Near-end crosstalk (NEXT) and far-end crosstalk (FEXT) are defined for twisted cable pairs [1], while both FFT and Baum-Liu-Tesche equations have been applied to analyze crosstalk in bidirectional coaxial cables[2]. In contrast, I wish to simulate a model for straight unidirectional wires in a multi-lead cable.
My goal is to visualize the crosstalk and how it affects the "pure" signal. I only have access to a compound signal and not the individual leads. However, this is a broadband signal, where I can isolate the narrowband components and the momentary phase.
How can I study crosstalk in multiple pairs of straight wires that share direction of current, but with varying phase*?
* The phase is neither exclusively odd nor even, but is subject to continuous phase shifts. The wires belong to circuits with independent power sources.
Edit: Resources
These are the materials I am working with right now:
I also found this resource interesting:
"The error rate in multipair cable systems is affected by the
transmitted signal, level, cable loss, crosstalk, and the required
receiver signal-to-noise ratio."—Digital Transmission Systems, Ch.
7.8.
I still haven't figured out if the equation 7.25 applies though. It involves "the transfer function of the baseband code".
In extension to "straight", these types are also of interest:
|| parallel
\/ fanning
X crossing
> < kissing
<< bending
Question moved from physics.stacksexchange.com
Best Answer
You can estimate the mutual inductance between an aggressor loop (signal wire + return wire) and a victim loop (signal wire + return wire). Given that and the current signal risetimes and amplitudes you can easily calculate the crosstalk effect on the victime wire.
If the two loops share the same return path, you should also consider IR drops on the return path due to the aggressor signal as a crosstalk on the victim channel.
Edit As Andy points out, there is also capacitive (electric field) coupling to consider. AFAIK, this is rarely the dominant source of crosstalk in the type of cables you seem to be asking about, although it could become an issue if the cable and/or its terminations are badly designed.
Parallel wires is going to be the worst case scenario. All of your other scenarios are going to have lower crosstalk than a parallel pair separated by the smallest amount of separation in a non-parallel pair.
Wires crossing at 90 degrees will have the least crosstalk.