SPCforTestSystem§GaugeR&R1有多少個(gè)黑點(diǎn)？2這些橫線是不是平行的？3是靜的還是動(dòng)的？4三個(gè)正方形5幾何謎題6沒(méi)有適當(dāng)?shù)牧繙y(cè)工具就無(wú)法管理品質(zhì)!!7比較式絕對(duì)式待測(cè)量轉(zhuǎn)換訊息標(biāo)準(zhǔn)值顯示量測(cè)儀量測(cè)之種類待測(cè)量標(biāo)準(zhǔn)量轉(zhuǎn)換訊息轉(zhuǎn)換訊息比測(cè)儀顯示轉(zhuǎn)換放大8BasicConceptsEveryProcessproducea“Product”.EveryproductpossessesQualities(Features)EveryQualityFeatureCanbemeasuredTotalVariation=ProductVariation+MeasurementVariationSomeVariationInherentinSystemDesign.SomeVariationisduetoaFaultyperformanceoftheSystem(s).9ComponentsofMeasurementSystemEquipmentorGageTypeofGage:Attribute:go-nogo,visionsystemsVariable:calipers,probe,coordinatemeasurementmachinesUnitofMeasurement-discriminationAtleast1/10of6*processstandarddeviation(6s)Otherrules:atleast1/10oftoleranceOperator&OperatingInstructionsPartlocatingororientationschemeagagemustbeabletoconsistentlylocatethepartbeingmeasured.10TermsYouShouldKnow?Measurementsystem:

Thecollectionofoperations,procedures,gagesandpersonnelusetoobtainmeasurementsofworkpiececharacteristics?Gage:

Adeviceusedtoobtainmeasurements?Repeatability:

Repeatability(Precision)Thevariationinmeasurementobtainedwithonemeasurementinstrumentwhenusedbyoneoperatortomeasure.)anidenticalcharacteristiconthesamepart.(EV:EquipmentVariation?Reproducibility:

Thevariationintheaverageofthemeasurementsmadebydifferentoperatorsusingthesamemeasurementinstrumentwhenmeasuringidenticalcharacteristicsonthesamepart.(AV:AppraiserVariation)11TermsYouShouldKnow?Bias:

Thedifferencebetweentheobservedaverageofmeasurementsandthereferencevalue.?PercentR&R:

Thepercentageofprocessvariationrelatedtothemeasurementsystemforrepeatabilityandreproducibility?Stability:

Thetotalvariationinthemeasurementsobtainedwithameasurementsystemonthesamemasterorpartswhenmeasuringsinglecharacteristicsoveranextendedtimeperiod.?Linearity:

Thedifferenceinthebiasvaluesthroughtheexpectedoperatingrangeofthegage.121.AccuracyDifferencebetweenthetrueaverageandtheobservedaverage(trueaveragemaybeobtainedusingamoreprecisemeasuringtool)132.ReproducibilityVariationinaverageofmeasurementsmadebydifferentoperatorsusingthesamegagemeasuringthesamepart.AppraiserVariation(AV)143.Repeatability(Precision)Randomvariationinmeasurementswhenoneoperatorusesthesamegagetomeasurethesamepartseveraltimes.EquipmentVariation(EV)154.StabilityDifferenceintheaverageofatleast2setsofmeasurementsobtainedwithagageasaresultoftime.16PrecisionandAccuracy17AccuracyandStabilityMeasuresamepartsusingamorepreciseinstrument(Accuracy).

Measuresamepartsusingthesamegageatalaterdate(StabilityorDrift).18MeasurementSystemVariables19PossibleSourcesofProcess

Variation20SourcesofMeasurementErrorSystematicVariations1.accuracy:impropercalibrationofequipment2.reproducibility:differentoperatorsusingsameequipmentwithdifferenttechniquesRandomVariations3.repeatability:(forexample,inconsistentpartlocating)PeriodicVariations4.stability:wear,deterioration,environmentalconditions21TheGR&RDesignModelSupposemultipleoperators,measuremultiplepartsseveraltimeseach.Then:

Intermsofthegageperformance,between-operatorreproducibility

isduetotheoperatoreffectandtheoperator*partinteraction.Within-operator

repeatability

isfromthepureerrorterm.22TheComponentsofVariance23MeasurementError–GageR&R24AveragesANOVAanalysisbeginswiththecalculationofallrelevantaverages.Here’sanexample:25TypesofGageErrorandAnalysis

Methods:Accuracy–GageAccuracy(Bias)StudyGageLinearityStudy.Repeatability&ReproducibilityGageR&R(Methods:X-bar/RangeorANOVA)RepeatedMeasurementStudyRepeatability–Static&DynamicStudyStability–RepeatedMeasurements26VariancePartitioningThetotalvarianceoftheGR&Rdataispartitionedintocomponentsaccordingtothemodel.Variancesadd,soallpartitionedpiecessumtothetotaldatavariance!27TwoWayAnovaTable28GageR&RAcceptanceTodetermineacceptability,aratioistypicallyformedcomparinggagevariationtoeither:GageR&R%ofToleranceWidthGageR&R%ofTotalVariation6s-represents99.73%ofvaluesifnormaldistribution5.15s-represents99%ofvaluesifnormaldistribution29CriteriaforEvaluatingGageR&RGageR&R%-TVorR&R%-ToleranceifR&R%is:<10%=measurementsystemisacceptable10-30%=measurementsystemmaybeacceptablebaseduponimportanceofapplication,gagecost,costofrepairs..>30%=measurementsystemneedsimprovement30ConductingaGageR&RStudy1.Plantheapproach–Whatisnatureofoperatorandmeasurementinteraction?–Reproducibilitycanbenegligiblesometimes.2.Select3welltrainedoperators3.Select10partsfortest,Selectsamplepartsshouldrepresententirenormalproductionrange.(Prefercoverthedistributionof+/-4sigma)4.Repeatmeasureeachparts3times.5.Takemeasurementsinrandomorder.6.Takeblindmeasurement?Personmeasuringdoesn’tknowwhichpieceisbeingmeasured,whatpreviousmeasurementswere,orwhatotherappraisersfound.31ConductingaGageR&RStudy7.Recordallmeasurementsonworksheet.8.Calculatetrialrangeforeachpieceforeachoperatorandrecordvalues,calculateR&R(squarerootofthesumofthesquaresoftherepeatabilityandreproducibility).9.Keepthestandardparts&itsreadingfor“EquipmentCorrelation“purpose32GuidelinesForGaugeR&RAcceptanceUnder10%Error:Acceptable10%~30%Error:MaybeAcceptabledependingontheimportanceoftheapplication.Over30%error:Improvementisneeded33GaugeR&RStudy34GaugeR&RStudy35GaugeR&RStudy36GaugeR&RStudy37GaugeR&RStudy38AnovaMethodExample39EquipmentCalibrationBySimpleLinearityModel(Y=mX+b)

40SelectionofStandardUnitsSetupabout10pcsstandardunitswhichcancovertheoverallspreadofnormalproduction.Prefer2rejectunitbeyondupper/lowerspec.AgingallunitsFindstandardreading41RegressionAnalysis:EquipmentvsStandardMeasureStandardunitoneachindividualtestequipmentseveraltimes.TaketheaverageofmeasurementdataforeachunitRegressionAnalysis;EquipmentReadingvsStandardReading.Ifcorrelationcoefficient>90%,thencalculatetheslopeandoff-setpertestm/c.Ifcorrelationcoefficient<90%,findthereason.TestResult=ActualReadingXSlope+Offset42MeasurementDataofStandardSets

@eachTestMachine43RegressionAnalysis44AddTranslationKeyAfterCorrection=ReadingXSlope+OffsetTestresultbecomecloseramongtestmachines.Slope/OffsetperTestEquipment45EquipmentLongTermStabilityTestEquipmentlongtermstabilitycanbecheckedbytesting1~3workingstandardunitspereachtimebeforeoperatorusethattestmachineandplotthetestresultsonaSPCchart.46ComputingKappa

(GR&RTestForvisualinspector)47KappaCoefficient(Cohen,1960)SzBpOtherni.pi.Sz106.53(78).39104120.6BP2228.14(15).0751060.3Other2126.03(2).0120.1n.j1305020200p.j.65.25.11Rater2Rater

1pi.xp.i .39 .075 .01?Highreliabilityrequiresthatthefrequenciesalongthediagonalshouldbe>chanceandoffdiagonalfrequenciesshouldbe<chance.?Usemarginalfrequencies/probabilitiestoestimatechanceagreement.

Proportionagreementobserved,po=∑ipii=1/n∑iniiProportionagreementexpectedbychance,pc=∑ipi.xp.i

po-

pc

1-pcKappa,K=po=.53+.14.03=.7pc=.39+.075+.01=.475

.7-.475

K=1-.475

=.429

K=1,perfectagreementK=0,chanceagreementK<0,agreementworsethanchance.48KappaCoefficient(Cohen,1960)Proportionagreementobserved,po=∑ipii=1/n∑iniiProportionagreementexpectedbychance,pc=∑ipi.xp.i49Select20goodand20badpartsHaveeachinspectorinspecteachofthepartsandrecordwhetherit“Passed”or“Failed”;capturedataina“Kap