Average and Trigger windows show different currents

[Saxman58]

I find that the low current readings in the Average and Trigger windows can vary significantly.  I understand that the sample rates are much different, but does that also affect the current readings and offset?  For example, in the screenshot below, you can see that the sleep current after an advertisement reads about 140 uA on the Average window, but only 17 uA (and closer to expected) on the Trigger window.  These aren't the same time period, but it consistent for all the periods just after the advertising that I looked at.

I tried firmware with nothing but a system shutdown in main():

    NRF_POWER->SYSTEMOFF = 1;
    for(;;){}

This showed near zero current (+/- 0.6 uA) in the Average window, but a few uA in the Trigger window.

My setup is a PPK on top of the nRF52833 Dev Kit with power from the attached USB cable.  The app is the v3.5 Mac version.  I've run the PPK through the DK board interface with the PPK COM switch set to DK, and with the PPK attached to a separate DK board through the Debug In connector with the COM switch set to EXT.  (I wonder if the latter causes any problems with ground loops with two USB cables and the Debug ribbon cable between boards.)  The results are similar between the setups, with the two windows not agreeing on the current value for the quiet and flat sleep periods.

Any suggestions for how to get the numbers match better, or reasons why they don't?  And if this is an inherent problem, which numbers should I believe?

[Kenneth]

Can you try to push "ctrl+alt+shift+A", this should pop up some additional controls. Try to push "reset calibration".

In general the average should be more precise, since trigger is mainly meant to give a plot of the curve you are looking for. Though the difference her was very big I agree, it may have been damaged.

[Saxman58]

Thanks, I was wondering how to see the resistor calibration and switch levels, so now I know.  I reset the calibration as you suggested.

I did find that experimenting with the different options for the electrical setup gave me a combination with more consistency between the windows, and more believable results.  I'm now using an external supply for less noise, but still leaving the USB connected for programming the DK board, and another DK board to the Debug In on the PPK board to get the data.  That lets me switch between programming/debugging mode and power sampling mode very quickly.  I just have to remember to exit the debugger before I start the Power Profiler sampling, and vice versa.

[Kenneth]

Sounds good. I do believe that you should be able to run 'nrfjprog --debugreset' from command line to exit debug mode without disconnecting the debugger/programmer.

[Saxman58]

Actually, I find that just clicking the Stop button on Segger Studio is enough.  I don't physically have to disconnect anything, or flip any board switches, which is great.