Kevin Sharpe's decreased Roadster range

[Kevin Sharpe]

BBC Radio 4 "You and Yours" Consumer Affairs Programme now available online. Includes interview with two Tesla Roadster owners concerned about range degradation (start at 21:38)

http://www.bbc.co.uk/programmes/b04gcdt8

 

[tomsax]

Thanks Kevin. Based on that log file, I'm not seeing any issues with brick 8, its SOC and voltage are in line with all other bricks. Your SOC is being limited by bricks 29 and 72, both of which are 0.02V below the rest of the pack. Bricks 45, 46, 73 and 74 are close behind and also lower than average. Based on this data it does look like the capacity loss is due to normal degradation, and not the result of brick 8 being an outlier.

We'd expect all of the bricks to be within 0.02V, that's what the brick balancing algorithm does.

What does the SOC column represent? It doesn't look like Ah, despite the name of the file. There could be a scaling factor, and the brick numbering scheme could be different, but the ratio of the average to the minimum doesn't agree with the log file entry from the same day, and it's odd that there are two bricks with the same minimum "SOC" value, but one brick is singled out consistently for years in the log file indicating there is not another very close brick.

 

[tezco]

...Unfortunately, EV owners are driving around on the equivalent of an eighth of a tank of gas, so more accurate numbers are essential since we are operating much closer to the edge....

Rather misleading since an EV is also far more efficient, which is why you can get 260 miles from "an eighth of a tank of gas".

You may be correct as far as the 85 S is concerned, but 100,000 Leaf owners are much closer to the "eighth of a tank" analogy I pulled out of the air. I'm getting close to Kevin's situation with my Leaf and my commute is only 30 miles round trip. When I ordered the car in advance of production I carefully studied the EPA's winter cycle figure and what little was known about pack degradation for EV's. Unfortunately, the EPA's winter cycle "MPG" figure is exceedingly overly optimistic when compared to my worst case scenarios, and the Leaf's rate of degradation also seems to be worse than I had predicted. Perfect storm, and I had spent a lot of time researching what was known at the time. California drivers own the largest numbers of EV's and therefore tend to dominate the forums, but their experiences do not help cold climate drivers for the most part, and that is where EV performance falls like a rock.

 

[djp]

We'd expect all of the bricks to be within 0.02V, that's what the brick balancing algorithm does.

What does the SOC column represent? It doesn't look like Ah, despite the name of the file. There could be a scaling factor, and the brick numbering scheme could be different, but the ratio of the average to the minimum doesn't agree with the log file entry from the same day, and it's odd that there are two bricks with the same minimum "SOC" value, but one brick is singled out consistently for years in the log file indicating there's isn't another very close brick.

The SOC is state of charge with binary point scaling. I haven't taken a close look at the units yet. The brick numbering is the same across all data sources on my car, including the ahr.log. I agree Kevin's ahr.log is inconsistent with his original post.

The ahr.log has two snapshots of the SOC and voltage data (prev and pres), both of which have similar distributions in Kevin's file. The 'present' data is a higher SOC and more well balanced than the 'previous' data, but they both show the same low bricks. Brick 8 isn't an outlier in either snapshot, and there is a large group of low bricks (including 29, 45, 72 and 73).

 

[ElectricLove]

And you knew this buying the car, right?

Let me respond to this as an EV fanatic, I'm guessing most Tesla drivers are driving/owning their 1st Electric Vehicle with their Tesla. For me, that isn't even close, I've owned closer to 10 in my life and have worked on most of them, I've built (in part) a few as well! As an EV DIY fanatic I obviously know all about battery degradation and can tell you that I've replaced many a battery pack sooner than I would have liked to or predicted would have been the case, I am completely knowledgable about that fact and consider it to be the #1 (perhaps really the ONLY thing besides "range anxiety") NEGATIVE side to Electric Vehicles and will lead to the demise of an Electric future if not properly addressed by the industry. As an EV fanatic, I'm going to drive Electric regardless, I'm already sold on the idea but when I consider the "masses" and a future where any significant market share (ie. more than 1% of vehicles sold) are 100% electric powered I realize this "inconvenience" is not gonna fly, "most" people are not going to be interested in buying a vehicle which lacks a warranty on performance (ie. it will perform as well or at least X% as well in 5 years as it does the day you bought it). The other big players in the EV world are addressing this, at least in part, by offering warranties on their batteries in regards to battery degradation (really "range" degradation to the consumer since the average Joe doesn't care if their battery is 100kwh or 50kwh just so long as they can drive 250 miles each and every time they have a full charge).

My sense of Kevin is that he is also an EV fanatic, he totally understands batteries and that they do degrade, however just like me I think he is passionate about the stance that; without the EV manufacturers providing a guarantee on performance of their vehicles in the long run, the widespread adoption of EV technology will fail! (to be what it could be) An example of where EV tech is losing is: Toyota has one of the "greenest" images of any auto manufacturer and they didn't renew any contracts with Tesla and have stated they are investing in a "Fuel Cell" vehicle future, Toyota would have been a huge difference maker in the EV world but I think the life-cycle of the batteries is preventing them from buying into the concept. The battery degradation issue is perhaps THE major hurdle that needs to be addressed for our EV future, range is being addressed (bigger batteries, lots of new charging stations, J1772 standard, SC, ChaDeMo, etc) and I think the "fight" for the best range is doing great harm to the degradation issue. Tesla should never have released range estimates that are based on 90% DOD (Depth of Discharge), its not a safe duty-cycle for the longevity of the batteries, its not right to tell people "our car will get you 240 miles on a charge" and then under your breath say "but we really hope you don't actually intend to drive it that far very often or it won't be able to give you that same performance 4, 5 or even 8 years from now"...

Honestly, my feeling is that these batteries should have been absolutely limited to using 70% of their initial capacity maximum and then over time increasing the percentage of the capacity to maintain the "usable kwh" accordingly.
Example:
70% of 100kwh when purchased gives you 70kwh.
5 years later the capacity of the pack may be down to 90kwh (maybe better b/c 70% duty cycle results in slower degradation) and the computer is now allowing 77.7% of that, .777*90=70, so the driver is still getting the exact same usable capacity and range as day 1... This means HAPPY CONSUMER and no one feels cheated! In this system it would be reasonable to expect that at the end of the 8yr/100K mile warranty period the driver would still have the exact same range as brand new!

I think for the average Joe out there this needs to be addressed, Tesla is relying on their buyers to be intelligent enough to do all the engineering themselves, to not actually use 100% of the battery like Tesla is allowing (just so they can claim higher range at purchase). Again, I think this is all an artifact of "range anxiety" being the #1 priority and "good engineering" being back-burnered; safety nets should be incorporated in every EV! This is the reason I also have a Volt, I stand behind what they did by only allowing for 65% of their battery to be used, this is why you can have one with 100K EV miles on it and the EV range is exactly the same as it was new, if you simply gave the Volt a 100kwh pack it would be able to drive 250 EV miles easily and be able to confidently do so 100K+ and 8+ years down the road...

Now, to the warranty point, the way the manufacturer solves this potential issue is by simply providing a warranty on battery performance, its up to them to engineer a product that their consumer can use as intended and perform its function as promised and within the boundaries of the warranty coverage. If Tesla builds a car that can destroy itself in 50K miles doing exactly what they claimed it could do, they are responsible for that and should be replacing Kevin's pack. When you buy something you should be able to use it exactly as you are told you can use it through advertising and salespersons (ie. go 0-60 in 3.9s, drive 240 miles on a charge... in other words: drive fast and drive far!) and shouldn't expect it to be unable to perform the same way just 50K miles and 5 years later, the warranty is supposed to give people confidence that it will perform the same way for X number of years and X number of miles or the manufacturer is promising to fix it for you... Just because Tesla has a warranty with a huge weasel clause doesn't mean it is ethical (in the business sense) to exercise that weasel clause.

 

[dsm363]

Hiding 30% of the battery is something GM might do but not Tesla. Tesla does need to do a better job of education on degradation but I'd rather use as much of the range I paid for as possible.

 

[qwk]

We'd expect all of the bricks to be within 0.02V, that's what the brick balancing algorithm does.

What does the SOC column represent? It doesn't look like Ah, despite the name of the file. There could be a scaling factor, and the brick numbering scheme could be different, but the ratio of the average to the minimum doesn't agree with the log file entry from the same day, and it's odd that there are two bricks with the same minimum "SOC" value, but one brick is singled out consistently for years in the log file indicating there's isn't another very close brick.

You are right. The latest logs he posted don't make much sense with respect to the prior logs. The information is so different, it's like looking logs from two different cars(this is not an accusation BTW). I really don't know what to make of it.

 

[ElectricLove]

Hiding 30% of the battery is something GM might do but not Tesla. Tesla does need to do a better job of education on degradation but I'd rather use as much of the range I paid for as possible.

That would be true for me to, I prefer to manage it myself BUT think of the average Joe buying a car... You bought a Roadster and a Model S, you obviously are more intelligent and responsible than 95% of people out there, you have earned a living that allows you to buy not one but 2 $100K+ cars and have the luxury of knowing if you needed to replace a battery pack for another $40K, you probably could and it wouldn't be a big deal... You are not the "average consumer", you are far above average, you do "the research", you know your stuff, you are not the one that needs this protection. This is fine if Tesla wants to sell 2500 high-end sports cars to enthusiasts, you are the right buyer for that...

However, Tesla's goal is widespread adoption of their vehicles, they decided it wasn't enough to stay niche and just sell to guys like you and I, and because of that I feel they need to execute things slightly differently; the average Joe buyer needs Tesla to protect them from themselves, they need that protection. GM understands that, they aren't going to build a car they want to sell to everyone and not enforce good engineering and safety-nets on the vehicles. I think this is something Tesla needs to learn. As you are invested in the stock and probably very wealthy because of that you can imagine how much more money you will have if Tesla is successful at selling millions of vehicles and gaining real market-share among the "big boys"; they are very small right now...

 

[Chickenlittle]

BBC Radio 4 "You and Yours" Consumer Affairs Programme now available online. Includes interview with two Tesla Roadster owners concerned about range degradation (start at 21:38)

http://www.bbc.co.uk/programmes/b04gcdt8

I think we have an answer as to why it is 19% down. In the bbc report he says his commute was 180 miles and he can only charge to 190 miles now. I strongly suspect he was range charging for his daily commute something the manual advises against to save the battery performance. Have asked in three posts whether he was doing excessive range charging and no answer.

 

[tezco]

Hiding 30% of the battery is something GM might do but not Tesla. Tesla does need to do a better job of education on degradation but I'd rather use as much of the range I paid for as possible.

As an EV enthusiast, I'd agree; however, the EV companies will need to get more skin in the game (ie realistic warranties), and they need to quit heavily advertising range figures that are not sustainable in the real world. EV manufacturers need to jettison the EPA and develop new ways to educate the uninitiated so that their expectations match the realities of battery chemistry. The EPA is stuck trying to force metrics developed to measure ICE on EV's, and it isn't working.

For example, we know that range is greatly affected as the temperature falls. We need to see a graph showing your expected range at 30, 20, 10, 0, -10, -20 and -30 degrees when new, and at 3, 5, 7 and 10 years down the road at the lowest battery capacity warrantied.

 

[tomsax]

You are right. The latest logs he posted don't make much sense with respect to the prior logs. The information is so different, it's like looking logs from two different cars(this is not an accusation BTW). I really don't know what to make of it.

To be clear, we aren't comparing different logs, we are comparing different pieces of data in the same log.

- - - Updated - - -

The SOC is state of charge with binary point scaling. I haven't taken a close look at the units yet. The brick numbering is the same across all data sources on my car, including the ahr.log. I agree Kevin's ahr.log is inconsistent with his original post.

The ahr.log has two snapshots of the SOC and voltage data (prev and pres), both of which have similar distributions in Kevin's file. The 'present' data is a higher SOC and more well balanced than the 'previous' data, but they both show the same low bricks. Brick 8 isn't an outlier in either snapshot, and there is a large group of low bricks (including 29, 45, 72 and 73).

So is the SOC value relative to some reference energy or Ah level? Or is it SOC relative to that brick's current estimated energy or Ah rating? If it's the latter, then it's totally useless for comparing the relative health of the bricks.

 

[pharma5]

BBC Radio 4 "You and Yours" Consumer Affairs Programme now available online. Includes interview with two Tesla Roadster owners concerned about range degradation (start at 21:38)

http://www.bbc.co.uk/programmes/b04gcdt8

A better listen than I was expecting after leaving this thread. Thought it might be pants, and that we'd have to put you in your local Tower Bridge lockup for a spell.:wink:

Is Alex the same gent on the bottom of your OP?

Alexander Sims (51,000 miles) - “Had 190 miles when brand new, now 50,000 miles later I still get 182 on a charge... Amazing”

Sounds like you both entered with your eyes open wrt battery deg, and the piece clearly acknowledged that the public recognize the phenomenon (Samantha: "...most electric car drivers are aware of this..."). It sounded less about, pick a phrase (gross injustice?), and more about how EV manufacturers can improve further.

For you, 20-25 real miles of range lost, unable to do the 180 mile roundtrip commute nonstop.

A minor point missing on the potential 'backward step' to petrol:
No other EV company has made a vehicle to meet your all-EV range need. If Nissan or Renault had a long range car, you might have an alternative in the marketplace with a warranty design to your liking. (Or they might similarly be reluctant to offer limits if they were on the cutting edge of long-range all-EV driving). However, each has short-range EVs only, and the market has forced them to provide more attractive warranty products (or battery lease) sooner to guard the limited range of their products and the higher cost sensitivity of their customer segment (which has the most EV competition).

As Tesla moves downmarket, and other manufacturers move upmarket, consumer buying decisions should take care of it.

- - - Updated - - -

... and they need to quit heavily advertising range figures that are not sustainable in the real world. EV manufacturers need to jettison the EPA and develop new ways to educate the uninitiated so that their expectations match the realities of battery chemistry.

+1, the advertising/communication of range needs to improve and simplify.

 

[djp]

Tesla battery range degradation forcing return to gas

So is the SOC value relative to some reference energy or Ah level? Or is it SOC relative to that brick's current estimated energy or Ah rating? If it's the latter, then it's totally useless for comparing the relative health of the bricks.

In my car the ahr.log SOC values correspond to the SOC MIN and SOC MAX on the VDS but with a ~15% offset. It looks like it's measured relative to a Standard Charge, with 85% true SOC being recorded as 100% and Range Charge as over 100%.

The SOC MIN determines SOC LIM, which in turn is used to calculate the displayed ideal range, so has a direct effect on the range estimate.

On my car the logs and VDS are all consistent with the same bricks showing low no matter the source, and the same outliers at mid charge or Standard Charge.

 

[dsm363]

As an EV enthusiast, I'd agree; however, the EV companies will need to get more skin in the game (ie realistic warranties), and they need to quit heavily advertising range figures that are not sustainable in the real world. EV manufacturers need to jettison the EPA and develop new ways to educate the uninitiated so that their expectations match the realities of battery chemistry. The EPA is stuck trying to force metrics developed to measure ICE on EV's, and it isn't working.

For example, we know that range is greatly affected as the temperature falls. We need to see a graph showing your expected range at 30, 20, 10, 0, -10, -20 and -30 degrees when new, and at 3, 5, 7 and 10 years down the road at the lowest battery capacity warrantied.

Eventually Tesla will have the luxury of adding bigger padding to the pack but selling a $80,000+ car with a 185 mile range might mean Tesla isn't around to make the Model 3.

 

[JRP3]

BBC Radio 4 "You and Yours" Consumer Affairs Programme now available online. Includes interview with two Tesla Roadster owners concerned about range degradation (start at 21:38)

http://www.bbc.co.uk/programmes/b04gcdt8

I can't say that was the scathing report you were implying.

We're now at the point where the range is 199 miles. What that means realistically is I can no longer use my car for the 180 mile round trip

So the car has 19 miles more range than you need, but you can't use it?

 

[stopcrazypp]

I've received a couple of comments about your analysis that I wanted to pass on...

"I don't think the log you posted recently doesn't provide the information for what they are looking for, a weak battery brick will charge and top of first.. then with balancing bleeding, and more charging topoffs most other bricks will start matching an full SOC. To get an idea what happening you need to get two snapshots, where one right after a charge to full, then drive until almost empty then take another snapshot to compare the two. To me the snapshot log you posted basically just showed the car fully top offed i.e. nearly full SOC on every brick, but we don't know the Amp hours that are available in each brick."

I will record the two logs requested and anything else that will be helpful.

I didn't have time to take a peek until today, but I would suggest the same thing too after having taken a look. Having a reading at a low SOC (near empty) will give the best picture because the weakest bricks will hit the lower voltages first and at that point of the discharge curve the slope of the curve is the highest so it gives the best SOC estimate (near full is the other point also).

It also eliminates the possibility of the log being "useless" as tom is suggesting in the case that the SOC is measuring a relative number (as in 0-100% SOC) rather than an absolute number (similar to Ah or CAC). This is because the pack is top balanced (meaning all bricks are at 100% SOC at full range charge), so as it is discharged, the weakest bricks will hit lower SOC first before the others.

However, what djp says is true too. If #8 is a weak brick is should still show up somewhat as lower SOC than the others even at a Standard charge. Right now from your log it appears perfectly fine.

- - - Updated - - -

So is the SOC value relative to some reference energy or Ah level? Or is it SOC relative to that brick's current estimated energy or Ah rating? If it's the latter, then it's totally useless for comparing the relative health of the bricks.

Even if it is relative rather than absolute SOC it is not totally useless. This is because bricks are "top balanced" (meaning all of them are at 100% SOC at the same max voltage point). As the bricks discharge, the weakest bricks will reach lower SOC first. And the reason the weakest brick "limits" the pack as a whole is because it reaches 0% SOC first before the stronger ones do (so the stronger ones still have some SOC left even when the pack BMS will refuse to let you discharge further).

This will however require Kevin to take a log at a near empty state (as his friend suggested).

 

[qwk]

To be clear, we aren't comparing different logs, we are comparing different pieces of data in the same logs.

I realize that, but brick #8(the problem brick), appears fine in the last log information posted. This is why Tesla does the bleed test in addition to looking at the logs, so they find out what is going on with much better accuracy.

 

[Kevin Sharpe]

I can't say that was the scathing report you were implying.

I am not anti-Tesla but think the whole issue of range degradation and the lack of warranties is important. You will find the mainstream press will focus more on the consumer rights aspect of this story (mainly with Model S owners) and that's exactly how it should be.

- - - Updated - - -

I really don't know what to make of it.

iirc the files I posted were from the most recent logs and were taken just after the car was returned from Tesla following the recent breakdown (yet another 440V controller). Tesla had the car for 8 weeks and obviously it was not driven much in that time. I've given Tom access to all the logs I have which basically cover the period from the battery 'repair' in December 2011 to date. I've also requested logs from Tesla for the period prior to the battery repair but they haven't responded to that request.

I will produce the 'empty' and 'full' logs when I get a moment (hopefully this week).

- - - Updated - - -

Is Alex the same gent on the bottom of your OP?

Yes, he's one of the six UK Roadster owners who's contact details I passed across to the BBC.

- - - Updated - - -

For you, 20-25 real miles of range lost, unable to do the 180 mile roundtrip commute nonstop.

You're mixing up 'ideal' and 'real' miles. My car now has a range of 199 'ideal' miles and the roundtrip is 180 'real' miles. If I drove at 55mph without hills, wind, heat, or cooling I could theoretically make it... on a real day, on real roads, I won't.

- - - Updated - - -

No other EV company has made a vehicle to meet your all-EV range need.

Correct, that's why we run PHEV and have two long range BEV conversions in the works.

- - - Updated - - -

I strongly suspect he was range charging for his daily commute something the manual advises against to save the battery performance.

My 'commute' is a twice monthly (sometime more) 180 mile roundtrip. I range charge just before I leave on the drive leaving the car at 100% for less than five minutes. I have been using the cool down feature of OVMS since it was implemented a year or so ago.

Remember that my range issues began following the battery 'repair' in 2011.

At no time has Tesla ever suggested that my ownership behaviour has accelerated the range degradation that I have experienced. I have no idea why my pack is so different from other UK cars that I've seen.

 

[CowboyR]

I think we have an answer as to why it is 19% down. In the bbc report he says his commute was 180 miles and he can only charge to 190 miles now. I strongly suspect he was range charging for his daily commute something the manual advises against to save the battery performance. Have asked in three posts whether he was doing excessive range charging and no answer.

My 'commute' is a twice monthly (sometime more) 180 mile roundtrip. I range charge just before I leave on the drive leaving the car at 100% for less than five minutes. I have been using the cool down feature of OVMS since it was implemented a year or so ago.

Remember that my range issues began following the battery 'repair' in 2011.

At no time has Tesla ever suggested that my ownership behaviour has accelerated the range degradation that I have experienced. I have no idea why my pack is so different from other UK cars that I've seen.

this message comes up every time you try to range charge a roadster and needs to be cleared before you can continue...

 

[Kevin Sharpe]

this message comes up every time you try to range charge a roadster and needs to be cleared before you can continue...

Correct... can you give me a link to Tesla's definition of "frequent"?

As I said previously, we have no evidence that my car use pattern has negatively impacted the battery range, and Tesla (who've been monitoring my car since day one) have never suggested this is a factor.

By all means blame me as an owner/driver if it makes you feel better but personally I'm more interested in understanding why my car is in the top 4 of the PiA study.