IEC 60068-2 Vibration Tests: What 'Pass' Means and What It Doesn't

The most common conversation we have with customers about vibration testing starts with "we passed 60068" and ends with us asking which sub-test, which profile, which severity level, and along which axes. The IEC 60068-2 family is not a single test — it is a toolbox of around fifty individual procedures, each with optional severities, durations, and pass criteria. "We passed 60068" without the rest of that sentence tells us nothing.

For mechanical and electromechanical electronics, the four sub-tests that matter most often are:

• 60068-2-6 (Fc) — sinusoidal vibration, frequency-swept. The classical resonance-finding test. Pass means no functional failure and no structural damage during a swept-sine excursion.
• 60068-2-64 (Fh) — broadband random vibration. The realistic-environment test that matches transport and in-service conditions for most industrial gear.
• 60068-2-27 (Ea) — mechanical shock, half-sine pulses. Single-event survival.
• 60068-2-29 (Eb) / Fda — repetitive bump or wide-band random shock for transport simulation.

"A report that says 'passed Fc 10–500 Hz, 0.5g, 2 hours per axis' tells you exactly what the design survived. A report that just says 'passed IEC 60068' tells you the test lab printed a sticker. We've seen both." — Pioneer Horizon test team

The rest of this article walks the profile selection process, the mounting-decoupling tricks customers most often skip, and the seven pre-test checks we run on every unit before it goes on the shaker.

Fc (60068-2-6) sweeps a sinusoidal excitation through a frequency band to find resonances, then dwells at each resonance to verify the structure can survive prolonged excitation at its weakest mode. The profile is defined by four numbers: frequency range, acceleration amplitude (or displacement at low frequencies), sweep rate, and endurance time per axis.

The standard severity table

IEC 60068-2-6 lists severity classes from "extra mild" (1.5g, 10–55 Hz) to "extra severe" (10g, 10–2000 Hz). For typical industrial cabinet-mounted electronics, we recommend testing at:

• Frequency range: 10–500 Hz, with a 0.75 mm peak-to-peak displacement below 60 Hz, transitioning to 5g constant acceleration above 60 Hz.
• Sweep rate: 1 octave per minute logarithmic.
• Duration: 10 sweeps per axis (about 75 minutes per axis at the rate above), all three orthogonal axes.

Resonance dwell

After the initial sweep, the test plan should call out resonance-search dwell times at each identified resonance — typically 10 minutes at each frequency where the structure shows a Q-factor above 5. Skipping the dwell is the most common way a test plan "passes" but fails the field environment, because real transport runs produce minutes-to-hours of dwell time at the dominant rail or road frequency.

Two profiles we hold customers to

1. Industrial cabinet (default): 10–500 Hz, 5g, 10 sweeps per axis + 10 min dwell at each resonance with Q>5.
2. Vehicle-mounted: 10–2000 Hz, 10g, 20 sweeps per axis + 30 min dwell at each resonance. This is the closest IEC equivalent to typical automotive vibration profiles.

If your end-use environment is harsher than industrial cabinet — vehicles, rail, rotating machinery — Fc alone is insufficient. You also need Fh (random) and one of the shock procedures.

Real environments are broadband random, not sinusoidal. Fh (60068-2-64) excites the device under test (DUT) with a power spectral density (PSD) profile shaped to match the expected service environment. PSD is in g²/Hz, integrated over the frequency band to give an overall grms value.

PSD profile selection

The right PSD depends on the service environment:

• General-purpose industrial: 10–500 Hz flat at 0.02 g²/Hz → 3.16 grms overall. 30 minutes per axis.
• Vehicle-mounted (commercial): 5–500 Hz shaped per MIL-STD-810 transport profile or equivalent → ~7 grms. 60 minutes per axis.
• Rotating-machinery payload: 10–2000 Hz, shaped with a notch at the machine's known fundamental → 10–14 grms. 60 minutes per axis.

Why grms can mislead

A 5 grms profile concentrated at 50–200 Hz is much more damaging than a 5 grms profile spread evenly over 10–2000 Hz, even though they share the same overall grms. The PSD shape matters as much as the integral. Always insist your test plan specifies the PSD plot, not just the grms number, and that the plot matches the service environment within ±3 dB at the dominant peaks.

The fatigue-equivalence trap

Customers sometimes ask to run a shorter Fh test at higher grms to "save time on the shaker." This works in theory via the Miner's-rule fatigue-damage exponent — typically n = 4 for electronic structures — but the trade-off is brittle. A 4× time compression requires √4 ≈ 1.41× higher grms only if the fatigue exponent is exactly 4, which it usually isn't. We do not recommend time-compressed Fh as a substitute for full-duration testing.

The single most common mistake we see in customer-supplied test plans is the mounting fixture. The IEC 60068-2 procedures assume the DUT is rigidly coupled to the shaker table such that the input vibration reaches the DUT mounting feet without attenuation or amplification. In practice, a poorly-designed fixture can amplify input by 3–6 dB at fixture resonances or attenuate by 10+ dB at antiresonances. Either way, the test result is meaningless.

Six fixture rules we hold to

• Fixture stiffness: the first natural frequency of the fixture must exceed the test band ceiling by at least 1.5×. For a 10–500 Hz test, the fixture should not resonate below 750 Hz. We typically machine fixtures from 25–40 mm aluminium plate.
• Mass ratio: fixture-plus-DUT mass should not exceed 50% of the shaker's rated payload. Above that, the controller cannot maintain the PSD shape at high frequencies.
• Mounting plane: the DUT mounts to the fixture in the same way it mounts to the customer's product — same screw torque, same washer stack, same isolators if any. Replacing rubber isolators with rigid mounts to "be conservative" actually invalidates the test.
• Control accelerometer: placed on the fixture as close as possible to the DUT mounting interface, not on the shaker armature.
• Monitor accelerometers: at least one on the DUT at the expected highest-amplitude location, plus one at the centre of any large PCB.
• Test all three orthogonal axes: failing to test all three is the single most common deficiency in customer test reports we audit.

What customers most often forget

Cable strain relief. Connectors that work fine at static load fail in vibration if the cable mass is allowed to swing as a cantilever. We add cable clamps within 50 mm of every connector body before the DUT goes on the shaker, and we measure the clamp position back into the design files so production builds match.

Before any product goes on our shaker, we run a seven-point check. Skipping any one of these will, on average, double the failure rate at the formal test.

1. Modal pre-screen. A low-amplitude (0.2g) sweep at the test rate to identify resonances without inducing damage. Frequencies and Q-factors recorded.
2. Functional baseline. All product self-tests, sensor calibrations, and communication links exercised and recorded. Anything that worked before the test should work during and after.
3. Visual inspection. Conformal coat, solder joints, connector seating, mechanical fasteners with torque marks.
4. Fastener torque audit. Every screw to spec, witness-marked. A loose M3 on the heatsink is the classic post-test "failure" that is really a build-quality issue.
5. Cable management. Strain reliefs verified, harness routing matches production.
6. Power and load conditions. Test at worst-case load, not at idle. Some failure modes only appear when the high-current rail is active.
7. Instrumentation calibration check. Control and monitor accelerometers cross-validated within 5%.

Reading the report

A defensible IEC 60068-2 test report should contain the following — if any item is missing, the report does not constitute proof of compliance:

• Test sub-test reference (e.g., "60068-2-6 Test Fc" or "60068-2-64 Test Fh, Method 1").
• Severity profile (frequency range, amplitude, sweep rate or PSD plot).
• Duration per axis and number of axes tested.
• Mounting description and fixture diagram.
• Functional pass criteria — what 'pass' means for your product, defined before the test.
• Pre-test, during-test, and post-test measurements with timestamps.
• Failures and their resolution, if any.

If you would like our test team to review your existing vibration plan — or to author one from scratch against your service environment — share the mechanical envelope and target environment and we will return a costed test plan within a week.