Firmware or degradation?

[insaneoctane]

I'm seeing a pretty noticeable loss in range just recently. Is it firmware or actual degradation? See the chart below. Orange triangles represent a new firmware adoption. The big loss happened right at the 40K mileage mark. Blue line is extrapolated range (left axis) and grey line is mileage (right axis).

 

[640k]

too hard to correlate firmware as a direct cause. based on your metrics, the data plots are too variable to reliably forecast any noticeable trending. each update seems to cause a ~5 mile range fluctuation that could be dependent on a lot more than just charging your battery. use, battery temp, amount of time on the charger vs. driving, etc., could all have significant impact to what you're plotting.

the last little dip is a bit concerning, but i'm curious to see what happens after a month, similar to the rest of your graph.

Also - you have no plot points to demo seasonality. did your battery charge points perform the same way around the same time of year? did you drive the same distance and same manner (wh/mi) as 12 months ago?

realizing you're in a relatively weather-stable area of the country, i'm not convinced by what i see.

 

[insaneoctane]

It's colder in Southern Cali, but would that really effect constant-based EPA / estimated range when fully charged?

 

[AlanSubie4Life]

would that really effect constant-based EPA / estimated range when fully charged?

Yes, temperature is accounted for, and quite quickly, too. I've seen a temperature related 6-7 mile change overnight. Fully warmed battery starting at an extrapolated 304 miles, overnight lost 6 rated miles (to 298 miles extrapolated) when cold soaked to 35 degrees, but stayed at the same 89% SoC. The same day, after returning to 90 degree temperatures with a good long drive, I was back to 304 miles at 100%.

If you're capturing temperature data with your captures it would be good to plot them alongside. It probably does not explain the variation entirely.

 

[KenC]

I'm seeing a pretty noticeable loss in range just recently. Is it firmware or actual degradation? See the chart below. Orange triangles represent a new firmware adoption. The big loss happened right at the 40K mileage mark. Blue line is extrapolated range (left axis) and grey line is mileage (right axis).

View attachment 489797

That's alot of driving. We've seen people in warm states, Arizona and Florida, show temperature correlated range drops. It may not necessarily be the nominal temp that causes it, but the temperature delta. I'm in a cold state, and I definitely have been showing temperature correlated range drop.

I think this chart is @Allistah's, and he's in Arizona, if I remember correctly. As you can see, the temp is on the right axis, and he sees 100F down to 62F, and his rated range estimate seems pretty well-correlated.

 

[madreag]

Driving habit isn't taken into account when calculating the displayed range. The Model 3 BMS uses the lowest brick capacity (this takes degradation into account) and the average brick voltage (balancing the bricks can help here) to calculate range. Only extremely cold temps are supposed to affect the calculations, very minimally.

Basically the only thing you can do is try to balance the pack... Unfortunately there is no quick way to do this. The BMS uses bleed resistors to very slowly lower the highest brick voltage (0.001 V per day). It will not start balancing until the battery SOC is greater than 85%. In order to balance the pack you should let your Model 3 sit idle as much a possible for several days/few weeks with a SOC over 90%.

The above chart showing a temperature correlation surprises me. As you can see below, this specific (also from Arizona) Model 3's range has been recovering even though temperatures have been dropping.

 

[petteriusa]

I'm at a pretty low odometer level right now but also noted a sharp drop off a few weeks ago. South Florida location so although it's been "cool" by no means cold. Doesn't seem to correlate to a firmware change either. Not very concerned yet, but interesting it was a fairly steep 2% drop all of a sudden and continues for now at least a slower decline. Updated to 40.2.1 on 16DEC last two charges on that firmware.

 

[Gasaraki]

Looks like seasonal change to me. It dropped from an average of around 316 to about 310 in Dec.

 

[KenC]

Driving habit isn't taken into account when calculating the displayed range. The Model 3 BMS uses the lowest brick capacity (this takes degradation into account) and the average brick voltage (balancing the bricks can help here) to calculate range. Only extremely cold temps are supposed to affect the calculations, very minimally.

Basically the only thing you can do is try to balance the pack... Unfortunately there is no quick way to do this. The BMS uses bleed resistors to very slowly lower the highest brick voltage (0.001 V per day). It will not start balancing until the battery SOC is greater than 85%. In order to balance the pack you should let your Model 3 sit idle as much a possible for several days/few weeks with a SOC over 90%.

The above chart showing a temperature correlation surprises me. As you can see below, this specific (also from Arizona) Model 3's range has been recovering even though temperatures have been dropping.View attachment 490529

Mine seems extremely highly correlated with temperature. This is from the past year, from January to December. I added moving averages in red and blue and you can see how well they match up. Got my car in December, and the rated range bounced around as low as 287. Spring hit and the rated range started to go up. Was around 310miles all Summer. Then Fall hit, and my rated range started to drop.

 

[insaneoctane]

Per @AlanSubie4Life, here's an updated plot with outside temp. It would only show the temp in my garage at the end of the charge cycle, so it's probably colder "outside". I had to remove the mileage to use the second axis for temp, but it's the same data as before, starting at around 25K and ending around 40K....

 

[AlanSubie4Life]

Per @AlanSubie4Life, here's an updated plot with outside temp. It would only show the temp in my garage at the end of the charge cycle, so it's probably colder "outside". I had to remove the mileage to use the second axis for temp, but it's the same data as before, starting at around 25K and ending around 40K....

View attachment 490676

Cool. As expected, it is definitely correlated, but does not fully explain the recent drop.

You could probably do a correlation plot and determine the correlation coefficient. Probably like 0.7 or something.

Correlation is not causation of course - though it probably is in this case.

More historical data would be good but sounds like it might not be available.

I do have some sneaking suspicion that the first couple % or degradation is “hidden” by Tesla. So the first winter you might not notice a drop. I have no proof of this though - I just can’t explain CAN bus readbacks showing over 77kWh (these kWh may be larger than a true calibrated kWh, BTW) on new cars without thinking that you can have 1kWh or so of degradation before there is an impact on the rated miles. Would need to see more data from new and aged cars to prove/disprove this hypothesis. (I used to not think this was the case but having a hard time explaining every datapoint otherwise.)

 

[KenC]

Seems to show enough correlation to me, given the constraints on when you get the temp reading. Mine are based upon the local Degree Day data. A number of the recent bigger drops, seem to match when the temp has a big drop as well.

 

[Zoomit]

Driving habit isn't taken into account when calculating the displayed range. The Model 3 BMS uses the lowest brick capacity (this takes degradation into account) and the average brick voltage (balancing the bricks can help here) to calculate range. Only extremely cold temps are supposed to affect the calculations, very minimally.

Basically the only thing you can do is try to balance the pack... Unfortunately there is no quick way to do this. The BMS uses bleed resistors to very slowly lower the highest brick voltage (0.001 V per day). It will not start balancing until the battery SOC is greater than 85%. In order to balance the pack you should let your Model 3 sit idle as much a possible for several days/few weeks with a SOC over 90%.

The above chart showing a temperature correlation surprises me. As you can see below, this specific (also from Arizona) Model 3's range has been recovering even though temperatures have been dropping.View attachment 490529

This is very specific guidance, that I personally haven’t seen before. What’s your source?

 

[madreag]

This is very specific guidance, that I personally haven’t seen before. What’s your source?

Someone had leaked a Model 3 Tesla service manual a while back on Reddit (post was deleted). It has some good information regarding how the displayed range is calculated and brick balancing.