Parametric Study of Thermal Performance of a Plastic Ball Grid Array, Single Package Technology for Automotive Applications

Abstract

Abstract Thermal performance of a three chip ball grid array single package technology (SPT) solution has been evaluated under horizontal natural convection condition for under-the-hood automotive applications by solving a conjugate heat transfer problem to determine the maximum junction temperatures as a function of the ambient temperature and the material parameters. The resulting conjugate heat transfer problem is solved using the methods of computational fluid dynamics (CFD). The SPT solution provides packaging of all the die on a single, wire-bonded, plastic ball grid array (PBGA), four layer, BT substrate. All the die are encapsulated in a single mold compound block. The SPT is attached to a 1.52 mm thick, four-layer (with two solid internal copper planes), FR4 printed wiring board (PWB). The multi-dimensional heat transfer effects in the vias and the C5s are taken into account through separate sub-model approach and the effective conductivity is used in the CFD model. The actual stack-ups of the BT substrate and the PWB are used in the CFD analysis. The following die power dissipations are considered: 0.715 W for the one of the dies, U1, 0.3575 W each for the other two dies, U2 and U3. Radiative loss from the exposed surfaces of the package and the PWB to the ambient is included. The following parametric ranges are investigated: ambient temperature: 23, 105 & 125 °C, thermal conductivity of the die attach: 0.5 to 3.7 W/(m K), thermal conductivity of the mold compound: 0.2 to 50 W/(m K), and mold compound and PWB surface emissivities: 0 to 0.8. Since the objective here is the assessment of stand-alone package level thermal performance of the SPT, it is assumed that no other components are dissipating power on the PWB. Transient conjugate problem is also solved for power up of the package initially at an ambient temperature of 125 °C for a power dissipation of 7W. CFD simulations of the transient have been carried out for 7 s after the die is powered up. Based on the results of these analyses, it is concluded that all the three die in the stand-alone SPT operate below a maximum junction temperature of 150 °C for the ranges of parameters investigated. It is noted that the difference between the maximum junction temperature and the ambient temperature decreases with increase in the ambient temperature. The radiative loss to the ambient decreases the maximum junction temperature as much as 3.5 °C, which can be significant in the under-the-hood automotive applications. The maximum junction temperature is nearly independent of the thermal conductivities of the die attach material and the mold compound. It is also shown that the increase in the substrate thermal conductivity (i.e., with increase in the number of vias) reduces the maximum junction temperature significantly — by as much as 10 °C. The maximum junction temperatures become nearly independent of the substrate thermal conductivity for values above 5 W/(m K). The results also show that the temperature distribution of on the surface of the mold compound is very non-uniform and the spreading resistance between the die and the mold compound surface may be a significant portion of the junction to case resistance. For 7 W power dissipation in 125 °C ambient, the die temperature at the end of the transient exceed the maximum allowed (for steady state) temperature of 150 °C. Since this is for a short period of time, it may not pose any reliability problems.