5
\$\begingroup\$

Before asking the questions, I have read App notes for TI, Cypress and Atmel. Those really helped me clear basics of RF Antenna designs. This is my first time with any RF PCB design, so before embarking on this journey I read as much as I can. Following things are clear to me.

Things I know:

  • Relations bewteen Trace Width, Dielectric Thickness, Trace height, Er => Calculate Impedance
  • No ground plane under Antenna
  • via stitching to GND plane.
  • RF IC -> Balun N/W -> PCB antenna or Chip Antenna
  • Balun's Single Ended Impedance = 50 \$ \Omega \$. So Adjust trace width to match the impedance to 50 \$ \Omega \$.

Things I noticed in any RF IC (BLE) Datasheet:

  • IC's antenna pins are differential in nature and we need to convert it to Single ended to make it work with PCB Antenna.
  • Impedance of these differential pins need not to be 50 \$ \Omega \$. I read somewhere that mostly differential impedance is around 100 \$ \Omega \$. Does that mean that I have to adjust the trace width to match the impedance to 100 \$ \Omega \$? Or I can have any trace width?

Things I Dont Know:

  • Trace Width from Balun's output pin to Feed point of antenna?
  • Should IFA's Trace width needs to be consistence?
  • Trace Width from RF IC antenna pins to balun ? Does it matter? Should I Calculate width of the trace (trace between IC's antenna pins to balun's antenna pins) based on 100 \$ \Omega \$?

I'm just trying to develop a BLE module based on nRF51822. Not much information is available for pcb design guidelines. I did download the reference design and studied that, but I have few questions about that. This is purely for learning purpose. I know that some of the questions are really silly, but for a beginner it matters.

\$\endgroup\$
  • 1
    \$\begingroup\$ Just FYI a 100 ohm differential line still uses 50 ohm traces, the 100 ohms is the total round trip impedance, i.e. 50 ohms up one leg plus 50 ohms down the other leg as opposed to 50 ohms up one leg and bugger all back through the ground plane. \$\endgroup\$ – Sam Sep 30 '16 at 22:18
  • \$\begingroup\$ Hi, So that means i should calculate the trace width based on 50 ohm only. From RF IC's antenna pins to balun to pcb antenna, trace width remains constant for 50 ohm , right? \$\endgroup\$ – Joana Rigbi Oct 1 '16 at 19:36
  • \$\begingroup\$ @Sam Are you certain of that? Differential traces are not simply two 50 ohm traces side by side, as far as I know. There is capacitive and inductive coupling taking place that needs to be considered. While at any one time one trace will be the return leg, it's not the same design as when the ground plane is the return path. \$\endgroup\$ – Brendan Simpson Oct 1 '16 at 19:51
  • 2
    \$\begingroup\$ @BrendanSimpson yes brendan, I know that that's not always the case and that the single ended impedances can be wildly different from the differential impedances, it will even change slightly with frequency as microstrips exhibit some dispersion. But we're trying to keep things simple here and a loosely coupled differential trace is a reasonable approximation of two single ended 50ohm lines when the spacing is a couple times the trace width. \$\endgroup\$ – Sam Oct 1 '16 at 21:58
3
\$\begingroup\$

Impedance of these differential pins need not to be 50 Ω. I read somewhere that mostly differential impedance is around 100 Ω. Does that mean that I have to adjust the trace width to match the impedance to 100 Ω? Or I can have any trace width?

The differential 100Ω output can be converted to single ended 50Ω by using a balun.

Trace Width from Balun's output pin to Feed point of antenna?

Just like all the other RF traces, design it for 50Ω characteristic impedance.

I'm just trying to develop a BLE module based on nRF51822

BLE just uses 2.4GHz and there are dozens if not hundreds of examples of 2.4GHz PCB antenna designs. Heck, you could probably just open up a device and measure the antenna with calipers and get a pretty decent result. Just focus on impedance controlled traces, proper antenna feedline, via stitching and other rules you've mentioned and it should be ok. Also consider putting an SMA or some other type of antenna connector in addition to your antenna to test it in case the PCB antenna doesn't work.

\$\endgroup\$
  • 1
    \$\begingroup\$ You answered things I already know. Not useful at all. What should be the trace impedance from RF IC's antenna pins to balun? Should I calculate the width based on 100ohm or 50? How about IFA's trace width? I saw different trace width in multiple design, why? \$\endgroup\$ – Joana Rigbi Oct 1 '16 at 18:06
  • 1
    \$\begingroup\$ Re: "Not useful at all", You should probably be less critical of people providing you free answers on a free site. Since I'm feeling patient today, the 100Ω differential trace design is different than conventional transmission line designs. I hate to say "just google it" but it's really too involved to describe here, and others have already done a better job explaining it. The 100Ω differential is not simply 2 50Ω tracks side by side, thus the width might be different. Furthermore, the width might depend on lots of other factors. \$\endgroup\$ – Brendan Simpson Oct 1 '16 at 19:51

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.