EARLY TRANSISTOR HISTORY AT RCA Gerald B. Herzog
Oral History – Jerry Herzog (Continued)	Oral History – Jerry Herzog (Continued)
The Request for Proposal defined large scale integration as the equivalent of 100 gates interconnected on a single semiconductor chip. While RCA and its semiconductor division were in no position to be able to fabricate such arrays at the time we bid on the Request for Proposal, we did have the advantage of having developed some CMOS memory arrays by that time, and I also had the computer aided design group. We put together a proposal that said we would develop Emitter Coupled Logic Arrays for logic and Complementary Symmetry Arrays for memory for an Air Force computer that would do navigational programs. At the time of writing the proposal the semiconductor people at Somerville argued that we should bid Resistor Transistor Logic, because that was the simplest form of logic. However, I felt that since the performance of RTL logic was very poor, it didn’t make any sense to make arrays that would not be functional in RCA computers. Instead, I opted to try to fabricate the most complicated form of logic arrays – Emitter Coupled Logic, or ECL. During this contract, the people working under my supervision made some significant contributions that would have industry wide impact for larger arrays in the future. Andy Dingwall proposed that a lot of the damage to the photo-resist causing faults in the integrated circuits was due to the mask being pressed against the photo-resist material.	He devised a projection printing system where the light source was elevated about eight feet above the mask itself, and the mask was separated from the semiconductor wafer so as not to impact the photo-resist material. This resulted in substantially improved yields for our integrated circuits. Al Medwin then pointed out that using projection optics would be even better, and got hold of an RCA projection system that used Schmidt Optics. He used projection printing to image the pattern on to the semiconductor wafer, which again gave great improvements in semiconductor yield. In one of our quarterly reports to the Air Force, we pointed out the improved results from projection printing and suggested the Air Force fund an industry effort to develop projection printing. This led to a contract with Perkin Elmer for projection printing equipment, which became standard later in the industry. Still, getting high yields of gates with as many as several hundred transistors in the array was not easy. Finally, after a little more than three years, we did deliver a working navigational computer that had large semiconductor memory arrays of complementary MOS as well as arrays of high speed Emitter Coupled Logic. Go To Herzog Oral History, Page 6
COPYRIGHT © 2004 by Jack Ward. All Rights Reserved. http://www.transistormuseum.com/ PAGE 5