The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> State transition matrix from PSS simulation
https://designers-guide.org/forum/YaBB.pl?num=1085480887

Message started by jlgonzalez on May 25^th, 2004, 3:28am

Title: State transition matrix from PSS simulation
Post by jlgonzalez on May 25^th, 2004, 3:28am

I'm revising some theoretical works about phase noise in oscillators (particularly the works of A. Demir and Franz. X. Kaertner). I would like to calculate the state transition matrix for a circuit. SpectreRF PSS analysis calculates it when solving the steady-state solution of a periodically time varying circuit. Is it possible to obtain this matrix from the simulator?

Title: Re: State transition matrix from PSS simulation
Post by Ken Kundert on May 25^th, 2004, 9:08am

Sorry, no. You are correct in that it is calculated and used internally, but there is no facility available for outputing it.

-Ken

Title: Re: State transition matrix from PSS simulation
Post by jlgonzalez on May 27^th, 2004, 9:40am

Thanks for the information. I would like to formulate another question.

My concern is in autonomous circuits. The theoretical work of A. Demir (and others) has shown that a linearized perturbation analysis of this LPTV system of equations is not correct for phase noise in oscillators, because the phase noise can be very large even for very small perturbations.
I have read the theory section of the SpectreRF user guide about Oscillator Noise Analysis and it seems to me that the analysis is done using the linearization technique, instead of the non-linear perturbation analysis proposed by A. Demir. Is that correct?

I have done some simulations in a VCO by inserting voltage sources at several nodes of the circuit. The idea was to analyze which of the nodes is responsible of the most important contribution to phase noise. I'm looking only at the transfer function corresponding to up-conversion of low frequency components to the sidebands of the fundamental of the oscillator. I run a PSS and a PXF analysis. The transfer functions that I found (corresponding to the harmonic -1) grow unbounded as the offset frequency approaches 0. If I multiply this transfer functions by a noise PSD at the perturbed nodes I expect to find the same result that by running a PNOISE analysis with only the -1 sideband. Is that correct?