Guarantee-Reliability-with-Vibration-Simulation-an

GuaranteeingReliabilitywithVibrationSimulationandTestingDr.NathanBlattau©2004-2007©2004-2010oNathanBlattau,Ph.D.-SeniorVicePresidentHasbeeninvolvedinthepackagingandreliabilityofelectronicequipmentformorethantenyears.Hisspecialtiesincludebestpracticesindesignforreliability,robustnessofPb-free,failureanalysis,acceleratedtestplandevelopment,finiteelementanalysis,solderjointreliability,fracture,andfatiguemechanicsofmaterials..9000VirginiaManorRd.Ste.290,BeltsvilleMD20705|301-474-0607|©2004-2007©2004-2010oHighcyclefatigue(HCF)duetomechanicalstressinducedbyvibrationoMillionsofcyclestofailureoSmallchangesinstresshavelargeimpactsontimetofailureoAccordingtoU.SAir-Forcestatistics20percentofallfailuresobservedinelectronicequipmentareduetovibrationproblemsVibrationFatigueSteinbergD.S.Vibrationanalysisforelectronicequipment.JohnWiley&Sons,2000.3©2004-2007©2004-20104ReduceCostsbyImprovingReliabilityUpfrontDesigninginReliability,EarlierisCheaper4©2004-2007©2004-20105LifetimeundermechanicalcyclingisdividedintotworegimesLowcyclefatigue(LCF)Highcyclefatigue(HCF)LCFisdrivenbyinelasticstrain(Coffin-Manson)HCFisdrivenbyelasticstrain(Basquin)VibrationFatiguebffeNE2cffpN2-0.5c-0.7;1.4-1/c2-0.05b-0.12;8-1/b205©2004-2007©2004-2010Vibration–HarmonicandRandomoSinglefrequencyoRandomvibrationisacontinuousspectrumoffrequenciesANINTRODUCTIONTORANDOMVIBRATION–TomIrvineMIL-STD-810G6©2004-2007©2004-2010oExposuretovibrationloadscanresultinhighlyvariableresultsoVibrationloadscanvarybyordersofmagnitude(e.g.,0.001g2/Hzto1g2/Hz)oTimetofailureisverysensitivetovibrationloads(tfW4)oVerybroadrangeofvibrationenvironmentsoMIL-STD-810lists3manufacturingcategories,8transportationcategories,12operationalcategories,and2supplementalcategoriesMechanicalLoads(Vibration)7©2004-2007©2004-2010TypicalVibrationLevelsHarmonicSteinbergD.S.Vibrationanalysisforelectronicequipment.JohnWiley&Sons,2000.RandomMIL-STD-810GFigure514.6C-1USHighwaytruckvibrationexposure1hourisequivalentto1000miles8©2004-2007©2004-2010oRandomVibrationo9.8to28Grmso0.07to0.5G2/HzoNaturalFrequencyo72HzoResultsoWithBGA’s,SnPbsolderalwaysoutperformedlead-freeoLessconclusiveforleadless/leadedpartsExcessiveVibrationAllBGA-225Woodrow,IPC/APEX20069©2004-2007©2004-2010FatigueExponent–HighCycleFatigue?cfieldtesttestfieldNNoAssumethesolderstrainisdirectlyproportionaltotheboardlevelstrainoCoffin–Manson10Fittodataprovidesc=1.5Exponentvalueistoolow;representativeoflow-cyclefatigueHigh-cyclefatigueexponentistypically4to6orhigherMIL-STD-810,SteinbergLowcyclefatiguebehaviorcannotbeextrapolatedtoHCFbehavior©2004-2007©2004-2010oHCFfailurestypicallyoccurintheleadorsolderjointThroughHoleSolderInterconnectVibrationFatigueComponentMotionBoardbending11©2004-2007©2004-2010SurfaceMountVibrationFatigue–BoardBendingSnPbfailure@1200µεSnAgCufailure@1200µε12©2004-2007©2004-2010SnPbFatigueCrackoVeryfinecracksoSecondarymicro-cracksoEvidenceofphasecoarsening13©2004-2007©2004-2010SnAgCuFatigueCrackWelldefinedcrackpathShrinkagecrackprovidedcrackinitiationsiteShrinkageCrack14UsingSimulation©2004-2007©2004-2010SherlockDesignAnalysisSoftware16©2004-2007©2004-2010PartsList17©2004-2007©2004-2010BoundaryConditions(MountPoints)18©2004-2007©2004-2010BoardProperties(Stackup)19©2004-2007©2004-2010TestVehiclesandModelingFiniteElementAnalysisinSherlock20©2004-2007©2004-2010o1stNaturalFrequency:159.45HzSherlock-ModalAnalysis21©2004-2007©2004-2010ModalAnalysisOneaccelerometerlocatedatthecentermisses3frequencymodes22©2004-2007©2004-2010oBoardlevelstrain(atcomponents)oGreen:predicted(peakvalues)oPurple:experimental(averagedoverstraingageregion)BoardStrainAnalysis23©2004-2007©2004-2010BoardorComponentMotion-FEA24©2004-2007©2004-2010oSimulationoIdentifycriticalcomponentspriorduringplandevelopmentoIdentifycriticalfrequenciesorboardresponsesoIdentifylocationsforaccelerometerplacementoMountingconsiderationoBoundaryconditionsoMountingconfigurationsSimulationDuringTestPlanDevelopment25©2004-2007©2004-2010oHALTchambersutilizerepetitiveshock(RS)togeneratevibrationinputoTypeofrandomvibrationoPneumatichammersstrikethechambertableageneratethevibrationoHALTvibrationisnotdesignedtoreplicatefieldenvironmentsoDesignedtoexposeweaklinksoRapidassessmentoInput/OutputistypicallydisplayedasaGrmsvalueoWhatdoesthismean?oIsitsuitableforsimulation,canitbeusedinSherlock?HALTExample-Vibration26©2004-2007©2004-2010oHALTvibrationisspecifiedbyaGrmsleveloWhatdoesthisrepresentoTherootmeansquareacceleration(Grms)isthesquarerootoftheareaundertheASD(accelerationspectraldensity)curveinthefrequencydomainoTheGrmsvalueistypicallyusedtoexpresstheoverallenergyofaparticularrandomvibrationeventoCanitbeusedforsimulation?oNooAstraightGrmsvaluecanrepresentaninfinitenumberofaccelerationandfrequencycombinationsoThetimehistoryofthetableexcitationmustbecapturedandprocessedintoausableformatforthesimulationHALTChamberSettings27©2004-2007©2004-2010HALT-RepetitiveShockInducedVibrationInputHALTlevelsetto33Grms28©2004-2007©2004-2010Vibratio