6σ

6σ
SIX SIGMA – BLACK BELT
LEVEL -1 TRAINING MODULE - 5

World Class Manufacturing Solutions (WCMS)

6σ

Contact details for continuous support

2

 Name of the mentor: Ram narayan

 Email : ram11270@[Link]
 Mobile: +919822401826

World Class Manufacturing Solutions

6σ

Training dates
3

Program Faculty Sep Oct Nov Dec Jan Feb March

2019  2019 2019 2019 2020 2020 2020
Six sigma Ram 12,13,14   26,27,28 18,19 22,23,24 20,21 18
Black Belt (M1) (M2) (M3) (M4) (M5) (M6)
training
Project Ram   3,4,5   16,17 20,21 6,17
facilitation Narayan
and Review in
shop floor

Version : [Link] 2018

World Class Manufacturing Solutions
6σ

Table of Contents
4

No. Topic Slide

numbers
1 Tool # 13 – Variation analysis 6-70

2 Tool # 14 – Simplex method 71-74

World Class Manufacturing Solutions

6σ

Six sigma phases

5
 Process followed is
-Identify and State the Problem
-Identify the Impact Rating of the Problem
1 Define (D) -Identify Team Leader and Members
-Plan for the remaining phases
-Identify Baseline and Target

-Understand the Problem

15 Measure (M) -Observe the process, collect and analyze data
-Identify Suspected Sources of Variation (SSV’s) for the problem

45 Analyze (A) -Pin-point the real reasons creating the problem

-Validate the pin-pointed causes

45 Improve (I) -Identify solutions, evaluate and Implement

-Monitor and control the pinpointed causes

30 Control (C ) -Monitor and control the Problem

World Class Manufacturing Solutions

6σ

Chapter - 1
6

 Variation analysis

1D

World Class Manufacturing Solutions

 6σ

Steps in Variation analysis

Start C
Slide: 7

Identify the parameter

S1 S4 Collect Data
that requires variation analysis

Identify are there multiple Do Multi-Vari analysis if there

S2 S5
Product/process streams Are multiple streams

S3 Decide Sampling Is the Part to Part

Variation
S6
highest
No

Yes B
A
World Class Manufacturing Solutions
 6σ

Steps in Variation analysis

Slide: 8

A B

Check for consistency of Implement solutions to correct

S7
Part to Part variation S8 The process

C
Estimate the Part to Part
S9
variation

Check the shape of the

S10
Distribution for normality

WCMS
World Class Manufacturing Solutions
 6σ

Steps in Variation analysis

Slide: 9

S14 Monitor the process using

Estimate the rejection
X-bar/R chart or
S11 Percentage due to
Pre-control chart (as decided
Part to Part variation
In Step 12)

Do Six Sigma analysis and

S12 End
Decide the monitoring method

Calculate control limits for

S13 Average if the decision is
to use X-bar and R chart

WCMS
World Class Manufacturing Solutions
 6σ

S3
Decide Sampling
Slide: 10

 We have to decide the sampling for two things

 How much should be the number of parts within the time block
 How many such time blocks we need to proceed with the analysis
 Sampling rules (With time block)
 There should be either 3 or 5 parts collected
“Continuously” from the process within the time block.
Collect 3 parts continuously from the process, if the
inspection time is more (eg: checking with height gage,
CMM, perthometer). Collect 5 parts continuously from
the process, if the inspection time is less (eg: checking
with vernier caliper, micrometer, dial snap gage)

WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling
Slide: 11

 Sampling rules
 There should be minimum 25 time blocks which should cover a time period of
at-least 4 hours of production
 If there are any unplanned events (eg: breakdown, power failure), it should be
captured between time blocks. There should be one time block immediately
after the unplanned event
 If there are planned events (eg: wheel dressing, offset), it should be captured
between time blocks. In case of planned events, there should be one time
block before the event and one time block after the event
 There should not be any events captured in the data within the time block
 If any unplanned event happens within the time block, then the grouping
should be modified as we discussed during the Multi-vari analysis
 Do not delete any data for the purpose of analysis.
 Note: Even though we are planning to collect 3 or 5 parts within time block,
due to the occurrence of unplanned events, the number of data within the time
block may vary. This is done to ensure that events are captured between time
blocks

WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling
Slide: 12

 Sampling rules
 When we are have to collect data for 25 time blocks, the question is should we
leave any time interval between two time blocks or not
 This decision depends on the Cycle time of the process
 All events occurring between the time blocks should be recorded

T1 T2 T3 T4 T5 T6

WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling
Slide: 13

 Sampling rules
 If the Cycle time per piece is >= 3 mins, we have to collect parts continuously from the process.
There will not be any time left between time blocks
 If the Cycle time per piece is < 3 mins, we have to collect leave some time between the time blocks
 When we leave the time between the time blocks, we have to ensure that the time difference between
time blocks are not equal and are totally random
 This is done to ensure that, if there are any cyclical variations, it is captured in the data

T1 T2 T3 T4 T5 T6

WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling – Rules summary

Slide: 14

 3 or 5 parts continuously collected within the time block

 Minimum 25 time blocks
 Minimum time duration is 4 hours
 Events should be between time blocks
 One time block immediately after unplanned event(s)
 One time block before and after planned event(s)
 If the CT is >=3 mins, do not leave any time interval between
time blocks, otherwise leave time interval between time
blocks
 When leaving the time interval, the time interval should not
be same between the time blocks. It should be random
 Record all the unplanned and planned events happening
between the time blocks
WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling – Example

Slide: 15

P1 P2 P1 P2

31.82 31.81 31.81 31.8

T1 31.8 31.8 T7 31.83 31.81
31.8 31.82 31.82 31.82
31.81 31.83 31.82 31.81
T2 31.8 31.81 T8 31.83 31.81
31.80.06 31.82 31.83 31.82 31.83
31.79 31.8 31.78 31.81
T3 31.8 31.79 T9 31.81 31.8
31.79 31.79 31.81 31.8
31.82 31.81 31.82 31.8
T4 31.81 31.8 T10 31.83 31.81
31.79 31.82 31.83 31.81
31.81 31.81 31.84 31.82
T5 31.84 31.81 T11 31.81 31.82
31.82 31.81 31.8 31.81
Parameter: Center distance 31.83 31.82 31.83 31.82
T6 31.82 31.81 T12 31.8 31.82
Spec: 31.8 +/- 0.06 mm 31.8 31.8 31.81 31.8

WCMS
World Class Manufacturing Solutions
 6σ

Decide Sampling – Some examples

Slide: 16

 In a grinding machine, the wheel dressing is done

every 15 numbers and the wheel change is done
every 450 numbers. The shift changes every 8
hours and every 2 hours the input material lot is
changed. The cycle time of the process is 10 secs.
How will you decide the grouping to capture Time
to Time variation??

WCMS
World Class Manufacturing Solutions
 6σ

S4 Collect Data
Slide: 17

 Collect data as per the sampling decided in step

#3
 The instrument selected for measurement should have

a Least count (Resolution) of <= 10% of the tolerance

 The decimal place in the data collected should always

be in multiple of the least count (eg: if the LC is

0.005mm, then the data decimal should always be in
multiples of 0.005)
 Regroup the data if required due to unplanned events,

as discussed in step # 3

World Class Manufacturing Solutions

 6σ

S5
Multi-Vari analysis
Slide: 18

 Do Multi-vari analysis if the data contains multiple

streams.
 Multi-vari analysis is not required, if the data does

not contain multiple streams

 Refer to slide numbers 450 – 453 of Module – 1 for

Multi-vari analysis

World Class Manufacturing Solutions

 6σ

S5
Multi-Vari analysis
Slide: 19

Finding Part to Part variation

P1 P2 P1 P2

31.82 31.81 31.81 31.8

T1 31.8 31.8 T7 31.83 31.81
31.8 31.82 31.82 31.82
0.020 0.020 0.020 0.020
0.000 0.000
31.81 31.83 31.82 31.81
T2 31.8 31.81 T8 31.83 31.81
31.82 31.83 31.82 31.83
0.020 0.020 0.010 0.020
0.000 -0.010
31.79 31.8 31.78 31.81
T3 31.8 31.79 T9 31.81 31.8
31.79 31.79 31.81 31.8
0.010 0.010 0.030 0.010
0.000 0.020
31.82 31.81 31.82 31.8
T4 31.81 31.8 T10 31.83 31.81
31.79 31.82 31.83 31.81
0.030 0.020 0.010 0.010
0.010 0.000
31.81 31.81 31.84 31.82
T5 31.84 31.81 T11 31.81 31.82
31.82 31.81 31.8 31.81
0.030 0.000 0.040 0.010
0.030 0.030
31.83 31.82 31.83 31.82
T6 31.82 31.81 T12 31.8 31.82
31.8 31.8 31.81 31.8
0.030 0.020 0.030 0.020
0.010 0.010

World Class Manufacturing Solutions

 6σ

S5
Multi-Vari analysis
Slide: 20

Finding fixture to fixture variation

P1 P2 P1 P2

31.82 31.81 31.81 31.8

T1 31.8 31.8 T7 31.83 31.81
31.8 31.82 31.82 31.82
31.807 31.810 31.820 31.810
0.003 0.010
31.81 31.83 31.82 31.81
T2 31.8 31.81 T8 31.83 31.81
31.82 31.83 31.82 31.83
31.810 31.823 31.823 31.817
0.013 0.007
31.79 31.8 31.78 31.81
T3 31.8 31.79 T9 31.81 31.8
31.79 31.79 31.81 31.8
31.793 31.793 31.800 31.803
0.000 0.003
31.82 31.81 31.82 31.8
T4 31.81 31.8 T10 31.83 31.81
31.79 31.82 31.83 31.81
31.807 31.810 31.827 31.807
0.003 0.020
31.81 31.81 31.84 31.82
T5 31.84 31.81 T11 31.81 31.82
31.82 31.81 31.8 31.81
31.823 31.810 31.817 31.817
0.013 0.000
31.83 31.82 31.83 31.82
T6 31.82 31.81 T12 31.8 31.82
31.8 31.8 31.81 31.8
31.817 31.810 31.813 31.813

World Class Manufacturing Solutions

 6σ

S5
Multi-Vari analysis
Slide: 21

Finding Time to Time variation 0.027 mm

P1 P2 P1 P2

31.82 31.81 31.81 31.8

T1 31.8 31.8 T7 31.83 31.81
31.8 31.82 31.82 31.82
0.020 0.020 0.020 0.020
31.808 31.815
31.81 31.83 31.82 31.81
T2 31.8 31.81 T8 31.83 31.81
31.82 31.83 31.82 31.83
0.020 0.020 0.010 0.020
31.817 31.820
31.79 31.8 31.78 31.81
T3 31.8 31.79 T9 31.81 31.8
31.79 31.79 31.81 31.8
0.010 0.010 0.030 0.010
31.793 31.802
31.82 31.81 31.82 31.8
T4 31.81 31.8 T10 31.83 31.81
31.79 31.82 31.83 31.81
0.030 0.020 0.010 0.010
31.808 31.817
31.81 31.81 31.84 31.82
T5 31.84 31.81 T11 31.81 31.82
31.82 31.81 31.8 31.81
0.030 0.000 0.040 0.010
31.817 31.817
31.83 31.82 31.83 31.82
T6 31.82 31.81 T12 31.8 31.82
31.8 31.8 31.81 31.8
0.030 0.020 0.030 0.020
31.813 31.813

World Class Manufacturing Solutions

 6σ

Multi-Vari analysis
Slide: 22

 Part to Part variation – 0.040 mm

 Fixture to Fixture variation – 0.020 mm

 Time to Time variation – 0.027 mm

 Is Part to Part variation highest ??

 The answer is ‘Yes’, hence we can proceed to Step

-7

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
S7
variation
Slide: 23

 Why do we have to check the consistency of Part to Part

variation?
 Because, we can estimate the part to part variation only when it is
consistent
 Why do we have to estimate the Part to Part variation?
 Because, the part to part variation calculated in step # 5 is only actual
value based on 4 hours data and we cannot declare that this is the
variation for ever. (Eg: in the earlier example, we cannot declare that
0.040mm is the part to part variation for ever in the machining process
which generates the center distance in the wiper housing)
 From this step till Step – 13, use the Variation analysis format
provided along with the training material

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation
Slide: 24

 Proceed to this step only if the Part to Part variation is highest in Multi-
vari analysis
 This step checks the consistency of the Part to Part variation, so that we
can “estimate the Part to Part variation”, if it is consistent
 Consistency is checked through three filters
 Check whether all the ranges are within the control limits (first check for
consistency)
 Check whether the stratification level is >3

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation
Slide: 25

 Check # 1
 Part to Part variation is reflected in the “Range”
 Consistency of the Range is checked based on the assumption that
Range follows “Chi-square distribution” (2)
 If the Range is not consistent, do DOE to identify the source of
variation
 The reasons for inconsistency are
Machine
Measurement
Process design
Operator
Input material

Six sigma BB training Ver:3.00

WCMS
World Class Manufacturing Solutions
 6σ
Consistency of Part to Part
variation
Slide: 26

Variation Analysis

Part Details
Gage
Characteristic Center distance Part number WP101 number D101

Unit of Part Description Gage

Measurement mm Wiper housing description Dial gage
Date Gage Least
Target Value 31.8 10th January 2004 count 0.01mm
  Gage R&R
Tolerance 0.12   value 7.20%
USL
Gage Bias
and
Study dates and Uncertainity
31.86 Shift 8th January, A shift (%) As per calibration report
LSL Any other
31.74 details  

Data Grouping and Sample details

Number of Groups (Number of
time blocks x number of streams)
24
Number of data in each group 3

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation 3

Slide: 27

60
UCL
50

40

30
R-BAR
20

10

Part to Part variation is consistent

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation 3

Slide: 28

An example of inconsistency in part to part variation

R Chart for Total -

1
10
1 11 1
1
1 1
Sample Range

11 11 1 1 1
1 1 11 1
1 1 1 1 1
UCL=5.476
5

R=2.732

0 LCL=0

0 100 200 300

Sample Number

World Class Manufacturing Solutions

 6σ

Consistency of Part to Part

Slide: 29
variation
 Check # 2
R Chart for C1
 Check whether stratification
level is >3 2 UCL=2.059

 If the stratification level is

Sample Range
<=3, then the instrument is
not capturing the part to part 1

variation, which is due to very R=0.8

less part to part variation

 In this situation, stop the 0 LCL=0

analysis, no statistical control 0 10 20

is required for this parameter Sample Number

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation
Slide: 30

 Check # 3
 Check whether is there any of R Chart for C1
the following patterns 0.2 UCL=0.1950
7 points continuously
increasing or decreasing

Sample Range
7 points on one side of the
0.1
average range R=0.07576

 If we get any of the these

patterns, then also the part to 0.0 LCL=0
part variation is inconsistent,
which could be due to any of 0 10 20
Sample Number
30

the 5 reasons we had

discussed and we have to do
DOE

World Class Manufacturing Solutions

 6σ
Consistency of Part to Part
variation
Slide: 31

Checking the consistency of part to part variation (Step 7)

Average Range (R-bar) (Round off to one decimal more than the
data) 0.0192

Upper Control Limit (UCL) = D4*R-bar (Round off to the same

decimal as data) 0.05

Lower Control Limit (LCL) = D3*R-bar 0

Yes / No
Is the Part to Part variation consistent
If the Part to Part variation is not consistent, STOP, do not proceed further. Plan for DOE

Is the Range Chart plotted Yes / No

Are the Stratification levels more than 3     Yes / No

If the stratification level is <=3, then the Part to Part variation is very less and the parameter does not require any monitoring. STOP. Do
not proceed further

Are there 7 consequtive points increasing, decreasing and one Yes/NO If yes write  
side of the Mean range causes (if
possible)

World Class Manufacturing Solutions

 6σ

S9 Estimating Part to Part variation

Slide: 32

 Why do we have to estimate when the actual part to

part variation information is available ??
 Because, the actual part to part variation is the
variation identified only for 4 hours production
 Our objective is we have to estimate the variation for
every in the process
 This estimation is done “only once” in the process

World Class Manufacturing Solutions

 6σ

Estimating Part to Part variation

Slide: 33

  is estimated using only Range and not the entire data .

Why??
 Estimation of Part to Part variation is given by
  = R-bar/d2
 What variation does this “” represent
  calculated using R-bar, will indicate only part to part variation
 What variation will the “” calculated for the entire data
represent
  calculated using the entire data, will indicate the total variation (Part
to part, time to time and stream to stream)
 Since, our objective is to estimate only the Part to Part
variation, we should always estimate  using R-bar/d2.

World Class Manufacturing Solutions

 6σ
Z - table

Slide: 34

Six sigma BB training Ver:3.00

WCMS
World Class Manufacturing Solutions
 6σ

Estimating Part to Part variation

Slide: 35

 Example of estimation.
 Two estimations are done on the ambient temperature
that will be tomorrow
A) 39 degrees
B) 32-45 degrees
 Which estimation is most likely to come true the next
day ??
 This is known as the “Confidence level”.
 For estimating the variation, we should fix the
confidence level with which we are going to estimate.

World Class Manufacturing Solutions

 6σ

Estimating Part to Part variation

Slide: 36

X-bar

Xi Xi

Area to the right

Area to the left

Z(Xi) = (Xi-X-bar)/

World Class Manufacturing Solutions

 6σ

Estimating Part to Part variation

Slide: 37

 Confidence level for estimating the part to part

variation is 99.73%
 Refer to ‘z’ table and tell the value of ‘z’

Confidence level Z
55
60
75
99

World Class Manufacturing Solutions

 6σ

Estimating Part to Part variation

Slide: 38

 = 0.011 mm
6  = 0.066 mm

Estimated part to part variation at 99.73% CL is 66

microns
We already know that the actual part to part variation is
40 microns

World Class Manufacturing Solutions

 6σ

S10 Checking for Normality

Slide: 39

 Do Normality test in Minitab

World Class Manufacturing Solutions

 6σ

S11 Estimating Rejection percentage

Slide: 40

Do Monte-carlo Simulation as learnt in the last module

World Class Manufacturing Solutions

 6σ

Slide: 41
Estimating Rejection percentage

Average 31.813 Z (usl) 4.27

 0.011 Z (lsl) -8.18
6 0.066 Estimated Rejection 0
above USL

USL 31.86 Estimated Rejection 0

below LSL

LSL 31.74 Cpk (usl) 1.42

Cpk (lsl) 2.72

World Class Manufacturing Solutions WCMS

 6σ

Estimating Rejection percentage

Slide: 42

 This rejection percentage is estimated “Only once” for a

parameter in the process.
 Since Cpk is nothing but the estimated rejection percentage
above and below specifications, Cpk has to be “estimated only
once”
 Our time and effort should be spent on monitoring and
controlling the process either using pre-control or X-bar and
Range chart
 Do not waste the resources in collecting data and calculating
the Cpk at some frequency, instead we should use the
resources for monitoring , controlling and doing Process
improvement

World Class Manufacturing Solutions

 6σ

Estimating Rejection percentage

Slide: 43

 If we get high rejection on one Process Capability Analysis for RESPONSE

side and low rejection on another USL

Process Data
30.3020
LSL USL

side, we have to recalculate the Z

Target *
LSL 30.2180
Mean 30.2902
Sample N 60

and estimate the rejection using StDev (Within) 0.0081327

“Drawing Target” as the average Potential (Within) Capability

[Link]
[Link]
1.45
1.45

instead of using the “Actual [Link]

Cpk
8.88
0.48

30.22 30.24 30.26 30.28 30.30 30.32

Process average” as the average.

Cpm *

Observed Performance Expected Performance

% < LSL 0.00 % < LSL 0.00
% > USL 13.33 % > USL 7.34

There is a provision given in the

% Total 13.33 % Total 7.34

Process Capability Analysis for RESPONSE
VA format for doing this LSL USL
Process Data
USL 30.302
Target *
LSL 30.218
Mean 30.260
Sample N 60
StDev (Within) 0.0081327

Potential (Within) Capability

[Link] 5.03
[Link] 5.16
[Link] 5.16
Cpk 1.72

Cpm * 30.22 30.24 30.26 30.28 30.30 30.32

Observed Performance Expected Performance

% < LSL 0.00 % < LSL 0.00
% > USL 13.33 % > USL 0.00
% Total 13.33 % Total 0.00

World Class Manufacturing Solutions

 6σ

S12
Six sigma analysis
Slide: 44

 This is the last step of the analysis part

 Six sigma analysis is done to find out what type of statistical
monitoring and control is required for the parameter analyzed
 The following type of monitoring and control are decided
based on this analysis
 Pre-control chart
 X-bar and Range chart
 X-bar and Range chart with 100% inspection
 Only 100% inspection
 The type of monitoring is decided based on comparison of 6
with the tolerance of the parameter analyzed

WCMS
World Class Manufacturing Solutions
 6σ M
2
3-2

Six sigma analysis

S12
Slide: 45

 There are three possible conditions

 6  <= 75% Tolerance
 6  > 75% and <= 100% of Tolerance
 6  > Tolerance
 The explanation of each of these conditions is
given in the following slides

Six sigma BB training Ver:3.00

WCMS
World Class Manufacturing Solutions
 6σ

Six Sigma analysis

Slide: 46

 6  <= 75% Tolerance

Upper Specification Limit

B
Sigma Sigma

C
E

Six sigma BB training Ver:3.00

Lower Specification Limit
World Class Manufacturing Solutions
 6σ

Six Sigma analysis

Slide: 47

 Analysis – Situation – 1
 Estimated Part to Part variation is less than the tolerance
 With the Positions A,B or C, all the products manufactured will meet the
specification as long as the Process stays in Control (Range Chart)
 In this situation, process need not be centered at the target. There can be a drift
in the actual process setting from the target as long as the rejection is Zero.
How much drift is allowed to have zero rejection can be calculated
 In this situation, the type of control that can be decided are
 Pre-control chart (When the actual part to part variation is <=50% of
tolerance)
 X-bar and Range chart (When the actual part to part variation is > 50% of
the tolerance)
 In this situation, there should not be any 100% inspection done for the
parameter. If 100% inspection is currently done, it should be removed

World Class Manufacturing Solutions

 6σ

Six Sigma analysis

Slide: 48

 6  > 75% and <= 100% of Tolerance

Upper Specification Limit

B
Sigma Sigma

Lower Specification Limit

World Class Manufacturing Solutions

 6σ

Six Sigma analysis

Slide: 49

 Analysis – Situation – 2
 This is a cat on the wall situation. The estimated part to part variation
is equal to the tolerance
 In this situation, the “actual part to part variation” will be slightly less
than the tolerance
 The following type of control to be used in this situation
 X-bar and Range chart with 100% inspection

 Improvement actions based on DOE to reduce the part to part

variation

World Class Manufacturing Solutions

 6σ

Six Sigma analysis

Slide: 50

 6  > Tolerance

Upper Specification Limit

B
Sigma Sigma

Lower Specification Limit

World Class Manufacturing Solutions

 6σ

Six Sigma analysis

Slide: 51

 Analysis – Situation – 3
 Estimated Part to Part variation is more than tolerance
 This is the worst situation
 In this situation, the following type of control should be
selected
 Only 100% inspection
 Use DOE tools to reduce the Part to Part variation

World Class Manufacturing Solutions

 6σ

Six sigma analysis

Slide: 52

For the Wiper housing , Center distance parameter

6  <= 75% of Tolerance
Actual Part to Part variation (0.040mm) is < 50%

of tolerance (0.060mm)
Hence the type of statistical monitoring and

control required is “Pre-control limit”

World Class Manufacturing Solutions

 6σ

Calculation of control limits for average

Slide: 53

 Calculate the control limits for average only when the type
of monitoring selected is X-bar and Range chart
 These limits will be used for controlling the Time to Time
variation during monitoring
 UCL = Average + A2*R-bar
 LCL = Average – A2*R-bar
 Do not calculate these limits, when the monitoring method
selected is “Pre-control chart”

World Class Manufacturing Solutions

 6σ

Calculation of control limits for average

Slide: 54

 Fundamental rules for calculation of the limits for

average are
 Averages of the data will always follow normal
distribution
 (x-bar) = /n (n is the data in each group)

World Class Manufacturing Solutions

 6σ

On-line monitoring
Slide: 55

 The type of on-line statistical monitoring and control that are

selected in Step 12 are
 Pre-control chart
 X-bar and Range chart
 X-bar and Range chart with 100% inspection
 These monitoring has to be done for ever for that parameter
analysed in the process.
 The frequency of monitoring will be decided based on the
type of parameter and the method of monitoring
 On-line monitoring has to be done by “Operator” (Person who
is daily running the process)
 The following slides explain how each of these monitoring
should be done

World Class Manufacturing Solutions

 6σ

On-line monitoring
Slide: 56

 For Both pre-control and X-bar-R charts, follow

the following guidelines for deciding the timing
between groups (time blocks)
 One group before and after planned events
 One group after Unplanned events
 Timing between groups should be random

World Class Manufacturing Solutions

 6σ

On-line monitoring – Pre-control chart

Slide: 57

 Calculate the Pre-control limits using the following

rules
 Drawing target is the middle line
 50% of the tolerance is the “Green Zone”
 Divide this 50% equally above and below the middle
line (above 25% and below 25%)
 Draw the USL and the LSL line
 Top 25% above green zone and the bottom 25%
below green zone will be the “Yellow” zone
 Above USL and below LSL are the “Red Zones”

World Class Manufacturing Solutions

 6σ

On-line monitoring - Pre-control chart

Slide: 58

Yellow Zone, 25 %of

Tolerance

Green Zone, 50% of

Target Tolerance

Yellow Zone, 25% of

Tolerance

WCMS
World Class Manufacturing Solutions
 6σ

On-line monitoring - Pre-control chart

Slide: 59

31.86

31.83

Target 31.80

31.77

31.74

WCMS
World Class Manufacturing Solutions
 6σ

Online monitoring – Pre-control chart

Slide: 60

 Rule for Operation

The process is setup approved when the readings of three
consecutive jobs fall in the green zone
After that two jobs are sampled continuously, the decision
to adjust the process is based on the following guidelines
If two points lie in the green zone, process should not be
adjusted
If one point lie in the green zone and one in yellow zone
process should not be adjusted
If two points fall in the same yellow zone, the process has to be
adjusted to bring to the target
If one point fall in one yellow zone and another in another
yellow zone, process should be stopped and the reason for
increased part to part variation investigated

World Class Manufacturing Solutions

 6σ

On-line monitoring - Pre-control chart - Example

Slide: 61
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

INSERT
CHANGE

OFFSET

World Class Manufacturing Solutions

 6σ

On-line monitoring - Pre-control chart - Example

Slide: 62

World Class Manufacturing Solutions

 6σ

On-line monitoring – X-bar, Range chart

Slide: 63

 In Pre-control chart, we have seen that both the part to part

and time to time variations are monitored using the same chart
 In X-bar and Range chart, part to part and time to time
variations are monitored separately.
 Range monitors part to part variation
 Average monitors time to time variation
 Since we are tracking the part to part and time to time
variations using range and average, in this monitoring method,
the operator has to do the calculation of the range and the
average
 Because of the calculations, x-bar and range monitoring is
slightly cumbersome compared to pre-control chart

World Class Manufacturing Solutions

 6σ

On-line monitoring – X-bar, Range chart

Slide: 64

 Collect the same number of parts continuously as done during the

Variation analysis (eg: if for VA we had collected 3 parts we have to
collect only 3 continuous parts for monitoring also)
 Calculate the average and Range
 Plot these values in the average and Range chart
 If the average and range lies within the control limit, the Process setup is
OK and production can be started
 After sometime, sample again the same number of parts continuously and
calculate the average and range
 If both average and range are within the limit, continue production
 If average is outside the limit, adjust the process to bring it to mean
 If range is outside the limit, stop the process and investigate the reasons
along with the process owner

World Class Manufacturing Solutions

 6σ

On-line monitoring – Range chart

Slide: 65

Range monitors Part to Part variation

If the Range goes outside the Upper limit,
this indicates that the Part to Part variation has
R Chart for C11
become inconsistent.
In this situation, Operator should stop the
UCL=0.02083
process and get the help of Process owner to 0.02

investigate the reasons for inconsistency

(Machine, Material, Process design, Operator,

Sample Range
Measurement)
0.01
R=0.008093

0.00 LCL=0

0 10 20 30 40
Sample Number

World Class Manufacturing Solutions

 6σ

On-line monitoring – Average chart

Slide: 66

Average chart monitors Time to Time

variation which is correctable by the Operator
himself X-bar Chart for C11
The average chart tells whether the Time to 17.008
Time variation is happening only due to Part to UCL=17.01
Part variation or it is also happening due to
some events

Sample Mean
The Process has to be adjusted and brought
back to the Mean by the Operator under the 16.998 Mean=17.00
following conditions
 Average goes outside the control limits

LCL=16.99
16.988

0 10 20 30 40
Sample Number

World Class Manufacturing Solutions

 6σ

On-line monitoring – Die dominant

(Metal forming) processes
Slide: 67

 For Die dominant processes (eg: forging, injection molding,

blow molding, pressure die casting), generally the part to part
variation will be very less.
 In these processes, statistical monitoring and control is
required only to know two aspects
 What should be Die design dimension to start with
 When the Die should be replaced or sent for maintenance
 In such processes, the application of SPC is used only for
these two aspects
 The next slides explain how to design the starting die
dimension and when to replace the die

World Class Manufacturing Solutions

 6σ

On-line monitoring – Die dominant

(Metal forming) processes
Slide: 68

 How to decide the starting die dimension

 Step – 1
 Do VA for a parameter which will be an indicative dimension of the die wear. Follow the
same methodology discussed earlier for doing the VA
 Step – 2
 Let us take for example, the P-P variation of the parameter is 0.018mm and the tolerance
is 0.1mm (+/- 0.05mm)
 Minimum ‘Z’ value for getting almost zero rejection is 4. Generally considering the FOS
(Factor of Safety), Z value is taken as 5 for calculating
 Z (usl) = USL – X-bar/Sigma
 Z (lsl) = LSL – X-bar/Sigma
 If P-P is 0.018mm, Sigma is 0.003mm
 Substituting Z = 5,  = 0.003 and USL = 0.05, in the above formula, we can find out X-
bar = 0.035
 This means that the limits for controlling this parameter will be +0.035mm to -0.035 mm.
 Step – 3
 Use these limits, to design and manufacture the die dimension and replace the die when
the parameter reaches the limit calculated

World Class Manufacturing Solutions

 6σ

Constants used in calculations

Slide: 69

 Constants used in VA
Number of A2 D3 D4 d2
observations in
a group
2 1.88 0 3.27 1.128
3 1.02 0 2.57 1.693
4 0.73 0 2.28 2.059
5 0.58 0 2.11 2.326
6 0.48 0 2.00 2.534
7 0.42 0.08 1.92 2.704
8 0.37 0.14 1.86 2.847
9 0.34 0.18 1.82 2.970
10 0.31 0.22 1.78 3.078

World Class Manufacturing Solutions

 6σ

Z - table
Slide: 70

Z Rejection
0 – 0.5 50 – 31 %
0.6 – 1.0 27 – 16 %
1.1 – 1.5 13 – 6.5 %
1.6 – 2.0 5.5 – 2.2 %
2.1 – 2.5 17,000 – 6,200 ppm

2.6 – 3.0 4,700 – 1350 ppm

3.1 – 3.5 970 – 230 ppm

< 3.5 < 230 ppm

World Class Manufacturing Solutions

6σ

Chapter - 2
71

 Simplex method

2D

World Class Manufacturing Solutions

6σ

Simplex

When to be used ?

• To find out best possible optimal settings for process having 2 design parameters

driving the process

World Class Manufacturing Solutions

6σ

Simplex

What is required?(Data collection):

• Data of the four corner settings (response) of the graph shown in the next slide
• Data of the center setting (response)as shown in the next slide
• Following is the example of optimizing the drill life by varying Speed and feed
• We need to monitor the life without detiorating the quality parameters.
• We need to move in the direction we are getting more life as shown.
• And stop till we get the center trial in the box as the optimal setting.

World Class Manufacturing Solutions

6σ

[Link]

[Link] need to continue trials till the

590 [Link] need to move in the direction
center setting is the optimal setting
where we are getting more life
570
[Link] we get Max tool
550 life 700 at Speed 525 and
feed 425 from simplex tool
530 This is the final optimal
Life 670
510 Life 650
(480,450) (480,500)
setting.

490
Life 600 Life 620
470 (450,400) (465,475) Life 630
Feed

(450,550)
450 (425,435)
(450,450) (450,500)
Life 630 Life 690 (425,525)
430 Life 500 Life 700
(400,550)
410 (420,400) (420,450) (400,500)
Life 610 Life 640
Life 400 Life 450
390
370
[Link] need to collect Response tool life at 5 settings
350
350 370 390 410 430 450 470 490 510 530 550 570 590
Speed

World Class Manufacturing Solutions