ZS EV Mk2 51kWh LFP Battery - Not Balancing Massive Cell Deviation

[Peter WA]

I've got the Mk2 ZS EV with 51kWh LFP battery.

Despite being LFP I usually stop charging around 90%. In part because it is better for the battery, and mostly because I really hate it that KERS does not work properly with a fully battery.

Only once every month or two I charge to 100%.

I would expect that after reaching 100% the car spends some time balancing the cells. But there's little to nothing of that happening. Yesterday I noticed a cell voltage difference of ~260mV at the end of charge. Highest cell was 3.64V, lowest cell 3.38V.

I'm using OVMS to monitor these things. Here's a chart. After reaching 100% the charge current dropped to 9A, where it stayed for maybe 10 minutes, then less than 5 minutes at 3A (which should be the balaning current), then it went straight to zero. No proper balancing, despite massive imbalance between cells.

I realise that hardly anyone will have OVMS. So the main question to the group: can you see your MK2 spend time balancing? Meaning: the charger remains active, with a relatively low current, after the battery reaches 100%.

Or, if you happen to have an OBD2 dongle, and the app you use shows cell voltages: can you please check the cell voltage difference when you are at 100% fully charged?

 

[DB Go Rex]

Do you actually have a noticeable real world issue with your range/battery performance?

 

[fred12345]

Or, if you happen to have an OBD2 dongle, and the app you use shows cell voltages: can you please check the cell voltage difference when you are at 100% fully charged?

I use 3kW home plug charger which shows the current in real time in app and it definitely balances. Sometimes it takes quite a long time, probably over an hour at about 1kW or so... (From memory) . I do it once every 2-3 months but I don't have individual cells reading.

 

[Peter WA]

Do you actually have a noticeable real world issue with your range/battery performance?

The estimated range (GOM) is about 20km lower than it used to be. Could be imbalance, could be degradation, or could be a change in driving habits. Very hard to tell.

The problem with cell imbalance is that you'll only have an actual problem when your battery goes close to empty. What will happen then is that GOM over-estimates the range and suddenly the BMS shuts down the battery to protect the lowest cell from discharging to a level that would cause permanent damage. Leaving me stranded somewhere in the middle of nowhere...

Quite likely the reverse is happening right now when charging. The highest cell reaches the absolute maximum threshold of 3.65V and the BMS makes sure charging is stopped completely to protect it. Leaving other cells under-charged and over time growing the imbalance.

The fact that my battery is not balancing is a fault that will lead to more severe problems if not addressed. The point of this post is to find out whether my car is the exception (specific problem of the car), or whether it is a more wide spread issue.

 

[hermitdave]

My only experience with LFP is with home battery I installed 1.5 years ago. In early days I tried to limit charge to 90% with once a week full charge for balancing and started noticing substantial imbalance.

Having talked to the supplier and then reading the manufacturer documentation, I realised that the balacing kicked off once the cells hit 3.35V. The issue was that the a couple weren't hitting that and the rest wouldn't kick off the balancing.

Recently when we were looking at ZS EV and I was gravitating towards SE for LFP, I read up more on car LFP and according to Tesla and Ford (MACH E is LFP) the LFP packs should fully charged regularly. Once a month might not be enough IMO

If you look at the SoC / voltage curve it is near flat between 85%-20% so fully charging the pack takes each cell to max voltage allowing BMS to calculate health correctly.

Looking at your graph it seems like the one at low voltage is under 3.4V. Not sure what voltage the BMS kicks off balancing. 12:40 it seems that charge is complete and the battery is not charging / balancing.

I suggest you fully charge it for a week days / weeks and see if the deviation drops.

 

[fred12345]

I realised there is cell voltage in my car scanner app but no idea what it all means , good or bad ... Also, can't remember if this was at 100% charge or not ...

 

[hermitdave]

Decent video

 

[DB Go Rex]

The estimated range (GOM) is about 20km lower than it used to be. Could be imbalance, could be degradation, or could be a change in driving habits. Very hard to tell.

The problem with cell imbalance is that you'll only have an actual problem when your battery goes close to empty. What will happen then is that GOM over-estimates the range and suddenly the BMS shuts down the battery to protect the lowest cell from discharging to a level that would cause permanent damage. Leaving me stranded somewhere in the middle of nowhere...

Quite likely the reverse is happening right now when charging. The highest cell reaches the absolute maximum threshold of 3.65V and the BMS makes sure charging is stopped completely to protect it. Leaving other cells under-charged and over time growing the imbalance.

The fact that my battery is not balancing is a fault that will lead to more severe problems if not addressed. The point of this post is to find out whether my car is the exception (specific problem of the car), or whether it is a more wide spread issue.

I don’t think your estimated reduction in range is that significant. The BMS knows if it needs to do the equalisation charge.

 

[hermitdave]

Agreed GoM estimation is all about your driving style

 

[Peter WA]

I don’t think your estimated reduction in range is that significant. The BMS knows if it needs to do the equalisation charge.

You assume everything works as it should. Yes, the BMS should know if it needs to balance cells.

With a cell deviation as large as the one I experience, there is no doubt it should actually balance. Each and every time I charge to 100%. Until the cell deviation is reduced significantly.

It doesn't do that. It could be a software issue, after all, the BMS is mostly software. Or it could be a hardware issue. I.e. a balance cable not connected properly, or a resistor faulty.

Just for reference: I have built a home battery from individual LFP cells. 16kWh. It runs flawlessly, and I routinely achieve a cell deviation of less than 5mV at a cell voltage of 3.45. The reason to mention that: I know quite exactly how LFP batteries work, and what result can be achieved by proper balancing.

Unfortunately the car is very much a black box with very little information available. I don't even know whether it has an active balancer or only a passive one. Or what voltage it is supposed to start balancing at.

Even more unfortunate is that dealers and their workshops are mostly clueless.

Looks like I might have to wait until I get stuck with a battery that went from 10% to flat in an instant.

 

[hermitdave]

You assume everything works as it should. Yes, the BMS should know if it needs to balance cells.

With a cell deviation as large as the one I experience, there is no doubt it should actually balance. Each and every time I charge to 100%. Until the cell deviation is reduced significantly.

It doesn't do that. It could be a software issue, after all, the BMS is mostly software. Or it could be a hardware issue. I.e. a balance cable not connected properly, or a resistor faulty.

Just for reference: I have built a home battery from individual LFP cells. 16kWh. It runs flawlessly, and I routinely achieve a cell deviation of less than 5mV at a cell voltage of 3.45. The reason to mention that: I know quite exactly how LFP batteries work, and what result can be achieved by proper balancing.

Unfortunately the car is very much a black box with very little information available. I don't even know whether it has an active balancer or only a passive one. Or what voltage it is supposed to start balancing at.

Even more unfortunate is that dealers and their workshops are mostly clueless.

Looks like I might have to wait until I get stuck with a battery that went from 10% to flat in an instant.

From my experience with Leaf 30 and later with me messing up the balance on LFP I learned ignorance is bliss. Charge to 100% and don't over think.
Model Y I charge to 80% as it is recommended but MG will charge at needed to 100%

 

[Peter WA]

with me messing up the balance on LFP

What happened?

You might be right, I just found an old screenshot, from almost 2 years ago, when the cell imbalance looked pretty much the same as it is now. So while the car doesn't balance the cells, the imbalance doesn't seem to get worse.

At the same time my GOM range has deteriorated a lot more in the last few days. Until recently my range estimates (scaled to 100% charge) typically were over 350km. Right now it is down to 309km. Average power consumption has remained steady at 15kWh/100km for quite a while now.

 

[hmadsen]

What happened?

You might be right, I just found an old screenshot, from almost 2 years ago, when the cell imbalance looked pretty much the same as it is now. So while the car doesn't balance the cells, the imbalance doesn't seem to get worse.

At the same time my GOM range has deteriorated a lot more in the last few days. Until recently my range estimates (scaled to 100% charge) typically were over 350km. Right now it is down to 309km. Average power consumption has remained steady at 15kWh/100km for quite a while now.


View attachment 35039

Is this something you calculate based on some entities ?

If so, which ones, what's the calculus and what does these graphs show exactly ?

 

[DB Go Rex]

You assume everything works as it should. Yes, the BMS should know if it needs to balance cells.

With a cell deviation as large as the one I experience, there is no doubt it should actually balance. Each and every time I charge to 100%. Until the cell deviation is reduced significantly.

It doesn't do that. It could be a software issue, after all, the BMS is mostly software. Or it could be a hardware issue. I.e. a balance cable not connected properly, or a resistor faulty.

Just for reference: I have built a home battery from individual LFP cells. 16kWh. It runs flawlessly, and I routinely achieve a cell deviation of less than 5mV at a cell voltage of 3.45. The reason to mention that: I know quite exactly how LFP batteries work, and what result can be achieved by proper balancing.

Unfortunately the car is very much a black box with very little information available. I don't even know whether it has an active balancer or only a passive one. Or what voltage it is supposed to start balancing at.

Even more unfortunate is that dealers and their workshops are mostly clueless.

Looks like I might have to wait until I get stuck with a battery that went from 10% to flat in an instant.

Your evidence that you have a problem is a bit lacking that’s all I’m saying…

 

[Peter WA]

Is this something you calculate based on some entities ?

The BMS reports cell voltages via OBD2. OVMS reads out that information. The most helpful level of detail is to look at the highest and lowest cell voltage, labelled Max and Min in the first chart.

The difference between them is the cell deviation (calculated). In case of my battery that blows out to 270mV when charging stops. Meaning the highest cell voltage is 3.65V, and the lowest only 3.38V.

Just for reference: my home battery is made of 16 LFP cells. I charge them daily to a maximum voltage of 3.45V (a much safer cut-off than 3.65V). At that top voltage they get balanced, and achieve a cell voltage deviation of less than 5mV every single day. Some days as little as 2mV, so highest cell 3.451 and lowest cell 3.449. That is what a well balanced battery should be like.

I do realise that a LFP battery only holds a very small amount of charge between 3.45 and 3.65V, perhaps 0.2% of the total capacity. If I was to charge my home battery to 3.65V the cell deviation might blow out to 100mV or more.

At the same time, because they voltage curve goes very flat quite quickly, the battery already holds a significant amount of charge between 3.35V and 3.45V. Around 2% of the total capacity. Since only one cell runs away to 3.65V and most are still only 3.37V, that means nearly 2% of the battery capacity is not used.

That in itself is only a minor nuisance, but these kind of differences tend to build up. I remember that the Mk1 ZS EV had a software bug where people were losing 20 or 30% of their battery capacity to cell imbalance over time. And after a software update fixed the bug, it took weeks of daily charging to 100% for the balance to be restored and capacity recovered...

 

[Peter WA]

Ok, here's an idea to see whether my car is the odd one out. Or whether this is still "normal".

Right after a full charge to 100%, when first turning on the car, what voltage does your car report? Please only reply if you have a 51kWh LFP battery (Standard range Mk2 ZS EV).

 

[Outbackozzie]

Betweeen 410 and 415v.

 

[pollenface]

I think it's good to actually use the brakes regularly, keep the rotors shiny, I charge to 100% once or twice a week. Never paid attention to battery balance/SOH. I just charge it and drive it, I kinda don't want to know although do try to limit high SOC during summer. 99% of the time it's charged at 6A/1.5kW.

 

[T1 Terry]

With 12 yrs plus experience with over 250 systems to collect the data from .... you have a poor connection some where between the cell and the voltage monitoring device.

If you can see every cell voltage, the give away that a bad connection is the problem, the cell either side of the low cell is also outside the voltage range of the other cells.
The reason, the voltage monitoring device is reading the difference between the wire on the positive of the cell before, and the positive of the cell it is measuring the voltage across, this positive wire is connected to the series link to the next cell, so it also provides the negative voltage for the next cell while providing the positive voltage for the cell being measured ..... sounds like gobildigoop I know, and it leads to all sorts of misinformation being sent to the user ..... and that is why it's not a viewable screen in the standard software.

When the battery connections are easily accessible, you can check the voltage being displayed against the readings on a multi meter, if they are different, you can use the negative probe and positive probe either side of each connection .... if you see voltage being displayed, that is where the high resistance problem lies .....

Unfortunately, at the moment, we are reliant of the wiring and connections between cell being perfect .... or we are going to see B/S information.

The passive balancing is at cell level, it can't be across the voltage sensing wires because that would throw the voltage reading way out of kilter and meaningless ....

If the reading as consistently low, you don't see the low cell dropping close to 0V with a near fully drained battery under load, the cell voltage is fine, just the readings are inaccurate.

A down and dirty work around, if you can actually see each cell voltage, add the low voltage cell to the voltage of the cell before and the cell after, divide by 3, is that fairly close to the average cell voltage? If it is, the problem is poor resistance connection in the voltage measuring section, if the three voltages average to a voltage lower than the rest of the pack, you have a cell voltage problem ......

T1 Terry

 

[hermitdave]

Having dealt with a neighbour’s out of balance pack before, the only solution was to open up and then bring up the low voltage cells to match the rest of the pack.
Specialist garages like Cleevely can help with EVs but IMO the pack top voltage is 3.65 volt as BMS is continuously trying to top up the low voltage cells. IMO balancing isn’t kicking off.