# Loop-gain calculation

1. ## Re: Loop-gain calculation

For more info on this horribly inefficient and outdated method of analysis, see
Apparently the professor did a bad job in conveying the purpose of the analysis. But I absolutely disagree about "inefficient and outdated". Unfortunately many analog designers never learned to analyze feedback topologies...

Here it is. Negative feedback looks fancier than your negative feed-forward ("C").
There are different ways to assign the signs in the scheme. I personally prefer positive A and B factors, having the negative feedback sign located in the substractor only. Regarding C sign, it's an inverting amplifier circuit, I think it makes more sense to switch C sign than both A and B. But I won't insist on my version.

•

2. ## Re: Loop-gain calculation

Originally Posted by mtwieg
I had a professor in undergrad who absolutely insisted that simple circuits like this be analyzed in terms of shunt/series feedback systems. What fun we had, spending half an hour analyzing calculating the gain of an inverting opamp amplifier, which could be solved in seconds using KVL and KCL.

For more info on this horribly inefficient and outdated method of analysis, see: http://cc.ee.nchu.edu.tw/~aiclab/tea...cs3/lect08.pdf
I can tell that up to this days that method is still shown (I know because I am young... very young), so it is not "outdated".

Unfortunately many analog designers never learned to analyze feedback topologies...
And what do they do ? They do mesh analysis like mtwieg is suggesting ?

3. ## Re: Loop-gain calculation

A loop connection would have to include R1, i.e. connection to the LHS of R1, then the roll-off with freq will be included in any "loop" connection.

You can model the internals all you like - but to present them as the complete loop is, respectfully, misleading, and may lead to mis-understandings if other engineers rely on the definition.

•

4. ## Re: Loop-gain calculation

As already mentioned, the circuit can be used in a way that doesn't establish feedback across the outer terminals. If the amplifier cell is an OP IC with high gain and poles, nobody would doubt the usefulness of loop gain calculation for the internal feedback loop, e.g. to determine circuit stability.

Although loop gain can be calculated for the internal feedback loop of this "simple circuit", it's not necessarily a useful analysis step. In so far I understand why mtwieg feels reminded to fruitless exercises during his study.

5. ## Re: Loop-gain calculation

Originally Posted by Easy peasy
A loop connection would have to include R1, i.e. connection to the LHS of R1, then the roll-off with freq will be included in any "loop" connection.
If you would read again the contributions in post#5 and post#19 you will see that - of course - R1 is part of the feedback loop. Remember the feedback scheme of the clasical opamp inverter.

Originally Posted by Easy peasy
You can model the internals all you like - but to present them as the complete loop is, respectfully, misleading, and may lead to mis-understandings if other engineers rely on the definition.
I cannot understand why you speak about "internals" ? There is part/chip called CMOS-Inverter - and this part is used (similar to an opamp) as a linear amplifier.
And for this purpose it needs an external negative feedback path (like the opamp) which contains two separate parts called "resistor".
I repeat: This is a classical external feedback scheme.
What "definition" are you speaking of? This feedback circuit is in full accordance with the definition of a negative feedback path !!
Can you mention one single reason why the shown circuit would not represent a typical system with negative feedback?
I am afraid, this superfluous discussion could confuse some beginners...

6. ## Re: Loop-gain calculation

I saw the posts that included R1, very good up to a point, but if you are seeking the input to output transfer function you have to include the RC roll off ( C from the gates) to get a measure of the performance over freq - just using R's is a bit misleading.

Not an issue if it is just being used in a digital fashion - but for AC linear use the freq roll off has to be considered ...

•

7. ## Re: Loop-gain calculation

Originally Posted by Easy peasy
I saw the posts that included R1, very good up to a point, but if you are seeking the input to output transfer function you have to include the RC roll off ( C from the gates) to get a measure of the performance over freq - just using R's is a bit misleading.
Not an issue if it is just being used in a digital fashion - but for AC linear use the freq roll off has to be considered ...
Just a short comment:
Each active device exhibits a typical roll-off chracteristics for higher frequencies. This is a general rule ...and nobody has denied it.
However, up to now we were discussing the more fundamental question: Feedback yes/no because of your statement in post#9: "...there is no loop in the ckt...".

--[[ ]]--