1965 · Monolithic Operational Amplifier · Work 856 of 1,871
An analogue circuit on a laboratory bench in California, a small board of brown pressboard with rows of brass solder eyelets, in its middle a tiny round metal can nine millimetres across of plain smooth unmarked polished silver with a flat top and a rolled rim, below it a white insulating ring and eight long thin gold wire legs spreading out to the eyelets, around it small resistors with coloured bands and blue disc capacitors, a hand in a rolled-up white shirt sleeve turning a tiny trimmer with a small screwdriver, a soldering iron in its wire stand, a coil of hook-up wire, lamp light from the left.
What it was used for
At the introduction of the µA709 Fairchild states its purpose: a general purpose amplifier in d-c servo systems, in high-impedance analog computers and in low-level instrumentation, plus the generation of special linear and nonlinear functions. A 1967 advertisement promotes the commercial version as an off-the-shelf part for industrial applications. The predecessor µA702 was too expensive for that: Thomas H. Lee writes that its initial price of approximately 150 to 300 dollars limited potential sales to military and aerospace customers. In December 1965 the 709 cost 75 dollars in quantities of 1 to 24 and 50 dollars for 100 to 999; in July 1967 Fairchild advertised less than 5 dollars in quantities of 10,000.
What was new
Against the µA702, Fairchild's published figures gave the 709 an open-loop gain of typically 45,000 where, according to Thomas H. Lee, the 702's typical value exceeded 3,000; Lee adds roughly 1 MHz of bandwidth and the plus/minus 15 V supplies that had just become standard for discrete solid-state amplifiers. It was not the first: in 1963 H. C. Lin at Westinghouse built a custom operational amplifier for Minuteman II, and the Computer History Museum calls the 702 of 1964 the first widely used commercial product of its kind; module amplifiers came from Nexus Research Labs in 1962, then from Burr-Brown and Philbrick. The barb: 3 gain stages had to be frequency-compensated externally, which is why Widlar brought every high-impedance node out to a pin - Lee adds drily that sometimes the user even succeeded. Internal compensation arrived only with the µA741 of 1968.
What the sources leave open: The gain figures diverge: Fairchild's 1967 advertisement gives a typical 45,000 at 25 °C and ±15 V, Thomas H. Lee about 60,000 without conditions; the difference remains open. The introduction in November 1965 rests on Lee alone; the earliest contemporary record is 27 December 1965. Numbers produced or sold of the 709 are not documented.