DANBURY, Conn. -- The main mirror in the Hubble Space Telescope was flawed during its manufacture because a measuring rod probably was installed upside down in the device used to help determine the shape of that mirror, NASA investigators concluded yesterday.
The six-member Hubble Optical Systems Board of Investigation found "strong evidence" that a 1.3-millimeter spacing error in the testing device, known as a reflective null corrector, occurred because the inverted measuring rod did not fit snugly into an aluminum cap.
A technician at the manufacturer, then known as Perkin-Elmer Corp., apparently put the flat end -- and not the finely polished rounded end -- of the rod into the cap, accounting for a 1.3-millimeter spacing error between a mirror and a lens in the test device, investigators said.
The spacing error in the test device caused a flaw in the Hubble's 94-inch primary mirror called spherical aberration, resulting in blurry images being transmitted from the $1.5 billion telescope and greatly reducing the orbiting observatory's capabilities.
Dr. Lew Allen Jr., who chairs the board of investigation, called the error in the testing device "painful and tragic" yesterday and said that "adequate prudence was not taken" in relying on the reflective null as the sole test for the mirror's shape.
Double checks of the mirror's shape were recommended by engineers employed by the company, now known as Hughes Danbury Optical Systems Inc., but were rejected at the time, said Dr. Allen, who is director of NASA's Jet Propulsion Laboratory at the California Institute of Technology in Pasadena.
Asked whether the National Aeronautics and Space Administration had the final say in rejecting the double checks, he said it did not appear the recommendations ever got that far.
"These recommendations were not accepted within the Perkin-Elmer structure because of time and cost," he said.
From the beginning, the Hubble project ran behind schedule and had huge cost overruns. At the time those double checks were recommended, "there were enormous pressures" on the company, forcing it to operate "in a crisis mode," Dr. Allen said.
Repeatedly throughout the manufacturing process, other types of tests were performed on the mirror test setup that revealed problems with the mirror's surface. But those test results were discounted because of the confidence in the null, which showed no problem, investigators have found.
Dr. Allen said the board would address the questions of why the error was not recognized at the time in the panel's final report, now expected in early November.
He did say that neither the company nor NASA should have placed so much confidence in the reflective null corrector as the sole testing device used to check the mirror's surface without independently verifying the results with other tests.
Meanwhile yesterday, Charles J. Pellerin Jr., NASA's director of astrophysics, said he has asked the space agency's inspector general to investigate the management of the Hubble contract to determine where the weaknesses were.
Dr. Pellerin also said $3.2 million in bonuses for which the company is eligible for on-orbit performance of the telescope has been held up and may never be paid. That determination is up to NASA's lawyers and contract administrators to make, he said.
The investigative board, which reports to Dr. Pellerin, concluded two days of work behind closed doors at the company with a news conference yesterday to explain the role of the measuring rod and cap in the mirror's flaw.
The metal rod -- an exact instrument 210.75 millimeters long -- was used to determine where to position a "field lens" in the null corrector.
The aluminum end cap was supposed to fit snugly on the rounded top of the measuring rod, but because it probably was installed upside down, it did not.
The cap had a pinhole in its center that was designed to hold a beam of laser light centered on the rod's end.
The laser was focused to hit the rod end at a point between the two spherical mirrors inside the null by being bounced off the bottom mirror.
But, instead of hitting the tip of the rod, investigators believe the laser hit the mispositioned cap at a point on it where anti-reflective coating had worn off.
The rod was designed to allow for the precise placement of the field lens, but because it was 1.3 millimeters too long, it caused the mirror to be polished to the wrong shape.