+ Post New Thread
Results 1 to 17 of 17
  1. #1
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Half-mode SIW Simulation

    Hello,

    I am having a problem on simulating a half-mode SIW in HFSS.
    Since we cut the width by half, the resulting structure has cut-off frequency double the original one.
    But all papers said they have very close responses.
    Any idea how should I simulate the half-mode structure?

    Thank you for your help!

  2. #2
    Newbie level 6
    Points: 296, Level: 3

    Join Date
    Aug 2018
    Posts
    12
    Helped
    2 / 2
    Points
    296
    Level
    3

    Re: Half-mode SIW Simulation

    hello, the em fields are similar to that in HF siw. If you cut the siw into two identical ones, the fields almost have no change if w/h is very large.



  3. #3
    Full Member level 3
    Points: 1,203, Level: 7

    Join Date
    Oct 2017
    Posts
    160
    Helped
    26 / 26
    Points
    1,203
    Level
    7

    Re: Half-mode SIW Simulation

    Could you post an image of your model? You probably have conducting wall on the plane of cut, so it halves waveguide width and doubles frequency. In this case solution would be removing conductive wall on that side.


    1 members found this post helpful.

  4. #4
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Thanks for your answers!
    Please see two images of the structure

    Click image for larger version. 

Name:	Top.JPG 
Views:	5 
Size:	34.4 KB 
ID:	156349
    Click image for larger version. 

Name:	3D.JPG 
Views:	4 
Size:	131.9 KB 
ID:	156350

    Quote Originally Posted by Georgy.Moshkin View Post
    In this case solution would be removing conductive wall on that side.
    Could you please tell me how to do it?
    Thank you for your help!



    •   AltAdvertisement

        
       

  5. #5
    Full Member level 3
    Points: 1,203, Level: 7

    Join Date
    Oct 2017
    Posts
    160
    Helped
    26 / 26
    Points
    1,203
    Level
    7

    Re: Half-mode SIW Simulation

    I am not familiar with HFSS, just guessing from your image: check this vertical thin wall is not conductor, ensure there are no any hidden walls inside substrate on this side:
    Click image for larger version. 

Name:	hmsiw.png 
Views:	3 
Size:	296.9 KB 
ID:	156351
    What is outer material on this wall? PML? or maybe HFSS puts PEC (conductor) by default?



  6. #6
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Quote Originally Posted by Georgy.Moshkin View Post
    check this vertical thin wall is not conductor, ensure there are no any hidden walls inside substrate on this side:
    This wall is the substrate, no conductor here except at the bottom of the wall, there is a bottom horizontal ground (which is a conductor).

    Quote Originally Posted by Georgy.Moshkin View Post
    What is outer material on this wall? PML? or maybe HFSS puts PEC (conductor) by default?
    I didn't define any boundaries or excitations to this wall. I tried to draw a vertical rectangle there and define it as "perfect H" and "symmetry perfect H" but the responses were almost the same.
    Based on papers, the high ration of w/h will act as a PMC by default.



  7. #7
    Full Member level 3
    Points: 1,203, Level: 7

    Join Date
    Oct 2017
    Posts
    160
    Helped
    26 / 26
    Points
    1,203
    Level
    7

    Re: Half-mode SIW Simulation

    You may try this:
    1. extend dielectric substrate on this wall (near Y axis), only conductor is cut but not dielectric.
    2. check EM-fields distribution in HFSS, result must look similar to this image:
    Click image for larger version. 

Name:	hmsiw2.png 
Views:	0 
Size:	109.6 KB 
ID:	156355
    (image from Yuanqing Wang, Wei Hong "Half Mode Substrate Integrated Waveguide (HMSIW) Bandpass Filter")
    As you can see those field distribution "blobs" are cut by half. If they are stretched along Y axis, then boundary is wrong.

    We need someone who familiar with HFSS. It is unclear how top and bottom surfaces of substrate are handled, what boundaries are put by HFSS for this model. EM-filed distribution image may help to understand this problem.



  8. #8
    Advanced Member level 4
    Points: 6,895, Level: 19
    Achievements:
    7 years registered

    Join Date
    Jun 2012
    Location
    Edmonton, Canada
    Posts
    1,059
    Helped
    302 / 302
    Points
    6,895
    Level
    19

    Re: Half-mode SIW Simulation

    Can you please show us:

    The fields of the mode(s) you are exciting.
    The surrounding radiation box (extent of the simulation domain).



    •   AltAdvertisement

        
       

  9. #9
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Quote Originally Posted by PlanarMetamaterials View Post
    The fields of the mode(s) you are exciting.
    Click image for larger version. 

Name:	Port 1 Field.JPG 
Views:	1 
Size:	52.2 KB 
ID:	156381
    Click image for larger version. 

Name:	E_Field.JPG 
Views:	1 
Size:	105.8 KB 
ID:	156382


    Quote Originally Posted by PlanarMetamaterials View Post
    The surrounding radiation box (extent of the simulation domain).
    Click image for larger version. 

Name:	Radiation Box1.JPG 
Views:	2 
Size:	48.0 KB 
ID:	156383
    Click image for larger version. 

Name:	Radiation Box2.JPG 
Views:	0 
Size:	132.0 KB 
ID:	156384



  10. #10
    Advanced Member level 4
    Points: 6,895, Level: 19
    Achievements:
    7 years registered

    Join Date
    Jun 2012
    Location
    Edmonton, Canada
    Posts
    1,059
    Helped
    302 / 302
    Points
    6,895
    Level
    19

    Re: Half-mode SIW Simulation

    Thank you.

    The setup and solution appear to be correct -- so what is the issue, exactly?

    Can you plot the S-parameters of this setup and that of a similar full-mode SIW?



  11. #11
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Thank you for your feedback.
    Here are the S-parameters of the full-mode SIW (calculated at 50 ohms port impedance) using two different tapered widths.
    Click image for larger version. 

Name:	S_full.JPG 
Views:	1 
Size:	170.6 KB 
ID:	156385

    Next, S-parameters for the HMSIW
    Click image for larger version. 

Name:	S_HM.JPG 
Views:	1 
Size:	171.0 KB 
ID:	156386

    My main problem is that when I simulated these two structures without the tapered microstrip feedlines, the cut-off frequency will be doubled!
    But as you can see in Fig. 7 in the paper below, it should be the same.
    http://http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.728.5011&rep=rep1&type=pdf



  12. #12
    Advanced Member level 4
    Points: 6,895, Level: 19
    Achievements:
    7 years registered

    Join Date
    Jun 2012
    Location
    Edmonton, Canada
    Posts
    1,059
    Helped
    302 / 302
    Points
    6,895
    Level
    19

    Re: Half-mode SIW Simulation

    Quote Originally Posted by abuzaid10 View Post
    Here are the S-parameters of the full-mode SIW (calculated at 50 ohms port impedance) using two different tapered widths.
    They look to be about the same (cutoff around 5 GHz).

    Quote Originally Posted by abuzaid10 View Post
    when I simulated these two structures without the tapered microstrip feedlines, the cut-off frequency will be doubled!
    So you showed us the simulations that worked, but not the ones that didn't work. How are we supposed to help you? Please show us the same data and figures for the cases that didn't work as you expected.



  13. #13
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Thank you for your reply.
    Here is the structures and S-parameters as well as to beta (which shows the cut-off frequency) of the Half-mode and full-mode structures.

    Full-mode:
    Click image for larger version. 

Name:	FM_structure.JPG 
Views:	1 
Size:	51.8 KB 
ID:	156403 Click image for larger version. 

Name:	FM_Port1.JPG 
Views:	1 
Size:	18.6 KB 
ID:	156404
    Click image for larger version. 

Name:	FM_response.JPG 
Views:	2 
Size:	142.9 KB 
ID:	156405 Click image for larger version. 

Name:	FM_beta.JPG 
Views:	4 
Size:	131.9 KB 
ID:	156406

    Half-mode:
    Click image for larger version. 

Name:	HM_structure.JPG 
Views:	2 
Size:	39.3 KB 
ID:	156411 Click image for larger version. 

Name:	HM_Port1.JPG 
Views:	1 
Size:	14.7 KB 
ID:	156408
    Click image for larger version. 

Name:	HM_response.JPG 
Views:	2 
Size:	152.7 KB 
ID:	156409 Click image for larger version. 

Name:	HM_beta.JPG 
Views:	4 
Size:	146.0 KB 
ID:	156410



  14. #14
    Advanced Member level 4
    Points: 6,895, Level: 19
    Achievements:
    7 years registered

    Join Date
    Jun 2012
    Location
    Edmonton, Canada
    Posts
    1,059
    Helped
    302 / 302
    Points
    6,895
    Level
    19

    Re: Half-mode SIW Simulation

    Alright, so your half-mode data is not self-consistent: the S-parameters clearly indicate that the cutoff frequency is around 5 GHz, but the Beta plot is showing that you have a cutoff around 10 GHz.

    Quite possibly it's a simulation error in the wave port: sometimes they will solve for one mode at the convergence frequency, but another will be excited in your frequency sweep. What is your simulation setup (convergence frequencies, etc.)?

    Could you please show the port fields on both ports at 5.8 GHz and 11 GHz. Alternatively, upload the HFSS file and I can take a look at it.


    1 members found this post helpful.

  15. #15
    Junior Member level 1
    Points: 717, Level: 5

    Join Date
    May 2016
    Posts
    16
    Helped
    0 / 0
    Points
    717
    Level
    5

    Re: Half-mode SIW Simulation

    Thank you very much for your help.

    Please find attached the HFSS files of the half-mode and full-mode structures.
    SIW.rar



    •   AltAdvertisement

        
       

  16. #16
    Full Member level 3
    Points: 1,203, Level: 7

    Join Date
    Oct 2017
    Posts
    160
    Helped
    26 / 26
    Points
    1,203
    Level
    7

    Re: Half-mode SIW Simulation

    From S-plot it is obvious that cutoff is around 5GHz, but it is not matched. Your Half-mode SIW S-plot shows big insertion loss around 5GHz. Probably placing waveguide port directly on HMSIW edge results in big mismatch, because port width is halved, but "magnetic wall" and symmetry does not act for port the same way as for waveguide. Solution would be to change something in port configuration, or to use transition as in your earlier post.


    1 members found this post helpful.

  17. #17
    Advanced Member level 4
    Points: 6,895, Level: 19
    Achievements:
    7 years registered

    Join Date
    Jun 2012
    Location
    Edmonton, Canada
    Posts
    1,059
    Helped
    302 / 302
    Points
    6,895
    Level
    19

    Re: Half-mode SIW Simulation

    Click image for larger version. 

Name:	Screen Shot 2019-11-12 at 9.15.33 AM.png 
Views:	2 
Size:	115.6 KB 
ID:	156471

    Ok, so look at this figure. Your port is exciting a full mode in the HM-SIW, which is why it shows cutoff around 10 GHz.

    There's not an obvious reason why this is happening, it's just HFSS not getting the correct mode. You can probably resolve this by changing the convergence frequency (you have it at 6 GHz, I'd set it lower to about 4).

    Also, the port won't include fringing fields coming out of the HM open edge. I'd recommend enlarging your port such that it encompasses more of the modal fields.



    Quote Originally Posted by Georgy.Moshkin View Post
    From S-plot it is obvious that cutoff is around 5GHz, but it is not matched. Your Half-mode SIW S-plot shows big insertion loss around 5GHz. Probably placing waveguide port directly on HMSIW edge results in big mismatch, because port width is halved, but "magnetic wall" and symmetry does not act for port the same way as for waveguide. Solution would be to change something in port configuration, or to use transition as in your earlier post.
    Indeed, this is why HFSS introduced "symmetry boundaries", which are just perfect E/H, but also modify port impedance.


    1 members found this post helpful.

--[[ ]]--