Statistical diagnosis and parametric yield analysis of liquid crystalline polymer RF dualband filters

Abstract

An efficient statistical diagnosis and yield analysis methodology for RF passive circuit layouts such as bandpass filters is presented. The circuits are composed of quasi-lumped embedded inductors and capacitors in multilayer laminate type organic substrate like Liquid Crystalline Polymer (LCP). In this approach, statistical variations of layout parameters are mapped to performance measures through circuit and EM simulations based on fractional factorial arrays. Statistical distributions of performance parameters are computed from sensitivity functions derived from these simulations. The probability density and sensitivity functions are utilized as a diagnosis tool to estimate design parameters of RF circuit layouts for a given measured performance. The concept has been demonstrated by analysis of a RF dualband filter.