nRF52DK usb related power consumption issues

[essele]

Hi,

I have two questions about the nRF52DK power consumption when powered via USB .. one is for interest, but the other one appears to be an issue.

I have run some comparative current measurements (ble_app_template, with power optimisation and 5s advertising) and I see an average (500 readings at 12NPLC auto-zero off) current consumption of 4.1uA when using the coin cell and 6.7 uA when powered via USB.

I can remember reading another post somewhere (can't find it now though) that suggested USB power would increase the consumption by a couple of uA so it seems this is to be expected, however I am keen to understand why? My assumption is that it's something to do with the various pins that are switched differently if USB is connected? Is that correct?

Then the problem ... if I'm USB powered and I unplug and replug the USB connected (so de-powering and re-powering the nRF52DK) I then see an increase of about 12uA in current consumption pretty consistently, it's not affected by reseting (button) or re-plugging, it stays that way until you re-"build and run" at which point it drops back down to the normal numbers again.

I'm assuming this is to do with some kind of debugger default state ... is this expected?

Many thanks,

Lee.

[JONATHAN LL]

Hi Lee,

re-"build and run"

 Do you have to re-build? or is does just re flashing\running again work?

Can you also share witch version of the 52dk you have?

Regards,
Jonathan

[essele]

Hi Jonathon,

I'm not sure how I tell what version of the 52dk I have ... it's a couple of days old. The main IC label shows "PCA10040 2.0.0 2020.39".

In the J-Link properties is shows hardware version 1.0.

I don't need to actually do a rebuild ... I can also "connect j-link" followed by "disconnect j-link" and it also sorts itself out.

It's triggered by unplugging the USB or switching off, but not the reset button.

Cheers,

Lee.

[JONATHAN LL]

Hi Lee, 

The number 2.0.0 is the revision\version of the board. PCA10040 is the product and 2020.39 is the production date, year and week number. 

But to the current measurement, it is not reliable to measure current on the DK with the USB connected.  You can read more about how to correctly set it up on our infocenter.

Summarized in short it is:

• Do not use the USB connector to power the DK.
  
• Current measurements is unreliable when a serial terminal is connected to the Virtual COM port.
• After programming the device, the USB must be disconnected and the development kit power cycled to reset the debugger chip before current measurement.
  

The issues you are seeing is related to the debugger chip not being reset properly(Must be power cycled) and  the use of USB. 

Regards,
Jonathan

[essele]

Thanks for the reply Jonathan.

I was trying to understand a bit more about "why" ... as it's an absolute nightmare to keep switching between the coin cell and USB power during the development cycle ... and I'm definitely at a stage where I need to try many things to see what the impact on power consumption is.

Ignoring the debugger reset issue, which is easily avoided, it certainly doesn't look "unreliable", it just looks a couple of uA high .. if I understand why, then I can built that into my development cycle without too many concerns.

It does seem like a bit of a crazy limitation for a dev board for this kind of device though, that also (on the surface) makes no sense at all.

Lee.

[JONATHAN LL]

Hi Lee,

Do not have any papers or documents on the why part, but to remove the switching battery\usb source during the development cycle you can use a external debugger on the debug in, that way you can power the device from a coin cell and debug at the same time. 

I will update you if i find anything on the why part. But generally the reason the USB is inaccurate/unreliable is due to switching noise from the USB. 
Regards,
Jonathan