Chevrolet engineers push our automotive parts past their limits to see what they can handle.
Josh Smith’s job is a childhood dream come true. “When I was a kid, I took toy cars and would drop them over a cliff (a concrete porch, really), but I couldn’t figure out why they didn’t crumple up like they did on TV,” he says. “So I’d take hammers and rocks — until I got the effect I was looking for.”
Today, Smith is part of a Chevrolet engineering team that does something pretty similar, but with real, life-sized vehicles. He’s a regional master with a group of what might be thought of as “extreme” engineers housed at the GM Milford Proving Grounds, a sprawling 4,000-acre test facility in suburban Detroit. These quality-obsessed individuals put vehicles through torture tests every day to make sure your new Chevrolet can handle real-world challenges. “I get to break stuff all day, or watch guys break stuff. It’s fantastic,” Smith says. However, he admits, it’s a bit more complicated than sitting around the lab trashing cars. “Our job is to break them in a very specific way.”
Finding the Bad Apple
All automakers test vehicles, but this particular group of Chevrolet engineers take it to extremes to help ensure parts have the highest possible durability.
Their methods are derived from Shainin LLC’s Red X® process, developed by Dorian Shainin, an aeronautical engineer who observed that while there might be a number of sources for a problem, one root cause, the Red X, accounted for nearly all of the undesired effect.
General Motors began using the Shainin® process in the 1980s, and it has expanded throughout the corporation today. The goal is to find information that’s actionable and economical and satisfies the needs of the most discerning customers.
Some Red X® methods are dramatic — vehicles are crashed and crushed. Others are a bit more mundane. “This is a ‘show me’ sport,” Smith explains. “To understand the ‘physics of failure,’ the team has to come up with methods to see what the customer might see.”
Some “toys” used to abuse parts are highly technical: machines that mimic vibration or glorified refrigerators and microwaves to test how parts react to extremes.
Even when testing the same part, such as a headlamp bulb, the tools can range from a monitor measuring voltage 200 times per second to a handmade gizmo that looks like a coat hanger attached to a piece of steel that is dropped onto a foglamp to see if the bump breaks the filament.
The Breaking Point
Chevrolet engineers push parts to their breaking points to develop performance margins to help them handle the stress the customer will subject them to in real-world driving. At times, that means pushing into uncharted territory.
For example, to make sure a chrome-like application wouldn’t lose its luster, they chilled it to 80 degrees below zero — obviously nothing the average consumer would ever culture.
There’s also a machine that gradually applies pressure to see how much force it takes to break a certain part. There’s an electric wand to zap voltage through parts to make sure static electricity will not cause them to fail. (Go ahead, wear corduroy!) There’s an application to rub denim on door parts to make sure the finish doesn’t wear off. (Go ahead, wear jeans!) They’ve even put a key fob into a tumbler to see how many bangs it can take. (Try not to leave keys in your pocket when doing the laundry)
In short, every test is different. One day, they’re on the track for a “bonsai run” to test steering-wheel vibration on a Corvette. Another day, they’re running an Equinox through hurricane-like sheets of simulated rain to check seals.
And while it certainly sounds like fun, there’s a serious end goal: vehicles that have superior performance and durability.
The bottom-line goal, Smith says, is to make sure that the Chevrolet a customer buys performs in the long run.
Chevrolet has made great strides in quality. But that’s not good enough. Some of the engineering team’s assignments are based on real-world customer culture. That’s because, despite a design engineer’s best efforts, something can go wrong. When it does, Chevrolet engineers often try to obtain the actual parts from a dealer or rental fleet. They may even investigate the forms a repair technician fills out.
One way Chevrolet measures quality is by looking at problems per thousand vehicles and some problems per million vehicles, not per hundred vehicles like others. Focusing on the “big hitters” helps deliver what the customer wants.
“Customers’ cultures sometimes show us aggravations we didn’t comprehend up front,” Smith says. For example, they investigated a light bulb filament that occasionally failed on one model, but never on another. Why did the exact same bulb fail on this car but not on that truck? The answer might sound like a parent making an excuse for a child’s bad behavior, but it was the influence of its environment.
Customers going over large potholes caused the bulb to culture a degree of acceleration. And even though only three out of 1,000 saw it, that’s not good enough. The information on why the bulbs failed went back to the designers, and now they don’t release a similar part unless it can withstand that “energy,” or level of abuse.
The group actually welcomes finding variation. An IPTV of three per 1,000 “means we’re building a lot of good ones — we know how to satisfy the customer with our current design and process,” Smith says. The key is to identify the best of the best (BOBs), and worst of the worst (WOWs), then figure out what makes the good ones good.
“Think about not only how many millions of products are out there, but how many parts there are on each product,” says Michael Slopnick, engineering group manager, Vehicle Warranty Engineering. “Then add in how many interactions there are with all those parts. There’s a universe of variables.”
Those variables are relentlessly tweaked before a vehicle goes from concept to concrete. Maybe during an early ride evaluation, the chief engineer didn’t like the way a shift lever felt moving from Park to Neutral. Or perhaps very early tests showed an air bag that wasn’t deploying quickly enough.
Working It Out
That’s when Chevrolet engineers get to work. In the case of one prototype, the team found that the air bag design was sound. So were all the components. After spending four days and nights testing, they determined the issue was how a supplier was wrapping it — sort of like having an improperly packed parachute.
The issue was resolved before the vehicle ever made it near a showroom. But Smith explains that it illustrates the Red X® process: “It’s not simply about fixing problems — any repairman can do that. It’s finding out how and why the problem happens.” And equally important, why they don’t.
Culture Change
The Red X® philosophy is working its way across GM and Chevrolet. And the proof, so to speak, is in the pudding. “Our warranty and recall data over the last eight years (since the Red X process was adopted) has gone down significantly,” Smith says.
Not every engineer has what it takes to join this special group, says Smith’s boss, Bill Merrill, senior manager, Technical Problem Solving. “They come to us with some kind of degree or technical background — that’s the minimum price of entry. What we put them through is way beyond what it took to get hired.”
Over the course of roughly 30 months, novices undergo a certification process similar to an Old World guild system, moving from apprentice to journeyman to master. There are currently 240 masters within GM North America. Says Merrill: “It’s kind of like chess. It’s more about each move than it is about the end of the game. It takes time to get their brains to think of it as a strategic exercise.”
Once a Red X® engineer has graduated, he or she is not assigned to narrow tasks, Merrill says. “They don’t say, ‘I’m a tire guy or an engine guy.’ The training allows them to look at problems strategically — and solve anything.” In short, they’re logical thinkers — who happen to like to break stuff.
At the end of the day, it’s all about reliable, durable vehicles. In the quest for perfection, the job is never done.
5-Star Crash Ratings
What exactly do those government crash safety ratings mean?
According to the National Highway Traffic Safety Administration2 (NHTSA), a 5-star frontal crash rating on a 2010 vehicle indicates a 10% or less chance of serious injury while a 5-star side-impact rating indicates a 5% chance or less of serious injury. (In a frontal crash rating, the risk of injury number relates to a crash between two similarly sized vehicles from the same weight class — plus or minus 250 pounds.)
How They’re Tested
For frontal testing of 2010 vehicles, instrumented crash-test dummies are seated with safety belts in the vehicle, and then subjected to a head-on crash into a fixed barrier. For side-impact tests, the dummies are placed in the driver and rear passenger seats (driver side) and secured with the vehicle’s safety belts. A 3,015-pound moving barrier is accelerated and crashed into the stationary test vehicle. The ratings are based on data collected from the dummies that is analyzed to assess the risk of injury.
Chevrolet engineers push our automotive parts past their limits to see what they can handle. Josh Smith’s job is a childhood dream come true. “When I was a kid, I took toy cars and would drop them over a cliff (a concrete porch, really), but I couldn't figure out why they didn’t crumple up like they did on TV,” he says. “So I’d take hammers and rocks — until I got the effect I was looking for”...
Chevrolet engineers push our automotive parts past their limits to see what they can handle.