Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Simulating Dielectric Interface in HFSS

Status
Not open for further replies.

TomTurbo

Newbie
Joined
May 22, 2016
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
64
Hi there,

I am currently trying to simulate a reflection problem at a dielectric interface inside a parallel-plate waveguide using HFSS. Unfortunately I haven't found a proper way to define the source(s) to find the reflection coefficient for oblique incidence angles.

The top view of the interface (inside the PPW) is shown below. For normal incidence, I defined two wave ports (shown below) which eventually gave me the reflection coefficient (in terms of S11) that I expected.

Problem0.png
Problem1.png



But things start to get troubling when I try to simulate the transition for oblique incidence angles. I have though of different models to do this. The first one would be to still use 2 wave ports (see 1), with port 1 at the defined incidence angle and port 2 according to the angle of refraction that is found from Snell's law.

My second thought was to just use port 1 and put a radiation or PML boundary condition (2), since the reflection coefficient is the only parameter I would like to extract.

Problem2.png
Problem3.png

Unfortunately in both configuration the results seem to be strongly dependent upon the dimensions W and Ld, which I thought should have no impact (this is the case for the 2-port model with normal incidence).

Is there something that I am missing?

A last thought was to use a plane-wave excitation as shown below

Problem4.png

However, I think in that case the extraction of the reflection coefficient would be a bit more cumbersome. Perhaps someone has an idea on how this could be done, in particular on how to deal with the incident, scattered and total fields?

Any help would be greatly appreciated
Cheers!
 

1607056430810.png

Is there a reason you have to do this in HFSS? If you are using plane waves, and no curved surfaces I don't see why you can't solve for the reflection as a loaded transmission line problem. I faced a similar issue trying to determine the necessary thickness for Mylar beam splitters at THz frequencies.
Above is a simple model to represent your setup. I made up a few numbers as a start. I choose the relative dielectric for the dielectric as 4 and the relative dielectric for the matching layer as 9.
The port impedance in this case is 188, the impedance of the slab dielectric wave guide. The characteristic impedance of the matching layer in this case is 126 ohms, its length would be determined by the angle of incidence and thickness of the matching layer. The load impedance is air, 377.
Hopefully that might work, and save you a lot of time.
Sami
 

Hi Sami,

Thanks so much for your help, and sorry for the late reply. This would be definitely a much easier approach for a uniform matching layer. In fact I did the calculation once for this 3-layer structure (a closed form expression for the reflection coefficient is given for example in Born and Wolf, 1999, p. 65).

Unfortunately, I did not mention in my post that the matching layer I would like to use is actually not a uniform dielectric but a more practical transition, for example, by drilling holes into the substrate or by chamfering it as shown below in the sideview. So for this reason I decided to do this in HFSS.

edaboard.png


Sorry again for not having mentioned this earlier.

Have a nice day
 

Most likely, judging by your second group of figures, you are exciting higher-order modes in your PPW that are sensitive to parameters W and Ld.

Your setup is quite sub-optimal, since the PPW supports different modes that you are trying to excite. What you want is to keep the ports flat, replace the transverse PMC boundaries with Master-Slave pairs, and use Floquet ports to excite waves at certain angles.

There are many examples of using Floquet ports for this type of setup in the documentation and on youtube.

Good Luck!
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top