Cell voltage difference a function of SOC, explanations?

[Peter WA]

Just thought I'd document today's charging session, from 18% to 88%.

I charge with an openEVSE, the current is limited by the available current from my PV system.

The first figure is the most interesting one, the difference between the lowest and highest cell in mV.

Who's got a theory to explain why it first went down to as low as 4mV at about 25% charge, then started rising to as much as 24mV at about 55% SOC, before dropping back down to ~16mV where it usually stays until the battery is 100% SOC?

 

[JodyS21]

One could make some observations from the data etc, but to be honest I think it’s just due to it being a chemical reaction etc that will mean no cell will be perfectly identical.
I would guess that the higher the rate of charge, the more chance individual cells will store power at different rates. Which kinda matches your graph profile a bit.

 

[Peter WA]

I'll run another test in the next days, charging over night from the grid, with a constant current. Let's see.

 

[Peter WA]

Ok, here the observations from the constant-current charge. Exactly the same pattern of cell difference. It appears to be directly linked to SOC.

It could be cell chemistry, but to be honest, I have my doubts.

Alternate hypothesis: BMS software, once the SOC gets to ~55%, actively kicks in to keep cell differences at bay.

Is the MG ZS EV with current BMS now equalising during the charge, as we know Tesla does?

Does that mean the hours of equalising at the end of a charge are a waste of time, energy, and probably put unnecessary strain on a mostly already balanced battery, reducing expected life for no reason at all?

 

[Kithmo]

On the old BMS, 456V, I balance once a month, never use rapids and my battery cells are always 40mV out after a balance. I see figures of 20-30mV for most of you guys. Do have anything to worry about ?

 

[JodyS21]

Ok, here the observations from the constant-current charge. Exactly the same pattern of cell difference. It appears to be directly linked to SOC.

It could be cell chemistry, but to be honest, I have my doubts.

Alternate hypothesis: BMS software, once the SOC gets to ~55%, actively kicks in to keep cell differences at bay.

Is the MG ZS EV with current BMS now equalising during the charge, as we know Tesla does?

Does that mean the hours of equalising at the end of a charge are a waste of time, energy, and probably put unnecessary strain on a mostly already balanced battery, reducing expected life for no reason at all?


View attachment 4761View attachment 4762View attachment 4763

Hmm yes it is pretty much identical, peaks at 50% ish.

I wonder if it's a certain cell pack that has the difference.
Can you generate the same graph but showing each of the cell packs?
I'm sure each cell will react slightly different, I'm very much doubt they test each cell in the factory and then selectively match cells up.

I've not got OVMS but could read the data using an app, not sure it easily gives you graphs etc for this cell pack level data, I maybe able to log it to a file and then produce some similar charging graphs. Would be interesting to see if every car (cell pack) is different or not.

 

[JodyS21]

On the old BMS, 456V, I balance once a month, never use rapids and my battery cells are always 40mV out after a balance. I see figures of 20-30mV for most of you guys. Do have anything to worry about ?

Most people's data I've seem do have around 20mV difference after a balance. So yours is a bit out of the normal range, not sure if it's really anything to worry about particularly though.
You could try looking to see if it is just a particular cell pack that has the larger than normal difference.

 

[Peter WA]

Can you generate the same graph but showing each of the cell packs?

Best I easily got was this here from another charge session that peaked at 24mV difference:

max 3.804,3.803,3.805,3.804,3.805,3.808,3.802,3.805,3.805
min 3.798,3.789,3.792,3.798,3.794,3.796,3.784,3.797,3.796

that means the 9 cell packs had differences of:
6,14,13,6,11,12,18,8,9 mV

a little later when it dropped back to 20mV I got:
max 3.837,3.835,3.837,3.837,3.838,3.84,3.836,3.838,3.839
min 3.832,3.824,3.827,3.831,3.829,3.829,3.82,3.831,3.829
5,11,10,6,9,11,16,7,10 mV

it appears both the difference within each pack and between the packs was getting smaller.

I'll collect a few more data points, but with no big trips planned the coming weeks not sure I get to the 'sweet spot'

 

[aleks57]

Best I easily got was this here from another charge session that peaked at 24mV difference:

max 3.804,3.803,3.805,3.804,3.805,3.808,3.802,3.805,3.805
min 3.798,3.789,3.792,3.798,3.794,3.796,3.784,3.797,3.796

that means the 9 cell packs had differences of:
6,14,13,6,11,12,18,8,9 mV

a little later when it dropped back to 20mV I got:
max 3.837,3.835,3.837,3.837,3.838,3.84,3.836,3.838,3.839
min 3.832,3.824,3.827,3.831,3.829,3.829,3.82,3.831,3.829
5,11,10,6,9,11,16,7,10 mV

it appears both the difference within each pack and between the packs was getting smaller.

I'll collect a few more data points, but with no big trips planned the coming weeks not sure I get to the 'sweet spot'

I also have a 7 block has an imbalance of 18-19mV. This is the largest value.

 

[Kithmo]

Most people's data I've seem do have around 20mV difference after a balance. So yours is a bit out of the normal range, not sure if it's really anything to worry about particularly though.
You could try looking to see if it is just a particular cell pack that has the larger than normal difference.

I'm not overly concerned but just as a matter of interest.
The Torque app shows each single cell to be 4.2V but only displays to one decimal place, but the ZS EV app only shows the max (4.22V) and minimum (4.178V) so I can't see which individual is low or high.
Could it be the highest is too high at 4.22V or would it be the lowest at 4.178V is too low ?

 

[Deleted member 2439]

I’m not being funny - but exactly why are you doing this? Aren’t you causing yourself undue stress over something that you can’t do very much about?

 

[Kithmo]

I’m not being funny - but exactly why are you doing this? Aren’t you causing yourself undue stress over something that you can’t do very much about?

I'm not stressed, just interested.
Honest.

 

[Dave53]

I’m not being funny - but exactly why are you doing this? Aren’t you causing yourself undue stress over something that you can’t do very much about?

They're having fun let them get on with it, they also come up with things that are of intrest to rhe members regarding their cars

 

[Kithmo]

TBH, I only downloaded the app and used it because I already had a Bluetooth OBDII adaptor and most of the non battery data doesn't mean anything to me, if I hadn't already got the adaptor, I wouldn't have bought one.
I already had Torque Pro as I used it to find which cells were duff on SWMBO's old Prius so I could replace them to get rid of the red triangle of death before selling/trading it in.
I pretty much use the car for what I consider an ideal situation for an EV with limited range, short journeys and apart from this little dabble with the app (3 times in the last year), I just drive it, charge it when it gets low and drive it some more.
I don't even look at the screen with the Voltage, Amperage and such on, I only use the GOM and battery meter to check if I need to charge it and occasionally I look at the m/kWh figures to see how my driving is, although I don't really need to do that as I just drive it normally like I would an ICE car so I'm not going to adjust to improve it anyway.

 

[Deleted member 2439]

TBH, I only downloaded the app and used it because I already had a Bluetooth OBDII adaptor and most of the non battery data doesn't mean anything to me, if I hadn't already got the adaptor, I wouldn't have bought one.
I already had Torque Pro as I used it to find which cells were duff on SWMBO's old Prius so I could replace them to get rid of the red triangle of death before selling/trading it in.
I pretty much use the car for what I consider an ideal situation for an EV with limited range, short journeys and apart from this little dabble with the app (3 times in the last year), I just drive it, charge it when it gets low and drive it some more.
I don't even look at the screen with the Voltage, Amperage and such on, I only use the GOM and battery meter to check if I need to charge it and occasionally I look at the m/kWh figures to see how my driving is, although I don't really need to do that as I just drive it normally like I would an ICE car so I'm not going to adjust to improve it anyway.

Fair enough - same as me really. I’ve got a Bluetooth ODB dongle that I’ve used on ICE cars to look at stuff - but have deliberately kept clear of plugging it into the ZS for fear of becoming “bogged down“ in “stuff”.

 

[JodyS21]

Fair enough - same as me really. I’ve got a Bluetooth ODB dongle that I’ve used on ICE cars to look at stuff - but have deliberately kept clear of plugging it into the ZS for fear of becoming “bogged down“ in “stuff”.

Are you not interested in knowing if you have the latest version of the software? & SOH?

 

[JodyS21]

I'm not overly concerned but just as a matter of interest.
The Torque app shows each single cell to be 4.2V but only displays to one decimal place, but the ZS EV app only shows the max (4.22V) and minimum (4.178V) so I can't see which individual is low or high.
Could it be the highest is too high at 4.22V or would it be the lowest at 4.178V is too low ?

It seems to be impossible to get the voltage of each individual (108) cell.
All you can obtain is the voltage range within each of the 9 battery packs, so the minimum voltage out of those 12 cells and the maximum. It could be 11 of them are 4.178V and 1 is 4.22V or, 1 is 4.178V and 11 4.22V or, anything in-between.
I presume their idea is that only whole battery packs are changed should there be an issue, rather than individual cells within them, so they only need to know at that level....

The voltage data is provided to 3 decimal places, so anything shown less accurately would be due to an app deciding to do rounding etc rather than show it in full.
(Some apps are only setup to show you the overall minimum and overall maximum voltage, rather than do it per battery pack)

I've had "fully charged/balanced" cell voltages that range historically between 4.171 maximum and 4.139 minimum, highest imbalance instance I've seen is 0.020V.
Interesting that I've never got near the 4.22V you're seeing (I've always been on the 450v BMS).

 

[Deleted member 2439]

Are you not interested in knowing if you have the latest version of the software? & SOH?

Not particularly - being totally honest. Don’t forget that we are moving on from the ZS - in another 6 weeks or so. If we were keeping it then maybe …..

 

[Peter WA]

I’m not being funny - but exactly why are you doing this? Aren’t you causing yourself undue stress over something that you can’t do very much about?

I enjoy understanding stuff. Anything that keeps my brain active in a variety of topics, not just regarding the ZS EV. You could perhaps compare it to someone doing Sudoku puzzles, except here there's the off chance of discovering something that will actually make a practical difference.

I have no idea where this particular discussion might lead. But looking at the cell voltage differences in general I have for myself come to the conclusion that with the BMS version that is installed in my car right now, I won't do regular balancing sessions in future.

Each balancing session uses between 1 and 1.5kWh of electricity, which appears to be a complete waste of energy. Balancing also takes place at a nominal 100% charge level, which is very close to the voltage where batteries can suffer permanent damage if they get even a little bit over-charged.

People who only drive their cars for a few years might not care, but I tend to drive my cars until they are no longer economically viable (repair bills for the old one exceed the loss of value a newer one would suffer). Keeping the battery in top shape well beyond the warranty period is the best proactive thing I can do to get the most out of this car.

 

[Kithmo]

It seems to be impossible to get the voltage of each individual (108) cell.
All you can obtain is the voltage range within each of the 9 battery packs, so the minimum voltage out of those 12 cells and the maximum. It could be 11 of them are 4.178V and 1 is 4.22V or, 1 is 4.178V and 11 4.22V or, anything in-between.
I presume their idea is that only whole battery packs are changed should there be an issue, rather than individual cells within them, so they only need to know at that level....

The voltage data is provided to 3 decimal places, so anything shown less accurately would be due to an app deciding to do rounding etc rather than show it in full.
(Some apps are only setup to show you the overall minimum and overall maximum voltage, rather than do it per battery pack)

I've had "fully charged/balanced" cell voltages that range historically between 4.171 maximum and 4.139 minimum, highest imbalance instance I've seen is 0.020V.
Interesting that I've never got near the 4.22V you're seeing (I've always been on the 450v BMS).

The Torque Pro app makes it look better though, where they're all the same at 4.2V.