How I Recovered Half of my Battery's Lost Capacity

[AlanSubie4Life]

Would the routines @AlanSubie4Life recommends

I’m not sure exactly what you are referring to here. I don’t think there is a lot that you can do to recover capacity, though it is possible to nudge the estimate a bit, some of the time.

 

[dmurphy]

I’m not sure exactly what you are referring to here. I don’t think there is a lot that you can do to recover capacity, though it is possible to nudge the estimate a bit, some of the time.

Right, that's what I meant. Some of the monitoring and "battery hygiene" you've recommended over the years. Not recovering capacity but the general charging recommendations ... I look at you as the expert in understanding the ranging, charging and capacity algorithms. ;-)

 

[KenC]

Another check in and data point for me. As I’ve mentioned in prior posts I charge to 90% because I just want max power lol. 8/2018 build, performance, currently 46k miles..had a dip down which always seems to happen after an update…but range at 100% has gone back up to 282–285. I’m OG Stats user so I had alot of variability in the data points due to him using temperature adjust range. I’m not sure when he switched it. I actually stopped opening the app to capture data since it was so buggy. Only start using it again when I notice a dip in range. That’s why there’s less data points at around 23k and 40k lol.


View attachment 765928

There are two SOC apis that an app can use. Up until January 2021, the developer used the one that was temperature sensitive, so the Battery Health chart showed seasonal dips during wintertime. After January 2021, the developer used the other SOC api, the one that is not temperature sensitive, so there's no more wintertime dip. You can clearly see the temperature dip when there's a blue snowflake day. Before Jan 2021, you'd get dots way below normal. Now, you shouldn't get any dots due to cold temps. As you can see, my chart is very horizontal since Jan 2021, without the temperature variation. Whatever drops I get now, will be calendar aging and usage aging.

 

[AlanSubie4Life]

so while my car came with 310 miles of range, and with its current range of 309,

Insanity! 30k miles, coming up on four years old...minimal capacity loss (less than 3%). Do you keep it in a frozen bat cave under the Maine taiga?

 

[KenC]

Insanity! 30k miles, coming up on four years old...minimal capacity loss (less than 3%). Do you keep it in a frozen bat cave under the Maine taiga?

30,300 miles, 40 months old, and 0.3% estimated range drop, and yes, it's in the Bat cave, aka my carport. Remember, this is the car delivered to me with Zero SOC.

 

[AlanSubie4Life]

and 0.3% estimated range drop

I said less than 3% since you’re a hair under 76kWh and you may have started with as high as 78kWh (so no way to know what has actually been lost). But definitely basically zero visible range loss. Certainly close enough to zero as compared to nearly any other vehicle of that vintage!

 

[Candleflame]

30,300 miles, 40 months old, and 0.3% estimated range drop, and yes, it's in the Bat cave, aka my carport. Remember, this is the car delivered to me with Zero SOC.
View attachment 766222

ahh i remember you. i wonder if its the cold weather.
0% isnt harmful afaik for the battery anyway, seems to be a myth tesla has perpetuated so people dont run out of range. Seems the best way to protect the battery is to discharge it to 0.01% and store it that way. Technically (though obviously not feasable).

 

[KenC]

ahh i remember you. i wonder if its the cold weather.
0% isnt harmful afaik for the battery anyway, seems to be a myth tesla has perpetuated so people dont run out of range. Seems the best way to protect the battery is to discharge it to 0.01% and store it that way. Technically (though obviously not feasable).

I wondered about whether cold weather slowed deg, too; so I did a poll a long time ago, asking people with low deg, where they lived, and there did not seem to be any discernible pattern. Of course, it's a small sample size, but you would think if cold weather slowed deg, that perhaps we'd have heard something more substantial from people in northern climes, like Canada, Norway, etc., and I haven't heard anything. Has anyone else?

 

[scottf200]

Anyone else seeing 2021.44.* Firmware Version ranges raise mileage estimates? Charging methods and weekday range of 80 to 30 to 80 ... are the same.
Top graph below is from the Google sheet chart and the lower corresponding (purple line focus cone) is the corresponding data in the TeslaFI battery chart.
In below top chart, the [Col C sheet] or "FV +200" chart legend and blue vert lines in top chart values would be a value of
244 and equate to 2021.44.XX.Y (200+44)
232 and equate to 2021.32.XX.Y (200+32)
with the reason to quickly put it on the same scale as the mileage range.

 

[AAKEE]

I wondered about whether cold weather slowed deg, too; so I did a poll a long time ago, asking people with low deg, where they lived, and there did not seem to be any discernible pattern. Of course, it's a small sample size, but you would think if cold weather slowed deg, that perhaps we'd have heard something more substantial from people in northern climes, like Canada, Norway, etc., and I haven't heard anything. Has anyone else?

Cold weather (keeping the battery cold during storage) would slow the degradation as lower temperature lowers the rate of calendar aging.
Hot WX should increase the calendar aging.

Talking/writing about degradation in model 3 in a Swedish facebook group I was jumped on hard stating that there is cases with model 3 and range loss/appearent degradation of more than 10%.
In that thread there wasnt really anyobe close to 10% range loss.
I would guess that deeper search into this will show a clear connection between the climate and the degradadation.
Also, I would expect my very low degradation be higher if I didnt live close to the artic circle.

 

[Dave EV]

Talking/writing about degradation in model 3 in a Swedish facebook group I was jumped on hard stating that there is cases with model 3 and range loss/appearent degradation of more than 10%.
In that thread there wasnt really anyobe close to 10% range loss.
I would guess that deeper search into this will show a clear connection between the climate and the degradadation.
Also, I would expect my very low degradation be higher if I didnt live close to the artic circle.

Yep - calendar life losses can be approximated using Arrhenius' equation - every 10C rise in temperature the rate of capacity loss doubles.

 

[KenC]

Cold weather (keeping the battery cold during storage) would slow the degradation as lower temperature lowers the rate of calendar aging.
Hot WX should increase the calendar aging.

Talking/writing about degradation in model 3 in a Swedish facebook group I was jumped on hard stating that there is cases with model 3 and range loss/appearent degradation of more than 10%.
In that thread there wasnt really anyobe close to 10% range loss.
I would guess that deeper search into this will show a clear connection between the climate and the degradadation.
Also, I would expect my very low degradation be higher if I didnt live close to the artic circle.

I did ask the Stats developer about this, since he would have tons of data that could be separated by latitude. Of course, it took 2yrs of lobbying before he changed the SOC api he used, so I'm not optimistic that he'll do it any time soon.

 

[Candleflame]

heats kills batteries. I leave my car at 50% all the time and the battery sits at 30C or more all year around and I have 16% degradation.
Someone else who lived with me in the tropics sits at 90% all the time and he has 21% degradation. I have actually now moved away from the tropics and its considerably cooler now.

Im sure it is the heat day and night.

 

[PianoAl]

Why isn't it enough to, say, do one OCV measurement at 20% and one at 95%?

 

[Zoomit]

Why isn't it enough to, say, do one OCV measurement at 20% and one at 95%?

My perspective:

1. The BMS probably uses a Kalman filter that needs multiple calibration inputs to affect the battery level prediction algorithm.
2. The discharge voltage curve for a Li-ion battery is non-linear so inputs at various voltages, especially at the high/low extremes, are the most effective calibration inputs.

 

[Candleflame]

the BMS definetly recalibrates in chunks. thats why youtube tests where people discharge to 0% or less and charge to 90% or 100% are dumb.
the bms doesnt work like that. What happens is that one day you wake up and suddenly the car has 10km less range (or indeed 15km more range).

What I would love to know what the voltage readings at 5%/20%/50% are for people who experience a recalibration with restoration of i.e. 20km of range.
Did they always have the extra range available when looking at the voltage?

 

[afty]

I ended up doing the battery recalibration recommended in this thread as a side effect of a road trip last week. Now I am seeing a full 310 miles of range on my 3 year old Long Range AWD with 22k miles. Before the trip I believe it was reporting around 295 miles.

 

[AlanSubie4Life]

I ended up doing the battery recalibration recommended in this thread as a side effect of a road trip last week. Now I am seeing a full 310 miles of range on my 3 year old Long Range AWD with 22k miles. Before the trip I believe it was reporting around 295 miles.

Let us know how it looks when it is charged to 70-80%. Most likely this is extrapolation error since you are extrapolating from 10%. Though some recovery of rated range is of course possible. 295 is great for that vehicle, in any case.

 

[Sumguy]

Wow that's a pain to attach a picture using a mobile device.

Screenshot from scanmytesla.

I've been watching my SOC (%state of charge) vs. "SoC expected" lately. If I keep the car awake, the numbers track together as the battery depletes. If I let the car do a long 3hr+ deep sleep the numbers spread apart by a percent that would target my excess range loss. Makes me think that SOC is looking at the sleep cell voltages to figure out where it actually is in capacity of the battery, vs where SOC expected is guessing it is at.

Since my SOC is higher than the expected value it gives peace of mind that I still have battery capacity hidden from me for that 1 trip a year where I actually try to use >95% of battery capacity . Also I will see range estimate recovery over a long period of time.

Each 90->20٪ charge cycle i do it seems like the CAC (amp hour capacity) climbs by 1 amphour. I don't see large recalculation of capacity, just a slow recovery of guesstimate range.

 

[Twiglett]

I have a question for the knowledgable folks here.
I've noticed that when I come back from a longer drive, the reported %age and range "recovers" an hour or so later after the drive.
This last 50 mile drive came back with 37% reported, then 30-40 minutes later its reporting 42%

Max range has been hovering between 260-270 miles (2018 LR) with very consistent driving patterns so guessing BMS is really having a hard time figuring things out.