IRAS satellite to be 'revived' for tests.

The Infrared Astronomy Satellite (IRAS), which in 1983 performed the first survey of infrared sources across the entire sky before finally running out of its vital cooling fluid, is taking on a new job. It will make no more astronomical observations--loss of the cryogenic coolant that enabled its detectors to record faint heat emissions from stars, comets and even possible planetary systems in formation has left it effectively blind. Beginning in March, however, its still-operating computer and other systems are to be used in tests of several backup operating methods that might benefit future satellites but were never needed by the successful IRAS.

Soon after IRAS's launching on Jan. 25, 1983, for example, its controllers found that its fine-pointing sun sensor (which helped it maintain its orientation in space) was sending out spurious signals that the computer would interpret as evidence that the satellite was pointing in the wrong direction. Two such "glitches" per second could--and repeatedly did--cause the computer to switch over automatically from its programmable memory to a preprogrammed section containing safety instructions, but they would also render the changing astronomical observations impossible. The sensor was locating the sun correctly, but the glitches kept causing the memory switchover, the IRAS engineers worked around the problem by simply restocking the programmable part of the memory with instructions to ignore any such glitches that did not appear in a consecutive string a six or more.

Also helping the satellite hold its position, however, was a sensor that tracked the position of the earth's limb, or horizon, and which never necessitated any such problem solving. In the upcoming tests, the engineers will evaluate and automatic backup system that should come into play whenever the horizon sensor goes awry: IRAS carries a magnetometer that monitors the earth's magnetic field, including its direction, and the goal is to check out an algorithm in the computer that can automatically feed the magnetometer's directional measurements into the satellite's positioning system if the horizon sensor becomes unreliable.

The test series has been instigated by the European Space Agency (ESA), which was not responsible for IRAS but which is seeking experience in operating reprogrammble satellites such as its own planned Infrared Space Observatory. For the tests, ESA has contracted with Fokker B.V. in the Netherlands (IRAS's builder), the Dutch National Aerospace Laboratory (which wrote its computer software) and Britain's Rutherford Appleton Laboratory (which controls it from the ground).

In another malfunction that never actually appeared during IRAS's main mission, a "serious software failure," or programming error, will be deliberately loaded into the programmable memory so that the satellite starts to tumble. If the built-in safety provisions work as they are supposed to, the other ("read-only") memory will take over automatically to bring the craft to a safe orientation.

The other test in the series is more for use when everything is going right. To protect the mission against the possibility that IRAS's whole main computer might malfunction, it was provided with a spare, complete with programmable and read-only memories of its own. The test is merely to link electronically the programmable twice the capacity for data storage and long sequences of computer commands.

IRAS ran out of coolant on Nov. 22, 1983, and for a while there was thought of resupplying it via the space shuttle. The idea was scrubbed largely because of cost, says U.S. IRAS project manager Gael Squibb of Jet Propulsion Laboratory in Pasadena, Calif., since the mission's all-sky survey was done, and selected targets for restudy would be better observed anyway by other facilities to come.