Common-Mode distortion – by Dave Kimber

Feedback summing point common-mode gain


Negative feedback, when properly used, can significantly improve the performance of an amplifier. The emphasis must be on "properly used", as improper use of feedback can make things worse. A common mistake is to use too little feedback, either because the open-loop amplifier has high distortion or because of a mistaken belief that high amounts of feedback are necessarily a bad idea. In addition, stability will always be an issue.

A factor which is rarely mentioned, and often overlooked, is the behaviour of the summing point. The standard theory assumes that the amplifier responds solely to the difference between the input signal and the feedback signal. Any other influence is regarded as distortion which is reduced by the feedback. There is a flaw in this theory, although fortunately it turns out in most cases to have only a small effect.

In many cases the input and feedback signals are not directly subtracted, with the difference then fed into the amplifier, but applied to different input points. A typical solid-state amplifier will start with a long-tail pair (LTP), with input on one side and feedback on the other. A valve amplifier may have feedback applied at the cathode of the input stage. In both cases the main effect of the input circuit is to respond to the difference between the input and feedback signals, but there will also be a smaller response to the sum of these signals. This common-mode response will alter the closed-loop gain, and may introduce distortion which is not reduced by feedback.

Feedback theory

Feedback amplifier
As usual, A is the forward open-loop gain of amplifier. b is the feedback fraction. Vin and Vout are the input and output voltages. The extra factor, often omitted, is c which is the common-mode rejection ratio (CMRR). First find the output in terms of the differential and common-mode inputs:

Continues HERE