|
Making accurate estimates of the operating temperature of encapsulated IC's in its end application environment, which is necessary to predict and ensure product reliability, usually consists of characterizing the thermal resistance of the IC package. One common way of characterizing thermal resistance is with Theta JA, a single lumped constant describing thermal resistance between the chip or die junction and the ambient environment. Unfortunately, Theta JA is deemed to be too dependent on mounting and environmental conditions. Another thermal constant, Theta JC, which is the thermal resistance between the die junction and the IC case or package, is also frequently encountered in thermal resistance characterizations. Although this constant is not affected by environment, it is only valid for an isothermal case or package surface. This study deals with two other known models for package thermal resistance, namely, the Bar-Cohen and the Lasance constant models. Validation and analysis was accomplished using detailed 2D and 3D Finite Element Models. The study concluded that the Bar-Cohen model is accurate in characterizing the thermal resistance between the junction and the package when its case is maintained at a constant temperature or is under convective boundary. It will introduce more error in characterizing a board-mounted package where the conduction of heat through the leads is dominant. Lasance's modified model is an improvement over the Bar-Cohen model with its introduction of coupling resistors. However, the single node leads representation will introduce error when a board with non-uniform metallization is encountered. More lead nodes with coupling resistors will help increase the accuracy. A single lumped thermal resistance for a board-mounted package, Theta JL, which is the thermal resistance between the junction and the leads, is relatively independent from the environment and can be used for quick determination of the junction temperature in application environment.
|
