Airbox and PML settings

[JohnFid]

Hello, I am trying to create a bandgap diagram in eigenmode for the structure below:



but I don't know if I set the correct size of the airbox above the cell and the PML.

The size of the first airbox is 6h and for the PML I use the recommended value of thickness but
the minimum frequency of my setup solution and 6h as the minimum radiation distance.
Can anyone verify that these values are correct?
Thank you.

 

[khaled.ayman]

Hi,

The air box size should be at a distance of \[\lambda\]/4 from your structure.

 

[JohnFid]

I use 500MHz as minimum frequency to my set analysis. Do I have to set the air box size 0.15m?
What about the PML thickness and the minimum radiation distance?
Thank you.

 

[rodgerwxh]

your wavelength at 500MHz is 0.6m while your geometry size is only 8mm, that is saying this is a very E-small problem (geo size is much smaller than wavelength). if you use 0.25 lamda as the thickness of the airbox, imaging how big will be your problem domain. I think if the PML is well designed, it can be put really close to the the object. but a robust PML is certainly not there yet. Try ABC and see if the result makes sense, since your problem is almost quasistatic.

 

[JohnFid]

My design has eigenmode as solution type and so I can't use ABC. What do you mean "PML is well designed"? Do you suggest to use the default formulas of PML Wizard?

 

[rodgerwxh]

Common PML usage experience is to put it 1/4 wavelength away from the object. The reason for this is that regular PML is proven that it can not absorb evanescent waves at all. That is why 1/4 wavelength is suggested and the fact behind it is to decay the evanescent waves sufficiently. Well designed PML means complex-frequency-shifted PML. it can absorb evanescent waves and thus can be put really close to the object, and the 1/4 wavelength suggestion does not need to be followed. I am not an user of CST or HFSS, thus I can not say what type of PML they are using. But papers has clearly discussed about this for quite a while.

(https://www.google.com/url?sa=t&rct...=wDUDBzWVfSl05qsC5MWD6Q&bvm=bv.45645796,d.dmQ)

 

[PlanarMetamaterials]

I can confirm that in HFSS for eigenmode simulation of planar structures, having the PML 1/4 of a wavelength from the structure gives the same results (or as much as is discernible) as having it to within 1/30 of a wavelength. As far as I can tell from the information provided, your setup looks correct.

 

[JohnFid]

First of all, thank you for your answers.
For the PML, do I have to use the default values provided by the wizard or I must change them according to the design (minimum frequency, thicknesss and radiation distance)?

 

[PlanarMetamaterials]

I typically use default values - although if the radiating distance is too small I generally try to increase the size of my vacuum. Usually it's pretty close to what I need anyways.

 

[JohnFid]

Increase the size to match the radiating distance?
Another question is how to set the ground correctly. In some examples, ground is a perfect conductor box of some thickness. In the above project, I use a plane material assigned to Perfect E (plus Infinite Ground Plane).
What do you suggest?

 

[PlanarMetamaterials]

Increase the size to match the radiating distance?

Yes, plus a bit of a safety region, say 5%.

What do you suggest?

A volumetric PEC behaves the same on the surfaces as a plane PEC. Personally, I'd go with a volumetric box so that later you can change it to a copper material for increased accuracy.