Shuttle launch date up in the air as repairs are ordered
BY WILLIAM HARWOOD
STORY WRITTEN FOR CBS NEWS "SPACE PLACE" & USED WITH PERMISSION
Posted: December 27, 2007
NASA managers today cleared engineers to remove the external components of a suspect feed-through connector built into the wall of the shuttle Atlantis' external tank in a bid to fix intermittent electrical problems with low-level engine cutoff - ECO - sensors that derailed launch attempts Dec. 6 and 9. The external fittings will be replaced with soldered pins and sockets like those developed and successfully flown by tank builder Lockheed Martin for its Centaur rocket stages.
Shuttle Program Manager Wayne Hale said today the work likely will delay Atlantis' launch "a few days to a couple of weeks" beyond the previous Jan. 10 target. But that "no-earlier-than" date was little more than a placeholder intended to ensure the launch team enjoyed a few days off over the Christmas holiday. As such, it was not based on any actual repair schedule.
NASA has now settled on a course of action, but Hale said today he was not ready to discuss when Atlantis might be ready for a third launch try. During an afternoon teleconference, he told reporters "I'm not going to make a launch date announcement ... because we're in the middle of troubleshooting and repair. Until that gets a little bit further along, I actually have no valid dates to give you."
"To avoid what I think would be a totally misleading headline along the lines of 'NASA delays the space shuttle again,' we're just not going to give you a launch date because that, in fact, would not be accurate," he said.
But sources familiar with the discussion said the feed-through connector replacement and subsequent testing could delay launch to the last week in January when all is said and done, and that assumes the work goes smoothly.
Atlantis was grounded Dec. 6 and 9 when intermittent failures of ECO sensors at the base of the hydrogen tank occurred during fueling and later, draining. A fifth sensor, which indicates when the tank is 5 percent full, also malfunctioned when the tank was drained.
The ECO sensors are part of a backup system intended to make sure the shuttle's main engines don't inadvertently drain a tank dry after some other problem - a leak, for example, or an improper hydrogen-oxygen mixture ratio - used up propellant at faster than normal rates. An engine running out of hydrogen during normal operation likely would suffer a catastrophic failure.
The wires that carry signals from all four ECO sensors and the 5 percent sensor pass through the same connector in the wall of the external tank. The three-part connector features a pass-through fitting with male pins, embedded in glass, on both sides. Wires from the sensors inside the tank terminate in a female connector inside the tank that is plugged into the male pins of the pass-through. A similar female socket plugs into the pass-through on the outside of the tank.
The central feed-through fitting features 37 pins on each side, only 10 of which are actually used. Each pin is roughly two-and-a-half inches long and one-sixteenth of an inch in diameter. The 37-socket connectors that attach to each side of the central feed-through are a little more than an inch across, Hale said.
Based on data collected during a fueling test Dec. 18, engineers believe the problem involves gaps in pins and sockets on the external side of the feed-through connector when the system is chilled to ultra-low temperatures. They believe the sensors themselves are healthy and that only two circuits are actually experiencing problems: ECO sensors 1 and 3.
"We have allowed the team that did the troubleshooting to very thoroughly go through all of that data," Hale said today. "They have told us they are sure that the problems that we're seeing reside in that series of connectors (on the external side of the interface). Where exactly in that series of connectors is a little bit open to interpretation.
"We would like to get the whole thing out and send it to a laboratory for bench analysis. Unfortunately, as we discussed today, that's probably not a really good option. So what we decided to do today is to tell the team to take the next step and remove the pass-through plug and the external connector and some length of the wiring on the outside as a unit and send that back to the lab."
The removed components will be bench tested under cryogenic conditions in an attempt to duplicate the observed failure mode. In the meantime, replacement hardware featuring soldered pins and sockets will be installed on the tank at the launch pad. The fix is similar to changes Lockheed Martin made several years ago to correct problems with sensor circuitry in the Centaur stage used by unmanned Atlas rockets.
The internal connector that plugs into the pass-through pins from the inside of the shuttle tank does not have enough slack in the wires leading to the ECO sensors to permit the internal connector's removal. It will be visually inspected, however, and possibly X-rayed to check its integrity. But Hale agreed that NASA will face the possibility that the repairs ordered today might not fix the problem.
"It is a possibility that the internal connector is involved," he said. "However, all the physics-based discussion of the kinds of things that can happen point to something happening on the external connector. The Centaur problem was a problem with the external connector. And so while it's not a slam dunk, guaranteed situation, the preponderance of evidence indicates that the external connector is the problem."
Engineers are keeping their fingers crossed because "getting the internal connector out is much more invasive and there's much greater risk of flight hardware damage because you have to actually go into the hydrogen tank to get to the internal connector to replace it," Hale said. "We can pull it out just far enough to do a visual inspection, perhaps take some X-rays, but you cannot replace that connector from the outside of the tank. To get into the hydrogen tank, then you have to roll back to the VAB (Vehicle Assembly Building) and do all the work that's involved there.
"Since the preponderance of evidence to a very high degree point to the external connector - we can get to that, we know how to fix it, that's been our experience with our sister program - that's where we're going to concentrate our efforts. Of course, if we're unsuccessful we will come back and look at the internal connector again."