Strictly seen: A calibration is only valid at the time of implementation. However: In practice, the question more frequently posed is how often should a recalibration actually be carried out. In general, the operator is responsible for the determination of calibration intervals.
If internal specifications for recalibrations are available, e.g. in the quality management manual, then these will be the official source for recalibration schedules. For some applications, such specifications may also be present in general standards (e.g. ISO 376 for force measuring instruments or certification of vehicle exhaust gas emissions).
For those measurements where the highest demands are placed on precisely known measuring properties, it must be noted that the calibration can only make valid statements at the time of implementation. Consequently, an extremely complicated process is required: A calibration must be implemented both before and after every important measurement.
Examples of such applications with high requirements are measurements within the framework of DakkS calibration equipment accreditation or measurements within the framework of motor vehicle certification according to exhaust gas regulations.
If a more pragmatic view is followed in industrial practice, as explicitly recommended in ISO 10012, it is of course more sensible to allow a greater number of measurements or a specific time interval between two calibrations.
If the deviations measured during a calibration compared to the previous calibration lie within the metrological requirements, then the measurement results obtained with the measuring equipment are justifiable. If, however, the deviations are greater, then the question is whether measurements are only meaningful to a limited extent and should they be repeated.
The decision on how long a calibration interval should be must therefore take into consideration how high the costs are, on the one hand, for more frequent calibration (including time lost) and, on the other hand, for possibly worthless measurement results, re-measurements, recall actions, etc.
An important aspect here is also the probability for changes in the measuring properties that may result in significant deviations in calibration results between one calibration and the next. Qualitatively, it is easy to ascertain that certain conditions may require more frequent calibrations, e.g. such as high operating hour values (shift operation), extreme temperature conditions, long-term alternating load operations with transducers, dirt and moisture.
However, to produce quantitative statements about the measuring equipment used with the help of manufacturer data would require comprehensive statistical data for each type of transducer or measurement electronics that is normally not available. The operator can instead obtain a very good idea, by continuously tracking the calibration results, of the long-term behavior of the equipment that he uses under the operating conditions valid for the application. In other words: If a measuring amplifier is used in a test bench where the operating conditions are hard and the costs high, and measurement results are shown at a later stage to be untrustworthy, it may be sensible to implement a recalibration after 6 or even 3 months.
If, however, it is clear after the first or second recalibration, that the measuring properties remain stable, it is then probable that the measuring amplifier will also remain stable and one can then decide to lengthen the calibration interval. Such a procedure for determining under which conditions the calibration interval can be lengthened should be a part of the QM system. And it should at least also equally handle the shortening of calibration intervals, e.g. due to wear or drift behavior.
Comparison measurements of several calibrated test instruments among each other are another decision-making aid for adapting calibration intervals, For example, if a test laboratory uses several force transducers and has the equipment needed for comparison measurements on hand. Such comparisons can show whether a calibration interval that, initially, may have been rather generously calculated should be shortened in an individual case.
The significance of the operating conditions naturally means that a recalibration should be implemented in all cases where a measurement tool is subject to stresses that lie outside the intended use. These range from greater overloads, falling down, extreme temperature conditions to interventions in the equipment for repair purposes.