ISO/IEC 17025 clause 7.6.2: evaluate uncertainty for every calibration
Clause 7.6.2 requires a laboratory performing calibrations to evaluate measurement uncertainty for all calibrations, including those of its own equipment. The evaluation must reflect the calibration actually performed. A reference standard's certificate is one source of information, not automatically the uncertainty of the complete result.
What clause 7.6.2 requires in practice
Clause 7.6.2 is short and firm. A laboratory that performs calibrations evaluates measurement uncertainty for all of them, and that includes calibrations of its own equipment. It builds on the general rule that significant contributions to uncertainty, including any from sampling, are identified and taken into account with suitable analysis.
For a calibration laboratory the practical outputs are an evaluation for each type of calibration it offers, reflecting the range, method and conditions actually used, and the records showing how it was built. The reference standard's own uncertainty is one input among others: repeatability, resolution, the item being calibrated, environment and any corrections can all matter. The result then supports what is stated on the certificate. Calling an internal check a calibration, or the reverse, changes what is expected, so the activity must be described accurately.
Define the measurement and significant influences
Establish what is being measured, the method, relevant range and conditions. Identify contributions under clause 7.6.1, including those arising from sampling where significant. Consider the reference, the item being calibrated, the measurement process and relevant environmental or other effects using appropriate technical analysis.
The applicable contributions and analysis depend on the activity. Do not copy a generic uncertainty budget simply because it uses the same instrument family. A technically competent person must decide what is significant and how the evidence supports the evaluation.
Why clause 7.6.2 matters to the laboratory
A calibration certificate is used by someone else as the foundation for their own measurements, so errors in the stated uncertainty pass downstream and multiply. A laboratory that understates its uncertainty risks giving customers false confidence, while one that overstates it may lose work or fail to show it can meet a customer's needs. Equally, internal calibrations underpin every test result a testing laboratory produces, so skipping the evaluation weakens everything built on those instruments. An accurate, well-reasoned evaluation is also a competitive signal: customers and accreditation bodies can see that the figures on the certificate come from understanding, not from copying.
How to implement clause 7.6.2
Build one evaluation per calibration type and make it reusable without making it careless. Keep evaluations simple enough that the calibration staff themselves can explain them to a customer or assessor.
- Describe each calibration service by quantity, range, method and conditions, including on-site work if it is offered.
- For each, list the contributions: reference, repeatability, resolution, environment, item behaviour and the calculation itself.
- Gather evidence for each from certificates, repeated measurements and method knowledge, checking that it applies to the range and conditions in use.
- Combine the contributions using a documented approach and have a competent reviewer approve the result.
- Include in-house calibrations in the same system, with the same level of care, and check the outcome against how the instrument will be used.
- Control the calculation file or software, and record the version used on each job.
- Revisit evaluations when equipment, environment or method change, and compare them with monitoring and comparison data.
Use evidence appropriate to the calibration
Reference certificates, repeatability information, resolution, correction data, environmental evidence and method knowledge may all contribute where relevant. Check that the information applies to the range and conditions used. The calculation should be understandable and controlled so another competent reviewer can follow its assumptions and inputs.
An established evaluation can support a defined method and range where its assumptions and influencing factors remain controlled. That does not remove the requirement to evaluate uncertainty for the calibrations performed. Consider changes in equipment, environment, method or the item that could make the existing basis unsuitable.
Apply the same technical principle to in-house calibration
A laboratory calibrating its own equipment is still performing calibration. Internal use does not exempt that work from uncertainty evaluation or the relevant competence, traceability and record requirements. Distinguish calibration from a limited intermediate check; the name placed on an internal form should reflect the activity actually performed.
Connect the evaluation with reporting under clause 7.8.4 and any conformity statement under clause 7.8.6. A small numerical uncertainty is not inherently better if it excludes significant influences. The useful result is a justified uncertainty appropriate to the service and the customer's need.
Two illustrative uncertainty decisions
A calibration laboratory uses a highly accurate reference instrument at a customer's factory. It considers the actual on-site conditions and measurement process rather than reporting the reference certificate's uncertainty unchanged. The environmental contribution may differ from work in the permanent laboratory.
A testing laboratory calibrates one of its own measuring devices internally. It evaluates the uncertainty of that calibration and checks whether the outcome supports the device's intended use. Calling the work “internal only” does not remove the technical obligation.
Evidence and common mistakes
Retain technical records sufficient to explain the evaluation, inputs, assumptions and connection to results. An uncertainty worksheet is one possible tool, not a compulsory layout. Applicable accreditation policies and technical guidance may add detail relevant to the laboratory's field and recognised capability.
Avoid using the reference uncertainty alone, omitting significant contributions because data are inconvenient, or claiming the same uncertainty for every range and location without support. Review calculation tools and changes through the information-management controls; an unnoticed unit or formula error can undermine an otherwise careful analysis.
Frequently asked questions
Are in-house calibrations exempt from uncertainty evaluation?
No. Clause 7.6.2 expressly includes calibration of the laboratory's own equipment. The relevant technical requirements apply to that work as well.
Can we report the uncertainty from our reference certificate?
Not automatically. That uncertainty contributes to the evaluation, but the laboratory must consider the other significant influences in the calibration it actually performs.
Does ISO/IEC 17025 prescribe one uncertainty spreadsheet?
No. Use appropriate analysis and technically justified evidence for the activity, with controlled calculations and records that explain the result and its uncertainty.
Turn this requirement into reliable laboratory practice
Sterling can help organise the competence, records and implementation work supporting your calibration uncertainty evaluations. Bring a representative calibration process to a free consultation; authorised technical specialists remain responsible for the measurement model and conclusions.
Build the team's understanding through Measurement Uncertainty (ISO/IEC 17025:2017). Discuss the relevant scope, employee learning and agreed access to Effivity ISO & GRC automation software with Sterling; laboratory technical decisions remain with competent, authorised personnel.
ISO/IEC 17025 requirements hub · Uncertainty overview · Traceability · Calibration certificates