Estimation of the statistical variation of crosstalk in wiring harnesses

Abstract

Analyzing interference problems in vehicle wiring harnesses requires fast and accurate methods of approximating crosstalk. Worst-case approximations using lumped element models are fast and easy to use, but run the risk of overestimating problems. Statistical methods that account for the random variation of wire position help prevent overdesign, but are often difficult and time-consuming to apply and lack a clear link between problems and their cause. Here we investigate the use of simple lumped-element models to predict the statistical variation of crosstalk in wire harness bundles. Models are based on lumped-element approximations, where inductance and capacitance values are calculated for a single bundle cross-section, and only the circuit position is varied. Accuracy was evaluated by comparing results to numerical simulations. The method does a good job of quickly predicting the reasonable worst-case values of crosstalk due to inductive or capacitive coupling.