Hi everyone

I have the following system:

• Inverter: Kodak OG-PLUS (Axpert King) 5.48RM - 5kW 48V
• Batteries: 2x parallel Fox ESS LV2600 2.56kWh 90%DoD rated
• Inverter & BMS comms over RS485, running Pylontech protocol - BMS has control over charging
• RPi running Solar Assistant for monitoring and additional control
• 6x JA Solar 385W mono panels (3S2P)

I am very perplexed by the sudden SOC changes that my batteries exhibit during discharge to less than 50% SOC and the subsequent charge back to 100%.  The issue sounds similar to that described in this thread and is illustrated in the graphs that follow.

Referring to the Solar Assistant plot of the latest occurrence below, point A is the point at which load-shedding started and the batteries began to discharge.  The batteries discharged as expected from A to B.  However, between B and C, the SOC dropped from 46% to 11% within 3 minutes under a constant load of only 420W.  (For some reason the 11% isn't shown on the scale of this graph from Solar Assistant, but when I zoom in on the data I can see it.  Higher resolution plots can be found in the other attached files.)

Solar Assistant is set to shutdown the inverter at SOC less than 20%, which it did, and the values shown between C and D are therefore irrelevant - they are just the last the last recorded values.  The inverter remained off until grid power returned at D, at which point the batteries reported a 0% SOC.

From D to E, normal charging occurred until the SOC suddenly shot up from 64% to 100% between E and F, again within the time-span of around 3 minutes!

So I lose 46% capacity within 3 minutes (assuming SOC went all the way to 0 while Solar Assistant wasn't recording) and re-gain that same capacity in the same amount of time.  Something is not right!

This is not the first occurrence of this behaviour either.  I do not generally discharge the batteries to less than 50% SOC, but the one time I did previously, a similar thing happened:

The rest of the time, with less intense discharging, the batteries behave as expected. 

One thing I have noticed, however, is a fairly high (to me, at least) self-discharge rate of around 5% SOC over 9 hours even with no load on the batteries:

Does this seem normal?  Can anyone shed light on these issues?  Has anyone had similar experiences?

The system is less than 3 months old and has never seen more than 2.5kW of load.  The batteries have not been in any way abused, yet almost 50% of my battery capacity is unusable.  Out of the 5kWh of installed capacity, I've never been able to get out more than 3kWh - extremely frustrating and misleading!

Initially I thought that the BMS's might not be properly calibrated and that the issue would fix itself over a few cycles, but that has not been the case.  I've also charged both batteries individually to 100% before hooking them up in parallel but also to no avail.  To me it seems like one of the cells might be defective or imbalanced and/or that there is a BMS fault.  I'm keen to hear the opinions and suggestions of others.

For more information, attached you will find images of all the Solar Assistant data, from 16 days down to 8 hours surrounding the most recent incident.

Thanks in advance!

Thoughts:
In which documentation do they specify that the Pylontech Protocol is 100% compatible with this battery, or is this a rebranded Pylontech battery. Could be a mismatch with protocol because of a different BMS. Also I don't see the bulk and float voltage values so one can compare that to what Pylontech specifies.
Perhaps an intermittent connection problem in the BMS cable?
Do these sudden spurious SOC changes occur when a certain condition takes place, i.e. right after loadshedding starts or ends or only during hours without solar energy? i.e. is there a pattern or is just random?  
Have you run a test with voltage control instead of BMS Protocol to see if the same spurious anomaly occurs?
 

Edited November 17, 20223 yr by zsde

Yeah, I have become an advocate for using "User-Defined" battery parameters just because coms can be a little hit or miss... I prefer setting the parameters myself so I know exactly what the battery is doing. To be fair, someone else may be better suited to help with some of the more programming/coms side stuff... 

Otherwise, is it possible the sudden SOC drop has something to do with the PV dropping out? Is it possible that the PV was propping up/charging the batteries, causing them to read with a higher voltage and as soon as it dropped so did the SOC/voltage?

1 hour ago, zsde said:

In which documentation do they specify that the Pylontech Protocol is 100% compatible with this battery, or is this a rebranded Pylontech battery. Could be a mismatch with protocol because of a different BMS.

The battery documentation actually makes no mention of any protocol whatsoever.  However, the RS-485 wiring is the same as Pylontech and the supplier (Segen Solar) confirmed that the PYL protocol should be used in the inverter.  I suppose I could try other protocols and see what happens.

1 hour ago, zsde said:

Also I don't see the bulk and float voltage values so one can compare that to what Pylontech specifies.

The BMS sets both the Bulk and Float Charge parameters in the inverter to 56.0V.  This is the extent of the information offered by the manual on the topic:

1 hour ago, zsde said:

Perhaps an intermittent connection problem in the BMS cable?

I don't think so, it complains loudly and quickly if any comms die.

1 hour ago, zsde said:

Do these sudden spurious SOC changes occur when a certain condition takes place, i.e. right after loadshedding starts or ends or only during hours without solar energy? i.e. is there a pattern or is just random?  

There is no pattern, apart from the fact that it only happens when there has been an extended discharge.  SOC can cycle happily above 40-50% with no such occurrences, but every time it continues to discharge below 40%, the apparent remaining capacity just vanishes, though not always at the same exact SOC.

1 hour ago, zsde said:

Have you run a test with voltage control instead of BMS Protocol to see if the same spurious anomaly occurs?

I haven't, mainly because of the lack of info in the manual and hoped the batteries would know best.  I can give it a try, what values would you recommend?

Edited November 17, 20223 yr by electr0n

17 minutes ago, Psy said:

"User-Defined" battery parameters

Do you have any recommended values?  Are you not concerned about balancing problems?

18 minutes ago, Psy said:

Otherwise, is it possible the sudden SOC drop has something to do with the PV dropping out? Is it possible that the PV was propping up/charging the batteries, causing them to read with a higher voltage and as soon as it dropped so did the SOC/voltage?

I don't *think* so, based on the fact that the first time it happened, my PV wasn't even installed yet, and the next time, the drop occurred after sunset when PV generation had already ceased:

Sudden SoC jumps like that means that you are reaching an SoC limit event - full voltage reached with current below threshold, or empty voltage reached with low current. Or a cell reached a similar condition. It boils down to the BMS configured capacity and SoH being far off from the actual. This is likely caused by a dead(severely underperforming) cell or cells. Sometimes it can be caused by unbalanced cells, but the fact that you are seeing it both while charging and discharging indicates a bigger issue. Either way, you should contact your supplier for an RMA.

Edited November 17, 20223 yr by P1000

Thanks, that was my feeling too.  Not looking forward to the process though!

15 hours ago, electr0n said:

Do you have any recommended values?  Are you not concerned about balancing problems?

So the BMS will still perform a couple functions, even without coms. So the over voltage, over current, short circuit and balancing should be handled anyway. It does depend on the battery and how they setup the BMS. I am not familiar with the  ESS LV2600 at all, so I cannot really confirm that but I do know thats how a couple other brands do things. Oh and just use the specs offered by the battery manual/spec-sheet. The manufacturer should give you the Bulk and float voltages along with the recommended charging currents. (As long as your cables etc support it)

15 hours ago, electr0n said:

I don't *think* so, based on the fact that the first time it happened, my PV wasn't even installed yet, and the next time, the drop occurred after sunset when PV generation had already ceased:

Well that's annoying... The only thing that I can think of is what @P1000mentioned.  If the battery is old, it could be a sign of deterioration... When batteries get older, they can sometimes see exponential SOC loss once below a certain level. It is very possible that there is an issue with the battery, but the only way to test it is with a bank load tester that will draw like 20A over a couple hours and then track the voltage for big jumps like this. Does it happen at the same/similar SOC every time?

 

Hi. I have read this with interest.
I'm getting very similar with the Growatt GBLI6532 battery. Tonight, no PV input, the load was largely consistent; SoC was at 20%, turned on the toaster which immediately added a large load and immediately SoC dropped to 9%. Not the first time I have seen this. I don't think there is a correlation with a large load being added, but the last two times it has happened a load was added. I'm going to investigate with my installer but would be really interested with any updates you might have. Thanks.