Low power SAADC 394uA

[fhfs]

Hi,

I am having some trouble getting optimal power consumption while using the SAADC.

I've tried running the example given here 2766.nRF5_SDK17.0.2_ble_app_beacon_lp_saadc.zip. I've changed the LF_SRC in sdk_config.h to RC.

I am using a BMD-341 with SDK version 17.0.2

I've checked that the hf clock is not running. But I am still measuring 394uA with the powerprofiler which seems high.

Do you have any ideas?

[Jørgen Holmefjord]

Hi,

The BMD-341 contains a nRF52840 chip, while the project you have linked are for nRF52832 only (PCA10040 - nRF52 DK). Have you ported the project to pca10056/nRF52840?

Have you made any other modifications to the example?

The peak current also looks a bit odd to me. The expected TX current of the radio (when sending the beacons) in nRF52840 should be closer to 6-6.5mA, while in your screenshot the maximum is below 4mA. Are you seeing the beacons being sent from the board? Do you have another PPK that you can use to double-check the current measurements?

Best regards,
Jørgen

[Jørgen Holmefjord]

I created my own project and tested this on a nRF52840 DK with RC oscillator enabled for 32.768 kHz clock. I'm seeing about 90 uA average current with advertising enabled:

Can you try out this project on your board? ble_app_beacon_lp_saadc_nrf52840.zip

[fhfs]

I used the "main.c" and "sdk_config.h" from your example. Just to clarify, I build using the makefile and not using SES.

Here are the power profiler results without BSP

And without advertising and BSP

Why is the average current in the selected part of the graph around 150uA?

I was hoping since the SAADC was not used except for every 5 seconds, the power consumption would go well below 50uA.

[Jørgen Holmefjord]

I can't think of anything in the chip that will consume ~150 uA. How have you connected the PPK to your module/custom board? Is there something else on your board that could consume the additional current? Your plots does not look like they are caused by the functions in the provided example alone.

[fhfs]

This must be the answer, I tested it on a Nordic development board and I'm getting better results