3.0 Battery Longevity

[Marius]

In this thread most discussions are about the CAC value. However if I read the info from David Tesla has altered something in the software bringing the CAC value back to 215.
I have an original pack that did 200.000 miles and is now on CAC 124. However I have noticed that probably because of aging that after a drive I park my car for instance with an ideal mile range of 100 miles and 1 hour later this drops to 80 miles and 1 day later comes back to 95 miles. As far as I know the software calculates the ideal range based on the voltage of the battery cells. The CAC value is amp-hour and is a calculation based on parameters that I do not know.
What would be also interested to know is if the ideal range drops in the same way as the CAC value. New 3.0 pack seems to have 340 miles and CAC 215 which means CAC 190 should show an ideal range of 300 mile. Is this indeed the maximum range now?
If the ideal mile range is significantly higher, than we maybe can conclude that the algoritme of calculating CAC value has some faults and that the pack is much healthier than we think now based on CAC value.
So before I consider to purchase the 3.0 pack, my question is if the ideal range is going down with the same speed as the CAC value?
The ideal range determines the usability of the pack for distance driving and the CAC value is just a parameter that is calculated by the software.

 

[slcasner]

Standard mode? Standard mode hides the bottom 12%-ish of the battery capacity from the ideal miles calculation, so 8% SoC should already be underwater in standard mode, no? Those numbers are about right for Range mode.

When I say 8% I mean my visual interpolation of the fill level indicated by the battery display symbol on the VDS when the standard mode is selected. Similarly, the 22 ideal miles indicates how much more range is available in standard mode. For these long trips I charge in range mode but I always switch to standard mode while driving. That means I start out with the display pegged at more than 100%, but that's not a problem.

 

[bolosky]

What would be also interested to know is if the ideal range drops in the same way as the CAC value. New 3.0 pack seems to have 340 miles and CAC 215 which means CAC 190 should show an ideal range of 300 mile. Is this indeed the maximum range now?

Yes. My CAC is at 192 now, and the last time I range charged it (last week) it was at something like 308-310 at full charge.

The real question in my mind is whether the decline will continue at the same rate, or if it will level off. Continuing at the same rate (of roughly one milli Amp-Hour per mile and 25 mAh/day) means that the pack will wear out pretty quickly. If, on the other hand, it sticks at 90% of initial capacity for a long time, or at least descends much more slowly, then that's OK. My car is hinting that it may be at a plateau, but it's still much too early to say for sure. And I'd also like to see if other cars do the same thing, which will take a few months at least since I'm driving mine much more than anyone else in the data set.

 

[ecarfan]

The real question in my mind is whether the decline will continue at the same rate, or if it will level off. Continuing at the same rate (of roughly one milli Amp-Hour per mile and 25 mAh/day) means that the pack will wear out pretty quickly. If, on the other hand, it sticks at 90% of initial capacity for a long time, or at least descends much more slowly, then that's OK. My car is hinting that it may be at a plateau, but it's still much too early to say for sure. And I'd also like to see if other cars do the same thing, which will take a few months at least since I'm driving mine much more than anyone else in the data set.

Yes that is the big question here: are some of the 3.0 cars getting close to the state where their capacity loss starts to plateau or not? Honestly, I'm optimistic that the answer to that questions is "yes". But we likely won't know for another year or so. Again, kudos to @bolosky for taking the time to process the submitted data and provide us with clear visuals on how our new batteries are doing!

 

[dhrivnak]

When I first got my car the range charge was 351 and today is 343. But I guess that is still better than any other car out there.

 

[hurd300403]

When I first got my car the range charge was 351 and today is 343. But I guess that is still better than any other car out there.

Mine was the same but it quickly plummeted the more I’ve driven it. Still happy as a clam, but hoping it plateaus soon!

 

[ecarfan]

When I first got my car the range charge was 351 and today is 343. But I guess that is still better than any other car out there.

That’s pretty damn good. Did you send your data to @bolosky to incorporate into the 3.0 battery graphs he has posted?

 

[dhrivnak]

That’s pretty damn good. Did you send your data to @bolosky to incorporate into the 3.0 battery graphs he has posted?

Yes sir! And I plan to send an update about every 6 months.

 

[bolosky]

Adding new data for #660, very low mileage car. It saw several big steps down in CAC in this data, including a drop from 210.4 to 206.9 on 9/18 after a ~250 mile drive and then a second one from 204.7 to 202.4 after driving less than 20 miles. Combining the two, it's much less of an outlier than it had been, no longer being way above the others on the per day chart. It also continues the trend of very low mileage cars having these big steps downward occasionally. My guess is that little driving means little variation in SOC means there's little data to estimate CAC, so the system waits and then when it figures something out it has to make a big correction. I assume that the cells themselves are much more smooth, so the reported range is probably bigger than what you actually have.

I'm not updating my car (#670), since it's only been a couple of weeks, but I've driven another 600 miles and the CAC is in the 191 range, so it's still on the plateau that started around 19K miles.

Thanks, hurd3004003!

 

[ecarfan]

Looking at the graph for my data (425) and thinking about how in the past 3 months I have made multiple 200 mile trips which I range charge for (probably about 7 range charges during that time) I’m thinking that the range charging is what has caused my CAC to rise.

To me a big question with the 3.0 battery is, are the cells significantly more tolerant (experience less degradation) of moderately frequent range charging than the original cells? My guess is “yes”. But we are years away from knowing the answer to that question.

 

[JRP3]

Could also be software and/or BMS related.

 

[bolosky]

Looking at the graph for my data (425) and thinking about how in the past 3 months I have made multiple 200 mile trips which I range charge for (probably about 7 range charges during that time) I’m thinking that the range charging is what has caused my CAC to rise.

To me a big question with the 3.0 battery is, are the cells significantly more tolerant (experience less degradation) of moderately frequent range charging than the original cells? My guess is “yes”. But we are years away from knowing the answer to that question.

I think that range charging often increases CAC. I know that it pretty much always does with my car. That said, it's important to remember the difference between CAC and actual battery capacity. CAC is the BMS's estimate of the true capacity. My suspicion is that while range charging increases this estimate, it does not increase the true capacity. It's just that the BMS is under estimating how much charge it can get in to the battery. Likewise, deep discharges seem to reduce CAC. While they may well disproportionately reduce true capacity, again I think that most of the effect we're seeing is the BMS estimate rather than cells themselves.

To find out whether range charging effects real capacity, we need to have lots of data over a long time where we can compare battery capacity to the number/frequency of range charges. I don't think that we're really anywhere close to that yet, there's still way too much noise in the signal, and too few range charges in the data set to tell. I do it more frequently than others, and it's still under 5% of the time.

My guess (and it's only a guess) is that whatever effect there is isn't all that big given the relative infrequency of range charges, so if you think you'll need it, use it. One range charge every once in a while most likely doesn't matter that much.

Now, if we really want to find out, the best test would be to get a new 3.0 and just always range charge it. Anyone want to blow $30K to help improve the state of our data set?

 

[ecarfan]

My suspicion is that while range charging increases this estimate, it does not increase the true capacity. It's just that the BMS is under estimating how much charge it can get in to the battery. Likewise, deep discharges seem to reduce CAC. While they may well disproportionately reduce true capacity, again I think that most of the effect we're seeing is the BMS estimate rather than cells themselves.

Agreed, and I did not mean to imply in my post that I thought that range charging was — at this point in my 3.0 battery’s life — effecting the true capacity of my battery.

To find out whether range charging effects real capacity, we need to have lots of data over a long time where we can compare battery capacity to the number/frequency of range charges. I don't think that we're really anywhere close to that yet

Agree again, which is why I said that we are years away from knowing whether that is true or not.

One range charge every once in a while most likely doesn't matter that much.

If we only knew what “every once in a while” really meant...

 

[dpeilow]

I'm wondering if there is a part of the algorithm that is left over from the pre-3.0 days that thinks if the estimated miles is above 190, it must be a range mode charge and to drop the CAC accordingly.

Then when you do a real range mode charge now, it gets top balancing information and overrules this, hence raising the CAC slightly.

 

[simonog]

@bolosky : do you want another data set? I fitted the 3.0 battery in March in a 2.x car. The car has probably only done 400-500 miles since then though I use it most days.

 

[bolosky]

@bolosky : do you want another data set? I fitted the 3.0 battery in March in a 2.x car. The car has probably only done 400-500 miles since then though I use it most days.

Yes! The more data the better, regardless of the duration or mileage. I'll send you a PM with instructions.

Thanks!

 

[bolosky]

Adding our 11th car, #209. It looks pretty middle-of-the pack. It's got one big jump down without a long drive around 2100 miles, but we've seen lots of those now so they're not especially surprising. Usually, they seem to happen on the very low mileage cars, which this isn't, but it's not too out of the ordinary.

I've got the happy problem of running out of really distinct colors. I did the best that I could, her red is a little darker than mine.

Thanks, msjulie!

 

[bolosky]

Adding #541 (EU). It's a pretty typical low-mileage car: Looking better than average on the by-days graph and worse on the by-miles, which is just what you'd expect if there's a combined effect from each.

I also updated mine (which is only a couple of weeks' data, but why not?). Things are still holding around a CAC of 191, which has been true now for about two months and ~3K miles. So far, it's still consistent with a significant flattening at about 10% loss, but still hardly conclusive. One interesting thing is that I ran the car down to ~20 ideal miles (in standard mode), and didn't see a drop in CAC like I previously had. I'm not sure whether this is because I didn't range charge it first, or because I did it over several days rather than all at once (bad weather caused me to park the car at the bottom of the hill and not recharge), or because it's at the flat point in the curve.

Thanks simonog!

 

[ChargeIt]

We picked up our Roadster today from the SC with the new battery. Forwarded the data to @bolosky, our pack is starting with a CAC of 214.86Ah.

We were very happy with the process, originally it was going to take until mid January to get the pack in from CA, but another battery became available first week of December. Only took the techs 3 days to swap in the new battery.

 

[bolosky]

It's been a while since the last update, but Dave Denhart (#181) sent me his logs, and I'd been sitting on an update from dhrivnak (#255) since late November, so I decided it was time. I also updated mine (#670).

Probably the biggest news here is that the plateau that I was hoping to see at CAC ~191 seems not to be there. Both #670 and #181 have broken though that, and continued their steady decline. Dave's battery is now about midway between a new 3.0 battery and a new original battery in about a year and a half and 16K miles. Mine's pretty close with a little less time and ~9K more miles. Though suspect that if Dave range charged a number of times that his CAC would go up by quite a bit.