4
\$\begingroup\$

Can anybody recommend some literature or explain some rules of thumb for tapering planar waveguides? As I understand correct tapering is needed for example for bended waveguides or like in the picture below to increase magnetic ac field of the line in the tapered middle and avoid step discontinuities. Now I can put this geometry in something like HFSS, CST, Sonnet and optimize tapering to get best impedance match, less attentuation. So in general, is it better to make the tapering in steps like in the picture or why not go directly from large ground line width to the smallest one in the middle with only one tapering section, here it's two

enter image description here

\$\endgroup\$

1 Answer 1

Reset to default
3
\$\begingroup\$

Tapering to a reasonable degree is preferable to avoid sharp nodes for a VSWR mismatch. Now, if you taper for too long, then you just have a large amount of line that is a sub-optimal impedance. It really is a matter of what you're shooting for in terms of insertion loss/VSWR/space tradeoffs. I always consult the microwaves101 website for nice rules of thumb but I usually go 5-10% of a wavelength if I have the room. It looks like you have a good two step taper there but you haven't provided any sizing so I can't be sure.

Sometimes, the circuit might want a capacitive stub on either side of an inductive length such as this.

Note: My experience tends to be in the 26 GHz and below range so please take my advice with a grain of salt.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks for your answer +1, my range is 1-3 GHz, signal line width in the middle of the above waveguide 20 micron, gap width 12 micron. Can these tapering effects of VSWR be accurately modeled by EM software packages or can accurate impedance match and needed VSWR mismatch only be reached by a design-measurement-redesign cycle? \$\endgroup\$
    – James Last
    Jul 8, 2013 at 17:50
  • \$\begingroup\$ You should be able to model this in EM software packages. Have you tried this in HFSS or Sonnet? I have more experience with AWR ("2.5" dimensional EM analysis) but if you need the 3D, HFSS should be able to handle what you throw at it. \$\endgroup\$
    – scld
    Jul 8, 2013 at 18:47

Your Answer

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

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