nrf21540 gain always 20dB

[Neunz]

Hello,

We are testing the radio_test sample on a nrf21540dk board with ncs-2.3. 
We want to change the fem tx_gain but we always get the same RSSI value on the nrf connect RSSI VIEWER application. 

It's like we always operate with a 20dB gain. 

We added the following configuration to the radio_test sample : 

CONFIG_FEM=y
CONFIG_MPSL=y 
CONFIG_MPSL_FEM=y
CONFIG_MPSL_FEM_NRF21540_GPIO_SPI=y
CONFIG_MPSL_FEM_ONLY=y
CONFIG_MPSL_FEM_NRF21540_RUNTIME_PA_GAIN_CONTROL=y
CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB=10

CONFIG_RADIO_TEST_POWER_CONTROL_AUTOMATIC=n

Then we used the following commands : 

start_channel 1
end_channel 80
time_on_channel 50
fem tx_gain 10
output_power pos0dBm
start_tx_sweep

Are we missing something ? 

[Bendik Heiskel]

Hi,

The TX gain control in the radio test samples is working correctly with the config options you have provided. To confirm this here is the output power measured using a spectrum analyzer:

If fact, the only config options needed are:

CONFIG_FEM=y
CONFIG_RADIO_TEST_POWER_CONTROL_AUTOMATIC=n

RSSI is not a good measure of the output power from a device, since it is dependent on external factors,and the receiving device.

What are the RSSI values you are seeing with 10dB and 20dB gain?

 

Best regards,

Bendik

[Neunz]

Hi Bendik Heiskel ,

I checked again with only the two config lines. I always get arround -33 dB in RSSI Viewer. 

The board I use to check the RSSI is a nrf52840DK. 

Both nrf21540DK and 52840DK are on the same table. 

We also have a custom hardware with a nrf21540 and we ran the same code in conducted emission testing. 

In both case (10dB and 20dB) we measure between 15 and 20dB (the cable and connections loss are not taken in account) so it's like we are always emitting with 20dB gain. 

If we force the MODE pin to high (with a hardware connection) we effectively see a drop of arround 10dB. 

[Bendik Heiskel]

The nRF52840 have a absolute maximum input power of 10dBm, meaning it can be damaged by input powers of 10dBm or higher. The receiver will be saturated before this, so you won't really be able to measure a difference in RSSI between 20 and 10dBm output power from the nRF21540DK.

Do you have a spectrum analyzer available to measure the output power?

Here is the radio test FW nRF21540 I used for the measurement I showed earlier: nRF21540dk_radio_test_NCS_v2.3.0.hex

[Neunz]

Thank you, yes we have a spectrum analyser but we don't have the test probe to connect to the SWF connector to the spectrum analyser. 

We will purchase one and test this firmware. 

I come back with the result next week.

Can you test the nrf/samples/peripheral/radio_test from the following west manifest ? 

manifest:
    self:
      path: applications
  
    remotes:
      - name: ncs
        url-base: https://github.com/nrfconnect
  
    projects:
      - name: nrf
        repo-path: sdk-nrf
        remote: ncs
        revision: v2.3-branch
        import:
          # Import only necessary modules
          name-allowlist:
            - cmsis
            - fatfs
            - hal_nordic
            - littlefs
            - segger
  
            - zephyr
            - nrfxlib

Regards, 

[Bendik Heiskel]

Sorry for the delay.

That's the version of the radio test sample I used for testing previously.

[Neunz]

Hi,

Thank you, with the spectrum analyzer we see that the power was wright on the nrf21540DK. The same code wasn't working with our custom hardware because of protection elements on the SPI bus.