I am Sivaramakrishna. My query is about the numerical/simulation errors in pnoise/pss analysis of VCO.
I am designing a (CMOS LC-) VCO and Interesting I am getting different phase-noise for the same circuit with with same settings, such as initial conditions, and integration methods(it was suggested that I need to use *trap* integration method in-order to get correct results).
Its varying around 1-to-2 dBc. I feel that this variation is significant as it is in log scale.
The only way I can, at least to the best of my knowledge, see if the numerical/simulation noise is playing any role. Could someone shed light on it.
Thanks and Regards
If this is a strongly nonlinear circuit (e.g. ring oscillator), then pss-shooting is appropriate. Otherwise, I would use hb (harmonic balance).
Please check out your setup - very important to get correct results. In my blog:
I mention the new SpectreRF design workshops, which include Simulating Oscillators: LC and Ring Oscillators.
The workshops take you through step by step the best way to set up your oscillator to get accurate results. There is also a design database in the hierarchy. The setup instructions are in Appendix A of the MMSIM12.1.1 or 13.1 SpectreRF User Guide.