The transducer is the element that transforms the mechanical vibration into an analog electrical signal, to be processed, measured and analyzed. Taking into account its constructive principle, there are vibration displacement, velocity and acceleration transducers, each of them more or less suitable for every specific industrial application.

All transducers must be accurate when taking amplitude readings, offering repeatability (two input signals with the same amplitude generate the same transducer voltage output). Vibration transducers must also be very accurate regarding the frequency information contained in the mechanical signal. This is fundamental because in many mechanical defects, the relationship between the frequencies associated to the defect and the rotating frequency of the shaft used as a reference provides the analyst with precise information to determine the nature of the mechanical defect originating the vibration.

The different types of transducers respond to different vibration physical quantities, as shown in the following table:

Displacement measurement values are especially suitable for low frequency vibration or when the analyst needs to know precisely the movement of a rotor and not just the part of the vibration transmitted to the bearing support. These measurements are taken directly with displacement transducers and are an indicator of the stress or strain the machine is suffering.

Velocity readings are generally those of wider use, since velocity is an indicator of the level of vibration severity and is proportional to the level of fatigue that a mechanical system is sustaining. These can be taken with a seismic velocity sensor, although accelerometers are more often used for their better frequency response and lower cost. The accelerometer signal is processed to be converted to velocity units.

Acceleration is the parameter that gives the best indication of the internal forces associated with a particular source of vibration (F = m × a).

The procedure of converting a displacement signal to velocity, or a velocity one to acceleration is equivalent to mathematical differentiation. Conversely, the acceleration to velocity or velocity-to-displacement conversion is mathematical integration. It is possible to carry out these operations with instruments that measure vibration and in this way convert the data from any system of units to any other. From a practical point of view differentiation is a noisy procedure in itself, and used in very rare occasions. Integration, on the other hand is carried out very accurately, with a cheap electronic circuit or via software. That is one of the main reasons why the accelerometer is the standard transducer for vibration measurements, since its output signal can be easily integrated once or twice to obtain velocity or displacement. The integration is not suitable for signals with a very low frequency (typically below 1 Hz), since in this area the noise level progressively increases and the accuracy of the integration procedure greatly reduces.

Most commercially available integrators work well above 1 Hz, which is low enough for almost all industrial machinery diagnosis applications.