Is this SoH OK?

[Data]

The BMS will use coulomb counting, rather than voltage for SOC calculation.

Because the upper knee is so steep the delta will be quite large (i.e. compared to NMC cells) when cell voltages are over 3.45V per cell, despite them being relatively well balanced.

Graph below is no-load voltage vs. SoC for typical LiFePO4 cells

That's perfectly correct Everest. Good info.

 

[Everest]

Anyone care to explain

My understanding, which is from LiFePO4-based home energy storage balancing (so may be different with EVs) is that balancing is based on cell voltages, but with LFP can only commence when the cells are above a certain threshold (typically about 3400mV) as below that voltage is not a good indicator of state of charge.

However, those voltages will only be reached when there is a charge current being applied, even if that current is small (e.g. my home cells will be charging with a current of only 4A at that point).

Once the charge current is removed, the cells will soon settle to a lower voltage (say around 3600-3800mV). Hence balancing - at least for home storage systems - is performed when cells exceed (say) 3400mV AND when a charge current is still present.

The challenge, therefore, is to be able to balance enough, in the short time there is before the highest voltage cell reach the designer's maximum voltage (likely to be between 3500mV and 3650mV (the latter being the absolute max recommended cell voltage for LFP))

So, in response to your specific points... my 2p worth would be...

Balancing (at least the way I understand it) commences when charging is completed (I.e. when the first cell reaches the cut-off voltage).

I suspect that it will be done during the upper SOC levels but during charging, rather than when complete.

Balancing therefore works with cell voltages (not SoC) which can be determined quite accurately, despite them being in the upper knee?

Absolutely based on voltage, but because, rather than despite them being in the upper knee.

Happy to be corrected by those on here who are more familiar with EV BMS's compared to home-storage BMSs though (e.g. @T1 Terry & @Coulomb)

 

[Data]

Balancing (at least the way I understand it) commences when charging is completed (I.e. when the first cell reaches the cut-off voltage).

Passive balancing:
The BMS then determines the lowest cell voltage and turns on the blalancing resistors of every cell above that level and thus ‘bleeds off’ excessive capacity (turning it into heat) until they are all equal.

Active balancing:
Excess capacity of the highest level cells is used to bring up the lower levels cells until an equilibrium is reached.

Balancing therefore works with cell voltages (not SoC) which can be determined quite accurately, despite them being in the upper knee?

None of these methods require additional energy from the EVSE therefore I’m puzzled why there seems to be a continuous (albeit reduced) power draw after charging is completed.

Both methods would also suggest that when balancing is completed, all cells would end up at a lower level than cut off voltage.

Anyone care to explain where I’m going wrong?

Cell voltages determine state of charge. I think you are going wrong because you are assuming that balancing brings all cells up to the same level. It doesn't always, even though that is the aim it isn't possible when some of the cells have started degrading. Cells degrade, charge & discharge all at differing levels & with differing timings. As such the BMS has to assume a median level of charge & will display that median charge. This is why as the battery degrades it can become more unbalanced, the range shown especially as the battery depletes nearer the lower end of soc can suddenly go from say 35 miles to 11 miles in the space of 3 or 4 minutes leaving the driver in panic, thinking they will not make it to the next charging station. Initially this whole process is happening in a very small way but can increase over long periods of time.

Let's not forget our batteries are very robust & last a long time whether they are LFP or NMC. They don't break easy! No one should be worrying to any degree if you are following the manufacturers instructions. The engineers who design the batteries know what they are doing & talking about.

 

[MickeySw]

Cell voltages determine state of charge. I think you are going wrong because you are assuming that balancing brings all cells up to the same level.

I’m pretty certain that (near) perfect top balancing is possible with degraded cells; SoH of my 15 year old EV is around 80%, yet top balancing is within a few milivolts??

It doesn't always, even though that is the aim it isn't possible when some of the cells have started degrading. Cells degrade, charge & discharge all at differing levels & with differing timings. As such the BMS has to assume a median level of charge & will display that median charge.

A HV battery is usually made up of a bunch of individual cells all connected in series.

The current going through each cell is therefore the same. Degraded cells will have a lower capacity than the rest, however their voltage/SoC ratio hasn’t changed.

Their voltage will therefore drop faster and the imbalance will become larger and ‘fall off the cliff’ rapidly at the lower end of the curve.

When changing, the opposite happens, degraded cells' voltages increase faster and once they reach cut off voltage, they will have more or less caught up with the good ones, any difference can be corrected with balancing.

During a BMS calibration, coulomb counting (integrating current over time) is used to measure the energy added until the 1st (usually the weakest) cell reaches the cut off voltage.

Let's not forget our batteries are very robust & last a long time whether they are LFP or NMC.

Agreed, but I still like to know how this is achieved exactly..

 

[Data]

I’m pretty certain that (near) perfect top balancing is possible with degraded cells; SoH of my 15 year old EV is around 80%, yet top balancing is within a few milivolts??

A HV battery is usually made up of a bunch of individual cells all connected in series.

The current going through each cell is therefore the same. Degraded cells will have a lower capacity than the rest, however their voltage/SoC ratio hasn’t changed.

Their voltage will therefore drop faster and the imbalance will become larger and ‘fall off the cliff’ rapidly at the lower end of the curve.

When changing, the opposite happens, degraded cell’s voltages increases faster and once they reach cut off voltage, they will have more or less caught up with the good ones, any difference can be corrected with balancing.

During a BMS calibration, coulomb counting (integrating current over time) is used to measure the energy added until the 1st (usually the weakest) cell reaches the cut off voltage.


Agreed, but I still like to know how this is achieved exactly..

Not quite sure that you have quite a full grasp of EV batteries Mickey. No disrespect intended. However, in response to your question relating to how good long battery life is achieved, I refer you to my earlier post. It's simple.

 

[doubledroz]

Any advice on how an EV ‘noob’ with no diagnostic tools is going to verify that the ‘proper equalisation charge’ has completed correctly?

If you have access to the iSmart app for your car, if you're at 100% charge, the Completion Time is --:-- and Charging Power is --, the car still has the "Stop Charging" button in the Charging Management page, you're doing an equalisation charge.

If your button says "Start Charging" (or similar - sorry, I forgot the exact wording last time) then you've stopped charging, and the equalisation charge either didn't happen, or has completed.

doubledroz has already stated he has been unable to carry out an equalisation charge on his 3yr old ZS with an LFP battery pack.

Not quite true - I interrupted the first equalisation after a bit, then a couple of days later I tried again from a high charge state and it did equalise, and then my latest charge it did not.

I'm charging again today from about 65% so I'm full for the weekend - I took a screenshot of my cell voltage delta, and I'll post the before and after images once I'm done a little later in the day!

 

[T1 Terry]

Interesting. There is at least a lot more understood now than there was back in 2011 when I started long term testing and BMS design for our house battery systems.
Only two states of charge are a known, 100% SOC is when a cell will hold 3.5vdc or better for 24 hrs, 0% SOC is when the cell passes below 3vdc while under the prescribed test load that the capacity testing was carried out.
Each manufacturer seems to have their own figure for this and no doubt matches a sweet spot in the particular cell build. Generally accepted has been 0.5C or C2 load, discharged from 100% to zero % over 2 hrs under a constant load.

There are two types of balancing recognised, top balancing, where all the cells reach the same voltage above 3.45v, and bottom balancing where all the cells hit 2.8v at the same time.

These two different balancing method use two totally different charging regimes.
Top Balancing
The only method that actually does do what is claimed, is active balancing and top end balanced charging current less than the ability of the active balancer to shift capacity from the high cell/s to the low cell/s.
At the moment, I don't know of any EV manufacturer that uses this method because of cost and complexity .... so, top balancing will always be a compromise and probably never actually get all the cells up to 100% capacity .... cost v return results in probably 98% SOC as best .....

Bottom Balancing
This involves getting all the cells while under the test load at the same voltage. This was the preferred for DIY EV builders for many yrears. The idea is, no cell gets dragged down to a voltage that will damage it beyond recovery. Drag the voltage to 0 VDC while the cells around it are still at better than 2.8V will cause the current from the higher cells to be dragged through this 0V cell, effectively reverse current flow and the cell is dead in seconds.
The only way to recharge using this bottom balancing method is the stop charging as soon as a cell sees between 3.6 VDC and 3.8 VDC.

When individual cell voltage monitoring and control became relatively standard practice, bottom balancing became redundant really.


To clear up any misunderstandings or conceptions.
An LFP cell that holds 3.5 VDC or better for 12 hours is 100% charged, this is not 100% of the advertised capacity, it is just the maximum amount of capacity that cell can hold at that particular time.
An LFP cell at 2.8V under load is at 0% capacity, that is not the same as 0 volts which is the point the cell is destroyed .....
Resting voltage has nothing to do with anything other than 100% SOC, there no such thing as an SOC% scale based on cell voltage, that is just guess work.
SOC measured by capacity used v capacity returned is close, not an exact science but close.

T1 Terry