Rapid simulation of the statistical variation of crosstalk in cable harness bundles

Abstract

Accurate assessment of crosstalk problems in cable harnesses requires simulation methods that account for statistical variation in harness parameters. Use of the T-parameter method to rapidly estimate the statistical variation of crosstalk in cable harness bundles is outlined. Crosstalk is calculated by segmenting the harness into a cascade of equal length multi-conductor transmission lines (MTLs). Methods to test the influence of the statistical variation in wire positions, height, and number of twists are proposed using both T-parameter simulation and analytical methods. The analytical method uses the mean and variance of crosstalk estimated for fixed harness parameter values to estimate the mean and variance when parameters vary with known probability density functions. The analytical method may thus save additional computation time, since simulations are only required for a relatively small number of fixed parameter values. More importantly, the analytical method may give additional insight into the cause of problems. A comparison of the simulated and analytical results shows the analytical method can accurately estimate statistical variation for the parameters studied.