Floquet's ports in CST

[arxlan]

See the Line Length Versus Phase Curves in Fig 3. You will find out that these curves are for without ground plane
and then with ground plane separated by 0.1 lambda and 0.25 lambda. First i have to achieve results without ground plane
then with ground plane for both values of lambda.

 

[chameleon]

I'm assuming phase wrapping, if it exists, is not deceiving your perception?

---------- Post added at 21:51 ---------- Previous post was at 21:51 ----------

I saw the Figure later, you are right, they also checked 0.1 lambda

 

[arxlan]

Didn't get you. What do you mean by phase wrapping?

 

[chameleon]

What I mean is, since the phase repeats itself at every 360 degrees, sometimes you only see phase results in the +-180 range or 0-360 range. So the phase goes down to -180 degree and at next step instead of showing or plotting a value of -181 degree, it's plotted as 179degrees. (-181+360)

 

[arxlan]

I got it now. So the need of time is that i should simulate using "periodic boundary conditions" with open (add space) in Zmin and Zmax direction.

 

[chameleon]

As I said, if you have ground plane, you don't have to put Zmin open add space, you can use just Et=0 at Zmin.

 

[arxlan]

Now i got to know what exactly to assign the boundary conditions. Once again thanks to you. Will let you know about the results soon

 

[chameleon]

Good Luck!

 

[arxlan]

Dear Chameleon,

I have assigned the Floquet Port's to the designed structure. I'm not sure about zmin and zmax reference distance. Will you please look it for me?
Should i send you the model?

 

[chameleon]

That part, I'm not quite sure. The phase is usually tricky especially in the absolute sense. You may try setting your reference plane to top surface of your structure.

 

[arxlan]

It is to inform you that i'm able to achieve almost similar kind of results without ground plane.
Now i got another query. Hope you can help me this time again.
For the case of ground plane separated by 0.1 lambda, what should i do ??

 

[chameleon]

So ground plane results don't match for both ground plane separations? It can be because of the reference plane setup or something else.
0.1 lambda ground plane distance is tricky. What is your mesh size? Make sure your ground plane and structure are not on the same cube (i.e on different sides of the same cube, which could happen if your mesh size is also lambda/10). So I would suggest having at least three mesh cells between them. You can do it by placing a vacuum right above ground plane, having a thickness of 0.1lambda. Then under local mesh properties you can select mesh size as 0.1lambda/3 in the z direction (if I remember it correctly, your open directions) (you don't have to use extend limits).

 

[arxlan]

Actually the results are matching but the problem is, that normally when we
use ground plane we subtract the effect of ground plane from all the phase curves, but here the
results are matching without subtracting. If i subtract the ground effect then results will no
more be matching. That's why I'm not sure weather to subtract the effect of ground or not.
And if there's no need to subtract the ground then what will be the justification that why don't we
subtract it from the phase curves.

 

[chameleon]

Can you give a little more information on what you mean by "ground effect subtraction"?

 

[arxlan]

Ground effect subtraction:

In order to plot the phase curves, we need to subtract the ground from all the phase plots. For e.g, if center frequency is 10GHz.
By removing all other layers just simulate the ground with the excitation port and plot it's phase curve.
After this subtracting this from all other phase curves.

If at 10GHz, phase(degrees) is -500
and the value of ground phase is -30
then actual phase will be = -500-(-30)= -470

 

[chameleon]

Why do you need to do this? To me, actually, subtraction needs justification, instead of trying to justify not subracting. If you want to see the effect without ground plane, you already have it. This way you take the direct ground plane reflection out, but leave the coupling terms in there. Ground plane is now a part of your structure now. In your reference paper, are the results given after subtraction?

If you are going to use the reflectarray with the ground plane, then you should be checking and designing in the presence of the ground plane.

 

[arxlan]

Reference paper didn't do the subtraction as the ground plane is away from the structure. I have read in some paper
that subtraction is required in case of ground plane at zero level i.e attached with your structure.

 

[chameleon]

Aren't you trying to match your results to reference paper?

 

[arxlan]

I simulated this structure to get the results match with the paper results. and this is without ground plane subtraction.

 

[chameleon]

Ok, so you have your results matched to paper.
Now the question is "why aren't we subtracting the ground plane?" (the paper didn't subtract it either)

The other paper/papers may have subtracted it to show some certain effect they wanted to describe, or to show how this structure is different than a simple plain ground plane.

Think about this case: You have a ground plane at z =0. You simulated everything and got some phase data. Now you add another ground plane at z =- 0.1 lambda (assuming you are exciting from +z side). And you get new results. Since there's nothing passing through first ground plane, the effect of second plane will be zero, you will get exact same results, except numerical errors. Would subtracting the phase information from the simulation of second ground plane only make sense?