Automated Design of a Second Order Sigma Delta Modulator using Success History Based Adaptive Differential Evolution Algorithm

The optimization based automated analog design approach requires a global optimization technique that should satisfy both high success rate and high convergence rate for minimizing multiple design iterations. The traditional population-based optimization algorithms meet both the requirements. However, their performance depends on the values of the control parameters. Tuning the control parameters for a particular design is a tedious task. In this paper, a relatively new variant of Differential Evolution (DE) algorithm namely Success History based Adaptive Differential Evolution (SHADE) algorithm is applied to automate circuit design using simulator-in-loop optimization approach. The SHADE algorithm adapts its control parameters during the optimization run. The performance of the SHADE algorithm is compared with the DE on the success rate and convergence rate. The SHADE is applied for automatic design of a second order Sigma-Delta Modulator (SDM) to maximize its Signal to Noise Ratio (SNR) in a two-stage optimization process. In the first stage, a system level optimization is performed to obtain the integrator's specifications and in the second stage, the (two stage and folded cascode) OPAMPs are designed for the required specification posed by the first stage. TSMC's 180 CMOS process technology devices are considered for the SDM design with the supply voltage of 3.3 volts. The DC, AC, and transient simulation results confirm the effectiveness of the synthesized design.