HFSS impedance/current issues

I am trying to duplicate the results in "Absorbing Frequency-Selective-Surface for the mm-Wave Range" by Sakran et al. for the coupling between two FSS antennas. I have recreated the structure in the attached HFSS file. But my results are incorrect. I get exactly zero coupling between antennas (it should be -30dB, low but not zero). Looking at the currents I see that zero current is flowing in the antennas. When I replace the 2D impedance boundary with a Perfect-E boundary, current flows and coupling is as expected. Apparently there is some problem with the definition of the impedance boundary, or of the lumped port which feeds it. But I can't figure out what. Does anyone have ideas?
Thanks!

Hi, I have a similar problem.

As example, the structure is a basic dipole antenna, two cylinders linked by a short sheet where there is the lumped port. (same as the tutorials from Ansoft)
If I set an impedance boundary at the end of cylinders, on the circle surface in contact with the lumped port, zero current is flowing (all results are zero), but if the impedance boundary is on the all cylinder except the previous circle face, the current is flowing, results calculated and expected match quite good.

So, what's wrong with impedance boundaries in contact with lumped port ?

The answer I got from Ansys is that due to a bug in the solver, it does not recognize relatively high-impedance surfaces as conductors, and does not feed them any current. This is a bug in HFSS 12 and should be fixed in HFSS 13.

The workaround is to add a thin strip of perfect conductor at the edge of the port. Thus current will flow. If the strip is thin enough, it will not significantly affect the overall impedance. (If it is too thin though, it will cause mesh problems.)