The hall effect sensor is quite noisy, and I am trying to find where the noise comes from. The first I tried today is to measure the power spectrum and coherence among three hall effect sensors (in the vertical direction). Here is the result:

(the unit for the power spectrum density is in digital volt per root hertz.)

I do not quite understand why there is almost no coherence (apart from the 60Hz power line), even though the power spectra are almost identical among these sensors.

Can someone shed some light where the issue is? Is the noise non-stationary or what?

--------Another measurement with fewer average---------------------

It seems that when the number of average is small, the coherence is large, an indication of non-stationary?