Tesla Model S Survey by Plug In America

[Andyw2100]

Thanks for the note, and the report!

Sure!

Looking at what you've done a little more, I'm wondering if you've considered putting together anything to attempt to show an analysis of pack degradation. You have so much data collected I expect you might be able to see some meaningful trends. It might be as simple as creating a graph or something for each different pack type and car model, showing each entry of mileage and 90% and 100% charge. So I guess there would be two graphs for each unique battery / model combination--one showing 90% charge and mileage and the other showing 100% charge and mileage. Or perhaps the 100% and 90% graphs could be combined. And I guess time perhaps should be worked in somehow. This is not my area of expertise at all, obviously. It just seems like you have a lot of data on this that would be interesting to see.

Just an idea, anyway.

Thanks for having started this!

 

[tomsax]

Looking at what you've done a little more, I'm wondering if you've considered putting together anything to attempt to show an analysis of pack degradation.

Yes, definitely. See my post on November 13th letting people know getting new and updated reports would help. I plan to write up something similar to the Roadster report from 2013. Check that out for a preview of what the Model S report will cover.

 

[brianman]

I see 8 out of 141 entries that report values over 135. I'll bet you're right about that being amps instead of kW. I'll either follow with those owners or exclude those values from any analysis I do in the future.

Are any of the values over 120 kW legitimate?

I thought 135 kW was the latest limit for pedestal pairs, not for individual cars.

 

[William13]

Although I participate in both True Delta and Plug in America data collection they are inherently skewed. The participants self select to participate and this prevents the ability to predict the entire population accurately. People who have problems may be much more likely to volunteer their data.

The only way to determine the scope of the problem is to search through the financials to get a handle on the costs of warranty work. This is also much better than saying the DU "failed" and not saying the same thing if an ICE has any of a multitude of problems (starter, plugs, wiring harness, congealing oil, PCV, oxygen sensor, etc.). In a Tesla there is just a drive unit changed en mass.

I'm not saying there is no utility to the surveys but to generalize or extrapolate even with"good math" as Tom attests is unfortunately a case of garbage in, garbage out... The survey may reflect the larger population but it may not.

The early 2012 sample may be large enough to extrapolate from as you have more likely random participants (early adopters ) who are less likely to report only if problems. (My belief.)

It does make awareness.

 

[dhanson865]

PlugInAmerica has been around longer and is more well known.

Truedelta 2004 and only dedicated to reliability data and pricing for cars

Plug In America 2005 or later, when did they start doing any sort of reliability surveys? Maybe in 2012 for the Leaf? I'm not sure what their focus is but it seems to include much outside the focus of reliability data.


You might argue that PIA is more well known in the Tesla community but I'm not familiar with them and I've been dealing with Truedelta since the get go. From TTAC to priuschat to mynissanleaf and other car sites I've seen truedelta as the only open data for many brands of cars.

 

[tomsax]

Are any of the values over 120 kW legitimate?

I thought 135 kW was the latest limit for pedestal pairs, not for individual cars.

That's a good question. There have been rumors of people charging above 120 kW for quite a while. There are four reports above 120 kW on the survey from people claiming to be Go4IT, jdw, and Lloyd on TMC.

Can anyone confirm charging above 120 kW? A screenshot would be nice.

 

[tomsax]

Although I participate in both True Delta and Plug in America data collection they are inherently skewed. The participants self select to participate and this prevents the ability to predict the entire population accurately. People who have problems may be much more likely to volunteer their data.

There is definitely the risk of selection bias, but it's far from certain. In fact, there's pretty good reason to think the Plug In America surveys are not strongly affected.

Plug In America got involved in surveys and battery studies when we started hearing reports from Leaf owners in hot climates were losing capacity at a much higher rate than those in milder climates. That situation was quite severe and there was concern regarding the potential for owners with problems to be more likely to participate in the survey. When I wrote up the results, I sent a preview copy to Nissan for their comments before making it public. When they announced adding a capacity warranty to all Leafs a few weeks later, they linked to the Plug In America study report and recommended that all Leaf owners read it. So, clearly they didn't think the survey data suffered from selection bias over representing battery capacity loss.

For the Roadster battery study, I was able to pull data from two sources: the Plug In America survey and anonymous data collected via OVMS. Presumably, the owners who have installed OVMS on their Roadsters would not have the same selection bias, if any, as the self-selected survey participants. Those two data sets yielded very similar results for battery capacity loss vs. miles driven. That helps bolster confidence that the Roadster survey didn't suffer from any significant selection bias.

 

[wraithnot]

That's a good question. There have been rumors of people charging above 120 kW for quite a while. There are four reports above 120 kW on the survey from people claiming to be Go4IT, jdw, and Lloyd on TMC.

Can anyone confirm charging above 120 kW? A screenshot would be nice.

This was taken December 20th, 2013 at Tejon Ranch. 346 A x 358 V = 124 kW. Tesla may have backed off on the power a bit because I've haven't gotten above 120 kW on more recent trips.

 

[William13]

Tomsax, the Nissan data was likely accurate. The Roadster data does not have selection bias. However the issue is a much bigger and more random event namely "drivetrain failure". This is not a slow, inevitable progression. Selection bias is much more likely with this sort of issue than anxiety about battery capacity loss.

I'm still not saying the data isn't useful or accurate but that it is not as reliably accurate.

I appreciate all the work you do.

 

[tomsax]

Tomsax, the Nissan data was likely accurate. The Roadster data does not have selection bias. However the issue is a much bigger and more random event namely "drivetrain failure". This is not a slow, inevitable progression. Selection bias is much more likely with this sort of issue than anxiety about battery capacity loss.

I don't mean to belabor this point, but I disagree.

Unlike the Roadster and Model S, the Leaf doesn't have instrumentation to show battery capacity in an energy unit (rated/ideal miles). Instead, it has a range estimate based on recent driving, which fluctuates with driving, traffic and weather making it pretty useless for knowing anything about pack capacity. For capacity feedback, it has 12 bars with the first bar counting 15% and the others 7.5% each. So, it's quite an event when that first capacity bar disappears. Especially before Nissan added a capacity warranty, losing that first bar was a big and seemingly random event about which there was nothing that could be done no matter how quickly it happened, even within the first 10,000 miles.

For the Roadster, the failure rate of the power electronics module (PEM) was in the same ballpark as the Model S drive unit failure. If I combine that with motor replacements to be similar to the Model S drive unit, the survey shows 17 of 73, 23%, 2010 Roadster had the motor and/or PEM replaced. That's very similar to the 22.6% of 2013 Model S drive unit replacements.

 

[tomsax]

Live Charts Now Available

I just added a page of live charts based on the current data set to the Model S survey site.

Also, from the individual vehicle pages, there's a link to view a chart of that vehicle's reports overlaid on the most current reports for all of the other vehicles in the survey.

Some interesting vehicle report charts:

Vehicle 2
Vehicle 39
Vehicle 42
Vehicle 197

 

[KJD]

Thank you for the charts Tom, excellent work as always.

The Tesla battery packs seem to hold up much better than the LEAF battery packs. I guess in this case you get what you pay for.

 

[Andyw2100]

This is not a complaint, but am I correct that as a P85D owner, trying to compare my battery to the graph is pretty useless because the D's rated range is lower than other 85's rated ranges? I did, in fact already make the comparison, and my battery falls well below the average, but I believe that's because an average 85 started with more rated range than a P85D did. That's correct, right?

If so, any chance of breaking out the P85Ds from the other 85 batteries? That would actually provide more meaningful data for both P85D owners and the owners of the other 85s.

 

[tomsax]

I could create a separate group for each model, but there are a lot of them: 40, 60, 70D, 85, P85, 85D, P85D, 90D, P90D. Is that list right? There's no 60D, 70 or 90? The 85 and P85 seem to have the same EPA rating, but that's still 8 different groups. The 85D and 90D have the same EPA rating (270), but it seems weird to group them together. Likewise for the P85D and P90D (253).

Another way to do it would be to show all of the vehicles divided by their EPA range, so showing a capacity percent instead of rated range.

Opinions?

 

[Andyw2100]

Another way to do it would be to show all of the vehicles divided by their EPA range, so showing a capacity percent instead of rated range.

Opinions?

The percentage of capacity might be a good solution, in that it should cover all the models. The current graphs are potentially misleading for people who don't realize that matching pack sizes don't necessarily have matching rated mileage capacities.

 

[DCGOO]

You don't even allow the non-P, 90D (which I believe is very common these days)

 

[tomsax]

You don't even allow the non-P, 90D (which I believe is very common these days)

That was a bug. It's fixed.

Thanks for letting me know.

I also added the P85DL and P90DL.

If anyone filled out the survey and your model wasn't available, let me know and I'll fix it for you.

 

[tomsax]

Charts Update

There's a new version of the charts pages available.

The pack capacity chart is now broken up into finer granularity so the dual drive and performance variants are in different groups. Not all of the groups have enough data to make a trend line meaningful, so only some of the groups have them. On the full charts page there's a new table that shows the vehicle counts by all of the various models.

Full Survey Charts

On the battery capacity chart, hover your mouse over a group name to highlight those vehicles.

On the individual vehicle pages, there's a link to view a chart of that vehicle's reports overlaid on the most current reports for all of the other vehicles in the survey.

Some interesting vehicle report charts:

Vehicle 2
Vehicle 39
Vehicle 42
Vehicle 197

 

[Pilot_51]

I could create a separate group for each model, but there are a lot of them: 40, 60, 70D, 85, P85, 85D, P85D, 90D, P90D. Is that list right? There's no 60D, 70 or 90?

There is a 70 currently available in the Design Studio.

 

[dhanson865]

I could create a separate group for each model, but there are a lot of them: 40, 60, 70D, 85, P85, 85D, P85D, 90D, P90D. Is that list right? There's no 60D, 70 or 90? The 85 and P85 seem to have the same EPA rating, but that's still 8 different groups. The 85D and 90D have the same EPA rating (270), but it seems weird to group them together. Likewise for the P85D and P90D (253).

Another way to do it would be to show all of the vehicles divided by their EPA range, so showing a capacity percent instead of rated range.

Opinions?

There is a 70 without the D, there is not a 60D or 90 without the D.