Space Travel and Exploration

views updated

Space Travel and Exploration

Many people believe that the rapid increase in the capability of computers was largely due to the demands of the American Space Program. It is indeed true that the National Aeronautics and Space Administration (NASA) did pressure the computer industry for more speed and more memory in its ground based systems, but the computers in the Apollo spacecraft that traveled to the Moon and those flying in the later space shuttles are relics of a by-gone era. This conservative attitude is the result of NASA's unwillingness to risk human lives: old technologies have a known track record while there is implicit risk in an untested technology. However, even the unpiloted spacecraft use mostly obsolete computers, though NASA has allowed chances to be taken with their memory technologies. Nevertheless, NASA has been a good customer of the computing industry, demanding new techniques even while using older hardware.

Early Piloted Spacecraft Computers

When astronaut John Glenn rode his Mercury-Atlas rocket into orbit and into history in February 1962, not a single computer rode with him. The Mercury spacecraft was a ballistic object, like a warhead, just going fast enough for a few orbits of the Earth. It could not change its orbit, nor could it maneuver in any way except to fire its retro-rockets to return to Earth. In essence, it did not need a computer.

The Apollo Program was a piloted American space flight effort designed to fulfill U.S. President John F. Kennedy's 1961 promise to go to the Moon and back by the end of that decade. To help develop technologies for the piloted Apollo spacecraft , there were several programs involving unpiloted lunar spacecraft. One of them, the Ranger series of lunar impact probes, was ground-commanded. The probes were only designed to hit the Moon, taking pictures all the way down. The command to start the camera going was the last command that spacecraft received.

The Apollo spacecraft , designed to make a soft landing, was an entirely different matter. The soft landing could not be commanded from the Earth, owing to the fact that radio waves take about three seconds to make a round trip. This is why an on-board computer was necessary. The later decision to use a two-part spacecraft, one part landing and one part staying in orbit around the Moon, and the realization that making a re-entry to Earth at 40,230 kilometers (25,000 miles) per hour would be very difficult with a tired crew, sealed the decision. Several years before the Apollo spacecraft flew, the interim Gemini Program, using a Mercury spacecraft widened to accommodate a crew of two, had a computer put on board to practice rendezvous and computer-controlled descents that would later be used on the Moon mission.

This computer, designed and built by IBM, was a small, custom-made machine with only 4,000 words of memory. More storage was provided by a magnetic tape similar to the ones especially made for the Orbiting Astronomical Observatory flight series. This computer hardly stretched the state of the art, but performed flawlessly.

Nevertheless, if there were any problems, a Gemini spacecraft could land in minutes. An Apollo spacecraft could take days to get back. NASA was worried about reliability. The easiest way to assure reliability was to carry multiple copies of the same computer, so there is always a spare if one computer fails. However, as in all of aeronautics and astronautics, the biggest considerations are size, power, and weight. A redundant system would negatively affect all three, adding the problems of redundancy management. NASA therefore decided to use only one full-sized computer in each part of the spacecraft. One part would always be in touch with the Earth (except when its orbit passed the far side of the Moon). So, big IBM-built ground computers could back it up. The lander, since it could not be commanded from the ground, would carry a tiny computer as a back up that could only put the lander in a parking orbit. The command module would then have to rendezvous and save the crew.

NASA got away with using one computer by extensive testing and knowledge of the history of the various parts. Early in the program, when the requirements for lunar rendezvous and other maneuvers were levied, Massachusetts Institute of Technology's Instrumentation Lab, the designers, asked NASA if the lab could make the machine more capable by using the first mass-produced integrated circuit in its construction. This was a four-logic-gate chip. Initially, NASA balked because this represented new technology, but it relented after assurances by MIT that the actual use of this technology was still years away. There were no in-flight failures of the Apollo computer.

The Space Shuttle

The huge size of the orbiter part of the space transportation system suggested the idea that considerable weight could be saved if digital flight controls were used instead of mechanical linkages. Since this system would be used for active flight control, more than one computer was needed. The failure of the one computer would cause the loss of an expensive spacecraft and its crew. Two computers were not enough either, as it would be difficult to tell, in case of malfunction, which one of the two had failed. Both machines would be indicating that the other one was at fault! Similarly, three machines would devolve to that standoff if one failed. So four computers became the minimum, with a fifth machine added and a sixth carried in a locker as a spare unconnected to anything else.

In the 1970s, after the shuttle design was frozen, a capable airborne computer called the IBM AP-101 was procured for it. This machine is also used in the B-52, B-1, and FB-111 bombers as a navigation and weapons delivery computer. The AP-101 originally had 106,000 words of memory and ran at one million instructions per second, much less capability than a typical desktop computer today.

After the loss of the orbiter Challenger, the memory was increased to 256K of CMOS (complimentary metal-oxide semiconductor), and the processor was fit on one chip. By continuing to use the same processor, NASA avoided a prohibitively expensive rewrite of the software. These changes were being developed before the disaster, but were delayed by safety considerations.

Unpiloted Spacecraft

The use of CMOS memories, despite their low power consumption, was a breakthrough in piloted spacecraft, which were limited to the fifty-year old technology of core memories until late in the 1990s. Meantime, the Jet Propulsion Laboratory (JPL), operated by the California Institute of Technology and funded by NASA, had been using CMOS memories since the Voyager interplanetary probe, launched in 1977.

Traditionally, NASA had always been more comfortable using innovative computer technology in unpiloted probes rather than in piloted spacecraft. Unpiloted spacecraft operated computers mostly in pairs. The first time real general-purpose computers were used in a probe was the Viking lander on Mars in 1976. It contained two command computers and two lander computers. Voyager had three pairs of computers: command, attitude, and data formatting. The command computers, with 4,000 words of memory, were a design similar to those on Viking. The attitude control computers, also with 4,000 words, were a little faster, but incomparably slower than computers today. The data formatting computers, which take experiment and camera data and prepare it for transmission to Earth, were designed by a JPL engineer as part of his graduate work in electrical engineering.

Beginning with the Galileo Jupiter probe, JPL has gone to distributed systems of processors and larger memories; however, they are usually obsolete by launch time. It seems that NASA's motto for computer applications is "Good Enough Is Good Enough."

see also National Aeronautics and Space Administration (NASA).

James E. Tomayko