Demonstrating ultra-high component reliability

Abstract

The ever-increasing desire/need to produce high reliability systems of long life has lead to ultra-high reliability requirements for supplied components and assemblies. Requirements such as 99% survivors at 5 years or 90% survivors after 20 years are very difficult to demonstrate in the laboratory in a reasonable period of time. What alternatives exist to assure that the desired component reliability is present? The purpose of this paper is to outline a coherent series of steps that lead to the ability to demonstrate conformance with ultra-high reliability requirements of components and assemblies. This process begins with a statement defining ultra-high reliability requirements. Next, a short discussion about the ability to achieve ultra-high reliability component requirements sets the stage for examples. The first example is mechanical and the second is an electronic assembly. Both are explored in detail in order to show the full problem and why simple life test approaches may fail to assure the desired reliability is present. During the paper, consideration for product development systems, safety, warranty issues, quality and system reliability concerns are briefly discussed. These two examples explore a series of steps that can achieve and/or demonstrate the desired reliability without resorting to "brute-force", costly, time-consuming and/or possibly fallible test approaches. Lastly, the paper completes the examples by offering a systemic approach that can serve as a road map for most situations and companies.