A receiving bench in an avionics workshop in California, on a pale grey table under a magnifier lamp a clear plastic box lying open on a bed of black grainy foam, inside it five small cream-white plastic carriers two centimetres long with notches and square cut-outs, in each carrier a tiny grey square package with seven thin gold ribbons running out to each side in parallel grooves, the lids of four packages open to show a small chip ringed by very fine bond wires, tweezers in a hand in a white shirt sleeve lifting a sixth loose package by its ribbons, its plain grey underside toward the lens, cool fluorescent light.
What it was used for
Integrated diode-transistor logic was, in the Computer History Museum's account, the first high-volume product line of the IC industry. The flat package was meant for soldering or welding onto flat boards, not for plugging in: Texas Instruments' 1967/68 catalogue specifies gold-plated leads that need no further cleaning in soldered or welded assemblies, with 14 leads and a standard lead length of 0.175 inch. A NASA screening programme (contract NAS 5-9639, Grumman for Goddard) tested Signetics SE-DTL and Fairchild DTµL parts in a 10-month programme (report April 1967), 5 pieces per type; TO-5 cans made up only 5 % of the test lot, the remainder were flat packages.
What was new
Against resistor-transistor logic, DTL lifted the constraint on the number of inputs but left the speed problem in place; that is how the Semiconductor History Museum of Japan puts it. The predecessor was not integrated: in the IBM 1409 of 1959 the stage consisted of a discrete diode and a discrete transistor. Two makers raise differently cut claims to priority - Signetics, for its SE 100 series, the first implementation of DTL as an integrated circuit, announced at the IEEE Show in March 1962; Fairchild, for DTµL, the first DTL circuit expressly designed for integrated circuit technology. In time Signetics came first; according to the Semiconductor History Museum of Japan, Fairchild brought out the 930 series only afterwards. According to the Computer History Museum it overtook the SE100 in 1964 with better noise immunity and lower cost, and a Fairchild sales man recalls in the museum's oral history panel that Chuck Harwood later told him the deep price cut almost sank Signetics.
What the sources leave open: No civilian buyer of DTL flat packs in 1964 is documented; every source found names the military, spaceflight and computer building. Unit numbers and unit prices are missing. The dates for the Signetics SE 100 series rest only on a company chronology written by a participant; no data sheet for that series could be found. The numerical values therefore come from Texas Instruments and Fairchild.