Voltage Conversion Mixer gain ..Which one is the right one?

Hi all,
I have an interesting question regarding conversion gain. I have just simulated a single balanced active mixer and I get three different values of conversion gain..Yeah true..its crazy..Its quite obvious that I am doing something wrong and so I need some suggestions from you all.

1. I computed the conversion gain through this expression (gm_in*(2/pi)*R_L)
2. Performed a transient analysis and compared RF input voltage level and IF output voltage level.
3. Performed a Harmonic analysis with LO power level sweep and plotted the conversion gain across LO power sweep

I get three different values for these different approaches. Could you please guide me with this.

P.S: I also have issues performing Harmonic analysis in ADS beyond a certain value for R_L..I am not really sure why that is.I tried re-setting all the convergence parameters though.

Dan

1-Mathematical expressions are always different than real world
2-Conversion Gain is always defined as in "Power" not Voltage level..
3-This one probably most consistant one...

Thanks

1.but the magnitude of difference is quite considerable and not so small enough to make that compromise
2. but in all the literatures they mentioned usually the term voltage conversion gain
3. thanks for it but why is it not possible to perform harmonic analysis after a certain level. do you have any suggestions on it.

Dan

1-Since the frequency is increasing, this magnitude or difference will be pretty high due to simplification.In real wold, there are many unseen parameters in simulation models.In additional to this, physical implementation and measurments will also bring you somehow differences.
2-Voltage Gain is "generally" not used in RF and Microwave because of loads are generally complex.When you check a datasheet, you will see Power Gain under some circumtances.
3-HB analysis is always possible and preferable.Terminate your mixer with 50 Ohm and drive it with 50 Ohm LO and RF/IF sources and specify Power Gain in your report, everybody will understand.Otherwise, if you use Voltage Gain nobody will known which load you have used.Because Voltage Gain load is dependable.

thanks so much for the reply.
points 2 and 3 are really helpful to know more about mixers

but, dont get me wrong, i cant understand point 1 as theory cant be so deviating from what I see in my simulation. may be i am doing something wrong too in calculations..if the frequencies are the reason, how about AM mixer..that frequency isnt that high enough i guess..

though i agree to your point 2, just a curious question ... can i simulate the mixer with voltage sources for RF and LO input.
if so can i compute the conversion gain with open load at the output, provided if I want to have the mixer as a standalone block rather than in a system.?