How often should an instrument be calibrated?
There is no fixed calibration interval. It depends on how critical the instrument is, its drift history (as-found and as-left records), the manufacturer's advice, process risk and quality requirements. Start from the manufacturer's advice, lengthen the interval only with stable records, and shorten it when results fall outside tolerance.
Why there is no fixed interval
Calibration compares an instrument's reading with a known reference and corrects the reading if needed. How often this is done depends on how far the instrument drifts over time. That drift differs between instruments and between processes. A critical measurement does not have to follow the same calendar as a low-risk one that has stayed stable for years.
Each calibration record holds two values. As-found is the error when the instrument arrives from the field. As-left is the error after any adjustment. The share of as-found results that stay within tolerance shows whether the current interval is adequate. Until your own records exist, the manufacturer's recommendation is only a starting point.
Example: reviewing a 12-month interval
A pressure transmitter starts with a 12-month interval (hypothetical). Three consecutive calibrations show as-found errors within tolerance and a small trend:
- Calibration 1: as-found within tolerance
- Calibration 2: as-found within tolerance
- Calibration 3: as-found within tolerance, records stable
If this history is documented, the interval can be extended from 12 months to 18 months. If the same instrument later falls outside tolerance, shortening the interval from 12 months to 6 months may be sensible. Three clean records show a direction, but they may not be enough for a long interval alone. Write down the basis of the decision together with the measurement uncertainty and the tolerance band.
Common mistakes
The most frequent mistake is fixing the schedule from the first year and never reviewing it. The second is not recording the as-found result or the adjustment made; without them, nobody can tell whether the interval is right. The third is ignoring the uncertainty of the reference standard. The reference should be clearly better than the instrument under test, and a common guideline is a test uncertainty ratio of 4:1. Finally, process changes such as a new product, a different temperature range or a chemical change alter the conditions the instrument sees, so the interval should be reviewed again.
Follow-up questions
What if an instrument was out of tolerance? What about past readings?
First, determine the as-found error and estimate when the deviation probably started. Reassess the product and process records from that period, with the quality function, and document the result as an impact assessment. Records with no impact should also be kept, with the reason.
Can instrument monitoring reduce how often I calibrate?
It helps, but it does not replace calibration against a reference. Redundant sensor comparison and known-state checks can show how far two readings have moved apart, and that evidence supports an interval decision. To correct a systematic error directly, however, the instrument must be calibrated.
Sources
Example numbers are hypothetical, not measured field results.