Rectangular waveguide pipe thickness

Hello all,

I'm very new with HFSS and I've a few questions about the design of a rectangular waveguide with a dielectric inside.


Is it better to assign a boundary condition of finite conductivity with certain thickness around the dielectric or is better to design yourself the square metal pipe and fill it with a dielectric material?
The results I obtain are very different!


Thank you in advance!

If you need to take loss into account, then you are going to need to use finite conductivity.

But if you can ignore losses, then a waveguide needs to consist of nothing more than the dielectric, as the background material is PEC - perfect electrical conductor. So all you need is a box (for rectangular waveguide) or a cylinder (for round waveguide). You don't need to draw the metal, as that is already there. If you draw just a box made of some dielectric, then that becomes a waveguide with infinitely thick walls of infinite conductivity. But that is probably a pretty good approximation to what you want.

I think you need to describe in a bit more detail what you are doing with this waveguide, as I assume what you want to ultimately model is a bit more complex. That might dictate another approach.

Check out the example of a waveguide power splitter in the HFSS manual. You will see how one only needs to draw the vacuum as a box. There is no need to mess around with definining the waveguide wall thickness.

Dave

Thanks for answering so fast!

I'm trying to model the heating of a very dielectric material inside the waveguide, so I should use the finite conductivity option as a boundary layer? I have also tried to design the pipe and the results are very different.
As a background I've done a big cube with vacuum.

Thank you!

JGG said:

Thanks for answering so fast!

I'm trying to model the heating of a very dielectric material inside the waveguide, so I should use the finite conductivity option as a boundary layer? I have also tried to design the pipe and the results are very different.
As a background I've done a big cube with vacuum.

Thank you!

Click to expand...

What do you mean by "very dielectric material"? If you mean very lossy, then I'd say you have even less need to worry about the conductivity of the waveguide walls, as the losses will be due to the material will dominate.


Dave

A block of water for example, very lossy
I now what you mean, the problem is that I find very different volume losses designing all de rectangular conductor and puting water inside or not designing the rectangular conductor and puting only water. Is it normal?

What you are saying does not sound right. If the material is very lossy like water, then this will dominate the metal losses, so there's no reason to get a different result if you just use a dielectric, or if you make life complicated and use a metal too. The background material being PEC will mean there will be no practical difference.

I guess the most likely issue is how you have set up the port - you should have used a waveport. I don't know for sure, but I think if you draw the metal, you should select a face on the dielectric and add the waveport there. Then do the same at the other end of the waveguide. Either way, make sure the waveport covers the dielectric in the waveguide. It will probably mess things up if the waveport includes anything other than the dielectric. (The other side should be PEC).

I suspect if you zip up your .hfss file, which has both models in the one file, and plots of the results, someone can see exactly what you mean. They will have to run it on HFSS, but once they run it they will see what you see.

There's probably a theoretical result for the attenuation of a waveguide with a material of a conductivity of X. I don't know that there is, but I would imagine it is the sort of thing someone might have worked out. If so, you will have something to compared your HFSS results with.

Dave