|
The Junction Transistor With Permission, Bell
Laboratories RECORD, August 1951 This reprinted from SMEC 'Vintage Electrics'
Vol. #2, 1990
Significant advances in the development of the transistor, a tiny amplifying device which has been called the first serious rival of the vacuum tube, were recently announced by Bell Telephone Laboratories. The transistor was invented here three years ago. Most important of these advances is the construction of operating samples of a radically new type of transistor which has astonishing properties never before achieved in an amplifying device.
Its inventor, William Shockley, who initiated and directed the research leading to the original transistor, predicted the new type more than two years ago, as a result of complex theoretical studies he carried on as part of the Laboratories’ broad investigations into transistor physics.
Performance of this new transistor is described in technical articles published in the July issues of the Bell System Technical Journal, the Physical Review, and the Proceedings of the Institute of Radio Engineers. Associated with Dr. Shockley, and co-authors of the papers are Morgan Sparks, and G. K. Teal, who built the first of the new type transistors, and R. L. Wallace, Jr. and W. J. Pietenpol, who have been working on their development.
Development work on the original type of transistor has been so successful that this type will be put into trial use in the Bell System. The Laboratories have made transistors of this original type which are as uniform in performance as vacuum tubes. This achievement is a very significant advance over the state of transistor work two years ago. At that time transistors were highly variable in their characteristics and of uncertain reliability.
As a result of extensive development under the direction of J. A. Morton, the problems involved in reliability and reproducibility are now understood and it is expected that regular production can be started. Transistors have been produced which can withstand shock and vibration better than any known vacuum tube and they are expected to have a service life considerably longer than that of commercial vacuum tubes in current use. The transistor can now be designed for a great many specific functions, and its ranges of performance have been extended to include a wide variety of applications which at present require commercial tubes of the vacuum type.
The transistor is also expected to find future application in telephone apparatus where the use of vacuum tubes is now impractical, for example, in the complex switching mechanisms which are the basis of the dial system.
These and other advances, which may be expected to have an important effect on the entire field of electronics, are the result of intensive laboratory work on the original device. This is known as a "point contact" transistor, and consists essentially of two hair-thin wires resting on a tiny speck of germanium, which is a semiconducting metallic element. There is no glass envelope, no vacuum, and no heating element to cause a warm-up delay. The entire apparatus is housed in a metal cylinder about the size of a .22 caliber shell, although it may also be housed in a much smaller space for certain applications.
But in addition to making vast improvements in this original transistor, the researchers have now developed a radically new and in many ways more effective type of amplifier called a junction transistor. This is described in the technical papers, together with some developments in circuits for using it. Extremely efficient and rugged, this junction type is in the form of a small bead, about half the size of a pea.
The junction transistor has no point contacts, which in the original transistor corresponded to the terminals of a vacuum tube. Instead, it consists of a tiny rod-shaped piece of germanium, treated so that it embodies a thin electrically positive layer sandwiched between the two electrically negative ends.
The junction transistor derives its name from the two "junctions" between the negative ends and the positive layer. It differs markedly from the point contact type in which the contacts at the points play an essential role.
Power consumption of this new type of transistor is remarkably low. The signal level often found in modern electronic equipment is about one millionth of a watt. But a full watt is ordinarily used to amplify this signal by conventional vacuum tubes. This is about like sending a 12-car freight train, locomotive and all, to carry a pound of butter. The new transistor, unlike any earlier amplifier, can be operated on about a millionth of a watt, which is just sufficient to carry the signal without waste.
Meanwhile, rapid strides are being made in readying the original point contact transistor, with its recent refinements and improvements, for actual commercial use in the Bell System. The transistor will be used in equipment manufactured by the Western Electric Company for the nationwide long distance dialing program.
Transistors, when the invention was first announced, were demonstrated as amplifiers for telephone and television circuits, and to provide the functions of detection and amplification such as are found in an ordinary radio set. A short while later, Bell scientists invented a type which served as a photoelectric device. These and many other applications have been under continued study at the Laboratories.
Cooperating with Dr. Shockley in the original research were the inventors of the point contact transistor, John Bardeen and Walter Brattain. Dr. Shockley, who initiated and directed this general research program, is now in charge of the Laboratories’ investigations in the broad field of transistor physics, and is the author of a recently published book "Electrons and Holes in Semiconductors."
Perhaps the most remarkable feature of these transistors is their ability to operate with exceedingly small power consumption. The best example of this to date is an audio oscillator which requires for a power supply only 6 microamperes at 0.1 volt. This represents 0.6 microwatt of power which contrasts sharply with the million or more microwatts required to heat the cathode of an ordinary receiving-type tube. The power handling capacity, and particularly the efficiency, on the other hand, are high. The design can readily be varied to permit the required amount of power dissipation up to at least two watts. Furthermore, the static characteristics are so nearly ideal that Class A efficiencies of 48 or 49 out of a possible 50 per cent can be realized. The efficiencies for Class B and Class C operation are correspondingly high, reaching as much as 98 per cent.
Not only is the transistor a highly efficient unit, but it is compact and rugged. The transistor is enclosed in a hard plastic bead about 3/16 inch in diameter. Inside the bead three electrical connections are fastened to the germanium and are brought out as "pigtails" through the bead. This gives a very sturdy unit which readily withstands severe shock tests. The input and output impedances are always positive, whether the transistor is connected grounded-emitter, grounded-base or grounded-collector. This permits a great deal of freedom in circuit design and makes it possible, by choosing the appropriate connection, to obtain a considerable variety of input and output impedances.
Vibration tests in the audio frequency range produce no measurable microphonic noise.
Other salient characteristics of the new junction-type transistor are its relatively low noise figure and its high gain. The noise figure is 1000 times less than that of its predecessor.
While studies indicate that collector capacitance limits the frequency response at full gain to a few kilocycles, it is possible by using a suitable impedance mismatch to maintain the frequency response flat to at least one megacycle while still obtaining a useful amount of gain.
At 1000 cps, most of the units measured so far have a noise figure between 10 and 20 db. Power gains of the order of 40 to 50 db per stage have been obtained.
These devices are still undergoing exploratory development. While more complete information on the properties which may be achieved will be available after further development, the results to date are encouraging.
SMEC NOTE
It must have been exciting for K. D. Smith to work with all of the
people mentioned in this story. There are a vast number of history makers
that even may be retired and living only a mile from our residence. It
is all of our jobs to document the history of their achievements while
they still are able to share them with us. - EAS
SMEC UPDATE.
While K.D. Smith was involved with the development of the grown junction
transistor at Bell Laboratories, he had the job of making up development
project status reports. The sheet, that is seen above, is a reduced example
showing at what phase of manufacture a M-1752 transistor would fail.
When you get to the bottom line of this chart, what you will see
is that there was a 6% yield! In talking with Howard Dicken, Ivan Saddler,
Bob Ryder and Jim Early about this low yield, I was told that for an early
development report, the figures shown were normal or better than they had
seen on some projects!
The first 6 of the reports, in whole, are on file at the museum for
you to examine! Not only are there charts, but also pages of production
figures, as well as the list of which people at Bell Laboratories, and
on the outside, were to get samples of these development devices. The list
reads like a book of Who's Who!
( need to add chart)
Left-Grown Junction Right-Point Contact Transistors
William Shockley explains grown junction theory
Gordon Teal Whose efforts with grown single crystals made the grown junction transistors.
Morgan Sparks, who made the first grown junction transistor at Bell Telephone Laboratories
R. L. Wallace (left) balances an AC bridge to measure the impedance of a transistor as W. J. Pietenpol records readings.
|
