uCurrent gold noisy waveform on oscilloscope

[hi_dyb]

Hi all,
I have set up my uCurrent as shown in the diagram. I understand that I can't use the ground clips on my probes as the circuit is floating and I'd be introducing an unwanted ground point in my circuit, so am trying to measure the difference between my two probes, with the ground leads floating. The only problem is that my oscilloscope (Analog Discovery 2) trace is so noisy and I just can't figure out the cause. The DUT is (so far) just a bare-bones microcontroller setup with a single LED connected to an output pin, so I'm expecting a current in the tens of milliamps range.

I've tried searching this up but haven't found anything that could help me as of yet. It seems that when I use my multimeter (UT61E), the reading seems pretty stable at about 12mA.
Any help would be much appreciated. Attached is my trace and a diagram of my setup (the unlabelled box is the uCurrent, sorry). Many thanks.

[SilverSolder]

The two Minus connectors on the uCurrent are connected internally, so it is really designed to "live" on the Minus side of the load you are measuring.

So, I would try putting the uCurrent in series with the negative lead coming from the DUT,  instead of the positive lead, and see how that works out for you.  Now, you can use the scope negative clip on the probe and use just a single channel to do the whole job.

[hi_dyb]

Hi, thanks for your reply.
I've just tried reconfiguring the circuit as you described, putting the uCurrent on the ground return path and connecting the oscilloscope probe to the + terminal and the probe ground connection to the - terminal. Again I'm sure it works okay as I get a stable output on my multimeter however I still get so much noise on my oscilloscope trace that the reading is useless .

[SilverSolder]

The noise was reduced a lot compared to the previous screenshot, so things are going the right direction!

Have you tried scoping the 3.3V output on your power supply (in AC coupling mode), to see if that's where the noise is coming from?

[nez]

I'm no expert, but what is the noise you see if you just short the scope probe's ground to itself?  Using the ground clip, my scope in x10 mode is similarly noisy to your graph, while if I use the little mini spring ground instead of the big clip, the noise improves from ~16 mV peak-to-peak to ~8 mV (again, just shorted with no DUT).

Maybe this a case where your test setup could use extra EMF shielding?  I'll have to track down a cookie tin to make a shield and try out for myself sometime.

I also wonder about using x1 vs x10 mode to improve noise -- although you'll have to consider if the reduced bandwidth of x1 mode is enough to characterize your DUT's current fluctuations.  I'm not really sure, but maybe worth trying.


Edit:  Come to think of it, the bandwidth of the uCurrent is probably way lower than any x1 mode limit, so nevermind that!

[Kean]

Apart from your measurement technique, both your PSU and the microcontroller could be major noise sources.  To better see where the noise is coming from, I suggest you try using a battery as the power source and a resistive load as the DUT.  Chose the resistor value and power rating to simulate a similar load based on the battery voltage.  Then try your original PSU with the resistive load.  See how the readings compare.