When Your Calibration Record Becomes Legal Evidence

You keep calibration records because IATF 16949 clause 7.1.5.2 and ISO/IEC 17025 tell you to, and because a surveillance auditor will sample them. That has been the whole job for years: satisfy the standard, pass the audit, move on. But on 9 December 2026 the reason you keep those records changes underneath you, and most quality teams have not noticed.

That is the transposition deadline for the revised EU Product Liability Directive (Directive (EU) 2024/2853), which entered into force in December 2024. It is not a metrology standard, so it never lands on a quality manager’s desk. It is product-liability law — and it quietly rewrites what a calibration record is for. From that date, in every Member State, your as-found/as-left data, your certificates, and your traceability chain stop being purely audit paperwork and become potential evidence in a courtroom.

What actually changed

Two provisions matter for anyone who measures parts before they ship.

First, disclosure of evidence. Under the revised directive a national court can order a defendant manufacturer to disclose relevant evidence that is at its disposal, where a claimant has presented a plausible claim. If you are ordered to disclose and you cannot — the records do not exist, are incomplete, or cannot be shown to be authentic — the court may presume the product was defective. The gap in your evidence is no longer a neutral fact; it becomes an inference against you.

Second, the shifted burden of proof. The old regime put the full weight on the injured claimant to prove the defect and the causal link. The revised directive introduces rebuttable presumptions that ease that burden, including where the claimant faces excessive difficulty due to technical or scientific complexity, or where the defendant has failed to comply with a disclosure order. In practice, the manufacturer increasingly has to affirmatively demonstrate that the product conformed — and the primary way you demonstrate conformance is by showing the measurements were sound and the instruments that took them were in tolerance and traceable.

There is a third detail that turns this from a legal footnote into an operational problem: time. The liability window runs ten years from the moment a product is placed on the market, and the revised directive extends this to twenty-five years for latent personal injuries that surface slowly. Your calibration records for a part you ship this quarter may need to be produced, intact and authentic, a decade or more from now.

A worked example

A supplier machines a safety-relevant component and ships a batch in early 2027. Three years later a field failure leads to a personal-injury claim, and the claim reaches a court in a Member State that has transposed the directive. The claimant’s counsel argues the part was out of specification.

The court orders disclosure. To rebut the presumption of defect, the manufacturer needs to show, for the production date in 2027, that the caliper and the CMM used at final inspection were within calibration, that each had a passing as-left result, that both were traceable to a national metrology institute, and that the inspection actually recorded the part as conforming. That is a specific claim about specific instruments on a specific day, three years in the past.

If those facts live in a shared spreadsheet, the manufacturer has three problems. The spreadsheet is editable, so its rows carry no proof they were not changed after the fact — a point opposing counsel will make in one sentence. It has gaps, because someone left, a tab was overwritten, a due date slipped and nobody logged it. And it cannot be searched to the day and instrument fast enough to meet a court’s disclosure window. Each of those weaknesses feeds the disclosure-failure presumption. The manufacturer is now arguing about the reliability of its own records instead of the quality of its part.

"We already keep records for audits — isn’t that enough?"

This is the fair objection, and the answer is no, because an audit and a disclosure order test different things. An auditor samples your process, accepts reasonable evidence of control, and works with you. A court operates adversarially: it tests whether a particular record existed on a particular date and has not been altered since. A spreadsheet that passes a friendly Tuesday-morning surveillance audit can fail completely against that test, because it was never built to prove authenticity or completeness — only to be present.

The bar that the revised directive quietly raises is therefore not "do you have records" but "can you produce a complete, tamper-evident, instrument-and-date-specific trail, on demand, for up to twenty-five years." Those are different systems. One is a document. The other is an audit trail.

What an evidence-grade trail looks like

This is where the calibration record has to stop being a file and become a controlled, immutable history. Concretely, an evidence-grade system holds every instrument’s identity and its calibration due date; every calibration event with as-found and as-left values, pass/fail decision, and the attached certificate; the traceability chain back to a national standard; and the full custody trail of where the instrument was and who used it. Crucially, those entries are append-only — corrections are recorded as new events, never silent overwrites — so the record can answer the only question a court cares about: this instrument, in tolerance, traceable, on this date, and this history has not been rewritten.

This is exactly the shape of what PrüfDesk was built to hold: instruments registered with auto-calculated due dates, pass/fail records with as-found/as-left and certificates, procedure templates for consistent evidence, and a full immutable possession and calibration trail — hosted in the EU, where your data residency and your court both sit. Configurable overdue blocking means an instrument that has slipped past its due date is caught before it takes a measurement that later cannot be defended.

The one thing to do before 9 December 2026

Treat the transposition date as a hard deadline for one question: if a court ordered disclosure of the calibration evidence behind a part we shipped, could we produce a complete, unalterable, date-and-instrument-specific trail — and will it still be retrievable in ten years? If the honest answer is "it’s in a spreadsheet," you have a window, measured in one calibration cycle, to move that evidence into a system built to survive the question. Getting the trail in place before transposition is far cheaper than discovering its gaps under a disclosure order.

If you want to see what an audit-ready, evidence-grade calibration trail looks like on your own instruments, PrüfDesk is built for exactly this — and you can have it in place before the deadline rather than after the claim.