Can't set wave port polarization?

Hi, I've set up a five mode circular wave port. The particular ratio I have is 85% mode 1 and 15% mode 2 with mode 2 offset 180 degrees. However, no matter where I set the integration line, the overall polarization seems random. In fact, when I tried multiple times to rotate the rest of the model to have the polarization aligned, without even touching the excitation properties, the polarization still changes at the wave port. The part of the model near the wave port is cylindrically symmetrical, so that shouldn't change anything. Does anyone know why this is happening or how to fix it?

It might help to know what software you're using.

Oh right duh. I'm using HFSS.

The integration line should be fixing the polarization. Have your tried placing a seed mesh on the wave port surface, or increasing the port refinement (in analysis setup)? The initial mesh may not be seeing something that is quite circular enough.

Hm, this didn't work. Even without a mesh operation, it still gave the correct modes for a circle, just not polarized right overall.

Circular waveguides are particularly mode rich and often give me trouble getting the various modes aligned. In the modal port set up there are several choices to get things going. I find the third choice "Align modes analytically using coordinate system" to be most useful. The basic problem is that there is no preferred direction in a circular guide and the control of the direction of each is hard to define.

More fun is to be had as the order of the various higher modes is not clear beforehand. Look at the resulting modes carefully so you are confident that you have what you want.