Eliminate Coil Wire Breakage on Passive Anti-Theft System (PATS)

Size: px

Start display at page:

Download "Eliminate Coil Wire Breakage on Passive Anti-Theft System (PATS)"

Nancy Weaver
5 years ago
Views:

1 Eliminate Coil Wire Breakage on Passive Anti-Theft System (PATS) 1

2 Six Sigma Team Passive Anti-Theft System Black Belt Candidate: Sam. H. OEM Master Black Belt: Tom R. OEM Project Champion: Greg B. - OEM Black Belt: Wayne T. - OMRON Project Champion: Warren. W. OMRON Process Owner: Tom S. OMRON Financial Analyst: Bill N. OMRON 2

DEFINE VOICE OF THE CUSTOMER Electronics PATS Functional Area Name Passive Anti-Theft System Wayne Tollefsen PROJECT CLASSIFICATION: Passive Anti-Theft System (PATS).

3 DEFINE VOICE OF THE CUSTOMER Electronics PATS Functional Area Name Passive Anti-Theft System Wayne Tollefsen PROJECT CLASSIFICATION: Passive Anti-Theft System (PATS). TREND CHARTS and BREAKDOWN OF ISSUE: New component 55 parts (50 Plant & 5 Warranty) returned for no function due to broken coil wire. Y=f(x) CASCADE: See Fishbone Diagram The wire breaks here when the PATS is flexed during assembly at the customer location. VOICE OF THE CUSTOMER: Parts may not operate when compressed into the Key Lock Cylinder during assembly. Failure of the part to operate after assembly has been determined to be due to a broken coil wire. Failures have been observed at both the assembly plant and in warranty. CTQ STATEMENT (Customer Requirement): Part must function as designed. Part should be robust to multiple platform applications. DEFECT DEFINITION for Y (Engineering Metric): Parts Inoperable. When the part is compressed into the Key Lock Cylinder the coil wire on the outside location breaks. For analysis purposes we will use only one opportunity per part. COST OF POOR QUALITY: $900,000 PROBLEM STATEMENT, SCOPE, AND GOAL: 55 parts have been rejected due no function, 50 plant and 5 warranty returns. In each case, no function was due to broken coil wire. 3

4 MEASURE CTQ (y) CAPABILITY Broken Coil Wire on PATS C & E Diagram RED Indicates primary areas of investigation. 4

5 ANALYZE y=f(x) Potential Critical Inputs to be investigated: Parts handling Coil Winder Coil Pin deflection to increase coil wire elongation via process modification. Increase coil wire elongation by optimizing Coil Winding process using Design of Experiments. Plastic Coil Design Remove plastic strain relief. 5

6 ANALYZE y=f(x) Process Flow: OMRON RED Indicates primary areas of investigation. 6

Customer Tier 1 Parts test OK Tier 1 3XX: Key Lock Cylinder without

7 ANALYZE y=f(x) Process Flow: Supplier(s) to End Customer Defective parts are encountered at Tier 1 during assembly into Instrument Panel. Customer Tier 1 Parts test OK Tier 1 3XX: Key Lock Cylinder without ridge Tier 2 Parts test OK OMRON Parts test OK 2XX: Key Lock Cylinder with ridge 7

8 ANALYZE y=f(x) PATS Key Lock Cylinders used on 3XX & 2XX 6.98mm 3.34mm 3XX: 6.98mm of deflection 2XX: 3.34mm of deflection 8

9 ANALYZE y=f(x) Coil Pin deflection during coil winding process. 0mm 1mm 2mm Deflection Data Summary Average Elongation Std. Dev. 1mm 6.34mm 2.62mm 2mm 7.38mm 1.84mm Ref Capability Studies on slides 13 & 14. 9

10 ANALYZE y=f(x) PATS Coil Winder -- Coil Pin 1mm Process Data LSL Target USL USL Target LSL Within Overall Mean Sample N 24 StDev (Within) StDev (Overall) Potential (Within) Capability Cp 0.87 CPU 0.90 CPL 0.85 Cpk 0.85 Cpm Pp PPU PPL Ppk Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance PPM < LSL PPM > USL PPM Total PPM < LSL PPM > USL PPM Total PPM < LSL PPM > USL PPM Total

11 ANALYZE y=f(x) PATS Coil Winder -- Coil Pin 2mm Process Data LSL Target USL USL Target LSL Mean Within Overall Sample N StDev (Within) StDev (Overall) Potential (Within) Capability Cp 0.93 CPU CPL Cpk 0.72 Cpm Pp PPU PPL Ppk Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance PPM < LSL PPM > USL PPM Total PPM < LSL PPM > USL PPM Total PPM < LSL PPM > USL PPM Total

12 ANALYZE y=f(x) Coil Pin Deflection Experiment Conclusion After conducting the experiment on Coil Pin Deflection it was determined that there was no way to assure the integrity of the solder joint between Coil Pin and PCB when the Coil Pin is bent 2mm. Therefore the team decided not to change the Coil Winding process to include Coil Pin Deflection 12

13 ANALYZE y=f(x) Potential Critical Inputs to be investigated: Parts handling Coil Winder Coil Pin deflection to increase coil wire elongation via process modification = No Effect on Failure Mode Increase coil wire elongation by optimizing Coil Winding process using Design of Experiments. Plastic Coil Design Remove plastic strain relief. 13

14 IMPROVE y=f(x) The team further speculated that the coil wire elongation could be increased by optimizing the Coil Winder process. Combining brainstorming and process knowledge, the team identified three (3) factors and levels: Tension (g) Speed 1 (rpm) Speed 2 (rpm) Note: Level 0 is a center point, which represents standard operating condition. 14

15 IMPROVE y=f(x) Each of the three factors was studied at two levels so the arrangement is a (2 x 2 x 2) or 2 3 = 8 run factorial design. The eight combinations are shown in the following table with the factor levels coded by plus and minus signs. 15

16 IMPROVE y=f(x) Standard Order Tension Speed 1 Speed A 2 3 factorial design coded in standard order with center point. Note: run 0 is a center point, which represents standard operating condition. 16

17 IMPROVE y=f(x) The eight-factor combinations and center point can be conveniently represented geometrically by the vertices of a cube. If you imagine the center of the cube to be the origin of a three dimensional coordinate system, then the eight-factor combinations can be identified by runs 1 through 8. 17

18 IMPROVE y=f(x) Centerpoint Factorial Point Cube Plot for Standard Run Order Speed Speed Tension 18

19 IMPROVE y=f(x) Data collected from 2-Level, 3 Factor Experiment with center points: Std Order Tension (g) Speed 1(rpm) Speed 2(rpm) Elongation Note: run 0 is a center point, which represents standard operating condition. 19

20 IMPROVE y=f(x) Centerpoint Factorial Point Cube Plot for Elongation PATS Coil Winder DOE -- Phase Speed Tension Speed 2 20

21 IMPROVE y=f(x) The main effects, (i.e. average rate of change over the factor levels studied), are calculated as follows: Factors: Names: A B C Tension Speed 1 Speed 2 Std Run Order Average 170g 190g Avg Effect The main effects for factor A, Tension, is the difference between the average at the high level, (190g) and the low level (170g), i.e mm 5.890mm = mm. Elongation is decreased on the average 0.774mm when Tension is increased from 170g and 190g. The main effects, are summarized on the following slide. 21

22 IMPROVE y=f(x) PATS Coil Winder DOE Main Effects Plots Centerpoint Centerpoint Average = 5.370g Elongation Tension (g) Speed 1 (rpm) Speed 2 (rpm) The slope of the line drawn between the two levels is indicative of the significance level of each factor. Clearly Tension is considerably more significant that either Speed 1 or Speed 2. 22

23 IMPROVE y=f(x) The interaction effects are calculated as follows: AB AC BC ABC Interaction Effects The Interaction effects, are summarized on the following slide. 23

24 IMPROVE y=f(x) Centerpoint PATS Coil Winder DOE Interaction Plot Tension Speed Speed 2 24

25 IMPROVE y=f(x) The effects and p-values are summarized below using a normalized T-statistic Estimated Effects and Coefficients for Elongation (coded units) Term Effect Coef SE Coef T P Constant Tension Speed Speed Tension*Speed Tension*Speed Speed 1*Speed Tension*Speed 1*Speed Ct Pt The probability values (p-values) for Tension and the 3-way interaction between Tension, Speed 1 and Speed 2 are less than 5% and deemed statistically significant effects. Center points (Ct Pt) representing standard operating conditions, were not significant. 25

26 IMPROVE y=f(x) A Pareto chart of the effects vs. Decision Limit demonstrates that Effect A, Tension, and the ABC interaction are significant at the 5% level. Effects Pareto Chart Effect Values Effects Decision Limits Tension ABC AB BC Speed 1 AC Speed 2 Factors 26

27 IMPROVE y=f(x) The main effects plot for Tension demonstrates that increased elongation is obtained from a low Tension setting, 170g. Since the other significant effect is the three way interaction, we must now evaluate the effect of Tension on the other two factors, Speed 1 and Speed 2. 27

28 IMPROVE y=f(x) An interaction is best represented by a response plot, which provides a visual depiction of the response surface. Setting Tension at its low level, 170g, produces the following Response Plot (3D) and Contour Plot (2D) for Speed 1 and Speed 2. 28

29 IMPROVE y=f(x) PATS Coil Winder DOE Response Surface Plot Elongation Speed Speed 2 Hold values: Tension:

30 IMPROVE y=f(x) PATS Coil Winder DOE Contour Plot Speed Speed Hold values: Tension:

31 IMPROVE y=f(x) Careful analysis of the Response Plot and Contour Plot demonstrates that increased elongation can be obtained from the following process setting: Factor A, Tension, set at 170g. Process settings for Factors B and C, Speed 1 and Speed 2 respectively, will be determined by running a confirmation experiment. 31

32 IMPROVE y=f(x) A 2-level, 2 Factor confirmation experiment was conducted keeping Tension, Factor A at 170g and varying Speed 1 and Speed 2 as follows. Speed 1(rpm) Speed 2 (rpm) Note: midpoints, (500 & 1100) are center points, which represents standard operating condition. 32

33 IMPROVE y=f(x) 2 factors at 2 levels result in the following 4-run design matrix. Standard Order Speed 1 Speed Note: run 0 is a center point, which represents standard operating condition. 33

34 IMPROVE y=f(x) Data collected from 2-Level, 2 Factor Confirmation Experiment with center points: Standard Order Speed 1 Speed 2 Elongation Note: run 0 = center point 34

35 IMPROVE y=f(x) Centerpoint Factorial Point Square Plot (data means) for Confirmation DOE Speed Speed 1 35

36 IMPROVE y=f(x) The main effects, (i.e.average rate of change over the factor levels studied), are calculated as follows: Speed 1 Speed 2 AB Run Average Avg Effect

37 Centerpoint Center point Avg = IMPROVE y=f(x) Main Effects Plot (data means) for Confirmation DOE Spindle Speed 1 Speed 2 37

38 IMPROVE y=f(x) The effects and p-values are summarized below using a normalized T-statistic Term Effect Coef SE Coef T P Constant Speed Speed Speed 1*Speed Ct Pt The probability values (p-values) for Speed 1, Speed 2 and the Interaction are greater than 5% and cannot be considered statistically significant. Center points (Ct Pt) representing standard operating conditions, were not significant. 38

39 IMPROVE y=f(x) A Pareto chart of the effects vs. Decision Limit demonstrates that neither Speed 1, Speed 2 or the Interaction Effects are significant at the 5% level. PATS Coil Winder Confirmation DOE Effects Pareto Effect Value Effects Decision Limit Speed 1 AB Speed 2 Factors 39

40 IMPROVE y=f(x) PATS Coil Winder Confirmation DOE Response Surface Plot of Elongation Elongation Speed Speed 2 40

41 Speed IMPROVE y=f(x) PATS Coil Winder Confirmation DOE Contour Plot of Elongation Speed 1 Optimum setting for maximum elongation 41

42 IMPROVE y=f(x) Predicted Optimum of Response Surface = mm Optimal D Hi Cur Lo Speed Speed [200.0] [850.0] Elongati Maximum y = d =

43 IMPROVE y=f(x) PATS Coil Winder DOE Conclusions: Improved elongation performance is achieved with 1.Tension at low level, (170g); 2.Speed 1 at low level (200 rpm); 3.Speed 2 at high level (850 rpm); However, achieving elongation which exceeds 7mm is not possible with the PATS coil winder. 43

44 ANALYZE y=f(x) Potential Critical Inputs to be investigated: Parts handling Develop work instructions depicting proper handling of PATS, (OEM responsible) Coil Winder Coil Pin deflection to increase coil wire elongation via process modification = No Effect on Failure Mode Increase coil wire elongation by optimizing Coil Winding process using Design of Experiments = No Effect on Failure Mode Plastic Coil Design Remove plastic strain relief, (Omron responsible) 44

45 CONTROL X s Control 1: Parts Handling Work Instructions 45

46 CONTROL X s Control 1: Parts Handling; Squeezing the PATS Coil and the Key Lock Cylinder together. 46

47 Next Steps Team Recommendations Design change to remove the outer strain relief tab from the plastic spool part and then use an insert pin to mimic the outer strain relief tab during coil winding and solder. Design change completed, DV testing completed with zero defects reported. Safe launch at Tier 2 and Tier 1, 135 parts assembled, zero defects reported. Safe launch completed at OEM Assembly Plant, 100 vehicles assembled, zero defects reported. 47

48 Conclusion Congratulations to the Team Each and every team member contributed to the success of this project. 48

EC Timer Box Product Design for Six Sigma

EC Timer Box Product Design for Six Sigma Application of Product Customers: Whirlpool, Bosch, AEG, and others 2 Goals To Decrease the reject rate by a factor of 10 from 10% to less than 1% 3 Test machine