Lucas-TVS Ltd

Process Evolution
through

Integration of Shainin and Taguchi

- A Case Study in

Alternator Manufacturing
by

Dr N Ravichandran
Chief Executive Officer
Lucas-TVS
Lucas TVS Ltd
Padi, Chennai 600 050, India

Dr. N. Ravichandran

Lucas-TVS Ltd

Contents
1.0 Overview of Lucas-TVS
2.0 Introduction about Alternator
3.0 Problem definition
4.0 Diagnostic approach – Shainin Method
50
5.0 Root Cause - Validation
6.0 Improvement approach – Taguchi Method
7.0 Result Validation
80
8.0 Standardisation
9.0 Conclusion
10.0 Overview of Shainin Tools
Dr. N. Ravichandran
2
Lucas-TVS Ltd

1 0 Overview
1.0 O e ie of Lucas
L cas - TVS

Dr. N. Ravichandran
3

Lucas-TVS Ltd

1.1 Lucas–TVS Auto Electrical plants

9 Established 1961, originally a joint venture

between Lucas Plc UK and TVS, wholly
owned since 2001
9 Four decades of leadership on Indian Market
9 7 plants in India, main plant in Chennai with
2600 employees
9 Product development capability: 80% of
revenue from In-house developed products
9 Technical Collaboration
• Mitsubishi Electric: Geared Starters /
Internal Fan Alternators
• Denso: Ignition Systems, Two Wheeler
Starters
• YDK Japan: Blower Motors

Dr. N. Ravichandran
4
Lucas-TVS Ltd

1.1 Other Plants

Dr. N. Ravichandran
5

Lucas-TVS Ltd

1.2 Product Range

Annual Units
Products
((In Lacs))
Alternators 25.30

Starters 33.90

Two Wheeler Starters 38.00

Wipers 22.10

Compressor Motors 4.70

Ignition Coils 13.50

Dr. N. Ravichandran
6
Lucas-TVS Ltd

1.3 Blue Chip Customer Base

PASSENGER CARS COMMERCIAL VEHICLES

Dr. N. Ravichandran
7

Lucas-TVS Ltd

1.3 Blue Chip Customer Base

PICK UPS & SUVs BUSES & COACHES TWO-WHEELERS

TRACTORS ENGINES CONSTRUCTION

Dr. N. Ravichandran
8
Lucas-TVS Ltd

1.4 Recognitions
g & Awards
TS 16949 OHSAS 18001 ISO 14001 Deming Award

JIT Innovation Award from JIT Management Lab, Tokyo

JIT Grand Prix Award from JIT Management Lab, Tokyo (Thrice)
Frost & Sullivan – Platinum Award for Manufacturing Excellence
BIS – Rajiv Gandhi National Award
Energy Conservation Award

Dr. N. Ravichandran
9

Lucas-TVS Ltd

1.4 Recognitions & Customers Awards

• Maruti Suzuki Best Warranty Improvements 2009
• Maruti Suzuki - VA/VE Award, 2006
• Maruti Suzuki - Vendor Performance Award for Quality, 2004-05
Q lit
Quality • Ford - Q1 Award, August 2006
• Hyundai Motor India - Best Quality Performance Award , 2008-09
• Hyundai Motor India - 100 PPM Award, 2003
• Mahindra - Best Quality Performance Award (FES ),
) 2006-07
2006 07

• Maruti Suzuki Vendor Upgradation Award, 2011

• Maruti Suzuki Best Vendor Upgradation Award, 2010
• Maruti Suzuki Best Vendor Upgradation Award, 2009
• Maruti Suzuki Best Supplier Support Award, 2008
• Maruti Suzuki - Best Supplier Support Award, 2007-08
• Maruti Suzuki - Best Vendor Award,
Award 2006
• Maruti Suzuki - Superior Kaizen Performance Award, 2004
Performance
• Mahindra & Mahindra – Annual Commodity Award, 2011
• Ashok Leyland
y - Outstanding
g Performance in Management,
g 2007-08
• Cummins India - Excellent Performance Award, 2007+2008
• Cummins India - Best Performer Award, 2003
• Honda Motorcycles and Scooter India - Achievement Award 2008-09
• Honda Motorcycles and Scooter India - Best Supplier Award, 2005
• Hyundai Motor India - Overall Best Performance, 2004
• Tata Motors - Enduring Relations Excelling Together Award, 2008
Dr. N. Ravichandran
10
Lucas-TVS Ltd

Case Study -

2.0 Introduction about Alternator

Dr. N. Ravichandran
11

Lucas-TVS Ltd

2.0 Alternator - Introduction

Dr. N. Ravichandran
 Lucas-TVS Ltd

30 P
3.0 Problem
bl Definition
D fi iti

Dr. N. Ravichandran
13

Lucas-TVS Ltd

3.1 Case Study – Problem Definition

Problem Statement

¾ Poor First Pass Yield during Pilot Production Trial Run of a

New Product Introduced

Expected Yield Actual Yield

> 95 % < 85 %

Pilot run rejection data

Pilot run Pareto of Rejection
100
Qty Produced = 200 30
87.5 90.625
93.75 96.875
100
25 80
Qty Rejected = 32
mm%

20
Rej no

60
15
Cum

40
10
88% of rejection 5 28 1 1 1 1 20
due to 0 0
Initial Cut
C t in Failure
Fail e Initial cut in WL Not on Mild glow Bind Pulley
Damage

Dr. N. Ravichandran
Lucas-TVS Ltd

3.2 Case Study – Problem Definition

Wh is
What i Initial
I i i l Cut
C InI F
Failure
il ?
¾ When alternator is generating sufficient Threshold Current – the warning lamp on the Dash Board is OFF –
indicating to the driver proper functioning of alternator. The failure to produce this current is called Initial cut in
failure – Warning Lamp Not OFF

Dr. N. Ravichandran

Lucas-TVS Ltd

4 0 Diagnostic Approach – Shainin Method

4.0

Dr. N. Ravichandran
16
Lucas-TVS Ltd

4.1 Case Study – Cause Analysis

Cause & Effect Diagram:

Stator Assembly Process

Regulator screw not Ineffective soldering of stator leads

Id oversize
ti ht
tighten
Stator phase nut not tighten
Resistance imbalance Excess basket OD

Rpm Meter not working

Improper soldering
on stator leads
Warning lamp
NOT OFF
Rotor rubbing
Slip ring tag broken Insul bush damaged

R
Rotor wire
i cut Positive diode shorting
Carbon brush broken
Slip ring burr
Regulator A terminal damaged

ROTOR Rectifier and regulator

Dr. N. Ravichandran

Lucas-TVS Ltd

4.2 Case Study – Cause Analysis

Di
Diagnostic
i AApproach:
h
Conventional Approach Shainin Clue Generating Approach

¾ Check Conformance to Standards

& Specifications ¾ Identify the Cause from the existing
Good & Bad
¾ Identify the cause by conducting
Fresh experiments ¾ Select BOB – Best of Best &
WOW - Worst of Worst from the
¾ Fresh
F hE Experiments
i t conducted
d t d withith
existing lot
pre determined levels for
Specified factors - to avoid failure
phenomenon
¾ Conduct Designed Experiments
using these BOB & WOW to
¾ The factors and their levels are Identifyy the Culprit
p cause.
chosen based on Experience &
Knowledge

Si
Since Pil
Pilott R
Run - all
ll the
th parts
t have
h been
b checked
h k d for
f conformance
f prior
i tot assembly.
bl
Hence, to cut short time & effort Shainin Approach is Preferred
Dr. N. Ravichandran
Lucas-TVS Ltd

4.3 Case Study – Shainin Approach

Shainin Component Search : This is a diagnostic study by Elimination

¾ A pair of BOB & WOW Selected

¾ First phase of Elimination – Identify Assembly Process If D/d Ratio < 1.25
which is contributing OR
Using D/d Ratio by disassembling &
reassembling Twice - BOB / WOW
Constituent Parts If D/d Ratio > 1.25

¾ Second phase of Elimination – Identify

which part is contributing by Swapping Good becomes BAD
When the
parts by pre determined Priority between AND Culprit part is
BOB & WOW. Bad becomes Good interchanged

The Decision of transition from Good to Bad & vice versa is

not based on Design Specification limit but by new control
limits determined for each ppair of BOB & WOW

¾ Capping Run – confirming the Finding

Dr. N. Ravichandran

Lucas-TVS Ltd

4.4 Case Study – Shainin Approach

Shainin Component Search: Elimination Phase 1

BOB WOW
WL Current O/P WL Current O/P

Initial assembly 0 31
0.31 0 08
0.08
After 1st re-assembly 0.39 0.06
Specification of
After 2nd re-assembly 0.28 0.05 WL current O/P
Median 0.33 0.065 > 0.12 amps
Range 0.11 0.03

D / d Test:
Difference between the medians (D): 0.33 – 0.065 = 0.265
Average range (d) : (0.11+0.03) / 2 = 0.07

D/d = 0.265 / 0.07 = 3.785 > 1.25

Inference: Assembly Process is not Culprit

Dr. N. Ravichandran
Lucas-TVS Ltd

4.4 Case Study

y – Shainin Approach
pp

Shainin Component Search: Elimination Phase 2

Control limits for Swapping : This is not same as Specification

Control limit BOB = Median of BOB +/- (2.776/1.81)d

BOB :- Min = 0.2227, Max = 0.4375;
C
Control
l limit
li i WOW = Median
M di off WOW +/-
/ (2.776/1.81)d
(2 776/1 81)d
WOW- Min = -0.0423, Max = 0.1723

Priority for Swapping : This is based on Knowledge & Experience

1. Rotor assemblyy
2. Regulator Rectifier Assembly
3. Stator Assembly

Dr. N. Ravichandran

Lucas-TVS Ltd

4.4 Case Study – Shainin Approach

Shainin Component Search: Out put measured with parts interchanged

First
1 0.32 0.06
Reaaseembly
Second
2 0 28
0.28 0 05
0.05 Hence Stator
reassembly
ROTOR is the Culprit
3 0.23 0.06 RED X
INTERCHANGE
Bring back to
4 0.25 0.05
original
RECTIFIER &
5 0.29 0.08
REGULATOR
Bring back to Inference:
6 0.32 0.06
original
Total Reversal
7 STATOR ASSY 0.06 0.32
when Stator is
Bring back to Interchanged
8 0.3 0.07
original
9 Capping Run 0.07 0.31

Dr. N. Ravichandran
Lucas-TVS Ltd

4.5 Case Study – Shainin Approach

Shainin Paired Comparison : To Identify the Product Feature Contributing to the Defect

¾ Six pairs of BOB & WOW Selected

¾ These pairs are compared in all aspects – whether specified in the design or not
¾ Significance of each feature compared is decided using “ Tukey End Count Test”

The response of each feature compared is arranged in either ascending or

descending order. The no. of continuous good or bad at either end is called the
top & bottom end counts.
The Significance level of each feature is based on the total end count which is
sum of top & bottom count :

If the Total End count is

6 then Confidence level = 90%
7 then Confidence level = 95%
10 then Confidence level = 99%
12 then Confidence level = 99.7%
Dr. N. Ravichandran

Lucas-TVS Ltd

4.5 Case Study

y – Shainin Approach
pp
Shainin Paired Comparison :

Stator
Stator Stator Stator
S.No. Basket ID ovality
Basket ID Overhang ID
OD Features Compared
G1 119.10 99.2 15.24 95.39 0.07
G2 119.20 99.16 15.35 95.37 0.06
G3 118.95 99.17 15.39 95.38 0.065
G4 119.18 99.21 15.49 95.37 0.055
G5 119.14 99.36 15.65 95.39 0.06
G6 118.98 99.24 15.47 95.38 0.065
B1 118.93 99.14 15.45 95.44 0.05
B2 118.99 99.2 15.33 95.43 0.065
B3 119.1 99.18 15.21 95.44 0.06
B4 119.13 99.32 15.31 95.43 0.07
B5 119.17 99.22 15.27 95.425 0.065
B6 119 08
119.08 99 29
99.29 15 55
15.55 95 44
95.44 0 065
0.065

Dr. N. Ravichandran
 Lucas-TVS Ltd

4.5 Case Study – Shainin Approach

Shainin Paired Comparison : Response Arranged in Descending order

Stator Stator Stator

S No
S.No. Stator ID ID ovality
Basket OD Basket ID Overhang
G2 119.2 G5 99.36 G5 15.65 B1 95.44 G1 0.07
G4 119.18 B4 99.32 B6 15.55 B3 95.44 B4 0.07
B5 119.17 B6 99.29 G4 15.49 B6 95.44 G3 0.065
G5 119 14
119.14 G6 99 24
99.24 G6 15 47
15.47 B2 95 43
95.43 G6 0 065
0.065
B4 119.13 B5 99.22 B1 15.45 B4 95.43 B2 0.065
G1 119.1 G4 99.21 G3 15.39 B5 95.425 B5 0.065
B3 119.1 G1 99.2 G2 15.35 G1 95.39 B6 0.065
B6 119 08
119.08 B2 99 2
99.2 B2 15 33
15.33 G5 95 39
95.39 G2 0 06
0.06
B2 118.99 B3 99.18 B4 15.31 G3 95.38 G5 0.06
G6 118.98 G3 99.17 B5 15.27 G6 95.38 B3 0.06
G3 118.95 G2 99.16 G1 15.24 G2 95.37 G4 0.055
B1 118.95 B1 99.14 B3 15.21 G4 95.37 B1 0.05
Top end count 2 1 1 6 0.5
Bottom end count 0.5 1 1 6 1
Total end count 2.5 2 2 12 1.5

Inference:
¾ Stator ID is the only feature contributing significant @ 99.7 % confidence
¾ If the ID of the stator core is near to the top limit of the speciation – the defect occurs
ID Spec: 95.33 to 95.44 mm dia
Dr. N. Ravichandran

Lucas-TVS Ltd

5.1 Case Study – Shainin Validation

Shainin B vs C Test : To validate the finding

¾ Six pairs of units build fresh to confirm the finding

Six presumed Assembled the significant part conforming to (WOW) Bad unit
Bad units With values
l & other
th partst random
d

Six p
presumed Assembled the significant part conforming to (BOB) Good unit
Good units With values & other parts random

¾ Tukey test applied to the response of these pairs:

Six presumed If Results

i t
into Actual Bad units
Bad units
Validates
the earlier
Six presumed If Results Finding
d g
into Actual Good units
Good units
Dr. N. Ravichandran
Lucas-TVS Ltd

5 0 Root Cause - Validation

5.0

Dr. N. Ravichandran
27

Lucas-TVS Ltd

5.1 Case Study – Shainin Validation

Shainin B vs C Test :
Response from Suspected “C” and
Tukey Test on Response
Better “B” Process

C1 95.44 NOT OK B1 95.38 OK

Process ( B )
nos

er
with Current ( C ) and 6 n

are tightened in random orde

C5 95 44
95.44 NOT OK B2 95.39
95 39 OK
6 / 6 OK

C3 95.435 NOT OK B3 95.375 OK

C6 95.435 NOT OK B4 95.385 OK
C2 95.43 NOT OK B5 95.4 OK
ed Better P

C4 95.43 NOT OK B6 95.37 OK

B5 95.4 OK C1 95.44 NOT OK
T OK

B2 95.39 OK C2 95.43 NOT OK

with claime

6 / 6 NOT

B4 95.385 OK C3 95.435 NOT OK

B1 95.38 OK C4 95.43 NOT OK
6nos

B3 95.375 OK C5 95.44 NOT OK

w
6

B6 95 37
95.37 OK C6 95 435
95.435 NOT OK

Inference: If the ID of the stator core is near to the top limit of the speciation – the defect
occurs – this is validated.
validated

Thus the Root cause of the defect is Validated

Dr. N. Ravichandran
 Lucas-TVS Ltd

5.2 Case Study – Root cause

Stator ID Variation – Root cause :
Process Capability of Stator ID:

Cp
p = 1.44 Cpk
p = 1.25

Root Cause:
• Stators having ID near to the design target are Good.
• Whereas ID near to the top limit but still within specification are leading to defect
• And Stator ID p
population
p spreads
p up
p to the top
p limit of the specification
p

Dr. N. Ravichandran

Lucas-TVS Ltd

6 0 Improvement approach – Taguchi Method

6.0

Dr. N. Ravichandran
30
 Lucas-TVS Ltd

6.1 Case Study – Improvement

Stator ID Manufacturing Process : It is a coining process done in 250 ton
hydraulic press
Top die
Ram pressure
ID bunk Bottom Ejector

Die Cushion
pressure
Component loading Ram down During Dwell time Ram up

Factors affecting variation / Consistency in Stator ID :

1. Ram Pressure
It is decided to conduct
2. Dwell time
experiment at different
3. Diecushion pressure levels of these known
4. Initial pack thickness of stator before coining factors to identify
significance
5. Coil hardness of stator material
Dr. N. Ravichandran

Lucas-TVS Ltd

6.2 Case Study – Improvement

Improving Stator ID :
Selecting Levels of Factors & Response :

Factor Level 1 Level 2

A R Pressure
Ram P 100 bar
b 160 bar
b
B Dwell Time 3 Sec 5 Sec
C Hardness 112 Bhn 119 Bhn
D Die Cushion ejection pressure 5 Bar 20 bar
0.78 mm * 30 0.81 mm * 31
E Pack Thickness layers
y layers
y
Interactions considered
AB Ram pressure & Dwell time
AC R pressure & Hardness
Ram H d

¾ Levels of factors and their interactions are selected based on domain

knowledge and experience
¾ It is decided to measure Stator ID and ID ovality as Response
Dr. N. Ravichandran
 Lucas-TVS Ltd

6.2 Case Study – Improvement

Improving Stator ID :
Taguchi L8 OA design is chosen – based on the no. of factors, their
Experiment Design : levels / interactions and Deg. Of freedom
2
B
2 B
3 3
1 5
1 5 4 4
A C A C

6 7 D D E
6 7
STANDARD LINEAR GRAPH MODIFIED LINEAR GRAPH
A B C D E

Factors Ram Pr Dwell time 1*2 Hardness 1*4 Die cushion Initial p
pack
Column No. 1 2 3 4 5 6 7
Exp 1 1 1 1 1 1 1 1
Exp 2 1 1 1 2 2 2 2
E
Exp 3 1 2 2 1 1 2 2
Exp 4 1 2 2 2 2 1 1
Exp 5 2 1 2 1 2 1 2
Exp 6 2 1 2 2 1 2 1
Exp 7 2 2 1 1 2 2 1
Exp 8 2 2 1 2 1 1 2
Dr. N. Ravichandran

Lucas-TVS Ltd

6.2 Case Study – Improvement

Improving Stator ID :
Experiment Response Table :
Factors Response
Initiall pack

Stator ID ID Ovality
cusion
Dwell time

dness
m Pr

(mm) (mm)
1**2

1**4
Ram

Hard

Die c

95 33 - 95
95.33 95.44
44 01M
0.1 Max

Replication Replication
Col mn
Column 1 2 3 4 5 6 7 A B A B

Exp 1 1 1 1 1 1 1 1 95.30 95.2 0.075 0.04

Exp 2 1 1 1 2 2 2 2 95.29 95.38 0.07 0.025 Burr noticed

Exp 3 1 2 2 1 1 2 2 95.38 95.325 0.035 0.05 Burr noticed

Exp 4 1 2 2 2 2 1 1 95.40 95.39 0.055 0.045

Exp 5 2 1 2 1 2 1 2 95.33 95.29 0.065 0.065
Exp 6 2 1 2 2 1 2 1 95.00 95.315 0.045 0.05 Burr noticed

Exp 7 2 2 1 1 2 2 1 95.39
95 39 95.38
95 38 0.04
0 04 0.045
0 045 Burr noticed
Exp 8 2 2 1 2 1 1 2 95.33 95.34 0.04 0.04
Dr. N. Ravichandran
 Lucas-TVS Ltd

6.3 Case Study – Improvement

Experiment
i ANOVA
O Table
bl :
ANOVA TABLE- Stator ID
Mean
So rce
Source SS DOF Square F0 Ftabld
Ram pressure 0.0004 1.0000 0.0004 0.6677
Dwell time 0.0062 1.0000 0.0062 11.7774 Only Dwell time is
1&2 0.0000 1.0000 0.0000 0.0267 Significant for Stator ID
Hardness 0 0000
0.0000 1 0000
1.0000 0 0000
0.0000 0 0030
0.0030
5.3200
1&4 0.0015 1.0000 0.0015 2.8516
Die cushion 0.0000 1.0000 0.0000 0.0742
Initial Thickness 0.0003 1.0000 0.0003 0.5015
Error 0 0042
0.0042 8 0000
8.0000 0 0005
0.0005 1 0000
1.0000
Total 0.0126 15.0000

ANOVA TABLE - ID Ovality

Mean
Source SS DOF Square F0 Ftabld
Ram pressure 0.0000 1.0000 0.0000 0.0069
Dwell time 0.0005 1.0000 0.0005 1.9931
1&2 0 0001
0.0001 1 0000
1.0000 0 0001
0.0001 0 3379
0.3379
Hardness 0.0001 1.0000 0.0001 0.5586
5.3200
1&4 0.0001 1.0000 0.0001 0.3379
Die cushion 0.0003 1.0000 0.0003 1.1655 No factor is Significant for
Initial
t a Thickness
c ess 0.0000
0 0000 1.0000
0000 0.0000
0 0000 0.0069
0 0069 ID Ovalityy
Error 0.0018 8.0000 0.0002 1.0000
Total 0.0028 15.0000
Dr. N. Ravichandran

Lucas-TVS Ltd

6.4 Case Study – Improvement

Selecting Optimum Levels :

95 36
95.36 ¾ From the Response
p graph
g p the optimum
p level of the significant
g factor is
95 26
95.26
chosen
Dwell time

¾ Dwell time of 5 Sec yields response closer to the design target

¾ The
Th levels
l l off other
h factors
f are chosen
h b studying
by t d i th the response
table and based on Technical and economic feasibility etc..
T1 T2
3 sec 5 Sec ¾ It is observed at Die cushion pressure of 20 bar – burr is noticed hence
not chosen
¾ Further the hardness & initial pack are noise factors within the specified
tolerance band and are not significant – hence the allowable tolerance
band is chosen as optimum.

Recommended levels of factors :

Ram pressure = 160 bar

Dwell time = 5 sec
Die cushion pressure = 5 bar
Dr. N. Ravichandran
Lucas-TVS Ltd

7 0 Result Validation
7.0

Dr. N. Ravichandran
37

Lucas-TVS Ltd

7.1 Case Study

y – Result Validation
Confirmatory experiment at Optimum Levels (Stator manufacturing) :

¾ An experiment run at optimum level & process capability observed

Cp Cpk
Before 1.44 1.25
After 2.02 1.75

Inference:
The selected Optimum levels have yielded the desired reduction in variation

Dr. N. Ravichandran
 Lucas-TVS Ltd

7.2 Case Study – Result Validation

Confirmatory Second Production Trial Run (Alternator Assembly):
¾ Alternators assembled with stators manufactured from new process –
with
ith some stators
stato s selected nea
near to the top limit of the imp
improved
o ed pop
population
lation
Before After
Improvement Improvement
First pass Yield 84 % 98 % Improving Stator ID
Qty Produced (Nos) 200 200 Process Capability
Qty Rejected (Nos) 32 4 has Eliminated top ranked defect
and has improved
Failure Modes
First Pass Yield
Initial Cut in Failure - W lamp 28 Zero of the Assembly
Not OFF (Nos)
W lamp Not ON( Nos) 1 1
Mild Glow (Nos) 1 1
Bind ( Nos) 1
Pulley Damage ( Nos) 1
Through bolt damage 1
Brush Broken 1
Dr. N. Ravichandran

Lucas-TVS Ltd

8 0 Standardisation
8.0

Dr. N. Ravichandran
40
Lucas-TVS Ltd

8.1 Case Study

y – Standardizing
g Improvement
p
Process Standards Updated:

Dr. N. Ravichandran

Lucas-TVS Ltd

9 0 Conclusion
9.0

Dr. N. Ravichandran
42
 Lucas-TVS Ltd

9.1 Case Study

y – Conclusion
Summary:

¾ Initial Production trial run of a new product – yielded low first pass yield.

¾ Major contribution of defect being from Initial Cut in Failure – Warning Lamp Not Off

¾ Cause & effect diagram indicated contribution from both process as well as parts
parts.

¾ Shainin Clue Generating Experiments conducted - to quickly funnel down to the culprit

¾ Variation from the design target of Stator ID is identified as the root cause

¾ Taguchi L8 OA conducted to optimize the factors affecting the Stator ID

¾ The optimum levels improved the process capability

¾ The improved process capability Eliminated the Defect

¾ Second production trial run confirmed the findings

– Paved way for PPAP and Production Ramp up without delay

Dr. N. Ravichandran

Lucas-TVS Ltd
EOQ 2011

9.2 Case Study

y – Conclusion
Inference :

¾ A new product with new performance standards – demands new process standards.

¾ A process standard acceptable to previous products – not acceptable to new product

¾ Shainin clue Generating tools – enhances the process of identifying the unknown cause

¾ Shainin Clue generating tools – questions the design specification also

¾ Shainin Tools generates clue from the available product & process

¾ Taguchi OA enhances the optimization process once the unknown cause is identified

Future Study :

¾ Taguchi Parameter designed experiment is to be conducted – to make the process more robust
– irrespective of variation in noise factors.

¾ A study
t d for
f iintegrating
t ti various
i problem
bl solving
l i tools
t l are required.
i d

Dr. N. Ravichandran
Lucas-TVS Ltd

10 0 Overview of Shainin Tools

10.0

Dr. N. Ravichandran
45

Lucas-TVS Ltd

10.1 Overview of Shainin Tools

DETECTIVE APPROACH TO SOLUTION

FACTS FROM
ENGINEERS’
PRODUCT / C bi d with
Combined ith D t t
Detect SOLUTION
PROCESS KNOWLEDGE

Dr. N. Ravichandran
Lucas-TVS Ltd

10.2 Overview of Shainin Tools

COMPONENT
MULTI - VARI SEARCH
TO FILTER CYCLICAL ; TO FILTER PRODUCT PARAMETERS WHEN
POSITIONAL & TEMPORAL COMPONENTS ARE INTERCHANGEABLE
VARIATION
PAIRED
COMPARISON
TO FILTER PRODUCT
PARAMETERS
CONCENTRATION PRODUCT
CHART PROCESS SEARCH
TO FILTER VARIATION TO FILTER PROCESS
WITHIN UNIT PARAMETERS

Dr. N. Ravichandran
WHEN VARIABLES ARE UNKNOWN & MANY

Lucas-TVS Ltd

10.3 Overview of Shainin Tools

VARIABLES
SEARCH FULL FACTORIAL

5 TO 20 VARIABLES 4 OR FEWER VARIABLES

B vs C
TEST
ONE VARIABLE

WHEN VARIABLES ARE KNOWN

Dr. N. Ravichandran
Lucas-TVS Ltd

10.4 Overview of Shainin Tools

B vs C
TEST

TURNING THE PROBLEM ON & OFF

WHEN ROOT CAUSE IS KNOWN

Dr. N. Ravichandran

Lucas-TVS Ltd

10.5 Overview of Shainin Tools

RSM
SCATTER PLOTS RESPONSE SURFACE
METHODOLOGY
WITH NO INTERACTION OF WITH INTERACTION OF
FACTORS FACTORS

TIGHTEN THE TOLERANCES OF IMPORTANT VARIABLES

OPEN UP TOLERANCES OF UNIMPORTANT VARIABLES
Dr. N. Ravichandran
Lucas-TVS Ltd

10 6 Overview of Shainin Tools

10.6
Salient Learning Points

¾ MOST UNSUSPECTED CAUSES ARE REVEALED

USING CLUE GENERATING TOOLS.

¾ CONFORMANCE TO SPECIFICATION DOES NOT

MEAN FREE OF DEFECT.

Dr. N. Ravichandran

Lucas-TVS Ltd

Dr. N. Ravichandran