We recently tested a big piece of medical equipment in our vibration testing facility, and it failed. The failure itself was interesting. Only one component, a small flexible connector for an LCD, had problems: After the test, no display. It turned out that the connector was mounted in a rigid bracket. Removing the bracket eliminated the problem by allowing the connector itself to damp out vibrations.
Just another day in the testing lab. Some design engineer thought adding a bracket would reduce vulnerability to vibration when, in fact, it did the exact opposite. But situations like this are not unusual. Vibration testing often reveals just this kind of fuzzy thinking.
Problem is that a failure during a vibration test can only point out weak points in a design if the test has been run correctly. Developers can’t wonder whether a failure was the fault of the design or the fault of a test that really didn’t reflect real-world conditions. Yet many developers today find themselves in the position of having to learn and understand vibration testing and take a major role in testing their own designs. The reason is that a drop in military business, corporate downsizing, and attrition all have sent many of the test industry’s best engineers to greener pastures.
The good news is that today’s turnkey, computerized vibration-control systems let tests proceed relatively quickly and efficiently. Even an individual with limited technical experience can grasp the key points to program a vibration test.
The bad news is that programming and running a test is one thing, but conducting a test correctly is quite another. Referring to test specifications isn’t much help, though there have been numerous test specs drafted over the last two decades to keep pace with advances in vibration testing. These specifications often do little more than simply specify the vibration parameters applied to test samples. There are several issues over and above test specifications that go into conducting an effective test and getting accurate results.
Planning on planning
Developers who find themselves planning vibration tests should remember why they are doing the test in the first place. Though it may seem obvious, vibration tests determine product robustness and how well the product will stand up in the environment it will see. So developers should plan tests with some idea of the conditions the product will experience in the field.
There are small, portable data acquisition systems available that can help measure these conditions. The Institute of Environmental Sciences and Technology in Mount Prospect, Ill., has published a recommended practice, IEST-DTE-RP-012, “Dynamic Data Acquisition and Analysis Handbook,” an excellent explanation of how to properly gather and interpret dynamic data.
Short list for judging equipment needs
Frequency range and waveforms needed for tests
The idea of gathering data and analyzing it may sound routine, but there is really more to it than meets the eye. As a recent example, consider the experience of one manufacturer that measured the g levels seen in the environment where it planned to deploy a new product.
The measurements were taken correctly. They revealed a few huge transient spikes and much lower levels of broadband random g forces.
It could have been a mistake, in this instance, to use the momentary spikes as the maximum points for a design or testing envelope. Indeed, such a test criteria could represent an environment that the sample would never need to meet. The product could be overdesigned and overtested, making development costs soar.
The mistake was in failing to distinguish momentary, transient events from recurring, continuous levels. Disregarding the transient spikes would have kept a lid on development costs without sacrificing product reliability.
That’s not to say transient spikes are completely ignored. Shock testing will reveal a product’s resistance to these spikes. In general, developers should verify any vibration test criteria by correlating laboratory failures to historical field returns.
Step-stress testing is an often-used approach for “shaking out” any weak links. It vibrates a sample at a certain level for a prescribed period of time while operators monitor the device for proper performance. If everything is satisfactory at one level of vibration, the test ratchets up to higher and higher levels until something fails.
This type of step-stressing typically takes place during the design and development phase to ensure adequate robustness in the product. The approach also tends to be iterative, with a “test, analyze, and fix” philosophy employed until robustness hits desired levels.
A third common objective for vibration testing is to screen production assemblies. The screening can identify marginal designs and manufacturing defects. The IEST has published documents covering different approaches and philosophies on environmental stress screening (ESS). Its IEST-PR-RP-001, “Management and Technical Guidelines for the ESS Process,” addresses traditional ESS testing such as random vibration with single-axis electrodynamic shakers. It also discusses other approaches to ESS such as repetitive shock machines, multiaxis electrodynamic shakers, temperature cycling chambers, and others. All have merits and drawbacks and, of course, the goals and objectives of the screening program determine which is best.
Fixturing and setup
Settling on the best vibration spectrum is only the first aspect of a good testing program. There are numerous other issues that go into properly conducting a vibration test. Perhaps the most important is fixturing. An improper fixture can severely compromise the test. It sometimes even can induce unwanted resonances into the test sample.
Many vibration test specifications fail to address fixturing in any degree of detail. As a result, it is not uncommon to see different test results coming out of two different test laboratories — though both use the same vibration spectrum, vibration controller, and vibration shaker. The IEST has prepared a document which addresses vibration and shock test fixturing: IEST-DTE-RP-013, “Shock and Vibration Fixturing.” The document presents many of the issues and questions that developers must address when fixturing for vibration testing.
There are several key areas that demand consideration when developing test fixtures. The first is whether the sample will be hard mounted to the shaker table or attached in a way that resembles the actual field installation. Developers generally choose hard mounting when they know the actual field conditions pretty well. On the other hand, they’ll go with field-installation mounting when they understand the field conditions of the complete system in which the specific part will be used.
Some test fixtures are designed to hold more than one sample. This brings up the question of how many samples a multiple-position fixture can effectively hold. It is something to consider when screening or testing large quantities of samples for any significant period of time. Also important is whether the fixture permits quick and easy changeover of the test sample, and whether the fixture imparts the desired test spectrum.
One way to evaluate a fixture design is by conducting a fixture survey or modal analysis. A fixture survey employs a dummy load outfitted with accelerometers and mounted onto the fixture. Accelerometer outputs reveal resonances and antiresonances of the fixture both within and outside the test frequency bandwidth. They show whether or not the fixture will amplify or dampen shaker table vibrations.
Problem is, it’s difficult to develop an optimal test fixture with minimal resonances. So manufacturers often must devise work-arounds to minimize problems.
For example, a simple notching of the input vibration spectrum might counter the effects of unwanted resonances. Use of limiting accelerometers might be another approach. Placed on the test sample, limiting accelerometers can provide feedback that limits the input vibration levels. The idea is to keep the sample from seeing vibration levels exceeding its spec. Still, the best way to ensure accurate tests is to develop a test fixture which doesn’t experience these problems in the first place.
The fixture designer and fabricator both must understand the objectives of vibration testing. Also critical is that they know the vibration levels the sample sees. For instance, consider random vibration testing for electronic assemblies. The vibration-control accelerometer typically sits near the point where the test specimen attaches to the holding fixture. The holding fixture is, in turn, rigidly secured to a shaker table.
Circuit boards subjected to vibration are likely to flex in a couple of different modes. The center of the board will likely see the largest deviation, with an amplitude that depends on the physical board dimensions and their relationship to the wavelength of the input vibration frequencies.
For example, the board may resonate at a vibration frequency whose half-wavelength corresponds to the distance between two board supports. Besides resonating at this first-harmonic vibration mode, higher modes may be present as well, again depending on the input vibration frequencies and board dimensions.
The amount of deviation a given spot on the board experiences is, of course, equal to a superposition of the first, second, and other harmonic vibration modes that are present. So the amplitude of vibration at given points on the test sample depends on not only the input levels at the fixture attachment points, but also on the frequency of the input, the configuration of the test sample itself, and the fixturing arrangement.
It can be a challenge to design a test fixture for large, heavy machinery. At the right vibration frequency, the sheer weight of the test sample can cause even the strongest-looking fixture to flex and move. For example, thick 3⁄8-in. aluminum will flex at frequencies of 2 and 3 kHz, potentially leading to amplification of vibration inputs at these frequencies. The outcome will be inaccurate results, or a test which cannot even be completed.
Often, manufacturers will choose to design and fabricate the test fixture themselves in an attempt to economize. But it’s quite difficult to develop a test fixture that cleanly transfers shaker table vibrations to a test sample over a wide frequency spectrum. “Homemade” fixtures frequently turn out to have unacceptable resonances and must often be redesigned or rebuilt from scratch. This sort of debacle can end up costing more money than just bringing in a qualified fixture producer from the start.
There is a common misconception that a test fixture ought be inexpensive to make. But it’s not uncommon to spend half the testing budget and allotted time on just designing and fabricating the test fixture. In some cases, the fixture costs more than the test itself.
A few examples illustrate the importance of proper fixture design. A manufacturer of heat exchangers for diesel engines mandated a testing protocol using a rigid, “hard mount” fixture. Engineers there designed and built the fixture themselves. But the heat exchangers saw a high failure rate in the subsequent vibration testing.
An examination of the testing environment uncovered two key problems: First, the fixture was amplifying certain test frequencies. Thus the test sample saw vibration levels that exceeded those input to the test fixture. Second, the manufacturer had specified vibration levels far greater than the heat exchanger would ever experience in the field. A redesigned fixture and more realistic test limits produced useful results.
In another case, a manufacturer of base station radio-frequency antennas wanted to accurately simulate field conditions. The antennas — up to 12 ft long — had to be free to resonate. So a rigidly mounted fixture holding the antenna throughout its length wouldn’t simulate real life.
There was another factor that complicated testing. Developers needed to evaluate the antenna’s electrical characteristics while the vibration testing took place.
The solution was a specially designed fixture that let the antenna move, vibrate, and resonate as if buffeted by wind gusts. The fixture also held the antenna only at its base and was equipped with RF connectors that let engineers monitor antenna electrical qualities during the testing.
What to use in a new vibration-test install
It is no small undertaking to procure and install a vibration test system. The process of selecting equipment is complicated by the availability of different types of shakers with varying qualities and accuracies. They work with numerous software packages and vibration controllers, and with a wide range of transducers, signal conditioning equipment, and cabling. All are important parts of a test system.
In addition, it can be very expensive to put together a system that can handle a diversity of dynamic testing requirements. Even a small, dedicated unit can run into the tens of thousands of dollars.
The selection process demands the services of a trained individual who can assess the variety of vibration testing equipment on the market in light of the required performance. One document that may help in this endeavor is IEST-DTE-RP-009, “Vibration Exciter Selection,” published by the IEST. It addresses buying vibration equipment to get the most bang for the buck
Special considerations apply when choosing vibration equipment for large assemblies. Obviously this equipment needs sufficient force-pound capabilities. In addition, because accelerometers provide only localized control, tests of large assemblies require multiple control accelerometers to detect and control multiple vibrational modes.
The first step in a testing regime generally consists of programming the required test spectrum into a vibration controller. There are numerous types of vibration inputs that can be applied. They include sine, random, mixed mode, quasi-random, repetitive shock and others.
Operators who program a vibration control system must know the language and units used in testing. Sophisticated control systems give the operator many ways of controlling a simple vibration input. Typical parameters specified when programming a vibration test include the spectrum breakpoints, the number of accelerometers, and their sensitivity. Also detailed is the overall control strategy and the duration of the test. Sine tests require a specification of the sweep rate, while random test regimes should spell out drive clipping levels, peak and rms G-levels, displacements, and velocities.
The penalty for incorrectly specified parameters is test results that may have little relationship to real-world conditions. For example, the maker of large equipment used in anesthesiology requested a vibration test over a flat spectrum with an overall level of 1 G rms. The frequency bandwidth was 5 Hz to 2 kHz. When testing began, the test sample literally began to fly apart.
Subsequent analysis revealed the flat test spectrum was the problem. In a hospital, the anesthesiology equipment would never see 1 G rms over the frequency bandwidth used in the test. Reshaping the test spectrum to decrease the low-frequency weighting let the equipment pass its vibration test.
Quick punchlist: Designing test fixtures
- Test-sample size, quantity, mass, and geometry
- Rigid mount versus simulated-field installation.
- Level of displacement, frequency, and bandwidth of tests
- Identification of any test item resonances
- Is the fixture resonance-free within the test frequency bandwidth?
- Does the fixture install easily on the shaker table and change over easily to another axis?
- Does the fixture facilitate mounting in three axes?
- Three main components of fixture cost: design + materials + machining
Edited by Leland Teschler.