What is the best way to multiplex antennas with one reader IC?

[smithderek2000]

Hello,

I have an application whereby I need to use one reader IC with 10 antennas. It does not need to read very quickly, once per second or so. What is the best way to do this? Each antenna will be about 1.5" (38mm) spacing in a line. So the length of the farthest antenna is about 15" (380mm) from the IC. Application is like a vending machine. Should I:

a) build a antenna multiplexer from relays or RF switches right at the IC?

or

b) put each relay right next to the antenna?

Or is there another way?

Read distance is low, about 20mm to a tag approximately 20mm in diameter.

Thanks,

Derek

[Travis Palmer · Accepted solution]

Hello Derek,

In general for this application i would recommend the ST25R3916(B).

Having the mux / relay between EMC filter and NFC IC bears the risk of EMC problems.

The RFO outputs quite a fast rectangular signal. Thinking about Fourier, the rectangular signal is build out of multiple sine waves at multiple frequencies. The EMC filter removes / attenuates all of them and only the 13.56MHz sine will pass.

From this perspective, putting your switching circuit after the EMC filter seems to me the safest approach. The Vpp voltage there will be also lower than the one at the Antenna. You can have different CP and CS in order to adapt the matching of the different antennas.

You would now have many possibilities.

1. mux the differential signal after the EMC Filter (10x differential mux)
2. use the IC single ended and use two mux connected to each branch (2x5x single ended mux)
3. like 2 but use the 50Ohm approach (cable + machting + 50Ohm antenna) (2x5x single ended 50Ohm mux) (https://eds.st.com/nfc-tuningcircuit/ => ST25R3916 - single ended with cable)

To have less losses, your mux should be as low ohmic as possible.

You could try to connect the RFI between series cap and the EMC filter. You then would not need to switch the Rx path.

In my opinion, 2 and 3 are the most reason ways to go. 3 is for sure the most expensive one (due to the 50ohm coaxial cable) but most likely the most reliable one.

Approach number 2 should also work and there should be a big quantity of mux available. You can also build your own, based on relays. Best would be if the mux is not loading the antenna, if it is not selected (coupling and loading between selected antenna and un-selected antenna nearby).

Hopefully i answered all your questions.

br Travis

[smithderek2000]

Thank you for your response. We were going to use the ST25R3911 but now I'm looking at the '3916(B). The only downside of the '3916(B) is that there are no evaluation boards available so it makes it difficult to test our proposed antenna and mux combinations. All the antennas will be PCB traces; little 15mm diameter coils that we're testing now. I'm leaning towards the relay option: it's very low loss and minimizes load on the antenna line when disconnected. For the 50 ohm option I looked into RF muxes and there are quite a few. Interestingly, these are all rated for minimum of 100MHz, and here we're operating at 13.56MHz.

I agree, best place for the mux is after the EMC filter; basically right before the CP & CS tuned for each antenna.

[Travis Palmer]

Hello Derek,

The ST25R3916B boards are brand new and should be available soon.

Whatever you develop on the ST25R3911B-DISCO will also work on ST25R3916(B) boards.

br Travis

[MMont.8]

Hello, I'm developing a solution similar to this one with 5 antennas instead. What is the best way to route the RFI path? Since it's difficult to 100 per cent match the lenght of the RFI differential paths in all 5 antennas, do you think this may be a source of problem if there is a slight mismatch in both tracks?

[smithderek2000]

Hello,

I have it working with one ST253916B rfic and sixteen (16) PCB antennas. It's working very well. I used DPDT relays, Omron G6K-2F-TR-DC5, because they were low loss. The system is pretty resilient. I didn't do any length matching. I basically have a long bus, about 23 inches long with the transceiver in the middle and relays every 1.5" on either side. I haven't done any tuning or anything yet but it's working great. I can read all 16 channels in less than 10 seconds. If you need to read faster than that then you can refactor the software to use an RTOS and get it very fast. However, It's quite a bit of work according to someone I know who has done it. Also, because relays eventually wear down, you don't want to constantly scan the channels. You can also hear the relays a little bit, clicky clacky.

The relays are a bit pricey ($3 each) so for cost reduction I'm going to look into a Skyworks antenna switcher. But most of those aren't designed to work down at 13MHz.

[smithderek2000]

Hello,

I have it working with one ST253916B rfic and sixteen (16) PCB antennas. It's working very well. I used DPDT relays, Omron G6K-2F-TR-DC5, because they were low loss. The system is pretty resilient. I didn't do any length matching. I basically have a long bus, about 23 inches long with the transceiver in the middle and relays every 1.5" on either side. I haven't done any tuning or anything yet but it's working great. I can read all 16 channels in less than 10 seconds. If you need to read faster than that then you can refactor the software to use an RTOS and get it very fast. However, It's quite a bit of work according to someone I know who has done it. Also, because relays eventually wear down, you don't want to constantly scan the channels. You can also hear the relays a little bit, clicky clacky.

The relays are a bit pricey ($3 each) so for cost reduction I'm going to look into a Skyworks antenna switcher. But most of those aren't designed to work down at 13MHz.

[Travis Palmer]

Hello Smithderek2000,

thanks for sharing your experience. I think you did a quite nice application.

Same could be done with some multiplexer ICs or discrete.

Unfortunately finding a mux which is capable of dealing with the high voltages and power requirements is quite hard. Also the better their RF characteristics is, the higher the price is.

BR Travis

[Travis Palmer]

Hello MMont.8,

I think you do not need to match the length of all RFI tracks, only the once which are enabled at the same time. E.g. if antenna 1 is active, only the RFIs of Antenna 1 are enabled via the mux/relay. Those two RFI lines have to be approximately the same length to each other.

(depended if you have chosen a single ended or differential approach).

br Travis

[MMont.8]

Hello again smithderek2000 and Travis Palmer, thank you for sharing your help, i really appreciate it! I'm using the differential approach with mosfets controlled by a microcontroller to switch each antennas path on and off the circuit. The mosfets i chose have high VDS and small on resistance. I was worried about the routing of RFI because i connected the RFI between series cap and the EMC filter, so in this approach all the antennas "share" the same RFI path even though only one will be on at a time. I'm going to try and match the lines the best way i can and test it. Thank you for both your anwsers!