The Apollo 11 had a number of separate modules, each with their own function during the voyage. The command module, which was the conical module at the tip of the spacecraft, was used to house the crew and made the entire trip from the earth to the moon and back again. The command module was attached to a cylindrical service module, which contained fuel and jets for driving propulsion of the spacecraft through space once the main stages of the Saturn V rocket were discarded. The service module made the trip from Earth to the moon, but did not land back on Earth. Attached to the service module was the lunar lander module which contained the lunar lander. The lunar lander was used to actually land on the moon while the service module remained in orbit around the moon; in this manner relatively little fuel had to be spent to launch the lunar lander from the moon’s surface back into orbit where it would dock with the command and service module.
The command module had one AGC which was connected to the jets, radar system and other sensors of the spacecraft and provided a digital autopilot. The lunar lander also had an AGC, with essentially the same hardware but programmed with different software to control the lander.
Any defects or errors in either of these AGCs could result a crash or the crew becoming trapped in space. Once the spacecraft left Earth’s protective atmosphere, the risk of errors in either of the AGCs increased dramatically as in space the AGCs were exposed to large amounts of radiation which could cause changes to bits in memory. Several solutions to this problem were proposed, such as a triple redundant system in which all calculations were carried out redundantly by three identical units and a voting mechanism would choose a result if the same result was returned by at least two of the units. This had as a disadvantage that the weight of three computers had to be carried on both the lunar lander and the control module. As one AGC weighed in at about 32 kg, this would add up quickly.
Another proposed solution was to build in error checking codes and fault detection units in the computer. US patent application no. 679,055 which was filed in October of 1967, describes a “self testing and repairing computer”, which has many useful features, including error detection, rollback of a program until just before an error was detected, the use of self-correcting error codes, and automatic swapping out or deactivation of defect parts of a computer . You can find the corresponding patent by entering “US3517171” in the textbox at worldwide.espacenet.com.
In the end, neither of these solutions was used in the Apollo 11, but instead a sturdy casing was provided around the AGCs to withstand the radiation. To further improve the astronauts’ safety and peace of mind a manual override was provided for most functionality of the spacecraft. Nonetheless, many of the error detection and correction schemes developed for the Apollo 11 were used in later unmanned space missions for which a manual override was not an option. For those interested, a great presentation on the AGC can be found on https://media.ccc.de/v/34c3-9064-the_ultimate_apollo_guidance_computer_talk.
This blog post is part of a series of 'Man on the Moon' articles to honour the 50th anniversary of the Apollo 11 moon landing.