Issue link: https://iconnect007.uberflip.com/i/253291
44 SMT Magazine • February 2014 ple test points under consistent test conditions. Figure 2 illustrates examples of repeatable and non-repeatable processes. Thus, for the purpos- es of specifications, achieving smaller repeat- ability numbers is better because it indicates tighter groupings or a smaller range within the test data distribution. Repeatability differs from accuracy in that it is concerned with variations in achieved results relative to each other within a given sample size. In a fluid dispensing system, accuracy is the ability of the equipment to dispense to a target position. An accurate system is one that dis- penses exactly at the target location. In order to measure accuracy, though, we really mea- sure the individual test offset error between the achieved result and the target (ideal), as shown in Figure 1. Thus, accuracy specifica- tions are really expressions of the system in- accuracy and achieving a smaller specification means a system is more accurate. When speci- fying accuracy, however, it should not be based on a single test result, but many test results. As such, it is common practice to use the aver- age (mean) error amongst a range of test data to express the accuracy of the entire test data distribution. Accuracy and repeatability are closely re- lated but are two independent specifications. It is possible for a system to be repeatable but not accurate and also possible for a system to be accurate but not repeatable. As previ- ously mentioned, an ideal system is one that is both accurate and repeatable, as illustrated in Figure 3. accuracy and repeatability in Specifications: Standard Deviation, Cp and Cpk Accuracy and repeatability need to be placed in context to a specification to be meaningful. Typically, accuracy and repeatability specifica- tions are bounded by an upper specification limit (USL) and a lower specification limit (LSL). Standard deviation, Cp and Cpk, describe how accurate and repeatable a set of test results are contained within the USL and LSL. The standard deviation, σ (sigma), is a sta- tistical measure of the difference between indi- vidual data points to the average value of the sample of data and indicates spread of mea- sured data. Standard deviation is calculated based on the following formula: Figure 2: repeatability. Figure 3: Accurate and repeatable process. FEATUrE MaKING SeNSe OF aCCuraCy, rePeaTabILITy aND SPeCIFICaTION continues