In the first quarter of 1958, Lockheed System Engineering conceived and proposed to the US Air Force a concept for control of the entire Weapon System 117L (WS-117L) satellite network. The Air Force approved the concept and ordered its rapid implementation in April 1958. The task would take all of two years in the doing and would result in the Satellite Test Centerin the Lockheed plant complex in Suunyvale. However, the first Discoverer launch was scheduled for December 28, 1958 -- a scant seven months away. (The first scheduled launch was eventually slipped until January 21 1959*.) Therefore, an Interim Development Control Center (IDCC) was required for the early operations until the STC could be completed.
Fortunately, Lockheed had a large computer room housing two Univac 1103A computers that supported both administrative and research computational needs of the company. These were located in the headquarters building in the Stanford Industrial Park on Page Mill Road in Palo Alto, California. The only real hope of supporting the early launches was to use the 1103A computers.The 1103A machines were quickly upgraded to the 1103AF version by the addition of the newly developed Univac floating point package. Five paper-tape receiver/transmitter terminals and two teletype (TTY) units were installed for direct communications to the tracking stations.
A three-man team designed, developed, integrated, tested and operated this hardware/software system in support of the first Discoverer flights. The newly developed software was first demonstrated in October 1958 and was first used operationally in January 1959. It had the ability to do ephemeris generation, fitting to tracking data for orbital determination, impact point prediction and command generation.
By today's standards, the hardware configuration of the IDCC seems very meager. Today's personal computers (PC) possess a far greater power and capability than those old 1103AF machines. But in 1958, the Univac 1103AF was among the most powerful of this final generation of vacuum-tube computers. It was not until 1959 that the first transistorized computers appeared on the scene -- and these had numerous reliability problems. The first three large-scale transistorized computers were the IBM 7090, the Philco S-2000, and the Control Data Corporation's CDC-1604. For many reasons the CDC 1604 was selected for use in the STC -- but that's another story.
The 1103AF computers originally used in the IDCC used a 36-bit word. Each computer had 8,000 words of magnetic core storage and 32,000 words of drum memory (in PC equivalents: 36 kilobytes of random access memory (RAM) and a 144 kilobyte floppy disk -- about half as much as a full blown Radio Shack TRS-80 of the late 1970's!). In addition, each had eight high speed magnetic tape transports, a typewriter output, a very flexible central control console and paper-tape input/output (I/O). Printing was done off-line on two systems, each having a magnetic tape transport and a line printer.
Computer operations for the IDCC was via the five paper-tape terminals and the two TTY units already mentioned. The paper-tape terminals were used to send pre-pass tracking data to the designated Remote Tracking Station (RTS) for satellite acquisition and lock-on. Once locked on to the satellite transmitter, the RTS antenna's auto tracking system generated the actual measured tracking data for transmission back to the IDCC via the paper-tape terminals after the satellite pass. Once the tape was received, it was read into the 1103AF for orbital fitting and ephemeris generation.
Satellite commands were transmitted to the appropriate RTS via the TTY terminals prior to the satellite pass. Commands were manually entered into the RTS command console and transmitted to the satellite at the command of the IDCC Operations Controller -- usually by voice line, though sometimes it was necessary to employ the TTY. Verification of the actual command load achieved was sent back to the IDCC by voice, in real-time and post-pass via TTY for manual entry into the 1103AF through the controlconsole.
The only on-line rapid access memory was the 8,000 words of core memory and the 32,000 words of drum memory. This provided a total of 40,000 36-bit words, or 1.44 Mega Bits (not Bytes!) Some typical execution times for Floating Point (FP) operations are shown in Table I and are compared to a not-so-state-of-the-art 80386 PC running at 32 M Hz (all times in micro seconds).
When compared to today's PC's, we can see we are orders of magnitude in both memory capacity and performance over the early Univac 1103AF. Even a modest PC with an 80386 processor chip running at 32 M Hz will likely have 16 mega bytes (128 mega bits) of RAM and a 540 mega byte hard disk drive.
The technology has come a long way in increasing computer speeds and memory capacity with dramatic reduction in physical size and cost -- not considering the cost of the facilities needed to house, power and cool the older machines. It only serves to emphasize the remarkable achievements of the early engineers and software developers who's pioneering work and vision led to the creation of the many satellite control centers we know today.
* While the first launch was scheduled in January 1959, this launch was unsuccessful. Prior to Thor engine ignition, the ascent sequence timer was inadvertently started which caused the Thor-Agena separation sequence to start on the pad. The countdown was halted and a new Thor/Agena was readied for launch in February 1959.
Frank Druding, like many other young men of that era began his space associated work with Lockheed Missile and Space Company in Sunnyvale, California in 1957. He continued to work at LMSC until 1961 when he left to form Mellonics, a corporation in the forefront of software development for the Air Force's space effort. In 1970 Mellonics was acquired by the company now know as Loral and Frank continued his space activity there as Vice President. Currently Frank is retired and living in the quaint seaside village of Mendocino in Northern California.