New M3 Performance 8.4% degradation?

[AlanSubie4Life]

The classic faults of Tessie. Does not know or see the capacity.
All M3P 21 had the Panna 82.1kWh.

So youre at 73.8 kWh right now, seen from the range.

I think this is the approximate supporting info:
80.6kWh*289/315 = 73.9kWh

(Nominal 82.1kWh pack had degradation threshold of 80.6 or 80.7kWh.)

8-9% loss (might be as high as 10%)

0.936*78.8kWh = 73.8kWH (I think this is the calc that Tessie is doing - it’s only arriving at (measuring as directly as it can) the correct capacity though - both 78.8kWh and 0.936 are wrong.)

 

[AAKEE]

I think this is the approximate supporting info:
80.6kWh*289/315 = 73.9kWh

(Nominal 82.1kWh pack had degradation threshold of 80.6 or 80.7kWh.)

8-9% loss (might be as high as 10%)

0.936*78.8kWh = 73.8kWH (I think this is the calc that Tessie is doing - it’s only arriving at (measuring as directly as it can) the correct capacity though - both 78.8kWh and 0.936 are wrong.)

78.8 is a very common number.

The LG M50 ”5L” pack has 78.8 as the FPWN.

 

[AlanSubie4Life]

78.8 is a very common number.

The LG M50 ”5L” pack has 78.8 as the FPWN.

Yes. I meant wrong for this specific vehicle (US Performance (2021 - though no variation in US based on model year yet)). I guess Tessie can’t determine the correct initial value from the info available to them.

So they use 78.8kWh (which is wrong here), and measure 73.8kWh (correct) then calculate 93.6% (73.8/78.8 - incorrect). I think.

One out of three correct is not too shabby.

 

[AAKEE]

Yes. I meant wrong for this specific vehicle (US Performance (2021 - though no variation in US based on model year yet)). I guess Tessie can’t determine the correct initial value from the info available to them.

Yep, we agree totally .

The OTA / API data does not contain capacity numbers, only range and SOC. (Teslafi works with range as the degradation number).

Tessie tries to calculate capacity from charging (in a way the same way as I could forecast the real capacity on my Plaid when I got it with a grossly wrong estimated capacity by the BMS). It works fine for the sctual capacity but initial capacity is not that easy. Not all have Tessie from day one.
And a BMS that has not hotten the time to adjust the estimate might give wrong numbers if you get Tessie from day one.

So they use 78.8kWh (which is wrong here), and measure 73.8kWh (correct) then calculate 93.6% (73.8/78.8 - incorrect). I think.

One out of three correct is not too shabby.

Well, its a mess!

For many people it fool them to think they have very little degradation. People still mess up “usable capacity” which was changed to “capacity” recently = good thing.
Its extremely common in degradation discussions on facebook groups that prople think they have 1-2% degradation when the real BMS estimate is 7-10% or so.

Discussions about using slightly lower SOC do not start well when people always doing 90% says “I only have 1.5% and I charge to 90% which is the reason for my low degradation”.
In the end they find they was fooled by Tessie and also the charging tip Tessie gives on their home page. I havent checked it recently but they said 90% was the thing to reduce degradation.

 

[Gauss Guzzler]

3rd party apps like Tessie can only guesstimate battery condition because they don't have any way to directly measure it.
If you really want to know your actual battery condition, you have to tap the button on the service mode screen to do Tesla's built-in battery health test.

 

[Steve446]

Yep, we agree totally .

The OTA / API data does not contain capacity numbers, only range and SOC. (Teslafi works with range as the degradation number).

Tessie tries to calculate capacity from charging (in a way the same way as I could forecast the real capacity on my Plaid when I got it with a grossly wrong estimated capacity by the BMS). It works fine for the sctual capacity but initial capacity is not that easy. Not all have Tessie from day one.
And a BMS that has not hotten the time to adjust the estimate might give wrong numbers if you get Tessie from day one.


Well, its a mess!

For many people it fool them to think they have very little degradation. People still mess up “usable capacity” which was changed to “capacity” recently = good thing.
Its extremely common in degradation discussions on facebook groups that prople think they have 1-2% degradation when the real BMS estimate is 7-10% or so.

Discussions about using slightly lower SOC do not start well when people always doing 90% says “I only have 1.5% and I charge to 90% which is the reason for my low degradation”.
In the end they find they was fooled by Tessie and also the charging tip Tessie gives on their home page. I havent checked it recently but they said 90% was the thing to reduce degradation.

Definitely a mess! TeslaFi is no better since it also uses range. I find range a very noisy signal compared to NFP which I have used to estimate "degradation". Using my range data in TeslaFi would give a "degradation" of about 1.5%.

I really think its quite difficult to estimate "degradation" with any level of accuracy since there are many factors contributing errors of calculation. My exercise below shows this.

I tried to perform a reality check for my monthly average NFP value by totalizing the amount of kWh added for a displayed zero to 100% charge and comparing it to capacity estimated from NFP. I get this.

Estimated capacity kWh from NFP by SMT	72.4		72.5 +/- 0.21 SD		Daily value (L) & monthly average (R)
Buffer kWh	3.26		3.27
Usable capacity (display 0 to 100%)	69.14		69.23		Corresponds to capacity calculated from NFP-buffer
kWh Added over 2.7 d (7, 7 kW charges)	72.21		72.21		ie 7 separate charges were used to get from 0 to 100%
Phantom Drain (TeslaFi / my data), 2.7d	2.2		1.89		Phantom Drain by my observation for a typical November is 0.7 kWh/d
Usable capacity added after Phantom Drain	70		70.32		Corresponds to capacity calculated from kWh added - Phantom Drain

The weak points of my estimation are that the 0 to 100 was done over 2.7 d and so I tried to compensate for that by correcting for Phantom Drain during that time (without which could lead to an overestimation at next charge). Other errors could be value rounding and equipment and software reporting errors. On the basis of this exercise, I would conclude that the "degradation" is ~ 73.3/74.5 = 1.7% for the kWh added calculation and 72.5/74.5= 2.7% using NFP. My FPWN is 74.5 and all-time average SoC over 26 months is ~35%. Well maybe somewhere between 2 to 3%, maybe not who knows! One thing is certain - the only way is down!

I understand why many EV drivers focus on "degradation". In this case, my exercise made me think of a vulgar but wonderful phrase sometimes used in spoken French, "on essaye d'enculer une mouche" (we are trying to "b*****" a fly") by literal translation Sorry mod. for the last phrase, I don't want to break forum rules but just claim it's in context.