Mannerisms

Faggin joined Fairchild in 1968, the year two of its founders, Bob Noyce and Gordon Moore, went off to found Intel.

At Fairchild, Faggin pioneered the development of silicon gate MOS technology and designed the world's first commercial chip to use it - an analogue multiplexer with decoding logic called the 3708.

Faggin hungered to be at the forefront of semiconductor technology. "I guess I was trained to be a masochist -  trying to impress my father or something - with a need to do something people haven't done before".

That hunger brought him to Intel in the Spring of 1970. "If you wanted to do wonderful things", he says, "Intel was the place to be".

"When I joined Intel it only had a hundred people. It wasn't making any money. It was struggling to become a viable company. The semiconductor memory business which Intel pioneered wasn't coming on as fast as expected. Times were not good and that opened the door to Busicom."

Busicom had offered Intel a contract for custom chips for a calculator. Cash-strapped and productless in 1969 the Intel management had taken it on. "I was hired from Fairchild to implement the project" says Faggin.

"Intel was delinquent on the contract when I joined the company," he remembers.

Shortly after he started at Intel Faggin recalls a Busicom engineer, Masatoshi Shima, coming over from Japan.

"Shima was furious when he found that no work had been done since his visit approximately six months earlier," says Faggin, "he kept saying: 'I came here to check. There is nothing to check. This is just an idea.'"

Busicom's schedule for introducing its calculator had been compromised.

The situation spurred Faggin to Herculean efforts." I worked like a madman - 12 to 16 hours a day", recalls Faggin, "the early Intel was a tough environment - very results-oriented - a 'get it done' mentality. You had to perform you had to make it happen you couldn't coast even for a week. People got burned out but they were a bunch of good guys - the density of good guys was very high."

"I had nothing to do with the argument about the architecture" recounts Faggin, "Hoff (Ted Hoff, the inventor of the microprocessor) had created an architecture that could be implemented within the technological limitations of the day. Not much was known about designing random logic using silicon gate technology. You had to play very close to the physics of the devices in those days. We were trying to use as few transistors as we could. You could only integrate three or four thousand transistors on a chip in those days, so we couldn't be profligate."

"To keep the circuits simple I had to use bootstrap loads - no one at Intel thought it was possible to use bootstrap loads with silicon gate technology." When Faggin demonstrated their feasibility Intel adopted them for its memory chips.

Hoff's architecture envisioned four chips. The first chip to reach silicon was a 2K ROM with a 4-bit mask-programmable I/O port - dubbed 4001. It was first fabbed in October 1970 and "worked perfectly". Second out of the bag was a 320-bit RAM with a 4-bit output port. Called the 4002 it came out in November with "only one minor error". The same month the 4003 came out - a 10-bit serial-in parallel-out shift register to be used as an I/O expander - and "worked perfectly."

Then around Christmas came a "major disappointment" with the fourth chip - the 4004. The all-important 4-bit central processing unit (CPU) - turned out to be unusable. One of the masking layers had been omitted. Three weeks later they tried to make the 4004 again.

"My hands were trembling as I loaded the two inch wafer into the probe station" remembers Faggin, "it was late at night and I was alone in the lab."

 As he tested the wafer bit by bit his spirits rose. "My excitement mounted as I found various areas of the circuit working. By 3a.m. I went home in a strange state of exhaustion and excitement".

A few days later verification was completed and "only a few minor errors had been found." For Faggin it was a supreme moment. "I was elated. All that work had suddenly paid off in a moment of intense satisfaction."

NEXT FRIDAY, August 1st: Marketing the 4004 by Ted Hoff