TABLE OF CONTENTS ESS planning objectives require a quantitative specification of what the ESS program is to achieve. It is not enough to say "precipitate early failures" without specifying the reliability of the process. The ESS objective is thus to meet some reliability requirement at the time of product delivery. Given that the product was designed right to start with, ESS cannot improve the product beyond its designed-in reliability, but ESS can bring it close to it and do so at minimum cost. Cost may involve the screening process, its duration, cost of early failures (repairs and replacements) and penalties paid for incomplete screening, as well as the cost of failing "good" parts in the assembly that may occur during screening because of the elevated stress levels.
Two parameters appear to be important in the economical design of ESS programs:
The degree of contamination; i.e., the proportion of parts with latent defects contained initially in an otherwise well-designed system.
The predictive life distribution of parts in the assembly which exhibit early failures.
Both are statistical concepts. They govern the ESS process and determine the rate at which the purification proceeds. Both can be estimated from prior experience supplemented with current data.
Systems consist of components. Component populations contain latent defects. These combine into system latent defects that cause early failures (so-called infant mortality failures). In addition to component latent defects, production flaws are introduced into the assemblies at each phase of the manufacturing process.
