nRF9160dk current measurement with oscilloscope - too much noise

[crasspm]

Hello,

I'm trying to use an oscilloscope to profile my current consumption on the nRF9160DK.  I'm setting up according to the spec here:  https://infocenter.nordicsemi.com/topic/ug_nrf91_dk/UG/nrf91_DK/hw_measure_oscilloscope.html

However, I am seeing a lot of noise and can't discern my signal in the noise.  My scope is set to the following function:  abs(A-B)/0.5, which should be the current (absolute value of channel A minus channel B [voltage across resistor R60] divided by 0.5 [value of resistor soldered into R60]).

I'm wondering if the recommended 0.5 ohms for R60 is too low?  I see that 10 ohms is recommended when using an oscilloscope for profiling current with the nRF52840DK.

I'm using heavy averaging, 1x scope probe, high resolution etc.

Thanks!

[Stian Røed Hafskjold]

Hi, the problem with using the differential probe method is that the measurement resistor is fixed (i.e. fixed dynamic range), as opposed to other more advanced equipment that changes the measurement resistor based on the current load. This means that you have to find a resistor value that does not cause a too high voltage drop (when the current is high), but still high enough for the voltage drop to be measurable by the scope (when the current it low).

The current range for the nRF91 is from 1 µA (sleep) to 500 mA (worst case 23dBm at some channels)

This means that a 10 Ohm measurement resistor would give you a voltage drop of 5V (@ 500mA), which would of course make the nRF91 stop working. So that's why we recommend using a 0.5 Ohm resistor; it would give a maximum voltage drop of 250mV. But then again measuring 1 µA over a 0.5 Ohm resistor would require the scope to accurately measure 0.5 µV. This would probably just give you noise. Actually even measuring 1 mA (=0.5mV @0.5 Ohm) might give you lots of noise if you don't have a good scope.

If we assume that you won't get the highest peak currents (I usually see maximum 150 mA when measuring in live networks) and you set the voltage to 5V, you can maybe try with a 5 Ohm resistor. It will give you a 750mV drop at 150 mA, so it will stay above the minimum 3GPP approved voltage (3.3V). With a 5 Ohm resistor you should be able to measure accurately down in the 10 uA range if you have a good scope.

Other tips:

• Grounding of the probes is important. Connect them to the antenna connector or similar directly on the PCB. Not just through a wire connected to a header.
• Try and see if your scope has a bandwidth limiting filter, either analog on the input, or digital, or both. Measuring a current profile does not require a high bandwidth; higher bandwidth = more noise.
• 1x probes gives less noise than 10x probes
• Set the dynamic range on the scope (volt /div) on each probe as low as possible without clipping the highest currents
• Divide the measurements into separate tasks. E.g. measure radio currents first with 0.5 Ohm, then change the resistor to 10 Ohm and concentrate on the idle currents.
• Or, use a miltimeter instead to measure the lowest idle currents (the cheap handheld ones are usually the best and are accurate down to 1 uA usually)
• USB can cause a lot of noise. Disconnect the USB cable and power the DK from a good stable power supply on the external supply header

From my experience it is hard to get accurate results using the differential probe method on a scope. And I would recommend buying a power analyzer instead, with automatic dynamic range switching.

[crasspm]

Got it.  Those power analyzers are not cheap!  I wonder if it makes sense to use something like a current sense amplifier like this:  https://www.maximintegrated.com/en/products/analog/amplifiers/MAX9934.html

Thoughts?

[Stian Røed Hafskjold]

Sorry, I don't have any experience with those. One of my colleagues has had success in using something called Otii Arc. It's a lot cheaper than a conventional lab power analyzer. You can check out his blog here, where he uses it to measure radio and PSM current: https://devzone.nordicsemi.com/nordic/nordic-blog/b/blog/posts/measuring-psm-sleep-current-on-the-nrf9160-dk