Definition of calibration
Calibration is a process in which the deviation of a sensor or measuring system from a reference standard is determined and documented.
The measuring system consists, for example, of a force sensor and a display device, such as a measuring amplifier (GA) with an interface or a display. Calibration must generally be metrologically traceable. Traceability ensures that the measurement results are comparable with national and international standards.
During calibration, each measured value is assigned a deviation from the target value and a measurement uncertainty. The measurement uncertainty indicates the probability (usually 95%) that the measured values lie within the measurement uncertainty interval. The measurement uncertainty itself is composed of the laboratory's measurement uncertainty with its references and the measurement uncertainties determined during the calibration process according to a validated calibration procedure. The laboratory's measurement uncertainty is also reported in the calibration certificate as "best measurement capability" (BMC) or "lowest quantifiable measurement uncertainty."
Adjustment
Adjusting a measuring system involves an intervention that modifies the system. Such an intervention might involve minimizing display deviation, for example, by adjusting a trim potentiometer in analog systems, by setting a scaling factor in the display unit, or by triggering a calibration function in digital devices.
Adjustment is not part of calibration unless explicitly requested or if it is the initial delivery of a measuring system.
During adjustment, the device's states before and after the adjustment are recorded and documented.
Declaration of conformity
The conformity statement verifies the calibration result against a specification.
The conformity statement is an optional part of the calibration process.
The specification can be provided by the customer, for example, or taken from a data sheet.
A conformity statement can only be issued if a decision rule has also been defined.
Sensors with force display in N or kN
1. An absolute value is specified as the tolerance for the display deviation in each load level. For example, "The maximum deviation in each load level must not exceed 10 N."
2. A relative value, based on the actual value (%RD), is specified as the tolerance for the display deviation. "The maximum deviation in each load level must not exceed 1% of the displayed value."
3. A relative value, based on the final value (%FS), is specified as the tolerance for the display deviation. "The maximum deviation in each load level must not exceed 1% of the final value."
Sensors with display in mV/V, or V, or mA
One way to assess the conformity of a sensor with a display in mV/V, V, or mA is, for example, by comparing it to the accuracy class specified in the datasheet.
In these cases, the following are used for comparison:
- either the repeatability (span b') or the reference accuracy (span b) (expressed as a percentage of the actual value)
- and the linearity deviation (expressed as a percentage of the full-scale value).
The values specified in the calibration certificate are compared to the accuracy class as the tolerance.
Decision rule
When comparing deviations against a tolerance, we must determine whether and how the measurement uncertainties specified in the calibration certificate are to be considered in the tolerance specification.
Unless otherwise agreed, we compare against the tolerance without considering the measurement uncertainty. This is our decision rule "B".
Alternatively, you can specify that decision rule "A" is applied. In this case, the tolerance is reduced by the measurement uncertainty. In this case, the conformity statement is valid with a probability of at least 95%.
If decision rule "B" is applied, the conformity statement is valid with a probability of at least 50%.