a fundamental but tricky question on feedback

[sutapanaki]

There were many good and true things said above about this type of configuration and its analysis. I think the difficulty the initial poster had was with the impossibility of fitting the inverting amplifier topology within the traditional feedback block diagram, because it doesn't really fit. The non-inverting amplifier, of course fits readily in it. I can never remember the classification of feedbacks in shunt-shunt, shunt-series, etc. I prefer an approach that can analyze any given circuit. It is also true that there is a feed-forward path here, but in case of opamps, with high enough gain and at sufficiently low frequencies, I don't think we have to usually bother about it.
In the attached file I show how this non-inverting amplifier can fit in the general feed-back topology. From that block diagram for very large gain, one can get the approximate formula of -R2/R1.

 

[LvW]

Of course, I agree with sutapanaki.
However, the whole story can be explained in three simple sentences:

1.) The general feedback formula is based on a model which has a differential input (subtractor) followed by the gain stage with a feedback block β) and this model fits the noninverting opamp since the input voltage is applied directly to the "subtractor" input.

2.) If there is a voltage divider Hf=R2/(R1+R2) in front of the differential input (subtractor) the input signal is reduced by the factor Hf before it reaches the classical model input.(Hint: The divider formula assumes opamp output zero - in accordance with the superposition theorem).

3.) Thus: Vout=(1/ β)*Hf*Vin=-R2/R1*Vin (with β>>negative because of neg. feedback)

 

[yschuang]

Maybe the following link is helpful, which says about feedback.

**broken link removed**