Calibration Certificate
& Traceability Record FAQ

Traceability to a national standard means the calibration was performed using a reference standard whose accuracy is linked, through an unbroken chain of comparisons, back to a recognised national or international metrology institute — so the measurement can be trusted as accurate relative to a globally consistent reference, not just internally consistent within one lab.

In practice, this means the calibration certificate should identify the reference standard used and its own calibration status, which in turn traces back through the calibration hierarchy to a body such as a national metrology institute. Without this chain, a calibration certificate confirms internal consistency but not genuine accuracy against a recognised standard.

ISO/IEC 17025 accreditation is not an explicit regulatory mandate, but it's the industry-standard way to demonstrate that a calibration lab's traceability, uncertainty calculations, and quality processes meet a recognised benchmark — using a non-accredited lab is possible, but the organisation then carries the burden of independently justifying the lab's competence and traceability chain.

An inspection-ready calibration certificate should let a reviewer confirm exactly which instrument was calibrated, against what reference standard, with what results, and whether it passed within defined tolerance — vague certificates that only state "calibration passed" without supporting data are a common and easily avoidable gap.

• Instrument identification: make, model, and serial number.
• Calibration date and, where applicable, the due date for the next calibration.
• Reference standard(s) used and their own traceability and calibration status.
• As-found and as-left measurement results at each tested point.
• Stated tolerance or acceptance criteria and measurement uncertainty.
• Name of the calibrating technician or laboratory and, if applicable, accreditation number.

A calibration certificate must reference the exact instrument by serial number, not just the model — a model number alone doesn't confirm which specific physical unit was actually calibrated, especially where multiple identical instruments exist across a facility.

Calibration compares an instrument's readings against a traceable reference standard across its measurement range and typically results in a documented adjustment or correction factor if needed; verification is a lighter-weight check, often just confirming the instrument still reads correctly at one or a few points, without necessarily adjusting it or establishing full traceability.

GxP use of an instrument generally requires full calibration at defined intervals, with verification sometimes used as an interim check between full calibrations for higher-use or higher-risk instruments — verification alone is not usually accepted as a substitute for periodic calibration.

No — an instrument with an expired calibration should not be used to generate or verify GxP data until it is recalibrated; using it anyway means any resulting data carries unverified measurement accuracy and is a common and serious inspection finding.

The practical control is a system that physically or procedurally prevents use of instruments past their calibration due date — such as a visible due-date label combined with an equipment status check before use — rather than relying on the user to remember to check.

This should be treated as a deviation requiring an impact assessment of every GxP record generated by that instrument since its last known-good calibration, to determine whether the data is still reliable or needs to be invalidated, reprocessed, or independently verified by another means.

The scope of the look-back depends on how far out of date the calibration was and how the instrument has historically performed — an instrument with a strong calibration history that was only slightly overdue presents a very different risk than one that was months overdue with a history of drift.

Yes — any instrument used to generate or verify data supporting a qualification decision, including FAT, SAT, IQ, OQ, or PQ, needs the same calibration rigor and traceability as a production instrument, because the qualification conclusion depends just as much on that measurement being accurate.

A FAT-stage calibration certificate for the equipment's own built-in instruments can often be carried forward to SAT if the certificate is still within its valid period and the instrument hasn't been affected by transport or reinstallation — but any external test equipment brought in specifically for SAT testing needs its own valid, current calibration independent of what was used at FAT.

An instrument's built-in self-calibration or auto-cal function can be part of an acceptable calibration program, but it should itself be validated and periodically verified against an independent, traceable external standard — relying entirely on an instrument's internal self-check without any independent external verification is generally not considered sufficient on its own.

An out-of-tolerance finding should trigger a deviation and a risk-based impact assessment of every GxP record the instrument was used to generate since its last successful calibration, to determine whether the degree of drift found actually affected the validity of that data or the decisions made using it.

The look-back should extend to the instrument's last known in-tolerance calibration, covering every GxP activity performed with the instrument in that window, with the depth of investigation scaled to the magnitude of the out-of-tolerance drift and the criticality of the measurements affected.

Yes — the reference standards used to calibrate a GxP instrument must themselves have documented, current traceability back to a recognised national or international standard; an instrument calibrated against an uncalibrated or undocumented reference standard has no genuine traceability, regardless of how thorough the calibration process itself looked.

No universal regulatory interval exists — calibration frequency should be determined by a risk-based assessment considering the instrument's criticality, historical drift performance, manufacturer recommendations, and usage frequency, and defined in the site's calibration program rather than fixed by regulation.

The calibration certificate should state the acceptance tolerance the instrument was assessed against, the actual as-found and as-left readings at each test point, and the measurement uncertainty of the calibration itself — without uncertainty information, it's not possible to properly assess whether a borderline result is genuinely within tolerance.

The validation protocol or batch record should reference the specific instrument by serial number and its calibration status or certificate number at the time of use, so a reviewer can trace forward from the protocol to the exact calibration record that was in effect, rather than having to infer which calibration cycle applied.

Shared instruments need a single, centrally tracked calibration record accessible to every system or site that uses them, with a clear process for checking calibration status before each use and flagging the instrument as unavailable if it's due or overdue for recalibration, regardless of which site currently has physical custody of it.

Calibration certificates and traceability records should be retained for at least as long as the GxP records generated using that instrument during the calibration cycle they support, since they're the evidence that those records were produced using an instrument known to be accurate at the time.

Inspectors typically pick an instrument used in a critical process step, confirm its calibration was current and in tolerance at the time it was used, trace the reference standard back through the calibration chain, and check whether any historical out-of-tolerance findings were properly investigated for downstream data impact.

A frequent finding is an instrument with a valid-looking calibration sticker but no accessible underlying certificate, or a certificate that doesn't clearly show as-found results — both make it impossible to confirm the instrument was actually within tolerance before use, not just currently in-date.

The most common mistake is treating calibration as a scheduling problem — tracking due dates — without equally tracking traceability and out-of-tolerance impact: an instrument can be perfectly on schedule and still generate a serious compliance gap if its reference standard's traceability can't be demonstrated or if a past out-of-tolerance result was never properly assessed for its effect on the data it produced.