Difference in the reactive part between loadpull simulation and large signal sparams

In SpectreRF, there is an option to run large signal s-parameters. I noticed that the reactive part of the impedance looking into the output port is often quite different from the reactive part of a loadpull simulation, especially at mmwave frequencies. I understand why the real part will be (and needs to!) be different between the two simulations but never got why the reactive part is different.

What do you want to mean ?

Former is a setting by you.
Latter is set internally by simulator.
So what do you compare ?

Looking into the amplifier output is what I meant. Sorry for the confusion

How do you evaluate output impedance in load pull simulation ?
Show me equations you use.

And drive directions are different between LSSP and LoadPull.
Former is reverse drive.
Latter is forward drive.
Do you understand ?

See https://www.edaboard.com/showthread.php?357126#5

By output impedance, I mean the impedance that the amplifier needs to see for optimum performance (power ,gain, PAE etc).


Let me re-phrase my question. Why is there a difference between the imaginary part of the optimum impedance from loadpull and the conjugate of the imaginary part of the impedance?

And how do I measure the output impedance seen looking into a PA?

1. I can use sp simulation in SpectreRF and look at ZM2 (of course I have ensured that the input port has the correct impedance along with the proper matching network). But at large signals, checking output impedance is not fully valid.
2. I can use large signal sp simulation. At low output powers (where the amplifier is more in 'small signal'), the large signal s-parameter results and the small signal s-parameters are very close.

My amplifier is a differential structure with feedforward cancellation (I check this by checking s12).

The way I understand is that the capacitance at the output of the amplifier needs to be tuned out. Which means that reactance of the optimum impedance from the loadpull simulations should be close to the complex conjugate of the reactance part of the output impedance of the amplifier.

So what am I missing here?

It is not an output impedance.
It is load impedance.

Use correct terminology.

Can you understand port impedance, load impedance and output impedance correctly ?

What do you want to mean by "the impedance"?

If you mean output impedance correrctly, there is no direct method.
See https://www.edaboard.com/showthread.php?357126#5

It is very natural result, since it is small signal analysis.
What do you want to mean ?
And sp analysis is not SpectreRF.

It is very natural result.

Still you can not understand drive ditrection at all.
LoadPull drives amplifier in forward direction.
LSSP for S22 and S12 drives amplifier in reverse direction.

More over, boundary conditions(=termination conditions) are different between LoadPull and LSSP, even if drive directions are same.
So voltage and current are different.

Surely see https://www.edaboard.com/showthread.php?357126#5

Can you point to some documents that talk in somewhat more details about the difference in termination conditions?







Can you point to some documents that talk in somewhat more details about the difference in termination conditions?

Answer my question.
It is too elementary thing.
It is mathematical basic about uniqueness of solution.

Before termination condition, can you understand drive direction's difference ?

This is a termination condition of input side.

You don't care about termination condition of output side at all.

Can you understand termination condition of output side for LoadPull and LSSP ?

I think you set 50ohm for LSSP.
However it is not 50ohm for LoadPull.
Can you understand ?

vivekroy,

You confuse a thing..
Driving an amplifier naturally and finding internal source impedance ( The Impedance which is found by Load Pull max. Power Output ) is NOT same as driving this amplifier externally by a source with 50 Ohm internal impedance by whatever the signal level is.