nRF91 sleep current high

[Petri1]

Hello,

I'm writing an application for nRF91 series and I have the functionality mostly working but the "sleep" current is high, about 4-7mA. I assume that this means that there is no power saving on.

I could work with relatively high sleep current, around 600µA but the lower the better.

I'm using an nRF9160-DK with a PPK2 at the moment.

My application is currently using UART, GPIO, I2C, and the modem but I'm supposed to add ADC as well. The chip also has GPS but I don't use it.


My application calls k_sleep() from the main function. According to this discussion it is ok.   K_sleep from main function

I think all the examples I've seen return from the main function and call k_sleep from spawned threads.

The k_sleep() call seems to result in correct amount of delay.

Could someone provide me with a minimalist configuration and code example to get the chip including the modem to sleep periodically? I could use that to validate my measurement setup and I suppose I could use that to figure out what's wrong with my application.

I think I would prefer CFUN=4 or CFUN=0 over PSM because as I understand it PSM is in some rare occasions not supported and slight extra power usage is not a problem. Or maybe I'm wrong about PSM.


Petri

[Asbjørn]

Hi,

I think you'd see some value in running the guide here and getting a benchmark:  Getting started with current measurements on the nRF9160

You should see something in the order of a few uA from the examples there and the methods could be incorporated into your code.

Best regards

Asbjørn

[Petri1]

Thanks for the link.

I used the UDP example from the link and I got ~3µA.
I'm not looking to use UDPs in this case but SMSs instead.

When I modified the UDP example to send SMS, the sleep current was again several milliamps. I probably missed some change in the prj.conf.

Does the PSM work with SMS? Could you provide required changes to the SMS example to make it sleep with k_sleep or similar? Alternatively could you help me with required changes to the UDP example to make it send SMSs instead of UDPs and still sleep properly?

Thank you

Petri

[Petri1]

Hello,
I did get the SMS example to sleep with ~700µA current with just k_sleep.
I can probably make it consume less with disabling serial port etc.

-Petri