The transistor was invented in late 1947 by Bardeen, Brattain and Shockley. By 1951 it had been rendered producible by intense joint effort between Bell Telephone Laboratories in Murray Hill, NJ and Western Electric in Allentown, PA. It was in the best interest of the U.S. Military establishment and the Bell System to foster a nation-wide transistor industry. To this end, a most exciting symposium was held at Murray Hill on September 17-21, 1951, to whet the appetites of prospective transistor makers and users. To this end, the Symposium succeeded admirably.
One note of explanation: In 1951, the point-contact transistor (now extinct) was king of the hill above the one-megahertz frequency range, and for switching applications. As photolithography and processing technology improved over the years, junction-type devices took over leadership in all applications. The fact that the Symposium dwelt largely on point-contact devices in no way diminishes itís importance in sparking the transistor industry.
At the time, I was a solid-state device-circuit research engineer with Hughes Aircraft Company in Culver City, CA. To my delight, I was one of several Hughes people assigned to attend the BTL Seminar and to evaluate the feasibility and business potential of transistors.
For a price of $25,000 we were permitted to attend the Symposium, and to later receive a large (792 pages) bound volume elegantly and simply titled "The Transistor". This volume came to be called "the Twenty-Five Thousand Dollar Book". Ma Bell didnít reveal her device processing secrets at the Symposium; her deepest secrets would be revealed only to those companies which signed up as patent licensees.
The symposium was held in the BTL Murray Hill auditorium. It was very well organized, with opening remarks by top Bell people and H. A. Affel, program chairman. Attendees were electrified by an announcement that production-quality point-contact transistor samples were available to qualified semiconductor makers and systems users. Junction transistors were expected to be available by the end of the year.
Introduction to the tutorial sessions was by human dynamo Jack Morton. There followed an array of excellent presentations on exciting topics by world-class technologists. Audience interest was piqued by early presentations, and grew into a wave of enthusiasm as visions of possible new device, circuit and systems worlds unfolded. (In reality, the transistor industry was destined to attain even far greater heights than were implied by the BTL presentations.) The result of this well-choreographed Bell onslaught was that almost all attendees were most anxious to get a piece of the action. We were also able to mix with the stellar cast of BTL technologists at lunch and other off-hours functions, and toured such areas as the huge vacuum-tube digital computer and the remains of the old electric-relay machine. We had very little access to the device and process labs, since we werenít yet patent partners.
A brief summary of the Symposium is necessary for the proper historical recording of the event. I have tried to condense my personal notes from an intensive array of presentations into the fewest possible words:
Jack Morton, head of Bellís semiconductor device programs, noted that there were three prime early limitations in transistors. These were poor reproducibility, reliability and designability. "Designability" covered a host of limitations ranging from basic design knowledge to not-yet understood phenomena in gain, noise, frequency and power. Jack then stated (perhaps with some bravado) that these limitations had been overcome, at least to the level needed for practical systems applications at moderate temperatures. It was apparent that small size, low power drain, small heat generation, innately long life (70,000+ hours) and high shock/vibration capabilities spelled the doom of vacuum tubes in many systems applications. Transistors had their own unique capabilities; one must start with the systems requirements and design around transistors, rather than by trying to substitute transistors one-for-one for vacuum tubes. Jack said that a major unknown was in accurate cost projections for transistor devices. The sense of Jack Mortonís remarks was that it was high time to push BTLís transistor technology into the outside world.
Gerald Pearson gave a short dissertation on semiconductor theory. (Bill Shockley had asserted his independence by choosing not to appear, a source of irritation to Jack Morton.)
Morgan Sparks discussed transistor theory, and made some previously mysterious phenomena more understandable. For instance, he discussed the P-N hook effect in NPNP structures, and then applied it to the high current gain that occurred in point-contact transistors.
Next, R. M. (Bob) Ryder described the widely variant characteristics of point-contact and junction transistors, and described the effective characterization of these devices.
As a change of pace, John N. Shive discussed Phototransistors, in which high sensitivity was attained by using photons to generate base-region minority carriers.
Bob Ryder then discussed point-contact amplifier behavior, based on his earlier discussion. He showed that these apparently unstable devices could be properly harnessed if one just took the proper vantage point in circuit design.
Gordon Raisbeck discussed the duality concept, based on the fact that point-contact transistor circuit theory behaves a great deal like vacuum tube theory if you just interchange the roles of voltage and current in the circuit matrix equations. This at least made the old vacuum-tube guys more comfortable.
Junction amplifiers were then discussed by R. L. (Bob) Wallace. This excellent paper related circuit design approaches to the characteristics distributions and temperature sensitivities of junction transistors. Old-line circuit designers were pleased to find that junction transistor circuits are straightforward and inherently stable. The circuit hacks in the crowd were pleased to find that common-emitter junction transistor circuits behave much like common-cathode vacuum tubes, if you donít look too close.
Bob Ryder summarized the relative performance of point-contact and junction transistors. At that state of the art (remember, weíre talking 1951), junction transistors were best in micropower low-frequency, low distortion, low-noise, high-gain, stable applications, and lent well to automatic gain control. Point-contact transistors were best as high frequency amplifiers (MHz regions) and as switching devices.
At this point the Symposium swung over to actual systems circuits, presuming that we had learned what was said in the earlier presentations on transistors and their immediate circuit environments.
R. S. (Bob) Caruthers discussed systems applications of transistors. This covered real, applications-proven amplifier circuits with actual circuit values. We had been pulled into the "real world" of transistor circuits.
Gordon Raisbeck then covered transistor oscillator circuit principles, both conventional (e.g. Hartley) and negative-resistance versions.
Bob Caruthers treated practical oscillator circuits, plus special novel circuits such as single-supply iterated amplifiers. One more stretch of the imagination!
Gordon Raisbeck discussed modulator circuits for point-contact transistors. He noted that the emitter junction acts as a diode modulator, so that the transistor can modulate signals far beyond the transistor cut-off, as long as the beat frequency is within the transistor frequency capability.
Bob Caruthers noted that germanium could replace copper oxide in balanced modulators, and that transistors would result in modulators with gain.
At this point the action swung to pulse and digital transistor applications. The Bell System was interested in transistors for call switching, for control, and for superior digital computer systems. Many of us at the Symposium caught the smell of blood, because of applicability to our own computer and control systems for both military and civilian applications. Others could care less, being limited in vision to the analog world which prevailed at the time. To the visionaries, here was a way to way to create practical digital systems that were impossible by the use of elephantine vacuum tube technology. However, none of us imagined the semiconductor revolution that was really to take place over the next forty years.
First, J. H. Felker fascinated us with the applicability of transistors to high-speed digital computers. He stated that the prime objective was practically infinite reliability, closely followed by low power consumption (hence minimal heat removal), small size and minimal weight. He blocked out computer subsystems elements, and gave instances of practical circuits. Point-contact subsystems had been operated over a year with no failures, and with complete interchangeability of modules. The new world was demonstrably here!
A. E. Anderson and J. R. Davey gave a set of four papers on the application of point-contact transistors to pulse and switching circuits. The discussions ranged all the way from device characteristics through flip-flops and switching circuits to detailed practical circuits. The capper was a subsystem (by 1951 criteria); an analog-to-digital converter!
R.E. (Bob) Yeager described optical encoder circuits for going from shaft-position Gray code to binary output. J. R. Harris then made the logical transition into practical shift registers.
Jack Morton gave a supercharged summary talk, reiterating his opening remarks, with their implication that it was high time to take BTLís transistor technology to the outside world. We believed! We believed!