SPCのCPK实战篇

1SPC及CPK實戰篇21﹑六西格碼在控制中的應用★廣義的六西格碼是一個完善的管理工具﹐是品質管理系統的一種說法﹔★單純的六西格碼控制應用﹐是一種質量控制的目標﹐請看以下說法﹐正確與否﹕۩我公司已經實施了六西格碼管理﹔۩我公司達到了六西格碼的要求﹔۩通過六西格碼管理﹐公司品質上了一個台階﹔………一﹑六西格碼32﹑六西格碼與不良率:±Kr百分比(%)百萬分缺點數±1r±2r±3r±4r±5r±6r68.2695.4599.7399.993799.99994399.9999998317400455002700630.570.002(規格中心不偏移)常態分配規格下限規格上限-6r-5r-4r-3r-2r-1r+6r+1r+2r+3r+4r+5rX424±Kr合格百分比(%)百萬分缺點數±1r±2r±3r±4r±5r±6r30.2369.1393.3299.379099.9767099.9996606977003087006681062102333.4規格中心值往左、右移動1.5r常態分配規格下限規格上限-6r-5r-4r-3r-2r-1r+6r+1r+2r+3r+4r+5rX1.5r1.5r4353控制方式与6控制方式的比较：6???兩個影響量﹐﹑1﹑的變化﹖2﹑的變化﹖21e-222F(x)=(x-)7二﹑SPC與CPK1﹑區別與聯系CPK﹕判定短期的制程能力﹐可以用來解析目前的制程能力或用來檢驗制程能力是否在管制狀況下﹔SPK﹕長期的制程能力﹐主要用在制程管制上。兩者配合使用﹐應該做到﹕短期=長期82﹑影響量----標准差Sigma樣173.673.672.076.576.872.076.873.673.673.673.673.670.470.472.0272.073.673.675.273.673.672.076.876.878.472.073.670.470.472.0UCLCLLCL本376.875.272.076.876.873.676.878.476.878.473.675.270.472.073.676.774.171.4測4UCLCLLCL222.4222.4217.6228.5227.2219.2225.6228.8227.2230.4219.2222.4211.2212.8217.66.62.60.074.174.172.576.275.773.175.276.375.776.873.174.170.470.972.5ΣX=4.81.61.61.63.21.64.84.83.24.81.61.60.01.61.6ΣR=X=R=1.960.690.85製品名稱TRX-7規格標準溫度時間群組數大小製造部門SMT期間49周上限USL225803管制項目爐溫中心限CL22075總組數機別回焊爐測溫板TRX-7-1實板測量單位℃&sec下限LSL2157015測定者姜雲霞時間時間管制ΣＸ日期2004.11.29時間1110.838.4RＸ時間74.12.6R管制圖CP=預估不良率(PPM)製程能力分析時間Sigma=CPK=X圖R圖備注及原因追查1﹑時間的數據離散性較大﹐Sigma=1.96﹐因此要從測量儀器﹑影響時間的設置來分析解決。時間X管缺制圖平均0.02.04.06.08.0123456789101112131415UCLCLLCLR69717375777981123456789101112131415UCLCLLCLXUSLLSLSL實例(Sigma較大﹐離散性問題)103﹑影響量----中心偏移量11CPK實例分析(離散性較小﹑中心偏移量影響)製品R圖名稱管制項目測量單位日期/時間批號12345678910111213141516171819202122232425ΣＸ＝15.8967樣10.1310.1350.1270.1270.1310.1350.1350.1270.1270.1310.1270.1310.1310.1350.1270.1350.1310.1270.1380.1310.1310.1270.1310.1350.127ΣＲ＝0.241本20.1350.1310.1310.1240.1240.1310.1270.1310.1270.1350.1310.1240.1270.1310.1270.1270.1240.1240.1350.1240.1270.1270.1270.1310.131測30.130.1270.120.120.1240.1310.1270.1240.1240.1270.1310.1270.1240.1270.120.120.120.1240.1310.1270.1270.1240.120.1310.131定40.1270.1270.1240.1310.1270.1350.1310.120.120.1270.1270.1310.1270.1310.120.1240.1240.1270.1270.1310.130.1270.1240.120.124值50.1240.1310.1240.1270.120.1270.1240.1160.120.120.1240.1270.1240.1240.120.130.1240.1240.120.1270.1310.1270.1240.1270.1310.6440.6510.6260.6290.6260.6590.6440.6180.6180.640.640.640.6330.6480.6140.6330.6230.6260.6510.640.6470.6320.6260.6440.6440.1290.130.1250.1260.1250.1320.1290.1240.1240.1280.1280.1280.1270.12960.1230.1270.1250.1250.130.1280.1290.1260.1250.1290.1290.0110.0080.0110.0110.0110.0080.0110.0150.0070.0150.0070.0070.0070.0110.0070.0150.0110.0030.0180.0070.0040.0030.0110.0150.007Ｘ＝0.1272R=0.0101﹑中心偏移量為0.0072﹐Sigma值為0.0024﹐通過計算SPC的管制范圍為0.011﹐可見控制范圍已經相當小﹐說明每次測量的數據都集中在一個區域﹐波動很小。因此CPK值十分合理﹐為1.81﹔2﹑但管制上下限均在規格中心到規格上限的范圍內﹐且管制下限還沒有達到規格中心。可以考慮調整規格﹐并作進一步跟蹤。R管制圖Sigma=0.0024CPK=1.8112備註及原因追查:CP=2.8242R平均X管制圖預估不良率(PPM)製程能力分析ΣＸＮ＝25Ｘ合　　　計量測數值的判定條件USL藍色LSL紅色測定者姜雲霞日期2004.11.16印刷機鋼板厚度0.100mmmm下限LSL0.125下限LCL0.1220.000中心限CL0.1270.010機別錫膏厚度中心限CL0.12總組數期間47周上限USL0.145上限UCL0.1330.020UT220V規格標準群組數大小管制X圖製造部門SMT00.010.020.0312345678910111213141516171819202122232425UCLCLLCLR0.0980.1030.1080.1130.1180.1230.1280.1330.13812345678910111213141516171819202122232425UCLXLCLLSLUSLCLSL124﹑CPK改善的方向及目標135﹑管制图的判断在計量值﹐服從常態分布的生產特性下﹐最常用的管制圖是Xbar-R管制圖。常用的控制图a、Xbar–R控制图（计量值、正态分布）b、不合格品百分率p控制图（计件值、二项分布）c、不合格品数np控制图(计件值、二项分布)d、缺陷率u控制图(计点值、泊松分布）e、缺陷数c控制图(计点值、泊松分布）f、DPMO控制图(计点值、泊松分布）141414.214.414.614.81515.215.415.615.8折線圖1!5.1﹑稳定原则----異常處理原则一:不可超出管制界限1514.214.414.614.81515.215.415.6折線圖15.1﹑稳定原则----預防處理原则二:不可呈现规律变动连续7点上升或下降呈现规律性波动,具有显著的波动周期1614.214.414.614.81515.215.415.6東部5.1﹑稳定原则----預防處理原则二:不可呈现规律变动连续7点在管制界限一侧连续11点中,有10点在管制界限一侧(间断)175.1﹑稳定原则----預防處理原则三:不可过于集中分布显著多于2/3的点集中在CL附近:即90%的数据在1δ线以内显著少于40%的数据落在1δ线以内数据集中分布在2δ到管制线之间(超过2/3)3点中,有两点接近一侧之管制界限7点中,有3点以上接近控制界限原则四:符合过程能力要求185.2﹑长期控制符合以下条件,可将管制图用于长期控制:过程稳定符合过程能力的要求再收集15组数据,无异常情形利用已经算出之管制界限进行管制按要求收集数据,并进行管制分析/控制在下列情况下,重新计算管制界限:初次过程能力研究,异常消除时对4M1E进行重大改进时产品修正时经过长期运行后19现象可能原因处理分布不稳定计算错误数据未层别查清原因,重新计算存在特殊原因查清原因,予以消除过程能力不足过于分散平均值偏移采用管理性/技术性措施,进行改进5.3﹑異常處理可以采取的措施有效后,标准化处理20三﹑制程管制實例1﹑CPK解析目的﹕通過集中量測解析目前的制程能力﹐得出制程管控要求﹐對制程管控要求進行分析﹐解析出合理的制程管控上下限。實例分析﹕以解析錫膏厚度的管控上下限為例﹕集中量測25組錫膏厚度數據﹔每組由5個量測點的數據﹔計算CPK值﹐判定是否合理﹔計算上下管制線。211.1﹑集中量測25組錫膏厚度數據次數測量點1234567891011121310.1510.140.1480.1440.1510.1480.1370.140.1370.1480.1440.1370.13720.1440.1370.140.140.1440.1440.140.1440.1330.1440.1370.1440.12930.1440.1480.1370.1330.1440.140.1370.140.1330.1370.1330.140.13340.1330.140.1440.140.140.140.1440.1370.1330.140.1290.140.13350.1440.1370.140.1440.1440.1440.140.140.1370.1370.140.1370.133AVG0.1430.1400.1420.1400.1450.1430.1400.1400.1350.1410.1370.1400.133次數測量點14151617181920212223242510.1370.140.1510.140.1440.140.1480.1480.1440.1440.140.13320.1440.1330.140.1330.1440.1370.140.1330.1480.1480.1440.12930.1330.1330.1440.1330.1330.1440.1370.1330.1370.1440.140.13340.1440.1290.1330.1330.140.1440.1480.1290.140.1370.1440.13350.140.1440.1510.1480.140.1330.140.140.140.1440.140.148AVG0.1400.1360.1440.1370.1400.1400.1430.1370.1420.1430.1420.1350.139832TotalAv