waveguide slot antenna

Hi everyone!

I've serious problems with the simulation of a waveguide slot antenna.
It has a single slot on the Broadwall of the waveguide and the frequency is 10GHz. it has one port, located at λ/2 of the slot, and is terminated by a short circuit at 3*λ/4 from the slot.
I'm trying to find the resonant length of the slot for a fixed offset, by parametric sweeps of the length, to get a 0 degree for the angle of the admittance mattrix.

normally, as written in the ansoft hfss 9.1 help documents, a boundary of λ/4 from the slot should do. But the angle of admittance, which I'm very sensitive on, varies a lot with changes of the boundary.

I've used some seed meshing on the slot and the boundary, but it doesn't help much.

How could i get a reliable answer???

Plzzz...help me!!

thanks in advance

sara

Hi,
what about to use PML instead of radation BC.

Pml are less sensitive to the angle of incidence and allow you to calculate
antenna parametr

Hope this help

Rugby

I suspect the problem is not on radiation boundary. Something wrong in other part of your setup.

The boundary cover of λ/4 from the slot seems too small, you must set one-λ at least. No problem in HFSS to get all parameters of you desired.

Hi

with a λ bounadry, it'll take years to solve the problem and it'll take too much memory, and the meshs won't get as dense as it must.

In a Ansoft 2003 Global seminars, there was a section about generating a slot design curve for a broadwall displaced slot design. The curve showed the resonant length of wave guide slot vs. offset distance. The boundary was stated as λ/4. It had done sweeps for a fixed air box and waveguide structure, with just the change of slot's offset and length. with a little change in the air box, everything will change.
I asked the writter of the seminar, and he answered this, with no more explanation or reply to my mail:
" You are right that you set 1/4 lambda of radiation box. But you did not
tell me other settings when solve: Delta S, number of passes, etc.
I have some comments for your case(wave guide slot antenna):
1. make seed mesh on radiation boundary and the slot. For the radiation
boundary, the default mesh is 1/3 lambda in HFSS v9. Please make 1/6 or 1/8
lambda seeding mesh( for your case, 10GHz, 4mm or 5mm is OK). For the slot,
you'd better make about 1/10 of the slot length seeding.
2. build a "virtual object" around the slot to make the mesh better( not so
long and thin mesh)


What about pmls? Are responses with PMLs really accurate? and how should i choose it's settings?

Thanks a lot
sara

Hi

To a single slot, it'll not take too much time & memory if set a λ bounadry. PML is good for pattern computation. If you only want S-parameters, ABC is accurate enough. Build a "virtual object" around the slot is a tradeoff way.

thanks!
XuQing