Volume 1 › Chapter 1 › 1.1
Chapter 1: What a Vibration Test Is For

1.1 Why Hardware Sees Vibration: One Life, Start to Finish

about 9 minutes · 2 figures · quiz of 5 · Slides

After this section you can: trace a piece of hardware from the factory to the end of its life, name every vibration environment it meets along the way, and say what each one does to it.

A vibration test is a question about one moment in a piece of hardware's life. Before you can ask it well, you have to be able to see the whole life.

The Department of War (the Department of Defense, by its older name) has a name for that whole life. The Stockpile-to-Target Sequence, the STS, is the ordered list of everything that happens to a weapon from the day it is built to the day it is used or retired, with the environments written down at every step. MIL-STD-810 calls the same idea the life cycle environmental profile, and commercial programs often say mission profile or use profile; this book uses the STS because it is the most complete answer anyone has written to the question this chapter asks. It is worth knowing about even if you will never test a weapon, because every product has an STS. Most industries never wrote theirs down.

The STS is not a vibration document. It covers every environment the article meets: temperature and thermal cycling, humidity, altitude and pressure, salt fog, sand and dust, electromagnetic interference, radiation, and the abnormal and hostile environments (fire, drop, crush, immersion, attack) in which the design must stay safe and must not function. Vibration and shock are two columns in a wide table. This series covers those two columns and leaves the rest to the disciplines that own them. But you should know the table exists, because the vibration test is usually one of a set, and the set has to agree with itself: the same article in the same configuration and revision, one sequence of steps (each test covering the steps where its own environment occurs), the same requirement document.

Walk a warhead through it. It is manufactured and assembled, and along the way it is handled: carts, forklifts, hoists, the bump of a pallet set down a little hard, the floor vibration from machinery running in the next bay. It is packaged and shipped, by truck over road, by rail, by cargo aircraft, by ship, and each of those is a different environment with a different character and duration. Road and rail are broadband and go on for days. An aircraft's engines put tones into everything bolted to the airframe. It goes into storage and sits quietly for years, then gets handled all over again at every inspection. It goes back for maintenance and surveillance, which means disassembly, bench handling, and transport in both directions. It is mated to its platform with hoists and carts and bolted down. It is carried, under a wing or in a bay (captive carry), where engine tones and broadband buffet (the shaking from turbulent airflow over the airframe) run together for hours at a time; or it rides in a missile, where the same thing happens faster. And it is delivered: launch, staging, free flight, separation and pyrotechnic events, reentry, terminal maneuvers. Every one of those steps is an environment with a level, a duration, and a character.

The STS sorts those environments into two columns. Normal environments are the ones the design must survive and keep working through. Abnormal environments are the accidents: the design must survive them safely, but it is not expected to work afterward. Vibration testing lives almost entirely in the normal column. The profiles a lab runs descend from the normal STS steps by way of measurement, enveloping, and the standards, which is where Section 1.3 starts.

Now take the weapon away and the list is still there. A satellite: factory, truck to the range, launch, on-orbit quiet, thruster firings. A vehicle part: assembly line, then a decade of road. A shipboard sensor: machinery and sea state for years. The steps are different in every case. The discipline of writing them down in order, with the environment at each one, is the same, and it is the first thing a test engineer owes the hardware.

Each environment has a character, and the character is what will decide the test type. Road, rail, and boost (the powered climb of launch) are broadband and stationary: the energy is spread across frequency, and the level and the shape of the spectrum hold steady for a long time (stationary is the statistician's word for that). Aircraft carriage and machinery are tonal and mixed: strong lines at the engine or shaft orders (an order is a multiple of a shaft's rotation rate: a shaft turning at 6,000 rpm puts its first order at 100 Hz and its second at 200 Hz) sitting on a broadband floor. Separation, pyrotechnic events, drops, and gunfire are transient: a lot of energy in a very short time. Real environments combine these, and real tests have to cover the combinations. As a first rule, a broadband environment becomes a random test (all frequencies driven at once, shaped to the profile's spectrum), a tonal one a sine test (one frequency at a time, swept or held) or sine on random (tones laid over a random floor; some controllers call it mixed mode), and a transient one a shock test (a short pulse, or a measured event reproduced); the test types, and how a profile is built for each, are the subject of Volume 2.

What vibration does to hardware along the way is a short list that covers most of what you will ever see. Fatigue at resonances (the frequencies at which a part's own motion builds up). Fretting and wear at joints and contacts (fretting is the wear from small, repeated sliding between surfaces held together). Connector and solder failures. Relative motion between parts that were meant to stay put. Joints that loosen. Most failures happen at a resonance, because that is where the response accelerations, forces and strains are largest, and most start at or near a stress concentration: a mechanical joint, a sharp corner, a hole. Many are cumulative: a joint loosens a little at a time, a structural member fatigues, an electrical connection degrades until it fails. And damage started in one environment often shows in a later one. The crack that opens during carriage may have started on the truck.

Decide. Before anything else, place the test on the article's sequence. Which step, what character, how long. If nobody can say which step the test represents, the test is not ready to run.

FactorybroadbandTransportbroadbandStoragequietMaintenancehandlingMatinghandlingCarriagetonal + broadbandLaunchbroadbandSeparationtransientGray bars: relative duration of the step, seconds to years, not to scale. Blue: the character of the vibration at that step.
Figure 1.2. The sequence as a timeline, factory to target, with the character and duration of the environment at each step.
StepEnvironmentCharacterDuration scaleTest type it points toSets up
Manufacture, assemblyhandling, floor vibrationtransient, low broadbandminutesshock; low-level randomloosening and contact damage, mostly at joints
Transportroad, rail, aircraft, shipbroadband; tonal on aircrafthours to daysrandom; sine or sine on randomfatigue at resonances, fretting
Storagequiet, then handling at inspectiontransientyears, then minutesshockmaterial set, creep, then loosening
Maintenance, surveillancebench handling, transporttransient, broadbandhoursshock, randomconnector and joint wear, handling errors
Matinghoists, carts, bolt-uptransientminutesshockbolts under- or over-torqued
Carriageengine tones on broadband buffettonal and mixedhourssine or sine on randomfatigue at resonances, especially at stress concentrations
Launch, boostbroadband, high levelbroadband, stationaryminutesrandomfatigue, more relative motion between components
Separation, terminalpyrotechnic, aerodynamictransientmilliseconds to secondsshockbrittle failure, latch and contact chatter, electronic failure
Figure 1.3. The same steps as a table: environment, character, duration scale, the test type it points to, and the failure mode each tends to set up. These are the vibration and shock columns of a wider table.

Common Questions

Our hardware is not a weapon, and nobody here has written its sequence down. Who do I ask which step a test represents?

Start with the requirement: it is supposed to name the step on the sequence (1.5, The requirement), so the question goes to whoever wrote it, usually through the customer. The vibration test is usually one of a set, so the thermal or climatic requirement for the same article may already list the steps under another name, such as a mission profile or a use profile. If none of them does, the lab asks before it schedules, the same as for any other gap in the requirement.

If damage is set up earlier in the life, should the tests run in the order of the sequence, on one article?

The chapter does not set a rule for that. Its point is that damage builds through the life, so a test of one step on a fresh article does not carry what the earlier steps left behind, and someone should decide on purpose whether that matters for this article. Which tests run on which article, and in what order, is settled when the set of tests is planned, and the plan has its own chapter in Volume 2.

When an environment has tones on a broadband floor, do I run one combined test or two?

Real tests have to cover the combinations, and 1.1 gives only the first rule for which test type a character points to. Whether a program runs one combined test or separate ones, and how each profile is built, is the subject of the test types in Volume 2. The requirement names the test type; the lab's part is to check that it fits the character of the step it stands for.

Quiz: Try It

Open book. 5 questions drawn from this section's pool. 4 correct passes the quiz. Nothing is recorded here; in the course your passing attempt opens the next section.