The rules of the Rue Morgue: a decade later

Abstract

The paper could limit itself to repeat the complaint that originated the first "Rules of the Rue Morgue", maybe updating the scenario of the many end users currently exposed to the risk of failed failure analyses. Nevertheless, some constructive proposals will be also pointed out, as those exposed by a recent paper (Cassanelli et al., 2005) that, dealing with the challenges in system reliability predictions, proposed some shortcuts to include even few field data into that process, and to include F.A. findings, when reliable, to skip cumbersome (and often not available) extraction of reliability parameters by statistical data. More specifically, in both the reported B and C cases the sudden occurrence of the failure mode was not related to any sudden firing of the root failure mechanisms, but other hidden roots have been identified, with completely different corrective actions with respect to the first interpretations. There is a simple and "correct" conclusion to this result: by means of thorough analyses, the first specimen (IGBT) was indicted for some higher sensitivity to latch-up, and the second (CMOS) to external EMI-induced ESD events. This could move to correct the corresponding pi factors employed for calculating the actual failure rate lambda drawing a physically sound shortcut to the estimation of the reliability parameters for some critical devices of a given electronic system