This week, we take on a couple of oddities. The first is the Leaf range loss issue. This has been widely reported elsewhere with an inordinate amount of information I’ve found implausible. So this week we went to Tony Williams, organizer of the group in Phoenix purporting to refute a Nissan claim that there was nothing wrong with the batteries.
This call cuts so many ways I’m loathe to try to make sense of it. One thing is certain, a number of Leaf owners, otherwise delighted with the car, are very unhappy over the range issue and Nissan’s response to it.
This is enough to represent a setback to the entire movement to electric cars, and tends to confirm the fears of those who hesitate to purchase electric vehicles because of “range anxiety.”. A most unfortunate turn of events.
Nissan has been stoically obtuse in dealing with it. Part of this is the very nature of being a large corporation. In building large numbers of anything, correcting errors must needs be involves large numbers. 20,000 is not a huge number in terms of automobille manufacture, but it is still a pretty expensive number if replacing the batteries in each of them is the fix. And once they agree to replace a few battery packs, ALL Leaf owners everywhere will want theirs replaced as well on any event where they don’t get 100 miles on a charge – Nissan hoisted on the petard of their own inflated range claims.
So it is difficult to work out the real from the perceptual, and at some point perception BECOMES reality in these matters.
For us, and our viewers, we actually need to know what is happening in reality, not perceptually. There is little we can do about Nissan and the broad damage caused to the mission by their cars and their reaction. But is there anything in here we can learn from the issuance of a large number of cars and a large number of batteries into real world applications? Anything we could learn and use for examples?
So far not much. The obvious opportunity to me would be actual evidence of decreased battery capacity in hot climates. Even the Chinese engineers tell me that the cycle life would be diminished at 45C compared to 25C. But when I press for some test data on that, they first refuse, and then finally admit they haven’t actually tested it with cycle tests.
I kind of implied in the video that there is no theory or mechanism for cycle life deterioration in heat. That is not PRECISELY so. There IS a bit of a theory. It COULD be true. I can find no one who has tested it and the tests of Professor Jay Whitacre on A123 cells seems to imply MUCH longer cycle lives for these LiFePo4 cells than previously predicted. But there IS a theory.
When the battery is first manufactured, lithium exists in two places and two forms. LIthium Iron Phosphate in the cathode LiFePo4 and Lithium Carbonate in the electrolyte LiPf6. And the battery has no charge at all.
The battery is initially charged very carefully and HOW it is initially charged is in all cases secret sauce because it has a great deal to do with battery performance later. Electrons are of course pumped OFF the cathode and ON to the anode externally by a power supply. This separates an electron from the outer valence band of the Lithium in the cathode, which then loses its place in the crystal structure and deintercalates or oxidizes. As a cation (Lithium ion positive) It then migrates to the Anode side where it picks UP an electron from the anode current collector to again become lithium metal. That electron is what allows it to hold hands with the carbon crystal lattice and INTERCALATE into that structure. In doing so, the addition of the electron to the anode and removal of one from the cathode makes the anode two electrons more negative than the cathode.
As this process continues, the anode becomes increasingly negative with respect to the cathode, until all the lithium is extracted from the cathode and appears as lithium metal in the anode. At that point the battery is fully charged.
During this first formative charge, there are some consequences in the breakdown of the organic solvents used as the electrolyte. This chemical breakdown causes the deposition of a flexible layer of by product on the SURFACE of the lithium/carbon anode. This is called the Solid Electrolyte Interphase layer or SEI. This sounds bad, but it is not really all bad. The addition of lithium metal to the carbon crystal matrix causes signficant changes in volume to the structure. It is somewhat brittle and this volumetric change causes it to fracture. The SEI layer acts as kind of an epoxy cement, that holds the carbon particles in place on the copper current collector. That’a GOOD thing. But it does slow the migration of lithium ions INTO the carbon structure and OUT of the carbon structure and so presents a portion of the diffusion delay limiting the amount of power we can produce.
Though most of this happens on the FIRST charge, there continues to be a MUCH lesser decay of solvents on subsequent charges. And so this SEI layer thickens very slowly – further limiting power. Somehow, it is felt this also translates to CAPACITY though I’ve always been a little unclear how. I understand the diffusion delay increasess, and I understand total instantaneous power might decrease, but it is felt that this also plays a role in capacity reduction.
And the theory is that this breakdown of electrolytic solvents is accelerated with heat.
Can’t find ANY test data to indicate that. I understand that this MAKES SENSE to a lot of people. To me, that implies it MIGHT be true, despite so many instances of things making sense and not being true that I have a personal bias that if it DOES make sense it probably is a great sign that it ISN’T true.
So the Leaf / Phoenix association takes on a bit of interest to me personally. Might be some data there.
Arizona and Florida have long been obvious havens for electric venicles, and it has never worked out very well. That is because Pb cells – lead acid, absorbed glass mat, whatever, have DRAMATICALLY shortened life cycles at elevated temperatures. If you get 3 years from a lead pack in Oregon, you MIGHT get a full year in Florida. But you won’t in Arizona. This is widely known and thoroughly proven. But I fear that knowledge has leaked into THIS situation, and thus making sense of the resulting conversation is problematical. How many ways to say it. Li batteries are NOT an improved form of Pb battery. They are entirely different devices, with entirely different chemistries, sharing nothing but an affiliation for electrons at one end and an abhorrence of them at the other.
The controversy took on a point when Andrw Palmer, Nissan Executive Vice President who has been involved with Nissans electric vehicle projects for many years, made a public statement that they did not have a battery problem, but most likely had an instrumentation problem.
Enraged that Nissan was apparently dissing their concerns, the Nissonoids of Arizona, led by Tony Williams of San Diego, decided to do a SCIENTIFIC TEST to PROVE that the Nissan did in fact have a BATTERY PROBLME. Then they embarked on a test of RANGE. This kind of muddled thinking drives me crazy. A battery is a battery and range is range. If you think they are the SAME thing, you should try to design SCIENTIFIC TESTS. They tried to control a ridiculous number of variables like tire pressure, wind direction, elevation, temperature, the weight of the car, even ballasting the weight of the driver. But they THEN used Nissan’s instrumentation to determine range.
By ALSO using the GIDMETER, they think they have elminated that as well. Not really. The Nissan BMS monitors voltage, energy usage, and battery SOC with a Battery Management System. This system reports results on the bars displays and of course using the graphic TURTLE. More accurately, it reports it as data on a CANbus and of course the display translates that data to graphics. The GIDmeter takes the SAME CANbus data and displays it in cruder, but more complete form, as numbers. In both cases, you are using the data from the BMS.
But pack voltage, indeed individual cell voltages I think, are available on the CANbus.
And so they did a RANGE test, terminated by a TURTLE, to determine the range of the cars. All ended with a turtle, but the pack voltages at the end of the test varied WILDLY, from a low of 295 volts to a high of 352 volts. I would posit that the batteries at 352 were not at the end of charge at all, and the cells at 295 very well may have been beyond it. In other words, they have a very fickle TURTLE on the display.
And it papears that Mr. Palmer was likely correct that they have an instrumentation problem, not a battery problem. And the Phoenix group is most likelyreading messages from God in cloud formations, leaving my hopes for some actual data deriving from a number of vehicles operating in the Phoenix heat rather lying in a heap – dashed in disappointment. We don’t know. And Nissan doesn’t know. And nothing in any of this will likely tell us. A number of Nissan Leaf’s HAVE been operating in Phoenix for over a year, and we have NO results to report. Gads.
Of course, long time viewers know MY biased conclusion. The BMS – AGAIN. You most likely have some perfectly operation automobiles with perfectly capable battery packs crippled and disabled by Yet Another Bad BMS Design. YABBD. Should something be done about it? Obviously. Now what would Jack do?????
Should Nissan be the one to take the financial hit? Or should the buyers of Nissans’ cars take the hit. I vote Nissan. But they can do whatever they think is in their own best interest.
My hope is that the more savvy viewers of EVTV now have some information. I’m sorry to report you still have no data. But you now know WHY you don’t have any data and how NOT to do a battery test. It has hard to do conclusive battery testing with turtles. Bars are not much better. And representing the SAME faulty data numerically offers only slight improvement, in that we were able to get FOUR terminating voltages anyway. Enough to ascertain that the testing methodology was comically wretched.
We struggle with such testing ourselves and with instrumentation ourselves. You will note a couple of failures here at EVTV. First, I apparently pulled the mic cord on the 70 mph range test of the Cadillac Escalade. Too bad. I spumed forth many precious Rickardisms on that drive lost to us all forever. As we are still two hours and forty two minutes long, I suspect you’ll forgive me when I summarize them thusly – 834 wH per mile.
WATT HOURS? Now where did THAT come from? Don’t we usually measure amp hours? Yes we do. Why do we do that? Because that’s the only instruments we have and can afford in an EV. Watt Hours have always been preferable in that they represent in a better way the total energy use accounting for both voltage and current, while the voltage varies widely during the drive. But we haven’t had a good wattmeter to use. BEFORE you crawl up my skirt on Cycle Analyst, I HAVE one, I HAVE tested it, and it cannot even measure current accurately, so any wH it would derive would also be inaccurate.
I have worked for several months with an instrumentation company in CHina to modify one of their current meters to use a 12v supply for example, and to display WATT HOURS, VOLTAGE< and CURRENT simultaneously on the same screen. It winds up being a LARGISH meter to display all that in large numbers I can read. It is a little complicated. And it was a little difficult to exlpain to them the DC nature of our world and our ongoing disinterest in power factor and RMS values. But we got er done and I got my prototype. They are anxiously awaiting a quantity order. one wee problem. I never really specified that it measure current in BOTH directions. As we are using a shunt, it never occured to me that their meter would measure in ONE direction only. But in going from AC to DC that's what they did, and they do not think it's economically viable to correct this little feature. I think it makes the meter nearly worthless. Oh, not entirely. I just have to zero it out after every drive. And it doesn't count back down while charging. Since the JLD404 does, by using BOTH it works reasonably well. But not well enough to be a product I'm afraid. And at the price, really a disappointment. I feel foolish to have never brought it up. Both sides are worn out from trying to communicate a feature set for a meter in Chinglish. So it has kind of died as a product. So when beating up Nissan, we have our own little design setbacks. Unlike Nissan, we are afforded the LUXURY of being able to offer a FULL REFUND to everyone that bought this killowatt meter. Of course, Im the only one who has and unfortunately the Chinese vendor is not as generous as I am so I eat the cost of the one meter delivered I'm afraid. So I commiserate with Nissan at the cruelties of the market place. But would encourage them to address the problem with some much simpler but more accurate instrumentation. I would also note that they TOP balance and the 352 volt car might INDEED be dead, with one short cell that the BMS has noted and shut down the car. BOTTOM BALANCE INSTEAD if you must. But in all things, I think Nissan should strive for a collegial and consensual relationship with their early customers, and avoid allowing this to become an adversarial relationship over ducats. With 20,000 cars down, and a new car plant AND a new battery plant coming online in Tennessee, the stakes are actually quite high and the lessons learned bill really pretty low by comparison. I like Teslas approach of gathering battery performance data by cell phone from their cars. Less guess work. More knowledge. Maybe we'll fund John Hardy to do some cycle life testing using a styrofoam cooler and a light bulb. Or maybe I'll do it. Jack Rickard