Generalized Polynomial Chaos Based Stochastic Collocation on the Uncertainty Quantification of CMOS Active Filter Circuits

In today’s nanometer-era semiconductor manufacturing technology, quantification of the effects of component-level parameter uncertainties on system operation has become in-dispensible. Well-known brute-force Monte Carlo is still a popular uncertainty quantification technique, but it requires high computational power, rendering it insufficient for the analysis of complex systems. On the other hand, Generalized Polynomial Chaos based stochastic spectral techniques are able to achieve the Monte Carlo accuracy with much less effort in certain situations. In this study, we compute the stochastic characterizations of several multi-component active filter circuits with the gPC-based stochastic collocation technique utilizing our Stokhos-based MAT-LAB/C++ toolbox and present performance comparisons with Monte Carlo along with intuitive and insightful comments.