A test a decade ago revealed an apparent flaw in the main mirror of the blurry-eyed Hubble Space Telescope, but key NASA officials and outside experts charged with overseeing the project say they were never informed of the problem.
Tests on the telescope's main mirror in 1981 uncovered a defect called spherical aberration, according to scientists investigating the problem and an optical expert who worked on the mirrors at Perkin-Elmer Corp., the company that designed, built and tested the telescope's optical system. The expert requested that his name not be used because of the ongoing investigation.
Spherical aberration is the same type of defect that now hobbles the telescope, whose vision of distant stars and galaxies is blurred, greatly reducing the observatory's scientific usefulness.
The test that detected the aberration was discounted because another testing device believed to be more sophisticated found no such flaw, according to the former Perkin-Elmer employee. Details of the discrepancy between the two tests were first reported by the Hartford Courant.
'Never Any Question'
Key officials at the National Aeronautics and Space Administration responsible for overseeing the construction of the telescope said last week they knew nothing of the different test results.
"There was never any question of that nature," said NASA Deputy Associate Administrator Sam Keller, who came to the project in 1983 and was responsible for reviewing all the components of the troubled program.
"I do not recall any difference between the two tests, nor do my records, which are extensive, indicate any knowledge of a difference in the tests," said C. Robert O'Dell, formerly NASA's Hubble project scientist and now a professor at Rice University in Houston. Hubble project chief engineer Jean Olivier also said he was not aware of the problem.
"I didn't know anything about it, either," said William Fastie, a former member of the Hubble Science Working Group and a retired professor at Johns Hopkins University. Fastie and Daniel Schroeder of Beloit College in Wisconsin were brought into the project as special optical consultants. Fastie said he believed Schroeder also knew nothing of the problem. Schroeder could not be reached for comment.
NASA officials said they knew of no one at the agency who was aware of the discrepancy in the test results. Former employees of Perkin-Elmer, on the other hand, said that NASA representatives were informed at the time.
Test Devices Examined
Last week, experts deployed by an independent Hubble investigation led by Lew Allen, director of the Jet Propulsion Laboratory in California, zeroed in on the equipment used to test the shape and smoothness of the telescope's main mirror.
The two mirrors aboard Hubble gather starlight, concentrating and focusing it at a spot at the rear of the spacecraft where cameras and spectrographs analyze and assemble the images before beaming them back to Earth.
Like a sleuth in a parlor mystery, the investigation team has assembled the leading suspects and is now putting them to the test at the former Perkin-Elmer facility in Connecticut. Perkin-Elmer was purchased last year and is now called Hughes Danbury Optical Systems.
At present, suspicions are focused on a pair of devices called null correctors, gadgets composed of lenses and mirrors and resembling nothing so much as a magician's prop. (But like any good mystery, it is possible the null correctors are innocent and that the optical equivalent of the butler is responsible).
Evaluating the Mirror
To test a mirror, a beam of light, more often a laser beam, passes through the null lens or lenses and then strikes the large mirror being tested. Then the light returns to its source, where it should produce a proper image. If the image is right, the mirror is considered well-shaped and properly curved.
Null correctors were used to aid the Perkin-Elmer team as it ground and polished the glass mirror. The equipment was also employed to certify that the Hubble mirror attained the exacting specifications set by NASA officials, who until recently liked to call the Hubble mirrors the most perfect ever made.
The null correctors were of two types. The device deemed most sensitive was a reflective null corrector, composed of a pair of small mirrors and a focusing lens. The device was thought to be state-of-the-art and was designed by Abe Offner of Perkin-Elmer, who wrote the chapter on null correctors in the mirror maker's bible, a book called Optical Shop Testing.
According to scientists involved in the Hubble investigation, Offner's reflective null corrector had been used several times on other mirrors that Perkin-Elmer made for spy satellites, which are similar in design to the Hubble. Therefore the curvature of the small mirrors housed in the device is thought to be accurate.
But the optics within the corrector can be adjusted, depending on the size of the mirror to be tested. Recent calculations show that a small error in the spacing between the null mirrors could have produced an image that led mirror-makers to grind a spherical abberation into the Hubble mirror. It is also possible that the small focusing lens used in the reflective system could have been the wrong shape or itself somehow flawed.
The second test device is called a refractive null corrector. Composed not of mirrors but a pair of lenses, this device is generally believed to be less sophisticated. It was this refractive null corrector, however, that found a possible flaw in the mirror, sources said.
The two null correctors obtained different results, indicating that either the Hubble mirror or one of the test devices was flawed. But it appears that the discrepancy was ignored. No more tests were done.
By discounting the results of one test, the engineers at Perkin-Elmer were, in essence, relying on a single test to assure the mirror was perfect. Optical experts say now that Perkin-Elmer and NASA should have challenged the mirror with at least two or three independent tests.
NASA's Keller said that at the time, the space agency considered the mirrors the least of its problems. "The mirrors were not perceived as being that tough of a job," Keller said. He added that Perkin-Elmer was thought of as "managerially weak" but "technically excellent." The company had built components of a number of spy satellites using similar optics and presumably similar tests. But since that work is highly classified, word of its success or failure has never been made public.
As for the failure to obtain independent tests of the mirrors, Keller said in retrospect that additional tests should have been required. Perkin-Elmer's bid did not include independent tests, although a losing bid submitted by Eastman Kodak did. According to documents obtained by the Associated Press, Perkin-Elmer was awarded the optics contract for $64 million against a Kodak bid of $100 million. Because of massive cost overuns, Perkin-Elmer was eventually paid $451 million.
"It would seem in retrospect that to test the mirrors in just one way would be a fairly serious defect," Keller said.
The key test that may have led to the Hubble Space Telescope's blurry vision may have involved this kind of test, called a null reflector. At the mirror maker's factory, an apparatus like this was used. The Hubble's main mirror was at the bottom and the test apparatus above.
A laser beam shines through a hole in the null corrector. The beam is split, strikes the farther mirror and is reflected back up toward the nearer mirror. From there it goes back down through a lens and onto the large mirror being tested.
The big mirror reflects the beam back along the same paths. A sensor (not shown) above the null corrector combines the returning light with a sample of the original laser light. The two light sources interact to produce a pattern that tells engineers whether the mirror has the correct shape.
If the distance between these two mirrors is slightly off, it can lead mirror makers to grind the mirror being tested into the wrong shape. Small focusing lens could also be the source of the problem.