Peak current when the nRF52 is in the SYSTEM ON mode

[Sylvain]

Hello,

I have done some consumption measurement but I have a strange behaviour while the nRF52840 is in the SYSTEM ON mode.

For the measurement, I send advertising packet each 1 second and then I put the nRF52840 in the SYSTEM ON mode. When the system is in SYSTEM ON mode I measure peak, but I don't know why. Are there any internal process in SYSTEM ON mode that generate these peaks ?

Plot of the measurement:

Test setup :

  - Segger Embedded Studio - SDK V14.2 SoftDevices 5
  - Windows
  - PCA10056 - nRF52840-PDK

[Sigurd]

Hi,

The SoftDevice will continue to advertise in System ON mode. You can read more about System ON mode here. What you see in the plot is first the pre-processing and HFXO ramp, then you have 3 radio TX peaks when you advertise on the 3 advertising channels. You can also see this using our Online Power Profiler.

[Sylvain]

Sigurd,

I know how the System ON mode and how the advertising work. You don't answer my question. Maybe it was not clear. So, my question is about the peak that occur when my mcu is in System On mode. So, I talk about the peaks after the data transmission. In resume, why I measure peaks between 31.5 and 31.88 s and also 32.2 and 32.5 (MCU is in System On mode between these intervals).

[Sigurd]

I tested your code on a nRF52840-PDK, and it didn't show any of these peaks here.

Image from Nordic PPK:

* If the PDK is connected to a USB-hub, could you try to connect it directly to the computer?

* Have you connected any external sensors to the PDK ?

* I see that you have plotted the measurement in Matlab or something similar, did you see the same peaks directly on the Tektronix DMM7510 display also?

[Sylvain]

I'll perfom new measurements in the following days to be sure of the test condition.

What's the sampling frequency of the Nordic PPK ?

Yes, I saw the same result on the DMM7510.

[Sigurd]

The sampling frequency for the PPK is 77 kHz.

[Sylvain]

Unfortunately, I think the dynamic of the PPK is not wide enough to measure the same behaviour as me. Furthermore, the slow sampling frequency probably smooths the signal.

I have the same behaviour as previously. My sampling rate is 1 Mbps. I expect a sort of refreshment from the nRF52 to maintain this System On mode. I think is like that with the CC2642 ...

I'm just asking myself if this behaviour is normal because it changes the average consumption ...

[Sigurd]

I also tested the main.c you posted with a KeySight N6705B also, and it didn’t show these peaks.

There is something called refresh spikes in the internal regulator, but that is not what is shown in your plot. The behavior you see is not normal.

It could maybe be USB noise. Could you do a test where you power the board with “External supply”(P21) or with a battery instead.

[Sylvain]

I performed measures with external power supply and now I have good results. I guess the noise comes from the USB power supply as you previously mention.

As you can see on the graph, I have some peaks at a frequency of approximately 38 Hz. Could be the internal regulator?

Thanks a lot for your support.

Sylvain.

[Sigurd]

Hi,

Sylvain said:

I guess the noise comes from the USB power supply as you previously mention.

Yes, this looks alot better.

Sylvain said:

As you can see on the graph, I have some peaks at a frequency of approximately 38 Hz. Could be the internal regulator?

Yes, 38 Hz is consistent with refresh spikes. So yes, this is the refresh spikes of the internal regulator.

[Sylvain]

Thank you Sigurd !