A comprehensive linear-regression-based Procedure for inductor parameter extraction

Abstract

A novel procedure to extract all parameters from a compact double-π model for on-chip inductors has been presented. The major method used in the process is linear regression, which avoids iteration and complexity. The methodology presented here is used to extract parameters from the measured scattering matrices of symmetric inductors fabricated with a 0.18µm process. The comparisons between the predictions of the model based on the extracted parameters and the measurements demonstrate the effectiveness and accuracy of the proposed procedure over a broad frequency range even beyond the self-resonant frequencies(SRF)