Calibration: keeping gauges tied to a national standard and handling the day one fails

Traceability is a chain, and every link has to be documented

A calibration certificate is only meaningful if the reference used to perform it was itself calibrated against something higher, continuing up to national measurement institutes and the international system of units coordinated through the International Bureau of Weights and Measures. Certificates that state a result without identifying the reference and its own status break the chain and prove very little. When reviewing a supplier's certificates, that is the thing to look for. It is also what distinguishes a genuine calibration from an adjustment performed by somebody with a better instrument and good intentions.

Setting an interval is a risk decision, not a tradition

Intervals are usually inherited rather than reasoned. A more defensible approach uses the history: instruments that consistently return within tolerance can move to a longer cycle, those that have drifted stay short, and anything dropped, used heavily, or exposed to coolant and vibration gets treated as a special case. Usage matters more than the calendar for hand tools that live in a toolbox. Document why each interval was chosen, because the alternative is a schedule nobody can justify that simultaneously over-calibrates the stable equipment and under-calibrates the instruments doing the difficult work.

The out-of-tolerance recall is the part everyone dreads

An instrument found outside its limits calls into question every acceptance decision made with it since it was last known good. That review is the real reason calibration exists, and it needs a written route: identify the affected characteristics, work out which product was judged with that gauge, assess whether the observed error was large enough to change a verdict, and decide whether anything already dispatched needs the customer told. Doing this well depends entirely on knowing which gauge was used where, which is a record most plants only discover they lack at the moment they need it.

In-house, accredited laboratory, or manufacturer

Simple equipment can be checked internally against reference artefacts, provided the references are themselves calibrated externally and the person doing it is competent and independent of the pressure to pass things. Accredited laboratories operate under arrangements assessed against the international standard for testing and calibration laboratories, ISO 17025, and their certificates carry weight in a dispute for that reason. Instrument makers handle their own specialised equipment. The choice is a cost and credibility trade-off: internal work is cheaper and faster, external work is what a customer will accept without argument.

Status has to be visible where the gauge is used

Every plant that has been audited has been asked to prove an instrument in use was in calibration on the day it was used. The system that answers this reliably is dull and physical: a unique identifier on each gauge, a label showing status, a register that anyone can query, and a mechanism that removes overdue items from service rather than merely flagging them. Personal instruments brought in by employees are the classic gap, along with items that live at a machine and never appear on the register because nobody ever booked them in.

Frequently asked questions

How do we decide how often an instrument needs calibrating?

Base it on evidence rather than habit. Track what each instrument's results have looked like across previous calibrations, and lengthen the cycle for the ones that never move while shortening it for those that drift or get rough treatment. Factor in how the tool is used: an instrument in a toolbox on a wet machine deserves more attention than one that stays in a controlled room. Record the reasoning, so the schedule can be defended and adjusted rather than simply repeated.

What must happen when a gauge comes back out of tolerance?

Treat it as a nonconformance affecting product, not merely an equipment issue. Establish what was judged with that instrument since it was last confirmed good, assess whether the measured error was sufficient to change any acceptance decision, and decide on containment for anything affected, including material already delivered. Record the assessment and its conclusion even where nothing needs recalling, since the reasoning is exactly what an auditor or a customer will ask to see.

Do calibrations have to be done by an accredited laboratory?

Not universally, and internal calibration against externally calibrated references is common and defensible for straightforward equipment. Accreditation matters where a customer contract calls for it, where the measurement is technically demanding, and wherever a result may end up in a commercial dispute, because an accredited certificate removes an argument about competence. Accreditation bodies operate under international mutual recognition arrangements, which is what makes such certificates portable across borders.