contoh kasus six sigma dipublikasi oleh
MENURUNKAN CACAT NO POWER PADA PRODUKSI
TV
DENGAN METODA SIX SIGMA
DI PT LG ELECTRONICS INDONESIA
A R I P I N
2002 - 21 – 140
TEKNIK INDUSTRI
UNIVERSITAS INDONUSA ESA UNGGUL
2004
Konsep Six Sigma
Six Sigma : Suatu sistem yang flexibel untuk memperbaiki kinerja
yang mendukung sukses secara terus-menerus.
Six Sigma dibangun berdasarkan banyak ide manajemen yang penting dan
praktek terbaik dari metoda sebelumnya
Sigma adalah ukuran satuan statistik yang menggambarkan
kemampuan suatu proses dan ukuran nilai sigma dinyatakan dalam
DPU (Defect Per Unit) atau PPM
Dapat dikatakan bahwa proses dengan nilai sigma lebih tinggi
dalam proses akan mempunyai cacat yang lebih sedikit
Makin bertambah nilai sigma,
makin berkurang biaya kualitas dan waktu siklus
Konsep Six Sigma
Menggunakan Z sebagai satuan ukuran kemampuan proses
•Z =3
Sample’s average :
LSL
n
USL
=
X=
Xi
i=1
n
Sample’s standard deviation :
n
( Xi - X )
1σ
2σ
=s=
3σ
2
i =1
n-1
Nilai Z (sigma level)
Z =
Kemampuan 3s
•Z =6
x -
USL
LSL
Bila keragaman / variasi turun, maka
kemungkinan cacat juga turun,
maka kemampuan (capability)
1σ 2σ 3σ 4σ 5σ 6σ
Kemampuan 6s
meningkat
Konsep Six Sigma
DEFINE
• Standardize
solution
• Establish standard
measure to
maintain
performance
CONTROL
• Develop ideas to cut
of root causes
• Optimation solution
• Measure solution
IMPROVE
• Identify Customer’s CTQ
• Develop a team charter
• Defne and build a process
map
• Validate problem
/ process
• Measure key
problem
• Current
MEASURE condition
• Screening potential
factor
• Identify vital few root
causes
• Validate hypothesis
ANALYZE
Konsep Six Sigma
Geser
menuju
Target
Tepat namun tidak akurat
Shifting/bergeser ke Target
& Mereduksi Variasi
T
USL
USL
T
USL
μ
USL
μ
Object
dari 6
adalah
Akurat namun tidak tepat
Shift(geser)
ke target
T
Menurunkan
Variasi
USL
USL
μ
Turunkan
Variasi
Aplikasi detail Six Sigma
Perbaikan dengan Metoda Six Sigma ( DMAIC )
PT LG ELECTRONICS INDONESIA
REDUCE SMPS TRANSFORMER
SHORT ( NO POWER )
2004 06 14
Incoming Quality Control
Project Registration Sheet
Theme
Name
Project
Period
Quantity
Target
REDUCE SMPS TRANSFORMER
SHORT ( NO POWER )
Theme
Measurement February 3 ~ 6, 2004
Outline
Analysis
February 9 ~ 21, 2004
Improvement
February 24 ~ 28, 2004
Control
March ~ May, 2004
Current
Target
Improvement
Ratio (%)
840 PPM
0 PPM
100 %
6
Anticipation Short Test as check point at line
Effect
production of Samwha and IQC LGEIN.
Status
&
Problem
• Material Checking method
• Quality System set up
WIRE HOLE
WIRING CROSS
CHECKING PROSSES
Name
Member
Role
Aripin
$29000
Customer’s
Needs
Definition
Good reliability part
Telp.
QA
212
100%
Amount
Of target
(US$/Year)
$0
Before
Great Company Great People
Dept
2 - 22
After
Fun Innovation By New Culture
Problem Situation
Management policy said we have reduce field claim.
Data field claim show No Power is cause biggest
than other.
No power is TV Set have no supply voltage. It show by not light of Indicator power and no display on tube.
No power can caused by parts, like power cord, SMPS Transformer, IC, FBT, Transistor and others.
SMPS Transformer have function as divider power voltage to several sub circuit or part like FBT, IC, and
others.
If SMPS not function, it can’t supply voltage to parts or sub circuit .
Defect Analysis by No Power
Data field claim 2003
No Power
5 WORST SYMPTOM 29%
IN 2003
No Sound
Knowledge
Picture NG
CPT NG
10%
15%
Sound
SoundNo
NG
15%
A/V NG
No Power
42%
8.4%
Screen
16%
8%
27%
IC
5.2%
Crystal
3.4%
FBT
2.0%
Others
840 PPM
100%
13%
Diode
Resistor
70.2%
6170VMCBO1L
15%
Transistor
12%
Target
29%
SMPS
Fuse
14%
Defect Analysis by Model
26.3%
6170VMCBO1K
0PPM
Others
2.0%
3.6%
3.5%
Before
3 - 22
After
Theme Selection
Big Y
Little y1
Little y2
Improve
Improve
TV
TV Quality
Quality
Attribute
Attribute
30%
30% SCR
SCR
Reducing
Reducing
Knowledge
Part
TYPE
PIC
Statu
s
Crack, Ball
Opt. Skill, Ability
Wire Connection
Handling
Handling
Opt. Skill, Ability
Wrong Connection
Handling
1P1P
LQC
Finished
1P1P
LQC
On Progress
1P1P
1P1P
IQC
On Progress
Production On Progress
IIC
1P1P
Production Finished
Production On Progress
IIC
IQC/Prod
Finished
AGC Circuit
1P1P
IQC
On Progress
Opt. Skill, Ability
1P1P
Production On Progress
Antenna
How to Use
No connection
1P1P
OQC
On progress
How to Use
OQC
Product Spec
Electricity
How to used
Plug Connection
1P1P
1P1P
On Progress
On Progress
Environment
Environment
IC
Transistor
CPT
Chassis
Soldering
Process
Little y3
Adjustment
Speaker Connection
PWB Crack
Loosen Part
Bad Wiring Connection
PWB Short
Front Cabinet Dented
AGC Circuit
Bad System Connector
FBT
Tuner
Speaker
SMPS Trans
Reliability
Screen Focus, DYC,PRT
ABL Short
AGC, No Ruster
No Sound Problem
No Power
Fuse
Capacitor
Un Connect
Reliability
Reliability
Resistor
4 - 22
1P1P
OQC
OQC
1P1P
1P1P
1P1P
OQC
IQC
IQC
On Progress
On Progress
Finished
1P1P
IQC
Finished
IIC
1P1P
IQC
IQC
Finished
On Progress
1P1P
IIC
IQC
IQC
Finished
Finished
1P1P
IQC
On Progress
On Progress
Process Mapping
SAMWHA
Process
LGEIN IQC
LGEIN LQC
LGEIN OQC
•R&D
• Appearance
• Patrol
• ELT
• Purchase
• Dimension
• Performance
• LTLT
•IQC
• Structure
Final
Inspect
• Reliability
Product
•Production
• Performance
• LQC
• Assembling
• OQC
• Reliable Part
High defect
SMPS no power
At LGEIN production
Find Out what is
the Main Cause of
Problem
Selecting Best
Condition
for reliable Part
• OS&D
5 - 22
What is SMPS Transformer ?
9
7
SMPS
8
10
5
1
Output of SMPS (Voltage)
• Pin 9 - 8 : 24 ~ 25
• Pin 10 - 11 : 10 ~ 12
• Pin 14 - 13 : 110 ~ 115
• Pin 13 - 12 : 13 ~ 14
• Pin 1 – 2 : 7 ~ 10
14
13
SMPS Component :
• Core
• Bobbin
• Wire
• Insulation
• Tape
• Shield
12
2
Potential Short
Area
Short can be happen if one of wire on potential area
get direct connect.
6 - 22
Clarification Of Y
Kind of defect by SMPS
MARKING NG
APPEARANCE
Master
Signal
RUST
ISOLATION NG
SMPS
DEFECT
DIMENSION
Test
Signal
Test Signal
P/P OVER
NG SIGNAL DISPLAY
WIRE SHORT
PERFORMANCE
Master
Signal=
OK SIGNAL DISPLAY
No
Power
BROKEN
SMPS Transformer Characteristic
Display of Surge Comparison Tester ( ST-215 ) will
show two signal both master and part test.
If form signal different that’s part NG
and if same is OK.
TURNS
Demonstration of Data gathering
Display
Indicator
TV Set Power OK
TV Set No Power
7 - 22
Measurement (Gage R&R)
Instrument : Surge Comparison Tester ( ST-215 )
Agreement Analysis
21
22
23
24
25
Attribute
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Try 1
Pass
Pass
Pass
Fail
Try 2
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Within Appraisers
Fail
Pass
Pass
Fail
Fail
OPR 2
Try 1
Pass
Pass
Pass
Fail
Try 2
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Fail
Pass
Pass
Fail
Fail
Try 1
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Try 2
Pass
Pass
Pass
Fail
Within Appraisers
Fail
Pass
Pass
Fail
Fail
96
96
:
# Matched
Percent
25
25
100.00
(86.09, 100.00)
OPR-2
25
25
100.00
(86.09, 100.00)
OPR-3
25
25
100.00
(86.09, 100.00)
88
88
OPR-1
25
25
100.00
(86.09, 100.00)
OPR-2
25
25
100.00
(86.09, 100.00)
OPR-3
25
25
100.00
(86.09, 100.00)
Assessment Agreement
# Matched
Percent
25
100.00
OPR-3
Measurement System are
Acceptable
All appraisers' assessments agree with each other.
25
OPR-2
Appraiser
95 % CI
Between Appraisers :
# Inspected
OPR-1
Instrument is valid and reliable
Appraiser's assessment across trials agrees with the known standard
Percent
OPR-3
Operator is good in inspection
Each Appraiser vs Standard :
# Matched
86
OPR-2
Appraiser
Assessment Agreement : 100 %
95 % CI
# Inspected
92
90
OPR-1
# Inspected
94
90
86
OPR-1
Appraiser
98
92
95.0% C I
P ercent
100
98
94
JA NUA RY 15. 2004
A RIPIN
SHORT TEST
Appraiser vs Standard
95.0% C I
P ercent
100
Appraiser agrees with himself across trial
Appraiser
Date of study :
Reported by :
Name of product:
Misc:
Assessment Agreement
OPR 3
Percent
Sample
1
2
3
4
OPR 1
Percent
Known Population
95 % CI
(86.09, 100.00)
8 - 22
Capability Process (Current)
Based on data October ~ December ‘03, defect SMPS 124 ea from 147595 ea.
• Defect Op. = 1 ( Wire Short )
• DPU
= 124 : 147595
• DPO
= 124 : ( 147595 x 1 ) = 8.4E-04
0.04
• DPMO = DPO x 1.000.000 = 840 PPM
From Z Normal table, we get :
DPO
= 8.4E-04
Z Value
= 3.1 + 0.04
= 3.14
Our Condition now :
Sigma level
= 3.14+ 1.5
= 4.64
3.1
9 - 22
8.45E-04
Estimate Cause Analysis
By Analyzing all potential problem that could have significant effect for SMPS Performance,
thru Brainstorming and screening, we succeed to collect some root cause for the problem
Potential Factor
Dielectric Breakdown
wire
Material
Short
(No Power)
Hole
Core
Varnish
Press
Turn wiring
Space
Home pin
Finishing
Method
10 - 22
Selecting Vital Factor
X1 = Dielectric Breakdown
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between D/B below 2250 V and D/B above 2250 V)
Ho : P≠P ( There are different Proportion Between D/B below 2250 V and above 2250 V)
D/B : 2250 V
Sample
2250
Decision
Sample
2250
Decision
1
2235
FAIL
14
1850
FAIL
2
1990
FAIL
15
2230
FAIL
Sample
X
N
Sample p
3
2010
FAIL
16
1990
FAIL
4
2240
FAIL
17
2010
FAIL
1
0
25
0.000000
5
2125
FAIL
18
1890
FAIL
2
19
25
0.760000
6
2150
FAIL
19
2110
FAIL
Difference = p (1) - p (2)
7
2250
FAIL
20
2225
FAIL
8
1990
FAIL
21
2000
FAIL
Estimate for difference:
FAIL
22
2010
FAIL
95% CI for difference:
23
1995
FAIL
9
2000
10
2230
11
2000
FAIL
Sample
2250
1985
FAIL
1960 1
FAIL5430
2
4795
12
13
D/B : 2250
V
FAIL
24
Decision
25
PASS
1890
FAIL
Sample
2250
1885
FAIL
14
4300
Test and CI for Two Proportions
-0.76
(-0.927414, -0.592586)
Test for difference = 0 (vs not = 0):
Z = -8.90
Decision
PASS
PASS
15
5800
PASS
3
4500
FAIL
16
7800
FAIL
4
5350
PASS
17
7900
PASS
5
6000
FAIL
18
4600
FAIL
6
5100
PASS
19
5800
PASS
7
5700
PASS
20
5500
PASS
8
5700
PASS
21
7500
PASS
9
4500
PASS
22
7900
PASS
10
5000
PASS
23
5600
PASS
11
4100
PASS
24
4500
FAIL
12
4250
FAIL
25
6500
PASS
13
5600
PASS
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Dielectric breakdown
has an effect or SMPS reliability.
Dielectric Breakdown is Vital Few factor
11 - 22
Selecting Vital Factor
X2 = Pin Hole Wire
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between Pin Hole Below 5 and Above 5)
Ho : P≠P ( There are different Proportion Between Pin Hole Below 5 and Above 5)
P/H : 5
Test and CI for Two Proportions
Sample
5
Decision
Sample
5
Decision
1
0
PASS
14
1
PASS
Sample
2
0
PASS
15
1
PASS
3
1
PASS
16
0
PASS
4
0
PASS
17
0
PASS
5
0
PASS
18
0
PASS
Difference = p (1) - p (2)
6
2
PASS
19
1
PASS
Estimate for difference:
7
1
1
PASS
8
3
9
P/H : 205
PASS
1
PASS
Sample
PASS 1
21
5
22
7
1
Decision
0
PASS
10
1
PASS 2
11
0
PASS 3
23
7
24
8
12
3
PASS 4
13
0
PASS 5
6
X
N
Sample p
1
25
25
1.000000
2
14
25
0.560000
95% CI for difference:
PASS
Sample
PASS14
5
Decision
6
PASS
0
PASS
0
FAIL
PASS15
PASS16
7
PASS
7
FAIL
25
7
1
FAIL
PASS17
6
PASS
6
PASS
18
7
PASS
6
FAIL
19
8
FAIL
7
8
PASS
20
8
FAIL
8
8
FAIL
21
6
PASS
9
6
PASS
22
7
PASS
10
9
PASS
23
9
FAIL
11
10
FAIL
24
7
PASS
12
11
FAIL
25
8
PASS
13
10
FAIL
0.44
(0.245420, 0.634580)
Test for difference = 0 (vs not = 0):
Z = 3.76
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Pin Hole Wire
has an effect or SMPS reliability.
Pin Hole Wire is Vital Few factor
12 - 22
Selecting Vital Factor
X3 = Pressure
Hypothesis Testing
H0 : P 1 = P 2 (There are no different Proportion Between varnish and no varnish )
H1 : P 1 ≠ P 2 (There are different Proportion Between varnish and no varnish )
Test and CI for Two Proportions
Sample
Varnish
X
N
Sample p
1
22
25
0.880000
2
21
25
0.840000
Sam ple
Varnish
Decision
Sam ple
Varnish
Decision
1
V
PASS
14
V
PASS
2
V
PASS
15
V
PASS
3
V
PASS
16
V
PASS
Difference = p (1) - p (2)
4
V
PASS
17
V
PASS
5
V
PASS
18
V
PASS
Estimate for difference:
6
V
FAIL
19
V
PASS
7
V
V
PASS
8
V
V
Non Vr.22 DecisionV
NV 23
PASS V
PASS
Sample
FAIL
9
V
10
V
11
V
12
V
13
V
PASS
20
No
Varnish
PASS
21
Sample
PASS
1
PASS
95% CI for difference:
(-0.152036, 0.232036)
Test for difference = 0 (vs not = 0):
Non Vr.
Decision
14
PASS
15
PASS
NV
PASS
NV
PASS
2
FAIL
3
PASS
NV 24
NV 25
PASS V
PASS
NV
PASS
16
PASS
17
NV
4
PASS
5
NV
PASS
NV
PASS
18
NV
PASS
6
NV
PASS
19
NV
PASS
7
NV
PASS
20
NV
FAIL
8
NV
PASS
21
NV
PASS
FAIL V
0.04
9
NV
PASS
22
NV
PASS
10
NV
FAIL
23
NV
FAIL
11
NV
PASS
24
NV
PASS
12
NV
PASS
25
NV
PASS
13
NV
PASS
Z = 0.41
P-Value = 0.683
P Value > 0.05, Receive H0,
Varnish
13 - 22
not significant as Vital Factor.
Selecting Vital Factor
X4 = Space
Hypothesis Testing
H0 : P 1 = P 2 (There are no different Proportion Between space and no space )
H1 : P 1 ≠ P 2 (There are different Proportion Between space and no space )
No Space
Test and CI for Two Proportions
X
N
Sample p
1
22
25
0.880000
PASS
2
24
25
0.960000
S
PASS
Difference = p (1) - p (2)
S
PASS
S
PASS
Sample
Space
Decision
Sample
Space
Decision
1
S
PASS
14
S
PASS
2
S
PASS
15
S
PASS
3
S
PASS
16
S
4
S
PASS
17
5
S
PASS
18
6
S
PASS
7
S
8
S
PASS
Sample
PASS 1
9
S
10
S
11
S
12
S
13
S
Space19
No20Space
21NS
22NS
23NS
24NS
25NS
S
Decision
S
PASS
S
PASS
S
PASS
S
PASS
S
PASS
6
NS
7
8
PASSSample
PASS 14
No Space
Decision
NS
PASS
NS
PASS
NS
PASS
PASS 17
PASS 18
NS
PASS
NS
PASS
PASS
19
NS
PASS
NS
PASS
20
NS
PASS
NS
PASS
21
NS
PASS
9
NS
PASS
22
NS
PASS
10
NS
PASS
23
NS
PASS
11
NS
PASS
24
NS
PASS
12
NS
PASS
25
NS
PASS
13
NS
PASS
PASS 4
PASS 5
No Space
Estimate for difference:
PASS 15
PASS 16
PASS 2
PASS 3
PASS
Sample
14 - 22
95% CI for difference:
Space
- 0.08
(-0.152036, 0.232036)
Test for difference = 0 (vs not = 0):
Z = 0.41
P-Value = 0.297
P Value > 0.05, Receive H0,
Space not significant as Vital Factor.
Selecting Vital Factor
X5 = Finishing Method (Winding Method)
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between Parallel and Cross)
Ho : P≠P ( There are different Proportion Between Parallel and Cross)
Analyzing Finishing Method by Comparing whether Parallel winding and Cross Winding contribute a significant
effect for wire short.
Parallel
Test and CI for Two Proportions
Sample
Cross
Decision
Sample
Cross
Decision
1
C
PASS
14
C
PASS
Sample
2
C
PASS
15
C
FAIL
3
C
FAIL
16
C
FAIL
4
C
PASS
17
C
PASS
5
C
PASS
18
C
PASS
6
C
19
C
FAIL
7
C
FAIL
Cross
8
C
9
C
10
C
11
C
12
13
FAIL
PASS Parallel
Sam ple
C
PASS
C
SampleFAIL Parallel
22PASS
23PASS
C
C
P
C
3 PASS
4 FAIL
P
24PASS
25PASS
C
5 PASS
P
6
P
7
8
Decision
14 FAIL
15 FAIL
P
PASS
P
PASS
16 PASS
17 PASS
P
PASS
P
PASS
PASS
18
P
PASS
FAIL
19
P
PASS
P
PASS
20
P
PASS
P
PASS
21
P
PASS
9
P
PASS
22
P
PASS
10
P
PASS
23
P
PASS
11
P
PASS
24
P
PASS
12
P
PASS
25
P
PASS
13
P
PASS
P
C
C
Sample p
1
14
25
0.560000
2
24
25
0.960000
Cross
Estimate for difference:
20
P
N
Parallel
Difference = p (1) - p (2)
21
Decision
1 PASS
2 FAIL
X
95% CI for difference:
0.190807)
-0.4
(-0.609193, -
Test for difference =0(vs not = 0):
Z =-3.75
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Finishing Method
has an effect or SMPS reliability.
Finishing Method is Vital Few factor
15 - 22
Selecting Vital Factor
Result hypothesis test, get vital few factor
Dielectric
Breakdown
Vital View
Hole
Vital View
wire
Material
Short
(No Power)
Core
Varnish
Press
Turn wiring
Space
Home pin
Cross
Method
Need improved
Not need improved
16 - 22
Vital View
Selecting Optimum Condition
In this Step, we try to find the optimum combination of all factors which have a
significant effect on SMPS Transformer Reliability.
• Dielectric Breakdown ( ↑ 2250 V and ↓ 2250V)
Cube Plot (data means) for PASS ( % )
• Wire Hole (↑ 5 and ↓5 / on 5 Meters)
0
• Method Finishing Home Pin ( Cross and parallel)
50
0
40
H +
Opt.
Condition
DB
DB DB DB +
DB +
DB +
DB +
DB DB -
PH
PH +
PH +
PH +
PH +
PH PH PH PH -
MTD
P
C
P
C
C
P
P
C
Pass %
0
0
50
40
70
100
0
0
H
0
100
P
0
MTD
70
H -
C
DB -
DB
DB +
Optimum condition :
Dielectric Breakdown : ↑ 2250 V
Pin Hole : ↓ 5 hole / 5 m
Method Finishing : Parallel
17 - 22
Improvement
Improvement on Finishing Method at Samwha Process
Before :
Change Wire Winding
Finishing wire of pin 12 is upper ( entered to home pin )
After :
╝
╚
4 M Change Report
8
Cross Method
Finishing wire of pin 10 is upper ( entered to home pin )
Finishing wire of pin 12 is lower then pin 10.
Parallel Method
18 - 22
Improvement
Improvement on Inspection Checking Method
IQC Samwha
Before
Before
After
After
Supplier
Supplier
Delivery
Delivery
Incoming
Sampling
Inspection
(Conventional)
OK
Storage Area
Incoming
IQC
• Appearance
• Size
NG
Sampling
Content
Content
IQC
• Appearance
Size
• 5 m each roll wire
( test hole & D
Breakdown )
• Involving new item inspection as an
new inspection check point :
• Check hole wire by Phenolphthalein
Inspection
( > 5 hole/5m, NG )
• Dielectric Breakdown test
Check hole
&D
Breakdown
NG
OK
Storage Area
19 - 22
Maintain CTQ
To keep good condition, we do :
Dielectric breakdown and hole wire control at wire vendor ( monthly )
Wire hole test before to production
Item CTQ Control
QC and Process Irregular Audit at Supplier
Total Inspection Part (100% Short Test)
CTQ
D/B
Inductance
Resistance
Control
Control
Control
I Chart of D/B
I Chart of INDUCTANCE
UCL=549.08
4
155.0
_
X=529.2
530
520
Individual Value
_
X=5.316
5
UCL=158.14
157.5
540
6
Individual Value
Individual Value
I Chart of RESISTANCE
550
UCL=7.135
7
_
X=152.4
152.5
150.0
147.5
LCL=3.497
LCL=146.66
510
3
2
4
6
8
10
12
Observation
14
16
18
20
LCL=509.32
2
4
6
8
10
12
Observation
14
20 - 22
16
18
20
145.0
2
4
6
8
10
12
Observation
14
16
18
20
Saving Cost
Part production on March ~ May ‘04 are 245915 EA, SMPS Transformer defect: 1 ea
DPU
= 1 : 184259 = 5.427E-6
▶
DPMO = 4 PPM , Sigma Level = 5.97
Condition part Oct ~ Dec’03
Condition part Mar ~ May ’04
No Power by SMPS
(70.2%
)
840 PPM
4.64
4.64
Improv
e
423
BOIL
(26.3%
)
56
BOIK
OTHE
RS
Part No
Defect PPM
Defect PPM
1128
No Power by SMPS
4 PPM
5.97
5.97
BOIL
Part No
SAVING COST
HISTORY :
October ~ December ‘03 :
Analysis cost : $ 1000
$29000
Hold, rework & sorter cost : $ 1950
Material cost : $ 4000
Loss time prod : $ 300
$0
Before
BOIK
After
Total : $ 7250 ( 3 Month )
SAVING COST : $ 29,000 / YEAR
21 - 22
OTHER
S
THANK YOU
22 - 22
TV
DENGAN METODA SIX SIGMA
DI PT LG ELECTRONICS INDONESIA
A R I P I N
2002 - 21 – 140
TEKNIK INDUSTRI
UNIVERSITAS INDONUSA ESA UNGGUL
2004
Konsep Six Sigma
Six Sigma : Suatu sistem yang flexibel untuk memperbaiki kinerja
yang mendukung sukses secara terus-menerus.
Six Sigma dibangun berdasarkan banyak ide manajemen yang penting dan
praktek terbaik dari metoda sebelumnya
Sigma adalah ukuran satuan statistik yang menggambarkan
kemampuan suatu proses dan ukuran nilai sigma dinyatakan dalam
DPU (Defect Per Unit) atau PPM
Dapat dikatakan bahwa proses dengan nilai sigma lebih tinggi
dalam proses akan mempunyai cacat yang lebih sedikit
Makin bertambah nilai sigma,
makin berkurang biaya kualitas dan waktu siklus
Konsep Six Sigma
Menggunakan Z sebagai satuan ukuran kemampuan proses
•Z =3
Sample’s average :
LSL
n
USL
=
X=
Xi
i=1
n
Sample’s standard deviation :
n
( Xi - X )
1σ
2σ
=s=
3σ
2
i =1
n-1
Nilai Z (sigma level)
Z =
Kemampuan 3s
•Z =6
x -
USL
LSL
Bila keragaman / variasi turun, maka
kemungkinan cacat juga turun,
maka kemampuan (capability)
1σ 2σ 3σ 4σ 5σ 6σ
Kemampuan 6s
meningkat
Konsep Six Sigma
DEFINE
• Standardize
solution
• Establish standard
measure to
maintain
performance
CONTROL
• Develop ideas to cut
of root causes
• Optimation solution
• Measure solution
IMPROVE
• Identify Customer’s CTQ
• Develop a team charter
• Defne and build a process
map
• Validate problem
/ process
• Measure key
problem
• Current
MEASURE condition
• Screening potential
factor
• Identify vital few root
causes
• Validate hypothesis
ANALYZE
Konsep Six Sigma
Geser
menuju
Target
Tepat namun tidak akurat
Shifting/bergeser ke Target
& Mereduksi Variasi
T
USL
USL
T
USL
μ
USL
μ
Object
dari 6
adalah
Akurat namun tidak tepat
Shift(geser)
ke target
T
Menurunkan
Variasi
USL
USL
μ
Turunkan
Variasi
Aplikasi detail Six Sigma
Perbaikan dengan Metoda Six Sigma ( DMAIC )
PT LG ELECTRONICS INDONESIA
REDUCE SMPS TRANSFORMER
SHORT ( NO POWER )
2004 06 14
Incoming Quality Control
Project Registration Sheet
Theme
Name
Project
Period
Quantity
Target
REDUCE SMPS TRANSFORMER
SHORT ( NO POWER )
Theme
Measurement February 3 ~ 6, 2004
Outline
Analysis
February 9 ~ 21, 2004
Improvement
February 24 ~ 28, 2004
Control
March ~ May, 2004
Current
Target
Improvement
Ratio (%)
840 PPM
0 PPM
100 %
6
Anticipation Short Test as check point at line
Effect
production of Samwha and IQC LGEIN.
Status
&
Problem
• Material Checking method
• Quality System set up
WIRE HOLE
WIRING CROSS
CHECKING PROSSES
Name
Member
Role
Aripin
$29000
Customer’s
Needs
Definition
Good reliability part
Telp.
QA
212
100%
Amount
Of target
(US$/Year)
$0
Before
Great Company Great People
Dept
2 - 22
After
Fun Innovation By New Culture
Problem Situation
Management policy said we have reduce field claim.
Data field claim show No Power is cause biggest
than other.
No power is TV Set have no supply voltage. It show by not light of Indicator power and no display on tube.
No power can caused by parts, like power cord, SMPS Transformer, IC, FBT, Transistor and others.
SMPS Transformer have function as divider power voltage to several sub circuit or part like FBT, IC, and
others.
If SMPS not function, it can’t supply voltage to parts or sub circuit .
Defect Analysis by No Power
Data field claim 2003
No Power
5 WORST SYMPTOM 29%
IN 2003
No Sound
Knowledge
Picture NG
CPT NG
10%
15%
Sound
SoundNo
NG
15%
A/V NG
No Power
42%
8.4%
Screen
16%
8%
27%
IC
5.2%
Crystal
3.4%
FBT
2.0%
Others
840 PPM
100%
13%
Diode
Resistor
70.2%
6170VMCBO1L
15%
Transistor
12%
Target
29%
SMPS
Fuse
14%
Defect Analysis by Model
26.3%
6170VMCBO1K
0PPM
Others
2.0%
3.6%
3.5%
Before
3 - 22
After
Theme Selection
Big Y
Little y1
Little y2
Improve
Improve
TV
TV Quality
Quality
Attribute
Attribute
30%
30% SCR
SCR
Reducing
Reducing
Knowledge
Part
TYPE
PIC
Statu
s
Crack, Ball
Opt. Skill, Ability
Wire Connection
Handling
Handling
Opt. Skill, Ability
Wrong Connection
Handling
1P1P
LQC
Finished
1P1P
LQC
On Progress
1P1P
1P1P
IQC
On Progress
Production On Progress
IIC
1P1P
Production Finished
Production On Progress
IIC
IQC/Prod
Finished
AGC Circuit
1P1P
IQC
On Progress
Opt. Skill, Ability
1P1P
Production On Progress
Antenna
How to Use
No connection
1P1P
OQC
On progress
How to Use
OQC
Product Spec
Electricity
How to used
Plug Connection
1P1P
1P1P
On Progress
On Progress
Environment
Environment
IC
Transistor
CPT
Chassis
Soldering
Process
Little y3
Adjustment
Speaker Connection
PWB Crack
Loosen Part
Bad Wiring Connection
PWB Short
Front Cabinet Dented
AGC Circuit
Bad System Connector
FBT
Tuner
Speaker
SMPS Trans
Reliability
Screen Focus, DYC,PRT
ABL Short
AGC, No Ruster
No Sound Problem
No Power
Fuse
Capacitor
Un Connect
Reliability
Reliability
Resistor
4 - 22
1P1P
OQC
OQC
1P1P
1P1P
1P1P
OQC
IQC
IQC
On Progress
On Progress
Finished
1P1P
IQC
Finished
IIC
1P1P
IQC
IQC
Finished
On Progress
1P1P
IIC
IQC
IQC
Finished
Finished
1P1P
IQC
On Progress
On Progress
Process Mapping
SAMWHA
Process
LGEIN IQC
LGEIN LQC
LGEIN OQC
•R&D
• Appearance
• Patrol
• ELT
• Purchase
• Dimension
• Performance
• LTLT
•IQC
• Structure
Final
Inspect
• Reliability
Product
•Production
• Performance
• LQC
• Assembling
• OQC
• Reliable Part
High defect
SMPS no power
At LGEIN production
Find Out what is
the Main Cause of
Problem
Selecting Best
Condition
for reliable Part
• OS&D
5 - 22
What is SMPS Transformer ?
9
7
SMPS
8
10
5
1
Output of SMPS (Voltage)
• Pin 9 - 8 : 24 ~ 25
• Pin 10 - 11 : 10 ~ 12
• Pin 14 - 13 : 110 ~ 115
• Pin 13 - 12 : 13 ~ 14
• Pin 1 – 2 : 7 ~ 10
14
13
SMPS Component :
• Core
• Bobbin
• Wire
• Insulation
• Tape
• Shield
12
2
Potential Short
Area
Short can be happen if one of wire on potential area
get direct connect.
6 - 22
Clarification Of Y
Kind of defect by SMPS
MARKING NG
APPEARANCE
Master
Signal
RUST
ISOLATION NG
SMPS
DEFECT
DIMENSION
Test
Signal
Test Signal
P/P OVER
NG SIGNAL DISPLAY
WIRE SHORT
PERFORMANCE
Master
Signal=
OK SIGNAL DISPLAY
No
Power
BROKEN
SMPS Transformer Characteristic
Display of Surge Comparison Tester ( ST-215 ) will
show two signal both master and part test.
If form signal different that’s part NG
and if same is OK.
TURNS
Demonstration of Data gathering
Display
Indicator
TV Set Power OK
TV Set No Power
7 - 22
Measurement (Gage R&R)
Instrument : Surge Comparison Tester ( ST-215 )
Agreement Analysis
21
22
23
24
25
Attribute
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Try 1
Pass
Pass
Pass
Fail
Try 2
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Within Appraisers
Fail
Pass
Pass
Fail
Fail
OPR 2
Try 1
Pass
Pass
Pass
Fail
Try 2
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Fail
Pass
Pass
Fail
Fail
Try 1
Pass
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Try 2
Pass
Pass
Pass
Fail
Within Appraisers
Fail
Pass
Pass
Fail
Fail
96
96
:
# Matched
Percent
25
25
100.00
(86.09, 100.00)
OPR-2
25
25
100.00
(86.09, 100.00)
OPR-3
25
25
100.00
(86.09, 100.00)
88
88
OPR-1
25
25
100.00
(86.09, 100.00)
OPR-2
25
25
100.00
(86.09, 100.00)
OPR-3
25
25
100.00
(86.09, 100.00)
Assessment Agreement
# Matched
Percent
25
100.00
OPR-3
Measurement System are
Acceptable
All appraisers' assessments agree with each other.
25
OPR-2
Appraiser
95 % CI
Between Appraisers :
# Inspected
OPR-1
Instrument is valid and reliable
Appraiser's assessment across trials agrees with the known standard
Percent
OPR-3
Operator is good in inspection
Each Appraiser vs Standard :
# Matched
86
OPR-2
Appraiser
Assessment Agreement : 100 %
95 % CI
# Inspected
92
90
OPR-1
# Inspected
94
90
86
OPR-1
Appraiser
98
92
95.0% C I
P ercent
100
98
94
JA NUA RY 15. 2004
A RIPIN
SHORT TEST
Appraiser vs Standard
95.0% C I
P ercent
100
Appraiser agrees with himself across trial
Appraiser
Date of study :
Reported by :
Name of product:
Misc:
Assessment Agreement
OPR 3
Percent
Sample
1
2
3
4
OPR 1
Percent
Known Population
95 % CI
(86.09, 100.00)
8 - 22
Capability Process (Current)
Based on data October ~ December ‘03, defect SMPS 124 ea from 147595 ea.
• Defect Op. = 1 ( Wire Short )
• DPU
= 124 : 147595
• DPO
= 124 : ( 147595 x 1 ) = 8.4E-04
0.04
• DPMO = DPO x 1.000.000 = 840 PPM
From Z Normal table, we get :
DPO
= 8.4E-04
Z Value
= 3.1 + 0.04
= 3.14
Our Condition now :
Sigma level
= 3.14+ 1.5
= 4.64
3.1
9 - 22
8.45E-04
Estimate Cause Analysis
By Analyzing all potential problem that could have significant effect for SMPS Performance,
thru Brainstorming and screening, we succeed to collect some root cause for the problem
Potential Factor
Dielectric Breakdown
wire
Material
Short
(No Power)
Hole
Core
Varnish
Press
Turn wiring
Space
Home pin
Finishing
Method
10 - 22
Selecting Vital Factor
X1 = Dielectric Breakdown
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between D/B below 2250 V and D/B above 2250 V)
Ho : P≠P ( There are different Proportion Between D/B below 2250 V and above 2250 V)
D/B : 2250 V
Sample
2250
Decision
Sample
2250
Decision
1
2235
FAIL
14
1850
FAIL
2
1990
FAIL
15
2230
FAIL
Sample
X
N
Sample p
3
2010
FAIL
16
1990
FAIL
4
2240
FAIL
17
2010
FAIL
1
0
25
0.000000
5
2125
FAIL
18
1890
FAIL
2
19
25
0.760000
6
2150
FAIL
19
2110
FAIL
Difference = p (1) - p (2)
7
2250
FAIL
20
2225
FAIL
8
1990
FAIL
21
2000
FAIL
Estimate for difference:
FAIL
22
2010
FAIL
95% CI for difference:
23
1995
FAIL
9
2000
10
2230
11
2000
FAIL
Sample
2250
1985
FAIL
1960 1
FAIL5430
2
4795
12
13
D/B : 2250
V
FAIL
24
Decision
25
PASS
1890
FAIL
Sample
2250
1885
FAIL
14
4300
Test and CI for Two Proportions
-0.76
(-0.927414, -0.592586)
Test for difference = 0 (vs not = 0):
Z = -8.90
Decision
PASS
PASS
15
5800
PASS
3
4500
FAIL
16
7800
FAIL
4
5350
PASS
17
7900
PASS
5
6000
FAIL
18
4600
FAIL
6
5100
PASS
19
5800
PASS
7
5700
PASS
20
5500
PASS
8
5700
PASS
21
7500
PASS
9
4500
PASS
22
7900
PASS
10
5000
PASS
23
5600
PASS
11
4100
PASS
24
4500
FAIL
12
4250
FAIL
25
6500
PASS
13
5600
PASS
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Dielectric breakdown
has an effect or SMPS reliability.
Dielectric Breakdown is Vital Few factor
11 - 22
Selecting Vital Factor
X2 = Pin Hole Wire
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between Pin Hole Below 5 and Above 5)
Ho : P≠P ( There are different Proportion Between Pin Hole Below 5 and Above 5)
P/H : 5
Test and CI for Two Proportions
Sample
5
Decision
Sample
5
Decision
1
0
PASS
14
1
PASS
Sample
2
0
PASS
15
1
PASS
3
1
PASS
16
0
PASS
4
0
PASS
17
0
PASS
5
0
PASS
18
0
PASS
Difference = p (1) - p (2)
6
2
PASS
19
1
PASS
Estimate for difference:
7
1
1
PASS
8
3
9
P/H : 205
PASS
1
PASS
Sample
PASS 1
21
5
22
7
1
Decision
0
PASS
10
1
PASS 2
11
0
PASS 3
23
7
24
8
12
3
PASS 4
13
0
PASS 5
6
X
N
Sample p
1
25
25
1.000000
2
14
25
0.560000
95% CI for difference:
PASS
Sample
PASS14
5
Decision
6
PASS
0
PASS
0
FAIL
PASS15
PASS16
7
PASS
7
FAIL
25
7
1
FAIL
PASS17
6
PASS
6
PASS
18
7
PASS
6
FAIL
19
8
FAIL
7
8
PASS
20
8
FAIL
8
8
FAIL
21
6
PASS
9
6
PASS
22
7
PASS
10
9
PASS
23
9
FAIL
11
10
FAIL
24
7
PASS
12
11
FAIL
25
8
PASS
13
10
FAIL
0.44
(0.245420, 0.634580)
Test for difference = 0 (vs not = 0):
Z = 3.76
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Pin Hole Wire
has an effect or SMPS reliability.
Pin Hole Wire is Vital Few factor
12 - 22
Selecting Vital Factor
X3 = Pressure
Hypothesis Testing
H0 : P 1 = P 2 (There are no different Proportion Between varnish and no varnish )
H1 : P 1 ≠ P 2 (There are different Proportion Between varnish and no varnish )
Test and CI for Two Proportions
Sample
Varnish
X
N
Sample p
1
22
25
0.880000
2
21
25
0.840000
Sam ple
Varnish
Decision
Sam ple
Varnish
Decision
1
V
PASS
14
V
PASS
2
V
PASS
15
V
PASS
3
V
PASS
16
V
PASS
Difference = p (1) - p (2)
4
V
PASS
17
V
PASS
5
V
PASS
18
V
PASS
Estimate for difference:
6
V
FAIL
19
V
PASS
7
V
V
PASS
8
V
V
Non Vr.22 DecisionV
NV 23
PASS V
PASS
Sample
FAIL
9
V
10
V
11
V
12
V
13
V
PASS
20
No
Varnish
PASS
21
Sample
PASS
1
PASS
95% CI for difference:
(-0.152036, 0.232036)
Test for difference = 0 (vs not = 0):
Non Vr.
Decision
14
PASS
15
PASS
NV
PASS
NV
PASS
2
FAIL
3
PASS
NV 24
NV 25
PASS V
PASS
NV
PASS
16
PASS
17
NV
4
PASS
5
NV
PASS
NV
PASS
18
NV
PASS
6
NV
PASS
19
NV
PASS
7
NV
PASS
20
NV
FAIL
8
NV
PASS
21
NV
PASS
FAIL V
0.04
9
NV
PASS
22
NV
PASS
10
NV
FAIL
23
NV
FAIL
11
NV
PASS
24
NV
PASS
12
NV
PASS
25
NV
PASS
13
NV
PASS
Z = 0.41
P-Value = 0.683
P Value > 0.05, Receive H0,
Varnish
13 - 22
not significant as Vital Factor.
Selecting Vital Factor
X4 = Space
Hypothesis Testing
H0 : P 1 = P 2 (There are no different Proportion Between space and no space )
H1 : P 1 ≠ P 2 (There are different Proportion Between space and no space )
No Space
Test and CI for Two Proportions
X
N
Sample p
1
22
25
0.880000
PASS
2
24
25
0.960000
S
PASS
Difference = p (1) - p (2)
S
PASS
S
PASS
Sample
Space
Decision
Sample
Space
Decision
1
S
PASS
14
S
PASS
2
S
PASS
15
S
PASS
3
S
PASS
16
S
4
S
PASS
17
5
S
PASS
18
6
S
PASS
7
S
8
S
PASS
Sample
PASS 1
9
S
10
S
11
S
12
S
13
S
Space19
No20Space
21NS
22NS
23NS
24NS
25NS
S
Decision
S
PASS
S
PASS
S
PASS
S
PASS
S
PASS
6
NS
7
8
PASSSample
PASS 14
No Space
Decision
NS
PASS
NS
PASS
NS
PASS
PASS 17
PASS 18
NS
PASS
NS
PASS
PASS
19
NS
PASS
NS
PASS
20
NS
PASS
NS
PASS
21
NS
PASS
9
NS
PASS
22
NS
PASS
10
NS
PASS
23
NS
PASS
11
NS
PASS
24
NS
PASS
12
NS
PASS
25
NS
PASS
13
NS
PASS
PASS 4
PASS 5
No Space
Estimate for difference:
PASS 15
PASS 16
PASS 2
PASS 3
PASS
Sample
14 - 22
95% CI for difference:
Space
- 0.08
(-0.152036, 0.232036)
Test for difference = 0 (vs not = 0):
Z = 0.41
P-Value = 0.297
P Value > 0.05, Receive H0,
Space not significant as Vital Factor.
Selecting Vital Factor
X5 = Finishing Method (Winding Method)
Hypothesis Testing
Ho : P P 2 ( There are no different Proportion Between Parallel and Cross)
Ho : P≠P ( There are different Proportion Between Parallel and Cross)
Analyzing Finishing Method by Comparing whether Parallel winding and Cross Winding contribute a significant
effect for wire short.
Parallel
Test and CI for Two Proportions
Sample
Cross
Decision
Sample
Cross
Decision
1
C
PASS
14
C
PASS
Sample
2
C
PASS
15
C
FAIL
3
C
FAIL
16
C
FAIL
4
C
PASS
17
C
PASS
5
C
PASS
18
C
PASS
6
C
19
C
FAIL
7
C
FAIL
Cross
8
C
9
C
10
C
11
C
12
13
FAIL
PASS Parallel
Sam ple
C
PASS
C
SampleFAIL Parallel
22PASS
23PASS
C
C
P
C
3 PASS
4 FAIL
P
24PASS
25PASS
C
5 PASS
P
6
P
7
8
Decision
14 FAIL
15 FAIL
P
PASS
P
PASS
16 PASS
17 PASS
P
PASS
P
PASS
PASS
18
P
PASS
FAIL
19
P
PASS
P
PASS
20
P
PASS
P
PASS
21
P
PASS
9
P
PASS
22
P
PASS
10
P
PASS
23
P
PASS
11
P
PASS
24
P
PASS
12
P
PASS
25
P
PASS
13
P
PASS
P
C
C
Sample p
1
14
25
0.560000
2
24
25
0.960000
Cross
Estimate for difference:
20
P
N
Parallel
Difference = p (1) - p (2)
21
Decision
1 PASS
2 FAIL
X
95% CI for difference:
0.190807)
-0.4
(-0.609193, -
Test for difference =0(vs not = 0):
Z =-3.75
P-Value = 0.000
- Since P-Value < 0.05, Reject Ho
- It’s mean that Finishing Method
has an effect or SMPS reliability.
Finishing Method is Vital Few factor
15 - 22
Selecting Vital Factor
Result hypothesis test, get vital few factor
Dielectric
Breakdown
Vital View
Hole
Vital View
wire
Material
Short
(No Power)
Core
Varnish
Press
Turn wiring
Space
Home pin
Cross
Method
Need improved
Not need improved
16 - 22
Vital View
Selecting Optimum Condition
In this Step, we try to find the optimum combination of all factors which have a
significant effect on SMPS Transformer Reliability.
• Dielectric Breakdown ( ↑ 2250 V and ↓ 2250V)
Cube Plot (data means) for PASS ( % )
• Wire Hole (↑ 5 and ↓5 / on 5 Meters)
0
• Method Finishing Home Pin ( Cross and parallel)
50
0
40
H +
Opt.
Condition
DB
DB DB DB +
DB +
DB +
DB +
DB DB -
PH
PH +
PH +
PH +
PH +
PH PH PH PH -
MTD
P
C
P
C
C
P
P
C
Pass %
0
0
50
40
70
100
0
0
H
0
100
P
0
MTD
70
H -
C
DB -
DB
DB +
Optimum condition :
Dielectric Breakdown : ↑ 2250 V
Pin Hole : ↓ 5 hole / 5 m
Method Finishing : Parallel
17 - 22
Improvement
Improvement on Finishing Method at Samwha Process
Before :
Change Wire Winding
Finishing wire of pin 12 is upper ( entered to home pin )
After :
╝
╚
4 M Change Report
8
Cross Method
Finishing wire of pin 10 is upper ( entered to home pin )
Finishing wire of pin 12 is lower then pin 10.
Parallel Method
18 - 22
Improvement
Improvement on Inspection Checking Method
IQC Samwha
Before
Before
After
After
Supplier
Supplier
Delivery
Delivery
Incoming
Sampling
Inspection
(Conventional)
OK
Storage Area
Incoming
IQC
• Appearance
• Size
NG
Sampling
Content
Content
IQC
• Appearance
Size
• 5 m each roll wire
( test hole & D
Breakdown )
• Involving new item inspection as an
new inspection check point :
• Check hole wire by Phenolphthalein
Inspection
( > 5 hole/5m, NG )
• Dielectric Breakdown test
Check hole
&D
Breakdown
NG
OK
Storage Area
19 - 22
Maintain CTQ
To keep good condition, we do :
Dielectric breakdown and hole wire control at wire vendor ( monthly )
Wire hole test before to production
Item CTQ Control
QC and Process Irregular Audit at Supplier
Total Inspection Part (100% Short Test)
CTQ
D/B
Inductance
Resistance
Control
Control
Control
I Chart of D/B
I Chart of INDUCTANCE
UCL=549.08
4
155.0
_
X=529.2
530
520
Individual Value
_
X=5.316
5
UCL=158.14
157.5
540
6
Individual Value
Individual Value
I Chart of RESISTANCE
550
UCL=7.135
7
_
X=152.4
152.5
150.0
147.5
LCL=3.497
LCL=146.66
510
3
2
4
6
8
10
12
Observation
14
16
18
20
LCL=509.32
2
4
6
8
10
12
Observation
14
20 - 22
16
18
20
145.0
2
4
6
8
10
12
Observation
14
16
18
20
Saving Cost
Part production on March ~ May ‘04 are 245915 EA, SMPS Transformer defect: 1 ea
DPU
= 1 : 184259 = 5.427E-6
▶
DPMO = 4 PPM , Sigma Level = 5.97
Condition part Oct ~ Dec’03
Condition part Mar ~ May ’04
No Power by SMPS
(70.2%
)
840 PPM
4.64
4.64
Improv
e
423
BOIL
(26.3%
)
56
BOIK
OTHE
RS
Part No
Defect PPM
Defect PPM
1128
No Power by SMPS
4 PPM
5.97
5.97
BOIL
Part No
SAVING COST
HISTORY :
October ~ December ‘03 :
Analysis cost : $ 1000
$29000
Hold, rework & sorter cost : $ 1950
Material cost : $ 4000
Loss time prod : $ 300
$0
Before
BOIK
After
Total : $ 7250 ( 3 Month )
SAVING COST : $ 29,000 / YEAR
21 - 22
OTHER
S
THANK YOU
22 - 22