The ProboStat itself is a passive component and does not originate noise, so it comes from somewhere else.

Noise – instabilities in the electrical signals and readings - may have several sources. Electromagnetic fields may induce voltage and currents in the sample and measuring circuit. The source may be communication signals (radio, TV etc.), fields from displays, lamps, and neighboring scientific equipment, and from the electric currents used to heat or cool the sample. Noise from external sources thus covers a wide frequency range, from radio-frequency to hum (50 or 60 Hz). Such noise can thus be reduced by appropriate filtering or screening.

Many measurements, instruments and circuitry have limited high frequency response, and are thus insensitive to radio frequency noise.

External noise has to pass an impedance to enter the sample or measuring circuit whereafter it experiences the impedance of the sample and circuitry; it is attenuated by the voltage divider consisting of the transfer impedance and sample+circuitry impedance. Therefore, the lower the impedance of the sample and the circuitry, the smaller the residual external noise that enter the measurement: A sample with low impedance shortcircuits noise.

Noise may also come from the sample itself. This is thermally generated noise – fluctuations in the concentrations and energy of charge carriers. Such noise is generated with a statistical distribution in frequency and level and thus in principle covers all frequencies.

Finally, noise is also part of the measuring instruments, for the same reasons as above. Thus, as the measured value approaches zero, some level of noise will inevitably remain, depending on the quality of the instrument.

For more information see the ProboStat manual.