In metrology, drift can be defined as the variation caused in the output of an instrument, which is not caused by any change in the input. Drift in a measuring instrument is mainly caused by internal temperature variations and lack of component stability. It also represents a static characteristic of an instrument.
A change in the zero output of a measuring instrument caused by a change in the ambient temperature is known as the thermal zero shift. Thermal sensitivity is defined as the change in the sensitivity of a measuring instrument because of temperature variations. These errors can be minimized by maintaining a constant ambient temperature during the course of a measurement and/or by frequently calibrating the measuring instrument as the ambient temperature changes. Drift may be classified into three categories:
- Zero drift: if the whole calibration gradually shifts due to slippage, permanent set, or due to undue warming up of electronic tube circuits, zero drift sets in. In other words: the zero drift is the deviation of the index from the zero position, that is from the origin of the graduation curve.
- Span drift or sensitivity drift: if there is a proportional change in the indication all along the upward scale, the drifts are called span drift or sensitivity drift.
- Zonal drift: in case the drift occurs only a portion of the span of an instrument, it is called zonal drift.