A question about the isolation of a 40db coupler.Help!!!!~

--------------------------------------------------------------------------------

These days I'm trying on a 40db coupler and the isolation is respected to be greater than 25db,while by using HFSS9.0 I got only 20db.How to improve the fact?
I used a stripline coupler,a very simple model.you can see from the photo.
Hope someone nice can help me ~~please.
Another,how to perform the mesh operation?I mean in a model some part for example the area around the stripline I want the mesh to be a bit more,and other part far away the stripline in vacuum to be less.
Thanks!~~~
The photo of the model is under below

The result of the model

id_guy:

I think you might want to check the documentation for pointers on seeing the mesh in advance. You will need a lot of meshes on the line, under the line, and extending some distance outside the lines for good convergence.

By the way, why not use a 2.5D simulator for a job like this? It's a lot easier to draw and simulate a planar circuit in a 2.5D simulator, and I think they will tend to converge to a good result faster and with less memory. There are even a couple of free ones out there that you could try that might work for you.

Just a thought.

--Max

Thanks,Max.
I have solved the coupler by adding some blots above the stripline which are not coutinuous.I am a new hand on the Electromagnetism.I wonder if you can recommend some 2.5D software and will help me a lot .:D

Hi id_guy:

I can think of a couple of 2.5D solvers you could try. You might consider looking at Ansoft Designer SV (http://www.ansoft.com/ansoftdesignersv/) or look at Sonnet Lite (http://www.sonnetsoftware.com/lite/). I think either of these might be a little easier for creating models and running simulations that are well converged for parts like this.

By the way, you might want to check the field distribution in your stripline ports to make sure that your port areas are large enough. Try expanding the sides of your ports to make them a little larger to make sure that you are capturing all the possible field lines. For a part like a -40 dB coupler, you need to make sure that you eliminate all possible errors; this could be sensitive to such errors (if they exist).

I think you found a good technique to improve your mesh. I have heard of people creating "hidden" bodies that are there only to influence the mesher in the coupled line region, in order to force more (and smaller) mesh elements.

Good luck with your design,

--Max

Hi,


for such circuits I suggest the following method...
1. Use the circuit simulators which gives best results & even you can optimize the circuit quickly for your desired specifications (use coupled & discontinity models effectively)...

&

2. Then final round of simulation use 2.5 D planar EM Simulator for result validation or accurate simulation...
.....................manju.............

Thanks for guys helped me .
I will try~~~~~~
By the way ,I am a Chinese,hehe