- the general motors team that developed the pioneering ev1 electric car had to invent new technologies that are now commonplace on evs. among them: low-rolling-resistance tires, keyless ignition, a heat pump for hvac, and regenerative braking.
- the impact prototype (more accurately a concept car) that led to the production program was unveiled at the la auto show in january 1990. the 1994 speed record car (pictured above), also called "impact," used a highly modified early-build development car.
- the adventure ended, sadly, in 1999 when production was stopped after only a few more than 1000 ev1 cars had been built. here's the inside story of the development process.
in the 1990s, the team working on the electric car that became the gm ev1 faced many challenges, foremost among them the means of extracting even marginally usable range. the car had the energy equivalent of about a half-gallon of gas stored in its 26 lead-acid propulsion batteries, and this problem required rethinking just about everything on the then-cutting-edge automobile.
while gm electric vehicles (later advanced technology vehicles div.) engineers were rethinking and reinventing virtually every component and system, our milford proving grounds team got to test and develop their fine work in full-vehicle form. i somehow scored the job of leading that team as vehicle test and develop manager and was blessed with the outstanding services of lead development engineer clive roberts (borrowed from lotus when gm owned it) and a trio of brilliant young engineers named marty freedman, garrett beauregard, and travis schwenke.
the list of technologies designed, developed, and put into production by that talented and tireless ev1 engineering force is truly impressive. on the ev1 were many industry firsts and many widely used today in both ice and ev vehicles. among the most significant were: power electronics design, packaging, and cooling directly related to today’s evs; electrohydraulic power steering (ehps), which soon led to electric power steering; heat pump hvac (the “grandfather” of today’s systems); low-rolling-resistance tires; inductive charging (now widely used for phones, electric toothbrushes, and other things); an electric-defrost windshield (virtually invisible embedded wiring to defog the glass); keyless ignition (the ev1 used a console keypad); electric brakes and parking brake; by-wire acceleration, braking, and gear selection; cabin temperature preconditioning; tire-pressure sensing; regenerative braking (including variable coast regen—which was on our early development cars but just two set levels selected by a shift lever button on production ev1 cars, due to legal issues with brake-light activation), and regen/friction brake blending; igbt (replacing mosfet) power inverter technology, and low-friction bearings, seals, and lubricants.
our clever test and development team also created innovative ways of doing their jobs. because dynamometer range tests—accelerating, coasting, and braking to follow a precise trace on a computer screen from 100 percent to zero state of charge—were long, tedious, and boring, garrett and travis developed a system to drive the car for them. "we had drive-by-wire and brake-by-wire, so mostly thanks to travis, we made that work," garrett recalls. "it was weird to stand behind a wall and watch the car do its thing, matching the trace very closely for repeatable tests, and we won an r&d award for that." a sort-of early precursor of autonomous driving . . . but with the car tied down on a dyno.
as a new team member, garrett vividly recalls his first meeting when engineer "chips" leung began presenting in chinese. "cold sweats! did i miss the language requirement?" he feared. then powertrain chief engineer jon bereisa told the story (in english) of the team's visit to china (because right-hand-drive ev1s were intended to be marketed there and elsewhere around the world), and the tale of a chinese vice premier getting in a prototype impact to drive it with chips. only then did that vip nervously admit that he had never before driven a vehicle.
direction sometimes came from a need noticed at the top. "one of our engineers spilled his big gulp on the console," recalls program executive director ken baker, "which fried the mechanical electronic switches for the gearshift mechanism. [program chief engineer] jim ellis then declared that we needed a design that would survive a big gulp dumped on it, which led to a big gulp test. the solution was a protective membrane over the switch body."
in the fall of 1993, when gm was in financial trouble and our program was officially "on hold," we did a series of media briefings and drives. we brought one publication at a time into the proving grounds, briefed them thoroughly on every engineering aspect, then let them test clive's proof of concept (poc) chassis development car on a fun route to the town of milford, michigan, where we recharged it for an afternoon run by another reporter who joined us for lunch. i also gave each one a thrill ride on a proving grounds hill course to show off the car's surprisingly competent handing. the resulting articles were highly positive. "we drive the world’s best electric car," gushed popular science on its january 1994 cover. "gm's hard-charging impact is practical, fun to drive, and a master stroke of engineering," echoed popular mechanics. even enthusiast magazines (including car and driver) were pleasantly surprised. and we heard later that those positive reviews helped the gm board decide to revive the program the next year.
vehicle chief engineer mike liedtke vividly recalls the day he hosted ace comedian, talk show host, and car collector jay leno's visit to the program. leno wanted to know everything about it, so he was thoroughly briefed and got a development-car demo drive, which he loved. then he decided that he desperately wanted to buy the first one. but gm told him, "sorry no sales, lease only," which made him mad. then, after gm steadfastly refused to sell him the first production ev1 (which was still called "impact" at the time), he made on-air jokes about the name: something like, "was crash-and-burn already taken?"
in warm weather (our lead-acid batteries lost substantial range in cold temperatures), i could drive ev1s nearly 60 miles home from the proving grounds and had 240-volt charging equipment there to get me back the next morning. one hot summer evening, i left in a car with an early nimh battery pack and, because nimh essentially doubled the range of the standard lead-acid, i felt comfortable diverting to a dinner in ann arbor. but i ran short of range on the way home that night and barely made the last several miles in painfully slow "limp home" mode with the lights off. that triggered an investigation that revealed that nimh batteries lost energy (and range) when hot and led directly to reengineering the ev1's battery tunnel to provide air cooling for the 1999 model's optional nimh packs. which, for that reason, were not offered in arizona.
achieving such surprisingly good dynamics on michelin’s skinny 50-psi low-rolling-resistance tires of the day was a substantial challenge. "the ev1 was difficult to tune because the battery pack put so much mass in the center," clive reports. "the first cars were a bit soft and floaty, probably because i put too much emphasis on the notorious california freeway hop and allowed the rear end too much vertical motion to avoid a feeling of tugging over the waves, thereby degrading the precision feel. they also had large-diameter rear dampers with a valve system using multiple discs, which gives an immense range of tuning choices, many thousands of possible combinations, so you took your best guess at the end of the available time. and the composite in the aluminum-composite rear axle links was flexible enough to give the rear axle a disturbing springy, non-precision feel in some dynamic situations. we couldn't change the links but were able to compensate to some extent. the later cars switched to a smaller, simpler damper using a spring-loaded valve with a large but finite number of options, and with this and better understanding, they felt more nimble and secure.
pch, pikes peak, and laguna seca
garrett recalls our la press drive, where all the ev1s there for the event took a long, slow drive through marina del rey and down the pacific coast highway into malibu the day before to maximize their range gauges in advance. "folks must have thought aliens had landed when they saw this caravan of unknown jellybeans going by," he laughs. and the day a group of police drove ev1s at milford. "i rode shotgun with several who had never driven an ev other than a golf cart," he relates. "one said he liked the acceleration and would want one for hunting drug dealers because its quietness would allow driving right up to them without their hearing it coming."
after the car was in production, clive and marty made a reconnaissance trip to pikes peak to explore the possibility of a competitive run up that famous mountain road. clive reported that if we should ever enter a vehicle, we should use a specialist driver. "it was no place for a beginner!" both also recall that clive lost his lunch on the way down. "i tried to kill the windshield camera before the altitude sickness took over," marty laughs.
and one of my fond memories was giving fast rides around monterey, california's laguna seca racetrack (which i knew well) in ev1s to 1997 ted (technology, entertainment, and design) conference attendees. it was great fun showing off their hot-lap handling, though we needed a couple of cars to do it since their batteries drained quickly at racetrack speeds.
in 1994, clive roberts, in his capacity as lead development engineer, took a prototype to the 7.7-mile fort stockton, texas, tire-test oval track. the impact electric development car he piloted—shown at the top of this story—was artfully modified and meticulously prepared throughout to enable an ev speed record. it had more power, supertall gearing, a lowered and stiffened suspension, and aerodynamic aids (including a long, wake-smoothing tail cone) that reduced its drag coefficient (cd) from 0.19 to an astounding 0.137.
yet it had stock brakes. "this was a straight-line activity, so we didn't want any extra mass," roberts relates. "but we were leaving the measured mile somewhere above 190 mph, and there was not a lot of room before needing to turn into the banking to complete the lap. the car would slow to about 160 mph before the brakes were completely gone and i had to turn and keep it in the top lane. adding to the excitement was the lack of guardrail, just a lot of texas over the top. when we got the car back to michigan, we found the brake pedal substantially bent!"
the process required two runs through the measured mile in both directions within an hour. "we set up the mile on the main straight," he continues, "then did a series of test runs to find the best point to start and the best entry speed onto the banking. we could do just one run before changing the battery. entry speed to the banking was in the 175-to-180-mph range. too slow would give insufficient time to accelerate to maximum speed for the trap; too fast would waste energy scrubbing off speed on the turn. and even the slightest lift would lose 2 to 3 mph."
but on march 11, 1994, clive and his crack support team got it done to the tune of 183.822 mph, a record for "street legal" evs that (we think) held until 2016 when it was topped by a battery powered corvette at 205.6 mph. and achieving that record for a (future) production electric car was greatly aided by its ground-breaking potpourri of lightweight components, including its bonded and riveted aluminum space frame (with some composite structural elements), aluminum control-arm front suspension, cast aluminum wheels and cast magnesium seat frame and steering column support. the 1997 production ev1 that emerged two years later weighed just 2970 pounds, of which some 1200 pounds was its 16.5-kwh pack of 27 advanced lead-acid batteries (26 propulsion, one for accessories).
it’s difficult to compare energy usages then to now since the measuring process has changed, but marty (who was in charge of efficiency development as well as nvh) recalls that our 1994 preview drive cars delivered something like 5.6 miles/kwh on average. that compares to 4.5 miles/kwh for a tesla model 3 long range and 4.0 for a chevy bolt today at their epa combined range figures, and closer to two miles/kwh for soon-to-come big truck evs. that '97–'99 ev1, while a marketplace failure due to high cost, low range, and just two seats, was a true technological triumph.