Metrological traceability in ISO 17025
Metrological traceability is the property of a measurement result that lets you link it to a reference, through a documented, unbroken chain of calibrations. Each link in that chain contributes to the measurement uncertainty. This is how it is defined in the VIM 3, definition 2.41 . The VIM is the international vocabulary of metrology, the science of measurement. ISO/IEC 17025 sets the requirement in section 6.5. The RvA (Dutch Accreditation Council) bases its policy on this subject on ILAC P10, the policy of the international accreditation body ILAC on traceability. The concept is linked to measurement uncertainty and to the calibration of your measuring equipment.
Traceability does not mean that a measurement is correct. It means that you can show where the number comes from, and how much uncertainty has built up along the way.
What does that mean in plain language?
Picture a scale in a bakery. It is calibrated with a weight that was checked by a calibration laboratory. That laboratory had its reference compared with a national standard (at or through a national metrology institute). That gives three links: national standard, laboratory reference, bakery scale.
| Link | What happens | What you can show |
|---|---|---|
| 1. National standard | The reference weight is traced to the SI unit (the International System of Units, here the kilogram) | The certificate of the national institute, in the Netherlands VSL |
| 2. Calibration laboratory | The bakery’s weight is compared with the reference | A calibration certificate with uncertainty |
| 3. Bakery | The scale is checked with the calibrated weight | A control record with date and result |
If one link breaks, you can no longer trace the result. That is the whole point of the chain. The list stays short, but every link must have paperwork.
What an auditor wants to see
Traceability is a question of evidence. The auditor picks a measuring instrument and asks you to show the chain. They look at four things. These follow the notes to the VIM definition.
An established calibration hierarchy
You decide in advance which measuring instrument relates to which reference. That hierarchy does not have to be a straight line. A network of references is allowed, as long as it is fixed and you can follow it. The VIM describes the calibration hierarchy as a requirement.
Documented references
Each reference must be described: what it is, which range it covers, when it was last calibrated, and by which laboratory. A calibration certificate without a date is useless.
Calibration intervals with justification
How often do you repeat a calibration? The answer belongs in your procedure, based on stability, use and earlier results. “Every year” with no reason is not a justification.
Uncertainty per link
The contribution of each link to the total uncertainty must be visible. The VIM says the effort must match the contribution to the result. Where there are several input quantities, the chain may also branch. An expensive reference for an unimportant measurement is therefore unnecessary.
Fictional example: a pH meter in a dairy company
A dairy company measures the pH of milk with a pH meter. The laboratory that carries out the measurement must be able to show that the meter is traceable.
| Element | Evidence (fictional) | What the auditor checks |
|---|---|---|
| Buffer solutions | Supplier certificates with nominal pH value and uncertainty | Expiry date and storage conditions |
| Calibration of the pH meter | Calibration report with two or three points, date, person who carried it out | Whether the range covers the measuring points |
| Calibration interval | Procedure: weekly with buffer solution, annually externally | Whether the interval fits the use |
| Deviations | Log of pH checks and any corrections | Whether deviations were followed up |
Note: the buffer solution is itself a reference with its own uncertainty. This link is often forgotten, because it “just comes out of the box”.
In most cases, a calibration certificate without an uncertainty is not usable as proof of traceability. When you buy one, always ask for the uncertainty, not only the statement of conformity.
Which evidence does ILAC P10 mention?
The VIM refers in a note to ILAC P10 for the elements by which you confirm traceability. Use them as a checklist for each link:
| Element | Example with the bakery scale (fictional) |
|---|---|
| Unbroken chain to an (inter)national standard | Weight, laboratory reference, national standard |
| Documented uncertainty | Uncertainty on the certificate of the weight |
| Documented procedure | Work instruction for the weekly check |
| Demonstrated competence | The calibration laboratory is accredited |
| Traceability to the SI | Certificate states the SI unit kilogram |
| Calibration intervals | Annually external, weekly with its own weight |
Common mistakes with metrological traceability
- Using “traceable” as a synonym for “certified”. A certificate says something about a measuring instrument at one point in time. Traceability says something about the chain behind it.
- Not calculating the chain. The laboratory knows the links, but nobody has written down the uncertainty per link.
- Forgetting software or reference material. A calculation model or a reference material can also be a link. It belongs in the chain.
- A hierarchy nobody knows. If the analyst cannot say which reference they use, there is a gap in the system.
How do you make the chain visible?
A simple approach is an overview table per measuring instrument. One line per instrument is enough to answer most questions.
- Measuring instrument and unique identification
- Linked reference and its calibration date
- Laboratory and certificate number of the reference
- Uncertainty of the calibration
- Next calibration date and justification of the interval
This table belongs with your measuring equipment management, not in a loose folder of certificates. In this way you find the answer to the auditor’s question within a minute. For the related uncertainty budget, read the page on measurement uncertainty.
Frequently asked questions
Is every measurement by an ISO 17025 laboratory traceable? It should be, wherever this is reasonably possible. For measurements without a usable reference, the laboratory describes an alternative and the related justification.
What is the difference with product traceability, as in food? That concerns the origin of a product along the supply chain. Metrological traceability concerns the link of a measured value back to a reference. In practice these terms are often mixed up, which causes confusion.
Must I have a national standard myself? No. Your laboratory does not have to calibrate at the national institute (VSL). It must, however, show that its reference can be traced to that standard through a calibration hierarchy.
What do you do if a reference can no longer be traced? Then the chain is broken. Take the measuring instrument out of use until there is a new, traceable reference. Also assess the results since the last valid calibration, and record that as a deviation. See nonconformities.