ThermalanalysisofLEDarray精.docx

1、ThermalanalysisofLEDarray精 Thermochimica Acta455(2007 2125Thermal analysis of LED array system with heat pipeLan Kim,Jong Hwa Choi,Sun Ho Jang,Moo Whan ShinDepartment of Materials Science&Engineering,Myong Ji University,38-2Nam-Dong,Yongin,Kyunggi449-728,Republic of KoreaAvailable online3December200

2、6AbstractThis paper reports on thermal characterization of high power LED arrays.Thermal transient methods are used to measure the junction temperature and calculate the thermal resistance.The emphasis is placed upon the investigation of junction temperature rise of LED array for a limited range of

3、boundary conditions which include design effect of heat pipe,convection condition,and ambient temperature.The junction temperatures of LED array with and without heat pipe at the same air velocity of7m/s were87.6C,and63.3C,respectively.The corresponding thermal resistances of LED array were measured

4、 to be1.8K/W and2.71K/W.It was found out that the measured junction temperatures and thermal resistance of LED array are increased with the input power and ambient temperature,but decreased with the air velocity.An analytical thermal model analogous with an equivalent parallel circuit system was pro

5、posed and was veried by comparison with experimental data.2006Elsevier B.V.All rights reserved.Keywords:LED array;Heat pipe;Junction temperature;Thermal resistance1.IntroductionLight emitting diode(LEDis a solid state semiconductor device that converts electrical energy into light.LEDs demon-strate

6、a number of benets compared to traditional incandescent lamps.Nowadays,high power LEDs are being investigated as replacements for cold cathodeuorescent lamp(CCFLin the LCD display backlights and head light lamp for automobiles 1,2.With further improvement LEDs have a great potential to become a new

7、illumination source.However,the real challenge is that the life time of LEDs still be easily shortened by heat; not only by the heat from ambient but also by the heat generated within the LED itself3,4.In addition,the performance of unit LED is known to signicantly depend on the systems where the LE

8、D packages are laid down.In an extreme case,the life time of unit LED package with an excellent thermal performance can be very short if the system around it has a poor thermal design.Therefore,effective thermal design and reliable thermal characterization of LED system are important for the unit LE

9、D package.Thermal characterization of LEDs in an array is very different from that of single LED package.The junction temper-ature of LED array will be signicantly inuenced by ambient temperature and side effect from multiple chips.It is gener-ally known that thermal behavior of LED array is affecte

10、d by Corresponding author.Tel.:+82313306465;fax:+82313306457.E-mail address:mwshinmju.ac.kr(M.W.Shin.more factors than in the case of unit package.There have been several reports on thermal characterization of LED packages with a single chip5,6.However,there have been no reports on the thermal analy

11、sis of LED array system so far to the best knowledge of the authors.In this paper,thermal behavior of LED array system is reported.Test chips are widely used to predict the junction tem-perature of array system with electronic devices(CPU,CMOS, etc.7,8.However,extra special test chips are not used i

12、n our experiment and LED itself is used as a test chip.The method can reduce several measuring parameters which can mislead a real junction temperature without destructing its electrical circuits.Thermal transient measurement was done using the so-called structure function9.Thermal characteristics o

13、f LED array with heat pipe and without heat pipe are compared under different ambient temperatures and forced convection conditions.2.Theoretical backgroundAt JEDEC Standard No.51-1,thermal resistance of a single semiconductor device is dened as:R JX=T JT XP H(1 where R JX is the thermal resistance

14、between device junction and the specic environment,T J the junction temperature of device in a steady state condition,T X the reference temperature for the0040-6031/$see front matter2006Elsevier B.V.All rights reserved. doi:10.1016/j.tca.2006.11.03122L.Kim et al./Thermochimica Acta455(20072125 speci

15、c environment,and P H the power dissipation in the device10.The equation is for a single chip package.Thermal resistanceof LED arrays which have multiple heat sources can be describedas the following relation using the average junction temperatureof LED array,T j,avg.ja-avg=T j,avgT ambP(2whereja-av

16、g is an average junction to ambient thermal resis-tance,P is the power dissipation of the entire packages,and T amb is the ambient temperature.The equation assumes that each LED mounted on the array exhibits the same thermal char-acteristics.Because the LEDs used in this experiment are of identical

17、geometry and power dissipation,the employment of the above equation is valid in our analysis.Total power dissipa-tion is calculated by the measured voltage and the input current. Temperature rise can be interpreted by the change of voltage drop in a following way for LED and the slope is known as a

18、K factor10;slope=d V Fd T J(3where d V F is the differential of forward bias voltage,and d T J is the differential of junction temperature.Expanding this theory to a series of multiple LEDs leads to the following expression for modied slope:slope total=d V F totald T=n d V Fd T=nslope(4Eq.(4indicate

19、s that the K factor for the LED array is easily dened from the slope for a unit LED package.The slope total for the array system is n times of the slope for a single LED package,slope.For an LED array,the nslope is a constant, so the total forward voltage of the LED array can be used as a temperatur

20、e sensitive parameter(TSP.3.ExperimentsCommercial GaN-based LEDs coated with yellow phosphor (Luxeon Vwere used for the fabrication of array in this experi-ment.LED arrays were prepared either with or without heat pipe. LED array is composed of six high power LEDs and mounted on 5cm7.5cm metal core

21、printed circuit board(MCPCBwith a 2.5cm pitch.Fig.1shows the schematic structure of LED array mounted on MCPCB and array system with heat pipe.The diameter of the heat pipe is1.27cm and the length is 30cm.LEDs in the array are electrically connected in series and simultaneously driven.Sensor current

22、 of20mA was used to detect the forward voltage of the array.Measurements were carried out by a thermal transient tester(T3ster.The theo-retical framework of the evaluation of the T3ster is based on a representation of the distributed RC networks.The structure functions are obtained by direct mathema

23、tical transformations from the cooling curve.After a calibration process,which deter-mines the ratio between the temperature and the forward voltage drop as a temperature sensitive parameter(TSP,cooling curveFig.1.Schematic structure of an LED array mounted on MCPCB. was obtained.The size of Al cham

24、ber used in this experiment is800mm140mm100mm.Theow rate of coolant in the chamber was optimized so that the ambient temperature of the samples was kept constant during the measurement.4.Results and discussionsFig.2is the forward voltage versus temperature plot obtained from the LED array without he

25、at pipe.The linearity between the voltage drop and temperature is the K factor.The K factor of the array is0.01969V/C at the sensor current of20mA.The K factor of LED array is six times of one LED,because there are six LEDs connected in series.Fig.3represents the derivative of thermal capacitance as

26、 a function of thermal resistance for the LED array without heat pipe under several convection conditions.Thermal capacitance varies directly with both specic heat and mass;it is the quantity of heat absorbed by the sample when its temperature rises1C.The peaks imply the material transitions in the

27、heatow path.The total thermal resistance of array is found to decrease with the velocity of airow.The thermal resistance is about 6.8C/W at the natural convection state and about2.8C/W at an air velocity of7m/s.At an air velocity above2m/s,the thermal resistance of array drops rapidly.It is worth wh

28、ile con-sidering the difference in thermal resistance between the unit Fig.2.Forward voltage vs.temperature plot showing a K factor of LED array.L.Kim et al./Thermochimica Acta 455(2007212523 Fig.3.Differential structure functions of LED array as a function of convectioncondition (without heat pipes

29、.LED package (8C/W 11and the value of 1.8C/W which was measured in this experiment.The deviation can be well explained by considering an equivalent thermal circuit of LED array as was described in Fig.4.Because MCPCB is one,thus the temperature of the MCPCB is assumed to be the same.The temperature

30、difference between the junction and the ambient is expressed for each LED chip as T J i,a = T J i,MCPCB + T MCPCB ,a=P i i + 6k =1P k 0=P i i +P total 0(5where T J i ,a is the temperature difference between the i th chip and ambient, T J i ,MCPCB the temperature difference between the i th chip and

31、the slug, T MCPCB,a the temperature difference between the MCPCB and the ambient,P i the input power of the i th chip,i the partial thermal resistance between the i th chip and the MCPCB,and 0is the partial thermal resistance between the MCPCB and ambient.Because the heat is generated atthe Fig.4.Eq

32、uivalent circuit of LED arrays investigated in this study.junction and ows to the environment through the MCPCB,P total is the sum of P i .Dening T J,avg =T j,avg J amb ,let us assume that T J,avg = T J i ,a (i =16.Applying Eq.(2into Eq.(5,leads to the total thermal resistance as follows:ja-avg =T J i,a P total =P i i +P total 0P total (6Applying Eq.(6into the LED array system with six LED pack-ages,it is simplied as ja-avg =(1/6P total i +P total 0total =1