nRF52840 continuous carrier TX test shows another peak +2MHz away from f_c, why?

[frank2608]

I'm running the nRF52840's radio (Raytec MDBT50Q-1MV2 module) in continuous carrier mode just to check some stuff with my RF spectrum analyzer. The analyzer shows a two peaks, one at f_center, and a smaller one +2MHz from f_center.

Where is this second peak coming from? And why is it upper sideband and nothing lower sideband?

Screenshot shows two differently positioned nRF52840 units performing the same test, but with different f_center, both of them have the same second peak at +2Mhz

The radio itself works fine, no bit errors when I actually send packets, RSSI look great. I am mainly just curious about this second peak, for educational purposes.

Thanks

[ketiljo]

That looks strange, I haven't seen that before. Spurious are usually symmetrically modulated into the carrier. Can it be an artifact from the instrument? Do you have access to another spectrum analyzer?

[jackstones092023]

I think you should check the mode of operation of the radio when you start the continuous carrier mode. According to the radio_test.h file, there are four possible modes for the radio test:

• RADIO_MODE_CARRIER_TEST - Unmodulated carrier test signal

• RADIO_MODE_CARRIER_DUTY_CYCLE_TEST - Modulated carrier test signal with duty cycle

• RADIO_MODE_CARRIER_ED_TEST - Energy detection test

• RADIO_MODE_CARRIER_MODULATED_TEST - Modulated carrier test signal

Depending on the mode you choose, the radio will behave differently and may produce different signals on the spectrum analyzer. For example, if you choose the modulated carrier mode, you will see a modulated signal with a certain bandwidth and frequency deviation. If you choose the energy detection mode, you will see a signal that depends on the energy threshold and the carrier sense mode.

duotrigordle

[frank2608]

The mode of operation does not affect continuous carrier test. During modulated tests, it happens with nRF, BLE, and IEEE802.15.4 modes, in all 1mbps, 2mbps, and 250kbps. It happens in both the bottom of the nRF52840's freq range, in the middle, and at the higher end of the nRF52840's freq range. Yes I have done an entire matrix of tests, at the request of Nordic Semiconductor, and the full set of test data has been submitted. I've uploaded a screenshot of just one out of the many test sessions (2410 MHz, modulated, nRF mode, 1mbps).

 

The nRF52840 does not feature a test mode register like the other chips in the family. The instructions from the SDK says that for nRF52840, continuous carrier is done by actually just not sending TASK_END when the packet is is done sending, and leaving the transmitter running (and obviously no shortcuts enabled). Modulated carrier tests are done by continuously writing to TASK_TXEN when the TX state exits, while the packet buffer being pointed to contains random data.

I don't think it's the instrument, this is a radiated test, I am unaware of any physics phenomenon that would indicate that this is a problem with my analyzer. Even if my analyzer's own oscillator had an unfiltered harmonic, it is scanning through a spectrum, so that harmonic would've shown up on literally every single frequency. Am I wrong? I think I can test again with a ESP32 just to make sure. That's all I have access to. (unless anybody knows if I can put a Raspberry Pi's Wi-Fi module in RF continuous carrier test mode)

That's the test setup. This is all for fun. I use the RF Explorer for crap like "why does Android Auto stop working beside this one public school when I drive by it" :-D

[frank2608]

ohhhhh crap

ESP32 is showing same behavior, it might be the spectrum analyzer

I have submitted a support ticket with RF Explorer about this

I am also out of devices that I can place into continuous carrier test mode

[frank2608]

RF Explorer has confirmed to me that this is from its own imaging oscillator at 1.8 MHz, and it will always be 30 dB lower than the peak detected, so if you are measuring anything 30 dB higher than the noise floor, there will be an artifact 1.8 MHz away

(only in the 2.4GHz band, it switches to different hardware in different bands)

Sorry about posting this then, not anything to do with the nRF52840, case closed I think