AN6121 questions - antenna trace matching and stubs created by inactive antennae?

[CaptainW]

I'm implementing a children's game board with 6 spaces to place nfc tags with 6 antennae multiplexed as described in AN6121 - which is a great application note and ideal for what I'm doing!   I've done rf and speed layout before but not anything for NFC and I'm a bit confused by a couple of things:

- looking at the photos of the board in AN6121 there seems to be a thick and thin trace going to each antenna, and also a thick trace on the back of the board,  none of these look like a differential trace to me, is it just not that critical in this application or am I missing something here?  I wondered if the thick traces on opposite sides of the boards were matched to each other?

- secondly I'm reading this "As shown in Figure 14, each antenna has its switching system close to the ports. This matching network is located directly on these ports."  to mean the switching should next to the loop antenna,  but this would leave 5 pairs of traces to nowhere and stubs are usually bad news with reflections etc, is this right?   

lastly i saw this

- "The path between the switching, matching network, and EMI filter should be equalized as much as possible."   The reader coils/antennae on this board are placed in a line,  so would i be correct to interpret this as a suggestion to zig zag the shorter lines to coils nearer the chip in order to get the length to match?  

Any help much appreciated - I'm really keen to get this working!

[Travis Palmer · Accepted solution]

Hello CaptainW,

1.) looking at the pictures, both traces seem to have the same thickness (after the selection circuit), as i would do and recommend it:

Bottom:

Top:

2.) Correct, but since NFC is basically DC (13.56MHz) compared to BT, WiFi, UWB or mm wave, it is not super important. The wavelength of NFC is around 22.1 meter.

3.) The ST25R NFC readers are outputting a square signal at their RFOs. The EMI filter is then filtering it and damps higher harmonics of the spectrum. It means that after the EMI filter, mainly the 13.56MHz frequency is present. I think my colleague put one trace on top and one trace on the bottom layer to have them as close as possible together, minimizing radiation.  Also due to the prototyping approach, we have been limited to 2 PCB layers only.

The AN can be seen as a proof of concept and does not perfectly implement all use-cases. Some applications may require that the switching circuit is located close to the ST25R, others can put it directly in front of the antenna. 

I think you have made some good and correct assumptions which are worth to be implemented if the application allows it. 

BR Travis

[CaptainW]

Thanks travis.  I have the luxury of 4 layers so i've routed them through the inside of the board and matched them at 50 ohms.  The board is being made and my fingers are crossed