Feko Reflection Coefficient
Altair Employee Hello, I have a question about Sparameters in Feko.
I have an infinite medium in x,y direction, in z direction it has the thickness d. The plane wave is normally incident on this medium.
I additionly calculated the Sparameters in Matlab with NRW Method, thought the permittivity is right, S Parameters dont match.
For example i have a medium(lossless) permittivity er = 20
S11 = (1 - Z^2) * Г / ((1 - Г^2) * Z^2) = -0.904 - j0.000 /abs = 0.904
S12 = 0.000 - j0.425 /abs = 0.425
(Г(refl coeff ideal for d = infinity) = (n2 - n1)/(n2 + n1)
n1 = 1, n2 = sqrt(er)
d = lambdas/4
Z = exp(-j ks d) = -j)
In feko i have at 10 lambda
S11 8.67303E-01 -159.80
S12 4.49503E-01 -87.71
though they shouldnt be different, i wonder why
https://arxiv.org/ftp/arxiv/papers/1206/1206.4801.pdf (NRW)
<?xml version="1.0" encoding="UTF-8"?>
Answers
-
Dear bobo44
From looking at your CADFEKO model I believe that you are getting your S11 and S12 directly from the (scattered) near field at -301.408 mm and 0 mm respectively.
In your model you are using the 'Coarse' meshing option. This does not appear to be sufficient. If I switch to the 'Fine' option I get values of:
S11 = (9.10981E-01, -160.01) and
S12 = (4.26543E-01, -69.28)
format = (magnitude, phase)
Your analytical values are:
S11 = (9.04E-01, 180) and
S12 = (4.25E-01, -90)
As such the values FEKO report for the 'Fine' mesh option shows a much closer agreement in the magnitude and about 20 degrees difference in phase.
This difference in phase I believe is caused by differences in where the dielectric layer is located and how reference plane(s) are defined between your FEKO model and the analytical solution. Note that your layer thickness d is equivalent to lambda0/17.888 which would be equivalent to 20.125 degrees of phase in free space.
In the attached model I changed the dielectric layer to be from Z=0 to Z=d (with the plane wave still traveling in the +Z direction) and I sampled the scattered fields at these points ( S11 at Z=0 and S12 at Z=d). Also, I decreased the mesh size even more. Using a 'Custom' mesh size of 1.25mm FEKO then gives values of:
S11 = (9.03410E-01, 180) and
S12 = (4.29548E-01, -90.12)
This model is attached here.
Please note that a much more efficient solution for this problem can be obtained by using the Infinite Multilayer Substrate rather than the Periodic Boundary Condition.
0