Temperature has been considered by many reliability engineers and system designers as a major contribution to the unreliability of electronic equipment.  As a result, temperature has often been lowered without fully understanding the impact of the cooling system reliability; the cost of temperature reduction in terms of dollars, weight and size; and any actual reliability improvement gained by lowering temperature.  In this paper, existing methodologies for modeling the temperature-dependence of microelectronic device failures are discussed, and the reasons why these can lead to misleading results are explained.  An alternate methodology is suggested at the end, to define the objectives of new design models which provide more realistic relationships between temperature and microelectronic failure mechanisms.  The new methodology is based on an increased level of understanding of the role of temperature, and is intended to help the reliability community break away form existing misleading paradigms.
 

Complete article is available to CALCE Consortium Members.

[Home Page] [Articles Page]