GM's hit and run: How a lawyer, mechanic, and engineer blew open the worst auto scandal in history

By Adam L. Penenberg , written on October 18, 2014

From The News Desk

As the sun was setting on a stormy Georgia day, Brooke Melton was 30 miles outside of Atlanta in her Chevy Cobalt. It was March 10, 2010, her birthday, and the 29-year-old pediatric nurse was on her way to her boyfriend’s to celebrate.

Melton had purchased the white GM Cobalt in 2005, the year the four-cylinder compact first rolled out of factories, and lately it had been giving her trouble. A week earlier the engine had unexpectedly shut off. Melton managed to pull over to the side of the road and restart it, but the incident shook her. She phoned her father, who advised her to bring the car in to the local dealership. So she wouldn’t forget, Melton scribbled a list of the problems in a notebook: “Key locking in the ignition,” she wrote. “Suddenly shutting off while driving and unable to turn vehicle.” Under “strange knocking sound” she underlined “ignition problems.” Mechanics at the dealership assured her nothing was wrong, and after cleaning the fuel injection gave Melton back her car a few days later with a clean bill of health.

As dusk bled into darkness early the next evening, she was driving north over a stick-straight section of Highway 9 at 58 mph when it happened again. The Cobalt’s engine shut off and the lights inside and outside the car dimmed. Melton hit the brakes, but no power from the engine meant no anti-lock brakes and no power steering. The car fell into a skid. Tires squealing, the Cobalt’s back end fishtailed, coming up on her left. Melton instinctively spun the wheel counterclockwise.

Three and a half seconds after the engine quit, Melton was the reluctant driver’s seat passenger of a car hydroplaning sideways across the centerline. In the southbound lane, bearing down at highway speed, was a gray Ford Focus driven by a 26-year-old man from nearby Acworth, GA, his two-year-old daughter strapped in the back.

The Focus plowed into the passenger side of Melton’s Cobalt: 3,000 pounds of steel, glass, plastic, and human smashing into 3,000 pounds of steel, glass, plastic, and human. While Brooke’s lap belt glued her waist to the seat, her shoulder harness went slack the instant the engine shut off. As the side of her car caved in, Melton’s torso, neck, and head whipped violently to the right, the force equivalent to falling from the 16th floor of a building. The Cobalt spun around and was heaved 15 feet down a hill, ending up backward in a creek swollen with rainwater. Rescuers found Melton slumped over the steering wheel, her body submerged up to her shoulders.

Brooke Melton was not the first to crash her GM car after the engine stalled. Over the course of a decade, dozens of people died and scores more were injured on American roads. Four months earlier, Hasaya Chansuthus was heading home to Nashville in her 2006 Cobalt when she sideswiped another car, her engine shut off and the Cobalt raced off the highway. She was killed when her head struck the steering wheel. In 2009, an 11-month old baby was paralyzed and his grandmother and aunt were killed in Pennsylvania after the Cobalt’s engine turned off without warning and was hit by another car. Earlier that year, 81-year-old Marie Sachse lost control of her 2004 Saturn Ion outside of St. Louis and died eight hours later from internal injuries. In 2006, 18-year-old Natasha Weigel was a few miles from home in eastern Wisconsin when the engine quit and the Cobalt shot off the highway, killing Weigel and another teenager.

Then there was the tragic story of Candice Anderson, who was motoring over a quiet Texas road in 2005 with her boyfriend when the engine stalled and she lost control of her Saturn Ion. Her boyfriend died in the accident and police, pointing to skid marks that indicated that she hadn’t attempted evasive maneuvers, charged Anderson, who suffered broken limbs and internal injuries in the crash, with intoxication manslaughter even though a blood test came back mostly negative (there were trace amounts of an anti-depressant). Three years later, Anderson avoided jail by pleading guilty to the lesser charge of criminally negligent homicide and paying a fine.

Countless articles have been written about General Motors and its massive recalls earlier this year. What hasn’t been fully told is how GM might have gotten away with multiple counts of consumercide were it not for the efforts of three men: a Georgia lawyer, a Mississippi mechanic, and a Florida engineer.

Working together, they unraveled the mystery of why Brooke Melton and scores of others suffered fatal or otherwise catastrophic accidents after their vehicles shut down without warning. 

What follows is their story.


Beth Melton answered the phone. On the other end of the line was a doctor, who told Beth that her daughter Brooke had been in a car accident. Brooke’s neck was broken and she suffered severe internal injuries. It was unlikely she would survive.

“I kept thinking ‘it’s her birthday,’” Beth Melton said later. “‘This can’t happen… It’s supposed to be a happy day.’”

By the time Beth and her husband Ken got to the hospital, Brooke was gone.

Ken bent down to kiss his daughter’s forehead. He felt a piece of him had died along with her. Overcome with grief, he was also furious. Brooke would never lose control on a road she had driven thousands of times. She had told him about the problems with her Cobalt. If not Brooke’s fault, it had to be the car.

He whispered in her ear: “I will vindicate you.”

Ten months after the Meltons buried their daughter an insurance adjuster contacted them, concerned that the driver of the other car might sue Brooke’s estate. He recommended they confer with an attorney, and gave them a name: Lance Cooper from nearby Marietta, Georgia.

During their initial February 2011 meeting in Cooper’s office, the Meltons produced a recall notice relating to a defect in the Cobalt’s power steering that had arrived in the mail after Brooke died. Despite the conclusions of the police accident report, which claimed that Brooke had been driving too fast for the road conditions and lost control when she hydroplaned over a sheet of water, the Meltons were adamant that the fault lie with the car.

Cooper, 51, believed this was possible. Two decades earlier, fresh out of Emory Law School, Cooper had taken on his first product liability case, involving a Ford Bronco II. A cultural touchstone, the Bronco II was not only infamous for having been the little brother to the Bronco, which ferried a fleeing O.J. Simpson as a battalion of police cars and news helicopters gave chase on national television, it also had a woeful safety record. Prone to rollovers, the Bronco II was so unstable that a 1989 Consumer Reports rated the car’s handling as “poor” in a test that simulated rapid lane changes and urged its readers to stay away. The Institute of Highway Safety called it the most deadly SUV on the road. The Bronco II was so dangerous that federal accident statistics reveal that 1 in 500 Ford Bronco IIs ever manufactured has rolled over and killed at least one person in the car.

“I kept thinking ‘it’s her birthday. This can’t happen… It’s supposed to be a happy day.'”

Despite the Bronco II’s tipsy reputation and woeful safety record, it was never recalled for stability reasons, and Ford opted not to redesign the SUV from the ground up. Instead the automaker instituted mostly cosmetic changes, perhaps the biggest being the name, which changed from Bronco II to the Explorer. To shave costs Ford manufactured the vehicles in the same factory and on the same too-narrow assembly lines as the Bronco. The Explorer was marginally better from a rollover perspective. Only 1 in every 2700 Explorers manufactured between 1991 and 2000 rolled over and killed someone in the car. Yet the Explorer remained top-heavy, its roof so weak that if you flipped one upside down and lowered it to the ground, it could cave in under its own weight.

While Ford banned its own drivers from performing stability tests after an Explorer rolled over on a test track, the company sold millions to an unsuspecting public. Cooper considered himself well-educated and informed, but until his first Bronco II case it had never occurred to him that the U.S. government would allow a dangerously unsafe vehicle to remain on America’s roads for almost 10 years.

“The Bronco II litigation opened my eyes,” Cooper says. Ford “had options to fix the stability problems and chose not to because of the cost. As a result, this unstable vehicle was sold to the public and there were literally thousands of rollovers.”

gmstory2 Brooke Melton scribbled a list of the problems with her Chevy Cobalt in a notebook

After Cooper settled with Ford, more accident cases came his way. Cooper handled Chrysler minivan “Liftgate” lawsuits involving a weak latch in the back that could pop open after impact at speeds under 20 mph, ultimately killing 41 people, mostly children, and sparred with tire companies. He litigated pickup truck roof crush cases and faulty seatbelts. But he didn’t always win. In the mid-1990s Cooper took his lumps on his first GM case, a minor accident that Cooper says caused the driver to lose control, resulting in a Chevy Blazer rolling over. Jurors bought General Motors’ defense: The driver caused the accident, and if the car went out of control it wasn’t a stability problem.

“Up until then, we’d been successful,” Cooper says. “It made me more careful in the cases we took.”

He also concluded that federal safety standards governing automobiles were woefully inadequate. Then again, “it’s the federal government,” he says, “so what do you expect?” Those staffing the National Highway Traffic and Safety Administration (NHTSA) are “well-intentioned, well-qualified people, but they’re usually underfunded and more often than not at the mercy of the manufacturers.”

Which is why he views the civil jury system as critical. Although attorneys are often justly lampooned, litigation has been more effective at shaping responsible business practices than government. It’s why trucks beep when they back up and farm machinery comes equipped with safety guards, why asbestos no longer poisons homes, schools and workplaces, and fast-food restaurants, aware of their super-sized liability, convinced meat packagers to clean up processing plants. When juries speak, Corporate America listens.

Of course, it’s hard to think of trial lawyers as brave foot soldiers in the war against corporate malfeasance. More often they are perceived as corrupt, dishonest, greedy, heartless hypocrites, ambulance chasers who are not only a drain on the economy but may be a threat to national security. Is it a coincidence that a majority of United States Senators list their occupation as attorney? Clad in $3,000 suits, some trial lawyers jet around in private planes, a hundred times richer than any judge. The public is convinced they cry crocodile tears for their crippled clients while gnawing on upwards of 40 percent of the take (plus expenses).

Still, as the old saying goes, everybody hates lawyers, except his own. What’s more, an attorney like Cooper works on contingency and is highly motivated since he only gets paid if he’s successful. If he loses, he gets nothing. He shoulders the risk of developing a case that can run anywhere from hundreds of thousands to millions of dollars and take years to resolve. It’s a sad fact that in the absence of effective government oversight trial lawyers like Cooper have become self-appointed national auto safety czars.

And so it was with Brooke Melton’s case. But before any of this could happen, Cooper needed to find an expert to find out why Brooke Melton’s car had crashed.


Charlie Miller has been a grease monkey for almost his entire life. Born in Tippo, a tiny town in the heart of the Mississippi Delta, Miller overhauled his first engine when he was 12. He put himself through a local college working as a mechanic while supporting an addiction to drag racing and after he graduated he become a professional hot rod driver. In 1974 he opened an auto repair shop in Merigold, but it was 15 years before he was asked to offer an expert opinion, and only because of happenstance. A member of his race team’s pit crew crashed a Ford Bronco into a ditch and the vehicle caught fire. The crewmate escaped the car and toppled into the ditch, where lighted gasoline poured into the water, burning his leg off and part of his hand.

While the man recuperated in a hospital, a friend of his -- an attorney -- retained a fuel systems expert from California to inspect the charred Bronco corpse for a possible lawsuit against Ford. The lawyer asked Miller to tag along, because anything he could contribute could help save their friend money. The out-of-town expert analyzed the fuel system and concluded that the carburetor overflowed and the needle valve stuck. This, he ventured, allowed the fuel to spurt out.

“Charlie,” the attorney asked, “is that what you think happened?”

“Well,” Miller replied, “not really.” He thought it was more likely a design flaw. Ford had routed the Bronco’s plastic fuel line underneath the vehicle, where it could easily get cut. Miller believed the gas leak originated from underneath the driver’s seat.

“Show me,” the lawyer said.

Miller did. There it was.

“Get the cameras back out,” the lawyer said. “We got to re-film.”

After that, insurance companies called, then plaintiffs’ attorneys whose cases Miller had either aided or eviscerated. Over the years Miller has investigated hundreds of car and truck carcasses, and never knew what he’d find until he dug into a car’s mechanicals. In his experience crashes could be caused by any number of things -- driver error, incomplete or incompetent maintenance, poor design, a faulty part, or just plain bad luck. Then came a December 2011 phone call from Lance Cooper, who got his name from a fellow attorney.

Cooper asked Miller to travel to Atlanta to conduct a thorough mechanical inspection of Brooke Melton’s car. When Miller arrived, he encountered one hell of a wreck. The Cobalt looked as if it had been tossed into a huge trash compacter that didn’t get to finish the job. The passenger side and roof were caved in, the front wheel pointed inward while the rear wheel jutted out, white paint was badly scraped, and the back fender hung off the rear like a tail that had been attached in the wrong place. Cooper had filled Miller in on the basic outline of the case. He knew what had happened to Brooke Melton, not that he allowed himself to think about it. He had a job to do.

Once he photographed the car from a variety of angles, Miller got down to business, plugging a diagnostic computer under the dash to determine what was and wasn’t working. He wanted to see if the Cobalt’s wiring had been damaged in the accident. It hadn’t. He checked the engine control module. It was fine. Then he turned to the body control module, a small computer that handles the electric windows, locks, defroster, and heater. It was OK, too. He looked at the antilock brakes, which were completely functional. The front airbag checked out, too. It had not deployed in the crash, but this wasn’t a head-on collision, and Brooke’s Cobalt didn’t come with side air bags. The only module Miller couldn’t access controlled the power steering. Still, both of its power sources were available, and the electric power steering module had voltage getting to it. Despite the recall notice Cooper shared with him, Miller didn’t believe power steering was to blame.

He knew what had happened to Brooke Melton, not that he allowed himself to think about it. He had a job to do.

Upon completing his baseline diagnostics, Miller dug up a copy of Brooke’s handwritten note. Two entries caught his eye: “Key locking in the ignition” and “Suddenly shutting off while driving, and unable to turn vehicle.”

That’s not right, Miller thought. Older cars were equipped with hydraulic assist power steering, so if the engine went dead, it became hard to turn. Not so with newer makes like the Cobalt, which came equipped with electric power steering. Whether the engine was running or not shouldn’t make any difference.

He reread the note. It was, he says, as if Brooke were sending him “a special message” from the grave. Something must have cut off the 12 volts to the system. Pawing through paperwork he found a printout of the data downloaded from the Cobalt’s black box. Virtually all new cars have black boxes, as do many earlier models, which automakers rely on to assess vehicle performance but can also identify safety problems. When Miller started the engine with the ignition switch in the ‘on’ position, however, everything worked fine.

gmstory3 The information captured by the “black box” showed how long before the crash the engine shut off.

One key piece of data involved the car’s speed and the engine’s rpms in the moments before the crash. Five seconds before the accident, the engine was turning at 2,048 rpms and the car was traveling at 58 mph. Four seconds before, the rpms were down to 1984 rpms while the speed held constant at 58 mph. Three seconds before the collision the car’s speed and engine rpms registered zero.

“It’s impossible for a vehicle to go from 58 to zero in one second, and also impossible for an engine to go from 1,984 rpm to zero in one second,” Miller says. “Typically, in a case like this, we see the rpm just slowly diminish. By this going to zero, this told me, as a mechanic, that the reporting status of the powertrain control module — that’s the module that reports rpm and mile per hour — had lost power.”

It was conceivable that something had disrupted the wiring harness. If Brooke had run over something hard and sharp, for instance, it could have shot into the car’s undercarriage and sliced a wire. If that were the case, however, Miller wouldn’t have been able to access that module; it would have been dead. With power put to it, all he had to do was flick the ignition switch to the ‘on’ position and everything worked perfectly. As he thought about it, the only thing that could have interrupted the power was the ignition switch, which must have turned off on its own accord.

Why would that be? When Miller rotated the switch he noted how easy it was to move. It didn’t feel like other ignition switches he had handled. It was like a ballpoint pen with too weak a spring. This lack of springiness, he deduced, could have allowed the ignition switch to jump out of the ‘on’ position with very little pressure.

He had isolated a potential culprit in the form of a defective part, but as far as he could tell it might be nothing more than a single bad ignition switch, a one-off. If 99 percent of 1 million parts a company manufactured were good, that meant 10,000 were bad. Brooke Melton could have gotten one of these. It could be that she was simply tragically unlucky.

Returning to Mississippi, Miller trekked to a nearby salvage yard, forked over $80 for a steering column from a 2005 Cobalt, and extracted the ignition switch. It looked “very similar to Miss Melton’s,” Miller says, meaning it also lacked torque. At his local GM dealer he purchased a new ignition switch for that same 2005 Cobalt. It also appeared identical to the two other AC Delco switches down to the part number (10392423). But when he played with it he found that the new switch possessed significantly more springiness.

To measure the torque would require ingenuity, since there existed no off-the-shelf tool for this, but Miller was an able improviser. He scrounged up his Bass Pro Shops fish scale and hooked it to the ignition key to weigh the force required to turn the key ‘off.’ After trying it several times on the switch from the salvage yard, he swapped it out of the steering column with the new part, which required twice the force to move the switch from the ‘on’ position. Miller snatched an old screwdriver he had ground into a makeshift tool to manually flip the switch on and off. Just as he suspected, the new switch was much harder to turn.

This gave Miller pause. While the ignition switches were manufactured six years apart, GM had given no indication that it had reengineered the part. If it had, the automaker would have stamped it with a new number. Given what he knew thus far, Miller was now sure the defect went beyond Brooke’s car. He thought it was conceivable a bad batch of parts had escaped from the factory. If so, other drivers -- Miller didn’t know how many – could be at risk.

Given the tools at his disposal in his garage, Miller had taken his investigation as far as he could. To learn more, he would need to x-ray the ignition switch from Brooke Melton’s car. Lance Cooper would have to sign off on it and GM’s representatives would have to agree with the testing protocol and be present to maintain the chain of custody and ensure the integrity of the evidence.

This would take time to orchestrate. Automakers are notorious for stretching out litigation. The strategy serves several purposes. The longer it takes, the more it weakens plaintiffs’ resolve and bleeds attorneys, who must front the costs of developing a lawsuit. It also means that plaintiffs sometimes die waiting for their cases to make it to trial. This also benefits automakers, since death cases cost far less than accident settlements, which take into account lost wages over a remaining lifetime plus the cost of providing lifelong medical care.

Indeed, a compensation expert retained by GM valued a deceased 17-year-old student at $2.2 million while a dead married 25-year-old father of two earning $46,000 a year was worth about $4 million. (If that same married father of two earned more, say, $75,000, he would be eligible for $5.1 million). Contrast these proposed payouts with those of a married 40-year-old paraplegic -- GM would be willing to offer $6.6 million – and a 10-year-old paraplegic, who GM estimated could receive $7.8 million.

Miller wasn’t surprised it took seven months before everyone could assemble at his garage to witness his sending off Brooke Melton’s ignition switch to the lab. Unfortunately, the x-rays didn’t reveal much. Miller couldn’t see a break in the spring or damage to the unit.

“Basically,” he told me, “what it told me was there shouldn’t be anything wrong with this switch from a physical failure standpoint.”

There was only one thing left to do: take the ignition switch apart and examine its internal components.

For this, Miller would need help.


McSwain Engineering of Pensacola, Florida is in the business of finding out ‘why.’ The firm’s 40 engineers, experts in failure analysis, comb through the wreckage of planes, helicopters, cars, trucks, even pipelines to figure out what happened. When TWA Flight 800 blew up off the coast of New York in 1996, killing the 230 passengers onboard, a McSwain investigation tamped down rumors of a missile strike by revealing that the fuel tank ignited. McSwain also conjured many of the fixes for the Boeing 737, which over the years has had a spate of mechanical and design problems. While the majority of McSwain’s caseload involves aviation, it has conducted hundreds of vehicle investigations, of which Charlie Miller has referred more than a dozen.

In October 2012, two and a half years after Brooke Melton’s crash, Miller traveled to Pensacola to meet with Lance Cooper and McSwain engineer Mark Hood. Toting Brooke Melton’s ignition switch assembly, Miller bore other gifts, too, in the form of damning GM paperwork. Wondering if GM had issued any recalls and service advisories or bulletins for ignition-related problems, he had accessed a database that mechanics subscribe to for repair information, parts catalogs, and online tools to help mechanics estimate labor and materials costs. It didn’t him take long to locate a technical bulletin relating to an ignition problem that GM sent to dealers in 2006, a year after the Cobalt debuted in showrooms.

Dated Oct. 25, 2006, the curiously worded “Bulletin #05-02-35-007A: ‘Information on Inadvertent Turning of Key Cylinder, Loss of Electrical System and No DTCs,’” covered seven GM makes and models -- the Pontiac G5, Pursuit, and Solstice, Saturn Ion and Sky, and Chevy HHR and Cobalt. It seemed to blame drivers for “inadvertently turn[ing] off the ignition due to low ignition key torque/effort.” GM singled out “short” drivers, as well as those who possessed a “large and/or heavy key chain,” whose “knee would contact the key chain while the vehicle was turning and the steering column was adjusted all the way down.” This affected short drivers, the advisory explained, because they were more likely to “have the seat positioned closer to the steering column.”

gmstory4 GM’s fix involved filling in the slot where drivers could hang extra keys, laminated ID cards, and the like.

Miller didn’t know how many short people in America drove GM cars but he recognized a serious design flaw. The bulletin even referenced the part’s “low ignition key torque/effort.” Its engineers must have signed off on the spring’s tension inside the ignition switch, and they were the ones who decided how close the seat could come to the steering column. Now a whole segment of the driving public faced the risk of the engine shutting off simply because of their God-given height. But that was only part of it. If a knee brushing a key chain was enough force to shut off an engine, so could hitting a pothole, running over road debris, or a minor fender bender. He wondered what fix GM would offer to fix that.

It quickly dawned on him the automaker didn’t plan to fix anything. Instead, GM instructed dealers to advise customers complaining of unexpected engine shutdowns to remove “unessential items from their key chain,” and provided a free insert for the key ring to prevent users from adding additional keys. “As a result,” the bulletin read, “the key ring cannot move up and down in the slot any longer – it can only rotate on the hole.” GM threw in a replacement key that didn’t jut out as far as the original, so that driver’s knees would be less likely to jostle the keys when they were in the ignition. Per recipient, these solutions, if they can be called that, ran GM under a dollar – less than the cost of a child’s birthday goody bag.

Continuing to dig, Miller tracked down another technical service bulletin in the March 2012 issue of MOTOR, a trade publication. In the section “Service Slants” the magazine compiled technical bulletins from manufacturers. One from GM addressed ignition keys that were hard to turn and could get stuck, a problem the automaker attributed to the lock cylinder. In her note, Brooke Melton had mentioned this very problem. The fix required installation of a redesigned lock cylinder, and was intended for the same cars listed in the other technical service advisory.

Miller wondered if GM was attempting to cover up the ignition switch’s low-torque problem by surreptitiously changing parts. With the tools and techniques available in McSwain’s spic-n-span lab, which housed electron microscopes, 3-D printers, and a dizzying array of special tools, he was confident that he and Mark Hood, who Miller had worked with on several other cases, could reverse engineer the ignition switch to find out what went wrong.

Even as a child, Mark Hood, now in his 50s, had a knack for taking things apart and being able to put them back together. After graduating from Auburn University in 1985 with a bachelor’s degree in materials engineering, he worked at the Naval Aviation Depot in Pensacola, where he tested parts used in the Navy Marine Corps. helicopter fleet, performing aircraft component failure analyses, accident investigation, materials processing, and nondestructive testing. Eventually he was promoted to Head of the Metallic Materials Engineering Branch before moving to a maker of automotive castings, working in the nondestructive testing program until joining McSwain in 1995.

GM had switched the switches but never told anyone, leaving millions of unsafe cars on the road.

The first task that Hood and Miller undertook in the McSwain lab was to confirm Miller’s torque measurements for the new ignition switch versus an “exemplar” from a 2005 Cobalt. Hood’s tools were a far cry from the simple fish scale that Miller relied on for his measurements. After constructing a fixture so that he could mount the ignition switch assembly and turn it on, he used another tool to measure and record the force required to turn the key. He arrived at the same conclusion: the old switch measured about half the torque of the newer switch.

“When you’re doubling the force, to me, from an engineering perspective, that’s significant,” Hood says.

Then Hood calculated the torque on the switch from Brooke Melton’s Cobalt. First he looked at the amount of force required to physically move the switch from one click position to another, then measured it electronically to spot any variation. There wasn’t. He and Miller built a light box so they could see, electrically, when the switch changed positions. All of this confirmed that there was a significant difference in torque between switches from 2005 and the newer ones. When Hood replaced Melton's ignition switch with a newer, store-bought switch, the new part worked much better.

From here, Hood, over several months, collected steering columns with ignition switches from salvage yards and GM dealers -- three from 2005 Cobalts, two from 2006, 2007, and 2008, and eight newer switches (post-2009). Hood also scrounged up switches from Saturn Ions, which were also listed on the technical service advisory that Miller found. Like other automakers, GM used the same part across other lines of cars. And sure enough, Hood found that the switch in the Ion was identical to the one in the Cobalt, although it was stamped with a different number.

Now it was time to cut into these switches to see how they worked. At McSwain, Hood had a machine shop and a capable machinist at his disposal. They relied on a precise technique to cut a half-inch by half-inch square into the side of a 2005 exemplar switch, because Hood didn’t want to do anything that could alter its inner workings. Peering inside at the “detent” plunger, he could see how it interacted with the plate inside the housing. Hood found that it took very little force to induce the switch to turn on and off. Then he took Brooke Melton’s switch apart and measured the components inside, particularly the spring force on the plunger. It was exactly the same.

Comparing the 2005 switches to those from 2006, 2007, 2008 and beyond, Hood noted significant differences. The springs in the newer switches (manufactured after 2008) were not only springier, they were longer, and so were the tubes holding them.

GM had switched the switches but never told anyone, leaving millions of unsafe cars on the road. Then someone or somebodies at the company tried to cover it up.

Now that attorney Lance Cooper knew what to look for, he filed a revised complaint almost three years to the day after Brooke Melton died. He accused GM of negligence in designing, testing, and manufacturing, as well as failing to warn consumers. Months earlier, Cooper had begun papering the auto company’s attorneys with requests for information on ignition switches, including any and all lawsuits and documents relating to the modified key in the service advisory. GM responded in January 2013 with a heap of material but objected to several discovery requests, such as drawings of the original switch and any revisions, until a judge ordered the carmaker to hand over “all responsive documents and materials.”

Cooper targeted 15 engineers to question, and their answers helped him fill in key pieces of the puzzle. In one deposition, GM’s head switch engineer on the Cobalt, Raymond DeGiorgio, admitted that he recognized differences between the original and replacement switches but couldn’t explain why it had been changed without GM or Delphi, the parts maker, modifying the identification number. In further testimony, Cooper learned of a GM engineer who had experienced the ignition shutdown problem during a test drive in 2004. Shortly after, in 2005, GM's own engineers concluded there was a problem with the switch.

GM changed the design of the ignition switch but kept the same part number. GM changed the design of the ignition switch but kept the same part number.

Perhaps the most damning testimony came from Gary Altman, GM’s program engineering manager for the Cobalt in 2004 and 2005. When Cooper asked if GM had “made a business decision not to fix this problem” then five months later sold Brooke Melton “a vehicle with the problem,” Altman replied, “That is what happened, yes.”

Three months later, GM and the Meltons settled for a reported $5 million. GM, Cooper says, never did produce all the documents the judge ordered.


Brooke Melton needn’t have died that night. She was killed by a corporation’s callous disregard for the safety of its customers, made worse by a regulatory agency reluctant to regulate. At least 26 others perished, and scores more were injured, and these numbers will almost certainly grow. Reuters, after sifting through government accident data, estimated that 74 people have died in crashes that reflect “key similarities.” Thus far, General Motor’s victim fund has received 1,330 filings seeking compensation, including 165 from families of people killed in GM cars.

At General Motors, the problems go back to the fall of 2002, when Ray DeGiorgio, an engineer who had been with the company for 23 years, made a fateful decision: He approved a switch that fell well below GM’s own specifications, specifically as it related to torque.

Evidence suggests he knew full well what he was doing. Earlier that year, an engineer from GM’s part supplier – Delphi -- had emailed DeGiorgio, informing him that the torque in the switch was insufficient. Delphi could increase it but this would risk other problems, including premature wear as well as potentially affect the switch’s electrical functions. The cost would be “nominal” but would take time to create, test, and validate the new design.

Faced with a deadline, DeGiorgio replied: “If increasing” the torque “will destroy the switch than (sic) do nothing. Maintain present course. Under no circumstances do we want to compromise the electrical performance of the switch.”

He signed the email “Ray (tired of the switch from hell) DeGiorgio.”

gmstory6 Email exchange between a Delphi engineer and Raymond DeGiorgio.

By this time, DeGiorgio had been working on the switch since 1999, and had encountered a multiplicity of challenges. The prototype was so glitchy its entire electrical system had to be overhauled. Then there was its low-torque, which DeGiorgio erroneously assumed wouldn’t harm a car’s performance. After the switch was put into production an entirely different complication surfaced. Complaints flooded GM when cars wouldn’t start in cold weather – an embarrassment for a company from Detroit.

Nevertheless, the ignition switch became standard equipment in the Saturn Ion, Chevy Cobalt, and other GM makes and models. It was a problem from before day one when GM employees, members of the automotive press, and customers discovered the engine would stall if they knocked the ignition keychain with a knee or drove over rough terrain.

A New York Times auto reviewer in June 2005 described his wife’s frightening Chevy Cobalt experience: “She was driving on a freeway when the car ‘just went dead,’ in her words. She recalled bumping her knee against the steering column just before the car shut off.” He brought the car into his local GM dealership, where mechanics found nothing wrong. Then he searched for other complaints and found a newspaper review dated two weeks earlier from The Daily Item of Sunbury, Pa: The reviewer, Gary Heller, wrote: “Unplanned engine shutdowns happened four times during a hard-driving test week. I never encountered anything like this in 37 years of driving. I hope I never do again.”

DeGiorgio kept the same part number for the revamped switch but never told anyone.

Despite the negative media coverage and lawsuits that began trickling in, no one at GM did anything. A company attorney told the New York Times that moving stalls in the Cobalt did not pose a safety risk. GM’s engineers assumed the faulty switch wasn’t dangerous because, as another GM engineer, Gary Altman, testified in a deposition, the cars “could still be maneuvered to the side of the road.”

What these engineers didn’t realize was that the Cobalt and Ion had been designed to shut off the airbags when the ignition key was moved to the “off” or “accessory” positions. They categorized the problem as one of “customer satisfaction,” not safety. The nation’s top auto regulator would later parrot this line of argument: “If a consumer can pull a car over to the side of the road and restart it,” David Friedman testified before Congress, safety is not a problem.

That didn’t mean no one inside GM was paying attention. A group of engineers tasked with safety issues replicated the moving stall problem in the Cobalt when GM’s Director of Product Investigations shut off the car with her knee. Afterward, a variety of GM committees proposed fixes but management rejected them all because they were deemed too costly because “none of the solutions represents an acceptable business case.” Instead GM issued a technical service bulletin to its dealers, advising them to tell customers who complained to remove extra objects from their key chains or plug the space with a free insert.

DeGiorgio viewed the cold weather start snafu a personal affront, and decided to revamp the switch. But he didn’t put much stock in the moving stall complaints because he assumed they would be fixed in the new design. He was right, since engine shutoff complaints practically disappeared after the new switch was introduced. Unfortunately, DeGiorgio kept the same part number for the revamped switch but never told anyone. This stymied several investigations into the switch, but when investigators confronted DeGiorgio, he claimed to remember nothing about any changes -- like he had in his deposition. (Later a document authorizing the changes surfaced with his signature on it.) GM, a global corporation with a market cap of $50 billion, never did figure out what was wrong with the switch.


gmstory7 Memo re: “Cost estimate to change the vehicle key for the Cobalt.”

As a scathing report commissioned by GM into the Cobalt switch debacle concluded, it took “an outside expert working for a plaintiff’s attorney” to take apart two switches to find “the cause it took GM years to determine.” That “expert” was actually two experts – Charlie Miller and Mark Hood, working for Lance Cooper.

The Melton case spooked GM’s legal team. The first time GM learned that the ignition switch had been altered in post-2008 vehicles was when Cooper provided the evidence to the automaker in April 2013. Three months later, an outside counsel to GM who evaluated the Melton case informed the company that the case, from the automaker’s perspective, was a “very poor trial candidate.” The plaintiffs “would argue that GM had known of the defect from the time that Cobalts first ‘rolled off the assembly line’ and yet it has ‘essentially done nothing to correct the problem for the last nine years.’” To avoid a “substantial adverse verdict” he recommended that the “case needs to be settled.”

While the Melton case settled in September 2013, GM, moving at LA rush hour traffic speed, took five months before recalling 800,000 Chevy Cobalts and Pontiac G5s. Cooper was incensed at the puny size of the recall, and dashed off a letter to the National Highway Traffic and Safety Administration (NHTSA).

“GM has failed to fulfill its legal obligation to report this defect to NHTSA within five days of discovering a defect,” Cooper wrote. “Nor has GM met its duty to include all defective vehicles in the recall. This recall is nine years overdue, and incomplete.”

A week later, GM added to the tally: 600,000 Chevrolet HHR, Pontiac Solstice, Saturn Ion, and Saturn Sky vehicles, and in March threw in another 824,000 cars GM sold in the US between 2008 and 2011. The following month GM found another flaw in the ignitions of the 2.6 million vehicles it had already recalled, and as a remedy announced it would replace an additional part.

With public and congressional pressure mounting, the automaker, under its new CEO, Mary Barra, claimed in a Congressional Hearing that it was turning over a new leaf, commissioning an investigation by attorney Anton Valukis, who lambasted the company for a pattern of “incompetence and neglect.” It absolved upper management of direct blame for the switch debacle but pointed to a see-no-evil, hear-no-evil corporate culture that valued inertia.

“Over a decade,” the 315-page report stated, “GM personnel failed to search for, share, or gather knowledge and that failure had serious consequences. There are multiple components to these failures, involving individual mistakes, organizational dysfunction, and systems inaccessible to some and impenetrable to many.”

The report mentions the “GM Nod,” which the company’s newish CEO, Mary Barra, explained was a staple of GM managers, nodding in agreement at steps that should be taken then doing nothing, and the “GM Salute”: where employees would sit through meetings with their arms folded, as if to say responsibility lay with anyone else but themselves.

Now, suddenly, GM has become proactive over safety. Over the course of the year the automaker has earmarked $1.3 billion to cover the cost of recalling 30 million other vehicles for a litany of troubles -- from defective seat belts (Saturn Outlook, GM Acadia) to faulty wiring (Corvette, Buick Lacrosse), software programming errors (Cadillac), shoddy switches (Camaro), busted airbags (Pontiac Vibe, Cadillac Escalade), and defective windshield wipers (Cadillac CTS).

It’s an unprecedented feat. Time calculated that if you took all of the cars that GM has recalled this year and lined them up bumper to bumper, you’d end up with a line that would wrap around the earth four times.

Stack them and you’d reach halfway to the moon. (Some, no doubt, would call that a good start.)


I phoned a spokesman for GM, James R. Cain, who called the report the automaker commissioned “sobering,” and told me the chastened company has truly changed. He claims safety is now the automaker’s number one priority, and ticked off a number of recent initiatives that GM has undertaken to streamline its decision-making -- reorganizing departments, beefing up safety inspectors in the field, encouraging better vehicle integration, and ensuring that problems are addressed before any car is sold. The automaker went back to the early 1990s to see if earlier models had faulty switches, and when it discovered some did, it issued more recalls.

“We want to make sure nothing like this ever happens again,” Cain says.

As for Raymond (“tired of the switch from hell”) DeGiorgio, he was placed on paid leave for two months before he was fired, along with 14 other GM personnel, including his colleague, Cobalt engineer Gary Altman. As the justice department continues to investigate it’s not out of the realm of possibilities that he and others at GM could face criminal charges.

There remains, however, a clear and present danger. Slightly more than a million of the 2.6 million recalled GM cars have been brought in for repairs, with a third of them belonging to drivers with children. This is a serious, potentially deadly problem. In March, there was a death linked to the faulty ignition switch in a car that had already been recalled, but GM didn’t have the parts in stock yet.

The company has sent multiple mailings, called car owners at home, taken out ads, and prodded dealers to do their own outreach as well as using social media to get the word out. GM employees have even been showing up at the homes of recalled car owners, offering free loaner vehicles while they effect repairs.

“Recalls are inconvenient,” Cain says, “but we’re trying to make it easy as possible for people, and we can do the job fairly quickly.”

Sadly, this new embrace of safety comes four years too late to help Brooke Melton, who died in a car accident that never should have happened. Her family is trying to reopen her case, alleging that the company acted fraudulently in concealing the ignition switch flaw. If DeGiorgio had not lied, Cooper argues, the Meltons would not have settled. While the link between the ignition switch and the airbags at a time when a driver and passengers need them most has received the most attention an airbag wouldn’t have saved Brooke Melton. She wasn’t killed in a frontal collision when the airbags didn’t deploy. She was T-boned at 60 mph, her car had no side airbags, and the Cobalt received some of the lowest ratings in side-impact crash tests.

If Lance Cooper is successful – a trial date is set for April 2016 -- it could lead to other settled cases being re-filed, complicating GM’s attempts to put the whole sordid, potentially criminal, mess behind it.

And the Meltons wouldn’t have it any other way. This time they say they won’t settle. They want their day in court. They want the world to know what GM did to their daughter.

[Photo credits: Feature image of Brooke Melton’s post-crash Cobalt and photos of her handwritten note, the printout from her Cobalt’s black box, car keys, and comparison of old and new ignition switches courtesy of The Cooper Firm; other images via the Valukis Report.]