Menu

Discussion Forum

New Discussion
Filter

Phase at output Port

Posted Nov 20, 2023, 10:42 a.m. EST 9 Replies

Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello,

I am simulating mode conversion ( TE10 to TE20) on a hollow curved waveguide of rectangular section. On the Port section I inserted Phase=0 for both ports (input and output). I don't understand why on Port 2 (the output Port) the field is always (almost) vanishing even if I change wavelength, i.e. exp(ikz_port2)=0 (where z is the direction of the waveguide axis). Intuitively I would expect that, based on the relation bewteen wavelength and length of the guide, on Port 2 the solution would change the value of the exponential part (i.e. the part propotional to exp(ikz)). Why this behaviour?

Thanks in advance

Attachments:

Reply
9 Replies Last Post Nov 29, 2023, 1:47 p.m. EST
Robert Koslover Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 20, 2023, 3:54 p.m. EST
Updated: 5 months ago Nov 20, 2023, 4:07 p.m. EST

What did you specify in terms of the detailed boundary conditions on your ports? If both TE10 and TE20 modes are above cutoff, then to model this properly, and assuming a clean launch of TE10 at one end, then you need to specify TE10 and TE20 port conditions on both your physical ports, but with distinct port numberings for each distinct mode.

Input port (to be overloaded with two port specifications): Port #1: Specify TE10, with Port "ON" (Note: Only this one is set to "ON"). Port #2 (select same physical port as Port #1!!): Specify TE20, with Port "Off" (This lets it absorb reflected waves in TE20 mode).

Output port (again, to be overloaded with two port specifications): Port #3: Specify TE10, with Port "OFF" (This will absorb TE10). Port #4 (select same physical port as Port #3!!): Specify TE20, with Port "OFF" (This will absorb TE20).

This allows your single excitation of a TE10 mode at your input port to accept: 1: Reflections, if any (to the input port) as a TE10 mode. 2: Reflections, if any (to the input port) as a TE20 mode. 3. Transmission (to the output port) as a TE10 mode. 4. Transmission (to the output port) as a TE20 mode.

The s parameter corresponding to TE20 output (at the output port), if given a TE10 input (at the input port), will then be emw.S41 . You can (and should) use the argument of that complex quantity to find the phase. Similarly, you can also quantify amplitudes and phases of the reflections and the transmitted TE10. To extend to even higher modes, specify more ports. (I've overloaded ports with up to dozens of modes, so I know this works.)

If this advice isn't enough, then I encourage you to post your .mph file to the forum.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
What did you specify in terms of the detailed boundary conditions on your ports? If both TE10 and TE20 modes are above cutoff, then to model this properly, and assuming a clean launch of TE10 at one end, then you need to specify TE10 and TE20 port conditions on *both* your *physical* ports, but with *distinct port numberings* *for each distinct mode*. Input port (to be overloaded with two port specifications): Port #1: Specify TE10, with Port "ON" (Note: Only this one is set to "ON"). Port #2 (select *same* physical port as Port #1!!): Specify TE20, with Port "Off" (This lets it absorb reflected waves in TE20 mode). Output port (again, to be overloaded with two port specifications): Port #3: Specify TE10, with Port "OFF" (This will absorb TE10). Port #4 (select *same* physical port as Port #3!!): Specify TE20, with Port "OFF" (This will absorb TE20). This allows your single excitation of a TE10 mode at your input port to accept: 1: Reflections, if any (to the input port) as a TE10 mode. 2: Reflections, if any (to the input port) as a TE20 mode. 3. Transmission (to the output port) as a TE10 mode. 4. Transmission (to the output port) as a TE20 mode. The s parameter corresponding to TE20 output (at the output port), if given a TE10 input (at the input port), will then be emw.S41 . You can (and should) use the argument of that complex quantity to find the phase. Similarly, you can also quantify amplitudes and phases of the reflections and the transmitted TE10. To extend to even higher modes, specify more ports. (I've overloaded ports with up to dozens of modes, so I know this works.) If this advice isn't enough, then I encourage you to post your .mph file to the forum.
Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 21, 2023, 6:34 a.m. EST

Hello,

I followed the procedure you indicated me, but unfortunately the emw.S41dB parameter is -inf (this would indicate zero trasmission of TE20 to output port and it seems strange). I attach the .mph model (I cleared mesh and solution but the model is very fast to run)

Hello, I followed the procedure you indicated me, but unfortunately the emw.S41dB parameter is -inf (this would indicate zero trasmission of TE20 to output port and it seems strange). I attach the .mph model (I cleared mesh and solution but the model is very fast to run)

Attachments:

Robert Koslover Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 21, 2023, 10:40 a.m. EST
Updated: 5 months ago Nov 21, 2023, 11:05 a.m. EST

In your waveguide, the cutoff frequency for the TE20 mode is 52.7804 GHz. You were running your model at 50 GHz. This is below the cutoff frequency, so TE20 will not propagate. Either make your waveguide wider or run your model at a higher frequency, if you want to allow the possibility of a TE20 mode to appear.

p.s. On a separate matter, I notice you are plotting sqrt((emw.Ez)^2+(emw.Ex)^2+(emw.Ez)^2). That is not the norm of the field, since E is complex. Instead, I would encourage you to plot emw.normE . If you want to "see" the wave in the guide, then (in your case) plot emw.Ez

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
In your waveguide, the cutoff frequency for the TE20 mode is 52.7804 GHz. You were running your model at 50 GHz. This is *below* the cutoff frequency, so TE20 will not propagate. Either make your waveguide wider or run your model at a higher frequency, if you want to allow the possibility of a TE20 mode to appear. p.s. On a separate matter, I notice you are plotting *sqrt((emw.Ez)^2+(emw.Ex)^2+(emw.Ez)^2)*. That is not the norm of the field, since E is complex. Instead, I would encourage you to plot *emw.normE* . If you want to "see" the wave in the guide, then (in your case) plot *emw.Ez*
Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 23, 2023, 3:11 a.m. EST

Thanks for your answer, I have a last question. With the settings that you have suggested me, the S (in dB units) matrix will have the following meaning: 1)S31 will quantify the attenuation of TE10 at output with respect to the TE10 excitation at input, so S31 close to zero is desired 2)S41 will quantify the transmission of TE20 at output with respect to the TE10 excitation at input, so S41 as negative as possible is desired 3)S11 will quantify the reflection of TE10 at input with respect to its excitation on same port, so S11 as negative as possible is desired Is my interpretation correct?

Thanks for your answer, I have a last question. With the settings that you have suggested me, the S (in dB units) matrix will have the following meaning: 1)S31 will quantify the attenuation of TE10 at output with respect to the TE10 excitation at input, so S31 close to zero is desired 2)S41 will quantify the transmission of TE20 at output with respect to the TE10 excitation at input, so S41 as negative as possible is desired 3)S11 will quantify the reflection of TE10 at input with respect to its excitation on same port, so S11 as negative as possible is desired Is my interpretation correct?
Robert Koslover Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 24, 2023, 12:39 p.m. EST
Updated: 5 months ago Nov 24, 2023, 12:38 p.m. EST

If your goal is to avoid mode conversion to TE20, then yes.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
If your goal is to avoid mode conversion to TE20, then yes.
Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 28, 2023, 6:22 a.m. EST

Hi,

I now would want to understand the attenuation of my modes due to Ohmic losses given by the material (copper) which surrounds the waveguide. For a first try I modeled a WR42 straight Wg of 0.5m and I used the Impedence boundary condition to simulate Ohmic Losses. From what I understood, the attenuation is quantified (for a given mode) by the beta attenuation coefficient. From the solution I obtain 4 beta coefficient (but I have two modes, i.e. TE10 and TE20), and the coefficient beta1 (which I expect to be related to TE10 attenuation) is very large resulting in a fast dissipation of TE10, which is not seen in the E-field pattern where TE10 seems dominating. Is my error in the interpretation of the beta coefficients or in some setting of the model? You find the .mph model attached

P.S: Using Impedence Boundary Condition I noticed that some properties of the built-in Copper material are named with a different name from the one required by impedance boundary conditions, so one can not select "From Material" for the properties

Hi, I now would want to understand the attenuation of my modes due to Ohmic losses given by the material (copper) which surrounds the waveguide. For a first try I modeled a WR42 straight Wg of 0.5m and I used the Impedence boundary condition to simulate Ohmic Losses. From what I understood, the attenuation is quantified (for a given mode) by the beta attenuation coefficient. From the solution I obtain 4 beta coefficient (but I have two modes, i.e. TE10 and TE20), and the coefficient beta1 (which I expect to be related to TE10 attenuation) is very large resulting in a fast dissipation of TE10, which is not seen in the E-field pattern where TE10 seems dominating. Is my error in the interpretation of the beta coefficients or in some setting of the model? You find the .mph model attached P.S: Using Impedence Boundary Condition I noticed that some properties of the built-in Copper material are named with a different name from the one required by impedance boundary conditions, so one can not select "From Material" for the properties
Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 29, 2023, 2:53 a.m. EST

Sorry I forgot the attachment, you find it here

Sorry I forgot the attachment, you find it here

Attachments:

Federico Ruffini
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 29, 2023, 12:51 p.m. EST

Hi,

I have, as always, other doubts. When adding a Port and specificying the mode, the equation box has a coefficient S which is given in terms of the integral of the field over the boundary of a physical surface (the boundary of the physical port I guess) . The integral contains a field (E1 for example if the port is port 1). Is that field the one of the mode specified on the port? The most important doubt is about the various coefficient S11, S21 ecc. How are they computed? I mean what is their expression? I guess they are related to the coefficients S reported on each port, but how? Is S_ij=S_i/S_j where S_i is the S reported on port i and S_j the S reported on port j? This informations are not reported in any COMSOL manual/guide. It means the physics underlying every option is not reported anywhere. How can I use properly the software if the meaning of each variable/ calculation/options are either not reported or reported in a cryptic way? Every scientist I spoke to said me that the mechanisms with which COMSOL works are a mistery and one can not fully understand the options. It may be very useful if a manual, containing a deep discussion of the RF module functioning, is released.

Hi, I have, as always, other doubts. When adding a Port and specificying the mode, the equation box has a coefficient S which is given in terms of the integral of the field over the boundary of a physical surface (the boundary of the physical port I guess) . The integral contains a field (E1 for example if the port is port 1). Is that field the one of the mode specified on the port? The most important doubt is about the various coefficient S11, S21 ecc. How are they computed? I mean what is their expression? I guess they are related to the coefficients S reported on each port, but how? Is S_ij=S_i/S_j where S_i is the S reported on port i and S_j the S reported on port j? This informations are not reported in any COMSOL manual/guide. It means the physics underlying every option is not reported anywhere. How can I use properly the software if the meaning of each variable/ calculation/options are either not reported or reported in a cryptic way? Every scientist I spoke to said me that the mechanisms with which COMSOL works are a mistery and one can not fully understand the options. It may be very useful if a manual, containing a deep discussion of the RF module functioning, is released.
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 5 months ago Nov 29, 2023, 1:47 p.m. EST
Updated: 5 months ago Nov 29, 2023, 2:10 p.m. EST

Hello Federico,

Hmmm. The mathematical definition of, eg, emw.S11, can be viewed in the GUI. Ctrl+F will take you there. See also page 85 and following of the RF Module User's Guide, version 6.2. You can access the documentation through File > Help > Documentation.

Best,

Jeff

-------------------
Jeff Hiller
Hello Federico, Hmmm. The mathematical definition of, eg, emw.S11, can be viewed in the GUI. Ctrl+F will take you there. See also page 85 and following of the RF Module User's Guide, version 6.2. You can access the documentation through File > Help > Documentation. Best, Jeff

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

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.

Suggested Content