Capacity Loss on 2011-2012 LEAFs

[Herm]

Surfingslovac I had to spoof my IP address so you could not track my down in Timbuktu, so disregard any odd results you may get..thanks for the link.

 

[surfingslovak]

Very funny, Herm. Any EVs in Timbuktu?

 

[azdre]

Azdre, Hyundai offers a lifetime warranty on the battery in their hybrid.. perhaps Kia does the same?.

What is the latest word on the Volt warranty?.. does it cover capacity or not?.. I thought they were required to cover it due to Federal emissions laws.

Yeah, and Nissan offered a 10 year warranty that's only exclusion was 'gradual' capacity loss and look where that got us. 30% gone in 16 months and being told it's gradual. The Civic Hybrid I'm sure has a great warranty that doesn't stop them from changing up some software and killing your efficiency. I really hope to be able to go a few years without having to talk to lawyers about my car.

I've heard the Volt warranty is good, but my husband and I are both very tall which makes the backseat as useful as the backseat in a Porsche, a great place for your gym bag.

 

[surfingslovak]

I really hope to be able to go a few years without having to talk to lawyers about my car.

Seriously. Or spend half a year on an owner forum debating the finer points of lithium-ion battery care.

 

[Joeviocoe]

I added a Florida Leaf driver to the tracking wiki
http://www.mynissanleaf.com/wiki/index.php?title=Real_World_Battery_Capacity_Loss" onclick="window.open(this.href);return false;

It was the first Florida Leaf driver to have lost a capacity bar, that I have been aware of so it was an easy choice to add to the wiki.
I have used the naming format of "PIA-xxx" with the 'x' as a placeholder for the survey vehicle ID.

Does anyone recognize any of the vehicles here??... and that has not already been otherwise added to the wiki under a MNL username?
http://www.pluginamerica.org/surveys/batteries/leaf/vehicles.php?order=bars

 

[tomsax]

I really hope to be able to go a few years without having to talk to lawyers about my car.

It's been 3 1/2 years for me, but that's a whole 'nother story.

 

[surfingslovak]

I added a Florida Leaf driver to the tracking wiki

Thank you, Joe. First report from Florida, that's quite significant. Many thanks for your earlier comment as well. I'm sure that everyone involved or affected appreciates it, I know I did.

 

[Weatherman]

I added a Florida Leaf driver to the tracking wiki
http://www.mynissanleaf.com/wiki/index.php?title=Real_World_Battery_Capacity_Loss" onclick="window.open(this.href);return false;

It was the first Florida Leaf driver to have lost a capacity bar, that I have been aware of so it was an easy choice to add to the wiki.
I have used the naming format of "PIA-xxx" with the 'x' as a placeholder for the survey vehicle ID.

I believe this is the same person referenced on the PIA list:

http://www.mynissanleaf.com/viewtopic.php?f=49&t=550&start=1370#p252949" onclick="window.open(this.href);return false;

pretty much right on schedule for the first battery capacity bar to fall on a South Florida LEAF. I'd expect to see a lot more reports from Florida over the next six months.

 

[Stoaty]

I believe this is the same person referenced on the PIA list:

http://www.mynissanleaf.com/viewtopic.php?f=49&t=550&start=1370#p252949" onclick="window.open(this.href);return false;

pretty much right on schedule for the first battery capacity bar to fall on a South Florida LEAF. I'd expect to see a lot more reports from Florida over the next six months.

Unfortunately, I agree. Hopefully Florida will not be quite as bad as the battery aging model predicts.

 

[Joeviocoe]

I added a Florida Leaf driver to the tracking wiki
http://www.mynissanleaf.com/wiki/index.php?title=Real_World_Battery_Capacity_Loss" onclick="window.open(this.href);return false;

It was the first Florida Leaf driver to have lost a capacity bar, that I have been aware of so it was an easy choice to add to the wiki.
I have used the naming format of "PIA-xxx" with the 'x' as a placeholder for the survey vehicle ID.

I believe this is the same person referenced on the PIA list:

http://www.mynissanleaf.com/viewtopic.php?f=49&t=550&start=1370#p252949" onclick="window.open(this.href);return false;

pretty much right on schedule for the first battery capacity bar to fall on a South Florida LEAF. I'd expect to see a lot more reports from Florida over the next six months.

I edited the wiki to reflect the username "braindonor" links are updated too.

I truly hope that this one is a statistical outlier. I've lived in Florida for 10 years. It's HOT. But I am still thinking it won't be anything like Arizona. The state will certainly lead the East coast with prematurely degraded packs, but I am thinking, it won't be AS BAD AS the desert Leaves. We're all watching closely.

 

[surfingslovak]

I truly hope that this one is a statistical outlier.

That's exactly what I thought when azdre posted her first post in the 500-page thread. Seven months and many scars later, I think differently.

I've lived in Florida for 10 years. It's HOT. But I am still thinking it won't be anything like Arizona.

Yes, agreed. Hopefully the work done offline and online here on MNL gives some of the information presented here a little more credibility than you gave them initially. That said, I would urge you to have a look at Stoaty's battery aging model, if you haven't seen it already. Although nobody can claim that it's authoritative, it is well-researched.

The state will certainly lead the East coast with prematurely degraded packs, but I am thinking, it won't be AS BAD AS the desert Leaves. We're all watching closely.

Indeed. Regarding your recent comments on Green Autoblog: I disagree with the assessment that early range degradation affects only 1% of US LEAFs. We had that argument before. In spades.

Although a lot will depend on the current and future geographic vehicle distribution, the current data points to as many as 10% bar losers within the first two years of ownership. Whether that's an acceptable number or not is for everyone to decide for themselves. The PIA study also shows that about 20% of LEAF owners report noticeable loss of range, based on subjective observation. This number includes the bar loser group, obviously. While the vast majority of owners will likely do well, this data set also shows that loss of range might be a real problem, and not some statistical artifact.

 

[Joeviocoe]

The state will certainly lead the East coast with prematurely degraded packs, but I am thinking, it won't be AS BAD AS the desert Leaves. We're all watching closely.

Indeed. Regarding your recent comments on Green Autoblog: I disagree with the assessment that early range degradation affects only 1% of US LEAFs. We had that argument before. In spades.

Although a lot will depend on the current and future geographic vehicle distribution, the current data points to as many as 10% bar losers within the first two years of ownership. Whether that's an acceptable number or not is for everyone to decide for themselves. The PIA study also shows that about 20% of LEAF owners report noticeable loss of range, based on subjective observation. This number includes the bar loser group, obviously. While the vast majority of owners will likely do well, this data set also shows that loss of range might be a real problem, and not some statistical artifact.

This is why I think the current PIA survey is too high a percentage.. and is destined to drop down as the sample gets larger.
http://www.pluginamerica.org/surveys/batteries/leaf/Leaf-Battery-Survey.pdf" onclick="window.open(this.href);return false; page 1

"One of the first places it was posted was on an MNL thread discussing the battery capacity loss issue, so early data was
probably oversampling owners who have lost battery capacity at a higher rate than the average owner."

This survey's sample size of 240 vehicles is MUCH too low to start making predictions on the true number of affected Leafs vs. total Leaf population. And it wasn't intended to make that prediction at all.. It CAN and DOES work well to identify correlation with charging habits and ambient local temperature. Which is quite valuable.

I truly do NOT think the number is as low as 1%... as reading my rehashed argument on ABG might have you believe. Just that it is our largest confirmed number. This is all very informal right now.
I am just trying to quell the premature inflation of this number. The problem is real... but on ABG, some are using it to scare away potential Leaf buyers who don't even live anywhere near a hot zone. Customers should be aware... but not scared with exaggerations.

The sample size must grow. And MNL and PIA can only do volunteer data collection, which is inherently going to skew the results toward drivers with more problems. Nissan would have to extract data from ALL Leafs sold, to get a complete picture... good luck with that. :roll:

 

[OrientExpress]

This survey's sample size of 240 vehicles is MUCH too low to start making predictions on the true number of affected Leafs vs. total Leaf population. And it wasn't intended to make that prediction at all.. It CAN and DOES work well to identify correlation with charging habits and ambient local temperature. Which is quite valuable.

The PIA survey is on its way to being a statistically valid survey for US LEAF owners and potentially LEAF owners globally. Today the US LEAF owner universe is close to 18K, and to get a sample result that has a 95% ±4 level of confidence, then the sample size needs to be about 580 responses. If the universe is expanded to all LEAFs sold world-wide (43K cars to date) then the sample size only needs to expand to about 600 responses.

The current PIA sample has a confidence level of 95% ±6.5, which is good enough to draw some broad conclusions, one of which being that the overwhelming majority of LEAF owners do not any have issues with their car's traction battery.

One variable that the survey really does not quantify in the survey is driving style which does vary widely, and is a major influencer in how a vehicle performs.

 

[surfingslovak]

Nissan would have to extract data from ALL Leafs sold, to get a complete picture... good luck with that. :roll:

Please see this video, where Andy alludes to the fact that Nissan has all the data required to drive an informed decision.

[youtube]http://www.youtube.com/watch?v=R1tfX7fRWPI[/youtube]


The study was circulated off-list before I posted it on MNL. It was subsequently shared in four regional Facebook groups, two of whom are 300 or more members strong. Whether the sampling is representative or not is another matter, and hopefully this will improve over time. What's interesting is that the preliminary results match fairly closely in their distribution of individual outcomes, what this board has collected on the Wiki. The Wiki data comes from several venues (Facebook, comments on various articles, etc).

Yes, both of these data collections hint at a fairly significant frequency of this problem in Southern locales. Given the strong and unexpected correlation between mean temperature at the place of residence and Gids, which was spotted many months ago, this type of outcome should not surprising.

I'm all for a level-headed approach to this issue, but trivializing it, as was attempted before, will not help solve it. Yes, working with incomplete and noisy data sets can be tedious and frustrating. Much like measuring usable battery capacity via "butt dynos", as you noted couple of months ago. If your aim is to expand the base of the PIA study, then I'm all for it.

 

[Weatherman]

I think we've reached a point where trying to draw conclusions based on national statistics is pointless. It's like a weather forecaster saying, "based on national statistics, the probability of the temperature getting above 80F this afternoon is less than 5%". That forecast is meaningless. If you are in Seattle, the probability is near zero. If you are in Miami, the probability is near 100%.

We already have a model, which provides a reasonable first guess as to what your capacity loss will be. Use it. It's easy. Just put in your location and the number of miles you expect to drive each year, and you'll get a good idea of how fast your LEAF battery will degrade.

 

[Joeviocoe]

This survey's sample size of 240 vehicles is MUCH too low to start making predictions on the true number of affected Leafs vs. total Leaf population. And it wasn't intended to make that prediction at all.. It CAN and DOES work well to identify correlation with charging habits and ambient local temperature. Which is quite valuable.

The PIA survey is on its way to being a statistically valid survey for US LEAF owners and potentially LEAF owners globally. Today the US LEAF owner universe is close to 18K, and to get a sample result that has a 95% ±4 level of confidence, then the sample size needs to be about 580 responses. If the universe is expanded to all LEAFs sold world-wide (43K cars to date) then the sample size only needs to expand to about 600 responses.

The current PIA sample has a confidence level of 95% ±6.5, which is good enough to draw some broad conclusions, one of which being that the overwhelming majority of LEAF owners do not any have issues with their car's traction battery.

One variable that the survey really does not quantify in the survey is driving style which does vary widely, and is a major influencer in how a vehicle performs.

I disagree. The confidence would be correct IF the sampling was truly random. But PIA's survey is volunteer based and advertised in places where Leaf owners with problems are a bit more likely to respond than owners with no problems. So I would not have such real confidence even at 600 responses.... not unless Leaf owners are chosen from random.

Nissan would have to extract data from ALL Leafs sold, to get a complete picture... good luck with that. :roll:

I'm all for a level-headed approach to this issue, but trivializing it, as was attempted before, will not help solve it. Yes, working with incomplete and noisy data sets can be tedious and frustrating. Much like measuring usable battery capacity via "butt dynos", as you noted couple of months ago. If your aim is to expand the base of the PIA study, then I'm all for it.

I completely agree. It may seem like I am trying to downplay (trivialize) the problem. I truly am not. I agree there is a significant failing here, and Nissan needs to correct it. The warranty is a start. I am simply trying to prevent the pendulum from swinging too far in the other direction... mass panic. ABG thrives on sensationalism, MNL is quite a bit more rational. So it is an uphill battle to bring any sanity there.

We already have a model, which provides a reasonable first guess as to what your capacity loss will be. Use it. It's easy. Just put in your location and the number of miles you expect to drive each year, and you'll get a good idea of how fast your LEAF battery will degrade.

This is the best we've got... and it is good, and hopefully getting better as data comes in. Thank you.

 

[Stoaty]

I think we've reached a point where trying to draw conclusions based on national statistics is pointless. It's like a weather forecaster saying, "based on national statistics, the probability of the temperature getting above 80F this afternoon is less than 5%". That forecast is meaningless. If you are in Seattle, the probability is near zero. If you are in Miami, the probability is near 100%.

We already have a model, which provides a reasonable first guess as to what your capacity loss will be. Use it. It's easy. Just put in your location and the number of miles you expect to drive each year, and you'll get a good idea of how fast your LEAF battery will degrade.

The study confirms that the climate you live in is the main predictor of the rate of battery capacity loss. We already know that in Arizona the loss of at least one battery capacity bar in a fairly short period of time is above 12%. In the pacific northwest, the northeast and the northern parts of the mid-west, no need to worry about it at all. While the battery aging model may not give an exact accurate prediction of capacity loss, it is very useful for determining where you should worry, and where you don't need to be concerned.

 

[OrientExpress]

I disagree. The confidence would be correct IF the sampling was truly random. But PIA's survey is volunteer based and advertised in places where Leaf owners with problems are a bit more likely to respond than owners with no problems. So I would not have such real confidence even at 600 responses.... not unless Leaf owners are chosen from random.

That is true that the initial sample is somewhat biased given the recruitment methods used, but that bias will even out to be more representative of a universal sample that was chosen randomly when you reach an optimum sample size for a given universe. For the global universe of LEAFs in service, than minimum number is 600.

 

[RegGuheert]

I disagree. The confidence would be correct IF the sampling was truly random. But PIA's survey is volunteer based and advertised in places where Leaf owners with problems are a bit more likely to respond than owners with no problems. So I would not have such real confidence even at 600 responses.... not unless Leaf owners are chosen from random.

That is true that the initial sample is somewhat biased given the recruitment methods used, but that bias will even out to be more representative of a universal sample that was chosen randomly when you reach an optimum sample size for a given universe. For the global universe of LEAFs in service, than minimum number is 600.

I also agree that the sampling for the PIA survey is not sufficiently random.

That said, why would Nissan executives point us to it if it were overly pessimistic. IMO, the answer is that they wouldn't.

 

[OrientExpress]

Without getting into probability theory, and sample gaming (or how many angels can stand on the head of a pin), suffice to say that the current sample size the PIA survey has a margin of ±6.5, which is fairly random (but not where it needs to be) but is an fairly accurate quick-look sample of LEAF universe trends. But to be accurate , the recognized confidence for valid randomness is ±4. That is why if the sample size, regardless of how the sample points are recruited grows to 600 the confidence in the survey grows to ±4 that it has a 95% probability of representing the universe. If you want a confidence of 95% ±3 then the sample size needs to grow to 1007, for a confidence of 95% ±2 then the sample size needs to be 2119, and for a confidence of 95% ±1, increase the sample to 6263. If you want a 99% ±1 then do a sample of 8647.

But in the end, a sample that yields a 95%±4 is the easiest, quickest, and a very accurate way to measure a universe, and will be extremely close to a sample scored at 99% ±1. Next time you watch a news program that has some political survey notice that the poll is most probably ±4.

Looking at the PIA survey as it sits today, the results are probably borderline pessimistic, with only a 90% (±6) response of no issues. I suspect that if a data set of the complete universe was reviewed, it would show a response closer to 92 ~ 97% ±4, so the PIA sample is in the ballpark. I say that because of experiences of working on similar analyses of technical issues with a consumer product.