I have the following setup:

  • Very small PCB - 25mm in diameter that hosts chip antenna;
  • I feed directly into antenna matching network with about 7.5cm coax via ufl connector;
  • There are no other components on the board as I'm testing antenna only.

I use miniVNA Tiny to get some meaningful vision into the quality of RF that was proposed by the vendor. I calibrated it at the end of coax, so that I basically measure the antenna only.

Scanning the entire supported range from 1MHz through 3GHz produces the following results:



Which looks too good to be true. The frequencies I'm interested in are around 900MHz, 1.5GHz, 2.5Ghz, ie LoRaWAN, GPS and WiFi, and it looks like all good.

I really would like to play the devil's advocate to make ensure this is nothing wrong with my setup and I'm reading the right values.

What concerns me is that no matter how I try to confuse the antenna either with a metal object nearby or anything similar, fundamentally the outcome is similar in terms of values/diagram.

Update A nice link that has lots of goodies at the bottom https://www.nonstopsystems.com/radio/frank_radio_coax-sw.htm

  • \$\begingroup\$ Can you give us a trace for S11? \$\endgroup\$
    – MAM
    Jan 12, 2021 at 23:26
  • \$\begingroup\$ @AdilMalik will do that later if it's possible with VNA I have on hands, need to remember how to use it properly. Thanks! \$\endgroup\$
    – crtag
    Jan 12, 2021 at 23:52

1 Answer 1


You literally say "I don't trust the swr I measured to represent my system" (and I agree. The worst possible antenna is a perfectly matched termination resistor. Which has perfect swr. I often wonder why vswr is so commonly oversold as metric, especially within the ham radio scene!).

You have a device that calls itself vna - so do some network analysis with it instead of just measuring reflections!

Use your chip antenna as an antenna: use a known to be OK on at least on band antenna as counterpart.

Test s12 with your chip antenna connected to port 1, and more than 1 wavelength away your known antenna, connected to port 2.

If performance isn't worse outside of the bands your other antenna supports, something is fishy.

  • \$\begingroup\$ Thanks for your comment! This does look like perfect load and this is quite confusing. I tried to use it as an antenna (still through the same coax feed) and it looks fine, at least better than no antenna at all, but I'm not sure if this is PCB ground plane playing the trick in a way. Another chip antenna on the same board produces rather different figure, though, it's tuned to a specific bandwidth. The one in the question is promoted as working in any frequency up to 8GHz, so that may answer the question. VNA I have is rather basic with a single port, so I have to see what I can milk from it. \$\endgroup\$
    – crtag
    Jan 12, 2021 at 23:58
  • \$\begingroup\$ Actually I read what you said once again and it's basically the right way to test antennae. That is the subject antenna can be tested two ways - being emitter on DET port and receiver on DUT port with a "standard" antenna on the other one. This should provide more meaningful input into RF behaviour. \$\endgroup\$
    – crtag
    Jan 14, 2021 at 23:21

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.