Output impedance measurement in amplifier

[lawfulgm]

Hi all,

I am a bit confused about this output impedance measurement. When we measure the output impedance, we short the input to ground and apply AC source to the output.
Supposed I want to measure simple NMOS input amplifier with PMOS load followed by a shunt-feedback amplifier's output impedance, should I ground DC bias of the input transistors to the ground as well? what about output ? when I apply AC voltage to the output, should I apply a proper DC voltage as well? or put an AC coupling capacitor b/w AC source and the output node?

I think what I should do is only short input AC source to the GND but apply proper DC bias voltage to the input, and place an AC coupling capacitor b/w the output node and AC source at the output.

What is the correct way to measure this? what about in transient situation then?

Thanks,

 

[KlausST]

Hi,

A picture could help to clarify what you mean.

When we measure the output impedance, we short the input to ground and apply AC source to the output.

Do you want to measure "open loop" or "closed loop" output impedance?

I'd say "closed loop" makes more sense.
For cosed loop you connect +IN to GND and -IN to output.

when I apply AC voltage to the output, should I apply a proper DC voltage as well? or put an AC coupling capacitor b/w AC source and the output node?

It depends. If you are interested in impedance at very low frequencies or DC, then you must not use a capacitor.

Klaus

 

[LvW]

I think what I should do is only short input AC source to the GND but apply proper DC bias voltage to the input, and place an AC coupling capacitor b/w the output node and AC source at the output.

Yes - of course, that`s correct.
The whole amplifier must be DC biased as under normal operational conditions.
Just the signal input must be zero (no open signal input node, but grounded).

 

[FvM]

A straightforward way to measure impedance in AC analysis is to connect an 1A AC current source to the respective node and read the AC voltage. Reasonable bias conditions must be set.

 

[lawfulgm]

Here is a schematic


- - - Updated - - -

Thanks a lot to FvM, LvM, and KlausST !
To KlausST : I want to measure open-loop output impedance measurement. My amplifier will never be used in feedback configuration.

 

[c_mitra]

The most relevant and appropriate measurement will be a differential measurement (dV/dI) under given conditions. The best way to represent the result will be a simple graph.

 

[dick_freebird]

In simulation, you might just apply a looped DC current
source and look at voltage deflection. On the bench I'd
suggest a series resistor and a supply that is swept from
just below to just above the natural common mode output
point, measure dI and dV, and there you go. Presuming
you were after DC output impedance and not impedance-
at-frequency-of-interest.

 

[FvM]

A DC sweep is only necessary if you don't know the correct bias point. If you know it, an AC measurement is the easier way to find the impedance. It's not actually frequency dependent, just choose a sufficient low frequency.

 

[c_mitra]

...Presuming you were after DC output impedance and not impedance-
at-frequency-of-interest.

Unless it is a simple DC amplifier, the DC output impedance will be of little use. In fact, dV/dI will directly give you the complex impedance if the V and I are not in phase. Of course you will need to have a quadrature detection (using part of the excitation input) and the complex output impedance will not only be a function of the input frequency, it will also depend on the signal bias.