How to obtaind reflection for plane wave excitation?

[irfan1]

I know this question has been asked before. But since there is no satisfactory solution I wanted to ask again. I know in MWS you can use probes to find the transmission. But to find reflection back from the sample you need to put the probes in front of the excitation. But MWS sends the plane wave excitation from the domain boundary. So it seems there is no way to find reflection in a plane wave simulation.

 

[svarun]

Hi Irfan,
I too had the same problems when characterizing reflection from PBG surfaces. I went through a pdf which describes how to do it. ( You might have seen it )

It says that you need to do two simulations , one with the exact structure and one by making all materials as PEC and then subtract the two E-fileds detected by the probes in each case and then find the phase of this result. Still not clear to me.

Due to exams I am not doing any simulations for the next 10 days. Will let you know of any development. Kindly do let me know if you do progress. Thanks a lot.

-svarun

 

[irfan1]

thank you very much for your reply. Could you please share that .pdf file?
Best regards

 

[RFSimulator]

Hi,

just some ideas about this. The problem is, that the probe detects the superposition of the incoming and the reflect wave. In order to get the reflected wave only you have to subtract the incoming wave from the total field.
I guess they are two ways of doing this.

- either using a reference solution of the structure without any material -> there is no reflected wave, the entire probe signal is just the “incoming” wave

- or analytic, the probe position -with respect to the boundary- and the amplitude of the incoming wave is known and this should allow you to determine reference solution analytically.

Hope this helps.

F.

 

[irfan1]

let me try this with a homogeneous material and with a perfect conductor in front of the incoming wave and see what I get. I beleive I have to use the phase information also to account for interference

 

[RFSimulator]

Hi,

Maybe there is an even simpler solution. If your probe is some distance away from your reflecting boundary, you might be able to separate the incoming and the reflected pulse directly in time domain. The incoming pulse might already be over when the reflected pulse arrives. I guess one could write a macro which generates a probe signal which only contains the reflected pulse. Using DFT and weighting the freq. response with the spectrum of the original pulse gives you the amplitude of the reflected wave.

O.K. Maybe it’s not easier, but it is certainly an elegant way ;-)

F.

 

[RFSimulator]

Hi

let me try this with a homogeneous material and with a perfect conductor in front of the incoming wave and see what I get. I beleive I have to use the phase information also to account for interference

Yes, this is what I meant. If you work in freq. domain, you have to subtract the complex representation (amplitude and phase) of the incoming wave from the total field. However,. In most cases the amplitude will be 1 anyway…

F.

 

[irfan1]

I wonder if someone have such a macro
What I am trying to do is to characterize an inhomegenous medium, such as a photonic crystal. I want to find permittivity and permeabilty(if some magnetic material is inserted in to the lattice). Thats why I need to find the complex reflection constant and also transmission constant to extract the permeability and permittivity.

 

[irfan1]

Just to let you know all... the receipe suggested by RFSimulator works well. Still something bothers me. When I simulate free space with a plane wave excitation to get probe amplitude equal to one I have to use a mesh size of 30 at least.

 

[Matt]

Irfan,

Check the number of samples in your probe setup. I have found that default setting is insufficient for an accurate result.

Also, did you know you can use the AR filter on the probe signals, useful possibly if you can position the filter after the incoming wave and before the reflected. I have not tried this, just a suggestion.

Matt

 

[RFSimulator]

Hi again,

nice to hear that this works well. How large is your structure in terms of wavelength. The amplitude error might be due to the effect of "numerical" dispersion. This dispersion can cause interferece because the wave travels at different speeds at the boundary and inside the structure.
Numerical dispersion is well known at a general problem for all field simulators HFSS as well as MWS.
In order to get accurate "phase" results you might need a dense mesh in either simulator.

F.

 

[irfan1]

Thank you for your replies

My structure is around one or two wavelengths in the propagation direction. But it is much smaller on the perpendicular directions.

 

[RFSimulator]

Hi,
for one or two wavelength, the dispersion error should be pretty small even for lambda/10 -lambda /20. Can you post your project here?

F.

 

[irfan1]

I have attached a sample test file that I have used to determine permittivity and permeability. There are two files. In one of them there is an alumina slab. In the second file alumina is replaced with vacuum. I ve obtained pretty accurate permittivty and permeability values from these simulations. For the purposes of easy calculation I have choosen a slab thickness smaller than free space wavelength in the direction of propagation.

 

[irfan1]

Just a side note. When there is surface modes or evanascent waves in the structure, I beleive one must place the probes sufficiently far away from the structure.

 

[RFSimulator]

Hi Irfan1,


just wondering: Why don't you use "WG" Port instead of a plane wave? If you use a WG Port and you use magnetic boundaries at Zmin/max and electric boundaries at Ymin/max the Port solver will use a plane wave as excitation. You immediately get the S-Parameter out of this calculation which corresponds directly to your reflection and transmission coefficients? There would be no need for any “calibration run”.

F.

 

[shahid78]

hi irfan

i am also interested in finding the bandgap of EBG surfaces using reflection phase method. i am just thinking after modeling the alumina and then free space how u find the permitivity and permeability. can you explain it in detail.
also can you please tell me the procedure of how to set up the model which template to use and which boundary conditions to use if i want to find out the bandgap of the EBG surface.

your guidance will be highly valuable for my research work.

thanks
shahid

 

[abdoeng]

shahid,see that link,it could help u

 

[cuongtran]

hi shahid, did you find out a good solution for tracing EBG bandgap without eigenmode? Please share ur methode. Thanks

 

[shahid78]

hi
i am trying to find a method using transient solver . i will share my comments once i am successful.

regards,
shahid