Drift

Drift is the slow, progressive change in an instrument's output or indication over time. Unlike a sudden failure or shift, drift is gradual and may go unnoticed unless the instrument is periodically recalibrated. Common causes include component aging, mechanical wear, spring fatigue, crystal aging (in frequency standards), and cumulative environmental exposure.

Drift is detected by comparing as-found calibration data across multiple calibration cycles. A consistent trend in one direction — readings gradually increasing or decreasing over successive calibrations — indicates drift. The rate of drift (units per time) is a key parameter for predicting when an instrument will go out of tolerance and for setting appropriate calibration intervals.

In calibration management, understanding drift is essential for interval optimization and risk management. If an instrument's drift rate is known and predictable, the calibration interval can be set so the instrument is recalibrated before it is expected to reach its tolerance limit. Mathematical models can extrapolate drift trends to predict the probability of out-of-tolerance conditions at any point in time. Instruments with excessive or unpredictable drift may need more frequent calibration, environmental protection, or eventual replacement.