How much distance for Floquet port in HFSS?
Well, the answer to this question is rather complicated. Yet the distance could be obtained easily for each structure, frequency, and incident angle.
As you know when a periodic surface is illuminated by a plane wave all Floquet modes will be excited. Depends on the frequency and incidence angle some higher order Floquet modes could be propagating away the periodic surface. If these higher order modes reach the Floquet port you should include them in the simulation in order to get right answers.
In the Floquet port wizard you should specify how many of the Floquet modes would contribute in simulation. These modes are indicated my two indices m and n. m=n=0 corresponds to the specular Floquet mode which you want to measure S-param for this mode. There is also a mode calculator in this wizard. Using the calculator you can determine that which mode is propagating from the periodic surface and which is evanescent away the surface. The calculator simply shows the attenuation of each mode. The attenuation of ?0 dB? shows a propagating mode and any positive attenuation shows the evanescent modes. The attenuation is calculated in dB/unit length of the HFSS project. Using the calculator you should make sure that all higher order modes are heavily attenuated when they reach to the Floquet port.
For example:
80dB attenuation for higher order mode m=0, n=1
17dB/unit length attenuation for the higher order mode m=0, n=1
HFSS unit length is mm
You should place the port at least 4.7mm away from the periodic surface.
There are still some points to clarify:
1. Usually the attenuation of higher order modes increases as their indices (i.e. m and n) becomes bigger. However, the specular mode (i.e. m=n=0) is always propagating. In addition, the attenuation of higher order modes will decrease as the incident angle increases. This means that for near grazing angles most of the higher modes might be propagating from the periodic surface (attenuation is zero).
2. In the case that you have propagating higher order modes (attenuation is zero modes), you need to select these mode in Floquet port wizard to incorporate them in simulation unless you get inaccurate answers.
Good luck
Hello,
I'm analyzing a periodic array that has higher modes propagating at an interested frequency using the Floquet ports in HFSS ver.11. I'm including all the propagating modes in the analysis, and I do mesh refinement including all the propagating modes as well. I also put the Floquet ports far enough so that the evanescent modes would have been attenuated heavily. This terns out to be computationally expensive. Could you tell me if you know any faster way to analyze such problem? For instance, can I use a plane wave incidence and PMLs with a smaller computational domain? If so, what are the implications of using PMLs?
Thank you,
Kazuko
As far as I know Floquet port is the best tool for analyzing such structures. If U need to reduce the computational complexity, relax the convergence criteria on the S-param, run the simulation, use smaller convergence criteria, and run the simulation again. Furthermore, U can use smaller air box and lower attenuation on the evanescent harmonics (e.g. 40dB). In the case that you run out memory, use solution setup> option> Enable iterative solver
Sincerely yours
Feri
Hello,
Could you please tell me how to change the incident angle of plane wave in Floquet port
Hi!
Can I ask you how do you know how much is the attenuation for higher order mode m=0, n=1?
Thank you very much
Hi
There is modes calculator to calculate the attenuation for each mode.
Suppose for my structure(Unit cell) no attenuation for m=0,n=1 mode.
Floquet port>mode setup>mode calculator
Thank you very much!
hello, could any one help me about the deembed distance? thanks.
Excuse me for my late answer,
The relative phase shift between the master and slave boundaries determines the incident angle (this relative phase shift in HFSS 12 can be change directly from the Floquet wizard, but in HFSS 11 you need to change them in the wizards of master/slave boundaries).
1. zero phase shift between the master/slave boundaries (both along x and y axis) means the incident angle is zero (normal incident)
2. 180 deg phase shift between maxter/slave boundaries along x axis and 0 deg phase shift between maxter/slave boundaries along y axis means that the wave incidents at grazing angle along x axis
3. Using the ideas in term 1 and 2 you can do a simple geometrical calculations to determine the proper phase shift for any other incident angle
Good Luck
- - - Updated - - -
See Microwave engineering by Pozar (P. 180 edition 3) under the title "A shift in reference plane".
By setting the de-embed distance equal to the distance between the port and the surface of your periodic structure. You can get the S param with the reference plane located at the surface of the periodic structure.
Good Luck
thanks for your help.
after reading ur article, I still dont know how to calculate the distance. Is there a formula to connect the distance to the frequence and incident angle? Why 4.7mm for the structure?