Service
Calibration intervals and ISO 17025 traceability
How to read a calibration certificate, compare uncertainty with tolerance, and set an interval that survives an audit.
A calibration certificate is not a sticker that says the instrument is good. It is a measurement result with a stated uncertainty, a traceability chain and a scope. The useful questions are whether the laboratory was accredited for the parameter and range you use, whether the uncertainty is small enough for the tolerance you are judging, and whether the interval matches the risk in your process. This note explains how to read the document and how to choose the interval without treating either as a formality.
Calibration, verification and accreditation
Calibration establishes the relationship between the instrument's indication and a reference under stated conditions. Verification checks that the instrument meets a specification. A certificate that reports as-found and as-left values is a calibration record; a pass/fail statement is a verification decision.
ISO/IEC 17025 is a requirement for testing and calibration laboratories, not a certification of the instrument. A CNAS-accredited laboratory has been assessed against that standard for a defined scope. The scope is the important part: it names the parameter, the range and the method. A certificate for voltage calibration does not automatically cover current, and a certificate for one current range does not automatically cover another. Check the scope against the ranges you actually use.
The traceability chain
Traceability means an unbroken chain of calibrations, each with a stated uncertainty, leading to a reference standard and ultimately to the SI. The certificate should identify the reference standards used, their calibration status and the method. If the chain is broken — an expired reference, a missing uncertainty statement, an unaccredited subcontractor — the number may still be useful, but it is not traceable in the sense a quality auditor means.
The battery test system's voltage and current channels are calibrated against traceable voltage and current standards in a CNAS-accredited laboratory following ISO/IEC 17025. The certificate states as-found and as-left values with uncertainty, and the certificate number is added when the document is issued. Those are the elements to look for; the laboratory's accreditation status is separate from the instrument's own accuracy specification.
Reading the uncertainty statement
An expanded uncertainty is usually reported at k = 2, which corresponds to roughly 95% coverage for a normal distribution. The statement should say what it applies to: a percentage of reading, a percentage of full scale, a fixed value, or a combination. The site's services page gives an example format of ±(0.02% RD + 0.01% FS) at k = 2, and explicitly labels it as an example rather than a specification. Treat any uncertainty number the same way: read the units and the range before comparing it with a tolerance.
Do not compare the calibration uncertainty directly with the datasheet accuracy. The datasheet accuracy is a performance limit for the instrument; the calibration uncertainty describes how well the calibration laboratory measured the instrument. A small calibration uncertainty with a large instrument error still means the instrument is out of tolerance. A large calibration uncertainty with a small instrument error means you cannot prove the instrument is in tolerance, even if it probably is.
Many quality systems use a test uncertainty ratio (TUR) of at least 4:1, meaning the tolerance is at least four times the expanded uncertainty. Where that ratio cannot be achieved, a guard band is used: the acceptance limit is tightened by the uncertainty so that a conforming decision is not driven by measurement noise. The exact rule belongs to your quality system and the standard you work to; the principle is to decide the rule before looking at the data.
| Certificate item | What to check | Why it matters |
|---|---|---|
| Laboratory and scope | Accredited for this parameter and range | Accreditation is scope-limited, not general |
| Reference standards | Identified, in calibration, traceable | Without a chain, the result is not traceable |
| As-found / as-left | Both reported, with units and range | Shows drift and the effect of adjustment |
| Uncertainty | Expanded, with k and units | Needed to judge conformity and guard bands |
| Method and conditions | Procedure, temperature, humidity | Calibration is only valid under stated conditions |
| Interval and date | Issue date and next due date | An expired certificate is a finding in an audit |
Choosing the interval
There is no universal interval. A short interval is appropriate when the instrument is critical to a release decision, when drift history shows movement, when the channel is used near its accuracy limit, or when the environment is harsh. A longer interval can be justified by stable history, moderate use and a conservative accuracy margin. The interval is agreed with the customer's quality system; the site does not impose one, because the right answer depends on the risk and the data.
Use the calibration history rather than a rule of thumb. Plot as-found error against time for each range. If the error is flat and well inside the tolerance, extending the interval is defensible. If the error is moving toward the limit, shorten it or investigate the cause. A repair, a shock, a move or a change in the measurement setup is a reason to recalibrate early, regardless of the calendar.
What to file
Keep the certificate, the scope, the uncertainty statement and the as-found/as-left data together. Record the instrument serial number and the channel or range in your asset register, not just the model. If the certificate covers a subset of channels, note which ones. When an auditor asks whether a result was produced on a calibrated instrument, the answer has to connect a specific channel and date to a specific certificate.
Takeaways
- ISO/IEC 17025 accreditation belongs to the laboratory and is limited to a scope; check that the scope covers the parameter and range you use.
- Read the uncertainty with its units and coverage factor, and compare it with the tolerance before deciding conformity — the datasheet accuracy is a different number.
- Set the interval from drift history and risk, not from a default; recalibrate early after repair, shock or a change in the measurement setup.
Request a configuration review
Tell us the channel count, current range and protocol. We will check the measurement class against the note and return a configuration, not a generic datasheet.
Request a configuration review