The velocity pickup and the accelerometer are the two families of case-mounted vibration sensor, and choosing between them shapes what a measurement channel can and cannot see. The velocity pickup is an older electromagnetic device that outputs velocity directly, while the accelerometer is a piezoelectric device that outputs acceleration and, where velocity is wanted, is integrated electronically. New installations overwhelmingly use accelerometers, and understanding why helps an engineer add vibration channels sensibly.
Velocity Pickup vs Accelerometer in one line: A velocity pickup is a self-generating electromagnetic sensor that outputs a velocity signal directly, while an accelerometer is a piezoelectric sensor that outputs acceleration and is electronically integrated when a velocity signal is needed. The accelerometer offers a wider frequency range, greater ruggedness, and smaller size, which is why most modern machinery-monitoring channels use an accelerometer with integration rather than a velocity pickup.
A velocity pickup is a moving-coil device: a coil suspended on soft springs moves relative to a permanent magnet as the case vibrates, inducing a voltage proportional to velocity. It is self-generating, meaning it needs no external power to produce a signal, and its output is already in velocity, which is convenient because casing vibration severity is commonly judged in velocity units. This directness is the historical appeal of the device.
A piezoelectric accelerometer instead uses a crystal that produces charge when a proof mass compresses it under acceleration, giving an output proportional to acceleration. It has no springs or moving coil, only the tiny elastic deflection of the crystal, so it is far more rugged and compact. Most modern accelerometers include a built-in amplifier that outputs a low-impedance voltage signal, making them easy to cable over distance and to interface to monitoring hardware.
Because the two speak in different units, getting velocity from an accelerometer requires integrating its acceleration signal once, which is done electronically in the sensor's conditioning or in the monitoring instrument. Integration is a well-understood operation, but it must be handled carefully at very low frequencies where small acceleration signals integrate into large, noisy velocity values. This is the main practical cost of the accelerometer's flexibility, and it is routinely managed in machinery-protection instruments.
Frequency range is where the accelerometer decisively wins. A stud-mounted accelerometer measures accurately from low frequencies up into the high kilohertz region, which is essential for detecting early bearing and gear defects that show up as high-frequency energy. A moving-coil velocity pickup is limited at both ends: it rolls off below its natural frequency, losing the low frequencies important on slow machines, and it has a comparatively low upper limit, missing the high-frequency defect tones an accelerometer captures.
Reliability and form factor also favor the accelerometer. The velocity pickup's springs and coil are delicate moving parts that wear and can fail, and the device is bulky and sensitive to mounting orientation because gravity acts on its sprung mass. The accelerometer has effectively no moving parts to wear out, is small and light enough to mount almost anywhere, and is far less orientation-sensitive. These are decisive advantages for permanent installations expected to run for years.
Sensitivity and cross-axis behavior round out the comparison. Accelerometers are available in a range of sensitivities to suit high-amplitude or low-amplitude machines, and good ones have low cross-axis sensitivity so they respond mainly to vibration along their intended axis. The velocity pickup offers less flexibility in these respects. The one lasting merit of the velocity pickup is its self-generating simplicity and direct velocity output, which is why a population of them remains in service even as new work standardizes on accelerometers.
When adding vibration channels to a monitoring system, the default modern choice is a piezoelectric accelerometer with electronic integration to velocity, because one sensor then serves both worlds: raw acceleration for high-frequency bearing and gear diagnosis, and integrated velocity for casing severity trending and alarms. This single-sensor coverage simplifies the I/O rack, since each bearing point needs only one transducer and its conditioning rather than separate devices for different frequency ranges.
The signal from a modern accelerometer suits industrial wiring well. Its built-in amplifier produces a low-impedance output that survives long cable runs to the monitoring hardware or SCADA input, and the conditioning module extracts the overall velocity, any band values, and the spectrum that stream into the historian. Where a velocity signal is required to match an existing standard or a legacy alarm setpoint, the integration is configured to deliver it, so the accelerometer coexists with older velocity-based thresholds.
There are still niches for the velocity pickup, chiefly replacing one in an existing installation where the mounting, wiring, and setpoints are all built around a velocity output and a like-for-like swap is simplest. But for a new channel feeding cloud or SCADA monitoring, the accelerometer's wider range, ruggedness, and small size make it the practical selection. The engineer's real decisions then become the mounting method, which sets the usable bandwidth, and how the integration and alarm units are configured, rather than which sensor family to use.
Accelerometers offer a much wider frequency range, reaching the high frequencies where early bearing and gear defects appear, and they are more rugged with no delicate moving parts to wear out. They are also small, light, and easy to cable. A velocity pickup's only lasting advantage is that it outputs velocity directly and needs no power, which is why new work standardizes on accelerometers with electronic integration.
Yes. An accelerometer outputs acceleration, and a velocity signal is obtained by integrating that output once, done electronically in the sensor conditioning or the monitoring instrument. This lets one accelerometer provide both raw acceleration for high-frequency diagnosis and integrated velocity for casing severity trending. The integration must be handled carefully at very low frequencies, which machinery-protection instruments are designed to do.
The main case is replacing an existing velocity pickup in an installation whose mounting, wiring, and alarm setpoints are all built around a velocity output, where a like-for-like swap is simplest. Its self-generating operation, needing no external power, can also be convenient in specific situations. For a new channel, though, an accelerometer with integration is almost always the better choice.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.