Efficient transient simulation of high-speed interconnects characterized by sampled data

Abstract

In this paper, we present an algorithm for efficient simulation of high-speed interconnects characterized by sampled data. The method constructs pole-zero models of arbitrary interconnects using robust rational approximations of the measured or simulated scattering parameters. In order to obtain accurate interpolations of the data over a wide frequency range, a set of powerful techniques is applied to deal with the resulting ill-conditioned Vandermonde-like approximation matrices. By utilizing the analytic properties of the scattering parameters, the algorithm efficiently generates multiport pole-residue models. The models are combined with the lumped/distributed components for direct time- or frequency-domain simulations. The method can easily be implemented into conventional simulators such as simulation program with integrated circuit emphasis (SPICE) and advanced statistical analysis program (ASTAP) or reduced-order modeling techniques such as asymptotic waveform evaluation (AWE), complex frequency hopping (CFH), and Padd approximation via Lancros Process (PVL) for transient simulation of high-speed interconnect networks. Examples of linear and nonlinear networks are given to demonstrate the validity and accuracy of the method.