Metrology & Calibration Management
Keeping measurement traceable to recognised standards: calibration intervals, tolerances, as-found readings, and the assessment of what a failed calibration means for everything measured since. A calibration failure is retrospective by nature — the instrument was out of tolerance for an unknown period, and every result it produced in that window is in question. That is why the as-found condition matters more than the as-left one.
What an explainer is not
A topic explainer is SPEQ’s synthesis of what a practice involves, cited to the standards that govern it. It does not reproduce their text, and it does not determine which of them apply to your product or process.
[ POSITION IN THE FRAMEWORK ]
7 DIMENSIONS · 23 LINKSCalibration failure is silent: an instrument out of tolerance does not stop producing results, it produces plausible ones, and everything released against them is in question until an assessment says otherwise.
06 · QUALITY MATURITY — METROLOGY & CALIBRATION MANAGEMENT, REACTIVE TO ADAPTIVE
Instruments are calibrated on a schedule by an external provider. Traceability is assumed from the certificate.
A programme exists with intervals and traceable standards, and out-of-tolerance results are handled by recalibrating and continuing.
Criticality drives interval and tolerance, out-of-tolerance triggers a documented backward impact assessment, and the calibration range covers the working range.
Drift is trended per instrument so intervals reflect behaviour, and measurement uncertainty is known where results sit near a specification limit.
The measurement system is understood as a whole — instrument, method, standard, analyst — so a marginal result can be interpreted rather than merely repeated.
SPEQ’s shared five-stage progression, labelled synthesis — not the FDA QMM rating scale. Where does your organization sit? Score your quality system →
07 · REGULATORY & EVIDENCE
GOVERNING STANDARDS · 5
Derived from the 5 standards SPEQ maps to this subject, across 4 regulatory bodies: FDA, EMA, USP, ISO.
RECORDS & OBJECTIVE EVIDENCE
- The calibration programme with criticality-based intervals and tolerances
- Traceability of reference standards to national or international references
- Out-of-tolerance investigations with backward impact assessment
- Drift trending and any interval adjusted because of it
- Measurement uncertainty determination where results are compared against limits
COMMON INSPECTION FINDINGS
- Out-of-tolerance findings closed by recalibration with no assessment of prior results
- Calibration performed over a range that excludes the working point
- Reference standards without traceable certification, or expired
- Intervals fixed by convention despite repeated drift at the same instrument
- Results reported near a limit with no consideration of measurement uncertainty
As-found is the data; as-left is the outcome
A calibration record showing only that the instrument was adjusted and now reads correctly has discarded the informative half. The as-found reading is what tells you whether the instrument had drifted, by how much, and in which direction — and therefore whether the interval is appropriate and whether previously reported results are affected.
Recording as-found systematically also converts calibration from a compliance activity into a data set. Drift trends per instrument reveal which ones are approaching the end of their useful life, which interval assumptions were wrong, and which installations are harder on instruments than expected. Very few sites analyse this, and the data is already being generated.
Intervals should be justified and adjusted
Calibration intervals are frequently set at annual by convention, applied uniformly, and never revisited. A risk-based interval considers the instrument’s criticality, its observed drift history, its operating environment and the consequence of an undetected excursion — which typically produces a spread rather than a single interval, with critical instruments checked more often and stable non-critical ones less.
Adjusting intervals in both directions is the part that requires evidence and nerve. Extending an interval on a stable instrument is defensible with drift data and indefensible without it, and it is what makes room in a programme to shorten the intervals that need shortening.
The out-of-tolerance assessment is a product-impact assessment
When an instrument is found out of tolerance, the affected period runs from the last successful calibration to the present, and every batch, test or release decision that relied on it is potentially in question. The assessment has to establish the magnitude and direction of the error, which measurements it affected, and whether any of them would have changed a decision.
Doing this quickly requires knowing in advance what the instrument feeds — which batches, which release tests, which environmental records. Sites that maintain that linkage answer in hours; sites that do not begin by reconstructing it while product waits. The linkage is a property of the asset register, which is another reason it is worth maintaining properly.
SPEQ interpretation — traceability is a chain, and chains have a weakest link
Traceability to a national or international standard is usually treated as a property of the calibration certificate. It is a chain: the working standard is calibrated against a reference standard, calibrated against a higher one, terminating at a national institute — and the measurement uncertainty accumulates along it.
Where the accumulated uncertainty approaches the tolerance being verified, the calibration is not establishing what it appears to establish. The practical rule most sites apply informally is a ratio between the tolerance and the uncertainty of the reference used, and it is worth stating explicitly, because the case where it fails — a tight tolerance verified with a barely-better reference — looks identical on the certificate to a case where it holds.
FREQUENTLY ASKED
Why does as-found matter more than as-left?
Because as-found tells you whether the instrument drifted, by how much and in which direction — which determines whether the interval is right and whether previously reported results are affected. A record showing only that the instrument was adjusted and now reads correctly has discarded the informative half.
How should calibration intervals be set?
From criticality, observed drift history, operating environment and the consequence of an undetected excursion — which produces a spread rather than a uniform annual default. Adjusting in both directions is the part requiring evidence: extending on a stable instrument is defensible with drift data and indefensible without it, and it makes room to shorten the ones that need it.
What does an out-of-tolerance finding require?
A product-impact assessment covering the period since the last successful calibration: the magnitude and direction of the error, which measurements it affected, and whether any would have changed a decision. Answering quickly depends on knowing in advance what the instrument feeds, which is a property of the asset register.
When is a traceable calibration not sufficient?
When the measurement uncertainty accumulated along the traceability chain approaches the tolerance being verified. A tight tolerance verified with a barely-better reference looks identical on the certificate to a case where the ratio is comfortable, which is why the acceptable ratio is worth stating explicitly rather than applying informally.