Learning Forum

[ixg90]

I'm trying to identify the source of a specific type of anomaly in the transmission (s21) parameter of a frequency selective surface. The anomoly can be seen at approximately 480 GHz below: 

 

 

I've set up a model using periodic boundary conditions and Floquet ports to simulate a single unit cell of a low-pass filter. The model uses a 'perfect conductor' for the material. Model is shown below: 

 

 

I've specified a 'Broadband' solution type with 0.5% Maximum Delta S as a convergence criteria: 

 

The discrete sweep goes from 1GHz to 600 GHz in 1GHz increments: 

 

I've verified that the mesh has met the convergence criteria: 

 

 

I've simulated similar frequency selective surfaces with several different geometries (and therefore cutoff frequencies). I've noticed similar artifacts in nearly all of these simulations. I'd appreciate any help in identifying the cause of these anomolies. 

[Faezeh Ladani]

Hi!
 
Check if you have enough number of Floquet modes for the ports (You can see how many modes are propagating through Mode Calculator). You may also try a plane wave incidence for excitation and two Rad boundaies on top and bottom (Select one, Right click on set as FSS reference), especially if you have many Floquet modes propagating at higher frequencies... The cell whatever size it is, is a very different structure in terms of wavelength to the lower and higher bands... At lower band you might only have fundamental modes, but for higher frequencies you will have more modes. You may read about it more on HFSS Online help. Simply search for Mode Calculator. 

Also, do you really need the structure for a band of 1-600GHz? Setting such a huge band for adaptive meshing might not be best...