Apollo at Draper
Draper designed and developed the guidance, navigation, and control systems for Apollo’s command and lunar modules. The software for the on-board computer was written and tested at the Lab. In addition, the Lab provided training for the astronauts who used the system, developed test procedures for the G&N systems, monitored the production of the systems by industry contractors, and supported the NASA mission controllers during the missions.
 
In August 1961 NASA awarded the Draper Laboratory, then the MIT Instrumentation Laboratory, a contract to design and develop the Apollo guidance, navigation and control system. Earlier work on a Mars Probe had provided the basic design. The system was intended to be autonomous, not dependent on ground control for critical mission events, and the astronaut crews were envisioned as active operators of the system, rather than passive observers.

The major system elements were:

- an inertial guidance unit (gyro stable platform with   accelerometers)
- a space sextant
- a general-purpose digital computer with a keyboard   and display unit for crew interface
- supporting electronics

Hardware
The major hardware elements of the system were the inertial system, the optical system, and the digital computer. The Apollo inertial system drew heavily on the Polaris Fleet Ballistic Missile guidance experience at the Lab. The Apollo unit was designed as a three gimbal system to achieve greater mechanical simplicity and a higher accuracy in the stellar alignment. The lack of a fourth gimbal introduced the possibility of gimbal lock at certain spacecraft attitudes which would necessitate a time-consuming realignment of the system. In the course of extensive flight experience, this presented no problems. Questions arose about the reliability of the system; a failure of a single gyro wheel would threaten mission success and jeopardize crew safety. The cumulative flight record of over 2,500 hours of operation without a system failure is a tribute to the design, testing, and operational techniques produced by the Lab.

Apollo Guidance Computer
The requirements for the on-board computer were for a highly reliable unit with sufficient memory and computational capacity, yet with minimal size, weight, and power consumption. The configuration chosen was a 16-bit parallel, general-purpose, real-time control computer, using integrated circuit logic. A single type of circuit logic, a three-input NOR gate, was used exclusively in the design, giving greater reliability. For the Block II design, there were 36,864 words of fixed (read-only) core rope memory and 2,048 words of erasable (read/write) memory. The nature of the core rope required freezing the flight computer programs early in order to provide time to manufacture the program. Although this may seem to have been a disadvantage, it eliminated the possibility of risky last-minute changes to the flight programs, and assured the delivery of thoroughly tested programs prior to flight.

An ingeniously designed display and keyboard unit provided input and output for the crew. The keyboard allowed the entry of the ten numeric digits plus seven other coded functions. The display included three five-digit numbers (plus sign) to indicate numeric data, and three two-digit numbers that identified the Verb, Noun, and Program selected. The verb-noun pairs permitted a language of action-object, such as “display-gimbal angles” or “load-star number.” The program number identified the major computational activity currently operating. This compact and relatively simple unit supplied a powerful and flexible means of communicating with the computer.

A key factor in developing the guidance system hardware was the self-imposed design decision that, as far as possible, the guidance system elements would be identical in the Command Module and the Lunar Module. This yielded great savings in manufacturing, testing, and crew training.

The Block I spacecraft flew with analog attitude control autopilots. In June 1964 the decision was made to use digital autopilots in the Block II system, which increased the demands on the flight computer.

Apollo Block II command Module GN&C Block Diagram, click image for interactive viewer.

Software Design and Development
Producing the software for the manned missions turned out to be a major effort. The flight programs for the early unmanned flights were produced by small, dedicated groups led by a chief engineer-programmer. For the manned flights, particularly the lunar missions, the number of program elements and the number of different programs that were needed increased dramatically.

The software requirements for each mission were documented in a multivolume Guidance System Operations Plan, which programmers used as the basis for writing the code. These requirements were reviewed and revised as the programmers gained experience. A comprehensive testing effort accompanied the release of each flight program, to assure the highest level of confidence in the system performance. The most detailed testing used a high fidelity all-digital simulation of the computer, spacecraft hardware, and mission environment. The programs were tested with real hardware in the System Test Lab. A hybrid (analog/digital) simulator that included mockups of the cockpits of the Command Module and Lunar Module allowed testing of crew procedures in real time. Still another simulator using real system hardware and a surplus radar mount to simulate spacecraft motion was located on the roof of the Lab building at 75 Cambridge Parkway so that sightings could be made on real stars.

Information condensed from "The History of Apollo, Onboard Guidance, Navigation, and Control," by David G. Hoag, published in the AIAA Journal of Guidance Control and Dynamics in 1982.