Well transient is always the most reliable simulation, but I am astonished by the difference. What accuracy are you running your PSS? Conservative? I heard there is some dB difference between conservative, moderate and liberal?

I have tried three accuracy defaults in PSS. The conversion gain is very close to each other.

I am using "Vout/Vin" as transient gain.

My friend told me his PAC result was the same as transient result in single balance mixer simulation.

Hi.

For PAC, do you use gain definition of "Vout/Vin" which is same definition for Transient Analysis ?

For simulator accuarcy or difficulty , there is no difference between single or double balanced as far as it can converge appropriately.

In general, the ambiguity in gain definition or input voltage definition  causes discrepancy between simulation results.

Some simple things to check:

1. Are you sure your transient analysis has reached steady state? If you only monitor peak voltage and the transient analysis has not reached stated state, your results will be misleading.

2. Are you sure you selected the right sideband options in your PAC analysis. I frequently require several attempts before I get it right.

3. A factor of EXACTLY 6 dB is very suspicious. It usually arises out of something simple like computing gain as the ratio of an output peak voltage to an input peak-to-peak voltage. I assume your transient and PAC circuits are one in the same so that you do not have a matching issue. For example, the port has a factor of two such that you only get the specified output when it drives a matched load. But if the circuits are the same, that should not be the issue.

4. Do you more than one PAC source? For example, if you cut and paste a source with PACmag=1, it is easy to forget that you now have two PAC sources when you meant to have only one.  Both will contribute to the output.