Generation of subassembly compact half models through experiment and modeling for hard drive thermal characterization

Abstract

The hard disk drive (HD) is a type of electronic subassembly that has significant potential for ease of thermal management through the supply chain (from IC provider to ultimate system designer) given the appropriate temperature reference points within the assembly. The challenge lies with understanding and quantifying the nature of the coupled heating of the entire subassembly in order to manage even a single device. Often the reality of fast turn measurements for an IC vendor on a prototype subassembly will not allow the luxury of an accurate worst-case set of power dissipation values for a laboratory data point. Also a measurement-based full compact model of the subassembly including all coupled heating terms may not be achievable due to the constraints of laboratory measurement time or the capability of the subassembly to be powered into different modes which allow for sufficient linear independence for such a formal treatment. In specific situations where the temperature of one particular device is of primary concern, a much-reduced "compact half model" (CHM) of the subassembly may be sufficient for the solution of the critical junction temperature. Such a model consists of only a single self-heating and single coupled heated term. While this model would necessarily have significantly less information than a formal compact model, the accuracy for many geometries may be sufficient for subassembly thermal performance assessment, package selection and roadmapping exercises. Limited measurement data sets will be considered as well as the role of choice of reference temperature from the point of view of model generation and ease of thermal management along the supply chain. The principle goal is the identification of the most straightforward method of a single device-in-subassembly performance characterization through prototype measurements.