6Sigma六西格瑪精益生產持續(xù)改善分析管理常用工具資料 Six_Sima_Project_Report_04_-_Improve_Phase

1、Y = F(X) 與 R(adj)-sq :Y(NTF%)=2.1514+1.185*House using times+0.9*Memory convert card using times+0.351*House using times* Memory convert card using timesR-Sq = 84.01% 目前製程能力 : (Zst) 1.09 Sigma 預期專案改善績效(Zst) : 3.0 Sigma最佳解決對策與改善計畫: settle production parameter as : a. CPU convert house using times = 1

2、750 times b. Memory convert card using times = 2200 times c. Operator shift = any oneR4 ImproveSummary階段開始日期: 06/16/2005階段完成日期: 08/15/2005專案類型 : DMAICConclusion / Issues / Next Step :To do the DOE with all possible factor afterA phase, and then confirm which is key factor toimpact NTF rate. 2. Based

3、 on Interaction Plot & Main Effect Plot & DOE analysis, we have got a result that CPU convert house and memory convert card using times will impact NTF rate. 3. In other words, NTF rate have 84.01% variability come from CPU convert house and memory convert card using times .4 . Based on Regression r

4、esult, we know that:Y(NTF%)=2.1514+1.185*House using times+0.9*Memory convert card using times+0.351*House using times* Memory convert card using times5. Based on DOE analysis, we have got a result that CPU convert house and memory convert card using times will impact NTF rate. But consider material

5、 cost, we will be settle production parameter as : a. CPU convert house using times = 1750 times b. Memory convert card using times = 2200 times c. Operator shift = any oneThis parameter is same as adjust current, so we will need modify using times SPEC6.Inspect sample 6400pcs, find P-value=0.000 0.

6、05 item (In red circle).I2-1DOE 實驗規(guī)劃Pg 6Factorial Fit: NTF （ %） versus House using time, Memory convert c Estimated Effects and Coefficients for NTF （ %） (coded units)Term Effect Coef SE Coef T PConstant 2.1514 0.1261 17.06 0.000House using times 2.3700 1.1850 0.1362 8.70 0.000Memory convert card us

7、ing times 1.8017 0.9008 0.1362 6.61 0.000House using times* 0.7017 0.3508 0.1362 2.58 0.017 Memory convert card using timesS = 0.667230 R-Sq = 84.01% R-Sq(adj) = 82.01%Analysis of Variance for NTF （ %） (coded units)Source DF Seq SS Adj SS Adj MS F PMain Effects 2 53.1774 53.1774 26.5887 59.72 0.0002

8、-Way Interactions 1 2.9540 2.9540 2.9540 6.64 0.017Residual Error 24 10.6847 10.6847 0.4452Curvature 1 0.1893 0.1893 0.1893 0.41 0.526Pure Error 23 10.4954 10.4954 0.4563Total 27 66.8161Main Effects P0.05I2-1DOE 實驗規(guī)劃Pg 7I2-1DOE 實驗規(guī)劃Pg 8有什麼發(fā)現? (運用工具找出什麼重要結果/結論之重點摘要?) 1.Based on Interaction Plot & Mai

9、n Effect Plot & DOE analysis, we have got a result that CPU convert house and memory convert card using times will impact NTF rate. 2. In other words, NTF rate have 84.01% variability come from CPU convert house and memory convert card using times . 3.Based on Regression result, we know that:Y(NTF%)

10、=2.1514+1.185*House using times+0.9*Memory convert card using times+0.351*House using times* Memory convert card using timesS = 0.667230 R-Sq = 84.01% R-Sq(adj) = 82.01%Y(NTF%)=2.1514+1.185*House using times+0.9*Memory convert card using times+0.351*House using times* Memory convert card using times

11、I2-1DOE 實驗規(guī)劃Pg 9Consider 95% PI :Consider 95% CI :Count cost base on 18.2K per month1.Optimal cost=6400*30/1880*70+6400*30/1000*30=RMB129092.Adjust cost=6400*30/1750*70+6400*30/2200*30=RMB102983.NTF cost=(6400*30*0.0141-6400*30*0.0076)*0.1*12=RMB1497Times=1880Times=1000Times=1750Times=2200Save cost/

12、Month=12909-10298-1497=RMB1113So save New parameter: CPU house times=1750Memory cards times=2200I2-1DOE 實驗規(guī)劃Pg 10有什麼發(fā)現? (運用工具找出什麼重要結果/結論之重點摘要?)1.Based on DOE analysis, we have got a result that CPU convert house and memory convert card using times will impact NTF rate. But consider material cost, we

13、 will be settle production parameter as : 1. CPU convert house using times = 1750 times 2. Memory convert card using times = 2200 times 3. Operator shift = any oneThis parameter is same as adjust current, so we will need modify using times SPECI3 Should-BeProcess我們想要知道什麼? (為何用此工具)有什麼發(fā)現? (運用工具找出什麼重要結

14、果/結論之重點摘要?)ICT TestPick up and place board on fixturePlace CPUPlace CPU coolerInsert MemoryInsert HDD cableInsert other cableInsert Test CardF/TOQCPackingNVASOPRepairNGShippingYesNVASOPInsert FDD cableBackouttest equipmentPlace board on linePassNGPassI4 專案改善績效評估我們想要知道什麼? (為何用此工具)有什麼發(fā)現? (運用工具找出什麼重要結果

15、/結論之重點摘要?)抽樣樣本數: n=6400假設檢定績效有效性: P-Value=0.000.05;接受Ha,NTF rate1.4%短期Sigma 水準(Zst)=3Test and CI for One Proportion Test of p = 0.014 vs p 0.014 95% UpperSample X N Sample p Bound Z-Value P-Value1 6 6400 0.000938 0.001567 -8.89 0.000Primary Metricsii fiii有什麼發(fā)現? (運用工具找出什麼重要結果/結論之重點摘要?)The DPPM is hig

16、h, there is much space to improve it.FinancialEstimate有什麼發(fā)現? (運用工具找出什麼重要結果/結論之重點摘要?)通過改善七大浪費的手法我們有節(jié)省了NTD9981.00I5 下一專案描述我們想要知道什麼? (為何用此工具)有什麼發(fā)現? (運用工具找出什麼重要結果/結論之重點摘要?) 專案計劃時程專案編號: G-Pro002 專案類型: DMAIC專案名稱: 提升 F2 test yield rate事業(yè)單位: CSBG G-Pro專案成員:Kano Chen. Jemmy Li. Annie Yeh. Kenny Liu. Hawk Yi. Jaron Zhou.專案核淮與啟動:定義D衡量M分析A改善I控制C起訖06/15/0506/30/0507/01/0508/01/0508/01/0509/01/0509/01/0510/01/0510/01/0511/01/05ChampionRichard Liang6 Exec.Parkson LaiBB / GBProcess OwnerJemmy LiFinancePatrick PaoMBBJack zhangConclusion / Issues / N