Random Vibration Testing for PCB Assemblies and Components
A board that survives a drop test can still fail on a shaker table, and a board that survives a shaker table can still fail in a vehicle. Vibration is a cumulative loading process, and the damage it does depends on the frequencies involved, the duration and the mechanical design of the assembly. Vibration testing is how those three are brought together into a measurable result.
Why Vibration Testing Is Necessary
Every structure has preferred frequencies at which it moves easily, and a printed circuit board is no exception. When an excitation matches one of those frequencies, the displacement at that point can be many times larger than the input, and the components mounted there experience amplified stress. Vibration testing identifies where that amplification occurs and whether the assembly can survive it.
The test is also a design tool. Comparing two board thicknesses, two mounting schemes or two component placements under identical excitation shows which choice reduces the response, and that information can be applied to the next design before a failure occurs in the field.
Random Versus Sine Vibration
Sine vibration applies energy at a single frequency at a time, sweeping through a range. It is used for resonance searching and for qualifying equipment that will encounter dominant tones, such as machinery with a rotating part. Because the energy is concentrated at one frequency at a time, it is an efficient way to find where a structure responds.
Random vibration applies energy across a broad band simultaneously, described by a power spectral density profile. This is more representative of transportation, engine vibration and most field environments, and it excites all the modes at once, which produces a more realistic stress distribution. Most qualification programmes use a resonance search with sine excitation followed by endurance testing with random excitation.
<img src="https://www.gopcba.com/wp-content/uploads/2024/09/QA_Inspection.webp" alt="PCB assembly mounted on a vibration shaker fixture with accelerometers attached” />
Resonance and Mode Shapes
A resonance search is performed at low level to identify the frequencies at which the assembly responds most strongly. Response is measured with accelerometers placed at representative locations, and the resulting plot shows the natural frequencies and the amplitude of the response at each. The level is kept low to avoid damaging the sample before the endurance test begins.
Knowing the frequency is not enough; the mode shape matters as much. A board that flexes along its length will stress components in the middle, while one that twists will load the corners. Mapping the mode shape with several accelerometers or with a scanning laser vibrometer shows where the strain concentrates, which is where the test should be monitored and where the inspection should focus afterwards.
Fatigue and Cumulative Damage
Vibration damages joints through fatigue, which is a cumulative process. Each cycle of stress consumes a fraction of the joint’s life, and the fractions add up regardless of whether the cycles arrive in one long run or several short ones. This is why duration matters as much as amplitude, and why a short test at high level cannot substitute for a longer test at the specified level.
The relationship between stress amplitude and cycles to failure is non-linear, so a modest increase in response amplitude produces a disproportionate reduction in life. That non-linearity is what makes resonance so dangerous: a component sitting at a mode peak can accumulate damage far faster than one a few millimetres away, even though both are on the same board.
Fixture Design and Its Influence
The fixture is part of the mechanical system and it changes the result. A stiff fixture transmits the specified input faithfully, while a flexible one absorbs energy, adds its own resonances and may either mask a board failure or introduce one. Verifying the fixture with a low-level survey before testing the product is standard practice.
Mounting points also matter. Reproducing the way the board is fixed in the product is essential, because a board supported at four corners behaves very differently from one clamped along two edges. Where the product mounting is not yet defined, testing should be repeated for the plausible variants rather than assuming the most convenient one. The interface between board and enclosure belongs to the same family of decisions as those described in this guide to board outline tolerance.
<img src="https://www.gopcba.com/wp-content/uploads/2026/08/rigidflex_pcb.png" alt="Fatigue crack in a solder joint after random vibration endurance testing” />
Monitoring During the Test
Continuous electrical monitoring detects intermittent opens as they occur, which is far more informative than an inspection after the run. A daisy-chained coupon routed through the critical joints gives a direct indication of joint integrity, and event detectors record the moment of each interruption along with the vibratory level at that time.
Structural monitoring with accelerometers provides the complementary information. Measuring the response at several points shows whether the amplification is growing during the test, which can indicate a joint loosening or a crack propagating. Changes in the response spectrum between the start and end of a test are a useful early warning even when continuity has not yet failed.
Failure Locations and Mechanisms
Solder joints fail by fatigue cracking, usually initiating at the interface between the solder and the component termination or the pad. The crack grows progressively and may take many thousands of cycles to become an open circuit, which is why the failure often appears late in a test with no warning signs beforehand.
Other failures include cracked component bodies, particularly on ceramic capacitors, fractured leads on through-hole parts, and damage to the laminate itself at mounting holes. Where the test causes a failure, a cross section of the joint or component is needed to confirm the mechanism before any design change is made. The resulting signatures overlap with those described in this overview of component tolerance and reliability.
Interpreting and Extrapolating Results
The results of a vibration test are specific to the profile, the duration and the fixture. Extrapolating from a passed test to a different environment requires assumptions about the relationship between stress and life that are rarely justified in full. Where the target environment is known, testing against a profile derived from measured data is far more defensible.
When a test produces an early failure, the useful question is which parameter drove it. Reducing the response by stiffening the board, moving a heavy component away from a mode peak or adding a support point all change the outcome. Re-testing after a single change confirms which measure was effective, and the record of that process is far more valuable than a single pass or fail result. Interpreting the evidence follows the same discipline as PCB quality judgement in production.
Specification and Qualification
A specification should define the excitation type, the profile or sweep range, the duration, the mounting configuration, the monitoring method and the acceptance criteria. It should also state whether the test is a screen or a qualification, because the required sample size and the interpretation of the results differ between the two.
Qualification normally tests the worst-case configuration, which is usually the largest board with the heaviest components mounted furthest from the supports. Testing only a convenient variant risks qualifying a configuration that will never be built while leaving the vulnerable one unverified.
FAQ
What is the difference between sine and random vibration testing? Sine testing applies one frequency at a time and is efficient for locating resonances and for tonal environments. Random testing applies a broad band of frequencies simultaneously, which excites all modes at once and is more representative of transportation and most field conditions.
Why is a resonance search run at low level? Because the purpose is to identify the natural frequencies and the mode shapes without consuming fatigue life before the endurance test begins. A search run at full level could damage the sample and invalidate the subsequent test.
Can a short high level test replace a long low level test? No. Fatigue damage is cumulative and the relationship between amplitude and cycles is non-linear, so a short intense exposure is not equivalent to a long mild one. The specified duration exists for a reason and substituting amplitude for time changes the failure mode that the test provokes.
