Current consumption drift when using internal RC clock to drive ble

[mark waldin]

I'm testing using the internal RC circuit to drive ble via the S132 softdevice.   Set up is with Nrf52DK and Power Profiler kit.  I got everything running and was measuring about 33uA whic was about 3uA above the average draw using the external crystal oscillator.  Over the course of 30 min to an hour, the current drifted up to 125uA.  It appears to me that the Softdevice is forcing frequent recalibrations at an increasing frequency based on the current profile.  I am having trouble understanding what is going on or what I can do to get the current draw stabilized.

[Sigurd]

Hi,

I have not been able to reproduce this, I tested for 1 hour and I did not see any drift. Could you try to reset the PPK calibration values? See this link.

[mark waldin]

It turns out that it was coincidental that I changed to RC and the drift started to happen.  As it turns out the drift occurs on both oscillator settings. I reduced my application to nothing more than turning on the bluetooth set to broadcast. Here is more detail on the situation:

1. turn on dk and start the Power Profiler and the avg consumption is 30uA.  Trigger shows nice 3 pulse train.

2. the avg consumption slowly drifts upward to about 122uA over a number of minutes.

3. stop and restart the averaging and the avg remains around 122uA.  The trigger window continues to show nice pulse form at the same amperage as at the start.

4. reset the resistor calibration.  no change.

5. turn on and off averaging. no change.

6. turn the 52DK off/on.  reconnect device in Power Profiler.  "Corrupt data detected" occurs in the log.  3 attempts and 3 failures.

7. reboot the ppk. connect to 52DK.  Hit start.  Profiler starts running.  Power consumption reported at around 118uA.

8. turn off DK and reboot  Power Profiler.  Connect to device.  Start.  Voltage regulator registers 4mV.  No recording going on.

9. disconnected from device.  reconnected to device.  voltage regulator shifts to 3v.  start device.  Recording resumes normally with 106uA being reported and drifting upward.

10. wait 20 minutes.  start DK. start  Power Profiler. Connect.  Average current draw registers 30uA.

[Sigurd]

It sounds like you might have a floating GPIO input pin.

Could you run this command, and post the result here?

nrfjprog --memrd 0x50000700 --n 0x80

[mark waldin]

c:\Python27>nrfjprog --memrd 0x50000700 --n 0x80 --family NRF52
0x50000700: 00000002 00000002 00000002 00000002 |................|
0x50000710: 00000002 00000002 00000002 00000002 |................|
0x50000720: 00000002 00000002 00000003 00000000 |................|
0x50000730: 00000000 0000000C 00000003 00000002 |................|
0x50000740: 00000002 00000002 00000002 00000002 |................|
0x50000750: 00000002 00000002 00000002 00000003 |................|
0x50000760: 00000003 00000002 00000002 00000002 |................|
0x50000770: 00000002 00000002 00000002 00000002 |................|

[mark waldin]

you were correct.  It was a floating gpio line.  That did the trick.  Thank you so much.