Finite Z component of E field for TE modes
Hi everyone,
I have been trying to simulate a waveguide junction problem with waveguide ports to calculate the field distribution on the aperture. I have a 5 circular waveguide problem where I use only TE modes at the small waveguide port. On the other end I have used a port with 0 excitation from all the modes to ensure an absorbing boundary condition.
The problem is that I still get a finite z component for the aperture E field. However, from theory we know that the aperture fields for a TE mode should have a 0 z component. This I have observed in only 1 waveguide problem too. This also happens when the discretization is lambda/25. Is that a numerical artifact? I am adding the cfx file also to this thread.
<?xml version="1.0" encoding="UTF-8"?>
Answers
-
The waveguide is way overmoded. Even though you apply the TE mode, because of the dimensions many modes are propagating.
From the OUT file:
TE 1 1 4.55156E+09 3.98372E+02 0.00000E+00 0.00000E+00 2.77479E+02 Propagating
TM 0 1 5.94492E+09 3.41078E+02 0.00000E+00 0.00000E+00 2.65650E+02 Propagating
TE 2 1 7.55032E+09 4.47366E+02 0.00000E+00 0.00000E+00 2.47090E+02 Propagating
TE 0 1 9.47229E+09 5.11613E+02 0.00000E+00 0.00000E+00 2.16061E+02 Propagating
TM 1 1 9.47229E+09 2.77408E+02 0.00000E+00 0.00000E+00 2.16061E+02 Propagating
TE 3 1 1.03857E+10 5.61795E+02 0.00000E+00 0.00000E+00 1.96761E+02 Propagating
TM 2 1 1.26957E+10 1.58787E+02 0.00000E+00 0.00000E+00 1.23672E+02 Propagating
TE 4 1 1.31454E+10 1.09505E+03 0.00000E+00 0.00000E+00 1.00945E+02 Propagating
TE 1 2 1.31798E+10 1.11703E+03 0.00000E+00 0.00000E+00 9.89584E+01 Propagating
TM 0 2 1.36461E+10 8.41727E+01 0.00000E+00 0.00000E+00 6.55583E+01 PropagatingIf the waveguide is made longer, at least the evanescent modes (again see OUT file) should have sufficiently died away over the length of the waveguide.
0 -
Altair Forum User said:
The waveguide is way overmoded. Even though you apply the TE mode, because of the dimensions many modes are propagating.
From the OUT file:
TE 1 1 4.55156E+09 3.98372E+02 0.00000E+00 0.00000E+00 2.77479E+02 Propagating
TM 0 1 5.94492E+09 3.41078E+02 0.00000E+00 0.00000E+00 2.65650E+02 Propagating
TE 2 1 7.55032E+09 4.47366E+02 0.00000E+00 0.00000E+00 2.47090E+02 Propagating
TE 0 1 9.47229E+09 5.11613E+02 0.00000E+00 0.00000E+00 2.16061E+02 Propagating
TM 1 1 9.47229E+09 2.77408E+02 0.00000E+00 0.00000E+00 2.16061E+02 Propagating
TE 3 1 1.03857E+10 5.61795E+02 0.00000E+00 0.00000E+00 1.96761E+02 Propagating
TM 2 1 1.26957E+10 1.58787E+02 0.00000E+00 0.00000E+00 1.23672E+02 Propagating
TE 4 1 1.31454E+10 1.09505E+03 0.00000E+00 0.00000E+00 1.00945E+02 Propagating
TE 1 2 1.31798E+10 1.11703E+03 0.00000E+00 0.00000E+00 9.89584E+01 Propagating
TM 0 2 1.36461E+10 8.41727E+01 0.00000E+00 0.00000E+00 6.55583E+01 PropagatingIf the waveguide is made longer, at least the evanescent modes (again see OUT file) should have sufficiently died away over the length of the waveguide.
Yes, I have all the other modes propagating as you mentioned. Therefore, I also tested when all these modes are active (excited) at the port and the result is different. And I even excited only the TM modes out of these modes and the result is again different. I have a theoretical model on MATLAB and I use z component for TE as 0 and for TM non-zero.
My only TM mode simulation in FEKO matches the results of MATLAB (all modes excited). Therefore, I expected that the only TE mode excitation in FEKO should give me almost 0 z component of electric field.
0 -
Your meshing should also be fine enough to avoid the warning:
WARNING 4682: Coarse triangular mesh of a waveguide port for representing the field distribution of the modes included in the mode expansion
0 -
Altair Forum User said:
Yes, I have all the other modes propagating as you mentioned. Therefore, I also tested when all these modes are active (excited) at the port and the result is different. And I even excited only the TM modes out of these modes and the result is again different. I have a theoretical model on MATLAB and I use z component for TE as 0 and for TM non-zero.
My only TM mode simulation in FEKO matches the results of MATLAB (all modes excited). Therefore, I expected that the only TE mode excitation in FEKO should give me almost 0 z component of electric field.
For example, take this example file I am attaching. In this I have excited only one TE mode at a frequency where the only TE mode is propagating. In this case also, in spite of having a mesh size of lambda/40, I still get some z component of E field. I guess because of this, when I add more modes, the z components due to only TE modes also get added.
<?xml version="1.0" encoding="UTF-8"?>
0
