Difference gain in FEKO and measure

Estelle NOGUES
Estelle NOGUES Altair Community Member
edited December 2022 in Community Q&A

Hello,

I simulate a MIFA antenna at 868 MHz as below.

image

The problem is that I measure a maximum gain of around 0 dB with FEKO and when I measure the gain with a VNA, I obtain a gain around -8 dB. The measurement has been made in the same conditions.

 

Does this difference can be explain with the loop on the top of the antenna which cancel currents and electric field ?

 

Thank you all

Tagged:

Best Answer

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited December 2022 Answer ✓

    Hi @Estelle NOGUES,

    Simulation with MoM and the mesh as I might recommend does indeed require a lot of memory and runtime. This is also the reason why it is usually not recommended to calculate complex PCB structures with MoM. For complex geometry (many small parts in relation to the wavelength) I would rather advise FEM.

    You can simulate the attached model and let me know if the results meet your expectations. You may have to refine the mesh further. FEM will not require a lot of memory / runtime.

    Best regards,
    Torben

Answers

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited November 2022

    Hi @Estelle NOGUES,

    If you like you may add the model (.cfx) here so I can check if everything is set up correctly.

    Best regards,
    Torben Voigt

  • Estelle NOGUES
    Estelle NOGUES Altair Community Member
    edited November 2022

    Hi @Torben Voigt, 

    You can find the file attached.

    Have a good day, 

    Estelle

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited November 2022

    Thanks! Which version of Feko are you using?

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited November 2022

    Hi @Estelle NOGUES ,

    Could you also attach the Matching_a868.cir file?

    Your meshing is very coarse. The metal faces are very close to each other, so they should be meshed finer. Try applying a mesh size of e.g. 5 to all metal faces (despite the ones you've already applied a refinement to).

    Also it may be necessary to use double precision, since the model is electrically quite small.

    Best regards,
    Torben

  • Estelle NOGUES
    Estelle NOGUES Altair Community Member
    edited November 2022

    Hi @Torben Voigt, 

    I do not use the matching circuit, it was for an old simulation. I display result using gain and not realised gain to not take account of the mismatch of the antenna.

    I try with a mesh size of 5 and the result is the same.

    The double precision simulation is too long, my pc can't run this.

    Best regards, 

    Estelle

  • Torben Voigt
    Torben Voigt
    Altair Employee
    edited December 2022 Answer ✓

    Hi @Estelle NOGUES,

    Simulation with MoM and the mesh as I might recommend does indeed require a lot of memory and runtime. This is also the reason why it is usually not recommended to calculate complex PCB structures with MoM. For complex geometry (many small parts in relation to the wavelength) I would rather advise FEM.

    You can simulate the attached model and let me know if the results meet your expectations. You may have to refine the mesh further. FEM will not require a lot of memory / runtime.

    Best regards,
    Torben

  • Estelle NOGUES
    Estelle NOGUES Altair Community Member
    edited December 2022

    Hi @Torben Voigt, 

    We did again the measurement and we realized that the coax cable varies the gain of 3 to 5 dB. Results obtain with Feko correspond with measurement.


    Thank you for you help and your advice as using FEM method instead of MoM.

    Best regards, 

    Estelle

Categories