February 4, 1971. Mission Commander Alan Shepard, the first American in space, has finally made it into Lunar orbit. His Lunar module pilot Stu Roosa stands at his side abord the Lunar Module Antares. The command module pilot Ed Mitchell is on a separate orbit in the Command Module Kitty Hawk, from which they detached several hours ago.
Al and Stu are going through the pre-descent checklist, getting the Lunar module ready for the powered descent and landing that is scheduled to occur several hours into the future. A light flashes on the control panel. It is the abort light. The automatic abort system has engaged. Fortunately, that system will not do anything unless the spacecraft is in its decent and landing mode.
“Antares, we are getting a signal that your abort system is engaged, please check the switch,” comes the voice from the capcom (capsule communicator) in Houston.
“Houston, we see the abort light, but the switch is not engaged,” Al announces. He takes a small cylindrical object out of a pocket on his space suit and taps the panel. The light goes off.
The voice from Houston says, “Antares, we are no longer reading the abort signal.”
“Houston, I tapped the panel with a cigar that I was keeping for after splashdown. There must be a ball of solder floating around behind the panel.” At that point, Al Shepard realized that they had a big problem. A small ball of metal solder, floating behind the panel in zero g, was intermittently completing a circuit which should only have closed if Al Shepard pressed the Abort button. If that ball of solder closed the circuit during the descent to the surface of the moon, the abort would engage, causing the LM ascent stage to fire its engines and detach from the descent stage. The attempt to land on the moon would be ruined, for the second time in a row. The previous year, three astronauts had nearly lost their lives due to an explosion of an oxygen tank in their service module.
The dilemma, how to bypass that abort button while still maintaining the ability to abort if they needed to. The button needed to be somehow physically locked out from affecting the operational and guidance computer. There was no way to open up the panel and vacuum out the piece of soldier, so they would have to lock out the switch programmatically within the computer, and because the programming of the computer was hard-wired, this was going to be a difficult solution to find.
The software team at the Massachusetts Institute of Technology jumped into action. They needed to find an elegant fix that would work around the problem. The first solution that they came up with had the problem that it could not be instituted until after powered descent began, and if the switch closed while Al was programming the fix in, the abort would happen. They needed a fix that could be set before powered descent began and which could be easily reversed if an actual abort needed to happen.
The solution resided in a single bit of Random-Access Memory within the guidance computer. During the design process, it had been determined that once an abort condition was met, and an abort happened, the abort condition would generally continue, but the actual physical abort process (severing the connections between the LM ascent and descent stages, firing the ascent engine, and switching the guidance system from the primary navigation guidance system (PNGS) to the abort guidance system (ags)) could all only be done once. Therefore, the computer already had a means for locking out the abort once it had already happened. When an abort was initiated, a single bit of data (ABORT = TRUE) would be set from zero to one. If this data point was equal to one, the computer could not initiate an abort because one was already underway. A procedure was quickly worked out for manually, through data entry on the guidance computer’s input panel, switching this bit of data to one. If they actually needed to abort, Al would have needed to trigger the abort manually by turning a dial to switch from PNGS to AGS, and then pressing a button (after removing a protective cover) to separate the stages, and then firing the ascent engine. He could do all that in about three seconds, so the prospect of an abort had become marginally more dangerous, a single button press changing into a three second process, but there was certainly no expectation of an abort.
The key sequence was entered, and the ABORT=TRUE bit was now on. At the proper place in the orbit, the descent engine was fired. During the several minutes of the descent and landing process, the Abort switch closed multiple times, and the computer blissfully ignored it. After a perfect landing, Al Shepard hit the longest drive in the history of golf, when he struck a golf ball and sent it sailing “miles and miles” in the low gravity and no atmosphere environment.
And that is the story of the only computer tech support call that was ever initiated at a distance of a quarter of a million miles.