One Giant Leap for Engineering Kind: How the Moon Landing Gave Rise to the IC

In 1961, President John F. Kennedy set in motion one of humanity's greatest technological achievements when he extended a historic challenge to NASA, newly created in 1958 from the former National Advisory Committee for Aeronautics (NACA).

“I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the Moon and returning him safely to the Earth.”

When Kennedy made his proclamation during a joint session of Congress, none of the mechanical hardware, electronics hardware, software, or infrastructure existed to achieve such a goal. At the time, only two human beings had ever been to space and returned safely. Yuri Gagarin, from the then Soviet Union, had completed one planetary orbit on April 12, 1961, and Alan Shepard, from the U.S., had taken a suborbital space flight on May 5 of the same year.

 

Earthrise from Apollo 8, orbiting the moon, December 24, 1968. Image used courtesy of NASA

 

Kennedy’s speech may have brought the goal into the public consciousness, but institutions and companies like MIT, IBM, and Fairchild Semiconductor were already inventing and improving the electronics technology that would make the mission happen. One of the key innovations necessary—the integrated circuit—was quietly being readied to leave the lab at Fairchild, and the MIT Instrumentation Lab (MIT/IL) was preparing a proposal that would use those chips to develop and build the Apollo Guidance Computer (AGC).

 

Competition for the Flight Computer Design

For the moon landing mission to succeed, NASA needed flight computers for the command module and lunar landing system. Nothing at the time was small, lightweight, or performant enough to take on the role of the flight computer. IBM created a design for the Saturn V rocket autopilot computer called the Launch Vehicle Digital Computer (LVDC). The LVDC used transistors and diodes fabricated in what IBM called a Unit Logic Device (ULD). The ULDs looked like ICs, but were much closer in design to a hybrid module; it included a transistor, two diodes, and two thick-film resistors mounted directly on a ceramic substrate.

 

Fairchild's planar dual three-input NOR gate from Block II AGC. Image used courtesy of Wikimedia Commons (Public domain)

 

MIL/IL did not believe a ULD-based design could meet the specifications for the command and lunar modules. They were convinced that an AGC built with the newly invented integrated circuits would be faster and more reliable than if built with discrete transistors or ULDs. 

 

Development of the Planar Integrated Circuit 

The first real integrated circuit came from Jack Kilby of Texas Instruments. In 1969, he demonstrated an oscillator circuit made from a single piece of transistor material. While hardly recognizable by today’s standards, it was the first complete functional integrated circuit. Its major drawback was that the individual components in the silicon—a transistor, a capacitor, and two resistors—had to be manually connected with microscopic gold wires.

The IC didn't resemble what they typically look like today until Fairchild cofounder, Jean Hoerni, developed the “planar” process, in which a layer of silicon dioxide is deposited on the surface of the chip to act as an insulator. By using an insulating layer, conductive traces can then be chemically deposited to connect the components in the chip. 

 

ICs Achieve Proof of Viability

Fairchild's Bob Norman visited MIT/IL in 1961 to demonstrate the new integrated circuits to the AGC team. Suitably impressed, Eldon C. Hall, the assistant director for the Apollo program, directed engineer David Hanley to order 100 of the chips. Hanley demonstrated a 2.5x increase in speed over discrete transistors. Hall then asked for and received approval from NASA to make the switch from transistors to ICs for the AGC design.

In the end, MIT/IL was given the task of designing and building the AGC while IBM supplied mainframes to be used on the ground in Houston, Texas. These mainframes were used as the primary navigation system, with the in-flight AGC taking a secondary role. By the time MIT/IL was ready to start building, Fairchild had developed a complete family line of different digital logic ICs. To limit the number of items to prove out, Hall elected to use a single type of component, a three-input NOR gate. Philco then took on the task of manufacturing the chip at volume production.

 

AGC: Designed as a General-Purpose Computer

MIT/IL developed the AGC as a general-purpose computer rather than as a mission-specific instrument. It was primarily built with the dual three-input NOR gate chips designed by Fairchild and is considered the first computer built using integrated circuits.

 

Apollo Guidance Computer, with DSKY Interface. Image used courtesy of Wikimedia Commons (Public domain)

 

In 1965, the Block II AGC used about 4,100 updated Fairchild chips with dual three-input NOR gates in a flat-pack chip. The dual gate design yielded a unit that used less power, offered more processing power, and cost less than the original Block I design. The Block II AGC was used successfully on all crewed Apollo missions.

 

NASA Pioneered Modern Software Engineering

As a multipurpose computer, AGC needed more advanced programming than was typical of the time. Software engineer Margaret Hamilton joined the team at MIT/IL to develop the mission software for the AGC. As director of the software engineering division, she oversaw a team designing and testing AGC's software. 

 

Margaret Hamilton with AGC software code listings. Image used courtesy of NASA

 

Prior to the AGC project, software development was not considered a science or a branch of engineering. Hamilton coined the term “software engineering” and, in addition to the task at hand, developed core processes that would later define the field as a discipline.

Hamilton and her team developed a simple interrupt-driven, real-time operating system with batch job scheduling, cooperative multitasking, and error handling for the AGC. The error handling was given a trial by fire minutes before Apollo 11 touched down on the moon when events were being triggered faster than the computer could deal with them. Due to the error handling feature, astronaut Buzz Aldrin was able to recover the computer, and the landing went on successfully.

 

IC Development Skyrockets

Integrated circuit technology advanced so fast that by the time the first humans boarded an Apollo flight, ICs containing hundreds of devices per chip were on the market. Meanwhile, the AGC was still using the original Fairchild design. It was during this time that Gordon Moore predicted Moore’s law, stating that the density of transistors in computer chips would double about every two years.

With the success of the lunar landings and the demonstrated reliability of integrated circuits in space, the technology’s viability was well established. Key figures from Fairchild and other major semiconductor companies involved in the AGC project went on to create the foundation of what is now known as Silicon Valley.