Unable to run UART BLE app example without debugger

[jerry2107]

Greetings. Finally ordered some NRF52832 board from HOLY IOT and spend few weeks trying to handle with them.

So, all about setup:

SDK: 15 (15.0.0_a53641a) @ Windows 7 x64

IDE: Segger (Embedded Studio)

SD: 132 (s132_nrf52_3.1.0)

HW: custom board from HOLY IOT (YJ-16048 NRF52832) Seems schematic is fully identical to DK 10040.

The trouble is - when i am programming example from SDK (path\nRF5_SDK_15.0.0_a53641a\examples\ble_peripheral\ble_app_uart\hex\ble_app_uart_pca10040_s132.hex) it is working without any problems with or without connected debugger. When I am programming thru J-link compilled version (without any changes in project, only open-build-run) - it is working only while debug session is running. After reseting power device going to hight current consumption (about 7 mA) and stuck - no advertising or UART communication. 

I spent a while here, and tried following:

- disable semihosting (in linker options i could not find anything like it);

- switch to internal RC (no result);

- disable RTT (one post said it could be blocking);

- disable\enable DCDC;

- reset RESETREAS register on start (as adviced in errata).

Any thought?
Thanks!

[Håkon Alseth]

Hi,

 

That module looks to have a "full reference design", meaning it has DCDC components, external 32k oscillator. 

Do other examples work just fine, for instance ble_app_hrs?

There is a currently known issue with newer versions of SES, and standard functions for input/output:

https://devzone.nordicsemi.com/f/nordic-q-a/34784/conflicting-types-for-_putchar/133776#133776

 

If you have removed retarget.c from your project, I suspect printf is redirected to a inbuilt debug print instead. This would only work if you enter debug mode. You could try to remove the calls to printf and see if it works better?

 

Best regards,

Håkon

[jerry2107]

Dear Håkon, thanks for fast answer!

Yes, i faced that problem with putchar, but instead of correct solution i just commented that part (yep, project was not touched, but retarget.c was modified and i forgot about that...).

Once i tried solution from link (proper type for parameter) - its became stable, example works exactly as expected.

Thanks a lot !

[jeremysf]

I am having the same issue with my own hardware design. USB CDC works great with the debugger is attached (i.e. JLinkExe is running). I can reset the device, and plug/un-plug the USB cable. When I quit JLinkExe, the USB device is still present and things work. If I unplug and replug the USB cable, (with JLinkExe not running) then there is no discovery.

I did not have retarget.c linked into my project, but I am using SEGGER RTT. I tried to disable logs via setting NRF_LOG_ENABLED to false (and I see no logs) but that had no effect on the issue.

[kalauffenburger]

jeremysf,

Were you able to resolve this problem?  I am seeing exactly the same problem you describe.  I also have a custom hardware and so working with the example application is not easily done without significant reconfiguration.

This problem did not occur using SDK nRF5_SDK_14.2.0_17b948a, but showed up when I ported the code to use SDK nRF5_SDK_15.1.0_a8c0c4d.

I'm not sure how to debug the problem other than using shotgun techniques.

--ken

[kalauffenburger]

I forgot to mention, I have removed calls to printf() and it still does not work.