Wave Port Size for Bandwidth?

Hi everyone,
I used the Solve Ports Only (HFSS) to determine the wave port size. It should allow only the fundamental mode to propagate.
Solution Freq.=14.25GHz (operation freq.);
Sweep Freq.=[10GHz-22GHz];

I want to see the transmission characteristics (including S-parameters later) in the sweep freq (bandwidth).
The wave port size should be only valid at 14.25GHz? Or, it should be all valid in the whole sweep bandwidth?

Waiting for your reply. THANK YOU!
Jianke

Hi jianke,

The waveport's physical size does not need to be in terms of wavelengths - so it can stay the same absolute size for different frequencies. As long as you have the correct mode for the solution frequency, the same mode will be excited for any swept frequency points.

Hi PlanarMetamaterials,
Great to see you again. Thanks a lot for your reply this time and your last help (in eigen-mode simulation).
I am sorry that I am not very sure what you said. I made a simulation comparation of waveport width between 6mm and 9.5mm (using HFSS Driven Modal).

When the waveport width is 9.5mm, we can see that the second higher mode appears at 14.8GHz, which is in the bandwidth from 10GHz to 22GHz. Does it affect the structure's transmission characteristics (including S-parameters) of the bandwidth from 10GHz to 22GHz?

Or, we should use the width of 6mm, which the second higher mode is out of the whole bandwidth? Thank You!

That is an interesting result.

The size of the waveport does affect how the port solver chooses the mode to converge on, which I believe is why you're getting that behaviour.

Have you set integration lines for each mode? If not, I would highly recommend doing so, as this may solve this problem. Also, have you confirmed that in both cases the second mode is actually the same mode?

Hi PlanarMetamaterials,
The above simulation results were obtained following the <HFSS Guidance_Microstrip Wave Port>. According to this guidance, I only set the fundamental mode to have the integration line to calibrate. The second mode of the two different portwidht may be not the same, but it actually happens next to the fundamental mode.

I got further simulation results using both two different waveport width in HFSS Driven Modal and Driven Terminal (see the attachment). Could you please help me check which simulation set is more reliable? Thanks again!
SIW_Air Radiation.doc

Hi jianke,

I can't really tell what exactly is happening from the information in your document, but its a good start. Can you plot just the second mode, and as well include the e-field vectors in the X-Z plane?

Hi PlanarMetamaterials,

Thanks so much for your reply again. The simulation results according to your request is attached.
This file only includes the simulation using two modes. However, you can compare it with my first document.
If you need further information, please let me konw. (The simulation files are under HFSS15.0)
Waiting for your suggestions. Thank you for your time!

A new simulation plot with
(a) same colour scales on the field plots to the the same for both modes (instead of on auto),
and b) plot the vectors instead of just the magnitude.

(a) more clear vectors (turn off "map size" and increase the arrow density)
(b) convergence criteria

Hi Jianke,

Thanks for the updated info. I have a few more questions for you. As well as the convergence info you posted, can you give me the post-run statistics? I.e. how many adaptive passes did your setup complete? What was the final Δs?

As well, what kind of excitation are you trying to model with this wave port? Is it a PPW? From what I can tell from your images, it looks like a higher-order even TM slot line mode is what is being excited, but it is a bit hard to tell.

Hi PlanarMetamaterials,

The simulation results include the final convergence info. are attached for your check. Thanks again for your reply.

Actually, I want the waveport to excite only one quasi-TEM at the end of microstrip line. And see transmission characteristics (including S-parameters) in the bandwidth (10GHz-22GHz). The center freq. is 14.25 GHz.

If I set the waveport width to 9.5mm, a second higher mode will appear at 14.8GHz. I am afraid that it will affect the transmission characteristics in bandwidth (especially from 14.8GHz-22GHz). Thus, I am thinking whether I should ensure only the fundamental mode will be excited in the whole bandwidth, corresponding to the waveport width of 6mm.

Thanks Jianke,

So the mode you want to excite is between the micro strip and the bottom cladding of the dielectric, but the higher order (surface-wave-like) modes exists primarily above the microstrip. In your first set of data, what weighting ratio were you using to excite the various modes? The microstrip mode should still be dominant over your entire bandwidth.



I need to correct my last post - it it slightly difficult to interpret the field vectors. Can you obtain a (final) set of images by zooming into the region of the dielectric (in between the microstrip's strip and ground), and plotting the field vectors (increasing the density appropriately)? I'm trying to tell what exactly the higher order modes are doing. Can you also plot the current density vectors on the top and bottom surfaces of the strip as viewed from the top?

Thanks :)

Thank you so much for your patience. The new simulation is attached.

The structure between two microtrip line is a substrate integrated waveguide (SIW), whose fundamental mode is TE10 mode. Maybe we can neglect the SIW and simplify the structure to a simple microtrip line (quasi-TEM). So, my question is which waveport width (9.5mm or 6mm) is more reliable (realistic) compared with measurement, since I want to see the quasi-TEM mode's transmission characteristics in bandwidth (especially from 14.8GHz-22GHz).

Hi Jianke,

Thanks for your patience :) I actually can't make out the vectors in your images, can you please increase the size of the vectors for the current density plots. As well, can you please zoom close to the port region in the dielectric (such that all we can really see is the dielectric), and plot the E-vectors in that region for the second mode.

I think I understand what you want, the problem is determining the physical feasibility of exciting the second mode. If, for example, this mode was impossible to excite with the mictrostripline feed, then you don't need to worry about it at all. However, we need to know the characteristics of this mode before we can make that judgement, hence my interest in figuring out its properties.

Hi PlanarMetamaterials,

Sorry for the inconvenience of my figures. The new figures are attached. Thank You.
I cannot plot the E-field for the second mode separately. I set the Number of modes (Port) to "2". I think the plotted E-field is combined with the fundamental mode and the second higher mode.

Why not? Can you not right click on "Field Overlays" and select "Edit Sources"? Your plots seem to confirm that you're only looking at the first mode, as they look very much like a microstripline mode (Let me know what your current settings are).

Thank you for your suggestions. I got the simulation results for the second mode.
The file only includes the E-field and current vectors of the second higher mode for waveport of 9.5mm and 6mm, respectively.
For the first (fundamental) mode, you can check the results in the file named "WavePort_Set6.doc".
If you need further information, please let me know. Thank You.

Hi jianke,

Thanks for the info. It really looks like that second mode is nothing but a surface wave mode excited by the waveport. The question you need to ask is: how likely is it that you will excite this mode? This will depend on how you are exciting the mictrostrip mode. Is it with an SMA connector? If this is the case, I would recommend setting up a simulation with a coaxial feed and exciting the coax mode on the connector.

As for your original documents, they seem to suggest that the the micropstripline mode will have a slightly different profile depending on the size of the waveport. Again, you can resolve this issue by properly simulating the resilistic excitation - i.e., the SMA connector or whatever you're going to be using.

So in summary, I think if you look at modelling your excitation source more realistically, you should be able to resolve your problem.

Good Luck

Hi PlanarMetamaterials,

Thanks again for your suggestions. The SMA connector will be used to excite the microstrip line.
Do you have some suggestions about the dimensions and design of the coaxial feed?
If available, could you please give me a *.hfss file about it? Thanks so much!

Hi jianke,

The coaxial feed should just be whatever the SMA connector ends in, I believe. You might just want to put the waveport on the end of the connector's coaxial end. When you get everything simulated with the connector, be sure to look at the fields and observe to what degree the higher order modes are being excited on the microstripline - obviously you need the microstrip mode to be dominant.

Unfortunately, I do not have an HFSS file. I imagine that some may be available elsewhere on the internet, though.

Good Luck.