# Effect of negative feedback on Amplifier

The above problem was asked in a test. I got the wrong answer because I solved it in the following way.

BW without FB = 1000 - 100 = 900

BW with FB = BW without FB * (1+A*beta) = 900 * 100 = 90kHz

I understand that both both the formulae are different and both can't be right at the same time. But my book mentions below 3 formulae bcz of which I got confused in the test.

BW with FB = BW without FB * (1+A*beta)

fH with FB = fH without FB * (1+A*beta)

fL with FB = fL without FB / (1+A*beta)

Please guide me on this, which formulae to use ?

If the amplifier is capacitively coupled this affects the lower frequencies that can pass. As originally quoted the low frequency cut-off point is 100 Hz and this is where I have a problem with the solution. No matter what extra applied feedback is introduced, that LF cut-off is going to remain at 100 Hz.

Think of an inverting op-amp configuration: -

The LF cut-off is dictated by C and R1 attached to the inverting input (a virtual earth) and you can reduce R2 as much as you want and it won't alter this fact.

How about a non-inverting amplifier: -

In this type of amplifier the LF cut-off is defined by C and Rbias and no amount of fiddling with R1 and R2 is going to alter this.

So, for the amplifiers I have shown above, the LF cut-off remains at 100 Hz.

For the HF cut-off, if the gain is reduced (by feedback) from 1000 (60 dB) to 10 (20 dB) and it is assumed that the gain-roll-off slope is 6 dB per octave (20 dB per decade) then you get a new HF cut-off of 100 kHz: -

• These guys ask just formula-based questions with no relation to the practical world. – Nikhil Kashyap Jan 22 '18 at 10:54
• @NikhilKashyap it is a badly conceived problem in my book and no-matter how hard I think, I can't see any scenario where LF cut-off is going to be affected. – Andy aka Jan 22 '18 at 12:04
• If the low-freq cutoff were INSIDE THE LOOP, behavior would change. – analogsystemsrf Jan 23 '18 at 19:09
• @analogsystemsrf yeah I know but can you think of a configuration like this? – Andy aka Jan 23 '18 at 19:16