1
\$\begingroup\$

I am trying to design a microstrip bandpass filter in ADS with a center frequency of 2.4GHz and a fractional BW of 10%. My results from the schematic look great: enter image description here

When I move to the layout and EM simulation, my results look nothing close to a bandpass filter. I am not sure if it is something with my setup or the physical layout. The schematic and layout is shown below: enter image description here

enter image description here

The results from the EM simulation are here: enter image description here

I was thinking it could be with how close the spacing is between the coupled lines, but if anyone can point me in the right direction it would be greatly appreciated

\$\endgroup\$
5
  • \$\begingroup\$ This is out of my field. but a guy that is true scientist makes those lines little bit different, surely there are also exact formulas, but I was not his student. Look for some his examples: s53mv.s5tech.net/cavity/cavity.html s53mv.s5tech.net/zifssb/block.html s53mv.s5tech.net/zifssb/lscband.html \$\endgroup\$ Commented Dec 27, 2020 at 0:24
  • \$\begingroup\$ I don't have the resources to use a cavity \$\endgroup\$
    – user267451
    Commented Dec 27, 2020 at 0:32
  • 1
    \$\begingroup\$ Your first and last couple sections look like they have no gap at all. I don't know why, but that is almost surely the problem. \$\endgroup\$
    – The Photon
    Commented Dec 27, 2020 at 0:41
  • \$\begingroup\$ Make sure the mesh size in your EM simulation is smaller than the 50 um gap you are trying to model. \$\endgroup\$
    – The Photon
    Commented Dec 27, 2020 at 0:44
  • \$\begingroup\$ Only the 1st link is about the cavity. It always uses at least 4 strips, two with small gap and between other two a larger gap. And also tells that different PCB materials behave different, everything is in boxed with a kind of absorber foam. \$\endgroup\$ Commented Dec 27, 2020 at 0:53

1 Answer 1

0
\$\begingroup\$

EM simulation is to pick cells per wave length. Typical filter centered at 10 GHz , then choose meshing at 20 Ghz and 50 cells per wavelegth

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.