Accelerometer calibration is the only way to confirm that a vibration sensor’s sensitivity is still what its datasheet says it is. Sensitivity drifts with age, temperature cycling, overload and rough handling, and a drifted accelerometer keeps producing plausible-looking data the whole time. For engineers running vibration tests, that is the trap. Nothing looks wrong until a result is challenged and there is no recent verification to stand behind it.
Why accelerometer sensitivity drifts and what it does to your data
An accelerometer is primarily characterised by its sensitivity, which is the ratio of electrical output to mechanical input. Every amplitude in your data is scaled by that number, so the entire measurement chain rests on it staying true.
It does not stay true on its own. Piezoelectric sensing elements age. Repeated temperature cycling shifts the behaviour of the element and its preload. A drop onto a concrete floor, or an overload well beyond the measuring range, can damage the sensing structure without leaving a visible mark. Cables, connectors and mounting threads wear with every installation. None of these failure modes switches the sensor off. The accelerometer keeps producing a signal, just not the signal it produced when it left the factory.
Vibration data rarely sits in a report unread. It feeds resonance identification, fatigue calculations, pass and fail decisions against a test specification, machine acceptance and product release. A sensitivity error scales every one of those results by the same factor, silently and in the same direction. That is what makes it dangerous, because the data stays internally consistent while being wrong.
The consequences surface late. A prototype passes a qualification test it should have failed, or fails one it should have passed. A structural test programme runs for weeks on a sensor that was overloaded in the first session. And when a result is questioned, by a client, a certifier or your own quality system, the first thing asked for is the calibration record. In Australia, that question carries extra weight, because the sensor often cannot be checked quickly. The nearest laboratory may be interstate and the manufacturer is an ocean away.
How comparison calibration works, in the laboratory and in the field
The working method for verifying an accelerometer is comparison calibration, sometimes called the back-to-back method. The sensor under test is mounted on a vibration exciter and driven at a controlled level across a range of frequencies. Its output is compared with the output of a reference transducer whose own calibration is traceable to a national metrology institute. The result is the measured sensitivity of the test sensor and its deviation across the frequency range, which either confirms the sensor is within tolerance or tells you plainly that it is not.
Nothing about that method requires a laboratory building. It requires a controlled excitation source, a traceable reference and a way to record the comparison, and all three now fit in a case.
Periodic laboratory calibration remains the backbone of a traceable measurement system, because it produces the certificate your quality records rely on. The weakness is everything that happens between certificates. A sensor overloaded in March is not scheduled to be seen until its calibration falls due, and every test it touches in the meantime inherits the doubt.
Field verification shortens that window. Checking a suspect accelerometer on site, the same day, tells you whether to keep testing or swap the sensor before more data is compromised. It also changes the economics of a test campaign. Instead of freighting sensors interstate or overseas and waiting, you verify in-house and reserve the laboratory round trip for the scheduled certification.
Whether an in-house check satisfies your quality system depends on that system’s requirements, and your calibration records should state which sensors were verified, against which reference and when. Many test programmes settle on a pairing of certified laboratory calibration at the scheduled interval with portable verification between intervals and after any suspected overload.
Selecting a portable vibration calibrator
Four questions do most of the selection work. The frequency range has to cover what you actually test, from low-frequency structural work up to high-frequency bearing and gear phenomena. The signal conditioning has to match your sensor fleet. A fleet of IEPE accelerometers has different needs from one that mixes charge-output sensors, DC voltage outputs and 4 mA to 20 mA transmitters. The internal reference must be traceable, because an unverifiable reference makes the whole exercise pointless. And if the calibrator is going to site, weight, battery life and data transfer stop being footnotes and start deciding whether the unit actually gets used.
Read the measurement uncertainty together with the conditions it applies under as well, because a figure quoted at laboratory temperature with a small payload will not be what you achieve at full payload on a hot site, and a reputable datasheet states both.
How often should accelerometers be calibrated? There is no universal interval. The manufacturer’s recommendation and your own quality system set the baseline, and the duty the sensor sees moves it. Rough handling, high temperatures or frequent transport justify shorter intervals, and any suspected overload justifies a check regardless of the schedule.
Can accelerometers be verified on site? Yes. A portable calibrator with a built-in exciter and a traceable reference performs a comparison calibration wherever the sensor is, including on a test rig between runs.
What sensor types can a portable calibrator check? That depends on its signal conditioning. Systems in this class verify accelerometers, proximity probes and vibration velocity sensors across the common output types (voltage, IEPE, charge and 4 mA to 20 mA), and can also check vibration meters and small vibration test beds on site.
Is a field check the same as a certified calibration? Not necessarily. The method is the same, but whether the result counts as a calibration in your records depends on your quality system’s traceability requirements. Treat field verification as the early-warning layer and keep the certified calibration on its schedule.
The CV-10 mobile calibration system
For engineers who need this capability in the field, the CV-10 mobile calibration system, available from Bestech Australia, is the configuration described above packed into one 9 kg case. It calibrates accelerometers, proximity probes and vibration velocity sensors on site, and also serves as a check for vibration meters and as a vibration test system for small devices. The reference transducer sits inside the vibration exciter, with traceability to NIST and PTB, and the unit ships with a PTB traceable calibration certificate (DAkkS).
The exciter drives from 5 Hz to 10 kHz at up to 200 m/s² sine peak and carries sensor payloads up to 900 g. Measurement uncertainty for accelerometer calibration is 1.5% up to 5 kHz and 3.5% from 5 kHz to 10 kHz under laboratory conditions, determined according to GUM with a coverage factor of k=2, and transverse motion is assessed according to ISO 16063-21. The integrated signal conditioner accepts voltage, charge, IEPE and 4 mA to 20 mA signals, with an amplifier module for PR transducers available as an option. Manual operation with presets is standard, while automatic stepped sine, automatic sweep and vibration measurement are options. The battery runs for up to 10 hours under typical conditions, calibration results export as CSV or XML over USB, and Ethernet support connects the unit to its companion PC software.
Visitors to our stand at Internoise 2026 in Adelaide saw the demonstration in person, with an accelerometer mounted, driven and verified against the reference transducer in minutes. Every vibration engineer should be able to say when a sensor was last verified, and what that means for the data it has produced since. That was the point of the demonstration.
Getting the measurement right starts before the first data point is recorded, with a sensor you can trust. If your accelerometer fleet is due for a health check, or you are weighing up field verification against laboratory round trips, have a chat with us about your application.
