After removing the ACIS from the spacecraft and the camera from the ISIM, the platform that holds the ACIS and HRC, no apparent reason for the door anomaly could be determined by visual inspection. The camera was then packaged and sent to Lockheed Martin Aerospace (LMA) where it was designed and built. It was backfilled with dry nitrogen while watching the separation between the door and the camera body using dial indicators to record the resiliency of the ``O''-ring as the force of about 240 pounds was removed. Initially, the door was in contact with the camera body (as determined by an electrical short between the two metal surfaces). As nitrogen flowed into the camera, the door and camera body separated with the separation increasing to about 15 mils when the pressure differential was reduced to zero. This proved that no cold weld had held the door shut in thermal vacuum, since the pressure differential there was also zero. The door was opened manually with only 23 inch pounds of torque. The actuator was removed and sent to Starsys, where it was examined to determine the state of the shear disk. The shear disk was found to have been ``functioned'', i.e burst. A meeting was held on July 7th between MSFC, LMA, Starsys, MIT and PSU to try to determine what tests should be done to determine the cause of the failure, to mitigate against future failure by redesign and/or procedure changes, and the testing required to ``prove'' that the ACIS door would open on orbit. A ``fault-tree'' was constructed by the assembled engineers and scientists and maintained by Alan Bean of MSFC. Leading candidates for the cause of the failure were: the ``O''-ring had frozen to the door which prevented it from opening since the ``O''-ring was captured in the camera body by a special groove that held it in place; the slow ramp-up of the voltage during early testing of the spacecraft had set the commands to open the door with a one atmosphere of pressure differential on it and thereby bursting the shear disk; and some ice had formed in the linkages that opened the door so that they were frozen in place/or the actuator had ice in it that jammed the actuator mechanism. The reason to suspect ice came from the fact that a great deal of water evaporated out of the optical bench during the test (5-6 liters were emptied from the cold traps at the end of the test). Since the camera was cold for much of this time, there was a suspicion that ice might form on some of the cold surfaces.
The engineers at LMA have performed numerous tests on the door and the actuator to try to find the cause of failure in thermal vacuum (TV) testing. It was found that the slow turn-on of voltage could have caused the door to try to open with a full atmosphere of pressure holding the door shut, but the subsequent tests of the temperature profile of the broken actuator plus microswitch behavior did not match the profile observed during the TV test making this explanation less likely, but not totally ruling it out. Slow turn-on of voltage is no longer part of the turn-on or testing procedures. Attempts were made to freeze the mechanisms using water and liquid nitrogen without success. An ``O''-ring was boiled in water for ten days to saturate it with moisture to see if it could be made to ``freeze'' the door shut at low temperature. Only a modest increase in adhesion of the ``O''-ring was measured and this adhesion was not repeatable. So far the cause of the door's failure to open remains a mystery.
In order to improve the chances of the door opening on orbit, the door will be opened warm (25C) to melt any ice that may have formed during the cold cycle required by lack of spacecraft power during orbital insertion. A potentiometer was added to the drive shaft that rotates to open the door, making it possible to determine if the door is opening when a measured amount of energy is delivered to the actuator. If the door is stuck (heaven forbid) the new procedure cuts off power to the actuator before the disk bursts. Then a series of temperature cycles will be initiated to try to break any stiction that may be holding the door shut. Tests are being conducted at LMA to determine if temperature cycling reduces stiction. An engineering unit has been evacuated and set on the shelf to reproduce the conditions expected prior to launch to see if the door will open. It will undergo vibration, vacuum and a cold cycle just as the flight unit is expected to be subjected to during the launch sequence. At the end of August the Flight Unit was put through a launch sequence to check the performance of the door after the modifications noted above were made. The door was opened three times with the camera warm after the door had been subjected to two weeks of cold, with no evidence of stiction. The door had only been closed for about two weeks prior to integration into the ISIM and the TV test. In the flight schedule the door will be closed for about two months prior to launch after the last time it is opened at the Cape. This length of time is required to check for leaks after the door has been opened - it takes about a month to get a good leak rate for the door, since the leak rate is quite low. The camera needs to be below about 10 torr to protect the optical blocking filter from the acoustic loads experienced during the launch on the Space Shuttle. We expect the ACIS door to open on orbit, based on all of our tests. But then, the door had always opened before the fateful TV test at TRW. The advantage with the new procedure is that it gives us more than one try to open the door and removes the possibility that ice could have caused the failure. The efforts to find the cause of the failure will continue at LMA, although with five months of work already expended on the problem, the chances don't look good for finding the cause in the near future.
- Gordon Garmire