AI Online


The Epsilon Solution

GM's fresh interpretation of the global-car gambit aimed at winning back lost mid-size customers.

General Motors has had quite enough of the punishment Honda, Toyota and others have been doling out for years. Throughout most of the 1990s, GM’s mid-size car market share bombed — from nearly half of that critical segment’s volume to less than a third. Since 1988, GM has squandered some four million sales to competitors. Something had to be done to keep the world’s largest vehicle manufacturer from devolving into a pickup truck and SUV specialist.

The something is Epsilon, a massive initiative that will eventually involve 1/6th of the cars GM builds around the world or 1.4-million units annually. Eight brands will sell Epsilon products. The program will eventually encompass over a dozen distinct body styles, up to eight manufacturing plants on three continents, stickers ranging from below $20,000 to over $40,000, and approximately 10,000 employees. European plants are already working overtime to fulfill demand for the first models out of the chute, Vectras sold by Opel and Vauxhall and Saab’s 9-3.

Epsilon is certainly the grandest worldcar umbrella in GM’s history but it’s hardly the first trip down global lane. In 1976, Chevrolet tried to stave off the Japanese with its Chevette, a subcompact spun off an Opel-engineered T-car platform. Two decades later, Cadillac took its turn with the Catera, a mid-size sport sedan riding on Opel Omega underpinnings. Both of those projects were miserable failures. Nevertheless, GM is betting on the global approach again, though this time with an obvious difference. Since Epsilon is a hundred times grander than any previous international project, the risks involved are staggering, even to GM.




Chevrolet’s Malibu rides on a 2,700 mm wheelbase version of the Epsilon platform. A long wheelbase version (Malibu Maxx) stretches to 2,852 mm by using different front and rear sections mated to a common center floor panel. All Epsilon cars share a front strut suspension and four-link rear with identical mounting points and wheel travel. Suspension tuning takes place through bushings, spring rates and shock calibration.

To find out how Epsilon can succeed in the wake of global car designs that failed, we tapped two experts on the subject. One of them, GM vehicle line executive Gene Stefanyshyn, was not only present at Epsilon’s 1997 creation, he’s also the architect of several of GM’s key global strategies.

Formulating the Epsilon solution began six years ago. Stefanyshyn notes, “The planning began in 1997 among German, Swedish and American engineers at the (Warren) Tech Center. The first step was acknowledging the Catera was not the way to do a global car design. We realized that bringing all parties to the table early was essential to success. We found that everybody’s got to be in the mood to make love.

It’s hard to gain someone’s attention when their next model change is five years out; they’ve got too many other things to worry about that are more urgent. So cadence is critically important.”

The timing plan for every Epsilon-based product was laid out at the very beginning of the project. Knowing exactly when each of the vehicles would roll-out globally was a major enabler for a smooth transition from product to product.

“Opel was first in line,” Stefanyshyn says. “So the Epsilon task force relocated to GM’s International Product Development Center in Russelsheim, Germany, but not before we had debated key elements such as basic dimensions, safety targets and suspension layouts. Opel started filling in the details with our guys in attendance. When the timing was right, Saab engineers moved to Germany to participate in the learning and to tailor Epsilon to their specific needs.”

Somewhat later, additionalU.S. engineers relocated to Europe. Lyndon Schneider, program engineering manager for Malibu and 20 or so others were in Germany for nearly three years.

Another early but critical task was defining exactly what is meant by a global car. Some makers use that term to mean a global brand selling the same products in far flung markets. But for GM, global does not constitute what Stefanyshyn likes to call a “Coca-Cola car.” Instead, Epsilon is a common architecture capable of yielding regional products. That’s because GM’s past attempts at the Coca-Cola approach have not worked.

“Living and working in various markets are essential to identify the major differences from one locale to another,” Stefanyshyn stresses. “Taxation, government regulations and customer tastes change when you cross borders and continents. So the GM plan is leveraging one very good global architecture and allowing various regions to tailor the design to suit their particular markets. The goal is achieving a high degree of technical competence rather than selling exactly the same car worldwide.”

Opel’s Vectra was the first Epsilon-based car in production, but the rollout plan for every iteration was decided before work on Vectra ever began. The car was engineered in GM’s International Product Development Center in Russelsheim, Germany. Engineers from all divisions selling Epsilon sat in during development.
The Epsilon body structure consists of a common center floor panel that mates to two different front sections and four distinct rear sections. That yields a total of four different wheelbases. Saab uses the 2675 mm dimension, the Chevy Malibu and Opel Vectra are at 2700 mm, the extended wheelbase Opel is 2840 mm, and the longer Malibu (five-door Maxx) is 2852 mm. GM says the wheelbase could be stretched a bit more if necessary and it can increase the wheel track dimensions if it needs to in the future. Rocker panels change to accommodate a wider track and bumper systems differ to suit market requirements. (Though Stefanyshyn barely acknowledges it, the Epsilon tree will also sprout all-wheel-drive branches for inevitable crossover products.)

All the cars use a layer build with modular body sides. The way the parts shingle on top of one another at the front hinge pillars differs slightly from country to country for a couple of reasons. One is that several of the parts are sized differently. For example, every Epsilon car has a distinct greenhouse and interior. Another difference is what Stefanyshyn calls “administrative heritage.” “Germany doesn’t always build things exactly the way we do here,” he conveys. Stefanyshyn makes it clear that GM’s notion of a global program hardly infers strict parts commonality. He explains, “A concern with commonality is where you stamp major parts such as Epsilon’s center floor pan. We could of course stamp them all in Germany for worldwide use but the payback for an extra die set versus all the shipping racks you’d need is about two months. We’ve got ample stamping capacity here so why not use it to avoid the cost of shipping parts all over the earth?

“Then there’s the fact that every press line has its own processing nuances. All our weld, hookup points, and body-mounting positions are the same across the Epsilon family, but each floor has certain special characteristics. One floor uses a 2 mil radius here while the corresponding part in Europe needs a 3 mil radius there — so in the end, the parts are different but only in subtle ways.”

He also notes that while GM’s dies are traditional in their size and weight, they’re very cost effective due to the tremendous strides the GM die organization has made in efficiency. It uses what is known internally as a blue line, which is the organization’s assessment of the costs per die, resulting in economies that some consider the best in the world for stamping tools. It also has the flexibility in Epsilon to use composite exterior skin.

Saab’s 9-3 (top) is a departure from other Epsilon models due to Saab’s insistence on body structure and vehicle safety. It uses the same platform but everything from the floorpan up is strengthened. The Pontiac G6 concept (above) is said to carry the styling cues for the Epsilon-based Grand Am.

Interestingly, Saturn, which would capitalize on Epsilon’s composite panel capability, is one of the eight Epsilon brands although GM has not yet officially confirmed its inclusion.

Stefanyshyn’s U.S. partner in crime, Greg Bellopatrick, is chief engineer for GM’s mid-size line of vehicles. He notes, “We believe there are varying degrees of commonality. In Epsilon, fuel tanks, rear cross members, steering knuckles, struts, lower control arms and engine cradles may differ slightly from market to market but they’re all interchangeable.” The front suspension is a strut package that always locates the same way and provides the same ride travel in every application. All of the Epsilon cars share one four-link independent rear suspension though each is tuned differently in terms of bushings, spring rates, shock absorber calibration, etc. The rear cradle can be mounted rigidly to the car or rubber isolated. But given all the features shared throughout Epsilon, it can accurately be called an architecture with a high degree of commonality.

“If you lined up all the Epsilon variations hoists to look underneath,” Bellopatrick offers, “you’d conclude they’re all the same. But because these are high volume products, the flexibility available in tooling allows us to change details as needed to tailor each product for its regional market.” Components like engine mounts, for example, are different to provide a major distinction between a Saab and a Chevrolet.

Also, the notion of pothole durability seems be very different around the globe. “Germans look at you in disbelief when you explain the need to conduct the brutal durability tests we run,” says Bellopatrick. They’re understandably reluctant to add unsprung mass to suspension components unless it’s absolutely necessary.

Since we have separate tooling and we don’t intend to ship parts across the ocean, we use the flexibility we have to tailor parts to specific market applications.”

Commonality in degrees also filters down to the supplier level. Stefanyshyn explains, “Faurecia supplies all of our seat frames. They’re stamping parts both here and in Europe. Saab and Opel need a lot of content in their seats but the Chevrolet Malibu is at a lower price point so it might have a manual versus a power lumbar adjustment. However, under the upholstery, the basic seat architecture is common.” In addition to body structures and chassis parts, there are electrical systems consider. Again, the priorities differ around the globe. Notes Stefanyshyn, “In America, we tend to pack all the computing capability in one body computer, using serial data link to communicate with remote modules. Our goal is to reduce the number of body computers used by GM from 40 or so today down to one with medium content and one with high content for the future.”

But that isn’t an optimum fit with the Opel Vectra which is sold as an affordable Chevrolet in export markets such as Brazil.

To accommodate those applications, Opel prefers to spread its computing power over multiple modules interlinked with two-way communication. That allows them to select cost-effective modules as needed, thereby avoiding the expense associated with one body computer that’s more powerful than necessary.

“Trying too hard to commonize electronic components also squanders the opportunity to take advantage of cost reductions occurring on practically a daily basis,” Stefanyshyn adds. “You’ve got to be fast on your feet in the infotainment area but that’s not possible when you commit to one electronic component to be shared by multiple car lines launching over a period of months or years.

Another matter is how electronic systems evolve. For example, electrically assisted power steering (EPS) wasn’t able to pass noise targets when the Opel Vectra was under development so passed on the technology. But Chevrolet had an extra 18 months to solve that concern, and was able to include EPS in the Malibu. In the final analysis, neither interchangeable nor common parts are the driving force behind Epsilon. Instead, the carrot is what GM calls technology transfer. “The keys,” reveals Stefanyshyn, “are speed and quality. Even though every Epsilon car requires its own validation process, by living under one umbrella we achieve higher quality answers more quickly.”

As an example, Stefanyshyn recalls how the Epsilon program assembled a safety integration team so that as Opel crashed its first Epsilon cars, Saab engineers were present to observe and learn so they were higher on the experience curve when their turn to conduct the same tests arrived. “In Europe, safety standards are geared to belted occupants,” he notes, “while in the U.S., federal standards specify unbelted occupants. So Opels comply one way and U.S. products must meet different standards.”

Opel Vectra
Vauxhall Vectra
Saab 9-3
Chevrolet Malibu
Pontiac Grand Am
Saturn LS
Buick (Speculative)
Holden (Speculative) 
Interestingly, Saab is in the middle because its cars have to satisfy both European and U.S. safety standards to sell in both markets. With Saab onboard and second in line in terms of cadence, Chevrolet gained valuable knowledge about crash performance well before testing commenced for the Chevy Malibu. Technology transfer is Epsilon’s major payoff opportunity. “The cycles of learning inherent to our common architecture allowed us to reduce the time and money invested in engineering development by 30 to 35 percent,” boasts Stefanyshyn. “Considering the fact that Epsilon will yield 14 body styles that should last two generations or roughly ten years, we believe that GM will surely get its money’s worth out of the billions invested here.”

The Man Behind Epsilon

GM mid-size vehicle line executive (VLE) brings global architecture experience to the Epsilon program.

Gene Stefanyshyn intends to transform GM’s mid-size segment.
Gene Stefanyshyn started his career in a junkyard. He fondly recalls, “As a teenager, I spent days in an Oshawa, Ontario, salvage yard that belonged to a friend’s father. We’d bombed around in cars, tore things apart, and rebuilt engines for fun. That’s where I starting loving cars.”

Stefanyshyn says his high school dream was to become a tool and die maker. He took the math and science classes required for college admission and also the hands-on technical courses. “When I was about to become an apprentice, an excellent machine shop teacher took me aside to tell me my abilities exceeded those required to be a tool and die maker,” he says. “They encouraged me to try for more. That’s when it occurred to me that perhaps I could become an engineer.”

The 47-year old Canadian joined GM as a General Motors Institute engineering student in 1976. Upon graduation, he took up corporate finance studies, earning an MBA in 1981. That led to a planning and competitive analysis assignment at GM of Canada while Rick Wagoner was the organization’s CFO. Subsequent assignments at Saab and Opel broadened Stefanyshyn’s horizons. Under Opel’s Peter Hanenberger, he helped lay a foundation for GM’s global architectures strategy. Before returning to North America in 1995, Stefanyshyn also created an engineering council for GM’s international operations, strategy for Asia Pacific business, and an engineering proposal for GM’s China joint venture partner.

Back in the U.S.A., Stefanyshyn’s assignment as GM’s mid-size VLE was to grab a range of under-performing models by the scruff. His global experience made him the perfect change agent for solving mid-size problems and integrating future models into GM’s global product portfolio.

Previous posts

Next posts

Sat. February 4th, 2023

Share this post