Fig. 21. A tuned-base oscillator with feedback from the collector winding. A better impedance match to the base of the transistor may be obtained by using the tapped inductor as shown in B.

for example, a series LC combination is inserted in both the base and collector leads, which together form the necessary resonant circuit. Fig, 20 shows an oscillator with the tank circuit connected between the collector and base with feedback taken oil the tank circuit by the in-phase feedback winding and applied to the emitter.

In the circuit in Fig. 21, the tuned tank circuit is placed between the base and emitter (through ac ground) and out of phase feedback picked up from the collector winding. In Fig, 21B a better impedance match to the base of the transistor is obtained by tapping down on the tuned-circuit inductance,

Most of these simple circuits have one very serious disadvantage—the frequency of oscillation is very dependent upon the collector resistance of the transistor. There is some dependence on the transistor characteristics in all oscillator circuits, but in the circuits of Fig. 20 and 21? the influence of the transistor predominates.

Colpitts or Hartley?

One of the big questions that invariably arises is what circuit to use in a specific application. In many cases the Clapp oscillator is chosen, particularly for VFO's, because in practice stability is somewhat easier to obtain. For other applications though, both the Hartley and Colpitts find favor. Between these two the choice is more difficult, However, as a rule of thumb, the Hartley is more satisfactory at the lower frequencies, while the Colpitts works best in the high-frequency and VHF range. The reasons for this are quite complex, but they can be explained fairly simply with a couple of block diagrams.

In all transistor oscillators where the tuned circuit is connected between the collector and emitter, the feedback network may be represented by X*, Xb and X* as shown in Fig. 22. Neglecting circuit losses, the only way that the voltage across Zi (the base input impedance) can be precisely 180° out of phase with Zl (collector impedance) is for the reactance X» to be opposite from the reactances Xb and X* and for X. to be equal to the sum of Xb and Xc. In the Colpitts oscillator Xb and Xc are inductors and X* is a capacitor. When the circuit losses are neglected, at resonance the input impedance of the feedback network in Fig. 22A appears as an infinitely large resistance*


Fig + 22. The current flow in the feedback paths of the Colpltts and Hartley oscillators is shown in B and C. The genera) case fs illustrated in A.

However, in a practical circuit there is base loading, series resistance in the tank coil, loading due to power being coupled from the circuit and the impedance is not purely resistive. In a practical Colpitis oscillator for example, the feedback circuit would be represented as shown in Fig. 22B; Zl is the output impedance of the transistor, while Zi is the input impedance, In this circuit currents ii and i:? are not exactly the same magnitude or of opposite phase as in the ideal lossless circuit. The loading of the base circuit (Zi) causes the current i* to fog the collector iriving voltage across Zl by less than 90°; hence the base driving voltage lags the collector driving voltage by something less than 180°.

On the other hand, in the Hartley circuit represented in Fig. 22C, the base loading causes the current is to lead the base driving current by less than 90and therefore the base driving voltage leads the collector voltage by less than 180°.

In the Hartley oscillator the effect of circuit losses and transit times are accumulative, but in the c olpitts circuit these effects tend to offset one another. The fact that the base driving voltage through -he Colpitis feedback circuit lags the collector voltage par-


Fig. 23, Connecting a zener diode across the tank circuit of a Clapp oscillator to obtain output stability.


Fig. 23, Connecting a zener diode across the tank circuit of a Clapp oscillator to obtain output stability.

tiallv compensates for the effects of transit lime. For this reason the ('olpitts oscillator is somewhat superior lo the Hartley circuit at the high and very high frequencies.

Oscillator requirements

The requirements on any oscillator circuit are varied, bill in addition to frequency stability, there are several important characteristics which serve to specifv the performance of any particular circuit. Perhaps most important of these are amplitude stability, harmonic content, output power level, efficiency and noise output.

Amplitude stability

It is usually desirable for the output signal to remain constant within certain limits as the transistor and other components age during operation. This is particularly a problem with variable frequency oscillators, but fortunately the output may be stabilized in most cases by one of the techniques described below. Theoretically, the amplitude of the voltages and currents in an oscillator will become infinite unless some limiting action occurs somewhere in the circuit. I he nonlinearities which w ill limit the amplitude ol the output in a practical circuit are:

1. Limitation of the available dc voltage or current by the capability of the supply.

2. Nonlinearities in the transistor,

3* Nonlinearities in external loads. This is true because the external loads are often a function of voltage and current; above a certain amplitude their values change so that the condition for oscillation is no longer satisfied.

In any case, the amplitude of oscillation will build up until it is liirtited by one of these three basic limiting mechanisms.

When the oscillator is designed for high efficiency and (lie output voltage nearly equals the supply voltage, variations in the supply will cause fluctuations in output power. These fluctuations may be eliminated by stabilizing the supply voltage with a zener diode.

Another technique wliich is slightly more sophisticated has been successfully applied to variable frequency oscillators (Fig, 23), Here the output is compared to a reference diode and the difference fed back to the transistor. Whenever the ac voltage across the capacitor in this Clapp oscillatoi exceeds a value determined by the variable resistor R3, the diode conducts a compensating base current and reduces the output amplitude.

Harmonic content

In many applications it is desirable to restrict the output power of the oscillator to one single frequency. In other cases harmonics are desirable for frequency multiplying. I here are always certain nun linearities in an oscillator circuit which give rise to signals as multiples of the fundamental. The harmonic content depends on many factors and is as difficult to control as is stability, but primarily it is dependent upon the non-linearities in the circuit and the filtering action of the tank capacitance, [f very low distortion is desired, push-pull operation in a two-transistor oscillator may be advised. On the other hand, nonlinearities may be deliberately used to produce frequency multiplication. rhis is done In incorporating another tank circuit into the oscillator which is tuned to the desired harmonic.

Output power level he maximum power output from an oscillator is important in many cases, as well as the maximum voltages and currents available within the limitations of stabilitv and har-

monic content. The requirements for frequency stability and harmonic content are closely connected with the power, voltage and current-handling characteristics of the transistor used in the circuit.

The conversion efficiency ol the transistor oscillator depends primarily on the class of operation and increases as you go from class A to B to C, However, the circuit must be initially biased somewhere in the active re gion to insure that the oscillator will be self-starting. With most transistors, efficiencies of about 50% in class A, 7H% in class B and 80-905S in class C may be expected.

A bypassed emitter resistor permits class C operation in a maimer somewhat similar to the . id-leak method used with vacuum-tube oscillators. An average voltage builds up across the emitter RC combination that provides reverse bias for the emitter diode, With an initial operating point near cutoff, rising oscillations will first result in clipping at the low-current end of the load line, and eventually the buildup will be limited by the nonlinearities at the high current end; the operating point will eventually He in the cutoff region.

The efficiency of an oscillator is reduced by the dc losses in the resistors of the bias circuit and is tied in very closely with the required operating point stability and ease in gelling I he oscillator started. AC losses in the resonant tank also reduce efficiency, and a high unloaded 0 in the tank circuit is desirable, o obtain high efficiency it is necessaiy in all classes of operation to utilize as much of I he available dc supply voltage as possible, with the peak ac collector voltage being equal to approximately 90% of the supply voltage.

In addition, the output power delivered by i:ie transistor to the tank and load must be high. This means that the load impedance seen by the transistor must be designed to be as close as possible to the matching impedance for !he transistor. If the output power from the oscillator is specified, then the supply voltage should be only a few percent higher than the ac voltage swing necessary to deliver I be required power into this approximately matched load impedance*

Unfortunately, the requirement for high efficiency will lead to low Q of the loaded tank circuit. This may cause poor frequency stability and a compromise must be found.

Noise output

In many applications it is very important to keep the noise power from the oscillator at a minimum. This is particularly true in VI i converters where a minimum of noise should be injected into the mixer.

Noise in (lie transistor also effects frequency stability—the initiation of oscillation is a result of thermal and other forms of noise shock-exciting the oscillator circuit,

0 0

Post a comment