## Advantages Of Inverse Feedback

IN AUDIO AMPLIFIERS: A beam power amplifier to be truly modern should incorporate inverse feedback. It is a commonly recognized fact that low plate resistance tubes such as the 2A3 are superior from the standpoint of low distortion and good quality. With inverse feedback the high plate resistance beam power tube may be made to take on the characteristics of the low-mu triode, yet retain most of its high power sensitivity. The important advantages obtained by the use of inverse feedback are as follows: first, reduction of wave form distortion; second, improvement of frequency response; third, reduction of hum; fourth, reduction of "hangover" effect; and fifth, feedback provides a convenient means of altering the frequency response or providing tone control action. The only disadvantage of inverse feedback lies in the fact that the gain is considerably reduced.

EXPLANATION OF INVERSE FEEDBACK: In the circuit of Fig. 1 a certain amount of the voltage developed in the plate circuit is fed back out of phase with the signal in the grid circuit. If without inverse feedback, a certain voltage Eo is developed across the output circuit with an ihput voltage Ei the gain of the stage is E0 divided by Ej. If now a certain percentage N of the voltage E0 is fed back to the grid circuit in such a way that the voltage is out of phase with the input voltage Ei the total input voltage to obtain an output voltage of E0 is (N E0 + Ei) and gain of the stage is E0/ (N Eo + Ej). The ratio N is the percentage of the output voltage which is fed back to the input circuit. It may be readily seen that if N is large, the gain of the stage depends more upon N than upon the circuit constants.

The ratio reduction in gain by the addition of inverse feedback may be readily determined by dividing the gain without feedback by the gain with feedback.

Fig. 1 illustrates a common method of applying inverse feedbaik across a single ended output stage. Fig. 2 shows a method of applying feedback across a push pull output stage using a resistor condenser network. The amount of inverse feedback is equal to Ri/(Rl+R2) assuming that the reactance of the condenser Ci is negligible over the operating frequencies. The application of inverse feedback need not be limited to one stage. The feedback loop can include several stages as employed in the 25 watt amplifier described on page 146. Here a portion of the output voltage from the secondary of the output transformer is applied back in the proper phase to the cathode circuit of the mixer stage. Thus all of the stages within the feedback loop obtain the above mentioned advantages that inverse feedback affords.

Fig. 1 illustrates a common method of applying inverse feedbaik across a single ended output stage. Fig. 2 shows a method of applying feedback across a push pull output stage using a resistor condenser network. The amount of inverse feedback is equal to Ri/(Rl+R2) assuming that the reactance of the condenser Ci is negligible over the operating frequencies. The application of inverse feedback need not be limited to one stage. The feedback loop can include several stages as employed in the 25 watt amplifier described on page 146. Here a portion of the output voltage from the secondary of the output transformer is applied back in the proper phase to the cathode circuit of the mixer stage. Thus all of the stages within the feedback loop obtain the above mentioned advantages that inverse feedback affords.

REDUCTION OF DISTORTION: As was pointed out in the preceding paragraph, an inverse feedback circuit feeds back a certain portion of the output voltage to the grid circuit. If distortion is introduced in the amplifier stage, a certain amount of the distorted voltage will be fed back into the grid circuit and this will tend to cancel out the distortion developed in the amplifier stage. If in the circuit of Fig. 1 a certain amount of distortion voltage B is present in the output circuit, the distortion voltage fed into the grid circuit will be N x B and this quantity multiplied by the gain of the stage will give the cancelling effect of the inverse feedback. The total distortion present in the output is then equal to the sum of the distortion without inverse feedback and the distortion cancelled by the inverse feedback. In other words, if b is the distortion without inverse feedback, the total distortion, B, with inverse feedback is equal to(b+B)xNxA, where A is the gain of the stage. Evaluating B gives the quantity b/( 1+NA). In other words the distortion is reduced by the ratio of 1/(1+NA).

REDUCTION OF PLATE RESISTANCE: In addition to the reduction in distortion obtained by inverse feedback, there is also a reduction in the effective plate resistance of the tubes. A high plate resistance is a definite disadvantage in the case of a power tube which operates into a speaker load which is more or less variable depending upon the impedance of the voice coil. In the circuit of Fig. 3, it may be easily seen that the voltage E developed across the load depends a great deal upon the actual value of R^ which is the reflected impedance of the voice coil. This is due to the fact that the signal current depends almost entirely upon the high plate resistance of the tube. Since the load resistance is low in comparison to the plate resistance, the voltage developed across the load is almost directly proportional to the impedance of the load which varies appreciably with change in frequency. In Fig. 4 it may be seen that the voltage across the load does not vary so much since the signal current depends both upon the load and upon the plate resistance of the tube. If the voice coil has an appreciable amount of reactance, the impedance rises with the frequency, causing distortion and giving an unnatural amount of "highs". The high plate resistance is unsuitable from another viewpoint, that of the amount of low frequency distortion which may be tolerated. This low frequency distortion is not due to the tubes which remain unchanged regardless of the frequency, but depends upon the magnetizing current in the output transformer. The magnetizing current is a distorted nonsinusoidal wave and this current, on flowing through the high plate resistance of the tube, develops a nonsinusoidal voltage drop across the tube which, when subtracted from the input signal, results in a distorted wave across the output. Unfortunately, most amplifiers today are measured for distortion at 400 cps where the magnetizing current is practically negligible. It is not uncommon to find beam power amplifiers without inverse feedback which have only 25 per cent of the rated power at 40 or 50 cycles. This low frequency distortion is particularly objectionable since all harmonics fall within the audible range. Inverse feedback effectively reduces the plate resistance so that the distorted voltage drop caused by the magnetizing current is exceedingly small, with the result that there is very little distortion across the output circuit. With a poor output transformer it is quite possible for the distortion to be as high as 30 per cent at 40 cycles without inverse feedback.

## Post a comment