Excessive current consumption when RTC running.

[Valerii_Ie]

Hi, I use s132 and SDK15 to work with nrf52832. 
In code I set up RTC and CC to trigger every N ticks. It works as expected and time periods counted correctly. However, when I measure current consumption, I notice current pulses about every N/4 ticks. No interrupt triggered except intended CC related. 
I've done an experiment and doubled N = N * 2; The period of current pulses doubled too. So they are related to RTC settings. Their current amplitude is about 6mA and duration 0.5-1.5ms. 
What can it be related to? How  to fix?

Thanks.

 

[Susheel Nuguru]

I am curious to know if there are any other parts of your code logic that depends on the define of N ticks? Can you show me your power consumption analyzer snapshot to see the spikes nature? 

[Valerii_Ie]

Hi Aryan, I found similar figure on devzone. https://devzone.nordicsemi.com/f/nordic-q-a/24324/power-consumption-spike 

Also I tried to disable all peripherals on board and to catch any retated interrupts. No result for now. If this spikes are normal operation of ldo oi wonder if any official doc can give detailed description of dependencies to make one understand how frequency of this spikes can be adjusted.

Thanks

[Einar Thorsrud]

Hi,

The spikes you are seeing do not look like the LDO refresh mode. Looking at the graph you posted, they seem to be about 20 ms apart and almost 1 ms long. This is way too wide and with a too long distance in time for refresh mode. It looks like you are using the CPU or similar. Can you make a plot that zooms in (in time) on a single spike? It would also be very useful to see your code.

[Valerii_Ie]

Hi Einar,
The distance between pulses is always near 17ms. I assumed that it RTC related because If I increase RTC sleep interval the distance between pulses increases proportionally. 
Please see attached zoomed plot. I'll start preparing code for revision.  

[Einar Thorsrud]

Hi,

The RTC should not cause this high peak current consumption (6 mA). Are you confident that the measurements are correct? Assuming they are, and as you see it varying depending on the RTC configuration, I wonder if you might be seeing some interrupts that you perhaps have not expected, and that you spend some time processing them. You could get a current consumption in that order if the CPU is active. Can you upload the code that you use to produce this behavior?