When the National Aeronautics and Space Administration flew the space shuttle Columbia to Cape Canaveral a few months ago, it sent 200 skilled workers from the shuttle factory Palmdale, Calif, there, too.

The reason is that the job of protecting Columbia against the blistering heat of reentry wasn't finished when it came time to move the shuttle to the Cape for testing. About half of the 33,000 foamed glass tiles that insulate Columbia had been installed. The task was going so slowly that fewer than 300 tiles a week were being fixed to the fuselage of the DC-9-sized shuttle.

"It's taking us a lot longer and a lot more work to put that thermal protection insulation on than was in the plan," NASA administrator Robert A. Frosch said recently in an interview "It has turned out to be a much worse job hanging those tiles on the Orbiter than we though it would be."

Ten thousand tiles still must be fastened on Columbia's aluminum frame, a task that probably will take until Christmas. That means Columbia can't be moved to Launch Pad 39A until early next year, which in turn means it won't fly until June 1980 at the soonest. Thus the first flight of the space shuttle will be at least seven months late and cost $220 million more than the space agency estimated in March.

"We started putting those tiles on two year ago," space shuttle program director Myron S. Malkin said in an interview. "We're going three shifts a day six days a week and we're currently getting 1.3 tiles per man per week."

The two-year history of fixing the tiles has been fraught with frustration. About a year ago, the tile makers at the Lockheed Corp. went on strike and production plummeted. Just about the time the supervisors brought production back to normal, the strike ended and the tile makers returned. They were so out of practice that production plummeted again until they warmed to the task.

The tiles begin bonded to Columbia are unique. Their starting material is river sand from Leseur, Minn., considered the purest in the United States. The sand is fired into glass in Richmond, Ind.; fiberized and strengthened in Chamco, Ky., then machined and sawed inti blanks in Palo Alto, Calif.

"Every tile is different and every te has to be precision-machined," Malkin said. "You can't stand even the tiniest gaps between the tiles because the heat on reentry would get in between the gaps and melt you."

The toughest tiles to install are the ones being put on Columbia right now. Each time a worker installs a tile, he fits a dummy tile to the space alongside the tile he's just put on, then sends the dummy to the factory at Palmdale. There a tracing machine mills out a real tile, which is then shipped back to Cape Canaveral to be fastened to the shuttle.

The tiles are fitted to the shuttle with a space-age glue painted on a nylon felt backing. Each tile is held gently but firmly against the skin of the shuttle with a tripod attached to a "strain isolation pad". The tiniest scratches cannot be tolerated; the spaces between tiles can be no more than 17/1000ths of an inch.

"We're talking about temperatures on reentry of 2,300 degrees on the surface of these tiles," Malkin said. "Any defect, any gap can burn a hole right tn the bird and depending on what's behind that hole, like a hydraulic cable you could lose the whole bird on landing."

Beside gaps and holes in the tiles, the space agency is worried that one or more tiles might come loose on lift-off, triggering a "zipper effect" that might unravel a whole string of title. Some engineers are so worried that Nasa is considering a suborbital test flight for the first shuttle.

"As of now, we are planning that our first flight will be an orbital flight but it's true the question of a subordital flight has been raised," NASA administration Froscg said. "The issue is a simple one. Are our predictions of the aerodynamic forces on ascent correct? Simply put, will the tiles stay on?"

If tiles come off during liftoff, the space shuttle still could survive a return to earth because the fierce reentry heat would not be encountered. Missing tile then could be replaced and the reasons they came off could be studied and problems correct.

There are strong arguments against a suborbital flight. First, the schedule of at least four test flights in orbit would be delayed by at least three months by a suborbital flight. NASA already figures it has used all of its delay time if it is to use the shuttle to get its and the Pentagon's satellites into orbit in the 1981-1983 time period.

There also are safety reasons to think twice about a suborbital flight, which does little more than simulate an aborted takeoff. The shuttle would fly about 400 miles downrange from Cape Canaveral, turn around and fly back to land on the runway at the Kennedy Space Center.

To do that, the shuttle's engines would have to fire for 823 seconds instead of the 520 seconds they would burn if the shuttle were going into orbit. The astronauts would then be asked to perform a series of difficult maneuvers to turn the shuttle around and get it back to Florida. Landing would be on a two-mile-along runway instead of 30- to 40-mile dried lakebeds in California's Mojave Desert.

"Anybody who has to land that thing for the first time doesn't like landing on the little strip," program director Malkin said. "Nobody likes to practice aborts until they have to do it".

The astronauts who would have to fly a suborbital flight are just as concerned about the engines as they are about the runway. A space shuttle engine was tested for the first time last week at a full 823 seconds, the time it would have to burn during a suborbital flight.

There has not yet been a test of three engines firing together for 823 seconds. In fact, there has been no successful test of all three engines firing at once. The last time a three -engine test was attempted all three engines cut off at 58 seconds into the test instead of running the 520 seconds it would take to put the shuttle into orbit. CAPTION: Picture, Workers fasten tiles to protect shuttle from intense heat on reentry from space. NASA Photo