All In One BMS Release 8

All, I have today received release 8 of the All In One BMS.
Based on 4 rounds of testing, this has been send as releasable to the public, I however still wish to test this in line with previous tests, to ensure it is not too different from tests 3 & 4 which were good.

Understand its high level, but rest assured progress is made, testers are providing some good data, from their system, which has led up to this release.

Once I get authorisation to release public, It will be pushed to the portal as a matter of urgency

Is there anything specifically being targeted for improvement? If it can address connectivity drop outs that would be excellent....

TheDragon (GivEnergy) I sent you an E-Mail but thought id double down on heads up on the off chance you are in tomorrow before 10am feel free to perform the update and calibration on my system, going to be out tomorrow and ill let the battery discharge just incase overnight. Its under 20p all day til 4pm and 6hrs is usually enough time for it to finish, I actually remembered to check pricing for tomorrow after getting back from Hospital.

Thanks TheDragon (GivEnergy) my update and battery calibration completed this afternoon.
Will report back if I see anything of note, fingers crossed your inbox remains empty 😃

treborUK Same here lol, although since the comms dropouts have eased since new dongle firmware was updated I only dropped a line when the SOC dropped to 0% and when I went consecutive days id post timeline of charge in/out and what the original full charge from flat was to see any variations at 100%.

They check every few days so thought id help out by pinpointing when anything out of the ordinary happened to save them a bit going through a few days of data 🙂

Not been good at home since last thursday, to and from hospital every day since so missed 2 emails with the next test firmware so hoping they are in tomorrow and can start it off before 10am, if not ill have to see what agile is like next week and ive been pretty absent from checking up and seeing how others are getting on on here.

How's this last round of testing going - are we any closer to its release and will it address the connectivity drop outs?

Jellybaby Is there new dongle firmware? How to I get that?

GaryS I don't know if its new but a different configuration was loaded to mine on 21st of november, was informed that a different dongle firmware was loaded that can auto reboot it if lost comms was detected but it was designed to be for comms between inverter-portal not dongle itself and portal but since then it had been rock solid unless my network switch needed power cycling but that is easy to figure out as the hue lighting app wont load any lights.

GaryS What do you need changing?

If you raise a ticket with support, so I can track your request

Is there a date in place for fireware update to take place??

Testing has been underway for over a week.
The results are not unanimously to release.
A small loss in capacity has been found, maybe due to cold but cannot rule out the firmware is to blame.
Idle voltages are higher than expected when flat. Indicating its not fully discharged.

Another round of tests are now required. Adding another week.

Please understand this has to be right, release half tested will just result in a different issue to the one already seen..

Paul.

TheDragon (GivEnergy)
Thank you for keeping us up to date, I think a lot of the frustration has been the lack of regular updatess as to the progress being made. Strogly agree that it is better to take more time and get the update thoroughly tested before release.

This has been released to installers.

Portal release in the new year

Does BMS 8 address the problem of the battery suddenly dropping to 4% (mine has being doing this for the last 10 days, typically dropping from 10-20% down to 4%) almost everyday? I think I saw this mentioned in another thread but cannot now find the relevant thread.

SteveUK The rate at which the SOC drops is a little more consistent but you may see faster drops than you would expect at some points (By this I mean in one hour period you might see it go from say 80-75 but next hour it may go from 75-65) as the SOC is out with what the remaining battery life you have left is but you shouldn't be seeing the behavior you have currently in regards to extreme flatlining and I have found that it does go down pretty well, ill post some samples below of days when it has emptied.

30th December

27th December

I am another one hoping this version fixes the SoC issue and not a cover up. They have had 7 chances.

My head is saying because there are 4 x batteries in the AIO (and possible 8 if I get another one) the SoC issue is more difficult so sort as if one or two batteries report there full and another is still charging is the software cleaver enough to compensate, also visaversa, if 2 are empty and 2 still have juice left does the software just throw the towel in and drops the % to 4% (Idle)

The Powerwall is one battery and can already have batteries 2 and 3 etc added, so what is the really issue? Or am I just on a different planet 😀👍

Here's a screenshot showing the type of drop I'm getting most days, in this case from 20% to 4% at 19.20 so we were powered by the grid for the r
est of the day.

Its the new Year

Daveb01 Powerwall 2 has NMC chemistry (as opposed to the AIO's LFP). One of the properties of NMC is that it's easier to estimate the state of charge.

Avenir
That's something I wasn't aware of. There seem to be a lot of positives with using LiFePo so hopefully the monitoring technology will continue to improve.
I did wonder about the monitoring kit they are using in the AIO. With the comparatively recent dramatic shortage of electronic components and devices they might have had to settle for whatever was available in order to get the Units out of the door, and perhaps future units will have more up to date, capable and sophisticated control firmware. If that was indeed the case, then it would be nice to think that we could all look forward to a visit from GivEnergy / Installers to swap out the monitoring system.

Avenir

Not so sure it's a reason for bad SOC we see here though, Many others manage it perfectly fine, Tesla cars with LFP seem to be okay for instance.

From a personal perspective I have a Anker LFP 2kwh battery with separate added extension 2kwh and it never loses a beat on its estimation, as logically it should be easier with separate batteries to spot and correct issues in the BMS.

DreadorUK I make no excuse for this debacle! Just responding to the person who asked how it is different from the Powerwall.

Many others do manage it perfectly fine, including my GE 9.5kwh – also LFP – which has been absolutely solid since installed in May.

DreadorUK Not so sure it's a reason for bad SOC we see here though, Many others manage it perfectly fine, Tesla cars with LFP seem to be okay for instance.

I've heard that Tesla recommend charging LFP up to 100%, which isn't something they do for NMC. Possibly because a full charge avoids this issue with tracking state of charge accurately ?

https://www.autoevolution.com/news/the-puzzling-reason-why-tesla-misleadingly-recommends-charging-the-lfp-batteries-to-100-187075.html

DD

Maybe for them, but not the issue with AIO I don't think, as I do that - I have no solar and charge 100% overnight (well I say 100% yesterday that 100% transpired to be 4.4 kWh)

Ready for the firmware update whenever it comes out. We have had a AIO for just over two weeks and it's not ever charged above 10.5kWh, despite being asked to charge to 100% overnight. Some of the readings are also rather odd, with the numbers not adding up.

Yes also awaiting this BMS firmware with anticipation. I get the drops to 4%. Today I've discharged 7.8kwh so far and sat at 26% remaining, which is not good 🙁. Just had a drop from 22-16% in 10 mins with no further load - lucky to make what was added to the battery let alone 13.5kwh.

Also battery pause doesnt appear to do anything in the app.

This one is interesting too. I keep having the numbers not add up. At nearly 1 in the morning the house is consuming 779W and there is 17W export (why???). That adds up to 796 W. Yet the battery is apparently discharging 883W. Nothing is smoking or getting hot so.... why???

SunnyWithAChanceOfPV There is always going to be a small amount on the grid side, just handshaknig with grid or it adjusting after something has spiked the use and died down and it isnt all the time but you often catch it.

Also in regards to battery discharging being more than house load you are seeing the home load without the power the inverter takes added on to do its thing, this can range from my personal experience between 80-140w and so adds around 2-2.5kwh to my old daily consumption.

Jellybaby

With the battery being advertised as "High system round trip efficiency – 92.5%", I would expect 13.5kWh import to result in just under 12.5kWk being available for usage/export. I.e. I would expect 1kWh, not 2.5kWh. Surely it's not that bad?

SunnyWithAChanceOfPV That figure is EXTREMELY misleading - and frankly that's pretty common in GE's "documentation", which is either accidentally inaccurate or deliberately misleading (you decide which).

The 92.5% figure relates solely to chemical losses within the battery during charge/discharge. Inverter losses of probably another 10% can get added to that.

Vestas
I subscribe to a few Solar PV / Battery channels on YouTube, one of which is " Tim and Kats Green Adventure "
A few months ago he was discussing the AIO and saying that on his recent visit to their headquarters they were hinting that there was sufficient overprovisioning in the AIO that some of this extra capacity might be made available to the end user in the near future. That's obviously not going to happen, unless they mean that we'll be able to use the promised 13.5kWh 😉

Tenkaykev Frankly I take anything "influencers" say as total bollox. They don't get the red carpet treatment/freebies from manufacturers for telling the truth 😉

NB - not pointing the finger at the channel you mentioned (no idea who they are), more a general rule of thumb.

Vestas To be fair there is possible overcapacity built in as the battery used to show it was a 15.96 kWh battery and 10% inverter loss should be around 14.5 so they already built in compensation, its thought thats how they could market it at 13.5 kWh usable - as no 13.5 base battery would ever give you 13.5 due to inverter loss.

The 4% is another thing and I don't know why thats there, I do plan to pester my installer around 'usable 13.5' but as my main issue is Im getting 70% SOC so that jump from 4.5 to 13 would be welcome for a starter.

Which still leaves me wondering why is it still not on the portal? Every day waiting costs those of us impacted money.

Hope to see this software release imminantly. My installer is aware I will be requesting a refund if it doesnt fix the issue completely. I've only had the promised 13.5kwh once since my system was installed in October.

DreadorUK I am sometimes seeing 14+ kWh charge/discharge according to the meter data tab so it can meet or exceed the specs for sure.

Think Paul said he was back in office today and would start the process for portal release as long a no issues had come up over the festive break with the latest version he was testing. So hopefully it will be days not weeks before it is out there.

Been on BMS8 for about a month now and last week was the first time where the battery didn't get a chance to fully discharge for around 1 week.
BMS 8 hides SoC drops better by detecting that the real SoC is lower and the % drops faster near the bottom end; but after a while, it will lose track and give a sudden SoC drop just like BMS 7

mine dropped from 15% SoC to 0% but the output got throttled to 500W when it hit 16%
The graph above had a total discharge of 8.2kWh, quite far off from 13.5kWh expected.
The upside is that after a SoC drop, it will charge back up to 100% taking _14kWh depending on how it feels.

Overall, BMS 8 has been 1000% better than BMS 7, but they haven't fully fixed the problem yet.

Blasteh This is me so far, I am currently on day 4, just got charged up on solar.

SunnyWithAChanceOfPV I track the kwh in and out every time the battery goes flat and ive found it on many occasions to be +/- 0.5kwh out when it finally goes flat.

This is where I stand today so far on day 4

Its how I used to track and plan my SOC drops on Firmware 6+7 and I used to be pretty damned close to spot on.

EDIT
Didn't even realise I replied to someone else in this thread lol, really should refrain from browsing when waking up til ive woken up a bit more after drinking coffee lol

Also when I was using all the battery in a day I pretty much got out what I put in every time, I guess it depends on how much you draw to how much the losses are, pretty consistently during the day between 6-900w draw (depending on if fridge freezer goes on)

Blasteh Overall, BMS 8 has been 1000% better than BMS 7, but they haven't fully fixed the problem yet.

Good to hear. When I raised the poor performance of the AIO to my installer, GE downgraded me back to v6. I am getting the occasional drops still, but I think the AIO is fully charging overnight - some days more than 15kWh going in according to the energy graph. And a few days of greater than 13.5kWh out. This is with a charge rate set to 4kW.

All In one BMS v8 has been authorised for public release.
Further tweaks are being made to accuracy for BMS9 in a few weeks, maybe a month or 2 to allow for testing.
SOC drops are gone.
Changes.

1. Compatible with CCS battery module;
2. Modify the SOC estimation logic; (This is the new OCV logic)
3. Incorporate Activate Battery & Remote Wake Up;
4. Fix Stuck BCU and lockup;
5. The balanced duty cycle is changed to 50% (For passive equalization of cells);
   OCV Logic operates on the charge cycle.
   Charge above 3600W, below 3300W OCV wont run in this version.

Congrats on the release many are waiting for. Would just make a polite request that GE really needs to sort out their release note language - this is technobabble - most people are not familiar with deep battery technology. Would be great to see this in plainer English so people with issues can see if a release addresses their problem.

I know I know you're rushing to get it out, just a suggestion 🙂

BMS v8 is out now and available for all users to install in their All-In-One system.

V8 addresses the issue where the battery charge rate would drop/increase suddenly. It also addresses the issue where the full battery capacity could not be utilised.

I dunno what the rest means so can't translate it 🙂

Thanks for the update.

TheDragon (GivEnergy) Charge above 3600W, below 3300W OCV wont run in this version.

Can I understand this comment better please?

Today I can set my AIO to charge at any rate between 0 and 6000W. Will this firmware therefore narrow the charge range to between 3300 and 3600W? So for arguments sake, in a cheap charge window of 3 hours, the max charge possible is now 3hrs x 3.6kW = 10.8kWh?

As per pacemaker above - would be helpful to know what all the TLAs mean.

TheDragon (GivEnergy) Brilliant! Will try asap. Does this force a recalibration full discharge and charge when you apply the update? (Just so I can time it sensibly around prices and usage)

26Left

You can set what charge rate you wish between 0 and 6000.

The self calibration maps are only coded in from 3600W upwards
CCS - cell type
BCU - Battery Control Unit
OCV - Open Circuit Voltage

Barns Yes, as per all auto batttery updates. But especially this one

pacemaker

I used to put basic speak, then got moaned at for no detail as to whats was fixed.
Noted.

Ill do both

Barns It should yes.

For paceMaker

1. The New BMS fixs the rapid fall in State of Charge readings, some users expierenced. By introducing a new method of self calibrating the charge level while charging at 3600W or higher (10 amps).
2. This BMS has remote diagnostic and restoration functions
3. This BMS fixes the rare situation the battery control unit would stop responding and no longer operate correctly.
4. This BMS now has added a cell balancing function, to help maintain battery cell balance.

TheDragon (GivEnergy)

Thanks for clarifying.

I misinterpreted your earlier comment as:

0----------------3300----3600----------6000
...............................<XXXXXX>........................... Allowable charge rates (using OCV logic)

But I think what you're saying is:

0----------------3300----3600----------6000
<XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX> Allowable charge rates
....................................... <XXXXXXXXXXXXXX> OCV logic works

TheDragon (GivEnergy)

Thank you TheDragon.

Is there an update for the Gateway software or just the EMS Battery?
Does the Gateway need an update? Don’t fix what ain’t broke.

Excellent. I'll wait for the cheap electric time tonight and then install the firmware update to see if it helps with the rapid capacity drop and the 10.5kWh max charge issues I'm seeing.

Separately, the installer just changed the communication settings over from WiFi to ethernet and I was surprised to see that the powerful WiFi access points from the battery & gateway were still present. Even if I'm not using them, having strong 2.4Ghz signals beaconing is not going to help the busy WiFi environment. When using ethernet I'd like the wifi off as this will also save (a small amount) of power. The Givenergy support person I spoke to yesterday was surprised that the battery & gateway were still broadcasting wifi when ethernet was enabled & working. He said "it sometimes happens" and that perhaps the firmware update will sort that. I'm not seeing anything communication/network related in the firmware update however.

SunnyWithAChanceOfPV as far as I know you can’t turn off the WiFi. The advice I was given was to login, set the SSIDs as hidden and set a password that isn’t the default serial number or whatever.

Tried to install this twice so far but unfortunately it gets into a restart inverter loop with five minutes between each restart.

Appears to loop for two hours before it stops the attempts and gives the firmware upgrade failed message.

Anything I can try to resolve this or is this one for the helpdesk tomorrow?

With the 3600+ limit for the correct for SOC value, what happens when I'm charging from solar at under this amount. Will it totally screw up the SOC until it goes down to 4% or would at overnight top-up at >3600 sort it out?

ma9mwah I've been been running BMS8 for almost a month now (part of the beta test group) and I've not noticed any issues with Solar charging for a few hrs a day at sub 3600W, but I have mostly been emptying the battery daily to my 10% reserve and charging to full overnight. Not had any SOC drops and still mostly putting circa 13kWh in overnight from my 10% reserve to full. Very pleased with the performance of the system now after a shaky start with BMS7.

Have just in process of upgrading software to BMS 8, started at 18.40. Took 9 mins to upgrade then the discharge started.

All good until 20:22:02 when the battery got to 0% and I was expecting it to start charging.
It seems the battery had other ideas and continued to discharge for another 30 mins.

Anyone else see this trend?
I thought it was going to stop at 4% and then start charging.
Is this 30 mins the 4%?
Any ideas what this means?

Daveb01 it needs to find the true 0% - based on voltages, when calibrating, so it's not unusual.

It'll get to 100% and carry on charging too - until it finds the actual 100% based on voltage.

When it's complete it'll then know the true 0% and 100% and hopefully charge and discharge to those levels.

If I’m not experiencing any battery issues, should I avoid upgrading @TheDragon (GivEnergy)?

treborUK AH. Not crazy about that at all. A hidden WiFi access point still pollutes the frequencies, causes congestion and is uses unnecessary power. Plus it seems to reset to default at a moment's notice, leaving my network with open access point (no default password) and the default 'admin/admin' username password for the network interface.

@TheDragon (GivEnergy) is there such a thing as a feature request, to have the option to properly turn off WiFi for better security, less wireless congestion & improved energy efficiency?

@Daveb01 , that looks like what I would expect. I.e. if my battery is only charging up to 10.5kwH, then either it does not know the correct top, or bottom, or both, so I would expect it to exceed it's 'Top' or 'Bottom' as it currently understands it. I'll be finding out myself soon enough. :-)

SunnyWithAChanceOfPV from memory I had to do a restart of the interface as part of the procedure to set/save the new password, not had any issues with it resetting back to default since.
Still an unwanted AP really but at least not open.

You can submit feedback / feature request through the portal - Leave Feedback

Waited till before off peak started prior to upgrade, battery update took around 12 minutes and confirmed successful. It is now near on 50 mins later and though the calibration said it is starting, discharge doesn't appear to have started yet 🙁

treborUK I managed to get a SoC drop on BMS 8, a week without a full discharge/charge cycle and really low solar charging caused it to think 8.2kWh was 100% capacity, so lost out about 30% of capacity until it SoC dropped and recharged back to true 100%

Blasteh After 5 days 100% for me was around 10kwh but there was no SOC drop for me, this is doing the math since last full charge

I am going under the assumption it isn't close to 0 charge left due to losses day to day.

Vez Woke up this morning, thought wha the hell, paused/unpaused started to discharge/calibrate. Interesting firmware update behaviour to leave device in some form of paused state - will revert back, still discharging now since 6am from 34%.

Updated the BMS last night - went well. From 0% to 100% 15.3 kWh charge.

All good here, though it was funny as I set it to go at 23:20 thinking the 4% would drop pretty quick and it did and started charging at 23:30 it charged for one and a half hours and thought it was at 100% (clearly not) then continued at 3 kWh until just gone 4am so based on charge speed alone thats 1.5 @ 5.1 and 3 @ 3 so about 16.5 kWh.

For context to maybe explain my data I was only getting about 4.2 kWh out of the battery with .1 removed every day pretty much for last month and a bit.

Now to see if it starts to degrade or not

SunnyWithAChanceOfPV the wifi if not used can be capped & hidden.

There is no option to turn the AP off.

New Dongles yes. After APR 24

Is it normal to have a BSM over voltage during the recalibration? It was at 100% SOC.

gbleeds sometimes, if its hit the upper limit, which a calibration can do if 100% was not reached before 330v

TheDragon (GivEnergy)

Thanks. It hit 230v at 100%.

gbleeds ok, it was finding true 100%, as it hit 100 too early

TheDragon (GivEnergy)
Not seeing the WiFi being "capped & hidden".

Not sure what you mean about new Dongles after 24 April. I have an AIO with built-in ethernet & wifi.

Not sure what I'm looking for.

I got an over-volt warning at 10:08, but the battery hit 100% at 09:12 at 324.82 volts, then rose to 329.44 volts at 09:48 and has been dropping slightly since. Yet the battery remains at 100% and is not discharging as expected.

SteveL I'm having this exact same issue as well. Have you managed to resolve this?

kckc I've had a look over the settings from the control panel and switched ECO off & on again & seen confirmation that it has been written. I've just sent a Restart Inverter command, so will see if it does anything.

I'm reluctant to send a "Reset to Default" command, as it often results in the two Wifi points becoming gaping security holes, i.e. the battery and gateway start broadcasting an open (no password) wifi access point into my IoT network, where the GE network settings have the default username & password. I'm rather appalled by what I've been reading in this community about the lax attitude to network security that Givenery have, especially as they have been repeatedly warned. I'd not be surprised to see it ending up in court at this rate, with someone who doesn't know about VLANs/segmenting networks and who put GE in their only network which was then opened wide up to any passer-by. :-/

SunnyWithAChanceOfPV I've actually already tried the "Reset to default" button but no joy. BMS Communication Fail and Battery Voltage Low comes up every time it restarts. I did wonder why the Wifi has become open, that explains it 😅

TheDragon (GivEnergy)

I'm a bit surprised the battery % isn't dropping. Firmware updated and it fully discharged at 06:45 and fully charged at 09:12 with max voltage at 09:48 and over-voltage warning at 10:08. It's been powering the house since 10:10, there is no grid power coming into it

The "Energy" Dashboard view is a little slow updating, as it's 12:26 and the 11:30 and 12:00 time periods aren't calculated yet.

kckc "did wonder why the Wifi has become open, that explains it 😅"

Yes. NOT funny. I wonder how many folks are going to be badly caught out by this security failure of GE. Very few are geeks on community boards that know about securing open networks. It's a serious matter.

The default 'factory reset' or firmware update should at the very least should have a default username & password set to access the battery & gateway wifi access points. Then there should be app & email notifications strongly urging users to change the access point passwords and network web access passwords on both devices, plus ideally encouraging them to changing the usernames as well.

SunnyWithAChanceOfPV my understanding is that when the battery gets calibrated it takes it all the way up to 15.98 kWh. Until the battery has discharged down to less than 13.5kWh it will show 100%.

kckc unfortunately my AIO is in an unrecoverable state after these failed attempts, spent some time this morning working with James from the support team and they are going to have to send an engineer out next week to swap the battery packs.

15.3kwh to full charge here, shame the firmware ended pausing my AIO last night :/

SunnyWithAChanceOfPV I'd not be surprised to see it ending up in court at this rate, with someone who doesn't know about VLANs/segmenting networks and who put GE in their only network which was then opened wide up to any passer-by. :-/

Well, yes quite.

ATM I have my AIO on my IOT vlan. But I am seriously considering creating a vlan specifically for it, as it is important enough to my home infrastructure to need protection, but has such poor security that it could easily act as a portal to its entire network. It should be the most secure IOT device, not the least!

PianSom

I have done just that.

vLAN 1 - UniFi
vLAN 2 - IoT
vLAN 2 - Energy (SolarEdge Inverter, AIO, Gateway, EV, Chrager, SmartPlugs)

I have also followed what was suggested in other posts:
Changed the WiFi name
Changed the PW
Changed the security protocol
Hide the WiFi

Do not join the Energy vLAN with WiFi, leave it as its own AP with its own IP range. If for any reason the LAN breaks you can still access the AIO via the AP. This is the best and most secure option you can have, without disabling the WiFi.

If anyone needs help please let me know and I will post the settings.

Daveb01 Been considering yet another VLAN (think my count is presently 5), but decided to leave in IOT realm, and tighten up other device security. First thing I did was lock down and hide the wifi, this should be the installers responsibility, but I had enough issues at the time with firewall permissions for all the new kit, as I run a very tight ship with what can and cant get access to the internet or LAN from inside or out.

Vez

I thought about putting my solar stuff on IoT vLAN as well, however it is not really an IoT device, plus my IoT vLAN has a setting to isolate all devices.

TheDragon (GivEnergy)

Thank you TheDragon.

Is there an update for the Gateway software or just the EMS Battery?
Does the Gateway need an update? Don’t fix what ain’t broke.

A quick update about the Battery Firmware Version 8 update. I performed the update and it seemed to work smoothly, with the app notifying me at various stages of progress (Update started, Update finished, Calibration started, Calibration finished). From the time the battery was down to 0%, the fill-up took about 3.5 hours, with the 'complete' message coming in a bit after. I miscalculated the timings, so don't yet have a clear/clean day to check, but from the calculations I've done it looks like the battery has taken in a lot more energy that it usually does. Normally it will max out at 10.5kWh, but it looks like it's charged up to 13.36 or 14.3 depending on what's compared. At the end of tomorrow I will have a 'clean' day to check total charge and how the final % discharge goes and will post results here, for you cautious/sensible 'wait and see' upgraders. :-)

Well mines already looking a bit flakey today already used up 10 kWh so far and showing only 11% left 🙁

Update topped out at discharge of 11 kWh and delivered 10.4 kWh only and now back to drawing from grid

Since I had BMS8 I’ve been tracking SOC at start of day and end of day, working out what that means in terms of stored power based on 100% being the reported 13.58kWh, then comparing with the daily Charge & Discharge values.

I have 30 days of data and the average difference based on SOC and calculation has been 0.6kWh. Usually the system SOC at end of day shows less than the calculation from start plus charge/discharge but have had 4 or 5 days where system has reported more than the calculation comes out at. Biggest difference either way has been 1.3kWh.
Assume the difference is ‘running cost’ conversions etc.
seems quite stable and still putting over 13kWh in on overnight charge from 10% SOC.

DreadorUK First day after calibration was always a little weird for me but it seemed to stabilise a bit

atm full charge for me this time around was 13.1kwh, total charge 14.3 and total discharge 9.4 and math wise charge left is 4.9kwh and its reporting 51%, quite under what it should be but as the energy goes down further I expect it to start falling a little quicker.

Jellybaby my last 11 overnight charges to full from reserve 7-10%

treborUK where are you pulling the data from. The app/portal or meter readings..?

treborUK I had not thought about keeping track of full charges, I think I might have to do that but aside from first day ive not gotten over 13.5 I don't think

couplands taking data from inverter METER DATA tab in portal.

Been looking at first SOC recorded of the day, any discharge recorded before my overnight charge starts at 0400, charge to full, then any charge or discharge through to the last recorded data for the day.

BMS 8 certainly seems to be an improvement although I don’t have much data to go on, looks like we are close the 13.5kwh rather than the 8-10kwh we’ve been getting up to now. Happier so far!

Just had a 32% Drop to Zero (it didnt even drop to the reserve @ 5%). 12.4kw Battery Discharge, 9.49kw to home. Can others explain how and where they are getting their stats please (mine are from the web portal widgets)?

Not convinced we are any nearer.

@TheDragon (GivEnergy) Can you assist?

Vez

I was going to answer but then I don't know how much you imported on solar for it to be contextually correct 🙂

Vez

To add, overnight, the battery added 10.14kw.

DreadorUK

Solar:

0.77 kwh to battery
2.2 kwh to home

For the battery to have been at Zero of the useable 13.6kwh, added 10.14kwh overnight, Im going to fall well short again.

Vez
I’ve found the only reliable data to use is “Meter Data” on the portal. Are you on BMS8?

Goshiki2

Thanks, why is the view totally different between mobile/PC browser. Either way, the data stacks up against where I am using the stats from. Yes BMS8.

Vez Please contact support to have the battery checked. Incase there is an acutal issue here.

Vez
In my experience data is all over the place. The daily and weekly reports are not worth generating as 90% of it is meaningless. The app reports a wildly different solar generation to the meter data. To be fair, I think their priority was getting BMS8 released so they’ve probably not been focussing on much else.

TheDragon (GivEnergy)

Have just send through an email request, at 0920hrs. forgot to add my forum username on it....doh doing now.

12.7kwh discharge yesterday to 5% so about on the money if the 5% is considered part of the 13.5kwh. I would still like to be able to use the whole 13.5kwh as advertised though.

1.7kwh discharge to 90% so far today which doesnt tally though.

Ben Taylor-Wright

3.0kw discharge to 86% so far today, at least yours is somewhat realistic!

Goshiki2

Oddly on my setup this figure is way off the mark showing 7.5 but it imported for 3 and a bit hours at 4 kWh and home page is as bad - So I go with time charging at charge rate for how much has gone in.

Oddly discharge and energy today figures are more realistic though

DreadorUK

Even using this logic of charge rate * time, the actual charge is still approximately the same for me.

Goshiki2

Ah, now that is good to know. I had a look at the Dashboard/App vs Meter Data

That’s quite a lot of difference between the different sources. I'll aim to use Meter Data from now on.
Hope they align the figures between Meter Data & Dashboard/App.

SunnyWithAChanceOfPV
An extreme example but look at this comparison for solar from the 1st December.

SunnyWithAChanceOfPV v7 is rubbish. I had the same problems when first installed back in October, Constantly got 10kWh or 11kWh rather than 13.5kWh.

After a while - and THREE calibrations - they said that issues had been found and I was rolled back on to v6 which was fine (until whole unit failure in December).

Also leakage to the grid was up to 100W on v7 - much better on v6.

Thinking about updating the Software to the latest edition and wondered about the optimum time. As I understand it the battery goes through a discharge / charge cycle. Battery is currently at 49% and it is 5:30pm so I can set the process in motion at 10:00pm and just leave it to do its own thing? or is it best to start the process earlier in case things get hung up. Any opinions?

Tenkaykev

It kinda depends how much you need it - if your getting close to maximum now Id advise to keep as you are.

If your getting SOC its then going to be down to how bad SOC is. Mine took 6 hours there abouts at full speed as it hit the 100% then took a few hours finding its way back to real 100% at half speed.

So if it fails and takes 4 hours to get to said fail point, then you wont have support if you start now, or later, which kinda indicates its best to wait till its cheaper to do so.

Again I would hold off unless you need to take it though

DreadorUK
Yes, I think I will hold off for the time being. I've mentally pencilled my battery capacity as 10kWh, it's been higher on a few occasions and there are a few other things around the house that need sorting out, so I'll leave it be until I've got a clearer head.

To add, unsurprisingly today we fell hugely short with 9.2kwh of battery discharge from 100% through to 5%, over 4kwh short of what should be delivered .

@TheDragon (GivEnergy) no sign of support responding to my email.

Sorry to hear the Vez - I would go to installer to discuss over going direct to GE as they will need to do any corrective work and its under their warranty at the moment.

For the first time ever mine actually hit mark today for the first time in 3 months.

If this continues I can then start moaning about this stolen 4% part of the store

13.11kwh output yesterday, so BMS 8 seems to be delivering consitantly better results so far.

GE were in contact late yesterday via mail. Going to start with a fresh upload of BMS8 and recalibrate, hopefully after the sun is no longer providing power.

With my usual typical luck, I waited up till midnight to start the upgrade so that it would charge in the 00:30-04:30 Go window. Battery showed as 10% so thought it wouldn't take too long to discharge. Spent 6 hours at 1kW discharging to then start charging at peak rate 🙁

At least its sunny today. 15.6kWh into the battery on calibration.

ma9mwah Last calibration I ran (hybrid rather than AIO) I was surprised just how long it spent exporting at a low rate after it had reached 0% - guess it was tentatively exploring how low it could go before it hit the minimum cell voltages.

If you kick the update off in the morning, will it charge up using solar + grid or just the grid?

DD does anyone know what the min/max voltages are for the AIO ?

ma9mwah
You can see the battery voltage on the power graph.
Select the power graph in the portal, then select the grey "Battery Charge" button. Click on "Battery Voltage" and it will be added to the graph. Mine varies from 308v @ 10% to 330v as it approaches 100%.

Decided to initiate the update last night as my battery was down to 5% at about 8:00pm. There was a period of inactivity before the process started, of interest was that the battery was exporting from 5% at 2930 watts for 90 minutes then continued exporting at about 1400 watts for a further 50 minutes. Just after midnight in started a charge cycle at about 3200 watts for 4h 40 minutes before dropping to about 990 watts for about 20 minutes. Charging finished about 05:00 and the reported battery charge since midnight is 15.4kWh. There was a short spell during charging where the immersion heater kicked in but this demand was met from the grid and didn't affect the battery charging.
A promising start.

Tenkaykev I also took the plunge last night to try and overlap calibration charging with the Octopus Go cheap window of 00:30-04:30.

I had been downgraded from v7 to v6 previously and so my AIO had been behaving better.

22:07 pushed the update button - SoC 49%
22:23 update completed, calibration started, discharging at 6kW
22:40 SoC 40%, discharging at 4.5kW
23:00 Soc 23%, discharging at 1.5kW
23:50 SoC 9%, discharging at 0.6kW
00:10 SoC 0%, charging (at my max rate of ₄.2kw)
03:05 SoC 95%, charging rate starts to taper
03:15 SoC 100%, charging at 3kW
03:40 SoC 100%, charging at 1kW
03:55 SoC 100%, charging complete
04:30 SoC 100%, dynamic discharging as scheduled

Energy graph reports

• 13.08kWh of discharging yesterday
• 14.52kWh of charging today

Externally mounted, and a cold night, so battery temperatures were on the low side. Started discharging at 6C, started charging at 13C, finished charging at 20*C.

Dashboard notification confirms firmware and calibration complete.

🤞

26Left
That's good to hear and as you say 🤞 My AIO is in the out house so it is sheltered, only a couple of metres from the back of the house. I too was a bit concerned as it is unheated. Just checked ant the temperature low point was about 11.5 degrees at its lowest late yesterday evening and then once the update cycle commenced it remained pretty constant in the 18 -23 degree range.
I've got a small ceramic fan heater from Screwfix, was going to couple it with a Smartplug so I could switch it on remotely, but I think it may not be necessary. Not sure if the reported temperature from the Battery is from the circuit boards or battery modules.

Vez Re-callibrated yesterday (2nd time on BMS 8).

3.8 KwH drained and at 86% battery :S

Im going to be a little sad today as Im going to have spare capacity which is a first for over a month - damn this update 🙂

I just need Octopus to agree to let me to export so I can dump the unused for a profit

Updated to 8 a few days ago, 2am when I was due to start Octopus charge at lowest rate it lost comms....

woofy

It is weird that it is charging up to 100% and the battery voltage is 334.59V and later 316V is still 100%. I know that the voltage will settle after charging, but that is too much I think. Also as on then on the same day 317V is 85%. And yesterday 316V is 38%....
Not sure this is how it should be or it is stg wrong.

Vez SOC is a bit out of whack and unreliable in giving an accurate representation of how much is left as when you are on your last few kwh it starts going down faster than usual but at least no massive SOC drops due to no charge left at all and massive capacity drops.

This is me so far

Day 2 of BMS 8 saw my system charge exactly 8.0 kWh going from 44% to 100%.

Morning,

After my second calibration under BMS 8 it appears I have a problem.

Reported discharge on 14.1KwH (I stayed up to observe this, and it wasnt done discharging either, though it was at this time down to _600wh. I have the data below taken at random points in the day:

Key Facts

• Battery Calibrated @ 3674w greater than that specified by @TheDragon (GivEnergy)
• 14.1kWh discharge with 13% battery remaining (and hadn't stopped discharging).
• Discharge Didn't stop at 5% battery reserve (see table column H)
• 12.1 Overnight charge to 100% - It is clear the charge logic takes perceived remaining percentage and charges accordingly - in my case this is a major issue.
• Today estimate capacity based on man maths 10KwH, with predicted sudden drop (see previous posts - same behaviour).

@TheDragon (GivEnergy) I will email support and my installer as I believe there is something fundamentally wrong with my system.

Tenkaykev Day 2 of BMS 8 saw my system charge exactly 8.0 kWh going from 44% to 100%.

That seems absolutely spot on including a little overhead for losses 🙂

Is anyone noticing more frequent inverter timeouts/WiFi drop outs on BMS 8? I had a patch applied by @TheDragon (GivEnergy) but since the update my inverter goes offline for 7-8 minutes frequently which is messing with my Home Assistant automations.

Vez Just to add on this, I did force the discharge, there was no sudden drop in SOC, but didnt expect it not to stop at 5%? Is this expected

Forgive my naivety with all of this. New to the game. I've done the firmware update and it's currently discharging those precious kWh. My question is, it's at 0% and still pushing exporting 500w at 0% for the last half hour/hour. Just wondering how long this goes on for before it goes to a charge cycle. We have a nice clear day here today and the solar is already pumping at 3kW so was hoping the charge cycle would be happening at peak solar hours.

Josh
Hi Josh,
When I did my software update a few days ago the batter reached 0% ( reported ) and carried on discharging for a further two hours before starting a charge cycle. For a while it was discharging at just under 3kWh so there was a fair bit of capacity left when 0% was being reported.

Tenkaykev Perfect! I don't ever remember this properly calibrating so we'll see. Just keep looking at my solar generation and it's all being exported...ah! As long as it's bumbling along nicely without issues it's all good.

Only other thing that's odd for me is my Gateway is always Idle and doesn't have any firmware.

Josh
Yes, mine too, it's always been orange under the " My Inverters " tab. Everything is working so I decided to leave well alone.

Josh
We've got gas CH but It's off at the moment. Bought a couple of electric heaters as it seems daft to have the CH boiler on when there's so much solar generation ( 3.5kWh at the moment)

Tenkaykev Hmmmm that's a good idea! Wonder if I can get them to be set off an export clamp like the iBoost on my cylinder!

Josh You'll need some logic to stop the battle for excess solar. I did consider putting in a storage heater to sink excess solar but since getting the AIO, I'm exporting very little as the iboost eats most of it after the battery has charged. Had to configure the iboost to 300w cut in to stop it conflicting with the AIO and then seeing/using the battery as excess solar though. Still running v6 on my AIO as it seems to be behaving well. I'll let V8 bed in a bit more I think before I do any upgrade.

I considered a Sola Diverter but having watched " Gary Does Solar " video on the subject I decided not to. I have a Smart Switch connected between the double pole isolator and my HW tank immersion heater. It's controlled from my phone and I have it scheduled to come on during the night when the Agile rates are most favourable. It's just a button press in the app on my phone to toggle on/off so when we're generating excess solar I switch it onfor a daytime boost. I purchased two heaters, a compact 2kW ceramic fan heater from Screwfix, and a 3 bar oscillating quartz heater from the same place. The fan heater is quite noisy so I've put that in the kitchen which is the coldest room in the house. I just give it a quick 20 minute blast first thing when I get up to make coffee. The quartz heater is a real hit, 3 x 400 watt bars each individually switchable. It's surprising how warm it makes our living room with the added bonus that we've been able to really cut back on the Gas CH hours.

Ultima357 I had to do the exact same thing with the iBoost. Was default 100w and always seemed to go off randomly so set to 500w and that stopped!

V6 seemed to only be causing me issues with disconnects. Been having them ever since install. Hoping this fixes or at least alleviates it. I will have a look at trying to change the channels on the Wireless AP again as this has solved other issues with IoT devices in the past, doing so with the GE devices always seems to default back to channel 11 on reboot.

I've been on BMS 8 for two weeks now and it's been working OK until last night. I had my first SOC crash from 22% to 0% and it only charged up 11.4kWh to 100% overnight.
![
](https://)

Tenkaykev mine as well, only upgrade option on firmware is the AIO nothing on the Gateway ....when it's actually on line....

The Black Cat Interesting that yours is diving off the edge at circa 271v. Mine has not really dipped below 290v and today did a dive from 24 to 5% at circa 302v. Only gave me 12.1kwh too whereas it normally is giving 13.5+. Will see what it charges up to tonight. V6 firmware BTW.

Jellybaby As a follow up to this im at around 2.5 mark again

Glad im still keeping manual track of this.

EDIT
Looks like I had a bit more in the battery than I should oO

Ultima357 Well last night it topped up 12.65kwh to 100%, so circa what it gave yesterday down to 5%. Beginning to wonder if this soc drop off is susceptible to the power being drawn. Eg, yesterday I was pulling power out continuously at 1kw to 4kw, peaking at 6 (kettle on), as it was a freezing morning and ASHP working full time. Cold this morning but not as bad so will see what happens today if I empty it.

Today saw my battery charge up in q couple of overnight slots. A total of 12kWh to go from 20% to 100%.

Dropped from 11% to 0% in the early hours this morning and only put 10.8 kWh in overnight to get to 100%. Looks to be on a slippery slope again.
![

Below are my battery status calculations made since the BMS 8 recalibration on 28th December. The percentage figures are based on a capacity of 13.5kWh, though when you look at GIVtcp, that seems to work on 15.96. All figures have been taken from the meter tab on the portal. The initial battery level figure of 14.7 is based on the amount of electricity used during the calibration to raise the battery from 0% after the discharge to when it finished charging at 100% (it was 16.4 kWh) minus the battery discharge for home use until midnight on the 28th.

The Black Cat I don't believe ive seen a 0% since going on 8 myself and I was part of test group, last night though battery got to 5% with -0.9kwh but tonight instead of finding 1kwh ive lost one lol

BMS8 has not allowed the math to be very consistent as it had been on previous ones.

13.8kWh charge last night from 5%.

More problems yesterday, that's 3 days in a row. The SOC went down to 5% about 21:20 last night and dropped us back onto the grid. However, yesterday we imported 10.8 kWh into the battery and only used 8.2 kWh from the battery. 5% of 13.5kWh is 0.68kWh, so at the minimum we are missing 1.92 kWh just from the energy put into the battery yesterday. The battery dropped from 57% at 6pm last night to 5% at 9:20pm but only used 2.8kWh in this period. Battery charged up to 100% overnight but only put in 11.6kWh, so expecting another stressful day!

Morning
My pattern these days with my AIO is that I will use all the battery during the day (most days) so I am cycling the battery every day, empty to full and back again.
Since BM8 I am getting 11.4 - 11.7kwh out of the battery on a regular basis - 12-13 goes in.
With what seems like so many variables and ways of interpretting the data, I never feel confident that I know exactly how much is coming out of the battery from the grid or solar etc.
Bring on summer

Early days but my feedback 9 days in no solar so fewer outside influences as such. I also 100% use whats available every day so far to make this consistent.

I have one issue where whilst battery charge W is being reported correctly capacity stored overall in kWh is not - however my battery to home figures do look about right for both W and overall kWh.

So my main check point is actually just the load time full charge end and subsequent trail off time

I also prefer to track this time side of things as I then use that during the day to estimate my expected capacity (due to no solar throwing the calculation)

What's interesting is I started to see the capacity drop in the way I was seeing with 7 aka daily decrease in charge but what I see now is nos sudden drop, so the battery seems to know it didn't put enough in and so correctly displays % down to 5%.

I was about to throw dummy out the pram but then it seems to have some checkpoint as you can see as it fixed itself so I have picked it up and washed it off and will wait a little longer.

Still early days and Im also not getting 13 kWh fed into the house (4-5% thing meaning no 13.5 kWh) but today Im mostly expecting 13 kWh with this mornings feed in.

Had my first issue with BMS8 yesterday. SOC dropped straight to 0% and then started charging at about 400W back up to 5%, so it's actually worse than BMS7. Very unhappy.

Alibloke 4 days on from the v8 upgrade and ditto.

SoC drop from 33% to 0% in the space of 5 mins.

AIO has been consistently charged overnight to full at a rate of approx 4.2kW. Intervening nights got as low as 54%, 44%, 29% and 27% without issue.

@dragon2905 Very frustrating - feels like we’re back to square 1.

Maybe I should go back to v6 (AGAIN).

I must say I am quite shocked that a few are having 0% drops or straight drops at all as ever since I was moved over to the 3-4 tests leading up to the 8 before release I didn't encounter straight up drops and only had the 0% drop on one of them and didn't see anybody else that had been on the 8 have any issues.

Jellybaby that's presumably exactly the frustration givenergy are facing: no matter how much testing they do, when they deploy, new problems manifest.

DD Indeed, from the sounds of the other people that were testing it the SOC drops seemed to have been all but gone, hopefully they didn't do anything from the BMS8 test version afterwards that may have had an adverse side effect.

The only problem I need sorting now atm is the inverter idling for 7m after stair chair is used and that is something only a GE Engineer can look at since its to do with their equipment, I have no idea why a stair chair that is powered by batteries which get charged at the top and bottom of the stairs makes the inverter go idle and to cut all readings except for load power but its not something my installer or a local sparkie can check out it has to be looked at by a GE Engineer.

One thing you might try if you have access to an AM / FM portable radio is to tune the radio between stations and listen for interference when the chair makes contact with the charging station in case the switching / charging circuitry is " noisy " and somehow affecting the inverters electronics and it takes a while to recover.

Yes DC motors can emit a lot of wideband radio frequency interference if they have brushes that spark.

Either that or the charger draws a big surge or makes a big spark when it connects.

nophead
@Tenkaykev
It doesn't happen right away, forgot to mention that, it happens give or take 3m after it reaches upstairs.

I actually tested it and brought it up manually twice in a row and waited before I noted when mum brought the chair up and when it beeped to indicate it had reached end and when it went offline and there is a margin of error on the time it took due to the refresh rate as I said.

If the power goes out in the house due to grid outage the chair can still be used as it runs on batteries in the chair, 2 of them side by side and the contacts up and downstairs are basically battery chargers.

It is 100% replicatable and hard to pinpoint time between any further as even with the app open sometimes it ticks up to 20s before it refreshes.

I load the app up and go by the refresh after it goes beep to indicate it has reached top of stairs.

I gave them a call today and the issue had been sent up the chain and would get back in touch but I feel like only a GE Engineer is going to be able to find out what interaction is causing this issue and if it can be resolved.

It could be something to do with the charging point upstairs but I don't want to spend several hundred quid to try and replace something that may not be a problem, maybe something could be adjusted equipment side to allow for whatever is going on but only they would know.

Are the charging points interchangeable? I.e. could you swap them to see if the fault follows or is due to more work being done on the uphill trip.

Can you isolate the power from the top charger and see if it doesn't occur and then reconnect it and see if it occurs then?

26Left
I wonder if your issue is being caused by the low temperature. Paul did allude to this earlier in this thread. LiFePo4 batteries do have lower capacity at reduces temperatures, not as much as you are seeing but if there is an issue with the low temp calibration in bms8 that might be at least part of it. My battery is at 16.8C and has no issues, is yours in a room that can be heated - just as a test.

nophead They are not, I was going to try and cover top one but I have no idea how long the batteries last without a charge and don't want mum to get trapped on way downstairs.

This is what upstairs looks like
https://media.discordapp.net/attachments/572589136456974337/1196581827972956321/20240115_222801.jpg?ex=65b826ab&is=65a5b1ab&hm=b6c21f4b1569083edf1e5fa34d4d43641c0b470421da4dbbadb29db724cf12f5&=&format=webp&width=1095&height=616

Looking at the size of it I am surprised it could cause a problem for the inverter.

nophead Indeed, its why its so confusing as to why, this is how I confirmed my original suspicion.

https://cdn.discordapp.com/attachments/572589136456974337/1196605753390268516/image.png?ex=65b83cf3&is=65a5c7f3&hm=6e1788b9cc49bd482217f2d74c60a591263cac43c02b280d59e04258502e0622&

That last bit is meant to say "Not sure where battery discharge bit come from as it was empty at 5% and had been running off grid as you see"

woofy it’s external mounted on a house wall. The battery temp hasn’t been below 5C.

Tenkaykev If Jellybaby uses VDSL2/G.fast (FTTC) as his internet connection then DC motors can cause a resync of the line if they're noisy. Depending on the router (and whether its G.fast) then 3 minutes to resync/bring back up the internet connection/wifi seems in the ballpark....

Whether that'd have an effect on the AIO - who knows. Easy enough to check if the stairlift is causing problems with the internet connection though.

Vestas
The Company that I used to work for did Sea Trials support for the RN/ MOD. Nothing ever surprises me when it comes to wierd stuff happening with regard to Electrical Systems / RF.

SteveL yeah mines been dead since November due to a stuck update. Engineer came out and couldn’t fix it. Waiting for a date for them to come and replace the AIO. Only had it a month when it broke. If I could send it back and get a Tesla power wall I would.

Has anyone else had issues with the SOC going the other way?

It charged up to 68% then it went straight to 100%. Strange! First time I have seen this. Not seen any SOC problems before.

Very Odd!

Josh can you post a picture of the battery page with the voltage on it. im guessing at the 68% the voltage hit 335ish

I had similar the other night, The SOC hung at 33% for 40 minutes whilst being charged at 2.2kW before it started to increase. Once it got to 63% it shot up to 100%.

Below is a screenshot of the battery page - battery voltage was 331.4V max before the SOC shot up.

Why do you suspect 335 ish??

32% to ZERO (not even the reserve) after 6.9 KwH discharge.

Mine refuses to go over 330V, in fact it will trigger an over-voltage error every time it gets to 330V and then it discharges for a bit before resuming charging.

Vez similar here - 31% to zero after less than 6kW out. That’s two in three days for me now.

ma9mwah

Here's the battery voltage page!

It seems the SoC issue has come back for some people?

What do people think, is it the way we use the Battery? Or something else.

1. Those that charge/discharge every day gets a “works for me”
2. Those that charge up each night (plus if solar during day tops it up), leave it in dynamic to supply house (there will be some battery left) then after a few days there is a SoC drop.
3. Something else?

Daveb01

I come under 1 and yes its working with no SOC yet for me but still not supplying advertised usable 13.5 kWh Im getting 12 kWh which for life of product will impact me by around £1,500 if not resolved.

Daveb01 I come under option 1. Easily use a fully charged AIO by lunchtime, unless I hold some back for saving sessions then it gets used up (exported) during the savings session. I’m not updating from bms 6 until I see more positives. I also charge overnight at half speed not sure if that’s relevant. I have seen SOC drops and jumps but not since winter kicked in.

I fall into category 2. After three bad days in a row, my battery has been OK since the 13th as there has been sufficient solar around to keep my battery topped up during the day and only going down to a minimum of 38% (according to GIV) at midnight.

26Left

To add more into, 13.6 KwH addeds to tha battery to 100% overnight....

After a discharge of 2.5 KwH today, Solar charged the Battery by 1.6KwH to reach 100%. So straight away -0.9 KwH difference without losses!

Daveb01 I'm definitely group 2. Mine is set to fully charge at 4.2kW each night. Most days we don't use the full amount, get a bit of solar top up on in the middle of sunny days. Seemed to be much better behaved on v6.

Vez hmmn. For me it only took 8.7kW from 0% to 100% overnight... I've set it to charge at full rate ahead of the 1600-1730 saving session, and then discharge 1730-1830 at full rate too. But I think a reversion to v6 and re-calibrate might be the best answer at this point.

Looking at the saving scheduled for this evening, I just logged onto the battery to see what was left as I noticed my smart meter display showing grid use and literally watched it go from 30% to 5% in just 2 minutes while it gave 0.019kwh. Maybe madness has got into V6 too. Only 10.24kwh from 100% too even with 0.57 solar charge during the day. Mine never dips below 292v at full discharge either which though set at 4% does not go lower than 5%.

This is all very grim. Is the consensus V8 isn't fixing the issues? I'm about to decide on my batteries for a new build and was thinking 2*AIO but the support for this isn't there yet and I can't see it coming out until these SOC issues are resolved.

26Left Did the battery only just discharge to 0% just before it started charging to 100%?
I get lower 100% capacity if the charging starts soon* after it finishes reaching 5%

*the battery hasn't had time to settle and stabilise voltage

Daveb01 1 and 2 are both true from my experience

I'm in category 1,

I have been on the beta program for version 8, and until 3 days ago all the SOC drop issues had gone away.

I've had SOC drops for the last 3 days in a row, the first thing that springs to mind of "what's changed" is the weather has been very cold with subzero temperatures and my battery is installed outside.

Been fully charging overnight due to low amount of solar being produced right now, with v8 and had no SOC or reducing capacity issues.
As the solar starts to increase I’ll be reducing the overnight charge to a lower SOC to leave room for charging from excess solar. Will monitor the system as I start to have more lower rate charging from solar and see how that impacts the SOC tracking.
Hopefully by the time the weather and solar improves to the point where I don’t need to charge from grid overnight the issue with lower charge rates and OCV calculation will have been fixed in another update and it will continue working as I want.

Daveb01 My last run was 5 days before it got a full discharge and so far no soc drops for me and I was on BMS8 as part of testers rollout.

Not having one I've not really kept an eye on this much but I really thought they'd got it this time, but it appears not. Which is a bit odd as the BMS is completely different to the old 51.2V ones.

One of the root causes has been identifed. Cold, its having far greater effect on the cells than anticipated. So calculating the calibration values based on lab test at 15C is not working as designed and tested. Granted it wasnt cold and even test AIOs outside could easily maintain 15C while charging.

Another bug also found, relating to the real time clock, not SOC related, but something to address, for an even smaller minority.
All this deep dive into BMS 8, is taking the battery engineers time, but its a priority to fix.

Hi All

I am on Battery Firmware version 8 with software version D0.609-A0.609.

2 days ago my battery went from 100% down to 8% discharging 12.8kWh. Yesterday it also went from 100% to 8% but only discharged 10.2kWh. To me that is a huge difference.

So far today it has discharged 1.7kWh but still shows as having 94% SOC.

Any thoughts?

TheDragon (GivEnergy) assuming they are in Shenzhen they may need to buy a big freezer to stuff one in!
it's rarely below 10'C over there.

So how was operating temperature range on the marketing material advised as -10c to 50c. So the product material and certification has not been validated?

TheDragon (GivEnergy) thanks for the update. Are those of us with an outside AIO installation better off going back to v6 in the meantime?

Having some strange behaviour with charging again. Solar is pumping today on my 6.8kW system and getting a nice 4kW. It's got to 68% again and acting strange and started to export to the grid with a few hundred watt still going to the battery. Never saw this on previous firmwares but never reach 100% charge before Firmware 8 I don't think.

Temperatures are cold and this morning the battery did go as low as 2.4*C this morning.

A pre-requisite for a battery install was to have outside.

And it's whizzed up to 100% again. Weird!

Another SOC drop in the early hours this morning from 24% to 0%, fortunately it was at 1:45am.

Battery is in the garage but it's temp was only 10 degrees when it crashed, well within the advertised operating conditions.

The trouble is I don't know where the bottom end of my battery is, so I'm having to guess if I can get through the 4pm-7pm peak flux period each day.
The figures above all come from the meter tab and represent the reported charge and discharge figures since the battery recalibration was run on 28th December when BMS 8 was installed.

The Black Cat Can't really figure out from that sheet but are you keeping track separately of in/out kwh, its what I still do despite having little problems so far as I still find the SOC to be out compared to charge left and whilst not as accurate as it used to be it is usually +/- 0-1kwh out each time it runs dry, sometimes it finds an extra kwh or sometimes it loses it but its not much.

If you don't do this it may aid in finding out when the SOC will drop for you or if you have enough to cover a certain period.

It is odd though, the 100% is what it took to full charge after depleting and what it is right this very minute (sun is far enough to not cover home load anymore so starting to discharge but still says 100%)

0.7kwh difference but that could be just because slow charge and SOC being at the upper end, it went from 2.5-2.8 whilst generating over 1kw and it was just slow discharging 75w all afternoon.

Hi all. New AIO BMS 8 SOC jumper here joining the community. 👋

@TheDragon (GivEnergy) I appreciate the transparency and the regular updates I can see you’ve been providing

TheDragon (GivEnergy) can this not be solved in a similar way as Tesla. Cold weather kit and heat mode? Albeit current cold weather will be a distant memory by Sunday probably. The reason I ask these questions is I am bout to buy some of the AIOs.

pfmorris57 and also as per nearly a year ago. Different solutions for this. Force discharge at a low rate to keep heat in AIO overnight

pfmorris57 It is slightly concerning! I don't have the space internally for a battery system and predominantly looked at the Tesla Powerwall. I was recommended GivEnergy by the installer and trusted his experience. The warranty and cycle count was the clincher.

I do feel a bit of a beta tester but hopeful that GivEnergy iron out the issues. Bit concerned my batteries aren't fully charging and when charged 8kWh they then show 100%.

I'm glad there is this space here and seems quite active!

Hi all found this today in graphs (extra tabs). There are loads but think these are the only extra ones needed. It shows the up and down effect of Voltage and battery temp etc.
Have a pay with the settings as some may work for you.

Press the button that looks like a chessboard and go to graphs in the cloud. Then click the small down arrow on right

Well the battery chemistry is definitely not as efficient at low temperature, particularly if you want to pull a heavy load so maybe it's not exactly surprising to see some dips in performance in cold weather. My PHEV BMW has pre-heaters activate before charging and cooling in hot weather too. Can't measure kwhr out of that but range basically halves in cold weather due to heaters, lights etc.

I still find it odd though that some of you are discharging down to circa 280v as mine, (V6) never dips below circa 292v. Mine is in my insulated attached garage, so not subject to the extreme weather.

Given the cell capacity is handily printed on the front on the AIO (52Ah) if your really really bored you could work out the capacity curve vs temp.
52Ah is such an odd size that you can probably assume it's a Gotion Cell given such slim pickings in that size, however you may spend more time learning Mandarin (or it'll be summer first!)

Daveb01 How dare you call the waffle a chess board 😃

It's a barbeque fish slice ...

Another shortfall in charge, 11.4 kwH to 100% from 5%....

After a full discharge gave me 8.6 KwH. It appears if you mix solar into the variables, then it really struggles to figure out charge.

But back on topic.... its a Rubix cube!

Vez

This is why I posted this comment above and I am 2 in the winter. Not had a SoC crash yet.

It seems the SoC issue has come back for some people?

What do people think, is it the way we use the Battery? Or something else.

1. Those that charge/discharge every day gets a “works for me”
2. Those that charge up each night (plus if solar during day tops it up), leave it in dynamic to supply house (there will be some battery left) then after a few days there is a SoC drop.
   2b. As above with export as required
3. Something else?

Daveb01
I’m in 2 but not had an SOC drop in 6 weeks since running v8. I am seeing some variation in battery voltage though, a lower SOC some days has a higher battery voltage, even taking into account whether it is idle or not. I’m still putting in around 13 kWh on overnight charge at 4500W from 10% though and it is sorting itself out.

I think temperature is impacting some systems, and maybe charge rate too, apart from a couple of days I’ve not had much solar above house load to charge at low rate for any length of time.

From what I’ve seen on the forums v8 is generally a good improvement, but there are still some issues that GE have mentioned with charge rate and more recently temperature. I’m sure GE are working these issues and I think there are less people impacted than were with v7.

I'm on BMS8 (also went to 0.609 from 0.605 at the same time), and it's much better. No SOC drops yet - my AIO is in the basement, so temperature is pretty stable if that's relevant - and I used to get them all the time previously. Also the amount I'm accidentally exporting is much lower (assuming the inverter is more responsive), and I don't know if this is related but I used to sometimes get RCD trips but it hasn't happened in the last few weeks since the update. All in all, I'm much happier on this version.

I have just gone into my 5th day and whilst the 21st reported 100% at just over 11kwh down from the 13.1 flat>full charge it varies day to day but this is where I stand right this minute, unit is inside in a cupboard.

I am hoping it runs dry for a fresh flat>full charge, I did set a 4kw charge last night down from 6 to spread it out a bit longer but for some reason its still ignoring anything I set and was charging at 4.5.

System went from ignoring discharge limits to ignoring both charge AND discharge limits.

EDIT
Added flat results, not too bad, my battery temp does stay above 20c though.

10.3 KwH Discharge (to 37%) and a top up of 8.8 to 100% - predict another day of shortfall ahead.

Took a dive off the high board today after giving just over 9kwh. Strange that the battery voltage was circa 300v when it did this and not at the usual 292v.

5.9KwH Discharge today till now and reading 36% after I have been out all day. Lets see todays SOC drop and shortfall. Will report back tomorrow.

We only got 8.97kwh out of it yesterday, which appears to be the first time since the update we've had it massively underperform. For the previous few days we haven't been totally discharging it (its been quite sunny and we've been out of the house), so I don't know if thats related.

My post above- so it is looking like number 2 situation is the one now causing the most SoC drops.

Number 1 is the most stable?

This could get worse as we move into more sunny times and we all start using the sun to charge up the battery?

I'm hoping the next update that fixes the known issue with lower charge rates will come before the better weather and I start getting prolonged periods of solar charging at low rates. Only had the odd hour so far with solar excess over house use.
Currently 48 days in on BMS8 with no SOC drops, no capacity issue and average nightly charge to full from 10% is 13.4 kWh.

As we start to get more solar production I'll be gradually reducing the overnight charge ceiling from 100% down to (hopefully) no charging from grid until next autumn.

Is a SOC jump something that people are seeing now? dont remember having this on BMS7. Suppose its better than a SOC drop. Could it be because it was charging via Solar?

I've just had my first SOC jump under BMS8, I've had a few SOC crashes before. I thought it was all going very nicely today with the solar charging my battery from an overnight charge to 60% and then bang, 80% to 100% and we are running around having to turn everything on as I'm still waiting for my export meter to be set up.

Looking at your battery voltage, 328.3v is about here it gets to a full charge, I think the 100% is correct and the 80% is out. Has the battery had a few charge/discharge sessions without reaching 4% or 100%.

woofy Yes, because I've had a few SOC drops from around 20% to 0% but it has been charged to 100% most nights because of the poor solar. It just makes it impossible to manage this battery. I normally try and stay towards the upper end of the battery so it doesn't crash on me during the peak flux period.

I’m having a really good time with BMS8 so far, my AIO was replaced on Monday because a previous update bricked it.

During its calibration it’s discharged 13.5kwh and then recharged 16.1kwh.

So far no SOC drops, remaining percent seems consistent with capacity. Twice this week we’ve got about 13kwh out of it when it reached 4% and charging back up to 100% it’s been putting about 13.5-14kwh back in.

Previous battery and BMS I could barely get 10-11kwh out of it.

We don’t have a massive solar array as the panels came with the new build so it’s only about a 1kw array so the solar never recharges the battery we just use it. We use most of the battery most days.

If it stays like this, I will never update again.

I also don’t charge it at the full 6kw. I charge it at 4kw.

JasonF Last two times I've charged mine at 6kw (octopus free power session) it has dropped Comms. I'm not sure if it was coincidence or something else but next time I hit it at full charge and if it drops out, then I'm sure it will be the cause.

Ultima357 my gateway and battery are connected to Ethernet and I’ve never experienced the comms dropping even when charging at 6kw

Caught 2 problems in a single night:
1) lower end SoC drops faster than actual energy discharged, it went from 21% to 5% and discharged 1.2kWh instead of ~2kWh - this effect is sometimes more severe, and when the battery has really lost track it results in a SoC drop.
2) charging from 5% to 100% after a recent discharge to 5% means it won't fully charge, I got 11.6kWh instead of 13.5kWh

click for full graph

A "quirk" with BMS 8 is that when the battery does do a real 100% charge, I get over-voltage errors, if it doesn't give these then the battery didn't reach true 100%

Here's an example where my AIO does charge to true 100%, there's a good 1+ hour gap between discharge to 5% and the start of the charge. You can see it does a bunch of little discharges near the end, these are the over-voltage errors when it hits 330V. This one charged from 5% to 100% with 13.7kWh

click for full graph

woofy "Looking at your battery voltage, 328.3v is about here it gets to a full charge, I think the 100% is correct and the 80% is out."
I calculated the kWh in the battery when it reached 100% and you were correct, it had reached 16kWh, so how can it be 20% out?
I've attached my spreadsheet of daily battery charge/discharge totals from the meter tab:-

I'm using 13.5kWh as the battery capacity, I'm not sure whether to use this or 15.96 as that's what GivTCP and Predbat seem to use.
I also had to make a downward adjustment of 5.2kWh on 18th January as the calculated capacity of battery charge/discharge was getting up to around 20kWh which was ridiculous and that seem to bring it in line with the reported SOC % until the SOC jump yesterday. I have double checked all the values reported in the meter tab since I moved to BMS 8 and re-calibrated on 28th December, so where did that 5kWh go?

My wife got home in the early hours yesterday and plugged her car in - as I hadn't planned for that Octopus Intelligent started a charge and the AIO dumped its entire charge into the car. Obviously a bit fraustrating but what was worse was it was only 8.96 kwh.

So we seem to be back to where we were before the BMS 8 update..

Ben Taylor-Wright With intelligent I schedule my car charge to be complete by the end of the cheap period, the same period as I set the battery to be on charge. This way the car can never suck the AIO dry.

Yeah I had this issue once when I first got the battery. If the car needs longer than that cheap period to charge and therefore the battery is outside of the charge times (this would be pre 11:30pm for me as I tell octopus to be done by 5:30) then it can consume the battery super fast. I have an automation set in home assistant to detect when my car starts to charge outside of the normal overnight cheap period and then it just disables the battery discharging until either the slot finishes or it reaches 11:30pm.

As a compromise between giving Octopus free reign over scheduling and avoiding draining the battery, I ask for charge to be done by 7am, when we wake, but set max battery discharge to 250W overnight. That's enough to cover typical load from 5:30 but if the car does charge (as it did this morning, until 6:30) it doesn't take too much from the battery. Eventually I'll get round to deploying some code to actually read the planned charging schedule.

I think it’s time for a party. Today my AIO gave me 13.5kwh, usable for the very first time! This is from 100% to the 5% cut off.

I’d been hovering around 13kwh.

So it is possible….

Let’s see how consistent it is.

Ultima357 yes we often do this however occasionally we need to charge the car right up to the 11am cutoff (she sometimes works nights). Often have to do quite a bit of juggling with all the apps!

Now here's a bit of a twist. I upgraded to V8 over the weekend and since then sunshine has been really good and kept it well topped up but today being grey, it bottomed out at 5% whilst cooking dinner this evening having given only 9.9kwh. I'm beginning to think that if you are pulling high power, circa 4.5kw in this case, it sees the big dip in battery voltage and prematurely flags empty. Then the twist. During the evening it jumped up to reporting 7% charge at 308v (same as at the steady state 5%). It had dropped to my usual 292v when it first hit 5%. Pre the dinner cooking load it looked about right, 8.4kwh given at 44%soc, but then the ASHP kicked in on top of cooking and it rapidly headed to 5%. So I thought, I'd set it to a scheduled discharge of 120w to use that last 2% but it didn't want to play. So I'll see what charge goes in to it tonight.

I think there is definitely something in speed of charge and discharge. I now charge at 4kw and get pretty good results. If I discharge at 6kw (at the end of the night to empty the battery before recharging) I don’t quite get out of it what I think I should.

Any battery will have less capacity if you charge or discharge more quickly because more energy is lost as heat in its internal resistance and power loss in a resistance is a square law, so it will be significant.

nophead Wouldn't that be built into the overall battery capacity though as it's sold as a 13.5 @ 6 so not fit for purpose if it can't get close to the capacity its marketing pushes.

Ultima357 It charged 13.5kwh. So maybe the bottom was correct and previous top wasn't.

JasonF

I got the AIO because it can charge/discharge at 6000w. If we are all going to reduce this rate I should have gone for the Powerwall & got two batteries today.

The other reason is, I really want 2 AIO’s and need it to charge at 6000w in the winter to be able to charge both in the cheap period on Flux (2-5am)

It needs to do “What is says on the tin” 😀👍

Okay so January how did everyones AIO do on 8?

Mine a lot better but still not hitting capacity (not once)

So down 1.4 kWh on average for the month (well month on 8)

I did note the last 20% was more flakey in not supplying 20% and it would degrade quicker than usage - so say house was pulling 2 kWh at said 20% remaining which on my average that would be 1.8 kWh (15%) but it gets down to 5% in around 20 mins (so not a SOC drop more a speedier decent)

Again for clarification no Solar just load over night then use during day

Pretty well I think.
I've been charging at 4600 kWh overnight.
I have my reserve set to 10% as we get quite regular power cuts and I felt that should cover EPS operation in most cases (now regular SOC drops seem a thing of the past I may review this and lower it).
Through January but ignoring the GivBack event day and the 3 or 4 few days where there was solar top-up and the battery didn't drain to reserve, I've averaged 13.3 kWh discharge to home per cycle. Highest 14.3, lowest 11.7. Only 2 days were below 13.

January was pretty consistent but this mornings charge was...... unexpected

Pretty consistent throughout the month but last night didn't even get close to 12.

I've had more problems yesterday and today. Yesterday it dropped quickly through out the evening from 55% to 5%, dropping us onto the grid but only 2.5kWh (about 18%) was discharged from the battery and today, it moved rapidly from 75% to 89% (about 1.89kWh) when only 0.1kWh was put into the battery.

My system seems to be behaving a lot lot better under the new firmware. I keep a general eye on it and unless I see a big discrepancy I don't dig too deep. Last night saw a scheduled overnigh charge of 12.4kWh from a reported starting SOC of 20%. I'm using Wonder Watt and am really pleased with the way it picks the cheapest overnight Agile slots. I also set it up for last nights Saving Session and it worked flawlessly.

Tenkaykev

My AIO is much better on this software version as well, but not got any automated system yet like Wonder Watt.

Do you think Octo Agile is better than Flux? I have left mine on Flux to see how it goes.
There is also the new Tracker.

I did not really want to keep changing every 6 month, but if your experience is a good one I may consider changing at the end of the summer 2024?

Octoprice says
Export best in order,
Intelligent Flux
Flux
Fixed
Agile

Import
Tracker
Agile
Flux

The trouble is what I would like can not be set as Flus needed import and export, it can not be split.

Daveb01
I was on Flux as my install was in November and it took a few weeks to get the export meter ready to go. I find Agile to be, so far a much better tariff, my bill as down almost 50% since switching;

Tenkaykev I was on Flux as my install was in November and it took a few weeks to get the export meter ready to go. I find Agile to be, so far a much better tariff, my bill as down almost 50% since switching;

I think flux is usually considered a tariff for the summer months, when generation exceeds consumption. At this time of year, most of us are net importers, and therefore focusing on the import prices is probably a win.

Daveb01 If your battery can last for the majority of your home load then id say agile over tracker as its pretty much always been below tracker during the wee hours.

I’ve had 0 drops on this firmware. Pretty consistently hitting 13.5kwh (according to the reporting numbers). It’s lasting me all day now without worry, which on firmware 7 it was not even getting close to all day, barely managing 10-11kwh.

I also this year have been keeping track and then at the end of Jan started noting down what was left after a flat > full > flat charge.

After a few days of drawing less than full power out of the battery due to mild weather, some sunshine and Octopus power ups, today was cold grey and miserable, the AIO gave me just 8.8kwh before declaring it was at 5%. Not impressed as it's now bottoming out at circa 300v and full charge at 328v ish. On V8 too. After quiting, it goes back up to 308v and declares 6% reserve too. So I've raised a problem with GivEnergy. Incidentally, on tonight's power up I again lost comms just after hitting the battery with 6kw charge and bumping other stuff up, so drawing just over 10kw. I'm beginning to think the cause is voltage drop as it dipped from 242 to just 223-225v. Anyone else have these issues? I am end of line in a small village so I guess voltage drop is expected. WiFi is not the issue as good signal level at the AIO.

I've been using a combination of the excellent " Inverter " app from bbcbasic.co.uk ( iOS only ) and the Givenergy portal to monitor the charge. I use the GE app to check the reported battery percentage at the start of the charge window, then the Inverter app to see how many kWh went into the battery overnight. It's been pretty consistent, last night for example the battery percentage was 58% at the start of charging, and the Inverter app reports 6.4 kWh of charge.

+1 for the Inverter app. Simple, but wonderful

Tenkaykev Wonder how accurate the Givenergy app compares to your app? I do wonder how accurate the Givenergy one is sometimes as it's not easy to see and sometimes doesn't keep up to date.

As I understand it the " Inverter " App reads directly from the inverter in almost real time. The main screen shows energy flow and there's a rolling 10 second timer display that counts up to ten seconds then rolls over to zero and starts again. That's the screen that I have displaying on an old iPad that sits on the sideboard as a real time display. I think that it updates more frequently as I've heard the microwave go on and seen the display update almost immediately, likewise with the thermostat on the fan heater. There are four simple screens, I noticed that the one that displays the charge / discharge slots is showing the discharge settings I entered on the Givenergy Portal for the last Octopus savings session.

Just for reference, that app albeit only showing up in Apple App Store is well supported across most platforms:
iOS: https://apps.apple.com/gb/app/inverter/id6443976497
Android: https://bbcbasic.co.uk/inverter.apk
Windows: https://bbcbasic.co.uk/inverter.exe
MacOS: https://bbcbasic.co.uk/inverter.dmg
Linux: https://bbcbasic.co.uk/inverter.zip
Raspberry Pi (32-bit): https://bbcbasic.co.uk/inverter-rpi.zip
Raspberry Pi (64-bit): https://bbcbasic.co.uk/inverter-rpi64.zip

Hi all,

I have started to see a pattern.

If I charge my AIO up between 2-5 and it runs out during the day (no sun), I get good results. This includes any export between 4-7.

If I charge my AIO up between 2-5, and it is sunny, then the SOC leaves Earth and goes to a different planet. Then after 2 days of this I get a huge SOC crash.

So the new software seems to not like a full charge, then discharge until sun comes up then charging, then discharging after sunset, then full charge. If that makes sense.

Here are the three days with app pictures.

Daveb01 It's been mentioned that charging at 4kW rather than 6kW seems to give more stable results, though I guess it's not quite enough to get all the way from 0% to 100% in hours. (But perhaps if it stops crashing to 0, it won't have to.)

DD

I understand what your saying but the specs say 6000w, and this is one of the reasons I got the AIO, if and when a second AIO is available then charging at a lower rate is not an option.

EV’s can charge at 7 kw and above so what’s the difference?

Daveb01 I've been watching the discharges too and some of the issue I think is if you take a heavy load out as it nears the last 30 to 40%. Taking a heavy load in this case dips the battery voltage and maybe that is the cause of the crash? That said, I don't know how the algorithm works but I suspect that it will be monitoring battery voltage. Yesterday mine only gave 8.8kwh and after it said game over at 5%, it afterwards creeps up to 6% and 308v. So not fully discharging?

Daveb01 I understand what your saying but the specs say 6000w, and this is one of the reasons I got the AIO, if and when a second AIO is available then charging at a lower rate is not an option.

I don't have an AIO so only repeating what I've read here. With 2xAIO, they'd charge in parallel, so that would be 8kW.

EV’s can charge at 7 kw and above so what’s the difference?

They can go a lot faster than that on DC! Are we heading towards 350kW or something stupid ? But they do have batteries even bigger than an AIO. (No, I'm in no way jealous !) And probably at a higher voltage.

But charging tends to slow down as the battery gets full. Which didn't obviously appear to be happening on your charge graph: I was looking to see if it slowed down as it neared 100%, but not obviously. Which I thought was interesting...

[An analogy I heard somewhere was of filling a bathtub with a firehose: when the bath is almost empty, you can throw in water at quite a high rate; but once it gets full, you have to throttle back the rate or you'll get water everywhere.]

Maybe if there's a big enough buffer at the top of the battery SoC, then it can still accept charge at 6kW even as it nears (reported) 100%.

The charging "sweet spot" of the AIO is mentioned on GE's latest youtube video https://www.youtube.com/watch?v=oChIkOkpf2I at 4:43. States "try to keep your charging rate above 3.6kW, ......... sweet spot is 3.8 to 4.2kW, as that will aid calibration". I recall the Dragon has said something similar on one of the Facebook groups.

I have a 4kWp array (2011) and not going to be charging via solar at 3.6kW at all going forward.

As with all other existing AIO customers, the battery sold to me had a charge rate of 6kW - anything less and we don't have what we've paid for.

Ultima357

Exactly the same behaviour, 4 good days, BMS has no idea what its doing with solar. The ONLY time it is reliable is if it does a full discharge.

It should manage itself, I got only 8.5KwH out of mine yesterday, DESPITE it being charge to 100% twice, once over night and once during the day (for a number of days running).

How do competitors like Tesla handle this?

Genuinely interested as AIO is the direct competitor.

Hook
Tesla Power walls are far more sophisticated. They have a liquid thermal management system for starters. The battery cells also use a different chemistry (Nickle manganese cobalt, IIRC), which is an easier chemistry to measure SoC. Having said that, their Model 3 and Model Y cars use the same chemistry as GE batteries and they most certainly do not exhibit SoC drops like Gae batteries do.

Zakalwe one thing I have heard about the Tesla LFP models is that Tesla recommend always charging to 100% because of the difficulty in tracking SoC. This lets the BMS reset its calibration.

DD
Yes, they do recommend charging to 100% regularly.
I've not heard of any cases where the SoC drops to zero on a regular basis, and especially not from levels like 35%. They clearly have a far better grasp on managing LFP battery technology than anything GE has ever exhibited.

So another sunny day down south, so will be looking at this issue again.

Can I confirm with you guys where to get the most accurate ready for battery charge please? I am looking at Meter Reading figures between 2-5, etc.

1. Between 2-5 am
2. Any top up via Solar
3. Discharge reading?

Daveb01 Well it's reasonably consistent wherever I look. Today had a power up session from octopus and charged up to 100% from 50% at 4.6kw. The app gives 7.47kwh input, meter reading from the cloud gives 7.4 and under the other readings from the top right expansion icon where you find the graphs, it gives AC charge. Oddly, this states 7.6kwh but that I take it is the input figure that accounts for losses (heat given off mainly) So it seems about right today after giving only 8.8 the other day. Warm day and a bit of sunshine means I'll not dip anywhere near full discharge today.

As a PS from one of my other notes above regarding Comms drop out, GE support updated the dongle firmware today so will have to see if that corrects it.

Hi all,

Battery upgrade to latest version on 6th Jan 2024.

Here is what I have found, not as consistent as I was expecting on the question of SOC when there is Solar.
(I have highlighted when there is a difference if solar was topping up the battery)
(In two cases the battery was topped up to 100% twice)

I have found another alarming statistic on the right hand column, is any one else seeing this gradual increase?

Daveb01

Im not sure I even trust that import data - mines the opposite I seem to be able to get 12 kWh out of the battery and only import 4 kWh maybe Im stealing yours?

DreadorUK

Are you looking at the Meter Data on AIO in Portal?

😀😡 you may be stealing mine (that will cost you 1 x beer)

Think I got the last column wrong that should be: (Total Grid Import Today)
I have listed (Grid Import Total) doh so it will go up (forever)

I agree, I only had a small solar array. I get great results from the battery when there’s no solar. Now the suns been out a bit, something goes wrong for sure.

Apparently the AIO doesn't handle slow charges well so kinda defeats the purpose of solar battery storage lol

This seems a tad more off than usual lol, from 5%

DD Yes, I have a Tesla M3 with an LFP battery. I've had no issues with SOC drops and have confidently driven it to 5% SOC before charging without issue. I wish the same could be said for our AIO.

zag164 The difference with cars is that for most of the time they are idle. That makes checking battery SoC a lot easier. It'll be interesting to see what happens when V2H/V2G becomes more commonplace - my money is on the same issues appearing...

Vestas

I have a Anker F2000 + Expansion battery (total 4kWh) and its depleted and charged daily the same way that the AIO is.

It pretty much runs the working from home office / snug every day and has the same load pretty much day in day out and gets down to about 15% - its rock solid in it's delivery and has not had any form of percentage drop.

However I will concede that it isn't doing 6 kWh nor does it have a huge capacity, so on the flip side to my own argument it is interesting that the power wall 3 seems to be taking ages to get to market (LFP)

DreadorUK it's just going through ENA approval now (PW3 and gateway)

Daveb01 Can you add "Battery Discharge today" to your sheet? will actually bear a little more relevance I feel 🙂

Vez

👍

I'm still keeping a daily eye on mine and it's been pretty consistent 🤞 As I'm on the Agile tariff I have Wonder Watt set up to pick the 5 cheapest slots after midnight and before 6:00 am. In the morning I use Give Cloud to see the reported SOC immediately before the first charging slot after midnight then look at the " Energy " page in the excellent " Inverter " app from bbcbasic.co.uk to read the charge from grid. I do a bit of mental arithmetic and only dig deeper if the numbers seem skewiff. This mornings readings show a charge of 8.1 kWh from 52% to 100%. I have set my overnight charge rate to 4.5 kWh.

I think they still got some work to do in regards to slower charging, mum decided to do wash and dryer load after charging yesterday and before sun came up and today I was actually surprised to find an overnight charge didn't go over 10kwh on total charge (it has not fully discharged yet)

Only change was sun charged a larger portion of the battery than it usually does since being on BMS8, will be interesting to see what leaving it to charge from solar day to day does and what the charge left will be when it fully empties which shouldn't be long.

This is after charging via solar, after last nights charge it was bang on 10kwh

Not sure why it dropped nearly 1kwh but solar has been low production today and with a base load of around 500w

Although last night with only 2-3kwh left the SOC was saying nearly 50% lol

On the pre-BMS9 now and I’m finding the SOC is wandering now I’m not fully charging overnight and draining fully each day to the reserve. Slow charge from solar and remaining largely in the 80-30 SOC range gives less voltage swing for it to work with I guess.
Had my first SOC drop since BMS8 the other night - 500w steady discharge and went from 22% to 0% instantly, followed by a self-charge to recover towards reserve.
Subsequently did a 3000W charge from 10% to full and it took 13.6 kWh so seems to have sorted itself out.
Might stop charging overnight routinely now and just charge up to full once it drops to reserve, to keep cycling the full range ? Hopefully shouldn’t be too many more weeks before grid charging isn’t necessary very often at all as solar should start doing the job.

That is if solar doesn't screw up the SOC too much as it fluctuates, 3.6kw+ charge is typically recommended here and I only have a 3.8kw array so won't be hitting that after base load is covered but if we can get a week of sun it would be interesting to see how the battery holds up over several days of just solar vs grid and how much I lose day to day.

EDIT
Since that kwh info I posted earlier battery discharge has gone from 4>5 so im down to 8.3 charge left and SOC on 95% oO Going to be an interesting watch as the day comes to an end.

treborUK
The battery voltage is of no use whatsoever in calculating the SOC. Look at the plot below which I've just included battery volts and SOC. Look at the 4 points where the SOC is at 90%, the voltage varies from 310v to 327v. At another point the voltage is 310v whilst the SOC is 83%, at the end the voltage is higher at 312v but the SOC is just 59%. The only way to track the SOC is by measuring the power going in and out, and that will always be subject to drift.
My battery is working well, but then I always 100% charge overnight which gives the BMS a reference point.

woofy absolutely agree, similar variations here. Voltage changes according to whether under load, charging and probably with temperature too.
When I was daily charging to 100% and dropping to 10% reserve I never had an SOC drop and reported SOC based on energy in/out was generally within 0.7 kWh of reported SOC at end of day.
Since I’ve been only charging to 75% overnight to leave room for morning solar and sometimes only dropping to 40% by end of day this difference between calculated and reported has been regularly 2 kWh daily with over 4kWh on the day of the SOC drop.

My feeling is that during winter with nightly charging and during summer with solar exercising the full range it will track SOC ok. For some periods in Autumn and Spring when the full range isn’t exercised for days at a time it will sometimes struggle and SOC will drift resulting in occasional SOC drops/ recalculations but as long as it recovers ok from these and usable capacity then remains to spec then I don’t have a big issue with it - which seems to be the case right now (still miles better than BMS7 was).

treborUK Well following the recent chat above I can report that V8 firmware is still not tracking well unless you go from full to empty. Yesterday, cold, grey after several days when milder temperatures and plenty of solar gain meant I never got anywhere near full discharge (and I'm still topping up to 100% on cheap rate overnight) I got just 8kwh out from 100% down to 5% when you nett off the brief input of a few sunny spells. Last night it pumped in 13.34kwh to charge it back up, so indicates that it's not the discharge that's incorrect but falsely reading full? Today again another even greyer day it delivered just over 14kwh (again after netting out PV charge) to 5% after the overnight 100% charge at 4.6kw. So be interesting to see what it puts back in the small hours overnight tonight. I charge from midnight to 5:30am leaving it floating at 100% until the end of cheap rate. Tomorrow I have a power up freebie from lunch time so won't be running it flat.

So basically I think it is OK on the full to empty cycling but when floating up and down with solar it goes adrift and more so if pulling 3kw plus as it nears the bottom. Voltage wise, mine cuts charging at around 329v then drifts back to steady state of 320v but I noted in the days of solar input, this drift back went to 317v, maybe indicative of not actually being 100%. Cuts out (5%) depending on load at around 300v now since the V8 upgrade. It used to be 292v under V6. Then drifts back up to 307v. As you say, using voltage to monitor soc is a loser and really needs a power in/power out modifier. GE stating that best soc tracking is done at 4.6kw charge input would suggest that they certainly haven't cracked this problem yet.

I'll post the overnight charge tomorrow morning.

Ultima357

So basically I think it is OK on the full to empty cycling but when floating up and down with solar it goes adrift

This is what I am thinking as well as its been pretty consistently charging near to original charge between days of sunlight but before I started manually discharging battery doing the math I should have 1.6kwh left but it was at 68% state of charge and whilst discharging it seems to be going down 1% a minute at just under 4kw discharge, wanted to do a slow discharge.

I am curious as to what charge left the math says when its done, the real test is going between sunny days in summer without charging at all from the grid.

This morning I went from 100% being 10kwh to 9.2 after topping up from solar oO

EDIT
Nice, it found an extra kwh from somewhere lol

Overnight charge 13.56kwh. Topped out at 330v ish before dropping back to 320v steady state. At present rate due to warm temperatures overnight it'll only be down to around 60% when I top it up again under free power slot from 12 to 15:30. Very little solar today. ASHP only pulling around 2kw today so a gentle discharge.
Noted that the set charge rate in the app is always exceeded by around 400w as an asides issue. It does make you wonder how accurate all the app and cloud figures are.

It's a miracle... 21% charged with almost no solar (no grid charge either) :-D

Mine has been consistently under charging on a daily basis until a SOC drop occurs (although not to full discharge).
For example, with the new calibration configuration, It took in 16.7 kWh as part of the calibration, I then used 4.6 kWh, and then followed an overnight charge of 2.4 kWh.

Solar topping the Battery back up to 100% has been very poor. For example, after 2.5 kWh usage the battery was filled to full with just 0.3 kWh.

I, as others have observed; if the battery is discharged daily.... If it doesn't doesn't, either by less usage, or by adding solar into the mix.....

Mine pretty much has been used and charged over night day in day out since the upgrade, and Im still not getting near 13.5 kWh my average is 12 kWh

So whilst I'm not getting SOC Im still not getting capacity.

Is there anyway to force a battery calibration?

My AIO only ever has around 10kWh generation to battery with a 4/5% reserve. Surely I should be generating around 12kWH to the battery when it hits 100%?

Josh Not come across it within the app or cloud. You could simply do a discharge to grid at say 500w to empty (it doesn't seem to reach full empty if pulling kWs out) and then do a grid charge at 4.6kw which is the preferred calibration rate. Mine puts in 13 to 14.5kwh in a grid charge. Most of us agree that solar random charging doesn't keep it in calibration.

Josh only your installer or GE can set a calibration off, you can ring them up and request it.

Slow solar charging really does throw the SOC out of whack a bit lol

Had a bit of solar charge on 1st and 3rd of march.

Jellybaby

Thank you for confirming my findings and let’s hope this can be addressed by GE.

God knows what my AIO will be like tomorrow as have had 3 x solar charges back up to the high 90’s as it’s been sunny then cloudy then sunny then cloudy then sunny, now raining. 😀🤷🏼‍♂️

Daveb01 It is weird how charge left at 100% gets lower and lower than what typically happens with rubbish days and overnight charging but then it ends up finding some extra kwh after the fact when battery gets a few kwh charge from solar lol

typically charge in/out is quite reliable unless im doing a saving session, will be interesting when the sequential sunny days start.

Not kept an eye on this as I thought it was pretty much OK now. Odd that PV still makes a difference, seeing as it's all AC anyway. Not sure if there is any plans for a Hybrid AIO with direct PV inputs that would fare any better?
Not a lot of "1 box does all" solutions on the market other than Tesla or Skyworth that I'm aware of so I'm sure one would be well received, especially if your short on wall space!

Watched Gary Does Solars video on the new Tesla Powerwall 3. It looks very sophisticated and with the simplified installation procedure I can see Installers pushng this to customers. I have a recent AIO install otherwise the Powerwall 3 would have been my choice were it available.

After a sunny day yesterday when solar topped it back up to 100% with 5.3kwh charge, the overnight charge this morning didn't hit peak voltage at 100% (328v). As a result, the total out yesterday netted after taking off the solar was 8.82 but only charged 7.68 overnight, hence confirming this drift down from full capacity with PV charging. Maybe also with charging from mains, but only when not from empty too. Note that yesterday it charged from empty to full on mains topping out at 332v. I'd expect the algorithm to always top out 100% charge at about the same voltage? (For clarity I charge from mains from midnight to 5:30am)

Tenkaykev A lot more expensive though....!

Ultima357
I'm not sure if 328v is a magic number at 100%. There's probably some complicated algorithm which includes charge current, temperature etc. I also charge midnight to 5:30am to 100%. For reference my graphs for today show a peak voltage of 326.6v at 93% charge and the voltage then reduces. By the time it reaches 100% the voltage is 320.5v.

hoggy Odd that PV still makes a difference, seeing as it's all AC anyway.

More PV input makes it a lot less likely the battery will hit the cutoff point/idle to rebalance, so the SoC drifts over time.

Same as the rest of the battery range in that respect.

Vestas fair point, I had forgot the SoC readjustment prefers idle moments. See what summer brings I guess.

hoggy It'd be good if someone in the thread with an AIO would confirm our assumption by maybe setting their cutoff SoC to something like 10% and trying to let the battery idle at that level for an hour or so before doing a full charge.

See if their charge/discharge figures reconcile if they do that regularly.

Hi all, here are my graphs from the days when I get topped up by sun more than once plus the day after (so you can see what happened.

Here is another interesting stat, the difference between battery charge and discharge.
The difference is over 100 kWh. (I understand there are losses between ac to dc and back again).

Also if you have an Octopus Mini you can get readings every 5 mins and I am seeing this while discharging for some reason, import as well? Any ideas (see pic below graph)

oO

Down only 12% after nearly 2kwh drop, im still getting what I should out of battery I just wish the SOC thing was just a little more accurate hehe

I've been finding that when I charge to a percentage below 100% overnight it generally puts more power into the battery than I expect. Once it charges to about 80% on solar during the day, it adjusts it's SOC % to about what I am expecting, which is good. However, because more power has been put into the battery overnight, I'm hitting 100% earlier than expected and I'm exporting to the grid at the standard export rate rather than at the peak rate from 4pm-7pm (I'm on Flux). Is anyone else finding this? I've got my charge rate set to 4.6kW, which I've seen someone mention before. Is this the best rate to charge at? Any suggestions on getting the right amount of charge in overnight?

Below is a graph of the reported SOC % against my calculated SOC % based on battery charge/discharge reading from the meter tab. Last night it charged from 53% to 80% but actually put 5.4kWh in as opposed to 3.7kWh (27%) expected, so you can see the gap at the start of the day and it has already closed that back to the expected SOC and I've hit 100% now and started exporting.

Tenkaykev The BBCbasic app is very useful, it reads the inverter meter data. Not sure how the GE app generates its summary numbers but it doesn't alway match up with the meter data on the portal.

I tend to use the BBC app mostly when at home, only use Iphone GE app when I am away.

The bbcbasic app is listed in the apple appstore under "Inverter" if anyone wants to try it for themselves.
Only downside is it does not keep historical data only one day(I use the portal meter data for historical results)

The Black Cat I've been finding that when I charge to a percentage below 100% overnight it generally puts more power into the battery than I expect

I don't quite understand what you're saying here. I think perhaps what you mean is that it is reporting a lower SoC than the true value, and that the reported SoC catches up later. Could you show the SoC during the overnight charging period - since power is constant, SoC ought to rise linearly. As an experiment, I'd try charging at constant power up to 100% and see what shape the SoC curve had. Perhaps you have old data from earlier in the year, if you have previously charged to 100% overnight?

If it is consistently underreporting SoC from overnight charging, I guess you just want to lower the target SoC so that better reflects the true SoC. Ie if your experiments suggest a true SoC of 50% is reported as 40%, just charge to an indicated 40% overnight?

LFP batteries have a very flat voltage curve, so it is hard to infer SoC from voltage. But GE seem particularly bad at it ?

Another experiment I'd try is to let the battery idle after overnight charge, and see if reported SoC changes over that time - it's even harder to infer SoC from voltage when battery is actively charging or discharging.

DD I think perhaps what you mean is that it is reporting a lower SoC than the true value, and that the reported SoC catches up later.

Below is the chart from yesterday and last nights charge:-

You can see that yesterday morning when the SOC% got to about 80% it rapidly amended itself to be in line with my expected SOC% and stayed in line with it all day. However, when it came to the overnight charge from 43% to 59% the reported SOC% is not reflecting the amount of power put into the battery and it won't readjust until it gets to around 80% today. For the SOC% to rise the 16% to 59% it actually put in 5.7kWh (42% of the battery).
Below are the charging periods from last night and it shows that the change in SOC% is not linear with the amount of power input which makes it difficult to control the amount of power going into the battery using SOC% as the governing factor. (I'm using Predbat, so it charged in several chucks last night):-
From 02:11-02:35 it put 2kWh in (15%) and SOC% changed from 43% to 45%

It then idled and from 02:41-02:51 it put in 0.8kWh (6%) and SOC% changed from 45% to 47%

From 02:56-03:11 it put in 1.2kWh (9%) and SOC% changed from 47% to 51%

It then idled from 03:11-04:41 and from 04:41-05:00 it put in 1.7kWh (13%) and SOC% changed from 51% to 59%

The Black Cat Okay, I assumed you were charging to a particular %SoC. But if predbat is in control, I assume it just adds and removes power on a timed schedule - wouldn't imposing an arbitrary upper SoC limit just get in its way ?

DD Okay, I assumed you were charging to a particular %SoC. But if predbat is in control, I assume it just adds and removes power on a timed schedule - wouldn't imposing an arbitrary upper SoC limit just get in its way ?

No, I think Predbat is just re-evaluating what SOC% to charge to as the night goes on based on PV forecast/Load changes. It uses SOC% as it's controlling mechanism.
The same thing happens when I put a charge to percent using the GivEnergy Portal "Timed Charge" setting, so it's nothing to do with Predbat just that the SOC% is unreliable.
I did try and use the "Timed Charge" on the Portal, leaving it at charge to 100% and use the time to control the amount of power I put into the battery, I think I had set something like a window of 04:30-05:00 to charge in. However, I was horrified to find that it was still charging from the grid at 06:30 when I got up. I don't know why it ignored the end time but I've not tried that again.

The Black Cat No, I think Predbat is just re-evaluating what SOC% to charge to as the night goes on based on PV forecast/Load changes. It uses SOC% as it's controlling mechanism.

Ah, okay, hadn't realised that. Interesting.
Might be possible to reimplement that to just use charging time.

Just to show that Predbat was not causing the problem, I used the "timed charge" on the portal last night. Predbat had planned to charge to 59% last night, so instead I set the "timed charge" from 02:00-05:00 to 49% as suggested by DD on the Portal. The SOC was 33% at 02:00 and it charged to 49% by 02:50 and then continued to trickle charge at about 30w until 05:00 when it was still at 49%. The meter data shows a charge of 4.1kWh was put into the battery, that's 30% not the 16% (2.2kWh) expected but in line with the 4.9kW charge rate shown on the battery tab. Why it's charging at 4.9kW I don't know as the charge rate on the portal shows 4640w, on the app shows 4633w and in the remote control section shows 4611w. How can you have 3 different figures for the same field?

Yesterday, the solar was not as good as expected, so my battery did not get to 80%, according to it's reported SOC, so it never adjusted to the amount of power that was actually input which means it's still under reporting.

The blue spikes on the graph were caused by my old laptop that runs Home Assistant grinding to a halt and needed several re-boots to get it back up and running. The blue calculated SOC takes the previous end of day SOC and applied the days charge/discharge movements from the meter tab, so it righted itself once the laptop was back up and running. As you can see, the reported SOC, which was the same as my calculated SOC the day before, see earlier post, never adjusted itself after the overnight charge yesterday morning as it never got to 80%. Yesterday, it was showing about 4kWh difference all day, today it as about 6kWh out, which is hopeless to get that far out in a couple of days.
My calculated SOC has been running unadjusted since 18th January and generally the reported SOC adjusts to this figure most days. Below is the graph for the last 7 days

The Black Cat That's really interesting... much clearer when charging at a constant rate. It looks as if there's a factor-of-2 error when charging: when discharging, the two curves seem to track each other quite well, but when charging, the blue line is going up at almost twice the rate. (Except on the last graph, around 4th Mar, the orange curve looks like it's dropping faster ?)

Your description is that, at a reported 80%, it suddenly jumps up with the measured, or true curve. Perhaps instead it's when the true curve reaches 100% that the reported curve catches up - that's really the point at which the voltage increases, and is therefore really the only significant measurement the BMS can make to correct its estimate. (Similarly down at the lower end, but if the reported curve gets down to 0, it may stop discharging without continuing down to find the point at which the measured voltage, and hence true charge level, drops.)

The Black Cat Why it's charging at 4.9kW I don't know as the charge rate on the portal shows 4640w, on the app shows 4633w and in the remote control section shows 4611w. How can you have 3 different figures for the same field?

I think it actually represents charging power internally as a % of 'C' (as in the rate at which it would fully charge in 1 hour). Or at least it does on my hybrid. (And it stores it as an integer, which means you can only store quantised values.)

In the activity log, if you select 'raw display', it shows you the internal number sent. Eg in my log I have a line:
2024-03-08 07:05:02 You Battery Charge Power Written Successfully 1143 476 476 ...
But if I turn on the raw display, it becomes
2024-03-08 07:05:02 You Battery Charge Power Written Successfully 12 5 5 ...

where 5 is 5% of 9.5kW or 475W.

The UI translates back from this internal value to a friendly power. Unfortunately, looks like the different interfaces don't quite agree on the scaling factor ? I hadn't noticed before, but sure enough, remote control and portal are showing 475, whereas the app and the log show 476. (My app asked for 476.)

The Black Cat my AIO backs off, does a brief discharge, and reports an overvoltage error when it hits 330V, interesting that yours is showing 330V and only at 49%?

I can confirm that my AIO doesn't behave the same as yours, I think yours has a fault or something. Maybe ask for a calibration?

My AIO loses track of the true stored energy when charging back up to 100% (solar or grid, and at any charge rate) when it hasn't been fully emptied to 5% and had a resting time of at least 1hr at 5%
It will charge between 12-13kWh and think it has reached 100%

if I fully empty my battery to 5% and let it sit for about an hour before a charge to 100% then it will put between 14-14.5kWh in to the battery and also hit 330V triggering an error.

this is what a full charge looks like on BMS 8, there's 14.6kWh charged from 5% to 100%
the wiggly bits at the end is the battery hitting 330V, backing off, and throwing up an over voltage error.

DD Perhaps instead it's when the true curve reaches 100% that the reported curve catches up - that's really the point at which the voltage increases, and is therefore really the only significant measurement the BMS can make to correct its estimate.

Generally the "true curve" is in the 90's when the reported curve catches up and as I say, it tends to be when the reported curve gets to around 80%.

DD (Similarly down at the lower end, but if the reported curve gets down to 0, it may stop discharging without continuing down to find the point at which the measured voltage, and hence true charge level, drops.)

I did had a drop to 5% about 10 days ago:-

It dropped rapidly from 40% to 5% when only 1.9kWh was actually discharged from the battery, dropping us back onto the grid. The "true" SOC at the end of the day should have been at 34%, that's 4.6kWh.
It all makes it impossible to manage this battery.

The Black Cat Do you have access to the voltage in the historical portal data? Are you able to make any sort of correlation between interesting changes to the voltage and the curves either syncing up or diverging? Voltage is one of the few things that the system can actually measure directly, rather than infer from history.

I don't think individual cell voltages are currently available for AIO.

DD Do you have access to the voltage in the historical portal data? Are you able to make any sort of correlation between interesting changes to the voltage and the curves either syncing up or diverging? Voltage is one of the few things that the system can actually measure directly, rather than infer from history.

Thanks for your interest. I'll have a look over the weekend

That's twice ive lowered this discharge rate as a test and twice its just gone back up

EDIT
Make that 3 times

Jellybaby presumably any changes ought to show up in either the portal activity log, if the are coming through portal, app, or API; or in the home assistant logs if you are doing controlling the inverter more directly that way.

If I'm reading the graph correctly, on the first occassion, after you reduced it, it went up first to 6kW, then spontaneously reduced to 5kW. So it does seem like there was an external agent that wanted it to be 5kW from that time.

My AIO went offline at 08:30 this morning, anyone else experiencing the same issue?

I was out and got back in at 12:00 ish and had to re-boot whole house as just an AIO reboot did not fix it.

DD nope

Daveb01 What do you mean reboot whole house? How did you reboot the AIO, if it was just the button on the side that doesn't do it every time but what I do is press button on side then flip the breaker next to button on the AIO then flip the switch in the Gateway for the AIO to cut its power.

After waiting a few seconds after the light goes out on the AIO I flip the switch again on the Gateway then AIO and press the button.

Can take up to around 5 minutes to fully come back online from my experience.

Daveb01 My AIO went offline at 08:30 this morning, anyone else experiencing the same issue?

Nope.

Maybe it’s just me, but I have never restarted my AIO. And never seen a comms fail. Been rock solid since install last Sep.

Jellybaby

I pressed the button on the side, then isolated the switch under the plastic cover, then flipped the AIO switch in the Gateway. Waited a while, then turned it all back on again in reverse order. Did not fix it.

Isolated the incoming mains to see if EPS would click in, it thought about it then whole house died. Turned isolator back on and everything came back on and back to normal.

I tested EPS and it’s now working.

Nothing seems to be in logs:

DD Do you have access to the voltage in the historical portal data? Are you able to make any sort of correlation between interesting changes to the voltage and the curves either syncing up or diverging? Voltage is one of the few things that the system can actually measure directly, rather than infer from history.

If we look at the data for 27th February when my battery dropped to 5%:-
Around 19:50 it was at 40%

Around 22:30 it had dropped to 5% but only 1.9kWh had been discharged from the battery

I've graphed the full day for 27th February where the reported SOC got to 100% on solar around13:00.

And 28th February as the reported SOC% just about righted itself

The Black Cat Hmm... no, I don't see anything obvious happening to the voltage when the reported SoC recovers.

DD Hmm... no, I don't see anything obvious happening to the voltage when the reported SoC recovers.

This is the graph from 7th March when the reported SOC% didn't get to 80% and didn't adjust itself causing a big difference in the reported SOC and the calculated SOC based on charge/discharge figures for the battery:-

I've been away and off forum for a while. Scanning what's happened since, it looks to me like the issues on AIO v8, for some at least, remain.

TheDragon (GivEnergy) All this deep dive into BMS 8, is taking the battery engineers time, but its a priority to fix.

@TheDragon (GivEnergy) 2+ months on - are we any closer to a new version to rectify?

26Left TheDragon (GivEnergy) thanks for the update. Are those of us with an outside AIO installation better off going back to v6 in the meantime?

@TheDragon (GivEnergy) Or an answer to this ^^^ please?

In the last two days I've seen big SOC drops again (and there have been at least 2 others over the last month). This suggests the useful range of my AIO is currently <8kWh. Inaccurate SOC tracking has probably been partially hidden since late January thanks to some PV generation here, thus not requiring full battery range. Now the weather has turned and solar generation is poor, the low useful range of my AIO has again been exposed.

Would appreciate any views on my options from here:
A/ Should I go back to V6 (and has anyone else gone back with a better outcome)?
~ OR ~
B/ Should I recalibrate on V8?
~ OR ~
C/ Should I sit on my hands and hope this magically gets better / wait for V9 with an uncertain release date?

FWIW my config has been steady throughout

• Externally mounted AIO
• Eco mode enabled
• Timed charge to 100% between 00:30 and 04:30 every morning at 4.2kW
• AIO updated to v8 and calibration complete overnight 9/10th Jan

Recent examples below. Thanks all!

10 March 2024

• charged to 100% from 66% took 4.71kWh - infers useful capacity of 4.71/(1-0.66) = 13.85kWh
• useful discharge 7.42kWh
• SOC drop 41% to 0%

11 March 2024

• charged from 2 to 100% took 7.55kW - infers useful capacity of 7.55 / (1-0.02) = 7.70kWh (!)
• useful discharge 7.85kWh
• SOC drop 21% to 0%

26Left Not seen any soc drops like yours since I went to V8 a couple of months ago. What I've found since is that during solar charging days where it never gets empty and can top up to 100% too, is that it gradually loses the upper limit and then doesn't fully charge (only gives say 8kwh to empty). You can see this on the cloud portal on battery data as it floats out at 100% at a lower voltage, typically 317v vs 320 when in peak condition. However, a full discharge followed by a full grid charge resets it. Try charging at 4.6kw. That was the dragons recommended level on V8. I've put in a support request on the solar anomaly. Under V8 mine has never allowed it to go below 5% at full discharge so suprised that you go to 0.

Ultima357 Interesting thanks for sharing. How did you go about achieving a full discharge / charge to reset it?

Did you ask GE for a calibration, force discharge / charge manually or just let the AIO do its usual thing via ECO / timed charge?

My above experience (suggests full discharge night of 10th, full charge 11th, full discharge again) didn't clear the issue straight away - on the 11th less than 8kWh went in overnight, and less than that came back out later that day. (13kWh went in early this am however so fingers crossed it has self-healed...)

26Left It just did its own recalibration # after I did the firmware upgrade that I timed to take advantage of the overnight cheap rate (Octopus Intelligent 11:30pm to 5:30am). And the "self heal" is just it emptying into house demand on grey cold days (we're fully electric ASHP) and then the overnight grid charge set at around 4.8kw. I charge from midnight to 5:30am regardless of charge state so it's floating at 100% after it tops out. Or, today I have a free power period, so set the overnight charge to just 75% and will then top to 100% for free during the so called octopus powerup event 11am to 1pm.
#I believe only givenergy or your installer can force a recalibration outside a firmware change.

Ultima357 You should be able to charge from empty in 3 hours. Have you looked at delaying the start of charging until 0230? From what I've read, balancing happens down when the battery is near empty, and the longer it remains idle, the more chance it gets.

But maybe that's just low-voltage batteries? I now delay charging until 0030, giving it an hour at a low state of charge, and the reported SoC does seem to drift upwards during that hour, as if it's doing something to its internal state. I need to start recording individual cell voltages during that time, though I'm aware that's not currently available on AIO.

Another experiment I'd suggest, based on some of the pictures earlier, is, on a sunny day, limiting the charging power so that it charges at a uniform rate (rather than fluctuating with the solar power available, even if it means exporting the excess), and see how the reported SoC compares to the energy going in. On one of the earlier graphs, it only seemed to grow at half the expected rate.

(Not that these things should be necessary if the firmware was working reliably...)

DD I charge from midnight because I'm also often charging the car and run the heating too (ASHP, passive house so little heat loss vs more likely to use daytime leccy at 4x the price when only daytime heating. All rooms zoned, underfloor into 150 tons of concrete heatsink, so pretty efficient). Batteries always drift back up after clocking off on full discharge, it's the battery chemistry. More so if flattened with a heavy load. Mine dip just below 300v but float back to 307v.
As for regulating PV charge, that's getting into micro management which I don't want to do and shouldn't have to. And with being an all electric household, I aim to use most of the generation and with the ASHP cutting in and out, it would be impossible to manage load manually.

Ultima357 As for regulating PV charge, that's getting into micro management which I don't want to do and shouldn't have to.

Of course... I was thinking only of one-time experiments to collect data, which might help GE actually fix these problems.

New record, can't even trust charge in/out anymore

The difference between the report SOC% and my calculation of what has been input/output from the battery has been growing over the last week as my battery until yesterday hadn't got above 80% during the week. On the 6th March the reported SOC% agreed with my calculated SOC% which had been monitoring the battery input/outputs since 18th January. As mentioned previously, my battery SOC% tends to adjust itself roughly in line with my calculated SOC% once it goes above 80%. However, over the last week, it had not got to 80% and the difference had been growing such that my calculated SOC was 6kWh higher than the reported SOC. Yesterday was the first time it got to 100% for a week on solar and the gap closed to about 3kWh but it has grown again since then, currently sitting at a difference of 6.5kWh. Most of the differences occur on the overnight charge. Generally, when I charge to a percentage overnight, more kWh are put into the battery than expected for the percent increase and when I charge based on a restricted time to put a specific amount of kWh into the battery, the SOC% does not increase as much as expected for the power input.
Below is a graph of the overnight charges for the last month which shows apart from 2 nights when a fuller charge was applied, they SOC% does not increase as much as the kWh that is put into the battery would suggest (the SOC% is all based on a battery size of 13.58).
The yellow bars are the amount of energy put into the battery based on the change in SOC% and the blue line is the amount of energy put into the battery based on the meter tab.

Over the last month the reported SOC has been on average 19% lower than the amount of energy put into the battery just for the overnight charges.
Does anyone else have the same sort of discrepancies?

The Black Cat have you reported this to GE? it's clearly broken.

Hi everyone. I just received a reply from GE to my support query regarding the inconsistent full capacity charging under PV charge (it slowly degrades to around 8kwh capacity) and have been informed that there is a firmware upgrade coming SOON. Yippee! Offer to be put on the beta trial that I've declined as I prefer a stable known environment with known problems to the fingers crossed approach. I guess if you want you can opt in to the beta testing.

The Black Cat Losses shouldn't exceed 4 to 5% so I'd call it in for a warranty check.

Ultima357 I contacted GE Support today, they offered me the BMS beta trial and I'm going to give that a go see what happens.

Ultima357 Wouldn't be so quick to dismiss the beta firmwares, for me they were WELL worth it and whilst on the BMS tests I consistently got at or near capacity instead of 1-3kwh after a 5-100% charge.

Jellybaby If I had huge problems I'd have to agree but mine is 95% OK so hence the reluctance to jump first.

Ultima357 Fair enough 🙂

Ultima357 I jumped, two days in, night and day difference, your issue sounds as mine; battery out and Solar In now correlate, as does over night charge - monitoring.....

Vez That's good to know. Let's hope that they have finally got it all sorted. Just need the dynamic discharge to work properly now so that you can limit the discharge when for instance, charging a car. In case you haven't tried this, it just gives up to 6kw regardless of the setting at the moment.

Ultima357 I run HA/GivTCP with MyEnergi Integration, and pause discharge if I need an emergency charge during the day (if the charger isnt in Eco+). Will be moving that over to the spare Shelly PM that I have now (as a result of moving to AIO) so I don't have to rely on the cloud for this.

I have just seen this Gateway Firmware. Anyone updated it or been offered the update yet?

https://kb.givenergy.cloud/article.php?id=131

Looks like one firm is getting ready to sell AIO 2-6 modules (No Gateway)?

https://www.segen.co.uk/product/givenergy-all-in-one-6-0kw-13-5-kwh-no-gateway/

https://batterygroup.co.uk/batteries-by-type/solar-storage/13930/givenergy-all-in-one-13.5kwh-battery-and-inverter

I can not update the Gateway Firmware as there in no button 🤷🏼‍♂️🙁

Q2, what is BTT
Q3, Is the SOC column measurement correct? At this point in time it says 49%
(Looks like this is the state of charge in real time)

Daveb01 Cant check here, no gateway in portal.

Daveb01 my GE shows the same FW.

Shortly after installation my gateway did update from 4 to 5 via the portal giving me an update option.

GivE confirmed that there is little for a use with the later updates so far. I assume they aren’t pushing them through the portal, as you say no update button.

Past couple of days solar has topped the battery up a few kwh in the afternoon, as of right now this is where I stand and im curious as to what the charge left will be when I hit 5%.

I am going to have a punt at guessing -1.6 but im not sure if it will deplete before charge period later.

EDIT
Looks like it didn't empty overnight before charge period started, as of right now after it topping back up to 100% from solar

I might force discharge if it doesn't deplete by midnight to see what charge left says when it hits 5%

EDIT 2
This is interesting

That's a first on BMS 8, went past 5% to 0%.

It took 1.7kwh to get back to 5%, seems a little much.

I keep an eye on my AIO using a couple of apps in tandem. In the morning I open the " Inverter " app from bbcbasic.co.uk which reads the inverter and tells me how much charge has gone into the battery overnight, I then open the Givenergy app and select the battery icon which then shows the graph of battery SOC and note the battery percentage at the start of charging. I then do a bit of mental arithmetic and only dig deeper if things seem way off. This morning I see that the battery had dropped to 68% at the start of the first overnight charging slot, and the Inverter app shows 5.5kWh of charge to bring it up to 100%.

I had the sudden drop in battery from 50% to 0%. This is getting ridiculous that it keeps doing this after ver 8.
What is going on guys??

Nik

Please see my post above on the 14th Feb. I am on fw v7 and noticed the difference once the sun started to re-charge the AIO durning the day back up to 90/100%.

I was hoping fw v8 would sort this issue out, so fingers crossed 🤞GE sort it.

Nik Would seem like it's a fault. Call it in to GE support. Check the battery voltage in the cloud portal too to see if that gives you any better idea of what happened. I don't get these issues on mine on V8.

Ultima357
Thanks will do

The changelogs for the next AIO releases (inverter and BMS) are publicly available on the knowledge base, if anyone is interested in seeing what's coming up

https://kb.givenergy.cloud/article.php?id=128
https://kb.givenergy.cloud/article.php?id=129

PianSom

Fix PV under 200W issue

nice!

PianSom

Here is the Gateway firmware releases (coming soon)

https://kb.givenergy.cloud/article.php?id=131

Daveb01
I queried this over on the beta forums a few weeks ago, and was told that there were no real user benefits in the changes between v5 and v8.

But there could be minor improvements in v9 when it comes.

Like everyone else, I'm still on v5.

Blasteh Also this will be welcomed by many:

5. Improve SOC calculation, to improve SOC tracking and capacity issues.

KevinB I'll believe it when I see it, they've been trying to fix the SoC in every version and it's clearly not fixed

I just want to be able to see 13.5kWh from the thing, still only averaging low 12 kWh

I'm wondering when the latest software will be released on the portal. My V8 is currently drifting away from full charge due to the solar charging. Now says full at just 316v, so probably only 8 to 9 kwhr stored.

I mentioned in another message that I monitor mine in an ad hoc manner using the GivEnergy app where I select the battery icon then scrub along the 3hr timeline to the start of the first charge window and take note of the reported battery percentage. I have the bbcbasic.co.uk " Inverter " app running as a permanent display on an old iPad which I check first thing in the morning before Solar charging kicks in. This shows how much the battery has charged from the grid. Last night for example the Givenergy App showed that the battery was at 74% at the start of the 1/2 hour charge window and it finished at 94% having taken 3kWh of charge. So if 20% is 3kWh that's about 15kWh reported total capacity. I realise it's not the most accurate method but it has been pretty consistent over the last few weeks at between 14-16 kWh extrapolated charge.

That %age is a little high lol

Should be around 40%, seems quite out of whack for less than 48hr period.

How are those on 9 finding it?

With 9 I’m finding that SOC does still wander a bit if the charge level doesn’t get near the top or bottom for a few days, but it does seem to sort itself out without the big SOC drops that we had with 7 and capacity when charged fully remains good.
I’ve stopped the in depth daily analysis as I’m happy that I understand the way the system is performing and it seems pretty stable now. Only look closely at the numbers if I see something in the graphs that grabs my attention.
Will track closely again for a while after the next firmware update.

Update now available, those that have done the beta is it a good time to upgrade or wait a bit longer?

I also guess this thread should be closed and a new BMS 9 started?

Daveb01
Been on the beta since early March. No reason not to upgrade from me.

I think I may have had one small SoC drop (memory fail) but otherwise all good.

No update for the Gateway though 🤷🏼‍♂️🙁

https://community.givenergy.cloud/d/4305-all-in-one-bms-release-9

Anyone brave enough to press the upgrade button yet?

Most will probably be waiting till its cheap rate....

So pressed the button, 1st time failed, but updated the Dongle from 204 to 205 and battery from v8 to v9.
Pressed again it took 12 mins.

Inverter Model
All-In-One

Software Version
D0.611-A0.611

Inverter Firmware Version
D0.611-A0.611

Battery Firmware Version
9

However was expecting a recalibration but it did not do one.

Daveb01 for me 9 has been far more reliable, much better usable capacity, no SOC drops and solar registers at lower than 200w.
All good as far as I am concerned

Mine failed mid way at battery upgrade so inverter is upgraded but battery still on 8 - tried about 20 times

Think its due to my setup phasing on and offline frequently that throws it

Oddly my stats look better before my charged figure for the day used to read as 3 kwh or there about showing as 10 which is about right (with the 30 mins before)

I will thinking about it move my charge window back to 12 o'clock as its easier to monitor

Updated dongle & inverter to 611, took about half an hour, all good.

PV issue below 200w doesn't look fixed for me in the latest update

deanjesper
The earlier betas were a bit more sensitive iirc, but my AIO currently recognises power levels just under 100W.

What is the minimum you are seeing?

PianSom for some reason it started showing at 100w and it's showing now 78w so I will report back this evening!

So it only reports when it hits 100w and then it'll report lower than 100w. The lowest I've seen is 59w. If it drops to 0w then it won't report again until it hits 100w, so it'll never go from say 0w to 20w, but it will go from 0w to 100w and then down to say 60w.

Very similar here:
In the mornings it appears to start around 100W

And in the evenings it drops off around 50W

Before the recent BMS upgrade I would typically see 200 W both morning & evening.

Mine was the same, plus I got about 30 mins extra load to house before the 200

I wanted to kick firmware update off at 11:45 then start charging around 12:15 but after nearly 40 minutes it failed, another 40m on hold to help desk to get it sorted then another 20-30m to fully discharge battery so had to up the charge rate to max and still didn't finish til 16:35 and I didn't get anything done today despite getting up at 10am lol

Been on battery for over half an hour now but still at 100%, total put in was 14.6.

EDIT
In "Meter Data" after the reset when doing firmware update it says 1.2kwh discharge from battery but in battery tab its still on 100% and the main page discharge is not going up and still says 100% as well.

Just looked at Gateway this morning and there is a SW update, who’s going to be first?

OK I will have a go 🤞🤞🤞

Done, took 4 mins, just a dongle update, no power cut to house.
So if you look on Inverter setting both dongles are on 205.

Daveb01
Good man 👍😁

Daveb01 Don't have gateway on my portal, maybe I should inquire about that.

On another note this is weird, got charging paused but when battery power is used for kettle the %age seems to go up

I just tried to apply the new firmware to my AIO; it failed twice. No indication of why it failed, or what I can do to resolve it. Ah well, I guess I'm staying on the previous version for a while.

KevinB

Everything is still the same v5 fw, only thing changed was the Dongle to 205, so don’t worry to much if it does not update.

KevinB Do you mean 8 to 9? If so I had this and just kept trying, it took eventually.

DreadorUK Yeah, 8 to 9. I was trying to time the calibration with today's cheap Agile period, but the ship has sailed on that. Hopefully next weekend will have some cheap daytime slots... 🤞

KevinB I had to call help desk to get battery sorted, was done nice and quick.

I can see this thread is sticky in GivEnergy Products section. Since the December 2023 when thread was opened, a few more threads has been opened with discussion about newer BMS releases, making this one obsolete. Notice no reply here since April 2024.
Can please admins @TheDragon (GivEnergy) unstick this thread.