The crew of the shuttle Columbia lost its only working link with the spacecraft's $150 million astronomical observatory early yesterday after a component apparently overheated, sending scientists and engineers at three ground facilities scrambling to salvage the last half of the trouble-plagued mission.

The National Aeronautics and Space Administration mission has been a frustrating ordeal for scientists who have waited through five years of delays, including four unsuccessful attempts to launch it this year.

The Astro-1 observatory includes three ultraviolet telescopes and the most powerful X-ray telescope sent into orbit. Scientists are attempting to study the most violent and energetic objects in the universe, such as quasars and supernovae, by measuring radiation that is emitted in wavelengths that do not penetrate Earth's atmosphere.

The only way astronauts inside the shuttle cabin can operate the three ultraviolet telescopes in the shuttle cargo bay, which is like a truckbed open to space, is with one of two systems. Each system includes a computer, a video screen and a keyboard -- similar to a personal computer outfit. At about 7:15 a.m., a keyboard in one of the systems shut itself down.

The display screen in the other system had shut itself down Sunday, nine hours after the early-morning launch of the mission. In both shutdowns, crew members reported that they thought they smelled something burning, according to Flight Director Walt Pennington, although smoke detectors found no smoke and the seven-man crew was said to be in no danger from fire.

Investigating with a flashlight, the astronauts discovered that the air-cooled system's intake vents were clogged with blue lint that may have caused overheating, according to Steve Elsner, head of the guidance and control system section at NASA's Johnson Space Flight Center in Houston.

The lint normally sprays out in a shower when the Velcro straps used to secure equipment aboard the shuttle are pulled apart. However, Elsner said, that problem is well known and the systems contain filters designed to prevent lint damage.

The astronauts vacuumed the lint out of the intake and at midafternoon tried turning the first system back on. But they smelled the same burning odor and quickly turned it off again, Pennington said. The electronic symptoms of the problems on the two systems are similar, he said, indicating the same problem may be common to both.

The astronauts said they "feel it {the burning odor} is associated with something electrical, not something like hot lint," Pennington said, although he added that the lint could have contributed to the problem.

The clogged vents are under a control panel in the aft flight deck of the orbiter.

Scientists and engineers at Johnson center; Marshall Space Flight Center in Huntsville, Ala., and Goddard Space Flight Center in Greenbelt spent the day developing new software commands to send from the ground to one of the onboard computers, which tells the telescopes what to do. Such commands normally would be entered by astronauts using keyboards on the shuttle.

"Believe me, we are into ground at this point that we have not simulated," said Mission Scientist Ted Gull, referring to the numerous practice runs the team went through in preparation for the mucH-delayed flight.

"And yet, there is not panic," he added. "We're not giving up on doing science."

The mission has been intertwined with a series of misfortunes besetting the space agency in recent years. First scheduled for launch in 1986, it was postponed by the Challenger accident that year. When it was tried again last April, Columbia sprung a hydrogen leak, the first of several that prompted NASA to ground the shuttle fleet for the summer at the same time it discovered a major flaw in the Hubble Space Telescope.

Gull and other officials said they do not know how difficult the unplanned method of operation will be or how much scientific observing time it will cost the mission, which has also suffered lesser glitches involving its pointing system and other components.

Once the computer has reacted to the commands from the ground, the astronauts will use a "joy stick" hand control to fine-tune the telescopes' aim, which must be extremely precise -- like focusing on a basketball at 35 miles away, Elsner said. The astronauts have had unexpected practice with the joy stick because the automated pointing control system has been out of order for much of the flight.

Management of the mission is unusually complicated for a shuttle flight. The scientific aspects are being managed at Marshall, using a European-built facility aboard the shuttle called Spacelab, which includes the troubled pointing system and the two broken control systems, to house and support the telescopes. Goddard is controlling one of the four telescopes. And shuttle mission control, which has the only channel for communication with the astronauts, is in Houston.

"There are a lot more people, or organizations, involved than on a standard flight," Elsner said.

Scientists have been working around the clock to minimize the effect of the problems and take advantage of brief, precious observing time. The mission is to end Tuesday.

"I'd be lying if I said I wasn't frustrated," said Mary Jane Taylor of the Wisconsin Ultraviolet Photo-Polarimeter Experiment (WUPPE), developed by the University of Wisconsin. "But I don't think anybody thought this {mission} would be error-free. We're doing something that has never been done before."

The experimental observatory is whirling around the planet at about 17,000 mph as it attempts to aim precisely at its faint, distant targets. It is also using untested new technology built by laboratory scientists.

Until yesterday, "the mission was going extremely well," said Goddard's Stephen Maran, a member of the Ultraviolet Imaging Telescope team. "We have been doing about as well as you have a right to expect . . . . Remember, these are experiments. These are built by little people, not big aerospace companies."

The Broad-Band X-Ray Telescope, developed by Goddard, is the only one of the four telescopes that is normally controlled from the ground and therefore was not affected by the latest problems. It "chugged along" doing its work all day yesterday, Maran said. Ironically, it was added to the mission as an afterthought when an unusually close supernova, or exploding star, burst in the heavens in 1987, promising a rich source of X-rays.

The scientists said they were just beginning to get the hang of overcoming the balky pointing system and other aspects of the operation, and had taken a number of observations never achieved before when the latest breakdown occurred.

"I don't know whether to smile ear-to-ear or cry," said William Blair of the Johns Hopkins Ultraviolet Telescope (HUT) team, which the day before had studied the brightest quasar in the sky -- one of its "top 10" targets -- an object thought to house a black hole.