Meeting with Nissan, Phoenix, Jan 8, 2013, 6pm, drinks prior

[palmermd]

If the LBC, related software or instrumentation can have such a profound effect on the autonomy, wouldn't this give more credence to calls for a true range warranty? You see, from an owner perspective, it matters little, if it's battery degradation or perhaps some other defect, that's restricting the range of the vehicle. I believe Tony made that point earlier last year, and this was one of the motivations for the range test on September 15.

I would agree, except this is hard to test for because of the time involved and the wide variance. You'd pretty much need to transport the car to a specific test track (or duplicate them all over the country) and drive it with a fixed profile. Really hard to quantify!

I believe the LBC can be fixed to work properly, and since it records all energy in/out, as well as knowing the health of each cell block, it will be able to accurately calculate any degradation. Right now there are clearly some issues Nissan must fix. Hopefully this will also give us a more accurate SoC, and thus the GoM will even benefit.

-Phil

which is why I suggested the dyno test. http://www.mynissanleaf.com/viewtopic.php?p=257088#p257088" onclick="window.open(this.href);return false; granted it is not perfect and will take an hour or more to perform, but you should be able to set up a protocol that can be repeated at any dyno anywhere in the world and the results should be comparable. The loading on the dyno will be set to a prescribed level to simulate the load driving say 65mph, and then drive the car at 65mph and see how far the car can go before VLBW.

 

[Ingineer]

which is why I suggested the dyno test. http://www.mynissanleaf.com/viewtopic.php?p=257088#p257088" onclick="window.open(this.href);return false; granted it is not perfect and will take an hour or more to perform, but you should be able to set up a protocol that can be repeated at any dyno anywhere in the world and the results should be comparable. The loading on the dyno will be set to a prescribed level to simulate the load driving say 65mph, and then drive the car at 65mph and see how far the car can go before VLBW.

That's a much better idea, but I still don't see it happening when there's a better way and it will have been recording every charge/discharge. I think Nissan's already decided in any event, so there's little point in this exercise.

-Phil

 

[Ingineer]

Would Leafscan help with this?

LEAFSCAN™ is only a tool to read information that the Leaf calculates using it's systems, so it's only as accurate as those systems are. Right now LEAFSCAN™ would only read the same loss of capacity as indicated by the capacity bars, as both are fed from the information the LBC (Battery Controller) is erroneously calculating. (Though it would give a higher resolution wrong number! =)

Once Nissan fixes the LBC's software, then LEAFSCAN™ will give an accurate degradation factor. (As accurate as the LBC can calculate)

-Phil

 

[surfingslovak]

I think Nissan's already decided in any event, so there's little point in this exercise.

Yes, that's what I have to assume as well. The only reason why I stirred the pot, and raised the issue of instrumentation was the feedback from owners leading up to and during the Town Hall meeting. While we can be reasonably certain that Nissan has done their homework in regards to degradation and battery life projections, I believe that little attention was given to actual driving performance of heat-affected cars.

I drove Randy's LEAF during the range test, and I have to admit that it behaved differently than my LEAF in a cooler part of the country. Gids seemed to decline more slowly than I would have anticipated at low SOC. If I recall correctly, Randy's LEAF got close to 20 miles after hitting LBW while going 62 mph (GPS) / 64 mph (dash) on the freeway, and it's still had some life left in it. That was pretty surprising to see, and I could not complete the test, because the car significantly outperformed Tony's original Gid-based range projection.

Azdre and others have mentioned something similar: their LEAFs have lost a lot of range, much more than what one would assume based on the degradation figure measured in Casa Grande. It's important to note that they would only cycle the battery between full charge and the low battery warning.

Interestingly, this "lost" range seemed to be "hiding" at low SOC. One would have to go significantly below the low battery warning, which many owners feel uncomfortable hitting on a daily basis. I don't believe that this has been acknowledged anywhere. It's a concern, because even with a battery capacity warranty in place, owners could have trouble getting adequate range from their cars.

 

[planet4ever]

I'd like to take one more stab at explaining the instrumentation hypothesis, without taking sides, because it seems to me that some people here don't "get" it, despite attempts by Phil and others to explain it. So, here are some hypothetical numbers, that are almost certainly not correct, but should make the concept clear:

Assume a new battery with 24kWh total capacity and an LBC (using Nissan's name for the battery controller) that limits charging to 22kWh maximum and 1kWh minimum. As users we would see a 21kWh capacity from "100%" to dead turtle.

Next, assume that at some time later the battery has degraded to 23kWh total capacity, but for some reason the LBC now sets the charging limit at 20kWh maximum and 2kWh minimum. This could be due either to a concern that the degraded battery needed more protection or to a programming error. Either way, as users we would now see an 18kWh capacity from "100%" to dead turtle.

All of our tests, whether Gids or wall kWh or range tests or dyno or whatever you can imagine, would show a capacity loss of 3/21 = 14%. In fact the true capacity loss would be 1/24 = 4%.

Ray

 

[Stoaty]

All of our tests, whether Gids or wall kWh or range tests or dyno or whatever you can imagine, would show a capacity loss of 3/21 = 14%. In fact the true capacity loss would be 1/24 = 4%.

Agreed... but just to be clear, Nissan's bench test would show 4% capacity loss. The fact that their bench test showed a 14% loss as the best result for the Phoenix Seven means a combination of accelerated battery degradation and instrumentation problems in that climate.

 

[TonyWilliams]

I believe Tony made that point earlier last year, and this was one of the motivations for the range test on September 15.

I would agree, except this is hard to test for because of the time involved and the wide variance. You'd pretty much need to transport the car to a specific test track (or duplicate them all over the country) and drive it with a fixed profile. Really hard to quantify!

Actually, it's not that hard to quantify or duplicate the efforts of our PHX range test anywhere, or any other well designed and documented protocol for a test. Anything that would substantially affect range (outside of a failure of the car) is documented in our test.

I certainly recommend a test track for testing; the one I phone called in Phoenix would cost over $5000 to rent for the day. That's not a big deal for every car manufacturer that has a private test track anyway. But, there's no shortage of facilities to run tests throughout the world.

A fixed driving profile; well, that's super easy (actually, 64mph on cruise control couldn't be easier). Ambient conditions; easy. So, where is the "wide variance"? Frankly, if there is a wide variance, you're not doing it right.

Like I've said many times, whether the battery is 100% of the range issue or 0%, or some number in between, the end consumer doesn't care. Only the actual vehicle performance matters.

We already know that the instruments are bad, and Nissan agrees, and we also already know that these batteries degrade substantially in heat (otherwise EVERY OTHER EV manufacturer probably wouldn't waste their time and money cooling the batteries). We also know that this particular chemistry is the MOST susceptible to heat degradation of all the current EV chemistries. This is public domain knowledge.

We also know it gets hotter than #%*? in Phoenix, and no so much in Seattle. Both have substantial volumes of LEAFs (actually, significantly more in Seattle), yet Seattle has a grand total of nobody with serious degradation (even with a car at 60,000 miles) and Phoenix and other hot areas.... not so good with no known cars anywhere close to 60,000 miles.

It would be a stretch beyond my threshold of common sense to think that the batteries are not a significant part of the problem. Edit 1: I'm not referring to the battery controls, but the physical chemical breakdown of the batteries to reduce their performance.

Edit 2: yes, dyno tests are fine, particularly since that's exactly how EPA derived range for the LEAF. Much easier to control the variables in a lab.

Edit 3: of course the electronic controls of the battery are no doubt not "perfect". Does a consumer care if the electronic battery controls are 2.4% responsible for their 30% degradation of vehicle range autonomy, or 0%, or 100%. No. But, again, if somebody is thinking there's no battery problem and its just electronic controls or instruments.... well, that's a stretch that all I've mentioned above won't let me jump to.

 

[Ingineer]

I'd like to take one more stab at explaining the instrumentation hypothesis, without taking sides, because it seems to me that some people here don't "get" it, despite attempts by Phil and others to explain it. So, here are some hypothetical numbers, that are almost certainly not correct, but should make the concept clear:

Thanks Ray, that's very well explained.

-Phil

 

[Ingineer]

Agreed... but just to be clear, Nissan's bench test would show 4% capacity loss. The fact that their bench test showed a 14% loss as the best result for the Phoenix Seven means a combination of accelerated battery degradation and instrumentation problems in that climate.

How do you know that for sure? Their bench test might have been relying on data obtained with Consult III+, which is obtained from the LBC, which is where the errors lie. I think Nissan now knows there is a problem with the LBC, as they have admitted.

I'm not saying you are wrong, I'm just wondering how you know for sure?

-Phil

 

[Ingineer]

I'll say it AGAIN for clarity:

I'm not saying there isn't degradation in the battery, in fact, I'm sure there is some. I'm saying that it's possible that this is being amplified by flawed logic in the LBC, and that once this is fixed, many cars may not have as much actual loss as they presently appear to.

-Phil

 

[surfingslovak]

Agreed... but just to be clear, Nissan's bench test would show 4% capacity loss. The fact that their bench test showed a 14% loss as the best result for the Phoenix Seven means a combination of accelerated battery degradation and instrumentation problems in that climate.

How do you know that for sure? Their bench test might have been relying on data obtained with Consult III+, which is obtained from the LBC, which is where the errors lie. I think Nissan now knows there is a problem with the LBC, as they have admitted.

I'm not saying you are wrong, I'm just wondering how you know for sure?

From what I can tell, the bench test was a bit more than a Consult III+ readout. I believe that it was a monitored discharge with a constant 5 kW load. Nissan kept some of the cars at Casa Grande for a long time, couple of weeks or more, and yet they put only few miles on the odometer. While the affected owners would be much more qualified to answer this question, I have every reason to believe that the batteries were put through their paces last July. If I recall correctly, we had this discussion on the 500+ page thread last year.

 

[TonyWilliams]

I'll say it AGAIN for clarity:

I'm not saying there isn't degradation in the battery, in fact, I'm sure there is some. I'm saying that it's possible that this is being amplified by flawed logic in the LBC, and that once this is fixed, many cars may not have as much actual loss as they presently appear to.

-Phil

Certainly a valid theory to test. I would like to have tested a car both before and after whatever firmware change they make to quantify any improvement in vehicle performance.

My hunch is that the end result will continue to be that physical degradation of the cells will be the overriding issue.

 

[TonyWilliams]

From what I can tell, the bench test was a bit more than a Consult III+ readout. I believe that it was a monitored discharge with a constant 5 kW load. Nissan kept some of the cars at Casa Grande for a long time, couple of weeks or more, and yet they put only few miles on the odometer. While the affected owners would be much more qualified to answer this question, I have every reason to believe that the batteries were put through their paces last July. If I recall correctly, we had this discussion on the 500+ page thread last year.

I most certainly would like to develop a similar test that we can administer (separate from Nissan) for a static load test to quantify nothing but battery performance.

 

[surfingslovak]

I most certainly would like to develop a similar test that we can administer (separate from Nissan) for a static load test to quantify nothing but battery performance.

Yes, I would like to get hold of this test protocol and all the pertinent details. It might be better than to try design something from scratch. Although I suspect that the 5 kW load is supposed to mimic the LA4 cycle, it might be good enough as a standardized test. One of the downsides would be a long discharge time for larger packs.

 

[Ingineer]

Why not use a larger load for a faster test? Seems like 1C is about right, and then would take less than an hour. Should be pretty easy to build a test load that heats water. That's about a small swimming pool heater's output. =)

-Phil

 

[Stoaty]

Agreed... but just to be clear, Nissan's bench test would show 4% capacity loss. The fact that their bench test showed a 14% loss as the best result for the Phoenix Seven means a combination of accelerated battery degradation and instrumentation problems in that climate.

How do you know that for sure? Their bench test might have been relying on data obtained with Consult III+, which is obtained from the LBC, which is where the errors lie. I think Nissan now knows there is a problem with the LBC, as they have admitted.

I'm not saying you are wrong, I'm just wondering how you know for sure?

I don't know for sure, we only have Nissan's report (filtered through the affected owners who reported what they were told by Nissan) to go on. However, if even Nissan can't remove the battery from the Leaf and test it and determine its actual capacity we have a much bigger problem here. If they don't have a test to accurately determine actual battery capacity, how can they say battery capacity loss is less than the instruments in the Leaf show?

 

[TonyWilliams]

Why not use a larger load for a faster test? Seems like 1C is about right, and then would take less than an hour. Should be pretty easy to build a test load that heats water. That's about a small swimming pool heater's output. =)

-Phil

My "quick charger guy" has a big load that he uses for testing the DC chargers up to about 75kW, I think.

That could be set up for whatever load. So, my next question is do we test the battery "dead" (no BMS, no safety from bricking) or otherwise?

 

[surfingslovak]

I'm not saying you are wrong, I'm just wondering how you know for sure?

I don't know for sure, we only have Nissan's report (filtered through the affected owners who reported what they were told by Nissan) to go on. However, if even Nissan can't remove the battery from the Leaf and test it and determine its actual capacity we have a much bigger problem here. If they don't have a test to accurately determine actual battery capacity, how can they say battery capacity loss is less than the instruments in the Leaf show?

Fair enough, although I assumed that we had some basic consensus. Let's just say that I have a bit more than just hearsay to go by, but please don't take my word for it. It would be best to have one of the AZ 7 re-confirm this if they were still following the forum.

That said, here is the graph TickTock, one of the most credible contributors on this forum, jotted down from memory after meeting with a Nissan engineer. This is supposed to represent battery capacity projection for different locations, and TickTock's data point from Casa Grande added in for easier reference.

Here's my receation of the graph I think is most pertinent to all owners/buyers to decide individual course of action. The actual dots are numbers I wrote down, I filled in the gaps from memory based on the shape.

Indeed there is a precipitous drop in capacity in the first 6 months and then the degradation slows and continues to slow over time. The key thing to notice is the 80% @ 5 years we've all been using to base our decisions is actually only the median value and not what a Phoenician can use as an "expected value".

These curves are based on average mileage per area: 7500 for Phoenix (I don't recall what the Boston number was except that it was higher then Phoenix). I drive almost 2X that which explains why I am sitting well below the Phoenix predicted curve at 84.7 @ 1.3years. This data is not LA4, but rather a 32.2A discharge rate to better match 55mph discharge rate.

The axis on the right was not on the graph I was presented. I added that based on the 84.7% SOH Nissan measured and the 74 mile range Tony measured shortly after so I can get a realistic idea how much range I will have at year 5. As you can see it will be well below the 65miles per day I require.

If you stick within the 7500 miles per year (~30 miles/day), you should get 65 miles after 5 years and still be quite happy since you only need 30.

 

[Ingineer]

My "quick charger guy" has a big load that he uses for testing the DC chargers up to about 75kW, I think.

That could be set up for whatever load. So, my next question is do we test the battery "dead" (no BMS, no safety from bricking) or otherwise?

The proper way is to power up the LBC and watch the Max Discharge current and then stop the test when it "turtles" (goes below the 1C test power).

-Phil

 

[TonyWilliams]

My "quick charger guy" has a big load that he uses for testing the DC chargers up to about 75kW, I think.

That could be set up for whatever load. So, my next question is do we test the battery "dead" (no BMS, no safety from bricking) or otherwise?

The proper way is to power up the LBC and watch the Max Discharge current and then stop the test when it "turtles" (goes below the 1C test power).

-Phil

Ok, that sounds logical. So, at 393.5 volts and 1C (24kW load), I should see about 60 amps, and as the voltage drops below 300 volts, around 80 amps.

Where's the best place to connect the load?