Battery health tracking SS

[AAKEE]

For the several ways:

Drove 100-0%, let the car stand for a while and check the exact SOC After ”sleeping” a little.
From this (delivered energy) we can calculate the total capacity quite well.

One year ago I did this with my M3P.
75kWh delivered, 0.41% SOC left after a short sleep ( about 0.3kWh) and the buffer 3.5kWh was there. Totally 78.8kWh capacity.

We can also let the car sleep with high SOC, record the exact SOC*, drive a longer drive, record the delivered energy and let it sleep for a while and then record the SOC* exactly.
Used energy/ delta SOC* = capacity.

*) SOC = true SOC, not the displayed.

 

[AAKEE]

Yes, it is the very basis of the discussion.

And the fact that you indicate that a longer drive helps the BMS goes to the very core of what I'm talking about.

BMS can not be trusted by itself. You have to know if the BMS has had an opportunity to update it's calculations.

And I don't believe that it is just a longer drive, it's also a charge to 100% and let it sit for a bit (maybe even a day or two to let all the balancing occur).

Then and only then, will the BMS provide a good guess. But again, it's STILL A GUESS!!

Just put the picture here. Its very easy and you do not need to cause any cycles. Take the photo after a daily /regular charge.

 

[ewoodrick]

Just put the picture here. Its very easy and you do not need to cause any cycles. Take the photo after a daily /regular charge.

Okay, show me the exact picture that you want and I may think about it.

 

[AlanSubie4Life]

Okay, show me the exact picture that you want and I may think about it.

Ideally as close to 100% as possible. But 80-90% is fine.

(This projects to 71.9kWh or so including buffer, but only from 55%, so some rounding error possible.)

Based on your earlier 293rmi number, I expect yours to be around 68-69kWh.

Then and only then, will the BMS provide a good guess. But again, it's STILL A GUESS!!

Have you ever tried to look at the difference? It’s rare for the BMS to be off by more than a percent or two vs. the metered energy, and the values by design are intended to converge long term. It’s pretty excellent; one of the many awesome things about Tesla.

 

[AAKEE]

Heres how it normally looks on a full charge:
Nominal remaning, the energy the BMS estimates is in the battery right now.
Nominal full pack= the estimated capacity of the battery.
SOC is 100.00% and both nominal remaining and nominal full pack is 81.1 kWh.

In most cases for me, the nominal remaining has overshot the nominal full pack, but only with 0.3-0.4kWh (about 2km range).

The BMS is good at estimating the battery capacity.

 

[KenBlub]

Just got a brand new Model 3 long range. My daily charge is 50%. Just for grins I want to use the energy chart to determine battery capacity. Going on a long trip soon and will charge to 100% before leaving. How soon after it hits 100% should I leave? @AAKEE you mention taking a “longer“ drive. How far is longer? I once had a lengthy conversation with the wife about our different interpretations of I’ll be home in a “few” minutes. Thanks

 

[AAKEE]

Just got a brand new Model 3 long range. My daily charge is 50%. Just for grins I want to use the energy chart to determine battery capacity.

The energy chart calc can be performed at *any time*. No need to charge, no need to drive.
The energy chart calculation involves making division with the SOC number.
SOC is rounded to whole percent on screen and this means that at close to 100%, the rounding is up to 0.5% which cause a error margin of 0.5%
At 10% SOC the rounding still is 0.5%, but the effective error margin will be 0.5/0.1= 5%. So the higher SOC the less error margin.

No need for driving at all. The earlier drive data in the cars memory works as well as a new drive.

Going on a long trip soon and will charge to 100% before leaving. How soon after it hits 100% should I leave? @AAKEE you mention taking a “longer“ drive. How far is longer? I once had a lengthy conversation with the wife about our different interpretations of I’ll be home in a “few” minutes.

The longer drive I wrote about earlier is for ”see through” the BMS calibration isnt really needed (but possible). What you do is look for the change in the SOC between just arriving after the long deive and maybe one hour afterwards. If there is no change in SOC, the BMS has a good calibration.
If the displayed SOC decreases like arriving with 40% and having 38% after one hour, the BMS estimation is off and overestimates the capacity.
If the SOC instead increases after the parking, the BMS underestimates the capacity.
You should read an follow this link below for doing a longer drive.
This is a description how to perform this test / calculation
It can be performed with precise SOC number (including digits) from scan my Tesla, and also can be performed using the displayed range (higher resolution than SOC). Displaying km gives the best resolution.

Background:
My new Plaid showed a nominal full pack of 95.3kWh when I bought it (the BMS had’nt had gotten to see the battery capacity. Full pack when new is 99.4 kWh, so it was quite low compared to the expected. From daily use, the NFP increased about 0.1 kWh each day.

I made that long drive (240km to work) and calculated the capacity to 98.0kWh. Same result for two drives.
In the end, a couple of weeks later the nominal full pack reached 98.0 kWh and still is about this number.

 

[AAKEE]

Just got a brand new Model 3 long range.

The model 3 (2021-2023) has a ”degradation threshold” at 79 kWh. The range will not drop before the battery is below that capacity, so then it depends which battery you have. All model 3 LR had the 82kWh panasonic for a couple of years, with the EPA range 358 miles.
The new model 3 is advertised with 333 miles or so. I do not know what battery that is yet.

If you have the 82kWh ”Panna”, its expected to find 79 kWh with the energy screen method as it can not ”see” beyond the 79kWh degradation threshold.

If you have the new 333 miles version, the result is very interresting.

 

[KenBlub]

The model 3 (2021-2023) has a ”degradation threshold” at 79 kWh. The range will not drop before the battery is below that capacity, so then it depends which battery you have. All model 3 LR had the 82kWh panasonic for a couple of years, with the EPA range 358 miles.
The new model 3 is advertised with 333 miles or so. I do not know what battery that is yet.

If you have the 82kWh ”Panna”, its expected to find 79 kWh with the energy screen method as it can not ”see” beyond the 79kWh degradation threshold.

If you have the new 333 miles version, the result is very interresting.

The salesperson at the store said he believed it was a 78.1 LG battery but the store was very busy and no research went into it. It is rated at 333. It was a holiday and I don’t think the service department was even open. Thanks for the response. Instead of doing it after 1 trip, can I do the calculation between charges assuming I just take shorter trips and don’t charge for a few days? Should be the same result shouldn’t it??

 

[KenBlub]

The energy chart calc can be performed at *any time*. No need to charge, no need to drive.
The energy chart calculation involves making division with the SOC number.
SOC is rounded to whole percent on screen and this means that at close to 100%, the rounding is up to 0.5% which cause a error margin of 0.5%
At 10% SOC the rounding still is 0.5%, but the effective error margin will be 0.5/0.1= 5%. So the higher SOC the less error margin.

No need for driving at all. The earlier drive data in the cars memory works as well as a new drive.



The longer drive I wrote about earlier is for ”see through” the BMS calibration isnt really needed (but possible). What you do is look for the change in the SOC between just arriving after the long deive and maybe one hour afterwards. If there is no change in SOC, the BMS has a good calibration.
If the displayed SOC decreases like arriving with 40% and having 38% after one hour, the BMS estimation is off and overestimates the capacity.
If the SOC instead increases after the parking, the BMS underestimates the capacity.
You should read an follow this link below for doing a longer drive.
This is a description how to perform this test / calculation
It can be performed with precise SOC number (including digits) from scan my Tesla, and also can be performed using the displayed range (higher resolution than SOC). Displaying km gives the best resolution.

Background:
My new Plaid showed a nominal full pack of 95.3kWh when I bought it (the BMS had’nt had gotten to see the battery capacity. Full pack when new is 99.4 kWh, so it was quite low compared to the expected. From daily use, the NFP increased about 0.1 kWh each day.

I made that long drive (240km to work) and calculated the capacity to 98.0kWh. Same result for two drives.
In the end, a couple of weeks later the nominal full pack reached 98.0 kWh and still is about this number.

Sorry. Link not attached

 

[AAKEE]

The salesperson at the store said he believed it was a 78.1 LG battery but the store was very busy and no research went into it. It is rated at 333. It was a holiday and I don’t think the service department was even open. Thanks for the response. Instead of doing it after 1 trip, can I do the calculation between charges assuming I just take shorter trips and don’t charge for a few days? Should be the same result shouldn’t it??

If it is the LG 78.8kWh (78.8 "Full pack when new when" checking the BMS-data with the Scan My Tesla) they usually starts low and inreases the capacity during the first month or months and ususally ends up at about 79kWh as the top value.

I have seen pretty low starting numbers and worried owners due to this on the Swedish forums (they see a low max range at 100% initially).
A explantion about the "increasing capacity and follow up always have showed about 79kWh.

If it is the LG 78.8kWh, it is really good. Low degradation, it is initially smaller than the Panasocin 82kWh but long term it will keep the capacity better and have better range. It charges a little slower on SuC V3. Not a big deal for most people.

So a capacity test if its brand new will most probably show a low number, at least as the BMS is concerned.

Just use the energy graph calculation to begin with. It will be good enough. No need to drive, no need for any special charge. But charge a little higher than 50% would be good to reduce the error margin.
The NMC (and most probably the NMCA like the LG78.8) have the "step" in calendar aging at about 62% true SOC so 60% displayed SOC and below is on the good side of the calendar aging step. (disclaimer that htere arent much calendar aging tests on these LG NMCA yet available).

 

[AAKEE]

So I asked for few Model S older cars to calculate expected degratation in the Swedish forum.

1. Model S P85+ 2014/07, charged to 85% and recharge at about 50%. 265K km (165K mi).
My automated forumals got the degradation to 20,5%, and the range should be 348-354km (depending on different sources fpr the rated consumption.
This car has 354km range today, according to the owner.
Actually I probably have used a average cell temp slightly to high as the car never is in a garage, which mainly keeps the battery cool in sweden. Adjusting the temp down 5C which seems probable will set the capacity and range of today spot on.

2. Model S70D 2015/10. Charges to 70% recharge T bout 40% 78K Km.
This car has 333km at 100% according to the owner and my calc set it at 10,8% degradation and 340km.

As it seems, the calc works rather fine on old model S as well.

 

[AlanSubie4Life]

You should read a follow this link below for doing a longer drive.

I think you intended to include a link to your driving procedure to meter the pack energy content in another post, but still don’t see the link?

Just use the energy graph calculation to begin with. It will be good enough.

@KenBlub - just note that this method “maxes out” at a certain point. Once your battery has enough energy to show the maximum rated miles for your particular vehicle, this energy screen method will not display a higher capacity (even if your pack capacity is higher).

Not really a big deal, and after a year or so not relevant at all usually, because then you will be below the degradation threshold, and showing rated mile loss.

In other words: if you are showing the max possible rated miles at 100%, this method gives you the lower bound for your pack energy content.

 

[AAKEE]

@AlanSubie4Life
Oops! Link comming, after dinner.

The LG ”always start low, and tops at best just above the 79-line so if they didnt change that, it will not matter yet (or barely ever).

 

[AlanSubie4Life]

The LG ”always start low, and tops at best just above the 79-line so if they didnt change that, it will not matter yet (or barely ever).

Yeah, just not clear on the constant and degradation threshold on the US 333-mile Model 3 Long Range, which seemed to be the vehicle in question. You’ve said in the past which pack that is; I would have to look it up. Can’t keep track in my head, need a cheat sheet (probably available online)!

 

[AAKEE]

Yeah, just not clear on the constant and degradation threshold on the US 333-mile Model 3 Long Range, which seemed to be the vehicle in question. You’ve said in the past which pack that is; I would have to look it up. Can’t keep track in my head, need a cheat sheet (probably available online)!

My opinion is that in EU there is no other degradation treshold for the 78.8.

 

[AAKEE]

This is a description how to perform this test / calculation

Something happen with that link:

This example

This example includes using displayed range instead of smt

 

[msinfo]

Yeah, just not clear on the constant and degradation threshold on the US 333-mile Model 3 Long Range, which seemed to be the vehicle in question. You’ve said in the past which pack that is; I would have to look it up. Can’t keep track in my head, need a cheat sheet (probably available online)!

Would 2023 MYLR also have LG battery? I always assumed Panasonic battery in my MYLR but of course that was just an assumption ...

You mention the "cheat sheet" it WOULD BE cool if one existed or somebody know of one and share it

 

[msinfo]

This example includes using displayed range instead of smt

There is a video

I thought was very interesting IMO. Also, lot's of good info in that VERY LONG discussion. Thank you for sharing.

 

[AAKEE]

Would 2023 MYLR also have LG battery? I always assumed Panasonic battery in my MYLR but of course that was just an assumption ...

It still has the Panna 82kWh according to the EPA tests on dis.epa.gov
I havent heard or checked if the range is changed for MYLR.

The 333 vs 358 miles indicate a bigger difference than 82.1 vs 78.8.

Mathematically it should be about 343-344 miles with the LG.

The 4680 in MY has quite low energy density despite 4680 should be up there.
I think Tesla changed to NMC in that pack.
Is it possibly a inhouse NMC 2170?

"cheat sheet”

There was someone producing these in Germany for the tff-forum.de, but they was often a little wrong in some way.