What an unbroken calibration chain looks like
Traceability is not a stamp on an instrument — it is a chain of comparisons. Each link is a calibration in which a device is compared with a more accurate one, and each link adds uncertainty. The chain ends at the SI definition of the unit.
| Level | Example (evidence) |
|---|---|
| SI unit | Volt, realised via the Josephson effect / Definition of the unit |
| National metrology institute (PTB, NIST, NPL…) | Primary voltage standard / Calibration certificate with uncertainty, CMC entry in the BIPM database |
| Accredited calibration laboratory | Reference multimeter calibrated against the national standard / Accredited certificate (ISO/IEC 17025) with scope covering range and uncertainty |
| Your laboratory | Working multimeter or data logger calibrated against the reference / Calibration certificate or internal calibration record |
| Your test result | Voltage measured on the test item / Test record referencing the instrument and its calibration status at the time |
If one link is missing — an expired certificate, a certificate that does not cover the measured range, a reference without uncertainty — the chain is broken, and so is the claim of traceability for every result that passed through that instrument in the meantime.
The routes clause 6.5 allows
Clause 6.5.2 names the ways in which a laboratory can ensure traceability to the SI. Which route fits depends on the quantity.
| Route | How it works |
|---|---|
| Competent calibration laboratory | Instruments are calibrated by a laboratory that demonstrates competence, impartiality and consistent traceability — in practice, accredited to ISO/IEC 17025 for the quantity, range and uncertainty you need. Typical use: Electrical, temperature, force, torque, pressure, dimensional, mass |
| Certified reference materials | A reference material from a competent producer (ISO 17034) with a certificate stating traceability to the SI is used to calibrate or check the method. Typical use: Chemical and material analysis, pH, conductivity, hardness blocks |
| Direct realisation of SI units | The laboratory realises the unit itself, realisation confirmed by direct or indirect comparison with national or international standards. Typical use: Rare — mostly national metrology institutes and a few highly specialised laboratories |
| Other references (6.5.3) | Where traceability to the SI is technically not possible, traceability to a reference such as certified reference materials, agreed methods or consensus standards, shown by interlaboratory comparison. Typical use: Qualitative or method-defined properties, e.g. some hardness scales or material grades |
What a calibration certificate must show for your chain
Receiving a certificate is not the end of the job. Review it before releasing the instrument for use:
- The calibrating laboratory's accreditation (body, number) and a scope that covers the quantity, range and uncertainty you need.
- Unique identification of the instrument — serial number matching the one in your equipment register.
- Measurement results with expanded uncertainty and coverage factor, not just "within tolerance".
- A statement of traceability to national or international standards, ideally naming the references used.
- Calibration date and conditions; the due date is set by you, not by the certificate (see the calibration intervals guide).
- As-found and as-left data where the instrument was adjusted, so that earlier results can be assessed.
Then compare the calibrated uncertainty with what your method needs. An instrument whose calibration uncertainty consumes most of your tolerance is traceable but not fit for the task — that belongs in the uncertainty budget.
Keeping traceability intact in daily work
- 1
Register every instrument that influences a result
Unique ID, serial number, range, location, owner. Anything that contributes to a reported value belongs in the register — including data loggers, thermocouples and fixtures with measuring function.
- 2
Define the metrological requirement per method
Which uncertainty does each measurement need? That determines the accuracy class of the instrument and the level of the calibration it requires.
- 3
Choose the calibration source
Select an accredited calibration laboratory whose scope matches, or justify an alternative route. Keep scope documents on file.
- 4
Review each certificate on receipt
Check the points above and record the review. Release the instrument only after the check.
- 5
Set intervals and block overdue equipment
Derive intervals from drift history, record the due date and make sure an overdue instrument cannot be used unnoticed.
- 6
Link instruments to the results they produced
Every test record should state which instruments were used and their calibration status at that time. This is what makes a chain auditable years later.
- 7
React to out-of-tolerance findings
If a calibration shows an instrument was out of tolerance, review all results since the last good calibration and decide on corrective action (clause 6.4.9 and 7.10).
Typical assessment findings
- Calibration certificate from a laboratory that is not accredited or whose scope does not cover the range used.
- Certificate has no uncertainty or no statement of traceability.
- An instrument is in use after its calibration due date.
- Test records do not show which instrument was used, so affected results cannot be identified after an out-of-tolerance finding.
- Internal reference standards are used for calibration without their own traceable calibration.
- The calibration uncertainty is not considered in the measurement uncertainty budget.
- Calibration results were never reviewed against the method requirements.
Nearly all of these are record-linking problems: instrument, calibration, test job and result are held in different files. When they live in one system, "which results used this instrument?" is a query rather than an investigation.