The following example is from a hoist gearbox input shaft cylindrical roller bearing. The first spectrum is a conventional vibration spectrum measured in the radial direction on the bearing housing. The frequency scaling is in orders of shaft speed of the bearing, and the markers on the spectral peaks indicate the harmonics of this speed :
This is a very complex, noisy spectrum, and the numerous strong harmonics indicate extreme looseness in the machine. The dotted cursors are set to 6.31 orders of run speed, and they correspond to an outer race fault frequency of the bearing. But, with this much looseness shown in the spectrum, it is difficult to say how badly the bearing is damaged, since looseness elsewhere in the machine could produce such a noisy spectrum.
The figure below shows the demodulated spectrum taken from the same measurement point:
This spectrum shows harmonics of the outer race bearing fault frequency rising about 25 dB above the noise floor. It is an excellent example of a severely damaged bearing spectrum. Note that the first harmonic is quite well defined in frequency, and the higher harmonics are progressively "smeared" in frequency. This indicates there is some jitter in the frequency of the fault tone, probably caused by excessive distributed damage to the outer race.
The next figure, below, contains photographs of the outer and inner races of the bearing: (The inner race picture is taken with more magnification than the outer race picture.)
The excessive damage to the outer race is impressive, and it is surprising that the inner race is damaged so little. The inner race is indented rather than spalled, whereas the outer race is heavily spalled.
Azima DLI provides products and services for Predictive Maintenance including vibration analysis instruments, monitoring and diagnostic software, and consulting for CBM programs.
© 2009 - Azima DLI