by John Hurll
Volume 16:4, Oct-Dec 2009
It is often the case that it will be necessary to provide information in calibration reports about whether or not the stated results for general purpose test and measurement equipment comply with a given specification. For measurement standards it is likely that the measured value and expanded uncertainty will be of more interest to the user, and specification compliance is less relevant. All measurement results are subject to uncertainty, which therefore has to be taken into account when assessing whether or not the “true value” really does lie within the specification limits. Uncertainty evaluation is normally performed as described in the GUM [1] and related documents [2, 3]. It is sometimes assumed that if the test accuracy ratio (TAR) or test uncertainty ratio (TUR) is better than a certain value, say 4:1 or 10:1 — then the uncertainty can be ignored and, providing the stated result is within the specification limits, compliance can be assumed. This is, at best, over-simplistic and is usually demonstrably incorrect, as will be shown in this article.
