Table of Contents

Computer glitches

During early tests, Canadarm2 continued to perform flawlessly. “It's very smooth, very precise,” reported Space Station astronaut Susan Voss. Then, a few weeks later, it fell prey to the most exasperating kind of problem: an intermittent computer glitch that afflicted the backup control system. The primary control system was unaffected so the arm remained capable of accomplishing the critical airlock installation procedure. However, until the computer problem was clearly understood, no one wanted to risk having the arm stuck in an unstable or dangerous position that might imperil the Station, the shuttle or their crews, so NASA delayed the launch of STS-104, which will carry the airlock into orbit.

What made everyone most uneasy was the mysterious disappearance of the glitch. “Nobody likes magic in the flight control world,” commented NASA flight director John Curry. “We want to understand what is causing the problem.”

It took weeks of troubleshooting, but engineers from the CSA, MDR and NASA finally tracked down the culprit—a faulty computer chip in the shoulder joint’s electronics box. It is similar to chips used throughout the Station and, as it turned out, the problem with this particular chip had been previously identified. A software patch was sent up to the Station to bypass the erroneous signal from the chip if it were to recur, clearing the way for a July 12 launch of STS-104.

Once the reason for the failure was understood, there was no concern about the arm’s ability to carry out the airlock installation, Middleton said: “It’s designed to be robust in the face of failures like this and we had many ways of working around it.” For example, he said: “software was being developed that would lock the shoulder joint in a way that would make the seven-degrees-of-freedom Canadarm2 operate like the six-degrees-of-freedom shuttle Canadarm. It would have been fine,” he said. “The astronauts would not have noticed any difference in the operation of the arm.”

Given the complexity of Canadarm2, this probably won’t be the last time problems crop up. Noting that Canadarm2 contains 20 computers, Sachdev said, “we didn’t have that level of sophistication and complexity in the original Canadarm, which was not as heavily digital. One normally expects, in complex systems, to get teething problems or what you might call infant mortality. So we were expecting a few hiccups and this may not be the end of it.”

“Some good did come out of the adventure: it gave engineering teams at the CSA, MDR and NASA a good shakedown cruise in troubleshooting. It was fortunate the problem did not require an instant solution, ” Middleton said. “We had time to work it. This helped to identify a lot of process-oriented things about how we should be working between ourselves, the CSA and NASA as problems come up. Now we’ll be able to take a look at how the whole thing functioned, because at some point, we’ll get failures where there is a time-critical problem and you’ve got to have your processes in place.”

Technical Challenges: Strong as an Ox; Gentle as a Lamb