8
\$\begingroup\$

I'm designing a small board featuring a transmitter in the 900 MHz band. Since RF is definitely not my strongest subject, I'm basically trying to follow all the datasheet recomendations regarding proper layout, PCB traces, etc. However, I find puzzling the reference to an "antenna matching network" I "might" need, according to both the transceiver (AT86RF212B) and the antenna specifications.

Consider the following picture, which can be found in the antenna's reference design:

Matching network in the antenna datasheet

The picture above mentions said matching network, but says nothing about how to design it.

My board currently looks like this; my traces are supposed to match the antenna impedance (50 ohm), and are really short (2.6 mm. from IC pins to balun, 6.3 mm. from balun to antenna, 11 mm. total):

enter image description here

The component in light blue is the antenna, ZT1 is the balun and U1 is the transceiver IC. The IC output is 100 ohm balanced, while the antenna needs a 50 ohm feed.

Given this, what is exactly an antenna matching network, why should I need one and how can I design it? I've tried searching the internet about it but the references I've found were too complex or too abstract for me.

\$\endgroup\$
  • 1
    \$\begingroup\$ If the antenna impedance is not matched to the transmitter, the driving signals will get reflected, and if they are being reflected, they are not being radiated. \$\endgroup\$ – apalopohapa Jun 3 '14 at 21:27
  • \$\begingroup\$ You design it by getting your hands on a Vector Network Analyzer (VNA) and using it to measure the impedance at the frequency of interest. With the help of a smith chart (although not required), you determine the LC component values to make it what you need it to be (50 ohms in this case). \$\endgroup\$ – apalopohapa Jun 3 '14 at 21:29
  • 3
    \$\begingroup\$ Because the antenna will not be a perfect 50 ohms. Even the enclosure you put it in will change its impedance. Without a VNA, I'd recommend replicating a reference design exactly (including PCB thickness and material), and using whatever matching network they used in it. \$\endgroup\$ – apalopohapa Jun 3 '14 at 21:42
  • 1
    \$\begingroup\$ If you scroll to the middle of the following link, you can see an example: lsr.com/white-papers/antenna-matching-within-an-enclosure \$\endgroup\$ – apalopohapa Jun 3 '14 at 21:50
  • 1
    \$\begingroup\$ I'm pretty new at the rf thing, too. I've seen this before, though, with one component in series with the rf trace and another in parallel (to ground). I haven't run across having two components in series... Anyway, calculate your traces, put in the two (or three) component footprints, leave the parallel location empty, and stuff the series locations with 0-Ohm shorts. This should get you close. Then, if the rf matching isn't good enough, you can learn how to build (and tune) the matching network, and you'll already have the pads available. \$\endgroup\$ – bitsmack Jun 3 '14 at 21:58
3
\$\begingroup\$

Honestly, you're probably OK. The trace length is short enough that it's really not too significant, the important part is the balun and transceiver IC output impedance, which is probably only somewhat close to 100 Ohms differential. You may lose some range, but the fact that you're using a chip antenna is probably more significant.

Putting footprints in to play with the match might help, but it might also open up a big can of worms as well.

\$\endgroup\$
  • \$\begingroup\$ Opening a can of worms is exactly my fear, indeed. The antenna, however, is helicoidal, not chip. I'll edit the question with a link to its data sheet. \$\endgroup\$ – Guillermo Prandi Jun 4 '14 at 9:15
  • \$\begingroup\$ so what would you do in my place? Would you skip the LC footprints or put them just in case? Thanks for your thoughts. \$\endgroup\$ – Guillermo Prandi Jun 4 '14 at 21:47
  • \$\begingroup\$ It's 6 of one, half dozen of the other. I'd put the footprints in personally, because I've done this before, and have to tools to fix it. If I didn't have a VNA, or access to a VNA, then I'd call it done, and move onto things I could fix. if you want to see how well it's doing, do an antenna range test and see how close you are to estimates. \$\endgroup\$ – rfdave Jun 5 '14 at 0:57

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.