You are on page 1of 25

Prepared By Checked By Approved By

TSP TSP TSP


ZAMRIE RH BOON HM THAM
QA S.O.M GM

MOHD. ZAMRI RH BOON HM THAM

TECK SEE PLASTIC


SDN. BHD

Project Number : 01
Project Name : To reduce LCD Chopin
Rear Cover Painting Defect
Green Belt Name : Mohd. Zamri
Champion : Mr. HM Tham
Department : QA
PROJECT MISSION
PROBLEM STATEMENT

1. Reject rate for Aug/Sept.2005 is 7.6 %


2. DPPM = 76,000
3. Many complaint from SDMA during LSR

MISSON STATEMENT

To reduce DPPM from 76000 to 19000


1. ( 75% improvement)

2. To reduce cost of poor quality & effects.

3. To minimize risk of NG parts skip to customer.


Project Drill Down Tree High Level Process Map - SIPOC

PROJECT STATEMENT Supplier


Supplier Input
Input Process
Process Output
Output Customer
Customer

Incoming Material
-Raw -PP & C Production -Outgoing
Part confirmation
TSP Big - Y TSP Mega material. -SOP
Spray
output inspection
-Part of -Specification Reject quantity -SDMA
Project parts -Semi painting
Production
-Equipme Finished Drying efficiency
TSP Small - Y TSP Q100 Project nt Goods Plant
-Sub -Special Production checker efficiency
store requirement
-Tooling Printing

Project Name : Assembly


To reduce LCD Chopin Inspection
Rear Cover Painting Packing
Defect End

Problem statement

SECONDARY PROCESS DEFECT TREND DEFECT CATEGORY BREAKDOWN


1800 20.00
Others, 10, 2%
Assembly, 6, 1%
1600 17.43 18.00
Moulding, 8, 1%
1400
16.00 Handling, 64, 12%
14.00
1200

1000
10.79 10.66 12.00
QUANTITY (PCS)

DEFECT %
10.00

800
7.95 7.62
6.75 8.00

600
5.45 6.09
6.00

400 4.01 4.00


Painting, 467 pcs = 84%
200 2.00

0 0.00 ( Data collection period : 15/8/05


1 2 3 4 5 6 7 8 9 until 9/9/05 )
Prod. Qty Rej.qty Def.%
( Date source : QC dept )
Project Justification
Expected
ResultCOPQ
Expected : Saving :
RM 106,756.67 / Year @ USD 28,093/ Year
(Baseline period Aug 05 until Sept 05) Data source PP&C & QC Dept

COPQ
Baseline DPPM : Target DPPM / 1000000 ]x Yearly production quantity x
Calculation
(Average Repair cost per unit + Labour Manhour Cost) x No. of manpower] =

76000 19000 / 100000 ]x 1,200,000 x [(RM 2.6077 + RM 15) x 1)


RM 106,756.67 / Year

Strategic Importance:
Enhance customer satisfaction
Good image Better business allocation
Profitable business
Maintain market competitiveness
Project Team
CHAMPION

Mr. HM Tham

Sr. Op. Manager

Mr. RH Boon

S-Green Belter
Mentor
Mr. Alfizal Mohd. Zamri

SUPPLIER &
CUSTOMER TECHNICAL PROCESS DATA ANALYSIS

Mr. Kamarudin Mr. Anand Mr. Cheng Ms. Azah


Control of incoming Control of tooling,
parts quality Data collection
,machine & manpower Process management
Liase with customer & Analyze data using
Analysis & improvement Secondary process
supplier for any quality statistical tool
on technical matter. improvement
issue.

Measure Baseline Performance


BASELINE
BASELINE (PERIOD : SEP 2005) TARGET
TARGET IMPROVEMENT
IMPROVEMENT
COPQ : USD 3153.49 / Month COPQ : USD 788.38 / Month COPQ Saving : USD 2365.11/Month
RM 11667.90 / Month RM 2918 / Month RM 8749.90 / Month
DPPM : 76000 DPPM : 19000 DPPM : 57000
Sigma Level : 2.93 Sigma Level : 3.57 Sigma Level : 0.64
Measure Baseline Performance
BASELINE TARGET
TARGET IMPROVEMENT
BASELINE (PERIOD : SEP 2005) IMPROVEMENT
COPQ : USD 3153.49 / Month COPQ : USD 788.38 / Month COPQ Saving : USD 2365.11/Month
RM 11667.90 / Month RM 2918 / Month RM 8749.90 / Month
DPPM : 76000 DPPM : 19000 DPPM : 57000
Sigma Level : 2.93 Sigma Level : 3.57 Sigma Level : 0.64
Problem statement

Pareto Chart of CHOPIN 17 RC DEFECT


600
100
500
80
400

Perc ent
Count

60
300

40
200

100 20

0 0
DEFECT

Count 132 69 69 67 62 53 42 22 12 21
Percent 24.0 12.6 12.6 12.2 11.3 9.7 7.7 4.0 2.2 3.8
Cum % 24.0 36.6 49.2 61.4 72.7 82.3 90.0 94.0 96.2 100.0

PAINTING DEFECT SAMPLE

Dust Overspray Sanding mark Rough surface Borderline Miss spray


Map Actual Process
INPUT START OUTPUT
V
A Incoming material Parts dusty & sanding mark
Bare part flashing requires N
to sanding V
A
N After secondary process
Overall molding part quality V
A Bare part check Sanding & parts quality
treatment
Paint material quality
Spray painting equipment
Spray booth condition V
A Good & NG part
Painting jig Spray painting
Spray method
N
V
Drying temperature V A Paint reliability
A Oven drying Rework
Drying time

Inspection method Input check Only good parts input to line


Judgment criteria N Segregate NG part to rework
V
V A
Incoming POP quality A Rubber foot assy
Assembly
Assembly method Button melting
V
A
Inspection method QC Inspection Good & NG part
N
Judgment criteria V V
A A
Packing Good part for delivery
Packing material condition
Judgment criteria
End
Map Actual Process

Injection Moulding Bare part Spray Input check

Oven Drying

Packing QC Inspection Button Melting Incoming Part


After Spray
Cause & Effect Diagram Function Deployment Matrix

Border line
Over spary

Rough surface
Customer

Sanding mark
DUST/FI OVER BORDER

Mis spray
Key

Dust
BER SPRAY LINE Process
Output
Masking condition Variable
Dust from paint due Improper masking due to
Improper paint filtering not fully cover
poor jig design Customer
Paint tank not clean Jig design Priority 8 8 8 8 8 8
Rank #
Hairline flashing support base touch Paint miss clogged at masking area
Air hose & gun dirty at parting line
Conveyor dirty
Key Process
Input Variable Association Table Rank
%
Rank
Spray method
Spray booth dusty Hanger dirty
Dust not properly Drying oven dusty Spray method
HIGH 1 1
10
1 1 1

22.1
Touch up without jig Spray method 0 0 0 0 0 480 4%
Clean after sanding Parts have high static Excess paint at jig PAINTI
charges NG NG cleaning after 1 1 11.4
sanding 0 0 0 0 5 6 248 4%

Poor sanding Paint viscosity DEFEC


1 1 11.0
method to thick T Masking NG 0 0 10 0 0 0 240 7%
Poor touch up pressure
Poor masking during 0
1
10 0 0
1

11.0
No proper clean Gun setting NG Jig design NG 0 0 240 7%
touch up
Wrong sand 1 1 11.0
paper Paint clog at jig 0
0 10 0 0 0
240 7%
grade 1000 Nozzle
Bare part 1
0 10 0
1
0
11.0
adjustment Gun setting NG 0 0 240 7%
poor
NG
SANDING incoming Poor paint filter 10 0 10 0 0 0 160 7.38%
ROUGH MISS
MARK quality
SURFACE SPRAY Hairline flashing 10 0 0 0 0 10 160 7.38%

Bare part NG 10 0 0 10 0 0 160 7.38%

KPIV & KPOV Preliminary Process FMEA


Customer Customer Customer Key Ra % Process
Potential Failure Modes Potential Failure SE Potential Causes of O
Current Process DE R
# Function C P
Key Process Output Priority Process Input nk (Step) (process defects) Effects (KPOVs) V Failure (KPIVs) C
Controls T N

Variable Rank # Variable 48


1 Incoming quality Flashing (Hairline) Dust fiber 8 Mold wear 6 Trimming 10 0

Dust 8 Poor touch up method 480 22.14% 2 Incoming quality Parting line high Dust fiber 8 Adjustment 4 QC inspection 6
19
2

20
24 11.07 3 Spray Rough surface Cosmetic reject 8 Spray method 5 QC inspection 5 0

Over spary 8 Masking NG 0 % 28


4 Spray Miss spray Cosmetic reject 8 Spray method 6 QC inspection 6 8

Border line 8 Jig design NG 240 11.07% 5 Spray Sanding mark Cosmetic reject 8 Spray method 4 QC inspection 8 25
6

Sanding mark 8 Paint clog at jig 240 11.07% 6 Spray Dust Cosmetic reject 8 High static charge 5 Ionizer 5
20
0

Rough surface 8 Gun setting NG 240 11.07% 7 Spray Sanding mark Cosmetic reject 8 Sanding material grade 5 Standardize grade 7 28
0

8 Spray Dust Cosmetic reject 8 Paint material 6 Paint filter 8 384

Mis spray 8 Poor paint filter 160 7.38%


9 Spray Dust Cosmetic reject 8 Spray gun 7 Service gun 9 504

Part Hairline flashing 160 7.38% 1


0 Spray Dust Cosmetic reject 8 Method 4 QC inspection 10
32
0

1
1 Spray Border line NG Cosmetic reject 8 Jig design 10 QC inspection 10 80
0

1 40
Spray Border line NG Cosmetic reject 8 Spray method 5 10
2 QC inspection 0
Measure Process Capability
For Convergence Dimension : Cp = 0.53, Cpk = 0.14 (short term)
Convergence Adjustment Process is CAPABLE

Process Capability Analysis for DAY

LSL USL
Process Data
USL 6
Within
Target *
LSL 0 Ov erall
Mean 5
Sample N 9
StDev (Within) 0.88652
StDev (Overall) 2.82532

Potential (Within) Capability


Cp 1.13
CPU 0.38
CPL 1.88
Cpk 0.38

-5 0 5 10 15
Cpm *

Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
Pp 0.35 PPM < LSL 0.00 PPM < LSL 0.01 PPM < LSL 38387.64
PPU 0.12 PPM > USL 333333.33 PPM > USL 129659.97 PPM > USL 361691.09
PPL 0.59 PPM Total 333333.33 PPM Total 129659.98 PPM Total 400078.73
Ppk 0.12
Determination of Vital Few Xs

Y (Painting defect) = f (x 1,x 2,x 3, x 8)

The Potential Factors are : To determine


the Vital Few
Factor 1, x 1 : Jig design NG
Xs
Factor 2, x 2 : Spray gun condition
Factor 3, x 3 : Incoming part flashing
Factor 4, x 4 : Spray method
Factor 5, x 5 : Paint filter NG
Factor 6, x 6 : Dust from spray method
Factor 7, x 7 : Miss spray from spray method
Factor 8, x 8 : Sanding material grade
Test of Theory 1

Theory : The Painting (over paint) defect is caused by X 1 jig design P-Value > 0.05 Reject Ho
P-Value < 0.05 Fail to reject Ho
Analysis Tool : 2 PROPORTION TEST
Ho : Over paint have a difference with jig design.
Ha : Over paint is have no difference with jig design.
Test and CI for Two Proportions overspray

Sample X N Sample p
1 401 500 0.802000
2 778 800 0.972500
Miss spray
Estimate for p(1) - p(2): -0.1705
95% CI for p(1) - p(2): (-0.207221, -0.133779)
Test for p(1) - p(2) = 0 (vs not = 0): Z = -9.10 P-Value = 0.000
Masking jig design
borderline
P-Value =0.00< 0.05 Fail to Reject Ho

Conclusion : The over paint HAVE DIFFERENCE with jig design. Thus, the X1 is the Vital X.

Test of Theory 2

Theory : The Painting defect (dust) is caused by X 2 spray gun condition P-Value > 0.05 Reject Ho
P-Value < 0.05 Fail to reject Ho
Analysis Tool : CHI SQUARE TEST
Ho : Painting defect is independent on spray gun filter.
H1 : Painting defect is not independent on spray gun filter.
Chi-Square Test: NO FILTERS, FILTER
Expected counts are printed below observed counts
NO FILTE FILTER Total
1 800 1200 2000
815.63 1184.37

2 56 43 99
40.37 58.63
Total 856 1243 2099
Dust
Chi-Sq = 0.299 + 0.206 + Spray gun
6.048 + 4.165 = 10.719DF = 1, P-Value = 0.001
P-Value =0.01 > 0.05 Fail to reject Ho

Conclusion : The painting dust is NOT independent on spray gun condition. Thus, the X1 is the Vital X.
Test of Theory 3

Theory : The Painting defect (dust) is caused by X 3 incoming part flashing P-Value > 0.05 Reject Ho

Analysis Tool : 2 PROPROTION TEST P-Value < 0.05 Fail to reject Ho

Ho : Painting defect have difference with part flashing.


Ha : Painting defect have no difference with part flashing.
Test and CI for Two Proportions

Sample X N Sample p
1 378 500 0.756000
2 983 1000 0.983000

Estimate for p(1) - p(2): -0.227 Dust


95% CI for p(1) - p(2): (-0.265489, -0.188511)
Test for p(1) - p(2) = 0 (vs not = 0): Z = -11.56
P-Value = 0.000

P-Value =0.000 < 0.05 Fail to reject Ho Part hairline flashing

Conclusion : The painting defect HAVE DIFFERENCE with part flashing. Thus, the X3 is the Vital X.

Test of Theory 4

P-Value > 0.05 Reject Ho


Theory : The Painting defect (dust) is caused by X 4 from spray method
P-Value < 0.05 Fail to reject Ho
Analysis Tool : CHI SQUARE TEST
Ho : Painting defect is independent on spray method
H1 : Painting defect is not independent on spray method
Expected counts are printed below observed counts
SPRAY METHOD SPRAY METHOD
NG OK Total
OK 17 20 37
18.50 18.50
NG 3 0 3
1.50 1.50
Total 20 20 40
Chi-Sq = 0.122 + 0.122 +1.500 + 1.500 = 3.243
DF = 1, P-Value = 0.072
Dust
Spray method
P-Value =0.072>0.05 Reject Ho

Conclusion : Test reject is independent on spray method. Thus, the X3 is NOT the vital X.
Test of Theory 5

Theory : The Painting defect (dust) is caused by X 5 paint filter size P-Value > 0.05 Reject Ho
P-Value < 0.05 Fail to reject Ho
Analysis Tool : CHI SQUARE TEST
Ho : Dust is independent on paint filter size .
H1 : Dust is not independent on paint filter size.
Chi-Square Test: FILTER SIZE 200 MICRON, 120 MICRON
Expected counts are printed below observed counts
OK NG Total
200 Mic 972 28 1000
937.50 62.50
100 Mic 903 97 1000
937.50 62.50
Total 1875 125 2000
Chi-Sq = 1.270 + 19.044 +
1.270 + 19.044 = 40.627
Filter mesh size Dust
DF = 1, P-Value = 0.000

P-Value =0.000<0.05 Fail to reject Ho

Conclusion : The painting defect is NOT INDEPENDENT on part flashing. Thus, the X3 is the Vital X.

Test of Theory 6

Theory : The Painting defect (dust) is caused by X 6 spray method P-Value > 0.05 Reject Ho

Analysis tool : 2 PROPORTION TEST P-Value < 0.05 Fail to reject Ho


Ho : Dust is independent on spray method.
H1 : Dust is not independent on spray method.
Test and CI for Two Proportions

Sample X N Sample p
1 311 500 0.622000
2 658 1000 0.658000

Estimate for p(1) - p(2): -0.036


95% CI for p(1) - p(2): (-0.0876802, 0.0156802)
Test for p(1) - p(2) = 0 (vs not = 0): Z = -1.37 P-Value = 0.172
Spray method Dust
P-Value =0.172 >0.05 Reject Ho

Conclusion : Test reject is independent on spray method.Thus, the X6 is NOT the vital X.
Test of Theory 7

Theory : The Painting defect (miss spray) is caused by X 7 spray method P-Value > 0.05 Reject Ho
P-Value < 0.05 Fail to reject Ho
Analysis tool : 2 PROPORTION TEST
Ho : Miss spray is independent on spray method.
H1 : Miss spray is not independent on spray method.
Test and CI for Two Proportions

Sample X N Sample p
1 997 1200 0.830833
2 1943 2400 0.809583

Estimate for p(1) - p(2): 0.02125


95% CI for p(1) - p(2): (-0.00514461, 0.0476446)
Test for p(1) - p(2) = 0 (vs not = 0): Z = 1.58 P-Value = 0.115
Spray method Miss spray
P-Value =0.115 >0.05 Reject Ho

Conclusion : Test reject is independent on spray method. Thus, the X7 is NOT the vital X.

Test ofTheory
Test of Theory
8 8
P-Value > 0.05 Reject Ho
Theory :The Painting defect (sanding mark & r.surface) is caused by X 8 sanding grade
P-Value < 0.05 Fail to reject Ho
Analysis Tool : CHI SQUARE TEST
Ho : Test reject is independent on sanding material grade.
H1 : Test reject is not independent on sanding material grade.
Chi-Square Test: SAND PAPER GRADE G 1000, G 1200
Expected counts are printed below observed counts
G 1000 G 1200 Total
OK 750 1008 1758
771.26 986.74
NG 84 59 143
62.74 80.26
Total 834 1067 1901
Chi-Sq = 0.586 + 0.458 +
7.207 + 5.633 = 13.885
DF = 1, P-Value = 0.000 Sanding mark
P-Value =0.00 >0.05 Fail reject Ho

Conclusion : Test reject is Not independent on sanding material grade. Thus, the X8 is the vital X.
Vital Few Xs

Y (Painting defects) = f (x 1, x 2, x 3, x 4, x 5, x 7)

The Vital Few Xs are:


Vital Factor 1, x1 : Jig design
Vital Factor 2, x2 : Spray gun condition
Vital Factor 3, x3 : Part flashing
Vital Factor 5, x5 : Paint material
Vital Factor 8, x8 : Sanding material
grade
Improvement #1 By Using Non-DOE
Improvement of Factor, X 1 Jig design NG ( Over spray):

Before After

Ventilation hole area not cover causing the paint mist Modify jig design cover the ventilation hole area to
Stick to part surface prevent paint mist from sticking to part surface

Improvement #2 By Using Non-DOE


Improvement of Factor, X 1 Jig design NG ( Over paint):

Before After

The part holder area touching the parting line area & Modify the jigRedesign of part holder area not to
causing to paint residue to stick at part. touch parting line area to eliminate painting defect.
Improvement #3 By Using Non-DOE
Improvement of Factor, X 1 Jig design NG
( Borderline)
Masking jig cover shorter
Before After Masking jig cover longer
Paint mist
Paint mist

The extended paint jig


Part cover prevent the
The paint filling Part
paint from Filling the
borderline borderline
Failure rate : 1.9% Failure rate : 0.7%

The jig design NGParting line not fully cover causing Modify the jigRedesign of parting line area coverage
the paint able to penetrate causing to borderline NG. to ensure paint cant penetrate during spray process.

Improvement #4 By Using Non-DOE


Improvement of Factor, X 2 Spray Gun Type

Before After

The spray gun type have no filter to prevent the dust Change spray gun that have filter to ensure paint
from sprayed to the part surface. application is fully filtered before spraying to the part.
Improvement #5 By Using Non-DOE
Improvement of Factor, X 5 paint filter
size
Before After

Failure rate : 1.9% Failure rate : 0.7%

Paint material contain rough particle that turn into Change filter mesh to smaller type ( 200mic) to
spray dust & filter mesh 120mic cant properly filter ensure rough particle in the paint is properly filtered

Improvement #6 By Using Non-DOE


Improvement of Factor, X 3 part flashing

Before After

Part have hairline flashing that will turn fiber after Repair the mold & ensure the part running is free from
attack by thinner in the paint during spray process. any hairline flashing
Improvement #7 By Using Non-DOE
Improvement of Factor, X 5 sand paper
grade
Failure rate : 1.02% Before After Failure rate :
0.03%

Part have many sanding mark & Sanding mark & rough surface
rough surface when using sand reduce after using sand paper
paper grade 1000 grade 1500
Summary of Improve Phase Control Plan

Summary of Improvement

UPDATE FMEA
Process Current D
Potential Failure Modes Potential Failure Effects Potential Causes of OC
Function SEV Process E RPN
(process defects) (KPOVs) Failure (KPIVs) C
(Step) Controls T

Incoming
Flashing (Hairline) Dust fiber 8 Mold wear 2 Repair mold 3 48
quality

Incoming
Parting line high Dust fiber 8 Adjustment 3 QC inspection 3 72
quality

Use 1500
Spray Rough surface Cosmetic reject 8 Spray method 3 3 72
sandpaper

Spray Miss spray Cosmetic reject 8 Spray method 1 QC inspection 5 40

Use 1500
Spray Sanding mark Cosmetic reject 8 Spray method 2 2 32
sandpaper

Spray Dust Cosmetic reject 8 High static charge 2 Ionizer 2 32

Sanding material Standardize


Spray Sanding mark Cosmetic reject 8 2 2 32
grade grade

Spray Dust Cosmetic reject 8 Paint material 3 Paint filter 5 120

Spray Dust Cosmetic reject 8 Spray gun 4 Change filter 4 128

Spray Dust Cosmetic reject 8 Method 4 QC inspection 2 64

Spray Border line NG Cosmetic reject 8 Jig design 3 Modify jig 4 96

Spray Border line NG Cosmetic reject 8 Spray method 2 QC inspection 4 64


Result of Control
COMPARISON BETWEEN BEFORE AND AFTER IMPROVEMENT

Process Capability Analysis for DAY


G-GAIN (NEW)

LSL USL LSL Target USL


Process Data Process Data
USL 6 USL 50.000
Within Within
Target 42.500
Target *
LSL 0 Overall LSL 35.000 Overall
Mean 5 Mean 42.375
Sample N 9 Sample N 100
StDev (Within) 0.88652 StDev (Within) 1.65429
StDev (Overall) 2.82532 StDev (Overall) 1.65429

Potential (Within) Capability Potential (Within) Capability


Cp 1.13 Cp 1.51
CPU 0.38 CPU 1.54
CPL 1.88 CPL 1.49
Cpk 0.38 Cpk 1.49
-5 0 5 10 15
Cpm * Cpm 1.51
35 37 39 41 43 45 47 49 51

Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
Pp 0.35 PPM < LSL 0.00 PPM < LSL 0.01 PPM < LSL 38387.64
Pp 1.51 PPM < LSL 0.00 PPM < LSL 4.13 PPM < LSL 4.13
PPU 0.12 PPM > USL 333333.33 PPM > USL 129659.97 PPM > USL 361691.09
PPU 1.54 PPM > USL 0.00 PPM > USL 2.02 PPM > USL 2.02
PPL 0.59 PPM Total 333333.33 PPM Total 129659.98 PPM Total 400078.73
PPL 1.49 PPM Total 0.00 PPM Total 6.16 PPM Total 6.16
Ppk 0.12
Ppk 1.49

Items Before After Improvement : Improve from 76,000 dppm to 17000


% 7.60 1.70 dppm

DPPM 76000 17000


Conclusion : The project carried out is on the right
Track.
PROJECT ACHIEVEMENT STATUS

6 Activity Duration

D M A I C

TARGET :DPPM 19000


ACTUAL : 170000

PROJECT ACHIEVEMENT
BASELINE TARGET ACTUAL IMPROVEMENT
ACHIEVEMENT

COPQ : USD 2341.08 /Month COPQ : USD 788.38 /Month COPQ : USD 985.75 COPQ Saving : USD 1355.33 / Month
RM 8896.03/Month RM 2918 / Month RM 3745.85 RM 5150.25 / Month
DPPM : 76000 DPPM : 19000 DPPM : 17000 DPPM : 57000
Sigma Level : 2.93 Sigma Level : 3.57 Sigma Level : 3.62 Sigma Level : 0.69

You might also like