ST25R3911B, range of air gap between transmitter and receiver antenna

K2R · ‎2023-10-12

Hello, commuity,

We want to estabilsh a data link between two devices over NFC. This is, because we want to power a kind of sensor device over the air via a gap of abt 1 mm.

The sensor consists of a ST25DV04KC and a STM32L031 with a low drop regulator.

On the other side or transmitter is a ST25R3911B driven by a STM32L031 (the controller will change later). Antenna matching network was designed by using the ST's Antenna Matching Tool. Anntenas itself are PCB coils of 4 windings of about 55mm diameter

Testing the power transmission we found, that this works fine in the range of more than 5mm. But the closer the antennas are, the more the field of the transmitting antenna drops, until there is no supply voltage available on the sensor board.

Are there limitations in the use of power transfer in this small ranges or did we miss any other restriction of this?

Thanks in advance

Kai

Travis Palmer · ‎2024-02-13

Hello Peter,

I hope, I do not miss any of your questions - otherwise please repeat them.

Yes, when being un-powered the RFO's are in high-Z and do not disturb the smith chart measurement.

Yes, if the device is un-powered also the AAT pins are in high-Z and you see the matching as without AAT-caps populated. That's, why we recommend to power the chip and put the driver in high-Z mode.

Your approach of measuring the smith chart and comparing powered and un-powered state of the chip are completely correct. But your smith chart look somehow distorted. If you still have an un-modified eval board ( I assume you are talking about the ST25R3911B-DISCO) can you please measure it un-powered. There is an AAT application note (AN4914) and the un-powered smith chart should look like figure 2 (or very similar).

I measured the antenna parameters (AN4974 - Figure 7) in the past. It should look very similar to figure 7.

Regarding your antenna measurement (first figure in your post): It looks distorted to me. Maybe there is a problem with the calibration of your VNA? Did you also compensate the cable?(Autoport extension)? Can you measure O-S-L as a reference? You could also measure a de-soldered inductor (like EMC inductor) and maybe a 10Ohm resistor. Also your antenna might have a resonance below 1MHz. Typically the curve starts in the left part ( = 0R spot) of the real axe. Yours is starting from the right side ( = infinite Ohm spot). Maybe your antenna is not a closed loop. You can also check the resistance with a normal multi meter.

Regarding your ST25R3911B-DISCO antenna measurement (second figure in your post):

It looks better, but your beginning of the curve (1MHz marker) is very high in the inductive region. Again there is a strange distortion.

Please let me know if your problem is solved.

br Travis

View solution in original post

K2R · ‎2023-10-23

No one an idea what we can do?

Ulysses HERNIOSUS · ‎2023-10-23

Hi,

which antennas are used on both sides? Can you share a picture, please?

Best Regards, Ulysses

K2R · ‎2023-10-23

Hi Ulysses,

find bitmaps of footprint attached. Both anntenas share the same shape: 4 windings with a diameter of about 65mm.
As far as I know, the matching has been done using the ST25R Antenna matching tool.
As written above, the necessary energy transfer does work, but even on distances more than 5mm. All below that lead to a dropping field, therefore to dropping voltage.

Regards Kai

Ulysses HERNIOSUS · ‎2023-10-23

Hi Kai,

It sounds you didn't do the matching and above is an effect of non-optimal matching.

I think you need to go back to your colleague having done the antenna matching. I think he will need to measure the antenna matching (not only calculate it) and re-match according to the recommendations in our AN.

In case he requires help here on the community he will need to share measurements and smith charts of the antenna in its target environment.

Also just the antenna rendering may not be enough if there are PCBs/metal/... close by. That was why I was asking also for actual photos.

Best Regards, Ulysses

K2R · ‎2023-10-23

Hi Ulysses,

okay, the last point was a misconception: in my test setup there is no metal but only the PCB-antennas.

So I will go to the hw developer and lead him to here for getting the needed support.

Regards Kai

Travis Palmer · ‎2023-10-25

Hello Kai,

could you share the smith chart of the (un-powered) reader with and without tag present?

I think your antenna matching is perfectly fine if no tag is in the field, but once the sensor tag is being put on top the matching de-tuns and very little power is transferred.

Thanks, BR Travis

K2R · ‎2023-10-25

Hello, Travis,

as soon as our measurement equipment will be avaiable to me, we will do the measuring and upload the files here

Regards Kai

Peter_H · ‎2023-12-04

Hello,

I'm a colleague of Kai, we work together on this NFC project.

We did several comparison measurements. In particular, we measured the antenna of the ST25R3911B-DISCO with and without matching network. Not only with our VNA, but also with a higher class instrument in an EMC-facility.
For some reason our results differ from the Smith-diagrams found in AN4974.

I compared the results (with and without matching network) of an unpowered, unmodified ST25R3911B-DISCO board with the results of the same board, where only the matching network and the antenna were visible for the VNA. (All other connections towards the NFC chip were "open". Also the 2x main output wires at the bottom side were cut, which have no 0R resistors in the signal path.)

After these comparison measurements we saw that the unpowered chip pins are "high-Z" and have almost no impact on the results of the antenna measurements. It is clear, that in this case the effect of the AAT capacitors was also not there (and so these are not trimmed the resonant frequency to the desired value).

The AN4974 recommends to use a powered board for the VNA measurements (page 41, chp. 7.4). However the board should be reconfigured to deactivate the RF output stage and disable any voltage output (which could cause damage to the VNA input).

So we itentionally measured always only an unpowered board to avoid any possibility of any damage to the VNA (e. g. in case of a not performed reconfiguration).

At the measurements with matching network, the VNA was connected according to AN4974 fig. 34
At the measurements without matching network, the VNA was connected to J402 and R401 and R402 were removed.

The pictures of our measurements:

Antenna with matching network (we expected a plot like figure 35 on page 42):

Antenna (of the evaluation board) without matching network:

My question is, why we see an entirely different plot at the first picture compared to figure 35 of AN4974.

If You could please do a measurement at J402 without R401 and R402, You should get the same plot as at the second picture.

This would be my starting point to clarify, where the difference coming from.

I will post soon the plots of our antenna without and with the "tag PCB".

Many thanks in advance for Your advices!

Best regards,

Peter

Travis Palmer · ‎2024-01-21

Hello Kai,

your receiver circuit is loading and de-tuning the antenna. You could start by measuring the smith chart when the receiver is in different distances (0mm, 5mm, 10mm, 20mm) to see the effect. Please feel free to share the measurement with the community.

Most likely the antennas have a strong coupling in 0mm or 1mm.

I see two possibilities:

1.) tune the antenna with the receiver within the desired operation distance (e.g. 1mm).

2.) increase the transmitter antenna to lower the coupling in close distance.

please let us know, if this helped.

br Travis