Radiation boundary in HFSS 13

Hi all,

I got the new HFSS13 and Antenna Design Kit (ADK). I see that ADK makes TWO airboxes one inside the other and places Radiation boundary only on th outside airbox. Why TWO airboxes? How big are they?

Thanx

The inner box is not needed. It was called a virtual airbox and was used in older versions on HFSS to create an inner surface to do field integrations for far field plots.

have fun

Yes. As HFSS manual said, better meshing is implemented on the inner virtual airbox. Therefore it should give more accurate simulation results of the far-field properties.

Back to the ADK vfurlan was talking about, when using the tool to create a dipole why is the radiation box a 3d rectangular shape? The geometry of the electrical structure is obvious a cylinder (wire dipole) wouldn't a cylindrical radiation boundary be a "convex" shape and hence most of the radiated waves would be absorbed by the radiating boundary (off incident normal waves are not absorbed as well)?

Version 13 is not new! Version 14 was released last month. I don't think there's been any updates to the antenna design kit, which only runs on Windows.

I believe the reason for the two boxes is related to improved meshing, as someone else said. As for the sizes, looking at the output from the antenna design kit might give you some idea, though I admit I'm not giving you a full answer there.

The radiation boundary must be at least a quarter wave from any part of the antenna - preferably a little more, though in directions where there's little radiation, I believe it can be a bit closer, but until you really know what you are doing, it's best not to get it any closer than lambda/4. Note, when computing lambda/4, you must use the longest wavelength you simulate over. So if the design frequency is 1 GHz, but you simulate from 300 MHz to 3 GHz, you must make the radiation boundary at least 250 mm from the antenna, since the longest wavelength in your simulations is 1 m.

---------- Post added at 08:29 ---------- Previous post was at 08:27 ----------

For what it is worth, I was designing a coaxial colinear, with the aid of an excellent Ansys engineer in the UK. He said I should use a cylinder for the boundary, so I would tend to agree with you there. However, a cylinder does present a concave boundary, not a convex one!

The other option is a perfectly matched layer, which can be closer (I think lambda/6), but needs to be a certain thickness. There is a document about antenna design from Ansys