Statistically Quality Design Statistically Quality Design Topics Topics – Four perspectives in quality improvement Four perspectives in quality improvement – Review DOE topics and terminologies Review DOE topics and terminologies – Implementation plan and procedure for Implementation plan and procedure for experimental design experimental design – Full factorial design and Yates’ Full factorial design and Yates’ algorithm algorithm – Full factorial design example: improving Full factorial design example: improving wave solder process at TDY company wave solder process at TDY company – Concepts and examples for conducting Concepts and examples for conducting Block and Latin square designs Block and Latin square designs
70
Embed
Statistically Quality Design Topics –Four perspectives in quality improvement –Review DOE topics and terminologies –Implementation plan and procedure for.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Four Perspectives in Quality Improvement Four Perspectives in Quality Improvement
1. Downstream Perspective: Customer’s Quality – such as: fuel consumption, noise, failure rates, pollution, etc.
3. Upstream: Quality of Design (robustness of objective function)
– good for design & development after product planning
4. Origin: Quality of Technology (robustness of technology )– good for technology development prior to product planning– functionality of generic function– example: Hook’s Law for spring
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Design of Experiments (Topics)Design of Experiments (Topics)CRDCRD
Full FactorialFull Factorial- all possible combinations of factors and levelsall possible combinations of factors and levels
Fractional FactorialFractional Factorial- assume some interaction will not occur, a factor is assignedassume some interaction will not occur, a factor is assigned
Latin SquareLatin Square- each level of each factor appears only once with each level of two other factorseach level of each factor appears only once with each level of two other factors
On-line V.S. Off-line Quality ControlOn-line V.S. Off-line Quality Control
Design of Experiments (Topics cont.)Design of Experiments (Topics cont.)
System DesignSystem Design– the selection of materials, parts, equipment, and process parametersthe selection of materials, parts, equipment, and process parameters
Parameter DesignParameter Design– study the effect of noise factor in DOEstudy the effect of noise factor in DOE
Tolerance DesignTolerance Design– the specification of appropriate tolerance, product and process parametersthe specification of appropriate tolerance, product and process parameters
Signal to Noise Ratio (S/N Ratio)Signal to Noise Ratio (S/N Ratio)– a comparison of the influence of control factors (signal) to that of a comparison of the influence of control factors (signal) to that of
noise factorsnoise factors
Orthogonal ArrayOrthogonal Array− a design where correlation between factors is zeroa design where correlation between factors is zero
Outer ArrayOuter Array– in parameter design to identify the combination of noise factorsin parameter design to identify the combination of noise factors
Quality CharacteristicQuality Characteristic– the response variable (output)the response variable (output)
InteractionInteraction– the combination of two factors generates a result that is different the combination of two factors generates a result that is different
from individual factor. from individual factor. – main effect vs. interaction effectmain effect vs. interaction effect
DOFDOF– independent piece of informationindependent piece of information
ResolutionResolution– number of letters in the shortest length in defining relationnumber of letters in the shortest length in defining relation– the lower the number, the more saturated the design isthe lower the number, the more saturated the design is
ANOVA (ANOVA ( 變異數分析變異數分析 ))
The method of analyzing data collected by CRD / RCBDANOVA equation
ANOVA Table
...... iijiij YYYYYY
i j
iiji j
i
i jij YYYYYY
2.
2...
2..
=> SST = SStrt + SSE
source of variationsource of variation d.f.d.f. SSSS MSMS FF
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Implementation Plan and Procedure for Implementation Plan and Procedure for Experimental DesignExperimental Design
Basically, a twelve-steps (procedure) approachfor conducting any experimental design can be divided into the following three stages:
Stage 1 (( 準備及設計選擇階段準備及設計選擇階段 ))
1.1. Define the problems and state the objective of the experimentDefine the problems and state the objective of the experiment
2.2. Select quality characteristic (response) and input variables Select quality characteristic (response) and input variables (factors)(factors)
3.3. Determine the desired number of runs and replicationsDetermine the desired number of runs and replications
4.4. Consider the randomization of runs during the selection of the best Consider the randomization of runs during the selection of the best design typedesign type
Stage 2 (( 實驗及分析資料階段實驗及分析資料階段 ))
1.1. Conduct the experiment and record the dataConduct the experiment and record the data
2.2. Analyze the data using analyze of mean, analysis of varianceAnalyze the data using analyze of mean, analysis of variance
3.3. Use Yates’ algorithm and normal probability plot to determine Use Yates’ algorithm and normal probability plot to determine the significant main and interaction effectsthe significant main and interaction effects
Stage 3 (( 建立預估模式及確認評估階建立預估模式及確認評估階段段 ))
1.1. Develop a fitted model using regression analysisDevelop a fitted model using regression analysis
2.2. Draw conclusion and make predictionDraw conclusion and make prediction
4.4. Assess results and make decisionAssess results and make decision
Steps for Experimental DesignSteps for Experimental Design
1.1. Statement of the problemStatement of the problem:_________________________________:_________________________________
(During this step you should estimate your current level of quality by way of (During this step you should estimate your current level of quality by way of Cpk, dpm, or total loss. This estimate will be compared with improvements Cpk, dpm, or total loss. This estimate will be compared with improvements found after Step 7.)found after Step 7.)
2.2. Objective of the experimentObjective of the experiment:_______________________________:_______________________________
3.3. Star DateStar Date:_____________ ; :_____________ ; End DateEnd Date:_____________:_____________
(also known as responses, dependent variables, or output variables). These (also known as responses, dependent variables, or output variables). These characteristics should be related to customer needs and expectations.characteristics should be related to customer needs and expectations.
ResponseResponse TypeType Anticipated RangeAnticipated Range How will you measure the How will you measure the response?response?
112233
Steps for Experimental Design (cont.)Steps for Experimental Design (cont.)
5.5. Select factorsSelect factors (also know as parameters or input variables) which are anticipated to have an (also know as parameters or input variables) which are anticipated to have an
effect on response.effect on response.
6.6. Determine the number of resources to be used in the experimentDetermine the number of resources to be used in the experiment (Consider the desired number, the cost per resource, time per experimental (Consider the desired number, the cost per resource, time per experimental
trial, and maximum allowable number of resources.)trial, and maximum allowable number of resources.)
7.7. Which design types and analysis strategies are appropriateWhich design types and analysis strategies are appropriate?? (Discuss advantage and disadvantages of each.)(Discuss advantage and disadvantages of each.)
FactorFactor TypeType Controllable or Controllable or NoiseNoise
Range of Range of InterestInterest LevelsLevels Anticipated Anticipated
Interactions WithInteractions WithHow How
MeasuredMeasured11
22
33
Steps for Experimental Design (cont.)Steps for Experimental Design (cont.)
8.8. Select the best design type and analysis strategy to suit your needsSelect the best design type and analysis strategy to suit your needs
9.9. Can all the runs be randomizedCan all the runs be randomized?__________________________?__________________________
Which factors are most difficult to randomize?________________Which factors are most difficult to randomize?________________
10.10. Conduct the experiment and record the dataConduct the experiment and record the data (Monitor both of these events for accuracy)(Monitor both of these events for accuracy)
11.11. Analyze the data, draw conclusions, mark predictions, and do Analyze the data, draw conclusions, mark predictions, and do confirmatory testsconfirmatory tests
12.12. Assess results and make decisionsAssess results and make decisions (Evaluate your new state of quality and compare with the quality level prior to the (Evaluate your new state of quality and compare with the quality level prior to the
improvement effort. If necessary, conduct more experimentation.)improvement effort. If necessary, conduct more experimentation.)
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Full Factorial Experimental DesignFull Factorial Experimental Design
Principles:
Random Number TableRandom Number Table
There are 400 digits in this random number There are 400 digits in this random number table, 3 appears 41 times.table, 3 appears 41 times.
3 3 Factors, 2 LevelsFactors, 2 Levels
Four dimensional visibility with
823 test combinations in a full factorial matrix
Label The CellsLabel The Cells
8 TestCombinations 32
(1)(1) aa
bb abab
c c acac
bcbc abcabc
A- A+
B-
B+
B-
B+
C-
C+
Yates’ NotationYates’ Notation
823 Test Combinations
Cell A B AB C AC BC ABC (1) - - + - + + - a + - - - - + + b - + - - + - + ab + + + - - - - c - - + + - - + ac + - - + + - - bc - + - + - + - abc + + + + + + +
Yates’ NotationYates’ Notation
1624 Test Combinations
Yates’ Work SessionYates’ Work Session
Y = yield strength , PSI
A, B and C are concentrations of 3 separate elements
5858
5656
3636
3939
5151
5353
3434
3232
5353
48 48
5454
5959
4949
4949
5555
6161
A- A+
B-
B+
B-
B+
C-
C+
Determine the size of each contrast using Yates’ algorithm
What combination of elements will give the highest yield strength?
The Yates’ AlgorithmThe Yates’ Algorithm
two variables; A, B number of variables, n = 2 number of columns, n = 2
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Full Factorial Experiment Example: Full Factorial Experiment Example: Improving Wave Solder Process at TeledyneImproving Wave Solder Process at Teledyne• Objective :
– To determine the effect of flux type and lead length on the DFDAU wave soldering (WS) defects
• Planted steps for statistically designed experiment (1) Select output variables, 2 factors, 2 levels and 8 Runs (2) Randomize the sequence of runs and labels 8 DFDAU boards (3) Select two touchup operators to check the WS defects
consistency (4) Iso-plot the major WS defects for top/rear sides to compare one
operator against another (5) Analyze the data using ANOVA table with interactions or using
Yates' algorithm (6) Plot/interpret the results and draw the conclusions
Wave Soldering Process Flow ChartWave Soldering Process Flow Chart
Statistically Designed ExperimentStatistically Designed Experiment
Run No. Flux Type Lead Length LabelY1 New(OA) Trimmed Leads abY2 Old(RMA) Std. Lead Length (1)Y3 Old(RMA) Std. Lead Length (1)Y4 New(OA) Std. Lead Length bY5 Old(RMA) Trimmed Leads a Y6 New(OA) Trimmed Leads abY7 New(OA) Std. Lead Length b Y8 Old(RMA) Trimmed Leads a
New Flux = Alpha # 857Old Flux = RMAStd. Lead Length (IC connect. point not trimmed; IC 接點處之引線未被切平 )Trimmed Leads ( about .045”)
Yates’ AlgorithmYates’ Algorithm
Notations
422
(1)(1) aa
bb abab
Std. Leads Trimmed Leads
A- A+
B-
B+
OldFlux
NewFlux
ANOVA Table
ContrastsCell A B AB(1) - - +
a + - - b - + -
ab + + +
22 = 4 Combinations
Operator #1
1212
10.510.599
99
7.57.566
77
8899
88
7766
Operator #2
1111
9.59.588
99
7755
77
7777
77
4.54.522
Avg. Operator
10.510.5
1010
9.59.5
7.57.5
7.257.25
77
88
7.5 7.5
77
77
5.755.75
4.54.5
Std. Leads Trm. Leads
Old Flux
New Flux
Std. Leads Trimmed Leads
Old Flux
New Flux
Old Flux
New Flux
ANOVA Analysis For Two Factorial ANOVA Analysis For Two Factorial ExperimentExperiment
Two way ANOVA for flux type and lead length
Interaction Plot for Flux Type and Lead Interaction Plot for Flux Type and Lead LengthLength
TrimmedStd
10
9
8
7
6
Lead Length_aver
Mea
n
newold
Type_averFlux
Interaction Plot for response_averData Means
2 2 Factor Full Factorial Experiment Factor Full Factorial Experiment Summary and ConclusionsSummary and Conclusions
Summary of Findings:• ISPLOT reveals that 2 operators
were fairly consistent in calling out VIA defects
• VIA defects consist of 77 % vs. 93 % of total defects, 1 vs. 2
• Only the rear VIA defects are considered for output measures
• Defect level : 2941 ppm (Trimmed leads)
• Defect level : 3959 ppm (Std. lead length)
Conclusions:• The ANOVA/Yates’s Analysis
indicates lead length to be the most significant factor
• Interaction between flux and lead length proven to be the least significant factor
• 26 % improvement can be expected if using the “ Trimmed Leads“
• 23 % improvement can be expected if using the “ OA “ Flux
• Optimal range for the board temperature needs to be further studied
TopicsTopics– Four perspectives in quality improvementFour perspectives in quality improvement– Review DOE topics and terminologiesReview DOE topics and terminologies– Implementation plan and procedure for Implementation plan and procedure for
experimental designexperimental design– Full factorial design and Yates’ algorithmFull factorial design and Yates’ algorithm– Full factorial design example: improving wave Full factorial design example: improving wave
solder process at TDY companysolder process at TDY company– Concepts and examples for conducting Block and Concepts and examples for conducting Block and
Latin square designsLatin square designs
Latin Square (Latin Square ( 拉丁方格 拉丁方格 ))
AA BB CC
BB CC AA
CC AA BB
Model rkjiY ijkkjiijk ...,1,,
Operators I, II, IIIProcesses 1, 2, 3Material Source A, B, C
Operators
ProcessesI II III
1
2
3
Greco-Latin SquareGreco-Latin Square
I II III
1
2
3
A
C
B
Operators I, II, IIIProcesses 1, 2, 3Material # 1 Source A, B, CMaterial # 2 Source ,,
B
C
A
C
B
A
Latin Square DesignLatin Square Design
By using a Latin square design, three sources of variation, A, B and C, can be investigated simultaneously, providing there is no interaction between the three factors and also that each of them has the same number of levels r.
For example, suppose each factor has four levels denoted by
.,,,,,,,,,, 432143214321 CCCCandBBBBAAAA If factor A is associated
with the rows of the table and B with the columns of the table then each levels of factor C must appear once in each row and once in each column. In order to achieve this a systematic cyclic pattern, it canbe set down for the C’s as shown in the table. To randomize the design,the allocation of the A’s and B’s to the rows and columns is then carried out at random.
4B 2B 1B 3B
2A
4A
3A
1A
1C
4C
3C
2C
2C
2C
2C4C
4C
4C
1C
1C
1C3C
3C
3C
jBiA
Latin Square ModelsLatin Square Models
Latin Square Model: rkjiY ijkkjiijk ...,1,, The α’s, β’s, γ’s and ε’s are mutually independent.
Analysis of variance for a Latin Square designAnalysis of variance for a Latin Square design
The total sum of squares is divided into four component parts, onefor each source of variation and one for the residual.
SSCSSBSSASSTSSE
FCYr
SSC
FCYr
SSB
FCYr
SSA
FCySST
N
YFCrNyY
k
J
i
ijk
ijk
....1
....1
....1
..
.....,,...
2
2
2
2
22
Here Yi… is the sum of over the r observations in which factor A is at level i, with similar interpretation for Y.j. and Y..k and Y… is the sum of all the r2 observations.
The analysis and test statistics are summarized in the following ANOVA table.
ANOVA Table for Latin Square
Source d.f S.S M.S F
Factor A r-1 SSA
Factor B r-1 SSB
Factor C r-1 SSC
Residual r2 - 3r + 2 SSE
Total r2 - 1 SST
22̂
23̂2
4̂2
1̂
21
22 ˆˆ
21
23 ˆˆ
21
24 ˆˆ
ExampleExample
The following 4 X 4 Latin Square in which the effects of three factors, farm, type of fertilizer applied, and method of application (C1, C2, C3 or C4) on the yield crop are being investigated.
FertilizerFertilizer
FarmFarm
1A
2A
3A
4A
1B 2B 3B 4B
433C
238C 133C
236C133C
332C 432C 237C435C337C
129C
332C432C
333C 133C 235C
To ease the calculations, the data can be coded by subtracting 33 from each observation. Then the row and column totals and the totals for each method of application are calculated. ( 扣除 33, 不致影響 ANOVA 分析 )
FertilizerFertilizer
1 2 3 4 1 2 3 4 Total Total
1 1
Farm 2Farm 2
33
44
22
88
44
-2-2
TotalTotal 4 2 10 -4 124 2 10 -4 12
MethodMethod
TotalTotal -4 14 2 10 12 -4 14 2 10 12
40C 30C 10C 22C
25C 10C 34C 41C
10C 23C 42C 31C
31C 41C24C 14C
1C 2C 3C 4C
2
45954..4]0214)4[(
25934..4])4(1024[
13922..4])2(482[
85994..)4...(2000
0.91612..
1612...
2222
2222
2222
22222
2
2
SSCSSBSSASSTSSE
FCSSC
FCSSB
FCSSA
FCSST
FC
rNY
The calculations necessary for testing the significant of the threefactors are summarized in the following ANOVA table.
Since the critical value are F0.99(3, 6) = 9.78 and F0.999(3, 6) = 23.70, the farm effect is significant at 1 % level. The type of fertilizer used and the method of application are both significant at the 0.1 % level.
Open the Open the Latin Square.mtwLatin Square.mtw worksheet worksheet
Latin Square ModelsLatin Square Models
Stat ANOVA General Linear Model
The farm effect, the type of fertilizer used and the method of application are significant at α= 0.05