can we define infinite substrate in HFSS?

[kae_jolie]

I'm working on a simulation in HFSS in which I need to define an infinite ground and an infinite substrate. I know how to define infinite ground in HFSS, but I don't know how to define infinite substrate in HFSS. I know it can be done in other EM simulation software like FEKO.

Thanks.

 

[kae_jolie]

I tried assigning "inf" for infinite dimension on the substrate box, but I got an error. So, I take it, HFSS does not support infinite substrate. How about HFSS-IE?

 

[volker_muehlhaus]

Obviously, you cannot simulate/mesh to numerical infinity. The infinite substrate in planar tools means that the substrate is homogeneous throughout the simulation area. In my opinion, in HFSS and other volume meshing solvers this means that the substrate is homogeneous up to the boundary conditions.

 

[kae_jolie]

But aren't all substrates homogeneous?

I modified the size of the substrate from something small like 60X60 to as big as 200x200. It took forever to run with the bigger substrate for the obvious reasons, but the S11 plots were different. From 100x100 to 200x200 substrate size, not much S11 difference is observed. Is it fair to say that at 100x100 substrate, the substrate would be considered infinite in HFSS?

I wonder how MoM solvers manage to simulate the infinite substrate. Don't they have boundary conditions, too?

 

[volker_muehlhaus]

But aren't all substrates homogeneous?

An example for "inhomogeneous" would be an antenna at the edge of the PCB, where you want to include the field effect of the finite dielectric size.

I modified the size of the substrate from something small like 60X60 to as big as 200x200. It took forever to run with the bigger substrate for the obvious reasons, but the S11 plots were different. From 100x100 to 200x200 substrate size, not much S11 difference is observed. Is it fair to say that at 100x100 substrate, the substrate would be considered infinite in HFSS?

I can't say, because I am not an HFSS user. My best guess is that your change can be from the boundary conditions as well - it does not need to be a substrate size effect.

I wonder how MoM solvers manage to simulate the infinite substrate. Don't they have boundary conditions, too?

In case of Momentum = open boundary MOM, the underlying equations assume infinite substrate. The solver does not actually discretize to infinity, because MOM meshes & solves for currents on the conductors.

 

[PlanarMetamaterials]

Hi Kae,

If the reason you're looking for infinite substrate is that you want to neglect surface wave scattering at the edge of the substrate, I've found that placing radiation boundaries along the side walls of the bounding vacuum (or whatever medium you're in) box, which is tight to the edges of the substrate, works just fine, although with the caveat that the substrate has to be fairly electrically long (≧ λ/4 between the edge and any radiator).

 

[kae_jolie]

Thanks, guys, for your input.
PlanarMetamaterials, when you say "works just fine", how could you prove that?

 

[PlanarMetamaterials]

PlanarMetamaterials, when you say "works just fine", how could you prove that?

When I look at the animated surface waves on the ground plane, there don't appear to be any reflections off of the boundary, as would be the case for an infinite substrate. I've also used this setup for the analysis of a patch antenna; there are no 'ripples' in the radiation pattern as one might expect to see on a finite substrate.