Sudden discharge of batteries

Has anyone else's 8.2kw batteries suddenly discharged this morning? Should I turn them off? AC coupled system. Thanks

sheltont And is there a way to stop them fully recharging? I'm about to go onto expensive rate of electric 🙁

Wow - exactly the same has just happened to us at 7am, any thoughts would be gratefully received on the root cause and how we can delay the re-charge until we're generating electricity from the array?

Mine seems ok at the moment. Can you not alter the battery settings in the app/portal to force it to not charge?

No idea, it's stopped charging at 5%!!! Hopefully the solar will pick it up and charge then today.

A few things, there were improvements in the latest battery firmware to help with problems like sudden drops Look at the software tab of the inverter page on the portal to see if there is a firmware upgrade available.

If it drops to 0 it will automatically pull from the grid until it has reached the reserve level that defaults to 4%

To only charge from solar set the inverter to eco/dynamic model enabled (and all other modes disabled).

Oh, ok, that makes sense. Thanks. Can't find any software updates pending, so hopefully this will recharge and be ok. Must have charged to the reserve then, then stopped. makes sense.

sheltont I'd get onto GE support, sounds very odd.

How do I contact GE support please? They still aren't working...

sheltont It says There was an error changing your settings: Inverter Offline. Please try again

This is sorted now, apparently the charge and discharge rate was set to 0 for some reason!

At 7 am on the 4th July my 8.2 kw battery just went from 40% to zero in seconds it didnt slowly discharge it just vanished in seconds and recovered over the daytime to 60%. most annoying to say the least

JohnDawson YEs - that's what happened to me. My settings all vanished as well. I had to do a factory reset then check the battery settings as they had a both a minimum and maximum discharge of 0 rather than 3000. Then it all seemed to work again, but that was my work around. Hopefully it won't happen again, but yes, a full discharge over seconds at my end too 🙁

@sheltont . GivEnergy support can be contacted by referring to the following link. With various options depending on which country you are in.
https://www.givenergy.co.uk/contactus

The sudden drop in your reported battery state of charge % (SOC%) is not a real discharge of that amount of energy as the power rate would be higher than the inverter and batteries are actually capable of. Instead it is a mismatch between the reported SOC and the actual SOC in the battery.

There is a new battery firmware version 3007 that is designed to overcome this issue.
This may be presented to you through the web portal as @locked has stated above, or you may need to contact GE support in order for them to apply the update.

It's done it again this morning with me.... is this me? I've raised it as a GE ticket

sheltont yes this has happened to me?????

Has happened to me today. 58% to 0%.

At the time I was force discharging so this does sound like the second battery (2x8.2) was empty and showing a false SOC....however...

...Looking at my battery I'm already on 3007 but I can't tell when it updated. As mine also happened at around 7am this morning I'm wondering if this was a forced software update - I hope it is the case and it's not a sign of issues withy cells. This could be wishful thinking of course.

Does any one know how to check when the last update happened - mine is showing -- on last update?

Lmcde3 If you go to My Inverter > the either General or Software there is a last software update/last update time field, check there

Alas it's just blank on mine. I've restarted a couple of time and it seems fine now...however I feel like I'm just waiting for the inevitable 'next time'. Will keep an eye on this. Thank for your replies 👍

I got a response from giv energy, they think there was an issue with the initial charge and calibration, so have now discharged and charged again. Waiting to see if the same thing happens again...

My 8.2 battery that has been running 'flawlessly' for the last 3 weeks decided to discharge from 38% to 0% at 6.40am today :/

Battery firmware has be on 3007 since day zero.....hopefully this is a one off?

Should have added, they couldn't do it when it was going to be sunny all the next day could they 🙁

Ra That is concerning with it being the latest public release firmware which was supposed to address this issue.

chris_p Really? That's not good news at all 🙁

I'll have to work out how to submit a ticket to GE and see what they say.

mines done it this morning at 6.00 40% to 0
also on 3007
now reads 4% waiting for the sun
first time since instal 13/4/22 that we have not had battery power

Been monitoring my 100% SOC battery voltage since installing 3007. Looked OK at first but it is now gradually dropping day by day. Doesn't look like the BMS update has fixed the problem. 😞

Same again for me last night - 28% to 0% instantly and this despite having at the battery on AC charge from 00:30-04:30 which didn't work for some reason.

I've reported to GE via bug report. Will let you know what I hear.

I just happened to check the portal for something only to see the batter dumping 2.1kW to the grid. Which was quite alarming. There's no reason for it to have done that. Seems to have stopped now before it got fully exhausted but worrying nonetheless.

Woodstock , your post reads as if you regularly hit 100% charge in order to record the battery voltage value.
Do you regularly use the full battery capacity down to 4% charge, or more so say down to 50% then back up to 100% each day.
Just trying to understand what battery usage regime is most likely to knock the SOC calibration out over a period of time, and which regime is least likely to have that effect.

Simon_C if it helps, from various posts both here and on Facebook its never been reported that the older 80% DOD 8.2 battery (that has 160AH cells inside as opposed to 200AH) does it.
So that would point in the direction of it being something related to the either the BMS struggling to account for the buffers or the newer cells have different voltage profiles of some sort the BMS just can't get its head around.
I've one of the old ones that's had over 7MWh pumped through it to date and it's never done it.

Simon_C In the summer months my 100% DOD 8.2kWh battery is charged to a displayed 100% SOC by lunchtime. The battery then sits at 100% with house load being met by solar output until early evening. Overnight the battery will discharge to 75%-85% displayed SOC before the cycle repeats. Charging is carried out with battery input limited to around 500W to keep the inverter temperature low (maximum temperature reached on the very hot two days last week was 43 degrees but more typically it peaks at around 34 degrees). The inverter is cooled solely by original equipment finning and there are no aftermarket fans in use.

What I have noticed since installing 3007 is that the displayed SOC though the day is better following battery usage and reports across a wider range than was previously the case. What does not seem to have been fixed is the long term accuracy of the 100% useable SOC daily starting point. As was previously the case the voltage at which the battery stabilises (after having been charged to a displayed 100%) is declining day by day. This points to the the BMS trying to keep track of long term battery input/output rather than resetting the 100% full point as happens during calibration. Running a battery calibration charges to a pre-set maximum voltage and uses this checkpoint as the true 100% capacity starting point (well over 8.2kWh). It is as if the BMS is then trying to keep track of battery input/output for ever more rather than establishing a new 100% useable capacity (8.2kWh) checkpoint each time the battery reaches a pre-set voltage a bit lower lower than that used for the 100% true capacity.

I have no idea how the BMS is programmed to work but it appears to me as if following of battery usage has been improved but without introducing a measured (rather than calculated) 8.2kWh reset. The rate at which the reported SOC drifts from reality has been slowed and better represents what is happening throughout the day but the point at which the battery stops charging because it thinks it has reached 8.2kWh capacity is continuing to drift downwards. It may take longer to get there but I foresee a sudden drop to 0% at some point in the future while all the while suffering a slow reduction in useable battery capacity.

Woodstock , thanks for a very descriptive analysis.

If this is the experience of a number of other users then GE should be taking it on board when further improving the BMS firmware.

It is much like software I have had to write in the past for control of valve and damper actuators that are operated on timed drive open and drive closed signals. Any mid-position is only ever possible to be estimated, and every opportunity to reset the calibration at a known 0 or 100% (end of travel) is taken.

Simon_C Yes, I would have expected the battery calibration to be reset each time the usable 4% or 100% limits are hit but this would appear to only happen at the voltages deemed to be true 0% and 100%. Perhaps GE could advise if this is the case?

Woodstock 100% agree here. It's as if we just need the system to say 'i think SOC is 4%\100% - let me check by trying to discharge/add more power'

This happened again last night for me 40% to 0% instantly.

My worry is the winter. If we need to go to 0% to get the recalibration then the batteries won't be working well enough.

Has anyone heard back from GE - I've raised a number of tickets but still no a word?

Lmcde3 My worry is the winter too.

I had 12 years worth of solar PV production/usage stats and sized my battery capacity on having sufficient reserve to cover a number of consecutive days when electricity usage exceeded solar PV production. Of course this was based upon having a consistent battery capacity of 8.2kWh and not an ever reducing figure somewhere between 8.2kWh and 0kWh; depending on when the battery (erroneously) thinks that it is full and stops charging!

When solar PV output is at a premium in the winter you need to be able to store as much as possible on the days when the sun does shine.

Just slinging my name on this thread to say, "me too". This is what I've been seeing and reporting to my installer and GE since the system was installed in August. Exactly the same sorts of observations -- seems to need to hit the 4%/100% to reset calibration resulting in sudden "fully empty" or "fully discharged" jumps.
The bigger problem is that I too am seeing a reduction in capacity over time. My pair of 2.6kWh batteries are currently delivering around 4kWh of usable capacity between those limits, even having had one battery replaced. Looks like the inverter is mapping ₁₈% on the individual to 6% overall SoC, and ₉₅% on the batteries to 100% overall SoC. That's corresponding to roughly 3.33V high and 3.2V low on the cells, or 53V to 51V overall.
GE are sadly being as unresponsive here as ever.

Ive called GE and as expected they've run a recalibration. Not expecting any different result but am part of the process for now at least.

They did say they have a new firmware version they could roll out but it's not fully tested yet. Perhaps a fix will come soon.

Lmcde3 yes there is another thread discussing this

TheDragon (GivEnergy) https://community.givenergy.cloud/d/713-firmware-3007-has-not-fixed-sudden-soc-drop-to-0-problem/39 post 32. Hopefully this is the fix we need. Cheers

Lmcde3 hopefully.

A few mostly multi battery are waiting for this to be fixed once and for all

Hopefully.

TheDragon (GivEnergy) hopefully. Saw again early this morning - sudden drop from 23% to 4%, which is frustrating as it looked like we’d make it to solar charge fine - and then all of a sudden we’re back on grid power.

GrahamRead Interesting that yours went from 23% to 4%. Does this mean that it respected the battery reserve/cutoff % limit? When I experience sudden drops it goes all the way to 0% and starts a forced charge. I had always taken this to be that it thought the battery had a SOC more than 4% higher than the actual SOC and it only stopped discharging when it hit the low voltage limit.

GrahamRead
What was the battery voltage doing before and after the drop? The voltage is the true indicator or the battery. SOC is calculated and a different approach to this is in testing by @Adam-Reynolds

Dropped from 41% to 0% again this morning. Beginning to get a little bit annoying now 🙁

TheDragon (GivEnergy) Sorry, been away, hence no reply. It did it again this morning - here's the battery snapshot (these must be average values - as I have two batteries). First col is the voltage.

Also again today. 41% to 5%.

The problem clearly comes up on the first day of bad weather when the discharge is greater.

Any update GE team please?

Lmcde3 Firmware 3007 has not fixed sudden SOC drop to 0% problem.

Lmcde3 search the thread above. Latest update 5 days ago. Hopefully some progress soon.

This has happened to us three times now. The batteries were delivered in a state of deep discharge and I'm concerned that we're now seeing the damage this has caused. If we have paid for 10.4kWh and we aren't getting that, I will be rejecting the goods. The other consideration is whether the system is fit for purpose. The purpose was to allow us to charge overnight and move our electricity use out of peak times. This behaviour jeopardises that goal and again gives us grounds to reject the goods. I do not expect to be a beta tester when I've paid several grand. Also frustrating that I couldn't find this forum prior to purchase, which would have allowed me to see some of these issues.

PhilS Also frustrating that I couldn't find this forum prior to purchase, which would have allowed me to see some of these issues.

You have my sympathy. This is not a new problem and people are quick to post when they experience this. Elsewhere on the community, there's information that suggests that Givenergy has grown from about 1500 customer installations in 2021 to over 10,000 now. That is phenomenal growth for any company. That so many of their batteries are exhibiting very frustrating performance issues is a reflection on the company (QA, materials, factory-testing, who knows?). However, only Givenergy knows how many of their installations have problems. They have access to this data because all of the plants transmit data to their cloud. Any that aren't sending data must be faulty by definition. Only Givenergy knows what proportion of their installed batteries are dropping to 0% in an instant. On the assumption that many who experience this have found this forum and are making their displeasure known, I'm guessing that there's a silent majority who are not experiencing this (ATM, fingers crossed).

What I'm alluding to is that they need to find a solution for those who have sudden drops in SoC, but that not everybody is seeing this. Is that because of the way the batteries are being cycled? Ours reach 100% SoC most days before midday at this time of year. I do disable discharge on one battery at night sometimes to allow our other inverter-connected battery to discharge to a lower SoC, but not everyone can do this. Getting the BMS to accurately track SoC is very difficult and drift does occur over time, but if the answer is to plan a full discharge regularly, the it'd be good for them to be explicit.

Tim Indeed, if I just need to ensure I'm doing a full charge and discharge ever X days, I can do that. There'd be some efficiency loss but it would still be saving significantly. Just an explicit "this is what you must do to avoid this issue" would be fine. There's always a discrepancy between what the status lights on the front of the batteries say and the reported percentage.

I'm wondering if it's a power factor issue. Residential power consumption does not account for power factor and therefore we simply don't care about it. It costs the supplier more money but we don't see the impact. Industrial systems must have power factor correction because they pay for the extra. So, is it possible that the inverter is measuring the power going in and out but it's not accounting for power factor. When I run a pump in the house, it's drawing about 7% more than is being registered because it's an inductive load. This would surely come off the battery and, when an early AM correlation between the BMS and the software reported capacity is performed, it finds more has been used and corrects?

I would expect the battery BMS power metering would be at the 51 V dc stage, so no power factor involved.
The DC voltage and charge/discharge current can be measured accurately, but it is more difficult to relate this to the battery %SOC as the voltage will be influenced by the value of the current, temperature and %SOC to varying degrees.

PhilS So, is it possible that the inverter is measuring the power going in and out but it's not accounting for power factor. When I run a pump in the house, it's drawing about 7% more than is being registered because it's an inductive load.

Sorry, but that isn't correct. The device/pump/whatever is not drawing any more load than is being registered. It's drawing more current than a perfectly resistive load (kettle, iron etc) would take for the same power, but no more load. Effectively, an inductive or capacitive load consumes a proportaion power and exports that amount twice per 50Hz cycle over and above the actual power consumption.

The power factor is the ratio of the power (Watts) over the VoltAmperes (Volts times Amps).

The battery power, on the other hand' is purely DC (albeit with varying voltage and current) and power factor doesn't come into it.

Cheers
Mark

Anyone know why the voltage range is so restricted?
Might need to get a battery shunt to get a far better idea about the power remaining

ispookie666 I don't think the voltage range is restricted, it's just the characteristic of the technology that between about say 20% and 90% charge, the voltage is constant. I'm assuming that the Givenergy BMS uses a shunt already to clock the current in and out to calculate SoC.

Resting voltage at 100% SOC each cell comes to 3.34 V which is definitely not the fully charged voltage. Wonder if givenergy are intentionally restricting the maximum SOC (something like derating) to provide better life.
I suspect this could be adding to the problem or maybe i'm talking porkies

ispookie666 intentionally restricting the maximum SOC

Check out their KB. Can't be sure whether this is the same policy for all sizes and types of battery as it hasn't been updated in a while.

What the portal says is 100% really isn’t true 100% if you go from cell voltages but the value required to get you your advertised capacity in kWh.
Anything with 100% DOD actually isn’t what it says on the tin. They are oversized cells so your not really charging to 100% at all.
Others just scale accordingly to get you advertised capacity.

There’s no way they’d offer the warranty they do if you were running these at full bore cycling.

I’m in the middle of building out a wiki with this kind of thing here

Mine all of the sudden 'loses charge' at least once a week. Not acceptable at all and if anyone from GivEnergy lurks in these forums I would be prepared to have a lot of people asking for their money back soon.

The batteries (or at least my battery) is not fit for purpose and as such I will be rejecting it soon and they can have it back.

Ra Once a week is a tough gig! Have you raised a fault report with Givenergy? Have you asked them to carry out a calibration to see if that helps? As it's the indication that's dropping, not the actually battery charge, a calibration to fully discharge and fully charge (from actual 0% to actual 100%) would seem to be a solution. The issue then becomes how to consume the energy discharged and try to charge from PV if possible.

Tim Hi Tim, the last time it did it was the morning of the 26th August and that day it was charged from 0% to 100%.

It has dropped again to 0% this morning, so thats 4 days it has managed to not to it the last time.

I have just sent GivEnergy an email (so I have proof of the date I have kicked off), will see what they say.

Technically my battery has been calibrated about 7/8 times since I had it and it was only comissioned at the very end of June!!

Tim The sudden drop of displayed SOC to 0% occurs when the battery low voltage alarm is triggered. At this point the battery IS fully discharged. As you say, there is no sudden discharge of the battery it is just that the over reading of the SOC % prevents the 4% battery cut off limit from operating when it should. When the battery hits the low voltage alarm (true 0% SOC) the system automatically initiates a recalibration and force charges to 4%. It then switches to the user charge settings which operate until battery capacity (actual not quoted) is reached.

Once the displayed SOC % has drifted upwards by more than 4% from reality then a sudden drop to 0% becomes possible. Over time (as the displayed SOC % drifts upwards still further) the amount of battery discharge required to initiate the sudden drop reduces.

I would be very interested to know what SOC % is displaying when Simon is suffering his sudden drops. Given their frequency I would suspect that his battery is deeply discharged most days and the drops are occurring from a relatively low displayed SOC %.

My displayed SOC % drifts upwards by around 7% a week but as my displayed SOC % rarely drops below 75% in the summer I go many weeks between drops. When they do occur, however, it is from a high displayed SOC % with an associated significant reduction in effective battery capacity!

There is a school of thought that those who empty their batteries every day simply do not notice the sudden drops from a displayed 4-5% SOC to 0% (and subsequent forced recharge to 4% / recalibration) which are likely happening 1-2 times a week. Their batteries never suffer any apparent loss of charging capacity due to being regularly recalibrated. If this is the case then I would question the effect on the battery of being frequently discharged down to a true 0% SOC, given that it was thought prudent to have an automatic cut off at 4%.

Woodstock More than happy to provide you the information if you tell me how to get to it? I assume you want more than just the displayed percentage before it drops?

Woodstock I found this page which looks like they have a good understanding of the battery technology. Looking at the "Discharging an LFP Battery" section towards the bottom, the chart suggests that 95% SoC (the value Givenergy sets the customer 100% at according to the KB) by interpolation, would be about 3.4V per cell and that the 15% SoC (the value Givenergy sets the customer 4% at) would be about 3.15V per cell. However, just below the chart is a table which shows how the voltage varies between zero current and 0.25C. This suggests that the voltage drops quite dramatically when the SoC approaches single digit values.

If the data on the web page is broadly correct, cell voltage is a good indicator of SoC above 90% (about 3.35V/cell) and below about 20% (3.225V/cell). Between these values, it's difficult to be accurate because the battery is made up of a number of cells and I presume is affected by battery temperature. While I accept that you can see a 7% per week drift upwards in the displayed SoC, I'm struggling to get my head around how a voltage which can be measured accurately, results in an indicated SoC drifting (upwards or down).

Keeping track of cell voltages is something I have on my to do list. I can see them dynamically, I just haven't got round to capturing them and the associated SoC in Influx yet.

Ra No, nothing complicated, just the last displayed SOC % value displayed on the graph line before it plummets to 0%.

Given the ongoing issues I initiated a battery discharge overnight last night as looking ahead we have a few days of bad weather when I'll need my full 16.4 capacity.

30%-0% SoC instand drop - although I was expecting this I am still disappointed.

Graham if you see this, I'd be happy to take the latest firmware even if it doesn't go below 7% SoC - it would at least be consistent. Any news?

Woodstock I've had a quick look back at my values.

34%
21%
41%
38%
7% after only charging to 59% the previous day from a drop
41%
27%

Also noticed this strange 'blip' a month ago where it was at 100% in the afternoon but dropped to 50% for about 10 minutes before jumping back to 100% again.

My main gripe is that I'm exporting energy at some points of the day because my battery is 'fully charged' when the reality is that its only 60% full 😤 This is fine when we have days upon days of sunshine but this is only going to annoy me more as we approach winter

Ra That's interesting; the percentage drops are higher than I would have expected for something that is happening weekly. I was anticipating values < 10%. Your SOC % drift must be occurring at a rate much greater than I have experienced thus far. Agree that loss of useable battery capacity is a major concern as we go into autumn/winter and hours of sunlight reduce.

Lmcde3 Great minds think alike! I initiated a deliberate battery discharge two days ago when I could see my 100% SOC battery voltages had declined significantly. I had a 31%-0% instant drop and like you, although I was expecting it, I was disappointed. Won't be long before not having a full 8.2kWh of useable battery capacity will start to be an issue.

It would be some relief to see this pop up in the 'known issues' section while we're waiting on the release too...at the moment it feels like we're being held in the dark.

Lmcde3 and we'll literally be in the dark soon 🙁