Optimal design of discrete-time delta sigma modulators

Optimal Design of Discrete-Time ∆Σ Modulators by Matthew Edward Jackson Dr. Peter A. Stubberud, Examination Committee Chair Professor of Electrical Engineering University of Nevada, Las Vegas In this thesis, optimal signal transfer functions (STFs) and noise transfer functions (NTFs) for discrete time delta sigma (∆Σ) modulators are determined. For a given oversampling rate (OSR), these STFs and NTFs are optimized with respect to a weighted combination of the ∆Σ modulator’s signal-to-noise ratio (SNR) and dynamic range (DR). This optimization problem is solved using a novel hybrid orthogonal genetic (HOG) algorithm that uses customized genetic operators to improve algorithm performance and accuracy when applied to multimodal, non-differentiable performance surfaces. To generate optimal system functions, the HOG algorithm is implemented as a constrained global optimizer to minimize cost functions which represent approximations of the system functions.