2022 m3 LFP RWD
70,000 kms (43,500 miles) in 18 months. 438 km(272 miles) range when new now at 420km (261 miles)
70,000 kms (43,500 miles) in 18 months. 438 km(272 miles) range when new now at 420km (261 miles)
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Model 3 SR+ LFP Battery Range, Degradation, etc Discussion
- Thread starter Baluchi
- Start date
I think its down to software the numbers are far too consistent on displayed range to be genuine, other side of the planet 16,000 klms 2022 LFP RWD same numbers 438 new, 420 now.......no complaints it does what is says on the label.
This is my hypothesis too, I noticed no difference in displayed range across a number of cars despite various charging habits - frequent supercharging, or charged at home to 100% daily V’s only once a week or less often. Plus no noticeable difference for the km’s travelled, just the age of the car is the only constant. Either LFP is largely immune to charging habits and number of cycles (which well might be the case) or the reduction in range is somewhat ‘software based’. (possibly even pre programmed)
It's probably the first. Everyone is seeing just calendar aging. LFP is very resistant to cyclic degradation---you need many many cycles, like a full cycle every day as is used in battery energy storage.
There is a little bit of cyclic aging but as LFP battery Teslas are pretty new most people are seeing just calendar aging effects. No I don't think there is any pre-programmed software range reduction---makes no sense.
If everyone is seeing the same thing it means the quality control in the CATL LFP cells is superb and there is little pack to pack variability. Even battery research papers show a fairly wide spread of aging rates from supposedly identical cells.
Just that calendar degradation is the main mechanism, as our member AAKEE has taught us over and over. When the cars are 10 years old then cyclic effects will start to be noticeable. calendar aging goes as sqrt(time) but cyclic will go linearly with time.
It could also mean Tesla has been adapting VW's diesel emissions tweak to range estimates and putting it down to the BMS algorithms.
It’s probably several reasons.
Even NMC/NCA has quite low cyclic aging for the average user (mostly 1/10 of the degradation we see. And calendar aging seem to be quite consistent when calculating the estimated capacity with average temp, SOC etc.
So LFP has cyclic aging as a even lower part of total degradation. The annual cyclic will be only a fraction of the thousands of cycles the LFP can do before reaching 20% cyclic aging, so way less than 0.5% per year for the normal user.
As we know the BMS has a hard time measuring SOC exact, it would be hard to measure and estimate the capacity with a high precision.
It is possible that the degradation (present capacity) is calculated partly different compared to NCA/NMC.
Even NMC/NCA has quite low cyclic aging for the average user (mostly 1/10 of the degradation we see. And calendar aging seem to be quite consistent when calculating the estimated capacity with average temp, SOC etc.
So LFP has cyclic aging as a even lower part of total degradation. The annual cyclic will be only a fraction of the thousands of cycles the LFP can do before reaching 20% cyclic aging, so way less than 0.5% per year for the normal user.
As we know the BMS has a hard time measuring SOC exact, it would be hard to measure and estimate the capacity with a high precision.
It is possible that the degradation (present capacity) is calculated partly different compared to NCA/NMC.
That's a good point. Probably most LFP owners follow the charge to 100% frequently enough the top voltage is well calibrated but only discharge into somewhere into the flat SoC vs voltage portion, so the capacity becomes very hard to measure.
So perhaps there really could be an 'assumed' calendar degradation curve applied between times when the true capacity could be adequately estimated. With aging the voltage vs SOC curve might shift a bit as well.
However, the stated rated range at 100% did not seem to change on my car after a road trip that started at 100% and dipped below 15% frequently (and below 10% a few times), where the voltage - SoC curve is non-flat.
You would need to let the car sleep at low SOC. Only having low SOC driving and charging asap will not show the true SOC to the BMS.
At any driving the SOC is only estimated.
And also one single time might not cause a noticable change of the capacity estimate, I had a gross underestimate, showing 75.7 kWh and did run a 100-0% using 75.2 kWh and had a little left (0.3kWh) , which was 79.0kWh in total. I let the car sleep a few hours before the charge but it didnlt change the capacity estimate more than 0.3 kWh or so. It took a few months for the estimate to reach 79.0, but it did.
2023 RWD, 8 months old, ~3500 miles, kept SOC as low as possible and only charge to 100% about once per month. 266 miles at 100% currently. Not as good as I hoped, but one thing I have working against me is living in a fairly warm climate... my car is garaged, but the garage gets up around 90 F on most days in the summer, which probably isn't great for the battery.
Yes! Absolutely.
My example was my M3P with not LFP, but the behaviour is the same:
- Can not measure SOC during drive
-Need to sleep to be able to measure SOC exact.
- One single low SOC will not enable a complete adjustment of the estimated capacity.
Actually, most probably worse for LFP as it is harder to measure the SOC with LFP.
Again identical numbers, has to be software I doubt the batteries are that universally consistant with so many variables.
Again I'm not complaining I get great range but I'd rather Tesla take the mystery out of their battery packs and be open about it.
Krisja
Active Member
So with a fairly low mileage after 16 months it still shows similar degradation to the rest of us.
This is fairly interesting. As others have mentioned, seems age is the major variable.
Will be interesting to post again after 6 months or so.
Similar threads
