An Apollo Computer Lands on the Moon in Naples, FL
December, 1972. 380,000km (235,000 miles) away from Earth, in a small, spidery looking spacecraft, a Navy pilot and a geologist descend to the surface of the Moon, flying between the massifs of the Taurus-Littrow valley, near the grey and powdery shores of the Sea of Serenity. In a cabin the size of a couple phone booths, the two astronauts are busy reading dials and communicating with Mission Control back on Earth. Perhaps most crucially, during their 15 minute descent from a few thousand kph across the planet’s surface to a standstill sitting on its face, they read data and key in codes on a small, calculator-looking keypad with an electroluminescent screen, whose only readout is three lines of five digit numbers each. This is the DSKY (pronounced “dis-key”), the keyboard and screen, of the Apollo Guidance Computer (AGC). On this day in December, the AGC, overseen and maintained by the astronauts, pilots its final descent to the surface of the Moon.
The AGC was designed at MIT for the Apollo spacecraft to navigate to the Moon, execute burns, course corrections, autopilot maneuvers, and read information about the spacecraft’s status, as well as other critical tasks. According to Apollo 15 commander Dave Scott, the AGC is the single most important invention NASA created during the Apollo era. Without it, Apollo could not go to the Moon. It extends further than that—the AGC was one of the first useful portable computers, about the size of a briefcase, that could multitask, juggling several functions simultaneously. It was the first ever fly-by-wire system implemented into a flying machine. It was the first computer to rely on integrated circuits, making it a close ancestor of the chips we use today. In the circles of software engineering, computer technology, and circuitry, the AGC is well-respected and cited as a landmark invention. In fact, the term “software engineering” was popularized by Margaret Hamilton, who was important to the AGC’s creation at MIT.
July, 2019. 380,000km away from the Moon, in a Naples, Florida neighborhood, a small group of computer scientists and spaceflight enthusiasts descend into a lamplit living room with the only operational Apollo Guidance Computer in existence. One year ago, the computer, its owner, Jimmie Loocke, and its chief operator, Mike Stewart, were at Spacefest IX in Tucson demonstrating that the AGC was in good enough condition to be fired up for the 50th anniversary of Apollo 11. Loocke’s team and the AGC pulled through after thousands of man-hours and personal financial investment (as a privately owned AGC, no support is received from NASA or others). Now, before taking the computer on a brief tour to the Cradle of Aviation museum in New York and then to Boston for presentation at the computer’s home, MIT, Jimmie Loocke brings the computer to the home of Eldon Hall, who is often credited with its development.
I arrive to an open door. The ten or so people have been going in and out, and the nook table is covered in tools, paperwork, and computer hardware that is 50+ years old. A foldout table against the wall by the kitchen is occupied with the hefty, metal AGC, a working replica of a DSKY, voltmeters, circuitry, and Mike Stewart’s laptop. On the other side of the room, Loocke and a few others are sitting with Eldon Hall, so I pull up a chair and catch up with Jimmie before he introduces me to Eldon. It’s a moment I was looking forward to during the three-hour drive down to remote southwest Florida.
Eldon Hall was one of the directors of the AGC project at Draper Labs, then part of MIT, through the 1960s. It was he who led the hardware team, and held a central role in directing the computer’s development. Hall is best known for being one of the earliest proponents of integrated circuits, to which he faced some scrutiny, with the field holding mixed opinions on their worth and usage at that time. Having successfully achieved their use in the Apollo Guidance Computer, he is viewed in field-specific circles as helping to usher in a new world of computing.
Now 96 years of age, Hall quietly watches a team of people circling his invention. I ask if he ever reflects on the modern world of technology, seeing his role in its development. He answers in the affirm, “I started the carriage down the hill, pushing it over the top, and it started running on ahead.” Developing that idea into any advice he may have for the current players, he recommends: “They can make things easier for old people to use. That’s about the only thing I’d like to tell them all.”
Moore’s law states that integrated circuits double in processing ability every two years. The AGC in Hall’s living room was built in 1965 for LTA-8, meaning the integrated circuit he is credited for introducing to practical usage has doubly accelerated in capability 26 times since then. That’s 2 to the power of 26. “Yes, Moore’s law,” he replies, “now, I just stay out of it.” Eldon’s quip is met with laughter from the group, and he jokes that it’s all his fault. We assure him that it’s not his fault, per say, but perhaps he did contribute at least a piece of the pie. “It wasn’t all me anyway. I was the one that drove the horses and pulled the wagon forward, trying to keep all of them pulling the same way. That’s the way I put it. I didn’t design that Apollo computer,” Hall reflects with humility, “I had the vision. It all had to be reliable.”
Reliable, in fact, it was. The AGC performed beautifully in Apollo, only having one hiccup during Apollo 14, which MIT successfully scrambled up a solution for as Alan Shepard and Edgar Mitchell prepared to land on the Moon. The Apollo 11 program alarms, though distracting to the landing process, were not a computer fault, rather, the computer reporting that it was receiving too much information to juggle at once. The AGC was designed to accommodate this condition. In his book, Journey to the Moon, Hall elaborates on the computer’s priorities. Programs running guidance and navigation would always be the first job the AGC prioritized.
Once the AGC was interfaced with the landing radar, which provided accurate, real-time altitude, it could land the Lunar Module on the surface by itself, theoretically. Astronauts introduced themselves into the computer’s operations in the final stage, picking a spot on the Moon without boulders or craters. This is the very definition of fly-by-wire: a pilot inputing commands into the computer, and the computer executing the action. When Armstrong was said to “take over” the LM’s controls, he actually instructed the computer to level off the descent until he could find a suitable place to set the Eagle down.
The AGC wasn’t Eldon Hall’s introduction to aerospace, however. He also worked on the Polaris missile, as well something called “the Mars Project,” which planned to do a fly-by of Mars and take photographs of the red planet, all before Apollo. “I remember one of the days before Apollo, Wernher von Braun came by the lab to talk to Dr. Draper. We were taking Wernher von Braun to the computer we had in the lab for the program. There was this little box. Balls would come out of one side of the box, bounce along the floor, and then fall through the floor. On the other side, balls would come in and bounce along, and fall in the floor,” Hall explains how MIT would measure a computer’s ability to multitask: balls in a pneumatic system would be “juggled” by the computer as a test, “so the screen was showing that. Von Braun’s comment was ‘that proves the computer is on the ball!’ It was the funniest thing.”
I move over to the group of dedicated experts orbiting the AGC in preparation for the demonstration to one of its parents. Ken Shirriff, a software engineer, recalls his introduction to the restoration project, being brought in by Mike Stewart alongside Carl Claunch and Marc Verdiell (aka Curious Marc). “The four of us ended up in this hotel room in Houston in November spending a week testing the AGC, and seeing if it was still runnable,” Shirriff explains. This is when Verdiell began filming the restoration and all its steps in great detail, now posted on his popular YouTube channel. Come June, the team would meet again at Verdiell’s laboratory in Silicon Valley. They worked 10 hour days for over two weeks, in a taxing and expensive process. The reward was seen when the AGC ran a self-diagnostic and brought up its fixed memory. At that moment, it was essentially ready to run programs again.
Hall and his wife, Grace, are brought over to the computer. Mike has it up, running, and ready to fly to the Moon. Eldon sets things in motion, and for the first time in decades, he punches commands into the DKSY. It’s a special moment. Even 50 years later, Hall’s memory of the computer is sharp, asking precise questions about the restoration. Then, the true excitement begins. The AGC is about to land on the Moon.
At least, that’s what it will think. In a feat of continued excellence, Mike Stewart configured the faithfully accurate Orbiter spaceflight simulator to interface with the AGC, instead of running its own emulation of the computer. Stewart, with a joystick in hand, will land a virtual Lunar Module on the Moon, using an actual AGC running the actual programs as it would in flight. With a handful of exceptions, no one alive is more knowledgeable on this computer than Stewart, and he has even learned to fly a virtual Lunar Module for what is about to happen for the first time since December 1972, when Gene Cernan and Harrison Schmitt landed in the Taurus-Littrow valley.
The lights are cut, and the dozen or so people present find ourselves inside a LM, sounded with the quiet hum of the AGC’s cooling fan. The Moon’s dark, forbidding surface passes outside the triangular window. Mike enters commands into the computer, and after a short calculation, the AGC fires up the descent engine to 10% for 26 seconds, called “ullage,” to allow the LM’s propellents to settle in their tanks. At the end of the countdown, thrust goes to 100%, and the LM’s velocity crawls downwards, lowering in altitude for a touchdown. Mike’s hands are off the stick. The AGC is flying once again.
Stewart narrates the descent and the AGC’s actions, and Carl Claunch translates into lay-terminology. Midway through the descent, the landing radar comes online, and Mike instructs the AGC to accept the data. The AGC now updates its perceived trajectory to the utmost accuracy possible. In this simulation, we are coming in a bit low. After no time at all, the AGC corrects this, and Stewart punches in a VERB 16 NOUN 68. This was the moment, in Apollo 11, the infamous 1202 alarm flashed on the DSKY, mission control in Houston, and MIT in Boston. Today, however, the rendezvous radar is turned off, and the descent continues unimpeded.
The AGC continues the descent, soon transferring into P64. “Pitchover!” The Lunar Module pitches forward, giving a downward view out the window, and the AGC reads a two digit number. Using the LPD in the LM’s window, Mike now knows where the AGC intends to land. Moments later, I catch the shadow in the window, saying, just as Aldrin did, “I have the shadow.” It feels as if the LM is riding down a highway, steady and controlled. Mike decides to designate a new landing spot a little farther forward. Then, he calls out “program alarm!” A 112 alarm has the team wondering what is happening. Later reference shows the AGC correctly sounded an alarm because it did not have alignment from the AOT (Apollo Optical Telescope), which we obviously do not have in this simulation! However, the AGC continues to prioritize landing, engaging P65, the autoland program. At this stage, as a matter of control, Apollo astronauts would enter P66, the attitude hold program, allowing them to control the LM’s rates and orientation via the AGC.
Dust begins to scatter. The altitude on the DSKY, sitting in Eldon Hall’s living room, shows just a hundred feet left. Both the spectators and the AGC wait for the contact indicator to light up bright blue as the LM blasts down into the dust cloud. “Contact Light!” The descent engine cuts, and the LM sets down into the regolith. The calls of 50 years ago are said once again here in Naples, Florida:
Mike Stewart: “ACA: out of detent.”
Buzz Aldrin: “Mode control: both auto.”
Mike Stewart: “Descent engine command override: off.”
Buzz Aldrin: “Engine arm: off. 04 13 is in.”
Neil Armstrong: “Tranquility Base, here…”
Mike Stewart: “…the Eagle has landed.”
The room is speechless. This is history. In this place, Apollo is tangible. What happened in the circuitry of that computer is the same as what happened 50 years ago on the Moon. Eldon Hall sits with his hand on his chin, taking in the moment. No one said it better than Jimmie Loocke, the man behind the project, when speaking of today: “A goal in powering up my AGC is to be able to bring it to Eldon, and let him see it working. Actually, let him use it. That’s what we’re doing today.” It’s no coincidence that Eldon Hall’s house is the very first stop on their brief tour.
The next activity is downloading code from a fixed memory module in Hall’s possession. It is one of 6 modules left containing AGC information, of which Don Eyles has another two. Like its home computer, after more than 50 years, this module, a massive, metal, unfloppy disk, works like it is new. The transistors, capacitors, and diodes perform their task. Claunch states to Hall about the AGC, “it’s a beautiful machine, Eldon. Even after all these years, it read 4 volts, just like it was supposed to.”
The afternoon at Eldon’s house ends with chills going down my spine. Here, in this nondescript neighborhood, a significant piece of spaceflight history came back to its developer and made him proud once again. The AGC is placed back into its massive vault-looking case, which requires two business class seats on an airplane: one for the case and one for its guardian. Having just seen this demonstration, I would personally handcuff myself to it anywhere near the ever-curious TSA. Jimmie tells me that on each airplane, the pilots and flight attendants are astounded to learn of its contents. It has two more flights to go—first to New York for a stop at the Cradle of Aviation museum, formerly Grumman’s Lunar Module assembly facility, and after a drive to MIT in Boston, it will fly home to Houston and the safe guardianship of Jimmie Loocke.
The team has yet to find a private donor for their hard work, time, and travels, despite the attention from the Wall Street Journal and private space collectors. Though operational, the AGC will require continued love. It is, at this point in the future (relative to the 1960s), a unique machine that is not replaceable. This group has taken it upon themselves to prevent that loss to History. Jimmie’s ultimate motivation is knowledge for future generations about where this crazy, digital, space-age world came from. Meeting Eldon Hall brought a piece of that to me. For a software engineer, computer designer, and even future space explorers, meeting this AGC can produce something of that same feeling.
Images and videos courtesy of the Author, except where noted otherwise.
Cover image - Jimmie Loocke’s AGC at the home of Eldon Hall
Boesing, Danny. "Samtec Assists in Restoring Apollo 11 Guidance Computer." News and Tribune. July 19, 2019. https://www.newsandtribune.com/news/samtec-assists-in-restoring-apollo-guidance-computer/article_4ee73c74-a9a2-11e9-b333-ffb26d720028.html.
Eyles, Don. “Tales from the Lunar Module Guidance Computer.” Paper presented at Guidance and Control Conference, AAS, Breckenridge, Colorado, February 6, 2004. https://www.doneyles.com/LM/Tales.html?fbclid=IwAR0HZIxwFRyPPQFLLH2wtoCam02IoGf7GxdD1GVEqooosMiMDFCN0crlIow.
Hall, Eldon (Lead designer, Apollo Guidance Computer). Interview with the author. Naples, FL. July 16, 2019.
Hall, Eldon. Journey to the Moon: The History of the Apollo Guidance Computer. Reston, VA: American Institute of Aeronautics and Astronautics, 1996.
Hotz, Robert Lee. "An Apollo Spacecraft Computer Is Brought Back to Life." The Wall Street Journal. July 14, 2019. https://www.wsj.com/articles/an-apollo-spacecraft-computer-is-brought-back-to-life-11563152761.
Jones, Eric M. "The First Lunar Landing." Apollo 11 Lunar Surface Journal, 1995. https://www.hq.nasa.gov/alsj/a11/a11.landing.html.
Loocke, Jimmie (Apollo Guidance Computer private owner). Conversation with the author. Naples, FL. July 16, 2019.
Loocke, Jimmie (Apollo Guidance Computer private owner). Telephone interview with the author. August 8, 2018.
Scott, David (former NASA astronaut). Conversation with the author. Spacefest IX, Tucson, AZ. July 7, 2018.
Shirriff, Ken (software engineer). Interview with the author. Naples, FL. July 16, 2019.