Extremely rare Apollo Guidance Computer prototype—the brain behind the lunar landing
Amazing Block II prototype Apollo Guidance Computer (AGC), measuring 24″ x 12.5″ x 6.5″, with its black NASA/Raytheon Co. label on the back, reading: "Apollo G&N System, Computer Assembly, Part No. 2003100-051, Serial No. RAY 8, Cont. No. NAS9-497, Designed by MIT Instrumentation Lab, Mfg by Raytheon Co." This is the eighth of fifteen prototype Block II computers (out of a total of 57 Block II AGCs produced). The computer is offered with six core rope memory modules, representing the partial software from three different Apollo Guidance Computer programs, as detailed below. Aside from these six fixed memory modules, the computer appears to exactly match its specified configuration, as detailed in the list of materials on the MIT Instrumentation Lab/Manned Spacecraft Center's Drawing No. 2003100: it retains twenty-four logic modules, five interface modules, one oscillator module, one alarm module, two erasable driver modules, one current switch module, one erasable memory module, two sense amplifier modules, one strand select module, two rope driver modules, and two power supplies. To the best of our knowledge, this is the very first AGC to appear at public auction.
Based on MIT Instrumentation Lab/Manned Spacecraft Center's Drawing No. 6003001, this AGC (S/N RAY 8) was originally assigned to Guidance & Navigation Subsystem 602, the guidance system for the very first production, man-rated Lunar Module, known as Lunar Test Article 8 (LTA-8). At some point, before LTA-8 underwent thermal vacuum chamber testing in Houston, this AGC was removed from LTA-8 and replaced with AGC S/N RAY 14. Nevertheless, AGC S/N RAY 8 received nearly all of the same changes that the thermal vacuum test AGCs did: it went through hardware revisions that brought it from Part No. 2003100-021 through 2003100-091, before being returned to its -051 state. The difference between the -051 and -091 configurations is the application of Engineering Change Procedure (ECP) 474, which is the addition of a grounding test connector cover. This AGC does not currently have this installed, so whoever took it off apparently dutifully updated the 'dash number' back to -051.
The Apollo Guidance Computer was a technical marvel: in the era of room-sized computers, NASA allocated one cubic foot on their spacecraft for the electric brain that would be responsible for guiding humans to the lunar surface and safely returning them home. It was up to the best and brightest at the MIT Instrumentation Lab to make it fit. Rather than using the large vacuum tubes or big discrete transistors typical in computers of the time, MIT engineers pioneered the application of integrated circuits—microchips—to accomplish the same task in a diminutive package. During 1963, the Instrumentation Lab consumed 60 percent of the integrated circuit production in the United States, and by 1964 Fairchild Industries had shipped more than 100,000 ICs for use in the Apollo program.
The AGC hardware was thus a combination of cutting-edge technology and old-school craftsmanship: while these innovative, mass-produced chips made their way into the AGC's logic modules, the computer's mission-critical software was stored in handmade 'rope memory,' contained inside its fixed memory modules, which could not be erased, altered, or corrupted. This rope memory required absolute precision and was sewn by workers recruited from local textile factories: copper wire was woven in and around ring-shaped magnetic cores, with each wire threaded through the core representing a binary "1," and each wire bypassing the core representing a "0." It took eight weeks for the workers to weave the memory for a single flight computer, at a cost of $15,000 per module.
The six core rope memory modules in this AGC are as follows:
Module P/N S/N Contained Program
B1 2003972-371 161 Sundance revision 292, the original (unflown) release of the Apollo 9 LM flight software
B2 2003972-521 243 Colossus revision 236, the original (flight-equivalent) release of the Apollo 8 CM flight software
B3 2003972-391 160 Sundance revision 292, the original (unflown) release of the Apollo 9 LM flight software
B4 2003972-541 242 Colossus revision 236, the original (flight-equivalent) release of the Apollo 8 CM flight software
B5 2003972-421 159 Sundance revision 292, the original (unflown) release of the Apollo 9 LM flight software
B6 2003972-641 237 Sundance revision 306, the final (flight-equivalent) release of the Apollo 9 LM flight software
Developed using a mix of assembly language and an interpreted mathematical language, the software contained on these modules was as innovative, and as important to mission success, as the pioneering hardware. Many of the design principles developed at MIT for coding the AGC became foundational to software engineering in general—particularly in the design of critical systems that rely on asynchronous software, priority scheduling, fault-tolerance, fly-by-wire capability, and human-in-the-loop decision making. These modules in particular hold a special significance, as they contain portions of the software equivalent to that which guided the first manned flight of the Command Module (on Apollo 8, man's first flight to the moon) and the first manned flight of the Lunar Module (on Apollo 9). It is worth noting that two Colossus 236 modules are a special case, as Colossus 237 is what actually flew on Apollo 8. However, these two modules do contain the as-flown software: Colossus 237 differed from Colossus 236 only in module B5, so when the program was re-released, only module B5 was remanufactured.
Includes an original AGC Handbook in its MIT Instrumentation Laboratory/Apollo Guidance and Navigation binder, containing detailed schematics and diagrams for the computer; an original oversized 34″ x 22″ translucent mylar print of the "PGNCS & Block II CSM Functional Interface Diagram," showing the flow of spacecraft devices and their interaction with the computer; and a signed copy of Sunburst and Luminary: An Apollo Memoir by Don Eyles, a software engineer at MIT’s Instrumentation Laboratory who worked on the AGC's code for the lunar landing and famously improvised a workaround for an abort signal on Apollo 14 that allowed the mission to continue. From the Don Eyles Apollo Computer Collection.