Eclipse Aviation turns to the automotive industry for inspiration in mass producing jets

Sept. 24, 2007 — -- Nearly a year after winning FAA type certification, Eclipse Aviation CEO Vern Raburn casts blame in a lot of directions when asked why his company has been able to deliver barely 50 small jets -- far short of the hundreds he had forecast. His suppliers let him down, he says, calling the performance of a recently discarded avionics system "just really, really, really bad." Some of his managers fell down on the job, failing to grasp the complexities of mass producing airplanes. "They talked the talk, but they could not walk the walk. They had no concept of what it meant."

But then Raburn points the finger back at himself, acknowledging that he oversold the revolutionary idea that airplanes could be produced in big batches as efficiently as personal computers (he made his name in the software industry in the 1980s). "I didn't have an appreciation for the difficulty of ensuring that all this stuff was built in conformity," he says. "The biggest mistake I made was assuming the supply chain would function with the same efficiency and reliability as it does in the technology business."

Forget, for a moment, the long-running debates over whether there's a market for thousands of very light jets (VLJs) or how many of Eclipse's 2,600 orders will see fruition. The company's immediate challenge remains proving it can mass produce its two-engine, four-passenger jets with consistent quality.

The few dozen 500s that Eclipse has managed to deliver are falling short of their promised functionality, thanks in part to a last-minute switch in avionics suppliers. GPS isn't fully enabled, there's no approach mode in autopilot, and the avionics don't have a flight management system. Raburn says the integration of four systems components -- which he declines to name -- is still failing at "massively unacceptable rates." He estimates it will take another five months to iron out all of the problems.

"Half the challenge is designing an airplane, but the other half is building a line that can manufacture them at a consistent quality level," says Ed Iacobucci, president/CEO of DayJet, a Florida-based air taxi venture that is Eclipse's biggest customer. "Frankly, some of the first airplanes were half hand-built. They didn't have all the processes nailed down."

Time is of the essence for Eclipse. The Albuquerque, N.M., company has raised nearly $1 billion in equity and debt since 1998 -- backers include Microsoft Chairman Bill Gates—and investors are eager to see a profit. In June, the collapse of a $200-million investment deal with a hedge fund fueled rumors of a possible bankruptcy. Raburn insists the 1,550-employee company was months away from having to shut down and says it has now raised enough capital to reach profitability -- if it performs to plan. With a list price of $1.6 million, the VLJ industry's lowest, meeting that plan will require delivering 600 aircraft a year just to break even, meaning Eclipse's current production rate of about one aircraft per day must be doubled by year-end. Current plans call for a increasing yet again to three fully assembled aircraft per day by the end of 2008.

Meanwhile, the nascent VLJ market that Raburn pioneered when he began taking aircraft orders in 2000 is growing more crowded. Cessna's Citation Mustang entered service earlier this year, Embraer's Phenom 100 is expected to follow in mid-2008, and Honda Aircraft Co. is on track to receive FAA certification of its HondaJet in 2010.

Part of Eclipse's challenge is to create a predictable production system with repeatable processes that places much more responsibility on suppliers than aircraft programs historically have -- not unlike what Boeing is attempting to do with its new 787 airliner. (That too is a work in progress.) The Eclipse 500's wings come from Japan, the nose from Chile, the engines and landing gear from Canada, the tail from the U.K. and the windshield from the U.S. All the pieces are shipped to Albuquerque and assembled by Eclipse. Raburn says the "vast majority" of his suppliers are on schedule and cost. But "some took the attitude, 'Just build it, push it out the door and [Eclipse] will catch any problems at the factory.' That introduces massive inefficiencies to the supply chain."

The problems are hardly new. Five years ago, after the Eclipse 500's first flight, the company had to scrap the aircraft's original engine because of poor performance. Pratt & Whitney Canada was signed to develop the 900-lb.-thrust 610F, but the program was set back more than two years. Then early this year -- four months after the jet received FAA type certification -- Eclipse parted ways with its main avionics supplier, Avidyne, replacing the Massachusetts company with five new vendors for the jet's Avio Total Aircraft Integration System (AW&ST Mar. 19/26, p. 109).

"Those kind of changes will just wreak havoc on a supply chain, especially if you're trying to set up for a modular kind of build," says Pete Wiese, director of CSC Consulting's aerospace and defense practice. Raburn says Avidyne's performance was poor; Avidyne did not return calls seeking comment.

A recent visit to Eclipse's assembly facilities next to Albuquerque's airport yields signs of an operation that is getting its act together. More than 50 aircraft are in various stages of assembly, and the pace of deliveries unquestionably is speeding up.

To revamp its production processes, Eclipse sought help from experts in the hypercompetitive automotive industry. In March, it hired Todd Fierro, a seasoned plant manager at the Ford Motor Co., as vice president of manufacturing operations. Fierro quickly retained The Productivity Team (TPT), an industrial engineering consultancy that helps high-volume manufacturers apply Lean and Six Sigma principles to accelerate throughput and, as a by-product, improve quality. The second-largest firm of its kind in Detroit, TPT counts among its clients automotive original equipment manufacturers (OEMs) and many of their suppliers.

Following a three-month study of Eclipse's production operation, TPT concluded that much of the work it has done in the auto industry was immediately transferable, according to Steve Nolan, TPT's program manager-in-residence at Eclipse. "The goal is for Eclipse to deliver product quicker without ever losing sight that quality is paramount," he says.

Step One was to gain a full understanding of Eclipse's current production processes and identify constraints. The assembly floor was reconfigured to a linear flow model that involves kitting on two parallel production lines, versus the previous discontinued flow manufacturing Eclipse had been using. Revamped processes have been rolled out in phases and will continue to be introduced through the end of 2007. Aircraft fuselages move from one station to the next, with problems eliminated at the source or stopped from moving to the next work station.

The change is significantly speeding up production and yielding greater capacity, with little or no corresponding increase in investment in infrastructure. Since TPT's involvement, capacity has quadrupled and is forecast to double in the next phase from what Eclipse has now.

The challenge now is to sustain the more efficient processes that have been put in place. In addition to working closely with Eclipse in the implementation of the new production, TPT also is working in a similar fashion with Eclipse's 10 "highest-impact" suppliers so they don't become impediments to the overarching goal of the jet reaching its optimum production rate.

"The Eclipse production system is analogous to any other high-volume producer," says David Kunselman, president and founder of TPT. Both he and Nolan have been struck by the interest and engagement of the workforce.

"This to me was a very pleasant surprise," says Nolan. "The people on the assembly lines have the spirit of continuous improvement, which is very important. If they follow and sustain the production processes that are being put in place, nothing should stop them from reaching volumes no other airframe OEM currently is achieving."

Adds Kunselman, "It's the same approach we've used with automotive companies. At Eclipse, we have a very clear line of sight of three finished aircraft per day by the end of 2008."

Even as it increases collaboration with vendors, Eclipse is trying to minimize the shock of supply chain disruptions by dual-sourcing more components and continuing to shed suppliers that aren't meeting schedule or quality requirements. Fierro says that if a vendor delivers a defective part, "we'll ship it back at their cost."

Iacobucci, who launched DayJet's air taxi service earlier this month with 12 Eclipse 500s on hand, says the "squawks" in the aircraft being delivered—incomplete avionics, air conditioning problems, wings that leak fuel and cosmetic problems—have declined from 74 in the first jet to fewer than 30. "The deliveries coming now are a totally different airplane," he says.

DayJet has firm orders for another 297 Eclipses over the next two years and placeholders for more than 1,000 through 2011. Iacobucci says he still has faith in Eclipse, even though the production problems forced DayJet to delay its service launch by two months. "We kind of figured there would be some slippage," he says. "I've never seen a new airplane that's been delivered that hasn't had one form of problem or another."

Resolving quality problems and making sure servicing and pilot training networks are in place will be crucial. Gerald Bernstein, a business aviation consultant with the Velocity Group in San Francisco, recently flew on a DayJet test run and was impressed how much room there was for his 6-ft., 5-in. frame. "I haven't heard any stories of mass cancellations, and they're going to market sooner than Embraer and HondaJet," Bernstein says. "So at this point it's a matter of how happy the new users are with the aircraft. If they say it was worth the wait, people will forget" the initial problems.

But skeptics continue to question Eclipse's financial assumptions. "They're selling a product at below the cost of production," argues Teal Group analyst Richard Aboulafia, a longtime critic. "They've justified it to investors on the basis of impossible production rates. There will be a day of reckoning."

For his part, Raburn, who as a teenager got his pilot's license before his first car, vacillates between contrition and defiance. "I feel pretty good about the way Eclipse has been run," he says. "I don't feel that I need to apologize to anybody for what has happened." And he hasn't lost his entrepreneurial streak, unveiling a new "Eclipse Concept Jet" at July's Oshkosh (Wis.) air show to test the emerging market for single-engine personal jets.

But he has toned down the rhetoric about running Eclipse like a technology business, touting his expertise as Microsoft's 18th employee and a senior executive at Lotus and Symantec.

In an interview last year, Raburn boasted that Eclipse's just-in-time business model meant it would build an airplane in four days and get paid by a customer 20 days before it had to pay suppliers for the parts, a practice famously employed by Dell with personal computers. "If that's called a dot.com manager, then I plead guilty." Today, he's more likely to compare Eclipse to an automotive line than a computer operation.