The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Pole Zero (PZ) analysis
https://designers-guide.org/forum/YaBB.pl?num=1302795231

Message started by analogrf on Apr 14^th, 2011, 8:33am

Title: Pole Zero (PZ) analysis
Post by analogrf on Apr 14^th, 2011, 8:33am

Hi all,

I have a complex system with many poles and zeros. I used spectre's PZ analysis to get the accurate locations for my circuit. However i want the PZ analysis to tell me the associated node of each pole or pole-zero pair.

Is it possible to get to know the associated node of each pole/zero too ??

Best Regards

ARF

Title: Re: Pole Zero (PZ) analysis
Post by Garrett.Neaves on Apr 14^th, 2011, 10:28am

analogrf wrote on Apr 14^th, 2011, 8:33am:

Hi all,
Is it possible to get to know the associated node of each pole/zero too ??

It can occur that a particular pole or zero is associated with a particular node or can be approximately associated with a particular node. However, in general, a specific pole or zero is not associated with a specific node. It is possible to derive transfer functions in symbolic factorized pole-zero form which allows the poles and zeros to be identified in terms of the circuit elements. The derivation is performed primarily with pencil and paper.

Title: Re: Pole Zero (PZ) analysis
Post by analogrf on Apr 14^th, 2011, 11:04am

Thanks indeed for the reply.

You can think of it as this: One can enter the derived small signal schematics in spectre to get a clue of where the poles and zeros are. So lets say the manual analysis reveals 4 poles and 2 zeroes, then spectre also plots the exact 4p and 2z using PZ analysis.
What i wanted to know was, if one can track which respective nodes are each of these coming from (now they are associated with the SS model nodes ! ) . Lets say when they are close to each other, its difficult to predict which one is coming from node x and which one from node y.

Best Regards,

AnRF

Title: Re: Pole Zero (PZ) analysis
Post by Garrett.Neaves on Apr 14^th, 2011, 12:37pm

analogrf wrote on Apr 14^th, 2011, 11:04am:

Thanks indeed for the reply.

What i wanted to know was, if one can track which respective nodes are each of these coming from (now they are associated with the SS model nodes ! ) .

The PZ simulation is always performed on a small signal model. This small signal model can be one created by the simulator from the DC operating point of a circuit with non-zero DC biasing. Or it can be the small signal equivalent circuit entered into the simulator by the user. In either case, if the user manually derives a transfer function, then the user will find that , in general, the poles and zeros of that transfer function do not have a one-to-one correspondence to individual nodes of the circuit. The poles and zeros are, in general, functions of any of the elements in the circuit.

A pole or zero in radians per second is equal to the reciprocal of a time constant. For example, a simple RC time constant. The R can be between a pair of nodes in the circuit and the C can be between a different pair of nodes. In this example, then, their are 4 nodes which correspond to the pole or zero. And therefore, I think the question of tracking which node is associated with a pole or zero can only be answered by explaining that the question contains an incorrect assumption.

There are situations in which both the R and the C are connected between a node and AC ground, and there are situations where this is an appropriate approximation, but, it is not the general case.

Title: Re: Pole Zero (PZ) analysis
Post by raja.cedt on Apr 14^th, 2011, 10:09pm

hi,
plot each nose phase shift, so this gives some clue.

Thanks.