Background
Today I have checked the input noise floor of the oscilloscope channel in Analog Discovery 3. By default, both oscilloscope channel inputs of the Analog Discovery 3 are not shielded. It is mentioned in this thread:
https://forum.digilent.com/topic/30475-analog-discovery-3-vs-pro-channel-crosstalk/
It makes sense as dupont wires or non-shielded adapters can catch electromagnetic or radio frequency interferences. Apart from that, the way the AD3 is powered could also influence noise floor. I focused mainly on the way of powering AD3.
Test setup
Analog Discovery 3 with connected BNC adapter without connected any wire to the adapter. Measured voltage peak to peak, RMS and FFT.
Scenario 1
The notebook is powered only by a internal battery with connected AD3.
Scenario 2
The notebook is powered only by a wall charger with connected AD3.
Scenario 3
The notebook is powered only by a internal battery with connected AD3, which is powered by an external charger via AUX input.
Scenario 4
The notebook is powered only by a powerbank with connected AD3.
Results
| Scenario | Vp2p_mean [mV] | Vrms_mean [mV] |
| 1 | 1.44 | 0.20 |
| 2 | 10.70 | 2.26 |
| 3 | 16.62 | 2.42 |
| 4 | 3.34 | 0.59 |
The best results are when AD3 is powered only by only from a notebook or an external powerbank.