How to choose a calibration laboratory

Six checks

• Accreditation: by a body within the ILAC MRA, for calibration (not only testing).
• Scope: covers temperature (or your quantity) across the range you need, with a CMC at or below the uncertainty you require.
• Method: comparison in bath or dry block; immersion; whether they calibrate loggers as complete systems.
• Certificate content: uncertainty, k, traceability statement, conditions, as-found and as-left values, decision rule.
• Out-of-tolerance handling: do they notify before adjustment? Do they report both states?
• Logistics: turnaround, packaging guidance, whether on-site calibration is offered and how its uncertainty differs.

ValiTrac's laboratory finder checks these criteria against a small, individually-cited catalog of published accreditation scopes; for a laboratory not yet in that catalog, the accreditation body's own directory remains the authoritative source.

Start from your tolerance, not their brochure

The selection runs backwards from the decision the instrument has to support. If a logger polices an 8.0 °C limit against a ± 0.5 °C specification, the conventional target is a calibration uncertainty of about a quarter of that — 0.12 °C — and any laboratory whose scope can plausibly deliver it once your logger's own resolution and repeatability are added is a candidate. That single calculation eliminates most of the field before anyone is phoned.

The question that tells you most

Ask how the reported uncertainty will be built up for your specific device. A laboratory that answers in a sentence — your logger's resolution contributes about 0.03, our chamber uniformity about 0.05, so expect around 0.13 at k = 2 — builds a budget per job. One that quotes its CMC and stops has probably not included your instrument in the budget at all, and ILAC-P14 says the reported uncertainty normally exceeds the CMC for exactly that reason.

Turnaround and the things that go wrong afterwards

• Agree the adjustment policy in the purchase order: no adjustment without contacting you, and as-found reported whether or not anything is adjusted.
• Agree the points. Accepting a laboratory's standard set gets you a certificate at temperatures convenient to generate rather than the ones your alarms sit at.
• Agree the decision rule, or ask for no conformity statement at all and make the judgement yourself against your own tolerance.
• Ask about turnaround honestly, including the queue rather than the bench time, and plan a spare instrument if coverage cannot lapse.
• Ask about packaging and transport. A reference PRT damaged in transit is a common and expensive failure, and who bears it should be settled before it happens.
• Ask whether they participate in proficiency testing or interlaboratory comparisons for your quantity. It is a reasonable due-diligence question and the answer is informative either way.