EARLY TRANSISTOR HISTORY AT RCA Herb Meisel
Oral History – Herb Meisel  (Continued) The lab director Ed Herold (mentioned in Kurshan's history) felt we should work on diffusion for the diode junction.  In retrospect he was right, as usual. We did do one experiment with diffusion but we didn't know how to proceed after the diffusion step. The program really got into high gear when I got back to Somerville and Lorne Armstrong took charge. Hank Harmon was also assigned. We built a good sized vacuum furnace using a  pyrex tube about 18" long and about 1 1/2 " in diameter. The heater strip was wound on the outside. We made good aluminum to silicon junctions, when we could get decent crystal. Because aluminum is difficult to make a connection to, we came up with a way of firing a small tungsten ribbon into the aluminum dot during the alloying step. This described in U.S. Patent # 2,878,432 entitled "Silicon Junction Devices", inventors: Lorne Armstrong, Henry Harmon  and me. The mechanical requirements were very high; one of the planned applications was in proximity fuses in artillery shells. I was part of a group with a goal of putting a germanium transistor mount in an element meeting the form factor with a tough height requirement ~0.050". The structure is explained pictorially in U.S.Patent #2,971,138 ; inventors Adam Pikor, Leonard Schork, Adolf Blicher, and me.  Some of the engineering solutions I worked out  were: figuring a temporary mount to get through connector wire solder, etch, rinse , bake, junction coat , test, cut of from the temporary mount, mount into cavity, develop an alumina loaded resin to fill cavity, hot solder seal with pure rosin flux. Samples passed temp cycling, life test, and 20,000 G centrifuge.	Oral History – Herb Meisel  (Continued) I became an Engineering Group Leader in 1961 with responsibility for production support of germanium power and drift transistors and development and support of low and medium frequency silicon power transistors. The responsibility for the single diffused power family involved upgrading the designs, cost-reduction, completing a Signal Corps Industrial Preparedness contract and aiding in the transfer of production from Somerville to the new Mountaintop, PA plant. The Signal Corps contract included Devices 13, 14, and 15 which subsequently evolved into the 2N1482, 2N1486, and 2N1490 families. The contract was completed with delivery of the required devices plus 8000 hour life test results with data entered (manually) onto data processing cards ( aka Hollerith or IBM cards). After several chip design changes and some significant package cost reductions the 2N1486 and 2N1490 migrated into the 2N3054 and 2N3055 which became industry workhorse standards. The package redesigns were done by Milt Grimes.  The 2N3055 was the first multi-amp silicon power transistor to sell for less than one dollar! It was a huge success in the power supply market. In 1965, the team of Design, Production, and Applications engineers who launched these devices got  RCA Electronic Components Achievement Awards, which were quite generous. I was also involved in the engineering support of the production 2N2102,  which was a triple diffused planar transistor with a usable gain out to 1 amp and a power dissipation of 1 watt. It was a more powerful version of a very successful Fairchild planar transistor. NOTE: This website has been moved.  Use this link to continue.
COPYRIGHT © 2001 by Jack Ward.  All Rights Reserved.  http://www.transistormuseum.com