Problems with the excitation (again)

[sandruska]

Hy:
I'm a student of telecomunications thar are using HFSS in her project. I would like to know why the results (like parameters S11 and S21of any structures) depends on length of the integration line using lumped ports.
In special, I am Analysing a coaxial to rectangle/ circular waveguide transitions and the results of the parameters S11 and S21 aren't corrects. I am using rectangles for aproximating the coaxials and I'm using like excitation, lumped ports (internal to waveguide). Waveguide material is vacuum, and I have assigned a Boundary type PEC to the external face of waveguide. With this, the results of parameters S11 and S21 in the range of 11-15 GHz are wrong in the order of -10dB, why?. Which is the more correct excitation for this model?.

Thank's.

 

[Wiley]

You will want use wave ports, not lumped ports. Lumped ports enforce a one volt drop along the rectangle. The problem you are solving is the coupling between two magnetic probes in a cavity. (If this does not make sense, let me know.)

To create the wave ports:
1.) select a face
2.) right mouse click, assign excitation
3.) define an impedance line from the bottom of the guide to the top

Changing to wave ports will yield a rectangular waveguide through-line. I ran this and got the expected S21 = 0dB.

 

[sandruska]

You will want use wave ports, not lumped ports. Lumped ports enforce a one volt drop along the rectangle. The problem you are solving is the coupling between two magnetic probes in a cavity. (If this does not make sense, let me know.)

To create the wave ports:
1.) select a face
2.) right mouse click, assign excitation
3.) define an impedance line from the bottom of the guide to the top

Changing to wave ports will yield a rectangular waveguide through-line. I ran this and got the expected S21 = 0dB.

But the problem that I'm trying to resolve is a coaxial to waveguide transition, not waveguide to coaxial transition..the excitation must to be in the rectangles.

 

[Wiley]

I don't understand. Reciprocity says that a coax-to-waveguide transition is the same as a waveguide-to-coax transition.

One port will need to be a rectangular waveport, the other will need to be a coax. You can approximate the coax by drawing only the part of the probe that extends past the guide wall and placing a lumped port between the probe and waveguide wall.

A more accurate method is to recess both the outer and inner conductor of the coax from the guide wall. And use a waveport to excite the coax.

 

[asdfaaa]

I don't understand. Reciprocity says that a coax-to-waveguide transition is the same as a waveguide-to-coax transition.

One port will need to be a rectangular waveport, the other will need to be a coax. You can approximate the coax by drawing only the part of the probe that extends past the guide wall and placing a lumped port between the probe and waveguide wall.

A more accurate method is to recess both the outer and inner conductor of the coax from the guide wall. And use a waveport to excite the coax.

Yes, this is my point also. I am also not clear what is the purpose of this transition, studying mode transition, if that must use wave port like this way. If only interested in S-parameter, input impedance, we can us a lumped port to replace the coaxial feed, accuracy is acceptable.

Best Regards,