My 32Mhz crystal isn't working on Nrf52840?

[Mitch996]

Hello, I connected, according to official layout design, a 32 Mhz crystal to my nrf52840. Using a 60 Mhz oscilliscope, I decided that the crystal isn't working (no sine wave form no matter what) and I think it has something to do with the fact that I haven't properly configured relevant registers to enable it.

So... what are the list of registers that I should mind and configure to get the xtal to work?

The nrf52840 OPS v0.5, as usual, gives very, very brief and general information therefore not so helpful.

[Jørn]

Hello Mitch996

The oscilloscope you are using does not have a 60MHz sampling rate, but a 60 MHz bandwidth. This means at 60MHz the oscilloscope will show a signal that is 3dB lower (half power) than what the signal actually is. The sampling rate is at 1GS/s which would theoretically allow you to measure up to 500MHz. I therefore believe this isn't the problem. Even if the sampling frequency was lower than twice the frequency you are trying to sample, you would see a signal, however aliased and quite possibly distorted.

The Softdevice should handle all the register settings needed for the HFCLK, as it takes control of the radio. So as long as you are using a softdevice you should not have to set any registers for the HFCLK.

Normal oscilloscope probes have around 15pF capacitance to ground which is applied in parallel to the circuit you measure on. The capacitances of the crystal affects its oscillating frequency and amplitude. I suspect the capacitance from the probe is affecting the crystal oscillator, preventing it from operating properly. The HFCLK is also normally only running when the radio is operating, or if specifically requested by the application, so it only runs for very short time intervals. The combination of these two things might make it a bit difficult to measure.

If you are able to run one of the examples, and successfully is able to advertise you know the radio is working, which means the high frequency crystal oscillator is working as the radio cannot work without it.

As George suggested you should try probes with a high-z to minimize its influence on the crystal oscillator.

Best regards

Jørn Frøysa

[Mitch996]

Thank you once again Jorn Froysa, you answer was very detailed and crystal clear, it enriched my knowledge base, thank you!

As you probably could recall, I can now run the program on my custom board successfully, although still experiencing some programming setbacks, so the crystal is working properly.