nRF52810 board will not start on its own. It needs the RTT Viewer to be connected, then it will start and run.

[epolicar]

nRF52810 board will not start on its own. It needs the RTT Viewer to be connected, then it will start and run. It has run on its own before with this configuration:

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y

CONFIG_CLOCK_CONTROL_NRF_K32SRC_XTAL=y

CONFIG_SEGGER_RTT_MODE_NO_BLOCK_SKIP=n

Does anyone have any idea? I have struggled with this for a week now.

PS:

My schematic follows the reference schematics in the datasheet, which means the RESET line is floating. I am convinced this is not an issue because I have another board from Holyiot (beacon) with also floating reset. This board starts and run on its own just fine using the same firmware (pin assignments re-mapped, of course).

Regards,

[JONATHAN LL]

This issue is common to see when the 32kHz crystal is missing, the RTT viewer start ste device in a debug mode so then it simulates it internally. 

Doas all SW not work or just yours? 

Have you tested something else?

Regards,
Jonathan

[epolicar]

All my SW does not work.

[epolicar]

JONATHAN LL said:

Have you tested som basic samples ?

Yes, I did. 

I made a bare metal code that blinks LED. This code starts up on itw own upon power up. Could the problem be in Zephyr start up code and configuration?

[JONATHAN LL]

Can you share a PDF file of the schematic, the resolution here was not great and a bit difficult to read. 


If you are able to flash and debug as well as rund at least some code on the device then I suspect the main issue is some configuration in SW. 

Might be related to the use of the LFXO and if it is used or not but can also be a HW issue. Is it mounted correctly ? 

Regards,
Jonathan

[epolicar]

Her is the schematic pdf:

beacon.pdf

[epolicar]

I cannot get a waveform from both LFXO and HFXO since the start. Could this be why, eventhough, I enabled internal RC? 

[JONATHAN LL]

Looks like the VDD on pin 32 has the wrong component value, should be 4.7uF but you have 100nF.

When a debugger attaches to the device, it forces power and clock domains to remain on, which can mask power-related issues in your application. That is what I am suspecting happens in your case. 

Also you do NOT have DCDC on this design so you need to make sure that the device only runs software that has DCDC disabled. 


See if the external crystals are mounted the correct way. the LFXO (32kHz) should work, but the HFXO(32MHz)  can be mounted wrong if rotated 90 degrees. So double check that. 

Maybe try to use the synth clock:

Belive these should do. 

CONFIG_CLOCK_CONTROL_NRF_K32RC_SYNTH=y

CONFIG_CLOCK_CONTROL_NRF_K32RC_XTAL=n

Regards,
Jonathan

[epolicar]

JONATHAN LL said:

See if the external crystals are mounted the correct way. the LFXO (32kHz) should work, but the HFXO(32MHz)  can be mounted wrong if rotated 90 degrees. So double check that. 

HFXO is mounted correctly.

JONATHAN LL said:

CONFIG_CLOCK_CONTROL_NRF_K32RC_SYNTH=y

CONFIG_CLOCK_CONTROL_NRF_K32RC_XTAL=n

These settings did not work.

[epolicar]

Even the Blinky example will not start on its own unless RTT viewe is connected.

I would appreciate a reply.

Regards,