Vibration and shock testing

What vibration and shock do to a part

Vibration applies a repeated force over time. It can loosen a fastener, fatigue a wire, crack a solder joint, wear a contact or rub a coating through to the substrate. The damage often starts at a resonance, where a small input produces a large movement.

Shock applies a short, high-force event: a drop, a jolt, a transport impact or a switching transient in a mechanical system. It can bend a bracket, shear a fastener, displace a contact or crack a housing in one event.

Both tests are about the real journey, not the bench. A part that travels by road, sea and forklift sees a mix of broad vibration, occasional shocks and handling drops. The test profile should reflect that mix, or the result says little about the field.

Sine testing

A sine test sweeps one frequency at a time, usually at a fixed acceleration, from low to high and back. The classic use is a resonance search. The response of the part is measured while the frequency sweeps, and a peak in the response shows where the part amplifies the input.

A resonant dwell holds the part at that frequency for a defined time. It is a focused fatigue test. If a bracket has a mode near a motor speed, a dwell at that mode will show whether the bracket can survive it.

Sine testing is easy to understand and easy to compare. Its weakness is that real service vibration is rarely a single frequency. A part can pass a sine sweep and still fail in service because the real input is broadband.

A sweep rate that is too fast can miss a sharp resonance. The response needs time to build at each frequency, so a slow sweep gives a clearer picture. If the specification fixes a fast rate, note that some modes may not show up.

Random testing

A random test applies many frequencies at once, with a defined power spectral density. The profile describes how energy is spread across the frequency range. The overall level is often quoted as an RMS acceleration, or grms.

Random vibration is closer to road, rail, aircraft and machinery inputs. It excites several modes at once and produces a more realistic fatigue load. It is also harder to describe and harder to compare, because two profiles can share a grms value and still differ in where the energy sits.

A random test needs a control accelerometer, a defined axis and a duration. The duration should reflect the service life, often through an agreed acceleration factor. Without those details, a report that says random vibration at a given grms is incomplete.

Check the spectrum, not just the grms. Two profiles can share an overall level while placing their energy at different frequencies. One may excite a critical mode that the other never reaches. The breakpoints, the slope and the tolerance on the control spectrum all belong in the specification.

Shock testing

A shock test applies a transient pulse, often a half-sine or a sawtooth, with a stated peak acceleration and duration. The response of the part is often described by a shock response spectrum, which shows how the pulse excites different natural frequencies.

Shock is not a scaled vibration test. A short pulse can excite a mode that a long random run never reaches, and it can produce a peak force that causes immediate yielding rather than gradual fatigue. The two tests find different failures.

For a product that is dropped or jolted in handling, a drop test may be more representative than a laboratory shock pulse. For a mounted component, a specified shock pulse with a defined axis is usually the better comparison.

The pulse shape matters. A half-sine pulse and a sawtooth pulse with the same peak and duration do not excite the same response. State the shape, the peak, the duration and the tolerance, or the laboratory will choose a convenient one.

Matching the test to the risk

The table sets out what each test is good at and what it misses. Use it to build a sequence rather than to choose a single test.

Fixtures and mounting

The fixture is the most common source of bad vibration data. A flexible fixture adds its own resonances. A loose clamp lets the part rattle. A mount that differs from the service mount changes the boundary conditions, so the modes move.

A good fixture is stiff, light and bolted to the table at the same points the part uses in service. Where that is not possible, document the difference and treat the result as comparative rather than absolute.

Measure the fixture as well as the part. An accelerometer on the table and another on the part shows whether the input is reaching the part as intended. Without that check, a pass may only prove that the fixture absorbed the energy.

A quick tap test can reveal a loose mount before the run starts. If the fixture rings at a frequency near the part's own mode, the two will interact and the result will be hard to interpret. Stiffen the fixture or change the mount rather than accepting the coupling.

Sequence, axes and duration

A part usually needs testing in three axes. The service load may be dominant in one axis, but a bracket or a wire can respond in another. State which axes were tested and which were not.

Sequencing matters when several tests are combined. Thermal cycling can change a material or a joint before vibration, and vibration can open a crack that humidity then corrodes. The order should reflect the service life or the qualification standard.

Duration should be justified. A short run may find a gross fault, but fatigue needs time. If an acceleration factor is used to shorten the test, record the factor and the basis for it.

The order of axes can matter too. A part that fails in the first axis may not be fit to continue. Inspect between axes, or state that the run is cumulative and that a failure in any axis ends the test.

Acceptance and post-test inspection

Acceptance should be defined before the run. It may be no change in contact resistance, no crack in a visual or penetrant check, no shift in a critical dimension, or a functional check at the end of each axis.

Inspect the parts after the test, not only during it. Looseness, fretting marks, cracked plating and deformed brackets can be easy to miss while the shaker is running. Take photographs and keep the failed parts.

A test that passed with no post-test inspection is a weak record. The question is not whether the part survived the shaker. It is whether the part is still fit for service after the load it will actually see.

Keep the photographs with the report. A hairline crack or a fretting mark is easier to show than to describe, and the next engineer will need the image more than the pass line.