Identification of Highly Efficient Delay-Rational Macromodels of Long Interconnects From Tabulated Frequency Data

Abstract

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> We introduce a novel formulation of black-box models for long multiconductor interconnects, together with an identification algorithm from tabulated scattering parameters. The fundamental assumption requires a modal decomposition matrix that does not depend on frequency. The model structure includes low-order rational coefficients with suitable delay operators. The latter are included in a feedback loop; therefore, all infinite signal reflections are automatically accounted for. This feature may lead to significant speed up factors during circuit-based transient simulation with respect to other state-of-the-art solutions. The model is cast in a delayed descriptor form, leading to a straightforward conversion to an equivalent SPICE-compatible netlist. Finally, a purely algebraic stability test is presented based on a linear matrix inequality. The very high efficiency of the proposed models is demonstrated through several application examples. </para>