Mars Devices Failed in Other Spacecraft
Transistors from the same lot that may have crippled the Mars Observer probe also broke down aboard two U.S. weather satellites and were pulled at the last minute from two key military satellites, according to aerospace industry documents and NASA officials.
The parts had all passed rigorous industry and military quality control tests, but a subsequent investigation this summer by NASA, the Air Force, Martin-Marietta and several other aerospace companies revealed that many of the tiny diodes were defective, with fractured or sloppy welds. Of the 71 transistors tested by NASA, 53 failed, according to a summary of the investigation prepared by TRW Inc.
In the report, investigators attributed the flaws to an “out of control” manufacturing process. They cited “poor mechanical attachment” of welded bonds that led to fracturing of the transistors.
The company involved--a Massachusetts electronics subcontractor called Unitrode Inc.--made the transistors at a Salem, Mass., plant in December, 1983, and sold about 700 to Frequency Electronics Inc. in Uniondale, N.Y., which manufactures satellite timing devices and clocks. NASA had said initially--and mistakenly--that FEI had manufactured the transistors.
Project engineers at the Jet Propulsion Laboratory in Pasadena, which is running the mission, said the two clocks aboard the missing Mars Observer spacecraft utilized the diodes and are now the leading suspects in the demise of the $980-million mission this week.
Officials at Unitrode’s headquarters in Billerica, Mass., did not respond to several requests for comment Friday. They instead issued a written statement saying, “We are aware of no data which suggest that any Unitrode product was the cause of the problems in re-establishing communications with the Mars Observer.”
The FEI timing devices are used aboard virtually every U.S. military and civilian satellite in orbit and many international satellites, company executives said Friday.
FEI President Martin Bloch said Friday that the company’s only reliability problems have involved the transistors from the single Unitrode manufacturing lot. He denied any suggestion that the on-board timing devices might have been responsible for the Observer failure.
“I think this is totally unwarranted,” said Bloch.
“The clock has gone through extensive testing here. It went through a year of testing at Martin-Marietta. It operated flawlessly on an 11-month cruise from Earth to Mars. There are probably a dozen other possibilities.
“We did everything to build a reliable clock,” he said. “If you have a latent failure in a component (the transistor), only God can predict that.”
In the meantime, flight controllers at JPL began what appeared to be their last maneuver to re-establish contact with the missing spacecraft. They sent a series of commands Friday that, if received by the probe, would trigger resumption of radio communications by 6 p.m. Sunday.
“With that option exercised, the team will have utilized all presently understood recovery approaches,” said JPL spokeswoman Diane Ainsworth. “They would then begin a ‘listen only’ vigil.”
Paula Clegget-Haleim, a spokeswoman at NASA headquarters in Washington, where an investigating panel is being assembled, said that it would be months before the agency had any formal conclusions about the cause of the Mars Observer failure.
“It is kind of premature of us to lay blame on anybody,” she said.
Some NASA officials were skeptical of the transistor theory, which states that two transistors--one in each of the Observer’s two clocks--would have to have failed. “The probability of there being two simultaneous failures just blows my mind,” said one agency satellite program official who asked not to be identified. “I’m not sure how good an explanation of the failure that is.”
The Unitrode transistor problem surfaced in June when an FEI timing device failed on-board the $62 million NOAA-13 weather satellite awaiting launch from Vandenberg Air Force Base. By that time, Observer--carrying transistors from the same batch--was nine months into its journey to Mars.
Larry Draper, deputy project manager of the meteorological satellite program at NASA’s Goddard Spaceflight Center, said Friday that the transistors also caused an earlier failure of an FEI temperature controller on a second weather satellite awaiting launch in December, 1989. All were replaced. Another failed FEI timing device, which had not yet been installed in the satellite, was also discarded but it did not involve the suspect transistors.
The NOAA-13 weather satellite, which was finally launched earlier this month, disappeared in orbit last Saturday, the same day the Mars Observer failed to resume communications with Earth. But NASA officials said they did not think the transistors were at fault because they had replaced the original transistors with some from a different lot.
“We not only switched from the manufacturing lot, we switched the manufacturer,” said NASA spokesman Jim Elliot at Goddard.
The transistors also were found in, and removed from, FEI devices aboard two Pentagon Milstar satellites awaiting launch. The Milstar satellites are designed to be the world’s most survivable satellite communications systems, capable of keeping up military communications even after nuclear war.
TRW engineers were part of the team that investigated the Milstar failures.
