NASA managers assess
shuttle repair options
BY WILLIAM HARWOOD
STORY WRITTEN FOR CBS NEWS "SPACE PLACE" & USED WITH PERMISSION
Posted: December 19, 2007
NASA managers today reviewed options for fixing suspect low-level engine cutoff - ECO - sensors in the shuttle Atlantis' fuel tank. No final decisions were made and potential launch dates were not discussed. But shuttle Program Manager Wayne Hale, officials said, ruled out any discussion of launching Atlantis as is and engineers were cleared to begin preparations for removing foam insulation around a suspect "feed-through" electrical connector that passes sensor data into and out of the huge tank.
ECO circuit malfunctions blocked two attempts to launch Atlantis on a long-awaited space station assembly mission Dec. 6 and 9. During a fueling test Tuesday, test instrumentation spliced into the ECO sensor circuitry traced the problem to the three-part feed-through connector where wiring from the four ECO sensors and a so-called 5-percent sensor passes through the wall of the tank.
While the exact location of the fault is still under assessment, engineers are hopeful additional testing and data analysis will show the problem is in the external part of the feed-through connector, the plug on the outside of the tank that is relatively easy to access at the launch pad.
If so, that part of the connector could be replaced. If subsequent testing in a laboratory setting under ultra-low temperature cryogenic conditions duplicates the previously seen failures, NASA could press ahead for a third attempt to get the shuttle off the ground as early as mid January.
The current "no-earlier-than" launch date is Jan. 10, but that target was announced before engineers had a good idea of what might be needed to fix the problem. It's not yet clear whether Atlantis can be ready by then even if the problem is, in fact, in the external part of the connector.
"That would obviously be the best scenario," said a NASA official. "If it's that, if you could somehow guarantee that's what it is, you don't have to be invasive to the tank. ... There's some optimism it's in the external part of the connector."
But if it turns out the problem is in the internal parts of the connector hardware, engineers would have to remove foam from the bottom of the tank, open an access port and work inside the structure. Such invasive work likely would require a rollback to the Vehicle Assembly Building, delaying Atlantis' flight well beyond Jan. 10.
Sources said some engineers believe it makes more sense to accept a longer launch delay now to implement a permanent fix, replacing the entire connector with a redesign of some sort that features soldered pins to eliminate temperature-induced open circuits once and for all.
But any such redesign would face extensive testing above and beyond the work needed to fix Atlantis' tank and launch would face a more significant delay.
Another school of thought says it makes more sense to attempt a connector fix on the launch pad that would get Atlantis safely into orbit as soon as possible and to implement a long-term fix in parallel.
"There's a strong push to fix this problem on the pad and get the launch off," said one senior manager familiar with the discussions.
The downside to that approach, he said, is the possibility the data indicating the fault lies in the external part of the connector is incorrect or is being misinterpreted. In that case, additional problems could develop during the next launch attempt, forcing NASA to change its launch commit criteria or order a lengthy stand down and roll back for a more extensive repair.
A wild card in the discussion is what to do about fixing a recent problem with a space station solar array drive motor that is limiting the power available to the outpost. It appears a spacewalk by the station crew may be needed to replace a so-called beta gimbal assembly motor before Atlantis can fly regardless of work to fix the ECO sensors. While the station repair work could be added to the Atlantis mission, it would require major replanning.
As of this writing, no decisions have been made on how to proceed with either problem.
In the near term, Hale approved plans to begin preparations for removing foam from around the connector and to continue data analysis and engineering reviews. Another meeting to re-assess preparations and discuss repair options is planned for Dec. 27.
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.
NASA's launch commit criteria call for three of the four ECO sensors to be working properly for a launch to proceed.
During fueling for a Dec. 6 launch try, ECO sensors 3 and 4 "failed wet" about 35 minutes after they were submerged in liquid hydrogen. After the launch was called off, another sensor that shows when the tank is 5 percent full failed wet. After the tank was drained, ECO sensor No. 1 failed wet as well.
During fueling for a second launch try Dec. 9, ECO sensor No. 3 failed wet 24 minutes after all four were submerged. In both cases, the sensors returned to normal operation after the tank was drained and temperatures rose.
During Tuesday's tanking test, sensors 2 and 3 showed open circuit failure indications for a few seconds before returning to normal operation. Sensor 1 failed shortly thereafter and stayed that way for the duration of the test. Sensor 2 worked properly throughout the tanking test but sensor 3 failed several hours after recovering from its initial open circuit indication.
Hale said Tuesday the time domain reflectometry instrumentation used during the fueling test exonerated the sensors themselves. The data indicate the only sensor circuits with real problems are ECO 1 and 3. The intermittent glitches with sensors 2 and 4 are believed to be related to the pass-through connector problem that has affected sensors 1 and 3.