Consistent drops in SOC every time vehicle is turned off and on.

Hello,

I'm looking for advice on strange behavior I was able to observe on my Kia Soul EV 2018 (30kWh, 80Ah battery). Every time I turn it off and on again the SOC drops around 2.5%. It looks like it must stay in off state at least for some time for this to happen, but I have not determined minimal duration yet. The drop is more or less the same value irrespective either it was due to quick visit to supermarket, overnight stay at home or whole day parking at office. I checked it with Soul EV Spy and could see corresponding increase of around 0.8kWh in accumulative discharge power. The accumulative operation time also increases. I could see it going up by value starting from time car was off till about 2 hours.
So it all looks like car is having some kind of secret life by itself. I wonder if that is known behavior and part of normal behavior or should I worry?

You don't give us enough info to make a guess.

Examples
Do you live in Northern Canada or Alaska? My guess - battery heater.
Do you live in Dubai? My guess - battery fan.

At what SOC was the car when this happened?
Do you live on top of a hill?
Have you looked at the data for the 12V battery? Is it being charged?
Have you looked at the cell voltages of the main battery. is re-balancing occurring.

Thanks a lot for answer. Here comes more context:
I live in Europe. Car came from official dealer, so most probably designed for this reqion (54 deg. latitude). Neither south nor north, moderate climate. So far the effect was noticed at temperatures around 0-10 centigrade. I have no results for other temperature range yet.
Landscape is mostly flat, no significant altitude variations all the places I go during the day.
I do not yet know how to collect information about 12V battery. Any tool for that (Soul EV Spy seems to have a lot of information but it is not always clear what each parameter means)? Anyway 800Wh would literally blow 12V battery apart.
Cell voltages are balanced well. At stationary I never saw the delta bigger than 0.02V and that seems to be just the resolution at which BMS is reporting.
I typically keep battery between 25% and 83%, occasionally letting it to charge for unlimited time to give BMS a chance to re-balance cells.
I do not know how to check if BMS is doing any re-balancing at particular moment. Any suggestion?

Okay. there is something strange going on.
The main battery is normally 'disconnected' while the car is off. Only the 12V system stays on.
So my suggestions for recharging the 12V system or re-balancing cells are not normal.
To recharge the 12V system the car needs to be on and have some load eg headlights.
Calibration or re-balancing occurs at the end of a charging session.

For your car something unusual is waking up the car and keeping it awake for two hours.
My guess today is that you have the climate timer set to turn on the car and warm it up.

Thanks for suggestions. Unfortunately it does not look like climate control is involved as there is no preheating schedule active. And it happens irrespective of time of the day. But maybe charging schedule could cause such behavior, I dunno, will try to check.
Currently I have 2 hypotheses:
1. After idle time BMS performs some kind of remaining charge recalibration by cell voltages.
2. BMS reports consumed energy (or part of it) with significant delay so it takes some idle time for those number to settle in BMS and reach display.

I wonder if pulling the safety plug for a night would be safe enough procedure? That would allow to narrow down this behavior to battery or other electronics of the car.

Have you read this thread :- Latest E400 BMS update appears to have reduced usable capacity.

Did the glitch you are noticing begin after the 'fire-safety' recall?

I own this car for less than a year and have not been at local dealer for maintenance yet. Online check at https://www.kia.com/eu/service/kia-overview/service-and-maintenance/kia-recalls/ shows no pending recalls for this car. So chances are it has upgraded BMS. Is there any way to check BMS firmware/hardware version?

I read that thread (very interesting) and found no similarities to my case. Maybe I missed something.

Your suggestion that car may be performing scheduled heating gave my an idea to check for possible software glitches related to that feature. Originally both climate scheduling slots were empty. I filled both with arbitrary data and marked both as OFF. I also turned OFF charging scheduling and manually turned heating off before leaving the car. After more than 2 hours SOC stayed the same 32.5% according to Soul EV Spy. Interestingly cumulative discharge power still went up by 0.2kWh. But cumulative charge power ALSO WENT UP - by 0.1kWh. So some BMS tricks are definitely involved here. As overall battery voltage also went up by 1.7V, so maybe BMS applies some corrections and recomputes kWh and Ah values a bit. Guess I will have to collect more statistics at different settings and surprise dealer with few tough questions.

A.K. said:

.... Guess I will have to collect more statistics at different settings and surprise dealer with few tough questions.

Click to expand...

If they have any sensible answers, they will be an amazing exception. Generally, such questions yield either blank incomprehension, or clearly invented nonsense.

We have no way of simply reading the ROMID of the BMS.
Obviously there is a PID for this because the KIA software does read it.
If some hacker groups have found it they have never shared it openly.

Kia has changed the BMS software progressively with each update.
None of the DIY OBD tools used are able to keep track of this.

I used to analyse OBD data but gave up once I realised the data could no longer be reliably compared with any other car. I have two 21015 Soul EVs. If one performs differently I have no way of knowing if that is a glitch or a feature of a different BMS.

Examples :-

From :- Analysis of a replacement battery. Note the glitch at the top of the SOC


From :- Comparing BMS data on the Soul EV with the Ray EV

Just like in my previous test SOC (BMS) did not increase in a straight line. Why the change between 38% and 44% where it goes up faster?
This time I do not believe it is badly calibrated, given that I have already done a calibration test a week before.


Also the Cumulative Energy counter shows the same anomaly. This explains why the Energy Counter cannot be used to measure the capacity of the battery


The Cumulative Current is the same.


To show this is not real. The Energy Draw does not show this effect.

IanL said:

If they have any sensible answers, they will be an amazing exception. Generally, such questions yield either blank incomprehension, or clearly invented nonsense.

Click to expand...

I misread this as cleverly invented nonsense and assumed you were making a dig at how nowadays people rely on AI generated text. It superficially looks good but is actually stupid.

Actually, it accords with personal experience as well as reports from others.

Just wanted to share some ideas I came with after collecting more data (must praise Soul EV Spy here) and throwing away few theories. The interesting thing I noticed was that despite SoC dropping slightly almost every time I turn car off and on the cell voltages never dropped (reported resolution is not good enough, so there is still some probability that my observation is not correct). Except one case, but that coincided with significant drop of ambient temperature. And strangely enough few times I even noticed power counter representing energy flowing INTO battery to increase after car was turned off and on (not enough to affect SoC, but number definitely changed). Taking all that into account so far most prevailing theory (from my point of view) is that every time car is started BMS (or other module) recomputes SoC by using cell voltages and adjusts reported power flow counters and SoC value accordingly.

That seems very logical.

JejuSoul said:

Have you read this thread :- Latest E400 BMS update appears to have reduced usable capacity.

Did the glitch you are noticing begin after the 'fire-safety' recall?

Click to expand...

Hello there,

I report the same issue as @A.K. : a SoC (both BMS and displayed, as read on Car Scanner) drop that can reach 6%, usually with the following pattern:
I unplug the car at home at 83% (that also changed from 80 to 83% with the latest recall, thanks @JejuSoul for the explanation on the other thread), drive to work, leave it at for eg. 53%, and pick it up 10 hours later at 47% SoC displayed.

A few further observations:
- Car Scanner reports ~2 kWh of additionnal cumulative energy discharge, so I first thought that my car was somehow charging something (12v ? heater ?), but it doesn't seem to be the case as I don't use the planned heating, and I read that the 12v cannot just get charged if the car is off. Furthermore, the aux battery readings seem to be very good (14v when started, 11.8v otherwise).
- I didn't check yet the HV battery voltage values before/after
- My range doesn't seem to be affected whatsoever...
- The drop seemed to increase lately with low temperatures (at least for the region where I live - southern France). Last summer it was ~3% drop.

My guess at the moment is that it's a software glitch, and that the ~2 kWh of energy spent reported by Car Scanner simply do not exist.
My car is, like @A.K. , an old one with a new HV battery. I am not currently at home, but have access to another Soul EV falling in the same category: MY2015, new HV battery installed in 2021, BMS/harness recall last year. I will check if it has the same issue.

If you have any idea of a test I could do, I would gladly try it

I have contacted regional KIA representative with that issue. They claimed it may be fixed by BMS update (according to their records my car was missing latest update and latest update is supposed to improve SOC computation). Car went through yearly technical maintenance at local dealer week ago including BMS update and according to SoulEVSpy BMS is still computing SOH (and drops still happen). So I'm still waiting to see how it ends.
Concerning reason of this strange behavior I was trying to collect information and identify possible correlations. The most promising one so far is how big the drop is dependency on how much energy was spent before drop happens. It seems to be linear enough (at least as long as ambient temperature is stable and with limited data I was able to collect).
That could indeed be explained by wrongly evaluated SOH assuming SOC is evaluated by energy spent while car is running and recomputed from battery voltage after car goes off for long enough time (about 2 hours in my estimation). And taking into account SOC vs voltage dependency may be actually quite complex I'm looking ahead with hope for end of SOH computation.