|April 1952 Radio & Television News|
These articles are scanned and OCRed from old editions of the Radio & Television News magazine. Here is a list of the Radio & Television News articles I have already posted. All copyrights are hereby acknowledged.
In April of 1952 when this article appeared in Radio & Television News magazine, the bipolar junction transistor (BJT) had only made it out of the experimental laboratory of Messrs. Bardeen, Shockley, and Brattain at Bell Labs a mere three years earlier in December of 1948. It did not take long for commercial production to begin. Along with being a great primer for anyone new to transistors, herein is also some background on how the now ubiquitous BJT schematic symbol was created. Interestingly, only Dr. William Shockley is mentioned, making me wonder whether the contributions of Dr. John Bardeen, and Dr. Walter H. Brattain was not widely publicized early on. Not to worry, though, because all three were duly recognized as recipients of the 1956 Nobel Prize in Physics.
H. S. Renne
Technical Editor Radio & Television News
Details on Bell Laboratories' newly-developed unit and a discussion of some potential uses.
There was a tremendous amount of excitement among electronic engineers when the first point-contact transistor was announced by Bell Telephone Laboratories some four years ago. This initial enthusiasm, however, was soon replaced by a great deal of hard work at the laboratories where efforts were made to iron out the "bugs" existing in the original design, and to ready the device for mass production. These goals have now been reached and reliable, uniform point-contact transistors are in limited production.
The original interest in transistors was given a new stimulus last July with the announcement of the junction transistor. Based on theoretical work at Bell Laboratories by Dr. William Shockley, this new device possessed qualities which seemed to indicate that in many respects it would prove to be superior to the point-contact type of transistor. Intensive research and development work is now in progress all over the country to determine its limitations and capabilities and to prepare it for mass production.
Transistors are constructed of a metal called germanium, which is classed as a semiconductor. Conduction of electricity through germanium can take place either by a .flow of electrons or by a flow of so-called "holes" which are, in effect, an absence of electrons. The result is similar to what would occur if an electron were replaced by a positive charge. The type of conduction is determined by very slight amounts of impurities in the germanium. If these impurities are such that conduction takes place by means of electrons, the germanium is of the "N" type. If conduction is by means of "holes," the germanium is known as the "P" type.
The basic construction of the junction transistor is shown in Fig. 1. It consists essentially of a very thin wafer of "P" type germanium cut from a single crystal and placed between two tiny bars of "N" type germanium, also cut from a single crystal. Because of this construction, the junction transistor is often referred to as the "NPN" transistor.
This new device has many characteristics which entitle it to be called the first serious rival to the vacuum tube. Its amplifying properties are in many respects superior to conventional tubes, and it is far more rugged with respect to shock and vibration than any known tube. It is much smaller (about half the size of a pea), and has an expected service life greatly exceeding that conventional tubes.
Fig. 3 shows the form of presentation which has been adopted for the junction transistor. The base is the center or "P" type section of germanium; the collector and emitter are the "N" type sections.
Power amplifications as high as 51 db (100,000 times) have been achieve! with specific units, and it is reasonable to assume that this figure can be approached or even exceeded in production units. The actual power output rating of the transistor itself is rather small, depending primarily on the cross-sectional area of the germanium at the junction. Most of the experimental units use a germanium rod about a sixty-fourth of an inch in diameter, and have an output rating on the order of 50 milliwatts, heat dissipation within the unit being the limiting factor. One transistor has been assembled with a cross-sectional area of a square centimeter which has a rating of two watts, and still higher-powered units are possible.
A variety of input and output impedances may be achieved by connecting the transistor in different ways. Fig. 2 shows the three possible connections, together with a corresponding practical circuit for each.
The arrangement in Fig. 2A is called the grounded-base circuit, characterized by a low input impedance and a high output impedance. Typical values would be from about 50 to 250 ohms for the former, and from 1.5 to 13.5 megohms for the latter. These values depend to quite a large extent on the construction of the transistor, as well as the amount and kinds of impurities present in the germanium. Power gains of 40 to 50 db can readily be achieved with this arrangement when impedances are matched, and appreciable gains can be obtained with a load resistance of only a few thousand ohms. One advantage of the latter arrangement is that the gain is almost completely independent of those transistor properties which vary from unit to unit. When amplifier stages of this type are cascaded, a matching transformer should be used.
The second type of connection, illustrated in Fig. 2B, is known as the grounded-emitter circuit. This is the most desirable circuit for many applications. Input impedance is higher and output impedance much lower than for the grounded-base arrangement, typical values being 250 to 1500 ohms for the input impedance, and 250,000 ohms to 1.5 megohms for the output impedance. Maximum available gain is over 50 db. In the practical circuit of Fig. 2B, the base will float at a certain potential. If a different potential is desired, various arrangements for biasing may be used. A two-stage amplifier utilizing this circuit and having a power gain of about 90 db is shown in Fig. 4. This amplifier is pictured at the right of Fig. 5.
The power gain available from the grounded-collector stage, shown in Fig. 2C, is rather low, about 15 to 20 db, but the circuit has certain other advantages which make it desirable. Input impedance is high although varying widely with the load resistance. Output impedance likewise depends upon input loading, but in general is quite low. Thus, this circuit arrangement takes on the characteristics of a cathode follower, and if the source impedance is on the order of a few thousand ohms, the output impedance may be 25 ohms or less. Audio enthusiasts would undoubtedly like to use this device to drive a voice coil direct, without using an output transformer!
The junction transistor is especially suited for use at very low power levels. Its efficiency is exceptionally high, closely approaching the theoretical maximum of 50% for Class A and 100% for Class C amplification. Thus, the input battery power required can be extremely small. For example, an audio oscillator (shown at left, Fig. 5) has been built which operates satisfactorily with an input power of only 0.08 microwatt, consisting of 50 millivolts at 1.5 microamperes! To demonstrate the small power required, Mr. R. L. Wallace, of Bell Laboratories, powered this oscillator with the output of a photovoltaic cell exposed to room illumination. In another experiment, he formed a battery by wrapping a dime in a piece of paper which he had previously moistened by chewing on it. The moist paper became one electrode of the battery, and the dime the other. In both cases, the output was sufficient to be audible from an ordinary headphone.
Mr. Wallace also performed some interesting calculations to show that this device could really be called a "flea-power" device. Assuming a dog flea weighing one milligram and jumping to a height of 50 centimeters, both of which are reasonable figures, he calculated that the flea, in making one such jump every minute, would use approximately the same amount of energy as the minimum required to keep the oscillator functioning! It is logical to assume, therefore, that satisfactory amplifiers can be built which would operate on comparably small amounts of input power.
It should not be supposed from the previous discussion that this transistor is a cure-all for everything. There are disadvantages which mayor may not be overcome in future development work. One is the problem of frequency response. Maximum gain of the junction transistor is attained at frequencies on the order of a few kilocycles, with the gain falling off rapidly as the frequency increases. (Usable gains up to one megacycle and more have been realized, however.) One factor affecting frequency response is the transit time of the electrons and "holes" through the "P" layer. Making this layer thinner should increase the frequency response, but it would also increase the capacity between the two "N" electrodes. A thickness of about a thousandth of an inch has been used for this "P" layer.
Another problem is that of the mass production of these items. The amount of impurities in the germanium must be controlled very carefully, as variations of as little as one part in one hundred million can change the properties of the transistor. Also, there is the mechanical problem of fastening leads to such small pieces of germanium. How this is done on present samples has not been revealed. The units must be made to have uniform characteristics in production, so that a transistor in a piece of equipment may be replaced by another transistor without re-engineering the whole circuit. Samples which have been tested have been found to be very temperature-sensitive, although operation is completely satisfactory over normal temperature ranges.
These problems are gradually being overcome and it is expected that production samples on a limited basis will be available not too many months hence.
Here is a device which most certainly will give the vacuum tube a run for its money - if not completely replace it - in the majority of applications. It has practically no heat loss, making. it useful in applications where very large numbers are required, such as electronic computers of various kinds. It is extremely rugged, reliable, and has an almost indefinite life. A typical value of transconductance is 33,000 micromhos per milliampere, and of amplification factor, 39,000. Since the output impedance of a cathode follower is 1/Gm, an output impedance of 5 ohms is possible with a current of 5 ma. The Gm is inversely proportional to the absolute temperature, so that fantastically large values can be achieved by supercooling. Its noise figure is better than some vacuum tubes, being as low as 10 db and less on some of the units which have been tested.
These characteristics are certainly sufficient to justify the excitement and anticipation brought about by the announcement that such a device had been developed. Lest anyone be misled, it would be well to emphasize that this device is not commercially available at the present time, but commercial production is definitely planned for the near future.
Posted January 5, 2016