Space shuttle managers, fully confident they have fixed a computer timimg mechanism that had forced them to scrub the launch Friday, today rescheduled the maiden flight of the Columbia for 7 a.m. Sunday.
The only thing that might delay astronauts John W. Young and Robert L. Crippen from piloting Columbia into space Sunday morning is the condition of the Florida skies.
Shuttle mission rules state that clouds must cover no more than half the sky, a rule put in place in case the astronauts must abort the launch and fly the shuttle back to an emergency landing on the three-mile-long runway at the Kennedy Space Center.
The space shuttle team found that at 7 a.m. today, the skies were too overcast for a launch. "The crew was up early morning flying their T38 jet trainers and they were concerned about the cloud cover," Launch Operations Director George F. Page said. "It would have been hard to launch on time this morning, and if it's the same way tomorrow, we'll just have to hold and see when it clears."
The skies cleared later this morning so that a launch could have taken place by 10 a.m., which is more than enough time for Young and Crippen to complete their 54 1/2-hour mission in Earth orbit and land Columbia during daylight hours at Edwards Air Force Base in California Tuesday.
There are six hours and 26 minutes of cushion time in the launch "window," meaning that Young and Crippen could take off as lat as 1:26 p.m. and still complete their flight in daylight hours.
If the weather cooperates the way a balky computer did today, there will be no difficulty getting Young and Crippen into space on time. The way computer engineers described it today, two of four prime computers that were on line Friday morning in the final minutes of the countdown were ordered to tell the backup computer what they were doing, and they did it at the wrong time.
"The primary computers were calling the back 40 milliseconds [40 thousandths of a second] too soon," said Richard B. Parten, deputy director of the data systems and analysis dividion at Houston's Johnson Space Center. "In effect, the backup computer was hanging up the phone and not listening because it wasn't hearing the ring on time."
Parten described the condition the backup computer was put in as "unique and very invidious." He said that if the backup computer cannot communicate with the primary computer "it is no-go for launch."
The backup computer must be communicating with the prime computers at liftoff because it is commanded to fly the shuttle into orbit if all four main computers should fail together. The backup computer is also needed when the shuttle re-enters the atmosphere and returns to Earth, but in orbit the astronauts would have time to switch a balky backup with a spare computer they carry on board.
Computer engineers said that, once they understood what had caused the computers to lose their time, it was easy to fix. Turning the computers off and on again is sometimes enough for them to regain their timing. Parten said the only time computers lose their timing is when they're turned on in a "cold-start" condition, which the backup computer was in during the final minutes of countdown.
Parten said that timing in the computers is such a precise procedure that if they start out missing connections they will go right on missing connections.
"If it's there, it's always there," Parten said. "If it's not there, it's never there."
Late Friday night, the computers aboard Columbia were turned back on for the first time since the scrub and appeared to work find. Tests through the night showed no loss of timing between prime and backup computers. The computers were left on today and will be left on through the countdown to prevent a recurrence of the timing problem.
One reason the computers were not "reinitialized" -- shut down and started again -- on the launch pad on Friday was that the shuttle's navigation system would have to be restarted and reset, a procedure that takes almost an hour. It would also have taken too much time to reload the computers with fresh programs on the pad -- a potentially risky procedure, because during the reloading the computers would not be monitoring the condition of the explosive hydrogen and oxygen fuel in the shuttle's huge external fuel tank.
"It's hard to explain a time skew like this," Parten said. "All I know is that if we had reinitialized the computers on the pad the chances are very high that this condition would have gone away."