How to improve ST25R3916 2nd and 3rd Harmonic filtering?

[Karl Jacobi]

With the ST25R3916, how can filtering of the 3rd harmonic (40.5 MHz) be improved? Following the procedure outlined in AN5276, a matching network was made. Performance of the device was verified, and emissions in the 13.56 MHz band are within FCC limits. However, the 3rd harmonic is above acceptable levels. Measurements with a near field probe on the tag show similar levels of the 3rd harmonic in the device I am using and the NUCLEO-NFC06A1 test board.

Changes that have attempted but did result in changes to the measured field on a near field probe:

Changing the cutoff frequency of the EMC filter (11 MHz, 13.8MHz, 15.5 MHz)

Changing the damping resistor

Changing the driver resistance

Changing regulator voltage

Are there other recommended approaches to reduce this harmonic?

[Travis Palmer]

Hello KJaco,

Have the measurements been done with the tag in the field?

What was the distance between tag antenna and reader antenna?

You can also try to select different type of EMC inductors. (e.g. shielded vs. wire wound or ferrite inductors).

please let me know.

BR Travis

[Karl Jacobi]

Hello Travis,

The measurements were made in an EMC chamber. Apologies for poor wording on my part, but the device is acting as in reader mode and no tag is present.

Thank you

[Karl Jacobi]

Hello Travis,

The measurements were made in an EMC chamber. Apologies for poor wording on my part, but the device is acting as in reader mode and no tag is present.

Thank you

[Travis Palmer · Accepted solution]

Hello Karl,

Sorry for my late response.

Looking at the reports available on fccid.io, the spurs of the 40.68MHz are around 30 to 32dBuV/m, which is 10 to 8dB below the limit.

The test report can be found here: https://fccid.io/YCPNFC06A1/Test-Report/Test-Report-4494604.pdf

If your device is operated by USB you may also try a different USB cable. Some cables have very bad EMC performance. (which is typically present at higher frequencies).

During development of the Discovery board, we have also seen that over filtering of the supply can cause such too high spurs. If you have additional ferrite BLMs in the supply you can try to bypass them.

Also changing the EMC filter inductor could help. Different type, different value, different brand. Theoretically a lower EMC cutoff should give better results on such harmonics. You can use QUCs and import the s-parameter file of the inductors to have a more precise simulation.

Also over driving the RFI pins could trigger the ESD diodes to get active. This can also cause lower harmonics to rise. You should check the voltage level of the RFI pins if they are according to the datasheet < 3V (best is 2.6V to have some margin).

As described in AN5322, the varicaps can also contribute to the low frequency harmonics. There are some hints in the application note, how to lower them. You may also try a board with fixed matching.

Please let me know, if this helped you with your problem.

BR Travis

[gte938x]

Hi, Travis,

I am the new engineer taking over Karl's NFC task.

On page 8 of the test report (EUT Radio Specifications), Transmitter Parameters section (b) lists many different modulation types. Then it says "A test mode allow to test all modulations at the same time".

I am wondering if this means the test mode was running with a different firmware. We are running  X-CUBE-NFC6 software tools from https://www.st.com/en/ecosystems/x-nucleo-nfc06a1.html#tools-software, and I do not think it allows testing all modulations at the same time.

Modulation differences should lead to different emission peak levels. For example, in NFC type A, Miller encoding (delay encoding) is used with an amplitude modulation at 100%, while in NFC type B, Manchester encoding is used with amplitude modulation only at 10%. The two different types of modulations have different peak levels.

In the 30MHz to 1GHz scan data (after page 40), there are multiples adjacent peaks in the band of upper 60MHz to 70MHz (a few are even in 80MHz) range, what are these peaks representing? They are not the harmonics of NFC carrier, and most importantly, we do not see this kind of "wide" band peaks during our test.

Also, on page 17, it describes the test method, why STM has to do peak measurement and quasi-peak measurement at different setups (3-meter in chamber vs 10-meter open area). To me, if it is an easy pass, it does not make sense to contribute extra labor and equipment to do testing at these level of complicated different setups. Quasi-peak measurement can run immediately after peak scan under the same setup at most test labs.

After all, let's ignore what our design is, we are now running X-NUCLEO-NFC06A1 stack on top of NUCLEO-L476RG board with STM X-CUBE-NFC6 demo software, and the emission scan is something like this (see below). How can we replicate the same result in https://fccid.io/YCPNFC06A1/Test-Report/Test-Report-4494604.pdf?

Thanks.

[gte938x]

Since the topic is different, I started a new topic at

https://community.st.com/t5/st25-nfc-rfid-tags-and-readers/st25r3916-radiated-emission-issue/m-p/622538

Thanks

Lu Dong