Practical Solutions to Machinery and Maintenance Vibration Problems
Chapter 13, Additional Vibration Sources
Section 2, Vibration Due to Mechanical Looseness and Rubs
Over many years, the frequency of vibration due to looseness has been reported as 2 x rpm. A machine's base that is rocking on the floor will bounce against the floor much like a chair with uneven legs. One end "bumps" at one-half of the cycle and the other end at the other half cycle, producing a vibration at twice the rotational speed. However, with the increased use of spectrum analyzers that readily show all the vibrations of the frequency spectrum, looseness has been reported to create ½ x rpm vibration as well. To differentiate this from vibration due to oil whirl, oil whirl creates vibration at slightly less than ½ x rpm. Looseness can also create vibration at other multiples of rpm such as ½ x, 1 ½ x, 2 x, 2 ½ x, 3 x, 3 ½ x and so on.
For the usual spectrum that is created by either mechanical looseness or a rub, the above multiples are not always present on the spectrum. Instead, only one or two may be present. For example, a bearing that is loose in its housing may produce only the ½ x rpm peak. In another situation, a loose bearing may produce a ½ x rpm peak plus one at 1 ½ x rpm. In still another situation, the ½ x peak may or may not be present and the 1 ½ x peak is missing, but the 2 ½ x and the 3 ½ x peaks are present. In other words, the various peaks that are related to ½ x rpm can be present but are usually sporadic. In situations of extreme looseness or rub, many more of these peaks will be present. Unfortunately, these same multiples could also be caused by a severe rub. These frequencies are only a warning of possible looseness or of a rub. The analyst may differentiate between the two by considering the history of the job and determining which is most likely.
There may be many situations when a part is actually loose, and yet the multiples as described do not appear on the spectrum. For looseness to be revealed on a spectrum, vibration at another frequency must be large enough to cause the loose part to move. For example, visualize a small, well-balanced motor placed on a simple table top and run at its service speed. Although the motor is not bolted down, absence of vibration due to unbalance or any other sources is not great enough to cause the loose motor to rattle on the table. Therefore, a vibration spectrum will not show the symptoms of looseness even though the motor is actually loose.
If additional unbalance is added to the motor pulley, the unbalance amplitude of 1 x rpm will increase and yet not be strong enough to rattle the weight of the rotor. Peaks that reveal looseness will still not be present. Step-by-step increases of unbalance will finally reach a point where the rotating unbalance force will cause the loose motor to momentarily lift slightly from the table, causing a rattle. Sometimes only a slight rattle will produce several of the peaks due to looseness. At other times, a much stronger rattle will produce either none of the peaks or only one or two. Yet the stronger the rattle, the stronger the possibility for more peaks. Why some peaks in the sequence are missing at different times is not presently understood.
Similar reasoning applies to a rub. It's so easy in a classroom to describe the peaks produced by a rub. Occasionally, a serious vibration problem is solved by analyzing these peaks, yet purposely producing a rub will not always reveal these peaks. At present, all that can be done is to get a general idea of what is happening, and as instruments and experiences improve, this subject will become more clear.
This textbook contains only part of the information in our Practical Solutions seminar.