On the Metamodeling of Analog Circuits with Discrete Components: Application to the optimal design of a multifunctional filter

Abstract

The proposed work deals with the optimal design of analog circuits encompassing discrete components and integrated circuits. The proposed approach mainly consists of two steps: firstly, a surrogate model is constructed and validated on the basis of an established database, then, the model is used as performance evaluator within a metaheuristic-based optimization routine. We consider a multifunction filter formed by three current conveyors (CCs), three discrete resistors and two discrete capacitors. AD 844 IC forms the CCs, and the IEC60063 lists of preferred values are used for the parameters' values of the passive components. This configuration simultaneousely realizes three kinds of filtering: low-pass, bandpass and high-pass. For each filter two performances are taken into consideration, namely, the cutoff frequency and the quality factor. Once validated, the constructed models are used as performance evaluators within a metaheuristic-based optimization algorithm to optimize the circuits' performances. Spice simulation results are provided to show viability of the proposed approach.