ThermalanalysisofLEDarray精Word格式.docx

ThermalanalysisofLEDarray精Word格式.docx

《ThermalanalysisofLEDarray精Word格式.docx》由会员分享，可在线阅读，更多相关《ThermalanalysisofLEDarray精Word格式.docx（12页珍藏版）》请在冰豆网上搜索。

Abstract

ThispaperreportsonthermalcharacterizationofhighpowerLEDarrays.Thermaltransientmethodsareusedtomeasurethejunctiontemperatureandcalculatethethermalresistance.TheemphasisisplacedupontheinvestigationofjunctiontemperatureriseofLEDarrayforalimitedrangeofboundaryconditionswhichincludedesigneffectofheatpipe,convectioncondition,andambienttemperature.ThejunctiontemperaturesofLEDarraywithandwithoutheatpipeatthesameairvelocityof7m/swere87.6◦C,and63.3◦C,respectively.ThecorrespondingthermalresistancesofLEDarrayweremeasuredtobe1.8K/Wand2.71K/W.ItwasfoundoutthatthemeasuredjunctiontemperaturesandthermalresistanceofLEDarrayareincreasedwiththeinputpowerandambienttemperature,butdecreasedwiththeairvelocity.Ananalyticalthermalmodelanalogouswithanequivalentparallelcircuitsystemwasproposedandwasverifiedbycomparisonwithexperimentaldata.

©

2006ElsevierB.V.Allrightsreserved.

Keywords:

LEDarray;

Heatpipe;

Junctiontemperature;

Thermalresistance

1.Introduction

Lightemittingdiode（LEDisasolidstatesemiconductordevicethatconvertselectricalenergyintolight.LEDsdemon-strateanumberofbenefitscomparedtotraditionalincandescentlamps.Nowadays,highpowerLEDsarebeinginvestigatedasreplacementsforcoldcathodefluorescentlamp（CCFLintheLCDdisplaybacklightsandheadlightlampforautomobiles[1,2].WithfurtherimprovementLEDshaveagreatpotentialtobecomeanewilluminationsource.However,therealchallengeisthatthelifetimeofLEDsstillbeeasilyshortenedbyheat;

notonlybytheheatfromambientbutalsobytheheatgeneratedwithintheLEDitself[3,4].Inaddition,theperformanceofunitLEDisknowntosignificantlydependonthesystemswheretheLEDpackagesarelaiddown.Inanextremecase,thelifetimeofunitLEDpackagewithanexcellentthermalperformancecanbeveryshortifthesystemaroundithasapoorthermaldesign.Therefore,effectivethermaldesignandreliablethermalcharacterizationofLEDsystemareimportantfortheunitLEDpackage.ThermalcharacterizationofLEDsinanarrayisverydifferentfromthatofsingleLEDpackage.Thejunctiontemper-atureofLEDarraywillbesignificantlyinfluencedbyambienttemperatureandsideeffectfrommultiplechips.Itisgener-allyknownthatthermalbehaviorofLEDarrayisaffectedby∗Correspondingauthor.Tel.:

+82313306465;

fax:

+82313306457.

E-mailaddress:

mwshin@mju.ac.kr（M.W.Shin.morefactorsthaninthecaseofunitpackage.TherehavebeenseveralreportsonthermalcharacterizationofLEDpackageswithasinglechip[5,6].However,therehavebeennoreportsonthethermalanalysisofLEDarraysystemsofartothebestknowledgeoftheauthors.

Inthispaper,thermalbehaviorofLEDarraysystemisreported.Testchipsarewidelyusedtopredictthejunctiontem-peratureofarraysystemwithelectronicdevices（CPU,CMOS,etc.[7,8].However,extraspecialtestchipsarenotusedinourexperimentandLEDitselfisusedasatestchip.Themethodcanreduceseveralmeasuringparameterswhichcanmisleadarealjunctiontemperaturewithoutdestructingitselectricalcircuits.

Thermaltransientmeasurementwasdoneusingtheso-calledstructurefunction[9].ThermalcharacteristicsofLEDarraywithheatpipeandwithoutheatpipearecomparedunderdifferentambienttemperaturesandforcedconvectionconditions.

2.Theoreticalbackground

AtJEDECStandardNo.51-1,thermalresistanceofasinglesemiconductordeviceisdefinedas:

RJX=

TJ−TX

PH（1whereRJXisthethermalresistancebetweendevicejunctionandthespecificenvironment,TJthejunctiontemperatureofdeviceinasteadystatecondition,TXthereferencetemperatureforthe

0040-6031/$–seefrontmatter©

2006ElsevierB.V.Allrightsreserved.doi:

10.1016/j.tca.2006.11.031

22L.Kimetal./ThermochimicaActa455（200721–25specificenvironment,andPHthepowerdissipationinthedevice

[10].

Theequationisforasinglechippackage.Thermalresistance

ofLEDarrayswhichhavemultipleheatsourcescanbedescribed

asthefollowingrelationusingtheaveragejunctiontemperature

ofLEDarray,Tj,avg.

θja-avg=Tj,avg−Tamb

P（2

whereθja-avgisanaveragejunctiontoambientthermalresis-tance,Pisthepowerdissipationoftheentirepackages,andTambistheambienttemperature.TheequationassumesthateachLEDmountedonthearrayexhibitsthesamethermalchar-acteristics.BecausetheLEDsusedinthisexperimentareofidenticalgeometryandpowerdissipation,theemploymentoftheaboveequationisvalidinouranalysis.Totalpowerdissipa-tioniscalculatedbythemeasuredvoltageandtheinputcurrent.TemperaturerisecanbeinterpretedbythechangeofvoltagedropinafollowingwayforLEDandtheslopeisknownasaKfactor[10];

slope=dVF

dTJ

（3

wheredVFisthedifferentialofforwardbiasvoltage,anddTJisthedifferentialofjunctiontemperature.

ExpandingthistheorytoaseriesofmultipleLEDsleadstothefollowingexpressionformodifiedslope:

slopetotal=dVFtotal

dT

=

ndVF

=n·

slope（4

Eq.（4indicatesthattheKfactorfortheLEDarrayiseasilydefinedfromtheslopeforaunitLEDpackage.TheslopetotalforthearraysystemisntimesoftheslopeforasingleLEDpackage,slope.ForanLEDarray,then·

slopeisaconstant,sothetotalforwardvoltageoftheLEDarraycanbeusedasatemperaturesensitiveparameter（TSP.

3.Experiments

CommercialGaN-basedLEDscoatedwithyellowphosphor（LuxeonVwereusedforthefabricationofarrayinthisexperi-ment.LEDarrayswerepreparedeitherwithorwithoutheatpipe.LEDarrayiscomposedofsixhighpowerLEDsandmountedon5cm×

7.5cmmetalcoreprintedcircuitboard（MCPCBwitha2.5cmpitch.Fig.1showstheschematicstructureofLEDarraymountedonMCPCBandarraysystemwithheatpipe.

Thediameteroftheheatpipeis1.27cmandthelengthis30cm.LEDsinthearrayareelectricallyconnectedinseriesandsimultaneouslydriven.Sensorcurrentof20mAwasusedtodetecttheforwardvoltageofthearray.Measurementswerecarriedoutbyathermaltransienttester（T3ster®

.Thetheo-reticalframeworkoftheevaluationoftheT3sterisbasedonarepresentationofthedistributedRCnetworks.Thestructurefunctionsareobtainedbydirectmathematicaltransformationsfromthecoolingcurve.Afteracalibrationprocess,whichdeter-minestheratiobetweenthetemperatureandtheforwardvoltagedropasatemperaturesensitiveparameter（TSP,cooling

curve

Fig.1.SchematicstructureofanLEDarraymountedonMCPCB.wasobtained.ThesizeofAlchamberusedinthisexperimentis800mm×

140mm×

100mm.Theflowrateofcoolantinthechamberwasoptimizedsothattheambienttemperatureofthesampleswaskeptconstantduringthemeasurement.

4.Resultsanddiscussions

Fig.2istheforwardvoltageversustemperatureplotobtainedfromtheLEDarraywithoutheatpipe.ThelinearitybetweenthevoltagedropandtemperatureistheKfactor.TheKfactorofthearrayis0.01969V/◦Catthesensorcurrentof20mA.

TheKfactorofLEDarrayissixtimesofoneLED,becausetherearesixLEDsconnectedinseries.Fig.3representsthederivativeofthermalcapacitanceasafunctionofthermalresistancefortheLEDarraywithoutheatpipeunderseveralconvectionconditions.Thermalcapacitancevariesdirectlywithbothspecificheatandmass;

itisthequantityofheatabsorbedbythesamplewhenitstemperaturerises1◦C.Thepeaksimplythematerialtransitionsintheheatflowpath.

Thetotalthermalresistanceofarrayisfoundtodecreasewiththevelocityofairflow.Thethermalresistanceisabout6.8◦C/Watthenaturalconvectionstateandabout2.8◦C/Watanairvelocityof7m/s.Atanairvelocityabove2m/s,thethermalresistanceofarraydropsrapidly.Itisworthwhilecon-sideringthedifferenceinthermalresistancebetweenthe

unitFig.2.Forwardvoltagevs.temperatureplotshowingaKfactorofLEDarray.

L.Kimetal./ThermochimicaActa455（200721–25

23

Fig.3.DifferentialstructurefunctionsofLEDarrayasafunctionofconvection

condition（withoutheatpipes.

LEDpackage（8◦C/W[11]andthevalueof1.8◦C/Wwhichwasmeasuredinthisexperiment.ThedeviationcanbewellexplainedbyconsideringanequivalentthermalcircuitofLEDarrayaswasdescribedinFig.4.

BecauseMCPCBisone,thusthetemperatureoftheMCPCBisassumedtobethesame.ThetemperaturedifferencebetweenthejunctionandtheambientisexpressedforeachLEDchipasTJi,a=TJi,MCPCB+TMCPCB,a

=Piθi+6

k=1

Pkθ0=Piθi+Ptotalθ0

（5

whereTJi,aisthetemperaturedifferencebetweentheithchipandambient,TJi,MCPCBthetemperaturedifferencebetweentheithchipandtheslug,TMCPCB,athetemperaturedifferencebetweentheMCPCBandtheambient,Pitheinputpoweroftheithchip,θithepartialthermalresistancebetweentheithchipandtheMCPCB,andθ0isthepartialthermalresistancebetweentheMCPCBandambient.Becausetheheatisgeneratedat

the

Fig.4.EquivalentcircuitofLEDarraysinvestigatedinthisstudy.

junctionandflowstotheenvironmentthroughtheMCPCB,PtotalisthesumofPi.DefiningTJ,avg=Tj,avg−Jamb,letusassumethatTJ,avg=TJi,a（i=1–6.ApplyingEq.（2intoEq.（5,leadstothetotalthermalresistanceasfollows:

θja-avg=

TJi,aPtotal=Piθi+Ptotalθ0

Ptotal（6ApplyingEq.（6intotheLEDarraysystemwithsixLEDpack-ages,itissimplifiedasθja-avg=

（1/6Ptotalθi+Ptotalθ0total=1