The underlying problem
A calibration is a photograph on a given date. Between one and the next, the instrument can change through wear, impacts, contamination or environmental variations, without anyone noticing. Ensuring validity between calibrations means detecting those changes when they occur, not when the interval is up.
Main tools
- Intermediate checks. Compare the instrument with a standard or stable reference at one or a few points of interest, more frequently than the full calibration.
- Check standards. Stable artifacts with a well-known value, measured repeatedly. A change in what the instrument reads on that artifact indicates drift.
- Control charts. Plotting verification results over time reveals trends before they cross the acceptance limit.
- Redundant instruments. Two sensors measuring the same thing should agree; a progressive divergence signals that one of them is moving.
- External comparisons. Taking part in proficiency tests or interlaboratory comparisons checks your own measurement against those of third parties.
- Repeat measurement of retained samples. Measure a retained sample again and compare with the previous value.
- Analysis of calibration history. The drift observed between consecutive calibrations indicates whether the interval is adequate.
What makes a verification useful
The check standard must be more stable than whatever is being monitored, and the verification must be done under similar conditions each time. Otherwise, changes in the instrument are confused with changes in the environment or the method.
How to interpret the result
| Observation | Likely reading |
|---|---|
| Sudden jump in the verification | Impact, failure or configuration change |
| Gradual drift in one direction | Aging or contamination |
| Erratic variation | Connection problem, noise or unstable conditions |
| Stable within the criterion | No evidence of change |
A verification does not replace calibration; it serves to decide, with evidence, whether the instrument remains reliable until the next one.