Which software is more accurate, HFSS or CST?

[ethen]

vivaldi .hfss

I am wonding this question:

Which one is more accurate between CST and HFSS ?

Currently I simulate 3D antenna problems ( dielectric resonator antennas) , and find that the results
such as s11 are quite different from the two softwares even so I input the same structure parameters including geometrical and electrical ,
usually, S11 from HFSS is about 15 better than one from CST.
So, I am wondering which results should be inputed for fabrication.

I guess there should be some important setting such as meshing, or others some things, but I have no experience.
Does anyone who has this expereince give your suggestion ? I understand my question may be too general to answer, but your any general experience for accurate solution in CST or HFSS is greatly appreciated !

 

[irfan1]

Re: HFSS and CST

it may be a good practise to use adaptive meshing in CST till you get the desired accuracy.

 

[wlcsp]

Re: HFSS and CST

HFSS uses FEM while MWS uses FDTD. The obvious difference lies on the way the mesh created. Common structures have sometimes different ways to have accurate mesh, and usually rule of thumb for this is provided by the company.

What kind of structure are you modeling?

 

[KMPA]

Re: HFSS and CST

Hi,
I have a same problem with HFSS and CST,
please visite this forum:

It is my vivaldi antenna simulation and meaturment results, My simulations carried out with HFSS, simulated S11 are not in agreement with meaturment results. But when eirp ( Thanks eirp!) simulated it with CST, we saw that CST results are good agreement with HFSS.

KMPA

 

[ethen]

Re: HFSS and CST

Hi, thanks a lot for all your help.

I have used your suggestion, but the difference between the two results is still quite large. I understand that it should be due to my settings for some parameters. but i don't know how to adjust which one of them. Now I attached two modes used in HFSS and CST, and a related paper. My question is below:

I hope to shife the band to around 1.8 ~ 2.5 GHz, and CST and HFSS could get the rough same results. ( ps: if slot width /2 =0.25 mm, and strip width/2=0.94 mm, the result from HFSS can basically match the above requirement. but the one from CST .

Could you please take a glance for my question and modify modes for me?

Thanks a lot again.

 

[RFSimulator]

Re: HFSS and CST

Hi Ethen,

your *.par file for the MWS is missing.

Best regards,

F.

 

[eirp]

Re: HFSS and CST

Hi Ethen,

your *.par file for the MWS is missing.

Best regards,

F.

Yes, noted it also.
However the structure can be retrieved
HFSS --> export SAT
MWS <-- import SAT

Simulating now

Cheers

eirp

 

[RFSimulator]

Re: HFSS and CST

Hi Eirp,

can you post the SAT file? I don't have HFSS

Thanks,

F.

 

[ethen]

Re: HFSS and CST

Hi , Eirp:


Thank you very much for your help. I am waiting for your results and modification to the structure.

 

[ethen]

Re: HFSS and CST

Hi, RF simulator:

Thank you very much for your interst

Sat is here, but how to edit inport shapes ? it seems that CST doesn't know sat file from HFSS, what commond can be used to edit this structure of sat file from HFSS ?

 

[ethen]

Re: HFSS and CST

I can't see my uploaded files, it is not permitted ?
:?:

Hi, Eirp, could you send your feedback to me by rao88168@hotmail.com, thank you very much.

Thank all who has been interested in helping me, please reach me by the above email. :wink:

 

[eirp]

Re: HFSS and CST

Here's the imported structure for MWS.
I hope I didn't make mistake in materials. I also had to make finite thickness on mstrip.
Sorry I don't have enough time to make more investigation on this structure

cheers

eirp

 

[RFSimulator]

Re: HFSS and CST

I did run Eirps structure with three different mesh. I increase the mesh globally but in particular along the M-Strip, inside the substrate and inside the feeding slot. As you can see, the result is pretty much converged. From a numerical point of view I would guarantee for the results.

Eirp, just a short comment from my side concerning your port definition. I would not extend the WG Port for a M-Strip to underneath the ground plane. This sometimes can lead to problems and is certainly not necessary.

Substrate and dielectric are modelled without losses. Sometimes this can have a large effect on the return loss.

F.