7. Flight Operation

ASCA was launched by the seventh M-3SII on Feb. 20, 1993, 8 days behind schedule.

Vehicle manufacturing started at industries in 1991 two years before launch. No modification was required to the vehicle system because of engineering maturity. Only a few engineering experiments were to be carried out, i.e., flight test of an inertial navigation system based on a fiber-optic gyro scope and attitude determination of the spin-stabilized 3rd stage by monitoring the markings on the aft-end of the stage via TV.

The launch date was set for Feb. 12, 1993. Vehicle on-site assembly was carried out in Dec. 1992. The spacecraft was transported to ISAS KSC on Jan. 21. It was docked to the vehicle and the final check of the whole system started and ended without trouble. At this stage, 321 personnel were in ISAS KSC: 116 from ISAS and other universities (including 3 foreign researchers) and 205 from various industries.

On the evening of Feb. 11, the injectant, a 55% aqueous solution of NaClO4 55%, was loaded on schedule to the 1st and 2nd stage liquid injection thrust vector control (LITVC) systems. Visual inspection then revealed slight leakage of 2nd stage injectant, and the liquid was reported to be oozing out through the O-ring of the solenoid valve unit. At this moment, the team was rather optimistic, anticipating from experience that pressurization of the injectant, as required for the flight, would solve the problem. But it did not.

Replacement of the unit was initially estimated to take weeks and it might disturb the already tuned surrounding systems. Since the observed leaking rate was too low to incur injectant shortage during the flight, the problem at hand was to determine whether it would increase to an intolerable level during the flight. It was quite likely that the in-flight on-off operation of the valve would move the O-ring to a position to seal the leak. But, after discussions through the night, the launch date was shifted two days to confirm this using a spare unit. Straightforward on-site valve activation was impossible because it would pour the injectant onto the vulnerable front-end of the 1st stage.

The test carried out at a facility of a vehicle prime integrator using intentionally damaged O-rings, showed that, once operated, the leak would be sealed even if only half of the O-ring remained. At this stage, the cause of the leak was attributed to O-ring shrinkage due to some possible physico-chemical action with the purging agent of the unit that had been used also in the functional test of the LITVC system before vehicle assembly. For this purpose, a "FREON substitute" had been newly introduced from this flight replacing FREON 113 because of its ozone depletion potential. This also explained the fact that all eight independent units showed similar leakage. Manufacturing and testing records were traced and revealed that an O-ring, once swollen by contact with the agent, shrank slightly below its original size when it dried.

Then, when the problem seemed to have been settled, the degradation of electrical isolation of the valve driving circuit was revealed, probably due to the penetration of the leaked injectant which is a weak electrolyte. This had been monitored as a possible side effect. On-site experiment showed that it could be disastrous, especially through the depletion of the on-board battery power.

The launch crew devised a procedure to alleviate the operational difficulty and shorten the time required for replacement. Finally, it was decided to replace all the units and the launch date was set on Feb. 20. The tuned units were brought to KSC and integrated to the vehicle two days before launch.

The launch was perfect.

Post flight experiments, however, tell a somewhat different story: after a swollen O-ring is mechanically constrained by the grooves, a permanent concave deformation remains leaving leak paths since, in the flight mode, the O-ring never swells in the injectant. On the selection of the new purging agent, its compatibility with the O-ring material had been carefully investigated, but this effect was overlooked because a long time exposure to the agent under the constrained condition experienced in the actual launch preparation process had not been tested.