制程能力分析

1製程能力分析2008Mayrev.012目錄頁•前言•Ca、Cp、Cpk介紹•Minitab操作-常態•Minitab操作-非常態•Minitab操作-Within/Between•Minitab操作-多組比較•廠內製程能力•補充•P3~P5•P6~P11•P12~P25•P26~P31•P32~P41•P42~P54•P55•P56~P603製程能力分析何謂製程能力?指製程產出的一致性製程能力分析是用來衡量製程的品質符合規格要求的程度→在產品生產週期內統計技術可用來協助製造前之開發活動、製程變異性之數量化、製程變異性相對於產品規格之分析及協助降低製程內變異性。這些工作稱為製程能力分析(Processcapabilityanalysis)。4製程能力分析我們一般將產品品質特性之6倍標準差範圍(μ3)當作是製程能力之量測。此範圍稱為自然允差界限(naturaltolerancelimits)或稱製程能力界限(processcapabilitylimits)C1自然公差下限自然公差上限99.73%5製程能力分析Q:何謂“精”與“準“精不準不精不準精準C1Frequency10.410.210.09.89.614121086420LSL=9.7USL=10.3HistogramofC1NormalC1Frequency10.410.09.69.28.88.48.014121086420USL=10.5USL=10.5LSL=8.4HistogramofC1NormalC1Frequency11.210.810.410.09.69.28.814121086420USL=11.1LSL=9HistogramofC1Normal6製程能力分析製程準確度Ca(Capabilityofaccuracy)：衡量製程平均值與規格中心值的一致性:製程平均值與規格中心值的差距→(與目標值接近的距離)2)()()(LSLUSLXCa規格公差規格中值測量平均值Q:Ca愈大愈?7製程能力分析製程準確度Ca(Capabilityofaccuracy)：12.5%25%50%100%A級B級C級D級規格中心值規格上限(下限)Ｘ(實績)Ｘ(實績)Ｘ(實績)8製程能力分析製程精密度Cp(Capabilityofprecision)：雙邊規格衡量製程變異滿足產品品質標準(規格公差)之程度→生產產品品質彼此間差異的程度)(6)(標準差規格公差sLSLUSLCpQ:Cp愈大愈?9製程能力分析製程精密度Cp(Capabilityofprecision)：雙邊規格單邊規格[1].上限規格(USL)的場合[2].下限規格(LSL)的場合sXUSLCpu3sLSLXCpl310製程能力分析Cpk，製程能力指數：Cpk是總合Ca和Cp二值之指數，其計算式為；Cpk=(1-│Ca│)xCpCminUSLx3s,xLSL3spkCUSLx3spkCxLSL3spkUSLandLSL:USLOnly:LSLOnly:11製程能力分析一般製程能力要求A級：1.33≦CpkB級：1.0≦Cpk1.33C級：Cpk1.0公司要求若品質項目屬於產品管制特性時其Cpk≧1.67一般產品品質:Cpk1.33製程藥液:Cpk112製程能力分析-Minitab操作練習品管人員量測成型板厚，其規格為0.8~3.2mm，今天收集30批，而每批量測一片板子的板厚，其數據如附件所示，試計算其CP、CPK及此30批的不良率。13製程能力分析-Minitab操作1.先檢定此數據是否為常態分佈答14THKPercent4.03.53.02.52.01.51.0999590807060504030201051Mean2.146StDev0.4835N30AD0.454P-Value0.253ProbabilityPlotofTHKNormal-95%CI製程能力分析-Minitab操作答PValue=0.2530.05所以此筆數據為常態分配15製程能力分析-Minitab操作答16製程能力分析-Minitab操作答17製程能力分析-Minitab操作答3.22.82.42.01.61.20.8LSLUSLLSL0.8Target*USL3.2SampleMean2.14582SampleN30StDev(Within)0.404785StDev(Overall)0.48767ProcessDataCp0.99CPL1.11CPU0.87Cpk0.87Pp0.82PPL0.92PPU0.72Ppk0.72Cpm*OverallCapabilityPotential(Within)CapabilityPPMLSL0.00PPMUSL0.00PPMTotal0.00ObservedPerformancePPMLSL442.46PPMUSL4603.10PPMTotal5045.55Exp.WithinPerformancePPMLSL2892.84PPMUSL15321.81PPMTotal18214.65Exp.OverallPerformanceWithinOverallProcessCapabilityofTHKCpk=0.72，其totalppm為1821518製程能力分析-Minitab操作Exercise製工工程師收集30批(一批僅量測一片獨立線路區的一點)內層線路數據，試幫他計算其Cpk及良率;規格為40.419製程能力分析-Minitab操作1.先確認是否為常態分佈答20製程能力分析-Minitab操作1.先確認是否為常態分佈答21製程能力分析-Minitab操作1.先確認是否為常態分佈答THKPercent5.55.04.54.03.5999590807060504030201051Mean4.327StDev0.3186N30AD0.188P-Value0.894ProbabilityPlotofTHKNormal-95%CIPvalue=0.8940.0522製程能力分析-Minitab操作2.再計算Cpk答23製程能力分析-Minitab操作2.再計算Cpk答24製程能力分析-Minitab操作2.再計算Cpk答25製程能力分析-Minitab操作2.再計算Cpk答5.04.84.64.44.24.03.83.6LSLUSLLSL3.6Target*USL4.4SampleMean4.32741SampleN30StDev(Within)0.310628StDev(Overall)0.321376ProcessDataCp0.43CPL0.78CPU0.08Cpk0.08Pp0.41PPL0.75PPU0.08Ppk0.08Cpm*OverallCapabilityPotential(Within)CapabilityPPMLSL0.00PPMUSL433333.33PPMTotal433333.33ObservedPerformancePPMLSL9597.51PPMUSL407607.56PPMTotal417205.07Exp.WithinPerformancePPMLSL11805.18PPMUSL410644.20PPMTotal422449.38Exp.OverallPerformanceWithinOverallProcessCapabilityofTHKCpk:0.08其良率為1-42%=58%26製程能力分析-非常態分配若數據不是常態分配則如何計算Cpk?例:今現場技術員收集每天從化驗室量得到測微蝕槽銅離子含量分析數據，試計算其製程能力:銅離子規格為100ml/L(file:非常態Cpk.mtw)27製程能力分析-非常態分配銅離子Percent1007550250-25-50999590807060504030201051Mean6.243StDev17.74N25AD5.996P-Value0.005ProbabilityPlotof銅離子Normal-95%CIPvalue=0.0050.05表示不屬於常態分配28製程能力分析-非常態分配29製程能力分析-非常態分配30製程能力分析-非常態分配31製程能力分析-非常態分配420-2-4USL*transformeddataUSL*4.60517SampleMean*0.0520481StDev(Within)*1.42262StDev(Overall)*1.80865LSL*Target*USL100SampleMean6.24333SampleN25StDev(Within)5.97944StDev(Overall)17.9244LSL**Target**AfterTransformationProcessDataCp*CPL*CPU1.07Cpk1.07Pp*PPL*PPU0.84Ppk0.84Cpm*OverallCapabilityPotential(Within)CapabilityPPMLSL*PPMUSL0.00PPMTotal0.00ObservedPerformancePPMLSL**PPMUSL*685.91PPMTotal685.91Exp.WithinPerformancePPMLSL**PPMUSL*5910.86PPMTotal5910.86Exp.OverallPerformanceWithinOverallProcessCapabilityof銅離子UsingBox-CoxTransformationWithLambda=0Cpk=0.84,不良率:5910ppm32製程能力分析今負責電鍍的品管工程師想瞭解其電鍍銅面銅厚度的製程能力，於是設計了一抽樣計劃為收集30批板子，每一批量測5片板子，其數據如附件所示(面銅厚度:204m)練習33製程能力分析1.先計算是否為常態分佈答dataPercent27.525.022.520.017.515.099.99995908070605040302010510.1Mean19.84StDev1.711N150AD0.822P-Value0.033ProbabilityPlotofdataNormal-95%CIPvalue=0.0330.05表示此不符合常態34製程能力分析1.非常態分佈使用box-cox再分析製程能力答為何要使用Between/within?35製程能力分析CapabilityAnalysis(Between/Within):此時變異需考量組內的變異以及組與組的變異THKtankPANEL32121181521181521181525.022.520.017.515.0BoxplotofTHKvstank,PANEL組與組差異組內差異36製程能力分析答37製程能力分析答38製程能力分析答242220181614LSLUSLLSL16Target*USL24SampleMean19.838SampleN150StDev(Between)0StDev(Within)1.78896StDev(B/W)1.78896StDev(Overall)1.71343ProcessDataCp0.75CPL0.72CPU0.78Cpk0.72Pp0.78PPL0.75PPU0.81Ppk0.75Cpm*OverallCapabilityB/WCapabilityPPMLSL20000.00PPMUSL20000.00PPMTotal40000.00ObservedPerformancePPMLSL15961.16PPMUSL9996.14PPMTotal25957.29Exp.B/WPerformancePPMLSL12546.90PPMUSL7569.21PPMTotal20116.11Exp.OverallPerformanceB/WOverallBetween/WithinCapabil