BLE difference between nrf51822 CDAB and QFAC

[unzeitig]

When I tried S110 examples like BLE proximity, beacon and hrs on QFAC, it works without any problem. Unfortunatelly I can not see any activity on pin VDD_PA of CDAB. Is there any difference between both chips from BLE point of view? How to simply debug it?

[Jørn]

There shouldn't be much difference between the two. There would be some difference in the required matching network, but there shouldn't be much difference outside that. Are you able to debug it with keil or another IDE? If you are you could set a breakpoint in the errorhandler and see if it gets called. I assume you are using a custom board? Could there be a hardware issue with your board?

[unzeitig]

I am using a custom board now after previous experiments with BLE400, I can not eliminate hardware issue but all ports of nrf CDAB are functional except BLE. Is there any possibility in SW to find out ANT1 or ANT2 are overloaded by BAL-NRF02D3? One thing remain me probably to unsolder BAL-NRF02D3 and measure on its pins.

[Jørn]

Unfortunately, I am not aware of any way to read if the balun is sabotaging anything. The BAL-NRF02D3 should work fine. Do keep in mind VDD_PA will only go high during radio activity, and will remain low otherwise. So it will remain low for most of the time. You could try to check whether the high frequency crystal oscillator is functioning as intended, the radio will not function without it. If it doesn't, there might be a problem with the mounting of the crystal on your board.

[unzeitig]

After removing of balun is result the same - VDD_PA is not on during radio activity. Regarding crystal oscillator - i am still begining with nfr51822 architecture - do you mean CPU is running from internal clock and BLE with high frequency crystal oscillator? Should be oscillation present on crystal terminals whole time when power is up or during radio activity only? It could be good direction.

[Jørn]

Yes, the CPU is typically running from the internal RC oscillator. The crystal oscillator will remain off when the radio is not active. The crystal oscillator signal is used as a reference signal for the 2.4GHz generation. Oscillations will only be present on the crystal terminals when the radio is active, however due to its sensitivity to capacitances you typically cannot measure it with an ordinary oscilloscope probe, as the additional capacitance from the probe would interfere. You could make a test with the timer modules, by choosing the crystal oscillator as the clock source for the HFCLK and using the timers to toggle a pin f.ex. It would be best to do this bare metal to remove any other error source. You can find some bare-metal examples here. You can take the timer example and add some code to initialize the HF XTAL.