Energy management in smart building with wire-bonded QFN64b electronic package subjected to air free convection. Experimental and numerical study

摘要

Energy management in smart building is ensured by electronic assemblies often equipped with quad flat no-lead (QFN) devices. The wire-bonded model with 64 leads, denoted as QFN64b used for the simultaneous control of various equipment generates a high power ranging between 0.1 and I W. When its cooling is performed by air natural convection, the junction temperature can reach high values exceeding the maximum recommended by the manufacturers. This causes its dysfunction and even destruction. The knowledge of its thermal state is needed to properly size the electronic assemblies equipped with this QFN64b and ensure correct operation. This is the objective of this work in which the thermal field is determined throughout the assembly which could be inclined by an angle varying between 0 (horizontal position) and 90 (vertical position), according to the intented application. The 3D numerical approach has been done by means of the finite volume method. The study shows that the tilt angle has little effect on the QFN64b's thermal state. The linear increase of the temperature versus the generated power confirms that the global thermal resistance is constant. Moreover, the temperature of the QFN64b sides is more sensitive to the generated power than its top surface. In the horizontal position, the average temperature of the sides may exceed that of the environment by about 32 K. Measurements were performed on a prototype for all the configurations processed numerically. Deviations with the calculated values are small, which validate the adopted model. Correlations are proposed to determine the temperature in various areas of the assembly, according to the generated power and the inclination angle. They optimize the thermal design of the wire-bonded QFN64b electronic package which is increasingly used in many engineering fields such as in the field of smart building considered in this survey. (C) 2017 Elsevier B.V. All rights reserved.