Calibration intervals: setting them, shortening them, defending them

Updated: 2026-08-13 · ISKRIT

An instrument register is not a list of dates. It is the evidence that every measurement you have reported was made with equipment known to be in tolerance at the time, and the interval is the assumption holding that claim together.

Where the interval comes from

The manufacturer suggests one. That suggestion assumes typical use in a typical environment, and neither may describe your shop. Three things justify changing it: the history of the instrument, the consequence of an undetected drift, and how hard the instrument is used.

The review itself is the deliverable. An interval extended because it was inconvenient reads identically in the register to one extended on five years of clean history, and only the written justification separates them.

What an out-of-tolerance result obliges

This is the part most registers handle badly. When an instrument comes back out of tolerance, the calibration certificate is not the end of it: every measurement made with that instrument since its previous calibration is now suspect. The obligation is to assess the impact on those results, decide whether anything reported has to be corrected or recalled, and record the decision.

A register that stores only the last certificate cannot answer the question, because the affected work is defined by the interval between two calibrations and by what the instrument was used for during it.

The fields an auditor asks for

Identification and serial number, location and custodian, the measurement range and the tolerance the instrument is held to, the date of the last calibration and the date the next is due, who performed it, the traceability of the standard used, and the certificate itself. Missing any one of them turns a finding into an argument.

Two are forgotten most often: the tolerance the instrument is held to, as distinct from its specification, and the traceability chain of the reference standard. Both are the difference between a register and a spreadsheet of dates.

Overdue is a status, not a delay

An instrument past its due date is not "nearly due". Until it is calibrated it is not fit for use, and results produced with it in the meantime carry the same impact assessment as an out-of-tolerance finding. Registers that warn only on the day of expiry guarantee this happens; a warning window measured in weeks is what prevents it.

The instrument register tool keeps the tolerance separately from the range, warns ahead of the due date rather than on it, and holds the calibration history so the impact of an out-of-tolerance result can be scoped to the right period.