Airplanes made out of beer cans

WASHINGTON — "We make our airplanes out of aluminum from beer cans, titanium from golf clubs, and electronics from video games," says Lloyd Condra, technical fellow at the Boeing Phantom Works in Seattle.

Nov 1st, 2002

By John Rhea

WASHINGTON — "We make our airplanes out of aluminum from beer cans, titanium from golf clubs, and electronics from video games," says Lloyd Condra, technical fellow at the Boeing Phantom Works in Seattle.

Condra, who is the scheduled keynote speaker at this year's Military & Aerospace Conference West (with COTSCON), to be held Dec. 10-11 at the San Diego Concourse, hastily adds that this is an "oversimplification."

But it dramatically illustrates a worsening problem: where are the avionics firms going to find the electronic components they need in a marketplace dominated by the large-volume commercial users?

Condra traces the evolution of this problem in three phases. It all began with former Defense Secretary William Perry's series of acquisition reforms launched in 1994 aimed at superseding military specifications with their commercial off-the-shelf (COTS) equivalents.

As has been demonstrated since then, this step had the beneficial effect of reducing component costs and giving the military users access to state-of-the-art technology.

The downside was its path to the second phase — diminishing manufacturing sources, or DMS, as the parts suppliers shifted their attention to the more lucrative telecommunications and other commercial market segments.

If that weren't bad enough, the current third phase is even more foreboding. Condra describes this phase as the normal evolutionary trend to greater parts complexity with resulting reduced feature sizes that go well below the 1-micron barrier breached more than a decade ago (much of the improvement, ironically, funded by the U.S. Department of Defense).

Condra calls this phase the nanometer challenge, noting that it is characterized by early device wear out (three to seven years), possible non-constant failure rates and common-cause failure rates, increased susceptibility to atmospheric radiation, uncontrolled device configuration changes, and potential inability of avionics system designers to use them in applications other than those for which they were developed.

"We survived the COTS challenge because the components were more capable and reliable than we had previously supposed," Condra notes. "We are coping with the DMS challenge by aggressive tactical, ad-hoc responses to individual incidences of DMS."

Grave challenge

The nanometer challenge raises the stakes. "This challenge is as grave as any since the beginning of the solid-state revolution 50 years ago," Condra says in his COTSCON presentation. "It must be solved strategically, not tactically." What the COTSCON audience will hear is derived from a pitch that Condra and his colleague Joe Chapman, have been making to the cognizant military agencies. Chapman is an independent consultant in Midland, Texas, who is studying the problem for the Defense Standardization office at Fort Belvoir, Va.

What Condra and Chapman have been seeking is basically recognition on the part of the agencies that there is a problem plus their top-level support in addressing it on a strategic basis.

The focal point of this effort is what they call an Integrated Aerospace Parts Acquisition Strategy, or IAPAS, that would essentially get everybody on the same page.

The avionics manufacturers themselves have banded together under an umbrella organization known as the Aerospace Vehicle System Institute (AVSI), which is based at Texas A&M University in College Station. Members include Boeing, Honeywell, TRW, Smith Industries, and Goodrich.

A similar organization, the Avionics Working Group, was formed five years ago to address this problem and did issue a management plan for uprating the specifications.

The idea behind the IAPAS, as Condra explains it, is to serve as a bridge between the component roadmaps generated by the Semiconductor Industries Association in San Jose, Calif., and the avionics roadmaps being produced by the individual companies in this market. Out of this interaction Condra hopes to achieve some degree of compatibility — or at least some kind of dialogue.

In their presentations to the agencies, Condra and Chapman list the solutions they think won't work: money, coercion, and (one that strikes me as particularly poignant) "appeals to patriotism."

Money won't work? In a free enterprise economy like ours? Well, not in a conventional sense, Condra says. This is going to take some outside-the-box thinking.

That thinking is still being defined, but some kind of economic incentives are going to be needed to persuade the device manufacturers to supply the needed parts at the required levels of reliability and in the generally small quantities inherent in avionics applications. "It's all about money," Condra says, and that means that the aerospace systems firms will eventually have to foot the bill.

However, Condra insists that COTS is here to stay. "We're not looking to recreate the mil spec system," he says.

On one point the AVSI people are adamant: avionics manufacturers do not have the tools, information, or knowledge to assure continued quality and reliability to COTS devices. Device manufacturer cooperation is required, they stress.

Who are those device manufacturers? Condra and Chapman are understandably cagey on this point because this is a sensitive issue for those firms. It would seem logical that all the usual suspects would have to have some involvement with this program: Intel, Texas Instruments, IBM, Motorola, Xilinx, AMD, Siliconix, etc.

Distant solution

Condra estimates that it may take as long as five to seven years to get this kind of cooperation and that five to ten companies will be need to achieve a sort of critical mass.

The problem certainly isn't solving itself. By the mid-1990s, after the transition to COTS parts, the AVSI team estimates that 60 percent of the parts became obsolete within five years. Today, as volume shipments soar and feature sizes decline — with accompanying concerns about gate oxide and metal migration problems — the AVSI team estimates that COTS parts in service will wear out in three to seven years.

I find this particularly alarming in view of the fact that the average lifetime for an automobile in this country is 12 years. I would feel even more alarmed if I were buying a new automobile today and knew that the electronic system would wear out before the vehicle had achieved half of its expected lifetime.

Condra responds that the 12 years only work out to about 4,000 hours of actual operations and that typical commercial aircraft operate for 5,000 to 6,000 hours a year — even when constructed out of leftover beer cans, golf clubs, and video games.

The AVSI team estimates that COTS.parts in service will wear out in three to seven years.

More in Computers