Microamp precision when measuring current with a scope

Question

Hi there. 
 I'd like to start measuring the current consumed by a custom board with an nRF51822 on it in various states using a scope. I've read the DK user guide advice on this, which says "the resistor should not be larger than 10 Ω". There's also this post: 
 devzone.nordicsemi.com/.../ 
 And this tutorial, which is helpful: 
 devzone.nordicsemi.com/.../ 
 My scope's going to show 10 mV per division at best. 
 If I use a 10 Ω shunt resistor as suggested, 1 mA will show up as 10 mV on the scope and I'll see that no problem. But I'd like to be able to measure this with microamp precision ideally, or certainly tens of microamps. I'm interested in my board's consumption with the Nordic in sys off with RAM retention and with an accelerometer programmed on the board to fire an interrupt on motion and I'm aiming for of the order of 100 µA. This is with an external DC DC converter for the supply to the nRF51822. 
 At 1 µA the scope will see 0.01 mV and I won't be able to read it.
At 10 µA the scope will see 0.1 mV and I probably still won't be able to read it. 
 Is it realistic to expect this precision? Why is the DK user guide advice to NOT use a bigger resistor? 
 The main reason I want to get down to this precision is that I have a GPIO pin set up as sense input, from the accelerometer, and I fear I may be falling foul of this issue on the revision 3 nRF51822: 
 devzone.nordicsemi.com/.../ 
 The other option is a Fluke 116 multimeter, which has microamp precision, but I'd like to get a view of this over time.

Stian Røed Hafskjold · Accepted Answer

Generally it is not easy to get high-precision measurements with an oscilloscope (this of course is dependent on how good the scope is). The tutorial is based around measuring BLE events and then the currents can reach somewhere around 10mA. Having a larger resistor will force you to use a lower sensitivity mode in order to avoid clipping, which will give you lower accuracy. Another reason to not have a larger resistor is to avoid having a too large voltage drop which will result in wrong measurements. Using a lower resistor value the common-mode currents, which can be caused by both external and internal factors, are more significant, hence also a lower accuracy. So, if you are certain that the current is low, for instance when only measuring system OFF, you can try with a higher resistor value, still using the highest sensitivity setting on you scope. It all depends on the dynamic range of your equipment. So it is a trade-off here. Feel free to try different resistor values to find the one that fits your measurements and equipment best. 
 As described in the tutorial, you should "calibrate" your measurements by connecting both probes to the same node (at the voltage level you are going to measure) to see how accurate the differential voltage is. If you get other than 0 diff voltage it will give you an indication on the accuracy limits. 
 If you have a "high bandwidth" setting on your scope you should enable this. Also using 1X probes is pretty critical to avoid too much noise. If the signal is periodic you can consider using averaging mode on the scope, which will reduce the noise factor. 
 I think using a ammeter is a better way to measure low currents, since they use integrating opamps over a specified time interval to integrate all the current that goes through. So as long as the "events" on the chip that draw current is periodic with an interval shorter than the integration interval on the ammeter, you will get a very accurate measurement. Testing here shows that "event" intervals shorter than 100ms will give good measurements on general purpose multimeters. You can consider changing your code to sample the accelerometer at shorter intervals than 100ms just for measurement purposes and then do some calculations.