Practical Solutions to Machinery and Maintenance Vibration Problems
Chapter 3, Detuning and Proving Resonance
Section 5, Resonance Bump Tests
Calculations for resonance frequencies or critical speeds are not usually as accurate as those obtained through simple bump tests. (Bump tests are sometimes called resonance "impact tests.") Bump tests, however, are almost always limited to determining the first critical frequency of the part bumped. Also, if the part tested is connected to another part that is relatively flexible, then it is sometimes difficult to be sure that the resulting vibrations are originating from the part being bumped rather than from the part to which it is connected. (However, some of these problems may be reduced if instruments, such as Fast Fourier Transformers (FFT) or Real Time Analyzers (RTA), are used.) Bump tests with such instruments use special procedures that are covered in their own instruction manuals.
Bump tests are also ideal for rotors, rolls, fan blades, beams, columns, floors, covers, pedestals and welded steel bases. They are also practical for relatively rigid parts, including cast iron or cast steel bearing supports and brackets.
This test is based on the principle that when a "spring system" or part is bumped or deflected in some way, it will vibrate for several cycles at its natural frequency. Its amplitude decreases with each cycle, but its frequency remains the same. The low frequency of the repeated bumps is not measured, only the frequency that is generated between bumps.
For expected resonant frequencies in the range of normal speed machines, such as under 5000 rpm, bumping with a rubber mallet is preferred over harder materials such as wood or plastic. For higher expected frequencies, the instrument manufacturer usually suggests when to use harder hammer heads such as plastic. For more sophisticated instruments, such as a Real Time Analyzer, a special calibrated "force hammer" with an internal transducer is also available. (It usually has two heads - rubber and plastic.) It is used to determine if a bump induced frequency is the resonance frequency of the part that is bumped, or if it is from another part that is also affected. Instructions differ for each Real Time Analyzer (RTA) manufacturer.
Bump Test Using Swept Filter Instruments
With traditional swept filter instruments, there are two methods of performing the test:
a) Filter Out/Off Method
With the machine shut down, the following are step-by-step procedures for each method.
Filter Out/Off Method
1. Turn filter switch to Out/Off position.
2. Turn analyzer unit selector switch to displacement or velocity setting.
3. Set amplitude scale switch to relatively sensitive position, resulting in amplitude reading of half a scale or higher. (Okay if meter goes off scale.)
4. Turn frequency switch to the scale in which the resonant frequency is expected (or in which troublesome vibration levels have been experienced).
5. Apply vibration pickup to the member to be bumped.
6. Repeatedly bump the part while adjusting amplitude setting for reading at approximately mid-scale or higher.
7. Observe the vibration frequency on the meter. As the blows are given, the needle should jump to the natural frequency of the part and hold at that point for a few seconds. Several bumps producing the same peak meter reading will usually confirm that the proper natural frequency has been obtained.
8. If while the bumps are given, the frequency meter stays at zero or at a maximum, then the frequency switch was not set to the correct range. Adjust the switch as required until an on-scale reading is obtained.
9. If repeated bumps do not produce the same frequency, then the Filter In/On Method described below will be required.
Filter In/On Method
1. Turn filter switch to Out/Off position. This is so that the amplitude reading will be on-scale when the filter is finally adjusted.
2. Follow Steps 2-6 of Filter Out/Off Method, keeping amplitude meter at approximately mid-scale or higher.
3. Turn filter switch to In/On position. Usually the meter reading will drop considerably as the instrument is probably not yet tuned to the proper frequency.
4. Repeatedly bump the part while adjusting the filter dial, until the amplitude meter readings increase to an approximate peak.
5. It is not always possible to get an exact peak and, therefore, an exact frequency reading. However, it should be within five to ten percent of the desired result. When the approximate peak occurs, note the frequency on the filter dial or the frequency meter. Notice that when approximately tuned to the proper frequency, the frequency meter will read much more steadily and hold for longer periods than before. This is due to the filter tuning out other background vibrations in the area.
Bump Tests Using FFT-Type Analyzers
FFT-type analyzers have various settings which determine the type of spectra obtained. The procedure for obtaining the resonance data is generally similar to the above, but the setup is slightly different, as detailed below.
1. The FMAX (maximum frequency) should be set to a value higher than the suspected frequency.
2. The lines of resolution should be set to an appropriate number such as 400. Values greater than 400 normally are not required unless extremely close accuracy is required. The more the lines of resolution, the longer it takes for the instrument to acquire the data. Most modern data collectors have the capability to generate a "live" time display. If this is used, then 100 lines is more appropriate. Despite the loss of accuracy due to the lower resolution, the results are usually acceptable for most situations.
If the analyzer has the capability to display continuously updated (running) spectra using "peak-hold-type" averaging, then its use is recommended for the best results. If "peak-hold" is not available, then the number of averages should be minimized, such as 4 to prevent excessive data collection time. If possible, select a "Hanning-type" window since this gives the greatest frequency accuracy.
This textbook contains only part of the information in our Practical Vibration Analysis seminar.