nRF52840 DK- Lower RF output power as expected

[Adricast]

Hello, 

I have some troubles measuring the output power of the nRF52840 DK. I used both radio test and DTM firmware and I have the same problem. I try to measure the maximum output power of the nRF52840 DK, that is 8 dBm as said in the datasheet but I only obtain 4.8 dBm as maximum output power. 

My setup is composed of:

-Spectrum Analyzer Agilent CSA N1996A

- nRF52840 DK

-RF test probe MXHS83QE3000 (referenced by Nordic for RF measurements): 

https://www.mouser.es/ProductDetail/Murata-Electronics/MXHS83QE3000?qs=QnQa2dTVfy7u6%252B1BCmlR%252Bw%3D%3D

-Coaxial SMA to N-Type:

https://www.mouser.es/datasheet/2/418/9/ENG_DS_CSD_SMAM_610_NM_A-3238627.pdf

I expect to have around 1 dB of losses as said in both datasheet connectors, so I expect to measure around 7 dBm of maximum power but I only get 4.8 dBm when I configure 8 dBm as output power and I don't know why. I tried radio test with no modulated and dtm with different types of packets but nothing changes. I hope someone could help me.

Thank you in advance.

 

[Jørgen Holmefjord]

Hi,

Have you tried measuring multiple boards? The RF power accuracy for nRF52840 is +/-4dB, which means that your measurement of 4.8 dBm (I assume this is without insertion loss?) is within specifications.

The MXHS83QE3000 should have a typical insertion loss between 0.5 - 1.0 dB according to this page. The second cable seems to have insertion loss of 0.39*sqrt(f) dB, where f is frequency in GHz, which should correspond to around 0.6 dB for a 2.4 GHz signal. Worst case, the loss can then be around 1.6 dB in your cables.

Best regards,
Jørgen

[Adricast]

Hello,

thanks for your reply. Yes, I tried with two differents nRF52840 DK boards and the measurements are around 4 dBm. I'm not considering the insertion loss of the cables, so the output power without losses would be 6 dBm more or less.  If I understood well the maximum output power is 8 dBm but because of the RF power accuracy is +/- 4 dB, I could measure between 12 and 4 dBm = 8 dBm+/- 4 dB?

Thank you in advance. 

Kind regards.

[Adricast]

Hello, 

 I configure DTM or radiotest configuration. For DTM I use the nRF Nordic App. I select the bluetooth channel (single), transmit power (8 dBm), physical layer, packet type and the lenght of the packet.  Something like this:

 Then I use the spectrum analyzer with two modes, channel mode and spectrum mode, and I measure in the selected bluetooth channel , 2440 GHz in this case. And this is what I measure (channel mode):

In this case the output power is around 4 dBm (it's not the nRF52840 DK), with the nRF52840 DK is 5 dBm but no more. I don´t know if I have to configure the spectrum analyzer in a different way o like this its ok. 

Thank you in advance. 

Kind regards. 

[Jørgen Holmefjord]

The DTM mode will transmit modulated packets (PRBS9), which can give lower output power than the constant carrier signal.

I would recommend that you test with the radio test sample with the following commands:

uart:~$ data_rate ble_1Mbit
Data rate: NRF_RADIO_MODE_BLE_1MBIT

uart:~$ start_channel 40
Start channel set to: 40

uart:~$ output_power pos8dBm
TX power: RADIO_TXPOWER_TXPOWER_Pos8dBm

uart:~$ start_tx_carrier
Requested tx output power: 8 dBm
Tx output power set to: 8 dBm
Start the TX carrier

Unfortunately, I do not have any experience with the spectrum analyzer you are using. Can you look at the spectrum mode and do a peak search for the signal at 2440MHz?

[Adricast]

Hello, 

we want to make a BLE module and it needs to pass a test report through DTM test, using as transmission the PRBS9 type packet, as it is a modulated packet, how much would the output power be reduced? Using the nR52840 I have measured an output power reduction of 1 to 2 dB compared to measure with a constant carrier signal, is this normal?

Thank you in advance.

Kind regards,

[Adricast]

Hello,

this is what I measured with the Spectrum Analyzer mode:

Constant carrier:

PRBS9:

Both cases at 2438 MHz with tx_power of 8 dBm.

These results are under specifications if we consider +-4 dB of RF power accuracy. So you can confirm that this is a normal behaviour of nRF52840?

Thank you in advance. 

Kind regards,

Adrián

[Jørgen Holmefjord]

Adricast said:

Both cases at 2438 MHz with tx_power of 8 dBm.

Do you see the same if you reduce the RBW, for instance to 100kHz?

Adricast said:

These results are under specifications if we consider +-4 dB of RF power accuracy.

If your measurements are are not compensated for cable loss (potentially ~1.6 dB) and you measure ~4.4 dBm, which would sum up to ~6 dBm output power from the board, this is well within the specification of +/-4dB accuracy.

Adricast said:

So you can confirm that this is a normal behaviour of nRF52840?

I can check the measurements with similar signals in our lab after easter, but since there will be chip to chip variations in output power, and you have a different setup, it might not make sense to compare measurements directly,

[Adricast]

Yes, I changed the Resolution BW of the SA to 100 KHz and I still see the same measurements (around 5 dBm).  Yes, if we add the losses because of the setup (1.6 dBs) we reach up to aorund 6.2 dBm

Thanks in advance. 

Kind regards.

[Jørgen Holmefjord]

I'm measuring ~5.7dBm on my board without cable loss compensation, which corresponds to ~7dBm for this board.