The Chip That Jack Built

Eleven years after Bell Labs built the transistor, engineers could design machines with thousands of them. Building those machines was the hard part. Every transistor, resistor and capacitor had to be wired and soldered to the next by hand, and every joint was one more chance to fail. Engineers called it the tyranny of numbers. Two Americans, working half a continent apart, broke it.

1. July 1958 · Dallas

   Alone in the lab

   Jack Kilby joins Texas Instruments in May after eleven years at Centralab in Milwaukee. In July the company shuts down for its summer vacation. Kilby is too new to have earned any time off, so he has the lab nearly to himself. On 24 July he writes in his notebook that resistors, capacitors, transistors and diodes could all be made in one slice of semiconductor. When his boss, Willis Adcock, gets back, Kilby asks to build one.

2. 12 September 1958 · Dallas

   The sine wave

   Kilby works in germanium, the material TI’s lab can etch into the parts he needs. A transistor, a capacitor and resistors are all formed in one bar 7/16 of an inch long, joined by fine gold wires. TI’s leaders gather at the bench: Mark Shepherd, Willis Adcock and others. Kilby hooks the chip to an oscilloscope and switches it on. A steady sine wave crosses the screen. The integrated circuit works. A week later he has a second one running.

   

3. January 1959 · Palo Alto, California

   The planar process

   At Fairchild Semiconductor, Swiss-born physicist Jean Hoerni has worked out the planar process, an idea he first wrote down in December 1957. Leave the skin of oxide on the silicon, and it protects the delicate junctions underneath. On 14 January he shows it to Robert Noyce and the company’s patent attorney. Nine days later Noyce writes in his notebook how to build a whole circuit inside silicon under that oxide, wired by metal lines laid right on top.

   Robert Noyce, lab notebook, 23 January 1959It would be desirable to make multiple devices on a single piece of silicon, in order to be able to make interconnections between devices as part of the manufacturing process, and thus reduce size, weight, etc. as well as cost per active element.
4. 6 February 1959 · Dallas

   Solid circuits

   Texas Instruments files Kilby’s patent for miniaturized electronic circuits, with every part of the circuit built into one body of semiconductor. The patent says there is no limit upon the complexity or configuration of circuits which can be made in this manner. In March TI shows the solid circuit to the public. Noyce files his own patent on 30 July, for the silicon version with its metal wiring on the oxide. His is granted first, on 25 April 1961.

5. 27 September 1960 · Palo Alto

   The first planar chip

   Noyce puts Jay Last, a fellow Fairchild founder, in charge of turning the idea into a product. Last’s team designs a flip-flop with four transistors and five resistors on one chip of silicon. A first version works on 26 May 1960. On 27 September the design meant for production works too, its parts isolated inside the silicon itself. Fairchild announces it as Micrologic in March 1961. That October, Texas Instruments brings out its own planar chips.

6. 1962 to 1965

   To the Moon and the missile fields

   The first chips cost many times what loose transistors did. Only buyers who needed electronics small, light and sparing with power could pay. NASA led the way. MIT designed the Apollo Guidance Computer around a single Fairchild logic chip, and the program bought about 200,000 of them at $20 to $30 each. It was the biggest buyer of chips through 1965. That year the Minuteman II missile, guided by circuits TI designed, took the lead.

7. 19 April 1965

   Cramming more components

   Gordon Moore, research director at Fairchild, plots the number of parts on the most economical chips each year since 1959. It has doubled every year. In the magazine Electronics he extends the line to 65,000 components on one chip by 1975. In 1975 he resets the pace to a doubling every two years. The industry adopts it as its yardstick, and Caltech professor Carver Mead gives it a name: Moore’s Law.

   Gordon Moore, Electronics, 19 April 1965Integrated circuits will lead to such wonders as home computers.
8. 15 November 1971

   A computer on a chip

   In July 1968 Noyce and Moore leave Fairchild and found Intel. The next year Busicom, a Japanese calculator maker, orders a set of custom chips. Ted Hoff and Stanley Mazor propose a general-purpose processor that software can program instead. Federico Faggin, working with Busicom’s Masatoshi Shima, fits it onto one chip of 2,300 transistors. On 15 November 1971 Intel announces the 4004, the world’s first microprocessor.

9. 10 December 2000 · Stockholm

   The Nobel Prize

   Jack Kilby, 77, receives the Nobel Prize in Physics for his part in the invention of the integrated circuit. Robert Noyce died in 1990, and the prize is never given after death. In his Nobel biography Kilby writes that he has no doubt they would have shared it. Over his career Kilby earns more than sixty patents, among them one for the handheld electronic calculator he helped invent at TI to show what chips could do.