Periodic steady-state (PSS) analysis directly computes the periodic steady-state response of a circuit. PSS convergence is required before performing other RF analyses such as phase noise. PSS analysis is performed by time-domain (for example, shooting) or frequency-domain (for example, harmonic balance) algorithms.

Traditional RF simulators often fail to achieve PSS convergence, especially on large, highly nonlinear circuits such as 5 GHz LC-tank VCOs implemented in nanometer-scale CMOS. The linear algorithms in these tools struggle to find a solution because they can only approximate the circuit's nonlinear behavior. The designer is forced to work around convergence problems by breaking up the circuit into smaller blocks, loosening accuracy tolerances, etc. These work-arounds result in wasted time and do not verify the circuit as it will be implemented in silicon. Even if convergence can be achieved, the linear algorithms produce inaccurate results for critical phase noise and jitter measurements.

RF FastSPICE is powered by the Stochastic Nonlinear Engine™ that quickly and efficiently solves the stochastic, nonlinear, partial differential equations that accurately model the nonlinear, time-varying behavior of circuits. RF FastSPICE contains both time- and frequency-domain algorithms to address the spectrum of periodic circuits. The result is robust PSS convergence, accurate noise analysis, and high performance - and less user frustration.

The table below shows some of the customer circuits in which designers have run PSS analysis using RF FastSPICE and achieved full SPICE accuracy at least 5X faster.