Here is the abstract you requested from the IMAPS_2009 technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.
|Determination of Optimal Component Spacing in a High Power Light Emitting Diode Array Assembly for Solid State Lighting|
|Keywords: LED, thermal management, simulation|
|The use of high power light emitting diode (LED) for solid state lighting (SSL) has become a tangible reality. Some commercial products have appeared in the market nowadays. Nevertheless, there are still quite a number of technical issues to be addressed in practice. One major issue encountered in this field comes with the decision in the spacing (pitch) between LEDs for optimum cost, optical and thermal performance. Utilization of high power LED usually accounts for a major portion in the cost for manufacturing a lighting system. The number of components employed thus directly affects the selling price of the product. A smaller number is favored for cost reduction. But this scheme requires each of them to take more power in order to achieve a desired lux level, which intensifies power concentration in the system. From the thermal performance point of view, this not only raises the thermal burden of each LED, the efficiency of heat dissipation from the system surface is also reduced. For optical performance, the efficacy degrades with power concentration, and the degradation gets deteriorated with the worsened thermal performance. Therefore, there exists a trade-off between cost and the system performance, and there requires an optimum spacing to achieve the compliance. This paper aims to address such an issue when the array is mounted on a printed circuit board (organic or metal-core) attached with a heat sink subject to natural convection cooling. The first part of the present study correlates separately the total cost, the efficacy and the maximum temperature with the spacing. The second part discusses the determination of the optimum spacing based on some customized criteria such as a lowest cost design. In the third part, the methodology is employed for the design of a LED tube for the drop-in replacement of a fluorescent lamp as a case study. A simple correlation between the total cost and the pitch can be formed by counting only the number of LEDs. The correlation between the maximum temperature and the pitch requires a rigorous thermal analysis. This is achieved by adopting a lumped network model which is derived by successive simplified thermal resistances. This model attempts to account for the effects of the pitch, the thermal vias, the copper planes and the heat sink design. Since the effects of various design parameters are analytically related, the required thermal correlation can be attained for different system designs. Finite element models are developed in this session for result validation. According to the specification of Philips Lumileds Rebel?LED, the efficacy can be related to the power input and the maximum chip (junction) temperature. Incorporating with the lumped model the correlation between the efficacy and the pitch can be set up. Since the LED operates within a range of power input and it fails as it exceeds the allowable junction temperature, there exist upper and lower bounds for the spacing. An optimum design thus guaranteed based on a user added-on criterion. As a demonstration, the methodology is applied to calculate the optimum spacing for the design of a LED tube, which is aimed to attain the lowest cost while with the optical performance maintained as the same as a 4ft T8 fluorescent lamp.|
|S. W. Ricky Lee, Senior Technical Officer