I'm sorry I misrepresented what I meant. I am not very good at English and I did not double check my question. I meant "power-supply noise" or "output ripple", not "industrial frequency interference".

Hi Tianxiang Zhou ,

there are two ways the 50 Hz of the power grid can be coupled to the input of oscilloscopes:

First, with common probes, it is coupled into the loop formed by the GND wire, the GND clamp, the section of the probe between the GND connection and the tip, and the path (or net) between signal point and GND point in the circuit you are measuring. You discovered already that with signal sources of low impedance there is no problem. If your application requires the measurement of weak sources with high impedance then you have to minimize the area covered by the loop mentioned above. The most simple approach to this end would be to provide a GND point in close neighborhood to the signal point, and to wrap the GND wire around the probe. If this isn't enough then you have to use probe sockets which can be soldered to the PCB. For example, see the last item on page 8 of the attached file.

In the second case, grounding the scope is exactly the cause of the coupling: Your scope is already grounded via its three-wire-plug. If the circuit you are measuring is grounded, too, then the power cable of the scope, the power cable of the power supply (or any other ground connection) of the circuit, and the coax cable of the probe form a large GND loop into which 50 Hz are coupled. The proper solution is to supply either the circuit or the scope via an isolation transformer (whose output socket has no GND connection).

Hope this helps.

Thanks, Mr. Fricke. I think your answer is useful, and I will try what your said next week. I bought some copper wire on an e-commerce platform and I am learning to construct a digital trap on MATLAB. I will let you know if I make any progress next week.

Again, sincere thanks to those who answered me. Today I learned how to use Matlab and used it to design a simple high-pass filter to filter out signals below 54Hz. The image "before" is the signal before filtering and "after" is the signal after processing. I compared the processed image with the signal obtained from normal measurement and found that the shape and peak are similar.

However, I think the advice you gave me is correct. Instead of dealing with the interference signal after the experiment, you should eliminate it at the source. I will ground the circuit as you suggested and I will let you know the success or failure afterwards.