GivEnergy) Currently my charge rate has dropped to 250W at 92% which would mean another couple of hours to hit 100%. That surely isn't right.
Simon_C I'm not @TheDragon (GivEnergy) nor on the beta, but my understanding is that there is no full discharge/charge calibration as has been the case in the past, rather a voltage measure and monitor and adjust SoC as the battery energy increases and decreases.
As I understand it, a full discharge/charge calibration is not needed with the 4074 BMS. It was not requested in the email I had from TheDragon (GivEnergy) when he sent me the file to load on our two batteries manually.
Having trawled back through this thread, I found this, which explains somewhat.
TheDragon (GivEnergy) a day ago
chris_p
I have some space for Gen1.
Best time to load the new BMS is when flat. Saves me having to sit and babysit the calibration
chris_p, Stevef1964, and tjsmithuk replied to this.
chris_p likes this.If going from old to new. Calibrate is good if your SoC isn't flat already.
If its at 50% when you switch from 3007 to 4074. SoC drops to 0.
It will self recover over time, but an enforced Calibrate is better and quicker
TheDragon (GivEnergy) , is there an optimal charge and discharge rate to configure before the calibration rather than just flat out in both directions?
Is the calibration cycle likely to be more accurate if the charge/discharge rates are lower than maximum with less volt drop variations on the interconnections? Say 1.0 kW rather than 2.6/3.0/3.6 kW.
Your thoughts would be appreciated on this.
judgepd My charge was flat out until 95% then it tailed off to 1.5kw, down to 650w at 98% for the last 2%
If you want BETA, ideal your current SoC is 4% or less.
As he BMS registers are erased, SoC will be 0% after upgrade of BMS.
So a calibration IS NEEDED. Otherwise it will take 3 to 4 days of cycles to learn.
Which will be quicker if you are already flat. Any Stock to Future upgrade I do, WILL get a Calibration.
If BMS already new 4xxxx series, no Cal will be needed
TheDragon (GivEnergy) Personally I would suggest that a set step by step proceedure is written and adhered to so that a standard upgrade process can be followed by all.
If followed by all then confidence that everything is working as expected will follow and enable any issues to be better spotted.
Currently there is so much "noise" that it must be a nightmare to wade through it all to separate the "chaff from the wheat."
TheDragon (GivEnergy) I’m a new G2 owner - system commissioned a couple of weeks ago. Two strings, 9 panels south facing on string 1, 3 east, 3 west on string 2 with Tigo optimisers. I’ve been seeing a battery drain issue once it reaches the reserve that causes it top top up for, the grid once or twice a night (support have advised this is a firmware issue being investigated) and saw the SOC issue for first time yesterday with a jump from 70-100%. Happy to join your beta test group if it helps
judgepd I left for 9.5kwh mine charging for a few extra hours after 100% this morning.
It took a full extra hour before the 'hunting' stopped. I assume it stopped as it found 100%, even though it dropped immediately to 98%.
This puts the 9.5kwh at 5 hours to fill on an AC3, rather than the typical 3- 3.5 hours prebeta. The last hour for me charged at full rate on Octopus Go.
I assume if I don't allow it the full 5 hours, the SOC will drift as it will never work out where 100% is without solar to top it up.
Jamesrich Is the "hunting" a feature of grid charging I wonder? I only charge from solar and have never seen it. As a full battery is approached my charging rate is throttled back in steps and a soft landing is made at 100% SOC (no rebound or hunting). Typically the throttled back charging period only lasts 10 to 20 minutes. (AC3 and 8.2kWh).
Woodstock I grid charge my 8.2/AC3 every night, and am not seeing any hunting (4074 firmware)
Possible issue on 4074, but I'm not sure if it was there on other beta versions. I lowered my overnight charge from 100% last night and noticed that once the battery hit 80% (what I'd set) it then started discharging to cover load even though it was still well within the charge period. Shouldn't it be drawing from the grid at that point until the charge period ends? I'm pretty sure that's what it always used to do.
judgepd I can not remember if this happened pre beta also. But I see spikes charging and discharging for the remainder of the charge period. Like holding that exact percentage is tricky. When charging my EV too the usage spikes can be huge. But it does hold the percentage set so I thought nothing of it.
THALL This is mine from last night, you can see that as soon as the battery reached 80% it then switched to cover the house load (can be seen from 1:45 to 2:55) until the percentage dropped then it pulled from the grid to top up (all while in the charge period which ended at 4:30.
judgepd Mine from the last time grid charged. Charge perioded ended at 0530. Notice 2 of the 4 spikes the battery is discharging to the grid! Only for a moment though.
Hook @TheDragon (GivEnergy) not sure if you saw my response from a couple of days ago regarding Gen 1 hybrid?
My serial is SA2226G445 is you still require it.
THALL this did happen pre beta, although the oscillations around the target seem significantly larger on beta.
I know Paul has confirmed a fix is being worked, along with the fail to completely pause the charge/discharge, which is linked
mpartington Yep.
It'll be on the test schedule I'm writing. Till such time it works and I stop shouting at the embedded engineers.
Just had my Chinese Chat app set up. So will soon have direct access.
Is that WeChat to the team in China?
If so make sure you remember to set a phone do no disturb schedule. I say this from experience 👍
TheDragon (GivEnergy) 8 hour time difference is going to make you work late or they start early :-) :-)
Anyone else with multiple batteries and not fully cycling still seeing increasingly large soc gaps opening up between the batteries? FW4074 applied 21st Mar. 14% difference at time of post
mpartington
mine are 10% apart at a reported 76% overall the master is 69% and the slave is 79%.
Fully charged each day with solar.
positor1 similar cycle to mine. They hit 100% everyday and idle there, but rarely get into the bottom half. Waiting for a poor solar day so I can see if I can make it crash
mine are 20% apart now on non-beta. Fully charged 2 days ago. I do hope this firmware gets sorted soon
Do we have an ETA on the update?
Russ Untill I get direct access to China, I cant control them, nor infuence.
But when I do, the 1st thing I will get is an update on all the oustanding issues.
Gen2 has some odd new ones
Had some work done today which involved a planned power cut.
'Free energy' is seemingly fixed. It no longer shows battery power lower than consumption.
If your seeing the issue, I would reccomend turning it off and on again.
.
mobile app showing batteries at 100%
no charge to them, more information shows both batteries at 97% ?
53.5v and 53.0 v
Jamesrich Have you had the battery idle at displayed 100% SOC yet?
positor1 That is correct. Inverter SoC reading, and BMS actual Soc reading. Depending on the battery in use, the level if difference (Buffer) is different
Woodstock no.. battery to home is now fairly consistently above 9kwh, but it always drops back to 98% immediately.
Maybe tonight will be different after the power cycle
Sorry to ask, but is the east/west issue 100% fixed?
Our array on any GE system will be pure E/W so even ignoring all the other problems if that’s even slightly problematic still it’s a problem.
Am i right in thinking my issues here are pretty typical
and that the beta seems to have mostly fixed them?
If so can I get in on the beta action? Gen 1
Jamesrich Battery idle at 98% will do if it doesn't hold at 100% (idle is the important bit). Interested to hear what the Inverter Power is reading (Portal/My Inverter/System Data). Thanks.
I'm on the new beta firmware for this now, so will be monitoring and compiling information around it over the next few days
mpartington now at 20% split, software is not liking partial cycles
mpartington will it correct itself when it gets under 50% ? (hope so!).
TX200 interesting to find out. It hasn't done it gracefully on previous iterations. TBH I'm half expecting a drop to zero from 30%. It's difficult to get down that low now we're getting some decent generation, but for test purposes hopefully the clouds will remain today. Might try and empty what's left during the first part of my cheap rate to find out
I'll post the results, good or bad
Today's test was to see what happens when the battery (8.2kWh) is idle at 100% displayed SOC and you isolate the inverter (AC3) from the mains.
Here are the results ...
The inverter stayed powered up and the dongle stayed steady blue light.
Battery Power ............ went from a steady 0W to a variable 3-5W.
Inverter Power ........... went from a steady 21W to a steady 94W.
EPS Output Power ... went from a steady 0W to a steady 94W.
(I do not have an EPS and have previously only ever seen the EPS Output Power occasionally show a random 1W).
So the Battery Power started registering a small amount (which may or may not be the actual discharge keeping the inverter powered up).
The Inverter Power started registering a significant output (which does not appear to be coming from the battery).
The EPS Output Power started registering a matching output (which seems unlikely as there is nowhere for the power to go).
???????????
When I reconnected the inverter to the mains the Battery/Inverter/EPS Output powers all reverted to their previous values.
One other small oddity is that when I reconnected the inverter to the mains it appeared to perform an inverter restart but it did not do the now expected mini charge to get the battery from idle 99% SOC back up to idle 100% SOC. I initiated a manual inverter restart and it did then do the mini charge back up to 100% SOC.
Woodstock interested if anyone has EPS and wants UPS mode enabled.
If it's working in under 20ms. Spec says 9
TheDragon (GivEnergy) I presume you mean auto EPS? Option 4 in the EPS options listed https://kb.givenergy.cloud/assets/EPS%20Configuration%20v2.pdf
I have option 3 which is manual switchover to supply full house load.
TheDragon (GivEnergy) I have option 3 as well, so probably not of much use here. Randomly- were you passing through Andover today?
TheDragon (GivEnergy)
I would but I'm not on the beta and don't want it
I was under the impression the UK grid had a minimum time to confirm isolation?
2x5.2kWh batteries
4074 BMS
As anticipated this resulted in my 3rd crash since increased solar generation means I'm not deep cycling everyday.
1st - 15th March (30% drop) 4075 - 8.3kWh back to full
2nd - 23rd March (23% drop) 4075 - 8.1kWh back to full
3rd - 26th March (20% drop) 4073 - 7.9kWh back to full
This is incredible repeatable, it cycles in the top 50% for a few days, a large split in battery SoC develops (it got to 25% yesterday) and then if I push a deep cycle it crashes. Starting to get a bit frustrating that it is impossible to know what each firmware iteration is meant to functionally address (although I'm sure this will improve going forwards with Paul at the helm), therefore performing tests like this may be a complete waste of time if no relevant changes to code were made.
Each time only 100Wh has been discharged in the 'crash' region (this was 22.50 to 23.20 yesterday), regardless of the gradient of the line. All adjustment to the 2x batteries SoC occurred within this time period. The OCV then had between 23.20 and 00.30, where there was no battery discharge. It did not recover any power and the battery SoC remained the same in the period.
Now onto the charge.... Well, it reached 90% at 3 am. It then took a further 1.5 hours to get to 100% and perform the top end calibration. I think it completed 'hunting 'about a minute before my charge period finished at 04.30. That seem awfully excessive. It also only added 7.9kWh back to full, by far the worst recovery performance I've seen. My individual battery SoC also remain about 10% apart at 100%.
Conclusion... It is undeniable that fantastic progress has been made. It may be correct those saying soc drops are long behind may be true for certain systems and usage profiles, however I can categorically say that is not the case for me. I can see all battery voltages etc. and don't see anything of concern with the hardware or installation. This all seems related to partial cycling, which was one of the main aggravating factors that kicked this development programme off, so I find it hard to believe we've got this far and it doesn't appear to have been adequately accommodated. It doesn't seem to use the top point enough to reset and I suspect China assume everyone uses their batteries under idealised conditions (full to empty with a nice idle period to run the OCV regularly). I'd love to see the internal test schedules, so see if it covers the partial cycle profile. Personally I'm amazed there have not been more reports, I can only assume testers are either still fully cycling, or are now not pushing the bottom limits, or maybe it's a multi battery issue.. over to you Giv!
Yes things are better, I was seeing drops from 50 % pre beta and now it's 20%, but is it fixed?! Not in my opinion.
I had the same issue last night (first time I've seen it). A short platteau in SOC and then a rapid drop (not as rapid as before as the line would be vertical)
What is interesting is that my displayed voltage is well under yours. The most I ever see is about 54.2V at peak but 49V when it is starting to get empty, yours is phased a good 2V above mine for some reason.
My battery doesn't get fully cycled very often, maybe once a week, so I am a prime candidate for SOC drops and have had them fairly regularly since the system was installed. I think it has been better since I started on the beta firmware program a while ago now, but I'm not sure if that is just down to my capacity being reduced so the buffer is bigger at either end which naturally helps with the SOC crashes.
W
mpartington assume testers are either still fully cycling, or are now not pushing the bottom limits
I can confirm I am not even close to emptying mine ever, the lowest I get to each morning is approx 70%
@mpartington @Ra thank you for examples.
I'm going to document them and add to my BMS 4074 report that I am beginning today.
Partial cycling caused SoC drift if cycled if they upper half.
Left unclecked, it results in a crash on the next time the batteries get low.
This is all evidence for the manual OCV at the users choosing. Anywhere in the SoC range
Paul.
Smk82 Those with dedicated circuits on EPS all the time, option 1 and 2, where UPS mode is relevant. Keeping lights, router network kit on.
TheDragon (GivEnergy) also add to that the batteries do not allign at 100% - I'd expect that to pull everything together, not leave a 10% gap.
Challenge I set to the team for the manual OCV is think carefully on how can that be implemented to minimise disruption to the end user? I think the team need to really think hard on this and not necessarily just go with the current solution if it could be improved. As I'm seeing this happen enough to cause a crash in just 4 days, I suspect I'm going to need to run something almost daily. What happened to all the look up tables that were meant to help- can they be expanded to cover a wider range for example...
That's the most frustrating thing for me (non-beta).
I have two 8.2 batteries. The voltages track REALLY closely to each other, but the SOC are all over the place. No full cycling - this time of year, rarely drop below 40%.
mpartington I can only assume testers are either still fully cycling
This is what has been frustrating for me, I don't fully cycle and after a few days from it crashing to 0% if I get below 20% at any point I can expect a subsequent crash. I feel my usage profile is ideal for testing that any SoC problems are addressed. But from what I can tell this new "speeded up response time" code that requires the additional inverter hardware is stopping my system from being a test candidate. Add to that we will be waiting for an engineer visit to test if the inverter will take the new firmware is very disappointing.
Please GE, for those of us with older hybrid Gen 1s, create a version of the firmware that just fixes the SoC problem, I only ever wanted that fixing. I accept the speed of ramp up/down. Please don't make us second class customers because of this.
chris_p seems like you have a similar profile to me, unfortunately the beta software (whilst better at stopping the really big jumps/drop) is still allowing the crash. TBH I'm running out of (easy) test time for this another few weeks and I doubt I'll naturally get below 40%, let alone 20% without really trying. It'll have to be a bit more staged
chris_p seems like you have a similar profile to me, unfortunately the beta software (whilst better at stopping the really big jumps/drop) is still allowing the crash. TBH I'm running out of (easy) test time for this another few weeks and I doubt I'll naturally get below 40%, let alone 20% without really trying. It'll have to be a bit more staged
mpartington It'll have to be a bit more staged
It will I'm sure.
TheDragon (GivEnergy) I have EPS, layout 1 from the sheet. My EPS sockets are not on all the time though - is that just a software setting?
Ra mine were BMS voltages, so I've recreated your chart. However it still shows the voltage difference to your system. Wonder if I've got your upper limit and you've got my bottom limit :-)
Staged it will indeed. Future FW issues with come with a TEST SCHEDULE to complete.
Something I wanted implementing long ago, so everyone has the same tests
kai Yes.
But you could test, plug a lamp in, kill the power and see if the lamp remains on, or goes off for 5s and comes on again. IF EPS 0s switching is enabled, the lamp SHOULD remain lit after a FW update. Previous versions have wiped this setting.
No rush, just when youve 5 mins.
The next run of FW isnt due yet, Ive just submitted Gen2 and BMS 4074. So likely next will Be Gen 1, with the fixes in Gen2 migrated and hopefully some of the other Gen 1 issues too
TheDragon (GivEnergy) Currently my EPS sockets provide no power when the grid is live, they only come on when it goes down. If you can change that remotely then I can test this evening.
kai The EPS circuit is ALWAYS live when the inverter is on. When the inverter is powered the relay connect EPS Out to the AC In.
When power fails, you can hear the click and EPS is powered from the internal inverter.
I too have an AC3 and this is how it works, I have 3 circuits on EPS, all live 24/7
TheDragon (GivEnergy) Will need to check that when home. Last time I tried this though my sockets were not live when mains power was on, and when we had a power cut the other week they were live (ran a lamp from them).
kai Thats mental, whats the point in that. How would your router remain powered.
Ive checked and UPS Mode is now enabled, so see if its made a difference
TheDragon (GivEnergy) That was my thought, seemed a bit of a waste of time really. I queried it with the installer and they assured me that was how it was supposed to function.
Thanks for making that change, will test later.
mpartington It looks that way!!! Mine does run on the lower side for sure.
mpartington You are not alone. My usage profile is similar to yours (typically cycling between 100% and 60% in the summer) and I am seeing similar behaviour. My base load is small so when I do stray below 50% I am seeing a lower half SOC% adjustment rather than a sudden drop. On the next sunny day I plan to discharge the battery at a higher rate and I expect to see a sudden drop towards the bottom of the range.
I have a feeling that the top end correction processing is being hampered by an over reliance on the accuracy of the displayed SOC%. When the displayed SOC% is significantly under reading this is caught by the high battery voltage backstop and (while it may result in a visible jump) it does correct the alignment. When the displayed SOC% is over reading the best that I get is a trickle charge for 5-10 minutes after the battery is showing 100%. I am guessing that the "hunting" behaviour that others report is what is supposed to happen but I have never seen it. It is perplexing that you are seeing the "hunting" and yet your SOC% is still over reading! Kind of looks like the searching for 100% is only taking place on one of your two batteries.
CALL FOR A GEN 1 Tester on D0.197-A0.197
DC Discharge Timing Ignored
Turn Eco Off
Set a Discharge time for 5 mins
Check if your inverter Exports at full whack for those 5 mins or not.
I am still seeing the 98 - 100% jump when fully charged. Is this still on the fix list? Doesn't appear to affect actual capacity so graphical error. But it is a little thing that could be fixed.....
AC3 289
BMS 4074
So has anyone tried a full discharge from say 50-60%? A real world expectation would be the SOC % drops equally until 4% and stops gracefully. If the OCV had not run for a while, will we see some steep drops?
THALL I am still seeing the 98 - 100% jump when fully charged.
I'm not on the beta but I see that. I don't think I've ever seen it report 99% charge, it always charges 98% jump to 100%, or discharges 100% jump to 98%
THALL Yes, its cosmetic and likely a rounding error
.
Way down the list as you can expect, but its stil there
chris_p this is exactly how I see myself testing this going forwards, if I can't rely on natural usage. I would hypothesis 3 to 4 days operating between say 50 and 100%, followed by a full discharge would start to uncover some steep drops on systems not running OCV during that time. The gradient of the line during the drop (adjustment) does try to disguise it's suddenness a little more now.
TheDragon (GivEnergy) Thats all I needed to know. Many thanks
THALL 
97 to 100 today , just for information
positor1 what I see too, little blips, hunts, realises, nope still full and stops again. Eventually remains at 100.
If we had the decimal places we could see if its predictable which one it settles on 98 or 100
positor1 How long between hitting 100% before the first drop to 98? I hit 100% at 1pm and it has stayed there, no hunting for the top. But it looks like your system was at 100% for 4 - 5 hours before it hunted?
TheDragon (GivEnergy) Still no power to the EPS sockets when mains is live.
kai Really daft question, the Mains in is on the right of the 2 sets of terminals, EPS on the left?
Was going to also ask repeat with discharge power set to 2000W instead of 3000, which is another bug.
Need to find another tester now. (🙁
THALL
10.45 full
14.15 first hunt then one at times:-
14.30
14.50
15.15
15.50
then stayed at 100% until 18.00
updated pic
TheDragon (GivEnergy) Looks like the EPS is hooked up to the correct side.
TheDragon (GivEnergy) Thanks!
FW 4074 is displaying a large amount of hunting with an overnight target SoC < 100%. Last night:
Target SoC 75% (10.4 kWh battery, so 7.3kWh expected, 8.2kWh with conversion losses)
Charge energy: 9.75kWh
Discharge energy: 1.53kWh
That's 1kWh per hour to hold the correct charge (BMS 3007 would hold the charge constant with less than a tenth of this consumption).
Hopefully already a known bug with a fix on the way?
positor1 That seems a long time at full before it starts hunting. I thought it was about 1 hour? But just checked mine from yesterday and 6 hours pinned at 100%, no hunting. I would be interested to see if your system did this 2 days on the trot.
THALL
indeed , i will keep an eye on it but today is dull weather so less likely to hit 100.
THALL Are you able to see similar to mine if you set a SoC of less than 100%? Lots of hunting on an AC3.
stevelewis Not for me. Mine is pegged at 100%. I have an AC3 with dual batteries. I will upload a pic once I sit down.
stevelewis that's one of the issues on the list to be fixed. It oscillates around around the SoC target when under 100%, rather than going idle. If you have a large load, you can get some wild swings like on your chart and lots of wasted throughout. I've set some HA automations to reduce charge/discharge rate to zero once reaches target (or near to target) as a stop gap.
This was mine AC3, wasted 4kWh throughput. It sits just below 80% as well
stevelewis See my post #1381 and the following dialogue. I agree the system is using a disproportionate amount of energy to maintain a battery reserve. Don't know if it's on the the "to do" list.
Woodstock Interesting contrast between high load and low load situations.
OK, Apologies for being quiet. I have, as of today now in direct contact with the FW writers. So we are now discussing issues and establishing a working routine.
For now, im going back to ground, nothing much to report yet.
Gen 2 HYBRID
Firmware update.watch out tomorrow Gen 2 testers. Email will go out with a new revision and 2 small tests. Nothing groundbreaking.
MPPT wake up tweak.
Export limitation fix.
Good morning
So last few days my battery has hit 100% and I have seen no hunting at the top end. Yesterday my system emptied for the first time in a while and see double hunting pretty quickly image attached. Shouldn't this hunting happen both ends?
Secondly, since joining Flux my data on the GE system is way out. Not sure if this is a firmware issue or data issue. But the GE data says I have pulled in 19.47kw from the grid. My smart meter says 14.7kw. This is way out. I will email GE support about this but wanted to mention it here as I am on the beta. Images available if needed.
Last full 24 hours here. Overnight charge hit 100% and stayed there which is good. Didn't get back to 100% on solar as it was a miserable day. System hit 5% at 21:20 and started drawing from the grid, then recovered back to 8% and started covering the house load for another 1 hour 40 minutes all while still on 8%. I did a discharge to the grid just before the end of the day to make sure it was completely empty before an overnight charge to test the capacity. It added 7.2kWh back to full which is still low at 88% of full capacity.
THALL , are you checking your smart meter against the pretty GE graphs and animations, or the GE metering data available as the last tab (I think) in the 'My Inverter' section?
My system has very different values for grid import and export, depending on where you look in the GE web portal.
Simon_C Energy Graph card, charts, mobile app.
The data is even more out now. Smart meter says 15kw GE Energy Graph card says grid in 25.89. I do not recall a time previously where the data has been so far out. Though it does not appear to be just GE, myenergi app says I have imported 27.5kw.
Don't get it. Only seen this since I joined flux, can my smart meter be wrong? Guess I will see once the bill arrives! 27kw grid on on flux is not cheap.
THALL For Octopus you can download raw data. Go to website Account, Explore My Energy Usage, Download your smart data. Should give you answers (or more questions :-) ).
BrianHere I am seeing this difference on todays usage, not sure I can get the data from octopus until tomorrow?
Im trying to find your inverter. The serial I have SA2209G014 doesnt seem to exist
TheDragon (GivEnergy) I have an AC3 with EPS. We haven't had many grid outages, but when we have my son moans about the outage (a few seconds) taking down his PC, so I'm sure he'd be happy if the PC could survive. Are you still looking for someone to test UPS?
Also BTW my battery is still taking 7.2-7.4kWh per night to charge to full from 4% (so is not regaining the lost capacity)
TheDragon (GivEnergy) Not sure if you saw my post 6 days ago, didn't see any response.
"I have a Gen2 5kwh Inv with single 9.5kwh battery and 15*420w panels all facing due south on 2 strings, 5 and 10 on each. Would this be of use for Beta?
If so then what would I need to do, just report any issues as I see them or will you expect me to follow set scripts and provide specific feedback.
Only had system a couple of weeks and still learning, so not sure I will be knowledgeable enough to be of use?"
Issues spotted so far....
1) battery charges to 100% but fails to stay there even if surplus of solar, battery discharges to 95% before starting to charge again, pattern repeats until solar subsides.
2) Export limit set by installer at 3.68kw as specified by DNO but portal shows exports up to 4.8kw. Not sure if reporting is wrong or actually exporting too much.
other than that, not aware of other issues and no experience of SOC jumps.
TheDragon (GivEnergy) Sorry Paul, it should be SA2209G104
THALL Yep.. Patience. Go have beers time :-) :-)
THALL You need to ask for an Octopus mini, then you can see live usage and totals on the same day.
judgepd On the waiting list :-(
andrew_bond and today's battery discharge figure is only 5.7kWh for the opposite direction (100% to 4%) - this is very disappointing! (73% of spec)
Ra Cool, thats make more sense, your already fully up top date.
Ive deleted the duplicate line on my spreadsheet
Sounds ideal, especially as the export limit fix came in today for Gen2
Whats the serial number, will get it loaded 2night
Lost capacity has been moved up the list, Ive even given China the limits "I" want to be used. So im sure there will be some bartering.
I calculate 5% lost top end and 10% bottom for the worst affected (8.2s)
TheDragon (GivEnergy) ED2301G097
Anything else thats been fixed or changed that would be on interest?
Solar efficeincy in low light (voltage). Export limit fixed, faster reaction to changes.
Your current SoC is 63%. If i upgrade the BMS now, it will dump the lot to grid and then recharge.
Please confirm you are ok with this?
TheDragon (GivEnergy) i'll start cooking....go for it
TheDragon (GivEnergy) Stabilised 100% SOC voltage = 52.8V? (That's what mine was under 3007 and it dropped to 52.6V under beta).
THALL , my grid import compared with Octopus SMETS2 meter is quite close, for the chunk of March that I have had the HY5-Gen2, though I do have an Owen Brothers OM115 with a solid core CT fitted for the grid metering rather than the standard GE EM115 with the split core CT. Also specifically for grid import, the daily totals from the energy graph match with the Inverter-Meter Data-Import Today values.
Where there is a difference is when comparing generation from the energy graph to the Inverter-Meter Data-Generation Today.
When totalling up the generation for the month so far, there is around 7-8% error between the reported values (both by GE).
Note - depending on the time of day capturing the energy graph totals, there can be a slight change by a few 0.01 kWh.
There must be 2 different methods used by GE to get to these totals for each day, but which one is to be believed?
THALL I worked out why my smart meter was so far out compared to the GE and myenergi dashboards. Turns out my son took an interest in the smart meter and pushed buttons. Changed the screen to co2. Problem solved, nothing to see here :-)
I had forgotten that I had set the system to charge to 50% so was surprised to see this image this morning.
However 50% is being ignore, charge starts at 57% and stops at 59%,
I can not make sense of the other spikes. number 2,3,4,5, and 6 say battery power -2862w, grid import/export/pv and load all zero.
anyone?
What is it about 01:00?
Another charge spike at 01:00!
Not playing around with Battery Cutoff settings this time (left at default 4%) but the displayed SOC% was down at 4%.
Battery had been idle for around 6 hours beforehand after a bottom end adjustment hump yesterday tea time. (Note a spike in demand stopped the adjustment in its tracks).
The charge spike apparently added 8% to the SOC (although I would have thought 3% to be nearer the mark).
What is it that kicks off these charge spikes and why always at 01:00?
positor1 What does the battery voltage curve look like?
TheDragon (GivEnergy) Hi, as far as i can tell, the firmware update hasn't yet taken place. Was there an issue or just a scheduling/timing thing. Thanks.
positor1 There is a known bug in the Charge to code in all versions of firmware. It is being fixed, I am hoping in the next release.
Charge to x% ignored, or oscillating. 2 seperate issues
Along with the Discharge Home Demand doesnt work as expected.
Woodstock there was a post from someone else on here whose charge kicked in no matter what for 5 mins at 00:00 or something. I think the idea was it’s something to do with the 2nd charge registers … I’ll find the thread.
hoggy Yes; you get the charge spikes at 01:00 if the Battery Cutoff is set to something other than the default 4%, but in this case it was set at the default 4%.
Battery Cutoff is no longer used in the beta software so we were hoping that, by simply removing the ability to change it, the issue would go away. Seems from the spike this morning that this is not necessarily the case and hence the need to know what processing kicks off the spike (and why it occurs at 01:00).
You sure, you're SoC is 97%
TheDragon (GivEnergy)
thank you. i will wait patiently 😃
Ok Its loaded. I see before you had 3.6kw restriction on your 5kw. I have kept this in place.
The battery is charging, it will hit the upper ceiling, I want to see if teh SoC will re-set without a calibration
TheDragon (GivEnergy) ok, I saw that the battery firmware had been applied, assume inverter firmware doesn't get updated. i will leave in your capable hands. What voltage is upper ceiling?
Jon-boy if gets to 57v, it will trigger SoC jump.
Normally I would calibrate, but as your was so high, I thought I will try to provoke a jump up and reset, instead of calibrate.
Saves you some charging power
Ok you are charged, it jumped UP as expected.
As the bottom isnt yet set, you will get a jump down at some point too.
Give it a few days to settle
TheDragon (GivEnergy) Appreciate that, thanks.
Question related to capacity and this new beta.
My 9.5 kWh battery charged from 5% threshold to 100% overnight, but only added 7.83kWh according to the GE data (from grid). So seems I am missing some capacity if its 100% DoD?
pacemaker Try discharging from 100% to 5% and see how much you get out! 😢
Woodstock Don't forget to wait a good hour once it hits 5% to see if it finds more.
Also finally seen some hunting at the top yesterday. First time since the latest firmware applied. 2 dips to 98% then back to 100% even though still had around 3kw of pv. Doesn't look like the sun will be out for long enough today to hit 100% again though.
I have raised the priority of capacity with China, 2 more engineers (Who I am now in direct contact with)are looking at ways to recover it, aswell as reducing the last 10% without overshooting the upper voltage limit.
Lots of evidence given to them, likewise their counter arguments, while valid, dont meet with my requirements.
Great to see you flagging this (example of the difference you've made to Givenergy already) - this is. Basically this is a potential sales feature that GE have forgotten to pitch in their marketing (where their competitors often are!)
This would be great to explore as thought this was in this originally from the documentation, but no one at GE seemed to know!
Have we got any documentation on this, then I could rewire a bit to test!
THALL Yes, there can be more hiding. The thing I cannot work out is why 1kWh always shows as a much bigger percentage of the battery capacity on the way down than it does on the way up. On my 8.2kWh battery a constant charge suggests around 7.5kWh capacity while a constant discharge can see the slope of the SOC% curve equate to 5.5kWh capacity!
Woodstock That is frustrating. Even allowing for 4% reserve and a bit of conversion loss you would expect charge/discharge to be around the same.
Woodstock yes I see similar. Seems quite a difference between energy in and out.
Her is the past 60 days daily totals. kWh recorded in (DC) kWh recorded out (DC)
AC coupled, 3 x 5.2. So expect some loss, double conversion
TheDragon (GivEnergy) So are you suggesting that for the batteries only, the total efficiency in and out, is 96.3% ? That feels too good, but perhaps my understanding of the conversion losses have been swayed by the total journey, when all of the inverter (in)efficiencies are included.
If this represents the energy lost as heat, in the battery (on the way in and then out again), then I am impressed.
TheDragon (GivEnergy) It is not the actual energy in versus energy out that concerns me but rather the fact that the SOC% curve behaves differently when charging versus discharging.
When charging, the curve displays as if the battery capacity is around the nominal value. When discharging, the curve displays as if the battery capacity is much less. This is particularly noticeable as my battery usage is tidal with a period of continuous charging followed by a period of continuous discharging. A kWh going in increases displayed SOC by around 13% while a kWh coming out reduces displayed SOC by up to 18%.
Woodstock Two possibilities:
1) The indicated SOC (however it is being calculated) may not have a linear relationship with the actual SOC. So this could result in the differences you are seeing.
2) If the battery voltage is being used, as part of the indicated SOC calculation, then it could be as simple as the charging voltage being higher than the idle voltage, which in turn is higher than the discharging voltage.
(I have always wondered if GE gas ever tried to adjust/compensate the battery voltage, as a function of the current flowing in and out. In theory, if the battery has a broadly stable internal resistance, then it should be possible to do this. But of course, other factors would also be at play at the same time e.g. temperature, settling time since last charge etc.)
@TheDragon (GivEnergy) How far away is the firmware from being ready? (approximately)
thanks
Ooh thats a tough question.
Depends at what point we call it.
Do we want 100% solution. 90? Or too hell with it, it's fine now.
I want 100, but give me at least one more iteration of each inverter and BMS before I make the call.
Could release now the BMS. But 10% would not be happy. The capacity and some odd SOC discrepancies needs resolving for 100% .
Inverter, Gen2 could go on the next run. Providing Home demand and oscillating fixed.
Gen1 and AC3 at least 1 more. Trial
TheDragon (GivEnergy)
i would prefer to wait and have 100%
Talking about SoC oddities....
Whilst we wait for the user defined OCV, is there a work around to force some alignment, without going to empty?
I set the reserve to 40% last night (between 00.30 and 04.30) to pause the battery charge/discharge at my target soc and stop oscillations (another work around, pending fix). I would have set to 30%, but chickened out given my system likes to crash at 20% and didn't want to be taking from grid during breakfast.
I seems to recall battery SoC needs to be below 40% (or was it 50%) and below a certain current for OCV to run. Because the split is so large, it's hard to get the high battery below 40%.
I was hoping to set a low soc target &reserve on nights with a good next day solar forecast and provoke OCV into running.
When the sun comes out, to cover my load, i do intend on a fast discharge to test if it crashes at 20%.
TheDragon (GivEnergy) I would be ok with 90%. On most systems it fixes the issue it was meant to. Plus hopefully there will be continued firmwares anyway as people find other issues, be it data, or some bright enginer spots a way to get 5% more efficiency. capacity, throughput etc etc. Plus the scale of this issue, it would be good press for the fix to role out. Fix for the SOC issue at least.
IMO even if 75% is better than what folks have today you should ship it. Minimum viable product. If it's a backwards step then you hold it until it's not.
Perfection is the enemy of perfectly adequate
someone wise, once
Agree. Aim for 100% but accept that you will need to compromise some time soon. Give customers the option to install it, but perhaps more importantly, also give them the option to roll back if it causes them additional issues.
Hey Paul, I know this was quite a generalised reply and perhaps I'm in the 'wouldn't be happy' group, but could you clarify your opinion of where we are wrt to downward SoC drops for systems that do not regularly do a full cycle. What I thought I had picked up in the last couple of weeks was that this is still very much an issue and could not be considered 'resolved'. Maybe I have missed some nuance here given the increased flow of feedback on here recently, so please correct me if I am mistaken on this. If not, what's the action plan here. Cheers 😉
TheDragon (GivEnergy) Aiming at 100% is a must but at it stands right now I'm getting 3-4 times a week large SoC drops 30-40% (biggest recorded was 55%) so temporary 90% fix would be more than welcome.
Unless you can query your DB to specifically pick out those with X SOC jumps a month I’d do this in batches.
There’s thousands of systems and your support teams going to get swamped if you just let it loose in one lump no matter how good it is.
I’m sure I’m not the only one who’s had a firmware update hang, having to wipe the dongle/usb and so on…
This at-least placates the worst whilst also making sure nothing weird happens before you try a rollout to what, 12k systems by now?
Try not to be Tesla, “fix one break one” gets pretty tiring…
hoggy I can confirm that you are not the only one to have had a firmware update hang!
Following months of ignoring any commitment to fixing the SOC/Battery capacity issue by GE I think there are still a large number of people who remain sceptical that this will be fully resolved - Paul you have put massive amounts of work into this and it’s both highly appreciated and somewhat reassuring, but we need convincing via the update.
On a positive note I’m happy with just about every other aspect of my 5kW Gen 2/9.5kWh system - but the battery capacity issue does worry me.
So getting to the point, while I don’t think it’s ever a good idea to release an update with known bugs - releasing an update which reinstates our capacity (permanently without needing GE intervention via calibrations) would reinstate confidence - even if a couple of new gremlins are introduced.
ToothyChris 1) is possible but I would have expected this to affect the charge and discharge curves similarly. So that, while they may not be accurate at a given point, they would be consistent with each other. This morning my SOC went from 82% to 69% and then back up from 69% to 82%. The battery showed as having discharged 0.6kWh on the way down and accepted 1.0kWh on the way back up, over the same range.
2) is also a possibility but I would have thought that this would to result in the battery full and empty points always being at variance with the reported SOC%. This is not the case.
Based purely on observation, both the SOC% charge and discharge curves are plausible and the reported battery charge amounts tally. The conclusion I am coming to is that the reported battery discharge figure is consistently under reading.
Had my own software company - large business stuff. Occasionally, because of the precise issues, we did a release with a "known outstanding issues" list. Clients who were pressing for specific items could then decided to use or not, but they were given the option. That was always appreciated.
Other point is that if the existing, in use, software historically includes the outstandings then it makes no difference does it...
TheDragon (GivEnergy)
Hi,I thought I would give some feedback since being updated to the Beta FW for my Gen 2 5kw inverter.
Here is a picture from App showing Max Power.
Additional photo showing max export power
TheDragon (GivEnergy) I've just seen this bug on my system last night. I have automation that reduces the SOC target if Solcast predicts enough gen (first time the forecast has been sunny enough for this since the beta was installed). This will happen more frequently from now on, so a fix for this would be welcome.
andrew_bond FYI since I'm not sure if the forum tells you which post I replied to! This is the charge/discharge issue when target lower than 100% is reached (you posted Jan 14th)
My jump at the top has grown, was 2% now 3% jumps :-(
Had a strange bug last night, set the AC charge to 30% on both inverters. In the evening one battery was at 82% the other at 18%.
As I'm suffering from cross charging I expected the higher charged battery to fill the lower to the 30% - it did and went beyond it, bypassing the 30% limit completely.
New GEN 2 FW. This is now a test production version D0.908-A0.908.
Loading onto the few test Inverter I have.
It is also fixing the Throttled PV seen in FW 906 when the battery is full.
As usual an email will go out
@Jon-boy keep an eye out for your inverter version changing and check again your issues please.
Ref your DNO limit, its set to 3600W right now max AC to grid
TheDragon (GivEnergy) . My HY5-gen2 can go on the beta test if you want, having just had my 2nd SoC drop from 48% to 0% after just 4 weeks from installation, and morning startup issues with East/West strings.
Currently 95% battery charge, but the car is plugged in for an overnight charge so will take anything dumped by the GE battery.
Please let me know if you want to apply it.
Simon_C Will need your Inverter serial.
I can load this evening, been a long day in the office, need some food
TheDragon (GivEnergy) , sorry I meant to paste the image in.
@TheDragon (GivEnergy)
Would it be possible to add my AC3 and 2x 9.5 kWh to the beta?
CE2233G285
Simon_C OK. Upgrade in progress
to D0.908/A0.908
BMS 4074
Calibrate to 100% will follow, skipping the lower SoC find, so a jump or 2 will follow
Edit. Done
TheDragon (GivEnergy) , many thanks.
I have seen the rapid charge to 100%. Was getting concerned seeing the inverter voltage to the battery hitting 58 V, and 55.9 V on the battery, but a couple of minutes later it was finished.
Simon_C That was a calibration, it wont normally go that high.
I too was having kittens seeing it pass 57v. Highest Ive seen
BMS Capacity Update - 5.2kWh sample battery tested in the China LAB.
Direct compare of the same battery with 3008 & 4074 BMS
Im waiting for approval to post the whole report. I have also requested the same lab test for a 8.2kWh
Doesnt detract from the fact I want the last 10% reduced speed to be a little faster, say last 5% reduced.
BETA TESTING COMMUNITY CALL FOR HELP
Does anyone wish to put together, from a user perspective, a list of functional checks that can be used for confirming functions and features. Then we as a community can ensure no stone is left unturned.
Some plugs up for grabs
TheDragon (GivEnergy)
Am I reading that right ?
So (almost regardless of BMS version) their 5.2kW battery only managed a DOD of approx 69% ? From a battery that was advertized as being 90% DOD on all the datasheets released throughout 2022 - even though none appear to have been actually labelled up as such.
ianfretwell It does seem to be that way, if so quite disappointing
@TheDragon (GivEnergy) Going to need the rest of that report asap. Or at least the finer details like was that a 0 - 100% charge, how old is the battery etc etc etc. Because if that is 1 a fairly new battery, 2 a FULL charge cycle then people with 5.2 batteries are going to be mighty unhappy about being mis-sold their equipment. Not to mention advertising as 80% DOD. Need more detail asap please mr dragon.
I would be keen to see lab data on the 8.2 and 9.5's
OK, just for info - I have 4 different 5.2 datasheets, and a couple of things have changed on them from the earliest:-
The ones dated 01/22 & 05/22 have 90% DOD, 60A Current & 54Kg weight.
The ones dated 12/22 & 03/23 have 80% DOD, 50A Current & 63Kg weight.
So it's gotten 10Kg heavier and yet (apparently) 10% less capable over the last year...
Simon_C Was getting concerned seeing the inverter voltage to the battery hitting 58 V, and 55.9 V on the battery, but a couple of minutes later it was finished.
Hi Simon, where do you see the Inverter DC voltage, in the GE system?
ianfretwell Ok so @TheDragon (GivEnergy) also need to know if it was one of the 90% DOD or 80% too please.
I've been down the 5.2 DOD many times and there appears no straight answer as to whether there ever was a 90% 5.2 beyond datasheets.
Some have 32cells inside (2P16S on the nameplate 32x 51AH cells giving 102AH) some of the very latest ones appear smaller which would suggest they may have changed something (perhaps back to 16 cells) I've done a few posts about them.
TheDragon (GivEnergy) probably won't have a complete list, but happy to chip in some ideas/proposals
hoggy I had seen a post (from a very well in installer) that the 90% was enabled via firmware. However that test (unless there is also a different hardware configuration) would seem to suggest not 😞
mpartington I don't think this is completely far-fetched as when my installers followed up the 80/90% DOD on my behalf they were finally told (by George Rawdling I was informed) :-
"they (GE) are developing a firmware to work in conjunction with this 90% DOD but at the moment it is not compatible with Beta firmware you will need to choose which one you would like to prioritise."
Which I always though was complete nonsense - why would anybody choose to have a less capable system...
ToothyChris , it may be just my interpretation, but here goes.
When using the App, hit the battery icon (bottom right). The title is 'Battery graph', and under it the Inverter serial number.
Under the 'Latest Data' is a voltage, which I presume is at the inverter.
If you then hit the 'More Information' text link, I get 1 battery listed with serial number and a voltage.
I presume this is the battery voltage.
If you then hit the battery animation, you get trend charts of battery voltage, %SoC and temperature.
The 2 voltages reported are different, and the highest value logically swaps over when a large charge or discharge cycle is active.
Thanks Simon_C
I have an older Android phone with a smaller screen, which means that both of those voltages are invisible to me.
As far as I know, the inverter DC voltage is not displayed (or downloadable) from the Portal anywhere, but your assumption may be correct.
I was interested, because with the voltages at both ends of the DC battery cables, combined with the DC current, we (and indeed GE) would be able to monitor the resistance - which would be useful to identify potentially bad terminations of the DC cables or a failing DC isolator. I wonder if GE does this as part of the commissioning checks?
ianfretwell I can see why they never did that. Given that the entire range of batteries is (if we are splitting hairs) 80% DOD then there would need to be some clever logic to switch certain packs off to get that last 10% out of a 90 DOD.
Alright if you have all 90 or all 80 but not so much mix & matchable...
THALL There all 80%
Even the 8.2 and 9.5 are 80%. But to market as 8.2/9.5 100% they are 20% larger inside
Hence I also asked for the lab test on an 8.2 for comparison.
TheDragon (GivEnergy) Ok I do not understand, you say all the 8.2's and 9.5's are 80%? but go to market at 100%? How? Why? And simply put are we getting 80% on the 5.2's and 100% on the 8.2/9.5's? Ignoring the 4% reserve.
It's just marketing "trick" really.
The 8.2 and 9.5 are bigger than they say they are (by 20%) so they can be called "100% DOD" when in reality they are 80% DOD of a bigger battery.
the 5.2 and below was never marketed like that so stayed at 80% DOD (they could just have said it was 100% DOD 4.2kWh battery.
I did a thing about them on the Wiki. NOTE: there are still arguments about if the 4% reserve comes out of the user part or the buffer, never got around to conclusively answering that so that part is more "maybe"
THALL 9.5's are actually 11 or 12's internally. 80% of 12 is 9.6. Hence if the battery will perform to 80% DOD in real terms, at can be marketed at 9.5 and 100% DOD.
cluelesscris Correct. Which is why the 8.2 is for most the better bang for buck. Which also meant they sold out 1st in the shortage of 2022
Firstly, thanks to everyone who has worked on the development of the fix for this problem. I couldn't devote the time that was needed to be part of the beta and I can now see just how much work was involved and appreciate everyone's efforts.
It looks like there is now something worth moving to. My 8.2kWh AC3 system drops to 0% around 4 times a month (6 times last month) and I'd like to see if we have now moved on. My inverter is CE2119G082 if its any use for testing and I appreciate that this is still a beta so there could still be problems or new problems to work through.
OK so usable on the 5.2 is 4.1. Lab testing got just under 3.6. So 10% loss from dc ac conversion?
Personally I like to have delivered what I bought. How the supplier acheive that is not my problem.. if they had to double the battery pack to give me what I ordered it makes no diff to me...
Now take the other case of EV range.... They deliver it as 300miiles, but it won't give that. It'll be minus 20%...
They should sell it as 240 miles range.
BrianHere Well... Many EVs can deliver the WLTP range under the right (ideal) circumstances, it's actually possibe to do better. Just like petrol or diesel vehicles. Not quite comparable to buying a storage battery with a clearly stated capacity -- subject to a suitable operating environment. Resistive heating losses will have more effect at higher power, charging or discharging, but the batteries should meet spec. in 'normal' use.
Hi was wonder when the new software is being released as I have multiple issues with throttling being one of them? (Sunny all day and when full the PV input just stopped and did not export excess to grid). It is also not using battery for home demand during and pulling from grid at times as below: 
@TheDragon (GivEnergy)
First time of setting to 60% charge overnight since being on the beta software for both inverter & batteries and this happened:
When it gets to 60% it is constantly oscillating between charging & discharging at the peak rate of charge & discharge during the timed charged period (23:30-05:30)
Will the latest beta software cure this?
Also noticed that the % charge of each battery can be up to 15% different at times.
Set it to 100% last night as dull day expected today and it was fine. Stayed at 100% until end of charge period at 05:30.
Woodlands This is fixed on Gen2 FW.
Tested good overnight last night by some of my testers
Gordon I also have the same issue on D0.906-A0.906. Once the Battery does the SOC jump to 100% bug then solar production tanks and I see a similar graph to yours. It also doesn't seem to want to use the battery to power the house when the load increases. 
SolSucker Ask for the Pre Release Gen2 FW
The BMS is still Beta, this will resolve this solar issue when the battery is full. the Battery jumping is resolved with the Battery FW, but still beta and not final.
@TheDragon (GivEnergy) what is the state of the FW for Gen1 inverters and 9.6kWh batteries, in addition to 3-4 a week SoC crashes of 10-40%+ I am now getting midscale SoC jumps and drops - I'm getting to a stage where I don't know where and when the SoC will happen and how big it's going to be.
thanks
alfwro Gen 1 will follow on from Gen2, as they are similar FW
TheDragon (GivEnergy) Is the intention for the older Gen 1 hybrids to have a firmware which just resolves the SoC issues and not require an engineer visit? As you can imagine the feeling is that this has gone on long enough and I don't want to be waiting for an engineer to visit to see if my inverter has the gubbings for the faster ramp up/down code.
TheDragon (GivEnergy) dare I ask when we might see the firmware?
How do you get the pre release firmware loaded ?
Tried something different last night. Previously I've seen my battery crash to zero (on 4074) when it hits around 20% after a few days of partial cycles. However this has usually happened at low discharge rates (approx 200W across 2 batteries).
Last night I discharged from 30% at full power and the SoC reduced linearly with discharge all the way to 4% !! It also seemed consistent with the expected capacity amount (based on @TheDragon (GivEnergy) test shared above). The resting voltage at empty was also much lower (51.2 vs 51.7V).
This makes me wonder if there is an issue with the logic detecting empty too early at low discharge rates (the opposite to the original issue).
Still seeing a battery split, but this was corrected. I was able to use the first 2 hours of my off peak rate for this experiment, as I only needed 50% charge for today. However I need the option to manually or automatically run OCV at higher soc for more flexibility.
FYI
Again, not part of the testing but very frustrated with a big drop last night. Do we know why it happens yet and can we do anything to mitigate this while waiting on an update.
Generation 2
Inverter Firmware Version
D0.904-A0.904
Battery Firmware Version
3007
Rodmac76 Gen 2 is almost ready, the BMS im in discussion with one last update.
Lab says capacity is comparable. I say not, so we agree to disagree, and they are going to better optimise.
Yesterday was a chinese bank holiday, so nothing was done, today, in work and quiet, which usually means busy
Matt, with split readings (Which self correct if given the chance) are you seeing capacity as i mentioned.
If I average out the days over 20 days, im not far off maybe 3.9-4.1 overall. per pack. (11.7-12.3) for my 3 packs outside at 15C. I have no grounds to argue with 5.2s.
Hence why i asked for the same lab test on an 8.2, hoping they realise charging 7.5 in is not acceptable
In an ideal world same FW for all batteries, but even if the final answer is 3052, 3082, 3095 I can live with that#
chris_p I believe the FW is going to detect if the part is missing and adjust.
That will be deffo discussed when we begin the next round, Gen 1 tests
TheDragon (GivEnergy) Oooh so no engineer visit needed now or am I confusing the two things?
chris_p they didn’t say HOW the firmware detects it. If it goes bang immediately then firmware reports back that it’s an incompatible inverter.
(joke btw before anyone piles on)
I got updated to D0.906-A0.906 / 4074 on Monday night with the battery at around 95% and only a calibration to 100%, not to 0%.
Since then the AC charge to % has had almost no hunting, but I have much reduced capacity from the battery (around 1.5 out of 9.5 kWh). Not good as I am getting periods of import at the peak tariff rate when the battery goes in to idle mode with 6% or less SoC even when the battery voltage is around 53 V.
Does this actually need a full calibration or calibration to the bottom end of allowed battery voltage to restore some more usable battery capacity?
hoggy
Found this in the inverter update code:
if (inverterTemp > 1000) {
portalNotification("Firmware probably not compatible");
}TheDragon (GivEnergy) When I Was on 3007 BMS, my 8.2 battery would take >9.5kWh from a zero SOC drop to fully charged. Now it is more like <8Kwh.
TheDragon (GivEnergy) so yesterday was a really good test for this.
The battery started at 100% and then discharged til empty, with low solar. It discharged 7.6kWh, but charged 0.2kWh (from solar)
Therefore I saw 7.4kWh net discharge against 7.18kWh from your tests, so marginally better. Unfortunately I only charged back to 50%, so can't give you a charge back to full value (that took 4.7kWh).
As the final voltage was much lower, I suspect I got more out than if I'd gently approached idle at low discharge rate.
Based on 7.4/96 x 26% = 2kWh (assuming last 4% is not accessible), then the amount I discharged from 30% (1.9kWh) seems spot on
Simon_C If there is idle period no.
Best solution is full calibration, but wastes your power if SoC is high
If you want I can kick of a calibration FULL this am, while weather is pants
Off to sign out of my old job, hand in IT and say goodbye today. But will get on it later, by then SoCshould be low and today if it remains idle, OCV should do what its designed to do, Adjust the SOC based on Open Circuit Voltage.
Then it should adjust upwards, releasing more power as it does.
TheDragon (GivEnergy) , yes please do the calibration.
My BEV is plugged in with capacity, so will take anything dumped from the battery.
There is some solar today that will help out with the charge, and I would rather get the battery capacity back sooner rather than later.
Inverter software version reported as different now.
I keep reading here that Gen2 FW is nearly ready. Is Gen1 being given lower priority? If that's not the case what is the status of Gen1 FW?
thanks
alfwro if I remember correctly, gen 1 AC should follow on fairly quickly after, very similar code.
Gen 1 hybrid more complicated, depends on the hardware installed.
Simon_C done
TheDragon (GivEnergy) , many thanks for the quick turnaround. It is reporting a battery discharge at 4.02 kW.
For my info, what battery voltage should I expect the discharge to finish at?
TheDragon (GivEnergy) although my 5.2 discharge matches the lab test, I'm still not sure why the big difference between what's going in and out, as its measuring both in DC, so only the losses are in the battery. Is 10% really reasonable?
The 500 W difference will include losses but is probably due to DC power feed to inverter and BMS. Would be interesting to know time period of test. 10 hours at 40 W could be reasonable.
Newman good point, presented like that it does seem reasonable - just all adds up!
mpartington More small batteries = less efficient than less bigger i would imagine too.
I will wait for the 8.2 lab tests
SolSucker have you managed to get anywhere with this ? I am still having same issues. Giv switched my inverter & battery firmware back to an older ref but it hasn’t changed anything.
Could it be semantics? A 10kWh pack sold as an 8kWh rated pack will have a 100% DoD vs stated value but is truly an 80% DoD on the true capacity.
kram I don't think it's semantics - the important point is the amount of charge/discharge seems to be less on the new firmware, and most importantly less than the marketing number. Before I put the beta on, I was charging about 8.5kWh into the battery, now with the beta it's more like 7.5kWh (DC charge figures; this is an 8.2kWh 100% battery)
HY5-Gen2 - D0.996-A0-906 - Battery 4074
What with the excellent solar yesterday, I had plenty of charge left in my GE battery before the start of my night AC charge period, so I set the target to 40%, expecting the night time house load to draw down from the battery as was the case on the original firmware.
This did not happen. Instead the GE battery did not discharge at all, and all the house load during the period was taken from the grid (albeit at the cheap rate). Also once the solar gen started at around 0630, it met the house load and once there was excess it was exported to grid. It was like the battery was in a paused state and would not charge or discharge.
Has anyone else seen this, or should I reset to default then re-apply my required settings?
Battery at 0130 = 80%
AC charge settings - 0130 to 0730 - Target 40%
At 0700 I changed the charging period to end at 0704 in order to draw down from the battery in readiness for another good solar gen day.
I am on the same FW and agree, if you schedule a charge to a set % then the battery is paused until then end of timed period.
During any defined AC charge period the battery will not supply the house regardless of SOC, it will all come from the grid.
Phil_H , that was not the case on the original firmware when the battery already has more charge than the target % value during the AC charge period.
There are many users confirming this over the last few months on this forum.
Phil_H That isn't true. On my current setup on the original firmware, if I set a charge rate of 35% overnight and my battery is over that level at the start of the timed charge, it will discharge to 35% then stop.
Phil_H So is the consensus that it is preferable to have it behave in the way of the current production firmware during a timed charge: if SoC>Charge to %, the battery supplies house load until they are equal when house load is supplied from grid until the end of the timed charge period?
I don't have a preference for what it does as I have not used the charge to yet but setting a charge window effectively disables eco mode which I understood meant the battery didn't discharge. If people have it working differently then I'm happy to be corrected.
Tim , I would say so.
If you have left over battery charge from the previous day (notional for free as from solar), then it is preferable to use that to supply the house even during the cheap period.
Also if there is a good solar day forecast for tomorrow, it is preferable to have capacity in the battery to receive it rather than the battery being full by 1030 in the morning and all the remainder for the rest of the day being exported.
All depends on each users tariffs, but it made sense to me on the production firmware.
Simon_C agreed
Tim if SoC>Charge to %, the battery supplies house load until they are equal
@TheDragon (GivEnergy) can this be flagged up to the developers? The logic on the existing firmware makes sense to me, rather than suspending battery discharge during a timed charge.
Tim OK, I will feed this back.
So with winter mode ON (AC Charge to%) If SOC > Set %% then discharge to this level and hold.
I already know about it also stops solar charge
I have dug out the 2 registers involved in tehis logic and added those to the feedback.
To clarify there now only 2 bugs left on Gen 2. Both relating to Set% and solar/discharge
TheDragon (GivEnergy) So with winter mode ON
I didn't think this was a "thing" any more. AFAIK the modes in the web portal are: ECO, Timed Charge, Timed Discharge, and Timed Export. If ECO and a timed charge schedule is enabled, the Charge to % can be anywhere between minimum 4% and 100%. With the current production hybrid firmware, when the inverter enters a timed charge period, if the SoC > Charge to %, the house is supplied from battery until SoC = Charge to %. If SoC < Charge to %, house load is supplied from grid and battery is charged from grid until SoC = Charge to %. In both cases, when S0C = Charge to %, the battery does nothing and the house load is supplied from grid until the end of the timed charge period when ECO mode takes over.
TheDragon (GivEnergy) I already know about it also stops solar charge
If solar charge is disabled during a timed charge, that is good to know. However, there will be people hoping to use the timed charge to hold their battery at a level until the afternoon export slot and if solar charge during that period is disabled, it will be a missed opportunity for them. Enabling solar charge during a timed charge will be of benefit, I'm sure.
Tim I didn't think this was a "thing" any more
It is to the firmware developers. Thats what they refer to it as.
HY5-Gen2 - D0.996-A0-906 - Battery 4074.
Another difference found from the production firmware is with the chosen Charge Power and what is then reported via the App in Home mode, all assuming there is excess solar during the day or an active AC charge period.
Previously whether using the App or Web to set the value, the live Charge Power seen on the App in Home Mode would match until you chose a value lower than 390 W when the battery would be charged at 390 W even when the chosen value was 0 W.
Now the reported live Charge Power is higher than whatever you choose and set.
With a 9.5 kWh battery the increments are every 95 W, so too many to list, but these screen grabs give the idea.
So as a short term workaround with good solar gen days, I will be reducing the time period for the AC charge to less than my cheap rate period.
This will allow the house load to draw down from the GE battery for some of the time, with a guesstimate of what the end SoC will be.
Simon_C I'm confused, looks like you're trying to control the overnight SoC by varying the charge rate (or time). Why not set the AC Charge SoC target instead?
stevelewis , because on this beta firmware during the AC charge period with the starting SoC > Target SoC, the battery is paused rather than allowing the house load to draw down until SoC = Target SoC.
I expect it works fine when the starting SoC < Target SoC, but have not had that scenario since the firmware was loaded.
Simon_C what beta version?
I'm on 4074 (Gen1 AC) and haven't seen this.
Last night I entered time charge at 84% with a target of 80% and the battery supplied the house until it got to 79% then it topped up a little, drained a little, etc exiting at 78%
TX200 That oscillating is what got fixed in the Gen2, however they now made it pause totally if the SOC is greater than teh Target SOC. Which is not what I want it to do.
China say turn off AC charge to % and it will do as you ask, but previously it worked as expected, so I want it to now.
Its just explaining that in english they can understand. I really need a chinese intepreteter to exiapl tothen they rite it in chinese.
Long term, the issue will be gone, as firmware will be written in the UK. But for the next year at least we are where ewe are
TheDragon (GivEnergy) I really need a chinese intepreteter to exiapl tothen they rite it in chinese.
No you don't! You just did. hehe
TX200 , my system is HY5-Gen2 - D0.996-A0-906 - Battery 4074.
I presume this change in control strategy is due to the updated inverter firmware rather than battery.
TheDragon (GivEnergy) , doing as the firmware developers ask and turning off the AC charge to % is not the way to go.
This was the way to control the SoC to a target value at the END of the charge period, regardless of the SoC at the beginning. If the GE battery needs a top up then it gets it, if there is left over charge from the previous day then it feeds the house until the SoC drops to the target value.
It was the way to cater for fluctuating solar generation each day, whilst always having a known size buffer of battery capacity to ride through the expensive tariff periods that follow the AC charge to %.
I hope you manage to get it through to them to revert to the previous strategy, as anything else is a backwards step. I will have to consider coming off the beta test and reverting to the production firmware if this feature stays this way.
Simon_C yeah as the Dragon says it's a Gen2 firmware issue. I'm Gen 1 and the latest firmware hasn't come to us yet, hence I haven't seen that issue yet.
The oscillating was annoying, so it sticking/pausing when at the right percentage makes sense.
Although as others have said above, it needs to exit the timed charge at the required percentage, and using the battery during timed charge to get down to that percentage can be useful (especially if it's likely to be a sunny day!)
With the new Gen2 beta firmware completely pauses the battery for the timed charge period, whatever the SoC happens to be at the start will be the same at the end.
Previously I could align the AC charge period exactly with my cheap rate tariff AND my car charging time period AND reduce the grid import if there was leftover solar battery charge from the previous day. Now it is impossible to automate as you are guessing what SoC there will be at the start of the AC charge period.
Surely during Summer periods we should be aiming for 0 grid import, as long as the battery is sized to cover the hours with no solar gen, and excluding car charging when it is required.
TheDragon (GivEnergy) Good to see you're pushing for the correct functionality.
There appears to be a lack of system thinking: achieving a commanded SoC should be implemented in the inverter's control loop, not an external wrapper. If anyone wants the alternate functionality, it's far easier to write a wrapper which sends a single command to the inverter to set a target of max(Current SoC, x) than one which repeatedly polls until the SoC has dropped to a target value.
Inverter ED2248G812 - 5.0hybrid Gen2 + 2x 5.0 Batteries
Put on to Beta last week D0.996-A0.908 due to problems with fully charged battery probs. Full battery probs sorted but……
Seem to be having same problem with inability to charge battery from the grid, although it did charge up to and maintained 20% this morning. Setting timed charge doesn’t charge battery, just keeps the battery level constant, and runs the house off Solar and grid. I’ve set charge for 02:00-05:00 so that I can use cheap Flux to power house fo4 3hrs and save what’s in the battery for later tomorrow morning. Hoping for a sunny day tomorrow as battery happy to be charged from Solar.
colinchatfield I see someone from support is working this. looks like the FW was loaded and the Grid code wasn't reset. See thats done now as of 11:30 this am
Why didn't you get the Battery FW too? Not idea, but not the end of the world
TheDragon (GivEnergy)
Hi Paul. FW for batteries supplied by Fraser, Anne duly installed on both batteries. Called help desk and they set the calibration going. Discharged fine, and recharged up to 100%. Primary battery stuck at 54% for 15-20 mins but then up to 95%. Tried both charging and discharging this morning and all seems to work. Noticed the charge to batteries (from Solar) slows during last 10% to prevent overspill. Saw you’d used the glass of water analogy on here, and had explained this to my partner in the same way a couple of days before. Great minds eh?
Thanks for your help
Colin
colinchatfield No worries, Im trying to get that 10% reduced to 5%
I understand and was asking for the reason that when you buy a 10kWh battery if it’s 80% DoD you’re probably buying 10kWh of cells.
If you buy a 10kWh battery at 100% DoD you would expect to be buying 12kWh of battery cells with 10 accessible.
If people had been seeing the latter I would have said that’s as expected.
If people (similar to you are) expect 10 but are seeing 8, then IMHO there is a problem.
The best I've had on the latest beta firmware is 7.4kWh in for an 8.2kWh battery (90% DOD that should be 100%). This capacity issue is the only thing still outstanding to be fixed for me.
I personally am ok if I got a 90% system as long as this is what they state it is
TheDragon (GivEnergy) what is the situation with the east/west arrays and the strings not being fully separated?
kram Our gen 2 retrofit is happening tomorrow so we're also very keen to understand the E/W split issue. I think there was comment a while back that it was "close" to sorted but more clarity needed please. Our gen 1 has performed well other than SoC blips and low charge/discharge rates and we're keen to move forwards and not take a step back. Maintaining the performance of the arrays is obviously very important.
I have upgraded to the following Firmware Inv: D0.996-AO.908 and Battery 4074 (3.6 Gen2 with 9.5kwh) and the whole system is working much better. The issue with solar pv throttling when battery if near full has gone, the grid import when battery is near fully and solar pv is working has gone. It seems to solved the issue stated in my post above. I am also on Go and managed to top up battery between 0030 and 0430 with no issues.
Gordon , can you please confirm for your timed AC charge, what your target %SoC value is, and also have you had the scenario where your %SoC is greater than your target at the start of the timed charge period?
Been away from this thread for a while as my recent (now previous) beta setup (4075) was finally working well on my Gen1 3.6 Hy + 5.2kWh 80%. I notice, however, that at some point around the end of March I seem to have been moved onto a new set of FW w/o notification (197-197, 4074) and now it doesn't honour my max AC charge % setting for my ECO7 timed charge. I have tried restarting the inverter, resetting settings to defaults and reconfiguring but it still goes to 100% regardless of my max % setting. I have used the Remote Control page on the Portal to read out the values from the Inverter and they match what I have configured, but the inverter seems to be just ignoring this. I have now resorted to limiting this indirectly by shortening the charge window. Anyone else see this on the "latest" FW? Known issue?
Something else that has been wrong for a while now - during the AC charge window, despite me configuring a 2600W (max) charge, it only charges the battery at 1750W for a while, then it changes it to 2425W for a shorter time before it stops charging. The transition point varies but is somewhere between 35-50% (53.8-54.1V). Anyone else get this? Known issue?
JaseSpace The battery will charges at 1750w if it's temp is below 20 degrees (I'd really like this value to be a bit lower - my loft mounted battery is nearly always below 20) - once it's over that it goes to 'full' speed - could that be your issue ?
ianfretwell thanks for that info. Looking at the data on the portal, yes that could explain it! I had seen reports of lower charge power at low temperatures before but I thought that was <<20C. Is this a "recent" change to the temperature threshold, perhaps part of the beta changes? As long as it is expected I don't mind this too much, but agree that 20C seems quite high. If that is what the battery needs then so be it.
However, I'm not happy about my inverter ignoring my AC charge max % limit. It was working fine before the recent beta FW upgrade. I am curious to know if others also see this.
JaseSpace, is your starting %SoC higher or lower than the target %SoC?
The control strategy appears to have changed on the beta firmware for starting %SoC > target %SoC.
See earlier in this thread.
https://community.givenergy.cloud/d/1342-new-givenergy-inverter-bms-firmware-beta-testing-chat/2006
Simon_C Pretty much always from lower SoC (reserve) at start of AC charge period.
Issue affecting 8.2 & 9.5 capacity has been identified.
New FW 4073 is being written today, i should have late tomorrow.
Watch this space
TheDragon (GivEnergy) This is my number one issue as it currently stands so this would be great. Many thanks.
TheDragon (GivEnergy), since having a full battery calibration on 7/4/23, I have had 2 small jumps to 100 %SoC, with general usage between a minimum of 50/60% and regularly hitting a maximum of 100%. Is this to be expected with the beta firmware?
When watching the Home Dashboard on the phone with the battery showing 100%, there are short periods of 20-30 seconds when the battery discharges even when it causes a grid export, sometimes around 300 W but seen at up to 1.2 kW at times. Is this the battery cell balancing taking place?
HY5-Gen2 - D0.996-A0.908 - Battery 4074.
GEN 1 New Beta released.
Below are the bugs to test. email to follow, once uploaded to the inverter im managing
Nominal Power Reads 0W (Register xxxx) – This is preventing some inverters outputting any power to grid. Pinv reads only a very low power. Battery charges ok from grid or solar. Sending debug 0x********** does not set this back to 3600W as Register is read only.
Oscillating around a charge to set point (Winter Mode). Winter mode will charge to %% then discharge, and charge again.
Discharge Power Ignored (Register ****) – Setting is not working. Inverter will discharge at full power, even if this is set to 0W.
Solar MPPT Power Out while Calibrating – If SOC Calibration on discharge (register ****, value x or x) the Solar should be OFF.
Export Limit Ignored – (Register ****) is not used. Inverter will send too much power to the grid
EPS Not working if Discharge power not 100% - (Register ****) If battery discharge is not 100% value, the EPS fails to operate. EPS should ignore register **** and output what is required.
TheDragon (GivEnergy) Gen 1, that’s great to hear. Are you able to add us to the Beta now then? That would be great as we are desperate to get past the Soc high voltage errors that have plagued us since last Julys FW update 👍 SD2148G201
Simon_C
Hi Simon, I have not had the situation where charge was greater than %. I usually change the % based on the next days weather forecast. Ie have set at 40% and then 60% for the last two days. My house has limited demand (100 to 120w) until about 7:30am and then the sun is starting to come up. Thanks
mattpdavey too old for this version. 2209 or later. The older ones need remedial work.
Still deciding at high level what to do with the older batch
TheDragon (GivEnergy) any widening of the voltage limits (including 5.2s) or other fixes or is this one just limited to the 8.2s & 9.5 capacity fix (which is most urgent)
Noticed when one battery empties on one inverter, the other battery doesn't always take over and pulls from the Grid. Pause and unpause from the app seems to resolve it but it's catching when it happens.
TheDragon (GivEnergy) thanks for confirming this. Frustrating for us as the system was installed a year ago, so not sure why our inverter was such an old version. A hardware lottery I didn’t know we were part of. But after 9 months of this same issue and still no resolution, I am decided now to begin a warranty claim. If we need to be upgraded to a later model then that needs to happen now. We have been more than patient considering this isn’t just the SOC jump issue the majority have & in effect is far worse. Not your department I’m sure but it’s credit to your updates and pushing on the FW that has meant we have held off that bit longer, so feel it right to let you know our intended course of action.
4074 has been applied but seems to have interrupted my flux export for an hour before I noticed it. Is there any chance @TheDragon (GivEnergy) that we could get an email before any changes - even if it's just so we can keep an eye on the system for that period?
Apologies. Realised 4074 was applied ages ago. For some reason at 5pm today my export stopped. Nothing in logs or notifications to say why. I noticed later on and restarted it.
No obvious reason why that I can see. Discharge was turned on at 16:00, then at 16:55-17:00 was stopped.
kai Nothing what so ever done to AC3s
Sorry, not me gov
mpartington
They promised me improved optimisation, whatever that means.
But yes the 8.2/9.5 issue is the deal breaker to release
Depending on how many of the old Gen 1 batch there are that need a visit (mine is one of them) maybe the easiest solution would be to change them out using a limited number of installers. Could GE then upgrade theold Gen 1's and resell as reburbished.
I would be concerned that the current installers seem to have varying degrees of competence/care/knowledge and I would be very nervious of any installer doing any work on my current set up. I think I was spoilt with the installer who did mine!!
When we have a meeting to discuss the way forward, I will mention this idea.
Maybe easier than a specialist installer to modify the inverter internals
OK Folks
4073 has landed in my inbox
Some serious changes, accounting for different cell manufacturers, not all packs are the same internal cells. Now I never knew that.
Item 4, not too sure on this translation. So will wait and see what this does.
Version 4073 optimization:
Bear with me, this is gonna take a while to roll out
TheDragon (GivEnergy) Out of interest, what is the capacity of the 8.2 charge set to under 4074?
I always thought they were Gotion brand cells inside given their weird capacity (51AH, 102AH) so this is news. Kind of puts the claim of "manufactured in house from powder to power" in dubious territory though. Hope they know which ones in which battery by serial.
I suspect having different brands on the old firmware / original problem would not have helped in the slightest with different curves and characteristics...although perhaps explains why mine has been fine all these years if its got the "right" cells in it...
Edit: although lucky everyone if those AH values are for the usable capacity.
That gives 8.8kWh and 10.2kWh.
hoggy I suspect having different brands on the old firmware / original problem would not have helped in the slightest with different curves and characteristics...although perhaps explains why mine has been fine all these years if its got the "right" cells in it...
....and I wouldn't be at all surprised to find they didn't deem it neccessary to tell anyone what they intended to do or then document it either!!! Hopefully I'm wrong and they have changed their habits of a lifetime!!
Before rolling out on 8.2 and 9.5. Im establishing if a calibration is needed.
So dont all jump in at once
EDIT. Calibration needed. So will be emailing out before hand to get this agreed. No Im not doing them at silly hours of the night either. Before anyone asks. LOL
TheDragon (GivEnergy) Lets wait for a nice sunny day 🙂
TheDragon (GivEnergy) Is it only 8.2 and 9.5 that need a re-cal? I have 5.2 80%. Your email doesn't differentiate so I am guessing all battery types require a re-cal?
Can the re-cal be done entirely from PV (if sufficient) or will it have to use AC regardless?
Notice you say that bugs to look out for on Gen1 rollout is obeying the charge and discharge power set point 0=0 power, does this mean there is now a fix and does it apply to AC3 too?
JaseSpace The engineer said the calibration would be to reset the capacity of the 8,2 and 9.5.
No reason why the 5.2 cant be done as well, I am going to do mine too
If you reply to the email, I will be able to find you and get the FW loaded up earlier
Had the latest firmware 4073 installed earlier and the calibration started by emptying the battery. It's now been sat at 4% charging at 291W for over 2 hours...is this normal?
judgepd Mine's doing the same at the moment (291W/290W)
Could be handy if it can charge slowly till the off peak rates kick in. 🤣🤣🤣
andrew_bond Over 4 hours now and not sure what it's doing. Had an alert from the app to say calibration was complete but the battery hasn't refilled. Not sure if my usual timed charge tonight will interrupt the calibration process.
andrew_bond My timed charge kicked in at 21:30 - now charging at full speed :-)
That worked out quite well then!
Ok, I've gone early. Was planning on managing calibration discharge rate to hit zero at 00.30, but if it's slow charging for the first bit then need to bring that forward
Calibration started, discharge was fine then started charging at full rate to 65%. At 65%, 40 mins ago it went idle. The house load is coming from the grid, battery is doing nothing :-(
@andrew_bond How long did it take from starting to charge to hitting full rate? Was yours 2 hours as well?
judgepd
@andrew_bond
@Dw
All have kicked off, once they got over the cell balance
DW yours was an odd one, had lost its grid code setting. Very odd, as we only updated BMS
Not updating any more tonight. I have a family too. SO having a break now guys and galls
Potential bug, or unknown behaviour.
Charging from AC, if you have an AC charge to % set. Im seeing inverters stop at this %.
If affected, set the % to 100 and see if it carries on the calibration.
This not wanted behaviour on the Inverter FW
TheDragon (GivEnergy) That fits - I do have an (inactive) AC charge to % set at 65% from when I was testing the last inverter beta. Thanks for resetting the grid code, now charging again and well on the way to 100%
TheDragon (GivEnergy) another potential calibration bug - My inverter hit the bottom limit at full speed @ 3kW, with no throttling. I thought previously it reduced charge rate as it approached empty.
It then started to charge back up at 487W, which seems higher than others above (maybe because I have 2 batteries, so it's higher?!). I did manually limit the charge rate in settings to around 522W for an hour until my off peak rate started and it reached this power quite quickly.
Mine finished ok (I had soc target set to 100%), however only 10kWh was added during the calibration (10.4kWh was expected as 2x5.2kWh), maybe related to battery temperature, although this was above 20deg during the charge period. So potentially a little down on capacity.
Expected voltages seem a little off based on one of your previous posts.
Peak battery voltages were 56.5V.
Minimum voltages were 45.7 V
TheDragon (GivEnergy) still seeing charge/discharge flutter with target set to 99% on 4073. It appeared for 15 mins on hitting target, 0.5kWh discharged in total. Then trickle charged to target. Could be an overshoot in the control loop.
Hi all, I have a new 3.6kW Gen1 with a 9.5kWh battery installed a few weeks ago and I experienced my first SoC drop last night - 17% to 0 in 5 minutes, recovering to 4% 10 minutes later. I'm on firmware D0.450-A0.451, BMS 3008, so happy to try the new beta if that improves things. Inverter serial is SA2237G358. Many thanks in advance!
mpartington my calibration started about 17:30, discharge took about 30 minutes, then it charged at 291W until my timed charge at 21:30 - so 2.5 hours of slow charge.
Hi Simon (not sure if this is what you were asking me). My system is performing really well with latest firmware. Only small bug is that I set my recharge % to 20% last night and my SOC was higher. (Test). For the charge time period (2 - 4am) the house load was pulled from the grid and logically I would have thought it would have pulled from the batter until the point at which the battery was below recharge %. I was awake at 3:30 and checked and noticed the bug, at which time I removed the time block / charge % and the system reverted to pulling from battery.
stevelewis I saw the same, though my target was set to 100%. It got to 99% and then started discharging at 421/422W, then every now and then would charge at full rate for a few seconds, and repeat; this went on for about 40 minutes, then it stopped doing this and looked idle until the end of timed charge (still reporting 99% the whole time)
My battery was 53.7V when it stopped charging; looks like it went down to about 46.9V when discharging (there were some spurious readings so hard to be sure). Total charge from 4% (maybe close to 5%) to 99% was 7.6kWh (8.2kWh battery), so still rather low.
Gordon, yes that is the issue I was highlighting.
On the factory firmware previously with starting %SoC > target %SoC, during the first part of the AC charge period the battery would discharge to meet the house load until the %SoC fell to the target %SoC. From then on the battery would do small charge/discharge cycles to maintain the %SoC at the target value and the house load would be largely drawn from the grid.
It was very useful for utilising left over solar gen from the previous day, whilst also providing an automated low limit for your battery %SoC in order to cover the expected solar gen and house load for the current day.
mpartington TheDragon (GivEnergy) My inverter (5Kw G2 + 9.5 battery) also hit the bottom limit at full speed @3kW (maybe because the house load was high at the time?). Looking at data reported in GivTCP:
A couple of strange anomalies in the GivTCP SOC reported from the battery and inverter:
Dw Those TCP plunges and holds at 2% are a glitch in TCP. SOC under 3% crashes the entity.
SO you might have an automation restarting TCP if this is seen, explaining the jumps back up later to where it should be.
@TheDragon (GivEnergy) I got a notification on my phone at the end of the discharge part of the cycle to say that calibration had completed (the same happened last time calibration was done)
This seems a bit premature if you look at your table of calibration stages.
I did look at the SOC_ForceAdjust register a few times during this, and it certainly doesn't match the expected behaviour in your table (but does explain the timing of the notification); it got to stage 5 before it finished discharging:
17:37 1 (discharging at max - SOC=10%)
18:49 5 (discharging at 9W - SOC=0%)
18:57 5 (discharging at 17W - SOC=0%)
19:14 0 (charging at 290W - SOC=1%)
So now I'm wondering - has my system really done a proper calibration?
Dw my capacity is wrong following calibration as well
Last time I calibrated (under old firmware) these were around 105. Design capacity is 102
andrew_bond time will tell.
Give a few cycles and then from flat. Do a full uninterrupted charge.
See what goes in.
Matt same for you, that value altered as the BMS learns. Mine didn't change from 102, 101.97 and 95.7
All should be 102Ah.
I'm going to give it 2 full cycles and re evaluate
Weirdly my battery has just dumped it’s remaining load to the grid. It’s been on eco since it was installed 2 months ago.
At around midnight it dumped the remaining 9% capacity to the grid. I only noticed it as my Bright meter registers negative draw.
My inverter is SA2226G445. I’m not on any beta firmware and look like I’m still on the standard release.
Any ideas?
Do the 9.5s actually still run at 80% DOD?
Going by the sticker on the side of 9523.2Wh usable - that makes the 100% design capacity 232.5AH which is an unusual value no matter which way you divide it up (i've managed to finally confirm from the one someone sold split in to parts on eBay they have 32 cells inside - 4 blocks of 8 cells)
I've a feeling they are actually 120AH ones (possible CATL) so more like 240AH full capacity although would have thought they'd be marketed at 9.8kWh packs then.
TheDragon (GivEnergy) Will keep an eye on it. It did put 10kWh in during calibration, which is better than the capacity now stored in the registers.
Looking at the calibration cycle, the Master had an 18% SoC drop at the bottom point and then a 17% jump to the top limit. The slave had no drops/jumps. Wonder if that's messed something up in the calculation. I seem to recall (and it may have changed since the early betas) that battery capacity flag wasn't actually used in the software anyway
Looks like 4073 has arrived - spotted it because the timed charge max charge rate wasn't obeyed (seen that with previous FW updates and have to slightly change the setting and save it to fix it).
Usual:
Last two charges:
Showing oscillation still?
Tests:
EPS test. Seems to work at a variety of discharge rates not just 100%
Had trouble getting the socket to work but think just a dodgy indicator (said it was live but wasn't). Perhaps button wasn't quite pressed in enough.
Did work okay in the end at a variety of discharge rates, plus tested the test/reset a few more times, and it seems better now.
Timed export. Seems to work as planned.
60w export before timed export started then 482-487w export at timed export period (expected amount, rather than full). At end of period it went to 58w export. Looks like that's because there is some output from the panels, not much, quite grey out!
Timed discharge. Seems to work as planned, but some issues getting back into ECO mode.
Eco mode now enabled. 0w showing on IHD. Some solar going to battery now instead of grid.
At timed discharge time, 471w export (as household usage is low). Switching on electric fan heater, pulling from grid (as expected due to low discharge rate).
Back to exporting a bit after turning fan off.
Back to 0w grid usage when timed discharge period ended.
I then disabled timed discharge - it also disabled eco mode. Appears enabled in the portal but logs say it was set to disabled and IHD reading small import from grid.
Had do disable eco mode and save, then enable and save to fix it.
Nominal power - assume okay, as have managed to discharge/export as above!
Solar MPPT power out during calibration - didn't calibrate (5.2Batt) so can't test. Although I assume this is for Hybrid only (AC Gen1)
Cheers
TX200 Thanks for your feedback.
EPS should also be faster, I've enabled 0s switching for you, now I know you've got EPS. PLEASE TEST switch speed
Oscillation is not yet fixed in AC3, thats an inverter update to 288 im waiting on.
The leack power when the sliders are at 0w, is known, there are new Pause features coming, with paused timers too. The sliders will eventually be 300w to max. Not 0W.
If you want 0w, pause.
Yes the MPPT off, for calibration is Hybrids only, both fixed.
The timed discharge stuss, is awaining a portal update, to make use of the timed pause function in the firmware. As well as the 10 charge and 10 x discharge schedules.
mpartington Yes could have, its not used no, only set on calibration.
TheDragon (GivEnergy)
Yep, EPS fast switching worked. Can't say I spotted the light going out on the extension lead when I threw the big switch and vice versa when I went back to mains.
My neighbours definitely noticed the power coming back on - had to run to the burglar alarm panel to reset it (think I need a new battery in that!!!)
TX200 Cool, I will leave UPS mode on, as a gift for being a beta tester
Positive observation - Early days, but my 2 batteries SoC are stuck together like glue. Previously would have started to drift apart a little by now.
Either the calibration (my first on the beta), or I suspect a change in 4073
Used the entire battery capacity last night down from 100% SOC. GivTCP battery discharge indicates that I used 8.7Kwh which took SOC down to 6% at 1.45am. The battery then went idle which was odd as reserve is set at 4%. At 6.50ish this morning, the battery started discharging again at roughly the same time as solar PV starting. Another 0.2Kwh discharged before solar covered house load. The good news - 8.9Kwh + 6% is around 9.5Kwh.
Dw Pure speculation but I may have worked out what caused the battery to go idle at 6% rather than 4%. I was looking at the Battery to House view in GivTCP and noticed it was spiking to 3000w every 20min or so overnight. The spikes were only one data point in GivTCP so for very short time period. Guessed this may be my Quooker tap keeping the water in the tank hot so experimented by turning it off last night ... no spikes in load overnight.
When I turned the Quooker back on this morning, I got the spike as expected but for a longer period of time (around 4 mins). It was accompanied by a significant drop in battery voltage. I'm guessing the same is happening during the shorter spikes, but I was just missing it in GivTCP as the voltage data point timing is not exactly the same as the battery to house view.
Is it possible that when the battery was at 6% SOC, the Quooker kicked in briefly and the battery voltage dropped to a level where the BMS thought it was below 4% SOC very briefly (between data points reported in GivTCP or on the portal). Even though the voltage and SOC recovered quickly, maybe it figured that it was at the 4% reserve level and went into idle/standby until the solar woke it up the next day.
Dw Thanks. Thats comforting to hear, hopefully today a few more are in the right ball park, 8.2s as well
@mpartington Yes there were SOC optimisation, the addition of 2 additional maps, 5 in total now. Based on cell manufacture.
Time will tell if this is what the issue was, as not all cells are made equal. The differing cell sources, is new to me.
Looking forward, wont be an issue once we manufacture in the UK, as all cells will be made from UK materials all the same, bar tolerance
My charging pattern is still the same after three cycles - 4% to 99% takes about 7.6kWh (i.e. 7.9 for 0-100%), and when it reaches 99% it does about 40 minutes of oscillating - nominal 421W charging with spikes of 3kW discharge (the SOC stays at 99%). Then it sits at 99% with no charge/discharge until the end of timed charge (it's hard to see on the portal at 5-min intervals but GivTCP shows it (there are also spurious spikes showing on the house load which I believe are not really happening): 
Before installing the beta f/w the amount needed for a full charge (calculated for 100%) averaged 8.5kWh, on 4074 it was 7.5, and on 4073 it's 7.9
andrew_bond sorry, just correcting myself: the grid spikes do look like they are really happening - I can see them on my Bright figures (which are recorded every minute)
I really wish the EMS was available. I asked for the new firmware to be applied first thing today as the sun is out and thankfully it was at 8:45am, however it had already charged from solar from 25 - 34%. Discharged finished 11:05 as it just kept finding more and more power. Its now 15:46 and only at 70%. Solar is down from 6kwh to 3.9 and dropping. Generated over 43kw today but can not capture it fast enough. The next couple of hours could be expensive, about to enter peak hours on Flux!
THALL Switch the consumer unit off and go to the pub for tea?
THALL is it not supposed to isolate the solar during discharge at calibration to avoid this ? I’m sure mine did last time.
Tim even more expensive :-)
@Leon_shaw I have an AC3. Thankfully the solar got the battery to 95%
However both batteries are at 100% but are still drawing 945w from the grid, has been for an hour. So this morning they dropped to 0% and kept discharging at full speed for just over 20 minutes and now at 100% full it is still charging at 945w for the last hour
THALL However both batteries are at 100% but are still drawing 945w from the grid, has been for an hour.
Ours took about 90 minutes once it got to 100% SOC, which appears to be the batteries balancing.
Woodlands Mine just finished. 1 hour 10 minutes pegged at 100%.
Looks like 18.34kw went into my 9.5+8.2
THALL Your took an age to rise to 56-57v. Was about to email you, as Ive been watching it ALL day
Hopefully they are good now. Calibration should not be needed again, unless we use a totally new map. Which i dont think is now needed.
andrew_bond How old are your batteries? 13.5MWh through them, capacity is down 5% looking t teh cell data
TheDragon (GivEnergy) I see it jumped to 57v pretty quick right near the end.
Do you mean no more calibrations ever? I am not against calibrations, I just wish the customer had control over it. Plus the EMS and a second AC3 would help :-)
TheDragon (GivEnergy) The kit was installed almost exactly 2 years ago - 21st April 2021
Latest firmware (4073), battery still discharging to cover load when during a timed charge period and then periodically charging up again. Presumably this is already on the list of things to fix?
4073 installed yesterday and some oddities on the calibration, but more on that later. My immediate problem is I am trying to run a complete discharge/charge cycle today but the battery is refusing to discharge past 25% SOC. I set a timed export to grid (3000W discharge power, battery reserve 4%, battery cutoff 4%) and the battery discharged as normal down to 25% and then abruptly went into an idle state of its own accord and is refusing to discharge any further.
Tried turning off timed export, turning on Eco mode and then reinstating a timed export but no joy. Battery voltage dropped to 48.5V at its lowest point and has recovered to 51.8V so far since being in an idle state. Is the battery actually empty and just over reading its SOC by 25%? Bit worrying if it's that far out immediately after a calibration!
Update: Disabled the timed export and enabled Eco mode. Plenty of excess solar so battery started charging again. Let the battery charge to 27% then reinstated the timed export. Battery discharged (blowing my G100 export limit) but stopped again as soon as it got to a displayed 25% SOC. Battery voltage once again stabilised at around 51.8V.
Now going to just let the battery charge in Eco mode and see what happens. If the displayed SOC is really over reading by 25% (21%?) then I expect there will be a sudden jump later in the day!
Woodstock Also seeing some issues following calibration and update to 4073 on 15th. I'm not 100% sure the calibration completed correctly, as it hit the bottom limit at full discharge speed and abruptly stopped, before charging. Normally it spends some time eeking out the last dregs of capacity - did it record bottom limit correctly?! The capacities recorded as part of the cal for each battery were well below the design capacity (previous post). Hoping these will correct with use
Scroll forward to today, first time I have neared bottom limit since update. The SoC has dropped from 30% to empty (2 x 5.2 kWh batteries) whilst discharging only 600Wh.
I left it for a couple of hours at idle and voltage also stabilised at 51.8V (so I would say you are empty). OCV didn't find any more capacity during that time. Now charging on solar, so curious to how much goes in.
It's disappointing that the SoC curve is so far out following a calibration. Hopefully the software will learn.
If you consider at 30%, I only had 600 Wh additional discharge, then that would not have been a long way off the 25% you are seeing. Difference being mine adjusted (crashed) to zero (4%).
mpartington Well the plot thickens on this one. My recharge to 100% SOC looked perfectly normal with a ramp down from 90% SOC and a stabilised 100% SOC voltage similar to 4074. There was no sudden jump in the displayed SOC % and the charge figure was consistent with the battery having charged from 25% to 100% (not 0% or 4% to 100%).
Now I'm really confused. Was the battery truly at 25% when it stopped discharging (if so why did it stop discharging)? Or was the battery empty when it stopped discharging, as the voltage implied (if so why did the SOC show 25%)?
I have another timed export under way so we will see where it stops discharging this time!
Well at least it is consistent!
Stopped discharging at 25% SOC again and battery voltage stabilised at 51.9V.
What on earth is going on?
(The battery charged 5.8 kWh going from 25% to 100% and discharged 4.8kWh going from 100% SOC to 25%).
Cut off doesn't exist in the new firmware does it?
Still available on the portal as an option, presume it'll go from there when the beta firmware becomes the live firmware?
Maybe something got messed up, might need a reset of some settings by GE?
Woodstock This firmware needed a calibration to reset the limits on 8.2s, maybe something has done wrong on yours, you mentioned oddities?
Mine only added 7.8kWh back to full (was hoping for at least 8.3kWh, assuming 80% DoD). Had a small 2% jump overall, nothing on primary and a 10% jump on the secondary battery to hit 100%.
Last 10% took 35 min and only added 0.5kWh.
Battery cell temp started at 12 and reached 20 Deg C.
TX200 I did wonder about that. The last non 4% value that battery cutoff had on my system was 25% on 12/3/23. This was when I was testing the mysterious spike charges that occur at 01:00. The cutoff value was reset to 4% on 12/3/23 and (prior to the 4073 calibration) the SOC% has only been down as far as 25% once since then. This was on the night of 29-30/3/23 when the SOC% went down to 4% and I did have an unexplained spike charge at 01:00.
Remote Control shows the battery cutoff is currently set to 4%. The Remote Control History shows the battery cutoff being reset from 25% to 4% on 12/3/23 and no changes since but, knowing Remote Control's propensity for "gas lighting", it is anybody's guess what the current value actually is!
mpartington My calibration did a full bore discharge down to 44V and then performed a SOC% Saint Vitus dance jumping between 0% and 5% (while chucking out a multitude of meaningless battery voltages and temperatures) for around 15 minutes. The battery then charged at a steady 2900W up to a displayed 100% SOC followed by a further 2 hours charging at 945/946W. The battery took in 9.4kWh (should have been 10.2 kWh?) during this but only achieved a maximum of 55.28V and a stabilised voltage of 52.5V.
What's the latest consensus, still on track? Seems to be a few glitches surfacing after a fairly positive few weeks? Any news of a formal production release and how it might be rolled out? Also any facility to roll back if issues are found? Lots of questions, apologies for that, but I think we'd all appreciate a status update if possible please.
Curiouser and curiouser.
The battery has now discharged below 25% SOC.
Was it objecting to the rapid discharge or possibly the timed export or the presence of solar PV or has an unseen setting changed somewhere?
Going to disable solar PV now to give it the best chance of discharging down to 4% SOC in the morning; watch this space!
Really odd one this morning. Battery went idle again at 6%. Quooker off so no spikes in load disproving my prior theory for why it stops at 6% rather than reserve value (4%). Pre beta I did have it set at 6% at some point but it has been 4% since on beta and I’m pretty certain 4074 was respecting the 4% reserve.
More strangely, after the battery stopped discharging it’s stayed at 6% for about 45mins but then gradually stepped down to 2% and stayed there … but the battery voltage was increasing at the same time the SOC was decreasing. Any ideas why the reported SOC would drop when voltage is increasing and no discharge happening?
It's been a while since mine got that low, but think mine always used to stop at 6%, although it did go lower over time, which I think might be the inverter powering itself from battery?
Still getting the same issue on 4073 of the charge & discharge oscillating when the SOC is set to less than 100% for the Timed Charge period. So is this the Inverter Firmware D0.289-A0.289 causing this?
Noticed it 4:30 this morning, so changed charge rate to a very low setting & SOC to 100% which stopped the oscillation.
Dw The beta software has a feature where it disconnects the batteries at empty and measures the voltage once it has stabilized (it used to wait about 30 mins after discharge finished, but may now be faster in 4073). It's my understanding that its then meant to align this to a SoC based on a voltage map, temperature etc. This can result in further discharge if the battery is less empty than expected. I assume in your case, it was maybe more empty and it adjusted the SoC downwards to match.
Woodlands apparently that's still to be fixed. Next beta hopefully 😀
Woodlands yes, fix for this hasn't made it to the AC3 firmware yet. Temporary was around it is to set the reserve to the SoC target, however this really needs home assistant to automate, as it needs to be set back to 4% at end of charge period. Otherwise battery will not discharge
... and now for good measure
an 01:00 charge spike.
Should be down to 4%/empty soon, so we shall see what happens then.
judgepd YES.. Hybrid fixed, AC3 will get fixed, only happens on a set % not 100
Woodlands YES.
289 hasnt got the Oscillating around a set point fixed. Only the hybrids. The AC3 is proving to be a PITA to fix
Woodlands Thanks for this clear example, Ive sent this to the firmware team, to remind them that 289 AC3 has this issue and I want it fixed
Some good news, we have had tested the Grid over voltage issues. Where the inverter over 253 goes into alarm.
With 289/289 it alarms at 262, voltage was raised in the lab to 265v, inverter went offline, alarm, as expected. Voltage reduced to 256 and it came back online, without a restart.
TheDragon (GivEnergy) Not sure this is correct, mine was set to 20% last night for my 3 hour cheap rate window and I have lots of spikes where it is dipping and charging from the grid. Or is this the oscillating issue on the AC3's?
THALL Yes that's currently correct, expected, wanted, no. 288/288 will fix it once released.
Ask those with hybrids, it gets to the set % and stops dead.
THALL Looks like AC3 hasn't been fixed...only hybrids.
Okie dokie. Well that will be a nice thing to fix because at first look I think it sucked in 500 - 700wh from the grid rather than just stopping.
TheDragon (GivEnergy) this won't introduce the issue hybrid users are seeing with the battery not discharging down to the target soc during a charge period will it?
(Where soc at the start of the charge period is higher than target)
mpartington this is still in debate.
What is the correct way
The firmware team say its correct now.
If you are trying to charge, a discharge doesn't make sense
TheDragon (GivEnergy) The way it's always done it is the correct way surely. If your target SOC is lower than your actual SOC then it should discharge. I can't see a reason to change this behaviour.
TheDragon (GivEnergy) Ask those with hybrids, it gets to the set % and stops dead.
But you're saying it will only do this if it's below the target SOC? If it's above it it just stops dead as soon as it enters the timed charge period?
TheDragon (GivEnergy) I struggle to see why there is any debate on this. If your current SOC is above the target SOC, by definition you won't be charging and you should allow the battery to discharge down to the target SOC (which it might not do). The two options are:
a) draw from the grid the whole period to achieve/maintain a target SOC or
b) only draw from from the grid during the period if you have to, to achieve/maintain a target SOC.
Going to a) would be a bad step backwards.
TheDragon (GivEnergy) If you are trying to charge, a discharge doesn't make sense
Can certainly see a justification for that: if you are considering forcing an import from grid, then it could be inferred that the grid is relatively cheap, and so it makes sense to power the house from the grid rather than the battery.
Say I'm on Go tariff, and I charge car from grid overnight. I might have a solar forecast thingy that predicts there won't be a lot of PV tomorrow, so I've set it to have a minimum SOC of 50% during the cheap rate. But I don't want to discharge the house battery into the car battery if it happens to be at 55% already.
It comes down to whether you label it "target SOC" or "minimum SOC".
DD so I've set it to have a minimum SOC of 50% during the cheap rate. But I don't want to discharge the house battery into the car battery if it happens to be at 55% already.
Are you saying you would want it to charge up to 50% but not discharge down to 50%?
DD I prefer the flip side for this myself. Say the weather is predicted to be very good tomorrow so will easily top up my battery, but I have my car plugged in for an overnight charge, but I still have a large amount left in my battery.
I currently set my SOC to say 20% so the battery contributes the remainder to both house load and being dumped into the car during the cheap rate period, which retaining sufficient to allow me to continue to run on batteries for the morning until the PV kicks in to recharge things. I really don’t want to lose this functionality (AC3) - apologies to admin as I’m not on the beta here
Smk82 100% what I would want too...as it operates at the moment.
target percentage at end of timed charge is what I want.
If it enters timed charge above that, use the battery to power the house until it reaches the percentage.
If it enters timed charge below that, charge up and power the house from the grid.
If you get two sunny days in a row, the battery might be full in the evening but needs to be emptier for the next day. Okay so you could just stick the washing machine on or something...
4monks I'm in agreement on this -- I don't want the house pulling from the grid if there's plenty in the battery AND I want to be able to determine the minimum charge level I leave the cheap period at.
TheDragon (GivEnergy) I think it would be a backward step to change the current method. At the moment it is easy to define a time window during which I am happy for the battery to charge on grid (if necessary), I can also specify a minimum SoC I want maintained during that period. This allows me to automatically allow the system to either top up my battery or allow it to discharge towards that target to create headroom for the next days solar avoiding pulling from the grid. I need to change 1 parameter to achieve my aim.
If you change the current logic, then I need to set 3 parameters every night (not just soc target)
How would Giv implement an alternative solution, as people will be mightly annoyed by this if the capability is just taken away.
judgepd Are you saying you would want it to charge up to 50% but not discharge down to 50%?
Yes, conceivably. Though it is of course more complicated:
So since only one of these can be implemented, I guess I'll need some external agent code to force it to behave the other way when needed.
(Actually, that said, with current Flux prices, it's still better to leave the house battery as-is, and then export the excess to the grid during the day, rather than discharge the battery to the house during the night: not only does it reduce cycling, but I earn more by importing overnight and exporting instead during the day.)
Smk82 I prefer the flip side for this myself. Say the weather is predicted to be very good tomorrow so will easily top up my battery, but I have my car plugged in for an overnight charge, but I still have a large amount left in my battery.
That's certainly one way to do things. I might instead choose not to cycle the house battery, but charge the car from tomorrow's excess solar instead.
DD my car isn’t at home during the day except at weekends when the sun is available to me, hence this way at least I get some benefit to the car (and reduces what I need to pull from the grid) Solar charge at weekends where possible.
TheDragon (GivEnergy) Just to comment that my AC3 is also oscillating when the target is 100% - it gets to 100 for a brief moment, then drops to 99 and will oscillate for about 40 minutes before stopping the oscillation (unsure why it stops though, since the SOC is still 99)
@mpartington
Send this back as feedback. Pack balancing without an OCV happening at idle, I have simulated this by part charging at night, when not really needed.
Similar to what you see.
Poor pack balance
TheDragon (GivEnergy) yep, looks familiar 🙂
Exactly how it progresses for me, starts off fine and splits get larger each day. Glad you've done that, starting to think it was just me....
I see you had a few little drops/jumps on calibration as well and a potential fast drop to empty on the 15th
Gen1 3.6 Hy + 5.2kWh 80%, 4074, 196-196
Just before 2pm today I decided to reduce my overnight max AC charge % down to 30% as expecting a bit more sun tomorrow and wanted to check the AC % limit setting was working properly (it wasn't for me previously). Here from the portal log you can see that was the only thing I changed today:
However, it seems that just after that, my system decided to dump my entire 100% SoC, then charge back up to just over 30% (approx my new setting) from a combination of PV and AC, then switch to PV + AC trickle. Now I am struggling to get back up to 100% SoC with what solar there is left in the day. That AC is costing me ECO7 peak rate so not happy!
Now I haven't requested a move to 4073 and the portal shows I am still on 4074, so although this looks like a calibration I don't think it is - unless someone from GE knows otherwise?
DD it's easy to set the SoC target to the current value to prevent discharge in the cheap period if that's what you want.
But I want the battery to continue in Eco mode until it drops to the target - just like it used to. To write code to do that is harder as it requires constant polling as the SoC reduces. The inverter's own control loop can do this far more easily.
JaseSpace So it carried on charging from AC so I decided to try and stop it. Restart Inverter - nope carried on. Reset to Defaults - just increased the charge power to max and got to 100% from AC - I guess this is because the 30% AC charge limit had been disabled.
So what on earth happened? It looks like my simple single parameter change at 2pm (which should have had no impact until my ECO7 charge period tomorrow AM) immediately triggered a process that looks very similar to a calibration cycle and drew from AC in my peak rate. Not what I wanted at all!
@TheDragon (GivEnergy) any ideas what happened here then (see earlier post above with more details of what I did).
TheDragon (GivEnergy) , now seeing repeated charge/discharge blips from the battery when %Soc is 100% and there is plenty of solar excess generation. Both at the tail end of the battery calibration yesterday and also today.
Is this expected operation, as in my mind it is just inefficient cycling of the battery instead of a complete pause at 100%?
HY5-Gen2 D0.996-A0.908 - Battery 4073.
Stevef1964
Also having an early serial number installed last August SD2148G069, I agree that it would make sense to just replace them rather than swap out certain internal parts, As an engineer there are reasons this is rarely done, as you risk creating issues later down the line, you never expected, as some times it takes time to show.
I'm not on the beta, but I wanted to add my input on the Timed Charge piece. The way it currently works is as it should be IMHO. I want to be able to set the desired SoC that I exit the TC period with. For example, my system generated 30kWh yesterday. My car was set to start charging at 23:30. I had a full battery until 18:30 which discharged to 85% by 23:30. I knew that there was to be full sun today, so I set the TC SoC to 20%*. The system then discharged into the car overnight, taking advantage of the free energy generated during the day to partially charge the car.
It would be patently ridiculous for the system to sit at 85% SoC, using grid import, to charge the car just because I had set a low target SoC.
*I would set a lower SoC if I could. However the stupid system that GE uses does not keep my twin batteries balanced. I consistently end up with the primary in the red at 4% and the secondary at >30%. The dopey system continues to draw from the primary even when there is loads in the secondary. 😡
Found a different way to prevent the oscillation effect with the settings I tried last night, as expecting a very sunny day today (Solcast Forecast of over 40kWh).
Set the reserve to 20% & Timed Charge period to 04:30-05:30 @ 100% SOC.
Resulted in the graph below:
Battery discharged until it reach 20% SOC, then grid took over supplying house load/car charging on low rate, then when the SOC dropped a few % below 20% it appeared to trickle charge at just under 300w until 04:30 when the timed charge period started charging it at max rate, ending up with about 30% SOC.
Zakalwe Also not on beta, but totally agree. In my case it is the secondary which is the lower by up to 25% at times.
I saw the same, but it calmed down on the second day.
stevelewis but more to the point, how can anyone design a system without a clearly-written set of requirements to define the required behaviour? As consumers, we shouldn't be having a debate about design intent!
stevelewis seems to be a language issue - UK company and Chinese firmware developers.
Hi all,
Long time lurker, finally on the beta (GIV-HY5.0, 8.2kWh battery; D0.996-A0.908, 4073) after suspected issues with mppt causing a weird ceiling around 2.8kW on a 4.92kWp system.
Now that I'm on the beta it's much improved. However, I'm still seeing a ceiling at around 3.8kW. I do see momentary peaks of about 4.4kW (which is around what I'd expect) but after the peak it drops way off and then slowly ramps back up to around 3.8kW again. There was no shading while this was happening as far as I could tell. Does anyone have any ideas what would cause this? Are there further mppt tweaks planned that would fix this or is it some other issue? I was previously on a Gen1 and it was happily working away with no apparent ceiling.
Another issue I've noticed with the beta is the apparent inability to disable battery discharge. When I select the feature in the app it throttles the battery to around 200 W, but it doesn't turn it off. This is a feature that worked very well on my Gen1.
Cheers,
James
JaseSpace Hmmm... I'm not on the trial. Lowered my 'charge up to' setting last night to 50% in light of the good forecast for today and it almost certainly wouldn't have got down to 50% anyway. Checking this morning -- that small change seems to have triggered an almost immediate discharge to grid -- and it went down all the way to zero. Then proceeded to charge all the way up to 100%, behaving as per calibration charge. Annoyed to say the least that the system has drawn 2 hours or so of expensive electricity and is now sitting at 100% so all solar is being dumped to the grid for a pittance!
pelowj I was due a retrofit of a gen 2 last week but it didn't happen. The plan was to immediately drop on to the beta on the basis that the E/W strings issue and other glitches have been fixed. But maybe not? I'm managing the odd SoC issues with the gen 1 reasonably well but still need the gen 2 output potential as we have dual ASHP going in. Is gen 2, particularly with E/W arrays, even viable yet? Still more clarity needed as it feels the testing has hiccuped again recently.
sponev My reason for pushing for the Gen2 was my requirement for wired network connection (work wifi at home) but I kind of regret it now! The Gen1 was working perfectly.
My array is a single SSW facing string so I'm not sure if I should be seeing these mppt issues.
-James
pelowj I have been experiencing this problem too. I didn’t notice it before installing beta although it becomes really apparent on clear days with constant sun and granted there wasn’t much of that prior to beta firmware for me. I have reported this and am awaiting feedback.
Leon_shaw Glad to see it's not just me. I was starting to wonder had a diode died on my array or something like that! What's your kWp?
SimonP Wow OK that sounds that same as me then! Since you aren't on the beta FW (trial?) and this is new unexpected behaviour, maybe this is a recently introduced issue with the portal/app interface then?
@TheDragon (GivEnergy) any clues?
pelowj I have 10 x JA 405 w panels so should be able to achieve up to 4kwh with no problems I have regularly gone over that all be it momentarily. System is only 4 weeks old and has been checked.
What panels do you have fitted ?
JaseSpace I can only assume that something went wrong with altering the inverter setting and the 'glitch' caused the calibration (like?) behaviour. Peculiar that it happened to you as well though, even though my incident was several hours later. I do hope that a 'bug' hasn't been introduced somehow into the portal software. This morning I've done a 'reset to defaults' then restored my usual settings (Eco and an overnight maintain %) and it seems to be behaving normally.
Leon_shaw 12x JA 410 W. I had a Gen1 since Sept 22 and it worked perfectly so I don't think it's a panel issue. Interesting that they're both JA panels though.
pelowj my data from yesterday has been sent for review to the FW team. Have to see what they come back with.
Like you say though if gen one is performing better than a gen 2 with it’s upgraded specs there is definitely something wrong.
Simon_C did you get any feedback on this. I am experiencing the same thing ?
pelowj Interesting indeed, 3 systems all do this all JA solar.
Other dont see it. The plot thickens.
If i had your inverter serial I could gather evidence for the engineers and add to the pot of ammo that something isnt right. Maybe just JA solar, but thats not for me to decide.
LAB PV SImulator peaks at 13A as expected. 
I am still seeing the 2% jump at the top even with the latest everything and I have just noticed the app reports both of my batteries as only 97% charged?
TheDragon (GivEnergy) ED2152R267 is my serial.
Today looks like it's turning into an absolutely perfect day, so some good data there already. The apparent ceiling around 3.8kW is already manifesting itself.
-James
pelowj Argh found you.
Yes if at the end of the day can you email this in, if you have any previous graphs of the previous FW on a clear day too please.
I think this issue isnt this FW, but something far deeper, im gathering data, so please if you can send in a today and a few days ago, with a clear day
Simon_C I too am seeing this behaviour - Only appeared since BMS 4074 was installed on 2nd April.
I too have Solar JA panels and 3.6 Gen 1 hybrid inverter.
Also BMS 4073 installed last Friday and the same behaviour is happening today.
ALL BETA application is closing, now, we wont be taking on any more.
Way too many as it is to manage. Plenty of data, plenty of weak points found in peoples kit and even the odd bad battery, booked engineers to, where the customer unaware they had cell issues.
CCourtney
This from today:
The discharge is always at 171W and the charge is always at 121W.
TheDragon (GivEnergy) Hmmm. We had agreed to load the beta to a gen 2 retrofit but that was delayed and now could be another week or so away. I'm concerned that there seem to be generation issues with gen 2 on the non-beta firmware, particularly E/W likes ours, and hence the retrofit may be a retrograde step?
Would you advise upgrading please or holding off? I've been pushing the installer really hard for the retrofit so holding off presents its own huge issues unfortunately. If you miss an opportunity with installation these days you could be scr@wed for months. We also have ASHP going in and need the extra output capacity.
Presumably we can still get the beta given the circumstances please? We are not on JA Solar panels. Thanks.
CCourtney I'm seeing this little jump at the end but not the repeating spikes (as yet). I don't have JA panels...
4073 first full charge (3% SOC to 100% SOC).
Charging throttled back at 85% SOC (55.2V) and spent a long time at 247W.
Got to 100% in the end having taken on board a total charge consistent with an 8.0kWh battery (as against 8.2kWh nominal).
Load spike (1784W) at 09:29 is genuine ( microwaving my porridge 🙂) but the 2327W load spike at 13:00 is an anomaly as nothing was turned on at this time. Could this be the reverse of the mysterious 01:00 charge spike?
Leon_shaw, no not as yet.
When watching the phone App in Home mode with 10 seconds updates, I appear to have a few minutes at 100% just trickle charging at 174 W, followed by a short period of battery discharge down to 98%. The cycle then repeats every couple of minutes, so is not accurately captured by the web portal graphs that take a snapshot once every 5 minutes.
The power value of the battery discharge appears to depend on the current household load, so at the moment with charging the car and 3.4 kW solar gen I am getting the odd discharge of around 1.5 kW which is then maxing out the 5kW rating of the inverter for a short while.
Note - Not JA panels, mine are REC Alpha Pure 410 W.
hhall
Another not on Beta, but totally agree with working as present. Gen 1 hybrid 3.6, twin 8.2 batteries can be about 10% out normally, but generally happy with setup. Still on inverter software DO.450.AO.499 - battery 3005 and no intention of upgrading till everything proved with new software.
Simon_C mines doing the same and has done so on 11 separate occasions. Basically every time the battery is fully charged since moving onto beta.
Calibration completed on 4073 beta
2 of 8.2 batteries
total ac charge 19.1 kwh
2.86 charge all the way to 90% then charge tailing off
94% soc 2.04 kw
97% soc = 1.38kw
98% soc =1.16kw
not seen any oscillations at the top end so far
Of course now I have the PV installed and JA panels the beta closes 😂
I can at least drop in with PV info as my E/W 6.3 split and those panels look to still be a discussion point.
Better day for me on 4073 and my 2x 5.2 kWh & AC3.
A few days I posted how the last 30% of my soc only discharged 600Wh. The battery discharge at the time was very low. Today, with the good weather I decided to test the capacity again. The battery discharged linearly (forced discharge) at 3kW all the way down to 4% with no sudden drops. The resting voltage at empty (after approx. 3 hours) was 51.55V (lower than 51.8 in the previous test)
Batteries then accepted 8.7kWh DC to full, a vast improvement over the 7.8kWh in previous test.
Open question for me - Is there a weakness that causes a fast SoC drop at low discharge rates as the battery enters the last 30%, or has the previous OCV now corrected the bottom limit? (difficult to answer).
Potential inference (if the former) would be that the map works well for high loads, but is not accurate for very low loads? As I have 2 batteries, this means my already low base load is split between the 2 batteries. Each could only be supplying 100W, does this confuse the OCV voltage map by pushing the voltage lower and cause a sudden drop in SoC? This mechanism would be more likely to hit those with multiple batteries (and low baseload).
Good test for Giv would be to run a single 5.2 battery and set a 150W (or lower discharge) from 30% soc to semi replicate. Easy to do in the lab.
CCourtney I’m seeing the same but only since 4073 installed. It has happened every day I’ve reached 100% from Solar since 4073 installed. I’ve looked back over history from several days with good sun from when 4074 was installed and did not see this oscillation after reaching 100%. 
TheDragon (GivEnergy) What email should I use to send this data in?
Im flat, so today its a good 5 hours+ of cheap, so will do a full uninterrupted charge. Capacity check
Simon_C I see this. China came back with the Upper OCV is filling the battery.
Full, check, not full, fill, check, and so on. It goes away with time as it learns
I had this and still occasionally do see the 174W trickle at the very top. Nothing to be alarmed at
TheDragon (GivEnergy), the trickle charge makes sense, but why are the sudden discharges occurring when there is no grid import to trigger them. The solar gen is significantly greater than the load.
It is not just me that have captured long periods with significant grid export, yet the battery has pulses of discharge when there is no need for them.
Charge finished. 12.4kWh
Exactly 80% DOD well 79.98
WIll now leave the charge signal on, till 1600, and use grid for the purpose of this test, to check for the bouning
TheDragon (GivEnergy) I get the logic on the trickle charge, check, etc but why are there repeating discharges as part of this process when solar generated is exceeding house load. Removing grid from image and zooming so it is easier to see repeating charge at 174w, discharge to grid at 300w once battery full and solar exporting.
Hello. I'm not a beta user but I' do appear to be experiencing the sudden drop to 0% issue as described in my reply to post at https://community.givenergy.cloud/d/713-firmware-3007-has-not-fixed-sudden-soc-drop-to-0-problem/2533
Is there any information I can supply which could help?
P3tro probably not, as this issue appears to be fixed in the beta firmware. Don't think anyone has seen it since.
There has been some issues with the beta firmware in terms of capacity and various other bugs. These are getting fixed bit by bit and hopefully there'll be a firmware roll out once that's done!
It sounds like getting that firmware onto everyone's systems also needs to be re thought as the beta process has been quite manual so far and the firmware is substantially different to the current firmware.
Additionally some older hybrid Gen1 devices may need a hardware change.
Dw Thanks.
This is weird, will have to now begin trending this, to see if its just A particular battery doing it, as my 5.2s aren't doing it
I see you and @Simon_C are both Gen 2. and both 1 x 9.5.
So thats a similarity already
Here is my issue entry, so I can add to it. The data from Dave has been submitted to the FW engineers already, they too are puzzled and cant replicate.
TX200 Correct. Testing the impact of Inverter only upgrade, & BMS only upgrade. So we can delineate the issues if they arise.
So far BMS new, Inverter Old looks ok.
Inverter new BMS 3009 looks ok, 3007 behaves odd
TheDragon (GivEnergy) But mine is a Gen 1 3.6 with a single 9.5kWh battery and the charge is 121W and the discharge is 171W. Are we all Solar JA panels?
CCourtney can you show as a picture, i missed your table post, till i scrolled way back up
TheDragon (GivEnergy)
BMS for this date was 4074.
Sorry couldn't show the column headers on this System data image.
I now have BMS 4073 and I have the same behavior. Do you want images for that?
CCourtney No, more importantly it was there in 4074, so 4073 wasn't to blame.
Its been passed to the coders to look at. I guess its some sort of upper OCV, but its allowing discharge when it really shouldn't. I'm more inclined this is inverter FW issue, on some common code between the 2 Hybrids.
Unless An AC coupled comes forward with similar. As i don't see it like this, I see 2 pulses maybe 3 then it stops. The OCV register in the inverter is set.
Here is a close up of lower OCV happening after a busy up and down the previous day with neither full nor flat. But lots of charge and discharge sessions. Even releasing 30 mins of power, previously on the old BMS would have gradually raised the actual real SOC. Over time developing into a jump
TheDragon (GivEnergy) given the solar generation is spikey as well at the same time, I'd agree with you that it's more likely to be inverter firmware related. Not seen on my AC3 on any BMS firmware
nor here on my ac3
CCourtney I’m still having the same issue every time battery reaches full charge. There isn’t an occasion it hasn’t done this since update. 11 times so.
Leon_shaw I’m still having the same issue
what equipment?
positor1 gen 2 3.6 Hybrid inverter + 9.5kwh batt
TheDragon (GivEnergy), I also had the charge/discharge oscillation with battery firmware 4074 along with 4073 as reported earlier in this thread.
https://community.givenergy.cloud/d/1342-new-givenergy-inverter-bms-firmware-beta-testing-chat/2168
No oscillation of the solar gen with my HY5 Gen2, unlike @CCourtney who has a 3.6 kW inverter along with @Leon_shaw .
This is the battery full calibration on 7/4/23 following the battery firmware change to 4074 on 3/4/23.
It is clear that the solar gen is solidly maintained along with a steady load, and it is the battery charge/discharge causing the fluctuations in the grid export with the battery at 100%.
Simon_C apologies I should have been clearer. I don’t have solar oscillation just the battery.
I have also noticed a difference in battery charge rate between grid and solar charging when a user limit is set.
With all charge power settings agreeing at 760/762 W including the App, I scheduled a short grid charge overnight, then recorded the resulting grid and solar maximum charge rates later the same day.
The grid charge was nominally 760 W, but the solar charge was around 980 W with no settings changed in the meantime.
With raw values shown
Leon_shaw, sorry my mistake.
But with the HY3.6 Gen2 inverter, you do have clipping of the solar gen when the battery is at 100% I think.
Simon_C yes it clips regardless at what level the battery is full or not. It seems to throttle Solar to 2800w on what should be best days for generation.
Leon_shaw I have had a discussion with the FW engineers, the lab results prove 13A is the limit. They cannot understand why you and 4 other see caps around 8A.
So all the data has been submitted to them, they will likely look at your data themselves, to see if they can identify this issue with Gen 2 and JA Solar panels.
I don't want this to delay the release, as the issues fixed are way more serious, but I also understand this apparent reduced performance isn't good either.
OK the data has been accepted, they can replicate it. Its an inverter FW issue.
They will fine-tune the battery current, what ever that means
Simon_C @TheDragon (GivEnergy) I've looked back over history and it turns out I also saw oscillation on 4074 as well but only on the first couple of days the battery reached 100% from solar after 4074 was loaded and the oscillation pattern was not as regular as with 4073.
With 4074 the oscillations stopped happening on 3rd April after around 4 hours. From then the battery remained at 100% but no oscillations - completely flat for the next 3.5 hours. After that, no oscillation at 100% for several days until 4073 loaded. Following 4073, I have had the oscillation every day when I've reached 100% for the whole time it remains at 100%... although none of those days have been as good as 3rd April for solar, so it hasn't sat at 100% for the same length of time as 4074 did on 3rd April.
mpartington Thanks, that would explain it.
Curiously it did the same again two days later, only this time stopped at 6%, waited 30min, adjusted to 2%, waited 4 hours and then adjusted to 7%.
It seems to get a bit confused when it gets to a low SOC.... I wonder if the voltage map isn't quite right at low SOC. Also still puzzled why it stops at 6% each time when the reserve is set to 4%.
TheDragon (GivEnergy) no it’s not good at all. It’s having a impact on the running of the system, time to charge battery, ability to export, payback time etc. It’s hitting directly in the pocket.
Another member @pelowj on this forum who has upgraded from a gen 1 to gen 2 inverter never experienced the issue with the gen 1 it only started when upgraded to gen 2.
For me this points to FW issue or something wrong with the inverter itself as it’s started after the upgrade. Let’s see what they find.
TheDragon (GivEnergy),
I wondered if I have avoided all or most of the string current issue by my choice of panels?
My REC Alpha Pure 410 produces 410 W with a Vmpp of 42.7 V and Impp of 9.61 A.
Whereas an example JA produces 410 W with a Vmpp of 31.45 V and Impp of 13.04 A.
My highest string current so far this year was 7.8 A, as shown on the power graph below, so close to the approx 8 A cap seen by others. Will have to wait for another good solar day to see if the current is capped.
Simon_C Whereas an example JA produces 410 W with a Vmpp of 31.45 V and Impp of 13.04 A
It does depend on the panel, mine are only 395 and obviously we're not normally up at the 1000w/m2 range, so the max still comes in under 13a even with perfect light.
I did see some peaks of 11a today but the weather has been all over the shop today so it's hard to judge anything. Tuesday looked to be a high output day when I looked earlier, so perhaps by then I can help share some other data, the scaffold might still be up though which is shading a little, on top of that it's E/W.
Here are the couple of points today it went over those numbers, not much use I imagine but it's not nothing. G2 5kw inverter, JA 54s31 395w, 8 on each string.
I'm interested in the notion that James didn't see the issue on G1, and that once on G2 it was improved with the beta FW.
dbt85 I am running Hyb5 Gen2, 2x9.5kW, JA 395s also (8+6 strings). I have seen peaks of 13a and sustained 12.8a on non-beta standard firmware. Here are a couple of recent days.
Simon_C You have nice bell curves on each string, not close to capping
4monks SO the JA 395s are fine, teh JA 410s seem not.
Very strange that a panel choice can have such different results.
Do you have the spec of your panels? To compare with the 410s
TheDragon (GivEnergy) Here you go. I am running on std firmware Hyb5 Gen2.
OK, something else has to be going on here. The plot thickens, but I am pushing for impedance testing on the MPPT to be done, hopefully china go get a PV simulator, which can simulate the various characteristics along the V/I curve
I see no reason why red capps, and green doesn't
I'm on JAM72S10-410/PR, which have a lmp of 9.82A, so it seems to me that the 13A limit couldn't be coming into play in my case.
It's worth reiterating that I was seeing a ceiling of around 2.7kW in the previous stable release I was on, but that this improved to 3.8kW in the beta.
It's also worth reiterating that I was using these panels previously with a Gen1 and there was no issue.
String current on the beta (perfect day):
String current on the previous stable version I was on (not a perfect day, but it wasn't bad. The ceiling around 7A is clear enough):
pelowj Just to give more context to my 2nd image:
Based on my rudimentary understanding, I would need to be seeing 500V to be pulling 4.92kW (12x410) from panels with a lmp of 9.82A?
So, I guess it's an issue with the mppt logic?
TheDragon (GivEnergy) So my Gen1 3.6 inverter is doing this oscillation at 100% SOC and bright sun as mentioned above but I am on JA Solar 365W panels. Here is the spec sheet that includes their electrical specs. Does this muddy the water even more or help narrow down the issue?
CCourtney Just to add to previous post I have two strings of 6 panels on each string
Do we want to spin this off into a JA thread where others might see it and be able to provide other reference points? I can get a post together with a specific format people can post their setups in if you like.
Tuesday looks to be a great day for generation so I'll be able to share my givtcp data once I have it. I have seen momentary peaks above 9A again today, but until I see sustained values over that I wouldn't want to say I do or don't suffer from the problem.
@pelowj do you still have the inverter number for your old G1? While you can't see the data from it Paul probably can.
dbt85 Good idea,
dbt85 Yep, sure do. It was S2229G096.
Good idea to spin it out into a new thread.
JA Panel discussion
https://community.givenergy.cloud/d/2186-the-ja-panels-discussion
@TheDragon (GivEnergy) Noticed this topic has dropped to the bottom of the front page of the forum in light of the new clipping issue on hybrids. Is the firmware close enough for general release to cure the SOC issue? Close enough to please the masses while other issues are being worked on?
Got to the bottom of why battery discharge was sometimes halting at 25% SOC.
It did turn out to be the "Battery Cutoff % Limit" setting that was causing it, but boy was it subtle!
I changed this setting to 25% quite a while ago when forcing the charge spike at 01:00 anomaly (the two are linked but still not exactly sure how). Many "reset to defaults" had been run since. The first of these showed a change from 25% to 4% and all subsequent ones showed no change (4% before and 4% after).
Battery discharge mysteriously stopped at 25% again today. I went into Remote and checked the "Battery Cutoff % Limit" setting (which was showing 4%). Took it one step further this time and performed a read on the setting. Hey presto it showed 25% and updated the displayed value!!!
I then manually changed the setting to 4% and battery discharge recommenced.
.
.
Lessons learnt ...
Reset to defaults does not reset the "Battery Cutoff % Limit" (even though it says it has and updates the displayed value).
When battery discharging in Eco mode the "Battery Cutoff % Limit" is ignored.
When battery discharging in Timed Export mode the hidden (actual?) "Battery Cutoff % Limit" is observed.
Looks like a partial/incomplete removal of the code associated with "Battery Cutoff % Limit has taken place. If this setting is redundant in the beta software then a careful search for all references in existing code needs to be carried out.
This raises the question of what to do at release with systems that have "Battery Cutoff % Limit" set to something other than 4%? If all references have been removed then it should not cause a problem. There does, however, need to be a release note advising users that "Battery Cutoff % Limit" is no longer a thing and has been removed. This is in case anyone is currently using it to control their system. (There have been people using it for system control. This is how the charge spike at 01:00 anomaly came to light.)
Hybrid gen2 995/995 address solar capped PV power on one string
BMS 4072 (3012 Pre Release) address low current OCV
Both going out to alpha test 1st
TheDragon (GivEnergy) Alpha testing? Is there a ideal length of time for testing? Are we potentially looking at around 10 weeks for a general release?
THALL No, I am not pushing out 4072 to loads of inverters manually if the designed isue is not fixed. So 3 inverter will have it for a day. then tomorrow I will put it wider.
As this is pre Release, I would read into its not 10 weeks away
TheDragon (GivEnergy) I am not suggesting pushing pre release to the masses, but other betas like I am on do appear to fix the original SOC issue and while the beta has found new issues it feel likes general release is very far away if the current goal is to fix all major issues that have been found with lengthy testing. I do value all the testing, but there are a lot of people waiting for the SOC issue to be fixed. Fixing the SOC for the masses would be good. Missing capacity is not as bad as the soc issue.
Plus lets be honest, when companies release a fix for an issue it does usually uncover other issues or make new ones. I do not feel the capacity loss issue is a deal breaker vs the SOC. The soc means missing out on 40% or even 100% of the battery vs the capacity issue losing 10%? 20?
THALL I agree with this and I think many others with this issue would who have the SoC bug. I lose 10%-20% capacity after a couple of days anyway but I don't know exactly where the SoC is so I end up like today running on peak rate energy. Missing capacity and other issues could be resolved in a few months time or push out the fix to only those affected by the SoC bug. At least it's some progress for the masses.
@TheDragon (GivEnergy) Would you be able to give us (none-beta users with the SoC bug) a time frame when we might see the fix?
Thanks
cbowen After the issues with firmware and the SOC bug I totally understand that GE want the next general release to be right. But given new issues springing up the goal posts could keep moving. I am thinking a general release of the beta that fixes the SOC with a document explaining possible new issues with that firmware. The SOC issue appears to be fixed with the Beta's. Current issues are missing capacity and a PV issue with hybrid inverters which are being worked on. Just food for thought really.
I think giving people the choice would be a good idea.
The upgrade page on the web portal would be better if
Woodstock I had exactly the same issue as you in January this year. My issue was that it would go no lower than 10% which I had put in as a figure some weeks before but subsequently thought I had changed it back to 4%
I too did multiple inverter resets and eventually stumbled upon the solution just as you did. Eyes popped out with amazement when the 10% figure finally popped out when I did the read on the setting! that was displaying 4% prior to the read!
CCourtney Just a random thought - could this be some kind of caching issue, perhaps just affecting some users/browsers?
If you can reproduce the 'incorrectly rendered' value, it might be interesting to try different browsers, to see if you get consistent results - or not.
Lets see how the Pre-Release goes for the SOC rapid drop for low users, which a few days ago was your (Royal you all hot potato)
@mpartington let me know how it goes.
2 others that have the rapid SOC decline at low power also updated
TheDragon (GivEnergy) will do.. for others, this is the issue. Steep drop in SoC at low discharge currents. Here my solar was almost covering base load, but 2x batteries contributing circa 100W -200W (50-100W ea).
Issue is understood by Giv, just need to test fix
THALL Current issues are missing capacity and a PV issue with hybrid inverters which are being worked on.
Wasn't there also the change in behaviour of the target SOC when doing a timed charge from AC ? (Pausing the battery when SOC is above the target value.) Or did that get fixed. (I've not been following too closely.)
DD, yes I have that issue with %SoC > Target SoC on a HY5 Gen2 = D0.996-A0.908, battery=4073.
It is not helpful when trying to plan car charging on alternate days, then having to enable/disable the AC charge to suit, and after forgetting to set correctly the GE battery charge is emptied in to the car.
Simon_C that is also fixed in 995.
Can leave AC charge On and charge to on at 50%
TheDragon (GivEnergy) Good result on that!
TheDragon (GivEnergy) Will this allow the battery to discharge down when the actual SOC is higher than the target SOC during a timed charge?
mpartington was just going to add is the 50% an example or a limit ?
mpartington example
TheDragon (GivEnergy), is 995 just referring to the inverter firmware?
So if I am currently at HY5 Gen2 = D0.996-A0.908, battery=4073, will changing to this '995' also involve another battery calibration?
Also is there any update on the battery charge rate being different for solar charge compared with grid charge as reported previously, or is it deemed a low priority issue?
https://community.givenergy.cloud/d/1342-new-givenergy-inverter-bms-firmware-beta-testing-chat/2225
Simon_C D0.995-A0.995 is Inverter only.
If you experience solar cap, its a priority for you. If not, its just the Set a charge to % and solar can carry on fix
TheDragon (GivEnergy) Is the oscillation bug fixed in the next version of the Inverter software?
That is , when the Timed Charge is set to less than 100% we will no longer get the issue below:
Woodlands Yes it will be. You have an AC coupled, which has been delayed due to other FW priorities. Hybrid Gen PV and AIO. But its being worked still.
One pair of hands they say
Quite suspicious of the 7% jump in SOC today in just 10 minutes.
6 days after 4073 loaded`
positor1
AS no opportunity for OCV to run at idle, it ran with low load as a last resort, this need a correction made, to compensate for the lower voltage while drawing a load, vs idle. This correction needs a tweak.
@mpartington was assisting this with data
So the SOC went down too far, so when charged you got close to the ceiling and a small jump seen, i look to be the same as the SOC drop. As shown in the 2 red boxs
BMS 4072 is being tested by a few affected, to see if the tweak is better. I can put on yours if you wish, no calibration needed
TheDragon (GivEnergy)
You certainly know this kit inside out!
This is the first time i have noticed it , I am happy to stay or try 4072, whatever would assist you and the community best 😃 go ahead and load if that is your pleasure.
It Is interesting that you point out the soc drop too, i had not noticed but this morning i was surprised to see 35% when i expected 45% ish. well done Dragon.
positor1 Done.
TheDragon (GivEnergy) thank you
Gen2 PV issue looks resolved,
But need a few more positive responses to be sure.
D0.995-A0.995
Fixes
Waiting fix, ability to discharge during a charging window if SOC > Set %.
This one took a couple of hours of me decoding logic to get the point across.
FIngers crossed this is the last one to nail
TheDragon (GivEnergy) Hi Paul, wondering if this issue would explain an oddity I saw today - a rapid jump from 95 to 100% without much input power and then once at 100% the battery started discharging to grid at a low but variable rate (80-170W) for 2 hours whilst SOC remained 100%.
Dw those little jumps and odd discharge are minor but seem fixed in 4072 (3012 pre release).
Loaded yours just now.
You are now on D0.995-A0.995
BMS 4072.
Both are final pre release versions
TheDragon (GivEnergy) Thanks Paul. It maybe a complete coincidence, but some good news. Last night with 4072 was the first time since 4073 was loaded where the battery stopped discharging at low SOC and then stayed at same SOC level overnight rather than adjusting down and up again. It still stopped at 6% rather than 4% which is a consistent oddity, but a minor niggle compared to all the issues that you've got fixed. Many thanks for your (and the team's) efforts on this.
Is this the same issue people with JA panels have been seeing? The oscillation?
Dw The battery on and off small amount was 995/995 fix.
I did both
You have what is to be the final release if I can some good PV results today from Leon and the other who have Kindy assist in tracking this down. NOT JA SOLAR, or TIGO. But relating to MPPT tracker settings
TheDragon (GivEnergy) Not sure if I've seen the same issue. Battery gets to 100% and then discharges for 20 minutes or so then fills again...rinse and repeat. If you think it is, happy to test 995 on CE2105G039.
Hi Paul
I've been a bit quiet for a while, I guess life got in the way. I had intended to provide some data on charge in, discharge out in early April. However daylight saving time, messed up the data, followed by multiple virgin media outages and my wife turning off the router twice.
Most days on 4074, we had not managed to use the full 9.5kwh on the AC3, on days where we drained it early the system managed to find extra capacity at the bottom, and resume discharge between 1 - 3 hours later. If the battery ran out early evening, it typically opportunity to do this, and usually this gave us around 9kwh of capacity for day.
On some days, we ran out of battery between around 23:00, thus I suspect the hunting at the bottom never got chance run and find anything more, before the next charge cycle at 00:30. Daily discharge was typically around 8 - 8.5kwh.
Since 4073, I'd say discharge capacity was down somewhat. We got a full 9.5kwh out following the calibration, but further attempts to charge / discharge yielded around 7 - 8kwh before flat. Even if it dropped to 4% early in the evening, we didn't observe any further lower adjustments / bonus discharges on 4073.
Throughout each evening, after the battery is 'full' during off-peak - we see behavior where the battery appears to be discharging to meet house load, when ideally it should be sat ideal. It will then charge itself back up to 100%, some of this appears to be the 'hunting'. Largely it seems to be confused on whether its meant to be charging / discharging. This looks similar to the oscillation AC3 bug that has been discussed. See example below from the 23/04/23
Now the interesting thing that's happened recently, is we've moved to Octopus Intelligent and adopted GivTCP. The idea was to charge the battery, whilst the car is charging / off peak. This involves an automation which Forces a charge, then cancels when we move back to peak rates.
When this is happening, the behavior of charge / oscillation is very different. The force charge got it to 100% and held it for 6 hours, without any noise / jumping around. See example that shows 26/04/23 (this was actually the evening of 25/04/23)
The surprising take away, is following this charge on 25/04/23, the next day 26/04/23 we achieved a record 9.4kwh out of the battery! The is a first time on 4073, and a first to see on any firmware without a calibration.
To rule out any variation in the approach, last night I let the battery charge as normal without 3rd party intervention, and saw the same behaviour. It held nicely at 100% with no jumping around. See the shot from the 27th
It appears something has changed for the better, perhaps the forced charge has activated some kind of behaviour that just makes charging work as intended?
Appreciate this is all a bit inconclusive, but I'll continue to monitor over the coming days.
Cheers
James
Jamesrich
Ok, 4072 is the newer pre release. no calibration needed. SO I will put it on early for you
judgepd 995 FW is for GEN2.
Still waiting on 288 for the AC coupled. Gave them till the end of the month to provide me the FW
TheDragon (GivEnergy) Ah, sorry, no worries.
Thanks Dragon, I'll see if this changes our experience.
Since the forced GivTCP charge, so far it would seem we are in a better place already. The question is why this triggered a change of fortune, or was it just a coincidence?
I didn't expect things to calm down until the expected AC3 inverter update. Will keep you posted.
OK, we have another minor tweak. We have put in a modified base OCV map for better relationship between SOC & voltage, from previous iterations of tweaks. I looked at the actual code, and there are some differences, but only under 20dC. Down at 5dC the differences are quite different. Likewise in these later FWs there are also calculations taking into account cycle effect on the cells.
So 4071 has been loaded to everyone with 4072 instead. No action needed.
Work is now underway to implement the BMS FW (3012) to the portal, which can be loaded independently of the Inverter FW, from recent tests.
I've seen no serious issues with new BMS on older Inverter FW, The OCV still happens, but has no way to remember if its been done already. Not the end of the world. @chris_p how are you finding this setup?
I will be recommending BMS can go out separate to inverter. But you won't benefit from all the new features in the inverter FW.
so 4071 becomes a general release called BMS FW (3012)? and will the moaning brigade get that soonish?
positor1 Im taking leave. LOL
TheDragon (GivEnergy) The OCV still happens, but has no way to remember if its been done already. Not the end of the world. @chris_p how are you finding this setup?
Nearly three days in and can't say I have seen any SoC adjustment occur looking at the portal graphs. The amount going out vs SoC drop seems to be low, more kW out per % than I would expect, e.g. yesterday I worked out it was 0.073kW per % giving 7.3kW capacity (from 8.2kW battery)
My SoC issue was always lost capacity at the top, I would have expected the OCV when the battery was idle at 100% to drop the SoC and allow it to start to find more capacity and fill up again? Maybe it hasn't been used enough since calibration yet?
TheDragon (GivEnergy) so when will that happen then?
I would agree that the latest iteration is looking very positive. I've seen no fast drops in SoC at low power, despite deliberately putting the system in the conditions I know trigger it.
Still some work to do on multipack soc drift (when not hitting bottom limits) but this appears cosmetic and can be worked on as a further improvement to not hold up release.
Just need to be clear in a release statement that the drift will still happen (but importantly will correct at bottom limit), and is being worked for future update.
4071 now showing here.
chris_p
Today is a not good sun day, if the battery goes to Idle, it should do its corrections there. Looked back over your data and since calibration its not managed to get that low. The voltage vs SOC is tracking ok though
TheDragon (GivEnergy) Good to hear, thanks for checking. I can put the washer on to drive it even lower if you think that would be a beneficial test?
chris_p Fill ya boots. Would be nice to see how an unmanaged OCV firmware reacts. But don't waste energy for the sake of it. That's my job
TheDragon (GivEnergy) It is on. I assume you can see individual cell voltages etc and so can assess remotely the general health of a battery? In March last year GE identified an issue by an engineer site visit, and subsequently replaced the battery.
TheDragon (GivEnergy) Work is now underway to implement the BMS FW (3012) to the portal, which can be loaded independently of the Inverter FW, from recent tests.
Paul do I read this correct, will this be availble to all users not just thsoe on the beat testing?
Stevef1964 You read it correct. When its published, it will be for all
TheDragon (GivEnergy) how much benefit will the BMS firmware upgrade give without the associated inverter FW? (I know some of this will be covered in detail in the release docs), but I’m just wondering in a general way
@
Looking back over data of inverters that have BMS only. I cant find any jumps in the month worth of data.
Installed 23rd March on a triple AC3 from that point, not a single jump or crash (That was BMS 4074)
So I would say even without Inverter matching, there is still benefits to running this separately.
I would just highlight that with the magnitude of changes I don’t think any new firmware should start with a 3.
6 is a lucky number 👍
Far easier for support and engineering plus installers to go do you have 3 series firmware, ok update.
kram OK. Point taken.
All STOP. Can't update on a 3.
Now we have to wait for 7 million years to get to a 6.
Lol.
Nargh
Silly question time. I have Gen. 1 inverter and 8.2kWh 100% dod battery.
I see no SOC or capacity issues at the moment. Would a firmware update be recommended once it’s released?
Oh and I should add at this point - Paul, I certainly don't expect or need a reply over the weekend or Monday - sometime is fine.
You need a break from all this crap for a while!
I would also like to put on record how much your input is valued by me and I'm sure many on here and the wider install base - even those that aren't members on here or facebook.
anglefire Long time lurker here with the exact same setup and questions as you. Tracked this thread for a long time and really appreciate everyone's hard work. Most of what gets said here baffles me but it's great to see a community coming together to resolve issues. Thanks everyone!
anglefire In your situation, I would fully adopt the old adage:
'If it ain't broken, don't fix it'
I was in a sililar camp with my Ac3 and 9.5, that was installed end on nov 2022. All good until january.
Now im lucky to get 7 units from the battery and see lots of % drops to zero.
Id suggest if you dont have it yet, you will soon.
ToothyChris 'If it ain't broken, don't fix it'
Best advice right here... I don't have the issues (yet), so gonna watch and wait.
Whereas I want to be involved in the beta (and have been), but will have to watch from the back for now. My inverter isn’t happy. Found the battery was powered off, and on restarting it I got the clicks of death. In the queue for a site visit now. Very unlikely to be related to the beta (it isn’t a hybrid), just bad luck I guess.
TX200
how old is your kit?
TX200 Any idea of cause? Environment, mains high/low, or just "the kit" ? I ask because I like "prevention is better that cure" and would happily install something to give added protection. We've all come to really appreciate solar panels and battery and with the, still, high energy costs it's a must to protect such an important asset.
positor1 installed just over a year ago.
BrianHere I'm wondering if it was high grid voltage. Seemed a little high around the time. Although still within the acceptable range.
I'm not sure if grid voltage is the actual voltage of the grid (that would make sense) or if it's the voltage at my connection to the grid (as obviously after the fault, I was exporting, and I understand the inverter has to raise the voltage to export).
I've read elsewhere components can fail due to heat. So perhaps a bad connection, creating heat, causing failure.
It's not in a loft, so doesn't get too warm this time of year. So not that kind of heat. The inverter was 41.3c at the peak of the overnight charge and 32.2c just before the system failed.
Guess we'll find out when they swap the board. Maybe.
I hear that newer installs have surge protection fitted on the AC input before the inverter. Not sure if that was recommended or advisable at the time mine was fitted, perhaps not.
TX200 So sad.
You were giving good feedback too.
TX200 not sure if grid voltage is the actual voltage of the grid
I wonder this too. I assume that the voltage reported in Grid Data on the inverter card is from the EM115 whereas the Inverter Data card voltage comes from the inverter itself. If so, Vinv > Vgrid during export and vice versa during off-peak charge from grid?
Tim yeah you might be right.
Current grid voltage for me is 246.4 and inverter voltage is 247.7 (it's not doing anything other than monitoring though. The export is from a separate solar inverter).
And yesterday, just before it all went wrong, grid card was 246.8 for two data points. Inverter card showed 248.4 then 248.5
Tim I think your assumption is correct. I seem to remember seeing that kind of voltage difference, during charging and discharging, although I can't imagine the voltage drop will be much, in most installations.
So what do the people who purchased 100% DoD batteries that deliver around 85% DoD get?
A further fix, a heavily discounted second battery and EMS? Free extra warranty on inverter?
I’m not having a go, but the data sheets need to be clear and something should be done for the users that have put up with this and then don’t get the capacity they paid for.
Also in the "doesn't seem to be broke, do I really need a fix?" camp - AC3 and 2 x 5.2 batteries that seem to behave reasonably well on old firmware.
System just carries on saving me money as it was designed to do..........................just wish there was more sunshine this Spring!
wilfleek January was a great month for us- but after that a bit meeh. April looks like it will be about 93% of last year (only had the system since dec 2021)
Last year we beat the calculated /predicted production by a decent margin. Unless it changes it’s going to be touch and go.
Having said that it’s still saving us a good amount so not really complaining 🤣
@TheDragon (GivEnergy) Hi Paul, not a big issue in the scheme of things, but just to flag that I'm still seeing the small charge / discharge oscillation once battery reaches 100% SOC on solar. It is not a regular as with the previous beta but is still happening. I remove house load from the image so you can see, but the load is consistently lower than solar gen throughout the time shown 
with a single adjustment seen there, surely that was an idle OCV (open circuit voltage) adjustment as you would expect?
I had six of these yesterday at idle but looking at the voltage at those times it is clear that was what was happening to mine. kept finding a little more charging room. from 53.29 to 54.18
bms 4071
positor1 Mine works the other way around and finds more discharge room at the bottom end.
Can be a bit disconcerting to see the battery discharging and the voltage dropping at the same time as the SOC% is increasing!
This time of the year I only get anywhere near the lower limit if I force a discharge and I have had a 5% upward SOC adjustment each time I have tried this. Just hoping the SOC% error is not cumulative as I could possibly be in line for a massive upward adjustment if I just leave things to their own devices through the summer.
Woodstock Can be a bit disconcerting to see the battery discharging and the voltage dropping at the same time as the SOC% is increasing!
indeed.
positor1 Agree re the little blip in SOC. It is the oscillating between charging from solar and exporting battery to grid over extended period of time once SOC reaches 100% that I was commenting on. Apart from maybe finding a bit more capacity to charge, I would expect battery to be idle once at 100% whilst solar exceeds load and is exporting energy. Instead it looks like it thinks it can store a bit more charge then realises it can’t so dumps it to the grid, thinks it can store more charge, dumps back to the grid, repeat until load exceeds solar. Not a big issue, just odd.
Can someone with a 8.2kW 100% DoD battery post a portal chart showing battery SoC %, voltage, battery power, house load and PV? I've been on the beta for battery only (no firmware update) for about a week and my voltage hardly changes. Below is the lowest my SoC has been to then full, at 19% it is 51.7v and at 100% (idle) 52.8v to 53.0v. Not quite the range I was expecting when looking at that chart that keeps floating around with the rough % = voltage table. Plus I've yet to see any noticeable OCV corrections going on:
chris_p I can post that with an 8.2 and 9.5 combo? Not a like for like though. But will post it if you want it.
THALL Thanks but not sure it would be of use unless you separate them out? Not sure you can without GivTCP?
chris_p Not at home to fire up my givtcp. But no I do not see a way to separate on the portal.
chris_p 2x5.2kWh batteries (beta battery and inverter firmware). Battery voltage range is very similar to yours.
This is my master battery from resting empty to resting full. Looks like it did a top point calibration, as usually the voltage peaks at around 54.5V
mpartington Thanks for that, very helpful. Seems like it is doing what it should be. Really want a poor day of solar again and try to push the battery right down to empty to see if the "crash" is avoided.
chris_p I did that to mine on Saturday as it was close to empty. I forced a discharge and charged back up using sunny skies 🙂
The only thing I'd like to do is do this when high drain items aren't being used so I can see the amount of energy the battery takes from 'empty' to 'full'
Ra Yeah that is another test I want to do as trying to work out from say 61% to 100% and dividing the kW in by % diff is it really a valid test?
chris_p I think I have the same parameters in my chart for yesterday. This is a Batt8.2 (100% DoD) at 3007 showing swings between <52 to >54V
I haven't seen 9.xkwh out of my battery since 4072 / 4071.
Seem to be achieving 8.5kwh before empty. Do I need to flatten my battery hours before the scheduled evening charge to see a lower OCV happen? Previously they typically occurred when I'd been sat on 4% between 1 - 3 hours.
Also 'free energy' has become a thing again recently. Where the amount of discharge is around 8% less than the 'house load'. Could this account for why my daily discharge appears to fall short of 9.5kwh? Could i be getting 9.5kwh out, but its misreported.
Resting 51.5 to 52.5 would be fairly normal.
Lowest under load shouldn't be going below 49v, or above on charge beyond 56v. There isn't much capacity in the regions outside the 15-95% area. This is also where the worst of the wear happens.
I occasionally see the sharp rise at 56v, this soon reduces charging to 0.1C and stops at 57v. preventing excessive cell voltages
Jamesrich I haven't seen 8.5kwh out of my 9.6 battery ... well never - last night I had biggest drop so far of 56%. If I fully discharge and charge my battery and if I am very lucky I might be able to get low 8kwh'ish provided I discharge it on the same day
Guess i must be lucky - I've seen 9.5kwh out of my battery. But mostly only after a calibration where it puts around 10.5kwh in. So I know its capable of it!
Most days I charge 9.5, and attain 8.5 kwh out.
Jamesrich yeah not good considering it is supposed to be 100%dod
Well if we can't achieve 9.5kwh through software, I would accept one of these as a gesture of good will.. 😆
:
Jumping on this, I'm in the motion of getting quotes for a givenergy system, is there a gen 3 inverter on its way? Can't seem to find any info on this but been sent a data sheet on it?
Gards I am not aware of a Gen 3 coming. AC3's are still Gen1. I guess their All In One might get referred too as Gen 3? MIGHT be due end of May. MIGHT
THALL this is what I thought, but was sent this?
Regarding the reduced discharge capacity people are seeing, based on the tests shared earlier in this thread, I was working on the basis that the capacity was based on the charge cycle and the discharge would always be less due to conversion losses. Does seem quite a large loss though. This was the 5.2 results
Gards Ha yeah, gen 3 hybrids. Difference is export to grid? Larger, heavier. I just thought the curved front units were like gen2 v2 but apparently not.
https://midsummerwholesale.co.uk/buy/giv-energy/givenergy-gen3-3.6
THALL 15A input for PV
94jdh Higher 5kw backup power when using PV and battery on the gen 3 5kw. I wonder why there isn't a new AC unit yet? If there will even be one. I did expect a 3600w ac3 to be honest.
mpartington Yes, and less than the 80% DOD accounts for on the 5.2kwh..
I understand my 9.5 is 100%. I also know its capable of 9.5kwh discharge, it will easily achieve that after a calibration where it puts 10.5kwh in.
Its also done close to 9.5kwh from time to time, usually its had the opportunity to 'find' extra capacity at the bottom. Like 9.4 / 9.3kwh. I'd be happy with that! It also managed it once after a manually (GivTCP triggered) forced charge last week, but not since.
Right now, I'm curious on whether I am actually getting 9.5kwh daily - but the discharge is under reported. Obviously there is no such thing as 'free energy' but my stats say otherwise!
my 9.6kwh battery drained so much yesterday that it took today 10.948kwh to charge it to 100% (luckily the sky is blue) and when it crashed cell voltages got as low as 2.495V. I am just wondering what is considered safe levels (for cell voltages) and what this SoC bug doing to the longevity of my battery?
alfwro 2.495v is a little low, as long as its not left like that, it will be fine.
Likewise 3.5-3.6v per cell isnt damaging, but dont leave it up there.
The BMS will self preserve
TheDragon (GivEnergy)
Hi Paul,
Any update on the ETA for 288 on the AC coupled? Last night I topped up by 0.954kWh, the system then spent 3.5 hours holding this SoC value using twice that amount of energy. I'll take a beta 288 if it'll help :-)
May the 4th be with you. A selection of beta testers (Already on 4071) will get this 1st, manually. As its just a version No change
This is now in the process of release validation. In the hands of the Web Devs now. Ive done all I can.
TheDragon (GivEnergy) New data screen too? Or is that screenshot something only you see as a GE employee?
THALL Looks like home assistant to me
@All.
Unless there are any issues, likely no, you've had 4 days to bring up any show stoppers, Version 3012 should be made available at the end of today. The 4th of May
May I thank everyone who has assisted in testing and given regular feedback. I am sure there will be the odd niggle going forward, despite a beta set of over 150 inverters, of all ages, types and configurations.
Inverter versions to make use of the new BMS FW, will follow
TheDragon (GivEnergy) Just to confirm, based on this the Upgrade button will appear so people have the choice to apply it?
TheDragon (GivEnergy) Will 3012 be available to all users, or just Beta Testers at this stage?
THALL Vi button, it will NOT be forced on you
TheDragon (GivEnergy) Mine seems absolutely fine but haven't had the chance to double double check battery capacity. It seems better than what it was but would like to confirm it is as specified. Need some more sun first 🙂
Many thanks for all of your efforts 😃
stevelewis I don't see that anymore with my AC3, not since the 24th of April. No clear idea what changed.
The only difference was I did a forced charge via GivTCP on the 25th, then the issue has not been observed since.
My daily discharge is down somewhat, possibly because part of my daily discharge was happening when the battery should have been idle!
I went from this to this...
TheDragon (GivEnergy) Huge milestone, well done Paul. Appreciate all your efforts on this.
I have two remaining issues : -
Capacity. May be 'cosmetic' due to miscalculation of discharge due to 'free energy'. Where battery discharge is reported lower than house usage, with nothing from grid to account for the delta.
Immediate drop from 100% to 98%, when exiting the charge window. Never see 99%.
Are these two issues likely to solved through battery or inverter upgrade?
Many thanks
James
Jamesrich Immediate drop from 100% to 98%, when exiting the charge window. Never see 99%.
I have something similar but certainly wouldn't describe it as a problem. I never, ever see 98% on the way up or the way down. Always goes from 97% to 99% or vice versa. I wonder if it's some rounding going on somewhere.
Jamesrich BMS update to 4072 26 Apr, 4071 27th Apr.
Is there release notes somewhere?
pacemaker Yes, going on the KB.
#
Thanks a lot. Other than the "Capacity fix for 8.2 and 9.5 batteries" the rest doesn't mean much to a non techie. Is there any layman's terms for what other "fixes or benefits" are coming?
TheDragon (GivEnergy) Apologies to @TheDragon (GivEnergy) but I couldn't resist
TheDragon (GivEnergy) Is battery calibration required when applying the firmware (for those coming from 3009 FW)?
Probably worth adding not to expect perfect soc allignment for multi packs just yet. Hopefully this can be improved in future releases. I'm seeing a 10% split, however they charge/discharge the same amount (just between different limits). I.e. one goes between 20 to 100% and the other 10 to 90%. It is cosmetic and does not affect capacity, hence why not holding up release. Voltages are perfectly matched and most importantly NO soc jumps/drops.
You can see the firmware performing re-allignment of the slave soc (at empty) at around 8 am on 3rd May, but unfortunately the split reappeared at full charge. Oddly they switch over after the re-allignment, so now my master reads higher.
judgepd You actually made me chuckle.
mpartington The remaining issue is in the formal announcement, which is going through approval channels now
TheDragon (GivEnergy) Yes, they happened 24 hours after the overnight wobbles resolved themselves. You applied the updates in response to me reporting good news. 😃
TheDragon (GivEnergy) Thanks for all your hard work and perseverance Paul.
Just checked mine it’s on 3012
Currently running Gen2 inverter firmware version D0.906-A0.906 and (9.5kWh) battery firmware 3009 and having some odd behaviour with approx 400watt limit on battery draw when the battery is over 80%. Hopefully this update will sort it out, though it was only installed this week, so it's hardly been an issue (we tend to draw less than 400 watt most of the time anyway).
Thanks to all who have helped, I'd like to find out how to become a Beta tester for the next firmware if anyone can let me know the procedure.
Spencers What is your Inverter No. Ive got some spare time
TheDragon (GivEnergy) You kindly recommended that our system be trialed on 4072 after our installer visit and contact with Jacob at the GE support team. Since that was applied......its been like having our old system back, no more high voltage alarms and no noticeable SOC jumps etc. What a HUGE RELIEF and quality of life improvement!
Just wondering now - do we need to be updated manually by yourself to the more recent BMS version, or just sit it out now and await the Gen1 invertor FW to go along with it? Thanks
Is this the number you need?
ED2253G499
TheDragon (GivEnergy) Many thanks for the important update Paul. Much appreciated.
Two (three) quick questions, if I may:
1) Does this new software employ any degree of learning? By which I mean, that some of the maps/parameters are adjusted to align to each and every battery pack (and continue to learn, as the packs age) - with the aim of better tracking the SOC, and reducing the frequency and size of SOC corrections? If the answer is yes, how long do you expect it to take, for the system to do most of its learning?
2) What is the largest SOC 'correction' that users should expect to see, in normal use (after that initial learning period)?
Thanks.
TheDragon (GivEnergy) Many thanks for all the work you and your team have put into getting this BMS bug fixed. And also thanks to all the other beta testers, that have helped.
Since the last couple of beta versions we have not noticed any SOC drops or jumps (more than a couple of %) so was quite happy to load up the release version 3012 onto both our batteries.
One thing I have noticed, with the increased monitoring I have been doing, is that the Primary Battery appears to be cycling more than the Secondary Battery (Prim: 350 cycles since installation, compared Sec: 299).
TheDragon (GivEnergy) Is it possible to switch over the Primary & Secondary just by swapping over the settings on the dip switches, or does the wiring need to be changed?
Woodlands Wiring needs changing too. As the Master connects to the inverter via 2 pin, the master to slave is a 4 pin.
Then move the DC cables over. Master does normally cycle more, on the previous FW versions. Easy 30 min job, for your installer if you wish
Woodlands to perhaps answer on behalf of Paul who probably has a s..t load of other stuff to do, changing the dip switches should be all you need to do. The electric has no way of knowing where it is on the network.
Unless it’s clever enough to be able to detect the impedance variations of a couple of feet of wire!
Spencers Is this the number you need?
ED2253G499
OK, FW loading now. Do you want it calibrating, or let the BMS do it itself over a number of days
anglefire Ignore me as Paul has answered and be the comms wiring is different. Which for modbus is highly unusual. Usually either 2 wire or 4 wire. Which are different handshaking protocols.
anglefire just looked at the manual and it’s not different comms - just the master is a different port on the master battery to the slaves. So still two wire comms 😀
Thanks. Let the BMS do it itself.
There is someone else from GE support that replied to an email I sent earlier so I hope I'm not causing a problem with two working on it? His initials are MB if it helps.
Now I'm confused! Was on 4071 and now 3012 has just arrived with no blue button or recalibration run. Are we not testing the rollout procedure?
Woodstock there is no recalibration required 4071 is the same version just renamed to 3012
Woodstock
there is no recalibration required IF you are being upgraded from a recent beta where you have had a recent calibration.
Non beta users will need to re calibrate.
Is the rollout happening now for non beta users and is the BMS firmware what will resolve the SOC issue? Managed to have a drop last night from 39% and I'm currently running on the grid again so quite keen to get this firmware on my system ASAP.
positor1 I appreciate that but it seems odd that the rollout procedure is being bypassed for beta testers and the first people to try it out will be the general user base.
Woodstock
Did you not get an email asking if you wanted it done for you or wait for the upgrade button (yesterday) assuming you are a beta tester?
Woodstock some might say it's not a true test of the rollout if you take someone on the beta to new prod.
Needs to be old prod to new prod test now. 😀
Leon_shaw Not the need for a recalibration or not that is confusing me; rather the fact that we are not receiving the new release in the same manner as the general user base will. Seems like a missed opportunity to test the rollout procedure.
Woodstock
you can do it as normal but you would then need another calibration as that will be an intrinsic part of the procedure, beta users had a recent calibration so that would be a waste.
TX200 Agreed, that is the ultimate test of the new firmware once in place. I just wonder about the mechanics of getting it there. 😀
positor1 No and yes I am.
Woodstock
If you want it done for you without a calibration email Mr landregan or wait for the upgrade button to appear.
positor1 I cannot "do it as normal" (no blue button) and not carrying out an intrinsic part of the procedure is precisely my point. I am not testing the rollout procedure. Maybe others have already done this?
Woodstock you could always ask to be downgraded to old firmware first. 😁
Thanks to @TheDragon (Givenergy) for doing my remote firmware update yesterday. 🙂
I chose not to have it calibrated and let the BMS do it over the next few days. The battery went from about 60% just before the update down to 3% after (in about 5 minutes), but I guess it can't have actually discharged that much, that quickly, so is part of the update process?
The battery charge figure increase during the evening (not actually charging as it showed as 'idle') and got to 21%, now showing 23% this morning, but was still 'idle' despite some excess solar, which was just going to grid. I set a timed force charge a few minutes ago which did nothing and the battery stayed showing 'idle' but when I re-saved the settings for charging from solar only, then it started charging the battery instead of exporting 600-700 watts excess to grid.
Tl;dr If yours isn't charging the battery from solar, try turning solar + grid charging on briefly, then back again in the App.
Confirmed with Ben Holt (GivEnergy live chat) that my G100 Export Limit is set to 3520W. My system is currently exporting at 4530W. Looks like previous beta software roll outs did not lose the limit setting but rather that the 3012 firmware is ignoring it. (This is why I wanted to perform the blue button live rollout procedure, so that I could test this).
Woodstock isn't that more likely to be an issue with the inverter firmware, rather than BMS? We're still waiting the next beta (or final) release for all the inverter types i believe
mpartington Correct, the export issue is Inverter FW
Still waiting for the next AC coupled one, other things keep getting bumped up the priority stack
@Spencers I will take a look at yours
The reason why beta is gonna need a manual update, is an issue with the portal.
Beta numbers are higher than production, so your version is higher than available, so wont see the update.
This is a blow to my roll out plan for sure. I have beta testers on Beta Inverter and regular BMS @miffy So they can still play their part.
Not seen any calls regarding update, so its either not published overnight as planned, or all it going ok.
TheDragon (GivEnergy) guess that makes sense - stops a beta user accidentally getting prod firmware installed.
I guess the other way would be for the version to be called (say) 3012A for alpha, 3012B for beta, 3012C for release candidate, 3012P for prod.
Might end up skipping a few numbers when going prod to prod, perhaps 3012P to 3015P (e.g. failed or buggy betas), but otherwise would also stop beta users rolling back to prod releases.
TheDragon (GivEnergy) well I’m a non-beta tester and I can confirm No update button has appeared yet. (Although might not be in the batches of people where it may have yet?)
Smk82 I will chase the Devs now!|!!!!
Brilliant thanks for that. I'd been out and came back to the battery showing 100% and all is well. Kettle on and power comes from the battery and not the grid (like it was before). We haven't got export sorted yet, so we're now giving 3kW away to the grid, but it'll get sorted eventually I'm sure.
One of your colleagues replied to an email I'd sent yesterday before I joined the forums, so I hope I haven't caused confusion with two people working on it. I explained that you'd done the remote update for me last night.
Another FW revision for Gen2
This is the discharging while in a charge period, but charge not needed.
During AC Charge window
If SOC < Set % - Inverter charge from grid to set % and stop
If SOC > Set % - Inverter will continue to discharge to set % and stop
Any one use this, i had a few complaints a while ago.
I can load it on for a quick test, only need a couple of pass results. This will then become the final release for Gen2
The fix will go in to AC coupled and Gen1
TheDragon (GivEnergy) Same for me. No sign of the update yet. (Im not a beta tester)
ProximusAl Devs chased.
T|ry again on an hour, and report back
Also a non-beta tester and no update button.
TheDragon (GivEnergy) So the beta testers who want to upgrade to the production version aren't going to get the opportunity to do this at a time of their choosing because the button wont appear for us? Have I understood that correctly?
I wanted to control it so that I could start it at an optimum time where the battery was low but that solar was available to charge it. Also that it would automatically do a recalibration as part of the process, since I still have an issue after several days of empty to full cycles where the cut off SOC is 5% and there is a constant missing capacity. Only getting 8.3kWh out of my 9.5kWh battery (12.5% shortfall)
CCourtney Testers had the option to hit the upgrade button (though sounds like this was not made available) or have the firmware applied by Mr Dragon and avoid another calibration. We were given the choice.
cbowen look again. Devs say its there
It sounds like if the BMS version is beta, you won't see the upgrade to prod BMS - because of the numbering.
Otherwise you can upgrade but will end up calibrating via the portal
Or ask Paul to upgrade for you, avoiding the need for a calibration (but only if going from beta to prod, not old prod to new prod - as a calibration IS needed for that).
Any beta I will do manually if the button wont appear, due to the way the portl looks at the versioning (This was a little shock to me i must say)
TheDragon (GivEnergy) Thanks Paul. It's now there and updating now.
TheDragon (GivEnergy) Yes the button is there for me, I had to move away from 'My Inverter' and back again to see it.
I've got the upgrade button, but do I do it now as my batteries are at 100% or wait until it is very low?
Currently generating 3.2Kwh
@TheDragon (GivEnergy) When I press the update firmware button, nothing happens.
If I do the upgrade will my battery discharge to the grid as though I was exporting?
TheDragon (GivEnergy) no button for us still either. As mentioned a few posts above, you had put us on 4072 as a result of our specific system issues. Don’t know if that meant we are now part of the beta group or something in between? Let me know what I need to do or if you can update us manualy? Thanks
mpartington Good thinking. It failed to register with me that this release was BMS only.
Stevef1964 Yes, it will fully empty. You can use it by turning things on, or it will go to the grid then fully recharge itself.
cbowen Same for me. Nothing happens when button pressed.
How do I know the update is in progress?
Just to confirm, those pressing update the system is not discharging? Has it been 5 minutes since pressing it?
Stevef1964 On you dashboard you will see you system discharging.
THALL I presume this will be delayed whilst the update is completed then the discharge will start?
THALL As it turns out that isn't a choice available to beta testers though since the button wont appear for us. So I will have to ask Paul for the upgrade but I do want a re-calibration at that time and was going to kick off the process if I had been able to do it precisely when conditions where right for me and not have to bother Paul to do the upgrade and re-calibration at say 06:00 am!
mattpdavey will be manual. I can do manually. whats the serial
How many simultaneous downloads of the BMS can be actioned by the portal? I ask because nothing seems to be happening.
Just to confirm, those pressing update and saying nothing is happening. Have you had a look your inverter and battery? While the update is going on they should be lit up like a Christmas tree with multi coloured flashing lights and emitting assorted clicks and whirrs. 😀
Woodstock Yeh, mines not.
TheDragon (GivEnergy) Thanks that would be grand!
Bold move to upgrade on a Friday afternoon of a bank holiday weekend! Think I'll wait a few days or so 🙂
Like this:-
Had update button
1) No acknowledgement anything has started when clicked.
2) Popup window stays on screen and only way to get rid of it is to Cancel - so did it start or not.
3) Main update btn still enabled if it had already started.
Don't have xmas tree lights on inverter (AC3).
All pretty much guess work atm...
Woodstock Ummm.. maybe I should not have pressed the button when I was not at home to oversee things.
I've clicked mine, and like others, it doesn't appear to do anything.
I'm waiting for 30 mins, and then I'll check the logs to see if anything is actually going on.
Can't be crawling up in my attic to look at the lights, my knee is knackered 🙂
( and I did click checkbox for saying I understand) :-)
I started mine over 30 mins ago and it's not done anything yet. No flashing lights on the inverter & battery when I checked. I wonder if too many are requesting at the same time?
Is this also battery discharging 200W even when app pause button is active? If so, happy to test.
I waited 30 mins and nothing had happened so tried the update button again. This time the update message goes away and there is a messsage on the screen saying the update is in progress. I guess the servers were too busy when I first tried.
yep - worked for me this time as well.
Same Here. Just clicked it again....and we're off!!! Lets pump that grid full of electrons 🙂 (And then suck them back up)
Yep, mines updated and now calibrating.
Im going to wait now until some of you have done yours.
I've still got the blue button so from ajcrewe posted mine did not start.
Is now the time we identify all the batteries that do not have a USB stick installed?
Firmware update itself took 7 mins. Now in calibrate...
Should have check my phone.
All the notifications are popping up..
Have been using webrowser to portal.
And mobile will advise when calib is complete ....
I couldn't wait so did mine.
The f/w update got to approx 68% then failed.
I'm on SA2208 inverter f/w but I'm told updating the BMS should not have been affected.
Paul, any ideas why?
So the button reappeared and so I just had to press it.
So, now the update has loaded and my inverter is restarting and waiting for firmware to update.
waiting...............
Battery just bombed to 0%.
So recalibration must have started.
Waiting...........
Stevef1964 This was a BMS firmware, not an inverter firmware. I am not sure why the inverter is restarting? What firmware versions were you on?
Now seeing battery discharging to the grid and house at 2.58Kwh and the panels are not generating (as indicated by the app)
THALL The inverter restarted for me too, just after downloading the firmware, then restart, then verifying firmware.
I can't be 100% the inverter did restart, maybe it meant battery restart and is just a wording issue?
BMS s/w now showing as 3012
Stevef1964 All good then :-)
THALL fingers crossed!!!
Not sure why the first attempt failed, but maybe thats just how it is, updates do fail for all sorts of reasons - meaning nothing to worry about
Stevef1964 Yes solar will shut off. SOC will go to 0%.
As new BMS vs old BMS is TOTALLY different SOC register and calc.
ProximusAl That's actually going to have to be thought about down the line if GE keep growing at the rate they have.
i.e now 15K customers decide to push the update button around the same time (say average Charge/Dis rate is 3kW over several hours) is 45MW - not too much of an issue (but only 5MW below the generally expected level where you need to report output to the BM system - basically they have to inform the Grid of what they are doing/generating so they can balance the grid to suit - no surprise generation or demand)
When it gets to 150k+ customers however starts to get a little bit more meaty so will be interesting to see how this is handled in the future by the Grid (and by all manufacturers...)
hoggy This brings me back to my security background. Why we take it serious.
Imaging if you could set 150k batteries to discharge at the same time. Grid would collapse.
So any vulnerability is taken very seriously and fixed.
TheDragon (GivEnergy) I now have a "spare" G1 Hybrid inverter to play with on my garage testbench.
I have some plans to make it suffer like I've done to your portal - as always I'll let you know how I get on when the Dongle shows up!
@TheDragon (GivEnergy) - seeing as this needs a recal, if I wait until the battery is almost empty on the morning of a soon to be sunny day, will the charge from PV count enough to stop me pulling too much in from the grid?
hoggy That's why I plugged in my car just before hitting the firmware update button.
My battery is currently spewing it's electrons into the car battery, and not the grid 🙂
Paul : at what point of the calibration cycle should we expect to see the SoC figure "come alive" ?
Does it start to increase from zero once the battery is drained or does it wait until it's fully charged again?
BrianHere it will stay at 0 till the charge cycle begins
Calibrate will be at 0% the whole time, voltage drops down to approx 45V. Stops. Charging will commence. SOC will increase.
SOC will get to 100% voltage continues to rise to 57V and balance begins. The SOC 100% point logged. CAL finished.
3012.
<50% and >20% (forced at 25%) SOC adjustment at idle (point no. 10 in the change log). Went from 25% to 13% SOC in this example.
_
_
TheDragon (GivEnergy) This is the discharging while in a charge period, but charge not needed.
During AC Charge window
If SOC < Set % - Inverter charge from grid to set % and stop
If SOC > Set % - Inverter will continue to discharge to set % and stopAny one use this, i had a few complaints a while ago.
I can load it on for a quick test, only need a couple of pass results. This will then become the final release for Gen2
The fix will go in to AC coupled and Gen1
TheDragon (GivEnergy) I know this beta is for Gen 2 inverters, but if you need a tester for AC3 I am willing to try it, as I find the above useful when wanting to limit the amount of overnight charging (on off-peak rate) if the next day is forecast to be sunny.
At the moment, if I set the SOC to less than 100%, I get the oscillations, so instead I am using the reserve setting and estimating how much energy will be used before the timed charge period, which seems to work quite well.
Battery still discharging, so nothing I can do for a while now. So going down the pub!!
Have a good weekend all.
I have hit go, but it's been stuck on 'verifying update - 95%' for about 15 minutes now.
One battery still showing 4 green lights, the other now showing none (both were at about 80% charge when I hit go).
Simon.
I have now refreshed the page and have no status at all and no Blue button.
Both batteries still showing in HA as 3007 with batt1 showing 79% and batt2 1%.
I knew I should have waited...
Woodlands Yes the fix will be replicated in Gen1 and AC3
Simkna mine did this. Don’t worry.
Close your browser, not refresh, then visit site again.
No difference...
What can I do @TheDragon (GivEnergy) ?
My system says it has started the recalibration but this is what the portal is showing for my battery. Currently just sitting at 51.1V discharging at about 10W. Inverter also showing as idle on the portal.
Not what I was expecting to happen. Is this OK?
Oh dear. New issue.
So after battery firmware, battery completely discharged as per calibration, and it started charging again…..but has stopped at 20%.
App shows charge to 100%, so I changed this to 99% then back to 100% but it’s still not charging as per the calib.
ajcrewe Just realised my post is the same issue as you. Think this is an issue. @TheDragon (GivEnergy)
Several calls with support later... Still on 3007.
Once last go, then may have do update manually via USB.
ProximusAl ill fix this. know why this happens.
One reason why the Inverter roll out is to be controlled and scripted. many steps post install to carry out
Ok Paul. Thanks. I can see the charge up to settings keep being modified, so I’ll leave well alone for you to sort.
I owe you a beer….or ten!
TheDragon (GivEnergy) Just got a notification saying calibration complete, but it obviously isn’t. Just FYI.
Battery is now charging above 20%
ProximusAl the inverter has been totally re-initialised, zeroed everything, which needs doing if the inverter is moved from 904/906 or older to the new generation.
All the registers are different, well many of them.
So the tel tale sign is it sticks at 20%
TheDragon (GivEnergy) No worries. The joy of beta testing.
Will I need to set up my timed charge again once battery is at 100%?
All I ever set was ECO + timed charge
Just had a message to say it failed, will try again, fingers crossed
ProximusAl yes, its currently 0000-0000 and off
TheDragon (GivEnergy) can you also work your magic on my system. CE2208G172
I hope the screenshots Ive taken upload.
Does this look right? I had to call support as I got stuck on 0%. They forced a charge to about 10%. It then behaved well, until it got to 23% then started discharging and solar kicked in. I'm now charging on solar only.
TheDragon (GivEnergy) Should now be on 2330-0530 100% 👍
Do you think I need a battery recalibration again, or will it be enough already when it hits 100%?
Really appreciate this.
@TheDragon (GivEnergy)
Hi Paul,
Adam gave up on the remote updates, and I have just successfully updated both batteries via USB.
He said to call back to get someone to run the calibration as there would be someone there until 7pm tonight - I just tried calling in, but get a message to call back as it's now an installer line only...
Any chance you could kick off a calibration on CE2048G033 for me?
Thanks
Simon.
Simkna Done
TheDragon (GivEnergy) it looks like I’ve got the same problem as @ProximusAl can you do the same fix for me please
ajcrewe yours is different.
discharge power was set to 0W. It cant discharge. So will reset it all and re-start the calibrations
Calibration in progress
TheDragon (GivEnergy) Battery now discharging and all looking good. Thanks
Bloody hell I go down the pub and all hell breaks loose!
Monitored sun from from pub garden approx 2kwh average. Just got home and recalibration completed and all looks well.
Maybe monitoring from pub garden should be included in process!!!
Yeah opened the website and there's 81 new posts in the beta thread alone.
I can't update mine yet. Hoping I'll be able to do it on Friday after the GE engineer has been. 🤞
Just averted a disaster! Run out of beer! But wine was chilled!!
Enjoy your weekend
TheDragon (GivEnergy) All the registers are different, well many of them.
Time for a stupid question... I understood that the various external agents and things use register numbers, rather than names, to read/write information (via modbus). If the registers have changed, does that mean the external agents will have to be updated, and will need to match version number ?
Or is it just some internal registers that have changed, rather than the public ones that are accessible via modbus ?
For those who have taken the plunge the discharge/charge is at full power? You can’t set a lower rate via the app or portal?
Smk82
you should be able to, have you tried and failed?
Smk82 I think the upgrade resets timed charge settings, so might reset charge rates too.
Probably worth trying to change them after the upgrade, e.g. when it starts the discharge?
DD reading between the lines (such that Paul's regular postings from home assistant) unless Mark has written a new GivTCP for those on Beta the registers that anyone cares about are likely to have remained the same.
TheDragon (GivEnergy) Calibration Complete and normal operation resumed. Phew!
I can’t stress enough my thanks to you for your help with this.
Absolutely brilliant, thank you so much for your help.
Full 9.5 batt all the way to zero and back to 100% with no probs... Just let it do it's own thing with no interventions or constraint settings....
Paul: If you got a fast response AC3 update I can have then it's CE2223G689 - I have no special configs in the registers and no export limitations so you can just reset in full and update if you find the mins - I promise, really, really, to buy you a few pints :-) :-)
@TheDragon (GivEnergy) So I’ve noticed some discrepancies and don’t really know where they are at so I’ve grabbed some screen shots.
I’m on the new 3012 BMS and the beta D0.994-A0.994 inverter.
The below screenshots were all taken at the same time, within a few seconds.
As you can hopefully see:
The portal has a discrepancy between consumption and battery of about 50W, give or take.
The app is accurate, but power differs from the other inverter app by nearly 50W.
I checked my usage using my Zappi CT clamp and it agrees with the lower value of 373W, and not the apps 430W
Seems we have a mysterious 50W thing going on.
Hope this helps in some way to find the possible issue.
Thought I'd give it a go on my very old pack.
Chose low and slow for the calibration (dropped to 400W below <10% to allow some cells to catch up)
Never had the SOC bug at all but just curious as to what if any changes it brings to operation.
TheDragon (GivEnergy)
It seems to have finished the discharge part of the calibration, but is now just sat idle. I need it to charge up again whilst energy is cheap tonight.
I have a question, how many investor restarts should we expect, mine has done 6 so far, lots of clicking and stuff going on.....
Just gone to check my batteries. One of them had turned itself off..
Turned back on, but still all idle.
Finally kicked back into action. Going to bed now...
@TheDragon (GivEnergy) I wonder if you could have a look at my system, I completed the update, got the expected messages (Firmware update complete and calibration started) but the portal says the battery firmware is 0 and one battery SOC dropped to zero after the update was applied, inverter number is CE2142G186 Thank you.
I have also noticed that the battery capacity is incorrect - instead of two 5.2 it is listed as - 10 Ah | 0.51 kWh | (2 Packs)
ProximusAl probably the inverter itself. On the app you can choose a calculation method, but not sure if the web portal offers that option.
@TheDragon (GivEnergy) A little update on my calibration. The batteries now show the original specification (sizes) and the original software version and the "update firmware" button has re-appeared, however - 
This morning nothing to report except that my system is running as before with seemingly no issues.
I intend to monitor it over the next few days.
The two batteries do seem to be slightly apart , 43% & 44%, but hopefully this will rectify over the next few days and then stay. I was led to believe this update would fix the batteries drifting apart, but I could be wrong.
So mines done. Have noticed my battery "full capacity" at 100% has dropped from 150ah to 140ah (from a design cap of 158ah)
Will see how that shakes out over next few days in terms of energy but if that's real then that's s bit of a drop.
@TheDragon (GivEnergy) i know this is supposed to learn as it goes, does that change get reflected in the "full capacity" as it goes along or is that just set once on the calibration?
TX200 If I turn off “Show Calculated Load” then consumption matches the other inverter app, but the 50w discrepancy just moves between battery and solar. I think “Show calculated Load” takes one figure and uses maths to add them all up so it looks better.
I mean 50w really isn’t an issue in the scheme of things, just reporting it as a finding in case something can be done about it.
hoggy It adjusts all the time, periodically
@ProximusAl
Inverter itself needs some juice.
Approx figured
AC3 20w
HY 30-50w.
hoggy mine are also reading low (102 design capacity) following a calibration a couple of weeks ago. I haven't seen any changes to the values being reported over that time. Maybe i dont run enough full cycles.
I suspect they have been incorrectly calculated (but aren't used), as last full charge I got 8.8kWh in my 2 x 5.2kWh batteries (so performing just fine).
The calibration was when earlier beta firmware was causing issues pushing the SoC down at low power levels, so I did see some soc jumps during calibration as it corrected, which I wonder messed up the calculation.
mpartington this value isnt used in the actual charge discharge.
but is indicated, calculated
TheDragon (GivEnergy) do I need to run another calibration or just wait it out (assume it needs full cycles to update)? Like I say it's not (appearing to) impacting anything, but as the new firmware considers aging (point 6 in release notes),
"Modify static discharge calibration look-up table as the capacity changes (Ages)"
I wonder if this might give me an artificially aged battery which uses wrong look up table.
mpartington cycle count is used for sure in the code i read.
Variables in use each used in various ways
Voltage
temperature
cycle count
cell type
cell manufacture
TheDragon (GivEnergy) cool, I'll ignore it then. I did hope it would use cycle count.
Mine all working perfectly since yesterday morning when @TheDragon (GivEnergy) reset my clamp. Charged up to 100% from the sunshine and was still around 80% this morning (I don't do a forced charge yet since I'm not set up on Flux yet).
I noticed on the App that our typical load is 40-50 watts higher than previously, but put it down to tolerances between what my smart meter would show before we had solar, but seeing the post above that our hybrid inverter takes 30-50 watts then I guess this makes sense.
Probably not news to those on here with more solar experience, but I was surprised to see that even on this wet morning we're generating more than we're using (600 watts to about 250 watts), so our battery is charging slowly. 🙂
Thanks for your help @TheDragon (GivEnergy)
TheDragon (GivEnergy) ah ok, if it’s using cycles I’m up near 1000 now.
No idea who made my cells.
TheDragon (GivEnergy) Is there any data supporting the real life degradation rate of the batteries, say per 1000 cycles? And what effect if any is this new BMS firmware having on DOD?
3012 SOC % adjustment.
Following yesterday's SOC adjustment at idle from 25% to 13%, I have had on the fly adjustments today from 31% to 15% and from 9% to 15%. Fingers crossed these will reduce in number and size over time. Wonder if the jumping around is coming from the spot look ups of SOC% or the tracking of battery charge/discharge in between (or both)?
TheDragon (GivEnergy) Heres another one for you on BMS 3012. You might say it’s normal, just I’ve never seen it before the firmware.
Battery ran out at 20:40 at night with no solar, but as you can see at 21:45 it decided it can continue to provide power.
Battery cut of is set to 4% but it stopped at 6%
Was not charged in between. It’s almost as if it cooled off, or warmed up to 8% on its own, and started delivering energy.
ProximusAl This is one of the major changes of the new firmware. In this case the battery wasn't as empty as the system thought. After performing an idle OCV correction, it recalculated the SoC and released the energy. It also should now have learnt this new bottom limit for next time.
Going to give mine a few weeks as it learns but it's misjudged / overextending mine at the bottom end on first try.
Took itself to 49.5V and 4% actual, 2% Inverter - needed 7.11kWh to fill it back up.
Another run this morning and it's pulled 6.4kWh out (which is pretty much spot on given the packs age) but just waiting to see if it decides to adjust and carry on again after OCV idle)
I'm guessing Mark is going to get to know the register for the "adjustment complete flag" in the BMS to bring into GivTCP at some point. 
hoggy guessing Mark is going to get to know the register for the "adjustment complete flag" in the BMS to bring into GivTCP at some point.
I hope so that would be a usual addition.
Morning All
Can't work out if this is a daft question but is there or will there be a set of pointers that people with much less experience than a lot of the extremely knowledgeable community on here can follow to see if we are affected by the issues that have been discussed in this thread?
From the outside my system appears to be working correctly but I am struggling to collate all the different points such as voltages, SOC levels etc. to work out if I am getting all of the functionality that I should be. Would be great to be able to follow something to check.
Maybe to make the point you could check the data on my system? Serial number is SA2142G260 and my versions are D0.450-A0.450 with battery 3005. An update is showing as available, although I don't know to what version it is offering.
Thanks in advance!
milkycgs It will only offer the latest version. Your choice as to whether to update or not.
Can someone remind me of the criteria for when the OCV will run. My 2 batteries are averaging about 10% difference between them and haven't really seen much sign of them wanting to close the gap. They haven't been full or empty for several days now so I guess this is the problem.
Woodstock 3012 SOC % adjustment update.
In addition to the 31% to 15% and 9% to 15% on the fly SOC adjustments yesterday, my system went on to perform a 5% to 10% adjustment when the battery was idle at "fully discharged". A total of 27% SOC adjustment in a single day.
Will continue to monitor to see if this reduces over time.
Hmm this seems a bit odd too. Constantly charging and discharging at the top end. I thought once it hit 100% it should idle. Solar is producing enough to run the house.
anglefire Thanks for that info.
ProximusAl You also need the next Gen FW on the inverter to see the full effects.
BMS only has no serious drawbacks, oly benefits. But these odd issues will happen, as the BMS & inverter aren't fully communicating. Many registers needed are not there.
Oy the basic OCV will work, as this is not under Inverter control and can run on its own
TheDragon (GivEnergy) You did pop me on the latest inverter beta the other day. Are you saying there is an even newer one?
Thanks for replying
TheDragon (GivEnergy) Hiya. I've been away. Have noticed the upgrade button is available but I'm gonna wait until I have an export tariff to hit it. Besides, I'm not running new s/w over a Bank Hol weekend!
Bit of an upgrade issue for me - wondering if anyone else has had it.
I'm now 2.5hrs post confirmation that firmware has been updated and both the Solar and Battery are at Idle and I'm importing from the grid.
Most worrying..the battery capacity is now showing only 0.51kwh on the cloud information. ....eeeek
Anyone else having or had the same?
Cheers
Lmcde3 have any of the switches tripped on the consumer or battery ?
Leon_shaw thanks Leon but all looking good there
Charged from 4% SOC to 100% SOC today but the (8.2kWh 100% DOD) battery only took 4.9kWh according to the portal ribbon?????
Woodstock Did a bit of digging and found that in "Meter Data" the "Battery Charge Today" figure reset from 2.9 to 0 at around 13:06 and the "Battery Charge Total" went down at the same time but not by as much???????? If you add the lost 2.9kWh to the displayed 4.9kWh then 7.8kWh is bit more like it.
Lmcde3 restart the inverter? (Via web portal, my inverter page, power symbol button.
Woodstock "Battery Charge Today" was reset to 0 at around the time I performed a "Reset to Defaults" (13:08) but can't say I have ever seen it have this effect before. The "Battery Charge Total" went down by 1.45kWh which is suspiciously exactly half of the 2.9kWh lost "Battery Charge Today".
TX200 no luck on that but did a hard switch off and reset the batteries and I'm now seeing charge. Will see how it goes and update. Cheers for the replies.
miffy If there is a previous failed update file stuck, or a USB corrupted you get issues. Otherwise its 30 mins of wait then calibration.
It was timed with good weather for the vast majority moaning about calibration cost.
Now I cant do right because irs a bank holiday and you should never release on a weekend. cant win. But not getting down from the complainers, far far far exceed by the successful and happy non crash users now.
TheDragon (GivEnergy)
A very happy 'customer' here. I uploaded the new firmware from 3007 at 3 am this morning, making use of a run of low Agile import rates, averaging 13p per kWh. My 8.2 kWh battery was on 56% with AC3.6 inverter.
Firmware update took less than 5 min. Followed by discharge to grid (at 15 p Octopus outgoing) to 0%. Took 1.5 hr
Followed by charging, which took 3.5 hr, finishing at 8 am with 100%, the final charging coming from a mixture of grid and early morning solar..
I reckon that 1 hour of charge/discharge equating to 30% of my 8.2 kWh battery.
I could see what was happening on the app (iOS).
Everything went so smoothly but I didn't touch anything during the calibration process and I was sleeping through most of it. I would recommend just letting the process do it thing. It does take several hours.
Massive thanks for your efforts, Paul.
I now look forwards to the inverter upgrade to complete the process.
Rob
TheDragon (GivEnergy) No worries. I have just had an email from Octopus saying my export tariff should be working on 10 May, so will hold off until after then; maybe next week if the weather improves. There's no reason for me to hurry installing the update, is there?
miffy I waited until my battery was flat - but I have a fairly high usage and not a big array so apart from very sunny days the battery doesn't fully charge - hence most days goes flat.
I'm assuming you are not in this camp and will always have spare in the battery?
TheDragon (GivEnergy) It was timed with good weather for the vast majority moaning about calibration cost.
Silly for people to moan about that. The button is yours to press when you want, just wait until the time is right for you. The fact it appeared doesn't mean people need to press it.
TheDragon (GivEnergy) Now I cant do right because irs a bank holiday and you should never release on a weekend. cant win. But not getting down from the complainers
I do think such a major eagerly anticipated update going on a Friday PM before bank holiday weekend is pretty risky IMO, [which is 20 years in software delivery to millions of customers, not trolling]. There's always risks things weren't found in testing [horrible things sometimes] and you might need your full team available. As always, you weigh the risks vs the benefits. Folks have been waiting 1.5+ years for this fix, a few more days would have not really hurt and might have saved your weekend. But that's all your decision, and you're owning it.
I wouldn't see this as complaining, it's just feedback.
pacemaker I do think such a major eagerly anticipated update going on a Friday PM before bank holiday weekend is pretty risky IMO, [which is 20 years in software delivery to millions of customers, not trolling]. There's always risks things weren't found in testing [horrible things sometimes] and you might need your full team available. As always, you weigh the risks vs the benefits. Folks have been waiting 1.5+ years for this fix, a few more days would have not really hurt and might have saved your weekend. But that's all your decision, and you're owning it.
I wouldn't see this as complaining, it's just feedback.
Perfectly said. I think most of us even talking about it have experience in this regard, either as the developer, the person rolling it out or the operator left to deal with the system not working when the engineers have all buggered off home.
I'm glad the update is out and upgrading fine for most (I've done 2 of the 5 batteries I have on site). My suggestion would be a post mortem once it's all said and done to look at what might be done differently or better with future updates, be that the time it took to get it, release strategy, beta testing organisation etc. The ultimate goal has to be to learn from every opportunity.
@TheDragon (GivEnergy) Big thank you for releasing the new BMS - made huge difference here.
dbt85 Think the Beta testing organisation etc is sorted with Paul onboard TBH. As for releasing over a bank holiday - not sure it makes that much of a difference - it has been so eagerly awaited by many that even a few days miffs some people - and given that Paul and Martin (I've not seen any other names on FB but I'm sure some more are about) have been around all weekend is credit to their dedication to sort as many out as possible. You can be as sure as eggs is eggs, that many would have started the process "out of hours" to benefit from cheap rate charging or simply being at work, that the team at GivEnergy would have been working stupid hours regardless.
anglefire I've a new installation - only fully working 14 April 2023. It's a 5.2kW system with a Gen 2 5kW hybrid inverter, and two 9.5kW batteries. Our average daily usage is 19kW hence the two batteries. Charging on solar only, since 14 April we've saved 80% on our electricity costs. [I have a spreadsheet!]
Been on Flux import for two days, and due to go on Flux export later this week. Up until now I've only charged the battery from solar, so most of the time the batteries do not get to 100% thanks to our crap weather, and more often than not they've gone flat before midnight. For the first time last night, I used the timer to force charge the batteries from 2-5am, which got them from 40% to 89%. We are due barely any solar today, no more than 9kW, and it's not looking great for the next few days either.
I'm not in any hurry to apply the update. Currently thinking about running it on Thursday 11th May, assuming my export tariff is up and running by then. I don't want to export to the grid for free if I can in any way help it!
miffy If you are going to have to export to grid due to your usage patterns, then I too would wait until I had at least an export tariff in place to get something for it!
New BETA Gen1 D0.195-A0.195 (Final Beta/Pre-Release)
Looking for single string PV users, who have seen a cap of PV generation of 2700W HY3.6 or 3700W HY5 and not saturated the AC side, ie battery charging and AC out, but PV still capped for no known reason (not 11A limit either)
TheDragon (GivEnergy) I have a single string PV (SD2142G119
), just upgraded to the current BMS, but I'm only capping at 3700W due to only have 9x410W panels if you want to test that.
Hi TheDragon (GivEnergy) I've seen some quirky SOC behaviour from the new 3012 firmware this evening - the first evening when the SOC dropped low. I have a reserve of 10% set and at 18:42 is hit 11% (which is where I'd normally expect it to stay). It was steady at 11% for nearly 30 minutes and the inverter state was idle as expected. Then suddenly the SOC started dropping like it was being drained, only it wasn't outputting and it was still idle. The SOC dropped to 5% where it it now stable again. Thought it was strange enough to bring to your attention, mildly inconvenient, but not show-stopping enough to interrupt your bank holiday weekend 🙂
SpeakToTheGeek from why I’ve read about the behaviour that is normal whilst it’s finding the limits. It does various calculations based on voltages temperatures and the like - when idle is most accurate and it sounds like yours was idle so it did some stuff and decided the true SOC was 5% not 10% so made an adjustment.
It should sort itself out eventually. Or that is what I am lead to believe.
anglefire @SpeakToTheGeek This is exactly correct.
At idle it uses this known state to check the cells actual voltage vs lookup tables and makes adjustments up or in your case down.
Here is an example of the other way. 5% idle, OCV decided it was nearer 7% and SOC rose up, didn't hit 8% to trigger the discharge circuit back on though. If it releases 3% more, it will open up and discharge again.
This is what 1min resolution data (Which will be for all at some point later) looks like
TheDragon (GivEnergy) That's quite neat, thanks for confirming. It's a lot better than the sudden drop that's for sure, not that I ever actually experienced that myself with recent firmwares. I occasionally got a jump from 95-100% at the top end when charging, but that never bothered me. Good to see it's learning its own capabilities! Thanks for all the hard work on this.
TheDragon (GivEnergy) Hi Paul. I have 1 string (but having extra panels in a months time on a separate string), Hybrid 3.6 Gen 1 SA2142G078, but not sure if I've seen the capping as it's so long since I've seen any decent solar!
SpeakToTheGeek Thanks
It is a continuous process, as each day the variables are different. You will see on multi packs them playing tag with each other as they adjust to common agreed SOC. Over time it reduces.
I take it the "old" actual pack SOC boundaries are still in force so it should sit between 15% and 95% actual?
Mine seems to go idle at 15% then adjust itself down after OCV - which I get is the entire point, spent all night again at <10% actual last night once it realised.
Although I'm up at 980 cycles - are the cell degradation offsets linear or a hard step at say 1000, 2000 etc...(in which case i suspect another week or so and it'll probably stop going under 15%
Batteries 3 4 and 5 are all updated and now charging for their calibration. All were near empty this morning so it was a perfect time. They are all throttled on their charge rate as 2 of them are on a 72p per kwh business rate and frankly that stings!
TheDragon (GivEnergy) How long should people let their systems settle/learn for, before the system has reached a reasonable equilibrium? In either days or number of cycles?
ToothyChris id say from testing 3-5 cycles
Just wondered what type of balancing is taking effect.
Top Balancing
Bottom Balancing
and whether this is Active or Passive Balancing
From a steady battery decline from 54% from 00:00 over night it suddenly dropped like a stone from 26% to 4% between 6:15 am and 6:45 am this morning.
Also note the power drop to 3w.
I assume this is all part of the cycling process?
Im still seeing my battery show 0 degrees C about once a day via the battery info on the app ( I'm assuming this is just a missing data point?)
Was 3012 going to fix this too?
Roberto Sounds like a BMS comms issue?
Roberto I would contact support about that.
THALL i was thinking it might get fixed via 3012, but maybe not. I'll give GE support a shout.
Roberto It is not a software issue I am aware of. Either a comms problem, physical connection, or a faulty sensor.
I'd just like to add my name to the list of happy punters.
I took a brave pill this morning and updated from 3007 to 3012 on my 8.1kWh battery.
I did have a small issue where my battery calibration didn't start automatically, but Adam from GE sorted that after a quick call. Calibration is now complete and I'm back up 'n running.
@TheDragon (GivEnergy) Take a bow, Paul. You can see on this thread when the sentiment turned from hostile, through frustrated to engaged. You were a key part of that by your proactive chasing and keeping us informed.
Thank you to you and the team for pushing this update through.
I did the update earlier today on my 8.2kwh battery and it all went perfectly The house was empty and I watched it do its thing from work. Came home to a nice full battery mostly from the sun with only a little bit of grid when the rain came.
Another happy customer checking in - updated my 9.5kWh battery yesterday from 3008 and its all gone flawlessly. Thanks GE and @TheDragon (GivEnergy).
Updated my 8.2+9.5 AC system last week which went smoothly.
Too much sun so far, so I've only had one bottom end adjustment. So tricky to judge capacity improvement, but looks promising.
But the SOC jumps have gone which is good. Just hoping that the batteries start to sync up a bit more as they are always at a different SOC.
THALL Thanks for the heads up. GE have confirmed a fault and have logged it as a level 2 priority, whatever that means. They are arranging an engineer.
However, should i also be contacting my installer? Maybe they should be fixing it?
Been 6 days since firmware update, 8.x battery has smoothly gone from 100 > 0 each day & vice versa not a single sudden drop to zero or jump.
It's too early to declare victory given the length of history of this issue I've got - but this is honestly looking good.
Any indication on when the G2 hybrid inverter firmware might be released? Im stuck with the no solar generation when battery is full bug.
Hi, following my upgrade from 3007 to 3012, I 've had a couple of instances on the app showing one of my 3 gen 1 batteries at 0% and red and the others OK and green , even though all batteries are charged. It resolves later and has affected the primary as well as the other 2
Any one else getting this? Thanks
I've had this too, but before the BMS upgrade. I had to make sure my system was connected to the internet, which it was, and refresh to make sure the system was talking ot the cloud portal.
Not had it since the upgrade.
Thanks, ironically I've only had since the upgrade, I'll keep watching...
Any news on when the inverter firmware release will be?
Hilditchhouse Do you have givtcp setup? It would worth checking that all the batteries got the 3012 upgrade. You can contact support for them to check this too. It could also be purely a data issue on the app. I would assume it is a app data issue.
I also had this once shortly after the update (2x9.5) and then it resolved itself. I believe it might be when one of the batteries is throttled to zero to allow the other(s) to catch up. In such a situation, perhaps the app is not reporting it correctly. In my situation, there were inconsistencies in how GivTCP, the portal and the app reported it.
Hi, thanks for reply, my first attempt to upgrade firmware failed after 2hrs, so Sam at GE started it and oversaw remotely, that it installed OK so am hopeful 3012 is on all 3 batteries. Shout out to Sam, great help on tech support desk. Calibration took forever , and so far no real issues....fingers crossed.
Hilditchhouse I seem to remember that The Dragon/Paul said something about this behavior recently. Something about a battery being switched off as part of the new process. I remember he made a joke about the battery being naughty, and needing to consider what it had done, or standing in a corner, or something. It may have been this thread or one of the others - but may be worth scanning through his posts from the last few days.
ToothyChris great thanks, I regularly needed to switch the batteries off to behave themselves so hopefully now a thing of the past, cheers
Upgraded yesterday to a gen 2 inverter and saw exactly the same issue. The slave 9.5 sat stubbornly at 0% (although at 53v) for several hours. A new calibration started late evening sorted it overnight. However we also see the SoC discrepancy others are reporting, already several percent different. The master battery also reported an alarmingly high temp of 78deg C during the calibration run but back to sensible numbers now. The app usage figures were also bizarrely high but also settled now. Eventful!
It was this thread: https://community.givenergy.cloud/d/2258-post-bms-update-data-anomolies
Thank you
4monks Thanks for finding that.
When is the firmware update for the inverters going to be released to the general population? The sooner the better. I have a Gen 2 inverter and when the battery is full the inverter does all sorts of crazy things which end up costing me money. e.g. pulling power from the grid when the battery is full and it's bright sun in the middle of the day yet the inverter is only showing 783w of solar generation when before the battery was full it was 4kw.
I've just had my DNO back so I'll soon be loosing money from missed export payments as well now.
Hilditchhouse In this screen, battery 2 has gone out of range of the other 2, or has a cell out of balance, its took itself offline to deal with it. It should return once the fault is rectified, most the BMS corrects the issue itself
Cool thank you
Andy C Have you carried out your SPC charts on actual achieved battery capacity since the 3012 FW upgrade? If so how are they looking?
Did I read somewhere that some Gen1 hybrids will need a hardware component upgrade to support the forthcoming inverter firmware? I have a failed inverter on site (there are 2 here) and I wondered if there was some way of doing the upgrade to both whilst the engineer is on site repairing the broken one?
MikeJ Only Gen1 hybrids pre-a certain serial number.
Ra I've tried searching here, but can't find the serial numbers it applies to. Hopefully the repair on the broken inverter will contain this fix, but I was hoping that if it was needed, I could get GivEnergy to sort the other one on site to avoid another visit.
Those before SA/SD2209 were the original ones in question however I'm pretty sure they've worked around it in software now.
hoggy Thanks! Both mine are pre that.
Just a small update from me. Not really had any issues on the beta or the current 3012. But today I have a 5% spike. Hit the 57v so will be interesting if this adjusts itself. I have not had the packs fully charged for a few days due to poor weather and higher usage.
CCourtney I just need to update the last few days then I'll post an update, thanks for the reminder!!
Here we go all, this is the data across the various battery management systems over the past few months (only data points with a meaningful battery charge):
Shows very nice migration in terms of increased battery capacity, but just as important showing the better control of battery capacity. I could have lumped 4071 and 3012 together, but it would have made little difference to the calculation of control limits. Mean battery capacity across 4071 and 3012 is 10.3 kWh, so by my calculation 95% of that is 9.75kWh.
Hope that's interesting for some!
Just thought a summary table of mean, standard deviation and %SD might also be interesting:
Andy C Impressive steady progress made through the various FW stages then Andy. Looks really good. Thanks for sharing.
I've seen the question asked a few times, but I don't think I've seen an answer yet. As the inverter testing appears to have drawn to a close, is there a date /timeframe for the Gen 2 inverter firmware release?
Andy C Out of interest, did you base your capacity measurement on energy in or energy out?
4monks As far as I know inverter testing is on going. Only the bms testing has finished. Gen 2 and AC inverters are still waiting for new firmwares, maybe more testing.
THALL Thanks.
What is the process to join the inverter beta firmware program?
Arg0t I believe it is now closed, they already have something like 150+ beta testers.
Just out of curiosity, was anybody on here lucky enough to get into the new 1 minute update testing? Saw the post on the FB group but was too slow. Was curious to see what it looks like.
thewhalw21 that did look good I missed it by mins lol
Woodstock energy in. At some point I mean to do an analysis on round trip efficiency, just the time to get round to it!!
Is east/west fixed in the inverter firmware now or still an issue?
kram I think it's fixed in beta firmware. Gen2 only I think, which will be one of the first to be released.
Noticed this spike at about 3:30 this afternoon. Coincided with the battery getting to 100% as far as I can see.
That would put GE firmly back in scope if it’s fixed. 👍
@TheDragon (GivEnergy)
Hi Another SOC drop this morning.
Inverter Firmware Version
D0.904-A0.904
Battery Firmware Version
3012
Gen 2 3.6 Inverter
Rodmac76 interesting.
Has the OCV had an oppertunity to run? Was a calibration done?
Im away all weekend so not working
kram yes. I have the new FW now.
I will sort next week getting it initially to the service desk. Then once the portal is ready, on there 2.
I need work doing to allow Inverter and BMS to load seperatly
Yes the update went through last week and completed a full calibration
Look forward to seeing the data and feedback. Enjoy the time off, well deserved I am certain 👍
Been a brilliant day here today, blue skys all day. Gave me a chance to produce a graph without too much cloud noise which I think is a good example of the "Before " behaviour of the current production software (D0.906-A0.906 ).
I have a 5kW Hybrid, Gen2, two strings, 11 panels ( 4.62 kW pmax) south and 6 panels ( 2.52 kW pmax) west.
Shows:
East/west strings slow start in the morning. ( My strings are South/west, but same effect )
Southerly string hitting the 0.75*5kW limit
Maximising PV by full 5kW AC plus Battery fill during the day. Good apart from south string throttling.
When battery is full, and MPPT starts throttling total PV to 5kW slow to stabilise. On a normal partially cloudy day, it never does.
Shading performance ( Hard shade). West string is 6 half cut panels, as soon as the bottom corner of one of the panels gets shaded, then the Power for that column drops to 1/2, bringing down the whole string, and when both "halves" of a column are shaded, then the production really drops away. The base level ( a few hundred watts) shown is with 5 Panels still in full sun. There seems no attempt by the MPPT to force the bypass diodes in to action, which I would expect if it was looking for global maximum.
Looking forward to trying the new firmware when released, to draw up the "After" picture.
More strange battery behaviour set to charge to 100% but only reached 91% and stopped until end of Go window.
May require another calibration? Sun day tomorrow.
anglefire Also had a high voltage alarm on the portal.
My homeassistant computer went down yesterday - so running sort of blind at the moment - only sort of because the GivEnrgy App works fine as does the portal, but its surprising how much I rely on it! Hopefully will get it back up today or tomorrow.
I'm having the same issue, my battery is maxing out at 93% despite being set to charge to 100% overnight. My battery isn't getting much below 50% at any point in the day now with me being on Octopus Flux.... could this be the issue?
Gen1 5.0 Hybrid with 100% DOD 8.2kwh
Large SoC adjustment today of 20% (on 3012) after forcing an idle OCV. From what I understand adjustments.l are only made if the battery is under 40%, so if like me (with this good weather), you cycle in the top 60%, still lots of opportunity for them to gradually drift, resulting in a large correction once below 40%. This happened only 7 days since the last OCV was forced.
Edit- on recharge, battery took 8.5kWh, so the asjusted SoC after the OCV ran was spot on. Conclusion, 20 % drift in 7 days.
Massive soc drop on 3012 ?
55% to 0% ?
The battery has hardly got into the lower half in the 9 days since the inverter was replaced, the battery was calibrated and then upgraded to 3012 from beta.
TX200 I have Gen 2 5kW inverter with 8.2kWh battery. I had issues similar to you from day 1 (installed Feb2023). It was fixed with inverter firmware updated from 904 to 906, battery firmware not updated yet, still on 3007
I have other issues I described yesterday but not that one anymore…
Hi, this is happening as I write this.
Soc dropping from 29% to 19% (and still dropping) whilst the app is showing the battery charging at just under 500W. See diagram below
This happened to me on the 10th May too
Is there a problem
Dom66 yeah but firmware 3012 was supposed to stop this from happening!
TX200 is there any documentation about what 3012 fixes/improves/adds?
So, my SoC went down to 13%, hovered there for a while and has now risen to 14%.
My two batteries where showing % quite a way apart but have now levelled at 24%.
So is this the new 3012 BMS doing its thing and re-aligning itself?
Dom66 it's all about the battery and calculating the state of charge (soc)
https://kb.givenergy.cloud/article.php?id=22
Good luck with the technical terms on here. I think I understand half of it and I have a degree and many years experience in IT. 🤣
Stevef1964
I believe so. doing what it should. 😃
Hi @TheDragon (GivEnergy) how do I do another battery calibration? can this be done through the software? or do we go through technical support? Hopefully fix my SOC drop and battery not going to 100% (within the Go window) today should be a good solar day.
TX200 Seen, and passed to the FW engineer to take a look. Seems usage always over 50% is causing issues we do not want to see, after so much hard work testing over and over again.
Rodmac76 whats your serial, I can set one off, or support can too
TheDragon (GivEnergy) ED2152G223 Inverter will this do or do I need anything from the battery?
TheDragon (GivEnergy) Seems usage always over 50% is causing issues we do not want to see
eeep, this is my use-case. I'll hold off and watch this space.
cheers, there will be an email ticket sitting in the queue to, submitted 0819
I've been trying to work out the ins and outs of the battery from the meter data on the portal. Not 100% sure of the best way to start (e.g. based on a percentage of charge, or calibration day, etc), but ...
Rodmac76 Thanks, will run this now for you and stick the newest inverter FW too while Im at it.
Its not ont he portal yet, as they cant deploy just inverter right now. So portal is needing some work, before Gen2 final 909 version is mass released
TheDragon (GivEnergy) Any chance I could be put on the laters inverter FW too please?
Rodmac76 calibration begun
ZmGahH6yQBe Mt army of beta testers will get it manually during this week.
Then when the portal is ready to deploy, i will test it on a few more, as deployment is not standard on This new Old to new FW jump.
SO needs Portal deployment testing
TX200 not sure what it's up to now, but think it rebooted (heard a long beep from downstairs).
Seemed to be exporting my solar to the grid and exporting 3w from the battery too.
I've toggled eco mode off and on which I think has sorted it.
TX200 technical description useful for some I imagine but not me. The previous upgrade 3007 (which I am on) had a better description which I hope I can find the 3012 version of. Perhaps @TheDragon (GivEnergy) can help - he seems to pop up everywhere with much wit and useful info…
TX200 ahh, the reboot was the support desk upgrading the inverter firmware after an issue the other day (system swapped to grid half way through a high load cycle on washing machine).
TX200 So they have taken you off the Inverter Beta FW?
TheDragon (GivEnergy) I think so?
Now D0.535-A0.535
It looks like 3012 is doing a worse job of keeping my batteries in sync than before the upgrade...
Simkna
The self calibrate doesn't run in the upper half of the SOC.
It runs in the lower half under low power use, and idle under 20% .
Lt them run down, idle and see what happens. Here is it on mine working faultless
I am discussing how they can also run some calibration above 50%, however the voltage tolerances are so tight, its accuracy will need checking.
TheDragon (GivEnergy)
My batteries rarely fall below 50% in summer, and are generally only idle @ 100%
Is there any way to force?
Its not for mine. See here: https://community.givenergy.cloud/d/2272-dual-battery-soc-discrepancy-after-firmware-update/59
TheDragon (GivEnergy) The self calibrate doesn't run in the upper half of the SOC.
I assume there is a technical reason for this? First couple of consecutive bad weather days for those of us who don't drop under 50% typically are going to see the SoC hit 0% quite quickly with the drift over a few months will mean we probably only have 60% full at best?
So is the new inverter firmware imminent ? Sorry just want to know as I basically get no gen after battery is full and my eddi is hungry ?
welshrat I called support and they applied new inverter firmware, everything up and running with new FW within 15 minutes.
Hi Dom66 is that the inverter firmware ? I already have the latest battery firmware and was told when I contacted support a few weeks back that it was still in Beta and their beta testing was now full and I would have to wait for general release ?
welshrat yes, latest inverter firmware. They have not advertised this as being ready but I was told by @TheDragon (GivEnergy) to contact support as it would fix my issues. I have not bothered with battery firmware (Support said this was OK) as it seems to be for multiple battery issues - I only have 1 battery. Support said another battery firmware is due soon anyway!
My inverter on web portal now says:
Inverter Firmware Version
D0.909-A0.909
Thanks Dom,
@TheDragon (GivEnergy) is this something I can now request ? I know you have a million things to lookout for so don't want to cause hassle by contacting support ?
Simkna
Could you not just switch off Eco mode, or use the UI to initiate a forced export?
I'm no export, but it seems to me the batteries don't like being constantly full or nearly so -- occasional moderate discharge is good. So perhaps schedule fortnightly a time where you either don't charge batteries from solar (ie Eco mode), or you force export to Grid, so as to take SoC down to e.g. 25%.
20% drift in 7 days for me. I was always worried that partial cycling would be the true test of this firmware success.
I'm going to mitigate whilst weather is good by discharging into car to around 30%. Should naturally idle for at least 2 hours and leave enough to see me through until solar kicks in.
TheDragon (GivEnergy) idle with all batteries under 40% forces the calibration for me. Any benefit going under 20% you suggested?
Proper regular top point calibration also has to be an option doesn't it? Should be able to tell if it's a significant way off 100%
mpartington idle with all batteries under 40% forces the calibration for me
I'm concerned the next time I get close to 40% it will crash to 0% which defeats the object of this whole SoC issue. Each day I want (and should have) my battery charging to it's maximum capacity. If the only way to achieve this is that mini calibration occurs at 100% then so be it. I'm ok with battery idles at 100% for 15 minutes, then system starts to evaluate for however long is needed and adjusts the SoC down if more capacity found, then can start charging to new 100% level. And so on...
justpassing You referred to a failure of the inverter MPPT to find the best output power when part shade occurs. I have had this, but not with a givenergy inverter - I have had an Aurora PVI 3.6 for nearly 12 years. On one string I had shade from a BT telegraph pole. I could correct the MPPT output by fully covering the most affected panel (with a cardboard sheet). This forced by-pass of the current and the MPPT would then adjust to the correct output for the other 8 panels in the string. When I then removed the cover the correct output would hold. However, as soon as a cloud came by temporarily reducing output, the MPPT would revert to the low peak.
I paid BT more than £1000 to remove the pole. That may seem OTT, but FiT tariffs were (are) good from that time. Of course removing the shade may not be an option for you. I can see that an MPPT could look for multiple peaks and choose the highest but I suspect that they assume that there is only one highest peak and lock onto that when they find it.
Thought id provide some feedback on the new Inverter firmware.
I asked support to have it upgraded this morning and within 15 minutes it was done.
The PV generation cap when battery above 90% preventing grid export bug has now gone and the inverter is much faster to respond to house loads.
All in all very happy, thanks to Adam and the team at GE 🙂
Efpv did you phone or message? They told me to wait till it was on the portal 😞
Yes please, how did you go about this as I get virtually no export or surplus for my Eddi after battery gets full
chris_p agreed, having experienced SOC from has high as 83% and regularly from over 50%, having this mechanism only kicking below 40% does defeat the object. So it back to, as it has been for far too long, “SoC crash” anxiety.
Dai I have not had any soc crashes since joining the beta or having 3012 installed. The OCV that runs is to address capacity issues and other things. I saw a weird 5% jump at the top about a week ago, I do see adjustments sometimes when the system is at idle and 100%, a couple of small dips to 98% and it squeezes in some more charge. But the SOC drop issues, the big ones are gone with the new firmware.
THALL But your batteries get cycled to quite low SoC on a regular basis?
Well I have emailed support in the vain hope they offer me the update.
I’ve not seen a SOC crash since applying 3012 but also not seen any evidence of calibration either, keep telling Mrs Dai to bake lots of cakes in an evening to use what’s left after the peak flux export period as struggling to get the SoC down below 30% with these long daylight days to find any evidence of a battery calibration going on to give a measure of reassurance that all is fit for purpose after so many crashes over a long time whilst we all awaited a fix.
Efpv ditto… all my issues are no more with v909 of inverter firmware for my Gen 2 5kW hybrid inverter. Everything behaves as expected. So happy!
justpassing I emailed support.
welshrat emailed support.
chris_p Nope, 17.7kw of battery at this time of year rarely go under 50 - 60% unless I dump it all into the car. Everything just seems smooth now.
Dai I have not seen a ocv kick in on the low end for a while, I have not been able to force it in the 2 tests I have done either. But in the colder weather at the start of the year on the beta I saw the system find capacity on its own several times. When my system is 100% for long periods have seen dips to 98% without cause and then a small amount of charge going in. Nothing major, but I had tons of solar generation and no load at the times I've seen it do it. Nothing like the large adjustments when the system is at low soc however.
Sadly I have had a reply saying I must wait a few weeks, not sure what you lot do to get this to happen :-( but really need version 909 as I want hot water and its our only source in the summer unless we light a fire.
THALL Nope, 17.7kw of battery at this time of year rarely go under 50 - 60% unless I dump it all into the car. Everything just seems smooth now.
Ah ok, so when was the last time you went sub 40%? I mean I could easily say there is no issue with my system since 3012 but I've not been below 40% yet.
mpartington Yes, it uses the Maps, vs calculations.
Over 50% calibration is being worked, as is the SOC to 0 on cell alarm.
This was the gamble, there was public pressure on this and other groups to release a 95% solution. Now the 5% are now finding their needs not met.
It is in work, but the inverter roll out is higher up right now
I think it was the right call. Move forwards and improve the last few bits after.
Plus getting the inverter updates out for hybrid in the sunny months seems the right thing to do next!
welshrat I emailed support about a week ago and was told no. But emailed last night and got a reply this morning and have been upgraded. Maybe just try again and explain your situation. You wouldn't have to be dealing with no hot water.
Ok I have emailed again. fingers crossed.
TheDragon (GivEnergy) one slight regression with 909. It does not appear to respect some of the values that are manually set.
For example battery charge power will see the inverter output at a higher value that the target set in the interface. 1140 set as the target charge limit, 1349 being pushed into the batteries.
Everything else appears to be working as desired and a few of my issues with startup resolved.
Arg0t OK so minor then. The portal is being re-worked for the Next Gen FW. To include scheduled Pause
TheDragon (GivEnergy) agree more an inconvenence but consistent if set and reported through the portal or through via HA integration.
Experienced similar thing yesterday where servicedesk setting 3.6kw as the max grid export resulted in only 3.3kw being exported, they had to set a higher value to overcome the overhead that appears to have been calculated in.
chris_p Ah last night because I left my zappi on eco so it drained a nice 1.5kw constantly. Dropped to 18%. It looks like the ovc ran, soc went from 18 to 19% well before the sun came out.
@TheDragon (GivEnergy) If something could be configured so it would hunt for capacity when the system is at 100% that would be rather nice! This time of year I expect a lot of systems hit 100% by mid day then sit there for 4 or 5 hours.
THALL On Sunday mine was 100% from 11.20am to 9.20pm, so yeah 100% hunt would be great as I bet there is missing capacity there by now.
chris_p To be honest as long as it finds or checks im happy. I saw it working when the system was low soc before I am pretty confident in it. My only remaining tiny issue should be fixed with the new AC firmware, the 2% at the top jump on the graphs. Actually apart from the 1 5% soc jump mentioned above I would say I am 99% happy with my GE kit again! Might even go as far as recommending to some people soon : : 😉
THALL 😲 Haha only kidding, once I see it for myself then I'll be happier too.
chris_p Chances are the bulk of us will not see it run until winter, when our systems drop to empty and are idle for a while. But I can show screenshots if that would help? I had one yesterday, hit 100% at 11:15, system dropped to 98% (no load at the time solar doing 4500w+) charged at 700w again for 5 minutes then at 100% it says it was charging at 145w until 12:00
Only issue remaining for me is the weird inverter behaviour when my system hits 100%. Great example here where it goes in to a 15 minute cycle of charging and discharging.
THALL No that's fine thanks, I've seen plenty of examples of it at work. I am tempted to force mine down to 20% with an export and leave it to idle and see what happens, just for my own curiosity and confirmation.
THALL Jumps like this? I've got a gen1 hybrid, not AC coupled
Ra Err what are you using for your load to go up and down in sync with your pv but keeping the top 1kw pv available? But yes I see those 2% blimps at the top, but not as many!
Not using anything, I have an iBoost as well but that has been in since the start and it never used to do it. In fact sometimes it doesn't do it at all....
A couple of days earlier -
judgepd I used to get that on my AC3, then I forced a charge using GivTCP. Never saw it again.
I assume this is the same as a normal forced charge, rather than on a schedule. I'd be interested to see what happens if you try that?
Jamesrich Just had a better look and over the last week it's done it less and less each time it hit 100% and today it didn't do it at all, it just dropped from 100% to 99% and then has sat there for 2 hours without charging or discharging...guess it's learned something maybe.
TheDragon (GivEnergy) definitely the right call to release. I suspect the 5% will be much larger as most people with oversized battery systems will be staying above 50% at the moment, so will be hidden during the good weather. Could be a spate of soc drop posts first bad day.
However, so long as it's already recognised as a weakness that needs addressing and further development is going on/planned (in reasonable timescales), I don't think there should be any real complaints. There are also ways to mitigate it, which there weren't with the old firmware, so in a much better place.
Thinking about it, if it's best to go below 20% (BMS) to peform an OCV, that is effectively empty.
mpartington Thinking about it, if it's best to go below 20% (BMS) to peform an OCV, that is effectively empty.
The problem I have with this is my SoC issue it missing capacity at top end. I'll get to 20%, the OCV kicks in and then drop until empty... How is that any use to me?!
I have the new firmware due to not getting any hot water and moaning a lot (Really sorry about that). I am over the moon. Normally after 12pm I would get nothing much happening. Since the update I have 10kwh into Eddi, piping hot water and over 3.5kwh exported and its only 3:50pm
Fabulous stuff
mpartington I'm going to be away next week so will be maximising battery export at peak rates without having to worry about drawing from the grid when the missus springs a load of washing on last minute. Mine hasn't been below 50% for a while now due to good weather so this will be a real test!
judgepd since installing 4071, I've had similar cycling at 100%, and like yours, after a few weeks mine also seems a bit better now. However, with lower target SOC (say 90%), when it hits that, it goes into manic cycling: charging at full rate, discharging to meet load - and at the same time seems to think the load is spiking in sync with this (there's no big fluctuation in house load, so I'm pretty sure this is a symptom); this (lower target behaviour) was happening on 4074 too, and still is on 3012. 
(8.2kWh battery with AC3 on D0.289-A0.289)
chris_p it means it’ll know it has more room for the next grid or solar charge.
andrew_bond pretty sure that charge discharge issue is known about and planned for the forthcoming inverter firmware update.
chris_p I'd have thought that drop should open up the capacity as it would then be operating between 4-100%, rather than effectively 20-100%, that's how mine seems to work. I forced a drop to zero from 20% just after 9am on the chart, it then filled the full capacity.
Worrying thing from my perspective is how quickly it seems to drift. I'm trying to do a manual ocv whenever there is a natural opportunity, something 99% of users won't do (and why would they think to)
welshrat Good to hear. I was updated this morning. Battery charges to 100% and stayed there and the solar generation was a nice curve throughout the day as you would expect rather than dropping off on a full battery. So much better with the new FW.
mpartington I suppose the difference is to open up capacity to fill (jump to 0% folk) vs those who open up capacity they didn't know they had (those with jump to 100%).
Yeah same here, seen the battery drop to 98% and charge a few times to 100% but it was doing that before this firmware so doubt its much of an issue. Everything else is working much better though, Eddi got 12kw, it Exported 4kw and the battery responded much better to house load this evening.
I can wait now for further improvements as they come.
chris_p what I don't really understand is how the battery works at top end. You would think it knows where full is, and if it starts to drift it would just reach 100% and realise it's not full yet and keep adding charge. Looking back, before my 20% downward adjustment, i can see my voltage at 100% had dropped. If I can see it, software can see it and be written to correct.
mpartington completely agree
Cells degrade/change performance with temperature and the full to empty window is small. If you calibrate at one temperature then you’ll need to base full on cell characteristics or you could overcharge at different environmentals
andrew_bond The oscillating at non 100% is yet to be addressed in AC3. Like it is in Gen2, soon to be Gen1 then AC3.
Other firmware engineer priorities are getting in the way, but i have weekly conversations, then ext being Saturday!|!!
kram The BMS tries to compensate for this, again its a calculation. As is everything with LFP its a best guess, unless you can get a true idle condition, which the vast majority wont have right now. The calculation window is being made bigger, so it can run over the whole SOC scale if the load is low.
This would never be a 100% solution, as there was too much public pressure to get it out to solve MOST of the issues. We could have gone back and forward 100 times, tweaking the map tables for each situation, but we dont have the time.
As a slow burn, this will get done
TheDragon (GivEnergy) Any joy with the oscillating at 100% or is this doing something that it's supposed to be?
I've dealt with commercial software my whole life and appreciate the hated question of "when will it be ready?" requires a politician to answer it. So I won't ask that :-) . So AC3 is the last on the list of current equipment and new equipment which now looks like months not weeks for real world delivery. What I require is the fast switching in the AC3 but support respond "wait".... All the other problems being fixed I don't really see because of continuous deep discharge/charge and only a single battery.... So question is "How to I entice support to update my AC3 with current fast-switch firmware?"
Thanks Paul (in advance).... :-)
BrianHere Send me your serial.
judgepd D288-A288 will address this, i am hoping, its in the list of bugs with 289
As others already begun to notice, the Gen2 FW is out and with support for case by case fixs, ie solar or battery issues.
Portal deployment is in the queue to put out
TheDragon (GivEnergy) Thanks for confirming
I have to say, even if it isn't perfect yet your firmware guys have done a lot with essentially little to work with on the BMS. I'm pretty sure the design hasn't changed since GE "birth" in 2018 (and perhaps even older if it also went in the Prime kit as well) so for what it now does on 5+year silicon @72Mhz is an achievement in itself.
TheDragon (GivEnergy) CE2223G689 [ how many pints you owed atm :-) ]
I Just wanted to show how much of a difference this firmware has made to my generation, take a look at the two graphs, the first showing solar drop off after battery full, the second is today's graph until just now.
Spikes at 7:30 am are my son using the electric shower. My wife and I curse this everyday, probably should be the other way around for a 14 year old :-)
TheDragon (GivEnergy) mines been idle at 100% most of the afternoon 😉
Good to hear it's on the list, my 100% voltage is dropping already. Looks like it needs an offline OCV every 4 days, but think I'll be lucky to achieve 1 a week with the good weather (now there's a complaint, too much sun 😆)
TheDragon (GivEnergy) unless you can get a true idle condition
And this I could never understand how any typical user is ever going to see this in the lower half for a sustained period unless it hits 4% limit. I'm either charging my battery, discharging it or it sits at 100% or 4%. It won't ever sit idle at 15%, when would it?
It looks like my inverter got a firmware update during yesterday to D0.909-A0.909, and the first I knew of any change was from the car charging session in the early hours of this morning. On previous days the battery has been going idle during the charge period and ignoring the charge to %age (with starting SOC > Target SOC).
This morning the first hour or so was drawing from the battery flat out until the SOC dropped to the target of 50%. Initially I thought the time clock was an hour out, but after checking the details and the reported firmware version now know the charge to % is now working as per the previous production releases before all the recent beta versions.
I did a manual override to draw down the battery to 45%, hence the brief spike around 0600.
Still on battery beta firmware 4073.
Can anyone confirm if a further battery calibration is required when this is updated to the general release v 3012?
Simon_C Pretty sure it isn't required if you're going from latest beta to production version.
welshrat the faster switching is a Good Thing. Just beware that the latest beta has a bug which causes the inverter to hunt around when doing an overnight charge. The other night mine consumed 3.5kWh trying to hold a target SoC. The work-around is an automation to set the charge and discharge power to zero when the target is achieved and then back to normal at the end of the off peak period.
TheDragon (GivEnergy) Paul, whats the latest regarding the updating of f/w for the Gen 1's that had a serial number below 2209 (I think) that need a engineer visit?
Is this still the case or, as I think I read somewhere in this long thread, a visit is now not needed.
Thanks in advance.
Any chance of a brief summary of whats been done, whats still to do, what systems have a fix, systems still waiting, those still needing visits and any other anomolies?
Stevef1964 Support have test FW.
Load TEST FW. Wait for the inverter to come back up, less all its logic, just a dumb unit. It will then report some errors or not.
This is the indication if safe to load the new GEN 1 or not.
TEST FAIL, load old FW
TEST PASS, Load new FW
BrianHere OK, loading, give it 20 mins.
No special registers you said?
Hmmmm ive rest them all back as they were. You are now on the 289 Beta. I will note the serial, so later on i can update it to 288 if thats next, or straight to 536
Thanks Steve, not charging from the grid at the moment and don't think we will be for a while so hopefully by then it will have been resolved.
TheDragon (GivEnergy) Thanks Paul. So I can gleen from this you are expecting some of the old Gen 1's to still need a visit. (Just hope mine's not one of them!)
So nice day forecast, lots of export expected. I decided to empty the batteries and do an idle OCV less that 20%. It idled for a bit, then did the adjustment in the purple period. No sudden drops, but it does seem to empty better if it's discharging at high power - maybe the limits are wider.
Always seems strange that the BMS reported SoC drops so low (this happens most times), given it's meant to be including lower unusable buffer in the figure. Certainly a resting voltage of 50.87 doesn't feel that it's gone into that 15% buffer at all?
mpartington When i get the beginnings of the next BMS, can I use you as an Alpha tester?
A few tweaks are in the pipeline, for an even smaller community
Stevef1964 Yes, a small number, but we CANNOT be sure, till 2209 onwards.
TheDragon (GivEnergy) I wasn't bothered about the Go settings or the 50% charge limit I put in overnight. Too simple to put back - hence "nothing special". [note to self: say "ok to wipe registers" next time]. Update when you want.
And 2 words Pauil.
Thank. You.
Any corporate forward facing department needs clones of you !! (a lot could learn from you)
TheDragon (GivEnergy) sure, no problem
Hi all, I've kept tabs on this discussion for a few days now as I had my system installed on 22 May. Its a 5KW Gen 2 inverter running D0.906-A0.906 firmware and a 9.5kwh battery with 3012 firmware connected to 2 strings of 5x410W panels with Tigo optimisers. 1 string faces east and the other west. I think its suffering an issue that this firmware is supposed to fix. 3 mornings out of the 4 the inverter appears to not 'wake up'. I've reset the inverter and power generation has immediately kicked in to life, going from sub 100 watts to several hundred/1kw. On the 25th I was driving to work and power had picked up by itself by the time I arrived. Power generation appears to approach 200-300 watts as the sun rises but then dies off, almost as though the inverter tries to draw too much power and the voltage drops which then seems to cause the inverter to 'go back to sleep' or possibly something happens with the MPPT process to cause a mis-identification of the voltage for maximum power production. As you will have no doubt guessed by now I'm no electrical.electronics engineer, I've just semi-understood the wikipedia page on how MPPT works!
Anyway, the important question from my perspective is should I hang on for this beta firmware to be out of beta and released for standard OTA update I can initiate via the cloud?
Lincs_Will Contact Giv support and ask them to install firmware 909, it overcomes this issue. I had the same problem and upgraded earlier in the week, it's been working perfectly since.
mpartington record 9.1kWh added after emptying earlier (2x5.2 batteries). Definitely tell the weather has warmed up (or maybe it did discharge a bit too deep 🤫)
mpartington Not still missing ~ 1kW?
chris_p they are unfortunately 80% DoD, so should be around 8.3kWh (although data sheet did say 90% when I ordered)
mpartington Ah makes sense, always forget about 80% DODs.
I had a good day of solar yesterday after a few good days apart from Thursday being not so good - so started the day with the battery at 4%.
No discharge (recorded) all day and put in exactly 8.2kWh to get to 100%.
Wouldn't mind trying the beta firmware test on my gen 1 3.6Kw hybrid - but not desperate and not over the weekend if that means Paul isn't out and about with the family!
Sorry yes I was saying you don’t always know where full is as full changes based on a number of factors.
Eg full isn’t easy either. 👍
Looks like on new firmware(3012) its even more interesting, as before, I had only jumps to 100% while battery is charging, now I get this while discharging at night, 
hmm...
One of my 9.5Kwhr batteries has taken itself off line sine 13.30 and shows 0% even though it was at 93% before it went off line. Bit annoying as the other 9.5 may not get me through to tomorrows solar.
Any ideas?
Roberto If you look in the web portal (not the app) do you have red badges? Do these show any BMS messages
ProximusAl out for bike ride will take a look when back....
No red badges. Turned battery off and back on. After 5 mins it was back.
No idea what was going on.
Been about a month, rock solid.
Last couple of weeks the battery has never actually reached zero, which is usually the point at which I can expect it to fail to zero if the next time it gets down to 20-40%... but nope. Last night, the water tank took the last 3kwh out of the battery before drawing from the grid and it was a nice smooth line down to zero.
Good work Dragon (and if there's a team: team).
I won't pretend that I'm happy it's taken 18 months from install to get a predictably functioning system, and clearly from this thread there's still some wrinkles with multi-battery systems - but clearly even for single batteries it's been a monster of a problem to fix and I can separate out annoyance at the product/company during that time from my impressed-ness at the technical skills required of the individuals involved in actually fixing it.
So you should definitely be giving yourself a massive "experience / achievement for my CV" pat on the back as far as I'm concerned.
Just checking my understanding of the new firmware after a SoC crash last night. As background, we had a new gen 2 inverter retrofitted to two existing 9.5 batteries around three weeks ago. Since then we've rarely been under 50% SoC due to good solar. However the last couple of days I've forced a discharge on Flux - yesterday's was to 35% leaving enough for the evening. However late evening the batteries started dumping from 23% and hit 4% over a half hour period. I ended up topping up a little on Flux night rate.
Is this due to not having cycled the battery enough please? If so, how many cycles before it comes under control and do I periodically need to force a deeper discharge?
The day before yesterday's crash I also saw the oddness shown below, again during a Flux discharge. However no crash this time and the SoC percentages did seem to normalise. Still some oddness happening it seems. Thanks for any guidelines.
looks like battery 2 in your display was playing all by itself that day. Discharging down while leaving battery 1 idle.
Normally only disconnects a battery on error. Checked today and they are back together playing happy families again
TheDragon (GivEnergy) Many thanks for checking. All looks OK at the moment. Should we have seen the SoC dump as noted in my post please or is this just the new firmware starting to get to grips?
So, this morning at 06:30 - 07:00 my SoC dropped from 40% to 4%. When looking at the individual batteries they are saying 15% & 16 %. It's now rising, but which do I believe the individual batteries or the SOC and I question if the Soc drop is really resolved.
Stevef1964 Similar to what we saw per my post and question a couple of posts back. I'm hoping it's just because we are only a couple of cycles in following the update. Are you cycling regularly to use the full charge range? We've not been due to good solar and quite a lot of capacity.
The OCV has done a self calibrate and based on voltage the true SOC was lower than the reported one, hence the downwards move.
Are you sing the newer Inverter FW too? When mated as a pair the best results are seen. BMS alone, will do OCV, when the opportunity arises, but doesnt do the varied charge/discharge powers
Paul, i'm not using the new inverter S/W as my inverter is one of the older ones that might need an engineering visit (SA2208G029)
I know you said there was a possibility to see if it did work but I didn't think it was on general release yet.
TheDragon (GivEnergy) Hi Paul, we are not on the new beta inverter firmware but happy to try it if that's still possible? We are on a gen 2, 5kw hybrid running 2 strings, approx 5kwp each, East/West. No worries if it's a case of waiting for the formal roll out. Thanks.
You have 909 already bud
Stevef1964 Customer services have the transformer test FW. They can load remote, test pass, load Gen1 new FW. Test fail, load old 499 version and engineer visit list. for future discussion
TheDragon (GivEnergy) Ah, sorry I had no idea; the installer who upgraded us from gen 1 a couple of weeks ago must have done that. Good!
TheDragon (GivEnergy) I'll get in touch with them. Thanks Paul
Had first anomaly today where generation stopped for 30 mins and I had to restart inverter, all ok since but thought I should post here in case its useful @TheDragon (GivEnergy)
Oh Dear!!! Just had an email response from Adam on the Help desk regarding the firmaware upgrade telling me that they do not have the new firmware and I should contact some chap called Paul!!!
I assumed the helpdesk and customer services are one and the same? But maybe not. If not can someone tell me what the customer services contact details are?
welshrat Ok thanks, I can see no reason for the PV to shut down for a few mins. If it happens again let me know, and we can compare today with the next occurrence
Stevef1964 There the same. Just many of them, some may not know, or know they could pass to 2nd line.
Ive sent them all a note, explaining the test firmware
TheDragon (GivEnergy) Lets hope Adam sees your note and decides to now act on it.
Stevef1964 he has, we discussed, i believe there is an option friday.
Adam has now responded. Wants to do the upgrade next week, which is fine.
He said my inverter would be offline for up to an hour.
Friday option is fine too. (He didn't tell me that though)
I did ask that he contacts me to agree a day/time and prior to starting. (Is that unreasonable?)
Can you tell me what I can expect to happen?
Will my batteries need a calibration?
Whats the process? Flickering lights, reboots, etc?
Knowing what I can expect to see during the upgrade would be nice.
Inverter will be flashed, with the test. It will become totally dumb, all logis stripped out.
The notifications section will be checked 20 mins later, if there are no errors, the test was successful.
New FW D195-A195 can be loaded and a note put in the commissioning section.
If the notifications throws loads of errors, the test has failed. This will be also logged and the previous FW will be loaded.
Serial noted on a database for engineer works.
The engineer side of things is yet to be decided how the logistics of this will run.
Tomuks and after 4 days after jump now straight drop from 40% to zero :/
So, no calibration of the batteries are required and no loss of charge?
Hi all. I've had the 909 inverter update yesterday (ED2248G186) at about 09:30 and the first nightly run after the upgrade, things are different....
I have a timed charge from 23:35 to 05:25 with a battery target SOC of 20%. So normally, before the upgrade on 906, the batteries decay to about 60-80% to meet the house load overnight, don't dip below 20%, don't charge and no grid use takes place overnight. After the 909 update, the batteries were at about 88% before the timed charge last night. Last night (first time on 909), it stayed at 88% (end of day SOC) and used grid until the end of the timed charge.
I remember timed charge was a point of contention; would the new inverter firmware behave as before, or would it hold charge and use the grid if the SOC was > the target SOC. Last night, for me, it appeared to do the latter. I thought the new inverter firmware was going to continue doing as before?
Image 1 shows 906 normal night behaviour before upgrade (00:00 - 05:25) and the start of new behaviour after 909 upgrade (23:35-00:00).
Image 2 show the continued new behaviour last night after the 909 upgrade (00:00 - 05:25)
In both cases, the timed charge is set the same - 23:35-05:25 target SOC 20%, ECO mode set.
4monks Thats strange as this exact SOC > Set % was tested and functioning correctly.
2 testers tried this.
Granted it was with as Set Level of 50%, but shouldn't make a difference
I'll feed back for checking again in the lab
Im going to factory reset your inverter, to get the same lab conditions. then put your charge settings back in
If it works tonight properly , then the 10 charge and discharge registers, newly added are getting in the way, as these are not set yet, nor accessible in the portal.
I may need to set these all up to some values.
TheDragon (GivEnergy) Thanks Paul. Here are the last few days GivTCP battery data - you can see the normal night decay before and the first 88% hold after the update to the right.
4monks perfect, so good results to compare to. see tonight if it changes. As a full hard reset was what we did in lab and beta tests. Roll out was a set of registers that needed setting , quicker and easier, but maybe there are more affected not listed by the engineers
I’m not sure if it is related to the new registers but I am seeing occasional spikes in export to the grid at around midnight on firmware 909 when a timed discharge is set during the day (evening in this case). The inverter shouldn’t be importing or exporting to the grid, it should only be supplying the house. It may be getting confused at the clock change from 23:59 to 00:00 or possibly a start and end time of 00:00. It doesn’t appear to always happen.
TheDragon (GivEnergy) Thanks Paul. Just to confirm, you've now reset it and set the values to the same as before?
Hi,
I have two 8.2 batteries hooked up to a 3.0 AC inverter.
The primary was installed last August and the Secondary last Wednesday.
Commissioning went well and everything has been running smoothly since ... until today.
My primary seems to have dropped its SOC down to zero from 100%
My primary battery is DY2219G082, on 1%
The secondary, on 100% is DY2218G702
Both batteries are on 3012
The primary has run smoothly since installation last August but it did experience 2 SOC drops to zero last October, running on 3007.
Can you please take a look.
I have tried to upload an image from imgbb but was stumped by the link. I tried them all!
Thanks
Rob
Rbor I have tried to upload an image from imgbb but was stumped by the link. I tried them all!
This forum has been updated to allow pasting into the reply dialog directly.
Thanks Tim,
This should be the image for my problem battery DY2219G082 Battery SOC
Rob
Hi All - I'm upgrading mine tonight (00:00 cheap rate period!) due to regular drops from 35% to 0% - currently have Invertor Firmware D0.534-A0.534 and Battery 3007.
Any advice before I go ahead?
Thanks Dave...
TheDragon (GivEnergy) Hi Dragon, I have an issue with my system using grid power when the battery has plenty of charge, cannot force discharge the battery neither. Could you take a look, please. EA2302G008
Re - the upgrade: I started the upgrade but it failed at 67% - I've emailed support to see what my next step is...
@TheDragon (GivEnergy) Congrats on all you have achieved so far, it really does seem like a night v day change
I would almost certainly be in the 5% that would still suffer from drops due to cycling above 50% most of the time. I appreciate it's now not top priority, but I'm not going to commit capital until I know this is solved and can calculate a reliable RoI. Any idea how long I will need to push back install date? I would like to install a GE AC system, but at this point it feels installing before all is fixed is buying potential hassle I don't need, which frankly seems dumb. So far there is no pressure from the installer, but I will walk vs install if I am pressured. This is in no way a criticism, just a perspective from a not desperate customer who has been watching from the sidelines for more the 6 months and wondering what to do
TheDragon (GivEnergy) Hi Paul - I thought it was back to "normal" behaviour last night i.e. target SOC 20%, SOC at 81% at start of Timed Charge, no grid import during timed charge, battery use down to 71% during timed charge period; working as expected.
However, I noticed the "Timed Charge" toggled was in the off position i.e. there was no timed charge active last night. I have reset the values to defaults, set all the values to the same as before the upgrade and will try again tonight.
Mostly - I'm not close to flattening my battery at this time of year, but when we do I find I'm missing a Kwh, I assume at the bottom.
Is it best to set a timed pause. Perhaps for 30 minutes between 23:30 - 00:00, when off peak. Will this be sufficient to allow the OCV to run? Did it have to be under a certain percentage for this to work?
I can then kick in the charge from midnight.
P.S I have GivTCP, so have the option to configure this.
davlon Do in the morning, when its going to be sunny, cost you less
Jamesrich Timed pause while under 40, better under 20, where it uses actaul cell maps.
After a few weeks of operation in the top 50%, I tried a deep discharge today to see how the displayed SOC compared with actual. The battery discharged down to a displayed 26% SOC and then throttled its discharge rate to 103W. After a short period of throttled discharge the SOC "realigned" itself down to 4%.
It would appear that (as suspected) the top end 100% SOC alignment on its own is insufficient to keep the SOC reading correctly. (Single 8.2kWh 100% DOD battery, 3012, AC-3.0, DO.289-AO.289).
It could be that the 100% SOC alignment is just inaccurate or possibly the 100% SOC point is being determined correctly but the constant small discharge registered by the inverter (while the battery is idle at 100% SOC) is upsetting the tracking of the SOC.
Will be interesting to see if the AC-3.0 firmware update improves things.
Woodstock Pretty much exactly the pattern mine followed. Infrequent discharge below 50% at the moment but a forced discharge caused a dump from low 20s to 4% unexpectedly. Still not sure how to address this - much lower discharge once per week?? Needs sorting though as still not right.
Sorry - quick check - I've been on the beta for a while, just checking which version we're on and if I'm on latest version:
Inverter : D0.909-A0.909
Battery: 4074
Many thanks, just being a bit lazy and not scanning through 100s of posts as been out of the country for a couple of weeks!
Woodstock yep, I'm worried about that too. Had one big drop since moving to 3012 (think it was 50% to zero or something like that).
I've tried to hit 100% with solar on a few days
Plus trying to use the battery (e.g. put on an evening wash load and/or by setting low timed charge percentages) some days.
Not seen another crash yet. Also not seen any lower half soc adjustments either.
Keeping fingers crossed.
I've gone back to prod firmware too (AC3). But willing to try the new AC3 beta when it's ready.
alphabeta77 I think you're on the latest prod firmware for the inverter. Very new firmware. To fix solar issues.
Plus beta firmware for the battery. Email giv support and ask them to swap you to 3012 BMS
But perhaps do that when the battery is almost empty in case they decide to do a calibration. It shouldnt be needed but you never know.
I tried another OCV this morning. The batteries were paused at about 06:00 with the secondary at 28% and the primary at 20%. The primary SoC climbed slightly to 23%. The secondary did a big adjustment, dropping to 21%.
At 08:18 I took it out of Pause and started to charge it via PV.
The Primary climbed steadily to 100% at 57.4V.
The Secondary lagged behind, climbed to an indicated 89% at 55.6V. The SoC then jumped to 100% peaking at 57V
Does this look right for a Pause?
Still shows discharging at a constant 107w.
Ac3, 289 firmware. 9.5, 3012 firmware
Jamesrich it's a bug in that version of beta firmware. Waiting for next iteration of AC3 firmware, which will handle the pause separately to using discharge rates.
Current workaround is to unselect eco mode, or you can set the reserve to the current soc.
4monks Paul - this has been ok for the last two days after a full reset and setting values again. One thing I have noticed is that at the end of the Timed Charge period with 909, there is a very brief "blip" at the transition from Timed Charge to not Time Charge (even when not charging) where the data indicates the grid is being used rather than the battery.
I've observed this each morning since the reset. It's as if the end of the Timed Charge event causes a very brief non-battery use. As there is little load at that point in time, it is barely noticeable.
4monks How strange. I will keep an eye on this, to see if there are any more occurrences.
I will begin gathering these minor bugs fort the next release
I will look to collate bugs for periodic releases once the new FW is out for everyone. I have found a few blips in other areas that need a tweak too.
I was hoping a beta pool of over 150 would bring most issues out, but seems not, but didn't help than maybe 10% of the testes actually ever sent anything back. Free loaders cant be helped, but doesnt help my testing neither.
What I have seen so far.
BMS 3012 is 80% ok on inverters with previous version of FW. 95% with the new FW.
Most issues are around remaining above 50% for days on end, allowing the SOC to drift as it did before, going un checked. Work in progress for this BMS 3013. (Allowing the self check calculations in the 20-80% range) Below 20% is the OFFLINE MAP based check.
New inverter FW allows top end topping off, old does not.
Gen2 909 - Needs a grid reset post update (One off), proving difficult to automate, setting the most affected registers is good for 99% of cases, easier to roll out. Those with odd behaviours after, service desk can easily reset grid code fully.
This introduces 10 charge and discharge registers, if these have any garbage in them, it causes these weirds symptoms,
AC3 - Version D289-A289 is the oldest beta, has AC charge oscillation still present, which needs fixing like Gen2 and 1.
- Expecting at least D288-A288 before final release of D536-A536 later in the month.
Gen3 - D303-A303 - Based on Gen2 (909-909), but compiled for the Gen3 CPU
AIO - D605-A605 - Current version. with a minor issues with reporting of meter data. Fixed in A606, not yet released.
Awaiting ENA to put the G99/G100 compliance on their portal.
Think that's all i know of on the FW. Have a happy Sunday.
All the above have at least the ability to PAUSE the battery, either charge or discharge. Gen2 and 3 also have this with timings, so can schedule a pause. Gen1 hasn't the register space for the timer. But the Portal automations will be able to schedule it
Can I ask which software team works on the AIO and gateway firmware?
Is it the same team as the beta firmware or the same team as the original firmware?
TheDragon (GivEnergy) maybe 10% of the testes actually ever sent anything back
I often felt that repeating what had been reported was wasting someones time but should beta testers report repeat faults?
kram Same team that is on the new series of FW.
So the AIO will have similar operation to the Gen3, just bigger and HV BMS instead of a LV. So the BMS is totally different.
Currently on prod AC3 firmware (CE2133G030), but happy to go back onto beta when the next beta release is ready for testing.
4monks I too have had these grid import blips a couple of times in the last few days. The house load is low at the time but in both instances the solar had just started prior to the blips.
CCourtney What time did you timed import end? Looks like 03:30?
4monks 03:30 on the 1st and 03:00 on the 3rd. In my case I think more significant is the coincidence with the first solar appearing
TheDragon (GivEnergy) This introduces 10 charge and discharge registers, if these have any garbage in them, it causes these weirds symptoms,
Oscillation of the battery when full,
Hi Paul, If i'm reading correctly Oscillation is a sign of garbage in the registers? I think you kindly cleared mine the other day after updating me (EA2302G822) to 909, but am still seeing oscillation. Sorry if the fix is still being worked on 🙂
thewhalw21 Strangely the oscillations seem to settle down after a few weeks. I was getting them every day for a couple of weeks but in the last week I've had 3 or 4 days with no oscillation.
Dw Yeah, Mine are very intermitent, some days its completely fine, other days theres loads. Dont envy the devs who have to fix this one.
Dw Mine have never gone away all but for around 4 hours after 909 was installed Paul says its know and they are working on it so waiting for next update hopefully brings a fix.
Morning.
I have a BMS test version looking to allow SoC alignment upto 80%. Hopefully to help those who always remain above 50%.
I would like half a dozen if possible to test. It is the same as 3012 but with the above.
Version 3096
AC3 with prod inverter firmware (D0.535-A0.535) willing to help. Had one soc crash post 3012
CE2133G030
Presume no need to recalibrate if just a BMS update? 🤞
Cheers.
TheDragon (GivEnergy) In the summer months I dip down to about 40-50%'ish every day.....but in the winter its a lot lower? Not sure if I would be of use or not?
TX200 ibdont think so.
But will double check
TheDragon (GivEnergy) so can I infer from that the Gen1 is pretty much at the limit on major FW advances from now on (that take up register space) other than what can be done remotely via portal?
Basically short of some register culling it’s Gen2 onwards that will be the main focus (although I’m not entirely sure what else could possibly be added!)
I knew that Gen2 processors changed from Gen 1 but Didn’t realise the processors changed across 2nd & 3rd gen as well. Handy to know.
Ra have you seen a dip in SoC while living over 50%
Then yes.
Next few days are a big grey for me so useless me trying it
TheDragon (GivEnergy)
I do not fit the parameter you specify as I drop to circa 35% and do not have any anomalies but happy to test if it helps.
TheDragon (GivEnergy) Older than 2209 ? Can I easily check if mine is older than that ? Also thanks for all your hard work and keeping people informed
TheDragon (GivEnergy) If there is no calibration, happy to take part. ED2248G186. 2x9.5. Typically >60% atm.
TheDragon (GivEnergy) Haven't seen any SOC issues (touch wood) yet....only niggle I have on-going is a perceived slight reduced capacity of the battery.
I've been away on holiday the last week so made most of exporting battery to flex the flux rates so battery got full cycle more or less for a week. Didn't get anywhere near 8kWh output to grid but hoped this cycling would improve the issue? Don't know if you remember but when we updated my firmware it got stuck and the calibration didn't complete, would this be the cause?
TheDragon (GivEnergy) yes please, I had a 30% drop last night after being away for a week
TheDragon (GivEnergy) Happy to give it a go. Not likely to go below 50% before October unless I force it. 🙂
@TheDragon (GivEnergy) Also happy to try it out if a calibration is not needed.
Our batteries are tracking a lot closer than they used to, but there is still a difference at the top & bottom of the curve.
The secondary battery never reaches 100% when the system goes idle at 100% and it appears to always be at a slightly higher percentage than the primary when at the bottom of the curve.
Therefore, the secondary does not appear to be using its full capacity.
Will the new beta software cure this or will we need to wait for the AC3 updated software?
As usual, you will need to send me the beta bms to upload manually. CE2105G137
positor1 should beta testers report repeat faults?
In my experience of beta testing, yes you should. It allows for fault <> system correlation, such as problem 'x' only seems to happen for hardware 'y' in scenario 'z'.
TheDragon (GivEnergy) Not sure if you have enough peeps now, but happy to be updated and test.
Example below over the first 5 days of June, with the SOC going down to 5% a couple of times, but still the difference between the two batteries.
Woodlands curious if it fixes that as well, I've a constant 10% split in the upper half. Either battery can be the highest (usually switch position following an idle.ocv). However I suspect that would need some additional code at the 100% point, and this only goes up to 80% (hope I'm wrong). Even yesterday's 'realignment' from 30% to 0 hasn't closed the gap
Woodstock OK i will load it for you
Woodlands If you aren't seeing SOC issues when always above 50%, I wont load this for you,
Just looking for a few affected existing testers
chris_p Are you seeing any SOC issues over 50%?
TheDragon (GivEnergy) Not that I've noticed, the amount going in vs % difference is often showing reduced capacity so I suspect SoC drift is occurring.
chris_p OK, as you have a Gen1 on older FW, its a useful test. Updating now
TheDragon (GivEnergy) ok dokey, lots of flashing of the green lightening bolt and wifi dongle looks like it is in disco!
TheDragon (GivEnergy) Was about 11% delta between the batteries. Varies between this and ok. Just checking if you are loading as I just notice master has just gone to 1%.
4monks All good - I see v 3069 on both now.
4monks Thanks, i was still waiting for the portal to update, but you saved me sat waiting.
Thanks.
Please give a week of testing, if you know the conditions before to trigger the drop, please try and repeat.
@chris_p your BMS is done too, see if on older FW it too can assist. Then Iwould like to test your inverter for upgrade
TheDragon (GivEnergy) Then Iwould like to test your inverter for upgrade
Ok dokey, let me know if you need me to do/test anything.
TheDragon (GivEnergy) , so just to clarify how this new firmware is expected to work. As it only goes up to 80%, then we will currently see no benefit when the batteries idle at 100% full - the most natural opportunity (hopefully a future enhancement). So is the alignment method currently relying on user intervention to pause the batteries below 80% , or does it run at low powers (presumably for most people this would either be in the morning where solar generation is similar to base load, or at night if house has a low baseload? Is a 30 min stabilisation window required before anything will happen?
mpartington It runs at low powers. like the under 50% one was designed to do and gradually keep it in check.
time will tell if this works, the BMS engineer as spent a few days testing himself, so we will see.
chris_p let me know if you need me to do/test anything.
For now ensure the SOC remains accurate while above 50%. To check it, you could pause check the voltage and un pause compares to the tables for LFP. in the rough ball park is good. drifting away is no good.
Not sure what happened here but noticed the battery was pretty much flat at 6% (it doesn't seem to go any lower than this) 5:30am grid kicked in as battery was dead then out of no where before solar gen battery % jumped to 15% ? This is the first time I have noticed anything like this ? currently on 909 and 3012.
After the 909 upgrade :-) ED2248G036
Is this the oscillations you refer to ? I am seeing it when the PV hits the 5kW limit, and the battery is full. Battery being cycled on a regular charge discharge cycle ? Call to support desk ?
Comparing with 906 below, my slow start is cleared, and the string1 limit at 3.75 is cleared. No help on the hard shading in the evening though, but maybe I was being over optimistic on that :-). I still expected the bypass diodes to be activated, and mptt to have reduced voltage at same current, rather than the voltage be held, but the current collapse.
Leon_shaw That looks like the OCV process where the battery makes a self assessment of voltage/capacity etc and resets the SOC appropriately.
justpassing Is this the oscillations you refer to ? I am seeing it when the PV hits the 5kW limit, and the battery is full. Battery being cycled on a regular charge discharge cycle ? Call to support desk ?
No need (Just wasting their time), Its on my minor bug list. It keeping the battery online and awake, but removing the hysteresis to stop the lazy 95-100 behaviour has caused this. Future it will be better optimised. Not everyone see's it, of the 40 Gen2 betas, 2 had it, so deemed minor and we will live with it till the dust settles and we can re-work it
On the Bypass diodes, are you sure there activating? As that would indeed drop the voltage by one full panel, unless half cut, then the current will drop by one half panel
Re: Oscillations. Is what I am seeing different from ? Just on with support at the moment asking for a "grid reset" so might be wasting everybody's time.
Gen2 909 - Needs a grid reset post update (One off), proving difficult to automate, setting the most affected registers is good for 99% of cases, easier to roll out. Those with odd behaviours after, service desk can easily reset grid code fully.
This introduces 10 charge and discharge registers, if these have any garbage in them, it causes these weirds symptoms,Oscillation of the battery when full,
This is probably off topic for the thread, but I would appreciate feedback on my understanding.
https://www.solarquotes.com.au/blog/half-cut-solar-cells-panels/ and https://www.recgroup.com/sites/default/files/documents/whitepaper_twinpeak_shading_properties_eng.pdf
In my case:
The Bypass diodes look to me from the graphs to never activate. I thought that that was a combination of the diodes and the inverter, with the mppt effectively choosing the best outcome from maintaining current and bypassing panels/losing voltage, or keeping all panels running at the current of the worst affected panel. So its the inverter that ultimately causes the diodes to bypass or not.
Bypass Diodes: Assuming a single string, with half cut panels, Peak output current is 10 amps, with 2 half columns of cells passing 5 amps each, 3 columns per Panel.
My understanding is that where the shading only affects half of one "column" of cells ( so typically 1/6th of the panel).
1) Drop the current , maintaining the voltage, the diodes do not cut in, whole string drops to the output current of the shaded section. You lose most of the strings generation.
2) drive 5 Amps, where bypass diode doesn't kick in, and 1/2 the column is still able to pass 5 amps, lose half the string power.
3) Drive 10 amps ( actually sink 10 amps) and one of the diodes will cut in, slightly dropping the voltage, but leaving you with most of the generation.
Looking at my results,
As the hard shading hits one corner of one panel, I see case 2.
As the shadow covers a whole column, I see case 1.
I do not see ( 3) the diodes cutting in, and the current being maintained
So when will 909 be released to all ?
TheDragon (GivEnergy), are these the described battery charge/discharge oscillations when the battery is at 100%?
Currently on inverter = 909, battery = 4073. Also should the battery be 3012 production release?
Simon_C they look like the same oscillations that I have been experiencing. Not sure if 3012 is out sure Paul can clarify for you.
Simon_C 3012 has been out a while.
Already minor tweak to SoC self calibration.
Those on the beta 4073,4 would have had an email to upgrade to 3012 manually.
Or you can go straight to the next iteration with wider SoC calibration limits of 20 to 80%. Under 20% is by offline MaP.
3069.
Had a downward alignment from 72% to 57% as the sun came up this morning. Waiting to see what happens next. At the moment the SOC is sticking at 57% while the battery is charging at 700W+.
Had an upward alignment first night after 3069 install.
4monks OK, top end adjustment was the top end self reset, hit the upper limit see the spike in voltage.
Not OCV, but a correction
TheDragon (GivEnergy) The total SOC now reads > either of the individual battery SOCs (73% total vs 72% & 71% individual). I have seen this before 3069 though.
TheDragon (GivEnergy) 3069.
Not looking good. Battery now charging at 1.4kW+ (54V) and the SOC% is still stuck at 57%. Doesn't look like 90%+ throttling is going to get a chance to kick in so I am hoping over voltage will catch it. If not I am going to have to shut down.
Battery now charging at 1.9kW+ (54.64V) and the SOC% is still stuck at 57%. Got to take the car for its MOT now so I am going to shut down as I daren't risk what might happen while I am unable to intervene. Will report back in a couple of hours when I turn on again.
Woodstock over voltage will prevent over 57.5v, wonder why Soc is stuck
TheDragon (GivEnergy) When I turned everything back on again (2hrs 10 minutes later) the battery SOC% had dropped from 57% to 53% and the stabilised battery voltage was showing 52.75V. This is the highest stabilised voltage I have seen since being on beta and would equate to a displayed 100% SOC in normal circumstances. The system went into a throttled charge (249W) and the displayed SOC% rose slowly over the next 1.5hrs to 62%.
Thinking that the throttled 249W might not be low enough to initiate a further top half SOC% alignment I reduced the charge rate manually to 81W and left it for a further 45 minutes. Nothing happened other than the SOC% going up a further 1% to 63%.
I have now put the battery into idle and will leave it to see what happens. As far as I can tell the battery SOC% is under reading by a considerable margin as the voltage looks to be stabilising around 52.7V.
I will give it another hour and post the dashboard graph for the day around 17:00.
3069 day 1 graphs. See previous posts today.
Looks like the SOC % would have been closer to reality without the downward alignment at 07:25.
I will resume ECO mode when the sun goes in and (given the current 63% displayed SOC) I would not be surprised to see a lower half upward alignment at some point overnight.
@TheDragon (GivEnergy) I’m having my solar array increased on Friday and my installer is replacing my earlier 3.6kW hybrid Gen1 (SA2142G078) for a 5kW Gen1 hybrid (SD2247G111). Do the 5kW Hybrid Gen 1 inverters require any future hardware modifications for the new firmware?
Thanks
cbowen the replacement is compatible
TheDragon (GivEnergy) Brilliant. Thank you.
any news when 909 will be released? to fix battery discharging issue. Currently battery discharging 200w to grid as well as 1.5kw from solar.
3069 day 2.
Battery was displaying just under 50% SOC (way too low) at sun up today and had not performed any alignments while in the 50%-80% range overnight.
I set a rapid timed export going which stopped when the displayed SOC reached 6%. A lower half adjustment brought the displayed SOC back up to 16% (still way too low) and I then put the system back into ECO mode.
The battery charged in line with available PV (plenty) until the displayed SOC reached 71%. At this point the charging throttled back to 248W. The battery continued to charge at around 250W until the displayed SOC reached 76% (55V) when the displayed SOC jumped to 100% and the battery went idle.
I believe the displayed SOC% is now correct, but it took a sudden jump to get there. No evidence yet of upper (or lower) half adjustments keeping the SOC% in line with the true state of charge.
TheDragon (GivEnergy) For now ensure the SOC remains accurate while above 50%. To check it, you could pause check the voltage and un pause compares to the tables for LFP. in the rough ball park is good. drifting away is no good.
This is my effort today to pause for 30+ mins at 60%, 70% and 80% (70% I missed so put the immersion on for a bit, hence the spike). It looks pretty consistent, but no sign of SOC adjustment, is that still a sub 50% thing? I will try to run it low tomorrow and force an adjustment.
Woodstock do you have this as a graph, pictures speak better than words
@Woodstock OK, sent all this to the code writer, to see what they come up with, no doubt they will also have a nosy at your data direct too, to look for patterns. Thanks for assisting
chris_p the 60% looks like 50%
70 65%
80 70%
SO not that far off from the stock table, which isnt temperature or cycle compensated
@TheDragon (GivEnergy) My system stopped Gen again for an hour today, no reason I can see, grid voltage all good.
Here is the graph below, I didnt do anything this time and it brought itself back but thought you should know
TheDragon (GivEnergy) Yeah doesn't look far off, hopefully I'll get to 30% ish tomorrow for further low end testing. At what point should the OCV kick in?
100% voltage is considerably way off what the LiFo chart indicates, closer to 70% @ 52.8V as opposed to 100% @ 54.4V.
At 100% idle, my 8.2 shows Inverter: 100% @ 53.15V, Battery: 97% @ 53.5V.
Dunno which one is right, but the battery graph never goes above 97% when inverter is at 100%.
I'm still not sure what I should be seeing with the new battery and inverter firmwares. I've cycled three times to sub 30% charge including one this morning sub 20% and with the batteries paused. At this point one battery fell from 8% to zero, the other didn't crash and held 19% when I stopped the force discharge.
I expected the charge % to converge but it didn't and I'm currently charging with one battery showing 38% and the other 19%. Voltages are 53.4 and 52.9 respectively. Is it working or does it need more cycles - confusing.
sponev 38% @ 53.4V doesn't sound right to me. When were they last calibrated by GE?
Edit: Just realised you said charging, so that voltage may be accurate. I assumed IDLE.
chris_p That looks pretty consistent with what I see on my system. Stabilised displayed 100% SOC voltage was about 52.8V pre beta and is around 52.6V post beta.
Not sure if the voltage chart shown takes into account the buffers at the top and bottom of the operating range. A displayed 100% SOC is 100% of the useable range but a lower percentage of the total battery capacity.
Woodstock Not sure if the voltage chart shown takes into account the buffers at the top and bottom of the operating range. A displayed 100% SOC is 100% of the useable range but a lower percentage of the total battery capacity.
Ah yes, good point. I've rarely seen anything over 53V at rest.
the chart is TRUE 100%. Our batteries are configured not to get to this normally. 95% is the target SoC. Lower is 15%
The useable is within this 80% range
for 8.2 and 9.5 this is the useable.
TheDragon (GivEnergy) I understood the buffer, but my idling voltage at full is also around 52.8V (good we all have similar results). Doesn't appear to change dramatically with temperature. The voltage chart however suggests this is only around 70% full, so even accounting for buffers it still seems on the low side.
From what you're saying 95% would be around 53.4V, don't think I've ever seen it that high and stabilised.
Confusing thing, despite all of this I do seem to get the advertised charge capacity into the battery.
mpartington Just to add more information, my system has been idle at 100% for several hours now and the 2 battery voltages are 53.73 and 53.697.
judgepd More recent calibration? Newer batteries? Maybe even different cell manufacturer?
Latest 3069 SOC and voltage.
3069 day 3.
After yesterday's sudden jump to 100% the battery stayed within the 80-100% displayed SOC range overnight and all of today. As a result of this an upper half adjustment was not expected and did not occur (good). Stabilised 100% SOC battery voltage has picked up a bit to around 52.7V but not sure if this is down to 3069 or possibly a change in battery temperature.
Will try and use a bit more power overnight tonight so that the battery is between 50% and 80% when the sun comes up tomorrow. Want to see if there is a repeat of the substantial (erroneous?) downward SOC% adjustment that occurred on day1.
ProximusAl Yes that was when charging. The batteries were full by noon today on solar and peaked at 53.4 and 53.7 respectively. The reported SoCs have also converged to within a few percent so it does appear things may be tightening up. I'll force a few more deeper discharges to see how it progresses, hopefully those smaller but still evident bottom end crashes go away completely.
judgepd makes you wonder how much use that table is. Looks from the high peak that I had a top point 'calibration', so unlikely to see it any higher at 100% for me. Must be different cells in the 5.2s
A little clouded by my change in use (changed to flux), but positive signs. Batteries tracking together as close as they ever had. Hopefully stays this way.
So are we ok to add a few more 3069 test cases?
I will put it out on FB for another users, as no more takers on these means
Hi Paul
Happy to give it a go. Typically not using our full 9.5kwh capacity at present, usually have around 40% left at the end of the day.
On the odd day when we do need it, seem to be missing 1kwh. When we were regularly maxing it out about 6 weeks ago, it was slowly finding more capacity and improving bit by bit. Seems to have drifted now.
CE2225G411
Cheers!
TheDragon (GivEnergy) yeah, install it here : CE2133G030
(Guessing no calibration needed. If there was, now might not be a good time to start it - heading towards full battery)
Our two batteries appear to be tracking quite closely now, after a few cycles of full to 4%.
TheDragon (GivEnergy) Happy to try the latest version of the BMS and/or Inverter software. CE2105G137
TheDragon (GivEnergy) Happy to try here CE2105G039.
TheDragon (GivEnergy) are you still only looking for those that ALWAYS keep battery SoC under 50%? Ours is usually below that level overnight but the level of solar currently means in daylight hours is rapidly goes back to 100% until forced export for flux, so not sure if our usage scenario is one you want ?
TheDragon (GivEnergy) It is the same for me. I export in the evening during the Flux period to make space for excess solar as I have a G98 DNO approval. The SOC level sometimes dips below 40% but not always. I haven’t spotted anything that resembles a SOC adjustment on 909 / 3012.
The system is a Gen 2 3.6 with a single 9.5kWh battery.
Is that any use?
I mentioned before that after 909 upgrade the firmware didn’t fully honour the set limits on battery activities. Set a charge limit of 1100 and it charges at 1400, similar experience with all charge values.
Noticed this evening it also applies on battery discharge, 3600w limit set on discharge yet it discharges at 4017w. Either that or the portal and modbus fib, which is possible as the numbers in the screenshot do not sum out.
Any risk pushing draw 10% higher than spec?
Arg0t 3.6kw AC.
Be about 4kw DC depends on grid voltage fighting against.
Would that explain the roughly 200w gap in the overall sum?
Grid looks ok, temps right up though.
TheDragon (GivEnergy) Very happy to go on the beta - 5.2 and 8,2 batteries on AC3.0 inverter - serial CE2031G070
TheDragon (GivEnergy) Happy to test, SD2142G119. I normally have mine set to charge to 100% during Go hours, but I can change it to 70% for overnight charging to see if it does any SOC adjustments.
Now on inverter v909, and battery 3069.
It seems like my broadband router needed a restart today around 1400 so I expected some missing data from around 1330 with a straight line between, but looking at the My Inverter - Meter Data it looks like the generation today and total values went backwards at that time along with a period of logging at 10 second intervals before returning to the usual 5 minutes. The import and export readings continued to increment as normal.
I thought my generation today was a bit low for how much sunshine we had today. No new firmware updates as far as I know.
I guess I will add the backwards value of 9.2 kWh to the total for today.
Any ideas why/how?
TheDragon (GivEnergy)
I'm happy to try it. CE2051G191
My 8.2 and 9.5 vary from a 2% difference to a 25% gap. The top end seems to be close now with no jumps and stable but the middle on the way up and down is all over the place.
TheDragon (GivEnergy)
happy to , as i said before I( dont think I fit the scenario though. 35% overnight is about the lowest, 100% most of the day till export on flux.
3069 day 4.
Middle of 50-80% SOC range at sun up and there was no repeat of the day 1 large downward adjustment. System behaved all day and is back to normal summer operation (staying in the top 50%).
The sudden jump to 100% on day 2 seems to have realigned the SOC% correctly and it is now a case of waiting to see if I can identify a top half adjustment taking place over the next week or two.
This release is a difficult one to test as if it is working correctly there should be little or nothing to see. Whether or not I see any top half adjustments; the acid test will be to force a full discharge after a couple of weeks to see if the displayed SOC% has kept in line.
Updated my 8.2kW (with AC3.0) to 3012 yesterday. Picked a (very) sunny day and the battery was at 55% when I started. Commenced at 0830 and it was all over and recharged by 1.30 pm. Used about 2 kWh so very affordable.
Will I notice any difference....?
👀
Simon_C
I think the reason Import and Export totals were not affected is because these values come from the EM115. Generation and Battery figures come from the Inverter, so I think the fault lies in the inverter firmware.
@TheDragon (GivEnergy) As promised I pushed my battery down today and got some interesting results:
Discharge showed at 4.6kW which seems bit low for what the % from 0:00 (76%) to now but not too bad. Excitingly it looks like the OCV ran at 8.25am and added 6% from 10% to 16%. But bit surprised it went up as my issue is crashing to 0%, so expected to say it was too high (100% should be 80% say). Left it idle for 1h 30ish and it has stopped adding capacity.
Portal shows 6.1kW charge to take it from (OCV adjusted) 16% to 100%... I make that 84% added, so (6.1 / 84) = 0.0726 * 100 = 7.26kW. Seems a big chunk missing, yes my 8.2kW battery is now 18 months old but would is be missing 1kW already?
Edit: Just want to add that I was pleasantly surprised that as the battery was exporting at full tilt, that it got to 10% when I paused, no crash to 0% - so that was very good to see. I appreciate a recent calibration would likely have achieved something similar without any new firmware, so I'm hoping it isn't just that which everyone is seeing and really it is masking the problem.
In case anyone else hasn't been paying attention, OCV = open circuit voltage
Ref the Dragon's post on 4th May, approx #2346
Another one for the notebook...
wrefordj Yea. When idle so not interrupting your use. If idle for 30mins, the BMS disconnects and begins checking battery voltage vs a map of cell manufacture stored.
Then with cycles, temperature, and other sorcery a SoC figure is calculated. The old one is adjusted to the new.
Battery tells inverter (if new FW aswell) OCV is complete successfully. Then gets on with its day.
Another less accurate mode is trying to do this while pulling power. So the voltage measured is offset by another map of volt drop, gain based on load, upto 4A. Then SoC adjusted.
Not as accurate. But better than nothing.
Latest version goes to do this allbthe way to 80%.
Thats a lot of data points in the table.
Memory constraints and voltage changes don't warrant it done over 80%.
TheDragon (GivEnergy) so the FW only does this is the battery is idle <50% (80% in the new beta). It won’t do it at 100%SOC?
TheDragon (GivEnergy) More than happy to give 3069 a try on mine. SA2202G173. Its an older 3.6 hybrid and an 8.2 battery. During summer its usually sat in the upper half. Have experienced 2 drops from 20% to 0 since being on 3012 when forcing a discharge. Have also seen a SOC increase when I forced an idle at 10%, gained another 4% back to 14%.
Thanks.
chris_p
I'm pretty sure that a loss of 1kW in 18 months would not be expected. I can't remember off the top of my head what the warranty terms say about the anticipated degradation.
Your figures are interesting, and it would be worth seeing whether things settle and you start getting something closer to the full capacity in and out of the battery during these tests. As I've said before, my concern was not what the percentage gauge reads, but whether it was symptomatic of the batteries not actually being to reliably store or discharge their full capacity.
bjb Yes, I agree. I do keep checking each morning with the % diff between 0:00 and 8:00 with how much discharge there has been and it never works out to 8kW. Charging seems to show more capacity but not always.
I had a weird one last night....Both batteries on 3012, and the battery where the SOC crashed was only commissioned on Friday. Thankfully after 6 hours, it decided it was full again.
Hello chaps, what is the latest inverter firmware for Gen1 3.6 Hybrids? I'm on D0.196-A0.195 at the moment with 3012 battery firmware.
I also wanted to check the best methodology of checking battery capacity, I don't believe I'm seeing the full output from my battery as I did some full discharges last week while I was on holiday to make the most of flux exports and the portal suggested I was in the low 7's rather than the 8.2kWh I'd expect?
Ra My Gen1 Hybrids are on D0.499-A0.499. Tech support updated mine last week. They said that's the latest available that isn't the new faster firmware that gen2 now has.
Cheers fellas, I'll contact the relevant people and ask the question.
ProximusAl looks like the crash battery had a cell voltage breach, in fact 2 cells
The FW will learn, or should do, that there are 2 cells slightly different to the rest.
MikeJ No there is Gen1 Beta nearing release
For Inverter SA/SD2209 or newer
Mike are your inverters serial numbers SA2209 or greater?
Asking as mine is a 2208 and needs a test done prior to loading the newer firmware.
TheDragon (GivEnergy) Sounds about right. I have a lazy cell in pack 1 and a couple of overenthusiastic buggers in my new pack. Starting to look like a Morris Dancers Maypole 🙂
ProximusAl Daft question, this was calibrated on loading 3012?
TheDragon (GivEnergy) Yes indeed. Calibrated last Friday.
Stevef1964 They are both early ones. 2106 and 2112.
After seeing some 909 firmware users having timed charge or discharge issues I thought I'd best test mine. It worked but as you can see it discharged at 4kw according to the app. I thought the limit was 3.6?
Gen 2 5kw inverter and 9.5kwh battery.
Lincs_Will what happened after discharge finished at 17:50 - did it continue to operate as it did before discharge or did it just use the grid I.e. no discharge at all.
Thanks Mike. Any issues during the upgrades or was it very straight forward?
I ask because I know of one that failed and it took the system down and now needs a visit, which is a few weeks away.
So nice to hear of a successful one - indeed two!!
Dom66 it didn't discharge at all and used the grid for excess power usage. Standard behaviour for my system on 906 firmware though so no change there.
Is the new AC3 firmware ready for beta testing?
3069 update - SOC vs voltage. Increased battery dump for a few days.
TheDragon (GivEnergy) any decision made yet what to do Gen 1 with serial number before this (like mine SA2205xxxx which has only been installed less than 8 months by an approved installer) ? Think I recall for previous on this thread, that we may be needing an engineer visit?
Gen1 3.6 Hy + 5.2kWh 80%, D0.195-A0.195, 3012
So I am now on the other side of BMS beta and but still in inverter FW beta (I think).
Now that we are having loads of sunshine my battery does not fully discharge overnight. I am now seeing very regular SOC jumps (very noticeable adjustments). I wasn't expecting to see these on the BMS version I have. 2 examples from yesterday and today:
Can anyone explain what is happening at these SOC jump points and is that now expected and OK, or is there still a SOC issue lurking?
I was on the beta but that entire time was during the winter/early spring months when I would always fully charge/discharge the battery. The last few weeks are showing new behaviour for me.
JaseSpace Wow, over 55V - did you get an over voltage alert too?
chris_p No alerts. Seems to be peaking at >55V most days (according the portal info). Is that bad?
JaseSpace I thought it was a bit high but mine has reached that a few times recently. Has your battery been idle (nothing in/out) for more than 30 mins whilst under 50% recently?
3069 day 9.
The odd behaviour seen on days 1 and 2 returned today. A massive top half over correction downwards, followed by a period of normal charging and then throttled (248W) charging back up to full.
At full the displayed SOC jumped from 70% to 100%.
Interestingly the 2 hours idle at 15-16% SOC did not initiate a lower half correction. Is there a restriction on the number of upper/lower half corrections that can take place within a given time? There was a clear disparity between the battery voltage at idle and the displayed SOC% which was not acted upon.
Post Script: Displayed 100% SOC voltage is stabilising at around 52.49V which seems a bit low?
chris_p Based on the portal data I would say not. The only time there has been a flat line on battery power is just after these large SOC jumps. Each of those idle periods were for 1.5hrs, but at this point the SOC was 87-89% before the jump and 100% after.
JaseSpace If I were you I'd try to get it down below 50% then hit the pause button in the GE app. Let it idle for an hour and see if it corrects the SOC.
chris_p Thanks for the suggestion. So is this a known deficiency then, with that suggested temporary work around?
chris_p does this pause below 50% trick work on 909 firmware too?
RHardie I think it is BMS specific, version 3012 onwards
RHardie Below 50% is when the battery will realign the SOC if left idle for long enough. Not great that currently it doesn't do this at 100% when most batteries will remain idle. I'm hoping GE look at this again.
Just losing my mind at myEnergi and many different answers to max PV input.
One option is to keep the extra kWp generation and move to a 5kW g2/g3 inverter and a 9.5 kW battery.
As it’s an east/west array I’ll need the new inverter firmware asap or it’s a no go due to string limitations.
Would I get on beta or get the new firmware installed (and give feedback) if I decide to go this way?
kram Sure. Just ask.
I have many FW variations for all types of our inverter.
The best are selected for wider production use.
Ever evolving, new features, fix new obscure bugs.
The more use they get the more chance of finding and therefore fixing issues
TheDragon (GivEnergy) is the new AC3 FW in a fit state for beta testing?
chris_p thanks chris, the problem I am having is that I am unable to pause the battery at any point. I tried again late last night when it was at 45% but it continues to discharge at between 97w-300w and pulls the rest from grid to meet house load. During the day when the battery gets to 100 percent it constantly cycles between charging at 196w then discharges at around 400w instead of going to idle. Any bright ideas on a fix? It's a gen2 hy5 ED2248G069. Interestingly I also tried a forced discharge and it was discharging at 4kw instead of 3.6kw.
RHardie Mine does the same...Same inverter.
If you want to force the batteries to idle, turn off ECO mode for the duration. This will idle the batteries.
I did an OCV just the other day when the batteries were about 20% SOC, and it took 2 hours for them to stabilise. I just flicked ECO back on.
Pausing the batteries gives me the same result as you, either a discharge creep or a charge creep.
ProximusAl cheers for that, I'll give it try today.
ProximusAl what a smart cat you are! Eco mode now off and battery is now paused and idle. I'll leave it a while and see what happens.
RHardie You ideally want the SOC under 40% to get the real benefit of OCV.
It doesn't do much above that....
ProximusAl bit of a development.. battery stayed paused and idle until my PV generation just went above 5kw which is the AC limit on the inverter. Instead of capping the PV it is now charging the battery at 228w. I am thinking this isnt a BMS issue and more like to be inverter firmware problem? Any thoughts?
RHardie I think someone else mentioned on another thread that this may be an issue for you. The thread where your battery is spiking at 100%, due to exceeding inverter max.
It's impossible for me to have this problem as I have 3.2kW of panels on a 5kW inverter.
It's probably a bug in the inverter firmware.
For you though, generally you would want to OCV at night when your batteries are below 40%, and therefore no chance of solar pushing you past the inverter limit.
ProximusAl cheers, I'll give that a go. I've got 8,250w of panels so maybe that is my problem
TheDragon (GivEnergy) Hi Paul. BMS is not learning. I’m sat here with £7000 of paperweights again. Any chance you can take a look please?
Following on the from the email, even though I have no issues at the moment AC3 289 and BMS 3012 I am happy to test newer firmware to confirm various bits. I now have a stupid amount of solar so not fussed about any calibrations providing it is started in the morning :-)
Slap me with whatever you need/want testing!
in response to email @TheDragon (GivEnergy)
happy to try the 3068 beta 😃
CE2219G122
TheDragon (GivEnergy)
Not sure if you want any more beta testers for gen2, or if only ac3.
If you are, then very pleased to join the beta gang. ED2248G036
I have a gen2 5.0
9.5 battery
11 X 42O, south
6 X 420 west.
On 909 (thank you ), from 906
Solved slow start in the morning with 2 strings.
Solved capping of largest string at 3.8
Introduced battery cycling when PV exceeds 5kW, and battery 100%.
I have also niggling issues with:
Only seeing about 72% battery out to battery in.
Apparently no benefit from bypass diodes during partial hard shading.
The niggles may be unrelated to FW.
ProximusAl 0730 this AM i loaded 3068 for you, in response to your email
This should do the self calibrate upto 80% under low load, as you dont do flat and OCV doesnt run
THALL Loaded
TheDragon (GivEnergy) Is 3068 better than 3069?
3069 is not working for me, I’ve had SOC corrections, top & bottom on 2 consecutive days now.
Inverter CE2122G263
Inverter firmware 289 on the other hand is the bees knees! Almost completely eliminated grid spikes!
TheDragon (GivEnergy) Just for clarity, will the OCV up to 80% only work online. Will it work if I idle the batteries?
Also happy to try the the newer AC3 beta firmware here @TheDragon (GivEnergy)
AC3 with 2 x 9.5kWh on 3012. All working OK at the moment - just a bit slower than I'd like on response times a times 🙂
Thanks
TheDragon (GivEnergy) Hmm unless this screen is lying it has not loaded. Are you able to confirm from your end? I can fire up givtcp once I get thome.
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
3012
Apologies if I've missed it - once I'm discharging on flux from 4pm I'm seeing a big drop off of generation from the PV. It's as though the generation is being curtailed in favour of the batteries, anyone else having this issue?
94jdh If you’re discharging at 3600w, and are subject to G98, then this would make sense. The batteries can’t charge as you’re emptying them, you’re hitting your export limit, so only option is to pullback on the solar as there’s nowhere for it to go.
ProximusAl I'd agree but I'm able to Discharge more than double that to the Grid.
Also, the PV output at the time was around 3kWh then dropped to 0.8kWh, so would have expected to the batteries to supplement the PV to max export not throttle the PV?
@TheDragon (GivEnergy)
I see 3068 loaded thank you
TheDragon (GivEnergy) Just to confirm, not loaded this end. I am wondering if it is the same issue before where my older battery updated fine but the new one failed?
THALL When Paul did mine it took a few attempts. Took immediately after I temporarily disabled GivTCP.
No idea if you use it, but it hogs inverter resources which was causing the update not to take.
ProximusAl I have it installed on a box but only fire it up when I want data to check bits of data. But I had a small issue applying 3012 where my newer battery failed to report back correctly. Just wondering if this is the same issue. As in I do have the new firmware but it reports incorrectly.
Might it be worth creating a thread per beta version and archiving on launch?
Could then pin a post with the fixes and enhancements planned for that release.
Following this thread over multiple products and multiple firmware versions is tough.
@TheDragon (GivEnergy) 3068 issue, fully charged but not discharging when demand from the house exceeds PV. Tried restarting both inverter and battery (by power button) but the most it is discharging is 3W according to the app.
chris_p @TheDragon (GivEnergy) Hit pause in the app, waited a few seconds then hit play and battery started to meet required demand. Will continue to monitor.
Does this make any sense? Quite a big difference between both batteries.
Both batteries on latest FW.
3069 day 12 / 3068 day 1.
... and again.
Didn't see anything in the 3068 release note about this issue having been addressed; 3067 maybe?
Time to count energy in/out and reference what is expected against the cell voltages for a does the SOC sound mental check?
https://batteryuniversity.com/article/bu-903-how-to-measure-state-of-charge
Link not intended for GE staff just a for info.
Im happy to be corrected but I don’t believe the BMS is capable of that. I don’t think it has a shunt or any current transformers to monitor it.
A look through Givenergy Modbus would suggest the BMS doesn’t feedback power or current (just voltages) so the power readings are being done on the inverter side (so you’d only be able to make it work if you just had 1 pack, you couldn’t differentiate the master from the slaves as to which pack was doing what)
I don’t know, but wouldn’t that be a rather large omission if the BMS can’t report power in and out.
Does the AIO have an improved bms or is it going to have similar issues?
kram Intersting article, thanks for that. Helped me to get a deeper understanding of what we are dealing with here.
One of the forum members actually installed a Victron Shunt close to the battery and was able to predict accurately when SoC drops were going to happen by comparing the readings from it to the GE reports. ie, the shunt gave a much more reliable measurement of the SoC over time.
No idea if the AIO has one, but it sounds like a very good question to be asking at this stage.
Interesting what @hoggy is suggesting that this could possibly be made to work at the inverter for single packs, as I would guess that many of those spending very little time idle below 60-70% are single battery setups (else they're a bit oversized in some cases)
Pretty certain it was @ispookie666 with the Victron shunt.
If the counting and comparison to voltage could fix single packs with the hardware that’s there (and that is the big if), maybe a firmware version that does one thing for single packs and another for multiple would be worthwhile.
AIO question kind of needs answering though.
This is an educated guess so here goes:
AIO is HV as a system BUT is modular and uses 76.8V 52AH packs (x4 to get to 307V nominal of the system). 3.99kWh each with 3.4kWh usable (so about 85% DOD)
Now that’s not a million miles away from 51.2V AND the protection & monitoring system in the original packs is a LAPIS (now Rohm) ML 5238. (There is then an Atmel processor that GE runs its own firmware on to make sense of these readings the Lapis takes)
This has a max recommended voltage capability of 80V. So it would technically still fit their application in the AIO.
What doesn’t is that that particular chip is for 16cells. (Yes I know some of the GE packs have 32 but those are in 2PS16 so it’s in a roundabout way still 16 cells to monitor)
The AIO packs I assume to have 24 cells inside as they are still LFP (so this fits the 76.8V pack voltage) so this would mean some interesting design compromises if it’s really just the same BMS.
TL: DR - I don’t think it’s the same BMS. Would be a brave (foolish) move if it was given known issues with the old one.
Also the modular packs are very compact so I can’t see where you’d cram 24cells + their old BMS.
Fun fact - 52AH LFP cells are actually quite rare so you can narrow down the manufacturer to just a few companies. (My personal guess is Gotion brand)
phoenix emphasis on the "could possibly" part.
I believe the Gen 1 is already constrained resource wise such that it's implementation of the new "pause" function will be done via the cloud instead of in hardware/firmware owing to it running out of memory / CPU.
ED2248G186 3069 "BMS over voltage" and "battery voltage high" notifications on 3069? on the 17th. Noticed you updated me to 3068. Not sure if this was before or after above inverter notifications.
Seeing some odd behaviour on the 3068 firmware. Batteries at 100% but the logs show a tiny constant charge of 145w then a short burst 300w discharge. Charge has been constant for over an hour, see 2 300w discharges 15 minutes apart. A little weird.
THALL Is your solar generating more than your inverter rating? I've seen it happen with that.
ProximusAl I have an AC3 so should be unaffected. But checking my pv inverters current solar is below their rated spec. All of my inverters are rated higher than the pv attached to them. Checking back I can not see this happening on 3012.
CE2122G011 AC3 plus 2x 5.2 battereis - had our first SoC drop today, after a month or more where the SoC hasn't been below 50%.
Currently on D0.534-A0.534 and 3007. Would it be worth updating the BMS firmware without the inverter?
wilfleek Yes. Though I think your ac3 is behind by 1 release as well, I think 535 was the last public.
D'awwwww, 2 months of trouble free since the update and today this happened:
It doesn't look like a 'sudden' drop to zero... but it kind of is. That spike is the kettle going on first thing, which apparently sucked 40% out of the battery. I suspect if this was the old firmware instead of being a slope that would be a drop. the effect is the same: I have no juice left for the rest of the day (and it's going to be cloudy so I'm looking at a full day on-grid booo). This also fits the 'drop to zero' experience I had before - after an extended period of time where the battery never reached zero due to lots of solar etc., the first time it is 'challenged' below 40% by a heavy draw it goes 'wait WHAT?!' and dies.
/sigh. I will call in to get a recalibration once the office opens but when o when will we get the ability to run a calibration ourselves?
dunklezhan calibration is something you don't really want to do too often as it pushes the cells to the extremes and that is not where you want to be for longevity.
As I understand it, the firmware should sort out the SOC as it now knows where it is again.
dunklezhan If you are on 3012 or later, have you tried switching off Eco mode for an hour or so when less than 40% SOC (better if less than 20%)? After at least 30 mins of battery being idle, BMS checks the open circuit cell voltages and corrects state of charge. If you leave it idle, I’ve seen it adjust a second time a while later. If you hit the reserve level and stay there for at least an hour from time to time it does this itself, but otherwise it seems you need to manually pause the battery using Eco mode to get it to correct. It would be better if it auto paused and sorted itself out, but in meantime worth doing manually once every few weeks to avoid the crashes…
Dw I have not. But I can.
anglefire I know, I know - but it's not like whenever I've called up to ask they have quibbled or investigated or something first, they've just checked current SOC and today's graph no different than I would and then done it on the spot.
Plus for some reason they can't schedule it except by setting themselves a reminder to trigger it manually - even though its the far future of the 2020s already! Crikey its no wonder we don't have flying cars. The point being you can only do it within office hours, the least convenient time, and if you could do it yourself at least you could control the timing of it.
OK folks,
EDIT - FW breaks local connection. Waiting for further test results.
SOC to 0% issue and oscillation when full fix. Gen 2,
Looking for a couple of repeat offenders for SOC dropping to 0 when the battery is full. 3068 helps prevent it, the inverter FW will stop it showing as a SOC drop
Inverter G684 & G071 & G186 already done
Need 5 more, who are ALREADY a tester
D910-A910 (20 Jun 2023)
1. Add G99NI & G99NI Certification Codes
2. Improve battery disconnect on battery overcurrent
3. Optimize CRC issues in 0x10 multiset register
4. Fix SOC to 0% on battery charge MOSFET disconnect
5. Optimize battery charge logic once full (Oscillation)
TheDragon (GivEnergy) count me in for this 🙂 ED2248G069
RHardie I dont see you on the beta tester list.
Therefore am unable to test pre-release FW without you agreeing to it. If it breaks your inverter, It will be me you come crying to.
v909 was well tested on many system before issuing to the support cell, wider use.
This is under 24hr old, so Im not that brave
If you agree to beta FW, may not have expected results, in the worst case ay even cause unexpected damage, which we will fix as a matter of urgency, we cannot guarantee it will be repaired in 2 days, like some feel is needed.
Yes its unlikely, but can and has happened.
Been burned once helping someone out. Now playing by the rules.
This is why we BETA test. FW cant connect locally. Something in the data had changed.
Saw the 100 to 0 on 3068 yesterday on one of the batteries. ED2248G186.
TheDragon (GivEnergy) mate I am so sorry! I somehow didnt notice you were offering the FW to existing testers until I got your reply. Sorry for wasting your time and I totally understand your reasons for not wanting to install on a non beta tester device. all the best 😀
TheDragon (GivEnergy) please check if I am seeing this scenario on 3068 on ED2248G186. Drop from 100 to 0 yesterday on one of the batteries.
3068 BMS. Paused the battery for over an hour to see if the OCV would run but no SOC adjustment so either SOC is tracking well, OCV didn't run or I've misunderstood what conditions it runs under.
chris_p I'm seeing similar Chris. I'm doing some logical testing of the 3068 BMS, and I tried to force an OCV and nothing happened over 90 minutes either <80%, but I'm keeping Paul informed of my findings on a daily basis, and the OCV is one I noticed, but haven't given it my full attention yet, as I'm monitoring the BMS mainly on keeping cell voltages in check.
4monks also local Comms lost. Last GivTCP update 3 hrs ago.
ProximusAl Perhaps I should be emailing, but I see he is still checking here too. I prefer the forum as it is much more open and those not on the beta can get a window into how it is going - which ultimately should build confidence...
chris_p I can understand, but I wouldn't want to fill the forum with boring daily reports of my progress, so find it easier to email.
I'm definitely seeing improvements in the cells with BMS 3068, it's just taking time to make sure the test conditions don't change radically (Which is somewhat difficult with unpredictable solar) 🙂
After a few more days of the tests I'm currently doing, I'll shall then concentrate on OCV and SoC testing.
4monks Yes you are now 3 of 3 noting this.
Did the SOC drop to 0 while you have D910-A910 on the inverter? This was one of the issue addressed?
I will revert the Inverter to 909 once this is identified, and await the lab testing focused on local comms
TheDragon (GivEnergy) no idea as I wasn't aware /missed that I was getting d910. I've issued a restart, but no local Comms still. Will leave for now as I'm away.
4monks ok, so I can at least gather some data from the oscillation battery power while full.
Thanks
@TheDragon (GivEnergy) Are there any specific tests you want 3068 users to do?
chris_p I just dont want to see any more over charged cells
If I see that, im happy.
The rest of 3068, wont be visible, such as clearing the corrupted files on the USB drive, preventing further updates. That needs lab testing somehow, to copy a corrupted file there and test the resposne to it
TheDragon (GivEnergy) What kind of voltages should we be looking out for?
chris_p You don’t want individual cells going over 3.6V, or ideally 3.5V. If they ever hit 3.7V, that’s when things get real funky with the inverter. This is the exact test plan I’m working on currently and 3068 seems to be learning….
Same for pack voltage. If it gets to 57V, then funky things happen.
The good news is, since 3068 on my install, we are seeing….so far…good results, and the BMS seems to be teaching the cells a lesson!
@TheDragon (GivEnergy) tried updating inverter and battery firmware this morning, update failed and inverter now "dead", could you take a look please? CE2122G011
wilfleek knock me down with a feather - inverter took its time but seems to have updated according to the portal, trying battery firmware update now, fingers crossed.
Hmm, not sure that has gone well, needs a check please.
wilfleek 3 batteries. Will take a while.
What FW did you do.
From version to new version?
TheDragon (GivEnergy) Only two batteries?
Inverter went from 534 to 535, batteries were on 3007
I have cancelled the auto update, its failed, but passed on the inverter.
DOing BMS (3068) for you manually, then calibration
Thank you Paul.
I seemingly have never been able to do a battery firmware update "automaitcally" - had to pull the USB memory sticks the last two time, clear out file fragments, and manually enter the new files. Any reason why that should be?
wilfleek OK, the transfer has finished, reboot and wait and see what happens.
Looking at the DIP switch settings, its most likely failing, as you are Slave 1 missing. You have module 1, and 3. Master, Slave 2. Did you previously have 3 batteries, one removed.
If not, the slave needs its DIP switch changing. But will give it another 15 mins before abandoning the upgrade remote
All done as per your instructions, mystery over the DIP switches - as installed, had been working OK.
Edit - calibration running now, so all good.
4monks Feel free to revert to 909 at your leisure. I'm still away but monitoring remotely. Thx.
Just found this, mentioning overcharging cells, is it relevant to my post ? justpassing
hazelst3 other views are available. It’s saving me a fortune.
anglefire Intermediate view: Currently, for my use case I can see that it might pay for itself over a decade if I watch it like a hawk and there are no hardware related issues. Not compelling enough yet, I value my time. Part of the reason I was considering it is that I expect the Net Zero drive to lead to more frequent brownouts, but if I can't trust it to be there when I need it, there is little point. Candles, woodburners and an emergency gas stove for the win atm. YMMV obviously. That's not to say I don't hope these issues are fully resolved. I do, but wont commit capital until they are, which actually seems to be moving a little further away at the moment, having felt within reaching distance a month or so ago. Currently disappointed, but not giving up hope yet.
anglefire You clearly work for GivEnergy.
hazelst3 nope. Not at all. This month I have generated nearly 500kwh and only pulled about 20kwh off the grid.
I’m not going to, but I could put my consumption figures off the octopus forum to prove it.
phoenix I watch it with home assistant and givtcp - but I’m a sad sod. I don’t control the inverter it just does its thing.
Saved me about £800 last year. So yes payback is somewhere between 8 and 10 years (depends on tariff - our current one is 24p fixed)
anglefire Just checked my last month and I generated 379 kWh on little 3.2kWh panels. I’ve spent 235 kWh on grid juice at a cost of £28.64 🙂 and that’s charging the batts overnight, plus daytime overspill.
My leccy bill went from £180 a month to £40 since having my install.
ProximusAl and that's without factoring in export payments. With Flux export rates I'd be supprised if some people weren't making money.
Apologies for the inappropriate comments from a fellow member of this group.
Removed
So can someone confirm the the AIO is using a different BMS/Cells to the hybrid/ac3 battery packs?
TheDragon (GivEnergy) Ha! Removes the context of my previous post, which was intended to be some sort of rational middle ground vs the deleted post.
Never mind 😉
kram yes. AIO uses a HV battery pack.
With each pack in the set having a BMS. Controlled by a BCU
TheDragon (GivEnergy) could you repost them as I missed them first time round!!!!
P.S So far this period I've exported 375Kwh and used 5.7Kwh, and no gas used!!
So they will be paying me!!!
@kram the AIO battery modules look like this:
hoggy Nice that they have included some drawers, will be very handy in my utility room 😆
Gen 1 5 hybrid.
2 x 8.2 batteries.
3012 battery firmware.
I seem to be back to square 1 again.
Massive gaps in the individual SOC. I have done (I presume correctly) OCV balancing. For example on Wednesday morning I put the batteries into Idle and left them for 2 hours. One was at 56% the other at 32%.
I'm back to one battery hitting 100% and the other flatlining at 92/93% and not accepting any further charge.
I'm seeing the batteries peaking at 57 volts. Some individual cells are hitting 3.6V
Am I expecting too much or am I doing something incorrect here?
Zakalwe My overview currently shows 100% but the individual batteries show 84% and zero respectively. I'm hoping the zero is a transient blip but whatever, there is still too much unpredictability unfortunately. I too can't get both batteries to 100%, one always lags by several percent. We've been trying to cycle below 40% but not often below 20%.
I think that you also need to put the batteries into idle when the SoC is low to allow it to run a OCV.
Just cycling the batteries makes no difference. I have regularly cycled from 100% to <20% as well as cycling 100% to 40-60%
The battery percentage thing in the app is just the app being typically useless. On both my Android devices I always kill the app off and then re-open it fresh. That's the only way that I can consistently get any sense out of it.
Zakalwe Thanks for that. So the OCV has to be forced? How is that done please? Sub 20% and then force idle? If so, how long for please? Seems a bit clunky and I guess most users wouldn't do or even know about this? Is there something in the pipeline to do this automatically? I think there is also beta firmware available which balances at the top end but I'm losing track a bit as I've been up to the eyes with my daughter's house refurb.
sponev So the OCV has to be forced?
That is my understanding, but I've asked the question.
sponev Sub 20% and then force idle?
Again, that's my understanding. I think that there is another firmware that can allow it at the top end.
sponev If so, how long for please?
No idea (hence me asking). I've let it idle for up to two hours with little visible effect.
Zakalwe sponev I've asked these questions numerous times but still no official answer. I posted a comment on a Facebook post by Paul saying would be nice for this to be automated, the response was along the lines of "baby steps" so I guess it is coming, but from my experience baby steps are slow and clumsy 😆
@TheDragon (GivEnergy)
Paul, can you please advise?
Thanks in advance.
SoC jump at the top end is back too.
We still seem to be getting the difference between the Primary & Secondary. This may be because we have an AC3 that still does not have the software that lets the two batteries balance. The Secondary never appears to get to 100% as the Primary always gets there first.
Graph for today below:
I think we are in the land of approximations and not absolutes and that there will always be some anomalies. This may be acceptable for the majority but it would be good to have confirmation so we can determine whether it will work for us or not. There's always the hope that ongoing work will tighten up the system as the firmware improves. But where are we and what can we reasonably expect please?
So I’ve drained both my batteries as far as the inverter will go, and turned off ECO so both batteries are now idle. I’ll give them 90 minutes and hoping to see an offline OCV.
EDIT: I should add, my individual cells dropped below 3.0 V which isn’t a great idea for LifePo4, and my lazy cell in pack 1 hit 2.8 V gulps
Well, this is promising. I can visibly see the SoC changing already.
As I hoped, the SoC for pack 2 has dropped from 12% to 8% and still going. Pretty sure pack 1 SoC is spot on.
I always suspected the SoC was too high on pack 2.
ProximusAl 12% and 48v is wrong. OCV will correct to near 0
TheDragon (GivEnergy) Agreed. OCV has reduced it to 5% and still going. I’m hoping it hits 0% to match pack 1.
TheDragon (GivEnergy) Looks like 4% is the best I’m getting. OCV has now flatlined, but that’s an 8% reduction which I’m glad of.
Tracking near perfect now!
Offline OCV is the way to go.
ProximusAl
"near perfect" ?
gosh you are a hard cat to please. haha
positor1 I say near perfect because both batteries at 92% but inverter says 93% 😀
I should add, this is the beta BMS 3068.
Don’t want to disillusion people that this was 3012.
i have
inverter 53%
batt 1 53%
batt 2 50%
beta 3068
no forced idle OCV
@TheDragon (GivEnergy)
Can you please detail the process to carry out the OCV and what one should expect?
Thank you.
Already seeing a drift 🙁
Zakalwe Many time this has been discussed
OCV Offline - Below 20% and Idle >30min & < 24hr since last (Very visible)
OCV Online - 3012 below 40% and under 4A per BMS, Idle is irrelevant, the SOC adjust as it is in use (Not really visible)
3013 Below 80% and under 4A per BMS, as above
3013 also has in it
stuck file delete routine - 3068
auto BMS reboot on crash - 3067
validated CRC, reverting has yet to be tested, may not be possible once crossing this line (Comply with regulations on modified code in transit) - 3067
3067 has not been issued out, till i get confirmation how to revert once its gone past this point. As the BMS from this point forward will not recognise an un-validated binary file
3069,8,7 are all betas between 3012 and 3013, if you werent aware
Coming in the next round of Inverter FW
Validated CRC on the data & bin - Not issued till GivTCP is patched
SOC to 0% on battery charge MOSFET disconnect fix
Oscillating while full battery power
Additional EV charger register control
TheDragon (GivEnergy) charger register control? I presume this is only applicable to the forthcoming GE charger?
Smk82 correct
3013 Below 80% and under 4A per BMS,
If my maths is right P=VA so that means no draw more than (52v x 4a) = 208 W.
My house idles at over 300 W so in theory I’m going to have to do off lines only?
TheDragon (GivEnergy) are there plans for third party charger monitoring? Not full support, but at least monitoring energy flows so it doesn’t all get lumped with house load? Or would that require API as well?
This information is always appreciated, and the summary even more so.
Is there some way to sticky this, or set up an FAQ. I know there is the unofficial wiki, which was a great initiative @hoggy , but even that cant be found by people new to the forum/givenergy system.
There are a number of standard questions ( that I also had when I started ) that come again and again. And this thread is fascinating, but covers 4 different hardware versions, multiple inverter and battery firmware versions. A simple FAQ and a "state of the nation" statement as above per hardware variant would cut down a lot of chatter and confusion.
This is meant to be a constructive comment not a complaint, as there is a lot of information, knowledge and help on this forum that is appreciated by all.
ProximusAl I read it as 4A per BMS (I.e per battery) think you’ve got 2 so does your background usage fit in the 8A window?
Smk82 there is/was going to be 2 different ways of doing that -
1 was using another EM115 and just giving it another ID and linking it to the existing modbus network.
2 was using the Givenergy API where you send it data and it displays on the portal.
Not sure of the current status of either method though (I know the API is half implemented but not the bit that shows on the portal)
hoggy If I can draw up to 416 W total, then yeah, I reckon so. Hard at the minute as I have fans, air con running etc to stop my family melting at night 🔥🔥
As Paul says, the online OCV may not be visible, and I’ve never seen it….only got results from Offline OCV which I don’t have an issue doing periodically. That’s got to be healthy for the batteries a full cycle every now and again.
ProximusAl single battery yes.
TheDragon (GivEnergy) OCV Offline - Below 20% and Idle >30min & < 24hr since last (Very visible)
OCV Online - 3012 below 40% and under 4A per BMS, Idle is irrelevant, the SOC adjust as it is in use (Not really visible)
TheDragon (GivEnergy) Is this supposed to work with the current software on our AC3 (Inverter Firmware Version D0.289-A0.289) or do we still need to wait for the updated AC3 firmware for it to balance between the two batteries? Our Secondary Battery never gets to 100% as the Primary & Inverter stop charging when they indicate 100%.
As you know, we are on the 3068 beta BMS, but have not seen the Secondary charge to its full capacity, despite plenty of solar generation.
Woodlands
i am ac3 289 and beta 3068. batteries seem balanced ( enough) here. maybe 2% less sometimes.
This is ours over the last 24 hours:
Difference at the top end was 9% and 12% difference at 12pm today.
The Secondary battery never appears to get to 100%, even though the Primary battery does.
Woodlands same for me at top end, I've given up and accepted it as a feature. Seems fine at the lower percentages. Doesn't seem to affect capacity. 
TheDragon (GivEnergy) Would it possible to display somewhere on the app when the SoC was last 'corrected' by the BMS? That would allow people to get to know their systems better and provide more confidence/warn that forcing an adjust may be advisable. I think for many that the unpredictability is almost as big of an issue as the drops. This may help to resolve that ...
The logical next step from this would be for the BMS to learn and warn, but that's a few steps ahead of here I think.
EDIT: Just spitballing, it's probably the 'Kwh throughput' without any 'correction' that is more relevant than 'time' as the former would be a better measure of 'oscillation without adjustment', which is how I understand the errors creep in and add up.
Hmm, mine is the opposite. After the OCV last night, top end is great, but it’s currently waaaay off the mark.
Woodlands I would not expect the BMS reading to be regular getting to 100.
Should be around 95%
Woodlands I would not expect the BMS reading to be regular getting to 100.
Should be around 95%
TheDragon (GivEnergy) Woodlands I would not expect the BMS reading to be regular getting to 100.
Should be around 95%
TheDragon (GivEnergy) But the Primary Battery does regularly get to 100% with the current Solar PV generation. However, the Secondary Battery never appears to get to 100% because the Primary always gets there first. I understood from your explanation above that the latest beta BMS would allow the batteries to balance and both achieve the same level of charge.
TheDragon (GivEnergy) that's what I would expect, but 100% primary and 90% secondary also in my case. Why does the firmware constantly report incorrectly if 95% is the intention? And I don't (fully) believe it's voltage drop because the secondary has longer wiring as is offered up as the cause.
I'm sure if I ran it to zero and ran an off line OCV, I could get that to flip and get the secondary reading 100% and primary 90% (lots of examples I've posted where that happened albeit slightly older beta firmware). This was an example from may that busted the voltage drop theory and shows its firmware driven. Clear switch after 3rd May OCV
Does seem to be more of a visual annoyance rather than causing capacity issues.
mpartington My battery lag flips. Usually the secondary showing around 94-95% but I have seen it the other way around with the primary lagging.
I forced my single 8.2kW battery to 10% this morning, offlined for over an hour and saw the OCV take it to 16%. I'm confident this part now works after two similar tests, just now it needs to be automated or scheduled as I believe that done frequently enough (weekly?) then it should keep the SoC in better check.
Thanks for replying. It's not really visible as it gets lost in the day-to-day "noise". Perhaps a "sticky" would be in order to help users keep track? It'd also be useful to show exactly what the expected results are.
I discharged mine to below 20%. I then left the batteries in idle for 2 hours. Yes, there was a SOC readjustment on G308. As soon as I re-enabled Eco mode G302 charged higher and faster. It also discharged faster, in essence it is doing all the heavy lifting. G302 went from 15% to 46% in the same time as G308 went from 12% to 29%. Later in the day G302 went from 22% to 60% whilst the other went from 22% to 39%.
Is this normal behaviour? I don't see any evidence of the batteries taking themselves off-line or tracking.
TheDragon (GivEnergy) If you are on the new 3012 BMS.
Let it go to idle and watch the packs OCV function level them. If the voltages are within 0,5v they will track.
If one is different and exceeds the allowed tolerance, it will go offline and leave you one battery. till they next meet voltage wise.
Again, if this is normal behaviour I'll let it lie.
Zakalwe that doesn't look right at all! Probably explains your big soc difference over the whole range. My 2 battery voltages are much tighter for example. Almost like you've a big voltage drop on one of them, as they do move together, but offset
chris_p There certainly seems to be something special about 16%. When my system self adjusts at the bottom end it nearly always corrects to 16%. Seems unlikely to me that the SOC drift would be that consistent.
Woodstock Yes and my previous OCV adjustment went from 10% to 16% too... Odd to say the least.
To "close the loop" a bit on this, the batteries have now hit 100%
Both batteries did a SoC Jump to get to 100%.
A bit more concerning is that both batteries peaked over 57V at 57.17V and 57.37V.
3068: Paused my battery at 36% for an hour this morning and no SoC changes. Guess it is close to correct. Voltage was reported as being between 51.8V and 52V during this hour of pause.
Crazy SOC graph today. This is on 3069. I was better off on 3012.
ChrisRowe At ~ 56V I would say that battery is full so the SoC did need "correcting".
chris_p I think the problem is the 'correction' from >60% down to <20% at about 8am. I've seen a couple of examples of this recently in this chat where it looks like the online <4A adjustment has happened and it's taken a huge chunk out of the SOC, only for it to jump back up by the same amount to 100% when it hits the voltage limit later in the day. Seems the online adjustment is not correcting for the voltage difference due to load correctly.
Phil_H I agree but without the battery power on the graph it could be load reducing it, the shallower decline vs the steeper incline suggests it "could" be but yeah unlikely unless it is one of the smaller batteries.
chris_p Given the voltage shows no sign of a drop during that period I don't think it's down to a large load being placed on it
Phil_H Yeah I'm guessing that is the case, some kind of "correction" has probably occurred.
ChrisRowe If you scroll back through this thread you will find several similar examples from my system. The initial downward adjustment is way too big and is followed later by a SOC jump as a full battery is approached. @TheDragon (GivEnergy) said that one of the firmware developers was going to monitor my system and investigate the cause but I have not yet seen any feedback on progress.
chris_p There’s no one at home this week, so no large load kicking in. This is the second day this week that there have been “corrections” in an empty house.
ChrisRowe Definitely needs an investigation then
there is a new AC3 version as if yesterday 288
So latest betas
AC3 288 ( i dont have the file yet)
Gen1 194 (Breaks local control so not rolled out)
Gen2 Final 910 (Not rolled out till local control updated, not validated and crc checked
Gen3 Final 304 (As above)
AIO 607 (As above)
BMS 3067 (3068 only issued, as 3067 has validated crc function, so cannot be reverted, so need to be sure)
All new FW has validated control functions, so GivTCP & our own Ap needs an update before this can go out
TheDragon (GivEnergy) any update on when the AC3 master/slave version for plants will be ready for beta?
TheDragon (GivEnergy) I got an updated GivTCP docker last night, but I’m about 1500 miles from home currently so probably not wise to do any beta testing for you currently.
TheDragon (GivEnergy) is Dewet still knocking about or is Mark on his own & got to redo Givenergy_Modbus himself for the CRC stuff?
Well he fixed one of my issues yesterday, so I guess so :-) (Dewet)
Despite running a couple of OCVs in the last few days, this system is not tracking the batteries or accurately estimating the true SOC
G302 looks to have topped at at an indicated 90% (56.053V)
The OCVs were performed when the voltages were less than 0.5V between them. At this stage, I will just have to accept that this is how this particular system performs. 3012 (in my case) most certainly has not fixed the SoC differences between the batteries nor has it fixed the estimation of true SoC in the batteries.
just for info.
My two discharged to a lower % than normal and after a good solar day are reporting
95% over all
100% batt1
87% batt 2
charging at 170 watts
A larger variance than i was expecting.
Finished the charging off with both batteries jumping the last bit to 100%
Sorry, I’m a little confused, did I understand your previous post incorrectly about the AIO having different firmware and BMS?
Am i late to the party or did we just get new settings to play with 🙂
TheDragon (GivEnergy) when you say breaks local control, is this because the discharge rate is higher than it should be when exporting?
Gen1 3.6 Hy, 5.2kWh 80%, D0.195-A0.195, 3068.
I seem to be getting very regular very noticeable SoC jumps now. Here is an example of today's:
Large drop from 29% to 8% early this morning followed by large jump from 84% to 100% mid-morning. Similar on other days and is now happening every few days. On the face of it the SoC jump issue is now MUCH worse that it ever was on my system before the original beta started, and 3068 has not solved any of these issues.
The AIO has a different data packet structure, validated CRC being one part of it.
So GivTCP is getting an update.
Those extra controls are for the new FW, if it says time out, its not available to you
Gen1's can turn pause on and off
Gen2 and onwards can schedule it
Sorry been quite, fixing some AIO FW remote upgrade issues, which were failing and causing bricked BCU's. Light is visible now, majority of affected units sorted out.
TheDragon (GivEnergy) Just remember, Even dragons need a rest 😉
@"TheDragon (GivEnergy)"
Is the. Inability to pause the gen2 inverters still a thing in all versions of firmware ? I'm on 909/3012, and I can't get below 281 watts charging when the sun is out, and the battery is below 100%. Not a major issue, but keeping up with what's fixed needs regular forum scans 😊
There seems to be a bit of a pattern developing now on 3069 where a couple of times a week I get a large SOC drop in the early hours, and then a similar jump to 100% early afternoon. This suggests to me that the earlier drop wasn’t justified.
Any solution for this please?
As i have had no responses or issues with SOC on 3068, It will become 3013 in a few days, then the default for those leaving the factory from next month,
TheDragon (GivEnergy). I personally have had no issues, but there are 2 reported above (with charts) in the last 3 days
@ChrisRowe and @JaseSpace
Edit, actually 1 may be on 3069
mpartington My problem is with 3069. Will 3013 be available to install from the portal @TheDragon (GivEnergy) ?
ChrisRowe yes, it would be. Although as you're on a beta, it may not be displayed as the beta is a higher number than production (unless that's been fixed). In which case it would need to be loaded by @TheDragon (GivEnergy).
Looking at your charts it should be a big improvement.
TheDragon (GivEnergy) Yes, no problems here with 3068 👍️
No problem here with 3068, but still have the issue that the secondary battery never charges to 100%.
TheDragon (GivEnergy) No SOC jumps with 3068. Also no OCV adjustments taking place. However still questioning the overall capacity - Full daily cycles for last two weeks Average 9.8 kWh going in daily and only 8.23 kWh coming out. (Inverter card \ meter data)
TheDragon (GivEnergy) Please see my last post above. I am seeing regular SOC jumps up and down on my system (3068).
CCourtney round trip losses converting from DC-AC? Yours comes out at 84% which is a little low (normally 85%-95%).
Cdent Do you really see 95% round trip on your battery ? From PV charge to AC discharge earlie rin the year I was seeing 72%, and its now crept up to 82%
Cdent Ok I understand there are round trip losses and that my battery losses are on the high side of normal but my understanding from what I have read here in the past was that the battery was actually oversized so that it could be expected to output 9.5kWh.
TheDragon (GivEnergy) That new battery FW - when upgrading from 2012 - is there a (yet another) calibration required?
TheDragon (GivEnergy) My last top half (overly large) downward SOC adjustment followed by a SOC jump to 100% occurred on 17/6/23 and spanned the installation of 3068. Nothing since, so looks promising.
No issues here
TheDragon (GivEnergy) I've seen deltas which vary between primary and secondary and an unexplained battery dump which happening right now. ED2248G186.
4monks I'm wondering if the registers are corrupt or somethimg in GivTCP 2.2.2 did something odd. I'm running Octopus Flux and normally dump the batteries between 4pm and 7pm but didn't today. I have a safety backstop that stops the dump at 7:05pm. This runs every day and this is the first run since the GivTCP 2.2.2 update. Since then, the batteries are discharging at max rate for no apparent reason. BMS 3068. Inverter Gen2 on 909 (regressed from beta). I can only limit this dump by scaling back the export limit to be slightly more than consumption. Tried a reset and reboot of inverter and GivTCP. It continues to dump....
I'm having my Gen1 swapped to a Gen3 5kW Hybrid in the next 7-10 days.
I've 2 x 8.2kWh batts on 3012.
If you want anyone to try any new invertor / battery firmware on this type of setup I'm happy to help.
I've no problem accessing the usb sticks in the batts if needed and have a USB A to A cable on standby!
are these two drops on 3068 significant? it is the first time my batteries have been this low since winter.
was surprised to see only 17% left this morning
one drop of 12% over 40 mins and another of 8% over 30 mins both with a consistent equal load.
TheDragon (GivEnergy) Hi - could you possibly take a look at ED2248G186 - battery dumping at full rate and now not pulling solar. Not sure if this relates to 3068 or the 909 rollback or maybe recent GivTCP update. Thx.
CCourtney if that's the case you're down by 2kWh input/1.3output so could indicate an issue
4monks If its not pulling solar, sounds like a calibration?
I have almost similar setup to you, including 909 rollback and on 3068, only difference is I'm using GivTCP docker (very latest 2.2.3 I think), and my system is operating ok.
ProximusAl Thx - does look like calibration but no indication why / how. Batteries both went down to 1% and then to idle. Solar started working again once discharge complete but no battery charge as yet. Batteries slowly creeping from 44.85 and 46.89 to 47.76 and 48.03 over the last 1.5 hours but but GivTCP and portal indicate 0w charge - any excess from solar is going to export not the batteries. Will leave it for the moment - hopefully @TheDragon (GivEnergy) can confirm what went on / is going on. I updated to GivTCP 2.2.3 late last night but no difference.
4monks Might be worth emailing @TheDragon (GivEnergy)
ProximusAl Not sure I have Paul's email?
4monks Hmm, seems there is no way to DM in the forum, and I wouldn't like to give it out without his permission.
ProximusAl Thx. I've emailed support - hopefully Paul or support will pick up.
4monks I did a forced charge to kick to 2% then reverted to ECO and "seems" to be behaving.... No idea what happened.
3068 replaced by 3013. No calibration and I didn't feel a thing. 🙂
Woodstock How did you get 3013?
ProximusAl I'm going through those with 3068 when I get spare 10mins
4monks I see no reason why the calibration was begun then stopped. The reason no charge begon after 0% as it wasn't server initiated.
See fraser in service desk had a look at it too.
TheDragon (GivEnergy) I’ve disabled GivTCP for now just in case a spare 10 minutes is soon 😝
I’m back home now.
TheDragon (GivEnergy) Thanks for sending me the 3013 BMS software to upload on our 2 batteries. It has successfully uploaded and both working OK.
I'll let you know if I spot any issues. Thanks
TheDragon (GivEnergy) I’m still on 3069. Can you update me please!
TheDragon (GivEnergy) Thanks. Spontaneous partial calibration it remains a mystery..
4monks I wonder if it could have been an engineer on site somewhere else. When my engineers were fitting my second battery, I heard him go "Oh S**t", and he admitted he had done the commission on the wrong job. 😛
ChrisRowe Done
TheDragon (GivEnergy) I've disabled GivTCP again Paul, if you have a spare 10 mins?
OK, updating now
TheDragon (GivEnergy) Same please - GivTCP stopped. Assume this is for 3013 update? Thx.
@TheDragon (GivEnergy) you can slap mine with 3013 if you wish. The bugs we last spoke about I think were all inverter firmware not bms?
@TheDragon (GivEnergy) Is the 3013 update going to be made available via the portal? Also any idea when the AC3 update will come out.
Thanks.
ajcrewe Yes it is, its in the queue for portal updates
@TheDragon (GivEnergy)
Something odd, inverter notification while discharging to the grid.
Battery Voltage Low 2023-07-05 18:31:03 2023-07-05 18:32:03
number CE2219G122
battery info for that period though is;
time volts power temp %
18:34:03 51.05 3145 24 48
18:33:31 51.07 3148 24 49
18:33:02 51.14 3139 24 49
18:32:03 52.07 20 24 50
18:31:08 36.41 11 25 49
18:31:03 39.6 12 25 49
18:30:03 50.53 3148 25 49
positor1 check fw revision.
You were on tonight's list
TheDragon (GivEnergy)
showing
Inverter Firmware Version
D0.289-A0.289
Battery Firmware Version
3013
Looks like I got 3013 at 18:12 yesterday. Slight odd soc drop when it applied. Soc drop 12% then bounced back up 9%.
@TheDragon (GivEnergy) Thats good I take it? I still have that 2% at the top issue and yes I may have left my system discharging to grid last night..... was poking around the app and may have fallen asleep!
THALL the little difference would be worries about, it can drop to 0 then bounce back up as the batteries reboot. Depends on where in the data pull cycle the reboot and reload happens.
Falling asleep, while meddling, done that, in fact i only half set charge last night, so only 1 inverter did a charge, so that will test my SOC balance and X-Charge script today getting themselves back re-aligned from 40% difference
TheDragon (GivEnergy) Where as I will be testing capacity in today ;-) Nice to see the system hunt for extra power at the bottom 4 times! It worked really well, found over 5% more power.
First day on 3013. Should I be worried about this drop or is it still finding its feet?(so to speak)
ChrisRowe That could be the first online OCV up to 80% I've seen.
If it doesn't happen again, I'd say that a good thing.
ProximusAl it’s just jumped from 72% to 100% in 10 minutes. Is that good?
ChrisRowe Wait for your next charge / discharge cycle to be sure. Might be nothing.
THALL So I started with 7% (flat) fully charged took 18kw? No idea of exact losses but lets say 10% conversion 16.2kw 1.2kw reserve so 17.4kw of a total 17.72? Without knowing loss I can't give exact figures. But looks pretty good to me! Hopefully no more 100% drains until the winter........ turns of eco mode on the zappi back to eco+ ;-)
Will new inverter firmware equalise battery use?
I've 2 8.2kW batteries the primary is showing approx. 25% more use (cycles) than the secondary.
I'm getting my Gen1 swapped for Gen3 - will future firmware updates aim to use the batteries more evenly or is it worth asking my installer to swap my batteries around to help even the wear?
3068: Had my big boy pants on today, pushed battery to "empty" and hit 4% with no SoC jump/crash/leap. It found another 4% of capacity so I drained that too and went back to 4% eventually. Paused again to see if would find more but alas, no more was found. Now charging back up to 100% to test capacity. All in all that was a decent test and I feel my SoC issues are gone?! Hope not spoken too soon...
Some additional thoughts:
chris_p Why did it take the OCV 1h30m+ to start to find more capacity?
Waiting for the temperature to drop?
Tim Waiting for the temperature to drop?
Possibly, but even then it waited 30 mins from the previous drop from 32c to 31c before any change in SoC occurred:
chris_p Now charging back up to 100% to test capacity.
96% (4% > 100%) (7.6 / 96) * 100 = 7.92kWh.... Not too bad for a 20 month 8.2? @TheDragon (GivEnergy)
chris_p not bad. you want 3013 yet, still on 3068, the last of my AC3, Gen1 hybrids on 3068
TheDragon (GivEnergy) Yeah whenever you have time, thanks.
TheDragon (GivEnergy) I'll take 3013 when you have a moment CE2105G039.
Biggest SoC jump that I've ever seen on any firmware to date
82% to 100%
chris_p Done, having a beer now
TheDragon (GivEnergy) Well deserved, enjoy!
judgepd Thank you @TheDragon (GivEnergy)
judgepd w
TheDragon (GivEnergy) where are we with the 1.5 gen hyrbrid updates?
I have a SA2226G which I volunteered for beta testing but haven’t heard anything for months.
Do we have a date for the firmware update?
I had a firmware update then battery started to charge to full. Was told firware issues then asked for a role back. Now my battery is nit charging off the grid at scheduled times
Another drop this morning on 3013. I paused it at 11% for half an hour to see if that would help stability. Not much sun around so unlikely to get much charge today. Would there be any benefit in forcing a charge up to 100% today?
I’ve no idea what happened to the voltage around 11am!
Is the 3013 update available to all? Can’t seem to see it in Inverter->Software page. Currently on 3012.
Since 3013 I am noticing neither of my batteries report 100% charge in the app. The 9.5 now tops out at 98% and my 8.2 says 94%. The app reports 100% total however so it is a little odd. For those concerned about SOC balance I noticed over the weekend over 35% difference between my batteries however they do eventually balance out. Charging up this morning both are reporting 74%.
Further to my reports from last week of frequent SoC "jumps" I noticed 2 anomalies again today. I am on 3068.
Some time around 6am the PV kicked in and this caused a rapid decrease of SoC from 29% to 9% over 25mins- no other noticeable reason for the sudden decline. Not calling this a jump, but certainly it is an unexplained gradient change. Then later at about 0925 the SoC jumped from 85% to 100% within the 5min resolution of the portal.
If I look back over previous days there are other occasions of the unexplained morning decline in SoC. Seems to be happening at least once every 3-4 days. Sometimes followed by a later jump, but not always.
Out of interest, any update on a firmware update for AC3s?
JaseSpace The decrease is initiated by the battery input/output dropping below the threshold for a lower half adjustment to take place. Favourite time for this to happen is the transition between overnight battery discharge and daytime solar battery charging. This is normal operation now and is part of keeping the displayed SOC in line with true SOC.
The error is that the downward adjustment is overly large which results in the SOC jump to 100% the next time the battery approaches full.
This is a known problem which you will find reported several times earlier in this thread.
Of concern is that you are still seeing this in 3068. Last time I had one of these was in 3069 (earlier release than 3068) and it was supposedly fixed in 3068/3013. I have not seen this occur in 3068 or 3013. Perhaps the fix was not applied until 3013. Guess you need to get 3013 installed and see if you still have the problem.
Unexplained battery discharge spikes just after midnight.
Started with 3069 release and has continued through 3068 and 3013. Has appeared six times now although I am not sure if it is actually happening every night and only coinciding with the data point now and again.
Anyone else seeing this?
Woodstock I had similar at one point but found one of my dc discharge time slots was somehow set to start at 00:01 thru to 23:59 so would got a discharge spike each midnight. Updated the discharge slot setting back to 00:00 to 23:59 ( & set to disabled) and thankfully the midnight peaks I stopped happening.
Woodstock Are you referring to discharge spikes at midnight? I’ve seen this on inverter FW 909.
I've seen a total of three similar events during July 23 but none prior to this.
My system is set to run on ECO only with no additional settings.
Woodstock What you describe here is exactly what happened this morning to me. Everything ticking away nicely, then the sun hits the panels and the SOC plummets. Not much sun around today so unlikely to see the corresponding jump later. This is on 3013. The battery had a full discharge/charge cycle at the weekend so the calibration shouldn’t be far out.
@TheDragon (GivEnergy) Amy solution for this in the pipeline?
This is new and unseen before.
Like the OCV has mis calculated toe SOC direction
Im starting to lean toward the BMS needs calibrating on change, even though its very similar to 3012.
This is very strange and not seen elsewhere
TheDragon (GivEnergy) But it had a full discharge/charge cycle on 3013 at the weekend?
TheDragon (GivEnergy) Not sure what you are referring to but if you are talking about a SoC drop when PV kicks in in the morning then I reported it yesterday, and effectively reported about a week ago (though didn't associate it with PV at the time). In fact I think I mentioned it on the previous beta version I was on too (losing track of what version when, sorry). So if you are talking about this then it has been around for a little while now on my system and others seem to be reporting similar. If you aren't referring to this, then I still have an issue that needs some attention.
FFS ... can't commit capital to this
TheDragon (GivEnergy) and I have a similar pair of SoC adjustments again today:
SoC decline from 26% to 15% over 20mins around the time that PV kicks in at 6am, then later SoC jump from 80% to 100% within the 5min portal resolution. As I mentioned before, this is happening every few days on my system (Gen1 3.6 Hy, 5.2kWh 80%, D0.195-A0.195, 30683013)
(Edit: Actually, having just checked, it looks like my system has now been updated to 3013 from 3068 in the last day or two w/o me noticing or being notified.)
Out of curiosity, is everyone who is seeing these occurrences on a hybrid system? I’ve seen a few minor drops/jumps before I updated to 3012 but only a couple of %, but nothing obvious since? (AC3 x2, 8.2 x 2 here)
Smk82 I've got an AC3 and 9.5 battery and it seems much better behaved since 3012. I'm not getting the sudden drops to 0 which I had regularly in the past. I can't see any obvious SoC updating happening on the plots so assume it is happening quite gently in the background. I'm just waiting for the AC3 update to increase reactions time.
TheDragon (GivEnergy)
"This is new and unseen before." This is not true, you are misinformed.
I have the same problem and its been ongoing for weeks and i have many emails between my Installers and also GivEnergy.
I have been told the issue is related to a mismatch with the battery and inverter firmware (I have emails from Giv stating this), but ridiculously there is no timeline for the new inverter firmware to be released.
My installers asked Giv for the battery to be replaced under the sales of goods act as I'm now 'losing' up to 27% SOC when the panel start charging the battery.
This happens when the battery has been 100% charged and drops to under 40%, then when the first panel output reaches around 0.3kwh.
This is inline with what ChrisRowe and JaseSpace are seeing I believe.
When this first started it was a 10% drop, then 12 then 14 then 21 and the last was 27%.
We have now had the battery updated to 1303 as the possible fix. But as JaseSpace is still have the same problem while on 1303 I can only hope it fixes the issue for me.
Some examples of the SOC drop happening. As you can see as times gone on the loss has become worse.
Joeh1051 What is the inverter serial< i cannot find from JoeH1051
Seeing quite a gap in SoC between Primary and Secondary battery on 3068 / 3013 on ED2248G186
Joeh1051 There is an important distinction here in that when your battery next approaches full it does not suffer a "sudden jump" in SOC. This suggests that the lower half adjustments you are seeing are doing a decent job of realigning the displayed and actual SOC.
Whether the displayed and actual SOC should get that far apart in the first place is another question.
Smk82 Mines AC Coupled. Single 8.2 battery.
Dai Thanks for that. Checked my discharge settings and performed a read on each of them (never believe what the portal says). The start and end time in each case was set to 00:00 (Eco mode enabled, DC discharge disabled).
As far as I am aware only the discharge 1 settings are used to control battery discharge so I am wondering if the discharge 2 registers have been repurposed. I may just try setting discharge 2 start and end to 00:00 and 23:59 to see what happens.
Bearing in mind past problems brought about by changing the settings of redundant registers (charge spikes at 01:00) I may live to regret it! 🙂
TheDragon (GivEnergy) is is this? CE2241G175
4monks I see the same on mine. Paul hinted a calibration may have been necessary between the firmware versions, but I've not been in the position to ask for one just yet. Batteries too full 🙂
ProximusAl Guess my phantom calibration didn't quite do the trick 😉
Woodstock Do you mean getting close to a full SOC it will jump up above what should be expected from the panel output?
If that is what you mean I'm not seeing that happen.
@"Joeh1051"#p33979Essentially, yes. If the displayed SOC downward adjustment has been over enthusiastic it can overshoot and change from displaying a SOC higher than actual SOC to displaying a SOC lower than actual SOC. When this happens the next time the battery gets close to full the system realises its mistake and the displayed SOC will jump to 100%.
You are not getting the sudden displayed SOC jump to 100% which leads me to believe that what you are seeing is not the problem under discussion here. It looks like your displayed SOC downward adjustment is (albeit rather coursely) correctly realigning the displayed SOC with the actual SOC.
In all cases it is the displayed SOC that suffers the rapid adjustment. The actual SOC will only change in line with the real battery charge/discharge. Think of the displayed SOC as being like the petrol gauge in your car mis-reading. It doesn't matter what the gauge says, the actual amount of petrol in the tank is only affected by how much you put in or take out.
Woodstock my situation seam exactly the same as ChrisRowe’s.
The same drop as soon as the panels start producing at first light.
To confirm though, are you saying the situation I have is ok and normal?!
Joeh1051 I could be wrong but my guess is that when PV is equal to house load, or thereabouts, the battery is idle for a brief period and the OCV runs allowing the battery to start to correct the displayed SoC. As Woodstock says it is probably over correcting and going too far. The coincidence with the sun coming up is probably the trigger to allow the battery to start to idle.
Just noticed a change in behaviour when the battery is at 100% with no charge/discharge cycling now, and on checking this morning the battery firmware has changed from 3068 to 3013. Inverter is ED2301G474.
@TheDragon (GivEnergy) , did you by chance update the battery firmware at around 1345 on 12/7/2023?
As I posted when 3013 went out. I was manually doing all the 3068 to 3013.
I had some time yesterday.
So managed to do the rest of my beta inverters.
So likely yep. Was me.
Interesting the cycling stopped as that fix was an inverter patch from. 909 to 910
chris_p Spot on with the idle period at sun up initiating the SOC adjustment (as I said earlier).
Over correcting too far was the problem under discussion but this does not seem to apply in @Joeh1051 's case as he is not getting the sudden jump in displayed SOC to 100% the next time the battery approaches full.
Joeh1051 OK and normal may be going a bit far but rather it is consistent with expected operation of the current firmware release. What is not consistent with the expected operation of the current firmware release is the over adjustment of the displayed SOC which was the topic under discussion (but does not apply in your case).
Your SOC adjustment at sun up is expected operation. The absence of a following jump to 100% as the battery next approaches full is also expected operation and suggests that you are not experiencing the over adjustment which was the topic of the discussion.
Woodstock Ok with you.
I just saw people posting the same issue as mine and assumed this was the right place to discuss it.
Should it be its own thread?
Hi Paul - I'm still on 3012 - and a beta 289 AC3. Did I get missed or were you not covering the AC3 folks ?
Thanks.
ianfretwell I think 288 is currently in test. @TheDragon (GivEnergy) Paul is there a plan to roll this out to those of us still on 289?
I've temporarily held off with my plan to pause the battery each night, as it currently discharges itself at 100wh.
TheDragon (GivEnergy) Hi Paul - given that we rolled back to Gen2 909, is there any news on 909's sequel? Thx
4monks 909's sequel is 911
Need App update, which is tested good. Waiting verification and publishing. Once done, I will issue 911 for update
911 fixes the oscillation of battery power when full and sunny.
Adds local protocol security controls
Adds UPS
Jamesrich Not unless you specifically want it. 287 is the newest AC3 DSP beta, with the correct logif for AC charge to %, allowing discharge if over target and charge to if under target (During the charge window)
So now all of our inverters follow the same logic. Its a target % not a charge to %
can see here dual AIO/AC3 inverters both obeying the logic, at 2330 the target was 50%, actual was 80, so it didnt just sit there, it allowed the discharge down to 50% then stop, till 0530.
TheDragon (GivEnergy) Strange behaviour last night. BMS 3013. Inverter 289.
The car was charging 23.30 - 05.30 and the battery had a target % of 60 for the same period.
Every time the battery reached 60%, it would discharge to the car, down to 59%, and then charge from the grid, back to 60%. Again and again and again. I was sadly awake at 02.30 and watched it happening in real time!
I’ve never noticed this behaviour before in 18 months that I’ve had car & battery
ChrisRowe the oscillating around a set point was seen in this version, it is improved in the next.
Just remember not to let Auto update kick in
TheDragon (GivEnergy) Sorry, what’s Auto update?
ChrisRowe The green button that appears on the portal, when it thinks there is a newer FW available.
TheDragon (GivEnergy) is this the master/slave AC3 version youve mentioned before? Or is that another iteration?
Or are thing’s playing nicely together due to HA?
TheDragon (GivEnergy) Thanks Paul
So my desired target is always 100%, until a few months ago it had trouble maintaining it with 'charge to'- but now its totally fine.
Its the pausing that is my only issue, it refuses to hold. I'm happy to try 287, assuming there are no issues?
I think I read something about issues with GivTCP not supporting it yet? That would be a showstopper for now.
TheDragon (GivEnergy) Ah, I see you’ve done it for me, Thankyou.
I now have DO287-AO288. I’ll ignore the “update now” button!
Oh dear.. sudden soc drop to zero while exporting
warning notification of low battery voltage
both batteries showing full LED
inverter flashing green circle, dongle solid blue.
Battery disconnector had tripped, has never done so before, not hot either.
a mystery!
Reset and now showing 82% and all appears normal.
positor1 what rate we’re you exporting at?
I wonder if the current draw dragged the voltage down and the correction algorithm is being a bit keen.
kram
full power plus solar generation but that wouldn't affect the dc isolator on the batteries that tripped.
Been operational without a trip for 11 months so quite baffling as to the cause.
Bad news, it's back!
Had a massive downward displayed SOC% adjustment this morning. Dropped from 71% to 16%. Will post the graph when I get the expected sudden jump back to 100%.
Disappointing as I haven't had one of these since 3069. Operation under 3068 and 3013 has been fine up until today. 😞
@TheDragon (GivEnergy) As you know I have been reporting noticeable SoC changes every few days on 3068 and now 3013.
It has been a poor PV day today and my battery has been sat around 50% most of the day (after an early cheap period boost). Then as I was starting to make dinner (almost exclusively on the gas hob) this evening I noticed the SoC plunge from about 50% to 13% in a very short period. The portal captures this as:
That is my biggest SoC jump yet and it happended in a different part of the day to usual. 3013 seems very unstable. MUCH worse for me than anything before the original beta started. I get some sort of large SoC change every 3-4 days.
Please to report that my Gen1 3.6 inverter with the older serial number has succesfully passed the firmware test and is on the latest version - Thanks Paul 🙂 - which also means the battery firmware was upped from 3012 to 3013 - no calibration seemed to be needed.
Seem to have a lot less ripple when load changes - but has been hard to tell as its p***g it down today!
As expected the displayed SOC% sudden jump materialised later in the day (61% to 100%).
It is interesting that the downward (over) adjustment stopped at 16% which seems to be a "magic" number. Bottom end upward adjustments, more often than not, also stop at 16%.
Don't know how significant the downward adjustment short pause at 51% was. Had the displayed SOC% stopped falling at this point it would have been about right!
Not beta related per se, but do you have the temperature and cell voltages at the time of the jumps?
Tried to force an OCV adjustment today, paused battery at 25% for 1h 20m but the battery never went idle, it just sat at 6w-10w discharge and no adjustment took place.
kram Don't have the individual cell voltages but battery voltage and temperature were as follows ...
If (when?) I get another massive downward adjustment of the displayed SOC% I intend to take the route of discharging to "empty" to see what sort of correction (if any) takes place at the bottom end. I expect it will correct upwards to 16% but it will be interesting to see if it gets any higher.
Am I seeing something that isn’t there?
That SoC adjustment seems to be closer to the temperature and completely unrelated to the near flat voltage.
Is this something switching in a lookup table because of a temperature banding?
kram Not sure I am following you. Near flat voltage? My battery has not been below 50% charge since 7th June.
The upper half SOC adjustment took place under 3013 rules because the displayed SOC was less than 80% and the battery charge/discharge power was below the pre-programmed threshold for long enough to kick it off. I have no problem with the displayed SOC% adjustment taking place when it did, rather that the scale of it was totally out of proportion to what was required. This left the displayed SOC% under reading by a considerable amount which led to the "sudden jump" the next time the battery approached full.
@TheDragon (GivEnergy) Looks like there may have been a regression on 911 in the behaviour under lower light ( morning and evening) .
The picture below shows 909 changing to 911 on afternoon of the 12th. The early late low light has a smooth progression up and down under 909, fixing the 906 slow start issue. But 911 seems to have the 906 behaviour. This is a two string system, south and west.
Firmware 906 :-(
Firmware 909 :-)
Firmware 911 :-(
Up to about 10 am the voltage is between 51.5 and 52v
It goes up as the temperature drops from 20-19 but the SoC moves at that time too.
I could be totally misreading the graphs though as the line colours are similar.
JaseSpace @TheDragon (GivEnergy) so after that late plunge from 50% to 13% leaving me with an apparently "depleted" battery for most of the evening meaning I was drawing from the grid:
here was the adjustment back up the following day, a jump from approx 59% to 100%!:
All previous occasions of drops/jump pairs have occurred within the same day, but this pair straddled 2 days and meant I was importing from the the grid for some time, for no real reason other than a SoC bug in the BMS.
JaseSpace what was the battery voltage.
So see if the SoC was tracking it.
Have you a cell issue?
TheDragon (GivEnergy) Battery voltage is plotted on the same charts already (although I missed it off the 1st chart above, but the original (partial day) version of that in my earlier post does have the voltage on it). Sorry I don't really know what I am looking for, just seeing the SoC jump around.
If you look at the voltages when you get the SoC jump they’re all over the place.
Daft question @TheDragon (GivEnergy).. when the bms cell balancing happens does it take a cell out of the battery or perhaps mean that as there is a draw whilst voltage is measured it can affect it?
I know I’m not an expert in this, but you look at the graphs from those with the big jumps and there are significant (for a LFP) voltage changes on some. If they’re not oddity, then could they be a misbehaving cell or a bad connection to the cell.
TheDragon (GivEnergy) This looks very reminiscent of earlier releases when I was able to produce sudden jumps to 100% displayed SOC at will by forcing a lower half adjustment. Is it possible that change control issues have resulted in a reversion to the earlier lower half adjustment code which was over adjusting the displayed SOC%?
(My battery has not been below 50% SOC for over a month so I would not have seen this).
Woodstock I'm no expert but the voltage curve as it 'jumps' to 100% looks like the cells are completely charged, looks almost like a calibration curve. At the other end it looks like your battery thinks it is empty but the voltage is still quite high?
Ra Exactly. The downward SOC adjustment is way bigger than it should be which leaves the system thinking that the battery is in a much lower SOC than it actually is. System realises its mistake when the battery fills and reaches the upper voltage limit. The displayed SOC then instantaneously adjusts to 100%.
This over adjustment was a feature of the early "lower half adjustment" releases. It seemed to have been fixed for lower half adjustments but reappeared in the upper half (less than 80%) adjustment when that was released.
Recent posts are suggesting that it is now back again in lower half adjustments too!
Woodstock I had seen a post somewhere that suggested GivEnergy were leaning towards requiring a calibration on the later version of the battery firmware. Will see if I can come across it again!
Ra I've had a calibration since being on 3013 (got another battery). And I'm only seeing SoC jump on one of the Batteries at the top end. I didn't have the SoC jumps before tho so no idea if the calibration actually helped.
TheDragon (GivEnergy) Still getting the oscillation on the latest inverter firmware. I promise I didn’t hit the update button!
And again last night. Anyone else on Inverter Firmware DO.287 AO.288 seeing similar?
ChrisRowe thanks.
Will relay this back, as AC3 is still at beta, there is time to re-visit
Not sure if a bug - both batteries on 3013, Gen 1 5kW Hybrid on D0.450-A0.451
Batteries calibrated by support yesterday.
Why would .1v difference and similar temps show 16% difference in percentage charge?
The secondary battery a little later was at the same voltage (52.1v) as the Primary when it showed 48%, the percentage on the secondary was 63%
TheDragon (GivEnergy) AC3 : v289
Had same oscillations since you put 289 on mine. The Go 00:30 to 04:30 period shows this behavour no matter what charge level I set. I also have my own ct's and software that record every 5 secs. What seems to be going on is that the FW continuously tries to maintain the charge level and switches to charge at 3Kw mode to get a tiny amount of charge and seems to waste power. A longer charge delta time would be more sensible and give a better charge pattern.
@TheDragon (GivEnergy) Here is a screenshot of my own stuff with fast recording. It dramatically shows the continuous charging spikes. There's a ton of my own other info on there. Fuzzy white at top is voltage, and red/pink is load and charging.
This shows a few days ago when I changed charge to 80%. Have own high load as well which is the big step, but the visual highlights nice steady charge rate- hit 80% and then spikey goes wild time.
I realise this is nothing to worry about as it is a snaphot in time but it would be nice to understand what causes them.
The different soc levels with multiple batteries, does that matter? Is it just a future wear and tear issue more cycles to 1 battery than the other?
THALL Doesn't matter as such - I just wondered if it was a bug and how the inverter would deal with it as the batteries got near being fully discharged?
The percentages shown should be much closer to each other in the example shown.
Edward Different cell ages, cycle count? many parameters.
Also new inverter FW, assists in the BMS. they are a pair, only half of the features (The basic ones) work with BMS alone.
THALL Yes, the master seems to clock up more cycles than the slaves.
TheDragon (GivEnergy) So purely a longer term issue? I do see this on my batteries, seen as much as 35% difference between the 2 but never considered it an "issue" but guess it could be once they are out of warranty. Degradation etc. Realised my 8.2 battery is 2 years old this month.....
THALL I guess it depends how the overall SOC is calculated and whether 1 battery showing a low SOC effects the overall usable capacity.
TheDragon (GivEnergy)
I’m just wondering if there is a very approximate timescale expected for the AC3 inverter firmware update to be available?
Many thanks
judgepd Can't say I have seen a capacity issue though I do not check regularly. The 289 3013 combo is rock solid for me.
... and yet again!
Once again a massive over adjustment of SOC% in the upper half. As promised, this time I discharged fully and let the battery sit idle. The predicted SOC adjustment back up to 16% SOC took place but it went no further. Consequently there was another displayed SOC "sudden jump" as the battery approached full.
So we had ...
Massive over adjustment of the displayed SOC from 72% to 16%.
Partial readjustment of the displayed SOC from 5% to 16%.
Final readjustment of the displayed SOC from 71% to 100%.
Why is the upper half displayed SOC adjustment getting it so wrong?
Why is the battery idle at "empty" adjustment only partially correcting this?
What is so special about a displayed SOC of 16%?
Hi Paul,
Been out of the loop a little with work, holiday and a refresh of the DC side of the install (as part of issues from F4S installation) - now two Gen3 5kW inverters.
It looks like I've been left behind on the battery BMS 4074 and likely the inverter firmware ( because of the change.
Would you mind updating as required at your convenience, thanks.
94jdh Maybe give @TheDragon (GivEnergy) a clue about the new inverter serial numbers?
Sorry, yes that might help! I thought he might just have access to the account from previous.
FD2311G460
FD2311G462
Thanks in advance
TheDragon (GivEnergy) Any news on this?
Is 909 -> 911 due out soon?
@TheDragon (GivEnergy) Is this still closed to new testers? Would like to help. I have an AC3 with 8.2kWh battery installed in Oct 21. It's been pretty rock solid and looking to replace it with an AIO. Before I do, I'm curious to see how the faster response times work. I like how the Tesla Powerwall literally imports nothing (quick response to demand) and hoping that the AC3/AIO might get close to this (particularly the AIO with it being behind a gateway).
I've done beta testing for various companies and aware of the risks that things may and can break.
Happy either way.
Thank You
Simon
simon_rb The AC3 is not as quick as a powerwall, I can confirm that! The AIO however should be..... I don't know anyone with one yet to test.
THALL I think the AC3 is held back by regulations, where as the AIO should be freed a little as behind a gateway.. that’s how I understand it and how the Powerwall can do what it does so well.
I’m hoping the AIO can be as good as the Powerwall, I don’t want to replace my AC3 with the AIO until I can find out how quick it does respond. Always get jealous when my friends PW has only imported 10watts all day lol.
The new FW i can send over, just let me have your serial number, ill put you on my managed list
TheDragon (GivEnergy) thank you 😊
CE2133G240
What happens now? Forgive my ignorance.. I assume it will update the battery software at the same time and discharge then fully recharge. Is that correct?
THALL also, I’m sure there is an option to turn off the grid so I’m guessing it needs to be quick to respond if you can turn the grid off.
simon_rb
Firmware loading.
Never tried with such old BMS. So we are both gonna learn.
BmS will need a calibration. So won't load that till you say so
Do you have EPS circuits wired? If so I'll enable the 0s switching
TheDragon (GivEnergy) I just noticed us using the grid so can confirm it’s loading.
My cheap rate is 21:30 - 01:30 so might aswell do it straight away.
Oh, I have a socket connected to the EPS but its only for power cuts and nothing plugged in at the moment.
The 0s could be useful in the future.
simon_rb OK. Once it's back up I'll load BMS and it'll reboot again.
TheDragon (GivEnergy) Brilliant, thank you. It’s come back already! That was a fast firmware update, if I wasn’t walking past the smart meter screen in the utility room, I wouldn’t have known.
simon_rb haha.
Inverter done. BMS next.
Then it will dump all power and recharge.
So use it or loose it
TheDragon (GivEnergy) I’ve just plugged the car in. I’m ready 🤣
Inverter 289 - 289
BMS 3013
It’s charging at 400watts 🤷♂️
Calibration begun. Now discharging at 3kw
Discharge till approx 46v.
Set 0%
Charge to 57v
Set 100%
Finish
TheDragon (GivEnergy) thank you!!
TheDragon (GivEnergy) is there anything in particular we should be looking out for with this version?
Anything new added to monitor? Thought I’d ask as I’m a little late to the party.
It was mentioned this version responds to demand quicker.. I’ll keep an eye on that 👍
simon_rb I am in the market for a second hand ac3..... just saying ;-)
THALL I’ll keep you up to date 😉
simon_rb cheers. I have tried the ones on ebay but they are unwilling to provide me the serial number..... I thought more would be for sale now the AIO is here
THALL I suppose that depends on how many people are upgrading. Battery are a significant investment so not many will have enough money to literally junk their existing system to replace it
Smk82 There seems to be several on the forum, I was surprised when I saw how many people were posting about their AIO installs. But I would also assume most let their installer take the old kit away..... I do regret not getting a new second AC3 when it was around £600. The current £1000+ is a tad painful!
@THALL I’ll likely have an 8.2 battery also available. All depends if my father in law wants to buy the system. (Inverter has 10 year warranty which is transferred according to GE). It’s currently 21 months old.
THALL considering even “kit” is worth something I would hope they would get some sort of deal for that as the installer could just pop it on eBay to sell it on. Suprised how many considering the economic issues at the moment.. people have money to burn apparently!
simon_rb the inverter warranty is only transferred if it remains in the same property as originally installed 
simon_rb Upgrading after 21 months? And yes family first. While I would not be against a 3rd battery I sadly do not have enough wall left to install it to and the next bit of wall is too far away.....
Cdent I have an email from GE stating that the Warranty transfers once it’s been installed and recommissioned by a authorised installer.
THALL yes, the AIO is too tempting
simon_rb hmm interesting, their warranty doc says one thing but they say different. As long as you have the email you should be covered then
I'm a bit out of the loop on versions...
If I do this today, will it update my 289 AC3 inverter firmware or just my battery?
Does 3012 to 3013 now trigger a calibration? If so I'll wait til offpeak. Cheers
@TheDragon (GivEnergy) Hey Paul,
I'm on D0.909-A0.909 / 4073 – I remember there was some talk about the official release not getting pushed to those on 40xx due to the version number being higher than release. Would it be possible to get me back on the official track with everything on ED2152R267? Unless there's a need for further testing at this stage?
Cheers,
jp
pelowj all updated to latest of everything.
You may need the app update of not already done
TheDragon (GivEnergy) legend! Cheers.
Jamesrich Interested to see how you got on. My 289 AC3 didn't take the portal update from 3012 to 3013, it was covers-off time. However, no calibration with 3012->3013
3012 to 3013 on a Gen 1 inverter with a Gen 2 9.5 battery simply results in the server attempting to reboot the inverter every 5 minutes.
VERY VERY BROKEN.
Tried updating last night after forcing discharge the battery in the evening! Software update from 3012 to 3013 started, got notification that is complete and that calibration has started, however system is very irrational . BMS voltage reads 7.5V and got all sorts of notifications in portal…attempted switching off everything and turning back on but didn’t resolve anything…also notice that one of the battery went completely off and once try to restart it it just switches off… @TheDragon (GivEnergy) is it possible to have a look at this when you can? Inverter SN: ED2248G519
Three attempts this morning to update to 3013 but all failing at 74% - "battery ARM" or similar referenced. Two 9.5s with a gen 2, 5kw inverter - reference is ED2301G782.
Any chance @TheDragon (GivEnergy) could take a look please or advise an alternative if busy. Appreciated. There's a bit of sunshine (at last) so was hoping to get it done and the recal without too much grid draw.
dorosde I see the calibration got stuck and the server halted it. I have restarted it at the charge cycle.
sponev Three attempts this morning to update to 3013 but all failing at 74% - "battery ARM" or similar referenced. Two 9.5s with a gen 2, 5kw inverter - reference is ED2301G782.
The file is stuck at 121/134 packets.
This will need some one to work with you to delete the stuck file on the internal dongle
TheDragon (GivEnergy) Thanks buddy! I will keep an eye and if I noticed anything strange going on I will let you know! Battery found it's 0% and now calibrating to find it's 100% I suppose? It will not repeat the whole discharge-charge cycle again after it finishes to 100% isn't?
dorosde No just the charge cycle
TheDragon (GivEnergy) Thanks for the quick response. I'm at home today and can look at this now. I'm OK with IT stuff/networks so should be able to follow guidance. What's the process please?
sponev Got the firmware and procedure now - thanks.
Hi @TheDragon (GivEnergy) is there any timeline yet for firmware upgrade for Gen1 Hybrid Inverter on pre 2208 inverter model? (SD2142G119)
Hi @TheDragon (GivEnergy) - tried the firmware battery update and its failed twice, can you advise - inverter ref is CE2131G177. I hope this fixes the ability to not charge the batter overnight?
Thanks
Just a note to all about doing FW update 2012 to 2013 with no issues or rebalancing. Call support to do it. I did. Takes a few mins with no problems. Saves SOOO much hassle..
Not on beta, but I will sling my twopenn'orth in any way. Update to to 3013 on Gen 1 AC3 x 2 5.2 Kwh batteries.
First attempt stopped at validating 90% restarted by support with same result- 3rd restart by support succeeded with some rather strange results. Calibration started by support and went to bed. This morning one battery fully charged other empty and not charging.
After a while I reset settings to default with no result. Shutdown for 15 minutes and restarted and no change. Phone support and the guy worked on it for about 20 minutes and got it going - both batteries showed empty and charging. Charged battery jumped to 100% after about 20 minutes and the second continued charging from Grid. I eventually got it to recognise 1.5 Kwh solar and charge.
10 hours later, Primary 99% and Secondary 93%. SoC is 100% and alternator is Idle.
Update would appear to have problems.
@TheDragon (GivEnergy) Hi Paul, I think my battery calibration didn’t finish the other day. It discharged fine but it didn’t fully charge and now it slows right down between 89% to 93% then will jump to 100%. Voltages are quite low I think too. I suspect it failed as I have it set to charge from 21:30 - 01:30 and suspect that it stopped charging due at 01:30. I thought the calibration would override but it didn’t.
SOC is showing 100% on the home page but 86% on the battery more info page.
Deej CE2131G177
will do manually for you.
Paul, I had the BMS upgrade message this morning 7:50am ish and and so went for it.
Not sure but nothing seems to have happened. The update started but nothing since.
the update gets as far as "Verifying update - 90%" and thats it
Could you take a look?
SA2208G029
Thanks in advance
TheDragon (GivEnergy) Sorry to bother you…After loading beta firmware (289) onto my 3 AC3’s for another round of GivTCP testing, it seems as though something has gone wrong this time with two of the inverters whereby the dongle just gives a flashing blue light and I can no longer access them. Is there any chance you’d be able to take a look? Apologies for the PM on FB too, wasn’t sure the best route to take! Feel free to message if you need me to do anything my end.
I just asked support@givenergy.co.uk to update my firmware without a cal, and it was done in minutes. All good.
Is there really a need to ask TheDragon to do it? I have assumed here, that he is an expensive person and the support team are better suited to deliver a remote update?
Just received a notification that the update failed. All seems ok though.
Roberto Some people cant do it themselves because they were on a Beta version and dont get the update button. TheDragon runs the Beta programs and support will not have the latest Beta versions.
Is there a new Inverter FW out? Currently on 909 on Inverter and 3013 on the batteries and just realised there is an update button for the Inverter now. Done the batteries a couple of days ago…
dorosde Can you push the button and let me know the result. 😃
ProximusAl On a Friday afternoon?...don't think so! I will wait until Monday and let you know :-) Haven't even seen the Update Log to know what is changing...
dorosde I'm pretty sure it will be 911 or 912 if it exists. I actually had 910 briefly, but it broke GivTCP.
I'm hoping all that is now resolved with the new App update, and hopefully the new GivTCP docker.
ProximusAl It will be useful to have a log though to see what are the changes as how my system is at the moment works brilliantly, after overcoming some problems during the 3013 battery firmware update.
dorosde Do you know what.......I'm gonna do it...
I'm sure our resident @TheDragon (GivEnergy) will be there to save the day if it all goes belly up.....
Back shortly with results....just doing an inverter reboot before proceeding......
dorosde Smooth sailing.....
D0.909-A0.909 updated to D0.911-A0.911.
It's a bit bum twitchy at 90% when it's verifying.....but just be patient.....
Now to see if GivTCP still works....
EDIT: GivTCP working perfectly. Happy chappy here....
I have not yet done the 3013 battery update, and am on 909 for the Gen2 inverter, so I thought I'd see what the portal said for me because in theory I have 2 potential updates. Answer is I get 1 update button and the following text 
So it seems I can only update both or neither, and they will both happen at the same time. Not saying this is a problem necessarily, but it does tell you what will be updated (battery/inverter)
Phil_H Given the issues some are having with the battery firmware, you are probably safer waiting until Monday as per what support said.
ProximusAl You know what? I fancy a bit of action on a Friday afternoon! I am updating as well and will get back with results shortly!
ProximusAl I work in software development, I'm not going anywhere near a Friday update, certainly not a Friday afternoon one with potential niggles!
ProximusAl Just a reminder for people hovering over the button, 911 will be a "breaking change" for anyone using givenergy_local in HA, or older GivTCP versions.
justpassing Yeah, I'm using the latest GivTCP docker with MQTT and it's all good.....
ProximusAl Yes this is why ive delyed 911, 305, gen 1 194, and AC3 287
all needed GivTCP & App updates to work locally.
90% for the last 3 minutes....
dorosde Yeah dont panic....Just be patient....As I said...bum twitchy
dorosde Updated completed! now on 911.
dorosde CALL 911
2319.........2319.....2319
Doh, il get my coat, its friday
@TheDragon (GivEnergy) hey, sorry to bother you on a Saturday. Since we did the firmware update, my battery hasn’t quite been behaving. I don’t think it finished the calibration. I had a notification that it had but it never finished the charging. Today the percentage has dropped from 75% to about 20% in the last 30mins, bizarrely the battery is actually charging at around 100 watts during this time. Help? 🤣 cheers
simon_rb a change in SoC at vert low power levels indicates an OCV adjustment has taken place.
The real SoC was lower than that reported.
See how it goes today.
TheDragon (GivEnergy) Looks like the battery is full now and is only charing at 240 watts with the rest being exported. Its been doing this at around 90% since the firmware update, however today the battery is at 41%.
simon_rb sounds like my issue
TheDragon (GivEnergy) This has all the hall marks of the upper half massive over correction of displayed SOC that I have been banging on about for some time time now. The real SOC may well have been lower than that reported but the correction applied is way more than the actual difference. I predict a sudden jump in displayed SOC when the battery next approaches full!
Woodstock Mine kept dropping at a few to several percent every 5 mins for 30 mins and went from 72% to 16%. It has since charged to 42% and won't charge anymore. I'm assuming the battery thinks that 42% is 100%.
simon_rb That's interesting. What voltage is the battery showing when it is stuck at 42%? There is something significant about 16% as this is where several are reporting the displayed SOC stopping during both the initial downward adjustment and also the subsequent recharge (if the battery is fully discharged and a bottom end correction takes place).
Woodstock From memory it was around 53v but I’d have to go back and check. At about midnight last night, it jumped to 100% from 50% after 3 hours charging at 240watts. Today it has behaved 🤷♂️
Hi would you please have a look at these, firmware and BMS:
Sorry, yes that might help! I thought he might just have access to the account from previous.
FD2311G460
FD2311G462
Thanks in advance
94jdh both ill update Inverter & battery. will take me 30 mins or so
Many thanks, wouldn't have bothered you but I don't have the update button from being on the beta.
G462 is all updated. Will need a calibration at some point
G460 the DSP is stuck in update, so its kinda sat half updated, but working. Im still working this one
@"TheDragon (Gi
That's great, thanks for the update - happy for calibration to run to suit you
94jdh OK, the lower one is now charging, from the discharge of the higher one
I suppose in this instance that's a benefit of cross charging!
So 287 is fine with the latest GivTCP?
If so I'm happy to have it Paul, if you get a moment. Currently on 289.
I assume I can't upgrade through the portal as its a higher version and still beta?
CE2225G411
Many thanks
James
Jamesrich Yes 287 and 288 fine on GivTCP. Tested ok.
Im running 287
TheDragon (GivEnergy) I updated my batteries, 9.5Kw gen 2, to 3013 without any problems but trying to update the two inverters (both HY3.6 gen2 ) has been problematical. EG2152G105 gets to 26%, hangs there for a bit and then fails the update. EG2152G169 gets to 53% and then fails too. Both systems are now reporting software versions as DO909-AO911 which looks like a partial upgrade has happened? Support say "looks like firmware is stuck" and I'm waiting on advice on what to do next. What is puzzling me is why both systems have failed, in slightly different stages of updating.
Any thoughts?
Barry
@TheDragon (GivEnergy) Your guidance will be much appreciated as always.....My Gen2 5kw Inv is on FW 909 and Bat is on FW 4074.
I was expecting to have to get you to assist at some point as 4074 is the old Beta and so I wasn't expecting the Update Button to appear.....But it has.
Will the update button take care of both Inv and Batt?
simon_rb Yep, that pretty much confirms you are suffering the upper half "over adjustment" of displayed SOC% that a few of us are seeing. Would be nice to know if this has been recognised as an issue and whether anyone is looking at it.
Great, thanks for confirming. I'll happily take 287 and see if my 289 issues are solved.
Many thanks
James
94jdh
You have one stuck file on the internal Dongle
In order to remove, can you source a USB-A to USB A data transfer cable?
Instructions Below
Fully power the inverter down (No lights)
insert the USB cable in the inverter & laptop
A usb drive will appear.
copy the msn.txt file to your desktop
Format the dongle FAT
replace the MSN.txt file back on the dongle
Eject the USB drive and remove the cable from the inverter.
Power the INverter back on, and we can try again to load the new DSP firmware
Yes, will do when back home thanks
Jon-boy it will only do the inverter. Portal doesn't recognise beta FW.
4074 to 3013 will be manual and a calibration
Spent more money on the import during the post update battery calibration than in the total import cost of the past 4 months. - thanks GE.
IanW wow!
Not got one, but ordered for delivery tomorrow
TheDragon (GivEnergy) Many thanks for confirming. Would you mind adding me to your list for the manual update please. ED2301G097. Anytime at your convenience would be great, dont worry about SOC. Thanks
Hi, USB turned up at lunch, so popped home and completed the above for you to try again when you get chance,
Thanks
Thanks, I see it's worked this time.
Jon-boy @TheDragon (GivEnergy) Thanks Paul, looks like you slipped in and did your stuff, I never felt a thing. No Calibration either. Much appreciated.
... cue another one! Down to the "magic" 16% SOC again.
I have checked back though to when the upper half (50%to 80% SOC) adjustment processing was added and every triggering (six so far) has over adjusted the SOC, resulting in a subsequent sudden jump to 100%.
The idea of adding an upper half alignment to cover systems that rarely drop below 50% SOC in the summer months is a good one but the execution is currently flawed. It would be interesting to know if anyone has had an upper half downward adjustment that has not resulted in a sudden jump of the displayed SOC% the next time the battery approached full. It would also be useful to know if anyone has seen an upper half adjustment that has been upwards rather than downwards. If so, was this followed by a sudden drop of displayed SOC as the battery approached the lower reserve?
All of my upper half displayed SOC% adjustments (so far) have been downwards and every one has preceded a sudden jump to 100%.
Woodstock I repeat my earlier request of more detailed logging for SoC corrections
Woodstock could you click on Battery Percentage button at top of graph and add Batt Voltage and Temp pls.
I want to see how it compares with mine. And do a zoom in to the jump up. I've seen occaional "anomaly" but after I added the other values it made a bit more sense. Ta.
Woodstock have you got the capability to check the individual call voltages? Worth ruling out you don't have any cells that are misbehaving.
chris_p Sure, what would you like to see?
Woodstock Ah not from you sorry, I was meaning as part of the inverter logs we can see in the portal. Be good to see "OCV started", "OCV adjustment +/- %" etc etc
BrianHere Here you go ...
The sudden jump is preceded by the charging ramp down that normally occurs at 90%+ SOC; so the system knows at this point that the battery is approaching full. It is just the displayed SOC% that is in error following the earlier erroneous top half adjustment.
If the displayed SOC% was for information only then there would not be so much of a problem but the system uses a peculiar hybridisation of measured battery voltage and a mistaken belief that the displayed SOC% is always correct to control operation. For example if the displayed SOC% is under reading the system will stop battery discharge early and you lose useable capacity.
mpartington Sadly not, but apart from a general drop in stabilised voltage at full (since beta) the total battery voltage has not shown any change or anomalies. Perhaps GE could check the cell voltages from their end to confirm?
TheDragon (GivEnergy) Thanks for sorting...sorry for delayed response Ive been out of the country for a few days. I notice the battery didn't discharge and recharge. Do I need to do a manual recalibration? Thanks
Woodstock What type of battery do you have and is it a single battery?
MKICampaign AC-3.0 later gen 1 inverter and single 100% DOD 8.2 kWh battery. About as vanilla flavoured as it gets! 😃
Woodstock The reason I ask is that the voltage fall-off after calibration ends looks larger than compared to mine. However I have a 9.5kWh battery.
I picked a nice sunny day in June with the battery relatively idle after charge as an example to show how my battery behaves for comparison with your battery. See the highlighted point on the graph for post calibration resting values.
The voltage fall-off post calibration on your battery could be causing the over correction. There are a few possibilities I can think of.
I think this is one for @TheDragon (GivEnergy).
Deej No, it should be ok without, a couple of cycles
TheDragon (GivEnergy) Even though the battery is not discharging to 0% as I cannot turn off the auto charging at night.
When is this function being fixed by the way?
MKICampaign Thanks for that. My stabilised voltage at 100% SOC was consistently 52.8V up to and including 3007 but since 3011 (4077,4075,4074,4073,4071,3069,3068 and 3013) it has been anywhere between 52.5V and 52.2V. Sometimes it doesn't stabilise at all and sets off on a long slow gradual discharge! Others also suffered reduced battery capacity (lower 100% SOC voltage) under beta testing and we initially put it down to the new top end processing being overly conservative with the charging ramp down and cut off. This improved some over time (cut off voltage increased a bit?) but I am still not back to 3007 levels. Perhaps there is more to this.
Either way the top half displayed SOC adjustment was put in place to cover the relatively small number of systems that stay above 50% SOC for long periods (weeks or months at a time). I am convinced that it is the top half SOC adjustment getting it badly wrong that is the cause of my problems and a few others have reported similar behaviour. I suspect that it is only because of a small number of systems with this usage pattern that we are not seeing a flood of queries/reports.
From my perspective the top half adjustment processing is causing more problems than it solves and my system would perform better without it. I have yet to see a successful top half correction and have so far had six spectacularly incorrect ones! My battery fully charges every day when it is operating only above 50% and I would hope that the top end processing includes a realignment of the displayed and actual SOC at 100%. This ought to make any additional top half correction redundant.
Woodstock Thanks. The occasional on mine is not the same as yours and mine is explainable once you dig in. Looking at your graphs and your comments I tend to agree that at 80% there is a "hole" in the software/hardware logic.
Over to P.L.
Hi Paul, sorry to bother you again but I did the firmware update yesterday and everything appeared to work fine, woken this morning to a sunny day and hardly any generation is going on.
Could you take a look please ?
Seems the battery isn't charging and the inverter is not supplying enough power to the house.
Ok seems to have started charging now so will see how it goes
EDIT, all seems fine so apologies for the panic, I was just worried it had reverted to the problems prior to the last firmware but all seems good.
Any answers to the issues raised here?
@TheDragon (GivEnergy) What is the plan of attack on fixing the frequent sudden SoC jumps that I and others have reported a few times above? I have been on holiday so may have missed some important info, but this thread seems to have gone inactive for a while.
I see I have been invited to a new beta forum which I have joined, but i can't see a thread in there that relates to my issues.
JaseSpace The beta users have gone to the new community.
The SoC to 0% is fixed, i think. I have 3099 beta testing it, as do a few others. Right now its not wide spread beta testing, as i am taking this much more slowly, as its an annoying but cosmetic issue. The battery will discharge when needed, its only a reporting issue
TheDragon (GivEnergy) "SoC to 0%" doesn't sound like my issue. I have "regular" SoC jumps down/up (but never to 0%). I last reported this issue on 16 July above with example screen grabs showing the issue (there were earlier reports too). Looking at recent plots it looks like it last occurred for me on 2 August so looks like it is triggered by specific weather/usage patterns as not occurring ar regularly in August (every few days in July). Others reported similar issues on their systems and they thought it was a bug in one of the SoC adjustment zones, but I never saw any followup to my or their findings. My concern is that the problem goes away due to weather/usage and gets harder to repro so not easy to prove it is fixed (if a candidate fix is forthcoming).
Please revisit my posts in this thread (last one was 16 July I think and trace back).
JaseSpace Ditto, ditto, ditto!
Woodstock , @JaseSpace after switching back to GO from flux (only a week ago) it's now also appearing on my system. Both days when the solar starts to take over from the battery. Both times.it has only happened on one battery
It had several hours to realign during the GO period if it needed to, where the battery was completely idle.
TheDragon (GivEnergy) Hi Paul
Not seen anything about a new community, but my AC3 inverter is still on 289. Happy to move to D202-A202 when you think its ready. CE2225G411
Cheers
James
There’s a new beta/forum?
My inverter seems to play nice on 289, however my battery SOC is all over the place. Prior to 3013 firmware, it was rock solid but I was on a very old firmware. I suspect it’s because the calibration potentially failed.. it discharged however it never completed a full charge, it stopped charging around 30% if I recall.
Is there a way to have another calibration done tonight @TheDragon (GivEnergy)
simon_rb There is a new beta forum but it is a sub section of a new GivEnergy forum that appears to be still under test. Not sure that I really want to be beta testing a new forum at the same time as trying to report issues with existing beta/live firmware!
Interesting that both you and mpartington are seeing the discharge to charge transition provoking the displayed SOC drop when in the bottom half (<50% displayed SOC). It has only ever happened in the top half (50%-80%) on my system and started when upper half adjustment was added. Looks like the SOC drop "correction" (over correction) can take place in the lower half too!
(This time of year my discharge to charge transition is always in the upper half, so not surprising that I have yet to experience the displayed SOC% over correction in the lower half also).
Well, thanks for staying over here and continuing to post feedback! As I am pretty certain that my usage pattern would be similar to yours, as well as you seeming to have a great ability to assess the firmware, I am using you as my benchmark and won't be committing any capital to this product unless you consider it well and truly fixed. Which sounds like it's still month's away as I have not heard a single comment from @TheDragon (GivEnergy) about your current issues, and the fact that others seem to be having it too (I may have missed it as I'm losing a bit of interest tbh), other than seemingly calling it an 'annoying cosmetic issue' yesterday. It isn't, it's much more fundamental than that for anyone relying on battery power to be there when they need it. Perhaps I misunderstood him ...
I think @TheDragon (GivEnergy) is saying the down/up showing is cosmetic because whilst the display shows it, the battery still discharges.
I could be wildly wrong and would wait for Paul to confirm. I can’t test myself as we are still waiting for our system, although fingers crossed.
Once it’s bedded on more than happy to beta test AIO things.
kram OK.
Cheers I do actually have some beta AIO.
So when your ready.
Yes one battery of 2 or 3 drops to 0. As it's disconnected itself. Reporting error.
When called to discharge. It does happily.
If your already on the offical beta program you would have had by now an invite to the new beta forum too..
How does this relate to the issues Woodstock and others are reporting on 3013 of drops to 16% or lower and then subsequent jumps in the top half ?
Sounds like we are talking about different things. This is not related to multiple batteries or the AIO, and does not sound cosmetic at all
phoenix Maybe its my fault for taking a week off and not reading the many 100s of comments all over.
I will go back and read woodstock's issue
I will create a 3013 BMS area in there for beta testing discussions, however 3013 was well tested in beta and dops you describe were not seen, else it wouldn't have been issued.
However no amount of beta testing of 100+ iterations of kit can fully catch everything and in this case we have clearly missed something key. The self calibration being allowed over such a wide range maybe the actual cause of th issue. So trying to allow this as an issue in 3012, has fixed that but made it worse in other ways.
Did you when 3068 was offered for testing request it? I am not allowed to just install BMS on inverters without authorisation. So if no reply to the call was recieved, it wouldnt have been installed, and data only used from those who did.
TheDragon (GivEnergy) Yes, I think there are two issues being discussed. Mine (and others here) was the subject of my recent DM which you kindly responded to. This is certainly more than cosmetic as the under reading SOC halts discharge early, effectively reducing the battery's capacity.
TheDragon (GivEnergy) Yes, 3068 was installed on my setup after a brief email exchange on Friday 16 June. At the time I was on 3012 and seeing SoC jumps, and I interpreted your response as suggesting the updated "self-calibration" in 3068 might solve my problems. Anyway, it didn't solve anything for me.
Recent history of my SoC jump reports on this thread:
I think this thread has tried to cover too many things at once (different kit, different BMS+Inverter betas) and things have got lost along the way. But hopefully you can see from my summary above I have been quite vocal about the very obvious issues on my system that have spanned a few FW releases now. I appeciate you are trying to cover lots of things and have made a big positive difference when joining Giv, so I don't want to be too negative. Is there a problem with how the forum updates get flagged inside Giv which makes it too easy to miss them? Is this why we now have a new beta forum perhaps? Will that be better in terms of making sure all valid feedback is considered?
Just to jump in here, since 3103, my SOC had been similar to @Woodstock issue. Mine maybe because calibration didn’t fully charge the battery though. Unsure. Prior to 3103 I was on a very old BMS, it started with a 1. That was rock solid and it’s only been since 3103 that I’ve had massive spikes. This was the other day. (I have mentioned this earlier in the thread, reporting again just in case it was missed TheDragon (GivEnergy) )
Well, you definitely deserved a holiday and I hope you enjoyed it
hhall yours looks a bit like mine above, as in the big drop was when the battery started to charge from solar..
I have just posted a few example plots on the new thread specifically for the SoC jump issue: https://community-beta.givenergy.cloud/t/bms-3013-discussions/207/8?u=jasespace
(start at this comment and a few more below that).
At first glance I would say that this is the same issue as @hhall and @Woodstock. However my issues have been occurring throughout 3012, 3068 and 3013.
I've read about the soc solar start issue and the "lost" comms between AC/Batt . Have experienced latter recently but not sure about first. Then I saw an "anomaly" in my own data that I store in SQL database. Cell batteries vols reported as 0.0. Odd , so started digging. This is what I found and I believe it explains the "lost" comms and the SoC issues. Conclusion first (always what you want to read :-) ). 3013 BMS.
Config.
AC3/9.5
289/3013
Never happened before 3013.
Loss of comms between AC/Batt
If does not auto-recover needs battery disconnect (side breaker) for >= 60secs.
Data:
Inverter
reported Lead Acid, once at 11:16:47 - then reported Lithium
Nothing in portal dashboard. Expect Inv only reports Lost Comms to server after specific params.
Battery
11:14:34+ reports all cell voltages as 0.0, but total volts changes.
Before and after this 10 mins everything good.
This started when solar was used to chage the batteries and recovered when it stopped charging. Why then and not the hundreds of other such situations - you would need to see workings of BMS firmware to discover that.
Some of this images are large because they are screenshots of the the SQL data but when trying to explain things the more the better I find. Zoom in.
Phone screenshots are from battery detail graph shows straight line soc jump and the other is the solar power. Correlation on times.
Rest are data screenshots. Shows the Battery Cells and you can see the sudden report of 0 volts when solar charging starts. Refer to all 3 to see time sequences. The big power one shows the PV gen was constant, when the SolarToBattery started strangeness of SoC jumps etc.
All shows there is a "software issue" in 3013 - just too many clues and "not before 3013" to think otherwise...
ps: for images you can do R/Click, open in new tab, go see it big.... depends on you Browser of course. Small phone screens could be a prob
Now I know what I'm looking for in the data and not relying on phone app or 10min dashboard this is a repeating event now. Easy to see and consistent in it's behaviour.. Sun going in/out so giving a lot of PV power movement...
Not having to do anything to reset - just watch and wait...
When you know what to look for you find all kinds of stufff...
This is batteries. Cell volts to 0.0 [which they are not as seen by total]. 56:10 shows it being caught in the act. My data capture is a blend of time and changes and you can see some of the cells report correctly and the others not.....
Other point is the recovery time. It came back on it's own, but not 10 mins - is 24 mins.
And at the 98%+ level.
@TheDragon (GivEnergy) - Paul : My battery started at 3009 back in Nov.2022 . If you want to do a "process of elimenation" here and put it back we can see if the problem dissappears. That'll give you the extra confidence it's a 3013 issue - but with Richard's rollback removing his problem I think evey finger is pointing at 3013.
Found another clue. Can easily find the prob now. This image is for the changes to the Battery Charge Rate.
The date/time match to the above problems with the power level change.
Same this morning, but did more digging. When you look at the raw data feed it's now easy to spot the issues, but when I look at the Dashboard there is nothing in the Notificaitons and the 10min update window hides everything.
So very possible this problem has been there a long time but nobody has spotted this (apart from Richard I think as he downgraded his BMS) as the issue is not fully logged and it begins to look like there is more chance of the inv/batt comms recovering over a 30 min period than needing a full battery disconnect from the inv.
I noticed that I had the option to update my system, which meant a battery discharge/recharge, so I did it this morning as it was due to be a nice day. That bit all worked out fine, but now I read on here about problems with 3013 firmware for my battery. Should I be concerned?
My inverter is still on D0.909-A0.909 but no option to update this; it was previously updated by one of the GivEnergy team remotely, so perhaps this is why?
Having only recently got our export sorted and also hadn't gone below 50% battery since the initial installation in May I though the update might be worth doing in case of issues when the battery is nearer to 0%. Otherwise I'd have left it alone. I don't know where I could find out an update log/history, so for future reference can anyone let me know this, then I can make an educated decision as to whether I need to bother for future updates?
Spencers, The BMS version history and descriptions are in the GE knowledgebase here:
https://kb.givenergy.cloud/article.php?id=22
Spencers Whether you have an issue with 3013 is not something that seems be able to be predictable.
I'm on it with a Gen 1 inverter on D0.194-A0.194 firmware and the battery is on 3013 and have seen none of the issues others have. Might be the way I use my system - its not a huge battery (8.2kWh) so I'm lucky to get enough charge in to last until the next day (I am not yet on a TOU tariff) - so it fully cycles quite a lot.
What is the latest beta software for AC3?
Thanks all, I've saved the change log history page for future reference.
I should have said I'm using a 5kW Gen 2 inverter and 9.5kWh battery. Fingers crossed I don't have any issues with 3013 on this set up.
As all the official testing chat is now not public..
A quick update for public, as i as I pass through
the BMS testing so far.
VERSION 3014 ( Beta 3095)
No timeline for release, but will be at least weeks of testing and slowly adding more testers to the pool, from my already long list. To see if any bugs creep back in
TheDragon (GivEnergy) I'm still on 3008 and D0.535-A0.535. There's an update waiting but I've held off. Should I wait until 3014 goes live? Will it need to update (and calibbrate) multiple times to get to 3014, or update straight to 3014? Thanks as ever!
EmmetBrown my view is unless you are having issues on your current version I would stick with your current version.
I'm on 3012 now and my system is behaving (it did suffer the SoC crashing on the previous version - 3009). so i'm sticking until they fix the problems with 3013.
Mind you any new version always has the possibility of introducing new issues whilst fixing the previous ones with any system. However, my view is that this system seems to have so many variants it wouldn't surprise me if some user suffered new issues when it goes to 3014.
Thanks for the update Paul.
what is the status of Gen 1 Inverter firmware?
TheDragon (GivEnergy) good to read that things are coming on.
Where is a second programmed charging slot on the web portal interface in the work priority menu
It is essential to use Octopus Cosy well without getting into coding of some sort.
And from where it is in the priority listing, when will it be available. Personal engineering estimation is acceptable!
alfwro and the gen 1 “plant” version too?
Thank you for the update Paul.
I notice the offer to upgrade to 3013 has been removed from my portal.
Do I assume correctly that this is because a newer, corrected version is in production?
Stevef1964 yes.
There is many moving parts in the firmware world this week.
BMS work to fix a niggle in 3095 (Pre 3014)
AIO work to fix a few bugs found over the last month
Gen1 final beta tests
AC coupled Final released to the devs for upload
Has there been any progress on the issue that @Woodstock has been highlighting? The bottom half over-adjustment followed hours later by an upward correction?
phoenix not seen in 3095.
the team are hard at work adjusting code. when there is a new version, the beta testers will be the 1st to know. those on 3095 already =, priority
TheDragon (GivEnergy) Paul, I'm running 194 on Gen 1 - has worked flawlessly for me except perhaps quicker SOC drops to 3% after a SOC adjustment upwards and subsequent further discharge - and then it might drop - probably due to temperature changes. But not significant.
So, if you want to load the next iteration, not a problem.
anglefire Sure thing, I can update to the D193-A193 version.
Ahead of the final release
Also sent you a private invite to the beta forum
I've given up after waiting 35 minutes for GE support to answer my call (25 minutes waiting yesterday).
Your support team moved me from 3013 to 3068 at their suggestion (I was ringing for something unrelated). This is the result from last night.
Can you please organise a move to 3012? I'm not wasting any more of my time listening to your on-hold music.
SD2137G122
Thanks in advance.
(Timed Charge set to charge to 100% from 23:30 to 05:30)
TheDragon (GivEnergy)
Thanks for picking this up so quickly....I can see you making some changes to the system now.
Hi all, getting a bit lost between how the software/hub refers to firmware versioning and how everyone in the forum does!
My firmware is: D0.909-A0.909 and I've noticed GivTCP is broken (few folk saying updating inverter firmware fixes).
Can anyone tell me how this translates to the 3095, 4095 style version numbers people are throwing around here - in short, am I on the latest version!?
I believe the big numbers are Beta BMS (3095 etc) which will eventually become the new BMS for all batteries (3014).
The other betas are for the inverter firmwares for which there are different ones for each type, AC/ Gen 1 DC/Gen 2 DC/ etc
alphabeta77 909 is your inverter firmware. Latest version is 911. In GivTcp you may need to toggle the old firmware option.
As for 3095, that's the BMS firmware version or your battery. 3013/3012 are the latest public releases. With higher numbers being beta firmware.
Ah really helpful - that makes sense, so sounds like I'm a couple of versions out of date - how do I upgrade the firmware for my inverter? Have been on the beta, so think done remotely in the past by GE?
alphabeta77 There is no auto update past 909 right now. 911 was removed as a precaution, as it looks to be partially linked to EPS relay faults.
4074 was an older beta from Pre 3012 days.
If you let me have the serial number, i can check and update to 3012. Or if you wish to remain in the beta, i did email to all my known beta users directly, i can load the 3094, beta of 3014
I’ve been out of the loop for a while - does anyone know how the testing for the AC3 firmware update is coming along?
I also seem to recall something about moving the beta chat to a new forum, but I can’t find any links to that chat.
Hi Paul TheDragon (GivEnergy)
I'm not sure if this topic is followed anymore however I'd really appreciate a battery calibration? I have discharged/charged from 5% a couple of times this week and only getting 7kWh at best from out 8.2kWh battery. When we updated the battery firmware last time, it didn't complete is calibration and suspect I'm missing it from the top end?
If you can run a calibration, 23:30 would be ideal if it can be set in advance. Thanks Paul.
CE2133G240
PS. Feel free to update the battery if there is a newer beta to test 😉
Anyone else been having SOC drops to 0% on 3015? I'd had none at all on the previous firmware but they're nearly a daily occurrence since getting on the new version.
Not experienced one yet. Tonight will be the first test, battery forecast to hit 5% before the next low rate slot.
pelowj nope. Not had soc changes on any firmware version tbf.
Had 3015 since it came out - and 3014 for a while too.
pelowj went down to 7% without an SOC correction.
I've run 3015 since the 24th and had days running on the top end of the battery, charge and discharge in the 60% to 100% range. Historically, this would see a correction the next time the batteries went below 30% or the batteries had chance to run OCV.
28th to 3rd running at high end, 3rd to 5th running low end with no adjustments. Balancing still sucks and see drift between the two batteries.
BrettDavis hey, did you get any response anywhere on the AC3 firmware status?
I'm on 3015 on my 9.5 GEN2 and it went down to 1% 44volts last night, im sure it would have hit 0% had i not noticed and kicked off a forced charge from grid. i've tried setting a higher reserve but the battery just ignores that and keeps discharging. i much preferred 3013
cbowen isnt that just the same hardware unit but with a few parameters changed/firmware update?
"replacing my earlier 3.6kW hybrid Gen1 (SA2142G078) for a 5kW Gen1 hybrid (SD2247G111)"
Hi Paul
@TheDragon (GivEnergy)
Thanks for upgrading the firmware on this previously, all worked great until yesterday and we seem to have an issue of no volts with 70 % charge not charging or discharging now.
We reset the inverter which hasn't fixed this issue we also spoke to support to see if they could upgrade the firmware to 3015 which we have other batteries on and appears to work well.
Can you please advise the best way forward, happy to stay on the beta version and provide feedback if we can get the system to wake up.
Thanks again.
I see the battery is on a older beta. I can upgrade to 3015 for you now.
Then give support a call when the battery is nearly flat (Best time) for a calibration to be run)
Dualelectrical Also, it looks like the DC MCB is down, as the battery is fully charged, but inverter can see no voltage
Been a while since there was activity on this thread. I am considering a purchase of the 9.5kW battery to go with the Gen 3 inverter.
I've tried to follow the two threads on this SOC issue but it's near impossible with so much content. Can anyone tell me if the issues are limited to certain hardware (Gen 1 vs Gen 3 for example)? Or if the latest firmware upgrades have resolved this?
It’s best to ask GE directly but I’ve not seen recent issues flagged. I’ve noticed that the people with issues are NOT quiet! So I’ve assumed all is well on the latest. But since we have a 9.5kWh that’s only ever had minor issues I’ve not been reading very closely nor updated yet.
It wasn’t version specific.
I'm still on the 3013 Battery firmware and D0.909-A0.909 Invertor firmware. I've ignored the update firmware option over the winter as it's been working well and I didn't want any new issues while we were away on an extended holiday.
How do I know what firmware it will update to as it doesn't say anywhere? I'd prefer to read up first before doing it and decide whether to leave it alone or not.
