FoxESS Community - Owners & Installers Forum

I have a 12 x 450W solar array coupled to a Foxess H1-5.0-E-G2-WL inverter with and attached EP12 battery - the battery version_BMS is 1.011, the inverter version_Master is 1.47, version_Slave is 1.04 and version_ARM is 1.35. The data logger is WL-H1-WL with software version 2.06.

The system was installed on 19th November 2025 and seems to work fine with, what seem to be to me a notable anomaly - the system seems unable to regard Reserve State of Charge as an immovable barrier in that the SoC sometimes decreases below it with no attempt by the system to top it up from the grid. The other thing is the there SoC seems to have a mind of its own - increasing or decreasing by 1% with not evident reason that I can see, i.e. not taking power from the grid when it increases (with no input from PV being evident - e.g. at night) and not generating power when it decreases.

What is going on? I have pictures of this if it helps.

Probably best to start by saying a battery is an electro chemical device that is affected by temperature changes which probably explains small % changes, better to think of it as a fuel gauge rather than a totally accurate %, if it’s jumping by 20% then it’s time to worry.

There are plenty of posts on here about LFP batteries, but to keep it short they have a very flat voltage from nearly full to nearly empty, and an EP12 battery has a considerable number of cells within it that have to establish balance.

Balancing occurs when the battery approaches 100% and 10% where the cell voltages rise and drop rapidly and so the battery has to be fully charged and discharged several times before all the cells have had the chance to balance and the battery manager establishes what constitutes the limits. The battery manager constructs the SoC or battery reserve it presents to you from all of the cell voltages within it, and so even being sat on a cold concrete floor or against a wall can affect the individual cells slightly.

The inverter always takes a small amount of power from the batteries (or PV when present) and whilst only a small drain, over time it will discharge the battery, and your batteries are supporting the house load when there is no PV present.

It’s good practice to fully charge cycle the battery at least once a month to maintain that balance, have you done any full charges on the system from grid, or you relying on what little solar is available at this time of year ?

It arrived on 19/11 with 50% charge in it which the BMS then gradually discharged down to the 10% indicated bottom limit.

The 20th was a so so solar day, but the PV charged the battery up to 29% indicated.

On the 21st I tried out the function which allowed the battery to charge from the grid overnight, that took it to 94% indicated - I didn't stop it at that value, I was fast asleep.

By about 10am on the 22nd it had decreased to 78% indicated due to house load and then, because it was a reasonable solar day charged up to 100% 13:23. At this point I set the reserve SoC to 50% having in mind that having a good state of charge on a new battery might be a good idea, however, on the 23rd I set the reserve SoC to 20% thinking it was not a good idea to let it drop all the way to 10%. So far so good, I hadn't noticed any of the 1% changes up to then. But on the 24th the 1% drops appeared - two of them spaced hours apart took the SoC down 18% (the battery warming function was still disabled then), the battery temperature was around 9C and it was not a good solar day.

The 25th was a nice day, the SoC got to 75% indicated.

The 26th was an even better day, the SoC got to 96% indicated.

The 27th finished at 27% indicated.

The 28th got to 62% indicated.

I've also tried lowering the reserve to 15% and now returned it to 20%, but currently, due to the 1% changes it's down to 16%.

I have seen it decide to draw from the grid to top up the battery but the SoC value seems not to be factor which causes its internal algorithms to make that decision.

I have a technological background in computers and comms, but nil in solar, inverters and LP batteries, so maybe I'm just worried well and very curious about wtf is going on.

Any help is greatly appreciated.

Best regards and thanks so far

Derrick.

Hello Derrick,

That's a fair old post there, but the answers I have echo what Dave has already mentioned.

The BMS presents to us a human readable % that is taken from a calculation of the many cells combined, the Voltage difference between empty and full is quite small, think 3.100V to 3.450V (as my close morning guess)

The battery chemistry is affected by temperature, and the "gained" SoC % can be 1-4% without any charge/discharge happening. As your battery is new, and maybe is still doing a calibration/balancing of the cells, the swing can be a few % as above.

If the battery is pulling from the grid, without a schedule, then the battery was cold, this is the BMS taking over control trying to keep the battery working (it goes offline if it drops too low)

As for your reserve value, I have mine set at 10% so that I can access the most kWh out of the battery, after 11 months of being installed, I have a 100% SoH (State of Health) battery. The only reason for a higher reserve, is if you are going to experience a power cut and want to load up to 100% SoC, and this only works if you have a working EPS circuit/gateway device that can switch you from on-grid to off-grid during power loss.

OK, thanks again. But, while I now understand a little more about the workings of the battery and the algorithms governing its operation I still don't have answer to the why of the Reserve SoC being broken with no action by the system to maintain it. It raises the question as to how inviolable is the 10% limit.

It's a journey - as an end-user it seems very opaque.

Not that I'm unhappy - I've been truly amazed at even seeing the system, on several instances over the time since installation, charge the battery to 100% while at the same time feeding our house .

Best regards,

Derrick.

It's nothing to worry about, the SoC (on any battery) is a virtual construct, the battery manager itself is working off internal cell voltages - if you install a home assistant you will be able to see these.

You could never ask the battery to go outside of it's internal protection limits, it can sometimes drop below minSoC which is essentially the battery manager getting it's SoC estimate wrong. The battery management system will cut power and disconnect the pack as soon as an individual cell voltage drops below it's cut limit, it doesn't use SoC for this - it provides that as an estimate for the user.

At that point the battery management will monitor the cells and allow what's called bottom balancing to take place, if any charge is needed it will do small low power maintenance charges, and it will reset it's internal counters so that it gets it more accurate next time.

A new battery needs the chance to understand the limits, to do that it must reach 100% and hold there, and be discharged to 10% and hold there - ideally a few times, thereafter the advice is to do that at least monthly or the SoC accuracy will fall.

I just see Dave has replied as I was typing, but I will post my own 2 pence worth.


My 10% minimum setting has been breached, usually when doing a discharge session and it can overshoot to say 8%.

It really comes down to battery cells Voltage can drop a little under, often if you sit back and do nothing the cell Voltage will recover a little and the 8% might increase to 9% as the cells bounce back.

I have home assistant running now on my system, you can see the actual cell mV values, not just the shown SoC % reported via system. During discharge you can see the cells will drop low, say for example 3.115mV, then if you stop the discharge, the cells recover back up to say 3.225mV.

So in short, the limits being breached is not an issue, if you are shooting past the limit by 4% it suggests your BMS has not learned the cells low and high values still.


That is how I interpret this behaviour, I would suggest at least once a month drop cells to indicated 10% and then back to 100% in one. Maybe someone has better suggestion on how to maintain battery accuracy in the SoC and keeping the SoH as good as can be.

For my own scheduler, I have 11% set as the target SoC% for discharging. It'll sometimes overshoot into 10%, but not beyond that.

When I first set up predbat it did manage to push the SoC% down to about 8% somehow if I remember correctly, I overrode it at that point. After changing the config of predbat's min_soc_keep to be the kWh equivalent to 15% and min_soc_reserve to the equivalent of 10%, and configuring min_soc_on_grid to be the predbat reserve, it doesn't do it any more.

Reply to all and thanks for the detailed explanations - I've not tried to sort out home assistant yet, but it sounds like something I will delve into being a bit of a detail merchant.

I'll follow the advice and do the 100% and 10% limit charge/discharge cycles at least once a month as suggested.

Best regards to all, lots of reading to do.

Derrick.

Hi Derrick, me again

I have made this little graph up from a discharge I did last night, to bottom out the Cells and help the BMS keep the SoC % on track.

It starts at 29% SoC, and I stop the Force Discharge at 12%, it then drops down gently to 10% from a 300W load (TV and base load) until it stopped supplying energy. Soon after it reaches 10%, it then started a charge from my off peak tariff.

What you might see as interesting, is the Cell Voltage droops down low during the discharge, and actually recovers as the SoC changes down from 12% to 10% as I was more gentle with the load.
Hopefully you find this input fascinating, and helpful.

Hi, what I find interesting is the exponential cell potential droop at max SoC - what causes that?

I think, (I may be wrong in this) that from my own many moons past experience, that the droop in the read cell potential during discharge is due to the battery's internal impedance and the drop across it. The exponential droop is seems more interesting to me.

Best regards and thanks for this look at the internals,

Derrick.

Hi again Derrick, hopefully a final graph for you to peruse.

I have a further image to show a little further on past the other graph, and this time I added Cell Low and High, so that you can see the Cells look nicely calibrated (very small delta between high and low)

You can also see the curve does flatten off nicely towards 05:30, and I then allow the battery to discharge. The first small spikes are caused by kettle going on, saucepan on at dinner time, and the end of the graph ripples are from the oven going on/off as needed.

Thanks again, that's very illustrative of the battery's internal working but the thing I find to be fascinating and a bit enigmatic is that exponential curve on the left hand side - memories from dim and distant courses on electronics remind me of capacitor discharge curves.

Best regards,

Derrick.