Pre-determining comparative tests and utilizing signal levels to perform accurate diagnostics

Standard diagnostic schemes don't do enough analysis to focus in on the actual cause of a test failure. Often measurements will be border-line on test sequences prior to an actual test failure. These border-line measurements can be used to aid in the determination of an actual fault. An actual fault and the associated test that should detect that fault can be deceiving. The specific test in question can pass but be right on the border-line. The failure might not show up in testing until a later test is performed. This later test assumes all the prior tests passed and therefore the circuitry associated with these prior tests is good. This is not necessarily the case if some of the output measurements prior to the actual failing test were right on the border-line. What can we do? Take advantage of the order in which the faults are simulated1. We should structure our TPSs such that a review of preliminary tests should be evaluated before the R/R component list is presented. The review of the preliminary test can be rather straight forward. We can look at signals that are within 8–10% of the lower or upper limit. We can then use an inter-related test scheme to evaluate the test(s) that can be associated with the actual failing test. This paper will use an example TPS and show how a test scheme and an evaluation scheme can be used to determine the PCOF. The paper will show actual measurements and how these measurements can be evaluated to determine the actual cause of a failure.