Reducing Idle Current for NRF 52 DK

[cooperg]

Hi,

I am power profiling an application on the NRF52 DK. The application uses SPI to communicate with the BMA400 accelerometer:

1. The program initializes the accelerometer to trigger an interrupt after every 8 samples (at 25 Hz this is 320 ms).
2. In the ISR, a semaphore is incremented so that a "read sensor" thread will run to read the samples from the sensor.

The code runs without errors (I am using NRF connect sdk V3.0) and I now I am trying to reduce the idle power consumption.

I use the PPK2 to profile the NRF52 DK in Ampere Meter mode. I see the following plot with spikes every ~320 ms as expected:

I want to try to lower the idle current as much as possible. When I zoom into the idle region, I see this:

There seem to be periodic spikes every ~8 ms which decay. I know that the BMA400 consumes around ~4 uA when active.

Does anyone know what the source of these spikes are? 

[cooperg]

I was not quite able to pinpoint the cause of these periodic spikes, but I think I was able to remove them and lower the idle power consumption even further by using the sense-edge-mask discussed in previous posts:

&gpio0 {
	status = "okay";
	sense-edge-mask = < 0xffffffff >;
};