I am planning on making a board which will have some 2.4GHz RF capabilities, and would like to know if there is a cheap way to tune an IFA cheaply.

If I had access to the right equipment I could easily tune the antenna, but at the moment a multi-thousand dollar VNA is a pipe dream.

Is there anything that can be made or used on the cheap to get a crude antenna tuning going for this board?

Also something I saw which was interesting, on the antenna design specification it claims that no external matching components are needed. I'm assuming it would still be desired to have a pi network to deal with detunings caused by anything else on the board or in the environment right?

  • \$\begingroup\$ Pi networks are used for impedance matching. \$\endgroup\$ – Leon Heller Nov 20 '14 at 16:35
  • \$\begingroup\$ Right and the design note from TI claims that the antenna given is exactly 50 ohms and therefore doesn't require external matching components. I read that as not requiring a pi network, but would it still be advisable to include one? \$\endgroup\$ – Nick Nov 20 '14 at 16:57
  • 1
    \$\begingroup\$ No, it wouldn't. \$\endgroup\$ – Leon Heller Nov 20 '14 at 18:08

The antenna design already includes a matching network. If you need to adjust it, you would fiddle with some of the dimensions given in the design. If you don't know how to do that, just leave it well enough alone.

Adding a second lumped-component matching network would probably add more problems than it would solve, and lead to greater losses than living with whatever mismatch you end up with by using the design as-is.

  • \$\begingroup\$ So without any tuning and keeping exactly to reference designs would you imagine that I could get at least a decent quality signal out of the antenna? \$\endgroup\$ – Nick Nov 20 '14 at 18:04
  • 1
    \$\begingroup\$ Yes, if you dont trust their docs you shouldn't use their chips ;) just do as the paper says \$\endgroup\$ – chwi Nov 20 '14 at 20:14

The antenna linked in the question is also in this document: http://www.ti.com/lit/an/swra496/swra496.pdf. At the top of page 38, it says

No matching components used but still recommend to use with matching components to account for variations in resonance caused by different sizes of GND plane or physical surroundings.

TI's reference design for the CC2650 uses this antenna and includes two component pads for antenna matching. I have also seen a similar recommendation on page 17 of NXP's document on PCB antennas.

Antennas should often be tuned for each product. To do so well seems to require a vector network analyzer and/or simulation. You don't want to make it worse. I'm probably going to try building a homemade VNA that I found (the design files are on Github).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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