SPCConceptsFundamentalSPCConcept

ShewhartControlChart

IntroductionAverageandRangeControlChartAverageandMovingRangeControlChartWesternRules

Cp/CpkIntroduction

RealAppliedContentsQualityManagementProgressSPCSTATISTICALPROCESSCONTROLAstructuredmathematicalmethodofanalyzingdataornumbersAsetofconditionsorcauseswhichworktogethertoproduceagivenresultTomakeaprocessbehavethewaywewantittobehaveWhatisSPC?VariousDefinitionsofSPC:

ControllingProcessesusingStatisticalTechniquesorMethodsItistheuseofStatisticalControlChartstomonitorandcontrolmanufacturingoperationsinrealtimeItistheuseofControlCharts&StatisticalTechniquestopreventaprocessfromgoingoutofcontrol…whichaffectstheproductyield…aswellas…quality/reliabilitylevelsItistheeliminationofunnaturalvariancetoimprovetheprocess[SpecialCause]SPCis...SPCis…

amethod

ofusingstatisticsto

identifypatterns

andtrends

inordertocontroltheprocess

DataCollectionProcessStatisticalMethodAbnormalStableProcessAnalysisActionsContinuousImprovementWhyarewedoingSPC?ImproveQuality/YieldsIncreaseProductLifeandReliabilityGreaterCustomerSatisfactionGovernment&CustomerRequirement

TobeMoreProfitableSPCPhilosophy..IImproveQualitybyremovingthecausesofproblemsinthesystem..inevitably..leadstoimprovedproductivityThepersondoingthejobismostknowledgeableaboutthejobPeoplewanttobeinvolvedanddotheirjobswellEverypersonwantstofeellikeavaluedcontributorSPCPhilosophy..IIAstructuredproblemsolvingprocessusingstatisticalmethodsandgraphicaltechniques…proceduresbettersolutionsthaninanunstructuredprocessGraphicalproblemsolvingtechniquesletyouknow:

?therelativeimportanceofproblemstobesolved?whetherthechangesmadehavehadthedesiredimpactAdversarialrelationshipbetweenlabor&managementiscounterproductive&outmodedEveryOrganizationhasundiscovered“gems”waitingtobedevelopedProductDesignProcessCustomerneedsQualityResponseNominalvalueConceptualFrameworkforSPCTheoriginsandnaturalofvariabilityConceptualFrameworkforSPCTheoriginsandnaturalofvariabilityManufacturingProcessRawMaterialToolConditionQualityResponseConceptualFrameworkforSPCTheoriginsandnaturalofvariabilityProductuseinthefieldTemperaturePowerQualityResponseConceptualFrameworkforSPCTheoriginsandnaturalofvariabilityProductuseOverTimeWearAgingQualityResponseProcessEvaluationOverTimeAccuracy&PrecisionAccuracyishowwellnumberscenteronyourtarget.Precisionishowalikethenumbersare.AccuratePreciseYesNoYesNo**********************************ProcessEvaluationOverTimeAccuracy&PrecisionTargetTargetHighAccuracy&highPrecisionLowAccuracy&highPrecisionHighAccuracy&lowPrecisionLowAccuracy&lowPrecisionProcessEvaluationOverTimeTimeProcessinstatisticalcontroloneuniquecausalsystemProcessEvaluationOverTimeTimeProcessintransitioninsearchofacausalsystemProcessEvaluationOverTimeTimeProcessnotinstatisticalcontrolpopulationchangesovertimeControlChart--BasicConceptOutputDataOutputDataOutputDataOutputDataOutputDatat1t3t4t5t2ProcessMeanUpperControlLimitLowerControlLimit☆將製程之之品質(zhì)特特性以圖圖形之方方式表示示，把量測所所得之資資料點繪繪製於以以時間為為橫軸之之圖形上上。管制圖必必須包括括中心線線(CenterLine)、、管制界限限(ControlLimit)，，並應(yīng)用管管制規(guī)則則(ControlRules)判斷之。ControlChart--BasicConceptLCLUCLCLControlChart--BasicConceptUSLLSLUCLLCLUSLLSLUCLLCLUSLLSLUCLLCLUCLLCLInSpecificationOutofSpecificationIncontrolOutofcontrolUSLLSLControlChart--BasicConceptAllprocessesaresubjecttotwofundamentaltypesoffaultsorproblems.Theyare:類別機遇原因(一般原因)ChanceCause(CommonCause)可歸屬原因(特殊原因)AssignableCause(SpecialCause)特性1.包括許多隨機存在之個別原因。

2.任何單一機遇原因僅導(dǎo)致微量變異(但若許多機遇原因彙總在一起，可能產(chǎn)生頗大之影響)。

1.包含一個或少數(shù)偶爾發(fā)生之原因。2.

任何單一可歸屬原因均可造成大量之製程變異。可能原因機器之微震原料之略微差異作業(yè)員之錯誤機器不正確之設(shè)定不合格原料

解說1.

消除製程中之機遇變異不符合經(jīng)濟(jì)原則。2.

當(dāng)僅有機遇變異出現(xiàn)時，製程處於可接受水準(zhǔn)；倘若仍有不合格品產(chǎn)生，則必需進(jìn)行基本製程改變或修定規(guī)格，以減少不合格品。3.

當(dāng)觀測值在管制界限內(nèi)時，表示製程不應(yīng)調(diào)整。4.當(dāng)僅有機遇變異時，製程相同穩(wěn)定，可以用抽樣程序預(yù)測產(chǎn)品品質(zhì)。

1.

可歸屬變異可以被偵測出；消除該變異通常均合乎經(jīng)濟(jì)原則。2.

當(dāng)有可歸屬原因出現(xiàn)時，製程未在可接受水準(zhǔn)。3.

當(dāng)觀測值超出管制界限時，通常表示製程應(yīng)予以調(diào)整和矯正。4.當(dāng)有可歸屬原因出現(xiàn)時，製程不夠穩(wěn)定，不宜以抽樣程序預(yù)測產(chǎn)品品質(zhì)。

☆DEFINITION:It’sasnapshotofyourprocessovertimeprovidingfurtherinsightastotheconsistencyandqualityofthatprocess.ControlChart--BasicConcept☆PURPOSE:AControlChartisatoolwhichallowsyoutodetermineifthe““Process”haschanged.Theupperandlowercontrollimits(whicharesetstatistically)furtherallowyoutodetermine,withadegreeofcertainty,thatthechangeisarealchange;i.e.,statisticallysignificantchange.ShewhartControlChart☆ShewhartControlChartThinking:InthecontextoftheShewhartcontrolchartmethodwhatwedoiscollectanamountofdata(say,ksamplesofsizeneach)thatwethinkissufficienttoallowustotreattheaverageofallthesedataasifitweretheactualprocessmean.☆ControlChartsisintroducedtoIndustrybyWalterShewhartinthe1920’’sShewhartControlChart☆計量量值(VariableData/MeasurableData)種類適用狀況

平均值與全距管制圖?可用以管制分組之計量數(shù)據(jù)，即每次同時取得幾組數(shù)據(jù)之工程，如長度、濃度、成份等?是掌握工程狀態(tài)最有效之管制圖

平均值與標(biāo)準(zhǔn)差管制圖?與相同，用於subgroupsize≧10時

中位數(shù)與全距管制圖?與相同，但檢出力較差個別值與移動全距管制圖?適用於找出不同subgroup間之變化時

ShewhartControlChart種類適用狀況

P-Chart不良率管制圖?其數(shù)據(jù)有兩種類別，如合格/不合格、通過/不通過等

NP-Chart不良數(shù)管制圖?與P-Chart相同，每組樣本需相同C-Chart

缺點數(shù)管制圖?每單位數(shù)內(nèi)缺點發(fā)生之次數(shù)U-Chart

單位缺點數(shù)管制圖?與C-Chart相同，但樣本大小不同☆計數(shù)數(shù)值(AttributeData/CountableData)☆管制制圖圖之之選選擇擇方方式式：請參參閱閱QS9000SPCManual,AppendixCConstructionControlChartDataCollectionRationalsamples(unbiasedandrandom)GroupsmeasurementVariationisattributableonlytoaconstantsystemofcommoncause.MinimizetheoccurrenceofspecialcauseswithinsampleCalculationNecessaryItemsControllimit(20~25subgroupsofdataarerequired)CentrallineEstablishControlChartProcessmonitorAlarmsystemDetectspecialcauseAverageandRangeControlChartTherearetwoprocesscharacteristicsthataregeneralinterestinthestatisticalprocesscontrol:ThemeanleveloftheprocessTheamountofvariationintheprocessAverageChartRangeChartTointerpretthepatternsinthecontrolchartAlwaysBeginwithRChart

X-RChartAverageandRangeControlChartSampleSize:nTotal:kSubgroupsAverageandRangeControlChartXChart:Step1:CalculateMeanandRangeofeachSubgroup

Where:MeanistheaveragevalueofeachsubgroupRangeisthedifferencebetweenMaxandMinvalueineachsubgroupMeanRangeAverageandRangeControlChartXChart:Step2:CalculateCenterline、UpperandLowercontrollimitasfollows:Centerline:Where:kisthenumberofsubgroupsisthemeanvalueofi’thsubgroupisthearithmeticaverageofallavailablesampleaverageAverageandRangeControlChartXChart:UpperandLowercontrollimitUpperControlLimit:LowerControlLimit:Where:RistheaverageofthesamplerangesAisaconstantwhichvaryingsamplesizenandtabulated2AverageandRangeControlChartRChart:Where:

kisthenumberofsubgroups

istheRangeofi’thsubgroup

isthearithmeticaverageofallavailablesamplerange

CalculateCenterline、

UpperandLowercontrollimit

asfollows:Centerline:AverageandRangeControlChartRChart:UpperandLowercontrollimitUpperControlLimit:LowerControlLimit:Where:RistheaverageofthesamplerangesDandDisaconstantwhichvaryingsamplesizenandtabulated34AverageandRangeControlChart

23.7601.8800.0003.2671.1280.0003.2672.6601.88032.3941.0230.0002.5681.6930.0002.5751.7721.18741.8800.7290.0002.2662.0590.0002.2821.4570.79651.5960.5770.0002.0892.3260.0002.1151.2900.691

61.4100.4830.0301.9702.5340.0002.0041.1840.54971.2770.4190.1181.8822.7040.0761.9241.1090.50981.1750.3730.1851.8152.8470.1361.8641.0540.43291.0940.3370.2391.7612.9700.1841.8161.0100.412101.0280.3080.2841.7163.0780.2231.7770.9750.363nThefollowingtablearetheconstantsfordeterminingfromRthe3SigmaControlLimitsforControlChartsandforestimatingtheprocessStandardDeviationfromRAverageandRangeControlChartCalculation:MeanRangeCenterline:=2.795Centerline:AverageChart=1.808RangeChart===AverageandRangeControlChartCalculation:XChart:UpperControlLimit:LowerControlLimit:RChart:UpperControlLimit:LowerControlLimit:AverageandRangeControlChartX-RChartRepresentationAverageandRangeControlChartOBJECTIVE☆Revealnatureofaprocess☆Revealsignificantchangeinprocess☆EarlydetectionofprocessingproblemsAverageandMovingRangeControlChartCalculation:MeanMovingRangeCenterline:=2.795Centerline:AverageChart=0.674MRChart===MovingRange()=Calculation:XChart:UpperControlLimit:LowerControlLimit:MRChart:UpperControlLimit:LowerControlLimit:AverageandMovingRangeControlChartAverageandMovingRangeControlChartX-MRChartRepresentationSpecialNotesManytimeswehaveprocesseswherethemajorvariationisnotWITHINthenormalsubgroup…butisBETWEENthenormalsubgroups.ThisisespeciallytrueinbatchoperationssuchasEpi,Diffusion,SolderDip,etc.Inthesecases,X&RorX&s,whicharebasedonWITHINsubgroupvariation,arenotadequateandwemustuseMovingRangeCharts(MRChart).AverageandMovingRangeControlChartWesternRules(ZoneRules)TestForInstabilityProcessMeanUCL(UpperControlLimit)LCL(LowerControlLimit)AABBCCZoneA?1/40ofpointswillbehere+1σX+2σ-1σ-2σ-3σ+3σZoneA?1/40ofpointswillbehereZoneB?1/7ofpointswillbehereZoneB?1/7ofpointswillbehereZoneC?1/3ofpointswillbeherePointOutsidePointOutsideZoneC?1/3ofpointswillbehere☆I(lǐng)nApplyingWesternRules…consideronly1/2ofcontrolchartatatime…i.e.,upperhalforlowerhalf☆Eachzoneisonesigma(σ)WesternRules(ZoneRules)Test1:ExtremePoints/OutofControl☆

AppliesbothforXandR

controlchart

☆

ProcessisOUTOFCONTROL…ifonepointfallsoutsidethecontrollimit(s)☆

SPECIALCAUSESarePRESENTatTHISPOINT

IMMEDIATEANALYSISoftheoperationisrequiredMarkOutofControlPoint(s)withandtheRuleNumberwhichhasbeenviolated.ProcessMeanUpperControlLimitLowerControlLimitAABBCC1WesternRules(ZoneRules)Test2:TwooutofThreePointsinZoneAorBeyond☆

AppliesforXchart

only☆

ProcessmaybeOUTOFCONTROL…SPECIALCAUSESmaybePresentProcessMeanUpperControlLimitLowerControlLimitAABBCC1,22WesternRules(ZoneRules)Test3:FouroutofFivePointsinZoneBorBeyond☆

AppliesforXchart

only☆

ProcessmaybeOUTOFCONTROLProcessMeanUpperControlLimitLowerControlLimitAABBCC3331WesternRules(ZoneRules)Test4:Successive8PointsAboveorBelowtheCenterline☆

AppliesbothforXandR

controlchart

☆

ProcessmaybeOUTOFCONTROLProcessMeanUpperControlLimitLowerControlLimitAABBCC4WesternRules(ZoneRules)Test5:LinearTrendIdentification☆

AppliesbothforXandR

controlchart

☆

ProcessmaybeOUTOFCONTROLifsevensuccessivepointsshowacontinuingincreaseordecrease,asystematictrendintheprocessissignaled.ProcessMeanUpperControlLimitLowerControlLimitAABBCC55WesternRules(ZoneRules)Test6:OscillatoryTrendIdentification☆

AppliesbothforXandR

controlchart

☆

ProcessmaybeOUTOFCONTROLif14successivepointsoscillateupanddown,asystematictrendintheprocessissignaled.ProcessMeanUpperControlLimitLowerControlLimitAABBCC6WesternRules(ZoneRules)Test7:AvoidanceofZoneCTest☆

AppliesforXchart

only☆

ProcessmaybeOUTOFCONTROLif8successivepointsoccurringoneithersideofthecenterlineavoidzoneCProcessMeanUpperControlLimitLowerControlLimitABCABC7WesternRules(ZoneRules)Test8:RuninZoneCTest☆

AppliesforXchart

only☆

ProcessmaybeOUTOFCONTROLif15successivepointsfallinzoneConly,to