Calibration Drift and a Class 2 Meter Under a Class 1 Permit
During a six-week survey of a metal recycler, a technician uses a handheld meter marked 'IEC 61672-1 Class 2' with an integrating function. The permit requires measurements 'with a Type 1 (Class 1) integrating sound level meter conforming to IEC 61672-1, field-checked with a calibrator before and after each measurement period.' The acoustical calibrator reads 94.0 dB at 1 kHz before the day's measurements and 94.7 dB after. Mid-day the batteries are replaced and the instrument clock resets to 2019-01-01; subsequent files carry that date. Working alone, the technician holds the meter at chest height roughly 0.5 m from his body and 2 m above the pavement because the tripod was left in the van. A light drizzle begins and the foam windscreen is visibly wet for the final three samples. The report presents results to 0.1 dB with no mention of the meter class, the calibration drift, the clock reset, or the rain.
1. Assess the validity of this data set. Address the calibration check, the meter class, the clock reset, and the operator-held microphone.
Read the worked answer
Calibration drift. A change of +0.7 dB across the measurement period exceeds the customary ±0.5 dB acceptance band. A one-directional drift indicates the instrument's sensitivity changed - temperature, humidity, a failing microphone or a low battery - and the timing of the change is unknown. You cannot repair it by subtracting 0.7 dB: the shift may have occurred in the first ten minutes or the last, and it may not have been monotonic. Where a result sits within about 1 dB of a limit, drift of this size can flip the compliance conclusion.
Meter class. The permit names a Class 1 instrument. A Class 2 unit does not satisfy the condition regardless of how well this particular meter performed; the data are nonconforming on the face of the permit, and the Class 2 tolerance envelope is wider than Class 1 across the frequency range.
Clock reset. Files dated 2019 cannot be tied to the complaint time or to a specific measurement without independent notes, which destroys both correlation and chain of custody.
Microphone handling. Holding the meter 0.5 m from the body allows body reflection and hand vibration, commonly adding 1–3 dB, worst at low frequency; 2 m height is nonstandard and changes ground effect. A wet foam windscreen unpredictably attenuates high frequencies and is not correctable.
2. What should the technician and the program do now, before any further measurements?
Read the worked answer
Stop the survey and quarantine the data. Actions:
- Verify the calibrator itself - a field check only compares meter to calibrator; the calibrator needs its own annual laboratory calibration (IEC 60942) for traceability to hold.
- Have the Class 2 meter serviced and checked against a Class 1 reference, and obtain a Class 1 instrument for any compliance survey.
- Re-establish the calibration history and file before/after records with the case.
- Set and verify the clock against a reference at the start of every measurement day, and record the check in the field notes.
- Tripod-mount the meter, microphone 1.2–1.5 m above ground and ≥3.5 m from reflecting surfaces, operator at least 0.5 m behind using a remote cable.
- Abort in rain; dry or replace a wetted windscreen rather than trying to correct for it.
- Photograph the setup and log the measurement point on a map so position is documented independently of the instrument clock.
3. How should the results be presented in the report, given the defects?
Read the worked answer
Report them as 'data not valid for compliance determination'. State the nonconformities explicitly - meter class, calibration drift, rain on the windscreen - and state that no correction was applied and why. Describe the replacement survey and its schedule.
Do not report to 0.1 dB with an implied accuracy the instrument and method cannot deliver; report to the nearest whole decibel or with a stated expanded uncertainty. Retain all raw files, the clock-reset evidence and the calibration records. A report that discloses its own defects is far stronger in a hearing than one that is impeached later, and withholding the defects would be a professional-conduct problem, not merely a technical one.