Edwin G. Millis


Oral History – Ed Millis



Two seconds later, after a bunch of ratcheting sounds from stepping switches, one of the 108 plastic drinking glasses would light up and the transistor would be tossed into it.  It was a swell machine except for one slight oversight on my part.  Only one person in the Lemmon Ave plant could change a burnt-out light bulb in it, and that was Grace Weatherall, a lovely and petite secretary.  Her slim hand alone could reach through the drinking glass support hole and unscrew the bulb.  She was a very good sport about it, considering there were six machines, each with 108 bulbs!


The problem of testing all the transistors TI was cranking out was beginning to be a major headache.  The testing system was unchanged from the beginning of production several years before.  Each separate test for a transistor required a test box and an operator to run it.  The test box told the operator if the transistor was good or bad, or occasionally, what range of goodness it was in.  So there were a lot of test boxes on a lot of tables with a lot of operators passing transistors from one to the other.   And we, in the test equipment group, were drowning in orders for more test boxes.  Things were going to get worse before they got better.  So we began brainstorming about this test problem and developed an idea for a modular test machine that would be built up like tinker toys – just keep adding on what you needed.  It could have almost any number of testing stations and sorting stations.




Oral History – Ed Millis



Each transistor would ride in its own little test block through testing stations until it came to a sorting station that checked the results of the tests and decided if it qualified for selection.  If it did, the transistor would be pulled from the block and put into a bucket.  A punched paper card travelled with the test block, keeping track of the passing and failing of each test.  The sorting station would look at the array of holes and decide the transistor’s fate.  We had pretty much worked out the details of how we thought the system ought to be built, but Jim called one last meeting before putting together a proposal to management.  We needed to think of a name for it.  We argued for an hour with no particular progress until Earl “Mac” McDonald, a terrific test set builder, declared, “How about ‘Centralized Automatic Tester?’.  We could call it the CAT machine.”  We all liked that a lot and the name was adopted.  It lasted a long, long time. Although not immediately accepted by management, this CAT system soon became the standard for all TI transistor production testing.  We couldn’t build them fast enough.  With CAT’s, and in 1961 the Super CAT’s, TI was able to greatly expand the volume of transistors that could be manufactured and tested.  The CAT and Super CAT probably had more impact on the production of transistors at TI than any machines our group designed. 



Millis Oral History, Page 5


COPYRIGHT © 2001 by Jack Ward.  All Rights Reserved.  http://www.transistormuseum.com

Page 4