Vibration Measurement
What this domain tests
The Vibration Measurement domain covers the chain from transducer to reported value: accelerometer selection, mounting, signal conditioning, frequency analysis, unit conversion, and comparison against ISO machinery criteria. Exam items typically give you a sensitivity, a mounting method, or a spectrum setting and ask what the correct result or the limitation is.
Transducers and signal conditioning
Piezoelectric accelerometers are specified by sensitivity in mV/g and by mounted resonance. Charge-output types require an external charge amplifier; IEPE/ICP types carry built-in electronics and run on constant current, typically 2–4 mA, with a bias of roughly 8–12 V that must be verified at the connector. A dead or shorted channel is usually a bias-voltage fault, not a sensitivity fault.
Mounting
Stiffness order: stud-mounted, then cyanoacrylate/adhesive, then magnet, then hand-held probe. A magnet mount is convenient but typically limits usable bandwidth to about 2 kHz and can introduce its own resonance. Mass loading matters: keep the transducer mass below roughly one-tenth of the effective mass of the structure at the point of attachment.
Analysis
FFT resolution is Δf = fs/N; choose lines and bandwidth so the spacing resolves the phenomenon. Use Hanning for continuous random signals, flat-top where amplitude accuracy matters, rectangular for transients. Velocity = a/(2πf) and displacement = v/(2πf); high-pass before integrating to stop drift. ISO 20816 (which replaced ISO 10816) rates machines on RMS velocity in mm/s over 10–1000 Hz using zones A–D.
Worked example
A 100 mV/g accelerometer reads 10 mV RMS at 50 Hz. Acceleration = 10/100 = 0.1 g = 0.981 m/s². Velocity = 0.981/(2π × 50) = 0.981/314.2 = 0.0031 m/s = 3.1 mm/s RMS. That is the value entered on the ISO 20816 chart.
Traps
Mixing peak and RMS (peak = 1.414 × RMS for a sine); using a magnet mount and then trusting high-frequency data; integrating without a high-pass; picking the wrong window and smearing amplitudes; ignoring cable whip and ground loops; reporting a single-channel overall level when the fault is directional.