Gen 1 8.2kwh Battery Capacity Issue at 4 years, down to 79% of design capacity

System is 4 years old in September and I've had an issue with battery capacity on and off for about 12 months now.

Normal usage cycle is to fully charge the battery daily overnight all year round, get the odd battery crash when the SoC has drifted but I'm ok with that as infrequent.

Ran a couple of full export cycles last week whilst we were away as no other loads on the system bar background, discharge and charge for both the cycles was a somewhat disappointing 6.5kWh or so or about 79% of design capacity.

This matches with data available via HA, Battery design capacity is reported as 158.72 and battery capacity is down at 120.07. Quick bit of maths chucks out a 6.48kWh value at 54v for the current amp hour value. A bit of digging in older documents shows the battery should have a full capacity of 182.65Ah!

Time taken to achieve charge and discharge matches the 6.5kwh or so of actual battery capacity at 2.6kw, cell voltage drift is around 0.2v maximum across the 16 cells.

1140 cycles and 15,000 kWh throughput over 4 years isn't crazy numbers, on firmware v3015 as was rolled back from 3020 due to battery capacity issues.

Going to run a calibration tonight and see what it kicks out but

Rolled back from 3020 to 3015 by whom?
I'd suggest that 15MWh over 4 years is pushing it.
Design usable capacity 4%-100% for the 8.2kWh is 160Ah at 51.2V
I'm at 162Ah at a cycle count of 1303 after 3.6 years and firmware 3020

You sure it was a 100% dod battery to start with?
Those values are suspiciously similar to my 80% DOD 8.2 battery.
It'll say on the sticker on the side if it's 80 or 100%

mrand31 GE rolled it back.

Averages out to 10.26kWh of throughput per day, obviously that isn't flat over a 12 month period but I've never used it to grid trade etc, just normal domestic stuff.

GE's 2021 warranty allows for 10MWh per 1kWh of capacity, around 22.5kWh per day over a 10 year period albeit this is not applicable to the 8.2kwh unit.
4 years is 2000 cycles under fair use section.
I'm under half their throughput limit per day and over a year and half under their cycle count too - not exactly 'pushing it' but only 9% off a warranty claim point with 6 years left to go.

See what it looks like post calibration.

hoggy 100% DoD on the sticker.

Did have an odd issue post install where the battery identified as a 5.2kWh variant but this was resolved post a GE tech visit.

Interesting that your 80% one is kicking out similar values, got the full capacity value from the old portal interface when that sort of info was displayed.

yandards fair enough if that's what it says.
What inverter firmware are you running - the fast firmware (100 series not the 400 series) took 20% capacity from mine no matter what the battery firmware was.

[unknown]
Inverter firmware is D0.450-A0.451, assuming that's 400 series.

Got the capacity back up to 159.3Ah after the calibration last night, about 9.1kWh of power in.

Cell 10 stands out as a bit of an outlier during the end of the calibration process, voltage down to under 2.6V (min) and up to 2.8V (max) values tracked in HA - the other 15 cells don't see that much of a split and the next nearest low is 2.9V a 0.3V split. 6 cells dip below 3V (min) out of the pack, under half of the pack.

Going to run an export to empty cycle tonight and see what it does as a validation step.

hoggy
Inverter firmware is D0.450-A0.451, assuming that's 400 series.

Got the capacity back up to 159.3Ah after the calibration last night, about 9.1kWh of power in.

Cell 10 stands out as a bit of an outlier during the end of the calibration process, voltage down to under 2.6V (min) and up to 2.8V (max) values tracked in HA - the other 15 cells don't see that much of a split and the next nearest low is 2.9V a 0.3V split. 6 cells dip below 3V (min) out of the pack, under half of the pack.

Going to run an export to empty cycle tonight and see what it does as a validation step.

yandards are you saying cell 10 doesn't move past 2.8V even when fully charged? Or just when it's at the very bottom of the calibration before recharging (there's always one runner cell at both ends so one of them running down to the Low Cell trigger isn't particularly worrying)
The balancer is aiming for 30mV between cells & runs when above 3.4V.

hoggy correct, its only at the back end of the discharge process it does this and its back up and aligned with the other cells now.

The other cells in the battery have a definite above 3.4V trend towards the end of the calibration process but cell 10 is just about scraping the 3.4V line, HA shows it doesn't quite there on this data trace:

The other cells also don't show the large min/max split and I've added cell 11 as this is one as a reference example of the other. Appreciate the cell 10 is under more stress so it likely to manifest the largest variations.

Downloaded Rubiks app so going to use that going forward alongside HA as it's easier to see what is going on.

Export down to 4% overnight recorded as 7.8kWh discharge total, charge 8.1kWh so about 95% of build capacity.

Not bad for a 4 year old battery although I would have hoped it stayed closer to design capacity given the overhead the 100% DoD batteries have.

Different cell at minimum, now onto monitoring via Rubiks app to see when it next needs a calibration to drag it back to near capacity.

yandards is this the first discharge down from after calibration? If so it's a false reading as the calibration over charges the pack to find its top end. If its already been discharged down before this run then ignore me.

hoggy yes it was, like most at this time of year I don't get anywhere near the bottom 50% of the battery pack.

I'll run another export to empty tonight and see what it does, thanks for the information.

Ran an export on Monday night, 7.54 kWh discharge and 8.1 kWh charge, so far better than the 6.5 kWh I was getting!

Back to monitoring it and seeing what it does.