WHOSE BRAINCHILD WAS THE BRAIN CHIP?
In 1969, year-old Intel Corp. was working feverishly to invent the dynamic random-access memory, or DRAM. Intel unveiled the chip in 1970. By the mid-1980s, however, Intel was forced to abandon this early success as Japanese rivals dumped DRAMs on the market.
So Intel was fortunate that in those early days it was also dragged into the race to design the first microprocessor. Researchers had been discussing the idea of a computer-on-a-chip since the mid-1960s, and if Intel had dallied, someone else would almost certainly have built one first.
What got Intel going was a visit by Japanese engineer Masatoshi Shima in April, 1969. He brought with him plans for a set of seven chips that his company, Busicom Corp., wanted for a desktop calculator. Intel took the job of turning the plans into silicon chips.
BREATHING SPELL. Initially, things moved slowly because Intel was preoccupied with DRAMs. Engineers repeatedly were taken off the Busicom project to work on DRAMs. That summer, though, Marcian E. ''Ted'' Hoff Jr. had a brainstorm. He quickly blocked out an alternative design for the new chip. He wanted to discard Shima's decimal design--specifically suited for things like calculators--in favor of a digital processor using binary logic. This processor would do math by executing three dozen 4-bit instructions. Busicom approved, but the project continued only to inch ahead.
Then, in November, 1969, a team led by Victor D. Poor at Datapoint Corp. (then called Computer Terminal Corp.) hatched a plan for an 8-bit programmable computer terminal. It contained the nucleus for all modern microprocessors. Poor asked both Intel and Texas Instruments Inc. to take a crack at it. That heightened Intel's interest. At TI, Gary W. Boone initially figured the processor would require a set of three chips. But his team switched to a single-chip approach after Boone heard that Intel was taking that tack.
In the spring of 1970, Shima made another visit to Intel. Disappointed by the lack of progress, he stayed on to lend a hand, working on the logic design of Hoff's digital chip while Hoff concentrated on the overall integration of the logic chip and three types of support chips. Pitching in were Stanley Mazor, who took charge of software matters, and Federico Faggin, who ran the program and engineered silicon circuits from Shima's blueprints. This is the team that was recently inducted into the National Inventor's Hall of Fame.
In March, 1971, Intel produced the first microprocessor and began deliveries to Busicom. Two months later, Texas Instruments built the first 8-bit microprocessor for Datapoint. Although Datapoint decided the chip didn't measure up, TI became the first chipmaker to publicly announce a microprocessor. Boone began adapting the technology to other TI chips, but he failed to persuade TI to promote the technology. That gave Intel a chance to catch up.
Intel's engineers eventually convinced the company's marketing chief that the Busicom chip had broad potential. In November, 1971, Intel announced its 4-bit chip family with ads proclaiming ''a new era of integrated electronics.'' Each fingernail-size 4004 chip packed as much computing power as the first electronic computer, ENIAC, which was almost as big as a two-car garage when it was built from 18,000 vacuum tubes in 1946.
With only 2,300 transistors, the 4004 microprocessor was primitive by today's standards. But it was the launchpad for Intel's eventual domination of the PC market. And even today, the 4004 is still blazing new trails. It's the brain of Pioneer 10, which was launched on Mar. 2, 1972, to gather the first close-up images of Jupiter--and is now exploring space beyond the solar system, where no microprocessor has gone before.
By Otis Port in New York
Updated June 14, 1997 by bwwebmaster
Copyright 1996, Bloomberg L.P.