NRF52840 TX radiation sideband exceeds standard

[Frank Yang]

Hi Nordic,

When I was using NRF52840 for FCC certification, the DTM 8dBm, 2Mbps sideband emission power exceeded the standard.

The antenna matching circuit is to use the parameters recommended by Nordic engineer.

In addition to reducing the transmit power, is there another way to modify the parameters and make the test pass?

In the actual test, I set DTM 8dBm, some PCBA output can reach 9.2dBm, some only 7.5, their consistency is not very good.Is this a chip difference or something else?

RF schematic

Sideband  exceeded the standard

RF BOM

C10	CAP CER 1PF±0.05pF 50V C0G/NP0 0402	Murata Electronics	GJM1555C1H1R0WB01D
C17	Cap Ceramic 1.2pF ±0.05pF 50V C0G 0.05pF Pad SMD 0402 125°C T/R	Murata Electronics	GJM1555C1H1R2WB01D
L3	FIXED IND 4.7NH +/-0.1NH 700MA 160MOHM 0402	Murata Electronics	LQG15HS4N7B02D
L9	FIXED IND 2.2NH 900MA 90MOHM SMD	Murata Electronics	LQG15HS2N2B02D

Top layer

Layer2

Layer3 GND

[ketiljo]

Hi Frank

If you're trying to pass the FCC band edge test with 2 Mbps at channel 2402 and 2480 MHz, it will never pass this. The modulation bandwidth is too high. These channels are only used for advertisement and only with 1 Mbps. So you  musk use 1 Mbps here. 

7.5to 9.2 dBm output power is well within spec. You can expect +/-1.5 dB.

Best regards,

Ketil Aas-Johansen 

[Frank Yang]

Hi Ketil,

Thanks for the quick reply！

I checked the test report of the previous project, and I found that it passed the radiation at 2483.5MHz. It was also tested at 2Mbps speed. At that time, the RF circuit used the parameters of the previous three components.

previous three components circuit

（C3=C4=1pF，L1=3.9nH）

2483.5MHz PASS test

I would like to know:

1. Will the modification of RF parameters affect the test results? NRF52840 v1.7 datasheet description as blow, so I just modified the parameters.

2.2483.5MHz is the guard frequency band, and the last channel is actually 2480MHz. Is this laboratory test justified?

3. The certification laboratory said that unless the customer declares that he will no longer use the 2402 and 2480 frequency bands, he needs to test the 2Mbps mode. This is the result of their consultation with several TCBs. Is such a request reasonable?

Because this project is ready for mass production, I don't want to modify the RF parameters, a lot of extra verification work is required, and additional production materials need to be prepared. Could you help me answer the above questions?

Thanks!

Frank Yang

[Frank Yang]

My personal Case ID when review this project sch&pcb：

• Case ID: 282951

[ketiljo]

Hi Frank

Frank Yang said:

1. Will the modification of RF parameters affect the test results? NRF52840 v1.7 datasheet description as blow, so I just modified the parameters.

There's no component changes that will affect the modulation bandwidth. 

What FW did you use when you tested the previous design? DTM Isn't able to setup a carrier with 100% duty cycle and the carrier turn off and on between the packets will cause the bandwidth to increase somewhat. The radio test example in the SDK is more suitable for such testing: 

NSC: https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/samples/peripheral/radio_test/README.html

nRF5: https://infocenter.nordicsemi.com/topic/sdk_nrf5_v17.1.0/nrf_radio_test_example.html

Frank Yang said:

2.2483.5MHz is the guard frequency band, and the last channel is actually 2480MHz. Is this laboratory test justified?

Yes, this is correct.

Frank Yang said:

3. The certification laboratory said that unless the customer declares that he will no longer use the 2402 and 2480 frequency bands, he needs to test the 2Mbps mode. This is the result of their consultation with several TCBs. Is such a request reasonable?

No, this is not necessary. You shall test with the actual data rate on the channels that are in use. If 2402 and 2480 uses 1 Mbps only, you test with that. Since the other channels can be used with 2 Mbps, you can test with this at channel 2404 and 2478. 

This is just a test issue, most probably not anything wrong with the design.

Cheers,

Ketil

[Frank Yang]

Hi Ketil，

Our DTM test firmware is based on the nRF_SDK_16.0.0. Thank you for the clarification, I will explain to the lab that there are some issues with their testing methodology.

Thanks！

Frank Yang