Philips SE Tool Box

Statistical Engineering Tools

sPM Contact Group
This Excel Workbook contains example spreadsheets which automate the calculations and graphic presentation of
various Shainin Statistical Engineering tools. Each spreadsheet is designed for a different tool application, and
can be extended (e.g. add more data points) or modified for related applications. An index of tools is given below.

It is suggested that when a tools is needed, it is copied to a new spreadshet and renamed appropriately (see
instructions below for copying). A recommended approach is to organize all experiments for one Statistical
Engineering problem into a single Excel workbook which contains separate sheets for each analysis step you
perform. Number successive sheets sequentially with an extension number, e.g. FPD2.1, Soln.3, ISO.2, MV4.4, etc.
This collected workbook will aid in creation of Management Reports and in preparation of SE Projects Reports for
Journeyman application or publication on the Philips Electronics Statistical Engineering Network, PE-SEN.

This Workbook makes use of Microsoft Office, including the embedding of objects in Word 6.0, PowerPoint 4.0a.
Instructions for editing each spreadsheet are given at the bottom of the sheet.

Please send comments, additions and corrections to the Editor, Walter N. Schreiner, Ph.D., Sr. Consultant and
Project Manager, CFT-North America, 21 Macaulay Rd., Katonah, NY 10536.
Phone: 914-962-5711, FAX: 914-962-0495

Sheet ID Tool Name and Description

BBvC Barrier B vs. C. B or C process must be closer to limit. Use Signif.xls instead.
BvC B vs. C. Rank Order ANOVA tests if B (Better) is better than C (Current). Use Signif.xls instead.
Clue Record of clues and explanation of clues by the Red / Pink Xs.
Corr Correlation plot between any two variables
CS Component Search. Compare good (BoB) and bad (WoW) parts to find differences.
CS-0 A Component Search tool for careful Stage 0, Stage Ia and Stage Ib analysis.
CS-Fac2 Stage 3 of Component Search for 2 factors. Quantify factor strengths with a DoE matrix.
CuSum CuSum Chart. Plots cumulative sum to detect changes in process centering.
ExSum Executive Summary. SE Project Report format with 5 sections.
Fac-2D 2-Dimen Full Factorial Analysis with significance test and interactions plots. Signals Red/Pink-X.
Fac-3D 3-Dimen Full Factorial Analysis with significance test and interactions plots. Signals Red/Pink-X.
Fac-4D 4-Dimen Full Factorial Analysis with significance test and interactions plots. Signals Red/Pink-X.
GYH Focussed Problem Definition. Histogram of actual problem data.
GYH-Cpt Focussed Problem Definition. Generic bell-shaped curve for annotation in PowerPoint.
ISO Isoplot. Measure 30 samples twice and determine the discrimination ratio.
MV-1 Multi-vari Chart. Single variable, multiple parts (up to 5) within units, between units, and over time.
MV-2 Multi-vari Chart. Two measurements within a unit (positions tracked), between units and over time.
MV-4 Multi-vari Chart. Four measurements within a unit (positions tracked), between units and over time.
MV-5 Multi-vari Chart. 4 categories : 4 meas/unit, 3 units/group, 3 groups/time, 6 time intervals
PC Paired Comparison. Table for up to 20 pairs and 32 candidate variables.
PDT Problem Definition Tree. Select a problem for analysis; start with Scrap, Rework and Returns.
Pr-Bl Blank SE Progress Report form. Used to report results of SE projects to PENAC Industrial Council
ProbPlt Probability Plot. Converts histogram to a Probability Plot and indicates curve shape.
Rand Random number generator for randomizing the order of statistical experiments.
ROA-2 Rank Order ANOVA significance test for Red and Pink-X variables from Fac-2D spreadsheet.
ROA-2S Same as ROA-2 but with separate table for all reverse Pink-X shuffles.
ROA-3 Rank Order ANOVA significance test for Red and Pink-X variables from Fac-3D spreadsheet.
Signif The 1999 format for all significance tests, such as BvC, BBvC, Tukey, etc.
Soln Solution Tree. Spreadsheet with outlined cells and lines to create a Solution Tree.
SP Scatter Plot. Find BoBs and WoWs by comparing Green Y to Factory Control variable for 50 parts.
Strat Strategy Diagram - for enumerating BoB / WoW contrasts, extracting Clues, and selecting leverage.
Task Task list for SE (or any) team. Use during meetings to record action items, who and when.
Tol P Tolerance Parallelogram. Converts a customer spec to a control limit on a factory variable.
Var S Variables Search. Identify critical variables from a list of suspects when list is 5 or more in length.
X-mR Simple X-Moving Range Chart for looking at a moving record of data. Up to 30 points.

Copyright Notice: Some pages in this tool box implement methods copyrighted by Shainin LLC
under license to Philips Electronics.

380330654.xls SE Tool Index Tool Ver. 2/27/99

 Philips Printed On: 03/01/2018

Problem or Project Definition Tree

Leader: Tree Date:
Champion:
Coach:
Team Members:
Problem Statement as Management Sees It

Tool / Reason to
Basis of Split
x% y% z% Eliminate Boxes

Consequences of
Scrap Rework Returns Pareto
the problem

Failure modes Pareto

Problem statement goes here for a PbDT.

Green-Y statement goes here for PjDT. State
all the focusing steps used on the tree.

Paste Green-Y Histogram for a Project

Definition Tree here

380330654.xls PDT Tool Ver. 8/12/99

 Philips Solution Tree Printed On: 03/01/2018

Problem Title
Leader:
Champion:
Coach: Paste Green-Y Histogram from Project Definition Tree here
Team Members:

Focused Green-Y statement goes here

Tool / Reason to
Basis of Split
Eliminate Boxes

Product / process Isoplot - Median of 3

DM DP gave DP/DM = 6.9
or measurement

Increasing family Within Piece Piece-to-Piece

Time-to-Time Multi-Vari
circles (same piece) (same time)

Red-X: abcd, confirmed at the xx% confidence level

380330654.xls Soln Tool Ver. 2/28/99

Philips Green-Y Histogram Printed On: 03/01/2018

Problem Title: Test Date: 22-Nov-99

Green-Y Statement: Read PjDT upwards, e.g. "Diameter of Part No. XYZ-272 made on Line 3 and measured at left end"
Meas System: Optical comparator
By:
Bin Upper Bin Raw
Limit Cnt Data
Focused Problem Definition Histogram
22.5 0 13
9 22 0 14.5
21.5 0 16.2
BoB's
8 21 0 13.2
20.5 0 14.8
Frequency

7 20 0 18.4
Desired 19.5 0 16.2
19 0 17.5
6
18.5 1 17.4
18 0 14.5
5
17.5 3 15.4
17 1 16.2
4
16.5 3 16.8
16 2 13.7
3 15.5 8 14.6
15 5 15.6
2 WoW's 14.5 4 14.2
WoW's
14 1 15.3
1 13.5 1 15.3
13 1 15.3
0 12.5 0 15.3
11
11.5
12
12.5
13
13.5
14
14.5
15
15.5
16
16.5

17.5
18
18.5
19

20
20.5

22
22.5
17

19.5

21
21.5
12 0 15.3
Green-Y 11.5 0 15.3
11 0 17.5
Raw Data Statistics Desired Distribution - Analytical Functions 15.3
Max: 18.4 Range: 5.4 Gaussian Exponential Poisson Other 14.6
Min: 13 Stdev: 1.254711 Plot? Y Plot? N Plot? N Plot? N 14.4
# Pts: 30 Average: 15.38667 Center 17 Offset 13 Mean 15 Parm #1 15.9
Stdev 6 Mean 2 Parm #2 15
Histogram Specifications Conclusions & Rationale Parm #3 14.9
Start: 11 USL: 14 Parm #4
Step: 0.5 LSL: 21 Parm #5

380330654.xls GY Tool Ver. 11/22/99

Philips Green-Y Histogram Printed On: 03/01/2018

380330654.xls GY Tool Ver. 11/22/99

Philips Printed On: 03/01/2018

Strategy Diagram
Project Title: LE Cambridge Machine Downtime
Date: 2/10/99
Leader: Walter Schreiner
Champion: Ken Ballschmieder
Coach: Dick Shainin
Team Members: David Keith
John Cooper

Contrasts Clues and Possible Leverage

BoB Incidents
WoW Incidents Problem varies with time. Not sure how to leverage this.
WoW Location

BoB Location
Concentration diagram with respect to feeding system identified confinement
WoW Location Clue. Little other leverage possible
WoW Head

BoB Head
WoW Head Heads are unique to the part being fed. No leverage.
WoW Machine

BoB Machine
Once confinement was identified as the location, this contrast could be
WoW Machine leveraged (LE vs. GNA machines).
WoW

380330654.xls Strategy Tool Ver. 2/27/99

Philips Isoplot Tool Printed On: 03/01/2018
Title : Problem TItle
Variable Name
Name [Units]
[Units] Green Y : Variable Name [Units]
Meas Sys:
70 By :
Date : DT =
dP:
dP: Pass
Pass 45° Measurement Sequence & Data Table
dT: N/A Max/2nd Sample # Trial 1 Trial 2 Order #2
1 55 54 1
65 2 58 59 2
3 53 54 3
4 58 58 4
5 52 52 5
6 60 60 6
60
7 56 56 7
8 49 49 8
9 43 43 9
10 59 59 10
11 57 55 11
Trial 2

55
12 60 60 12
13 59 59 13
14 50 49 14
2nd 15 46 48 15
50 16 51 50 16
17 60 61 17
18 51 51 18
19 50 52 19
20 50 51 20
45
Max 21 46 43 21


l 2
DM 2 22 53 55 22
DP = 23 52 51 23
2 2
24 55 55 24
25 55 55 25
40
40 45 50 55 60 65 70
26 47 47 26
27 55 56 27
Trial 1 28 54 54 28
29 59 59 29
Note: Perp distance to fit line is approximated by averaging y=f(x) and x=f(y). 30 62 61 30

Sausage Measurement Product/Process Discrimination Ratios Max 62 61

l = 26.4261 DM = 2.99767 DP = 18.565 l /DM = 8.81557 DP/DM = 6.2 Min 43 43
4sM = 3.09005 d2*sP = 17.1685 DT/DM = 0 Range 19 18
Conclusions and Rationale :

380330654.xls ISO Tool Ver. 9/2/99

 Philips Components Search Tool Printed On: 03/01/2018
Problem Title: Steel Strength Date:
Green Y: Strength [Kpsi] By:

Stage I Components Search - Decision Limits

Run # BoB WoW Comments Components Search
Part ID = R393f R387c 50.0
Decision
47.5
1 40 0 Initial Measurement Limits
BoB
2 35 2 First Reassembly 40.0
3 38 5 Second Reassembly
4 Third Reassembly 30.0 28.6
5 Fourth Reassembly
3 Stage I Group Size
20.0
38 2 Median
5 5 Range 11.5
36 Dm 10.0

5 R-bar
V1, V2 Passed Rule #1 (EC = 6) 0.0
1.49 Passed Rule #2 (Dm>V1*R-bar) WoW
-7.5
1.89 47.5 28.6 BoB Limits (V2*Rb) -10.0
11.5 -7.5 WoW Limits (V2*Rb) 1 2 3 4 5 A B C D E F G D,G

Names of Components to Swap

Stage II Components Search - Elimination Experiments (Expected Order of Decreasing Importance)
Run # Swap Rest BoB Rest WoW Result Component Name
4 A 40 5 Eliminate 0 A
5 B 35 0 Eliminate 0 B V1 & V2 Factors
6 C 35 5 Eliminate 0 C Stage 1
7 D 20 5 Partial 1 D Gp Size V1 V2
8 E 40 0 Eliminate 1 E 3 1.49 1.89
9 F 40 5 Eliminate 1 F 4 0.89 1.31
10 G 22 5 Partial 2 G 5 0.74 1.10
11 D,G 5 40 Full 2 H Using: 1.49 1.89
12 2 I
13 2 J
14 2 K
15 2 L
16 2 Conclusions & Rationale : Interaction. Use DoE tools for Stage III analysis.
17 2

380330654.xls CS -1 Tool Ver. 10/29/99

Philips Components Search Tool Printed On: 03/01/2018
18 2

380330654.xls CS -1 Tool Ver. 10/29/99

 Philips Components Search Stage 0, I(a), I(b) Printed On: 03/01/2018
Problem Title: Car Door Popping Noise By: Dick Shainin
Green Y: Vibration on Lever (SST scale) Date:

Stage 0 - Measurement Repeatability

Run # BoB WoW Comments Components Search - Stage 0, I(a) and I(b)
Part ID = R393f R387c 7.0 Decision
1 0 5 Initial Measurement WoW Limits
6.0 5.9
2 0 4 Second Measurement
3 0 5 Third Measurement 5.0

Stage I(a). Mark parts, reassemble same orientation 4.0 4.1

A 0 4 Door cover
B 0 5 Spring 3.0
C 0 5 Pin
2.0
D 0 4 Lever
E 1.0 BoB 0.9
F
G 0.0
H
-1.0 -0.9
Stage I(b). Reassemble, opposite orientation -2.0
A 0 5 Door cover 1 2 3 A B C D E F G H A B C D E F G H
B 5 0 Spring
C Pin Decision Limits From Stage 0 V1 & V2 Factors
D Lever 3 Stage I Group Size Stage 1
E 0 5 Median Gp Size V1 V2
F 0 1 Range 3 1.49 1.89
G 5 Dm 4 0.89 1.31
H 0.5 R-bar 5 0.74 1.10
Passed Rule #1 (End Ct = 6) Using: 1.49 1.89
Parts Orientation / Mis-orientation Table Passed Rule #2 (Dm>V1*R-bar)
Code Part Opposite orientation defined Failed Rule #3 dP/dM >= 6
A Door cover n/a 0.9 -0.9 BoB Limits (V2*Rb)
B Spring Cut end at 3:00 5.9 4.1 WoW Limits (V2*Rb)
C Pin Rotated 180 deg
D Lever n/a Conclusion and Rationale:
E Orientation of the cut end of the spring is the cause of the door popping noise.
F

380330654.xls CS - 2 Tool Ver. 8/12/99

Philips Components Search Stage 0, I(a), I(b) Printed On: 03/01/2018
G
H

380330654.xls CS - 2 Tool Ver. 8/12/99

Philips Components Search Stage III - 2 Variables Printed On: 03/01/2018

Problem Title: Low Temp Field Failure of Hourmeter Yates Table - 2 Variables Contrasts
Green Y: Operator Temperature [oC] Cell Y(Cell) 1 2 Name DY Result
By: Keki Bhote (1) 5 17.5 68.5
Date: d 12.5 51 33.5 D 16.75
g 12.5 7.5 33.5 G 16.75
Factors Analyzed and Level Settings dg 38.5 26 18.5 DG 9.25
Factor WoW BoB Variable Name Captured ?
D WoW BoB Mainframe Green-Y range FPD: 50 Repeatability = 10.5
G WoW BoB Numeral wheels Green-Y range DoE: 40 Red-X is : in P (A=P+R)
Red-X captured ? Yes Cell Scatter ? Unequal - Caution
DG(Low)
2 Variable Factorial Table Average: 12.5 Possible Spike Interaction !! Use Barrier BvC to confirm diagonal.
D(WoW) D(BoB)
0 5 5 Interaction Diagram
45
5 0 20
7 5 40
G(WoW) 5 0
35
0 5
G(WoW) 30
Cell Average: 25
G(BoB)
Medians: (1) 5 d 12.5 8.75
20 G(WoW)
Range: 7 15
20 40 35 15
5 35 40
10
37 40
G(BoB) 40 35 5
35 40 0
G(BoB) D(WoW) D(BoB)
Cell Average: 25.5
Medians: g 12.5 dg 38.5 Average Values at Factor Settings
Range: 15 5 D G DG
D(WoW) Average: D(BoB) Average: DG(High) WoW 8.75 8.75 12.5
8.75 25.5 Average: 21.75 BoB 25.5 25.5 21.75

Conclusions & Rationale :

380330654.xls CS-Fac2D Tool Ver. 8/12/99

Philips CuSum Chart Printed On: 03/01/2018
Problem Title: Weekly Sales - Product X Point # Meas Cusum
Y Variable: Sales [$000] 1 450 -48.69444
By: Don Wheeler 2 512 -35.38889
Date: 3 487 -47.08333
Weekly Sales - Product X 4 666 120.22222
200 900 5 394 15.527778
6 462 -21.16667
7 525 5.1388889
800 8 507 13.444444
0 9 585 99.75
10 341 -57.94444
700 11 295 -261.6389
12 537 -223.3333
-200 13 610 -112.0278
14 342 -268.7222
600
15 449 -318.4167
16 382 -435.1111
-400 17 429 -504.8056
500
18 275 -728.5
Sales [$000]

19 307 -920.1944
20 626 -792.8889
400 21 570 -721.5833
-600
22 510 -710.2778
23 461 -747.9722
300 24 260 -986.6667
-800 25 406 -1079.361
26 689 -889.0556
200 27 774 -613.75
28 480 -632.4444
-1000 29 634 -497.1389
100 30 711 -284.8333
31 481 -302.5278
32 539 -262.2222
-1200 0 33 663 -97.91667
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 35 36 37 34 446 -150.6111
Conclusions & Rationale : 35 718 68.694444
36 430 0

380330654.xls CuSum Tool Ver. 8/12/99

 Philips Multi-Vari Analysis (1D) Printed On: 03/01/2018

Time 8:30 AM 10:30 AM 12:30 PM 2:30 PM 4:30 PM

Within Unit Date
Unit ID 11 12 13 167 168 169 314 315 316 453 454 455
Tile Position #1 72 59 59 64 63 68 59 59 58 57 43 58
Tile Position #2 56 58 38 53 42 56 48 41 49 10 14 36
Tile Position #3 58 63 54 44 46 55 44 40 20 41 7 62
Tile Position #4 65 48 48 58 44 60 30 35 58 38 33 56
Tile Position #5 67 66 70 58 52 62 61 53 64 66 62 68
Min 56 48 38 44 42 55 30 35 20 10 7 36 0 0 0
Max 72 66 70 64 63 68 61 59 64 66 62 68 0 0 0
Range 16 18 32 20 21 13 31 24 44 56 55 32 0 0 0
Unit Average 66 57 59 58 48 61 45.5 44 61 52 47.5 62
Unit-to-Unit Range 9 13 17 14.5
Time Average 58.733 55 47.933 43.4

Multi-Vari Chart
One Measurement 80
Per Part, USL
USL
70
Multiple Parts
"Within" Units
60

Spec Limits:
50
USL=
70 40
LSL =
40 30 LSL
LSL

Variation Ranges: 20
Within Unit
56 10
Unit-to-Unit
17 0
Time-to-Time 8:30 AM 10:30 12:30 2:30 PM 4:30 PM
AM PM
15.3333333333

Problem Title : Tile Strength Conclusions & Rationale :

Green Y [Units] : Clevage Force [lbs]

380330654.xls MV-1 Tool Ver. 10/6/99

 Philips Multi-Vari Analysis (1D) Printed On: 03/01/2018

By : Keki Bhote
Date :

380330654.xls MV-1 Tool Ver. 10/6/99

Philips Multi-Vari Analysis (1D) Printed On: 03/01/2018

12:30 2:30 PM 4:30 PM 6:30 PM

PM

380330654.xls MV-1 Tool Ver. 10/6/99

Philips Multi-Vari Analysis (1D) Printed On: 03/01/2018

380330654.xls MV-1 Tool Ver. 10/6/99

 Philips Multi-Vari Analysis (4D) Printed On: 03/01/2018

Time 8:00 AM 9:00 AM 10:00 AM 11:00 AM 12:00 PM

Within Unit Date
Unit ID 21A 21B 21C 46A 46B 46C 92A 92B 92C 122A 122B 122C 147A 147B 147C
Left Diam Min  0.2503 0.2504 0.2501 0.2499 0.2498 0.25 0.2493 0.2492 0.2493 0.2505 0.2505 0.2506 0.2503 0.2499 0.25
Left Diam Max  0.2509 0.251 0.2508 0.2504 0.2503 0.2504 0.2496 0.2494 0.2497 0.251 0.2508 0.251 0.2506 0.2503 0.2505
Right Diam Max  0.2507 0.2506 0.2507 0.25 0.2499 0.25 0.2501 0.2491 0.2493 0.2505 0.2504 0.2506 0.2501 0.2498 0.2499
Right Diam Min  0.25 0.25 0.2498 0.2496 0.2497 0.2497 0.2486 0.2488 0.2488 0.25 0.2499 0.2503 0.2498 0.2495 0.2496
Range 0.0009 0.001 0.001 0.0008 0.0007 0.0007 0.0015 0.0006 0.0009 0.001 0.0009 0.0008 0.0008 0.0009 0.0009
Unit Average 0.2505 0.2505 0.2503 0.25 0.2499 0.25 0.2494 0.2491 0.2493 0.2505 0.2504 0.2506 0.2502 0.2498 0.25
Unit-to-Unit Range 0.0002 0.0001 0.0003 0.0002 0.0003
Time Average 0.2504 0.25 0.2493 0.2505 0.25

Multi-Vari Chart 0.252

Four
Measurements
Within Units USL
USL
0.251
Spec Limits:
USL=
0.251
LSL =
0.249 0.25

Variation Ranges:
Within Unit
0.00145 0.249
Unit-to-Unit
0.00035 LSL
LSL
Time-to-Time
0.0012458333
0.248
8:00 AM 9:00 AM 10:00 11:00 12:00
AM AM PM

Problem Title: Pratt-Whitney Engine Shaft Conclusions & Rationale :

Green Y: Rotor Shaft Diameter [in]
By: Dick Shainin

380330654.xls MV-4 Tool Ver. 10/6//99

 Philips Multi-Vari Analysis (4D) Printed On: 03/01/2018

Date:

380330654.xls MV-4 Tool Ver. 10/6//99

 Philips Multi-Vari Analysis (4D) Printed On: 03/01/2018

USL
USL

10:00 11:00 12:00 1:00 PM

AM AM PM

380330654.xls MV-4 Tool Ver. 10/6//99

Philips Multi-Vari Analysis (4D) Printed On: 03/01/2018

380330654.xls MV-4 Tool Ver. 10/6//99

 Philips Variables Search Tool Printed On: 03/01/2018
Problem Title: Steel Strength Date:
Green Y: Strength [Kpsi] By:
Stage I Variables Search Data
Run # Experiment Order BoB WoW
1 2=B, 1=W 65 47 Variables Search
90.0
2 4=B, 3=W 64 48 Decision
3 5=B, 6=W 69 45 Limits
4 80.0
BoB
5 54.6
Green-Y range on FPD: 45 70.0

Stage I Analysis Table 60.0

3 Stage I Group Size 39.4
65 47 Median 72.6
5 3 Range 50.0

18 Dm
4 R-bar 40.0 57.4
WoW
V1, V2 Passed Rule #1 (End Ct = 6)
1.49 Passed Rule #2 (Dm>V1*R-bar)
30.0
1.89 72.6 57.4 BoB Limits (V2*Rb) 1 2 3 4 5 A B& C D& B,D
54.6 39.4 WoW Limits (V2*Rb)
Red-X Captured in Stage I ? Yes Names of Variables to Test
(Expected Order of Decreasing Importance)
Variable Name BoB Value WoW Value
Stage II Variables Search - Elimination Experiments A Carbon
Run # Swap Rest BoB Rest WoW Result B Tungsten V1 & V2 Factors
4 A 66 48 Eliminate C Antimony Stage 1
5 B& 81 70 Partial 0 D Sulfur Gp Size V1 V2
6 C 64 49 Eliminate 1 E Fluorine 3 1.49 1.89
7 D& 86 68 Partial 1 F Boron 4 0.89 1.31
8 B,D 43 69 Full 2 G Oxygen 5 0.74 1.10
9 2 H Using: 1.49 1.89
10 2 I
11 2 J
12 2 K
13 2 L
14 2 Interaction. Use DoE tools for Stage III analysis.
15 2 Conclusions & Rationale : Optimization step recommended.

380330654.xls VS Tool Ver. 8/12/99

 Philips Variables Search Tool Printed On: 03/01/2018
16 2
2

380330654.xls VS Tool Ver. 8/12/99

 Philips Paired Comparisons Tool Printed On: 03/01/2018

Problem Title: Filter Paper Thickness Date:

Require 4 pairs = 88% Confidence Level Green Y: Flow Rate [cc/min]
By: Dorian Shainin
Results Columns
Candi- Counts Difference Pair #1 Pair #2 Pair #3
date ? Pairs <0 >0 Avg Sigma Characteristic Name [Units] Good Bad Diff Good Bad Diff Good Bad Diff
Y 2 0 2 0.3 0.14142 Y characteristic [Units] 1.2 1 0.2 1.4 1 0.4
Part ID#
Mechanical Variables
A 4 2 2 -4.45 5.72509 Dimension xxx [mm] 86.5 97.5 -11 91 98.5 -7.5 98.5 98 0.5
B 4 1 3 18.25 35.5282 Weight [gm] 5571 5512 59 5582 5608 -26 5608 5597 11
C Clue! 4 4 0 -2.36 0.80279 Stress [chart conversion] 22.8 25.85 -3.05 22.38 24.95 -2.57 22.2 23.4 -1.2
D
E
F
G
H
Attribute Observations
A Clue! 4 0 4 Broken pin yes yes yes
B 2 1 1 Capacitor bent to the left yes yes
C
D
E
F
Chemical Variables
A Clue! 4 4 0 -1.275 0.58523 pH 4.6 5.7 -1.1 4.2 6.2 -2 4.5 5.9 -1.4
B 4 1 3 5.75 7.32006 Viscosity [cP] 82 81 1 94 78 16 83 77 6
C
D
E
F
Other Types of Variables
A Clue! 4 4 0 -2 0.8165 Discoloration grade level 1 4 -3 2 4 -2 2 3 -1
B
C
D
E
F
Warning: Not all Green-Y values have been entered !!

380330654.xls PC Tool Ver. 8/12/99

Philips Paired Comparisons Tool Printed On: 03/01/2018

Conclusions & Rationale :

380330654.xls PC Tool Ver. 8/12/99

Philips Significance Test Printed On: 03/01/2018
Problem Title:
By:
Date:

A) Purpose: Prove that confinement with 20 deg cutoff is better than confinement with gradual slope

B) Required Confidence: 95% a priori (alpha-risk)

C) Green-Y: Percent jams out of 10 artificially stacked sockets

D) B's and C's Defined: B= Confinement with 20 deg leading edge

C= Confinement with gradually sloping leading edge

E) Test Type: B vs. C (6-pack)

Sample Size: # B's = 3 # C's = 3 Total Runs = 6
Tails Counted: 2
Required End Count: 6

F) Results: Raw Data as Collected Ranked Data End Counts

Production Green-Y Green-Y Sorted (after sorting on Y)
Run # Order Response Best-to-Worst Order B
1 C 60% 10% B 1
2 C 80% 10% B 2
3 B 10% 50% B 3
4 B 10% 60% C
5 C 60% 60% C
6 B 50% 80% C
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
End Count = 6 Pass

Confinement with 20 deg leading edge proved superior to Confinement with gradually
G) Conclusion:
sloping leading edge with 95% confidence

380330654.xls SigT Tool Ver. 8/12/99

 Philips Significance Test Printed On: 03/01/2018

ement with 20 deg cutoff is better than confinement with gradual slope

Count
entries
End Counts
(after sorting on Y)
C

6
5
4

with 20 deg leading edge proved superior to Confinement with gradually

sloping leading edge with 95% confidence

380330654.xls SigT Tool Ver. 8/12/99

Philips Full Factorial - 2 Variables Printed On: 03/01/2018

Problem Title: Potted Coil Yates Table - 2 Variables Contrasts

Green Y: Crack Length [0-100%] Cell Y(Cell) 1 2 Name DY Result
By: (1) 27.5 122.5 128.75
Date: a 95 6.25 66.25 A 33.125 Pale?
b 3.75 67.5 -116.25 B -58.125 Red-X
Factors Analyzed and Level Settings ab 2.5 -1.25 -68.75 AB -34.375 Pink?
Factor WoW BoB Variable Name Captured ?
A Type 1 Type 2 Epoxy material Green-Y range FPD: 95 Repeatability = 12.5
B Std Therm Hi Temp Bobbin Material Green-Y range DoE: 100 Red-X is : in P (A=P+R)
Red-X captured ? Yes Cell Scatter ? About equal
AB(Low)
2 Variable Factorial Table Average: 49.375 Use Rank-Order ANOVA to test factor significance
A(WoW) A(BoB)
20 90 Factor Strengths
30 100 100

20 100 90
B(WoW) 40 90 80

70
B(WoW)
60
Cell Average:
B(BoB)
Averages: (1) 27.5 a 95 61.25 50
B(WoW)
Range: 20 10 40
10 0 30
5 0
20
0 10
B(BoB) 0 0 10

0
B(BoB) A(WoW) A(BoB)
Cell Average: 3.125
Averages: b 3.75 ab 2.5 Average Values at Factor Settings
Range: 10 10 A B AB
A(WoW) Average: A(BoB) Average: AB(High) WoW 15.625 61.25 49.375
15.625 48.75 Average: 15 BoB 48.75 3.125 15

Conclusions & Rationale :

380330654.xls Fac-2D Tool Ver. 10/25/99

Philips Full Factorial - 2 Variables Printed On: 03/01/2018

Problem Title: HPS Short Life Yates Table - 2 Variables Contrasts

Green Y: PCA Wall Temperature Cell Y(Cell) 1 2 Name DY Result
By: Dan Gronemus (1) 1450.75 2913.75 5772.75
Date: 9-Aug-99 a 1463 2859 32.25 A 16.125 Pink
b 1419.5 12.25 -54.75 B -27.375 Red-X
Factors Analyzed and Level Settings ab 1439.5 20 7.75 AB 3.875
Factor WoW BoB Variable Name Captured ?
A Dome Base Lamp End Green-Y range FPD: 67 Repeatability = 9
B Thin Thick PCA Wall Thickness Green-Y range DoE: 50 Red-X is : in P (A=P+R)
Red-X captured ? Yes Cell Scatter ? About equal
AB(Low)
2 Variable Factorial Table Average: 1441.25 Use Rank-Order ANoVA tab to test factor significance
A(WoW) A(BoB)
1454 1466 Factor Strengths
1470
1452 1465
1451 1462 1460
B(WoW) 1446 1459
B(WoW) 1450
Cell Average: 1440
Averages: (1) 1450.75 a 1463 1456.875
B(BoB)
Range: 8 7 1430
B(WoW)
1423 1446
1420
1421 1440
1418 1440 1410
B(BoB) 1416 1432
1400
B(BoB)
Cell Average: 1429.5
1390
Averages: b 1419.5 ab 1439.5 A(WoW) Average Values at Factor Settings A(BoB)
Range: 7 14 A B AB
A(WoW) Average: A(BoB) Average: AB(High) WoW 1435.13 1456.88 1441.25
1435.125 1451.25 Average: 1445.125 BoB 1451.25 1429.5 1445.13

Conclusions & Rationale :

Wall thickness and lamp lamp end both affect wall temperature. Base end burned up, no fixture. So expect base end to be
hotter, which iit is. There is no significant interaction. Still need to show a link between short life and wall temperature.

380330654.xls Fac-2D (2) Tool Ver. 10/25/99

 Philips Rank-Order ANoVA - Fac-2D Printed On: 03/01/2018

Problem Title: HPS Short Life By: Dan Gronemus

Green Y: PCA Wall Temperature Date: 9-Aug-99
Required Confidence: 95% # Tails: 2 Required End Count = 9

Red-X Significance Pink-X Shuffle & Significance Pale Pink-X Shuffle & Significance
Sorted Sign Red-X Sorted Sign Pink-X Sorted Sign Pale-X
Y Cell B Removed Y(Pink) Cell A Removed Y(Pale) Cell n/a Removed
1416 b + 1416 1416 b - 1432.125 1431.625 a
1418 b + 1418 1418 b - 1434.125 1432 ab
1421 b + 1421 1418.625 (1) - 1434.75 1432.125 b
1423 b + 1423 1421 b - 1437.125 1434.125 b
1432 ab + 1432 1423 b - 1439.125 1434.625 a
1440 ab + 1440 1423.625 (1) - 1439.75 1434.75 (1)
1440 ab + 1440 1424.625 (1) - 1440.75 1437.125 b
1446 (1) - 1418.625 1426.625 (1) - 1442.75 1437.625 a
1446 ab + 1446 1431.625 a + 1431.625 1438.625 a
1451 (1) - 1423.625 1432 ab + 1432 1439.125 b
1452 (1) - 1424.625 1434.625 a + 1434.625 1439.75 (1)
1454 (1) - 1426.625 1437.625 a + 1437.625 1440 ab
1459 a - 1431.625 1438.625 a + 1438.625 1440 ab
1462 a - 1434.625 1440 ab + 1440 1440.75 (1)
1465 a - 1437.625 1440 ab + 1440 1442.75 (1)
1466 a - 1438.625 1446 ab + 1446 1446 ab

End Count = 14 End Count = 16 End Count = 0

95.0% Confidence? Yes 95.0% Confidence? Yes 95.0% Confidence? No
Conclusions & Rationale:
Both wall thickness and lamp end are significant at the 95% confidence level. This does not

380330654.xls ROA-2D Tool Ver. 10/26/99

Philips Rank-Order ANoVA - Fac-2D Printed On: 03/01/2018

mean these are the Red-X and Pink-X for the short life problem (link not yet made).

380330654.xls ROA-2D Tool Ver. 10/26/99

Philips Full Factorial - 3 Variables Printed On: 03/01/2018

Prob Title: Example Yates Table Contrasts

Green Y: Cell YCell 1 2 3 Name DY Result
By: (1) 36.5 53.5 113 216
Date: a 17 59.5 103 -26 A -6.50 PPPX?
b 33.5 48 -27 13 B 3.25
Factors Analyzed ab 26 55 1 -53 AB -13.25 Pink?
Factor WoW BoB Variable Name c 7.5 -19.5 6 -10 C -2.50
A 33.2 36.4 Ball weight [gm] ac 40.5 -7.5 7 28 AC 7.00 Pale?
B 22 28 Temperature [oC] bc 43.5 33 12 1 BC 0.25
C 156 180 Grease Viscosity [cP] abc 11.5 -32 -65 -77 ABC -19.25 Red-X
Captured ?
Green-Y range FPD 76 Repeatability = 3.5
3 Variable Factorial Table Green-Y range DoE 40 Red-X is : in P (A=P+R)
A(WoW) A(BoB) Red-X captured ? Yes Cell Scatter ? About equal
35 19 Use Rank-Order ANOVA to test Significance of Factors
B(WoW) 38 15
Cell Factor Strengths Diagram
C(WoW) (1) 36.5 a 17 Averages 40
3 4 Ranges
35
35 24 A
B(BoB) 32 28 30 B
Cell 25 C
b 33.5 ab 26 Averages AB
20
3 4 Ranges AC
10 42 15 BC
B(WoW) 5 39 10 AB
Cell C
5
C(BoB) c 7.5 ac 40.5 Averages
5 3 Ranges 0
45 10 WoW BoB

B(BoB) 42 13 Average Values at Factor Settings

Cell A B C AB AC BC ABC
bc 43.5 abc 11.5 Averages WoW 30.25 25.375 28.25 33.625 23.5 26.875 36.625
3 3 Ranges BoB 23.75 28.625 25.75 20.375 30.5 27.125 17.375

Conclusions & Rationale :

380330654.xls Fac-3D Tool Ver. 8/12/99

 Philips Full Factorial - 4 Variables Printed On: 03/01/2018
Factors Analyzed Problem Title:
Factor WoW BoB Variable Name Green Y: Main Effects
35
A 33.2 36.4 Ball weight [gm] By:
B 22 28 Temperature [oC] Date: 30
C 156 180 Grease Viscosity [cP] Captured? 25
A
D 3.7 4.1 Ball diameter [mm] Green-Y range FPD 32 20
B
Green-Y range DoE 25 15 C
Red-X captured ? Yes 10 D
4 Variable Factorial Table 5
B(WoW) B(BoB) Yates Table Contrasts 0
A(WoW) A(BoB) A(WoW) A(BoB) Cell YCell Name DY Result WoW BoB
15 16 35 16 (1) 16
17 15 37 15 a 15.5 A -0.56 2-Factor Strengths
35
C(WoW) (1) a b ab Cell Name b 36 B 0.19
30
16 15.5 36 15.5 Averages ab 15.5 AB -0.06 AB
25
D(WoW) 2 1 2 1 Ranges c 37.5 C 10.19 Pink? AC
20
39 34 39 14 ac 35.5 AC -0.31 AD
15
36 37 36 17 bc 37.5 BC -9.81 PPPX? BC
10
C(BoB) c ac bc abc Cell Name abc 15.5 ABC -0.06 BD
37.5 35.5 37.5 15.5 Averages d 17.5 D 1.19 5
CD
0
3 3 3 3 Ranges ad 17.5 AD 10.69 Red-X
WoW BoB
18 19 18 39 bd 17.5 BD 0.19
17 16 17 36 abd 37.5 ABD 9.94 Pale? 3 and 4 Factor Strengths
C(WoW) d ad bd abd Cell Name cd 36.5 CD -0.56 35
17.5 17.5 17.5 37.5 Averages acd 37 ACD 0.44 30
D(BoB) 1 3 1 3 Ranges bcd 17.5 BCD 0.19 25
37 36 19 37 abcd 37.5 ABCD -0.06 ABC
20
36 38 16 38 ABD
15
C(BoB) cd acd bcd abcd Cell Name Repeatability = 2.0625 ACD
10
36.5 37 17.5 37.5 Averages Red-X is : in P (A=P+R) BCD
5
1 2 3 1 Ranges Cell Scatter ? About equal ABCD
Use Rank-Order ANOVA to test 0
WoW BoB
Average Values at Factor Settings Significance of Factors
A B C D AB AC AD BC BD CD ABC ABD ACD BCD ABCD
WoW 27 26.625 21.625 26.125 26.75 26.875 21.375 31.625 26.625 27 26.75 21.75 26.5 26.625 26.75
BoB 26.4375 26.8125 31.8125 27.3125 26.6875 26.5625 32.0625 21.8125 26.8125 26.4375 26.6875 31.6875 26.9375 26.8125 26.6875

Conclusions & Rationale :

380330654.xls Se-Fac4D Tool Ver. 8/31/99

Philips B vs. C
B Condition C Condition Variable Name Problem Title:
0.025" 0.020" Metal thickness Green Y:
800 oC 1hr 750 oC 1hr Metal heat treatment By:
Date:

Table of B vs. C For One-Tailed Test

End Counts (B better than C, not just different. No overlap allowed.)
B vs. C Experiments (after sorting
Exp't Random on Y) Confidence # Random Distance between means [s] B vs. C
Number Number B or C Y B C Level Samples Beta Risk (assume s B = s C) Test
1 625 B 0.1 1 B C 50% 10% 5% Pass/Fail
2 747 B 0.12 2 99.9% 2 43 3.0 4.0 4.3
3 350 B 0.12 3 Supercritical 3 16 2.5 3.6 3.9
4 598 B 0.4 4 4 10 2.3 3.4 3.8
5 98 B 0.14 5 5 8 2.2 3.4 3.7
6 473 B 0.15 6 6 6 2.2 3.3 3.7
7 989 C 0.16
8 562 B 0.16 99% 2 13 2.3 3.4 3.8
9 718 B 0.18 Critical 3 7 2.1 3.2 3.6
10 298 C 0.25 4 5 2.0 3.1 3.5
11 146 C 0.27 5 4 2.0 3.1 3.5
12 969 B 0.32
13 69 C 0.35 9 95% 1 19 2.5 3.6 3.9
14 504 C 0.38 8 Important 2 5 1.7 3.0 3.4
15 19 C 0.43 7 3 3 1.6 2.9 3.3
16 94 C 0.46 6 4 3 1.7 3.0 3.4
17 257 C 0.52 5
18 770 C 0.58 4 90% 1 9 2.1 3.2 3.6
19 438 C 0.585 3 Moderate 2 3 1.4 2.7 3.2
20 945 C 0.59 2 3 2 1.4 2.7 3.2
21 404 C 0.6 1
22 115 Note: Use 1999 format, Se-Signif.xls, for reporting all significance tests
23 968 Conclusions & Rationale :
24 25
25 239

380330654.xls BvC Tool Ver. 12/31/98

Philips B vs. C
Total End Count = 15

380330654.xls BvC Tool Ver. 12/31/98

Philips Red-X Clue List Printed On: 03/01/2018

Clue List and Problem Resolution

Problem Title: Tektronix P220 Ringing Prints
Green Y: Print Ring - Sensory Score
By: Dorian Shainin

Red-X: Interaction in which 1262 ribbon polished the glass over the print dot heater elements and
ultimately causes the ribbon to stick to the head when printing dark lines.

Clue List Table

No. Clue Red-X Explanation
Ringing only occurs after about 1000 prints are Head is smoothed over time due to polishing effect
1 made from a new ribbon of abrasive in 1262 ribbon. Ribbon sticks to smooth
head.
2 Frequency of ringing is about 50 Hz Ribbon stretches until ribbon breaks loose
3 Always occurs at the start of a dark print bar Contact between head and ribbon is maximum
4 New printers don't ring Head is not smooth yet
5 Replacing printheads fixes problem temporarily Head is not smooth yet
On-going Reliability Test Dept. saw little ringing Alternating 1262 and 1187 ribbons prevents fully
6
using a variety of ribbon lots polishing the head.
7 Cleaning hybrid head has no effect Cleaning does not roughen the head surface
8 Surface buffing improved performance from 5 to 1 Buffing scratched the surface of a smooth head
9
10
11
12
13
14
15
16
17
18
19
20

380330654.xls Clue Tool Ver. 2/28/99

Philips Tolerance Parallelogram Printed On: 03/01/2018

Problem Title: Filter Paper Thickness Lo Lim Hi Lim

Green Y: Flow Rate [cc/min] 27 37 Max
Red-X: Filter Thickness [mils] 12.2 17.3 Point # Red-X Green-Y or 2nd
By: Dorian Shainin Vertical Scatter = 5.3086 1 10 25 2nd
Date: = 53% of Green-Y Tolerance 2 10 27
3 10 29
Parallelogram: Convert Customer Requirements to Red-X Tolerance Limits 4 10 26
5 10 28
50 6 10 27
7 10 30
8 10 28
9 25 42
45
10 25 40
11 25 39
12 25 41
40 37
37 13 25 43
14 25 40
15 25 41
16 25 43
Acceptabilit
Customers'
Flow Rate [cc/min]

35 17 13 28
Range

18 13 30
y

19 14 33
20 15 37 Max
30 21 15 31
22 15 34
23 16 31
27
27 24 16 33
25 Factory
25 17 32
Tolerance
12.2
12.2 Limits 17.3
17.3 26 17 35
27 19 35
20 28 19 37
5 10 15 20 25 30 29 20 37
Filter Thickness [mils] 30 21 39
Conclusions & Rationale : Max 25 43
Min 10 25 Max =
Range 15 18

380330654.xls Tol P Tool Ver. 8/12/99