Range Loss Over Time, What Can Be Expected, Efficiency, How to Maintain Battery Health

[AAKEE]

Have you left the car sleeping to invoke OCV when at 100%? Or you drove it directly to under 55%.
And when at 0% have you left the car sleeping some hour?

For my 100% SOC charges, most of them included a immediate drive. Some didnt really even stop charging self before I drove. Once my wife needed the car at 100% and I was not nearby so I did set it to charge to 100% and she left ~ 1 hour after the charged was completed.
( I do not think there is a big need for the BMS to see the 100% SOC at OVC as it is a well defined end point of charging, reaching 4.20V/ cell and have a predefined currebt as the stop charge level.)

At 0%:
I did one in the end of last summer, down to 0.39% and the car got a decent sleep with that SOC.
I have made two low SOC this year, down to cirka -2%. Both times the car got to sleep a few hours, and both times the NFP shifted down precisely as I hoped(as I was quite sure that the NFP was overestimated).
After the first time the NFP climbed back to the before value, 80.5kWh.
The second time was quite recently, and after that the car have been doing more big cycles than the usual 55% charge level, due to a new job and some trips from my wife.
The NFP is steady at around 79 now.

I did a drive down to zero (0.40%) and charged (SuC) up to 100% charge during a trip home, to get a number of the charge session Immediate charge at arrival, and did not sit until it stopped.
(Before: Nominal remaining 3.80
After: Nominal remaining 79.0 kWh
Charged 75.4 kWh).
This did strengthen my believe about my battery capacity is 79 kWh today which is in pair with my calculations from average SOC, average cell temp and average cycles etc.

 

[AAKEE]

The publications do show I imagine the worst case scenarios for battery degradation

I’m not sure which publications you refer to.
Some researchers use 50-60C in the tests and then try to relate to normal temperatures. But in many cases 50-60C kill the battery in a completely other way than normal temperstures do.

The degradation charts from most tests do not show the worst case scenario but the most probable scenario. The spread between individual cells in the same teat and cells in compareble tests are quite small. We are talking for example about a mean capacity loss of 5% ( 95% capacity left, with a spread between 94 and 96%).
Not really much more than that.

Cyclic aging show a higher spread in different tests, probably depending on how the cycles are exactly performed and the measuring methods for capacity.

I did a calculation when I got my M3P, I assumed a average SOC and average cell temp. Using the collected data from some good research reports I was able to make a estimate for my batterys degradation curve.
I did drive slightly more the first year but as we know that cyclic aging is small and my cycles (small + low SOC range) it actualky dont really matter. Time is the battery capacity theft.

Well, today my battery curve tell me that I should have lost about 3.5%. Its around 3% calendar aging or so, and around 0,5-1% cyclic aging. Numbers rounded from memory, add up to about 3.5%.
With this I should have 79.2kWh capacity which I think my efforts to find the capacity also shows.

These 79.2 or 79kWh “found in the car” is not worst case scenario but the expected capacity after 1.5 years with my average 34% SOC and average cell temp around 10C. It it not worst case. Instead it is not really plausible to have less calendar aging during these 1.5 years than other similar cells show in the research reports.
It is also not the “best case” scenario either.
If my BMS had said “NFP = 73kWh” I would have more or less called that bullshit as I know I had around 82kWh initially.

Some day, when I have the time I will make an internet site where you can calculate the expected capacity.

but, as has been mentioned by @AAKEE, by skipping over testing at lower temperatures, so far we have few hard data about the best conditions for longevity in the 0-20°C range.

There is quite plausible data from research about lower temperatures.
Rules of thumbs from my memory:
-Stored at -20C*, there was no or extremely little calendar aging.
-10-15C reduce the calendar aging with about 1/3 of the 25C values.

The hardest thing is not to find the rate of degradation for a SOC and temperature.
The hardest thing is to estimate you own cars battery average. What Ive found from my logged data is that the battery is hotter from driving than one could guess when no heat is drawn from the battery by the heatpump.
Also, having the car out in the sun probably heat the battery quite much. My samsung tablet doing the logging overheats and shuts down up here at the artic circle and we are still long away from the hot summer days. Actually, the snow just went away.
My guess is that some of the peoples cars experiencing high degradation have the main source of the high degradation coming from this.

 

[conv90]

For my 100% SOC charges, most of them included a immediate drive. Some didnt really even stop charging self before I drove. Once my wife needed the car at 100% and I was not nearby so I did set it to charge to 100% and she left ~ 1 hour after the charged was completed.
( I do not think there is a big need for the BMS to see the 100% SOC at OVC as it is a well defined end point of charging, reaching 4.20V/ cell and have a predefined currebt as the stop charge level.)

At 0%:
I did one in the end of last summer, down to 0.39% and the car got a decent sleep with that SOC.
I have made two low SOC this year, down to cirka -2%. Both times the car got to sleep a few hours, and both times the NFP shifted down precisely as I hoped(as I was quite sure that the NFP was overestimated).
After the first time the NFP climbed back to the before value, 80.5kWh.
The second time was quite recently, and after that the car have been doing more big cycles than the usual 55% charge level, due to a new job and some trips from my wife.
The NFP is steady at around 79 now.

I did a drive down to zero (0.40%) and charged (SuC) up to 100% charge during a trip home, to get a number of the charge session Immediate charge at arrival, and did not sit until it stopped.
(Before: Nominal remaining 3.80
After: Nominal remaining 79.0 kWh
Charged 75.4 kWh).
This did strengthen my believe about my battery capacity is 79 kWh today which is in pair with my calculations from average SOC, average cell temp and average cycles etc.

@AAKEE AAKE I really thank you. You are always so accurate.
I have to plan a similar drive to 0% let it sit some hours and charge to 100% to check the real NFP and Nominal Remaining of my car.
Today i reached 79,5 kWh..... It was 73,9 kWh from mid february to mid march.
I think that 73,9 was unrealistic and that 79,5 is unrealistic too.
My march 2021 M3P always followed your same routine since august/september 2021 (unless big charges only during long trips) and it always slept under 55%. Obviously there is a big difference in ambient temp because I'm in North Italy where in winter it goes under 0°C but it's a full month now that we are experiencing 28-30up to 35° C .
With your calculation I suppose I'm in the 76-77 range kWh considering a 4 or 5% degradation and considering that my car never had more than 79,6 at the beginning. (probably even the 79,6 figure of the beginning was unrealistic too)
Rounding to 80 kWh, 4% loss is 76,8kWh Nominal Full Pack.

 

[AAKEE]

@AAKEE AAKE I really thank you. You are always so accurate.
I have to plan a similar drive to 0% let it sit some hours and charge to 100% to check the real NFP and Nominal Remaining of my car.
Today i reached 79,5 kWh..... It was 73,9 kWh from mid february to mid march.
I think that 73,9 was unrealistic and that 79,5 is unrealistic too.
My march 2021 M3P always followed your same routine since august/september 2021 (unless big charges only during long trips) and it always slept under 55%. Obviously there is a big difference in ambient temp because I'm in North Italy where in winter it goes under 0°C but it's a full month now that we are experiencing 28-30up to 35° C .
With your calculation I suppose I'm in the 76-77 range kWh considering a 4 or 5% degradation and considering that my car never had more than 79,6 at the beginning. (probably even the 79,6 figure of the beginning was unrealistic too)
Rounding to 80 kWh, 4% loss is 76,8kWh Nominal Full Pack.

Yes, low SOC cycles seem to offset the NFP upwards, an in my case the offset is ”hidden below 0%.
My first drive to zero, well 0.39% i think, did not change the NFP. The two with around -2% did, however.

If you drive down to zero (or below) and let it sleep and then charge full, you might affect the NFP downwards. For me I had guite a few single digit SOC drives that did not lower the NFP.
I am happy to have the NFP where I thought it would be anyway. It doesnt matter if it goes up from the charging schedule or down for some reason.

I do not know why the NFP seems lower in general in MIC M3P.
I would guess that one part is the BMS underedtimating the ”new chemistry of 2170L”.
If you keep track of the cell temps so it is posdible to estimate the annual average cell temp( or in blocks for the seasons) its not that hard to estimate the degradation I think.

The degradation charts show about 3% for 9.6 months between 45-55% SOC and 25C. 3 x square root (12/9.6)= 3.35% first year. 14 months (right) = 3.6%
(My average SOC was 34% according to teslalogger last time I checked.)

3.6%, add about 0.5% for the cycles of the year = about 4-4.5%

Not easy to know where to begin( real capacity…) most M3P from us showed Around 80-80.5 but the EPA test had 82.1kWh output.
77 to 78 is realistic if the initial capacity was about 80 or a bit above…?

 

[Steve446]

The BMS guess the capacity, and it is no better than the software values put into the computer. This should not be mixed with the real capacity loss and the real degradation.
Lithium batteries in our cars will only change the real capacity in one way:
- Downwards!

The BMS calculation is affected by a lot of things, causing the range to go up and down.
You also can see that the NFP is more solid than the estimated range. The estimated range (estimated by teslafi or SMT?) seems to vary a lot more than the NFP.
This is because the app or teslafi etc makes an own calculation of the 100% range. This is not a true Tesla value and it vary despite tha fact that the NFP is not varying, so estimated range by an app or teslafi should be taken with a grain of salt.
Disregard short term variations from those range calcs. Use only NFP if you have that value. NFP is not exact but vary much less. Still, this is not true capacity as the capacity only really goes down, never up.

Down below: a period when my NFP was rock solid, no more than 0.2 kWh difference(mostly none). The 80.4-80.6kWh should mean a true range of 506-507km.
Teslafi shows 497-505 km mostly and a lot of variations despite the same NFP.
View attachment 810839

For sure the only way is down! That's what my NFP is doing and I'm not arguing that point!. But, indeed, I believe as you say, that the degradation can be a little bit less by following the guidance you push. At one point my NFP fell precipitously but resurrected upwards following a recalibration charge (7-95%). That fall seemed to correlate with a sustained period at low SoC ~20% and below 20°C average, but, I have no proof that that was the reason. Indeed, apart from this one event, I have normally seen about the same level of variance in the SMT NFP as your observation. I had wondered how to relate my NFP data to the discussion you had with Dr Chaos (but I'm no mathematician).

Regards my FRR, I used TeslaFi for 3 months before I installed SMT and the SMT values thereafter. By eye the SMT FRR is less noisy than the TeslaFi values.

 

[Steve446]

I’m not sure which publications you refer to.
Some researchers use 50-60C in the tests and then try to relate to normal temperatures. But in many cases 50-60C kill the battery in a completely other way than normal temperstures do.

The degradation charts from most tests do not show the worst case scenario but the most probable scenario. The spread between individual cells in the same teat and cells in compareble tests are quite small. We are talking for example about a mean capacity loss of 5% ( 95% capacity left, with a spread between 94 and 96%).
Not really much more than that.

Cyclic aging show a higher spread in different tests, probably depending on how the cycles are exactly performed and the measuring methods for capacity.

I did a calculation when I got my M3P, I assumed a average SOC and average cell temp. Using the collected data from some good research reports I was able to make a estimate for my batterys degradation curve.
I did drive slightly more the first year but as we know that cyclic aging is small and my cycles (small + low SOC range) it actualky dont really matter. Time is the battery capacity theft.

Well, today my battery curve tell me that I should have lost about 3.5%. Its around 3% calendar aging or so, and around 0,5-1% cyclic aging. Numbers rounded from memory, add up to about 3.5%.
With this I should have 79.2kWh capacity which I think my efforts to find the capacity also shows.

These 79.2 or 79kWh “found in the car” is not worst case scenario but the expected capacity after 1.5 years with my average 34% SOC and average cell temp around 10C. It it not worst case. Instead it is not really plausible to have less calendar aging during these 1.5 years than other similar cells show in the research reports.
It is also not the “best case” scenario either.
If my BMS had said “NFP = 73kWh” I would have more or less called that bullshit as I know I had around 82kWh initially.

Some day, when I have the time I will make an internet site where you can calculate the expected capacity.


There is quite plausible data from research about lower temperatures.
Rules of thumbs from my memory:
-Stored at -20C*, there was no or extremely little calendar aging.
-10-15C reduce the calendar aging with about 1/3 of the 25C values.

The hardest thing is not to find the rate of degradation for a SOC and temperature.
The hardest thing is to estimate you own cars battery average. What Ive found from my logged data is that the battery is hotter from driving than one could guess when no heat is drawn from the battery by the heatpump.
Also, having the car out in the sun probably heat the battery quite much. My samsung tablet doing the logging overheats and shuts down up here at the artic circle and we are still long away from the hot summer days. Actually, the snow just went away.
My guess is that some of the peoples cars experiencing high degradation have the main source of the high degradation coming from this.

The worst case I referred to was the data in the papers you showed on this forum where degradation was highest (SoC >55-70% and temperatures above ~35-40°C). That's my labelling of "worst case". I now understand that it might not be the same for everybody. In your spare time a site for calculating remaining capacity would be great BTW do you distinguish between (real) degradation as a process and capacity? isn't (real) degradation what implies a capacity after a certain time? Sorry if that seems like semantics, that's not intended.

 

[conv90]

For sure the only way is down! That's what my NFP is doing and I'm not arguing that point!. But, indeed, I believe as you say, that the degradation can be a little bit less by following the guidance you push. At one point my NFP fell precipitously but resurrected upwards following a recalibration charge (7-95%). That fall seemed to correlate with a sustained period at low SoC ~20% and below 20°C average, but, I have no proof that that was the reason. Indeed, apart from this one event, I have normally seen about the same level of variance in the SMT NFP as your observation. I had wondered how to relate my NFP data to the discussion you had with Dr Chaos (but I'm no mathematician).

Regards my FRR, I used TeslaFi for 3 months before I installed SMT and the SMT values thereafter. By eye the SMT FRR is less noisy than the TeslaFi values.

Could you quantify the magnitudo of the "fell" and the "resurrected" ?. And, at lower value, for how many time it stayed "low"? And when "resurrected" which level reached? anf for how many time it stayed "high"?
Just curious to know if you speak of a "resurrection" as mine MP3 that recovered ALL (not a part...) I mean ALL what lost in the first 1 year.

 

[DrChaos]

The worst case I referred to was the data in the papers you showed on this forum where degradation was highest (SoC >55-70% and temperatures above ~35-40°C). That's my labelling of "worst case". I now understand that it might not be the same for everybody. In your spare time a site for calculating remaining capacity would be great BTW do you distinguish between (real) degradation as a process and capacity? isn't (real) degradation what implies a capacity after a certain time? Sorry if that seems like semantics, that's not intended.

So you've hit upon the central experimental problem with batteries:

1) there isn't any easy objective measurement to determine the state of charge. There is no battery 'dipstick'. Only voltage can be measured but the relation to SOC is variable.
2) there isn't any objective measurement to determine capacity other than full charge and full discharge, monitoring every Coulomb and computing, and even then, the details and conditions how you do it change the number.

Since, in driving there isn't any #2, everything you see in a car, from the state of charge to the capacity is all guesses by electronics and approximate algorithms and fudge factors. They might be tuned to some battery in the lab once upon a time which isn't your actual battery.

So lots of stuff you see fluctuating on the screen is all random BS with little implication to what's actually happening with the molecules in the battery.

 

[conv90]

Yes, low SOC cycles seem to offset the NFP upwards, an in my case the offset is ”hidden below 0%.
My first drive to zero, well 0.39% i think, did not change the NFP. The two with around -2% did, however.

If you drive down to zero (or below) and let it sleep and then charge full, you might affect the NFP downwards. For me I had guite a few single digit SOC drives that did not lower the NFP.
I am happy to have the NFP where I thought it would be anyway. It doesnt matter if it goes up from the charging schedule or down for some reason.

I do not know why the NFP seems lower in general in MIC M3P.
I would guess that one part is the BMS underedtimating the ”new chemistry of 2170L”.
If you keep track of the cell temps so it is posdible to estimate the annual average cell temp( or in blocks for the seasons) its not that hard to estimate the degradation I think.

The degradation charts show about 3% for 9.6 months between 45-55% SOC and 25C. 3 x square root (12/9.6)= 3.35% first year. 14 months (right) = 3.6%
(My average SOC was 34% according to teslalogger last time I checked.)

3.6%, add about 0.5% for the cycles of the year = about 4-4.5%

Not easy to know where to begin( real capacity…) most M3P from us showed Around 80-80.5 but the EPA test had 82.1kWh output.
77 to 78 is realistic if the initial capacity was about 80 or a bit above…?

Sincerely i don't know the Average SoC and Average temp. A tesla Guy (when I was concerned about my NFP at ONLY 73.9kWh after 10 months) said it was around 37% Average SoC. Basically here the temp averages about 10 to 12 more °degrees than Sweden.
About MIC MP3 battery I don't understand if you speak about my march 2021 MP3 batt. Because I'm sure it's a USA made Panasonic 2170L.
About the real capacity of the beginning I never had values of NFP over 79,6. (one time at 100% I had 80,2 nominal remaining but returned to 79,6 only 2 km away from the charger).
[Just now , the NFP went 79,6 with 230-232 CAC]
This is really strange:
here the new updated graph:

 

[AlanSubie4Life]

Sincerely i don't know the Average SoC and Average temp. A tesla Guy (when I was concerned about my NFP at ONLY 73.9kWh after 10 months) said it was around 37% Average SoC. Basically here the temp averages about 10 to 12 more °degrees than Sweden.
About MIC MP3 battery I don't understand if you speak about my march 2021 MP3 batt. Because I'm sure it's a USA made Panasonic 2170L.
About the real capacity of the beginning I never had values of NFP over 79,6. (one time at 100% I had 80,2 nominal remaining but returned to 79,6 only 2 km away from the charger).
[Just now , the NFP went 79,6 with 230-232 CAC]
This is really strange:
here the new updated graph:
View attachment 811046

This must make you very happy! Now you just have to figure out whether it's really there! Won't be hard to determine, now; 5kWh is hard to hide!

Seems like there must have been a bug in the software only affecting very specific BMS situations. Or something...

 

[AAKEE]

About MIC MP3 battery I don't understand if you speak about my march 2021 MP3 batt. Because I'm sure it's a USA made Panasonic 2170L.

If your M3P is Fremont built you can disregard the MIC part.
MIC= made in China. I do not remember when they started shipping M3P from
China. But they have had the same 2170L battery, 82.1kWh spec.
I have seen a few, and they have not started at around 80-80.5, but rather around 77 or so.

If the average temp is 15C above mine, then maybe the battery has an average of around 25C?
As it seems, your car is a Fremont built one. I think you should use 80-81kWh as a starting point( maybe even 82kWh, as the
BMS in these cars seem to underestimate the initial capacity. I think the low values
is more or less a BMS software issue( in combination with the new chemistry etc).

81kWh is a good number, in the middle of lowest and highest probable number.
4% loss from that set you at 77.8kWh.
You should probably have a capacity around that number( + or - 1kWh?)

 

[Steve446]

Could you quantify the magnitudo of the "fell" and the "resurrected" ?. And, at lower value, for how many time it stayed "low"? And when "resurrected" which level reached? anf for how many time it stayed "high"?
Just curious to know if you speak of a "resurrection" as mine MP3 that recovered ALL (not a part...) I mean ALL what lost in the first 1 year.

Absolutely nothing like your experience which is incredible! In my case, going back about 4-5 weeks, the NFP started to drop more quickly than previously observed over the last 4 months: ~0.5 to 0.6 kWh, compared to the normal observed variance +/- 0.2.
Whilst the value of the drop was not huge, it was a consistent and "out of character" drop which when projected/trended to 12 months would give a much larger first year degradation as observed by NFP (and not MRR).
Leaving the car @ 10.5% SoC overnight completely reversed that drop for about 7 days when NFP again dropped a bit lower than before, so I did a BMS "calibration charge cycle" 7%-->95% which pushed the NFP back up again as before. Been there now for a couple of weeks.
You can see those changes in my previous post #4917 (SMT data graphed for NFP vs elapsed time).

I interpret all these data as a personal learning curve for me regards my Tesla battery ie am I seeing simple day to day variation or real changes since lots of real changes, however small, can be significant after some time. None of this stops me from enjoying the car though and learning from these forum pages and its knowledgeable contributors.

 

[GrandWazoo]

I see on this Wikipedia page that it can be anywhere from 54 to 82 kWh. But how do I find the size of my car’s battery?

 

[FlatSix911]

FYI - my monthly battery report for May 2022.

Get free monthly battery reports that give you insight into how your EV is performing and what you can do to minimize battery degradation.
(no affiliation with Recurrent)

Recurrent EV Battery Reports

Protect your EV battery with Owner Insights by Recurrent. This is a community of 20,000 drivers who foster EV adoption by demystifying range and battery health.

www.recurrentauto.com

 

[Kimmi]

Same question. 2021 LR AWD delivered December 29/12/2020.

 

[jjrandorin]

I see on this Wikipedia page that it can be anywhere from 54 to 82 kWh. But how do I find the size of my car’s battery?

By knowing what model you ordered, and / or calculating it using the cars screen with the thread stickied in the subforum you posted in:

Calculating Your Battery's Estimated Capacity Using the Car's Energy Screen

Re posting here and making this a sticky post so it doesnt get lost. @AlanSubie4Life originally posted: =========================================== (originally posted by @AlanSubie4Life. Original thread this appeared in can be found here: ( 2021 model Y scan my Tesla battery size)...

teslamotorsclub.com

 

[Dave EV]

Just now with 0,15kW load
View attachment 810269

Here's a few data points from my '18 LR RWD.

SOC   PackV MinV  AvgV  MaxV
46.7% 362   3.772 3.776 3.780
45.3% 361   3.758 3.762 3.766
35.9% 353   3.672 3.676 3.680

BTW, I am jealous of your CAC! Mine is quite a bit lower my min/avg/max is 195/197/199 rated range is currently around 292 miles. I do think the BMS estimate is off - my last couple charge to 50% have resulted in the car waking up to a SOC of 52%.

 

[NiNJaMoNKeH21]

Hi all, I am new here. Owner of a 2021 Model 3 Performance in US, took delivery October 2021. I am trying to figure out the size of the battery pack in my car, as I have been noticing my car's actual range is nowhere near the 315 mile EPA rating.

I looked under the car and got this number from the battery pack sticker: 1104422-00-W

I've done the formula to calculate battery capacity from this thread: 203 Wh/mi * 284 mi projected range / .80 (SoC) = 72065 Wh, which gives me 72kWh.

I am wondering if my car should have the 82kWh battery, or is it 75kWh?

 

[Steve446]

FYI - my monthly battery report for May 2022.

Get free monthly battery reports that give you insight into how your EV is performing and what you can do to minimize battery degradation.
(no affiliation with Recurrent)

Recurrent EV Battery Reports

Protect your EV battery with Owner Insights by Recurrent. This is a community of 20,000 drivers who foster EV adoption by demystifying range and battery health.

www.recurrentauto.com

View attachment 811597

View attachment 811598

That looks like an interesting site. Unfortunately it doesn't appear to work outside of the USA.

 

[3sr+buyer]

[Not sure why this post got moved here, since this is a UI question, not a battery health or range loss question.]

If the battery remaining display is switched to distance, it looks like it shows rated range based on the rated consumption and the estimated remaining charge in the battery.

On the energy display, it shows a solid line for the rated Wh/mile and a dashed line for your average in the last 30 miles (or whatever selected distance) of driving.

However, if the dashed line and solid line are in the same place (which seems to be 226 or 227 Wh/mile for 2022 RWD with 18" wheels), meaning that you are driving at exactly the rated consumption, the projected range that is displayed on the right side of the energy display is lower than the range shown next to the battery remaining display. Why would this be the case?