Semiconductors and junction diodes

The 1950s marked the beginning of a revolution in electronics. It started with the invention by William Shockley of the transistor, a minute three- terminal device which could switch, amplify and oscillate yet needed only a few microwatts of power; it was also robust and virtually everlasting. Inevitably the transistor replaced the electron tube (valve) in all except very high power applications.

The pace of the revolution was accelerated a decade later by the development of the integrated circuit or i.c. (popularly known as the silicon chip) in which transistors and other components are manu- factured and interconnected by the planar process (see Appendix A) to form amplifiers, signal stores and other functional units on a single silicon slice. The miniaturisation now possible is such that several million transistors can be accommodated on an i.c. less than 1 cm2

The applications of i.c.s seem boundless. They feature in activities as diverse as satellite communication and control of model railways. They are widely used in audio, video and radio equipment and they made possible the computers and microprocessors now universally employed in commerce and industry. Perhaps their most familiar applications are in digital watches, calculators and toys.

Mechanism of semiconduction

As the name suggests a semiconducting material is one with a conductivity lying between that of an insulator and that of a conductor: that is to say one for which the resistivity lies between, say 1012 -cm (a value typical of glass) and 10–6 -cm (approximately the value for copper). Typical values for the resistivity of a semiconducting material lie between 1 and 100 -cm.

Such a value of resistivity could, of course, be obtained by mixing a conductor and an insulator in suitable proportions but the resulting material would not be a semiconductor. Another essential feature of a semiconducting material is that its electrical resistance decreases with increase in temperature over a particular temperature range which is characteristic of the semiconductor. This behaviour contrasts with that of elemental metallic conductors for which the resistance increases with rise in temperature. This is illustrated in Fig. 1.1, which gives curves for a conductor and a semiconductor. The resistance of the conductor increases linearly, whereas that of the semiconductor decreases exponentially, as temperature rises. Over the significant temperature range the relationship between resistance and temperature for a semiconductor could be written

Principles of Transistor Circuits

465 kHz in spite of the need for alignment. They are also used at 10.7 MHz but are often supplemented by ceramic filters which have sharper cut-offs at the edges of the passband and do not require alignment. In television receivers and video recorders the i.f. selectivity is commonly provided by a surface-acoustic-wave filter which is manufactured to have the particular shape of passband required and needs no subsequent adjustment. A ceramic filter is used in the 6 MHz inter-carrier sound i.f. amplifier. The circuit diagram of an i.f. integrated circuit is far too complex to reproduce here but an illustration of the type of circuitry employed can be gained from the diagram of an early and very simple i.c. given in Fig. 10.16. This has only a single stage of amplification. TR1 and TR2 form a cascode amplifier and TR3, the a.g.c. transistor, is in parallel with TR2 and its output load. TR2 and TR3 are biased from the same point on the potential divider R1R2 and TR1 is biased from the diode D1 which ensures constancy of mean current in TR1. When TR3 is forward biased it shunts the output load and reduces the output-signal level. The shunting effect becomes more significant as the positive a.g.c. voltage on TR3 base is increased. This is an effective method of controlling gain and has the merit of leaving the input resistance and input capacitance of the amplifier substantially constant. A block diagram of a typical a.m./f.m. receiver i.c. together with some of the associated external circuitry is given in Fig. 10.17. The i.c. incorporates all the a.m. stages from aerial input to demodulated audio output, together with a four-stage f.m. amplifier/limiter and quadrature synchronous demodulator (see page 236) for f.m. reception. The complete circuit diagram of a receiver using a similar i.c. is given in Fig. 12.21.

Sinusoidal oscillators have widespread applications in electronics. To name only a few obvious examples, they provide the carrier source in transmitters and form part of the frequency changer in superhet receivers. They are used for erasing and biasing in magnetic recording and time the clock pulses in computers. Many electronic measuring instruments incorporate oscillators and a variable-frequency oscillator forms part of a phase-locked-loop detector. There are many different types of sinusoidal oscillator but all consist essentially of two parts:

MHartley oscillator

This is an example of a low-frequency oscillator using an LC circuit for frequency determination and a transistor supplying maintaining pulses. The circuit diagram (Fig. 11.4) shows a common-emitter amplifier with the LC circuit connected between collector and base, the centre tap of the inductor being effectively connected to the emitter (the power supply being regarded as having zero resistance). A common-emitter amplifier inverts its input signal and its output signal is inverted by the centre-point- earthed inductor before it is applied to the base. The circuit can thus be regarded as that of an amplifier which supplies its own input: in other words there is considerable positive feedback and this causes oscillation, the amplitude of the signal (at the resonance frequency of L1C1 ) building up rapidly. The resulting pulses of base current charge up Cb , the polarity of the voltage thus generated biasing the base negatively. As the signal amplitude grows, so does the voltage across Cb until equilibrium is reached when the losses from the LC circuit due to output loading, ohmic resistance and base current equal the energy supplied to it from the collector. In this final condition the transistor may well be cut off for most of each cycle of oscillation, taking a burst of base current (and collector

current) on each positive peak at the base. In the interval between successive peaks Cb begins to discharge through Rb but, if the time constant RbCb is large compared with the period of oscillation, little of the voltage across Cb is lost in the interval and Cb can be regarded as a source of steady negative bias. Such a biasing system is used in many oscillators. It has the advantage of compensating to some extent for any reduction in oscillation amplitude caused, for example, by increased output loading or a fall in supply voltage. Reduction in oscillation amplitude causes reduced bias so that the transistor takes larger current pulses which tend to maintain the amplitude. This is, in fact, an example of class C operation (see page 180).

Colpitts oscillator

It is significant that three connections were needed between tuned circuit and transistor in Fig. 11.4 to provide positive feedback. The emitter connection is effectively connected to the centre point of the inductor but it could equally well be connected to the capacitive branch of the LC circuit by using two equal capacitors in series as shown in Fig. 11.5(a). This oscillator uses a jfet with a resistor Rd in the drain circuit, the LC circuit being coupled to the drain by a capacitor Cd . Thus the LC circuit is shunt-fed by contrast with the direct feed of Fig. 11.4. The frequency-determining capacitors C1 and C2 are in parallel with the input and output capacitances of the transistor which therefore have an effect on oscillation frequency. The effect can be minimised by making C1 and C2 as large as possible. On the other hand if oscillation is required at a high frequency necessitating a very small tuning capacitance, the input and output capacitances of a transistor can be used as C1 and C2 , a small variable capacitor being connected across L for tuning as shown in Fig. 11.5(b). This technique is sometimes used in

v.h.f. and u.h.f. oscillators. Biasing is again automatically provided by Cb which is charged by pulses of gate current and discharges through Rb . To permit the moving vanes of the tuning capacitor (and hence the transistor base) to be earthed, an r.f. choke with high impedance at the operating frequency is included in the emitter circuit. All three oscillators described above operate in class C for large oscillation amplitudes. To obtain a sinusoidal waveform the output must be taken from the LC circuit, e.g. by a coil inductively coupled to the LC circuit as shown in Figs. 11.4 and 11.5. If the output is taken from the transistor itself, e.g. from a resistor in the emitter or source circuit, it would consist of a pulse train with a pulse repetition frequency equal to the resonance frequency of LC. The automatic bias system used in the above oscillators, relying as it does on the flow of base or gate current, clearly cannot be used with mosfets which do not take gate current. Instead bias can be provided by the potential-divider circuit of Fig. 6.7 and, for small oscillation amplitudes, the transistor operates in class A.

Aurther

S. W. Amos, BSc, CEng, MIEE

M. R. James, BSc, CEng, MIEE

Downlaod Book