Thermal Design for Six Sigma of an Imaging Device with Dynamic Thermal Management
Wendy Luiten · 2018
Imaging devices are a growth market in today's economy. They are used as the eyes of many remote surveillance applications, in manufacturing and in logistics operations. One of the advantages of remote vision applications is the possibility to operate outside the human temperature range and in harsh environments. This poses new challenges for the thermal management of these devices as electronics have to operate reliably at higher ambient in sealed enclosures. Dynamic Thermal Management (DTM) scenarios are extensively used both in image sensors and in video processors to prevent operation at overtemperature. At high system environments, the temperature difference between system ambient and maximum component temperature become small, and the error in the on-chip temperature sensors becomes important. Previous work has shown the applicability of Design for Six Sigma methodology in thermal design. The effects of variability have so far not been considered as by nature CFD simulations are free from measurement errors. New insight recognises the error in the DTM temperature sensor as a source of variability. In addition, the heat dissipations are not fixed but use case (software) dependant. The present work uses DfSS tools to investigate system level performance and reliability including the errors in the on-chip DTM sensor and variability of in-use heat dissipation. The work shows that the variability results a bandwidth rather than a sharp boundary around the usage envelope. If this bandwidth becomes too wide, behaviour becomes erratic and DTM can contribute to lower reliability.