power profiling nRF-9160 feather using nRF-PPK2?

dahoward

Any suggestions as to the best way to power profile the feather?
I thought maybe using PPK2 source meter mode through the JST battery connector, but I can’t seem to get the board to start without USB connected.
David

jaredwolff

dahoward you can change JMP2 which turns on the board at all times. (with either battery or USB connected)

Or, you should be able to power on the device using a press of the mode button. You will have to keep PS_EN asserted in firmware though for this to work. (P0.31)

Hidastelija

Hi,

I’ve experimented with the nRF9160 Feather and the board works great. However I have a bit of trouble measuring the current consumption.

I have a simple test FW that turns off uart0, puts the modem to PSM and keeps application processor idle. I’ve tried to measure the current consumption by connecting the nRF PPK2 in source mode to nRF9160 Feather V4 battery connector. The measurement result is very good: _80uA. However, when I add two capacitors (100uF+10uF) to the battery connector, the measurement result are not that good anymore: _351uA.

Have you seen similar behavior and do you use similar setup for measurements?

Ilkka

jaredwolff

Hidastelija what’s the reasoning for putting those on the battery? I’ve seen some similar behavior with supercaps causing higher than typical current. Are you using a power supply? What voltage are you using?

Hidastelija

jaredwolff
I’m using Nordic PPK2 in source mode. The output voltage was 4.084V when I measured the 80uA/351uA currents. The capacitors are not parallel to a battery, they are only connected to the battery connector to smooth the current consumption while measuring with PPK2.

Active sleep PSM floor current: With the capacitors I see 27 spikes in a 100ms measurement window and the max current in this window is 1.50mA. Without the capacitors the max current is 19.97mA and the current is switching 48 times per 100ms window.

I’ve done the same measurement using Nordic Thingy91. Also with Thingy the current graph changes a lot in a 100ms window, but the avg current doesn’t change much. Avg is about 185uA. Max currents are 270uA/17.48mA with/without capacitors.

There may be some losses in the capacitors, but I don’t think they should be significant. My assumption is that the feather current is changing so much at a high frequency that the PPK2 doesn’t measure correctly.

jaredwolff

Additional capacitance shouldn’t be necessary for the LM3281 that is on the nRF9160 Feather. It is likely adversely affecting the switching of the device. From the data sheet:

Minimize inductance between LM3281 pins (VIN, GND) and the LM3281 input bypass capacitor CIN for best performance. The LM3281 device and CIN capacitor should be placed to permit shortest possible top-metal routing for these connections.
Poor board layout can disrupt the performance of a DC-DC converter and surrounding circuitry by contributing to EMI, ground bounce, and resistive voltage loss in the traces resulting in poor regulation or instability. Poor layout can also result in re-flow problems leading to poor solder joints between the DSBGA package and board pads which can result in erratic or degraded performance of the converter…

I know that some folks want to add larger super caps in parallel with the battery which has also shown to muck with the switching regulator operation.

The LM3281 does switch up to 6MHz and the PPK2 samples at max 100KHz but in low power mode the regulator is in PFM mode which is much slower than PWM mode. The waveforms for the rail should show up as a sawtooth where the bulk of the current occurs during the switch on event. (you see current spikes during this time).

I’ll see if I can get some captures to explain better. I’m doing my own testing here since this is an issue that you’re not the only person to see.