What charged the battery?

I can see the battery % went up from 3% to 6% between 10:30 and 11:30.

However I can see it wasn't solar charging it:

Nor was it grid (which only charged as set between 12:30 - 4:30)

And in fact nothing seemed to be charging it:

So what explains the increase in % 10:30 - 11:30?

Thanks.

I've seen this behaviour too. I think it's the battery recalibrating itself, which it is able to do while at a low charge level and with a low load.

Yes if I let my battery discharge it will recover a few % two or three times and resume powering the house for more than an hour.

Seen the same too, going from 4% to 7% and then suddenly discharging again.

It's this system working on the battery that is adjusting your SOC.

hoggy "A thrilling update — we can now deliver you a solution to the SOC (State of Charge) bug"

That didn't age well did it? Reads like a complete load of bollox given the farce that was 3013, which in turn was released because 3012 didn't work either.

Edit - in fact its comedy gold.... "Though there can still be small jumps, these are both very rare and very small (for example, 95 to 100%."

Vestas yeah it was perhaps a celebration too soon in that respect. I'm unsure if it'll ever truly be nailed for all without some significant hardware change to the BMS itself if I'm honest. Just too many variables for a fix all (cell numbers, manufacturers, age/cycles)
3015 for example takes my old 80% DOD down to 5% (actual) for some reason after it's done some OCV stuff - ignoring what used to be 15%-95% buffers, so now have a 90%+ DOD weirdly.
It's its 3rd birthday tomorrow so it'll be getting its yearly discharge run / capacity check to see how it's doing.

SimonP I've seen this behaviour too. I think it's the battery recalibrating itself,
In that case I would expect a notification

hoggy It's this system working on the battery that is adjusting your SOC.

Again, surely there should be a notification ??

Pete Thomas It would be nice, but no. The idea is really that it all works perfectly in the background with no user intervention ever required... Hmmm.

SimonP The idea is really that it all works perfectly in the background with no user intervention ever required... Hmmm.

Nice idea but if they give you all these the stats (which is nice), the last thing you want is to be scratching your head over why the battery appeared to charged without any apparent input.

Did your installer put in a secret cable to the closest lamppost, check to see if it flickers every now and then 😀👍

Pete Thomas

There is quite a lot of discussion about this on various forums but from what I’ve learned and “simply put”…

LFP batteries have very different characteristics than other types of battery chemistry.
They have a non linear voltage throughout their charge/ discharge range. It’s a pretty flat line for most of the capacity range, unlike other chemistry batteries where you may see an almost straight line across most of the SOC range, or some type of predictable parabola.

It is notoriously difficult to assess the actual capacity remaining. I think it’s something like 70-80% of the voltage range is within a couple of 0.2 volts. ie if you just walked up to an unknown LFP battery and measured the battery voltage and it said 52 volts then the capacity of it could be 70% or it could be 30% ! The voltage also varies with temperature!

Because of this, displaying the “actual” percentage of an LFP battery is partly a guestimate or derived from input/output, look-up tables and some sort of accounting method combined with measurement.

When it’s charging and discharging, it indicates a slightly different voltage to when it’s at rest. And in addition when it’s discharging at different rates it’ll display a slightly different voltage too.

It was described to me once as kind of like measuring a fast flowing choppy river level from a boat: The measurement you get depends when and how choppy it is and how fast it’s flowing.
However if you measure it when the river is calm and flat it’s a lot more accurate. Same for LFP batteries, if you measure them when they are at rest, you get a far more accurate indication.

Solar/ house batteries are hardly ever at rest, so it makes it all the more difficult. When these batteries are “at rest” (‘fully’ discharged or charged - or at some manufacturer set ‘at rest’ parameter) this is when the system can attempt to get a reasonable guess at their true state of charge.

With your system what you are seeing is the battery re-calibrating itself to adjust the displayed percentage SOC (the ‘guesstimate’) closer to the actual SOC measured in the battery to correct your usable range of power and display it to you more accurately.

This is called the OCV calibration and it runs at certain times when the battery is at ‘idle’. Of course people don’t know this because it isn’t mentioned anywhere that the average customer would look.

We are all seeing this behaviour with the batteries and it’s “normal” for this type of GE battery (even though it’s a little bit frustrating and non-intuitive). We’d all just like to see it go from 0-100% and back again! It’d make planning much easier.

Previous firmware versions used different versions of this re-calibration method that produced different results (some a lot worse!). But I think that with the latest S/W versions GE are starting to Nail it down a lot better. (AIO excluded from that comment.)

You’ll see the SOC vary slightly at the top and bottom end / ‘at rest’.

This is my current understanding and I hope that explains what you’re seeing.

I run an automation to prevent the battery cells dropping below 3.1V because after v3013 I don't trust GE's firmware to manage the battery properly, Discharging cells to 2.8V (44.8V total) WILL damage them, no ifs or buts about that & that is what v3013 does to Gen2 batteries.

Having said that, what I've found (for me) is that at 20% battery load (600W) the cells will sit at 3.3V when 100% charged. When the automation kicks in at 3.1V then there's 10-12% left in the battery - GE's firmware/software usually thinks there's about 25-30% left but that's nonsense. 3.06V = 5-8% left.

So what I see is 0.24V drop on a 3.3V cell equates to 95% of the charge at around 600W discharge rate. 0.2V drop = 90%.

It is clearly possible to track SoC accurately - other manufacturers manage it on LFP cells used for the same purpose.

The hardware used for the BMS on GivEnergy batteries is simply not capable of this without making the battery unusable for at least an hour every day by pausing its operation to run internal OCV calibrations. Nothing is going to change that apart from new hardware.

That's the state of play.

Edit - oh and the internal OCV calibrations only run/work below certain charge levels, which complicates matters further....