Performance Capability Modeling And Optimization Of RF/Millimeter Wave Integrated Functions And Modules Using A Hybrid Statistical/Electromagnetic Technique That Includes Process Variations

摘要

Various uses of statistical tools combined with deterministic electromagnetic simulators in the analysis, design and optimization of RF and microwave systems are presented. First, the statistical methods are introduced, showing the advantages over the conventional techniques. The statistical tools used in the methodology include sources of variation tools such as ANOVA (Analysis of Variance), SPC (Statistical Process Control), and MC (Monte Carlo) simulation, to account for the process variability. Using this methodology, the developed transfer functions predict both the nominal values and the variation expected for system performance. This is of great value for complex 3D RF integrated modules, RF MEMS and reconfigurable systems, especially at high frequencies where the fabrication tolerances affect the system more due to the smaller circuit features. The methodology can be extended to predict performance of multi-level systems, for which the outputs of the lower-level system become the inputs of the higher-level system. The presented methodology is applied for the analysis, design and optimization of a benchmarking geometry of 60 GHz cavity filters in LTCC (low temperature coflred ceramic) technology.