MG4 51kWh LFP Battery Health

[Everest]

If anyone sees any other errors, please point them out

My 2p worth...

The 51kWh is actually a rounded up figure. The 400v can only be achieved by charging the 104 cells to 3.85V .....

Any cell voltage above 3.4V is surface charge, put it under load and that 0.45VDC over charge will vanish

3.85V is outside the safe working voltage range of LiFePO4 cells. Maximum charge voltage is 3.65V, not 3.85V. Above 3.65V, damage will occur due in part to some cell swelling and internal de-lamination that may take place.

So it looks like there really are 104 cells

That seems likely. I saw an image of the battery pack from a breaker that showed raised areas on the outer casing where, presumably, the cells were located. There were 4 rows of 9 areas on the outer edges and 2 rows of 8 areas in the middle. That comes to (4 x 9) + (2 x 8) = 52 positions where cells might be. Assuming 2 cells are situated in each raised area, that equals 104 cells.

If that is the case, then, as we know the nominal voltage of an LFP cells is defined at 3.2V, the the 'nominal' voltage of the MG4 SE SR will be 104 x 3.2V = 332.8V.

Assuming the quoted 51kWh capacity of the SE SR is correct and an 'actual' capacity, rather than 'usable' capacity, then we can deduce that each cell has a capacity of 51000Wh/3.2V/104 = just over 153Ah each.

I read (on here, I think) that some newer builds of the SE SR quote a capacity of 49kWh. I wonder if that is simply a change in metric from actual to useable capacity? i.e. assuming 4% actual remaining when 'usable' capacity has reached zero.

The whole calculation thing actually requires a standard for cell load testing and a definition for nominal voltage.

but IME 3.2V is considered the industry standard nominal voltage for LiFePO4.

FWIW, if anyone is interested, here is a graphic showing LiFePO4 voltage for different SOC and different charge/discharge rates.

Edited to add... not sure what happened to the graphic

 

[Coulomb]

3.85V is outside the safe working voltage range of LiFePO4 cells.

That was exactly Terry's point: the "400V" figure can't be accurate. Indeed, "400V" is just to distinguish from the very roughly double voltage that some high end models use, which is called "800V", even though it could actually be say 720V nominal, and perhaps 690V actual at a very low State of Charge.

I read (on here, I think) that some newer builds of the SE SR quote a capacity of 49kWh. I wonder if that is simply a change in metric from actual to useable capacity? i.e. assuming 4% actual remaining when 'usable' capacity has reached zero.

I don't believe that's the case, i.e. I think that those models actually have a smaller battery capacity designed to have the same nominal range as the 51kWh models. I think I read that this is to be achieved with a more advanced battery design.

 

[Everest]

I don't believe that's the case, i.e. I think that those models actually have a smaller battery capacity designed to have the same nominal range as the 51kWh models. I think I read that this is to be achieved with a more advanced battery design.

Sure, I was just musing

but ....to clarify... if they have a smaller battery capacity, do you mean they have a smaller actual capacity, but the same usable capacity? i.e. the better battery is utilised to a greater depth-of-discharge.

Otherwise, if they have an equally lower usable capacity, then achieving the same range could surely only be achieved with lower consumption by the car itself? Or am I missing something?

 

[wattmatters]

Its good to charge the LFP battery to 100% at it recalibrates itself (I believe), at least thats whats I've seen on my home 20kWh battery system - at home it gets a little lost if I dont charge it to 100% every few weeks.

While top balancing occurs and a BMS may reset to 100% SOC once voltage and current meet certain charging parameters, this does not calibrate the BMS for the purposes of state of health (SOH) measurement.

That requires a deep discharge and full recharge so the coulomb counter can measure the actual capacity of the pack. Unless a deep cycle has been performed recently, I would not place much faith in a SOH number.

Deep cycling about every three months is a manufacturer recommendation, as per page 293 of the MG4 user manual.

 

[Tig170]

Attached is mine from car scanner, At my recent service (via MG) I was told that my SoH was 100%

 

[Semerkand]

It looked pretty good to me for a car with 30,000 miles on it.

Any cell voltage above 3.4V is surface charge, put it under load and that 0.45VDC over charge will vanish, the first bit drawn off the cells will drop the voltage further to around 3.35V x 104 x 125Ah = 43.5kWh, and that capacity you can rely on for quite some time.

What I understand from what you said (please correct me if I'm talking nonsense) it is healthier not to charge a cell above 3.4V in terms of battery life. 3.4 x 125 x 104 = 44.2 kWh. Taking this value as reference, charging a lithium-ion battery with LFP cathode to 80% instead of fully charging it (excluding monthly balancing) will not charge a cell above 3.4V

 

[T1 Terry]

It looked pretty good to me for a car with 30,000 miles on it.


What I understand from what you said (please correct me if I'm talking nonsense) it is healthier not to charge a cell above 3.4V in terms of battery life. 3.4 x 125 x 104 = 44.2 kWh. Taking this value as reference, charging a lithium-ion battery with LFP cathode to 80% instead of fully charging it (excluding monthly balancing) will not charge a cell above 3.4V

It comes down to a charging speed issue really. You can't charge an LFP or LYP cell to 3.4V using a 3.4V charge limit, in easy to understand speak, you have to force the last bit in. These cells can be charged to 4V but you have to watch them like a hawk if/when you do this, once the cell is full, it's like a glass of beer, you can get a head to form above the beer, you can even get a slight dome of beer above the glass, but that last bit is right at the edge of breaking the surface tension, at the point, you can't put any more in without a bad result.

There is difference between charging voltage, surface charge voltage and fully charged voltage. This is all related to the internal resistance of that particular cell. If it has a very low internal resistance, the charging voltage will be within 0.05V at low current of the actual cell voltage when rested, not charging and zero load. That means, a fully charge cell resting at 3.4V will require a min of 3.45V for any current to flow into that cell ...... if anything more than a few mA trickle into that cell, it wasn't fully charged and the 3.4V was a surface charge ........

Normal charging, I always set the upper limit to 3.6V ± 0.05V, that allows for a 0.2 to 0.25V surface charge, but to qualify that, I built my own BMS system that allowed charging to resume once the high cell voltage dropped below the 3.6V mark for an adjustable length of time. This was to allow the dynamic balancing to move any excess capacity to the "not yet full cells" then charging would resume until another cell over voltage warning was detected, and the process would repeat a set time the voltage was above 3.5V x the number of cells, then the charging system would drop back to the 3.45V per cell average and maintain that while the sun was out because these systems were primarily solar charged.

If the charging current is high in regards to the C rate, that point of upper safe cell voltage can be raised all the way to 4V if you want, but internal thermal build up will result and that is the thing that causes internal cell damage, the heat, not the voltage.

I best stop now, I see a lot of eyes rolling back in heads It is a very complex subject, so many things come into play, but unless you plan to build LFP or LYP battery packs that you want to last in excess of 10 yrs .... you really don't need to know the finer details .... if you are building these battery packs .... "do your own research" as the Google knowledge interweb crazies would say .... but do it hands on, the truth is out there, but there is a lot of fertilizer as well ......

T1 Terry

 

[T1 Terry]

That was exactly Terry's point: the "400V" figure can't be accurate. Indeed, "400V" is just to distinguish from the very roughly double voltage that some high end models use, which is called "800V", even though it could actually be say 720V nominal, and perhaps 690V actual at a very low State of Charge.

I don't believe that's the case, i.e. I think that those models actually have a smaller battery capacity designed to have the same nominal range as the 51kWh models. I think I read that this is to be achieved with a more advanced battery design.

Possible, but I doubt it, I believe they would use all the capacity they could fit in the same battery box, then increase the claimed full battery range ..... isn't that one of those things that also only seem to exist in theory, does anyone actually drive a number of these vehicles a number of times and report the average?

To me, it would appear they are down grading the capacity to cover their butt, always better to claim less and deliver more, than the other way around ........ If they program in the lower figure, then battery degradation reduces or stops ..... and you can claim you have achieved something special, all the while using smoke and mirrors to hide the truth .... they were never 51kWh batteries.

T1 Terry

 

[T1 Terry]

Attached is mine from car scanner, At my recent service (via MG) I was told that my SoH was 100%



View attachment 32966

94% of 51kWh is around 48kWh isn't it ....... allowing a bit for a safety margin, not all mass produced cells are equal, that sounds closer to the mark, if you use the fully charged with some surface charge voltage of 3.6V per cell x 104 cells = 375V. 48kWh / 375VDC = 128Ah per cell, fairly close to what I posted early on .....

T1 Terry

 

[Coulomb]

Otherwise, if they have an equally lower usable capacity, then achieving the same range could surely only be achieved with lower consumption by the car itself? Or am I missing something?

My impression is that they were using batteries that are lighter per kWh, so that with the lighter 49kWh battery they could achieve the same range in the same body as a current 51kWh battery in the same body.

These might have been the so-called solid state batteries, or possibly "semi solid state".

 

[tsedge]

I think this thread is a good examole of why you can't trust manufacturer's SOH numbers. We have no idea how they are being calculated and there are many variables, they are not comparable across different models and chemistries let alone across makes. We can be sure they are not a true measure of battery capacity.

 

[T1 Terry]

I think this thread is a good examole of why you can't trust manufacturer's SOH numbers. We have no idea how they are being calculated and there are many variables, they are not comparable across different models and chemistries let alone across makes. We can be sure they are not a true measure of battery capacity.

The SOH at least lets you watch for the "degradation" to level out, at that stage you get a better picture of the true battery capacity. How many manufacturers told porkies about the hp of the old gas guzzlers, or MPG figures that required a tail wind down a greased mine shaft to achieve.

The "51" is just a number, call it a model number and you won't feel cheated, it's the 4 cyl with the small fuel tank model, we refer to them as "poverty packs" over this side of the globe. Doesn't have all the bells and whistles, doesn't go quite as far before it needs a refill ..... but they do go well all the same. Time will tell if the SOH levels out, then we might have a better idea of what the true capacity is, or will they continue to degrade showing either a poor choice of cell manufacturer or a poor BMS.
The fact the LFP cells can be safely charged to 100% regularly without damage .... how far they can be discharged and retain cycle life is yet to be seen, this chemistry doesn't like to be rapid charged if the cell is below 3V rested, but that also amounts to 100% DoD (Depth of Discharge).
If a cell doesn't recover to 3V or better rested, it has been discharged below its rated capacity ..... 0V does not mean fully discharged, it means murdered, 3V rested means fully discharged. The cell might drop to 2.5V under high load, yet return to 3V rested, if a cell drops below 2.5V under load, stop torturing it and limp to a charging station.

No idea if there is an app that shows the voltage of each cell, I've been up to my eyeballs in other stuff and haven't had a chance to play with my MG4 51 yet, maybe the New Yr will see a light at the end of the tunnel ... that isn't a train approaching

T1 Terry

 

[Tig170]

94% of 51kWh is around 48kWh isn't it ....... allowing a bit for a safety margin, not all mass produced cells are equal, that sounds closer to the mark, if you use the fully charged with some surface charge voltage of 3.6V per cell x 104 cells = 375V. 48kWh / 375VDC = 128Ah per cell, fairly close to what I posted early on .....

T1 Terry

... : that ties in with my Energy Consumption Since Last charge : i.e. if I charge to 100%, and then take a reading at 50% charge, I normally get 24kWh.

 

[wattmatters]

No idea if there is an app that shows the voltage of each cell

An OBD reader will show the minimum and maximum cell voltages (and delta).

 

[Tig170]

Has anyone seen this?

 

[T1 Terry]

An OBD reader will show the minimum and maximum cell voltages (and delta).

Is this for each cell, or just the lowest and highest cell and identify which cell it is? Will it record a graph of that for later analysis? The Junsi cell log 8 were great for doing this, but no longer available for the last 3 yrs or more .... I'll have to do some research to find a cell monitor with output alarm that can be programmed in a similar way so I can actually record what is happening while I'm driving.

T1 Terry

 

[wattmatters]

Is this for each cell, or just the lowest and highest cell and identify which cell it is?

Low and high cell voltage for the pack. It doesn't say which.

 

[Everest]

What I understand from what you said (please correct me if I'm talking nonsense) it is healthier not to charge a cell above 3.4V in terms of battery life. 3.4 x 125 x 104 = 44.2 kWh. Taking this value as reference, charging a lithium-ion battery with LFP cathode to 80% instead of fully charging it (excluding monthly balancing) will not charge a cell above 3.4V

I don't concur with that - for a number of reasons...

With regard to the capacity calculation... the kWh capacity of a battery pack is measured at the 'nominal' voltage of the cell, which for LFP is 3.2V. An LFP cell has a very non linear voltage to SoC curve (which I posted above, but didn't work - will repost a simpler one below). The curve is very flat between approx. 20% SOC at around 3.2V (no load) voltage and 3.4V (over 99% SOC). Hence you can't just multiply any specific voltage by the capacity (Ah) of the cells to get a capacity as the graph is not linear.

With regard to not charging cells above 3.4V... in terms of battery life and performance of the battery pack, it is important that individual cells are not charged above 3.65V. For that reason, the BMS will monitor individual cell voltages and typically stop charging once any individual cell reaches 3.65V or whatever voltage has been set by the manufacturer. For example MG may set that value to 3.6V or 3.55V maximum.

But due to small differences between cells (both in terms of capacity and internal resistance) not all cells will reach 100% at the same time without help. That 'help' is the balancing that is one of the key functions of the battery's BMS.

However, those small differences between the cells stored charge amount is only detectable when there is a significant voltage difference between the cells. Hence why balancing can only be performed when individual cells exceed 3.4V - i.e. they are in the upper knee of the SOC/V curve (see image below). This will correspond to a SOC of around 99%. And is why LFP cars have to be regularly slow-charged to 100% periodically to allow balancing to take place.

 

[mechfurkan]

what kind of obd dongle do you use?

vgate icar pro?


In my opinion, it's OK to lose high percentage of battery at the beginning of the use for LFP chemistry. Probably degradation will stop in 2 years unlike the NMC.

And also, SOH may depend to your last charge performance, decrease the battery at the minimum point and charge it to %100, your SOH could be changed.

 

[MeGatron4]

Thanks for the reply. I usually try to keep my car at 15-85%. However, I charged it to 100% 3-4 times in 2 months. I heard that LFP is not affected much if charged to 100%, but it is still a precaution. Idk.
So how do you charge yours? And is slow charging better for battery life? any idea?

Actually after watching this video:
The full charge isn't for battery health, it's to allow the software to accurately calculate your cars current state of charge as there's area of the charge whereby the voltage doesn't change much that the computer has to kind of guess your state of charge.
For battery health best practises skip to 10 minutes into the video, keeping the charge between 0% and 25% maintains the best battery health. The take away is the lower charge states you keep the car in, the longer the battery health stays healthy for LFP batteries