Phase Margin Gain Margin

[Debdut]

I have some difficulty is clearing some concepts.

Suppose there is a forward transfer function G(s) and feedback transfer function H(s). Is PM and GM calculated for G(s)H(s) or simply G(s) in order to determine the stability of the system?

For example in two-stage OpAmp we only calculate the PM for the open loop two-stage amplifier i.e. only G(s) to see whether it is having at least 60 degrees PM irrespective of any feedback path.

However in PLLs or Integer-N Synthesizers we calculate the PM of G(s)H(s), G(s) being the Phase-Frequency Detector, the Loop Filter and the Voltage Controlled Oscillator, H(s) being the Frequency Divider.

Reference to PM calculation of PLL I found in "Fourth-order PLL loop filter design technique with invariant natural frequency and phase margin" by I.V. Thompson and P.V. Brennan.

Are the above statements I made correct? Please help me out! :???:

 

[D.A.(Tony)Stewart]

Bode Plots measure the Open Loop Gain with the model that feedback on output Phase is 100%.

that is ... the feedback block can be put in the forward path of the basic loop with unity feedback.

thus G(s) = K/(s*(1+sT)) where K =Kd Ka Ko for detector amp and vco

 

[Debdut]

Thanks for replying SunnySkyguy.

Do you mean that for OpAmp we consider that the worst case feedback is the unity gain feedback and then calculate the PM for G(s)H(s) where H(s)=1?
Can we do the same for PLLs?