Loading Effect in Feedback

[Zhiheng Cao]

I am a student studying analog circuits and I am having hard time
understanding loading effect in feedback.

Some books say when we cut the loop, we need to load
the output of the opamp with feedback network beta and the opamp input
behind it, but we can set the voltage (by the feedback) at the opamp
input using an ideal voltage source.

Another book(Razavi's Analog CMOS book) says, that first
we calculate the input impedance of beta with beta's output open.
Then we load opamp output with this impedance. Then we calculate
the output impedance of beta with beta's input shorted to ground.
Then we add this impedance in series to the opamp input impedance.
This is only for voltage-voltage feedback.

Is the first method not as good as the second method or are they
essentially the same?

Another thing I have doubt is that when input/output impedance is
modified by feedback, I suppose we must include the impedance added by
feedback network beta, and then multiply (1+Ab) or 1/(1+Ab).
But for the first method, it seems that the impedance contributed
by feedback network beta at the opamp input is not taken into account,
and that different impedance is added at the output, since the output
is loaded with beta plus a opamp behind it, whereas in the second
method it is loaded with beta with its output open.

I am sorry if my explanation is too confusing.

Thank you for reading my post.

 

[Jim Thompson]

I am a student studying analog circuits and I am having hard time
understanding loading effect in feedback.

Some books say when we cut the loop, we need to load
the output of the opamp with feedback network beta and the opamp input
behind it, but we can set the voltage (by the feedback) at the opamp
input using an ideal voltage source.
[snip]

I am sorry if my explanation is too confusing.

Thank you for reading my post.

Look up Middlebrook's method, where the loop is NOT broken.

Then study the LOOPGAIN part on my website. The part implements the
Middlebrook procedure for use in a Spice simulator.

...Jim Thompson

 

[Zhiheng Cao]

Thanks for the pointer, but can one calculate
terminal impedance modified by feedback using the Middlebrook's
method? As far as I know that method is used to find out the
open loop gain Ab.

I am a student studying analog circuits and I am having hard time
understanding loading effect in feedback.

Some books say when we cut the loop, we need to load
the output of the opamp with feedback network beta and the opamp input
behind it, but we can set the voltage (by the feedback) at the opamp
input using an ideal voltage source.
[snip]

I am sorry if my explanation is too confusing.

Thank you for reading my post.

Look up Middlebrook's method, where the loop is NOT broken.

Then study the LOOPGAIN part on my website. The part implements the
Middlebrook procedure for use in a Spice simulator.

...Jim Thompson