Thermal characterization of electronic parts with compact models: interpretation, application, and the need for a paradigm shift

In recent years, renewed interest in the thermal characterization of electronic parts has been driven by the need for more accurate thermal analyses at board and system levels. A European consortium, partially funded by the European Community under the ESPRIT contract code-named DELPHI, has published a series of papers dealing with both numerical and experimental aspects of accurate numerical modelling and experimental validation. Their approach has spawned many discussions, amongst engineers and scientists alike, about the physical integrity of the compact models, about the accuracy required by end-users given the other unknowns at system level, and about the practical generation and implementation of the models. Another interesting area of endless discussion concerns the questionable notion of the thermal resistance as the analog of the electrical resistance. It is argued that the analogy between electrical and thermal resistances is only valid for unrealistic boundary conditions, so the terminology commonly used in the thermal characterization of electronic parts should be reconsidered. This paper hopes to set the stage for fundamental considerations about current and future ways of generating and presenting thermal data for electronic parts, which may possibly be of value for both component manufacturers and system designers.