• 1. Thru RTY InnovationAchieve10 million per yearLG Electronics / Digital Appliance Company LGETA MWO
    • 2. Vision 12.5 11.810,110.09.8(’04 Target)(’03, Million)TA(10.0) TA(11.1) Global No.1 MWO Division ( `05 Sales: M$ 385 ;Profit: 8 % )China Market Top 1 ( `05 China M/S : 45% )Cost CompetitionVisionBefore 2005Now As the basement of Profit TurnaroundFast InnovationFast GrowthGlobal PositionMWO Maker Benchmarking VisionVision1/19D MA I C
    • 3. Big Y Issue7.110.0`02`03CMOProductivity loss↑ Portfolio by CategoryHighLow PriceProfitHighQuality loss↑ Parts loss↑ Fast GrowthNew MDL Accelerate tool shift ( 0.7 ~1.8) Shorten the development period Increasing number of partsM unit / year30%CMO (0.7cu.ft)CMO (0.8~1.1)OTRCMO (0.7cu.ft)CMO (0.8~1.8)BrownerOTR2/19D MA I C
    • 4. InputOutputProcess Assy LinePressWeldingPaintingPurchase IQCOQCPCBSupplierCustomer Logistics system (Material, Product) Parts loss control Inspection System Box quality control Round Cavity strength PCB ICT FCT control Powder quality Gap defect . Product and Quality Strategy IR sensor quality Press line system Quick response system Outgoing Inspection system Material storage system Main IssueBig Y Issue 0.7Cu.ft latch adjustment Door ass’y trouble Anode abnormal temp. 0.7 UTC cavity Distortion P/Cord scratch Door dirty Production Process System Assembly LinePainting PCBMFG 6σ L/B improvement Powder thicknessBIG YLittle YAction WG Weld intensity Part qualityQuality YieldQuality up MFG 6σ Line balanceMFG 6σ Pressing PCB lead improvement Logistic Lead Time Quick ResponseLogistics TQ 6σ NWT MFG 6σRTY Innovation RTY↑TDRMFG 6σMFG 6σWeldingTQ 6σ 02.12Target12050%02.12Target3.52.7825%02.12Target3.52.5628%PCBPaintingLogisticsAss’yMiscellaneous ratioRTYzRTYz02.12Target3.52.7521%60RTYz10 million per year Assurance Target 027.010.0M units/year30%80859095RTYq60708090RTYp 3.0 3.5 4.0 4.5 5.0 5.5ABCDRTYz Portfolio3.5EA/Hour3/19D MA I C
    • 5. Define MeasureAnalysisImproveControlSupplierIQCPaintingPCBAss’yOQCCustomerBetter logisticsWorst logisticsIQCWaiting for ……?Stock Product stockAss’y linePainting linePCB lineNeck Point Plan change Everyday Stock is not enough Elevator is not enough Packing Time too long Broken Box Check time conflict Poor loading methodLogistics improvement Best practice-14/19D MA I C
    • 6. Vendor and supervisor meeting ImproveFocusInvoice UnloadingChecking Load toolsStock CommunicateLoading toolLoad wayStock manageinspect system●●●●●●●●●●●●●●●●Define MeasureAnalysisImproveControlVoice Of CustomerPlan change Everyday Stock is not enough Can not unload in stock Invoice check is difficult Elevator is not enough Packing Time too long Forklift is not enough irregulars training for loader Broken Box Waiting every procedure Check time conflict Poor loading method Poor loading vehicle Incoming check time Poor part quality Stack management●●●●●●●●●●●●●●●●Logistics improvement Best practice-1SupplierIQCPaintingPCBAss’yOQCCustomer5/19D MA I C
    • 7. Best practice-1Define MeasureAnalysisImproveControlWaiting for …Ppk=-1.01Miscellaneous ratioEA/hOne-way ANOVA: make invoice, invoice, inspection, check, stock, recover Analysis of Variance Source DF SS MS F P Factor 5 908047 181609 70.14 0.000 Error 138 357333 2589 Total 143 1265379 Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --+-----+-------+-----+- make inv 24 14.70 5.71 (--*-) invoice 24 21.98 10.96 (--*-) inspecti 24 15.06 2.41 (--*-) check 24 26.75 15.64 (-*--) stock 24 229.85 122.93 (--*-) recover 24 8.27 4.47 (--*--) --+-------+-------+-----+---- Pooled StDev = 50.89 0 80 160 240Make invoiceInspect Stock Recover partCheck unloadInvoice checkLogistics improvement SupplierIQCPaintingPCBAss’yOQCCustomer8:0010:0013:0015:0017:0019:00111551361271521101409169201026Box Inject Press Two-way ANOVA: waiting versus time, part Analysis of Variance for waiting Source DF SS MS F P time 5 23546 4709 10.20 0.001 part 2 2302 1151 2.49 0.132 Error 10 4619 462 Total 17 30467 Individual 95% CI time Mean -------+---------+---------+---------+---- 100 11 (------*------) 130 54 (------*-----) 150 11 (------*------) 170 3 (------*------) 190 0 (------*------) 800 100 (------*------) -------+---------+---------+---------+---- 0 40 80 120 Individual 95% CI part Mean --+---------+---------+---------+--------- box 29.2 (-----------*-----------) inject 16.5 (-----------*------------) press 44.2 (------------*-----------) --+---------+---------+---------+--------- 0.0 16.0 32.0 48.0One-way ANOVA: waitingp versus vendor Analysis of Variance for waitingp Source DF SS MS F P vendor 10 298536 29854 57.13 0.000 Error 91 47551 523 Total 101 346088 Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --+---------+---------+---------+---- daxian 7 99.60 0.40 (--*-) douxing 9 74.40 9.01 (-*--) geleide 9 223.80 0.00 (-*--) henan 7 89.60 0.60 (--*--) lianhe 9 197.60 28.93 (--*-) lianhua 18 188.60 48.58 (*-) nix 9 63.80 0.00 (--*-) sanyue 9 124.80 2.77 (--*-) xian an 9 89.00 0.00 (--*-) xianhe 9 92.80 0.69 (-*--) xushi 7 155.60 1.60 (--*--) --+---------+---------+---------+---- Pooled StDev = 22.86 60 120 180 2406/19D MA I C
    • 8. D MA I C 7/19Define MeasureAnalysisImproveControlIQCStock Product stockAss’y linePainting linePCB linePress stockSupport time controlLoad method Ppk=1.7720mmHVT container truckBe sure the way of turn around , enlarge the cartwheel 2 timesAdd pothook , make the container and load car can connect ,1246032Jan.TargetJun.Lead time(min.)Box stockBest practice-1Logistics improvement SupplierIQCPaintingPCBAss’yOQCCustomerBest Practice 2 Ass’y RTY innovation1. Cavity Prepare 2. Base Install 3. L/Board Fix 4. MGT Fix 5. S/Guide Fix 6. P/Cord Install 7. Controller Fix 8. Door Install09. Connection 1 10. Connection 2 11. Door Gap Adjust 12. Latch Adjustment 13. Lamp Install 14. Structure inspection 15. Function inspection 16. Aging 17. Life test 18. O/Case Install 19. O/Case Fix 20. Leakage test 21. Nameplate stickup 22. Last Inspection 23. Packing 24. Pack upPart and Function 234567891011121232119181613241415172022Appearance Through process mapping we found some procedure has problems99%95%99%99%99%98%82%98%99%93%83%99%99%84%98%98%99%100%
    • 9. Define MeasureAnalysisImproveControlForget self checkLatch Adjust D/FrameLatchL/BoardCavityManOld or newMethod Mould agingTotal inspectionburrMachine Parameter injectionMaterial Gage aging The wayPPK=0.06 Latch 3 mould is CTQOne-way ANOVA: latch1, latch2, latch3, latch4 Analysis of Variance Source DF SS MS F P Factor 3 53.18 17.73 3.40 0.020 Error 116 605.11 5.22 Total 119 658.30 Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ----+---------+---------+---------+-- latch1 30 100.92 0.94 (-------*-------) latch2 30 100.92 0.76 (-------*-------) latch3 30 102.27 4.30 (--------*-------) latch4 30 100.52 0.95 (-------*-------) ----+---------+---------+---------+-- Pooled StDev = 2.28 100.0 101.0 102.0 103.0Latch adjustment improvementBest practice-2SupplierIQCPaintingPCBAss’yOQCCustomer8/19D MA I C
    • 10. Define MeasureAnalysisImproveControlRegression Analysis: time versus latch b, latch b-p, latch b-l, latch The regression equation is time = - 0.24 + 0.0176 latch b + 0.0920 latch b-p + 0.278 latch b-l- 0.0148 latch Predictor Coef SE Coef T P Constant -0.243 3.481 -0.07 0.945 latch b 0.01756 0.03172 0.55 0.585 latch b-l 0.09197 0.02284 4.03 0.000 latch b-p 0.27811 0.02761 10.07 0.000 latch -0.01481 0.02400 -0.62 0.543 S = 0.02795 R-Sq = 81.9% R-Sq(adj) = 79.0%Regression Analysis: time versus latch b-p, latch b-l The regression equation is time = 0.019 + 0.0948 latch b-p + 0.278 latch b-l Predictor Coef SE Coef T P Constant 0.0192 0.6785 0.03 0.978 latch b- 0.09475 0.02186 4.33 0.000 latch b- 0.27777 0.02589 10.73 0.000 S = 0.02717 R-Sq = 81.6% R-Sq(adj) = 80.2%PPK=2.19 PoolingLatch adjustment improvementBest practice-2SupplierIQCPaintingPCBAss’yOQCCustomer9/19D MA I C
    • 11. Define MeasureAnalysisImproveControlPast filmHold Put filmPutTake offPress filmHold plateTurn aroundHang 2.80.91.01.01.81.30.71.02.0Sec.‘03 2.30.80.80.91.71.20.70.81.1Ppk= -0.31 ‘02 Chi-Sq = 5.692 + 0.008 + 0.011 + 0.000 + 0.077 + 0.000 + 6.151 + 0.008 + 2.628 + 0.003 + 1.681 + 0.002 + 5.037 + 0.007 + 0.001 + 0.000 + 1.080 + 0.001 + 0.699 + 0.001 + 0.001 + 0.000 = 23.089 DF = 10, P-Value = 0.010Scratch by Model AnalysisChi-Square Test: Defect2, Total2 Expected counts are printed below observed counts Defect2 Total2 Total 1 9 14783 14792 19.49 14772.51 2 5 14783 14788 19.48 14768.52 3 2 14784 14786 19.48 14766.52 4 62 14784 14846 19.56 14826.44 Total 78 59134 59212Chi-Sq = 5.642 + 0.007 + 10.764 + 0.014 + 15.683 + 0.021 + 92.114 + 0.122 = 124.367 DF = 3, P-Value = 0.000Scratch by Vendor AnalysisBest practice-2SupplierIQCPaintingPCBAss’yOQCCustomerDoor ass’y process improvement10/19D MA I C
    • 12. Define MeasureAnalysisImproveControlDoor ass’y process improvementAss’y line Door line P/ codeS/ guideAss’y line Door line P/ codeS/ guideAdd total inspector to check itVendor package improve Press film2.10.70.80.91.71.20.70.81.0Past filmPut filmPutTake offHold plateTurn aroundPpk=1.9HoldHangBest practice-2SupplierIQCPaintingPCBAss’yOQCCustomer11/19D MA I C
    • 13. Define MeasureAnalysisImproveControlStandard workVCD trainingStandard lineStandard sampleOther line Training Process Training for new employee When line stop , training Supervisor inspection inspect oneself.VOC researchStandard modifyProcess training SupplierIQCPaintingPCBAss’yOQCCustomerBest practice-212/19D MA I C Best Practice 3 Painting Process innovationRaw material Welding defectPainting defectPress LineStorage(L/P)Welding LineStoragePainting LineAssembling Line
    • 14. Define MeasureAnalysisImproveControlPainting line improvementProcess mappingHow to cause it (X)RackingPretreatmentSetting Water Dry OvenCoolingPowder SprayMaskingBakingInspectionJig Quality Position Position HumidityRobot Work Powder FlowTemperature TimeJudgmentTemperature Concentration Process mappingHow to cause it (X)Powder different Slow line speed Painting robot defect Program differentNew model new Problems Painting condition Powder condition SupplierIQCPaintingPCBAss’yOQCCustomerBest practice-3Estimated Regression Coefficients for thicknes Term Coef SE Coef T P Constant -348.0 326.545 -1.066 0.312 distance 4.0 5.569 0.717 0.050 gun speed 4.8 5.230 0.919 0.029 powder 90.0 103.561 0.869 0.015 distance*distance -0.0 0.044 -0.737 0.178 gun speed*gun speed -0.0 0.030 -0.802 0.341 powder*powder -9.8 12.169 -0.802 0.241 distance*gun speed -0.0 0.049 -0.393 0.403 distance*powder -0.1 0.980 -0.131 0.298 gun speed*powder -0.2 0.817 -0.282 0.183 S = 4.158 R-Sq = 92.4% R-Sq(adj) = 85.0%powderOptimize painting conditionGun speedDistance The best powder conditionepoxyFactors polyesterpigmentTi02level0.3000.03000.3050.03250.3100.03500.3150.03750.3200.04000.3250.3300.3300.3000.3050.3100.3150.3200.3250.3300.3300.3000.3050.3100.3150.3200.3250.3300.3300.03000.03250.03500.03750.0400additiveDistance 35 Gun speed 68 Powder 3.28I35mm68mm/sBeforeAfterR-1 GUN powder controlResult 3425450PPMJan.Jun.13/19D MA I C
    • 15. Estimated Regression Coefficients for cost_1 Term Coef SE Coef T P Constant 0.498799 1.06587 0.468 0.045 PG 0.016906 0.08130 0.208 0.037 PSP 0.016195 0.09905 0.164 0.002 voltage 0.003204 0.00554 0.579 0.029 S = 1.1027 R-Sq = 98.2% R-Sq(adj) = 92.6%14/19Analysis of Variance for intencity Source DF Seq SS Adj SS Adj MS F P Regression 10 399.405 399.4050 39.94050 16.54 0.003 Linear 5 370.783 42.4636 8.49272 3.52 0.097 Interaction 5 28.623 28.6225 5.72450 2.37 0.183 Residual Error 5 12.073 12.0725 2.41450 Total 15 411.478Define MeasureAnalysisImproveControlPpk=0.37Estimated Regression Coefficients for intencity Term Coef SE Coef T P Constant 0.525 34.493 0.015 0.088 voltage 0.045 0.105 0.432 0.042 PG -9.175 5.124 -1.791 0.033 PSP -7.250 5.298 -1.368 0.029 time 16.875 25.325 0.666 0.535 dianji 627.500 506.498 1.239 0.070 voltage*PG 0.034 0.016 2.188 0.080 voltage*PSP 0.026 0.016 1.673 0.155 voltage*time -0.065 0.078 -0.837 0.441 voltage*dianji -2.700 1.554 -1.738 0.043 PG*PSP 0.288 0.388 0.740 0.046 S = 1.554 R-Sq = 97.1% R-Sq(adj) = 91.2%PoolingPpk=1.19VoltagePGPSP02.1203.63.62.7521%RTYzPainting line improvementSupplierIQCPaintingPCBAss’yOQCCustomerBest practice-3D MA I C Best Practice 4 PCB line improvement
    • 16. Define MeasureAnalysisImproveControlPCB line improvement2.0mm4.0mmPpk =-0.74LEADHead Adjust ANVIL high(Head)1.2+0.5 1.5+0.1 SpecPre _CTQKnifeCUTTER12.5+0.2Part Eyelet 5mm10mmRTICTMAN Insert Auto Soldering ICTLED SOLDERFCTC/P servePCB serveK/M past SCREWFCTAuto InsertSolderAssemblageSupplierIQCPaintingPCBAss’yOQCCustomerBest practice-4Two-sample T for C1 vs C2 N Mean StDev SE Mean C1 10 2.1490 0.0152 0.0048 C2 10 2.3420 0.0257 0.0081 Difference = mu C1 - mu C2 Estimate for difference: -0.19300 95% CI for difference: (-0.21287, -0.17313) T-Test of difference = 0(vs not =): T-Value = -20.41 P-Value = 0.000 DF = 18Two-sample T for C1 vs C2 N Mean StDev SE Mean C1 10 2.1490 0.0166 0.0053 C2 10 2.3500 0.0205 0.0065 Difference = mu C1 - mu C2 Estimate for difference: -0.20100 95% CI for difference: (-0.21856, -0.18344) T-Test of difference = 0 (vs not =): T-Value = -24.04 P-Value = 0.000 DF = 18Two-sample T for C1 vs C2 N Mean StDev SE Mean C1 10 2.1430 0.0241 0.0076 C2 10 2.3410 0.0191 0.0060 Difference = mu C1 - mu C2 Estimate for difference: -0.19800 95% CI for difference: (-0.21842, -0.17758) T-Test of difference = 0 (vs not =): T-Value = -20.37 P-Value = 0.000 DF = 18 1.5 ± 0.11.2±0.212.5 ± 0.2HeadANVILCUTTER15/19D MA I C
    • 17. 16/19Define MeasureAnalysisImproveControlTerm Coef SE Coef T P Constant -325.8 534.135 -0.610 0.006 cutter 44.4 84.500 0.525 0.001 anvil -11.6 11.390 -1.017 0.033 head 77.4 57.426 1.347 0.208 cutter*counter -1.6 3.369 -0.467 0.651 anvil*anvil -0.3 0.135 -2.408 0.037 head*head -8.1 3.369 -2.408 0.037 cutter*anvil 1.0 0.904 1.051 0.318 cutter*head -4.3 4.521 -0.940 0.369 anvil*head 0.4 0.904 0.442 0.668 S = 0.1279 R-Sq = 83.0% R-Sq(adj) =79.7%Finding Optimized ConditionPpk =-0.74Ppk =2.17Before After Cutter Anvil Head 12.5 1.2 1.5PCB line improvementSupplierIQCPaintingPCBAss’yOQCCustomerBest practice-4D MA I C Best Practice 5 Part quality innovation
    • 18. 10 million TDR ActivityDefine MeasureAnalysisImproveControlVendor parts improvementVendor Vendor TDR meetingBefore Now Quality ? Vendor SupplierIQCPaintingPCBAss’yOQCCustomerBest practice-517/19D MA I C Box viscosity improveElectronics –timer improveZbench =0.08Zbench =0.68LOTRTN516534357251842%89.7%Jan.Jun.Jan.Jun.Ppk =0.05Ppk =1.23CTQ SYSTEM
    • 19. System ControlControl Lead timeRTYq8085909560708090RTYp 3.0 3.5 4.0 4.5 5.0 5.5ABCDRTYz PortfolioBar cord systemFeb.March.Apr.May.Jun.12296836432Quality ControlRTYFeb.March.Apr.May.Jun.3254224351410671939PPMMin.3.51D MA I C 10 million per year Quality Assurance `03.01 Before`03.06 Result32545141000Target(PPM)`03.01 Before`03.06 Result1203260Target75%↓(minute )Process QualityLead time$ 9,020 KProductivity`03.1 Before`03.6 ResultTarget( EA/M.h)10.8% 7.48.08.284%↓18/19
    • 20. Future PlanYft1Yft2Yft3Yft4Yft5Yft6Yft1Yft2Yft3Yft4Building 2B lineBlock Design Block EMS EMS Supporting Build OTR & Jumbo Line Support procedure cut 30%ProjectBeforeAfter AchieveLine lengthNumber of procedureLQC employeeT / Time99.7 m72.0 m27 % ↓63 37 41 % ↓5 4 15 Sec13 Sec13 % ↓Productivity (EA/ MHr)3.5 6.6 88 % ↑0 % ↓Main IssueDMS Building19/19D MA I C Fast Growth Fast InnovationThank YouLGETA MWO
    • 21. (本页无文本内容)