Where is stray current coming from in uA system. BLE Nano

Question

I am testing how low I can get the power consumption of the RedBear BLE Nano, which is a development board for the Nordic NRF51822 System on Chip that contains an ARM micrcocontroller, and BLE 4.0 mdoule.

I have produced some software that should have turned off all of the subsystems in the processor, and put the device into a sleep mode that is interrupted by a timer periodically.

This seems to work, I am powering the module form a bench supply and I have a multi-meter in series. When I have my finger touching the metal case that shields the Raytac module which is on the BLE Nano, the multimeter reads a constand 20uA, which is great. Then happily comes out of sleep mode every 10 seconds, flashes an LED and goes back to sleep, again at 20uA.

If I do not touch the metal shield, then I see this 20uA value climb to ~330uA very slowly (takes about 1-2 minutes), then descend down to ~200uA, where it seems to stabilise.

My bench supply only tells me the current draw in milliamps, so it's registering zero the whole time.

I have no idea why there is a difference when I am not touching the shielding, and whether the module is really only consuming ~20uA or closer to 200uA.

If anyone can explain which current draw is likely to be the actual one, and why there is a difference, that would be great.

Thanks

Answer 1

Stray EMI gets rectified from parasitic diodes on DMM and stored in stray capacitance. Common mode high Z EM fields can be in many volts range and on an unbalanced measurement ( low z ground, high Z signal ) result in the CM noise becoming a differential signal due to poor CMRR ( CM rejection ratio)

Solutions include raising CM impedance with a large Ferrite CM choke or BALUN around probe wire pair (clamshell or torroid) or grounding CM source of noise (shield)

Essentially this is measurement error but illustrates how Sensor amplifiers can also make measurement errors from loop antenna effects of long leads.