Who Invented the Computer? TRADIC

Who Invented the Computer? This is the thirteenth installment in our ongoing series.

The invention of the transistor in 1947 heralded the second generation of computers. In terms of power consumption and cooling, this tiny marvel was far superior to its first-generation predecessor, the vacuum tube. Whereas tubes required one watt or more of electricity to operate, transistors ran off a millionth of a watt.

Transistors also generated much less heat than tubes and could be placed closer together without overheating and burning out. Throughout the next decade, transistors would be constantly improved and, in the process, enable the development of computers that were smaller in size, more reliable, and much faster than earlier models.

The one downside to transistors was cost — $20 per unit compared to $1 for a vacuum tube. And while transistors were clearly the future of the industry, vacuum tubes continued to play an important role in a number of impressive computing accomplishments like the 1948 Manchester Small Scale Experimental Machine.

The Manchester was the first computer to electronically store and execute a program. Vacuum tubes were also a key element of the Whirlwind I in 1949, the first computer to use magnetic-core memory.

TRADIC

Much of the history of computers was a Cold War-esque contest between the United States and Great Britain. Although Americans would eventually dominate the industry, the Brits did leave their mark: In November 1953, the University of Manchester unveiled a prototype that was the world's first transistorized computer.

The computer utilized point-contact transistors consisting of a piece of germanium affixed to two wires in close proximity to one another — the "points." Configured of 92 transistors and 550 diodes, the Manchester apparatus was capable of computing 48-bit digits. There was however some disagreement among computer jocks on both sides of the Pond as to whether or not the Manchester device was "truly" transistorized, as vacuum tubes were used to power its inner clock.

Meanwhile, in the United States, Jean H. Felker, an electrical engineer at Bell Labs, used a new version of the point-contact transistor to build a regenerative amplifier able to operate a high-speed closed circuit in a solid-state digital computer using low power. Acknowledging Manchester's accomplishment, and excited with his amplifier, Felker figured that whatever the "mother country" could do, the colonies could do better.

At the time, the fledgling U.S. Air Force was looking for an improvement on their analog control units used for real-time navigation and bombing and wondered about the feasibility of using transistors to build a digital computer that would meet their needs.

Felker visited the Pentagon and pitched his idea for a high-speed, flexible digital computer and promised that Bell Labs would deliver a computer that would be "as reliable as a hammer." Felker's idea was given the green light, and by early 1954 he had constructed America's first transistorized computer: TRADIC (short for TRansistorized Airborne DIgital Computer.)

Containing 700 point-contact germanium transistors and more than 10,000 diodes, TRADIC could perform one million logical operations a second. It was not as fast as the fastest vacuum tube computer, but it was more reliable and smaller than any tube model in existence. It was also energy efficient, operating on less than 100 watts of power.

Flyable TRADIC

Once the Air Force knew that a solid-state computer could do what they needed, their next question was whether it could be installed it a plane. Felker said that yes, of course it could, and promised to deliver an amazingly compact 1 cubic-foot solid-state computer.

The biggest hurdle to transporting a solid-state computer in a military airplane was that it had to be able operate in an environment with an extreme range of temperatures: from a low of -55� C to a high of 55� C.

Felker's team was successful reconfiguring TRADIC's circuits to handle the temperature variances, but failed to meet the 1 cubic-foot specification. In their rush to build a flyable TRADIC, they did not allot time to work on miniaturization. Their version was considerably bulkier — approximately a dozen cubic feet — but still compact enough to fit in the cargo plane where it underwent successful testing.

Future developments

The TRADIC System would develop two additional and improved models: Leprechaun, a solid-state computer utilizing junction transistors and a stored program control (software) to further test the capabilities of airborne computing.

Leprechaun was designed to run much longer programs than the two earlier TRADIC iterations. As engineers often do, the development group referred to their machine as the "long program computer" and were soon using the acronym "LONPROCON." The acronym sounded similar to "leprechaun" and the name stuck.

The XMH-3 was the last TRADIC version. It was designed to operate as the bombing and navigation control unit for the B-68, an all-weather, high-speed tactical bomber that was still on the drawing board. Unfortunately for the XMH-3 team, budgetary constraints led the Air Force to cancel the project before a prototype was fully developed.

Advances in transistors enabled the engineers at Bell Labs to prove conclusively the versatility and operability of airborne solid-state computers. From here on out, transistors would be continuously  improved and used to further expand the horizons of computing possibilities.