I was cleaning up some of my reference material and came across a copy of the ESA board of investigation report into the Ariane 501 accident. I’ve added my own personal observations, as well as those of other commentators, to the report.
What I find interesting about this accident, is the role that untested, implicit and as it turned out incorrect assumptions, such as a fault hypothesis that assumed software to be perfect, played in the disaster. In re-reading the report I was also struck by the recommendation that sensors should always send ‘best effort’ data. This is a particulary pertinent issue when considering fail safe design strategies and the selection of voting algorithms for redundant sensor channels.
European Space Agency, “Ariane 5 – Flight 501 Failure“, Board of Inquiry Report, 19 July 1996, 18p. Original report accessed from http://www.esrin.esa.it/htdocs/tidc/Press/Press96/ariane5rep.html.