So You're Going To Witness A Vibration Test
by Wayne Tustin

This article is intended for the reader (here addressed as "you") who has been given the task of witnessing a vibration test. You want to focus your attention on important aspects of the test, but you are not sure which aspects are important and which you can more or less ignore.

This article is not intended to increase the knowledge of readers who are expert in vibration. Some experts may disagree with what I identify as important. No, this article is written for individuals having practically zero knowledge of vibration testing. By the way, the shaker is that big iron thing. Accelerometers are those little stainless steel things.

The article is comprised of questions, requests and possible rejoinders. The first dozen are intended for you to ask your boss (who gave you the assignment) or possibly your coworkers. Opinions on these topics, within your own agency, can be useful. Possibly some of your colleagues have witnessed similar vibration tests, possibly on similar hardware, possibly at the same test lab.

Questions before you leave the office
Why is vibration important to my agency?
Possibly production quantities of the hardware you will see tested (a prototype DUT or device under test) will encounter severe vibration in transportation and shipping as well as in actual service. Your agency does not want hardware failures. You should know something about the purpose and projected usage of the hardware under test.

Why are vibration tests necessary?
To physically demonstrate that the prototype hardware has more than adequate strength to survive the anticipate rigors of transportation, shipping and service.

Why is it necessary to witness tests?
Because a few unscrupulous suppliers have claimed they had conducted specified tests, when in fact they had eased up on test severity or even had not tested at all.

What is a vibration test?
You will probably see one or more DUT specimens being shaken under controlled conditions.

What equipment am I likely to encounter?
A shaker (really an overgrown loudspeaker) will shake the DUT(s) through an adapter (called a fixture). Vibratory motion will be sensed by one or more accelerometers. If a sinusoidal vibration test is taking place, a display will show the vibration frequency (which may be fixed or may be sweeping across a range of frequencies) as well as the magnitude of motion (displacement, velocity and/or acceleration). If a random vibration test is taking place, a display will show the range of frequencies and the spectral density within that range. A specialized computer will regulate motion intensity and provide the displays.

Are there any known weaknesses (of this specific product, of this specific manufacturer, or of this specific test lab) of which I should be aware? By asking this question, you are asking your superiors and your colleagues for a "heads up" warning that can give you a little extra strength.

Have we prior test reports concerning this specific product? What happened last time?

Who will receive the report of the test I'm to witness?

Will the DUT(s) be operating or non-operating during the vibration test? If operating, who is responsible for supplying loads and power, for supplying and monitoring input and output signals?

Who will decide whether the product passed or failed? I know that my job is primarily to witness that the test was conducted, but my observations may be useful to whomever decides what happens next.

Is this test of direct concern to our agency, or am I to witness the test on behalf of another agency? Does our "customer" agency have any specific or unusual requirements?

Is the test (that I'm to witness) described in a Military Test Standard such as MIL-STD-810F? Or another country's military standard? Or some commercial standard? If so, which specific section is to be followed? Glance at whatever document is identified, so that you have some prior knowledge of the probable test setup.

Requests at the test lab
Lacking vibration expertise, you won't be able to evaluate the test facility and equipment, other than by appearance. However, you can evaluate test personnel. Here are 15 requests you might use to "qualify" them. Consider, for instance, the question What is vibration? One answer to the question is this: "Vibration is mechanical oscillation or motion about a reference point of equilibrium." You may wish to download and print this and other terms from http://www.equipment-reliability.com/glossary.htm. Alternately, your library might have TEST Engineering & Management June-July 1995, in which some of these and other dynamics terms were defined. If you are not too serious a person, you may chuckle at this, by Ann Landers: "Vibration is a motion that can't make up its mind in which direction it is going."

At the test lab, instead of asking it as a question, "What is vibration?" you might phrase it as a slightly self-empowering request: "Before you commence to conduct a vibration test, I wonder if you could define the word 'vibration' for me." The implication here is "How can you be trusted to conduct a vibration test if you don't know what vibration is." A few more requests such as that might bring an experienced test engineer to take over, replacing a "greenhorn" running his first test.

Possibly you are thinking: what other requests can I make (that will be appropriate and pertinent and add to my own knowledge, yet not make it too obvious that I'm new at test witnessing)? How can I push test personnel a little beyond their "comfort zone"? How can I politely ask test personnel to convince me that they are conducting the test properly? A few of the following may seem appropriate, before or during or after the test. Don't make all your requests at once. Just "toss in" an occasional request. Don't worry unduly about whether or not you understand the questions you are asking. Above all, don't be afraid of test personnel. A decade ago, the average test conductor had 5+ years experience. Today it's more like 0.5. Or less. He may well be hoping, "Gee, I hope this test witness fellow doesn't ask me a lot of questions."

So - here we go.

Please show me graphs of the vibration (and possibly temperature) at individual sensors. Test personnel may be satisfied with an "average" graph. Remember the apocryphal story of the man who drowned in a lake of average depth six inches. Would test personnel enjoy placing one hand in boiling water at +100oC, with the other in ice water at 0oC? Of course not. Yet the average temperature is +50oC.

Please show me your oscilloscope, displaying various "time domain" signals. Ask: Where are these signals coming from?

Is that spectrum ('frequency domain" signal) averaged? Averaged between what and what? Please also show me the "unaveraged" spectrum.

Please show me that you know how to convert deciBel (dB) ratios into numerical ratios.

Please show me your certificates that verify calibration of your instruments, including all sensors.

Please show me documentation that each sensor channel been end-to-end verified.

Please show me evidence that this fixture has previously been experimentally evaluated.

How much lateral vibration is the product receiving?

Please show me where the sensors are located.

Who decided upon those locations? What were their criteria?

If accelerometers are located some distance away from the DUT, ask: Why are measurements not being taken close to the DUT? on the DUT?

Look at pretest photographs. Are the pictures clear? Were they taken from different angles? Do they clearly show the DUT? How the DUT was attached to the shaker? Are the various input and response accelerometers (and their sensitive axes) clearly shown and identified? Could the setup be duplicated, if necessary? With these pictures, will people who read the test report be able to visualize what took place?

Conclusion
Please don't be nasty. Or overbearing. Make your requests in a dignified manner, but let test people know that you expect them to give you what you ask for. Here is a very powerful phrase which you shouldn't have to use more than once. It is, "I'm sorry, but I won't be able to accept the results of this test."

This article has been published at Test Engineering and Management Magazine in Oct/Nov - 2001. The magazine's editor kindly gave us permission to post it here.