Kevin Sharpe's decreased Roadster range

[ecarfan]

Remember that during the whole period from new until 20th August 2014 (when we turned it off for legal reasons) Tesla had remote access to my car. They used this access to identify the battery fault in November 2011.

How does Tesla have "remote access" to a Roadster? It does not have the capability to connect to a cell data signal like the Model S. Was your car modified in some way? I've been trying to following this discussion from the beginning but I missed that part.

I am hoping to become a Roadster owner soon. This thread has served a valuable purpose: it reinforces the importance of the warning message displayed in the Roadster whenever the owner does a Range charge. While it has not been conclusively proven that the degradation Kevin's battery suffered was due to frequent Range charging, it is certainly quite possible that was the cause. His battery is on the low end of the capacities documented in the PIA data set but it is not outside the boundaries of that data set.

This dispute with an EV manufacturer over what amount of battery capacity loss qualifies for a warranty repair reminds me of trying to convince an ICE manufacturer to repair one of their engines under warranty because a cylinder or cylinders has suffered a loss of compression that the vehicle owner thinks is significant but the manufacturer does not, or a dispute over whether a brake rotor is warped or not, or whether a transmission makes excessive noise or not. To the vehicle owner their complaint is valid and the manufacturer is being intransigent, while the manufacturer believes that the warranty does not cover the claimed defect. The two parties interpret the warranty language differently and arrive at different conclusions after examining the vehicle parts in question.

 

[djp]

How does Tesla have "remote access" to a Roadster? It does not have the capability to connect to a cell data signal like the Model S.

The Roadster has a GSM unit which is used to send log files back to Tesla. It does have cell data for the VDS, but not for navigation or streaming music like the Model S.

 

[Krugerrand]

Remember that during the whole period from new until 20th August 2014 (when we turned it off for legal reasons) Tesla had remote access to my car. They used this access to identify the battery fault in November 2011. At no time did Tesla suggest that my ownership behaviour (charging, driving, storage, temperature, etc) was having any detrimental impact on my battery.

It is not Tesla's job to babysit your car usage. Seriously. Be a big boy and own up to your abusive charging and driving habits that have negatively affected the battery of your Roadster.

We don't expect GM/Ford/BMW/fill in the blank or any of their dealerships/franchises/representatives/fill in the blank to tell us that driving/maintaining our cars outside of the manual instructions is going to have a negative impact on the engine/brakes/transmission/fill in the blank. That's a given. Even if we've got it in for a repair and they know we've abused the vehicle.

It makes no sense for Tesla to be repair proactive in November 2011 and not all this time since. It's not how they historically do things. If they truly thought there was a problem at their end with the battery degradation of your Roadster, they'd have done everything in their power to rectify it (and probably just to shut up you up).

 

[Krugerrand]

This dispute with an EV manufacturer over what amount of battery capacity loss qualifies for a warranty repair reminds me of trying to convince an ICE manufacturer to repair one of their engines under warranty because a cylinder or cylinders has suffered a loss of compression that the vehicle owner thinks is significant but the manufacturer does not, or a dispute over whether a brake rotor is warped or not, or whether a transmission makes excessive noise or not. To the vehicle owner their complaint is valid and the manufacturer is being intransigent, while the manufacturer believes that the warranty does not cover the claimed defect. The two parties interpret the warranty language differently and arrive at different conclusions after examining the vehicle parts in question.

In any situation where there's a grey area, something written in a document isn't spelled out for a two year old to understand, one party is using a different dictionary for basic word definitions, or circumstances have changed as time passed such that the verbiage doesn't cover every new, possible scenario....there's still a right way and a wrong way to handle it.

Clearly if both parties will only be responsible (for their part) and reasonable, then a satisfactory conclusion can be arrived at with little to no blood shed, most times quite happily. However, the person/entity trying to get something replaced/fixed/repaired/fill in the blank is advised to take on the responsible and reasonable role first, particularly if they wish to increase their chances of resolution to their satisfaction. They set the tone for the negotiations and control them moving forward. And know that compromise is as good a friend as Google.

I have read Mr. Sharpe's postings on this forum for quite some time now and I'm always left with the same perception of him. I can well imagine what goes through the minds of the staff at the Tesla center that he frequents. We should send them some alcohol.

 

[abasile]

If I'm understanding correctly that the cell voltage following a Roadster "range" charge is 4.15 V, and assuming that Kevin started driving his car within a few hours of completion of each range charge, then I doubt that the ~96 range charges by themselves could explain the bulk of the capacity loss. To protect the battery, I'm assuming the Roadster's BMS is smart enough to taper the charge current as full charge is approached. As long as a Li-ion battery doesn't remain at full charge for too long, and it is not excessively hot while at full charge, the additional capacity loss generally shouldn't be very significant.

I'm more inclined to believe that deep discharges (below ~5%, or below ~10-15% under a heavy load) could be the culprit here. This is an easy way to accelerate capacity loss, and it could take significantly fewer than 96 deep discharges to accomplish this.

 

[tomsax]

A few notes on some recent posts...

Looking at the logs from my recent Range mode charges, the maximum brick voltage recorded in the logs is 4.18V.

Looking at the chart on this post, charging to 4.2V would lead to 85% of original capacity after 400 pack cycles. Assuming the pack started at 244 ideal miles and dropped linearly to 85% after 400 cycles, that would be 90,280 ideal miles. I would expect that to yield between 65,000 and 90,000 actual miles depending on driving style.

Kevin's pack is at 55,000 miles and 81% of original capacity.

 

[Doug_G]

The Roadster has a GSM unit which is used to send log files back to Tesla. It does have cell data for the VDS, but not for navigation or streaming music like the Model S.

Actually that is true for 2.0 and 2.5, but not 1.5 Roadsters.

 

[Chickenlittle]

I grow weary of this thread but will make the following observations
1. 2or 3 range charges have now become less than 2 a month. Won't quibble about the number still frequent
2. Tesla never told me it was an issue? As others have noted you had more than 90 written warnings and acknowledged each by clicking
3. Despite noticing your range decreasing 2 years ago (see first post in this thread), you persisted in this behavior
4. You seem to imply since the decrease in range was not noticed earlier that it couldn't have been this yet the warning doesn't state that it would be linear decrease in range
5. Fault the battery warrantee yet the warranties usually only apply if properly maintained. Not performing oil changes on an ICE car will void an engine warrantee as an example. You were warned each time you range charged. Also warned in manual
6. Even other companies warrantee wouldn't have covered your loss, didn't meet the level of loss
7. Wish you the best but attempting to "sharp" ( use bad publicity to get what you want) only comes back to haunt you. Now in writing your own admissions of frequent range charging.

it was all avoidable since you could have charged at your company destination with some planning.
good bye

 

[stopcrazypp]

A few notes on some recent posts...

Looking at the logs from my recent Range mode charges, the maximum brick voltage recorded in the logs is 4.18V.

Looking at the chart on this post, charging to 4.2V would lead to 85% of original capacity after 400 pack cycles. Assuming the pack started at 244 ideal miles and dropped linearly to 85% after 400 cycles, that would be 90,280 ideal miles. I would expect that to yield between 65,000 and 90,000 actual miles depending on driving style.

Kevin's pack is at 55,000 miles and 81% of original capacity.

I actually did the math a few days ago with cells that more closely resembles the Roadster (same capacity, the cells in your link are 950mAh so are completely different):

Roadster (most likely using Sanyo UR18650F 2400mAh, datasheet I linked is 2500mAh but can expect similar degradation):
72% (1800mAh/2500mAh) after 500 cycles or 105,350 ideal miles (245 ideal miles max, 86% linear approximation cycle multiplier)
https://www.pollin.de/shop/downloads/D271248D.PDF
...
For the Roadster's cells you would expect 81% at 339 cycles or 75,228 ideal miles, using a linear approximation. I looked at previous posts and Kevin said he gets about 150-160 miles of range in the real world. Using 160 miles works out to 81% at 49,128 real world miles.

I know it is 2400mAh cells because Martin Eberhard said so, and that Tesla used Sanyo cells in the Roadster because it was disclosed in SEC filings.

 

[magnet]

Imagine for a moment that you're at 80% after 55K miles and Tesla say that this is normal degradation. What options do you have?

Sound pretty normal to me. NCA cells drop to 75% after a few years and then stabilize and flatline at 70%. Mileage does not mean much. Calendar life means everything.

Imagine for a moment your car has 70% capacity after 20 years?

Have you established that brick 8 is in fact an outlier? Every pack will have a lowest capacity brick. Usually the brick with the lowest capacity when the car is new will be the brick with the lowest capacity after years of use. The fact brick 8 is always the lowest is of no concern unless that brick has significantly less capacity than the other bricks

The Mercedes (Tesla) warranty on the electric B class is 70% after 8 years

 

[Kevin Sharpe]

I've received some more detailed analysis from the people helping me behind the scenes... we now have day records from 28th May 2010 (Note: the data shows the permanent section has been reset twice). I will pull out the data in more detail as it becomes available.

One of the most 'encouraging' things is the rapid decline in range between the 16th April 2014 and 20th June 2014 (when the car broke down with yet another 440V controller failure). This confirms my 'gut feeling' and it's interesting to speculate what would have happened to the range if I'd been able to continue driving the car...

Brick Ah v Date

Brick Ah v Odo

The graphs show two interesting things: the average and minimum brick Ah correlate very closely, which supports the theory that the weakest brick affects the average more than we'd expect, and that I've experienced a sudden drop in Ah capacity (with corresponding reductions in CAC and range).

 

[djp]

Kevin - that's the same data as your original post, but in a graph instead of a table. It doesn't tell us anything new.

I assume you've been doing multiple log downloads over the years. Each one of those will have an ahr.log file. Can you post all of those, so we can see how the brick SOCs and voltages vary over charge level? There might be several bricks bringing down your CAC, not just brick #8.

 

[Yggdrasill]

One thing the graphs show is that the repair in November/December 2011 was completely irrelevant. Another thing the graph shows is that if brick 8 has been the weakest brick the whole time, replacing this brick may restore some of the capacity.

Also, if you look at the Ah vs odometer, you have had three fairly similar "bumps", where the Ah has increased, before decreasing again. When you compare the downslopes of these bumps with the Ah vs date graph, you can see that when the Ah has increased, you've driven little or moderately over a longer period, while when the Ah has dropped, you've driven a lot in a short amount of time. This is especially apparent on the last downtrend, where the Ah loss looks extreme on the Ah vs date graph, while it looks fairly normal on the Ah vs odometer graph.

I think this strengthens the proposition that the capacity loss is due to range charges and deep discharges.

Also, a final observation - if the pattern seen on the three similar bumps holds true, your range should increase if you drive the car moderately over the next few weeks. (Edit: Maybe the litium deposits inside the cells become uneven with intense use, and must be carefully cycled a few times to restore some of the capacity?)

 

[Kevin Sharpe]

One thing the graphs show is that the repair in November/December 2011 was completely irrelevant.

Here's the Brick Ah graph from 1st November 2011 to 1st June 2012 (Tesla repaired the battery at the end of November 2011);

Note that although the range did initially improve after the repair it quickly deteriorated and triggered my email on the 19th April 2012 (see OP). Also note that after the 1st June 2012 the Ah never returned to the level it had been before the repair.

 

[Yggdrasill]

Here's the Brick Ah graph from 1st November 2011 to 1st June 2012 (Tesla repaired the battery at the end of November 2011);

View attachment 59150

Note that although the range did initially improve after the repair it quickly deteriorated and triggered my email on the 19th April 2012 (see OP). Also note that after the 1st June 2012 the Ah never returned to the level it had been before the repair.

When you compare the two graphs, the upwards trend is when you drove ~3000 miles in ~3 months, while the downward trend is when you drove ~7000 miles in ~3 months. The range stops falling when you over the next ~3 months drive around ~3000 miles. In other words:

Intense use = range loss, careful use = no range loss.

 

[scaesare]

So if 7/8 bricks report some arbitrary value 50 and then one brick reports 30, don't you think that would signal that there is perhaps a defect with that one brick out of 8? I do. It's just statistics. If we are to determine there is abnormal degradation in a pack and all bricks report 50, well then I could see Tesla saying that's normal degradation (I.e. Too many range charges, deep cycling, etc). But 1/8 is something I would define as abnormal event.

I'm not the only one with this view. As I said before, there are several other, more knowledgeable people than I, that have suggested this methodology. It makes sense. It holds water.

Sorry I missed your response initially.

It might suggest defect... but those numbers in your example are pretty wide in terms variance with regard to the low brick... we don't know:

1) What the actual variance is in this case
2) What the spec for acceptable low-brick deviation is


What we DO know is that:

1) Tesla has examined the pack and said there is no defect
2) Tesla has stated that pack is operating within acceptable range
3) The current pack range is within earlier published guidance


So, again, I ask... why should Tesla replace this pack?

 

[wycolo]

Roadster Ploys:

Charge in Standard Mode but drive in Range Mode. For hyper-miler this just provides for continued VDS display as miles remaining approach zero, no actual advantage.

Charge in Range Mode but only partially (just before leaving): to the Miles you feel you really need on that particular trip. Thus for a 180 mile round trip a FULL Range Charge might only be used when worst-case winter conditions might occur. Most times it would be about 190>215 miles.

I wonder to what extent the OP used these to mitigate battery stress. My regular long route is ~180 miles and I always do this with the Roadster. Also record on calendar the start and finish miles "221/25" which pretty much says it all about the trip.
--

 

[Kevin Sharpe]

For hyper-miler this just provides for continued VDS display as miles remaining approach zero

When traveling 'long' distance I prefer to drive in range mode (even if I charged in standard mode) because I find the VDS displayed range more useful. I also always drive with OVMS displaying battery percentage and follow a GPS route so that I have a very good idea of range required to destination, SOC, traffic, etc.

Using this technique I have only twice in four years entered the zone where the car cannot calculate range (although OVMS is still very accurate) and that was just before reaching my destination. I have never run out of charge.

Charge in Range Mode but only partially (just before leaving): to the Miles you feel you really need on that particular trip.

Normally I will finish the range mode charge five minutes before I leave for the trip. In the past (before OVMS) this was more difficult if I was staying at a hotel and then I'd just slow the charge rate to finish 30 minutes before I left.

On very long trips where I need a second range charge I will usually leave when the current drops below 32A because I can't be bothered waiting.

 

[donauker]

Personally I avoid Range Mode driving if at all possible, especially in warm weather. Range mode driving allows the battery temperature to run significantly warmer to reduce the power overhead of pack cooling.

 

[Kevin Sharpe]

Personally I avoid Range Mode driving if at all possible, especially in warm weather. Range mode driving allows the battery temperature to run significantly warmer to reduce the power overhead of pack cooling.

We don't get 'warm' weather in the UK and I constantly monitor battery temp with OVMS... given on these trips I'm driving for range I'm normally driving conservatively at 55-65mph (my regular trip is mainly single lane roads with some freeway).

On the VDS the battery temperature never moved from the lowest blue indicator and I even asked Tesla if it was working.