HFSS meshing frequency and a diplexer

[paszczakojad]

Hi,

I'm trying to simulate a waveguide diplexer using HFSS. I found in some documents, that the solution frequency used during adaptive meshing of a filter should be inside the passband of that filter - otherwise the mesh is dense enough only near the ports.

But how to do this when simulating a diplexer that has different passbands for each arm? When I used frequency inside first passband I got perfect simulation of the first filter response and bad of the second one. Similarily when I used the other passband frequency. But I need both simultaneously.

The only way was using low lambda target for lambda refinement (0.05) or applying appropriate mesh operations to create dense mesh and disable adaptive meshing, but that created hundreds of thousands of tetrahedra and the simulation time was a few times longer (and still less accurate).

I thought about making adaptive meshes for each arm separately and merging them together, but I wanted to make some optimizations, so I need this process to be automatic. I wish there was a 'mesh operation' consisting of adaptive meshing of selected solids using specified solution frequency.

Best regards,

P.

 

[tallface65]

In these cases you can use dependent solves. Create a solution setup with the solution frequency in the lower band and set it to converge (no frequency sweep). Then create another setup that links to the mesh of the first and set the solution frequency to your middle frequency and let it converge with a mesh that starts with the converged mesh from the first setup. Do this for the next band as well with a solution setup that links to the mid band mesh. For the last band, add your frequency sweep and you should have excellent results! If you have questions, post your project and I will take a look. This is a nice strategy that has HFSS converge for all the bands of interest in a cumulative manner that does not require mesh operations. And the final mesh will be suitable and accurate for all the frequencies.

There are reasons that this occur and it is due to lack of energy in the arms that do not pass a frequency and hence there is no field gradients to converge to except in the arms where the given frequency is supported.

later

 

[venu.gongal]

Hi

I am facing problems with meshing in HFSS for a waveguide filter design

Can u suggest me any document of HFSS inorder to setup the meshing and its convergence..

I can share my file with u ..if required.

Thanks
Venu