A TASS reporter quoted Gudilin as saying that

“the platform of the cosmonaut emergency evacuation unit—this is where the system which ensures that the rocket’s gyroscopes are installed precisely is situ­ated—did not move away to a safe distance’’ [43].

Reporting on the scrub several days later, Aviation Week interpreted these statements as follows:

“The … launch attempt was scrubbed when the orbiter access arm on the launch pad’s left service structure failed to retract as commanded… The arm extends to the orbiter’s side hatch and allows cosmonauts and technicians to enter the vehicle on the pad. The arm also carries an umbilical connection which provides

ground support to the orbiter’s guidance and navigation system—specifically, its gyroscopes ’’ [44].

The platform in question actually was a 300 kg black plate mounted outside the intertank area of the core stage. It held three instruments needed for pre-launch azimuthal alignment of the core stage’s inertial guidance platforms. The plate, about the size of a small automobile, was installed on the core stage in the Energiya assembly building. After the rocket arrived on the pad, the plate was connected to a swing arm extending from the launch tower. This arm was situated several levels above the orbiter access arm. The retraction process took place in two steps: first, the plate had to disconnect itself from the intertank in three seconds’ time and only after that would the swing arm come into action to safely retract it from the rocket. What happened on 29 October 1988 was that the plate needed forty rather than three seconds to disengage. Sensing the sluggish movement of the plate, the rocket’s on­board computers stopped the countdown and no retraction signal was sent to the swing arm, which obediently remained in place.

Gudilin’s rather confusing statements, which created the impression that an arm had failed to retract from the rocket, led to some discussion at the State Commission the day after the scrub. Particularly unhappy with the confusion was Vladimir Barmin, the chief designer of the launch pad design bureau KBOM, which would have been held responsible if a swing arm problem had really been the culprit. However, the problem was with the plate disconnect mechanism, which was con­sidered the responsibility of the rocket team.

Despite the TASS statement on the 4-hour launch delay, another launch attempt later in the day was never seriously discussed. According to information released much later, the countdown could not be recycled for another attempt the same day if it was halted after the orbiter had switched to internal battery power at T — 80 seconds. Furthermore, by the time the countdown was halted, the umbilicals for thermostatic control of the hydrogen tank had been disconnected and temperatures inside the tank were slowly rising. Therefore, some 10 minutes after the hold was called and the problem had been identified, the team decided to delay the launch for several days and begin the lengthy process of draining the core stage and strap-ons.

Even that did not go entirely by the book. Draining of one strap-on booster’s liquid-oxygen tank went agonizingly slowly, a problem later attributed to a filter that had become clogged by contaminants in the oxidizer tank. Access to the filter at the launch pad was very difficult, causing fears the rocket might have to be rolled back to the assembly building, but in the end the problem was fixed on the pad. One eyewitness later said the clogged filter would probably have caused a catastrophe during launch. If true, the scrub had been a blessing. Also uncovered during post­scrub operations was that an accelerometer in Energiya’s tail section had been inadvertently mounted upside down [45].

The problem with the plate disconnect mechanism had been completely unex­pected. There had been no azimuth orientation system on the Energiya 6SL vehicle launched in May 1987 because there were less stringent requirements for precise orbital insertion of the Polyus payload. Neither had the problem surfaced during

the simulated countdowns at the pad in January-February and May-June. The causes of the mishap were investigated by a team headed by Vyacheslav Filin, a deputy of Gubanov. Part of the troubleshooting was to simulate the forces needed to detach the plate. This was done at the Energiya assembly building using the already assembled Energiya 2L vehicle and also a mock-up core stage intertank structure. In the end, the problem was traced to rubber dust covers on the plate that had somehow become sticky, possibly because of exposure to variable temperatures in the preceding months. In no time, engineers at the Progress plant in Kuybyshev redesigned the dust covers as well as the plate disconnect mechanism to ensure that the problem would not recur [46].