Transistor Museum Oral History Todd Page 6 GE Germanium Transistors

EARLY TRANSISTOR HISTORY AT GE

An Interview with Carl David Todd

Recollections from the First Days of Transistor Technology

Oral History – Carl David Todd (Continued)

Early GE Transistor Production:

We at GE knew that key to a profitable production line of transistors was automation. Manual labor costs would be reduced and more uniformity could be achieved. We had some top notch mechanical engineers who could take our verbal descriptions of crude sketches and turn them into a working physical machine in days, and sometimes just hours. The chemists added their part in developing etching procedures, etc. My first recollection of the overall yield was around 1%. With 18 cents per package, you can see that the price would have to be pretty high in order to make a profit. We assembled the transistors as a string of some 50 units that could be handled by the automation equipment, and I designed testing procedures that could predict pretty well whether the device would yield a saleable unit after packaging. That way, we did not waste an expensive package. A tribute to the overall process development was that when the high production line in Buffalo was closed, they were running over 99% yield of saleable transistors. The techniques used by GE for the fabrication process, particularly the application of the collector and emitter “dots” was totally different from that used by other manufacturers. It was less subject to contamination and yielded much greater uniformity. One thing to keep in mind about the early semiconductor years was that the theoreticians could develop a device in the laboratory, but the real magic occurred in bringing that device to production reality. There is a great difference between making one device or a half dozen and making 100,000 per week.

Oral History – Carl David Todd (Continued)

The GE 2N107 Hobbyist Transistor:

Well, that particular one – at the time our yield on transistors was pretty low. I remember the day we raised the yield from 1% to 2%, by changing our etching process. We had barrels and barrels of transistors that, let’s say, didn’t look normal. One particular time, we never could figure out what happened, but the h_FE factor was super high. Practically very one of the transistors in this big lot was over 180 to 200, and the normal limits (were much less) for the 2N45. So these suspect units would end up going into these barrels, they were out of spec so far that we wouldn’t sell them - we didn’t trust them. Well, we were gathering all these barrels, and I knew there were a lot of hobbyists, and granted there was the CK722, but GE didn’t have anything to offer, because the 2N45 was probably around $3.50 to $4.00 by the time it got to the hobbyist.

So, I talked to the marketing guys and the product planner, and I said “Why don’t we introduce a transistor with much lower specs than meets our normal (requirements), strictly as a hobbyist transistor.” Believe it or not, they weren’t interested at first. But I kept pressuring the guys, and they knew I was a hobbyist, so I finally got them convinced and we started selling them.

Go To Todd Oral History, Page 7

PAGE 6