In order for measurements to be meaningful, they have to be performed with traceability to a national or international measurement standard. Consequently, all laboratories performing either calibration or testing measurements, are required to import measurement traceability from an external calibration service provider. The formal definition of "traceability" is: "The property of a result of a measurement whereby it can be related to appropriate standards, generally international or national standards, through an unbroken chain of comparisons" [1]. Typically, traceability is imported by means of a measurement standard, calibrated by a calibration service provider, who assigns an uncertainty of measurement that is smaller than the required accuracy of the measurement standard in subsequent use. Since instruments cannot be calibrated at an infinite number of points, the assigned uncertainty is only applicable to those discrete points at which it was calibrated. Using the measurement standard at the same measurement points that were calibrated, conveniently allows the assigned (certificated) uncertainty of measurement to be used as the uncertainty of measurement associated with the imported traceability, ("Direct Traceability"). The measurement standard is, however, often used at points in between those calibrated. In these cases the certificated uncertainty cannot be used as the imported uncertainty of measurement, ("Indirect Traceability"). This paper discusses the actions necessary to estimate the uncertainty of measurement for indirect traceability, the implications of not adopting these actions by citing some examples and points out the risks associated with the commonplace practice of simply accepting the certificated uncertainty of measurement for the calibration measurement results, as the applicable uncertainty of measurement for those points in-between.
展开▼
