Stochastic evaluation of parameters variability on a terminated signal bus

This paper addresses the simulation of the effects on a high-speed data link of external factors due to fabrication tolerances or uncertain loading conditions. The proposed strategy operates in the frequency domain and amounts to generating a suitable set of stochastic models for the different blocks in which the data link can be decomposed. Each model is based on the expansion of the block chain parameter matrix in terms of orthogonal polynomials. This method turns out to be accurate and more efficient than alternative solutions like the Monte Carlo method in determining the system response sensitivity to parameters variability. The advantages of the proposed approach are demonstrated via the stochastic simulation of a PCB application example.