The attenuator is built with ideal resistor as shown here:
The two input signal are f1=1GHz and f2=1.1GHz respectively, and the input power is -30dBm for both input signal. [
Run QPSS, the result is that:
At 1G and 1.1G, the spectre are both -33dBm, that' correct. (Input is -30dBm, and attenuation is 3dBm, so it is -33dBm.)
For the modulation signal: 2*f1-f2=0.9G, and 2*f2-f1=1.2G, why the amplitudes of the two modulated signals are quite different? Thanks.
Your analysis should show a perfectly linear attenuator. I believe your problem lies with the size of the steps that the analysis engine uses - it is probably too coarse to give you equal sidebands. You might want to see if you can change the step size (but it will cost you time).