From karn Fri Mar 3 15:46:10 2000 Date: Fri, 3 Mar 2000 15:46:08 -0800 From: Phil Karn To: ev1-club@cco.caltech.edu Subject: My email to Ken Stewart last week Reply-to: karn@ka9q.ampr.org Here is the message I sent to Ken Stewart at GM a week ago. He called me on Monday afternoon to confirm receipt. At that time he could only offer me vague assurances that GM was aware of the problem and working on it. --Phil ------- Start of forwarded message ------- Date: Fri, 25 Feb 2000 09:21:55 -0800 From: Phil Karn To: kenneth.1.stewart@gm.com, mary.n.moore@gm.com Subject: EV1 fire in San Diego Reply-to: karn@ka9q.ampr.org Ken, As you probably know by now, 1997 EV1 VIN 428 was involved in a fire while it was parked in its owners' garage in the Tierrasanta area of San Diego on February 17, 2000. The car was destroyed and the garage and most of its contents were damaged beyond repair. The lessees of this car are Ron Brauer and Ruth Bygness. Ruth and I are Qualcomm employees as well as EV1 drivers; she's a supervisor of technical publications and I am a principal engineer with graduate degrees in electrical and computer engineering. Before the fire we regularly shared a charger at Qualcomm. Ruth invited me to their house on Wednesday of this week to observe as three GM employees, two State Farm insurance agents and a fire origins investigator retained by State Farm took a look at the car. I took many photos with my digital camera. I have given copies to Ruth, Steve Mackaig (the State Farm fire investigator) and to several of the technically-oriented EV1 drivers I know. I would be happy to provide copies to GM as well if they would prove useful in your investigation. Since the fire, I've been getting numerous first and second hand reports of other charge port incidents, though none were anywhere near as serious as last week's fire. There are reports of charge paddles melting into the ports (on EV1s and on Toyota RAV4 EVs with inductive charging) and of smoke issuing from an EV1 charge port. The latter incident fortunately was witnessed by the owner who took quick action to cut power and stop the fire. I personally experienced an EV1 charge port failure, though not in my own car (VIN 660). When I was considering an EV1 in May of 1998, Saturn of Kearny Mesa gave me a car for an extended test drive. After only one day, the charge coupler failed while it was charging in the rain at Scripps Hospital in Encinitas. I remember a burnt odor in the front of the car when I returned. Fortunately there was no fire. When Ron Brauer first observed the fire in his garage, he saw flames erupting from the center of the front bumper, i.e., from the precise region of the charge port. Of course, the investigation has just begun. But given the frequent occurrence of charge port incidents, Ron Brauer's observations of the fire, and my own observations of the remains of VIN 428, I personally feel that the initial evidence points very strongly to the inductive charge port or paddle as the cause. All of us are EXTREMELY concerned about the potential for similar fires in other vehicles. And most of all, we are seriously disturbed by the complete lack of any official announcement or warning to date from GM regarding this incident. After Ruth and Ron called me two days after the fire, I sent a brief message to the EV1 club mailing list (ev1-club@cco.caltech.edu) informing the list of the fire and suggesting a few precautions that drivers might want to take. I had mixed emotions sending this message, knowing that many people read the list and any negative news about the EV1 could be seized by opponents of EVs and used to seriously jeopardize the program. But I felt that a real hazard had been demonstrated and in my judgment it was imperative to spread the word. Nobody was killed in *this* fire. The owners were fortunate. Their garage was detached from their main house, and the fire started when they were home and awake. I shudder to think what would have happened had a similar fire started in my own (attached) garage, say sometime after the timer kicked on at midnight and I was asleep in bed. I strongly urge GM to *immediately* get on top of this issue, to review the fire safety issues associated with the EV1 and inductive charging in general and to issue any appropriate warnings and recalls as soon as possible. This review should be thorough and honest. It must go well beyond vague public statements that you're looking at the problem. To be meaningful I believe it should seriously address at least these questions: 1. Is inductive charging of EVs still a good idea? GM heavily touted the safety advantages (especially in the rain!) of inductive charging when it first announced the EV1. Now we have several years of real-world experience with both conductive and inductive Level 2 EV charging that can be used to reassess the relative safety, reliability, efficiency and cost of the two systems. I respectfully suggest that GM set aside its proprietary interest in inductive charging and do what's best for the future of EVs. 2. The quest for weight reduction in the EV1 led to the widespread use of nonmetals such as foam rubber, polystyrene, epoxy composites and other plastics that are both lighter and more flammable than metals like steel. It is clear that if these materials cannot be made fire-resistant, then greater care will have to be taken in their use. Fire propagation paths from possible ignition points should be modeled and studied. 3. What safety analyses are done in the design of high power electronics for EVs? Do the design engineers consider the possible consequences of a shorting failure in, say, every single electrolytic bypass capacitor, power rectifier or switching transistor? Or do they simply assume that these components won't ever fail in that way? When your power supply is a 312V battery pack fused at 500 amps, a failed component would have to draw more than 156 kilowatts from the battery to blow the fuse. These energy levels can do very serious damage very quickly. There is simply no room to cut design corners. 4. What kind of quality control testing is performed on the software in the EV1? Drivers regularly report to the mailing list all sorts of weird failure modes that are almost certainly due to the failure of the software designers to anticipate input sequences that are unusual but not impossible. The most recent example is a sequence of events that can consistently cause a Gen II car to fail to inhibit driving when the charging paddle is still inserted. Software is part of several safety-critical systems in the EV1, including battery charging and braking, and these lesser problems don't inspire a lot of confidence. I apologize for the length and detail of this note, but as you can see the situation is very serious. I am still an enthusiastic supporter of the EV1 and of EVs in general, and I really want to see them all succeed. I sincerely hope that GM will respond to this fire as NASA did to the Apollo 1 fire -- in a way that makes the program and the product much better as a result. Thank you for your time. Phil Karn