Our second installment in our "To the moon" series is about IBM and UNIVAC, the two primary computer systems integrators for the Apollo missions.
This article was originally written in 2009.
IBM's Real Time Computer Complex in Houston (Top) and NASA's Fresnedillas DSTN station in Madrid, Spain, (Bottom) one of the many Deep Space Tracking Network outposts using UNIVAC systems for pre-processing of telemetry downlink and command uplink data.
Sending men to the Moon was not just about building gigantic Saturn V rockets and shooting them off into space.
While it was the key component of the overall program, and where a large portion of the money for the moon missions were spent, the rocket would have had no nervous system and we would have had no support infrastructure for the Apollo if it wasn't for the efforts of two large systems integration firms -- International Business Machines and UNIVAC -- and the people who worked at those companies during the 1960s.
IBM as Ringmaster
As a system integrator IBM's involvement in the Apollo program was very broad, which ran the gamut from being the vendor and support contractor for ground based mission control computer systems at NASA, as well as having a relationship as a "Primary" contractor in the manufacture of the Saturn V itself. No other vendor, Boeing included, ran the gamut of touching virtually aspect of the Apollo program.
IBM's involvement in the US Space program dates back prior the 1950s, where a number of early IBM computer systems were used for missile trajectory and satellite orbital calculations. In the early 1960's, model 7090 systems were used by NASA to do initial calculations to determine thrust and trajectory characteristics needed to launch the Saturn V, even before the rocket hit the drawing boards.
During the Mercury project, which allowed NASA to prove the initial manned spaceflight technology later used on Gemini and Apollo, IBM systems were used to monitor in real time the telemetry data being received from John Glenn's Friendship 7.
These systems were a precursor to the Real Time Computer Complex (RTCC) which was later built at NASA's Manned Spaceflight Center in Houston, Texas for the Apollo missions, using redundant Model 7094 mainframe systems initially and were later replaced by twin System/360 Model 75 computers, which were the modern forerunners of today's zSeries systems as far as their very basic systems architecture is concerned (The IBM System z10 is still backwards compatible with software written for the S/360, S/370, and the S/390 through machine virtualization and software emulation).
The IBM mainframes in the RTCC, as the acronym implies, monitored the telemetry data from every aspect of an Apollo mission in real time, so there was no lag in processing between problem determination if a condition arose and actually solving it.
The guidance computer was used during pre-launch checkout and was capable of launch simulation, and did all of the booster guidance during the rocket's ascent phase, calculating trajectory 25 times a second and adjusting the control surfaces on the gimballed rocket engines to keep the Saturn V on course to orbit.
The guidance computer was also used to ignite the engine on the S-IVB third stage to insert and guide the spacecraft into Trans Lunar Injection (TLI).
The ring-shaped Saturn Instrument Unit, manufactured under contract by IBM, had 57 components and weighed 4,400 pounds. The 21-foot wide SIU, which was mounted above the S-IVB third stage and sat between the Command Module/Service Module spacecraft was the brains and nerve center of the Saturn V, providing all the major guidance functions as well as all of the sensors and instrumentation packages needed by the rocket.
After Trans-Lunar injection, navigational control was handed off to the Apollo Guidance Computer (AGC) in the Command Module, which was an embedded systems computer designed by MIT's Draper Laboratories. A second AGC was also installed in the Lunar Module.
IBM's legacy in the Space Program lives on in not just their continued involvement in NASA's current and future efforts with their computer systems and support services, but also in some of the software that was built for the Apollo program.
The IMS suite of hierarchical database management applications which is still an important part of IBM's mainframe software portfolio,was originally designed by IBM in conjunction with Rockwell and Caterpillar so that the huge Bill of Materials (BOM) for the Saturn V, which was composed of hundreds of thousands of parts could be inventoried and managed.
UNIVAC, Running Apollo's Communications Network at Goddard and Beyond
While IBM's mainframes were at the core of Mission Control, IBM was not the only computer company with a critical role in Apollo's mission. While the company no longer exists today in its original recognizable form and few people of my generation would even be familiar with the trademark, UNIVAC was a household name for computers during the 1960's, much like Dell, Apple or Hewlett-Packard is today.
UNIVAC, originally started as the Eckert-Mauchly Computer Corporation, was the descendant of the namesake founders work on the ENIAC, the first stored program digital computer which went online in 1947.
In 1952, a UNIVAC I computer made history by successfully predicting the outcome of the presidential election, live on national television with legendary anchorman Walter Cronkite reading the results. This feat brought the company into the limelight, and secured its future over the next two decades as a leader in business computer systems, as well as a producer of computer systems for the military and government.
While IBM's systems were at the core of the Real Time Computing Complex in Houston, more than just that single datacenter was required in order to process and send telemetry data to and from the Apollo spacecraft.
Earlier Mercury and Gemini were both near-Earth-orbit missions and used C-Band communications (among others) for their primary communications. Communications changes were needed for more distant missions.
A network of radio monitoring and transmitter stations, known as the Deep Space Tracking Network, or DSTN, was built in order to provide signal processing of incoming telemetry and outgoing command functions over Unified S-Band radio (USB) and to maintain constant communication with the Apollo space vehicle on its way to and from the Moon.
A number of smaller USB Apollo ground stations had 30-foot antenna for transmitting to and from the spacecraft while it was relatively close to the Earth. For deep space, three large stations which were relatively equidistant from each other had to be built, in Goldstone California, In Fresnedillas, Spain and in Honeysuckle Creek, Australia.
Each of the three DSTN stations were composed of huge 85 foot wide movable radio dishes, radio signal decoding equipment as well as a computer room which contained UNIVAC 1230 systems that processed a continuous 48kbps data stream from the Apollo spacecraft using itson board S-band radio transmitter on the Command and Lunar modules. Each of those spacecraft modules had the ability to demodulate (and remodulate) information for transmission back to Earth.
Each pulse code modulated data stream received and sent by each DSTN station had to be scrubbed, reformatted, pre-processed, and sent back and forth to NASA's Goddard Space Flight Center in Maryland through the NASCOM communications network via commercial UNIVAC 494 and 495 systems before sending it to Mission Control (MCC-H), the Command Communication and Telemetry System (CCATS) and the IBM Real Time Computer Complex (RTCC), which could only process data at a much slower 2Kbps.
Phil Ross (shown left in the blue golf shirt) was Section Head for Telemetry Programs for UNIVAC at the Goddard Flight Center during the Apollo missions. Listen to a Podcast with Jason Perlow and Phil Ross, former UNIVAC Apollo Telemetry Section Head
By 1966, UNIVAC, now a division of Sperry's Remington Rand defense contracting subsidiary, had a great deal of experience installing military hardened versions of the 1230 on naval vessels, which were known as the 642A/642B variants, and were originally used for missile fire control processing.
Compared to the IBM mainframes they were much smaller -- refrigerator sized -- and could also tolerate harsher conditions than their commercial UNIVAC or IBM counterparts, able to operate up to 140 degrees F and were conditioned against saltwater corrosion.
In 1966, NASA quickly brought UNIVAC's programmers and its computers in to run the data processing for telemetry and command data streams for the DSTN and for CCATS. By the time of the Apollo 11 mission in 1969, at least three dozen 642B variants of UNIVAC 1230s were employed at DSTN stations around the globe, including as far away as Australia, Spain and California.
While UNISYS still makes computer systems (focusing on large SMP x86-compatible systems with its ES7000 in a manufacturing partnership with NEC) and still does a great deal of contracting with the United States government and NASA, like IBM, much of its revenue stream is now highly services-focused, concentrating on business transformation and outsourcing. But its role as the brains at the core of NASA's communications network will never be forgotten.
Were you an IBM or UNIVAC employee during the Apollo era? Talk Back and Let Me Know.