On May 29, 1990, at approximately launch minus 6 hours, the STS-35 (Astro-1) launch was scrubbed due to a hydrogen leak detected in the area of the Orbiter/external tank 17-inch disconnect panel during the cryogenic fuel tanking operation. Additionally, high levels of hydrogen were detected in the orbiter aft compartment. The payload was safed and reported no malfunctions. Protective dust covers were installed on the payload optics and a continuous nitrogen purge was initiated on WUPPE and HUT. Following inconclusive vehicle leak tests at the pad, the BBXRT argon was topped off and solidified prior to STS-35 rollback to the VAB on June 12, 1990.
The Orbiter was demated and returned to the Orbiter Processing Facility (OPF) on June 15, 1990. Following inconclusive tests on the Orbiter and ET 17-inch disconnect assemblies, the OV-102 (Columbia) 17-inch disconnect assembly was replaced with an identical unit from the OV-105 vehicle presently under construction in California. Columbia was transferred back to the VAB on August 2, 1990 and rolled out to launch Pad A on August 9, 1990.
Additional vehicle pad hydrogen leak tests were deferred in favor of actual pre-launch tanking. STS-35 launch was then scheduled for a September 1, 1990, launch. On August 29, 1990, during KSC operations and nearly coincident with payload door closure on August 29, 1990, the 16 kbps telemetry signal from the BBXRT was lost. Subsequent troubleshooting indicated that the signal was absent on both the T-0 umbilical and on the Orbiter Payload Data Interleaver (PDI) loop. The T-0 umbilical is a collection of electrical and fluid connections between the orbiter and the Pad which is disconnected immediately before liftoff at the T minus 0 time in the launch sequence. The PDI is the onboard data/communications system that interfaces the orbiter data systems with the payload data systems. With BBXRT telemetry absent through the T-0 umbilical, no information on the status of the argon dewars (pressure, temperature) was available and consequently the BBXRT launch Commit Criteria could not be verified.
The failure was isolated to the Shuttle Payload of Opportunity (SPOC) avionics in the Medium Rate Multiplexer (MRM) box. A transistor in the Pulse Code Modulator (PCM) low data rate generator electronics shorted to ground, possibly due to a conductive contaminant. The MRM box was then replaced on September 1, 1990, and the SPOC/MRM interfaces verified. BBXRT signal was verified at both the T-0 umbilical and the PDI enabling verification of launch commit criteria and flight rules. The argon in the dewars was resolidified and payload closeout was completed at 3:00 pm. September 2, 1990.
STS-35 launch was attempted again September 5, 1990. This attempt was scrubbed due to the detection of a higher than allowable concentration of hydrogen in the Orbiter aft compartment, specifically recirculation pump #2. Change out of the pump was made in preparation of a September 18, 1990, launch. During change out, a crushed seal was discovered, and it was thought that this improperly installed seal was the source of the previous hydrogen leaks. Preparations were continued for a September 18 launch.
The launch of Columbia was again scrubbed on September 17, 1990, during tanking when hydrogen was detected again in the aft compartment. Astro-1 was again placed under protective cover and in "safed" condition while work on the orbiter was ongoing at the Pad.
On October 8, 1990, Columbia was transferred from Pad A to Pad B to allow the next scheduled launch (STS-38) access to the pad. Pad A was needed to support STS-38 due to cleanliness requirements that could not be met at Pad B. When a tropical storm threatened the KSC area, Columbia was then rolled back to the VAB on October 9, 1990, and then again out to Pad B on October 13, 1990. Following repairs to the orbiter, a tanking test was conducted on October 30, 1990. The results of the tanking test revealed a hydrogen concentration level far below launch commit criteria specifications. Columbia was, as a result of the tanking test, certified as flight ready.
An unspecified problem with the STS-38 payload, which was scheduled to launch just prior to Astro-1, caused a 1 to 2 week launch delay. This directly impacted the launch date for Astro-1 which had been serviced and monitored continally. The Astro-1 launch date was announced as November 30, compressing the time between Atlantis and Columbia launches to two weeks. Additionally, the Astro-1 scientists became concerned that a November launch would impact observing opportunities for some of the high priority Astro-1 targets and requested an early December launch.
On November 27, 1990, the Astro-1 (STS-35) launch date was officially set for December 2, 1990, at the Flight Readiness Review (FRR).
-Text taken from a NASA press release.