Sound level meter calibration: field check, periodic test and class

A sound level meter that reads 94.0 dB when the calibrator is fitted and still produces a survey nobody can defend is a common outcome, and the reason is usually not the meter. The field check verifies one level at one frequency through one path; it says nothing about the frequency weighting, the time weighting, the directional response of the microphone, or whether the windscreen that should have been fitted outdoors was in a drawer.

Three activities, one word

IEC 61672-1 contains the Class 1 and Class 2 specifications for sound level meters; IEC 61672-3:2013 describes the procedures for periodic testing of instruments designed to conform to the Class 1 or Class 2 specifications of the second edition of Part 1, and states that the purpose is to assure the user that performance conforms for a limited set of key tests under the environmental conditions in which those tests were performed. The phrase "limited set" is doing real work: a periodic test is not a full pattern evaluation and does not claim to be.

A periodic test certificate is evidence about the instrument on the day it was tested, over the subset of tests performed, at the conditions recorded. It is not a statement that the instrument meets every requirement of Part 1 — only pattern evaluation of the model addresses that.

The field check, done properly

An acoustic calibrator produces a known sound pressure level in a cavity that couples to the microphone, conventionally at 1 kHz and at a reference level of 94 dB, 114 dB, or both. Requirements for these devices are in IEC 60942, and the pairing matters: a Class 1 sound level meter requires a calibrator conforming to the Class 1 specifications of IEC 60942, while a Class 2 meter may use a Class 1 or a Class 2 calibrator.

• Check before and after the session. A before-only check cannot tell you whether the data you just collected is valid; an after-check that disagrees invalidates the session, which is precisely the information you need.
• Remove the windscreen and fit the calibrator to the bare microphone, as the calibrator's coupling was designed for.
• Apply the static pressure correction. A calibrator's output level depends on barometric pressure, and the correction — and whether the instrument applies it automatically — is in the calibrator's own documentation. At altitude or in unusual weather it is not negligible.
• Let both devices reach ambient temperature. A calibrator carried in a warm car and clamped onto a cold microphone is not yet producing its rated level.
• Record the as-found reading before adjusting anything, not just the as-left value.
• Remember the calibrator needs its own calibration. It is the reference in this chain, and an uncalibrated reference makes the daily check decorative.

Recording the as-found figure is the habit that turns a daily ritual into useful data. A sequence of as-found values drifting in one direction over months is a microphone ageing or a seal failing, visible long before any single check falls outside tolerance — the same argument as for any other instrument's as-found history.

The microphone is the instrument

Most of what goes wrong in acoustic measurement happens in the first centimetre. Measurement microphones are built with a defined response in a defined sound field, and using one in the wrong field or the wrong configuration changes the result in a way no calibration detects.

The failure mode practitioners hit most often is the windscreen. It is removed for the morning calibrator check, correctly, and then not refitted before an outdoor survey. The resulting data contains wind-induced pressure fluctuations at the diaphragm that are not sound at all but are recorded as low-frequency level, and the measurement will often pass every plausibility test anyone applies to it.

Class 1 and Class 2

The two classes in IEC 61672-1 are distinct specifications, with the Class 2 tolerances wider and the frequency range over which conformity is required narrower. They are not grades of the same instrument: a Class 2 meter is not a Class 1 meter with a looser certificate, and where a regulation, a contract or a method specifies Class 1, a Class 2 instrument does not satisfy it even if it happens to agree on the day.

The dependency runs through the whole chain. The calibrator has to match the class, the periodic test has to be against the class the instrument was designed to, and the uncertainty you assign to a measurement reflects the class. Changing to a different microphone or preamplifier can take a system outside the configuration its class statement was based on.

Frequency and time weighting

Two sets of weightings have to be right and are easy to confuse. Frequency weighting — A, C or Z — shapes the response across the spectrum, and A-weighting is the usual choice for occupational noise because it approximates the ear's sensitivity, while C-weighting is used for peak measurements. Time weighting — fast or slow — governs how quickly the displayed level follows a changing sound, and is separate from the integrating-averaging function that produces an equivalent continuous level over a period.

Reporting a figure without saying which weightings produced it is the acoustic equivalent of a pressure reading without saying whether it was gauge or absolute. The periodic test covers the weighting networks, which is one of the things the daily calibrator check at a single frequency cannot do.

Uncertainty in a noise survey

ISO 9612 specifies an engineering method for determining occupational noise exposure, offering different measurement strategies, and it includes methods for estimating the uncertainty of the result. The structure of that uncertainty is worth internalising, because the instrument is usually the smallest part of it.

• Sampling of the working day: how representative the measured periods were of the whole shift, usually the dominant contribution.
• Variability of the noise itself from day to day and task to task.
• Microphone position relative to the worker's ear, and the operator's influence on the field.
• The instrument's own class-dependent uncertainty.
• The calibrator's uncertainty, from its certificate.
• Residual error from an uncorrected static pressure or temperature deviation.

A laboratory that reports an exposure figure to a tenth of a decibel on the strength of a Class 1 certificate, having sampled twenty minutes of an eight-hour shift, has put its effort in the wrong place. The decision being made from the number — whether hearing protection is required, whether a limit is exceeded — rests on the sampling at least as much as on the instrument, and the acceptance criteria themselves come from the applicable national regulation rather than from the instrument standards.

ValiTrac AI is not an accredited calibration laboratory and confers no compliance. Exposure limits and the required testing intervals for sound level meters are set by the regulation applying where the work is done, and periodic testing has to be performed by a laboratory with the relevant capability in its own accredited scope.