[MetrologyFirst]

On the way to work this morning, listening to CNN on the XM radio. They do a peice on who killed the electric car! Guess who comes out sounding like the villain? Guess what the San Francisco guy was driving who they interviewed? A RAV4 EV. Way too superficial to be a 2-3 minute piece. No mention at all of any of the documented issues with the early generation electric cars. And no mention at all of the Volt. Only one guy's opinion of why "a company" would "crush" all their innocent electric cars. I expected more from Miles O'Brian, he does a nice job on the NASA coverage.

GM can't catch a break even on a show called "American Morning".

 

[darthvader420]

GM really does not deserve a break. If they had actually pushed for electric vehicles a decade ago like they are apparently doing now with the volt, the face of the automotive industry would be very different today. Instead, they crushed the EV1s and sold the battery rights to texaco/chevron and mass produced SUVs and other gas guzzlers. The problems with the NiMH batteries used in the late model EV1s and RAV4s are very overstated by GM and the oil companies, as evidenced by the RAV4s that are still driving today with 10+ year old batteries. So now we're stuck in a situation where we're waiting for lithium battery costs to come down before EVs can be widely adopted, great.

 

[kubel]

GM was the villain. What were they thinking crushing the EV1s? They could have sold them as-is to willing buyers. If they just stopped making them and didn't crush them, they wouldn't look so bad. But crushing them showed GMs intent. It's like giving someone a pair of cement shoes. They didn't want to just discontinue the car, they wanted to make them disappear. Why? They didn't want to show people that such a vehicle was possible, because it could stir up the same kind of legislation that they saw in California.

So, yeah, GM deserves every bit of flak it receives. And we need to keep an eye on GM to make sure they aren't conspiring to commit murder again. Once we see government forcing standards (as a result of seeing what EREV can do), GM will realize that such improvements will cost them billions of dollars, and GM will want to kill the Volt too, just like it killed the EV1. GM won't want to keep a product around that proves to the government that they can make fuel efficient and zero emission cars.

 

[darthvader420]

GM's huge investment in the volt is certainly encouraging, but we have to be honest. GM is desperate to change their corporate image. With the oil crisis looming they know they have to not only hop on board the "green bandwagon," they have to look like they are leading it. Look to their ad campaigns these days and this flexfuel fiasco. E85 ethanol, derived from corn, is an absolute joke environmentally and is driving up the price of food worldwide. But hey, it sure sounds nice, freedom from foreign oil, GM must not be so bad after all! Lots of people are saying that GM is trying to fix up their image to help convince the feds to give them a handout and they are probably right. GM dug its own grave when they abandoned their electric cars and pushed their SUVs as hard as they could. Now the GM executives have finally seen the writing on the wall! Too bad they weren't so forward thinking 10 years ago.

 

[Tom]

Numbers tell the story

What were they thinking crushing the EV1s?

The government standards imposed on GM forced it to use non-traditional thinking, rapid development, and low capital requirements as a CAFE strategy. It was an opportunity for a government partnership for a new transportation paradigm: but no partnership. The idea was ahead of its time in terms of government support, cost, battery technology, and reliability requirements. Let’s review inverter reliability.

The EV1 used a six smartpole (24 parallel MOSFET/pole) inverter. Reliability predictions back then were done using the old constant failure rate, parts count, part stress, design complexity, prediction procedure MIL-HDBK-217E, rather than the modern Physics-of-Failure methodology. The MIL-HDBK-217E procedure used stress models: Arrenhius temperature, Kemeny voltage acceleration, and Coffin-Mason fatigue. Based on this methodology (assuming that early infant mortality failures are screened and there is no wearout for less than 10 calendar years), using four nines/year single MOSFET 1990 reliability with 288 MOSFETS, the predicted inverter reliability (see the failure curve below) projects that 18% of the inverters would fail at the end of 7 years. Considering battery life and predicted inverter reliability, clearly GM would not want to sell EV1s that after 7 years would have 1 in 5 inverter failures in the field. The litigation would be astronomical. It would be cheaper to do the unthinkable and crush.

The failure rate of Power Devices has improved about 50X (see graph below) in the twenty years between 1990 and the 2010 start of Volt production, power density capability has been improved by the availability of IGBTs, and the life of light weight batteries has been extended to 7/10 years. The time for EVs is now.

 

[darthvader420]

The government standards imposed on GM forced it to use non-traditional thinking, rapid development, and low capital requirements as a CAFE strategy. It was an opportunity for a government partnership for a new transportation paradigm: but no partnership. The idea was ahead of its time in terms of government support, cost, battery technology, and reliability requirements. Let’s review inverter reliability.

The EV1 used a six smartpole (24 parallel MOSFET/pole) inverter. Reliability predictions back then were done using the old constant failure rate, parts count, part stress, design complexity, prediction procedure MIL-HDBK-217E, rather than the modern Physics-of-Failure methodology. The MIL-HDBK-217E procedure used stress models: Arrenhius temperature, Kemeny voltage acceleration, and Coffin-Mason fatigue. Based on this methodology (assuming that early infant mortality failures are screened and there is no wearout for less than 10 calendar years), using four nines/year single MOSFET 1990 reliability with 288 MOSFETS, the predicted inverter reliability (see the failure curve below) projects that 18% of the inverters would fail at the end of 7 years. Considering battery life and predicted inverter reliability, clearly GM would not want to sell EV1s that after 7 years would have 1 in 5 inverter failures in the field. The litigation would be astronomical. It would be cheaper to do the unthinkable and crush.

The failure rate of Power Devices has improved about 50X (see graph below) in the twenty years between 1990 and the 2010 start of Volt production, power density capability has been improved by the availability of IGBTs, and the life of light weight batteries has been extended to 7/10 years. The time for EVs is now.

Do the RAV4EVs have similar problems? I've seen a few anecdotes about 10 year old vehicles still running but no substantial data.

 

[Tom]

Statistical Projections

Do the RAV4EVs have similar problems? I've seen a few anecdotes about 10 year old vehicles still running but no substantial data.

With regard to the EV1s: The numbers are a statistical prediction per the above MIL standard, based on projection of failures due to manufacturing defects for 1990s technology. Similar to human longevity, there will be a spread of lifetimes. Assume the probability distribution of MOSFET failures is "exponential," then given the above assumptions, the projected inverter median (50% survival) failure operating life is 24 years. Consider a crude illustration, of the thousand or so EV1s built, one possible distribution is that 70 of them may run for 50 or more years, 900 for 24, and 30 for 5 years, for an average of ~24. The probability of a randomly selected vehicle running for 10 years is ~75%. MOSFETs also have wearout mechanisms, such as aluminum electromigration, which would give an upper end operating lifetime (my guess is ~ 20 years).

With regard to RAV4EVs, the same general principles apply, but I have no detailed knowledge of their development/design. They may have been designed more conservatively, used MOSFETs with fewer manufacturing defects, or manufactured at later years with more reliable components, or not.