by R. Passov
Eldon Hall spent the first seven years of his life climbing hills alongside Oregon’s Snake River, trailed by a faithful Shepard dog. He and his father “…went fishing in the mountains…” and “… slept outdoors while his mother, safely residing at home, worried about the poisonous snakes that might bite [them.]” In 1926, when Eldon was seven, his father passed. If not for that, Eldon would have carried on farming alongside the banks of the Snake.
Unable to hold onto the farm, Eldon’s mother took her three children across the river to Paytte, Idaho. There she married a subsistence farmer. While Eldon was tempered working as a farmhand, she held his dream of getting an education.
The day before Pearl Harbor, Eldon defied the odds by enrolling at the University of Washington. The war dried up funding. On the verge of dropping out, he joined the ROTC. In 1943 he was called to active duty. The few college credits he brought along gained entry into a newly formed “Army Specialized Training Program” that led to City College, NY where his days were “…filled with lectures, testing, military instruction, calisthenics, and some free time to tour the Big Apple.”
After 18 months he was sent across another river to Rutgers University to begin a “training program” in electrical engineering. Of the two hundred or so men who started the program, 65 finished. The top four graduates joined the Manhattan Project. Eldon graduated in 5th position.
After Rutgers, Eldon was sent to the Army Signal Corp for more studying and testing in what turned out to be another shot at the Manhattan Project. Once again, he missed the cut. After the war, the University of Washington had no use for his army credentials. Not wanting to start over, he managed to convince Eastern Nazarene College in Quincy, MA to credit his Rutgers course work.
By the end of two years he had a bachelor’s degree in mathematics, along with a master’s in physics from Boston University. There he met Grace, then on her way to a PhD in physics. In 1948 they began a marriage that would last sixty-eight years.
Armed with his new degrees, Eldon entered the graduate physics program at Harvard. He did well enough in his studies but his family had grown to include two children and his wife was closer to her PhD. He found himself thinking about work and happened to notice a flyer advertising for someone with experience in “random controls” to join the Instrumentations Lab at MIT.
* * *
…the preliminary evidence presented makes clear that the IBM computer is neither suitable nor capable of performing the Apollo mission as it has been defined. It is indeed unfortunate that the portion of the BELLCOMM staff which has been an enthusiastic supporter of the IBM computer did not discern these obvious deficiencies in the many months of their association with the IBM and MIT staffs.”
Critique of IBM Apollo Study Report
If you’re going to take on AT&T (BELLCOMM) and IBM in the same memo you’d better have the backing of your boss.
In 1932, Charles Stark Draper founded the MIT Instrumentations Laboratory. He was an Elon Musk type, except he remained an academic. Born in 1901, at 16 he entered university in Missouri, his birth place. Two years later he would earn a degree in psychology from Stanford. After Stanford, he drove east with the intention of continuing his studies at Harvard. Instead, by some accounts he still holds the record for the greatest number of graduate level courses taken at MIT.
Somewhere in the late 1920’s he watched James Doolittle, who would lead the first Allied bombing run over Tokyo, circle the MIT campus. Draper got hooked on flying and in 1932, convinced MIT to fund a laboratory focused on developing flight controls. In 1938, tired of watching Draper enroll in class after class, MIT awarded him a doctorate.
During WWII, Draper’s lab built the Mark 14 gunsight that, through an ingenious joining of a gyroscope, optics and a spring, caused the gunner to fire not directly at the target but ahead, into the direction of motion. The Mark 14 “… saved countless thousands of American lives.” After the war, though he continued the Lab’s focus on gunsights, Draper returned to his first love – flight.
In early February of 1953, The University of California at Los Angeles hosted a “…top-secret conference on inertial guidance.” The conferees were scheduled to debate the feasibility of flying a plane without the aid of a human pilot. Unbeknownst to the conferees, several years prior, Draper had succeeded in talking the Air Force into funding a different top-secret project.
On the morning of the conference, Draper landed at the Los Angeles airport in an B-29 Superfortress, guided along it’s 2250 nautical mile journey not by a pilot but instead by Space Inertial Reference Equipment (SPIRE) – 2,700 pounds of gyroscopes and analogue computers. The conference ended early.
Draper dedicated a carefully selected group of tinkers and scientists to working on navigation and guidance systems. When he listened to a Navy Admiral describing the need for a system that could guide a missile launched from a submarine to the upper atmosphere and back down to its target, without the need for telecommunications, he was ready.
He persuaded the Navy to fund the design of an “inertial guidance system” that could send anything anywhere, and only require setting the “initial values of a state vector” to – 0, 0 and 0. The Polaris system would do the rest.
Then, he persuaded Eldon to build the Polaris.
* * *
Discrete transistor components – first-generation integrated circuits – gave Eldon a shot at getting the workings of a gyroscope and an accelerometer into a small enough bundle to fit within a missile designed to launch from a submerged submarine. But to get the system to work, Eldon needed something that could translate readings from the real, continuous world into the language of machines. So, he conceived and built one of the world’s first analogue-to-digital converters.
With the success of the Polaris system in hand, in 1961 Dr. Draper lobbed his old protege, Robert Seamans, then head of technical support for NASA. In a letter to Seamans, Draper went beyond offering the services of his laboratory. He offered to “…formally volunteer … as a crew member on the Apollo mission to the moon…” He argued that his “ …flying experience, although non-professional has extended over some thirty-five years with much experimental work included…” And then added, “I realize that my 60 years is a negative factor in considering my request …but a [high-ranking General] tells me that this is no sure bar to my selection as a crew member.”
Draper saved his best shot for last: “If I am willing to hang my life on our equipment, the whole project will surely have the strongest possible motivation…”
The MIT Instrumentations Lab was awarded the contract to design the guidance system for the Apollo Space Crafts
* * *
A 1972 Bell Telephone Magazine article detailing BELLCOMM’s role says it existed to look over NASA’s shoulder – point out what they hadn’t thought of and make suggestions for doing the job better.”
One Giant Leap for Mankind. Soren Telfer, AT&T
…I won’t go into the details of that backup, but it got kind of messy at times, from my point of view.”
NASA administrators knew they needed a lot of help in getting to the moon. As well, they were keenly aware of what awaited them as the world’s largest procurement operation got underway. The line at the trough would be filled with strong-willed influencers.
The Saturn Booster – under development to launch military satellites – had already been chosen as the mission launch vehicle. IBM built the computer that would guide the Saturn. But the booster had only to arrive at an orbit around the moon, not find a particular spot.
AT&T/BELLCOMM managed the communications links between the three satellite tracking stations, equidistant around the Earth, that would follow the Moon missions. Somewhere along the way, AT&T and IBM, miffed at not getting the shot to design the device that would guide the Apollo crafts, decided to combine forces. Together, they attempted to convince NASA administrators of the risk of relying on Eldon’s guidance computer and of the need for a “backup.”
* * *
My education had not even mentioned digital computers. Switching circuits, binary arithmetic, and Boolean algebra were foreign to the vocabulary of physicists.
On October 1st, 1963, after having watched AT&T and IBM, over the course of many months, collude in an attempt to push his computer aside, Eldon made his case.
In his memo, Eldon acknowledges that the IBM computer has twice the memory as his Apollo Guidance Computer(AGC); then he explains to IBM how their computer actually works. The “two banks” of memory within the IBM computer, at first glance, appear to offer more than what is available within the AGC. However, Eldon notes, one bank is entirely redundant: it exists only to serve as a backup so the computer can meet the reliability requirements of the mission.
Then he slices further: He explains how the IBM memory bank is composed of three distinct “Blocks” of 256 bytes. A program in execution phase can only address one Block; to go beyond requires a special HOP instruction whereby a place in memory holds the instruction location while the computer looks for additional instructions or data in a memory location outside of the current Block. It’s as though all cars in both directions of a twelve lane highway must get off at the same intersection at almost the same time.
After correcting the simple math, Eldon shows the AGC to have up to twice the useful memory of the IBM competitor. He then shows how calculations such as multiplication or addition require many extra steps, such that upon proper reflection the expected compute time for mission-critial operations on the IBM computer would be many orders of magnitude greater than for the AGC.
But it gets worse: Since so little of the AGCs memory is addressable, it’s unfeasible to use “subroutines” – snippets of embedded code that are likely to be reused over and over. Code for input and output routines but as well for important vector mechanics which are critical for “explicit” guidance that finds its own way and has the ability to course correct.
Because of the nature of the Apollo Guidance problem, subroutine linkage is the only practical method of conserving storage capacity…On the other hand, subroutine computation is, perhaps, not needed for the mechanization of the so-called Saturn “path adaptive” guidance mode for which the IBM computer was designed…”
Unbeknownst to the IBM engineers, the “path adaptive” guidance math for the Saturn Rockets could not satisfy the needs of the Apollo Missions.
Subroutines would be necessary to execute the math within a machine that would cause the machine to know where it was – to know its coordinates in three dimensions, plus time – and to self-correct along the way to a target requiring an accuracy within “40 seconds of an arc.” The limitations of the IBM computer were such that it could only guarantee finding a target within “a 2.6 minute arc” – or, as measured on the Moon’s surface, within a quarter mile.
* * *
The one last shot by the BELLCOMM / IBM folks was at the construction of the computer. The Discrete Transistors in the Polaris were simple devices, made of germanium arsenide. Strung together in the proper order, connected by wires and solder, they constituted a logic gate through which electrons could flow or not, creating the instance of a ‘1’ or ‘0’ which, along with time, is all that is necessary to compute anything that can be computed.
While Dr. Draper wanted inventors and tinkers in his Instruments Lab, he left manufacturing to a network of collaborators. One consequence of this network was an introduction to Jack Kilby. Working at Texas Instruments, Kilby invented a process whereby the discrete components comprising a circuit could be combined onto a single piece of Silicon. But Kilby’s invention required wired interconnections to be completed outside of the process of manufacturing the Integrated Circuit (IC).
At the same time that Kilby invented his version of an IC, Gordon Moore and Robert Noyce, who together would go on to found Intel, one-upped Kilby. They developed the “planar” technology that would allow for a “monolithic Integrated Circuit” to be built such that all of the components of the logic gates, along with the various inter-connections, would be manufactured in the same process: One small piece of Silicon, containing tens, then hundreds and now, billions of logic gates.
After flying to Texas to meet with Kilby, Eldon went to California. There he met with Noyce who convinced him that the untested Integrated Circuit constructed with the newly invented planar technology would prove capable of allowing a “compact, 70 lb. computer” to guide the Apollo missions to the moon and back.
Against the combined advice of IBM and BELLCOMM, Hall built the AGC around an invention then known to less than twenty people. But just in case, Eldon created a modular design. Had an IC for some unknown reason failed, the astronauts had a stash of spares at the ready.
In 1973, hounded by student protests over its involvement with military technology, the Instruments Lab separated from the MIT campus. Renamed the Charles Stark Draper Lab, operating as a not-for profit, though surpassed by others in the race into space, it continues on.
After the Apollo missions, Eldon’s worked drifted toward studies and reviews with perhaps one notable exception: In 1976 he hooked an electric typewriter to a screen and memory, creating one of the first word processors. In 1987, he returned to farming.
After many years teaching physics, Grace returned to college, earning an MA in literature. In 1999, she published The Tempest As a Mystery Play: Uncovering Religious Sources of Shakespeare’s Most Spiritual Work.
* * *
- Critique of IBM Apollo Study Report – IBM # 63-928-129 – 1 Oct 1963. Eldon C. Hall1
- Journey To The Moon: The History of the Apollo Guidance Computer. Eldon C. Hall. American Institute of Aeronautics and Astronautics, Inc. 1801 Alexander Bell Dr., Reston, Va. 1996
- From the Farm to Pioneering with Digital Control Computers: An Autobiography. Eldon C. Hall. IEEE Annals of the History of Computing, April – June 2000
- Draper at 25: Innovation for the 21st C. Morgan with J. O’Conor and D. Hoag. Charles Stark Draper History
- Monolithic Concept and the Inventions of Integrated Circuits by Kilby and Noyce. Invited Paper, Nano Science and Technology Institute Annual Conference, Santa Clara, Ca. May 20 -24, 2007
- One Giant Leap for Mankind – and Still Reaching. Soren Telfer, Dr. AT&T Foundry Palo Alto, Ca. July 19th, 2019