Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

MASTER THREAD: 2021 Model 3 - Charge data, battery discussion etc

This site may earn commission on affiliate links.
Anyway in the interests of this discussion it would be cool if @inigoml retook his test case of 10-100% with SMT data this time. Obviously we would have to trust that the public charger was accurately metering the energy.

It’s certainly true that efficiencies are somewhat higher if you don’t charge to 100%, since the charging slows a bit at the end, and there may be rebalancing too. Could take an interim datapoint at 90% and continue the charge if you have the ability/time.

Today I've repeated the experiment going from 4 to 100%. It's slightly different because I recharged at home at 7.2Kw ( 220 x 32A ).
These are my values.... SMT says I've pushed 44.6kWh since TeslaFi says 50.81. So, charge efficiency is far away from what TeslaFi says.

On the other side, interesting SOC variations.... from 11.0 to 100% or 15.7 to 100% (including buffer). What does not make sense here is energy bar that shows only 4%. Perhaps BMS is miscalibrated? (remember I've a LFP battery)

1656418148740.png
 
  • Informative
Reactions: AlanSubie4Life
Today I've repeated the experiment going from 4 to 100%. It's slightly different because I recharged at home at 7.2Kw ( 220 x 32A ).
These are my values.... SMT says I've pushed 44.6kWh since TeslaFi says 50.81. So, charge efficiency is far away from what TeslaFi says.

On the other side, interesting SOC variations.... from 11.0 to 100% or 15.7 to 100% (including buffer). What does not make sense here is energy bar that shows only 4%. Perhaps BMS is miscalibrated? (remember I've a LFP battery)

View attachment 822079
Good information. Yeah not sure what to make of the SOC discrepancy. Weird that it said 4% on the screen in the car.

This efficiency is a bit lower than I would expect.

Tangent: There are ways to set user scaling on some of the TeslaFi parameters, by the way. Not sure if it will allow you to eliminate discrepancies though. Have to read their FAQ. I don’t use TeslaFi so won’t give any further details.
 
Last edited:
Good information. Yeah not sure what to make of the SOC discrepancy. Weird that it said 4% on the screen in the car.

This efficiency is a bit lower than I would expect.

Tangent: There are ways to set user scaling on some of the TeslaFi parameters, by the way. Not sure if it will allow to to eliminate discrepancies though. Have to read their FAQ. I don’t use TeslaFi so won’t give any further details.
Found in the settings
1656448494920.png1656448958479.png
 
Last edited:
  • Informative
Reactions: AlanSubie4Life
Good information. Yeah not sure what to make of the SOC discrepancy. Weird that it said 4% on the screen in the car.

This efficiency is a bit lower than I would expect.

Tangent: There are ways to set user scaling on some of the TeslaFi parameters, by the way. Not sure if it will allow you to eliminate discrepancies though. Have to read their FAQ. I don’t use TeslaFi so won’t give any further details.

Same calculation but assuming NFP in SMT excludes buffer....

1656500587217.png


I suspect that in a recent update Tesla (or SMT) has changed the value of "Nominal Full Pack" to exclude buffer. Of course, this is only a theory. Based in what?

Several cases from several sources:
- From Bjorn Nyland in a video of 55 kWh LPF with 100km. 53 kWh in SMT. In a previous one months earlier it was 55.3 kWh with a different 55 kWh car. In a car with 100km this is a HUGE loss.
- From a friend whose battery pack was replaced due to a malfunction. 80 kWh in SMT with the new battery (Panasonic NCA 82 kWh pack). Loosing 2kWh in a new battery seems improbable.
- My charge efficiency. 83% for 32A, 230V is really low and does not match with others have shown in the past (should be around 85-88% like in this case).
 
Same calculation but assuming NFP in SMT excludes buffer....
It always includes buffer. This has never changed. Usable Full Pack is without buffer.
55 kWh LPF
LFP packs have a different algorithm to calculate the energy buffer. If the car is regularly charged to 100% it behaves normal and uses 4,5% of NFP. Otherwise it asumes the capacity calculation is unreliable and increases the buffer up to 11kWh.
Loosing 2kWh in a new battery seems improbable.
Highly probable. Especially with the 82kWh packs from Panasonic. Those can be 74,# kWh when delivered up to 80'ish kWh. Barely any packs get to actual 82kWh. Some only after lots of calibration and tinkering with 60% charges.
 
LFP packs have a different algorithm to calculate the energy buffer. If the car is regularly charged to 100% it behaves normal and uses 4,5% of NFP. Otherwise it asumes the capacity calculation is unreliable and increases the buffer up to 11kWh.
WOW... this is too much!!! Where this info comes from??

The discrepancy between energy consumed and energy pushed into battery clearly does not match. Something is wrong there. As you say, it can be BMS (I recharge weekly unless I've a trip, twice at 100%, twice at 80-90%) or some changes concerning NFP (just speculation)... or heavy battery degradation (but only 10.000km and range still displays 415km / 257 miles from a total o 423km / 263 miles)
 
It always includes buffer. This has never changed. Usable Full Pack is without buffer.

LFP packs have a different algorithm to calculate the energy buffer. If the car is regularly charged to 100% it behaves normal and uses 4,5% of NFP. Otherwise it asumes the capacity calculation is unreliable and increases the buffer up to 11kWh.

Highly probable. Especially with the 82kWh packs from Panasonic. Those can be 74,# kWh when delivered up to 80'ish kWh. Barely any packs get to actual 82kWh. Some only after lots of calibration and tinkering with 60% charges.
I have seen my 2022 M3P with the Panasonic 82.1kWh pack go from an NFP in the low 79.x to 80.2 in the roughly 7 months I've owned it. I don't drive it much (total mileage under 1500), so I've kept the "parked" SoC usually between 20% and 35%.

I've put a high-capacity "Whole House" fan in the garage to pull hot air out, which usually manages to keep the garage at under 86F/30C in summer heat. The car was charged to 90% when it arrived (in early December) at the delivery center, and I've never charged it past 80% since, and have always driven it down to 50% or lower.

I've been trying to "take care" of the battery, especially the first year where calendar degradation is potentially high with high storage temp/SoC. The not-driving-much thing isn't really part of that, but it does allow me to keep the SoC lower than usual.
 
  • Like
Reactions: AAKEE and eivissa
The LFP started at 4,5% briefly, then cars started quitting with 10% SoC and above. Tesla soon increased the Buffer to 11% (not 11kWh as I wrote earlier). Later an update followed that introduced the variable buffer that ranges from the normal 4,5% to 11% depending on the BMS'es certainty of its own capacity.

I am not sure about the timing of it all as I've never owned a standard range Tesla, let alone the LFP battery pack.
 
Otherwise it asumes the capacity calculation is unreliable and increases the buffer up to 11kWh [% actually]

Great information. Thanks.

Seems to mostly explain what is seen above. Not sure what the meaning of SMT SOC and SOC expected (expected seems to be true SOC) is (it’s been explained to me before but I have forgotten). If I understand correctly (I may not, please correct!): Anyway it seems like an extra 6.5 % was allocated here taking it up to about 11% buffer. That would lead to about 4% displayed SOC.

8.4kWh, 11% of 53kWh is 5.8kWh, so 8.4kWh-5.8kWh = 2.6kWh of (53kWh-5.8kWh) is 5.5%, 2.6kWh of (0.955*53kWh) is 5%. Not sure which of the ways they would calculate the displayed SOC. Close to 4% anyway.

Confusing.

Same calculation but assuming NFP in SMT excludes buffer....
Don’t think you should do this, as discussed. NFP always includes the buffer, but sounds like the estimate can be off on these LFPs at low SOC. Still, you can trust the energy available shown by SMT is the system’s best estimate, even if it is not correct. So: It’s possible the system has less energy than the estimate at low SOC (hence the patches to the buffer on the LFPs), which would mean in reality you may have added more energy than the SMT delta. Which would yield better charging efficiency, perhaps closer to 85-90%.

You could try smaller additions of energy (say 50%), after a quick drive after a charge to 100% (not allowing time for BMS to drift off), and see whether the numbers are more reasonable.

In any case this still supports my claim that no more than 91-92% efficiency is likely possible. I guess one of these days I should check it in my car, though I’d have to just time the charge precisely, since I don’t have a dedicated meter. Also I don’t have SMT so would be multiple error sources.

I guess the SR+ (and RWD) vehicles getting 89% efficiency in EPA docs supports the idea that all vehicles, even the LR, are charged at 32A, 7.7kW, not 48A, for the 89% EPA number. Would be nice if they documented it. So at 48A 11.5kW it would be a bit higher efficiency, maybe 1-2% better. And yes the overnight loss in EPA test could knock 1-2% off these numbers. So maybe 93% best case?
 
Seems to mostly explain what is seen above. Not sure what the meaning of SMT SOC and SOC expected (expected seems to be true SOC) is (it’s been explained to me before but I have forgotten). If I understand correctly (I may not, please correct!): Anyway it seems like an extra 6.5 % was allocated here taking it up to about 11% buffer. That would lead to about 4% displayed SOC.
I think that

- SOC is without buffer.
- SOC expected is with buffer.

At least calculation matches exactly.
 
I think that

- SOC is without buffer.
- SOC expected is with buffer.

At least calculation matches exactly.
Its not.

The SOC and SOC expected changes the relationship from time to time. Most times there is only a few points delta.

I would guess the difference is something like the delta between the expected SOC after a drive and the measured due to the resting voltage.
The voltage can not be used during a drive as a absolut value as it differs much depending om the load. The most probable way to calculate on screen SOC during a drive is to use [ nominal remaining - consumed energy / nominal full pack] and correct for the buffer.
This will be a ”calculated SOC” that might differ from the true SOC that is defined as a voltage when the cell is at rest.
I would guess that the SOC vs SOC expected reflects this difference.
 
To late to edit my last post.
This is my M3P after it have been parked for >24h. Before this only a few very short drives duribg the last days, so the SOC and SOC expected should not differ much.
This, as the BMS have seen the OCV Voltages during long time.
1CDCC7D8-AD13-4939-8622-05C09BBAC490.jpeg



I guess that the Nominal Full Pack can be calculated like this:

-During a drive the used energy is subtracted from the nominal remaining, and converted to SOC (expected) for presentation on the screen.
-When the car have low power the voltage is measured to estimate the SOC, or connectors are open during a sleep the OCV is measured. The OCV defines the True SOC.
-SOC Expected is a estimate from the used energy and OCV measured is the true value for SOC.
-The SOC expected is compared to the SOC and is adjusted to be the same when the possibility is available to know the ”true” SOC by the OCV.
- If the SOC expected often show a lower number than the true SOC, the NFP is adjusted upwards and vice versa. This, as using 50% of the energy from a full charge should put the SOC at 50%.

Example: If the measured SOC is 52%, the used energy was not 50% but 48% —> this means that the capacity is higher than the current used value for NFP.

This way the capacity can be calculated and adjusted to roughly reflect the true capacity without the need for a full 100% to 0% cycle.
 
  • Like
Reactions: Steve446
Any US M3P 2021 or 2022 driver around, who is using Android based Scan My Tesla?

I would like to know what your cars maximum Current (Amps) is, when accelerating through 50-60mph.

European (Panasonic) cars stop at 1240 Amps and only increase slightly above that, when the SoC is getting very low.
 
Any US M3P 2021 or 2022 driver around, who is using Android based Scan My Tesla?

I would like to know what your cars maximum Current (Amps) is, when accelerating through 50-60mph.

European (Panasonic) cars stop at 1240 Amps and only increase slightly above that, when the SoC is getting very low.
Does it need to be a us M3P?
( I have the 2170L and android)