GA-based modeling of a non-ideal second order low-pass ΣΔ modulator

An approach is presented for the high-level simulation and synthesis of discrete-time modulators based on a simulation-based optimization strategy. The high-level synthesis approach determines both the optimum modulator topology and the required building block specifications, such that the system specifications mainly signal to noise ratio (SNR). A genetic-based differential evolution algorithm is used in combination with a fast dedicated behavioral simulator to realistically analyze and optimize the modulator performance. The approach has been implemented in a MATLAB Optimization tool. Simulation results are shown for both the analysis and optimization capabilities, illustrating the effectiveness of the approach. The selected range of optimized modulator topologies as a function of the modulator specifications for a wide range of values indicate the capabilities of and the performance range covered by the tool.