TeslaFi - Battery Degradation Reports (upload your data)

[AlanSubie4Life]

Oddly, my son's TM3 is showing 3.0 'Energy buffer'.

That's about what I would expect with a battery that has lost capacity to 65.6kWh. It's around 4.5% (I said 4.7%, but 4.5% is probably closer to correct given the 234Wh/245Wh ratio that is so well established), so about 2.95kWh. It's great to see the constant tracking so well to 4.5% of the full value with so much degradation - it confirms that is how it behaves.

Thanks for the confirming data. All makes sense. Pretty sure it is directly reported by the car, or it's calculated from two values (nominal remaining and "usable remaining" (which I've seen in some apps) ), which make the buffer size implicit.

 

[TimothyHW3]

The question is how do they hide it! 77.9 kWh should be more than 310 rated miles. What is the math behind it? I would have thought that they would just show greater than rated miles in those cases until the capacity drops. .

The math is simple - your car is capped at 310 miles, even though EPA calculated 322 or so. 78/3.22 is about 24.223 kWh/100km which is about 310 miles when you reach anywhere below 75.5kWh (75.5/24.223). By deliberatily caping the rated miles to 310 they are hiding that initial degradation from the math. Also, if you switch to km you should see 500km at full. But even though you see around 310 in miles the difference in km will be greater and you will already see a loss - at least 5km loss.

 

[TimothyHW3]

done!

Thank you. Is it possible to get the two lines, avg and the rated consumption, to match to see that number in km exactly?

 

[AlanSubie4Life]

the math is roughly 4.5% of usable

* You meant Nominal Full Pack.

Oddly you have a terrible degradation, but it could be just uncalibrated battery. I have never seen 10kWh loss on a LR (and it can't be MidRange, because they are at about 60kWh) That is about 13-15% degradatiom.

I keep saying that everyone is part of the distribution...

16% degradation after 40k miles is definitely a bit below average (there was an LR RWD car reported here with 116k miles or something with 308/325 remaining, which would be only 7.5% degradation from 78kWh). But 40k miles is a decent amount. Still, I'd be pretty annoyed to have 268 miles at 100% with that mileage.

 

[AlanSubie4Life]

By deliberatily caping the rated miles to 310 they are hiding that initial degradation from the math.

Yes, but there are a couple (at least) potential ways to handle this:

1) Stick at 310 rated miles for a while when you start driving (they definitely do not do this for Model 3).

2) Scale the energy content of each rated mile up to "fit in" the extra energy, until it degrades away. (appears to be what they do).

your car is capped at 310 miles, even though EPA calculated 322 or so

This is only applicable to the 2020 Model 3 Performance at this time. The rated miles for older AWD/Performance is 310 EPA miles (based directly on the EPA calculations from Tesla's dyno testing in 2018) and that's the framework Tesla uses for those models.

The 322 rated miles is accompanied by an actual efficiency improvement relative to prior results, and is not simply a modification (elimination) of the "hiding" of the extra 2kWh or so that you start with. I don't expect any changes to Tesla's treatment of this initial degradation with the 2020 Performance & AWD vehicles (still waiting for the final software updates to actually show the 322 mile range (not that it matters at all)! Should be coming soon.)

 

[TimothyHW3]

Maybe EPA has calculated the usable into the EPA rating. Or probably they take into consideration that battery capacity is not exact science and there are diviations. I have seen cars with 77 new and 78kWh new... No idea, either Tesla lowered the constant they received from EPA to make the 2kWh disapear or they just don't show anything above 310. Fact is that those 2kWh are not visible until 75.5kWh nominal full pack.

 

[AlanSubie4Life]

Fact is that those 2kWh are not visible until 75.5kWh nominal full pack.

I think the original question had no doubt about that. The question was exactly how it was handled.

Maybe EPA has calculated the usable into the EPA rating.

This stuff is all up to Tesla - they give EPA all the numbers from the Tesla tests, and the EPA just puts them in their spreadsheet and calls it good. The energy consumed in the EPA test is required to be all available for use to the end customer, however. And obviously the efficiency has to be in the right ballpark under the exact same conditions, on the dyno, as well.

One thing to keep in mind that the test articles have 1000-4000 miles on them at the time of the test so they may not be at "brand new" capacity. Which may be one reason why Tesla hides this initial drop. However, the measured energy in the test was 79kWh for the AWD vehicle. But, those kWh may not directly correspond to the "reported" kWh (they may not be calibrated the same way - the kWh in the car may contain slightly more energy than a true kWh). No idea on that one.

75.5kWh (75.5/24.223).

Fine tuning...but pretty sure once you drop below about 75.8kWh you will see 309mi or 499km

309.5rmi*245Wh/rmi = 75.83kWh

You tell me - saw in your other post you're in the low 75kWh range - so presumably you are no longer displaying 500km at a full charge, right? For 75.3kWh I guess I'd expect 496km rated or so. (75.3kWh/151.9Wh/rkm = 496rkm)

 

[TimothyHW3]

No, no longer 500km. It changed based on temp. I think it is 492 now, but it may get higher with temp rise.
The math on km is around nominal full kWh/153Wh/km

 

[AlanSubie4Life]

No, no longer 500km. It changed based on temp. I think it is 492 now, but it may get higher with temp rise.
The math on km is around nominal full kWh/153Wh/km

Hmm. So annoying. 153Wh/km is not equal to 245Wh/mi. Oh well. Anyway, using your value and your 492km number I'd expect your SMT value to be 75.3kWh. What is it, exactly, and how precise is your 492km number?

 

[SomeJoe7777]

You need an adapter to go from a unique one (of several) that Tesla is using to an OBDII. We got one for my sons car and he is using ScanMyTesla to display the data. Most details on the web that I know of are here: Diagnostic Port and Data Access - Tesla Owners Online

Ordered my cable and OBD2 adapter, will post BMS data from ScanMyTesla once I get everything set up. Because I have two vehicles, both logged with TeslaFi and both can be read by SMT, there should be some interesting data to correlate.

 

[ran349]

Yes, but there are a couple (at least) potential ways to handle this:

1) Stick at 310 rated miles for a while when you start driving (they definitely do not do this for Model 3).

2) Scale the energy content of each rated mile up to "fit in" the extra energy, until it degrades away. (appears to be what they do).

I don't see why they can't simply display extra miles in those cases. For example, 77.9 kWh would be 318 rated miles. It would give some people bonus miles for awhile, however short lived they were.
It would help balance out the much more common cases of people seeing less than 310 miles or whatever it is they expect from a new car.
No need to "hide" anything, or play with the constant temporarily.

 

[AlanSubie4Life]

I don't see why they can't simply display extra miles in those cases. For example, 77.9 kWh would be 318 rated miles. It would give some people bonus miles for awhile, however short lived they were.
It would help balance out the much more common cases of people seeing less than 310 miles or whatever it is they expect from a new car.
No need to "hide" anything, or play with the constant temporarily.

Yes, that's certainly the most obvious option and would probably mean the least confusion. You can imagine what would happen though...since there is probably some variability in initial start point, people would start looking for cars with the highest initial range and reject the others... "Why does my friend's car max out at 318 rated miles while mine maxes out at 314 rated miles?"

I'm more curious about what this means for warranty claims, since I haven't seen a lot of clear "plateauing" in the TeslaFi data posted so far. There's a pretty good argument that the claim should be relative to the initial capacity, not 310 miles. So that would be about 222 rated miles rather than a 217 mile limit. A small difference but not nothing.

Doing it the way you suggest would dispel this notion that cars are not losing capacity immediately, though. I imagine that cars are generally losing capacity the fastest when they are new. Everything would make a lot more sense!

 

[scottf200]

Would it be
It comes directly from the car and is calced by Tesla - it starts at 3.5 and gets lower with degradation - the math is roughly 4.5% of usable which in your case is exactly 3.0kWh.
65.6*4.5% is 2.95kWh which they round up to every 10th.
Oddly you have a terrible degradation, but it could be just uncalibrated battery. I have never seen 10kWh loss on a LR (and it can't be MidRange, because they are at about 60kWh) That is about 13-15% degradation.
Could you please post the All Tab screenshot in full with all the values? Or if you already have, maybe link me to it? And also maybe explain how you charge - could it be that your son charges 60/70%-90% daily and keeps it there?
Thank you!

This is my son's TM3. He drives about 110 miles / day. After looking at TeslaFI data he averages 40% battery usage. Use to charge to 75% (so down to 35%) and lately charges to 80% (so down to 40%). Actually I created an End_of_Day SOC chart for a little better accuracy but I started all the SOC Begin_of_Day using 80% but it still gives the gist even if off by 5%.

I realize you have a theory about not regularly going down to 20%. He does do that on road trips 2-4 times per year.

It is unclear what additional columns you are looking for in the 'All' tab. I don't have that. Here is the 'Battery' tab. NOTE: this is from last weekend when he charged to 100% right before doing some weekend errands. I have other charts showing the 100% charge lasted for 48 minutes until amps stopped flowing (ie. stopped charging).

Blue is Calculated 100% range.
Green is Firmware at the week but not release level (yyyy-wk-rlse).
Purple is the temp (F) plus 200 to scale it for the left Y axis.
2019.40.2.1 (aka 2019.777) seemed to be the biggest timing of the 100% range dropping.

 

[JBT66]

My October 2018 AWD is a bit over 10% loss at 15000 miles.
I charged daily at work to 90% for the first year and lowered that to 80% in October.
Most days it is plugged in all day long but never on weekends. I’ve used SuperChargers a hand full of times.

 

[TimothyHW3]

This is my son's TM3. He drives about 110 miles / day. After looking at TeslaFI data he averages 40% battery usage. Use to charge to 75% (so down to 35%) and lately charges to 80% (so down to 40%). Actually I created an End_of_Day SOC chart for a little better accuracy but I started all the SOC Begin_of_Day using 80% but it still gives the gist even if off by 5%.

I realize you have a theory about not regularly going down to 20%. He does do that on road trips 2-4 times per year.

It is unclear what additional columns you are looking for in the 'All' tab. I don't have that. Here is the 'Battery' tab. NOTE: this is from last weekend when he charged to 100% right before doing some weekend errands. I have other charts showing the 100% charge lasted for 48 minutes until amps stopped flowing (ie. stopped charging).


Blue is Calculated 100% range.
Green is Firmware at the week but not release level (yyyy-wk-rlse).
Purple is the temp (F) plus 200 to scale it for the left Y axis.
2019.40.2.1 (aka 2019.777) seemed to be the biggest timing of the 100% range dropping.

That is pretty decent information. So it seems your son charges 40%-80% and the cell imbalance seems to be sort of ok. He has almost 5x more AC charging so this 65kWh really could be that much of a degradation and not BMS imbalance...Strange at 40k miles.

The only thing that I can think of is that the temperatures are now very low. I see below 0C in your chart so maybe come spring time the kWh might get a little better. But I doubt by much.

This is the first so high degradation at these KM/Miles.

Is he driving a P, AWD or RWD LR and how long does he own the car for (I see some data from October 2018, was it then when he bought it?)?

The only other explanation I can find is that you have relatively high summer temps at above 30C - 90F so maybe this could also be an issue if you keep the car at high SOC% in the summer.

I dunno really, but it seems like a pretty high degradation at 40000 miles. I would speak to Tesla in the spring if I were you and have them check the car and see why the degradation is so high.

 

[AlanSubie4Life]

The math on km is around nominal full kWh/153Wh/km

For the record, I looked more carefully at this (it is obviously hard to get the precision, though I think more precision is easily possible with API access), and the value is 152Wh/rkm. (And presumably 244.6Wh/rmi). For now we can stick with the CAN bus/discharge constants being (1-0.045)*152Wh/rkm, etc., since the data we have seen does seem to suggest about a 4.5% buffer.

So I'd expect to see 498rkm or lower starting at around 75.8kWh, and 309rmi or lower starting at around 75.7kWh.

Anyway, using your value and your 492km number I'd expect your SMT value to be 75.3kWh.

And I would modify this to be 74.8kWh, not 75.3kWh. Anyway, easy enough to establish what the actual relationship is.

 

[TimothyHW3]

Hmm. So annoying. 153Wh/km is not equal to 245Wh/mi. Oh well. Anyway, using your value and your 492km number I'd expect your SMT value to be 75.3kWh. What is it, exactly, and how precise is your 492km number?

It is very precise and the number is indeed somewhere around 152-153.
Probably 152.5 rounded to 3 or something like that. I posted a screenshot a while back.
This is why I have hard time calculating in miles. Actually the math is 76.5 and below not 75.5 like I said.

And I actually tested this, when I was at around 76kWh it showed about 497km

Too many rounding errors you guys should change to km

 

[ran349]

It is very precise and the number is indeed somewhere around 152-153.
Probably 152.5 rounded to 3 or something like that. I posted a screenshot a while back.
This is why I have hard time calculating in miles. Actually the math is 76.5 and below not 75.5 like I said.

And I actually tested this, when I was at around 76kWh it showed about 497km

Too many rounding errors you guys should change to km

I think it's actually better to use miles, at least for US cars. The reason I say this is because I know that the API app provides the raw data in miles, and has to be converted to km by the app if the car is set to display km.
Also, I know my car's constant is exactly 287, an integer. I am able to verify that exact value by using the API miles value with 2 decimal places along with my TM-Spy data. The 287 number clearly does not convert to an integer for km. The conversion value for km is 178.3332.
Now maybe European cars do it the opposite way, and use a precise km constant that has to be converted to miles. If that is the case, then you should stick with km.

I have a question for you. Does SMT give a value for either rated miles or km? Or do you have to read the value from the car display?
I know TM-Spy does give a rated miles value for model S, but it only gives it to a rounded whole number. But I have been able to verify it is the same as the API value, except rounded. But I need the API number with the additional precision to verify the exact constant my car uses.

 

[TimothyHW3]

SMT doesn't gibe rated range, because there is no such value - this is an arbitrary value that the MCU calculates and displays on the screen based on the kWh and the constant. This is explained in the leaked inside documentation.

I know on S and X you can change between rated and some other GOM like value in the menu, I believe, so it might be different there.

 

[ran349]

SMT doesn't gibe rated range, because there is no such value - this is an arbitrary value that the MCU calculates and displays on the screen based on the kWh and the constant. This is explained in the leaked inside documentation.

I know on S and X you can change between rated and some other GOM like value in the menu, I believe, so it might be different there.

That's interesting because TM-Spy and SMT both provide rated miles for model S, so I'm surprised they don't provide it for the model 3. It's provided by the API for all cars, and it's the same number as OBD gives, only API gives it with 2 decimals.