why return loss results of HFSS 3D module and layout is different?

[nauma naseer]

Hi everyone I have design a 2x14 array by using HFSS 3D module and then exported the designed as dxf format and imported it in HFSS layout. When I compare both results they are very much different. The graphs are attached. In HFSS 3D module I have used terminal port for excitation while in layout I have used edge port , the radiation boundary conditions have been used in both. My question, what could be possible reason of this difference. which module 3D or layout gives more accurate results

 

[PlanarMetamaterials]

HFSS 3D Layout is not the same as HFSS. HFSS 3D layout is for laying out PCBs, and might have some issues with modelling antennas. Use just good old original HFSS.

 

[volker@muehlhaus]

HFSS 3D Layout is not the same as HFSS. HFSS 3D layout is for laying out PCBs, and might have some issues with modelling antennas. Use just good old original HFSS.

Sure? HFSS 3D Layout Design uses the same solver engine, at least that is my understanding based on using both for some workflow testcases. Using 3D layout workflow was recommended by HFSS support for my case, because the input data is layer-based and 3D Layout Design is the "better" HFSS user interface for that.

@nauma naseer You can check your 3D Layout Model by exporting it into an HFSS 3D model, and compare to the initial 3D model. Export is possible after you have the Analysis setup configured.

 

[PlanarMetamaterials]

Sure? HFSS 3D Layout Design uses the same solver engine, at least that is my understanding based on using both for some workflow testcases. Using 3D layout workflow was recommended by HFSS support for my case, because the input data is layer-based and 3D Layout Design is the "better" HFSS user interface for that.

Not particularly, no. But I'm not familiar with simulating antennas in 3D Layout -- if there's a discrepancy, I'd default to the original HFSS. I also find the excitation mechanisms not quite as useful in Layout, which looks like it might be the OP's issue.

 

[BigBoss]

Sometimes such programs can make scaling errors while they are converting.Have you ever mechanically measured/compared both two structures ??

 

[volker@muehlhaus]

Looking at the resonances, I think the substrate material might be wrong (default to air) in the HFSS 3D Layout Model.

 

[nauma naseer]

@volker@muehlhaus you are right that both have solvers, for that i try to compare the result for HFSS setup and planer setup. This time design are completely same but results are still different. @PlanarMetamaterials I also do prefer HFSS 3D module but for co-simulation like antenna performance on pcb requires layout module. Blue curve is the result achieved through HFSS 3d module while red and green through layout module

--- Updated Aug 18, 2021 ---

@volker@muehlhaus @PlanarMetamaterials I have tried the design in cst its result 's more resembling to plannerEM

 

[volker@muehlhaus]

@volker@muehlhaus you are right that both have solvers

I meant that both use the same HFSS solver engine, so you expect same results if your model is correct (identical). Only the user interface is different, so you there must be some difference in materials, boundaries or other settings.

If you upload the DXF and specify substrate data, I can run in another solver to double check your results.

 

[nauma naseer]

I meant that both use the same HFSS solver engine, so you expect same results if your model is correct (identical). Only the user interface is different, so you there must be some difference in materials, boundaries or other settings.

If you upload the DXF and specify substrate data, I can run in another solver to double check your results.

In 3D layout model there is option to use HFSS solved and Planer EM solver for the same design and configurations. the dxf of top layer is attached the RO4350 substrate with dimensions 98x15.6x0.508 mm with complete ground

 

[volker@muehlhaus]

There is no DXF attachment in your previous post.

 

[nauma naseer]

sorry for inconvenience

--- Updated Aug 18, 2021 ---

Sometimes such programs can make scaling errors while they are converting. Have you ever mechanically measured/compared both two structures ??

actually dxf was not supported for upload that why i have uploaded rar

 

[volker@muehlhaus]

Ok, I have simulated your DXF with two different solver methods:

• full 3D EM using FDTD with Enpire XPU
• planar MoM using ADS Momentum

and get very similar results. In both cases, a fine mesh was used. I have long experience with both tools and consider my results accurate, based on the DXF contents and your substrate specification.

Your DXF has some strange distortion, is that correct? The patches are not oriented exactly along the xy-axis, so I wonder if your DXF export from 3D was done at some angle/perspective? But anyway, based on the actual DXF contents, results are solver-independent.

 

[nauma naseer]

Ok, I have simulated your DXF with two different solver methods:

• full 3D EM using FDTD with Enpire XPU
• planar MoM using ADS Momentum

and get very similar results. In both cases, a fine mesh was used. I have long experience with both tools and consider my results accurate, based on the DXF contents and your substrate specification.

Your DXF has some strange distortion, is that correct? The patches are not oriented exactly along the xy-axis, so I wonder if your DXF export from 3D was done at some angle/perspective? But anyway, based on the actual DXF contents, results are solver-independent.

Thank you for your efforts. It is highly appreciated. Yes the orientation is not exactly aligned along xy axis they it is design intentionally like that