IPC TM-650 1.9A

Abstract

Tests performed on presumably identical samples under seemingly identical conditions do not always yield identical results. This is due in part to errors inherent in every measurement. During the development of a new test procedure or use of an existing test procedure, this variability must be understood and precautions taken to ensure that it is controlled to within necessary limits. Performance of this test method will help to estimate measurement error and troubleshoot possible causes. It can provide evidence that a new test procedure is suitable for use when submitted for review, or an existing test procedure is capable of measuring the applicable parameter. This method provides a simple, easy to use, standard procedure for determining the precision of a test method using the average and range method. It can be used on tests that involve measurements that yield continuous data. The calculations shown in this procedure are streamlined versions; useful for situations where up to five repeated readings are taken on each of up to 10 samples by up to 10 test laboratories, operators or test conditions. This procedure is not useful for measurements which result in binary data, such as pass-fail or go-no go results, or where more than five repeated measurements or more than ten laboratories or conditions are used. These situations are covered under other methods. (see 6.3) <br>This procedure is not useful for measurements which result in binary data, such as pass-fail or go-no go results, or where more than five repeated measurements or more than ten laboratories or conditions are used. These situations are covered under other methods. (see 6.3)