I'm trying to get a value of differential impedance on a cable on the E5080B. I tried using Smith Charts but the resulting chart is a mess and the TDR analysis of the impedance shows a pretty good representation but it changes with any movement in the cable. I don't know which values to trust or how to get an accurate reading. Any help is appreciated.

Normally a differential cable has 2 connects at each end and sometimes a ground. you have to connect each of the 2 connections to a port on the E5058 (you need a 4 port). For impedance only you can connect just one-end to 2 ports and terminate the other two ports in the nominal impedance. The cable impedance may not be constant with length, so the TDR response (or time domain, option 10) can be used instead. You need to set the parameter to Sdd11, and the reference impedance will be 100 ohms. You look at the "real" format of the reflection coeffcient, and you can set the marker display format to R+GB and read the impedance as a function of length directly. (In fact, this is not actually impedance as a function of length but reflection as a function of time, but for low-loss cables it gives a good measure of impedance).