Advanced computer aided design techniques for robust RF and microwave monolithic integrated circuits

This dissertation deals with the theoretical and practical aspects of design for quality and manufacturability for RF Integrated Circuits (RFICs) and Monolithic Microwave Integrated Circuits (MMICs). Novel contributions in the areas of linear and nonlinear statistical device modeling, yield estimation techniques, circuit performance variability minimization, circuit response surface modeling, and statistical circuit design are presented.; In the area of statistical device modeling two new approaches are demonstrated. First technique is based on the mathematical algorithm known as a Cholesky decomposition. Second approach is based on a combination of applied multivariate methods with heuristic techniques. These include Principal Component Analysis (PCA) and Factor Analysis (FA) in conjunction with Maximally Flat Quadratic Interpolation (MFQI) and Group Method of Data Handling (GMDH). Semiconductor data for pseudomorphic high electron mobility transistor (PHEMT) were utilized to validate the proposed modeling methods.; New and highly efficient yield estimation technique for MMICs was developed and advocated. Latin Hypercube Sampling (LHS) was used as an alternative to Primitive Monte Carlo (PMC) in yield analyses. Accurate yield estimate based on the LHS approach was obtained at the fraction of simulations usually required by the PMC technique.; MMICs performance variability minimization is addressed by a new two stage Computer Aided Design (CAD) methodology. The methodology uses Taguchi Orthogonal Array (OA) based Design of Experiments (DoE) and explores the effects of circuit designable (R, L, C, etc.) and random noise parameters (temperature, biases, imperfections of FET devices) on the circuit outputs. A 6 GHz active bandpass filter was utilized to demonstrate that the proposed technique is a relatively simple and effective one.; Response Surface Modeling (RSM) technique for RF and monolithic microwave ICs based on general purpose black-box modeling strategy combining interpolation and regressional methods was developed. The methodology is presented along with an example and some intuitive justification of the methodology.; Finally, a systematic statistical CAD methodology capable of bridging the gap between design and manufacturing for RF and monolithic microwave ICs was developed. An industrial circuit example 2.4 GHz RF/IF Up and Down Converter was utilized in testing the proposed CAD strategy. Simulation results have shown that methodology can improve yield and manufacturability of RFIC and MMICs.