1951 Bell Telephone Laboratories Transistor Symposium

From the November, 1951 issue of the Bell Telephone Record:

    In cooperation with the military services, Bell Telephone Laboratories recently held a five day symposium on Transistor electronics at the Murray Hill Laboratories in New Jersey, where the Transistor was invented a little more than three years ago. Nearly 300 guests attended the sessions which ran from September 17 to September 21, inclusive.

    In compiling the invitation list, the military services were asked by the Laboratories to nominate representatives of their own groups and of contractors whose work was of a nature that would benefit from more detailed information on recent progress in the Transistor field. More than 100 representatives from the Army, Navy, Air Force and government agencies attended the sessions. About a score of universities were also
represented.

    Nearly 100 representatives of a wide range of industrial companies also participated. Large and small, these companies covered the major part of the American electronic industry. The material presented was necessarily "restricted" and consequently all participants in the sessions were either United States citizens or specifically cleared by the military services.

    In opening the symposium, Dr. Kelly, stated that improved point contact Transistors are becoming available in small quantities through the Western Electric Company and that the new junction type Transistors are expected to be available for limited experimental purposes by the end of the year.

    He said the Laboratories proposed the symposium so that present knowledge of the Transistor and its circuit applications could be made
available to circuit and systems engineers wishing to explore their possible use, particularly in military projects. Accelerated application of Transistor devices in the interests of national security is expected to result.

    Material was presented at the symposium by nearly a score of Bell Laboratories scientists and engineers working on Transistor electronics.
Topics covered were basic Transistor physics and theory, the characteristics of experimental Transistors and Transistor circuitry.

    Arrangements for the Symposium were under the direction of a committee consisting of H. A. Affel, Chairman, A. Tradup, H. B. Ely, and T. N. Pole. Mr. Affel was in charge of the technical program and sessions; Mr. Tradup took care of the invitations in cooperation with the Military; and Mr. Ely and Mr. Pope were concerned with the general arrangements, luncheons, transportation, registration, etc. About 250 visitors were brought from New York to Murray Hill and returned each day in buses; these buses were also used to transport them to and from lunch each noon, at which they were guests of the Laboratories. F. E. Dorlon supervised the serving of the luncheons, which were furnished by a caterer.

    The size and length of this Symposium obviously required the assistance of many Laboratories people, too numerous to mention here.
Appreciation expressed by the visitors, however, for the opportunity to learn about the Transistor in its present stages of development, indicates that the efforts made in this undertaking were well worthwhile.

                        TRANSISTOR SYMPOSIUM AGENDA
                         Bell Telephone Laboratories
                               Murray Hill, N.J.
                               Sept. 17-21, 1951

Registration

Welcome

Introduction                                          J. W. McRae

Symposium Arrangements                                H. A. Affel

Present Status                                        J. A. Morton

Theory of Semi-Conductors                             G. L. Pearson

Transistor Theory                                     M. Sparks

Characteristics and Properties of Point Contact
Transistors Applicable in CW Transmission Circuits    R. M. Ryder

Basic Amplifier Properties of Transistors             R. M. Ryder

Photosensitive Properties of Transistors              J. N. Shive

Basic Point-Contact Amplifier Behavior                R. M. Ryder

Power Amplifiers and Duality                          C. Raisbeck

Amplifiers with Junction Transistors                  R. L. Wallace, Jr.

Resume of Transistor Characteristics in Their
Application to Amplifiers                             R. M. Ryder

Some System Applications of Transistor Amplifiers     R. S. Caruthers

General Oscillator Considerations                     G. Raisbeck

Experimental and Practical Applications of
Transistor Oscillators                                R. S. Caruthers

Inspection of Murray Hill Laboratories

General Modulator Considerations                      G. Raisbeck

Modulators in Carrier Telephone Systems               R. S. Caruthers

Application of Transistors in a High-Speed Computer   J. H. Felker

Transistor Characteristics and Properties Applicable
to Pulse and Control Circuits                         A. E. Anderson

Some Circuit Design Considerations                    J. R. Davey

Introduction to Transistor Building Blocks and
Assemblies                                            A. E. Anderson

Typical Building Blocks                               J. R. Davey

Binary Counter                                        R. L. Trent

Optical Encoder                                       R. E. Yaeger

Shift Register and Serial Adder                       J. R. Harris

Data on Experimental Transistor Types                 J. A. Morton

Concluding Remarks                                    H. A. Affel

Subject: The Present Status of Transistors and Transistor Applications                                                     Date: October 1, 1951

1. Transistor Symposium

The present memorandum is occasioned by the passage of an event which may prove to be a milestone in the electronics industry. The event was a week long transistor symposium held during September 17 through 21, 1951 at the Murray Hill Laboratories of Bell Telephone Laboratories, Inc. Arranged at the request of the Military Services for their contractors and service laboratory personnel, the symposium was intended to advise industry of the progress made to date in exploiting the transistor. That such a symposium should be held is direct evidence that in the minds of the military it is time that industry be advised of and become ready for transistor usage. Force is lent to this statement by the impending availability of Western Electric transistors. For the Bell Laboratories, the symposium was a progress report on what started as pure research on semiconductors and has become development and application of transistor electronics. It means that point contact transistor development and pilot production, junction transistor development and transistor circuit development have proceeded to the extent that large transistor usage is certain and that general application and improvement by the rest of the industry is in order. There is no doubt that from this day forward transistors will become increasingly important in development programs and ultimately in equipment designs. The several hundred engineers who attended the symposium will carry this story to their individual companies and those companies that are not, will soon be designing transistors or transistor circuits. That the situation is unusual there is no denying, for, over the years, it means the substitution of one active element for another in many circuits.

Before proceeding to an enumeration of transistor facts it is believed worth while to call attention to the fact that in the opinion of many Bell Laboratories scientists and engineers, as well as others, the present transistors are but the first devices in the new technological field of transistor electronics. Considered against the time background of the development of vacuum tubes, the statement can be really appreciated. Thus new techniques, new fields, rather than specific devices with limited characteristics are pictured from a broad point of view. There are undoubtedly other transistor electronic devices to come as well as significant reductions in the limitations of noise, frequency cutoff, gain and temperature dependence, In fact the latter factor, temperature dependence is the only factor having basic theoretical connotations. Not only is work required on the semiconductor material itself but circuit design to minimize or eliminate the dependence is possible and has essentially not begun.

A copy of the symposium program (Appendix I) is attached because it provides a quick appreciation of the extent of the material presented. Briefly, status, semiconductor theory, transistor theory, point contact characteristics, junction characteristics, CW applications and pulse applications were discussed. Copies of the symposium papers will be available at a later date (at least four weeks). Even more detail can be found in the references of Appendix II. Information on status given at the symposium will be presented in the paragraphs to follow.

The transistor invented in 1948 had a number of important limitations which prevented its general use. At the present time these limitations can be said to have been reduced but not eliminated. They have been reduced, as already indicated, to a point where serious application is appropriate. They are:

a.  Uniformity was poor - no two were alike.

b.  Reliability was poor - properties changed inexplicably with time and temperature.

c.  Designability was poor - properties were inadequate (noise figure, cutoff frequency, gain).

Because of the improvements of the last three years it can be said that:

a.  Theory can be used to design to desired characteristics, i.e., the electrical characteristics have been related to the physical properties and to manufacture, in some transistors.

b.  Point contact types are much further along (six months to one year at present rate of work). Both point contact and junction types can be made in the laboratory, but only in the point contact type is manufacture well understood and under control.

c.  Reliability has been greatly improved. There is still an important temperature dependence which is the greatest reliability problem. Other problems have been reduced to minor importance.

d. Properties have been improved, i.e., the range of performance extended through the use of the theory. The theory and the development of the junction transistor have greatly extended the breadth of application of transistor devices.

e.  Many problems remain, but use is now warranted. It is expected that rapid development of the art will take place as the bulk of the electronics industry works in the field.

It is early to compare a 3-1/2 year development with a 44 year old industry but the comparison must be made if usage is contemplated. The points of comparison are:

a.  Technical performance

    (1) CW or linear transmission
    (2) Pulse or non-linear applications

b.  Reliability

c.  Miniaturization

d.  Power consumption

Availability and cost are not points of comparison at the present time. The goal of quantity transistor production is a unit cost of less than a dollar.

In this area a number of figures of merit exist as a means of comparison

                             Vacuum Tube            Transistor
                             -----------            ----------

Gain per Stage                20-50 db               20-50 db

Noise Figure                   0-30 db     point contact 45 db at 1 kc
                                           decreasing 10 db/decade
                                           junction 10 db at 1 kc

Upper frequency limit        60,000 mc                30 mc

Power Output                10^6 watts     measured in watts at present,
                                           no intrinsic limitation

Gain x Bandwidth Product   100-1000 mc     100-1000 mc but fc = 30 mc
                                           and trading not always
                                           possible

Efficiency                 30% Class A,    Point contact 30% Class A
                          5% distortion    Junction 49% Class A
                                                    75% Class B
                                                    95% Class C

Generation                  0-50000 mc             0-VHF
(approx. 50 mc)

To summarize, from 0 to 10 mc, vacuum tubes and transistors are about equal. Above 30 mc, functions are best performed by tubes. In between is a disputed region probably favoring tubes at present.

By this category is meant pulse transmission and control including pulse generation, storage, delay, amplification, switching and computation. If the pulse width is less than 0.5 microsecond and the repetition rate less than 1 mc, transistors are the equal of tubes. Pulse rise times of 0.1 microsecond may be achieved quite easily. At faster speeds and shorter widths both transistors and tubes give trouble, with tubes somewhat to be preferred. The use of transistors in conjunction with diodes in pulse circuits results in simplified coupling problems.

Transistor life has been defined as the number of hours of operation until the gain decreases by 3 db. A life of 70,000 hours has been estimated. The figure came from several considerations including life tests on old transistors that have been running 24,000 hours. Vacuum tube life is 1000 to 25,000 hours usually near the lower end.

Both junction and point contact type transistors will withstand shocks of 20,000 g's.

No vibration modulation is produced in transistors that could be detected with 100 g over the audio band and 1000 g at selected frequencies.

An operating ambient temperature range of -20 C to +80 C is permissible for most transistors. Cycling from -30 C to +60 C does not affect characteristics. Performance with respect to temperature is definitely worse for transistors as compared to vacuum tubes.

Humidity does not affect transistors because they are sealed.

The volume required for a transistor is 1/50 to 1/2000 cu. in. in comparison to 1/8 to 1 cu. in. for vacuum tubes, a five or ten to one reduction at least. Transistors are small enough to return the burden of miniaturization to the designers of passive circuit elements.

Low power operation is one of the greatest of transistor advantages. Vacuum tubes require 25 to 1000 milliwatts heater power and a similar quantity for plate power. transistors have no heater hence no standby power but instant response. Operating power is 4 to 50 milliwatts for the point contact type and 1 to 100 microwatts for the junction type.

Low power consumption and high efficiency mean that real advantage can be taken of miniaturization. The density of transistors in equipment can be amazingly high.

In the process, of exploiting the transistor and in providing suitable patent coverage, Bell Telephone Laboratories has designed circuits to cover nearly all of the presently known vacuum tube applications. Hence, the extent to which transistors having the characteristics already described may be applied in typical electronic circuits was largely revealed at the symposium. It does not mean that all circuits and applications have already been invented. There is much to be done, particularly in the area of new applications, exploiting the peculiar characteristics of the transistor, and in improving bias circuits. Power supply circuits, if the advantage of low power use is to be maximized, need great effort.

Circuit design has been by three methods:

a.  Analogy of transistor to vacuum tubes in order to arrive at suggested circuits.

b.  Duality of transistor and vacuum tube in order to arrive at suggested circuits.

c. Independent or fresh approach

In the CW applications area, dc, audio, video, i-f and r-f amplifiers have been built. The highest frequency in any amplifier described was about 25 mc. Also oscillators, mixers, detectors, modulators, etc., have all been investigated theoretically and experimentally.

In the pulse or non-linear area, switches, gates, amplifiers, counters, storage elements, etc., have all been built and tested.

In both areas, not only have individual circuits been designed, but groups of circuits have been used in specific equipment applications. CW usage has been in telephone equipment. Pulse usage has been in computers and data transmission equipment. Bell Labs has emphasized micropower rather than high power transistors and circuitry. A rack full of transistors in computer circuits was shown which required 3 to 5 watts of power as well as an oscillator which ran from sunlight, body heat, pickup from an antenna, or a quarter-and-spit battery.

Again summarizing vacuum tubes and transistors today; transistors are the equal of tubes in pulse systems if the pulse repetition rates are less than 1 mc and pulse lengths greater than 0.5 microsecond. At faster speeds, tubes have a slight region of preference over transistors before both are in trouble. For CW applications, tubes and transistors are equivalent in performance up to 10 mc.

Above 30 mc tubes are to be preferred. In between lies a doubtful area where tubes are probably to be preferred unless miniaturization is all important. The temperature dependence, although not bad in the range -20 C to +80 C must always be checked in a given application. Transistors capable of power greater than several hundred milliwatts, although accomplished in the laboratory are not yet available.

Orders are now being accepted by Western Electric Company for transistors. Deliveries are still several months away and allocation will be by the Military Services. Copies of transistor characteristics will be circulated. A summary of types that will be available in the order of production as well as it is known is:

Point Contact Types:

WE Model  A1698    primarily switching use, cartridge style
Bell Labs M1689    primarily switching use, bead style
Bell Labs M1734    high frequency, switching or CW, cartridge style
Bell Labs M1725    minimum power drain, cartridge style
Bell Labs M1768    supersedes 1725
Bell Labs M1729    audio and carrier applications, cartridge style

Junction types:

Bell Labs M1752   NPN transistor, low voltage, high efficiency
Bell Labs M1740   p-n junction photocell
Bell Labs M1727   p-n junction diode, low forward impedance
Bell Labs M1728   p-n junction diode, high reverse impedance
Bell Labs M1754   p-n junction diode, high forward to back ratio
Bell Labs M1755   p-n junction diode, sharp break-away

Signed,
D. M. Powers

 Designation
WECO    RETMA    Description
------  -----    -------------------------
1689    2N23     point contact bead
1698    2N22     point contact cartridge
1723    2N21     point contact cartridge
1725             point contact cartridge
1729    2N25     point contact cartridge
1734             point contact cartridge
1752-1  2N27     n-p-n junction
1752-2  2N28     n-p-n junction
1752-3  2N29     n-p-n junction
1760    2N26     point contact bead
1768    2N24     point contact cartridge

The date, June 6, 1949, is only a year and a half after the invention of the device itself. Of course there were no type numbers beyond the generic "Type A" which simply meant that it was a cylindrical cartridge  style.