Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

number of modes exists in a waveguide

Please login with a confirmed email address before reporting spam

is there any possibility in COMSOL to find the number of modes exists in a rectangular waveguide in a certain range of frequencies and see those modes??
thanks you

7 Replies Last Post 25.06.2016, 13:51 GMT-4
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 18.06.2016, 19:18 GMT-4
Yes. I'm assuming you have the RF module. From the Model Wizard, choose 2D. Under Select Physics, choose Radio Frequency, and under that, choose Electromagnetic Waves, Frequency Domain. Click Add. Click on the Study arrow at the bottom. Under Preset Studies, click on Eigenfrequency. Click Done. Now create your rectangular waveguide cross-section, mesh, etc. You'll set up your search for eigenfrequencies when you get to the study settings. After you run it, you'll get a set of solutions, one for each eigenvalue it finds. Some of the solutions are just noisy junk (corresponding to complex numbers for eigenvalues...) but others will correspond to the correct eigenmodes. It should be pretty easy for you to tell the difference.

Yes. I'm assuming you have the RF module. From the Model Wizard, choose 2D. Under Select Physics, choose Radio Frequency, and under that, choose Electromagnetic Waves, Frequency Domain. Click Add. Click on the Study arrow at the bottom. Under Preset Studies, click on Eigenfrequency. Click Done. Now create your rectangular waveguide cross-section, mesh, etc. You'll set up your search for eigenfrequencies when you get to the study settings. After you run it, you'll get a set of solutions, one for each eigenvalue it finds. Some of the solutions are just noisy junk (corresponding to complex numbers for eigenvalues...) but others will correspond to the correct eigenmodes. It should be pretty easy for you to tell the difference.

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 18.06.2016, 19:43 GMT-4
should it be only in 2D, can't we find it in 3D model?
one more question:
is there a possibility to know which mode has the lowest loss (attenuation) while transmission??
also how can I distinguish between the modes which belong to TE mode and which is TM mode??
thanks a lot :)
should it be only in 2D, can't we find it in 3D model? one more question: is there a possibility to know which mode has the lowest loss (attenuation) while transmission?? also how can I distinguish between the modes which belong to TE mode and which is TM mode?? thanks a lot :)

Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 22.06.2016, 21:08 GMT-4
1. We generally call them "waveguide modes" when we refer to modes in waveguides, which are by definition idealized infinitely-long structures with constant cross-sections (generalized cylinders). If you are talking about finding eigenmodes in 3D, then you may be talking about "cavity modes". Is that what you meant?

2. Comsol Multiphysics can do that too. Same basic idea, just work in 3D instead.

3. Re TE vs. TM: At a minimum, you should be able to separate the two kinds by making arrow plots of the fields and inspecting the results.

Added later: You can compute losses too, if you make the walls not PEC. You *may* be able to do this in 2D (not sure). But you can certainly do it in 3D. Just make sure you have good quality input and output ports on your waveguide, and you'll have to use numerical ports, if they aren't clean rectangular or circular guides. That's a bit trickier, but I think there are some examples provided by Comsol for you to follow.

1. We generally call them "waveguide modes" when we refer to modes in waveguides, which are by definition idealized infinitely-long structures with constant cross-sections (generalized cylinders). If you are talking about finding eigenmodes in 3D, then you may be talking about "cavity modes". Is that what you meant? 2. Comsol Multiphysics can do that too. Same basic idea, just work in 3D instead. 3. Re TE vs. TM: At a minimum, you should be able to separate the two kinds by making arrow plots of the fields and inspecting the results. Added later: You can compute losses too, if you make the walls not PEC. You *may* be able to do this in 2D (not sure). But you can certainly do it in 3D. Just make sure you have good quality input and output ports on your waveguide, and you'll have to use numerical ports, if they aren't clean rectangular or circular guides. That's a bit trickier, but I think there are some examples provided by Comsol for you to follow.

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 23.06.2016, 04:14 GMT-4
1- yes, I mixed them up. How can I get the cavity modes?.

2- Assume we want to work on a dielectric slab waveguide with different materials for each layer two of them have complex permittivity, do we use the same steps or the applied physics are going to be different?.

3- I want also to draw a graph representing the group velocity of each mode separately as a function of the wavelengths.

4- regarding the TE and TM modes, after simulating and obtaining the eigen frequencies I got all the frequencies that could be existed in the wave guide without differentiating between to which mode it belongs and i got some other doubled frequencies just written (1), (2) after each frequency but actually I didn't understand why it's like that. I have attached you an image with that.
1- yes, I mixed them up. How can I get the cavity modes?. 2- Assume we want to work on a dielectric slab waveguide with different materials for each layer two of them have complex permittivity, do we use the same steps or the applied physics are going to be different?. 3- I want also to draw a graph representing the group velocity of each mode separately as a function of the wavelengths. 4- regarding the TE and TM modes, after simulating and obtaining the eigen frequencies I got all the frequencies that could be existed in the wave guide without differentiating between to which mode it belongs and i got some other doubled frequencies just written (1), (2) after each frequency but actually I didn't understand why it's like that. I have attached you an image with that.


Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 24.06.2016, 11:13 GMT-4
1 and 2. You should be able to find the cavity modes of a 3D rectangular cavity by creating a rectangular box in 3D, and applying the eigenvalue formalism much like you do in 2D. I don't think the basic process will be any different, regardless of what you put inside the box. Just make sure you specify your materials properly.

3. I'm not familiar with the idea of associating a group velocity with a cavity mode. A cavity mode in a rectangular cavity could be considered to be a standing wave, which is a superposition of a forward and reverse wave, in any particular direction. If the cavity is large enough, there can be many such modes, which can correspond to standing waves oriented in various ways. If the cavity is lossless, there is no transfer of energy in any direction at any speed and the bandwidth of each mode collapses to zero (i.e., supports discrete frequencies only) so the "group velocity" (which is defined from d(omega)/d(k)) becomes arguably meaningless anyway. Of course, on the other hand, you can associate a group velocity with a waveguide mode. So it seems, perhaps, that you are actually interested in waveguides, not cavities, after all? So, please clarify: Are you studying cavity modes (oscillating, but not traveling) of a resonant cavity or are you studying propagating modes in a waveguide (traveling waves)? In regard to computing attenuation, are you interested in that because you want to compute the cavity Q? Or are you instead interested in a traveling wave's attenuation with distance, as it travels along a waveguide?

4. The solver can find degenerate modes (more than one mode corresponding to an eigenvalue). You probably have some symmetries in your model that encourage this to occur. Note also that the solver can find "modes" which are not actually distinct modes at all (or even genuinely legitimate modes) but which had numerical properties that fell within a tolerance level that seemed to imply, to the solver, that they were modes. One can typically see the difference by looking at the field plots, if one is familiar with these things. The computer program, however, does not review its results in that manner, nor does it possess any artificial intelligence; it doesn't really know what results to discard and what to keep, if they meet the numerical criteria in its programming.
1 and 2. You should be able to find the cavity modes of a 3D rectangular cavity by creating a rectangular box in 3D, and applying the eigenvalue formalism much like you do in 2D. I don't think the basic process will be any different, regardless of what you put inside the box. Just make sure you specify your materials properly. 3. I'm not familiar with the idea of associating a group velocity with a cavity mode. A cavity mode in a rectangular cavity could be considered to be a standing wave, which is a superposition of a forward and reverse wave, in any particular direction. If the cavity is large enough, there can be many such modes, which can correspond to standing waves oriented in various ways. If the cavity is lossless, there is no transfer of energy in any direction at any speed and the bandwidth of each mode collapses to zero (i.e., supports discrete frequencies only) so the "group velocity" (which is defined from d(omega)/d(k)) becomes arguably meaningless anyway. Of course, on the other hand, you can associate a group velocity with a waveguide mode. So it seems, perhaps, that you are actually interested in waveguides, not cavities, after all? So, please clarify: Are you studying cavity modes (oscillating, but not traveling) of a resonant cavity or are you studying propagating modes in a waveguide (traveling waves)? In regard to computing attenuation, are you interested in that because you want to compute the cavity Q? Or are you instead interested in a traveling wave's attenuation with distance, as it travels along a waveguide? 4. The solver can find degenerate modes (more than one mode corresponding to an eigenvalue). You probably have some symmetries in your model that encourage this to occur. Note also that the solver can find "modes" which are not actually distinct modes at all (or even genuinely legitimate modes) but which had numerical properties that fell within a tolerance level that seemed to imply, to the solver, that they were modes. One can typically see the difference by looking at the field plots, if one is familiar with these things. The computer program, however, does not review its results in that manner, nor does it possess any artificial intelligence; it doesn't really know what results to discard and what to keep, if they meet the numerical criteria in its programming.

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 24.06.2016, 15:14 GMT-4
I'm working on a dielectric waveguide with complex permittivity (no PEC). what i have to study is the modes inside this waveguide and the group velocity of each mode in order to determine which mode has less dispersion and loss during transmission.
I'm working on a dielectric waveguide with complex permittivity (no PEC). what i have to study is the modes inside this waveguide and the group velocity of each mode in order to determine which mode has less dispersion and loss during transmission.

Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 25.06.2016, 13:51 GMT-4
Oh, ok! Based on what you just said, I think you would benefit from studying the details of the "Lossy circular waveguide" example provided by Comsol in the Comsol Application Library, which can be found in the "Transmission Lines and Waveguides" section of the "RF Module" part of the library. Good luck.



Oh, ok! Based on what you just said, I think you would benefit from studying the details of the "Lossy circular waveguide" example provided by Comsol in the Comsol Application Library, which can be found in the "Transmission Lines and Waveguides" section of the "RF Module" part of the library. Good luck.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.