Riding Rockets - Mike Mullane [224]
The nozzle damage turned out to be just one of the near misses for the STS-93 crew. Five seconds into flight an electrical system short circuit resulted in the failure of several black boxes controlling two of the SSMEs. Backup engine controllers, powered by a different electrical system, took over the control of those engines and there was no impact to their performance. But for eight and a half minutes, two ofColumbia’s engines were just one failure away from shutting down and forcing the crew into an ascent abort. The source of the short circuit was later isolated to an exposed wire.
Another shuttle near miss occurred on STS-112 when a circuit failure resulted in only one set of the hold-down bolt initiators firing at liftoff. In the launch sequence the hold-down bolts are exploded apart just milliseconds prior to SRB ignition so the rocket is completely free of the ground when the boosters ignite. Had the redundant initiators in the hold-down bolts not fired,Atlantis would have been still anchored to the pad at SRB ignition. The machine would have destroyed herself trying to rip free of the bolts.
STS-93 and STS-112 were saved by system redundancy, but there was another recurring problem on shuttle launches for which there was no redundancy to provide protection. Insulation foam was shedding from the gas tank and striking the orbiter. The phenomenon was first noted on STS-1 and was subsequently documented by photo imagery on sixty-four other shuttle missions. Hank Hartsfield and Mike Coats had observed it on our Zoo Crew flight in 1984. This foam-shedding anomaly was a violation of a design requirement, just as the pre-ChallengerSRB O-ring erosion had been a design violation. Nothing was supposed to hit our glass rocket, not even something as seemingly innocuous as the foam from the ET. But as hit shuttles kept returning to the Earth safely, engineers became ever more comfortable with accepting the design violation as nothing more than a maintenance issue—the foam strikes were requiring a handful of damaged tiles to be replaced between missions. The “normalization of deviance” phenomenon that had doomedChallenger in 1986 had returned to infect NASA and blind management to the seriousness of the foam loss problem. On January 16, 2003, eighty-two seconds into the flight ofColumbia, a briefcase-size piece of foam, weighing approximately one and a half pounds, shed from the ET and struck the Achilles’ heel of the shuttle heat shield, one of the wing leading-edge carbon panels. The impact blasted a hole of indeterminate size in that carbon. The damage had no effect on ascent andColumbia safely reached orbit. The site of the impact was not visible from the cockpit windows and the crew remained oblivious to the fact that their shuttle was mortally wounded. It could not survive reentry.
On the ground NASA engineers were aware of the foam strike—KSC cameras had recorded the incident. But these same engineers had no idea what, if any, damage had occurred and sinceColumbia was flying without a robot arm, they could not direct the crew to remotely survey the site (as we had been able to do on STS-27). A handful of engineers requested their management to ask the Department of Defense to use its photographic sources to acquire images of the impact site. Had these photos or a crew spacewalk determinedColumbia could not survive reentry,