rectangular waveguide error

niharika_kaja
niharika_kaja Altair Community Member
edited August 14 in Community Q&A

I have simulated two versions of  rectangular waveguide from 3-6GHz and just cannot seem to figure out what is causing a bad impedance mismatch. I want to improve the s11 of this file attached.

Tagged:

Answers

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited August 14

    Hi @Niharika Kaja,

    The first thing I notice is that the waveguide is closed at the rear end by a PEC wall. This means that all energy is reflected (s11 = 1):

    image

    It is better to use a second waveguide port here and define a SParameterConfiguraton, see attached model Rectangular_waveguide_3-6ghz_alt.cfx.

    By the way, you can use two symmetry planes if the waveguide is moved by b/2 in negative z-direction. This will reduce runtime and memory a lot:

    image

    An even more efficient method for such closed waveguides is FEM, see attached model Rectangular_waveguide_3-6ghz_FEM.cfx.

    You can see the differences between the s11 of MoM and FEM, which may seem large to you. Remember that in both cases s11 is approximately 0 or in the range of -80 dB. If you plot the difference between the results (abs(s11_MoM - s11_FEM)), you will see that it is a maximum of only 0.045:

    image

    image

    By refining the MoM mesh, the differences could become even smaller, but with FEM you have a reliable and fast variant.

    Best regards,
    Torben

  • niharika_kaja
    niharika_kaja Altair Community Member
    edited August 14
    Torben Voigt_20420 said:

    Hi @Niharika Kaja,

    The first thing I notice is that the waveguide is closed at the rear end by a PEC wall. This means that all energy is reflected (s11 = 1):

    image

    It is better to use a second waveguide port here and define a SParameterConfiguraton, see attached model Rectangular_waveguide_3-6ghz_alt.cfx.

    By the way, you can use two symmetry planes if the waveguide is moved by b/2 in negative z-direction. This will reduce runtime and memory a lot:

    image

    An even more efficient method for such closed waveguides is FEM, see attached model Rectangular_waveguide_3-6ghz_FEM.cfx.

    You can see the differences between the s11 of MoM and FEM, which may seem large to you. Remember that in both cases s11 is approximately 0 or in the range of -80 dB. If you plot the difference between the results (abs(s11_MoM - s11_FEM)), you will see that it is a maximum of only 0.045:

    image

    image

    By refining the MoM mesh, the differences could become even smaller, but with FEM you have a reliable and fast variant.

    Best regards,
    Torben

    Thank you, by making the above changes, I could obtain result with MoM.

Categories