A product can pass every functional test on the bench and still arrive at the customer broken, or work perfectly for four months and then start failing in the field. In both cases the culprit is usually the same: the product spent time being shaken. A pallet on a truck sees continuous random vibration for days; a device bolted to a tractor sees it for its entire life.
Vibration failures are frustrating because they are invisible in a static analysis. The part is strong enough, the stress is a fraction of yield, and it still cracks — because a structure driven at its natural frequency does not see the load you applied, but that load multiplied by ten or thirty.
Shock and Vibration Are Two Different Problems
They are named together and need opposite solutions, which is why products often survive one and fail the other.
Shock is a single brief high-magnitude event: a drop, a bump, a forklift setting a pallet down hard. Measured in peak g and duration, it breaks things by exceeding strength — a bracket yields, a solder joint cracks. The response is strength, compliance, and energy absorption at impact, the domain of designing a product to survive a drop test.
Vibration is continuous, lower magnitude, and cumulative. It breaks things by fatigue — millions of small stress cycles growing a crack from a stress riser — and by loosening what was held only by friction. A structure that survives a 50 g shock can fail under a 2 g input sustained for ten hours; see material fatigue in product design. Design and test for both: passing one tells you almost nothing about the other.
What the Transport and Service Environment Actually Looks Like
You cannot design for vibration without a description of the input. Fortunately standards bodies publish profiles you can design against.
| Standard | Covers | Typical use |
|---|---|---|
| ASTM D4169 | Distribution cycle for shipping units | The general US answer for "will it survive shipping" |
| ISTA series (1A, 2A, 3A, 6 variants) | Packaged product performance | Retail and parcel shipment; some carriers require it |
| MIL-STD-810, Method 514 / 516 | Vibration and shock in service | Vehicle-mounted, defense, rugged equipment |
| IEC 60068-2-6 / -2-27 | Sine, shock, random vibration | General electronics |
Truck transport is dominated by low frequencies — most energy below 100 Hz, with suspension and tire resonances peaking at 3 to 20 Hz where large assemblies like to resonate. Parcel shipment is shock-dominated instead, because the package is handled, dropped, and thrown more than it is driven. And service environments often deliver more total exposure over a product's life than transport ever did.
Pick the profile that matches how your product will really move and live, and write it into the requirements at the start. Retrofitting vibration robustness after tooling is committed is expensive.
Resonance Is the Number That Decides Everything
Every structure has natural frequencies. Drive it near one and the response is amplified by the quality factor Q — commonly 10 to 50 for a lightly damped structure, so a 1 g input becomes 20 g at the resonant part. Four rules follow:
- Know the first natural frequency of every significant subassembly — enclosure, mounted board, display, heavy component on a bracket. Modal analysis gives this quickly; see FEA simulation in product design.
- Keep it high. Push the first mode a factor of two or more above the dominant input frequencies. Stiffness raises it, mass lowers it; ribs, shorter unsupported spans, and more mounting points are the cheapest tools.
- Avoid coupling. Two subassemblies with similar natural frequencies feed each other.
- Watch heavy things on long arms. A transformer, battery, or heat sink cantilevered off a board is the classic failure. Support it to the chassis.
Circuit boards deserve special mention: an unsupported span with heavy components resonates in the low hundreds of hertz, cracking solder joints and lifting connectors. Mounting screw spacing is the main lever — see mounting a PCB in an enclosure.
Isolation and Damping
When you cannot design the resonance away, decouple the sensitive item from the input or absorb the energy. These are different mechanisms, and using the wrong one makes things worse.
Isolation puts a soft mount between the vibrating structure and the item, creating a low-frequency mass-spring system that filters out everything above it. The critical rule: isolation only works above roughly 1.4 times the mount's natural frequency — below that, and especially at it, the mount amplifies the input, so a careless isolator can make a product fail a test it previously passed. Put that frequency well below the lowest input you care about and verify the static deflection.
Damping converts vibrational energy to heat — viscoelastic pads, constrained-layer sheets, elastomeric washers, foam. It lowers amplification at resonance rather than moving it, and it is the most reliable way to quiet a panel that rings. Elastomer selection matters more than teams expect: durometer, compression set, and temperature all change stiffness, and a mount that works at 70°F can be twice as stiff at 0°F. The framework is in choosing durometer for rubber and TPE parts; the extreme case is gimbal and vibration isolation design.
The Details That Come Loose
Fatigue cracking is the dramatic failure; loosening is the common one.
- Threaded fasteners back out under transverse vibration however tight they were, unless something prevents it — prevailing-torque nuts, thread-locking adhesive, wedge washers, or enough clamp load that slip never occurs.
- Connectors fret: micro-motion wears through plating and builds resistance, producing intermittent faults that are maddening to diagnose. Latching connectors and strain relief are the fix, with cables secured close to the connector.
- Press fits walk over time; adhesive or a mechanical lock is cheap insurance.
- Anything on a long unsupported wire eventually fails at the flex point. Support it.
Testing: Sine, Random, and Shock
Three test types on a shaker table, each answering a different question. Sine sweep drives one frequency at a time with accelerometers on the product; it is diagnostic, finding your resonances and their amplification. Run it first and you already know where the product will fail. Random vibration drives all frequencies at once to a specified power spectral density, which is what real environments do — hours per axis, in all three axes, and the closest thing to a truthful simulation of a truck. Shock applies defined pulses (half-sine, sawtooth, or a specified response spectrum) in each axis and direction, covering handling, drops, and impacts.
Budget realistically: lab time for a full ISTA or ASTM D4169 sequence commonly runs $3,000 to $12,000, and MIL-STD-810 campaigns run higher. That is cheap next to a field failure, which is why this belongs in the plan described in reliability testing for a new product. Instrument the product so a failure tells you where and why.
The Package Is Part of the Design
For transport, product and packaging are one system, and it is usually cheaper to solve the problem in foam than in structure. A good cushion set changes the input by an order of magnitude — but only if its natural frequency is placed correctly. Foam that is too soft bottoms out on a drop; too stiff and it transmits everything. Cushion curves from the supplier, matched to product weight and bearing area, are the design tool.
Test the shipping configuration as it will really ship, including pallet patterns and stacking — not a bare unit on a table. Building a model specifically to test packaging and transit is standard practice.
Projects House designs mechanical hardware for environments that shake — modal analysis, isolation and damping, fastener and connector strategy, and test plans against ASTM, ISTA, and MIL-STD profiles. If your product is arriving damaged or failing in service, tell us what is breaking through our contact form.