I'll need to plot my car soon, but your range degradation looks very similar to what I have experienced with my Mid Range, also built in December of 2018. It's interesting to see the step changes rather than a smooth decline, which I would assume it's an indication of a change on the software end and rather than a true battery degradation.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Battery degradation - 2019.32.2.2
- Thread starter cherrybomb
- Start date
Based upon what we've seen already with @Allistah, who also is in a hot clime and the one vehicle in Florida, whose owner's name I can't recall, it's not the nominal temp that is correlated with BMS drift, but the delta, the change seems to trigger BMS drift. Notice, that you seem to show more drift, when the temp drop is greater. Not always, but more often than not, your bigger range drops happen right after a big temp drop.
So, that's about 6 of us who have shown some correlation between Rated Range estimates and temperature change.
The whole bottom buffer change is interesting, and hopefully more people with CAN bus data can confirm it.
This isn't the point of the thread.... we all know this already. The Range display, however, according to Tesla, is NOT adjusted for anything. It is supposed to be, again, according to Tesla, just current_battery_energy * EPA_rated_Wh/mi. From the above linked support page:
What we are discussing here is that there seems to be some mounting evidence that Tesla has changed the way the range is displayed, changed the amount of energy reserved in the buffer, changed the BMS to consider outside temp, and/or changed the range display to not include the buffer (which, in the past the range display DID include the buffer).
Edit: Typos/grammar
So one thing that doesn't make sense if this is just a temperature correlation... If this were true, then any new car would not display anywhere close the rated range when fully charged if purchased in the winter. But we know that new cars being delivered now, in general, charge to full EPA rated mileage.
Also, there are also many reports of folks still getting 310/325 for LR cars (or within a couple of miles) after many months/yr+ ownership.
SilverSp33d3r
No Longer Silver
Correct, range is based on constants however the bms has to “calculate” the energy in the pack to get total useable energy. This calculation is what is changing not the constants.
which is exactly what I said, no? ".....there seems to be some mounting evidence that Tesla has .....changed the BMS to consider outside temp...."
Looking at the CAN bus on my SR+, I can see exactly how it's calculating the range.
(Expected remaining kWh - Energy buffer kWh) / Discharge constant
The current value of Expected Remaining for my car is 31.1 kWh. Energy buffer is 2.3 kWh. The discharge constant for SR+ is 209 Wh/mi. Plugging them in gets me 137.8 miles. The range the car is displaying on the battery gauge? 138 miles.
I don't know why Tesla changed this to subtract the buffer, or why some cars still display the range without subtracting the buffer. But the math always works no matter my battery level.
(Expected remaining kWh - Energy buffer kWh) / Discharge constant
The current value of Expected Remaining for my car is 31.1 kWh. Energy buffer is 2.3 kWh. The discharge constant for SR+ is 209 Wh/mi. Plugging them in gets me 137.8 miles. The range the car is displaying on the battery gauge? 138 miles.
I don't know why Tesla changed this to subtract the buffer, or why some cars still display the range without subtracting the buffer. But the math always works no matter my battery level.
I suspect they changed it because it was technically displaying range you could never use because they tick off the miles faster to make it hit zero with leaving the buffer - so in other words, if you drove the rated Wh/mi, you would get less range than displayed.. By subtracting the buffer first, if you drive at the rated Wh/mi, you should get exactly the range displayed. They might be trying to make it "better" ...?
However, if over the course of a few recent updates Tesla made the following changes:
- Increased the buffer size
- started showing estimated range without the buffer
- Added some algorithm to account for ambient temp to the BMS
What doesn't fit though, is that would mean every car sold would immediately "lose" range at first update or they would never charge to the advertised range value even at new, which I think would be a lot more vocal.
Data Points:
Model: LR AWD - Stealth (SO FAST)
Build Date: 09/19
Current Miles: 4.3k/Miles
Current SOC to Miles: 90% SOC = ~274
Location: LA, California
Current Average Temperature: 50-60º (Colder Months)
Software/ Notes:
Typical Charge Patterns: 90% SOC 28/30 days in the month
I wish I would have found this thread earlier. Last night I drained my battery down to 3-5% and have been charging it up to 90% SOC from around 0100hrs. 6kW/21mi/h on my NEMA 10-30 which will equal ~12hrs of charging. I'll be done in about 1:30min and will report back.
Yes, I'm fully aware these are the most advanced batteries and performing a "battery recalibration" is either not needed or futile. But, hey, I had to try something it was getting extremely annoying.
I, unfortunately, have difficulty pinpointing if it's the colder California weather or the 36.2.1 update that caused this issue. It literally came out of the blue!
Hopefully, this helps the group!
Model: LR AWD - Stealth (SO FAST)
Build Date: 09/19
Current Miles: 4.3k/Miles
Current SOC to Miles: 90% SOC = ~274
Location: LA, California
Current Average Temperature: 50-60º (Colder Months)
Software/ Notes:
2019.36.2.1 - Noticed the range difference
2019.36.2.2 - Range difference still present
2019.36.2.3 - See Below
Average Commute: 20-40/miles a day2019.36.2.2 - Range difference still present
2019.36.2.3 - See Below
Typical Charge Patterns: 90% SOC 28/30 days in the month
I wish I would have found this thread earlier. Last night I drained my battery down to 3-5% and have been charging it up to 90% SOC from around 0100hrs. 6kW/21mi/h on my NEMA 10-30 which will equal ~12hrs of charging. I'll be done in about 1:30min and will report back.
Yes, I'm fully aware these are the most advanced batteries and performing a "battery recalibration" is either not needed or futile. But, hey, I had to try something it was getting extremely annoying.
I, unfortunately, have difficulty pinpointing if it's the colder California weather or the 36.2.1 update that caused this issue. It literally came out of the blue!
Hopefully, this helps the group!
You're right, @hcdavis3 has a LR-RWD, and I think he believes he's been getting 325 miles, but in the past month, his rated range is showing a few miles less than 2 months ago. But, his last reading was 324.5miles, so it's not deg, it appears to be BMS drift, related to ambient temperature change.
My car I believe is still showing 100% at 310 miles, a LR-AWD, but I get these lower numbers from time-2-time, 303 on Nov 14th, for example. Both of us believe we are showing zero deg.
It was only looking at @hcdavis3's graph that I noticed his looked like mine, and that we were in the same geographical region. That made me think that what we are seeing isn't true deg, but some sort of BMS drift, due to temperature swings. He shows peaks and valleys almost exactly where I show peaks and valleys. Too odd to be a coincidence. The only common link is regional ambient temps.
Now, neither of us are showing the sudden large drops, which may or may not be something different. There could be and probably are multiple things happening at once: temp change, software updates, change in buffers, change in constant, etc. That makes deciphering what is happening difficult. But in some cases, some of those variables are not a factor, making some insight possible. In @hcdavis3's case and my case, we don't believe were showing deg, since our change is so small, and the possible buffer change and constant change don't seem to be a factor since we're still hovering around 325 and 310, respectively. Which leaves a strongish correlation to temperature change.
Back to your first point, I don't think this is a nominal correlation, ie, if you live in colder temps, you have less range, that's something separate. I think it's the delta, that correlates to BMS drift. So, it doesn't matter if you're going from 100F to 60F, like @Allistah, or 80F to 30F like @hcdavis3 and myself. It's the delta that causes the BMS to start to drift, losing track of how it calculates the input and output of voltage. I thought it might be due to vampire drain effects, but that's just my theory.
Thanks Ken. I took delivery of my car in early March of this year so as the temperatures began to climb my chart makes sense. I agree with your conclusion. I have to believe that colder temperatures are the culprit here. The forums are now filled with posts about degradation and as temperatures drop the cold makes sense. Most of us haven’t been through a winter with our cars yet.
If you have your data from Stats, the battery health CSV file and phantom drain CSV file, I can see if I can match some of the peaks and valleys to my data to actually see if they are the same days. Would strengthen the theory.
Sorry Ken, I don’t have Stats. I used to but I couldn’t find a way to get it to stop polling. When I had it my car wouldn’t sleep.
I have... This'll be my second winter (took delivery in 9/2018). There was no range drop last winter; this seems to be an new thing (see the earlier dates in the attachment below).
For sure there is some temperature correlation. I think a lot of us are confused because the estimated range *should* be a constant multiplier. It's not supposed to change with temperature, unless Tesla silently changed something (which seems to be corroborated by a few plots people have done on this forum).
@morepizza - your estimated range plot looks a LOT like mine (posted on p5 of this thread). Although I have the mid range which did not get the range bump... so we were both getting “rated range” until ~June 2019 when it drop in several steps. Both of us are down about 20mi from “rated”.
I live in LA and you Live in CO, so we both have drastically different winters. This is looking more and more like Tesla has changed (either by mistake or on purpose) how the range display works.
I live in LA and you Live in CO, so we both have drastically different winters. This is looking more and more like Tesla has changed (either by mistake or on purpose) how the range display works.
Not saying you’re wrong, as this makes sense, but my car’s rated range dropped overnight from ~317 to 304, which is double what you’re suggesting
Model: LR Dual motor
Build Date: 10/19
Current Miles: 1.5k miles
Current SOC to Miles: 90% SOC = 279
Location: northern NJ
Software/ Notes:
2019.36.2.1 - Noticed the range difference
Average Commute: 20-40/miles a day
Typical Charge Patterns: 90% SOC 2x per week
The degradation started after the last firmware update. It’s worsened as it got colder.
12 mile drive from a 90% SOC (279) gets the battery down to 257. Almost 50% disappeared ?
Build Date: 10/19
Current Miles: 1.5k miles
Current SOC to Miles: 90% SOC = 279
Location: northern NJ
Software/ Notes:
2019.36.2.1 - Noticed the range difference
Average Commute: 20-40/miles a day
Typical Charge Patterns: 90% SOC 2x per week
The degradation started after the last firmware update. It’s worsened as it got colder.
12 mile drive from a 90% SOC (279) gets the battery down to 257. Almost 50% disappeared ?
Similar threads
