Is battery calibration pointless?!

[AAKEE]

Was test on individual cell? Or complete lack with Tesla BMS?

Individual cells.

But a BMS can not make anything against calendar aging and cyclic aging.
It can balance the pack by making the cells have similar voltage, other than that it can not stop cyclic aging and calendar aging.
For cyclic aging the BMS can Reduce the power when the SOC is low or the pack is cold, to reduce high wear but it can not hinder the regular aging.
Just in the same way the doctor can mot stop us from aging.

 

[AAKEE]

If the BMS believes that your battery is at 100%, it will stop charging (assume you've set it to 100%)... right?

Nope, the SOC is actually a cell voltage number. 4.20V is 100%, and BTW there is no top buffer or similar.
So the charging will continue until the highest cells are at 100%(4.20V) and from that there will be balancing by burning of voltage from the high cells.
This takes a while, so it will most probable not be done when the car says ”chsrhing finished”.

If the car has a battery with severe imbalance the total SOC will not be 100%, as the highest cells reach 100% (but some will be burned of) and the low cells might be at 9X %
The BMS most probably use the average cell voltage to calculate the SOC presented.

The cars do in general not balance the cells at lower SOC numbers. So if the car has not been charged full for a while it might have a little more imbalance, showing less than 100%.
If the var is left at about 90% or more the car balances the battery.
Doing charges to 100% a couple of times, or leaving the car slightly longer above 90% will most probably fix the imbalance and give a higher SOC reading including more energy to use on the drive, and a longer range.
But this part is not directly coupled to the BMS calibration.

Lithium batteries in general, do not increase the capacity by large cycles as we get when doing a BMS calibration. So we do not increase the real range.
BMS calibration is mostly recommended to people that are unhappy with the range reading/ appearent degradation.
As I wrote before, in cases with a big missmatch between the capacity and the real capacity the nav function will get more precise.

My M3P had a moment of severe misscalc of the capacity. I did not do a BMS calib, I just used the car as usual and eventually the BMS adjusted right.
I made a 100-0% measuring drive during the time of the BMS beeing off track, so I knew the numbers and new that the capacity was 79kWh, despite the BMS thought 75.7kWh.
I used 75kWh, still had 0.4kWh left or so, and 3.5 or so in buffer.
I did drive 513km despite the BMS was fairly off on that 100-0% drive and later drives when the Bms Was back in track did not indicate any more energy.

The dip can ve seen here:

 

[father_of_6]

If the car has a battery with severe imbalance the total SOC will not be 100%, as the highest cells reach 100% (but some will be burned of) and the low cells might be at 9X %
The BMS most probably use the average cell voltage to calculate the SOC presented.

So if I understand correctly, in this scenario you could set the car to 100%, but it would stop at e.g. 92% because of cell imbalance?

But given long enough charging, it would bleed off the high cells and eventually make it to 100%?

If that's true, why does the calibration routine require starting at a low SoC?

 

[father_of_6]

but it would stop at e.g. 92% because of cell imbalance?

"Stop" is the wrong word. Would charging get really slow, or essentially pause at 92% while the imbalanced cells are bled down?

Or does this happen fast enough that it would not impede the charging speed?

 

[AAKEE]

So if I understand correctly, in this scenario you could set the car to 100%, but it would stop at e.g. 92% because of cell imbalance?

When the cells with the highest voltage reach the max voltage, the charging stops.

The balancing process take time so it will say ”charging done” long before the balancing is finished if we have normal or higher imbalance.

There is no other credible reason for the charging not to reach 100% on the display.
( that the highest cells have reached 100%)

Normally when charging lithium batteries the maximum voltage has been held by the charger for a while, and the charging stops when the current has reduced to a low number.

But given long enough charging, it would bleed off the high cells and eventually make it to 100%?

No. Lithium battery chargers that balance by adding current to the low cells (much more complicated for a car with thousands of cells) can do this, and also be done with charging and balancing within 1 hour.

Tesla use resistors to bleed the cells with high voltage, this is a slow process.
The charging stops when the Tesla/BMS sees that the highest cells is full.
The balancing (or at least the last part of it) will be done after the charging stopped. This burns energy so the SOC will decrease during the balancing part, after the charge.

If that's true, why does the calibration routine require starting at a low SoC?

The BMS calibration routine is something completely different.

BMS calibration is about letting the BMS see the battery capacity. For this, we of course need to have both high and low SOC.
Also, we need to let the car sleep at both high and low SOC.
The real SOC number is only seen by the BMS when the car is sleeping ( = main battery disconnected). This means the main battery is at rest and the Open Circuit Voltage (OCV) can be read.
When the car falls into sleep after a drive or charge and the main battery disconnects, it takes time for the OCV to reach the true value. If you did drive, its lower and slowly goes up to the true value. If you did charge it goes down slowly.
The real SOC is only seen by the BMS by measuring the OCV when the car sleeps and the main batt is disconnected.
This is the reason for the couple of hours sleep needed after a charge when doing the BMS calib.
(The car can not measure the SOC while driving, or when charging so it is always a preliminary SOC, calculated from estimated capacity, SOC before charging, estimated energy on board minus the used energy or plus the charged energy(which in turn also is estimated)).

 

[Darmie]

I recommend reading this research report:
https://mediatum.ub.tum.de/doc/1355829/file.pdf

And this:
ShieldSquare Captcha

This might be a good read (actual model 3 cells), even if they had much higher charging and discharging rate than we have regularly in hour cars:
http://kth.diva-portal.org/smash/get/diva2:1755178/FULLTEXT01.pdf

Tesla model 3 cells:
https://research.chalmers.se/publication/535926/file/535926_Fulltext.pdf

Tesla model S cells, calendar and cyclic aging:
ShieldSquare Captcha

And this:
https://www.researchgate.net/profil...-side-reactions.pdf?origin=publication_detail

From the first report above, se have cyclic aging on Panasonic NCA cells:
They last about 750 FCE when cycled from 100-0%
750 FCE would be about 300K km
They last about 1000 FCE when cycled about 90% to 0% (400K km) and they last more than 1000 FCE when cycled 80-0%.
View attachment 981215

If 100% and 0% was very bad, they would not last for 750FCE and 300.000km or 200.000 miles.

I'm trying to muddle through the aging report.
Can you give me the Cliff notes on this chart? Is this indicating high C rating bad?

 

[AAKEE]

I'm trying to muddle through the aging report.
Can you give me the Cliff notes on this chart? Is this indicating high C rating bad?
View attachment 981611

Yes, in general higher current is worse.
Thats why Supercharging causes more wear than normal charging.

Reduction of current only reduce the degradation down to a certain current.

For normal temperatures this will be around 0.3-0.5C for NCA cells. Further reduction will not reduce the cyclic aging.

The picture show Panasonic NCA with 2.9Ah capacity, so 1A is about 0.3C

All AC home charging is below this so there will not be a reduced cyclic aging by reducing the charging power.

11kW is about 0.15C on a model 3/Y LR/P.
or 11% on a 100kWh S/X.

Also, remember that cyclic sging is the smaller part of degradation the first 5-8 years or so, for the normal user.
Its not bad to refuce cyclic aging, but if you need to choose, choose calendar aging.

I always charge with 11kW at home. This means I can let the car stand at a low SOC longer and charge later = less calendar aging.

 

[DJ Palmis]

I did run my M3P with the low SOC stategy for the 2.5 Years I had it.
Charged to 55% daily (as 55% gives the same calendar aging as the lowest charge setting, 50%).

Hi AAKEE,
Just a curiosity. I follow your posts with interest, as you may rember.
And you always say that you charge to 55% because is the same and so on.
But the displayed 55% is not the true SoC as you said. In service page, some versions ago, there was the true SoC value. I remember that at 54% the true soc was 58% (I don't remember well, sadly they removed this info from the LV page).
Question is: you just don't care (didn't care).. or the graphs (that states that calendar aging is worst after 55%) are based on the "relative soc"?

 

[AAKEE]

Hi AAKEE,
Just a curiosity. I follow your posts with interest, as you may rember.
And you always say that you charge to 55% because is the same and so on.
But the displayed 55% is not the true SoC as you said. In service page, some versions ago, there was the true SoC value. I remember that at 54% the true soc was 58% (I don't remember well, sadly they removed this info from the LV page).
Question is: you just don't care (didn't care).. or the graphs (that states that calendar aging is worst after 55%) are based on the "relative soc"?

Of course my 55% is adjusted for the true SOC, so safe
(Already answered exactly this a few days ago).

The true SOC limit for calendar aging invrease with SOC is around 57-58% on a new NCA cell (it sctually can differ if the chemidtry or amount of different chems on the anode or cathode).

Google calendar aging + central graphite peak for more info.

This ”limit” increases from calendar aging, so a calendar aged cell with 5% loss would have it somewhere around 60%.
But the ”limit” decreases by lithium plating, which we get from fast charging.
In general, I would guess the ”limit” goes up for more or less anybody, so 57% True SOC is safe.

The buffer is 4.5% below 0% displayed.
100% displayed = 100% true SOC.

This means the true SOC = displayed SOC + ((100- displayed SOC)x0.045.

So, 55% on screen is 57% true SOC (57.025% to be exact).

(I did double check this the first time I connected Scan my Tesla almost three years ago, so it is confirmed to be correct)

 

[DJ Palmis]

In general, I would guess the ”limit” goes up for more or less anybody, so 57% True SOC is safe.

..ok so I have to change my daily target SoC from 50% to 55%
I thought you meant 55% true soc, so 50% on screen!
Thank you, as always!

 

[AAKEE]

..ok so I have to change my daily target SoC from 50% to 55%
I thought you meant 55% true soc, so 50% on screen!
Thank you, as always!

Dashed line, at 57%.

In text

Here, the movement of the central graphite peak. It moves with the degratation rate. Low SOC + cold = slow or almost no movement.

55 is a good number!
About maximises the energi with still low calendar aging. (Lower SOC is in general better but the calendar aging is quite flat between 30-55%.)

 

[shahryaran]

Smaller cycles are better, thats no wuestion sbout it.

Low SOC is proved to be better.

I did run my M3P with the low SOC stategy for the 2.5 Years I had it.
Charged to 55% daily (as 55% gives the same calendar aging as the lowest charge setting, 50%).
I did charge late, thats so the car was just ready for the drive the next day.
Most often the car ended ip with 20-30% in the end of the day.
After I changed job, I have a long drive so I arrive with 10-15% and leave it like that for the week at work. I charge just before driving home.

My car had the lowest, or was among the cars with the lowest degradation in the Teslafi battery report for the same tesla.
I still had 492km range, out of 507km when I did the last full charge, and the average teslafi M3P with the same miles (66000km) had 463-465km.
My car often stood with low SOC, and I was down to 0% several times, and to -2% twice.

Interesting

 

[aerodyne]

Dashed line, at 57%.
View attachment 981703

In text
View attachment 981704

Here, the movement of the central graphite peak. It moves with the degratation rate. Low SOC + cold = slow or almost no movement.
View attachment 981705

55 is a good number!
About maximises the energi with still low calendar aging. (Lower SOC is in general better but the calendar aging is quite flat between 30-55%.)

For years, based on Elon tweet that 50-62% SoC was best for storage on the MS, I've kept the cars at 57%. Interesting that this agrees with you data and recommendation.

Recently with the new MS, I have reduced that to 50%.

There seems to be a very slight improvement on estimated range compared to the Teslafi fleet since, but it is too small to be conclusive.

One comment, Tesla now recommends charging to 80% for daily use on the MS, not 90%.

 

[AAKEE]

One comment, Tesla now recommends charging to 80% for daily use on the MS, not 90%.

One note is that before the change to 80%,
Teslas advice was “below 90%” for daily. Which in turn ment anything between 50–89 (or 90%).

The “Daily recommended” do not show unless you set the slider to 80% or above.
Which most certainly means that Tesla means that anything below 80% is OK as the daily charging level.

I have seen a lot of ideas around (not here on TMC though) that the new recommendation “80%” means 80% and nothing else…

 

[zoomer0056]

The “Daily recommended” do not show unless you set the slider to 80% or above.
Which most certainly means that Tesla means that anything below 80% is OK as the daily charging level

The daily upper recommended value shows on the charging screen slider as a delineation mark between trip and daily.

That little tick mark used to be at 90% and was changed to 80% in update 2023.26.

 

[stioc]

But what about the LFP? Tesla recommends we charge these bad boys to 100% at least once a week!

 

[shahryaran]

But what about the LFP? Tesla recommends we charge these bad boys to 100% at least once a week!

Refer to AAAKEE's posted documents

 

[stioc]

I've read a few long threads here in my short time. I guess the conclusion I arrived at is even with the LFP battery I should just prob charge to 60% (at work), get home with 48% then back again at work the next day with 35%. Rinse and repeat (35% -> 60%) daily. On weekends charge as required otherwise let it sit at a low SOC don't worry about what the BMS is reporting. Am I on the right track?

With only 1300 miles on my M3 and exactly one month later I already have a 0.2% degradation according to Tessie. Mostly L2 charging (6kw) to 80-90% SOC, a couple to 100% SOC and two supercharge sessions from approx 15%-85% SOC on the one and only long road trip so far.

 

[shahryaran]

I've read a few long threads here in my short time. I guess the conclusion I arrived at is even with the LFP battery I should just prob charge to 60% (at work), get home with 48% then back again at work the next day with 35%. Rinse and repeat (35% -> 60%) daily. On weekends charge as required otherwise let it sit at a low SOC don't worry about what the BMS is reporting. Am I on the right track?

With only 1300 miles on my M3 and exactly one month later I already have a 0.2% degradation according to Tessie. Mostly L2 charging (6kw) to 80-90% SOC, a couple to 100% SOC and two supercharge sessions from approx 15%-85% SOC on the one and only long road trip so far.

I think charging between 30-60 is a good balance between prolonging battery health and practicality.

 

[rjpjnk]

Yes, pointless.