Comperative analysis for the solution of led fixtures' heating problems usingvarious heatsink fin types

Since the introduction of incandescent light bulb by Thomas Edison, new and better methods of electric lighting is being researched through out our planet. Fluorescent lambs and LEDs (Light Emitting Diode) greatly increased the efficiency of electric lighting. However, the ever-increasing efficiency of LEDs are dependent on two key parameters: better semiconductors and junction point temperature. Both have been the topic of many studies and this study focuses on the latter. As the junction point temperature rises, the lumen output of the fixture decreases and its lifespan declines. To counter this effect, active and passive cooling methods can be used. Active methods such as forced air cooling is an expensive method used only for special circumstances. Most of the time, passive cooling via aluminium heatsink is preferred. Aluminium heatsink is connected with the PCB (Printed Circuit Board) which draws heat from the LED's junction point. The heat is transferred by conduction to the heatsink and dissipated to air. Various heatsink fins can be used for cooling of LEDs. There are many parameters that a manufacturer should take into considiration, such as the type of the fins, height, width and concentration of the fins and so on. So, this study tries to focus on those variables. For this task, SOLIDWORKS program has been used to create a basic LED fixture. SOLIDWORKS is a CAD program that have many add on tools, such as Flow Simulator. The LED fixture designed for this study includes the most essential parts of a fixture: A PCB (a metal core printed PCB is chosen from CREE labs), a 2.5 mm thick metal housing for the outer parts of the fixture which also functions as the base of the heatsink, a glass cover and of course a LED lamb. In this study Cree's XLAMP CXA 2011 model LED lamb is used. The reasons for this choice are that CXA 2011 is a preferred product for singular power LEDs and its ease of heat design. For this design, a metal core printed PCB is chosen from CREE labs. This particular PCB consists of 1.59 mm aluminium at the top, followed by 0.1 mm FR4, and 0.07 mm copper plate at the bottom which is connected with the LED. Electrical connections are ignored for the sake of simpler simulation and thermal calculations in Flow Simulator. In SOLIDWORKS, each material is predefined. The aluminium used for the heatsink and exterior metal housing is called 5032-H32. Copper, glass and FR4 are also defined in the SOLIDWORKS library. The Flow Simulation Wizard allows user to determine the conditions of each simulation. From this Wizard, the SI unit system, exterior analsys of the model, heat transfer in solids, the starting model and room temperature and air, as the default fluid is selected. After that point, each solid is defined by the Flow Simulator. Then the goal parameters are defined. For this study, Tmax(maximum junction point temperature) and Tmin(minimum junction point temperature) in the solid model are selected as goals. The Simulator begins calculating the variables and after each iteration converges to a value. When this value becomes constant after many iterations, the simulator stops and provides the user with results. In this study, a two way approach has been used. First, plate and pin fin heatsinks are compared. For this, an array of pin and plate fins are mounted on top of the base of the heatsink. For the plate fins, aluminium sheets of 55 mm lenght, 2 mm width and 35 mm height are used. For pin fins, aluminium pins with 2 mm base lenght and 35 mm height are used. 12 sets of simulations for each type of fins have been run on Flow Simulation. For the plate fins, arrays of 4 to 15 plates are used, while for the pin fins, arrays of 16 (4*4) to 225 (15*15) pins are used. Using the same room temperature (25 °C) and same fixture, each one has been compared for Tmax and Tmin.The results clearly showed that pin fins function better than plate fins. For the same effective surface area, pin fins provided a better cooling than plate fins. And generally, pin fins recorded lower temperatures than plate fins. It has also been monitored that, as the number of fins increase, the gap between the fins decline and junction temperatures increase. The reason of this, is the air flow between the fins slows down as the gap between fins decrease. So, there is an optimum number of fins a fixture should possess. This way, the first part of the study concludes; pin fins are more efficient that plate fins. The second part of the study, therefore, focuses on pin fins and their optimization. In this part, three parameters are taken into account: pin height, pin width and the number of pins. For pin height, 3 variables have been used: 25 mm, 35 mm and 45 mm. For pin width, also 3 variables have been used: 2 mm * 2 mm, 2.5 mm * 2.5 mm and 3 mm * 3 mm. In the previous simulations, for considering the number of pins, 12 sets of simulations were run. However, it has been observed that some of these are unnecessary, as too few fins result in very high temperatures and useless in forming a meaningful coraletion. The usage of too many fins also result in very high temperatures, as the gap between fins decrease to a point where the air flow slows so that even with increased effective surface area of aluminium heatsink, the dissipated heat decreases. So, instead of 12 sets of simulations, 2 of each extremities are omitted and 8 sets of simulations are run via Flow Simulation to determine the optimum junction temperatures of pin fins. 9 sets, each containing 8, of simulations are run for this part of the study. For each of these sets, an optimal value for number of fins are found. When the results are compared, a few considirations that a LED fixture designer could use have been harnessed. First, the optimum value has been achieved at the same number of fins when base pin width is constant, even if the height of the pin varies. This means that for a given pin width, there is an optimum number of pins a designer should use, regardless of the height of the pins. Second, the optimum value has been achieved at very similar Tmax values when the height of the pins are constant, even if the base width varies. This means that for a given pin height, a designer can achieve a fixed Tmax value, regardless of the width of the pins. And lastly, the optimum value has been achieved at very similar effective surface areas when the height of the pins are constant, even if the base width varies. This means that for a given pin height, a designer should not exceed a given effective surface area.