[SOLVED] Problem in simulating Butler Matrix phase in CST software.

[nor2]

Hi, I am simulating Butler Matrix in CST.
I have simulated each component of coupler, crossover and phase shifter separately and analyze the output phase.
But, when I combine every component into Butler Matrix, the total output phase is not the same as the total output phase manually from each component.
Anyone has experience this?
Anyone can share any thoughts in this? Thanks.

 

[psach27]

It could be that the couplings with the different parts of your model change that phase.
When you say "combine each component", are you meaning in a whole CST 3D model or by S-parameters blocks in a the circuit simulator?

 

[nor2]

Thanks psach27 for the reply
I mean it to be in the CST 3D model form, the output phase is different when all element is combined into Butler Matrix.

 

[psach27]

Maybe is due to couplings, as I mentioned before, or/and due to junctions beetween the different elements (for instance, between step impedance discontinuities or T-junctions). This can cause and extra phase change.

 

[c74236228]

if your 3D model is fine,try encrypting the CST mesh.Different grid divisions in CST can result in different phase.

 

[nor2]

psach27-noted, I didn't add any T-junction as I just connect using the individual elements to the final structure to see if any phase differs, yes, thanks, maybe due to coupling.​

In another note, I have set the boundary to be closed in the CST.​

​

c74236228-I have tried it until 4 x meshes increase and yes, the phase changes, around +/-10 degrees.​

I'm not sure from which setting I need to follow or maybe I'll check for the uniform mest thorughout the design. Thanks.

 

[c74236228]

if you use a 4 x meshes with a ± 10deg phase change,replace the simulation software HFSS， the simulation accuracy of CST is not high,which has nothing to do with your setting and is related to the CST grid division method

 

[psach27]

if you use a 4 x meshes with a ± 10deg phase change,replace the simulation software HFSS， the simulation accuracy of CST is not high,which has nothing to do with your setting and is related to the CST grid division method

What a huge phase shift! What mesh configuration had you use @nor2 ? and solver method?
Why do you think that HFSS is more exact @c74236228? Is tetrahedral mesh more exact? If this is the case, CST also has tetrahedra mesh generation.

 

[c74236228]

In most cases,the simulation results of CST and HFSS are the same,but some models,such as four-arm spiral or polyhedral antenna,are far different from the real objects by using CST simulation. If I change the mesh density or solver method,the difference is still relatively large.The simulation results of this type of antenna with HFSS are basically the same as the real thing.

 

[ferdem]

But, when I combine every component into Butler Matrix, the total output phase is not the same as the total output phase manually from each component.

You can try to combine the subblocks using a circuit simulator such as AWR. Import all s-parameters and construct combinations. Check whether the results are as expected. If they are not as expected, try to investigate how your blocks are sensitive to input/output impedances. Add attenuators between blocks to improve matching, see whether the phase response get closer to your expections. It is good to have a working circuit in AWR; however, most likely you need to improve the model in EM simulations.