Sensitivity analysis in the optimization of analog active filters by applying the richardson extrapolation

The key step of the analog active filter design is the optimal selection of component values due to manufactured series (E12, E24, E48, E96 and E192). In this paper, four simulation-based metaheuristics are applied to optimize four active filters using commercials available ICs as building blocks. The emphasis of this work is applying Richardson extrapolation-based sensitivity analysis in the optimization process of analog active filters. Indeed, Richardson extrapolation technique facilitates the calculation of the partial derivatives for the sensitivity using the simulation-based evaluation, without an explicit mathematical expression. Viability and benefits of the sensitivity analysis are highlighted. Monte Carlo analysis is performed in order to investigate robustness of the proposed sensitivity analysis of the active filters in case of component value variations due to specified tolerances of manufactured series.