What's New in Transistors
June 1954 Radio-Electronics

June 1954 Radio-Electronics [Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio-Electronics, published 1930-1988. All copyrights hereby acknowledged.

As with most new discoveries, advancements came quickly for transistors. A little more than six years after Messrs. Brattain, Shockley, and Bardeen announced their invention of a positive gain, point contact transistor, this article in Radio-Electronics magazine reports on the wonders of junction diodes and transistors that eliminate the mechanical interface of the "cat whisker" that was vulnerable to failure due to shock, vibration, and temperature changes. Note how closely spaced the patent numbers mentioned are for Sidney Darlington's compound transistor (aka a Darlington pair), Shockley's bistable transistor oscillator, Gordon Raisbeck's NPN-PNP balanced pair amplifier, and Robert Blakely's 3-terminal transistor mixer. Bell Labs, IBM, and the other big name research companies and universities were hot pursuit of the next big thing in semiconductor technology. I'm guessing lawyers were making almost as much money as the companies as lawsuits ran amok battling over intellectual property (IP) rights.

What's New in Transistors

Improvements in circuits and in the units themselves accelerate the progress of these newest and "hottest" things in electronics

By I. Queen

Transistors remain the hottest thing in electronics, new applications being continually brought forward. Transistors themselves as well as circuits for them are being improved. Recent developments include higher gain, lower noise, and greater stability. Transistors of the future may possibly be grouped or combined without need for transformers or coupling components. Grouping is possible because crystals may be N or P type, and thus complement each other.

William Shockley of the Bell Laboratories, famous for his work on transistors, has designed a bi-stable circuit (Fig. 1). It is assigned patent No. 2,655,609. NPN and PNP junction types are paired. R1 is the circuit load. Resistors R2 and R3, which may be 100 ohms, aid in providing a trigger action. Ordinarily we associate a trigger effect with point-contact transistors. Shockley has obtained the same result with the less expensive junction types.

A positive signal is applied. When it is low, current through load resistor R1 is small. The voltage drop across R2 and R3 is nearly zero. Since this drop determines the emitter bias for each transistor, each works near cutoff.

If the positive input voltage is increased, more current flows through the circuit. Resistors R2 and R3 produce a greater bias between emitter and base of each semiconductor. The bias is always in the forward direction for each transistor. More bias means more collector flow, and in turn, the emitter bias is increased still further. Soon, each transistor current reaches its saturation value where it remains, and the load current through R1 is maximum. The trigger returns to low conduction when the input voltage is lowered to near zero.

The crystal pair shown in Fig. 1 is equivalent to a single transistor with a current gain A/(1-A) where A is the gain of each individual unit. For example, if each has a gain of 0.9 then the equivalent transistor has a total gain of 9. The emitter of the equivalent transistor is E, its base is B, and its collector is C.

Another patent (No. 2,666,818) recently credited to Shockley is illustrated in Fig. 2. Again he pairs an NPN and PNP to obtain special effects. The result is a circuit that can handle relatively large amounts of power. As before, we show an arrow (in the emitter lead) pointing outward to indicate a NPN unit. The arrow pointing toward the crystal indicates a PNP. The two transistors V1 and V2 form a voltage divider across battery B1. Auxiliary battery B2 biases each transistor in the forward direction, that is, toward lower impedance.

When the input signal is zero, the transistors conduct equally. The output voltage is one-half of B1. Electrons flow out of the collector of V1, equal in amount to the holes drawn from the collector of V2. The same number of electrons are injected into the emitter of V1 as holes injected into the emitter of V2. Thus there is no need for a direct return path for each element.

When the input goes positive, each base receives the positive potential. V1 conductivity increases. At the same time V2 decreases in conductivity. The output voltage is decreased during this time. During the other half-cycle, this process reverses and the output voltage rises. Each transistor contributes toward the power output, yet the circuit needs no transformers, capacitors, or resistors.

Another NPN-PNP balanced pair appears in Fig. 3. It was invented by Gordon Raisbeck who assigned his patent (2,666,819) to the Bell Telephone Laboratories. As before, we indicate the NPN junction transistor by an emitter arrow pointing outward. The other transistor, V2 is a PNP type.

The circuit provides push-pull output without a transformer, and the input signal does not have to be balanced to ground. When the a.c. signal goes positive, the same bias is fed to both emitters. V1 conduction is lowered, while V2 is increased. The load receives power from each transistor. As in any push-pull arrangement, even harmonics are canceled out.

There is no d.c. return to the base. It is not necessary. At any given instant there are as many charges withdrawn from one base as are fed into the other. Base current is always zero. Fig. 3 is an amplifier, but may be connected to modulate, detect, or oscillate.

A single junction transistor may be slotted as shown in Fig. 4-a to form a compound unit. It is equivalent to a pair of transistors, yet formed from a single NPN junction crystal. This unit has a common collector, but separate emitters and bases. A Lead A connects one emitter (upper N region at the left) with the other base. Leads E, B, and C are connected to the terminals of the equivalent transistor.

The equivalent transistor (Fig. 4-b) has an unusually high alpha or current-gain factor. It is equal to 1 - (1 - A)². For example, if an unslotted transistor has an alpha of 0.9 the compound unit has an alpha of 0.99. Theoretical maximum for a junction crystal is unity.

With suitable slotting, the single NPN semiconductor can be made equivalent to a triple compound transistor, as shown in Fig. 4-c. In addition to the leads E, B, and C, two others are brought out for biasing purposes. R1, R2 are chosen for optimum gain and low idling current.

This compound transistor is credited to Sidney Darlington (Patent No. 2,663,806) and is assigned to Bell Labs.

One disadvantage of the previous transistor is its high collector current when emitter bias is zero. This represents a power loss and may be highly undesirable in some circuits. Bernard M. Oliver has discovered a means of solving the problem. It is disclosed in patent No. 2,663,830, assigned to Bell Telephone Laboratories. He finds that the idling current is minimized if the slots are cut as described here.

Fig. 5-a shows a double compound unit suitably slotted. The areas of the emitters (and bases) are unequal. The ratio should be 1:1-A, where A is the current gain of the unslotted semiconductor. If A is 0.9, the slotted areas should have a 10:1 ratio. Fig. 5-b illustrates a triple compound transistor slotted in accordance with this patent. Slot 1 is the first slot which gives the equivalent of two transistors. A second slot makes the unit equivalent to three separate transistors with a common collector C. Lead 1 connects the base of the first transistor with the emitter of the second. Lead 2 connects the base of the second transistor with the emitter of the third. The analog of this compound transistor is shown in Fig. 5-c. The area of V2 is smaller than that of V1. V3 is still smaller.

Other inventors have added to the usefulness of a point-contact transistor by using more than one emitter or base contact. A crystal tetrode has been invented by Robert T. Blakely, patent No. 2,666,150, and assigned to International Business Machines Corp. of New York (Fig. 6).

This tetrode gives the same effect as two separate transistors. Each emitter is fed from an input source, and each provides gain. Thus it is useful as a mixer.

Posted March 9, 2021