Monday, Dec. 10, 1990

Roots of The Hubble's Troubles

By David Bjerklie

Anyone can make a mistake. But when it sabotages a $1.5 billion project, the blunder is not easily forgiven. And when evidence of the mistake is repeatedly ignored until it is too late to fix the problem, then the episode becomes scandalous.

That was the case with the ill-fated Hubble Space Telescope, according to a remarkably frank investigative report issued last week by the National Aeronautics and Space Administration. The flaw that crippled the telescope's primary mirror was not obvious until after the instrument was launched last spring. Yet technicians at the company that made the mirror had indications of trouble long before the telescope went into space -- and apparently never told the design team about the disturbing signs. Meanwhile, managers at NASA who had responsibility for the Hubble project paid little attention to the details of the telescope's construction. "There were at least three cases where there was clear evidence that a problem had developed, and it was missed all three times," said Lew Allen, the director of NASA's Jet Propulsion Laboratory and the head of the six-member investigating committee that wrote the Hubble report.

There was more than enough time to catch the telescope's flaw. Rough grinding of the mirror began in 1978, final polishing was not finished until 1981, and the completed telescope sat on the ground for four years after the space shuttle program was disrupted by the Challenger explosion. The mirror's manufacturer, Connecticut-based Perkin-Elmer Corp.,* told NASA that the standards of precision established for the mirror were not only met but exceeded. The only problem was that the mirror had been painstakingly polished into the wrong shape.

To achieve the exacting specifications for the mirror, Perkin-Elmer used an optics template, a tubular array of smaller mirrors and lenses linked by connecting rods, to guide the grinding and polishing processes. When the Allen committee tested this template assembly, it found that there was a critical error of 1.3 mm (0.05 in.) in the placement of the template's components. The Hubble mirror was carefully fashioned to match exactly this error in the template.

What was worse, the company's reliance on this system was absolute. Though backup analyses pointed to a major flaw in the mirror, stated the report, these "indicators of error were discounted at the time as being themselves flawed." The evidence of the problem was never analyzed in detail by the engineers and scientists most qualified to do so. NASA accepted Perkin-Elmer's decision to rely solely on the precision of the template, when instead the space agency should have been alert "to the fragility of the process and the possibility of gross error."

Fortunately, the damage is not beyond repair. NASA plans to perform an in- orbit service call on the space telescope in 1993. In the meantime, pictures from the Hubble can be sharpened by computer enhancement. The telescope has taken some surprisingly good shots, including images of a gas jet streaming from a newborn star and a huge storm on Saturn.

But the Hubble cannot focus on distant stars nearly as sharply as had been expected. Its performance was compromised by mistakes that were easily avoidable. That is clear even without the benefit of optically perfect hindsight.

FOOTNOTE: *The company's optics division has since been sold to a General Motors subsidiary and renamed Hughes Danbury Optical Systems.

With reporting by Jerry Hannifin/Washington