Ask three engineers how often an instrument needs calibrating and at least two will say twelve months. It is the most common answer in the industry, and for a great many instruments it is a perfectly reasonable one. It is not, however, a rule. Nothing in any major standard says an instrument must be calibrated annually. What the standards actually require is that you can justify whatever interval you have chosen.
What actually determines the interval
Three things decide how often an instrument should come back to the laboratory, and none of them is the calendar.
- Drift. How far the instrument moves between certificates. This is measurable — it is the difference between the as-found reading this time and the as-left reading last time.
- Duty. A clamp meter that lives in a van and gets dropped twice a month is not the same instrument as one that sits on a bench in a temperature-controlled room.
- Consequence. What it costs if the reading is wrong. A tester used for a compliance certificate carries more risk than one used for a rough go/no-go check.
The first of those is the one most people never look at, and it is the one that does the real work. If you have three consecutive certificates showing an instrument arriving well inside tolerance, you have evidence to extend its interval. If it arrives out of tolerance, you have evidence to shorten it — and a much more urgent problem, which we come to below.
Why as-found data matters more than the pass stamp
When an instrument fails calibration, the certificate is not the important part. The important question is: what did we measure with it since the last time it was certified?
An out-of-tolerance as-found reading is not a calibration problem. It is a traceability problem affecting every measurement taken since the previous certificate.
This is why a certificate that records only "pass" is close to worthless, and why we record as-found values on every job before touching an adjustment. Without them you cannot perform a reverse traceability assessment, and you cannot answer the auditor's question about what happened to the work done with that instrument over the previous year.
A practical starting point
If you have no drift history at all, start at twelve months — the convention is a convention for a reason, and it is a defensible opening position. Then let the data move it. Review after three cycles. Instruments that arrive comfortably in tolerance three times running are candidates for eighteen or twenty-four months. Instruments that arrive marginal stay where they are. Instruments that fail go to six.
What you should not do is set every instrument in the building to the same interval and never revisit it. That is not a calibration programme; it is a purchase order that repeats itself.
