Is this SoH OK?

[wattmatters]

The MG owners manual isn't written for people to easily understand this process

It's not written for technical people either. It's full of translated information that is sometimes inconsistent.

It says that in our MG manuals (for the ZS page 9)

The MG4 manual is pretty ambiguous about the need for a (near) full discharge prior to an equalisation charge.

It just says a "full slow charge" is required, which can simply mean charging the car to full.

Further in it gives expected charge duration for slow charging which includes an equalisation charge at the end, but again it does not explicitly state that prior deep discharge is required for equalisation to occur.

 

[Data]

Yes, the ZS manual is similar although does mention iirc taking the soc down to just below the low battery warning to do a full & proper equalisation. On my NMC 72.6kWh battery that usually happens at 20% SoC. Anyway for clarification I went to my dealer & MG technical who told me to get the low battery warning then charge slowly non stop & let the BMS control the charge all the way to 100%. And don't interrupt it or it won't fully equalise.

 

[Everest]

My point being that very few people actually get issues with batteries. It is clear "it works ok" might be a better way of putting it. We know active balancing is better, but not really needed.

Agree - passive balancing works OK. And is probably safer in a (relatively) hostile environment like an EV. Active balancers can, and do, fail. When that failure happens due to poor connections and under-reading of a cell's voltage they can pump too much energy into a cell and cause over-voltage / bloating issues. That mode of failure is not possible with a passive balancer.

 

[MickeySw]

View attachment 44699

First of all thanks for sharing 'real life data', am I correct in assuming you used a granny EVSE with the aim to balance the cells? Did you by any chance check the result min/max cell voltages aferwards?

I would be grateful if some of the 'gurus' can help me understand what I'm seeing:

- for starters let's assume that there are no losses and all the energy goes into the HV battery (400VDC @100%SoC, 200Ah cell capacity).

• as long as SoC is below 100% the current into the battery is about 3.75A
• however after being 'fully charged' 3.75A is still 'flowing' for another 35min (2.2Ah)
• the current then sharply drops to about 0.6A afterwards for approx 25min (0.25Ah)
• the remaining 2.5h (balancing?) are averaging around 125mA (0.3Ah)

What triggers the initial current reduction:
a) the highest reported voltage cell is reaching cut-off voltage
b) the highest reported voltage cell is getting close to reaching cut-off voltage
c)?

At what stage are the parallel bleeding resistors (post #127) activated?
a) after SoC 100%
b) after 1st current drop
c) during the 2.5h (balancing) section
d)?

Assuming that to increase the cell voltage by 1mV, approx 150mAh of energy is required, there doesn't seem to be any scope to significantly 'lift' low reading cells during the balancing phase?

 

[Data]

Agree - passive balancing works OK. And is probably safer in a (relatively) hostile environment like an EV. Active balancer can, and do, fail. When that failure happens due to poor connections and under-reading of a cell's voltage they can pump too much energy into a cell and cause over-voltage / bloating issues. The mode of failure is not possible with a passive balancer.

The safety issue is actually a valid one Everest. I understand from my guru friend that it can be an issue especially as the batteries age, often triggered by corrosion. Passive has few drawbacks apparently. Interesting, & probably another reason on top of cost, that it's not very popular.

 

[Everest]

The safety issue is actually a valid one Everest

Yes, I didn't make it up - has been an issue for at least a couple of home energy storage battery systems on a solar forum I'm active on.

 

[Data]

First of all thanks for sharing 'real life data', am I correct in assuming you used a granny EVSE with the aim to balance the cells? Did you by any chance check the result min/max cell voltages aferwards?

I would be grateful if some of the 'gurus' can help me understand what I'm seeing:

- for starters let's assume that there are no losses and all the energy goes into the HV battery (400DC @100%SoC, 200Ah cell capacity).

• as long as SoC is below 100% the current into the battery is about 3.75A
• however after being 'fully charged' 3.75A is still 'flowing' for another 35min (2.2Ah)
• the current then sharply drops to about 0.6A afterwards for approx 25min (0.25Ah)
• the remaining 2.5h (balancing?) are averaging around 125mA (0.3Ah)

What triggers the initial current reduction:
a) the highest reported voltage cell is reaching cut-off voltage
b) the highest reported voltage cell is getting close to reaching cut-off voltage
c)?

At what stage are the parallel bleeding resistors (post #127) activated?
a) after SoC 100%
b) after 1st current drop
c) during the 2.5h (balancing) section
d)?

Assuming that to increase the cell voltage by 1mV, approx 150mAh of energy is required, there doesn't seem to be any scope to significantly 'lift' low reading cells during the balancing phase?

Are you concerned about your own battery Mickey?

As I understand it there can be variations in the answers to your questions. It can depend how the BMS has been setup (how it's software has been written). Also, how the battery has been built. Don't ask me about the battery build variations. That's the answer my German friends gave me. They research & develop EV batteries & have knowledge far above my pay grade.

Q. What triggers the initial current reduction?
A. Usually it's the highest cell set reaching a predetermined voltage. This may or may not be the cut off voltage. Often it is a voltage just below the cut off voltage. This is done so as not to stress the cells.

Q. At what stage are the parallel bleeding resistors activated?
A. The resistors can be activated at virtually any state of charge. This could be 60, 70 80% during the final phase of the charging as long as smart charging is disabled. In other words an equalisation charge (a partial equalisation) can occur at various soc. They can even operate when the car is NOT charging to prevent a battery becoming very unbalanced. I'm not totally sure if that is the case for our MG's but I believe it is the case. Most car EV batteries can do this apparently. Given that MG batteries are high quality I'd be surprised if that wasn't the case. There is no predetermined length of time for the equalisation period. If the car battery is unbalanced it may take an hour, or 30mins. Equally it might take 2hrs or more. My car takes anywhere from 12 mins to 30mins. It's usually not very out of balance.

Dunno if any of this helps but these questions are not really very relevant to knowing how to look after your battery. There are often no precise answers. It doesn't actually help to know this sort of stuff unless you are a developer or battery tech. My answers have been checked with a battery guru.

 

[wattmatters]

Note that the MG4 user manual does not distinguish between an equalisation methodology for LFP and NMC model cars.

While the fundamental principles are the same, the reality is the different models do it differently.

One example is since you can't set an upper charge SOC for an LFP battery model then it will only do an equalisation at a SOC of 100%. NMC models however will still do go through an equalisation charge phase at lower SOC levels.

 

[Data]

Good point Mr Matters!

 

[MickeySw]

Dunno if any of this helps but these questions are not really very relevant to knowing how to look after your battery.

I beg to differ, but let’s try something closer to home specifically the behaviour of your ZS:

One distinct difference between a ‘proper equalisation charge’ and ‘an equalisation charge’ is that the former will do a BMS calibration (although SoC won’t get better than 100%).

Your experience suggests that there is a difference in the way the cells are balanced, specifically that the ‘proper’ way will yield better results?


I know you don’t feel the need to look at actual cell data to quantify these differences but I would be delighted if you could provide the following data given that you keep a close eye on things anyway?

a) imbalance @current SoC I.e. min/max cell voltage, total battery voltage
b) imbalance, battery voltage @ low SoC (before calibration)
c) imbalance, battery voltage and SoC once balancing has concluded
d) EVSE data, the last two hours

While the fundamental principles are the same, the reality is the different models do it differently.
NMC models however will still do go through an equalisation charge phase at lower SOC levels.

It’s not just cell chemistry, I.e. MG5’s NMC batteries will only top balance @ 100% SoC?

 

[Data]

Mickey, my next "full & proper equalisation' won't be for another 6-8 weeks. I've just carried out one a few days ago. I don't keep records of my checks & these days I don't even bother to look at OBD2 data. There's no point as the car is performing correctly. I'll endeavour to do so next time & will let you know. Are you concerned about your own battery & are looking for comparisons?

Regarding your MG5. Are you saying your car won't carryout a partial top balance at say 80%? My car (which is also NMC) wouldn't do a partial equalisation at any SOC other than 100% either. That is until I realised I had to turn off smart charging which no one on here mentioned at the time. Once I did that I found the car would equalise (partially) at any SOC & fully at 100%.

What size is your battery?

 

[MickeySw]

Mickey, my next "full & proper equalisation' won't be for another 6-8 weeks. I've just carried out one a few days ago.

Your current values will do for comparison reasons

Regarding your MG5. Are you saying your car won't carryout a partial top balance at say 80%?

I have a ‘dumb’ EVSE and charge exclusively during the night (61kWh battery).

I usually limit SoC to 70% but charge to 100% (and leave it plugged in) about once a month before I head off on a long journey.

According to the MG5 manual, SoC needs to be 100% before balancing commences.

The current imbalance is about 13mV at all SoC levels, haven’t really seen that delta change, therefore I’m not sure if that’s too low to trigger balancing?

 

[T1 Terry]

First of all thanks for sharing 'real life data', am I correct in assuming you used a granny EVSE with the aim to balance the cells? Did you by any chance check the result min/max cell voltages aferwards?

I would be grateful if some of the 'gurus' can help me understand what I'm seeing:

- for starters let's assume that there are no losses and all the energy goes into the HV battery (400VDC @100%SoC, 200Ah cell capacity).

This one is unlikely, the 400vdc refers to and architecture, 400v, 800v 1,000v are the common ones in commercial EV's, DIY started as low as 72vdc.
As an example, the LFP MG4 51 is a 400v architecture, 104 cells, nominal voltage 3.2v, 330v, it will range between 3v @ 0% SOC, 312v and 3.6v @ 100% SOC 375vdc ..... but all of these voltages as still give or take a bit, absolutely fully balance at the end of charge would be 375vdc, and highly unlikely with resistor type balancing, it would be an absolute fluke for that to ever occur.

As the battery approaches 0% SOC as the BMS sees it, a warning will come up and performance will be reduced.

The same goes for 0% SOC, the battery should cut supply if any cell drops to 3v, this will be the cell with the least capacity at the time, the other cells will be higher than 3v, but as soon as the supply is cut, that 3v cell voltage will climb ......

this part is a guestimation by me, the supply is reconnected, but the available current is much reduced to get that last bit out of that low cell to get you home or at least off the road.

Passive balancing of the type used by the majority of EV makers that are built to a price, only top balance, so what is left will not be shared across all 104 cells, a good active balancer will, all but very slowly, they rely on voltage difference between high cell and low cell voltage to transfer capacity from one cell to another.

• as long as SoC is below 100% the current into the battery is about 3.75A

Probably not, the total supply current divided by the number of cells, less and electrical energy lost as heat while operating all the BMS and conversion from AC to DC and heat generated in any cabling .... and anything that is drawing power at the time, such as battery heating or cooling, charging the 12v battery, any lights using power out of the 12v battery, the air con if it's running, etc .... At 6 amps 240vac charging and the air con running, the supply is not keeping up with the demand ..... the losses are considerable

As the assumed SOC approaches 100%, the charge current will taper off to avoid cell voltage cut off over run. Cells with a higher internal resistance than others, will have a lower terminal voltage at a lower charge current, this means SOC while charging can not be assumed by voltage, there are a lot of factors involved.

• however after being 'fully charged' 3.75A is still 'flowing' for another 35min (2.2Ah)

Not likely, once a cell reaches the high voltage cut off, in these simple systems, charging stops, the power used from the mains is simply to power the BMS.

• the current then sharply drops to about 0.6A afterwards for approx 25min (0.25Ah)

No, in the cheaper systems, there is not trickle charging to attempt to reach 100% SOC.

• the remaining 2.5h (balancing?) are averaging around 125mA (0.3Ah)

The balancing resistor is only active on the high voltage cell/s to bring them back to 3.4v, any cells less than 3.4v, tough luck, maybe the owner will do another balance real soon to give those cells a chance to get closer and maybe exceed 3.4v so they can be balanced.
It is important to note here, with LFP cells, at a rested 3.4v, attempting to push even milliamps into the cell will increase the voltage to 3.45v min, the more current that is attempted to push into the cell, the higher the terminal voltage will climb.....
A cell at 3.5vdc, rested for 12 hrs, has a 100% capacity voltage of 3.4v, and a surface voltage of 0.10vdc, that has no real capacity, a 1 amp load will burn that surface voltage capacity off within 30 secs or so, you can not have a cell charged to 105%, only 100% plus a surface charge voltage .... I could explain how that is stored, but for another time.

That is the major questions answered all the rest as so variable, they can not be answered definitely .....

What triggers the initial current reduction:
a) the highest reported voltage cell is reaching cut-off voltage
b) the highest reported voltage cell is getting close to reaching cut-off voltage
c)?

At what stage are the parallel bleeding resistors (post #127) activated?
a) after SoC 100%
b) after 1st current drop
c) during the 2.5h (balancing) section
d)?

Assuming that to increase the cell voltage by 1mV, approx 150mAh of energy is required, there doesn't seem to be any scope to significantly 'lift' low reading cells during the balancing phase?

No, this is a wrong assumption, a cell voltage will increase 0.05v before any charging current will be pushed into the cell, just part of physics and the cell chemistry resistance to change ..... maybe that is why humans can be so stubborn at times, a lot of chemical reactions going on in one of them

I hope that helps and doesn't just add to the confusion, the learning curve is very steep for how LFP cell function compared to Lead acid or any other chemistry electricity storage device

T1 Terry

 

[MickeySw]

I hope that helps and doesn't just add to the confusion, the learning curve is very steep for how LFP cell function compared to Lead acid or any other chemistry electricity storage device

Appreciate the explanations re LFP, I should have mentioned that all my calculations/conclusions were based on NMC chemistry.

But from your reply I gather the following also applies for NMC cells:

• The BMS will stop the charging process when a cell comes close to cut-off voltage.
• The BMS then determines the lowest cell voltage and activates the bleed resistors of every cell above that threshold,
• Balancing concludes when all cells are at the same level.

Assuming above is correct this would mean

• bleed resistors are not activated during the charging process
• cells will never reach cut-off voltage
• multiple balancing sessions are required for best results

 

[T1 Terry]

Appreciate the explanations re LFP, I should have mentioned that all my calculations/conclusions were based on NMC chemistry.

But from your reply I gather the following also applies for NMC cells:

• The BMS will stop the charging process when a cell comes close to cut-off voltage.

It will taper the charging current, not stop until a cell reaches the preset cut off voltage, probably 4v, even though fully charged is 4.2v rested, they get hot real quick above 4v and have that nasty tendency to go into thermal run away and there is no catching them once that happens.

• The BMS then determines the lowest cell voltage and activates the bleed resistors of every cell above that threshold,

No, the BMS for each cell has a turn on voltage and a turn off voltage, probably around 3.9v turn on/ 3.85v turn off.
Considering nom. voltage is 3.7v, the balance resistor will be operating for the last part of charging, but maybe 15 mins at the absolute most. It will drag the high cell down to maybe 3.85v and allow it to rebound, possibly cutting back in if the voltage goes above 3.9v again .... depends how smart the BMS it ....

• Balancing concludes when all cells are at the same level.

Balancing ends when all cells are below the cut off balancing voltage, guestimated at 3.85v, any cell less than that voltage is not fully charged, better luck next time ....

Assuming above is correct this would mean

• bleed resistors are not activated during the charging process
• cells will never reach cut-off voltage
• multiple balancing sessions are required for best results

I think all of these have been dealt with in the previous responses.
I have a real dislike for the NMC chemistry, it is barely a step above LiPo used in model planes etc as far as safety. They will burst into flame if charged too fast, over voltage charged, pulled down below their safe cut off voltage and this can be caused by excessive discharge current when the cell is in a discharged state that it can't deliver that much current, most likely the reason the MG4 XPower derates performance under 50% SOC, better an owner complaining about loss of performance, than loss of a vehicle and needing new under wear when the battery pack bursts into flame.
The voltage NMC cells go into thermal runaway due to low voltage, depends on a heap of different factors, internal resistance is one of them, a 0.5C load might pull a weak cell down to 3v at 50% SOC, the same cell at a 2C load would drop below 2v and already be heating up, ready to release the fireworks ...
A good low resistance cell might show the same characteristics down as low as 20% SOC, they have to be treated as an unknown quantity if all safety parameters are to be considered, so the worst case is taken as the best case scenario, don't discharge below 20% unless the current draw is limited to the worst case scenario at a such a "low state of charge" or 20% remaining in the case of an LFP cell, low SOC is considered to be 5% or less with an LFP cell.

NMC are their own type of cat, balancing more than 8 LFP cells is like trying to herding feral cats using Dingoes, NMC are a bit like Leopards using Labradors and expecting to still have the full count of dogs once you have given up trying to get the Leopards into the enclosure .....

Top balance NMC cells to 4.2v at your own peril, like tap dancing in a mine field, you might get away with it once, maybe a few times, but your luck will run out ....

T1 Terry