Vibration test fixtures come in such a large variety of sizes and shapes that it is difficult to give general statements that can be useful for a particular design. However, in working with our customers who are newcomers to the field of vibration, we have found a common problem which we feel warrants pointing out. Customers with long experience in fixture designs and applications, will, of course, be aware of the problem to be discussed here and will have already overcome it.
Many newcomers approach fixture design from the viewpoint of static strength and stiffness. The fixture and load weight are estimated and multiplied by the ‘g’ level of the test. This yields the force transmitted. This force is usually quite modest in terms of static strengths so that the designer then proceeds to clamp the specimen to the fixture and the fixture to the shaker table with a few bolts and clamping sections which are entirely adequate to cope with these low static forces. This approach tails to account for the dynamic conditions that occur at the higher frequencies encountered in most vibration tests.
Usually, when a fixture is designed with only static loads in mind, a very severe major resonance occurs within the frequency range of the test. The clamping and bolting arrangement, while adequate statically, turns out to be very soft or compliant when analyzed dynamically. In other words, the mass of the test item and fixture is resonant on the soft springs consisting of the clamping arrangement to the shaker table. There are at least two major objections to this resonant condition.
The first is that the test item does not receive the correct “g” level as controlled by the accelerometer in the shaker table. Due to the resonance, the “g” level at the test item can be many times higher than the shaker “g” read on the meter. The results in severe over-testing, sometimes damaging the specimen.
The second is that due to the resonance, there usually occurs a frequency where extremely high amplifier power is required to maintain constant “g” at the shaker table. This can be observed by noting when the amplifier output current approaches the maximum value noted in the manual. This results in severe overdriving of the Unit Under-Test and fixture, high distortion, and unnecessary stresses on the vibration system components.
The cure is to design fixtures with the dynamic problems in mind. In practice, this means making the clamps and hold down bolts stiff relative to the masses involved, so that the resonance of the spring-mass system is as high as possible, preferably above the operating frequency.
A second cure is to control from an accelerometer located up on the fixture or test item. This is usually less convenient, for the accelerometer must be moved each time the fixtures are shifted on and off the table. It is sometimes the only solution, as for example, if the test item is so large that the resonance cannot be pushed above the operating frequency. With large test fixtures it is sometimes best to use a multiple accelerometer control scheme, (average or extreme).
The max length of the bolts used to tighten the fixture to the armature table or slip plate must be shorter than 20mm. If the bolts are longer, the bottom surface of the armature or slip plate will have protrusions because of the stress of the extra length of the bolts to make some problems.
The test items were small electrical components clamped to rectangular aluminum plates 5×5×12. These plates were then mounted on the fixture which was bolted permanently to the shaker table. The first design for the fixture was a welded box structure of 38 thick welded aluminum plates as in Figure 1. It was impossible to run the tests because of the severe resonance below 1000 Hz, and ‘g’ levels on the test items were over 10 times as great as measured by the shaker accelerometer.
The solution was the redesigned fixture shown in Figure 2. This fixture was a solid cube of magnesium with drilled holes to remove weight. Also more of the hold bolts into the shaker table were utilized. The resonance for this new fixture was moved up over 2kHz and no testing difficulties were found.
In this test specimen fitted into a cubical aluminum fixture. The fixture then was fastened at four points to a 14 inch thick aluminum adapter plate. The adapter plate was fastened at four other points to the spring and severe resonance occurred.
Two solutions were possible. The fixture could be redesigned to bolt directly to the shaker table, using more than four bolts if possible. The second was to redesign the adapter plate. It was made of magnesium about 1’’ thick. It was tied to the shaker table using all 13 shaker attachment points available on that shaker table. Additional bolts were added from the fixture to the plate.
Figure 3 and Figure 4 shows the previous and later design respectively in this case.