Model 3 SR+ LFP Battery Range, Degradation, etc Discussion

[karlos22]

Do you have that option where the brakes are applied to mimic regen ?

No i don’t. Car had behaved this way since day 1.

 

[AAKEE]

Regen is not battery capacity dependent, it is BMS dependent. Agree that 100% is not true 100% but the car can also be set to allow no regen above 90%. See case of Powerwall that will no charge battery until down to 60%. It is all software.

Of course the BMS is involved but the reason for the other chemistries (NCA/NMC) not taking regen at high SOC is that the battery cells never can be (must not be) exposed for a voltage higher than the maxiumum cell voltage, thats 4.20V/cell. If the battery is full it it charged to 4.20V/cell so even the slightest regen during the first km would increase the cell voltage to above 4.20V. Its a hard limit, and the BMS has to follow that.

For LFP's I do not know.

 

[AAKEE]

You can hear his statements starting at 37:50.

It was a general statement, and also included hes own home-battery.

 

[AAKEE]

There must be some sort of top buffer for the LFP as there is still a lot of regen when the battery is fully charged.

The actual amount I do not know, but clearly if it was 100% true SOC there would be zero regen.

Not necessarily.

It could be that the LFP's are not sensistive to higher voltage than the 100% end charging voltage, and also that they have a low internal resistance at 100% allowing to regen with good power without the need to increase the voltage high.

According to this post, with Scan my Tesla data from a fully charged LFP, the total votlage was 381V after the charge. 106 cells in series means 3.594V/cell in average. (the post also states CVMax 3.544 which doesnt seem to match really).
LFP pack and cell voltages at various states of charge


This is a CATL LFP cell close to the RWD cells (about the same size and capacity):
CATL 161Ah LiFePO4 Battery

We can see that if we have a resting voltage after the charge of 3.55-3.59V/cell, we are at or very very close to 100%. We can clearly see that we have reached the spike where the voltage increases rapidly but the charged energy does not.

95% would be in the range of 153Ah capacity, and 3.4V. The voltage would drop slightly after the charge is completed so we would expect around 3.4V or just below if it was at 95%.

For the regen possibillity it does not need to be true 95%. It could be a question about the LFP being capable of, and allowed to go over the max cell voltage of 3.65V, and/or relatively low internal resistance at 100%.

 

[crunchor]

So the best practice is still charging to 100% at least once a week but don't stay at 100%, get back to 70% or lower as soon as possible after hitting 100%?

 

[AAKEE]

So the best practice is still charging to 100% at least once a week but don't stay at 100%, get back to 70% or lower as soon as possible after hitting 100%?

Charging to 100% = resetting the energy counter in the BMS to 100% energy, a known value instead of calculated.

Staying at or below 70% will reduce the calendar aging - 100% is no crisis, opposite to most forum rumors.

Remember that the batteri will hold up anyway, by following the manual.

 

[crunchor]

Charging to 100% = resetting the energy counter in the BMS to 100% energy, a known value instead of calculated.

Staying at or below 70% will reduce the calendar aging - 100% is no crisis, opposite to most forum rumors.

Remember that the batteri will hold up anyway, by following the manual.

staying at 100% for days is no crisis too? Seem like the calendar aging speed up pretty quickly with 100% and 25 degrees to 35 degrees which is a more practical temperature range instead of 15 degrees.

 

[AAKEE]

staying at 100% for days is no crisis too? Seem like the calendar aging speed up pretty quickly with 100% and 25 degrees to 35 degrees which is a more practical temperature range instead of 15 degrees.

Did you see the picture in this post? LFP Calendar aging

That chart and this is not far from each other.
You can se that a doubled rate is not a crisis.
2 days at 100% is 4 at 40-60%.

LFP’s do not increase the internal redistance at 80-100% like the NMC abd NCA does. So from that point of view it is really just like doubled time above ~ 70%.

 

[karlos22]

staying at 100% for days is no crisis too? Seem like the calendar aging speed up pretty quickly with 100% and 25 degrees to 35 degrees which is a more practical temperature range instead of 15 degrees.

The charging to 100% myth for LFP is completely overblown.

Charge to 100% as often you need to, just do not leave it there for extremely long periods of time (weeks, months). Days is ok. Weeks or months not so much. But even so, as AAKEE demonstrated, that means an extra 3% degradation. Does it matter? If you didn't check the range display you would never even know.

Yes the optimal practice is to charge to 100% once a week and then drive the car to below 70% the remainder of the time. If you want to go even further, drive it below 40%! Where does the madness end?

It is very unlikely that you will charge your car to 100% and leave it like that all the time. Unless you have a wall charger in your garage and you can micro manage the battery for fun, I wouldn't worry too much about it!

The LFP is best for newcomers to EV's as it can be treated like a ICE vehicle - charge to full once a week and run it to a low % before charging again. Since LFP mostly degrades from calendar ageing, this is a better way to look after the battery (unlike NCA).

 

[crunchor]

The charging to 100% myth for LFP is completely overblown.

Charge to 100% as often you need to, just do not leave it there for extremely long periods of time (weeks, months). Days is ok. Weeks or months not so much. But even so, as AAKEE demonstrated, that means an extra 3% degradation. Does it matter? If you didn't check the range display you would never even know.

Yes the optimal practice is to charge to 100% once a week and then drive the car to below 70% the remainder of the time. If you want to go even further, drive it below 40%! Where does the madness end?

It is very unlikely that you will charge your car to 100% and leave it like that all the time. Unless you have a wall charger in your garage and you can micro manage the battery for fun, I wouldn't worry too much about it!

The LFP is best for newcomers to EV's as it can be treated like a ICE vehicle - charge to full once a week and run it to a low % before charging again. Since LFP mostly degrades from calendar ageing, this is a better way to look after the battery (unlike NCA).

I saw somewhere saying there is around 1% battery lose per day without constantly plugging in, so sounds good.

 

[Max Spaghetti]

It could be that the LFP's are not sensistive to higher voltage than the 100% end charging voltage, and also that they have a low internal resistance at 100% allowing to regen with good power without the need to increase the voltage high.

LFP's are very sensitive to a higher voltage. You must not charge above 3.65V or risk damaging the cells. Normally you charge at 3.65V until the charging current drops to a nominal cutoff value (not good to float charge at 3.65V) - the cells are then nominally at 100%. You can still charge again (e.g. regen), but only at 3.65V, and the charging current will be low (so not much regen available).

According to this post, with Scan my Tesla data from a fully charged LFP, the total votlage was 381V after the charge. 106 cells in series means 3.594V/cell in average. (the post also states CVMax 3.544 which doesnt seem to match really).

I agree that the figures in that post don't add up, so there is error either in the 381V or the individual cell voltages - I would dismiss the 381V and just look at the individual cell voltages because it is vitally important for a BMS to have accurate individual cell voltages.

We can see that if we have a resting voltage after the charge of 3.55-3.59V/cell, we are at or very very close to 100%.

That is the cell voltages after charging and parking (according to the post) but it won't be the resting voltage - the resting voltage of LFP at 100% would normally be close to 3.4V (or even high 3.3xV), so I don't think these cells have rested for very long. (There is no accurate resting voltage for 100% charged LFP cell - that is why we need to charge to find 100%, but it should be near 3.4V).

You can actually charge an LFP cell at 3.45V and it will still go to 100%. It just takes longer than charging at 3.65V. But maximum float charge after that (if you want to float - not recommended long term) is 3.4V.

The above is from my knowledge of LFP, but not Teslas specifically.

 

[Max Spaghetti]

I agree that the figures in that post don't add up, so there is error either in the 381V or the individual cell voltages - I would dismiss the 381V and just look at the individual cell voltages because it is vitally important for a BMS to have accurate individual cell voltages.

I have just come to the conclusion that LFP batteries have 108 cell bricks not 106.

I made a post in the Australia thread which I won’t cross post here, but my Battery voltage is 343V and divided by 108 that equates to my Cell Volt Average.

Also my Cell Volt Min brick ID coincidentally happens to be 108! …

 

[AAKEE]

Good info! Thanks!

I have a lot of lithium cells/packs but so far, not a single LFP

That is the cell voltages after charging and parking (according to the post) but it won't be the resting voltage - the resting voltage of LFP at 100% would normally be close to 3.4V (or even high 3.3xV), so I don't think these cells have rested for very long. (There is no accurate resting voltage for 100% charged LFP cell - that is why we need to charge to find 100%, but it should be near 3.4V).

Then we most probably have the solution to wy regen works at 100%:

If the cell voltage drops to 3.4V or so after a full charge without using energy, there would be a head room to feed the cells again with 3.65V/cell.

This will allow for regen at 100% SOC.

The above is from my knowledge of LFP, but not Teslas specifically.

I guess you agree that according that post and voltages there is no upper buffer and the cells are actually charged to 100%?

 

[AAKEE]

I have just come to the conclusion that LFP batteries have 108 cell bricks not 106.

I made a post in the Australia thread which I won’t cross post here, but my Battery voltage is 343V and divided by 108 that equates to my Cell Volt Average.

Also my Cell Volt Min brick ID coincidentally happens to be 108! …

What car, battery and make?

The 55kWh CATL LFP had 106S (quite sure, I think Sandy Munroe & co had a pack apart and said 106…)

108 would make sence…60kWh pack?

 

[Max Spaghetti]

What car, battery and make?

The 55kWh CATL LFP had 106S (quite sure, I think Sandy Munroe & co had a pack apart and said 106…)

108 would make sence…60kWh pack?

April 2023 Model 3 RWD Made in China 60.5kWh LFP pack.

 

[AAKEE]

This is a vid’ of a guy checking hi’s new Y (60.5 kWh CATL LFP)
Y LFP

Fylly charged with a bit of imbalance (havent gotten the chance to balance properly yet?)

The max pack voltage is 410V according to the BMS, this implies that the pack is allowed to be pushed to 410V (3.8 V/cell).
Max pack voltage has been correct on all cars i tested. I’m sure it does not mean to charge until the pack has 410V OCV…

Also, despite being at 100% (Scan my Tesla reports 101%, very common when the nominal remaining is > nominal full pack) the max regen is at 71kW.
The max regen and the actual maximum regen you can get have been agreeing when I checked them.

Highest cell at 3.49V

This picture confirms 108S BTW.
CAC x nominal voltage x cell count = nominal full pack
177 x 3.2 x 108 = 61.17 kWh

https://youtu.be/mxiOe-RMS2M?si=MeN5FLUeaoThmFUn

 

[Max Spaghetti]

I guess you agree that according that post and voltages there is no upper buffer and the cells are actually charged to 100%?

I agree. Same as NMC/NCA. I don't know what the guy in the video is talking about when he says there is a 5% top buffer, it makes no sense.

The max pack voltage is 410V according to the BMS, this implies that the pack is allowed to be pushed to 410V (3.8 V/cell).
Max pack voltage has been correct on all cars i tested. I’m sure it does not mean to charge until the pack has 410V OCV…

This looks correct - it appears that Tesla are happy to push the cells to 3.8V.

I don't know if they will normally charge the car up to that voltage - I look forward to watching SMT both with my 3.3kW UMC and also at a DCFC to see what the voltages and current are doing approaching 100% (and also the cell balancing).

But remember that charging at 3.8V won't add extra capacity. It just adds a small amount of extra surface charge which will dissipate into the battery at rest, and the resting cell voltage will still return to 3.4V or less. We would be talking about maybe 0.2% extra capacity, and causing additional degradation. So if Tesla do let it ramp up to 3.8V at a DCFC, I think it would be to help decrease charging time, not to provide additional capacity. I also don't know whether 3.8V is more acceptable during high temperature charging (preconditioned) - I haven't seen any data on that, but it could be plausible.

It could alternatively be plausible that they allow 3.8V during short term regen events - they may have determined that this doesn't have significant ill-effect. I would have to do some serious testing to work this out!

If you took it to the extreme and charged to 100% at the top of a mountain and then pointed the car down hill to regen all the way down, I think it would reach a point pretty quickly where it would prevent any further charge. It would be interesting to try this while watching SMT - need to find a charger on a mountain first! (And check the brake fluid).

I also don't know if it would try to dump energy via the heat pump and stator windings to still provide some "simulated regen" to maintain the single-pedal-driving effect. In theory it could, but I don't know if it's designed to do that. When pre-conditioning for supercharging, my RWD will use up to 8kW to heat the battery - some of that is the heat pump but I think most of it is via the rear stator winding. That could be 8kW of dump capacity for "simulated regen".

 

[AAKEE]

when he says there is a 5% top buffer, it makes no sense.

The usual misunderstanding.
For EV’s, when someone states something the odds seems better for it to be wrong

It could alternatively be plausible that they allow 3.8V during short term regen events - they may have determined that this doesn't have significant ill-effect. I would have to do some serious testing to work this out!

Yes, thats my take / guess.

For other types (NMC/NCA and lipos) they can take shorter burst of higher power co pared to the constant power.
Same seems acceptable for charging- regen uses much higher power than other charging high up in SOC.

If you took it to the extreme and charged to 100% at the top of a mountain and then pointed the car down hill to regen all the way down, I think it would reach a point pretty quickly where it would prevent any further charge. It would be interesting to try this while watching SMT - need to find a charger on a mountain first! (And check the brake fluid).

Yes, If we charge to 100% and accererate, there would be space in the pack energy wise to regen that energy back - but for the NCA/NMC they are so close to the maximum voltage that virtually no regen power can be pushed except exceeding the maximum voltage, even if there is space for the e ergy itself.

Of course, having 100% and going downhill can not give regeb that actually charges the battery so it either has to stop producing regen or crate a lot of heat.

I also don't know if it would try to dump energy via the heat pump and stator windings to still provide some "simulated regen" to maintain the single-pedal-driving effect. In theory it could, but I don't know if it's designed to do that.

I have seen the regen produce a little power when the battery was too cold (MSP) and the regen lower seemed to match the power needed to heat the cabin etc.

When pre-conditioning for supercharging, my RWD will use up to 8kW to heat the battery - some of that is the heat pump but I think most of it is via the rear stator winding. That could be 8kW of dump capacity for "simulated regen".

The stators have been using 3kW each at least before.
6kW battery power to the motors on my former M3P and 6kW to the heat pump (in lossy mode due to too cold and a cold battery), 12 kW battery power in total.

My MSP (Plaid) has been delivering 15-18 kW in the same condition.
Unfortunately the SMT did not have funtionality to track 3 motors, so I did not get any real data on it.
But I would guess 3x3 kW for the engines and 6kW for the heat pump makes 15 kW, so matches the 15kW seen.

 

[tgrobbo]

LFP cuts regen a LOT just fyi. Head down any half decent hill and the little green brake disk comes on, especially in cold weather. I assume it does start to blend in the brakes, but you can feel the car suddenly speed up when it reduces regen. It’s actually pretty annoying.

 

[AlanSubie4Life]

assume it does start to blend in the brakes

Have to turn that option on.