Before the piezoelectric accelerometer took over vibration measurement, the moving-coil velocity transducer was how casing vibration was sensed, and it is still found bolted to older turbomachinery today. It is elegantly simple: a coil and a magnet, one held by a spring, that generate a voltage directly from motion without any power supply or electronics. And it produces its output in the very units that vibration severity charts are written in. This guide explains how the moving-coil velocity transducer works, why its native velocity output is convenient, where its low-frequency and cross-axis limitations bite, and why it hangs on in legacy installations despite the accelerometer's dominance.
Velocity transducer in one line: A moving-coil velocity transducer is a self-generating vibration sensor in which a coil moves relative to a magnet as the machine vibrates, inducing a voltage that is directly proportional to the velocity of the vibration. Its native output is velocity, typically in millimeters per second, which matches the units of the widely used vibration severity standards without any conversion. It needs no external power, but it has a limited low-frequency response, is sensitive to orientation, and contains moving parts that wear, which is why the piezoelectric accelerometer has largely replaced it except on legacy turbomachinery.
The moving-coil velocity transducer relies on electromagnetic induction. Inside the sensor, a coil and a permanent magnet are arranged so that one of them is suspended on a soft spring while the other is fixed to the sensor case. When the machine vibrates, the case and the fixed element move with it, but the spring-suspended element lags behind because of its inertia, so the coil and magnet move relative to one another. That relative motion through a magnetic field induces a voltage in the coil, and by the physics of induction that voltage is proportional to the velocity of the relative motion. The sensor thus outputs a signal that directly represents vibration velocity.
The most important consequence of this design is that it is self-generating. The sensor produces its own output signal from the motion itself, so it needs no external power supply, no charge amplifier, and no built-in electronics. You can connect it and read a usable velocity signal with nothing more than a cable. This simplicity and robustness, and the fact that a purely mechanical-electromagnetic device has no electronics to fail, was a large part of its appeal when it was the standard sensor and is part of why so many remain in service.
The design is a spring-mass system, and that shapes its behavior. The suspended element must move freely relative to the case for the sensor to work, which is why it has real moving parts inside, unlike a solid-state accelerometer. The natural frequency of that internal spring-mass system sets the sensor's useful range: above it the sensor faithfully reports velocity, and near and below it the response falls away. This internal mechanical resonance is the root of both the sensor's convenience and its principal limitation.
The standout advantage of the moving-coil transducer is that it measures velocity natively, and velocity in millimeters per second is exactly the quantity that the common vibration severity guidelines use. An accelerometer, by contrast, measures acceleration and its signal must be electronically integrated to get velocity for comparison against those charts. The velocity transducer skips that step entirely, delivering a signal that can be compared to severity zones directly. For overall casing-vibration monitoring judged against velocity-based standards, that native match is genuinely convenient.
The main limitation is at low frequency. Because the sensor depends on its suspended element lagging the case, it only works well above its internal natural frequency, and below that frequency its output rolls off and no longer represents the true velocity. That makes the moving-coil transducer poor at very low frequencies, where slow machines and certain faults live, and it means the sensor simply cannot report vibration below its usable range. An accelerometer generally extends to much lower frequencies, which is one reason it is preferred where low-frequency content matters.
The sensor is also sensitive to how it is oriented, a cross-axis or mounting-orientation limitation. Because it contains a real suspended mass on a spring, gravity acts on that mass differently depending on whether the sensor points up, down, or sideways, which can shift its behavior with orientation, and the sensor is designed to respond along one axis so motion in other directions is not measured cleanly. Combined with the moving parts that wear over time and can change the sensor's characteristics, these traits make the moving-coil transducer less versatile than a small, orientation-tolerant, solid-state accelerometer.
Despite the accelerometer's advantages, moving-coil velocity transducers remain installed on a great deal of older turbomachinery, and they persist for practical reasons rather than technical superiority. Large turbines and compressors were fitted with these sensors when they were built, the sensors are wired into monitoring systems that expect a velocity signal, and the machines run for decades, so replacing a working sensor and its whole signal chain is disruptive and often unjustified while the existing setup does its job. An installed base that still works tends to stay installed.
There is also a genuine fit with how those machines are judged. Casing vibration on turbomachinery is commonly assessed against velocity-based severity criteria, and a sensor that outputs velocity natively feeds that assessment directly. For the overall-level casing monitoring these machines rely on, in the frequency range where the transducer performs well, it does exactly what is asked of it, and the native units remove a conversion step. So on legacy machines the sensor is not just tolerated but reasonably matched to the task.
When these sensors feed a modern monitoring picture, their velocity output becomes a measured value on the platform. A monitoring system reads the transducer's velocity signal, derives an overall level or spectrum from it, and reports it as a tag that a cloud platform such as Merobix can trend and alarm on alongside the rest of the machine's instrumentation. Whether the underlying sensor is a legacy moving-coil transducer or a modern accelerometer, the platform sees a vibration measurement it can watch remotely. That lets an operator keep an eye on decades-old turbomachinery, with its original velocity pickups, from the same screen as everything else, without ripping out sensors that are still doing their job.
A moving-coil velocity transducer measures the velocity of vibration and outputs it natively in velocity units, typically millimeters per second. It does this by inducing a voltage in a coil that moves relative to a magnet, and that induced voltage is directly proportional to the velocity of the motion. The native velocity output is convenient because the common vibration severity guidelines are written in velocity, so the signal can be compared against those charts without the integration step an accelerometer requires.
The piezoelectric accelerometer is smaller, more rugged, extends to much lower and higher frequencies, has no moving parts to wear, and is far less sensitive to mounting orientation. The moving-coil velocity transducer contains a suspended mass on a spring, so it rolls off below its internal natural frequency, is affected by which way it points, and its moving parts age. Accelerometers overcome those limits, so they have become the default sensor. Velocity transducers persist mainly on legacy turbomachinery already wired for them.
They were fitted when the machines were built, they are wired into monitoring systems that expect a velocity signal, and the machines run for decades, so replacing a working sensor and its whole signal chain is disruptive and rarely justified while it still works. They also fit the task well, because casing vibration on turbomachinery is often judged against velocity-based severity criteria, and a sensor with native velocity output feeds that assessment directly in the frequency range where it performs.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.