MG4 51kWh LFP Battery Health

Peter WA said:

You've got the JK Inverter BMS? That's the one I have. It happily talks with my Victron Multiplus II 48/5000 Inverter Charger via CAN bus. I run it in close-loop configuration (BMS is in the driver's seat, Inverter does what it is told with respect to target voltage and current) and have not had any issues yet.

Click to expand...

I have the older style non inverter version. It has a TTL port and they sell a RS485 adaptor, but it doesn't support Pylontech protocol, that my inverter expects. I'm told only inverter style JKs support that.

There is a project, that uses ESP32 as a middleware and takes whatever protocol this JK uses and converts it to Pylontech or whatever you need, but i don't want to experiment with that, to be honest.

The setup works well enough as it is, other than some edge cases. Those being when the pack is fully discharged and under low temperature protection, things on the inverter side gets quite whacky. The voltage on the pack shoots up, inverter switches to offgrid mode, discharges battery some more and switches back to grid. And can't combine solar+grid power in that sort of scenario and ignore the battery.

So my fix was; when BMS is in some sort of charge/discharge protection mode, I switch the inverter to a different mode, that prioritizes pv and grid consumption (and leaves battery alone) and switches back once BMS is ready to be discharged/charged.

Unless JK provides a firmware patch for more protocols, I might consider upgrading to the inverter style BMS to avoid such issues.

T1 Terry said:

if there is any deviation, the 50 amps x 3.65v = 182.5W, is required to burn off the incoming charge current to that cell will the other 3 cells in the battery catch up .......

Click to expand...

Usually the charge current drops very low when (passive) balancing. Usually into the mA range. That's why it will take a very long time. But instant wattage that needs to dissipate is a lot lower. Should be around 1W per cell that is high.

Peter WA said:

Usually the charge current drops very low when (passive) balancing. Usually into the mA range. That's why it will take a very long time. But instant wattage that needs to dissipate is a lot lower. Should be around 1W per cell that is high.

Click to expand...

To do a numeric example: let's say my worst cells are 10% out. That means they need to catch up 15Ah (150Ah cell capacity). At 200mA that would take 75 hours of balancing to get the low cells back up. Long, but feasible.

I'd still prefer it if my pack balances a lot earlier, when the lowest cell is just 1% below the highest or even less. Thanks to the steep voltage gradient that is very easy to see near 100%.

Peter WA said:

Usually the charge current drops very low when (passive) balancing. Usually into the mA range. That's why it will take a very long time. But instant wattage that needs to dissipate is a lot lower. Should be around 1W per cell that is high.

Click to expand...

No arguments there, an intelligent BMS communicates with the charging sources and throttles back the charge rate as the target voltage is reached.
We are talking about the BMS in this drop in battery that I linked to as an example of deception by omission, this BMS is the very basic dumb kind that can communicate with anything, relies on lossy balancing at a very low burn off rate and can not throttle the charge rate back. To stop the BMS allowing the cell to go into over voltage and destroying, if it can't throttle back the charger, it has to burn off the excess.
@hojnikb doesn't seem to understand the concept of deception by omission is often used by Chinese manufacturers in their promotions dept, not actually lying about a batteries capacity, but not saying how that capacity was measured.

A well known Chinese LFP prismatic cell manufacturer sells the product using a 0.2C load to measure capacity. This company was originally Thundersky, but when Winston left and took his patents with him, then they paid big $$ in an attempt to buy the technology, and got caught by the same deception by omission, and with the following to and fro, Winston ended up with the Thundersky name, the original company became Sinopoly and Winston became Winston Thundersky.
The standard accepted load rate for measuring capacity on an LFP cell, is 0.5C and this is what Winston Thundersky use, but in an attempt to compete, Sinopoly use 0.2C as their capacity testing load ..... the result, a Sinopoly 100Ah cell tested at 0.2C, will deliver 100Ah, but tested at 0.5C, it only delivers 80Ah ..... I put the graph up for the Winston 100Ah cell a while back, it shows that even at a 3C discharge rate, it will still just scrape in at the 100Ah capacity.

What I'm attempting to show here, is the claimed 156Ah for the MG4 51kWh battery, has no reference to how the 156Ah was obtained .... it could have been 1 amp for 156hrs till the resting voltage was 2.5V was achieved, that would still meet all the specs provided by MG about their battery.

A true and honest test would be at the average current draw to obtain the claimed range, on flat ground, then another range on mountainous roads, and a middle average load could be decided on ..... then test the cell at that load rate and publish a true Ah capacity.

What I'm saying, is, MG did not use this method to test for capacity, but rather came up with a number required, 156Ah, then designed a test procedure to obtain that result. The true capacity when measured at a realistic load, is some 10% to 20% less than claimed.

The SoH over time is showing that the claimed capacity isn't actually there, and over time, the SoH is dropping down to the true capacity ......

T1 Terry

@T1 Terry

We are talking about the BMS in this drop in battery that I linked to as an example of deception by omission, this BMS is the very basic dumb kind that can communicate with anything, relies on lossy balancing at a very low burn off rate and can not throttle the charge rate back.

Click to expand...

Pretty sure there's no drop in batteries, that can communicate with anything.

As for throttling the charge rate; there are batteries that can do that, even without any communication. These batteries are mostly designed for cars.

doesn't seem to understand the concept of deception by omission is often used by Chinese manufacturers in their promotions dept, not actually lying about a batteries capacity, but not saying how that capacity was measured.

Click to expand...

It's one thing for a no name manufacturer (or brand, as a lot of these batteries are repacks) to lie or omit specs but something entirely different for a big company like SAIC to lie about something so fundamental like raw battery capacity. Just becase these noname-ers lie, doesn't mean every other reputable chinese brand does it too.
Kinda souds like you're racist against Chinese?

What I'm attempting to show here, is the claimed 156Ah for the MG4 51kWh battery, has no reference to how the 156Ah was obtained .... it could have been 1 amp for 156hrs till the resting voltage was 2.5V was achieved, that would still meet all the specs provided by MG about their battery.

Click to expand...

Can you show us any other manufacturer, that does clain at what drain rate their car battery pack was tested at?

What I'm saying, is, MG did not use this method to test for capacity, but rather came up with a number required, 156Ah, then designed a test procedure to obtain that result. The true capacity when measured at a realistic load, is some 10% to 20% less than claimed.

Click to expand...

Actually, we don't know that. Those are just your unsubstantiated claims, that really don't hold much water in real world.

"the true" capacity would be to rip out the pack, drain it down to when BMS cuts off, charge it as it was charged in a car and then drain it back to 0%. The result kWh is your true usable capacity.

Anything else just gives us a really rough estimation and coming up to any conclusions is just plain wrong.

The SoH over time is showing that the claimed capacity isn't actually there, and over time, the SoH is dropping down to the true capacity ......

Click to expand...

The SoH we're getting from the app might not even be true, so again, getting any conclusions according to that is again just wrong.


How about this? Since you're so invested into this problem... Do a slow charge to 100% until the balancing stops, drive the car till dead (or close to it as possible) and then DC charge to 100%.

After you've done that, give us the exact kWh used by the charger, distance traveled and consumption reported by the car.

hojnikb said:

How about this? Since you're so invested into this problem... Do a slow charge to 100% until the balancing stops, drive the car till dead (or close to it as possible) and then DC charge to 100%.

After you've done that, give us the exact kWh used by the charger, distance traveled and consumption reported by the car.

Click to expand...

I did that last saturday. I went to 1.5% SOC and then charged at 6.5kW. The Wallbox said i charged 53.08 kWh. Assuming a 8% losses and the initial 1.5% SOC, a full charged battery is 49.57kWh, a 97.6% from the original 50.8kWh.

The car is almost two years old and 38,000 km.

Corderito said:

I did that last saturday. I went to 1.5% SOC and then charged at 6.5kW. The Wallbox said i charged 53.08 kWh. Assuming a 8% losses and the initial 1.5% SOC, a full charged battery is 49.57kWh, a 97.6% from the original 50.8kWh.

The car is almost two years old and 38,000 kms.

Click to expand...

So kinda difficult to have a ~42kWh usable battery pack, as T1 claims. Unless MG has Renault ZOE sort of charging losses

T1 Terry said:

Back on the 51 SOH, my theory about the true battery capacity ......

We arrived at The Royal Adelaide Hospital with 10% remaining battery and 2,000km on the clock, so we plugged into a Chargefox DC fast charger, this the invoice that was emailed to us
View attachment 35102

Quick math calculation, from 10% to 100% would require 90% of the capacity to be replaced, 37.83kWh. 90% x 1.1 =99.99% 37.83 x 1.11 = 41.99kWh ..... just saying.

T1 Terry

Click to expand...

Percentage you say.......
Sorry i not belive in percentage displayed by BMS after what i seen happens when battery calibration done on pack with huge imbalance.
If you use Car Scanner loock on high and low voltages of cells when you discharging lower then 10%, highly likely when you reach 0% you will see high delta but still high voltage (3.18 - 2,9V) on higest voltage cel and weak cell will be across 2.7-2.5Vl, in your case one or more cells not charged fully and after calibration pack lost 16%
I attached screenshoots of my pack discharged 3 months ago (sorry not MG4 but also LFP pack)
Highest cell have voltage 2.946 volt (not 0% discharge ) and 2 undercharged cells with lowes at 2.4 volt dictates nominal capacity of all pack less then it may hold because afterward BMS counted less ah charged back till first cell reached 3.75 volt and started display 1%= 1.46 ah, now loock at second screenshoot after 3 months at 94.2% BMS displays used 13 ah that way more then factory capasity per 1% (1,53 ah in my case) becaus now my pack balanced better, loock on third screenshoot

Peter WA said:

To do a numeric example: let's say my worst cells are 10% out. That means they need to catch up 15Ah (150Ah cell capacity). At 200mA that would take 75 hours of balancing to get the low cells back up. Long, but feasible.

I'd still prefer it if my pack balances a lot earlier, when the lowest cell is just 1% below the highest or even less. Thanks to the steep voltage gradient that is very easy to see near 100%.

Click to expand...

Yes that take in my case month to balance with passive balancers to reach delta 117-120 mV at moment when charging stops (before i balanced pack i seen also 360 mV), but think delta can be lower
My pack almost 2 years had high delta and i reach point of 93% SOH and recovered 3% by discharging pack to 2.0 v on lowest cell (BYD) and charging back to 100% slowly.
After i realised that balancing happens only when car active i understand that i screwed my battery by not letting car work enough hours bethwen charging sessions to do balancing , in my case at least when delta is high the balancers do their job better when battery discharges then when battery charges, in addition to that i used to frequently charge battery from 98-90% with 11 kW charger dayly stupidly think that after reaching 100% BMS balancing cell when car powered off.
Curently my pack have nominal 146.88 ah out of 153 ah of factory capacity.

From what I've experienced in the past, that fully discharged graph is telling you that you have either a high resistance connection or internal cell resistance, or you have a cell that has lost a lot more capacity than the others.
You can only top balance - at the end of charging, or bottom balance, at the end of discharging, you can't have both.
If the pack in top balanced, don't go down to 0% SOC cell voltage is more important, an absolute min. of 2.8v under load.

If the pack is bottom balanced, charging must stop when the highest cell voltage reaches 3.65v .... you can go as far as 3.85v if you are planning to use the pack immediately, but don't store it for any more than a few hrs at that peak voltage, you will cause the internal voltage in the cell to climb due to the internal heat generated causing the electrolyte to break down.
If a battery pack is bottom balanced, top balancing must be turned off, otherwise the bottom balancing will be affected creating a battery pack that isn't balanced at either end .....

T1 Terry

T1 Terry said:

From what I've experienced in the past, that fully discharged graph is telling you that you have either a high resistance connection or internal cell resistance, or you have a cell that has lost a lot more capacity than the others.
You can only top balance - at the end of charging, or bottom balance, at the end of discharging, you can't have both.
If the pack in top balanced, don't go down to 0% SOC cell voltage is more important, an absolute min. of 2.8v under load.

If the pack is bottom balanced, charging must stop when the highest cell voltage reaches 3.65v .... you can go as far as 3.85v if you are planning to use the pack immediately, but don't store it for any more than a few hrs at that peak voltage, you will cause the internal voltage in the cell to climb due to the internal heat generated causing the electrolyte to break down.
If a battery pack is bottom balanced, top balancing must be turned off, otherwise the bottom balancing will be affected creating a battery pack that isn't balanced at either end .....

T1 Terry

Click to expand...

We can`t disable either bottom (if it exist ) or top balancing in case of BYD at least, i think temperature play not last role when it below +25 degree because if temperature varies across pack when cold that greatly changes internal resistance and in my case almost always battery get temperature difference in same way because parked in same place same direction before charging.
Front of car get bit more air circulation and always hotter by 1-2 degrees on parking lot then back.
And also weak cell is in back of pack, that cell reaches higher voltage during charging because colder cell have higher internal resistance at lower SOC, in addition when balancer see high voltage it starts discharging that cell further making things worse.
I need to park in place where battery will be at same temperature in each cell after parking
Sadly i can`t heat up pack by using car menu (we not have control over thermal management), cold cells also not giving all they potential when we need it, probably also not reaching full capacity when we charging battery when it cold.
But main problem is unclear charging instruction from BYD !!!
They not restrict charging from high SOC and not describes how and when balancing of cells happens, that leads to most of low mielage drivers charging car that way that causes growing balancing errors, like charging from high SOC (95-99%) and charging way faster then is better for balancing and not letting car work enough time bethween charging sessions, adding to this story driving habbits of ferrary drivers that draws huge current from battery and using too heawy recuperation further destroys balancing, this progressing as low SOH of battery pack because during calibration BMS counts less energy then pack may hold.

The downside of continuous dynamic cell balancing I guess. The fix would be section cell heating and switchable cell balancing, not really something the average person could do to their own vehicle. Does BYD have a cell warranty or only complete pack capacity warranty?

T1 Terry