1964 · Transistor-Transistor Logic · Work 852 of 1,871
A small logic card on a grey steel test bench in a computer works in Massachusetts, an almost square circuit card of reddish-brown laminate seven centimetres across with fine copper tracks, on it five tiny flat rectangular packages of plain smooth unmarked gold, three in the upper row and two below, each with thin flat ribbon leads to left and right, along the top edge a plain black connector bar, along the bottom edge a row of gilded contact fingers, a hand in a white shirt sleeve sliding the card into a grey metal card cage of further cards, an oscilloscope with a dark round screen soft behind, fluorescent light from above.
What it was used for
The first practical application of TTL was the computer of a missile: the so-called Phoenix gate for the Guidance & Control Division of Litton Industries in Woodland Hills, working as a subcontractor to Hughes Aircraft on the Phoenix missile. In January 1963 Litton paid 300 dollars per device. Thomas A. Longo, who designed it at Sylvania, says in his Computer History Museum interview that most purchasers at the time were military and that the circuit got designed into all the military systems. Sylvania's data book of May 1966 then courts the computer industry; in 1972 Longo names the GE-600 series at Honeywell Information Systems as the only US mainframe into which standard SUHL was designed.
What was new
What TTL could do and DTL could not is stated in one sentence in Sylvania's data book: the inputs provide the logic capabilities and features of conventional diode-transistor logic, but offer considerably higher speed. Longo recalls building it without resistors and without diodes, with 7 transistors per gate; the data book, however, names a series resistor in the pull-up network of every circuit. The predecessor was discrete: in 1961 Ruegg and Beeson described multiple emitters tied together using individual components in Fairchild Technical Paper TP-21. At the same time James L. Buie was working at TRW Semiconductors: his patent 3,283,170 was filed on 8 September 1961 and granted on 1 November 1966, with 28 claims. According to the Computer History Museum, Buie patented the circuit and Longo, unaware of it, led the design of the first TTL family; Longo's account is a competitor's. The barb: Longo released the circuit diagram only against the 300-dollar order, Litton then put the circuit out for bids, 7 bids came in and 4 awards went out - after which Fairchild, TI and Transitron had it too.
What the sources leave open: No documented market-introduction date for TI's SN5400 was found; the year 1964 rests on the Computer History Museum's account and on Longo's recollection. The museum calls Buie a man of Pacific Semiconductor, while the patent names TRW Semiconductors as assignee; the relation between the two names could not be resolved. Almost all of the narrative about SUHL comes from Longo himself.