Floor current too high

[Artnia]

Hi,

SDK 3.0.2, nrf9160 (actinius icarus som), custom board, based on modules example. LTE-M with PSM.

Floor current is too high, at around 500 uA. There is a known issue on the custom board but approx. 450 uA are left. Attached is the behavior after reset and then zoom into where I would expect way less current consumption. Something is pulsing at a rate of ~7ms. 

This is the build config:

and the .config from output:

...
CONFIG_PM_DEVICE=y
CONFIG_PM_DEVICE_POWER_DOMAIN=y
CONFIG_PM_DEVICE_RUNTIME=y
...
# CONFIG_LOG is not set
# CONFIG_SERIAL is not set
...
CONFIG_TFM_LOG_LEVEL_SILENCE=y
...

Devicetree:

I have read https://docs.nordicsemi.com/bundle/ncs-latest/page/nrf/test_and_optimize/optimizing/power_general.html and several posts in the devzone but no clue where this could come from. When removing the nrf9160 from the custom board the current is as expected, so it is most likely something in the nrf9160.

[Hung Bui]

Hi Artnia, 

On your hardware, if you run a simple sample in the SDK do you see the same pulses ? Something without MCUBoot and TFM ? 

Have you tried to run the firmware on a DK ? 

[Artnia]

Hi Hung Bui,

thanks for you reply. I am not 100% sure but I guess the issue was that my software was build upon a DK example that used CONFIG_DK_LIBRARY and this internally set DK_LIBRARY_BUTTON_SCAN_INTERVAL to 10ms.

Is that valid? What does this 10ms interval actually do?

For me it works to completely disable the CONFIG_DK_LIBRARY and to work with manually created interrupts.

BR

[Hung Bui]

Hi Artnia, 
The DK_LIBRARY use a timer that wake the chip up every  DK_LIBRARY_BUTTON_SCAN_INTERVAL to check if the button is pressed or not. This will draw current as the CPU has to wake up. 
You can choose to turn off Button. Or not use this library and handle the button press using Port or Pin interrupt.