agilent ads software + coaxial probe feed
Maybe it is possible to include a coaxial connector/cable in the new version of ADS (Momentum), which has extended 3D capability. For older versions I guess it is possible, but it will probably be quite difficult to draw the structure.
Anyway, I don't think you need to include a coaxial connector/cable in your simulation. Typically, you get sufficieny accuracy by using local ports (not sure what they are called in ADS though) that you can put between the ground plane and the tip of the feed probe.
Finally, I've tried to do stuff in Momentum more times than I care to remember, and everytime I get dissapointed. The program just sucks. For instance, we tried looking at the transmission (S21) between two dipoles in space as a function of distance, expecting to see a 6 dB loss for every doubling of distance. Not so in Momentum. Below around -30 dB or so the program fails. In my view, this is simply horrible. It means there is something very very wrong with the way it solves the problem. As a reference, IE3D shows 6 dB increases loss down to about -100 dB (which is when I stopped the experiment). This is just one of a zillion bugs/errors in the program, so my advice is, if possible, to switch program.
Best regards,
Peter