Conclusion
In this project, the heat dissipation effect of high power LEDs implementing COB architecture was analyzed using infrared thermography methods and finite element analysis. Initially an existing LED implementing COB architecture was studied and the problem statement was identified. The heat sources at the center heat up faster and achieve a higher steady state temperature due to thermal crossover effect. Further experiments were performed to explore methods to minimize this thermal crossover effect. The results of the experiments can be summarized below
Under fixed power conditions, comparing the temperatures at the LED surface and the point at the center, it was concluded that a 4mm or a 5mm distance of separation was optimum for minimum thermal crossover effect while maintaining the compactness of the module. As power was increased, the temperature increased more rapidly as the distance between heat sources reduce. For distances under 4mm or 5mm, the temperature rise is exponential establishing the previous conclusion. Furthermore, 1W power is optimum for this heat source dimensions to prevent the module from heating above the critical junction temperature that causes failure.
Fixing power at 1W and varying the thermal resistance of the heat sink and varying the separation distances between the center of the substrate and the heat source, the temperature difference at each case indicated that 4mm is the ideal separation distance. For 4mm separation, the minimum thermal conductivity of the heat sink to minimize thermal conductivity must be 50W/mK. Below this value, the thermal crossover effect is significant which was indicated by a greater increase in the midpoint temperature than the heat source temperature with increase in thermal resistance of the heat sink. This conclusion was further substantiated by studying the heat flux distribution.
The major contribution of the project is to construct a guideline for designing LED modules that minimizes the thermal crossover effect and the junction temperature to 85°C. The maximum number of heat sources that can be arranged in a circular pattern for different values of power and thermal resistances were determined. Linear arrangement of heat sources is superior to a circular arrangement as seen in the current LED modules since it gives lower junction temperatures and thermal crossover effect. The rate of increase in thermal crossover effect with respect to thermal resistance is higher is in a circular arrangement pattern. In LED COB arrays, the LEDs are covered with a thin layer of silicone encapsulate for protection. Low thermal conductivity of silicone forces heat accumulation around the heat source as demonstrated in the last set of experiments. Although the increased thermal crossover effect has negligible effects on the arrangement if heat sources around the center, they cause a rise in junction temperature and hence their effect cannot be neglected.
In all the experiments performed, the center of the substrate has no heat sources. This was done to solve the heat accumulation at the center as stated in the problem statement drawn by studying the CITIZEN LED COB module. This center space can be used to clamp the COB module to the PCB substrate. Thus, we can conclude that heat sink properties and center distances are the 2 most dominant factors that affect thermal crossover effect in LED COB modules and hence affect the arrangement of heat sources.
The research has focused on studying the thermal dissipation in high power LED packages and ways to alleviate the thermal crossover effect that affects the junction temperatures. The discrepancy attained in the heating response comparison between he infrared thermography experiments and ANSYS thermal simulations show that practical cases might differ from the simulated and analytical models due to simplification of modeling, thermal contact resistance, forced convection, radiation effects etc. This study outlines a method to design an LED COB module by heat sink selection, heat sources arrangement and density, and power selection for heat sources. An analytical solution must be developed based on the experimental solution that reduces the selection to a few set of mathematical expressions that can be solved instantaneously.