Radc TR 85 194

Downloaded from http://www.everyspec.
com
RAMT-T315-194
MIierntm Reprt
in~c1 ber' 1985
RADC'MONE ECTROIVIUC RELIABILIT

~ NTEBOOK
'ELEOTE77
Hugqhes Aircraft Cormpwmy
* Analysis Cer~ter .OHTRI)
t7''J FIV37 P '. ý IEASE, DISTRIBUTIONl UNLIMITED
Best Available Copy

Downloaded from http://www.everyspec.com
b as been reviewed bv Lhc, ",, T ", .

.-•teas0aic to the National Technical I,•ornati'• ½..U
.. w•ll be releasable to the general public, incldi. O.
RAkDC-TR--85-194 has been r-vicwed and is provod ,
/.
JAI4LS A. COLLINS
Project Engineer
V. S. TUTdILL, Cclonel, UTSA,

Chi.ef Re lability & Compatibi-ity T-1Vf.GicrV
TD
FOR T HI C0'AIDIEA:: .
K your address has changed or if you wish t(o Ub rreovod f.rom tho, RA•Ci
-iailin• list, or if the addr(,ssee is no longer z-ploy!d by your or 1:.t:icv',.
a T) 0 j.ot.fy RADC (RBES) Griffiss AFB NY -341-5700. This will assist'. J,
"-,'i.r ',nt a current mailing list.
. ;,.n"o . ieS. this 'r pcrL unless Coui:ra c ... 2 ... ... -, ,: . ..
Best Available Copy

SEURITY CASiNiAION OF THIS PAGE
REPORT DOCUMENTATION PAGE

Ia. REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGS
UNCLASSIFIED N/A
12a.SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/ AVAILABILITY OF REPORT
N/A Approved for public release;
2b. DECLASSIFICATION I DOWNGRADING SCHEDULE distribution unl1mited.
N/A distributionunlimited.
4. PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S)
FR84-16-446 REV B RADC-TR-85-194
6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION
Hughes Aircraft Company (If applicable)
Ground Systems Group Rome Air Development Center (RBES)
6c. ADDRESS (City, State, and ZIPCode) 7b. ADDRESS (City, State, and ZIP Code)
PO Box 3310 Griffiss AFB NY 13441-5700
Fullerton CA 92634
Ba. NAME OF FUNDING/SPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
ORGANIZATION (if applicable)
Rome Air Development Center RBES F30602-82-C-0127
Bc. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS
PROGRAM PROJECT TASK WORK UNIT
Griffiss AFB NY 13441-5700 ELEMENT NO. NO. NO. ACCESSION NO.
62702F 2338 02 76
11 TITLE (include Security Classification)
RADC NONELECTRONIC RELIABILITY NOTEBOOK
12. PERSONAL AUTHOR(S) Ray E. Schafer, John E. Angus, Jack M. Finkelstein, Mal Yeras1l/iugnes
Aircraft Co.; Donald W. Fulton/Reliability Analysis Center (IITRI)
13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year, Month, Day) 11.PAGE COUNT
Interim FROM Jul 82 TO Aug 85 October 1985 588
16. SUPPLEMENTARY NOTATION
N/A -
17. COSATI CODES 18 SUBJECT TERMS (Continue on reverse if necessary and identify by blocknumber)
* FIELD GROUP SUB-GROUP DMechanicalReliability, StatisticalbNethods; BF-• Failure
14 04 'Watep .- Reliability Demonstration, Nonelectronic Reliability,
09 01 Relia lity Prediction, Reliability Specification, Components
79. ABSTRACT (Continue on reverse if necessary and identify byTiktxk number)
-----The objective-of this Notebookwas the collection, analysis, and presentation of
nonelectronic component failure data and the presentation of analytical methods that
form the state-of-the-art in nonelectronic reliability analysis. This report replaces
the former Nonelectronic Reliability Notebook (RADC-TR-75-22). i ]
This notebook is divided into six sections: Introduction,'--Applicable Statistical Methods

for Nonelectronic Reliability' Reliability Specifications; Special Afplication X-othods
for Reliability Prediction; Part F1ailure .Ciaracteristics; and ReliabilityWemonstration
Tests. 1r . J " 1I" l'' •
20 I)ISTRIBLJTION/AVAILABILITY OF ABSiRACT 21 ABSTRACT SECURITY CLASSIFICAIION

S J'JNCLASSI1'If D/UNLIMITED 1SAME AS RPT OTIC
) USERS UNCLASSIFIED
22a NA••IL ()I RESPONSIBLE INDIVIDUAL . 22b. TELEPHONE (Include Area Code) 22- OFi ICE SYM•IOL
JAMIES A. COLLINS (315) 330--2101 RADC (RBES)
DD FORM 1473, 84 MAR 83 APR edition may be used until exhausted, SECURITY CLASSIFICATION OF HIS PAGE
All other editions are obsolete. -2N C-LASS IFIED
Best Available Copy

%
S
PREFACE k
This Notebook is the result of research conducted at Hughes Aircraft

Company, Ground Systems Group, Fullerton, California, for Rome Air Develop-
"ment Center under contract number P30602-82-C-0127, covering the period
2. July 1982 through August 1985. The RADC project engineer for this effort
was Mr. James A. Collins (RADC/RBES). This research was undertaken within
the Systems Projects Section of the Systems Effectiveness Department of
"Hughesunder the direction and supervision of Dr. Ray E. Schafer until
his untimely death in September 1983. At that time, direction of the re-
o" search was taken over by Dr. John E. Angus with support and assistance
from Mr. Tom F. Pliska, Systems Effectiveness Department Assistant Manager,-.' .'
and Mr. Larry E. James, Systems Projects Section Head.
Several individuals made significant technical contributions to this *.
research. Dr. Mal Yerasi, working closely with Dr. Schafer, collected
and compiled the entire database of nonelectronic part failure data. The
statistical analyses and report generation for the Part Failure Characteristics ...
Section of the Notebook was undertaken by Dr. Angus with extensive computer - -
programming support from Mr. Shick P. Jue. Under a subcontract, Mr. Donald r
W. Fulton of RAC/IITRI (and past Rome project engineer on a previous edition
of this Notebook) wrote the section on Special Application Methods for
Reliability Prediction. Finally, the sections on Reliability Demonstration
and Specification were written by Dr. Angus with assistance from Mr. Jack
M. Finkelstein who also reviewed the entire Notebook. The Hughes report
number for this document is FR84-16-446 Rev B.
This document replaces RADC-TR-75-22, Nonelectronic Reliability Notebook.

Although RADC's interest in nonelectronic/mechanical components is limited -
to those used in electronic systems, this revised N'tebook contains failure
data and reliability methods pertaining to a variety of applications. /" '
......... . .......
I- ,-o.y o
Justiticatio~i Ave '-''a' .'-'_

4 -•..- 4
Avi a-r , -" o
Db
S--.-v...
TABLE OF CONTENTS
Page
SECTION 1.0 - INTRODUCTION . . . . . . . . . ........... . .. 1-1
SECTION 2.0 - APPLICABLE STATISTICAL METHODS FOR NONELECTRONIC

RLIABILITY ......... ....................... 2-1
2.1 Statistical Failure Models ........ ................ . 2-1

2-1
...... 2-2
2.1.1 The Hazard Rate Concept ..... ...................
the Exponential Distribution . . 2-
2.1.2 The Poisson Process and
............... ..... 2-3
2.1.2.1 The Poisson Process ...........
Distribution ...... ................ 2-3
2.1.2.2 The Exponential
...... . . ............... 2-5
2.1.3 The Weibull Distribution ..........
.................. .. 2-6
2.1.4 The Normal Distribution .........
.................
2.1.5 The Log-Normal Distribution ...........
2.1.6 Extreme Value Distribution ................
............................ .
2.1.7 Summary .................
2.2 Design of Statistical Experiments ...............
2.2.1 Introduction ........ .... .......................... 2-8
2.2.2 Determination of Stresses and Stress Levels. ........ 2-9
2.2.3 Statistical and Physical Test Design ................... 2-10
2.3 Fitting Failure Distributions ....... ................ .. 2-13
2.3.1 Introduction.............. . . . . . .......... 2-1
2.3.2 Small Sample Sizes (Kolmogorov-Smirnov) ... .......... .. 2-13
2.3.3 Large Sample Sizes (X? Test) . . . . . . ......... 2-16
2-21
2.4 Estimation Methods ............ ....................... 2-21
Distribution .................. 2-21
2.4.1 The Exponential
2.4.1.1 Analytical Point Estimation ............... ............ 2-22
2.4.1.2 Interval Estimation .....................
2.4.1.2.1 Two-Sided Confidence Limits ........... .............. 2-22
2.4.1.2.2 One-Sided Confidence Limits ........... .............. 2-23
2.4.2 The Weibull Distribution . .............. 2-25
2.4.2.1 Analytical Point Estimation ..... .............. ... 2-25
2.4.2.1.1 The Method of Least Squares ............... 2-25
2.4.2.1.2 Maximum Likelihood Method ..... .............. ... 2-27
2.4.2.2 Graphical Point Estimation ...... ............... .. 2-29
2.4.2.3 interval Estimation ......... .................... 2-33
2.4.2.3.1 Weibull Parameters ................ .................. 2-33
2.4.2.3.2 Reliability. ................. ..................... 2-34
2.4.2.3.2.1 Uncensored Samples . . . . . .. . . . . . . . . . . 2-3,4
2.4.2.3.2 Reliability ........... .................... ... 2-38
2.4.2.3.2.2 Censored Samples ................ .................. 2-38
2.4.3 The Normal Distribution ........... ..................... -4
2.4.3.1 Analytical Point Estimation ..... ............... .. 2-41
2.4.3.2 Graphical Point Estimation ...... ............... .. 2- / 3
2.4.3.3 Interval Estimation ................. ................... 2-45
2.4.3.3.1 Small Sample Sizes (a unknown) .... ............ .. 2-145
2.4.3.3.2 Large Sample Sizes (a unknown) .......... ............ 2-46
2.4.4 The Log-Normal Distribution ............... ................ 2-48
iii
TABLE OF CONTENTS (Continued)
2.4.5 The Extreme Value Distribution............. . . . . . 2-49

2.4.5.1 Analytical Point Eatimatioi . ............ 2-49
2.4.5.2 Graphical Point Estimation ........... ................. .. 2-51
2.4.5.3 Interval Estimation ....... .............. * . . . . . . . 2-53
2.4.6 Test for Increasing Failure Rate ......................... ... . 2-55
2.4.6.1 Distribution Free Test ......... ............... . . . . . 2-55
2.4.6.2 Test Based on Probability Limits and Weibull Assumptions . . 2-58 *
2.4.7 Outlier Tests ................. ........................ . 2-59

2.4.7.1 Early Failures ........................... . . . . . . . . 2-59 ' "
2.4.7.2 Late Failures ............. ...................... .. 2-61 ,
SECTION 3.0 - RELIABILITY SPECIFICATIONS
3.1 Introduction ................... .......................... .. 3-1
3.2 Reliability Specification for the Exponential Distribution. 3-2
3.3 Reliability Specification for the Weibull Distribution. ..... 3-3
3.4 Reliability Specification Without Respect to a Particular
Underlying Life Distribution ............. ................ 3-4
SECTION 4.0 - SPECIAL APPLICATION METHODS FOR RELIABILITY PREDICTION 4-].
4.1 Introduction ................... .......................... .. 4-i
4.2 Background for Reliability Prediction model Development . 4-1 •, --
4.3 Graphical Approaches to Reliability Prediction ... ......... .. 4-3
4.4 General Theory of Interference ......... ................. ... 4-4
4.5 Applications of Interference Theory to Reliability Prediction
Methodology ................. ........................... 4-10 .
4.5.1 Purpose ................ ......................... 4-10
4.5.2 Description of Method ................ ..................... 4-10
4.6 Application of Interference Theory for Normally Distributed
Strength and Normally Distributed Stress ..... .......... .. 4-13
4.6.1 Purpose of the Method ............. .................. ... 4-13
4.6.2 Description of the Method .............. .............. ... 4-13
4.6.3 Examples ................... ........................... ... 4-15 .
4.6.3.1 Shaft and Bushing Reliability ............. .............. 4-15
4.6.3.2 Lifting Eye Reliability ........... ................ ... 4-16 r
iv
-. z-1.
•
Who Wk.--
TABLE OF CONTENTS (Continued)
4.6.3.3 Diaphragm Reliability. ....................................... 4-17

4.7 Application of the Average Failure Rate Method for Grease
Lubricated Rolling Element Bearings. ......................... 4-18
4.7.1 Purpose of Method. ........................ .. .. .. .. .. 4-18
4.7.2 Description of the Method. ...................... .............. 4-18
4.7.3 Examples .. .................................................... 4-20
4.8 Reliability Prediction Method - Rolling Bearings Oil
Lubric ated.ed.. .. .. .................................... 4-21 -
4.8.1 Purpose of the Method. ............................ ............4-21Zv

4.8.2 Description of the Method. .... ................................ 4-21
4.9 Reliability Prediction Method - Spur Gear Systems. .. ............ 4-23
4.9.1 .....4-24
Purpose of the Method .. ..................................
4.9.2 Description of the Method and Example. ......................... 4-24
4.10 Reliability Prediction Method - Minimum Information ............. 4-37
4.10.1 Examples .. ................................................... 4-37
SE)CTION 5.0 - PART FAILURE CHARACTERISTICS. ............................... 5-1
5.1 Introduction. ................................ .................. 5-1
5.1.1 Use of Constant Failure Rate Analyses. ......................... 5-2
5.1.2 Use of Weibull Analyses. ....................................... 5-2
5.2 Constant Failure Rate Analysis .. ................................ 5-4
5.2.1 Index to Section 5.2 .. ....................................... -
5.3 Weibull Analyses - Project 1 .. .................................. 5-202
5.3.1 Weibull Analyses Summaries. .................................... 5-203
5.4 Weibull Analyses - Project 2 .. .................................. 5-232
3.4.1 Weibull Analyses Summaries .. .................................. 5-232
5.5 Weihull Analyses - Project 3 .. .................................. 5-270
5.5.1 Weibull Analyses Summaries .. .................................. 5-270
5.6 Part malfunction Data. .............................. ............5-284
SECTION 6.0 - RELIABILITY DE.IONSTRATTON TESTS .. .......................... 6-1
6.1 Introduction. .. .................................................6-1
. . . . *.. . .. . . . .. . . .
. .. .. . . .
6.. StatistLical Characteristics of a Reliabili.ty D~emonstration

Test. .............. ...........................
. . . . 6-1
6.1.2 Cost of Demonstration .. ...................................... 6-2
6.1.3 Sample Size Limitations. .. ................. . .. .. . .6-2
*6.1.4 Summary .. ....................................................6-3

6.2 Attributes Demonstration Tests .. I..........I..I...................6-3
6.2.1 Attributes Plans for Small Lots .. ............................ 6-3
6.2.2 Attributes Plans for Large Lots . . .. .. .. .. .. .. .. 6-6
6.2.3 Attributes Demonstration Test Plans for Large Lots
I(The Poisson Approximation Method). ......................... 6-9
*6.2.4 Attributes Sampling Using MIL-STD-105D. .... ..................6-11
6.2.5 Sequential Binomial Test Plans. ............ ..................6-13
6.3 Variables Demonstration Tests .. ................................6-15
c6.3.1 Introduction .. ............................................... 6-15
6.3.2 Time Truncated Demonstration Test Plans .. .................... 6-16*
6.3.2.1 Nonparametric Reliability Demonstration Test. ............... 6-16
*APPENDIX I- REFERENCES AND BIBLIOGRAPHY. . .....................
..........-
1
APPENDIX Il STATISTICAL TABLES............................................11-1
Iv
14
IL
LIST OF ILLUSTRATIONS
pFigurePae t
2.4.2.2 Graphical Point Estimation for the Weibull Distribution 2-32
2.4.2.3.2.1 Graphical Method for Interval Estimation of Reliability %'.)<-e%6
for the Weibull Distribution 2-35
2.4.3.2 Graphical Point Estimation for the Normal Distribution 2-44
2.4.5 Plot of Fitted Straight Line for Use With the Extreme
Value Distribution 2-54
4.4-1 Illustration of the Concept of Interference 4-7
4.4-2(A) Time Varying Strength Density 4-8
4.4-2(B) Time Varying Strength Density 4-9
4.6.2-1 Normalized Density Function for Excess Strength Overload 4-14 .-..
4.9-1 Planetary Gear Reduction .- 25
4.9-2 Planet Gear Forces 4-28-
6.3.2.1A Operating Characteristic Curve for Test Plan IA 6-20
6.3.2.2A Operating Characteristic Curve for Test Plan 2A 6-20
6.3.2.4A operating Characteristic Curve for Test Plan 4A 6-21
6.3.2.5A Operating Characteristic Curve for Test Plan SA 6-22
"6.3.2.6A Operating Characteristic Curve for Test Plan 6A 6-22

6 2 O a .-.. -
.3.2.10A
E'i Operating Characteristic Curve for Test Plan 10A 6-25 '.•.[•
"6.3.2.12A Operating Characteristic Curve for Test Plan 12A 6-25
vii
.. '.,. -,.,. o
* .. .. * *. * .. . * *** * . -.. .. -. . , -..
iLlS'l 0t ILLU61RATION.) kon~nueu)
xgure F age!,
o. J.2 1) peracin6 Characteristic Curve for Test Plan IJA 6-26
0).. b Qperatxng Characteristic Curve for Test Plan jI -2
t...b uperating Characteristic Curve for Test Plan 2B 6-27
b...b Uperatxng Characteristic Cuive for Test Plan 35 5-28

0.3.2.4b uperating Characteristic C.ar%!e for Test Plan 43 6-28
opera:in6 Cnar~cterisric C;urve 'or Test Pian 5o 6-29
0. 4.0 perz%,n6 (ýnaracteristi.o Lurve for TesL Plan t~b 6-29
7 ) Operatc.r,6 tCnaraC~erIStLI Curve tor Test Plan 78 6-30
0u)2o perat3.ng Characterist2.c c.urve for Test Plan db 6-30
.2.
Ib Ueracing --
naracteristic Cur~.e for lest Pian 9D 6-31
0.-.u uperati.n6 %unaracterisrliC Lurve for re !,inf I.Ub 6-31
O...1 .,eratin,,Cnaracteristlc Curve for Test Plan lib 6-32
D..~ Opera:in6 i-haracceristic t-urve for IeSL Plan L.ý 6- 32
1.. 0 peragin6 -..

naracc.erist3.c Curve for Test Plan D3 6-8
am=
V.
LIST OF TABLESAl
Table
1.1 Beginning Page index for Part Class Location in

Section 5 1-3
4.5.1 Parameters of Tabulated Probability Distributions as

Used in RALiC-TR-68-403 (Lipson, et.al, 1968) 4-11.
4.7.1 Quality Factors 4-19
L.4.7.2 K Grease Constant 4-19

g
4.8-1 Life Adjustment Factor of Reliability, al,

(From Source 1) 4-23 I,
-5.6.1 Part Malfunction Data S-285
6.3.2.1 Nonparametric Reliability Demonstration Test Plans 6-19
ix
1.0 INTRODUCTION
The purpose of the RADC Nonelectronic Reliability Notebook is twofold.

First, it serves as a reference doctument for the reliability characteristics
of the most commonly used, nonelectronic parts based on industry supplied
failure datal and secondly, to present the most useful reliability and
life data analysis methods applicable to nonelectronic parts. These
analysis methods are presented without regard to rigorous mathematical I-%
derivation and with an emphasis on making them accessible to reliability
practitioners possessing moderate statistical/mathematical training.
The suggested use of this Notebook is described by the table below

where each reliability task is associated with a section of this Notebook.
The use of section 5.0, Part Failure Characteristics, requires some olabora-
tion. In the majority of cases, the nonelectronic parts covered in this
Notebook are adequately described in the reliability sense by a constant
failure rate. Thus, mainly, section 5.0 will be used to look up a failure
rate for a particular device. Sometimes, however, either the part will
exhibit nonconstant failure rate, or the analyst will simply wish to
use a Weibull analysis of the part's reliability characteristics. For
thece purposes, section 5.0 also nresents Weibull analyses for selected
nonelectronic parts based on the availability of actual failure times
in the database. These Weibull analyses are based on data from three
different projects, two in the ground mobile application environment,
and one in the ground fixed environment. In many cases, the same part
type occurs in more than one Weibull analysis. In these cases, if it
*- is desired to use the Weibull analysis for modeling, the analysis in
"which the most failures were recorded should be used. In some instances,
the estimated parameters in the Weibull a" alyses of the same part type
will differ greatly. These differences are explained by differences
between project applications (even though the projects have the same
use environment) of the parts, and differences between parts of the same
name and type due to lack of data which would better characterize the
parts (i.e., two parts of the same generic part name and type can be, -..- -
nevertheless, different). As the results of these analyses indicate
the vast majority of the time, the constant failure rate tables will
be adequate. In spite of this result, the Weibull analyses have been
included for reference.
Reliability Task Nonelectronic Reliability

Notebook Sections(s)
Specification Section 3.0
Prediction Sioction 5.0, if part is represented

there;
Section 2.0, if failure data is available;
"Section 4.0, if no failure data is available.
"- Demonstration Section 6.0
,°-4
.• o.. o°
Section 2.0 of this Notebook describes the selection and application

of several failure distributions which are used for describing the life
characteristics of nonelectron~c parts, given part failure data. The remainder
of this section is devoted to methods of operating on failure data once
a failure distribution has been found to, or is assumed to, describe
nonelectronic part failure times. The general format used includes methods
of point and interval estimation for the reliability parameters of the
proven or assumed failure distribution based on empirical data.
Section 3.0 of this notebook presents guidelines and criteria for

specifying reliability for nonelectronic parts and equipments. Specifications *
appropriate for the nonparametric reliability demonstration test plans

presented in section 6.0 are included.
The next section of the notebook, section 4.0, addresses reliability

prediction, and is intended to supplement section 5.0. It gives rules
for using specific prediction models which are known to have application
to certain nonelectronic parts. This section is oriented towards strength
of alloys, grease and oil lubricated rolling bearings, and spur gear
systems. It explains and gives examples on the use of stress-strength
interference theory. A new subsection addressing reliability prediction
based on minimal vendor information and no life data is also included.
Section 5.0 is Part Failure Characteristics. This section describes

the results of the statistical analyses of failure data from more than - . "'
250 distinct nonelectronic parts collected from recent commercial and
military projects. This data was collected in-house (from operations
and maintenance reports) and from industry wide sources, all of whom
are aware of the importance of this Notebook. Tables, alphabetized by
part class/part type, are presented for easy reference to part failure
rates assuming that the part lives are exponentially distributed (as
in previous editions of this notebook, the majority of data available
included total operating time, and total number of failures only). For
parts for which the actual life times for each part under test were included
in the database, further tables are presented which describe the results
of testing the fit of the exponential and Weibull distributions. A quick I...*.*.*.-
reference index for locating the beginning page of the Tables for each
part class is presented in Table 1.1 in this introduction. The results
show that the exponentiil distribution is adequate for a large majority
of the nonelectronic parts for which its fit was tested. A small number •
of nonelectronic parts exhibited life times which were better described
by the Weibull distribution. Recommendations for approximating these
cases by the exponential distribution are presented. Part malfunction
data which was available when the part failure data was collected is
presented in Table 5.6.1. See the contract Final Technical Report describing
the study and investigation for more details on data and data analysis
dated April 1985, AD A157242).
L 9
(RADC-TR-85-66
Section 6.0 presents reliabilty demonstration test plans applicable

to nonelectronic parts. Both attributes and variables types of demonstration
plans are described. Attributes plans are used to demonstate whether
1-2
S.
::: ....
::::::-:::::......... ................ .......
. .. .... . . . .
...
. ....
_ . ..
. .......
.: :.: . . . . .. .. _. .. . . ......-.....-.......
. _..... -..-.. ....- ..... *-.-.-..-...
:.:.-- .. - . .
or not a finite population of items possesses an acceptable fraction

of items which have a particular attribute, while variables plans are
used to demonstrate whether or not a particular type of item possesses
an acceptable level of some pre-specifted reliability quantity. Whenever
the exponential distribution is judged appropriate for the life distribution
of an item of interest, test plans are documented in Mil-STD-781C and .......
Mil-RDBK-108 and are not reproduced in this section. When the exponential
distribution is not appropriate, the variables test plans of this section
are nonparametric (i.e., not dependent on the form of the underlying
life distribution) and based on one dimensional reliability specifications..=-
Because of this, these test plans are easy to design and use, and operating 1
characteristic curves can be developed and used.
TABLE 1.1
BEGINNING PAGE INDEX FOR PART CLASS LOCATION IN SECTION 5-- .4q :
PAGE NO. *PAGE NO.

PART CLASS EXPONENTIAL DIST WEIBULL DIST
ACCELEROMETER 5-10 5-206, 235

ACCUMULATOR 5-12
ACTUATOR 5-13 5-206, 235
AIR CONDITIONER 5-18
ANTENNA 5-19 5-236
AXLE 5-22 5-207, 236
AZIMUTH ENCODER 5-23 5-207, 237
BATTERY 5-24 5-208, 237
BEARING 5-25 5-208, 238
BEARING NUT 5-28
BELLOWS 5-29 5-210, 239
BELT 5-30 5-210, 240, 272
BINOCULAR 5-31 ." "
BLADE ASSEKBLY 5-32
BLOWERS & FANS 5-33 5-211, 241, 272
BOOT (DUST & MOISTURE) 5-36
BRAKE 5-36 5-211, 242
BRUSHES 5-37 5-212, 242
BURNER 5-38
BUSHINGS 5-39 5-243
CAM 5-40
CAMERA 5-40 5-273
CESIUM BEAM TUBE 5-41
CIRCUIT PROTECTION DEVICE 5-42 5-212, 243
'NOTE: A listing under this column indicates that a Weibull analysis for
one or more part types and environments under this part class is included
in Sections 5.3- 5.5.
1-3
......... ".
.. . . . . .. . . . . . . . . -.. . . . .-. . . . . . . . . :• •. . • . • . . . _ .. - _• : . - . . _:, _;.. . i. ' .- ..- ' " " .: : ' '. : '
PAGE NO. *PAGZ NO.

PART CLASS EXPONENTIAL DIST WEIBULL DIST -
CLUTCH 5-44 5-213, 244

COMPRESSOR 5-46
.OMPUTER MASS MEMORY 5-48 5-214, 245, 274
CONTROL TUBE ASSEMBLY 5-50 *N
CORD/CABLE 5-51
COUNTER 5-52
COUPLING 5-55 5-214
CRANKSHAFT 5-58 5-215, 245
CROSS HEAD 5-S9
DIAPNRAQ4S BURST 5-59 5-246
DIFFUSER 5-60
DISC ASSEMBLY 5-60
DISTILLATION UNIT 5-61
DRIVE 5-62 5-216, 247
DRIVE FOR COMPUTER
TAPES & DISCS 5-64 5-216, 246, 275 .'.
DRIVE ROD 5-67
DRUM 5-67 5-217, 248
DUCT 5-68 5-218, 248
ELECTRIC HEATERS 5-69 5-218, 249
ELECTROMECHANICAL TIMERS 5-70 5-249
ENGINES 5-71
FE-DHORN 5-71
FILTER 5-72 5-219, 250, 276
FITTINGS 5-74 5-219, 251.
F!LhSH LAMP 5-78 'l
FUSE HOLDER 5-79 5-252 \,
GAS DRYER DESICATOR 5-82 %?
GASKETS & SEALS 5-83 5-220, 253.-
GEAR 5-85 5-221, 254
GEAR BOX 5-91 5-222, 256
GEAR TRAIN 5-93
GENERATOR (ELECTRIC) 5-93
GENERATOR (GAS) 5-94
GLASS (SIGHT GAUGE) 5-94
GROMMET 5-95
GIMBALS 5-96
GYROSCOPE 5-97
HEAT EXCHANGERS 5-98 5-223, 256
HEATER 5-100
HEATER BLANKETS 5-101
*NOTE: A listing under this column indicates that a Weibull analysis for
in Sections 5.3 - 5.5.
1-4
S..- ... * .- .- - . . . . ,
"'- " " ' ' . ' i " "i " "o•" °' "" "i" -" ' "'.. S.'-°-'o"-'- S..-
°' '" ° -. '°.°-° . '.'. . . ." . . -". % -' "
IIT."
I _"
PAGE NO. PAGE NO.

PART CLASS ECPONENTIAL DIST WEIBULL DIST
HEATER, PLEX ELEMENT 5-101

HIGH SPEED PRINTER 5-102 5-223, 257, 277
HOSE 5-104 5-224, 257
HOUSING 5-107
INCINERATOR
INSTRUMENTS 5-107I
5-108 5-225, 258
SJOINT, MICROWAVE, ROTARY 5-116 5-225, 259 J
KEYBOARD 5-117 5-226, 259, 277..
"" KNOB 5-119 . *.
"- LAMP 5-120 -
LAMP HOLDER 5-121
LENS 5-122
LO3W SPEED PRINTER 5-122 5-278 I'I--
MANIFOLD 5-123PIE.7
METAL TUBING 5-124 5-260
MODULES 5-125 ." *.L
MOTOR GENERATOR SET 5-125

MOTOR, ELECTRIC 5-127 5-226, 260, 278
O-RING 5-132
PARTICLE SEPARATOR 5-132
PITCH HORN 5-133
PLOTTER 5-133
POWER CIRCUIT BREAKER 5-134 5-262
POWER SWITCH GEAR 5-136 5-227
PRECIPITATOR 5-137 .i
PRISM 5-137
PROPELLER 5-138
PROPORTIONING UNIT 5-138 .-
PULLEY 5-139 5-228, 262, 279
PUMP 5-140 5-263, 279
PURIFIER 5-145
QUILL ASSEMBLY 5-145
RADOME 5-146
REFRIGERATION PLANT 5-146
REGULATOR 5-147
RESILIENT MOUNT 5-151 5-264
RETAINING RING 5-153 5-228, 265
SEAL 5-154
SENSORS 5-155
*" SENSORS/TRANSDUCER/
"" TRANSMITTER 5-155 5-280
in Sections 5.3 - 5.5.
".1-5
+... .... .. .. . .. .. .. . .. .. .. .
::,,.... . . ... . . . .. . . . . .. . . . . . . . . .
.. . . .. . .. . .. . . . . . . . . . . . . .. . . . .. . . . . . . . . . . .
•... "..
. .... ..-.-...... .-... .-..........
...- .-....
%3y Downloaded from http://www.everyspec.com
PAGE NO. *PAGE NO.

PART CLASS EXPONENTIAL DIST WEIBULL DIST
SHAFT 5-162 5-229

SHOCK ABSORBERS 5-163 5-265
SLIP RING-BRUSH 5-165
SLIP RINGS 5-166
SOLENOIDS 5-167 5-229
SPRING 5-169
SPROCKET 5-171
STEAMBOILER 5-171
5-172
STOW PIN
SWITCH 5-172 5-230, 266, 281
SWITCHBOARD CONTROL 5-180
SYNCRO 5-181
SYNCRO ASSDEBLY 5-181
SYNCRO/RESOLVER 5-182
TACHOMETER 5-182
TANK 5-183
TELESCOPE 5-185 5-268
TERMINAL BOARDS 5-186
THERMOCOUPLE 5-187
TRACK BALL 5-188 5-269, 283
TRANSMISSION 5-189
TRUNNION ASSEMBLY 5-189
VALVE 5-190 5-231, 269
WASHER 5-196
WATER DEMINERALIZER 5-200
WINDLASS 5-201
in Sections 5.3- 5.5.
. . .........
1-6
.~~~~ .. .at
4i Downloaded from http://www.everyspec.com
2.0 APPLICABLE STATISTICAL METHODS FOR NONELECTRONIC RELIABILITY*
2.1 Statistical Failure Models.
2.1.1 The Hazard Rate Concept. The measure of an equipment's reliability

"is theinfrequency with which failures occur in time. A failure distribution
represents an attempt to describe mathematically the length of life of a
material, a structure, or a device. There are many physical Pauses that
individually or collectively may be responsible for the failure of a device at
any particular instant. The present state-of-the-art does not permit
isolation of these physical causes and mathematical account for all of them,
and, as a consequence, the choice of a failure distribution is still an art.
If one tries to rely on actual observations of time to failure to distinguish
among the various nonsymmetrical probability functions, one is still faced
with a difficulty because nonsymmetric distributions are significantly
different at the tails and actual observations are sparse, particularly at the
right-hand tail, because of limited sample size.
In view of these difficulties, it is often necessary to hypothesize the

type of failure distribution on the basis of knowledge of the physical failure
process. For example, fatigue failure of nonelectronic parts is usually
assumed to follow a Weibull probability distribution because the theoretical
development of this distribution was based on fatigue type failures.
One useful characteristic of failure distributions is the hazard rate or

failure rate.
Hazard rate, h(x) - f(x)

l-F(x)
where f(x) - density function
l-F(x) - reliability.
Hazard rate is the probability that a device already in service for time
x, will fail in the next instant of time, given no failure up to x.
Each absolutely continuous probability distribution can be characterized

by the hazard rate. Physical systems can also be classified in the same
manner. Thus the nature of the failure rates in a physical system suggests
the type of probability distribution to be assumed.
"*This entire section has been reprinted (with minor corrections) from
RADC-TR-75-22. A revised statistical methods section containing additional
advanced methods was done and is included as Appendix III of the Final Report
describing the study and investigation (RADC-TR-85-6 6 ).
2-I
16
To assist the choice of h(x) three types of failures generally have been
recognized as having a time characteristic. The first one, called the initial
failure, manifests itself shortly after time x - 0. The frequency of failures
of tnis type decreases during the initial period of operation. A good example
of this is the standard human mortality table, in which it is assumed that up
to the age of 10 years a child can die of hereditary defects, but having lived
past this age, it is almost free of such defects. The second type occurs
during the "chance failure period," in which the device exhibits a constant
failure rate, generally lower than during the initial period. The cause of
this failure is attributed to unusually severe and unpredictable environmental
conditions occurring during the operating time of the device. In the example
of human mortality tables, it is assumed that deaths between the ages of 10
.4. and J0 years are generally due to accidents. The third type is called the
wearout failure period, and is associated with the gradual depletion of a
material, or an accumulation of shocks, fatigue, and so on. In the human
mortality tables discussed before, after an age of 30 years an increasing
proportion of deaths are attributed to "old age." The three types of failures
have been classically represented by the "bathtub" curve, wherein each one of
' the three segments of the curve represents the three time periods of initial,
cnance, ana wearout.
The discussion in the previous paragraphs applies to the theory of Life

testing in general and may not apply strictly to every case where the life
'. characteristics of nonelectronic parts are involved. For example the wearout
process begins imediately in many types of nonelectronic parts.
It was stated before that given the functional form of h(x), the density
function t(x) and the cumulative distribution function F(x) could be easily
"" determined. The development of the foliowing two results is straightforward
,. and can be founa in Barlow and Proschan (1964). -.
1 -F(x) aexp ((h(X)dA 211
•.'.
anda t(x) -h(x) exp ('f h(x)d) t.I2
in the sections to follow a use will be made of this technique to develop

the commonly used failure distributions.
2.L.2 The Poisson Process and the Exponential Distribution. In

reliability studies, the exponential distribution plays a role analogous to
that of the normal Pistribution in other areas ot statistics. An acceptable
justification for the assumption of an exponential distribution to life
2-2
i .-
, ..•' .'.• ....-..
' .. -'.•
.. -..-'- .. ., ,• .. .. . .• -,.'.,...,. .' ., .,. .. .. .. . ., .. .. .. . .. .. .... .. / . I.
'
r.-
•,.• ... % •.. . . .... . ,,.

,° , ,. . .. . . .. • ,, ,.. .,. . .. ' . . . . ..
. . . . , .o o '""• ' . , • '• . - • " ' .
studies was initially discussed by Epstein in 1953. More recently a

mathematical argument has been advanced to support the plausibility of the
exponential as the failure law of complex equipment (Barlow a Proschan, 1964,
p. 16). Although many life distributions, especially those pertaining to the
nonelectronic devices, cannot be adequately aescribed by the exponential
distribution, an understanding ot the theory in the exponential case
facilitates the treatment of the more general cases. The desirability of the
exponential distribution is because of its simplicity and its inherent
association with the well developed theory of Poisson Processes (Feller,
l
1968). The applicability of the exponential distribution is limited because
of its lack of memory property; this property requires that the previous use
does not affect its future life length, and the exponential distribution is
the only continuous distribution with this property (Feller, 1968).
2.1.2.1 The Poisson Process. The exponential distribution corresponds to

a purely random failure pattern, and mathematically this means that whatever
is causing the failure occurs according to a Poisson Process with some
parameter X. 'as 2oisson probability law can be derived from rigorous
mathematical considerations, and the interested reader is referred to Feller
kl9bb). Briefly, the postulates of a Poisson Process are stated below.
Consider a system %or a unit) subjected to instantaneous changes due to

the occurrence of random events (shocks). All random events are assumed to be
of the same kind, and one is interested in their total number. Let Pm(t) be
the probanility that exactly m random events occur during a time interval of
length t.
The physical process that induces the occurrence of the random events is
characterizec by the tollowing two postulates:
i) the process is time homogeneous and the future occurrences of the
random event are independent of its past occurrences.
iO) the simultaneous occurrences of two or more events is excluded.
The above postulates lead to a system of differential equations for

Pm(t), which lead to
mt m m
2.1.2.2 The ixponential Distribution. Ihe probability density function

ot the exponential distribution can be obtained trom either the hazard rate
* concept, or by condidering tue waiting time between arrivals in a Poisson
Process. Consider the latter situation first.
Suppose that the device under consideration is subjected to an environment

in which shocks occur accoroing to the Poisson distribution, with a Poisson.
rate A. Ine device will fail only if a shock occurs and will not fail
otherwise. Let X be the lite of the device.
2-3
-. .-.
•#,.•..-.. .-.. -.. • - .. . .. . .-.-..- **.-.. . .... . ......... .- -. . ... . . .- .• . . '
........
•'-•'•
........••--•=•• ... .• ,:'..:.-.'.,.,........
..............
~~~~~~~~~~~~~~~~~. .. "..'. -."••.-.-.•"•Z.-.'.- .-
Let R(x) * Pr(X > x) = Pr [no shocks occur during (O,x)J
= e4x, by putting m 0.
Pr(X < x) 1 - e-Ax or
-,~x) = Xe-xx, x > O.
The same expression for the probability density function of X could be

obtained from the hazard rate concept, since the assumption of random shocks
with a constant Poisson rate X implies a constant failure rate h(x) " X,
for x > U.
Substituting h(x) X in equations 2.1.1 and 2.1.2 one has
F(x) I - e-x.
This section will be concluded by emphasizing tne fact that the

exponential distribution can be chosen as a failure distribution if and only
if tne assumption of a constant hazard rate can be justified. This assumption
implies that the failure of a device is not because ot its deterioration due
to wear, but is due to random shocks which occur according to the postulates
ot a PoLsson Process. This fact is of importance in nonelectronic parts
consideration, since invariably the failure of these is due to either a pure
wear or due to a combination of wear and shocks.
2.1.3 The Weibull Distribution. Recently, the Weibull distribution has
emerged to be a popular parametric family ot failure distributions. Its
applicability to a wide variety of failure situations was discussed by Weibull
(L951); it has been used to describe vacuum tube failure by Kao (1958) and a
ball bearing failure by Lieblein et. al. (1956). While the applicability of
the exponential distribution in limited because of the assumption of a
constant hazard rate, the family ot Weibull distributions can be written to
include tne increasing and the decreasing hazard rates as well. Since many
mechanical or electromechanical components have an increasing failure rate
(i.e., due to deterioration or wear), the Weibull distribution is more
palatable in describing the failure pattern of such devices.
If the hazard rate h(x) is some monotone function ot x, say if
h(x) (x-y) a
2-4
I:
=.e
- -.... . .-. -• . -.--• -.. .-. . . - .- % °- .. -. -. - % -. -. - -. -, . - -° - .° % - -. . - )-. - - -. . - - . .
I 646.:
* ~then equati~ons 2.1.1, and 2.1.2 give ...
F(x) 1- exp for x > y and
f(x)
-" (x~ exp - x > I
SU ouerwise.
•, u, and y are the shape, the scale, and the location parameters
respectively, in Section 5 of this notebook (and other sources) the scale
parameter is C1/0,.
The hazard rate for the Weibull distribution is increasing in (x-y) if

D > 1, and is independent of x if b I. When 0 a 1, the Weibull
distribution becomes the exponential distribution with location parameter y'
"and when P < 1, the Weibull distribution reduces to snat is called the
nyper exponential distribution. When 1 < I, the hazard rate decreases in
":.x-y), and such a hazard rate is useful in characterizing phenomenon such as
work haroenin6 or ocher pnenomenon associated with the improvement ot
relxaoility such ab deougging, etc.
Experience in the use of the Weibull distribution in describing the life

characteristics of nonelectronic parts leads to the conclusion that very often
the location parameter y can be assumed to be zero. This leads to tne
failure model referred to in many of the methods presented in this and
succeeding sections of this notebook as the two parameter Weibuli distribution.
2.1.4 Tne Normal Distribution. A fundamental derivation of this

distribution is not attempted here because of
d .s familiarity.
FR It X denotes the time to failure random variable ot a device which tails

* accoroing to the normal or Gaussian law, then the probability Jensity tunction
of X is given by
•()=e ___ 1/2 P -=<x < • '
V"o.. .- -,
- . . -,, . .
,- .. :..-.
Sand o are the parameters, commonly referred to as the mean and the
standard deviation respectively. The failure rate of the normal distribution

is increasind in x, and hence, this distribution can be used to char4cterize
wear. Since it is not possible to observe negative lifetimes, the use of this
distribution is limited to positive random variables.
2.L.5 The Log-Normal Distribution. The Log-Normal Distribution can

sometimes be used as a failure model when failure is due to fracture. Since
failures due to fracture occur quite commonly in practice, especially for
nonelectronic devices, a study of the Log-Normal Distribution is warranted.
are The Log-Normal
normally Distribution
distributed, implies
and hence, that betheeasily
it can logarithms
derivedof bythe
a lifetimes
simple •••
logarithmic transformation. it can also be derived more fundamentally by

consioering a physical process wherein failure is oue to fatigue cracks, and
tlhe interested reader is reterred to Kao (1965). The probability density
tunction of the Log-Normal Distribution is given by
f(x) I exp 1/2 , x > U, > U, -= < P < Q>
"mU ocierwise.
and a are the (usually) unknown parameters.
2.l.b Lxtreme Value Distribution. The Largest Extreme Value (L.A.V.)

aistribution is a two parameter right-skewed probability distribution, similar
7in appearance to the gamma, log-normal, or Weibull distribution. The L.E.V.
distribution has been successfully fitted to failure data, particularly where
failures are caused by fluctuation of a random load variable such as stress or
voltage. For example, a component with tensile strength Y is subjected to a
stress during each mission. Let X be the largest stress observed in n
missions. If n is large, the random variable, X, will have a L.E.V.
aistribution, and reliability for n missions will be given by P(X < Y),
assumin& tnat the theory of cumulative damage does not apply.
The L.L.V. distribution reliability function is
% x) I - exp (e-O(x-m) 1 > 0, -> < x < Go
and the density function is
t(x) a oeO(x-m) exp J(e1(x-M))
'1he Smallest hxtreme Value (S.E.V.) distribution is the "mirror image" of

tne L.E.V. distribution and represents the distribution of cne smallest
observation in a Large number of trials. It is unique amon6 the many lite
distributions available in that the probaDolIty distribution is skewed to Ltt:
2-b %
1.*
left. I.
One obvious application for the S.E.V. distribution is the "chain"
model, a series system of n components where X is the lowest strength among

the components. For large n, X has a S.E.V. distribution.
The reliability function is
R(x) - exp -eB(X-m) > 0, -00< x 0
and the density function
f(x) Bea(x-in) exp (eB(w-m))
2.1.7 Summary. In the preceding secticns several failure models were

proposed as possible candidates for describing the life characteristics of
nonelectronic parts. In practice, it is very difficult to identify a
particular model as the suitable one, because of the considerations given in
Section 2.1.1. However, some broad guidelines for the applicability of
"certain models were given in the other sections and these can be presented in
the table below.
Model Applicability Conditions Comments
Exponential Failure due to exactly I random Does not characterize wear

"shock
Systems comprised of many

components
Weibull Applicable under a variety of Characterizes wear or work

conditions, especially mechani- hardening
cal parts that fail due to wear
Normal Failures occur due to wearout Describes many life

processes as well as many
manufacturing processes
Log-Normal Failure due to fatigue cracks Characterizes wear
Extreme Value Failure due to extreme value of Corrosion is one example -A

some variable
The remainder of this section of the Notebook is devoted to methods of

operating on failure data once one of the previously discussed failure
distributions is found to or is assumed to describe nonelectronic part failure
times. The section is divided by failure distribution and follows a similar
pattern for each. Methods are described for calculating point estimates of
the reliability or of other parameters of the proven or assumed failure
distribution based on empirical data. Where they are available and have
utility for the users of the Notebook, graphical methods for estimating these
2-7
. . . . ..
. . ==
t :::=~~~~~~~~~~~~~~~~~.:.
.. =. . =.-.--
..".. =... =.=. . . ==.. .. . .. =. . . . .. . . . . .. .=.==========::::::::::::::::::.
::::::::::::::::::::::::
same values are presented. Tne general format used includes methods ot point p
for the reliability parameters ot each failure
ana interval estimation
distribution. humerical examples are presented ano the user ot the Notebook
is furnished with references for theoretical development and additional
examples of each of the methods presented.
2.2 Design of Statistical Experiments. in a

needed information
the
2.2.1 Introduction. generate
to state
Since the oi the art st failure data
it is
necessary
collection for nonelectronic parts does not give sufficient information for
proper analysis,
systematic manner.
In collecting failure data for estimating reliability characteristics tot

nonelectronic parts, one is frequently faced with insutficient information.
This problem arises from the manner in which the data is collected. It is
common practice to include only the operating time of the equipments or r
systems ana the number of failures observed during the operational period
covered by a failure report. Incividual part failure times are generally not
recorded.
This section of the Notebook, therefore, isdvte Odeciin n

* methodology or the principles to be followed in the event that one has the
opportunity to generate failure data for a nonelectronic part. It describes
in a logical manner the discrete guiaelines for settin6 up an experiment which *',
*• will yield an evaluation of the important factors or combinations of factors

which affect the lite characteristics of the parts of interest. ri
"The general steps for planning and conducting an experiment plus the
procedures to be followed to most efficiently analyze the results of a test
program for nonelectronic parts are outlined below. The remainder of this
subsection describes these steps in detail and references sources which cat be
used as patterns to be followed in generating and analyzing the types of
information which are cequired to allow a complete and detailed reliability
analysis of nonelectronic parts and of the operating and environmental
Stresses which affect tneir life characteristics.
When these guidelines are utilized in the generation and analysis of

reliability data for nonelectronic parts, a full and complete analysis or all
the data should be possible and no assumptions or guesses should be required
which might dilute the power or the conclusions that can be reached froin a
proper treatment of failure data.
The major steps discusseo in the succeeding pages are:
I. Uetermination of Stresses
2. Determination of Stress Levels
3. Statistical Teat Designs
4. Physical Test Designs
5. Analysis at Experimental Data
.
2-8 °. . "- -
" -. "-, . .-...
%'• •..
••"•
.
.'°,o- p. -....
. *-.-.'.,*-. .. ".. ".".
.- % °O,.
..-." *'. ".
..-,•,-..
" . -
.•. ". °•.
• V ..
.- .
... *.- .. . . ...
. . .
. . .-
.
-. -..
- ". .
...
..
.. " ..
,
."
*-
." . . . .
. . .
.0° - '- '.
~~~
-
" .o .%.
%°. ..' .. ....
•V '~~~~~~~~~
, ,' -.
2.2.2 Determination of Stresses and Stress Levels. In setting up an

experiment to generate reliability data on nonelectronic parts, those stresses
snoula be evaluated which experience and/or failure mode analysis indicates
have the greatest ettect on part life for the applications of interest. % W.N
Lven if a part is operated in a very benign environment, there are still

many factors competing in combination to cause deterioration of its life
characteristics. The rate of deterioration is a function ot the level or
concentration of a given stress. It is well-known that certain factors affect #r
a product more than others. For example, a tire's life is reduced by
vibration, radiation, corrosive agento, type of road surface, and temperature ""
to name but a few factors. It is well-known, however, that the effects of
temperature greatly exceed the effects of other environmental stresses to the
point that setting up an experiment to evaluate the life characteristics of
tois part without including this factor would render useless or distorted
results. Some factors work in combination in such a manner that their effect
together is greater than the sum of their individual effects acting
separately. For example, ozone increases the tendency of rubber to crack, but
ozone combined with high temperature creates an even greater amount of damage
in most cases. The selection of factors for an experiment then should not
only be directed toward including the most important factors affecting life
but should seek to apply them in combination since this allows the evaluation
of synergistic effects and more closely simulates actual operating conditions.
The goal of an experiment or test program to generate reliability

inrormation is therefore to determine the manner in which a part's life
characteristics vary over the envelope ot environmental and operating stresses
to which it will be subjected during normal application. If all the stresses
to which it is to be subjected during normal applications are included in the
experiment and are evaluated, the size and expense of the experiment required
w(uld be prohibitive in nearly every case. The decision as to which stresses
to include must therefore be based on experience, knowledge of failure theory,
historical data, failure mode analysis, or predicted values in order to yield
the mobt intormation for a reasonable expenditure ot monies and time.
In accition to being certain to include the most important operating or

environmental stresses for inclusion in a reliability data generation program, s.-..-.
xt is important to evaluate the effects of various levels of the stresses. it .-
is extected in most cases that very little degradation occurs in a so-called

laboratory environment. tiowever, the rates of deterioration for a given
stress usually vary in some systematic manner over the range of stresses to be
encountered in a given application. Therefore, it is desirable to investigate
what happens at several points in the environmental and operational envelope. .....
in determining the most significant stress levels to investigate there are

several points to consider. The first is to relate failure mode to stress
level. This can be done either by prior Knowledge or by running some short
time screening tests to locate the general screqc level at which failure mode
changes can be expected. If these itress levels can be located, they may even
be used in infLuencing tne operating or environmental limits to be specified
in order to effect a meaningful improvement in product life. Another point to
2-9
A..-....
.
consider in stress level selection is the concept of endurance limit - ,

associated with many nonelectronic parts. This assumes that below certain
stress levels life can be considered to be infinite or at least extremely
long. Therefore, in order to observe failures in a reasonable time these
stress levels must be avoided.
This presents the usual dilemma which can best be solved by step-stress "
testing as described by Dodson and Howard and by Prot. It consists of testing
a specimen or group of specimens for a fixed time at a fixed stress level.
The survivors are continued on test at the next increment of higher stress for
the same fixed time. This procedure continues until it is possible to select
several
be spacedstress levels that
suitable for evaluation.
can be The stressbutlevels
so closely no differences detected shouldshould
bei/;•. not
selected to sufficiently cover the spectrum of interest.
The objective of the test program is to generate a mathematical model that .-

demonstrates how life characteristics change as operational or operating -
stresses or combinations of stresses vary. In the case of nonelectronic parts *'.*..•
the dimension of time must also be included in the model since the probability
of failure increases as the part sees more service. • . .
In summary, this topic discusses the general ground rules for selecting
logical stresses and stress levels when setting up an experiment or test
program for evaluating the reliability of nonelectronic parts. The objectives
of the guidelines are the generation of more useful data for analysis than is
now generally available.
2.2.3 Statistical and Physical Test Design. The goal of the designer of del
an experiment is the generation of accurate and useful conclusions based on J"V
economical test program, efficient data collection methods and the proper
selection of statistical methods which will lead to the attainment of the
goals. 1
In generating or gathering reliability data on nonelectronic parts, it is
possible to simply put a part or group of parts on test at nominal operating
and environmental conditions and collect information on operating times and
failure times. From this, it will be possible to estimate the parameters of
the assumed failure distribution exhibited by the parts with the methods
described in Sections 2.3 and 2.4 of this Notebook.
-.
If a greater degree of diversity is desired because the part of interest

frequently may encounter stresses other than nominal in different applications--. ..
then it is probably wise to attempt to evaluate the effect of a ,'iven stress
on part life when applied at several different levels. Also, it is probably
prudent to investigate the effects of several stresses that are thought to be
major factors affecting part life again investigating each of these at several
levels. More importantly, it is reasonable to evaluate these stresses when
they are applied in combination since this comes the closest to simulating the
real life situation.
2-10
• -....
When it is desired to evaluate stresses applied in combination on a part
I
the most efficient type of statistical experiment to use is some form of
factorial design. It is true that it would be possible to evaluate the effect
of contact current on the life of a switch by holding all environmental and
operating conditions constant while varying the stress of interest, in this
case contact current over a desired range of values. This same procedure
could then be followed for actuation rate, vibration, temperature and the
infinite number of other operating and environmental stresses that could and
do affect the life of switches. The obvious result would be a series of tests
that would take a rather long time to perform. More importantly, however, is
the fact that there would be no measure of how two or more of the stresses
might act when both were at levels other than nominal. In other words, if a
synergistic effect was brought about by a given set of stresses or stress
levels, the aforementioned procedure would not be able to ovaluate it. The
solution to the problem lies in the use of some form of factorial experiment.
There are full factorial experiments in which every combination of stresses
and stress levels is tested and perhaps replicated. This type of experimental
design yields the maximum amount of information regarding main effects and the
effects of interactions of stresses. The price paid for the complete
information is paid in the cost of parts placed on test and in the test time
required to gather the requisite amount of data needed to perform the analysis.
Fractional factorial experiments can be performed in which some of the

test cells in the experiment are omitted in order to reduce test time and
expenses. The disadvantages associated with the omission of some stress
combinations is that the higher order interactions cannot be evaluated.
Therefore, the judgment as to which treatments to omit must be based on
experience or opinion. Naturally, those that are not felt to be significant
will be omitted. There are several other special types of experimental
designs which can be used such as central composites, latin squares and many '.-*-
more. For further information regarding the details of how to set up and ,....
analyze this type of experiment the reader is referred to "The Design and
Analysis of Industrial Experiments," by O.W. Davies, Hafner Publishing
Company, 1954. An example of an application of a full factorial experiment is
presented in RADC Technical Report 65-46 "Accelerated Reliability Test Methods
for Mechanical and Electromechanical Parts," July 1965. Its Defense
Documentation Center number is AD 621074. This report details all the steps
necessary in selecting the stresses, stress levels, the development of the
statistical test design, the physical test design and all the mathematical
analyses required to evaluate the effects of environmental and operating
stresses and their interactions on part life. Included also is the
methodology for fitting the failure distribution and for estimating the values
of the parameters of the Weibull distribution. Oe details presented in the
subject report will serve as a complete example *or designing and performing a
statistical experiment. The methodology used can be followed in almost a
step-by-step manner and can therefore be used in a similar manner as this
Notebook. An example of a central composite experimental design is Technical "lot
Report ECOM-01433-F "Multi-Pole Relay Evaluation Study," December 1968. The
central composite analysis yields a response surface in terms of regression
equations involving the main stress effects and certain interactions. .'%
2-11
- - *.-,.--.
With regard to the physical test design there are a few guidellhes that
should be followed. The test equipment for generating reliability data for
nonelectronic arts should simulate actual operating conditions as accurately
as iossible. In addition to this the test equipment should be economical to
build and use and should yield accurate ueasureaents of the parameters which
determine whether or not a part has failed. Finally, the physical test design
should include a simple and accurate data collection system. When a large
number of parts are tested the record-keeping problems can quickly become
quite complex. Therefore, automation of data collection is a commendable goal.
In summary, this section has presented the major types of statistical

designs that will yield accurate and useful data and refers the reader to
typical sources that will aid in the specification of useful, efficient and
economical statistical and physical test designs.
6.,
Isso•,
2-12.)-i
I,% ..
.....................................
. . . .. .. . . . . . . .. .. . . . . . . . . .
*.*.. .o
. . . . . . . . . . . . .
2-1-
2.J Fitting Failure Distributions.
2.3.1 Introduction. The two topics immediately following this one deal
with specific methods of fitting failure distributions. The placement of this
discussion in the overall table of contents is logical because when empirical
data are observed the first logical step in its analysis is to attempt to
determine the underlying distribution of failure times. While a methoa for
small sample sizes is presented as well as one for large sample sizes it is a
fact of life that must be accepted that tests based on small samples are
* simply not very powerful. Therefore, the methodology is presented here for
completeness but very likely a more logical approach is to first make an
assumption regarding the failure distribution based on engineering judgment or
on historical data or on knowledge of the failure characteristics of similar
parts. Once tne failure distribution has been assumed the test can be
performed for goodness-of-fit for that particular distribution. It the
Shypothesized distribution is shown not to fit, it is likely that the assumed
d istrioution was not the one from which the .amples were selected. If,
however, the goodness-of-fit test shows that the data could have come from the
hypothesized distribution, then it is likely that other tests for fit would
yield like results. %
In summary then, it must be realized that the tests presented in the next
two items have limitations. The only cure for these limitations is a larger
-aumber of observations. It tnis proves uneconomical or not feasible from the
standpoint of test time required to generate the desired number of failures,
then the only alternative is to use the results of small sample size analyses '.-.
with proper discretion.
"2.3.2 Small Sample Sizes (Kolmogorov-Smirnov). "
... "
I. when to Use
When failure times from a relatively small sample have been observed
and it is desired to determine the underlying distribution of failure
"times.
2. conditions for Use
• a. Usually nistorical data or engineering judgment suggests that part

failure times ot interest are from a given statistical failure
distribution. 'This test then follows the step of assuming a given
failure distribution and is useful to determine if empirical data
disproves tnis hypothesis.
b. The Kolmogorov-Smirnov test for goodness of fit is distribution free

ano can theretore be used regardless of the failure distribution that
the data is assumed to follow.
C. The discriminating ability of the statistical test is dependent on

sample size so naturally the larger the sample size the more reliable
the results. Where l.arge sample sizes are available the X2 Test
tor Goodness-of-Fit should be used. Where sample sizes are small the
Kolmogorov-Smirnov test provides some assurance.
2-13
d. Strictly speaking, this test method requires prior knowledge of the

parameters. If the parameters are estimated from the sample the exact
error risks are given in Lawless (1982). However, the values from a
Kolmogorov-Smirnov table will provide an adequate approximation in
most circumstances.
.Method Example
a. Observe and record part failure a. Given the following 20 failure

times* times in hours
92 640
130 700
233 71U
260 177U
320 830
325 1010
420 1020
430 12, ,
465 1330
518 1690
b. Assume a dxbtribution of b. Assume failure times are dis-

tailure times based on his- tributed according
parameter to the two
Weibull distribution. %
torical information or on
engineering juagmenc.
c. Estimate the parameters of the c. By the method of least squares

assumed distribution from the (see Section 2.4.2.1.1) the .
observed data. Weibull shape parameter (6)=l.5u
and the Weibull scale parameter .
(a)c2b400.
d. Ualculate the probability of d. For the Weibull distribution

failure for eacn observation the cumulative failure function
from the cumulative failure is -
function for the assumed
distribution. - ex /(- I' ) -• .oo) --
where x - observed failure time

o1.5 a Weibull shape parameter
a-2640U - Weibull scale parameter
F(x) a probability st tailure aL
or before time x.
2..*.4
,-1.-..-..
. . , - . . . . .
. . .. . . . .. .
. . . . . . .-. . .
•. .
3. Methoda bxample
d. (Continued)
For the 20 observations oi this

example, the probability of
failure at the respective x's is:
x (x_
92 0.03
130 0.05
233 0.12
260 0.14
320 0.Lt
325 0.19
420 0.26
430 0.27
465 0.30
518 0.24
640 0.43$
700 0.48
710 0.49 .
770 0.53
630 0.57
1020 O.bb
1280 0.80
1330 0.8.
1690 0.91
e. Calculate the percentile for e. For n-g0, ; gives the tollowing

each of (i) failure times by results:
the relationship r s
F(i) n and subtract tnese x IFmx) x
respective values from tnose -
of step d. above. Record the 0.03 0.05 0 0.02 O.03

absolute valua of the differ- 0.05 U.1.0 0.05 0.U.5 0.OU
ence. Also, shift i to i-1 0.12 U.15 0.10 0.04 0.O2
and compute the aifterences 0.14 0.20 0.15 0.0b 0.01
once again. 0.1 0.25 U.20 0.07 0.02
0.19 0.30 0.25 0.11 0.06
0.26 0.35 0.30 0.09 0.04
0.27 0.40 0.35 0.14 0.08
0.30 0.45 0.40 0.15 U.I0
0.34 0.50 0.45 0.1bO 0.11.
0.43 0.55 0.50 0.1. U.u7
0.48 0.60 U.55 0.1.2 U.u7
0.49 0.65 0.60 U.10 U.11.
0.53 0.70 0.U5 0.17 U.12
2-1.5
"."",".
S-. . . - : " •. •". . . . . . - . .- • • - . a",, , . . . .. ", ". .- ,•,>.N

3. Method Example
e. (Continued)
0.57 0.75 0.70 0.18 0.13

0.68 0.80 0.75 0.12 0.07
0.68 0.85 0.80 0.17 0.12
0.80 0.90 0.85 0.10 0.05
0.82 0.95 0.90 0.13 0.08
0.91 1.00 0.95 0.09 0.04
f. Compare the largest difference f. The largest difference in step

from step e. with a value at the e. was .18. From the Kolmogorov-
desired significance level in Smirnov Table for a significance
the Kolmogorov-Smirnov Tables vf .05 and for a sample of size
to test for goodness-of-fit. 20 a difference of greater than
If the tabled value is not ex- .29 must be observed before it
ceeded then it is not possible can be said that the data could
to reject the hypothesis that not have come from a Weibull
"- ~~the failure times are from the distribution with ý-l.5,".."
"assumed distribution. 0t=28400.
4. For Further Information
The example presented here illustrates how to test the hypothesis that
the failure data came from the Weibull distribution. The Kolmogorov-
Smirnov Test can also be used for other failure distributions by properly
estimating the parameters in step c. for the appropriate distribution and
by using the appropriate cumulative di4tribution function in step d.
Kolmogorov-Smirnov Tables are available on pages 321 and 322 of the
Handbook of Tables for Probability and Statistics, Edited by W.H. Beyer,
-• ~~published by the Chemical Rubber Company, Cleveland, Ohio, 1966, and ir. ,.
many texts on statistics.
2.3.3 Large Sample Sizes (X2 Test)
1. When to Use
When failure times are available from a relatively large sample and it
is desired to determine the underlying distribution of failure times.
2. Conditions for Use

a. In the statistical analysis of failure data it is common practice to
assume that failure times follow a given failure distribution family.
This assumption can be based on historical data or on engineering ..
judgment. This test for goodness of fit is used to determine if the
empirical data disproves the hypothesis of fit to the assumed
distribution.
2-16
b. The X 2 test for goodness-of-fit is asymptotically distribution

free and can therefore be used regardless of the failure distribution
that the data is assumed to tollow when samples are large.
c. This test is not directly dependent on sample size but on the number
ot intervals into which t•ne scale of failure times is divided with the
restriction that no interval should be so narrow that there are not at
least 5 theoretical failures within the interval. Therefore, the test
is only useful if a relatively large number of failures has been
observed. A.",t
'E -. -)
d. A table of x percentage points is required. .% .. 4,
3. Method Example
a. Observe and record part a. The following is the number of

failure times. cycles to failure for a group
of 50 relays on a life test: r1
1283 6820 1630b

1887 7733 17621
1b88$ 8025 17807
2357 81*5 2U747
:1137 b559 2199U
3606 8843 z3449
3752 9305 2894,
3914 9460 29254
4394. 9595 3082;4
439b 10247 3a319
4b65 11492 41554
5147 12913 42870U -
5350 12937 62890
5353 13210 6391U
5410 14633 6888b
553b 14840 73473
6499 14988
b. Assume a distribution of b. Assumc failure times are dis-

iaalure times based on his- tributed according to th, two
torical information or on parameter Weibull distribation.
engineering jucgment.
c. Estimate the parameters of the c. By the method of least squares -.
assumed distribution from the (see Section 2.4.2.1.1) the

observeo data. Weibull shape parameter t=l.21
and the Weibull scale parameter
cP127978.
:~.-f:j2..IKK.-..*::~-
$ :.. .. .*- **... .-.
. . . . . . .. . . . .. . . . .. . . . . .. . . . .. . .. . . . . .
2-17
3. Method Example
d. Divide the spectrum of failure d. Divide the relay cycles to

times into intervals of such failure into the following
intervals:
a width that the theoretical
number of failures in each in-

terval will be at least five. 0 - 4000
The width of intervals need not 4001 - 7200
be equal. 7201 - 13000
13001 - 18000 %
-
18001 - 25000
25001 - -
e. Calculate the theoretical number e. The expected number of failures

of failures for each interval, in each interval is obtained as
follows:
For the Weibull distribution

the cumulative failure func-
tion is
F(x) - 1 - exp - .'
where x - observed failure

times
8 - Weibull shape
parameter
a - Weibull scale
parameter
Then F(xn) - F(xnI) "

probability a failure time
falls within the interval.
Then for each interval the prob- .

ability of failure in that in-
terval multiplied by the sample
size - the theoretical number _
of failures for each interval.
2-18
.... %* **
* - -. <'. ,* .*: •. ••"
.-. .- "..*- . -'** * "-o .""" . "•". ." "" o ". "" .. .' " .. . " "" -"*" - "
-, .~
3. Method Example
e. (Continued)
(1) (2) (3) (4)

Upper Theoretical
"Boundary Failure 0
of F(xn) Frequency
Interval F(x) -F(x.,_1) (Col. 34x0)
"4000 O.b 0.1.6 8

7200 0.30 0.14 7
13000 0.52 u.2Y I
18000 U.bb 0.14 7
25000 O.,U U.14 7
S1.00 0.2u 10
"NOTE: The theoretical frequency

must not be less than • for
any interval.
f. Calculate the >2statistic f.

by the formula _____'_""
Upper
-. ( -F )2
Upper I L.
a 1F .
2 Boundary of
Interval F f I
4000
4ul 8 8 0
7200 7 10 1.29 .
where k - nuuoer of intervals 13000 11 12 U.11
f a observed frequency/ 18000 7 7 0
interval 25000 7 3 2.29
F - theoretical frequency/ 10 i0 0
interval
50 50 X2-3.69
g. Determine if the X2 statistic g. The degrees of freedom for this

indicates that the data could example are calculated as:
have come from the h~pothesized d.f. - (k-1) - p
distribution using X tables d.f. 0 (b-1) - 2 = 3
and (k-i) - p degrees of The value from the XZ table for
-1freedom. f abe o
freedom. J. degrees ot freedom and 0.05
level of significance is 7.b15.
where Since 3.69 does not exceed the
k = number of intervals tabled value, then the hypothe-
p a number of parameters sis that this data came from a
V estimated from data Weibull distribution cannot be
rejected at the 54 level of
significance.
. 2-19
• -'-.,-.".,",':..>.',-.2-.2...'2-,-."..".,..-,',/,:...,. ........................................................-.,,--..-..-..-.-..•......
-'" " "" " "....................................................................................,...".---.-...
• ;.•,-:-
4. For Further intormation
The example presented here illustrates how to test the hypothesis tn-v
tne ooservea failure data came from the Weibull distribution. "lae
test can also be usea for other distributions by properly estimating tne
parameters in step c. for the appropriate distribution and by using the
appropriate cumulative distribution function in step c. In step g. tne
selection of the 54 level of significance was arbitrary and will aepend on
tne researchers willingness to risk a wrong decision in re~ecting the
hypothesized distribution. There is also a risk of accepting the
distribution wrongly which for this test cannot be specified. 1here are
* several versions of 2 tables but the one used with this example is
x
from "New Tables of the Incomplete Gamma-Function Ratio and ot Percentage
Points of the (hi-Square and Beta Distributions," by H. Leon Harter,
Aerospace Research Laboratories, Office of Aerospace Research, United
States Air Force, 1964.
*1. *•,
-~o
"2-20
• ~~~~~~~~~~~~~~~~~......-.-...•..-..
:.,.......-_.-.. ,..-.,...•.
..-.......
..-..............-..-.. -.. '..-.......-.-............,.-,....-...... ........:...
N
2.4 Estimation Methods
2.4.1 The Exponential Distribution
2.4.1.1 Analytical Point Estimation
1. When to Use
To estimate the parameter e, mean-time-between-failures, in the

exponential distribution function
F(x) -1 -e -x/O
this computation may be performed. Also, the estimation of reliability is

described.
a. This estimation method may be used when units are selected at random
and placed on test, whether or not all units are allowed to fail, and
whether or not failed units are replaced.
b. No burn-in or wear-out type failures occur. Use of the exponential

distribution assumes a constant failure rate.
c. Total test time and total number of failures must be collected.
3. Method Example
F a. Sum together the test time

accumulated on each unit tested
to get the total test time.
Whether failed units are
replaced or not does not affect
the calculation, nor does it
a. Suppose 10 units are placed on
test for 80 hours and the
failed units are not replaced.
Failures occur at 20, 30, 35,
45, 70 and 75 hours. So, test
time accumulates as follows:
matter whether all units are
allowed to fail. Only compute Unit 1 20 hours
the total operating time of Unit 2 30 hours
parts on test. Unit 3 35 hours
Unit 4 45 hours
Unit 5 70 hoursN.
Unit 6 75 hours
Unit 7 80 hours
Unit 8 80 hours
Unit 9 80 hours
Unit 10 80 hours
>..>
Total Test Time - 595 hours ,..,
2-21
owl.
vi-m. . .
3.i Method Example
b. Divide total test time by total b. bince the total number o.

number of failures to get an failures is 6, divide
estimate of 6 a mean-time-
between-failures. 595-- =99.1 hours.
c. The reliability is given by c. The reliability for 30 hours is

estimated to be:
"R(x) ) -30
(1 /ROW0 *,e3/99.1
(30) 0.74
4. For Further intormation
Aoditional examples on the use of the exponential distribution are

"presentedin "Reliability Theory aria Practice," by Igor Bazovsky,
Prentice-Hall, 1961.
"2.4.1.2 Interval Estimation

I%.
2.4.1.2.1 Two-Sided Confidence Limits
I. When to Use LI
To compute upper and lower confidence limits on the exponential
distribution parameter t (mean-time-between-failures), this method is
usea.
2. uonoicions for Use
a. A confidence level, say l-a, must be specified.
b. T'otal test time and total number of failures must be collected,

wnether Qr not tailed units are replaced.
c. A Lable of X2 percentage points is required.
.3. Nethoo Example

a. It the test is failure trun- a. buppose J units ace placed on
cated, rather tnan time lite test and fail at 20, 30, 35,
truncated, then the lower 45, and 7U hours. It the 90%
"two-sided confidence limit is confidence limits are desired,
then
"2T .'- .
"2
Xkr, 1- O/2.t-" 1-ci 0.90
n 5
2-22
,-:.-.
•,.......-..-................
-.
3. tiethod Lxauxpl1e
a. (Continued) a. (continued)
where I 200 hours

T - total test time
1-i a- confiaence level •,

desired So, the lower two-sided confi-
dence, limit is
r a total number of
iailures .2
2 i2qunie r-z---- 2 x 200 1.8 5 ~
X2r, lI-W2 -/ quantle o 10,0.95
the chi-square dis-
tribution with
2r degrees of freedom.
b. The corresponding upper two- b. The upper two-sided confidence

sided confidence limit is limit is
2T 2 x 200
2 101.52
2
X2r,/2 X1 0 , 0 . 0 5 r.
4. For Further Iniormation
For a time truncated test, the lower two-sided confidence limit is

computeo with 2r + 2 degrees of rreeaom: I
2T
A2r+2, 1-W2
fne upper two-siaed contidence limit is the same as in a failure truncated

test, with 2r degrees ot freedom.
Additional examples demonstrating this method are presented in

"Reliability Theory and Practice" by Igor Bazovsky, Prentice-hall, 19bl.
2.4.1.2.2 One-Sided Confidence Limits
I. When to Use
Use this method to compute a lower one-sided confidence limit on the

exponential aistribution parameter e (mean-time-between-tailures).
2-23
. -.°. -. -
. . ..
S. - .. . : -. .
Z
W. ..onaitions for Use 2-v,
a. A confidence level, say l-(x, must be specified.
hN b. Total test time and total number of failures must be collected,

ry whether or not failed units are replaced.
c. Even if no failures have occurred, this method may be used.
d. A table of 2 percentage points is required.
3. Method Example
a. If the test ib failure trun- a. Suppose 5 units are placed on

cated, rather than time life test and fail at 20, 30, 351
truncated,* then the lower 45, and 7N hours. If the 90%
one-sided confiaence limit is lower confidence limit is
desired, then
l-0 * 0.90
X2 r, 1-a
where
T -200 hours Wv
T -total test time
e8 T/r = 4U hours
1-x a confidence level
desired So, the lower 90% one-sided
L
confidence limit is
r - total number of
failures
2 2 2 x 200 ~25.02
- 2 '•=
x 1-a quantile ot X 10 ,0 .9-
XZr, 1-x
the chi-square dis- ."
tribution with , IA,
Zr degrees of freedom.
-For a time truncated test, the lower

one-sided confidence limit is com-
puted with Zr + 2 degrees of
treedom.
2T
2
NOOTE: It no failures have occurred,

the lower one-sided contidence
iLM1C is T/(-lnc().
2-24
-. . . . . . . ...
%.. . . . . . . . . . . . . . . . . . . . . . . . .
Additional examples demonstrating this method are presented in

"keliabLlity Theory anO Practice" by Lgor Bazovsky, Prentice-Hall, 19b1.
1.4.2 the Weibuil Distribution
"2.4.2.1 Analytical Point Estimation
2.4.2.1.1 The Method of Least Squares
1. When to Use
Estimating Weibull snap* and scale parameters may be accomplished by

fitting a least squares line tu transformed Weibull data, provided that
the location parameter - is known or assumed. y is often assumed to
be zero. it not the transformation X' a X - y will reduce to the 2
parameter form. This section of the Notebook on the Weibull distribution
contains three methods of estimating the shape and scale parameters. The
method of least squares is basically a more accurate version of the graphi- ..
cal method. It takes more calculations to estimate u and ý than the
&raphical method and hence the added cost of these calculations must be
balanced against the costs associated with using a less accurate graphical
method and being subject to estimation error. While a computer helps -
reduce calculation time, the least squares method does not require as
complex a computer program as the maximum likelihood method although it
may not result in as accurate an estimate.
a. Failure times must be collected.
b. A computer ir helpful.
c. A table ot median ranks is required. It is provided in Appendix Ii,

Table 1.
3. hiethod Example
a. For n ordered Weibull failure a. Table I in the Appendix gives

times tl, t 2 , ... tn, find the median ranks for these
meoian rank values F(t) from Weibull failure times.
Taole I in the Appendix.
Failure Time Median Rank
92 Hours 0.0341
130 Hours 0.0831
233 Hours 0. 132.Z
260 Hours 0.181 'AN
320 Hours 0.2301
325 Hours 0.279J
42U Hours U.3283
430 Hours 0.3774
465 Hours U.426,'
518 Hours 0.4755
640 Hours U.5245
700 Hours 0.573b
f2-25
. . . . ..
•°~~~~........o-.....o............,.,,............... ...... '
S= • : :-- - -- -=• : -'• : .'= •- : . .. . ' • " "- -"'- *-'-* '- .'--'- '. *--*" , *.-, .
"3'3. tiethod Example
a. (Continued) , -
Failure Time Median ManK

710 Htours 0.6226
77U Hours 0.6717
830 Hours 0.7207
1010 Hours 0.769d
1020 Hours 0.8186
1280 hours 0.8678
1330 hours 0.91.69
b. or le
rniton b. 1690 hours 0.9659
a 2.36 x 10~
c. *Compute
jai ial x 01
da.lain byhadbumacppuete .
-in n
L 2
~
*UUTE:
Stp ~ .add ~ wudb egh
calcula.. .... by. ....

had.uacmue
program...o.....a...................
leas.sq..es..y..........aptd.t
Additional examples demonstrating the application of this technique to

the estimation of Weibull shape and scale parameters are given in RADC
TR65-46 "Accelerated Reliability Testing for Nonelectronic Parts", July
1965, AD b21074.
2.4.2.1.2 Maximum Liuelihood Method
1. When to Use
Use this method to estimate the shape (c) and scale (a) parameters of
the Weibull cumulative 4istribution function given by F(x)i-nexp(-xC/ •.).
The maximum likelihood method as its name implies is the most accurate
based on the observed data. However, the calculations required are the
most complex in that an iterative method must be used. If an iterative
computer program is utilized the disadvantages of the maximum likelihood
method are not so serious. This ertimation method must be used for
calculating aianr c if the interval estimation method of Section
2.4.2.3.1 is to be used for calculating confidence limits.
a. Failure times must be collected. N
b. An iterative computer program is practically essential for economical

use of this method.
*3. Ketnod Example
a. Make an initial estimate of the a. The 20 failure times given in

"shape parameter (possibly by Section 2.4.2.1.1 have been
the graphical method of Sec- plotted on Weibull probability
tion 2.4.2.2). Refer to it paper in Section 2.4.2.2. The
as C. graphical estimate of the shape
parameter is 1.5. Therefore, a.
tirst iteration of C is 1.5.
b. Using the first estimate of C,

eeosolve the maximum likelihood
equation:
b. With C a 1.5, the equation yields
F n 1.603. Since F 0 0, it is
necessary to proceed to Step c.'..:
Li
::F x i + E Ln x "'
n .(Xci ln xi)_rn
C LxC1
F'4'
2-27
A-...
.s""""
. " """ -'•" -;~''-""" - :" • :" "•-• " '"" .""" ' "" L , "" -"", .' '' , """•"""•" ,
3. Method Example
b. (Continued)
where
n - sample size
xi - ith failure time
It when the equation is solved

F w 0, then C represents the
maximum likelihood estimate ot
the shape parameter. It F 0 0
go to Step c.
c. Take the derivative (with respect c. Using C = 1.5, F' - -13.99.

to C) of the equation in Step b:
F4 i - -ln'x) [X In x] n
d. Set b - L - (F/F'). d. bh 1.5 + O.115 1.blS.
e. Set e a some small number, e. Suppose 2 place accuracy as

determine* by the accuracy desired, then set C 10-* .
desired in the answer. It
accuracy to K places is desired,
then set C - "(K4 ).
t. It I.'F' I > C, set C - b and f. IF/F' I 0115.

0
repeat Steps b. - t. U.115 > 10-3 so set L - b and
return to Step b. The iterative
process eventually gives
Z•- 1.62.
g. Now it is necessary to apply the g. From Appendix Table X1 the unbias-

unbiasing factor to the maximum in& factor for a sample size of
likelihood estimate. Appendix 2U is 0.931.
Table XI gives factors to be
muitiplies to the maximum like- Therefore
lihood estimate.
c 0.931(l.62) - 1.51.
2-28
=_
3. Method Example
h. To solve for the Weibull scale h. The estimate of the Weibull scale
parameter u parameter is
" /= 4.47 x 104
The statistical theory developing the use of this method is presented

in "Inferences on the Parameters of the Weibull Distribution," by Thoman,
Bain and Antle, Technometrics, Vol. i1, No. 3, August 1969, pp. 445-460.
:4.4.2.2 Graphical Point Estimation
1. When to Use ¾.
Estimates ot ti:e Weibull shape and scale parameters may be obtained %
graphically by using specially prepared Weibull probability paper. The
decision to use this method over those described in the two previous
topics should be based wholly on the accuracy desired. This method is the
least accurate but can be done quickly and easily.
L.%
b. Median rank tables are required. They are provided in the Appendix,
Table 1.
c. Weibull probability paper is required. See Figure 2.4.2.2.
3. Method Example
a. To plot the ith tailure time in a. As an example of plotting failure

a set ot n ordered failure times, times on Weibull probability
find the median rank plotting paper, consider a case in which
position on the left-hand ordi- 20 items are all tested to fail-
nate by consulting the table of ure; the 2V failure times, in
median ranks at n, i. To obtain ascending order, are given below
median ranKs for n greater than in the left-hand column. In the
twenty, the following formula right-hand column are the median
may be used: rank plotting positions for each
failure time, obtained from the
i - 0.3 table of median ranks for n 0 20
n + 0.4 ' in the Appendix, Table 1.
where Failure Times Median

(Hours) Ranks
i =order number of failure
92 0.0341
n number ot tailures. 130 0.0631
2-29
2.'-..-. - -. **
... , - -- - - -. -.. -.-. .-.-.-.-.-.-.-.-.-.--. . -. %' %*'* -.- * - ..- %. .. -% %.-.%....*
• .'.. , . • , .. .'. .' .' • - .. -. :.
*.-.-.. .. ". -. ' ,..". -. .- ~..'..-.' . •% .'• .. " . ' .
33. Method
Examele
a. (Uontinued)
Failure Times Median

(Hours) Ranks
233 0.1322
260 0.1812
320 0.2302
325 0.2793
420 0.328-1
430 0.3774
465 0.42b4
518 0.4755
640 u.5245
700 0.573b
710 0.6226
770
830
1010
0.6717
0.7207
0.769d
I
1020 .88186
1280 0.867b
1330 0.9169
1b90 0.9659
before plotting the data, it is

necessary to perform a transfor-
mation on the bottom scale to
accommodate the large failure
times. The _xis must be multi-
plied by 10 in oraer for the
failure data to fit on the
paper. So, the bottom scale is
properly labelea HOURb X 10-2.
b. The Weibull line is drawn b. The Weibull line, labeled t

through the plotted data by in Figure 2.4.2.2 is drawn as
using the last point plotted as described.
a reference point for a
straight-edge anu dividing the
rest of the points into two
equal groups above and below
the line.
C. To estimate 0, parallel to the c. The shape parameter 0 is esti-

Weibull line draw a line passing mated by drawing the line,
tnroughl the small circled point labeled X 2 , parallel to
on the paper. il and passing through
point A.
%4
2-JO
'4W
3. Method Exampl e
d. Horizontal projection of the d. The point where L2 intersects

point where this line intersects Z3, the principal ordinate, is
the principal ordinate to the projected horizontally to the
right-hand scale gives -0. The right-hand axis and -0 read off
I principal ordinate terminates

in 0.0 on the upper scale.
as -1.5. So, B - 1.5.
e. Sometimes in order to plot the e. To find a, the intersection of

failure data it is necessary to t1 and t3 is projected hori-
"convert the bottom scale to zo tally to the right-hand axis.
hand'e larger numbers. The The value read off the axis,
scales used on this axis are -2.9, is -In eK, and must be
oelecte
for the purpose of converted.
convenience in presenting tne
data on the graph. It the

bottom scale has been multiplied
by K, then read -1n aL at the -.
horizontal projection to the
right-hand axis of the intersec-
tion ot the Weibull line and
"the principal oroinate.
f. Find the value of oL by using •. The value of is found to .•.-
a calculator. The computed be 18.2.
-jLK is a coded value which
is dependent on the time scale
used.
g. To convert CK to an uncoded g. og is converted to an un-

state that is independent of the coded state by dividing by
time scale used on the probabil- KO. So, divide 18.2 by
ity paper, divide Ck by Kg, 10-2 giving
where o is the previously 18.2 = 18.2 x 102 x 1.5 ,-

obtained shape parameter. 16-20
- 1.82 x 10
2-31 .. ,
2-3-1
. . ...--... .
S2242-IP
100- PE00N FAILURE)
0 0
N Z '.-
_ __ __ _ __ _ _ _ _ _ _ _ _ -
a BA,
Iw
I -o
0 0
2-32
. . -------
If desired, the median ranks may be replaced by the unbiased estimate

i/(n+l), where
order number ofSi

failure
n - number ot failures.
2.4.2.3 Interval Estimation
2.4.2.3.1 Weibull Parameters
I. When to Use
Use this method to obtain confidence intervals on the shape (c) and
scale (W) parameters oi the two parameter Weibull cumulative
aistribution function given by F(x) -1 -exp(-(xC)/tI).
a. Point estimates of c and ctmust be made using the method of

Section 2.4.2.1.2.
b. Tne tables required in the calculation are provided in Appendix 11,
Tables V and VI.
* 3. Method Example
a. To compute 100(1-y) percent a. Suppose it is desired to compute

contidence limits on c, locate 90% confidence limits on the
in Table V the column labeled Weibull parameters for the exam-
2
with the value of y/ . Kead ple given in Section 2.4.2.1.2.
value at N - sam- Then N - 20, c - 1.51
otf the table
and • - 4.47 x From
ple size and call it Ll. Locate
the column labeled with the Table V, at N - 20 in the 0.05
value of I - y/2. Read oft the column the value of Ll is read
table value at N = sample size as 0.791 and at N - 20 in the
and call it L 2 . Then, tne 0.95 column tne value at L2 as
confidence interval on c is of 1.449. So the interval on S is
the form (Z/L 2 , Z/Ll), where (1.51/1.449, 1.51/0.791) - (1.04,
c was obtained by the method ot 1.91).
Section 2.4.2.1.2.
b. To compute lUU(Il-y) percent b. The 90% confidence interval on

confidence limits on a, locate a requires consulting Table VI:
in Table VI the column labeled
with the value of y/Z. Kead t1 - -0.428
oft the table value at N a sample
size and call it ti. Locate t 2 a 0.421
the column labeled with the value
2-33
. "
L La ..
3. Method Example
0. (Lontinued) 0. (Lontinuea)
at I - 1/2. Read off the table The interval on cL is given byI "-
value at N a sample size and call
it t 2 . Then, the confidence
interval on c is of the form .47 x 104 exp (-0.421),
Sexp 2 ),, exp (-t 1 ))4.47 x 10 exp (0.42b)
where a was ootained by the

method of Section 2.4.2.1.z.. (2.93 x 10-, b.b7 x 104)
4 4
4. For kurther Information
This method is from "Inferences on the Parameters of the Weibull Dis-

tribution," Thoman, bain and Antle, Technometrics, Vol. 11, No. 3, August
19b9, pp. 445-460.
2.4.2.3.2 Reliability
2.4.2.3.2.l Uncensored Samples
1. When to Use LK
Use this method to estimate 90U confidence limits on reliability for

the Weibull distribution.
a. A plot of the failure times must be prepared on Weibull probability

paper. See Section 2.4.2.2 for the methodology.
b. a table of 54 ranks and one of 954 ranks are required. These are .-.
provided in Appendix 11, Tables I and 111.
3. Method Example
a. Draw the Weibull line for the a. Refer to the example ot 2U fail-
observed daca, as described in ure times used in section 2.4.2.2.
Section 2.4.2.2. The Weibull line on Weibull proba-
bility paper is presented here
again, Figure 2.4.2.3.2.1.
b. Locate the median ranks on the b. The following median ranks, for
lett-nana axis, project tnem n - 20, have been marked oi- the
horizontally ana mark their Weibull line:
intersection with tne Weibull
line.
2-34
....-..-.--...
...
.
...... .. ....
:• -..........:•---..•:-= .:','-._•.-............ "............-....... . . .--
LN LN 1(0
- CC!
40
U. x
0 *0
2-35
3. Method Example
b. (Continued)
Median Rank Graph Notation
0.0341 A
0.0831 B
0.1322 C
0.1812 D
0.2302 E
0.2793 F
0.3283 G
0.3774 H
0.4264 I
0.4755 J
0.5245 K
0.5736 L
0.6226 M
0.6717 N
0.7207 0
0.7698 P
0.8188 Q
0.8678 R
0.9169 S
0.9659 T
c. For the jth plotting position c. The tables show a 5% rank value
of sample size - n, find the 5% of 0.0183 and a 95% rank value
and 95% ranks from Tables II of 0.2182 for n - 20, j - 2. So,
and III. Locate them on the these plots have been plotted on
left-hand axis of the Welbull the vertical line through point B. t_%
paper, then project them hori--
zontally. Their intersection
with a vertical line drawn
through the previously marked
point of intersection of the
jtf median rank and the
Weibull line gives the lower
and upper limits on reliability
for that point on the time
(lower) axis.
d. Continue plotting each pair of d. Curves are drawn through the

upper and lower limits for all plotted points.
n failures. Then connect the
points above the Weibull line
with a smooth curve. Likewise
connect the points below the
Welbull line, thus forming a ..-
band around the Weibull line,
narrowing at the center and
widening at both ends.
2-36
• " " -'. ,'.-.."

• .'. ~~~~........
•..'..'..'.-'. -. , '". "-..... -......... •.~... "" ?.L.• ' '
.........
...... . . ~~~... ........ ._. .. ,... .•.,..**..*.~.*... ..... ...- ••. .. : -- •:-• ,:--:::
J. Method Example
e. To find limits on reliability e. For example, the confidence

at a particular time t, locate limits on reliability
t on the lower axis and read 400 hours are 0.19 and at0.49.
oft upper and lower limits on
reliability from the two curves
by referring to the left-hand I
axis.
Tables 1I
and III for 5% and 95X ranks are from Electromechanical
Component keliability, May 1963, Chernowitz, et.al., RADC-TDR-63-295,
American Power Jet, Riogefield, N.J.
If confidence intervals other than the 90X for n < 20 given here are
desired, any percentile ranks may be obtained in the following manner for
sample sizes up to 50. Tables of the Incomplete Beta Function are
requirea. The notation used here is found in Tables of the Incomplete
Beta Function, K. Pearson, Cambridge University Press, 1956.
Method Example
a. For the jth failure ot n a. Suppose n 10 and it is required

failures, compute to find the 80% confidence limits
on reliability. Then the follow-
k- n - j + I ing table may be constructed:
For j > k let p- j, q k. k 4 p
For j < k let p - k, q - j. 1. 10 1 10

2 9 2 9
3 b 3
4 7 4 7
b 5 0
b5 5
7 4 4 7
LU I l -
b. To compute 100 (1-c) percent b. To Qompute 0B confidence limits,

contidence limits, locate values the O.l and 0.90 percentile
* Or lx(p,q) in Table I of the ranks are needed. So, in Table I •'. •.
reterenced tables, such that find the y - 0.10 and y - 0.90
CLI/ and l-cc/2 are approxi- percentile ranks for j-l on
"mated as closely as possible. pp. 22, 23, for p s 10 and q 1".
For j > k, the y percentile The value of x opposite 0.0946828
tank, where y = cis, 1-c0/2, in the table, the closest value
is given by the x in the left- to 0.1 given, is 0.79, and the
hand column on the same row value of x opposite 0.9043821,
as the value of Ix(p,q) - y. closest to 0.9, is 0.99. So, the
"2-37
.. . . . . . . . . . . . . . . . . . . . . . .
r7_...,
Method Lxample
b. (Continued) b. (Continued)
For j < k, the yfpercentile desirea percentile ranks for

rank is given by l-x, where x n - 1', j = 1 are 9U%: I - 0.79
again is in the left-hand column * 0.21; lO0: L - 0.99 = 0.01.
on the same row as the value of
lx(p,q) a y. The complete set of percentile
ranks follows, as computed from
Table 1.
l .2o o i.
1 0.21 0.01
2 U.40 0.06
3 0.51 U.09
4 0.55 U. l'.
5 0.65 0.2o
b 0.74 0.35
7 0.81 U.45
b U.91 .f
9 0.94 0.b
10 0.99 0.7v
2.4.2.3.2 Reliability
2.4.2.3.2.2 Censored Samples
I. Wnen to Use
This section descrioes a procedure for calculating a lower confidence

limit on reliability for parts which are known to have Weibult failure
distributions. Ihe method is applicable to both censored and uncensored
test data.
2. .onditions for Use

b. Certain tables are required in the calculation. They are provided in
Appendix II, Tables VII and VIll.
3. Metnod Example
a. For a test ot n items, r ot a. Suppose 2U parts are put on test

whicn are allowed to fail before and the test terminates at~er ,-.'e
termination of the test, order lotn failure. Then n = 2U,
the r failure times and for r = 10, and suppose the tollowin6 -
ia 1 to r, set Xi iith failure t 1 mes are ooserved:
failure time.
2-38
nlr'ý -- Am W-
3. Method Example
a. (Continued)
"X1
a 92 hours
X2 - 130 hours
SX3 - 233 hours
X4 - 260 hours
x - 320 hours
X .6325 hours
X7 - 420 hours
X8-430 hours*
- 465 hours
xa518 hours
b. To find the lower confidence b. Suppose it is desired to find

limit at time to, compute the 95Z lover confidence limit on
for i - 1 to r R(50), reliability at 50 hours.
Then in (50) w 3.91202 and . 4.
"Y-
j n Xi - in to
"" - 0.60977 '
-Y20.95551
Y3 1.53902
"4- 1.64866
',5¥5 1.85630
SY6 -1.87181
Y a 2.12823
Y8 - 2.15177
. -92.23002
Y10- 2.33796
c. Let p r/i and find values in c. Since p - 10/20, Table VII gives
Table VII for aj and bj, for the following values for ais
i = to r. and bi's. . .
a -0.04527 = -0.09198
a 2 - -0.04032 b2 = -0.09230
a3 -0.03371 b -0.09010
a4 - -0,02574 b4 - -0.08597
a 5- -0.01650 b 5 a -0.08013
2-39
k.R
3. Method Exampl
c. (Continued)
a6 -0.0059b bb a -0.07264
a7 - 0.00595 b7 - -0.06345
8" 0.01935 b8 a -O.U5246b
a9 a 0.03444. b9 a -0.03948
aio a 1.10777 blu * 0.66851
a. Compute d. Za (-0.04527)(0.b0977)
+ (-O.U4040j)(0.955l)
r + (-0.03371).(,53902)
+ (-U.Oz574)Ml.64666)
aa + (-U.0l650)(l.b5b3u)
1=01 + (-Oo.0596b) (1. 711)
+ (0,U0595)(2. 12823) "
r + (0.01935)(i.15177)
b(+. 10777)(2.33796)
bi~ =b2.5188
b 0 (-0.09198)(O.b0977)
+ (-0.09230)(U.95551)
+ (-0.0901u)(1.53902)
+ (-0.08597) (l.b48b6)
+ (-0.0680 13)(.856ýO) -
+ (-0.07264)(l.,7lbl)
+ (-O.Ob345)(2.1262)"
+ (-0.05246)(2.15077)
+ (-0,03948)(2,23002)
+ (0.66651)(2.3379b)':
0.517b
e. oUompute e. a 4.87
-a/b
P
t. To find the lower continence f. Reterring to Table VIii,

limit on reliability with con-
zidence coefficient y, use L*(Za/Zb) L*(4.b7)
lable VIII anc find the value
of L*(Za/Zb) in the rolumn U 0.939
with the desired y heading.
It is the exact lower confidence So, the lower 95A confidence
bound for K(to), reliability limit on reliability at 5U hours
at time to. is U.939.
I. I
2-40
.•:. .:....-...
....................
.,. .. . ................... ,.........................,......
-. . - - - .-.-. , ! -- • - - -. ' . : .'. . :! i " . .i • : ": : :
.'. • / " -. "o.' .. •. •"o." o °
-. o"
/. • ,. - . o-"o . . . . - o. ". .%. .• . . •- . . , ".-. - %'-, ." o "-°- o oo. o..4
3. hethoa Example
g. A point estimate oi R(to), g. Reliability at 50 hours is given

"A reliability at time to, is by
given by
e'p(-exp(-4.b7)) * 0.991.
PRto) •exp(-exp(-Za/Zb))
I
4. For Furcner Information
a. The meLhod given in from "An Exact Asymptotically

this section is
"efficienc Confidence Bound for Reliability in the Case ot the Weibull
"Distribution," Johnsand Lieberman, Technometrics, Vol. 8, No. 1,
February 1966, pp. 135-175. That paper also includes an estimation
method for obtaining a lower confidence limit on reliability for large
sample sizes.
b. A graphical technique for obtaining two-sided confidence limits on s
reliability is given in Section 2.4.2.3.2.1.
2.4.3 The Normal Distribution
1. When to Use
Use this method to obtain estimates of u and 0, the mean and

standard deviation of the Normal distribution. Xhe choice of this method
over the graphical method described in Lhe next topic is a matter ot the
accuracy desire*, witn this one yielding the most accurate estimate.
Failure times must be collected and data must be uncensored.
"3.method EaMv -l"
a. The sample mean, x, is an esti- a. buppose 20 units are tested to

mace of p and is given by failure and the following failure
times observed:
n
i "
175 hours
i-1 b95 hours
872 hours
1250 hours
where 1291 hours
1402 hours
x1 = ith tailure time 140. hours
1713 hours
n - sample size. 1741 hours
1893 hours
2025 hours
2115 hours
2-41
3. aethode -mp
a. (Continued)
2172 hours
2418 hours
2583 hours
2725 hours
2844 hours
2980 hours
3268 hours
3538 hours
Then n - 20, so
20
x - 1955.2 hours
20 K.
b. The sample standard deviation a b. The sample standard deviation ®r

is an estimate of a and is calculation gives
given by
s - 886.6 hours
n 1/2
(Xi x)
• "\i"(n
-1
where
xirn ith failure time
n - sample size
x a sample mean.
An alternate form, useful in

computer programming of this
method is
n n 2 1/2-- °m-./
(n- 1) n(n - 1)
K 4. For Further Information
Additional information regarding these estimation methods can be

obtained from any text on elementary statistics. ".
2-42
! !...
2.4.3.2 Graphical Point Estimation
1. When to Use
This method estimates V and o, the mean and standard deviation when
failure times are normally distributed. This method yields a less accurate
estimate than the method of the previous topic but requires very minimal
calculations.
a. Failure times must be collected, but may be censored.
b. Normal probability paper is required.
3. Method Example
a. On normal probabi.lty paper, plot a. The sample data used on the

the ith failure time in a sample example for Section 2.4.3.1 is '.: '.'
of n ordered failure times on repeated here, with the necessary
the lower axis vs i/(n+l) on plotting positions.
right-hand axis.
Plotting •..-
Position .-
Failure Time! L/n+l)
175 hours 0.05

695 thours 0.10
872 hours 0.14
1250 hours 0.19
1291 hours 0.24
1402 hours 0.29
1404 hours 0.33
1713 hours 0.38
1741 hours 0.43
1893 hours 0.48
2025 hours 0.52
2115 hours 0.57
2172 hours 0.62
2418 hours 0.67
2583 hours 0.71
2725 hours 0.76
2844 hours 0.81
2980 hours 0.86
3268 hours 0.90
3538 hours 0.95
b. Draw the Normal line of best b. Figure 2.4.3.2 is the plot of

fit through the plotted points this data on normal paper. The
by using the last point plotted normal line has been labeled i".
as a reference point for a
straight-edge and dividing the
rest of the points into two
2-43
RELIABILITY A. .
S. 99.99 99.8 99 9S 90 80 70 60 50 40 30 20 10 2 1 0.5 0.2 0.1 0.05 0.01
500 -
L
p: 1000 :zzz•
5O 0 --
S•20o00 -••:
2500 - I;-•
I/
u 3000 - m
3500 -I-
4000 1.
I I1111 'ii-I
0.01 0.15 0.1 0.2 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 99 99.899.9
PERCENT FAILED
Figure 2.4.3.2. Graphical Point Estimation for the Normal Distribution
2-44 r.
%L
3. Method Example
b. (Continued)
equal groups above and below

the line.
c. The mean, P, is estimated by c. The value of i is read off as

projecting the 50% probability 1950 hours.
of failure point on the right-
hand axis to the normal line
and then projecting that inter-
section point down to the lower
axis. The estimate of ., x,
is read off there.
d. The estimate of 0, s, is d. U - 2900 hours.

obtained by projecting the
intersection of the 842 proba-
bility of failure point on the
right-hand axis with the normal
line to the lower axis. Call
that point on the lower axis U.
e. Repeat Step d. with the 16% e. L - 1000 hours.

point. Call the noint L.
f. The estimate of a is f. The sample standard deviation, s,

-L is
-'U
"'2 U -L 2900-bo00, 950 hours
Additional examples of the use of this estimation method are presented

in most texts on elementary statistics.
2.4.3.3 Interval Estimation A
2.4.3.3.1 Small Sample Sizes (a unknown)
1. When to Use
Use this method to obtain confidence limits on the mean of the Normal
distribution for sample sizes < 30.
"2. Conditions for Use

a. Estimates of the mean and standard deviation must be available. See
Section 2.4.3.1 for method of computing.
b. A table of percentiles of Student's t distribution is required.
2-45
-'.
.-.. > : :-..-: ,, • .... ::, .. . .
J. Method Examele
a. To find two-sided* lO0(y) per a. Consider the 2U failure times

cent confidence limits on p, used as an example in Sec-
consult a table of the percen- tion 2.4.3.1,
tiles of the t distribution, for
the value of t 1+11/2,- where x 0 1955.2 hours
where c is Stu~ena with n-1
degrees of freedom. a - 686.6 hours
To obtain two-sided 95% confi-

dence limits on p, the value of
t is needed. From a table of .:i J•
percentiles of the t distribution
it is seen to be 2.093.
b. The two-sided confidence limits b, The confidence limits are then

are given by
886.6{
X1t(I+ y)/2,n-1 X ±n tsl~y)/2,~

//n44, l 1955.2 ± 2.093 x 8.4 -
-- = ~(1540.1, 2370.3) '-

where x a sample mean f (4. 23 3
Ssample standardIf it were desired to calculate a
lower one-sided 60% confidence
deviation limit on 1j,it would be given
n - sample size by |
*NOTE: One-sided 100(y) per cent 1955.2 - 0.861 x

4.47.
confidence limits are given
by~
by - 1784.
upper only x + t n l s/'
lower only x - t sl/Vn
4. For Further Information LX
Adoitional examples demonstrating this method are presented in

"keliability Handbook" edited by W. Grano Ireson, McGraw-hill, 1906.
1.4.1.3.2 Large Sample Sizes (a unknown)
1. When to Use
Use this method to obtain confidence limits on the mean of tne Normal
distribution for sample sizes > 30. If the standards deviation is
unknown, Stuaent's t distribution holds. however, for sample sizes of JU
or more the Normal distribution approximates the r distribution.
Z-4b
... -
2. Qonditions for Use
a. Estimates of tne mean and standard deviation must be available. See

Section 2.4.3.1 for method of computing.
b. A table of standardized normal variates is required.
3. Method Example
a. To find two-sided l0U(y) a. Suppose, after testing 50 items

confidence limits on U, con- to failure, the sample mean and
sult a table of standardized standard deviation are found to
normal deviates for the value be
of Z(1+,/2, where Z is the
stanoardized normal variate and x = 3780 hours
(U+Y)/ý is the area under the
curve to be found in the table. a = 1440 hours
by the methods of Section 3.4.3.1.

Suppose, it is desired to find
two-sided 90Z confidence on Li.
Then, from a table of standard
normal deviates, -. •
Z( 1+0 9 0 )/ 2 - 1.645.
b. The approximate two-sided conti- b. The confidence limits are then

dence limits are given by
3780 t 1.645 x -L4_0 (3445,4115)
x 1±'(+y")/2 a /n 7.07
wnere x - sample mean
s - sample standard
deviation
n sample size
c. One-sided IUU(y) percent con- c. If it were desired to calculate

fiaence limits are given by an upper one-sided 99ý4 confi-
dence limit on i, it would be
upper only x + Z s/v'n given by
1440
lower only x - sI/•n 3780 + 2.33 x T- 4255.
4. For Further Intormation
Additional examples describing the use of this method are given iti
"Reliability Handbook" edited by W. Grant ireson, McGraw-Hill, 1966.
2-47
Z.4.4
procedure
The Log-Normal Distribution.
is to first
To treat Log-Normal failure data, one
take the natural logarithm of each failure time.
The advantage of this procedure is that tables of the standardized Normal

deviates may be used in working with the logarithms of the failure times,
since they are Normally distributed.
Ihen
use the methods of Section 2.4.3, the Normal distribution, on the logarithms. U
An alternate procedure for treating Log-Normal data is to use Table IV in
Appendix 1I. Its advantage is that logarithms of the failure times do not
have to be computed.
4
Use of Table 1V
Method Example
a. For n Log-Normally distributed a. Consider the following set ot

failure times xi, compute 6 Log-Normally distributed fail-
ure times:
-xi
5.53 hours
n •5.70 hours
im 16.62 hours
7.61 hours Iv
and 8.33 hours
8.76 hours q"
n 1/2
2 Then
b.
= _______x
To find the I percentile value, b.

s )
= 7.09 hours
1.3b hours
To find the 5th percentile, that

located in Table IV the column column in the table is consulted
headed "y Percentile." Kead and the value located opposite
off the value in this column s/x - 1.36/7.09 0.19 is read
opposite the value of s/x in off as p U 0.72.
the left-hand column. Lall it p.
c. The y percentile is estimated c. The 5th percentile then is

by p1 . 0.72 x 7.09 - 5.1. This can be t.
interpreted as an estimate of the '-
reliability (R 0 . 9 5 ).
Additional Applications
a. It it is desired to find some percentile other than chose presented in
the table the following formula will be useful:
Let pi the desired percentile
Px =the mean of the observations which are assumed to be Log-

Normally distributed
:-..
2- 8 ;'-::-'-
."'".'-.-..'.'.-.'...'.....'. .'.-.'. .'. ." . . . .. e.. ,. . . .-.. ,,.,... .I.- '•, ,....-... .. ,.
2
the variance of the observations wk'ich are assumed to be
"XLog-Normally distributed
Z the standard Normal deviate associated with p.
Then
"• L1/2
W2 2 2 k..
Pi [ x]I/2 exp Z 10o
b. If the observations are in logarithmic form and a given percentile is

desired, the followin6 formula is used: 0
Let pi - desired percentile
py - the mean oi the loge xi's
ay M the standard deviation of the loge xi's
Zi - the standard Normal deviate associated with pi
Then
Pi exp(Py + zicy)
% For Further Information
Ine mathematical theory of this distribution is presented in "The

Log-Normal Distribution" by Aitchison and Brown, Cambridge Univetsity Press,
1957. The analysis methods and tables accompanying this method were developed
and prepared by J.G. Frost, Hughes Aircraft Company, Fullerton, California.
2.4.5 The Extreme Value Distribution
1. when to Use
Wnen the distribution of failure times follows the extreme value

aistribution, this method is applicable. The parameter estimates are
calculatea by the method of moments. A graphical method for parameter
estimation is presented in the next topic. The decision whether to use
the analytical or graphical method rests in the accuracy desired when
compared to the calculations to be performed in the method described here.
2-49°

Failure times must be known or assumed to follow the extreme value
distribution. Random failure times must be observed.
3. Method Example
a. Observe the failure times from a. Given the following 30 failure

a randomly selected sample. times from a process assuming a
largest extreme value (L.E.V.)
distribution:
Failure Time (Hours)
220 453
230 455
262 470
288 476
289 517
297 540
312 550
315 552
360 586
369 588 -
394 633
399 637
412 657
431 690
438 728
b. Calculate estimates of the b. Using the methods of Section

mean and standard deviation 2.4.3.1, calculate estimates
from the sample observations, of the mean
x 0 451.6
1..
and the standard deviation

s 139.0
c. Estimate the parameters 8 and c. For n- 3 0 , Gumbel gives the

m from the following formula: constants!
TN - 1.11238
N N
IN ft 53662..,
NN
2-SO
. . . . . . . . . . . . . . . . . • . .-. , o . . . -,-. " -.

•. . . . . . .. . . . . .
.-....... '-.-•. ...-.
"~~~~~~~.-.--.,,.,'."
"... -....... "."%.....-'......'"..-.."."..".""..,-.-..". •.-.....-.
- . -- • • Vr:
_-_ . . •. . ,.. ..-.
_- ,. . .. .... . ..
•~ .•, .- _ . , - ._.
.
S3. *Method Exampl e
c. (Continued) c. (Continued)
where a and Y are two con- Then

N N
stants which are functions of 1.11238
sample size only. A table of 008
these constants is available in
W
Gumbel, "Statistics of Extremes,"1941.6" .53662
University Press, 1 M 451.6 - .0i 384.5
Columbia
d. Estimate the reliability for d. The estimated reliability for a
any time t from the formula mission of 500 hours is
"Rt l-ee
"R(t) -
e(t-m) R(5
R(500) =
1-e-e'••";- .008(500-384.5
R(500) = .335
The example presente In this section is for the Largest Extreme Value .1.J
(L.E.V.) Distribution. The methods for the Smallest Extreme Value -
(S.E.V.) Distribution ar. essentially equivalent.
2.4.5.2 Graphical Point Estimation
1. When to Use
When the distribution of failure times is known or assumed to follow

the extreme value distribution, this method is applicable. This method
yields somewhat less accurate estimates than the anslytical method of
Section 2.4.5.1 but does not require the performance of calculations.
2. Conditions for Use "-

Either random or ordered censored observations may be used to estimate
the parameters.
3 Method Example
a. Collect failure timps from a a. Given 30 random failure times

randcm process and couple them from a process assuming an
with median ranks F(x) and L.E.V. distribution. Obtain
calculate median ranks for n- 3 0 using the
method of Section 2.4.2.2 and
in In -AI couple these with the failure
F )- times as fo'lows:
2-51
- . *, , -
3. Method Example
a. (Lontinued)
Median
Ranks in in -z- Failure Times

F(x) F(x) (Hours)
.023 1.33 220
.056 1.06 230
.089 .88 262
.122 .74 286
.155 .62 2b9
.188 .52 297
.220 .41 312
.253 .32 315
.28b .22 36U :2•
.319 .13 3b45
.352 .04 394 . '.
.365 -. 05 3994
.6 -.14 517
.451 -. 23 401
.484 -. 32 43b '.--•,
.549 -. 51 455 .'.-,%

•.5b2 -. bl 47U k,..-' .
.615 -. 77. 47b L 1-
.714 -1.09 550

.747 -1.23 551
.780 -1.39 58bb-,. -
.813 -1.57 5b6
.845 -1.78 6J33
.87b -2.04 637
.911 -2.38 657
.944 -2.86 69.
.977 -3.76 72o
o. Plot the failure times on the b. See Figure 2.4.5 for the graph
I of this data.
x axis ano In In - on the
F(x)
y axis and draw a Line ot best
fit through the points.
c. 6stimate parameters 0 and m C. From the graph the slope is

from the graph. -. O07b and the intercept is
Z.b5. Then
- the slope = .07b
Sm = the intercept
2-52
- -- - .- .- - .
*
B
3. Method Example
d. (Continued)
Therefore
-*2.85
-38 380
NOTE: These results agree

closely with the results
of the analytical method
Section 2.4.5.1.
The example presented in this section is for the Largest Extreme Value
(L.E.V.) Distribution. The method for the Smallest Extreme Value (S.E.V.)
Distribu, is essentially the same. Commercial graph paper is available %.•,
for plc. the extreme value distribution.
2.4.5..j interval Estimation
1. When to Use
When the distribution of failure times follows the extreme value
distribution, this method yields an approximate confidence interval or '-
lower confidence bound for reliability.

This method is applicable if it is assumed that the sample size, n, is

large and that the graphical estimates of parameters, B and m (see
Section 2.4.5.2) are maximum likelihood estimates. The actual maximum
"likelihood estimate values can be obtained only by iteration. "
"3. Method Examples
a. Obtain estimates for the param- a. From the example in Section

eters B and m using graphical 2.4.5.2
methods of Section 2.4.5.2. -
007 6 .-.
- 380
LIE b. Specify the mission time and b. For a mission time of 300 hours,
.. confidence desired, find a lower 95Z confidence bound".. .-
4., "on reliability.
I -. -,..
.. ° .
2 .... .-..
.............................
4 " ...
______ A+
b~lb
'lo
... ,., ., .
-2
-3
z o o+o•o o o .: 2.-.-'"
2-5
-1, .. < .-
S•.. - .- .- . . .- 0,
. . . . . . .. -. .. + - ,, .., - - - ,, -. -.
• ~Figure 2.4.5. Plot of"Fitted Straight Line for Use With the Extreme Value Distribution
3. Method Examples
c. Perform the following c. The calculations provide the

calculations: following results:
V a a(3-M) v - 0076(300-380)
6 Z .1.645
c a l-.05
cC2 (1.645) .
b =ff -
•2
where b -. 1- 4
x - mission time
N -sample size
ZI~ l-a quantile of the
standard normal
d istribution
d. Calculate a lower bound for v
-* Ly .423c2
(--- + Vo +
V*~
d. %# - -. 52
v+ .432+ (l- 2 )b
423
2
)"
e. Calculate a lower bound for e. The lower 95% confidence bound
reliability as follows: for reliability at 300 hours is
1R* e .52 -
R* -1- exp
R* -. 61
This example demonstrates the Largest Extreme Value Distribution. The

method for the Smallest Extreme Value Distribution is essentially the same.
2.4.6 Tests for Increasing Failure Rate
2.4.6.1 Distribution Free Test
1. When to Use
When a number of failure times have been observed and the probability
distribution is unknown this test may be used to determine if the *-
observations are from a distribution with a decreasing, constant or

increasing failure rate. The test is non-parametric and was originally
r proposed by M.G. Kendall (1938). The detailed method presented here is
from the work of Henry Mann (1945), and from Barlow and Proschan (1964).
2-55
.. ..........................
.•......-...........-.............................. .. .. -. > .. ,.-._. .? -o-...-. .
-.... :.--
:..•...-.- ... .- ? . -. ...-.. " "" -" °"'"'<" . "'. :"""
""' "
"- " '" "
"" "". . "" " " ' "" """ "°"" "'" "I" " " "i ""
2. (.onditLons for Use
Tne observed failure times must be arranged in ascending order. A

simple computer program will facilitate the calculations. lable & in
Appendix 11 is used in conjunction with the calcuLations to define the
values required to hypothesize whether the failure rate is decreasing
constant or increasing.
3. Method Examples
a. Arrange the observed failure a. Given the following 8 failure

times (xi's) in ascending times arranged in ascending ,.,
order. order:
2,6,9,12,14,16,17,18
b. Compute T. for each adjacent b. In this example, the Ti's are

pair of failure times as
follows: 11 - 2
T;4 a 6-2 a 4•
:
1L=x1
T2 x 2- x T3 9-6 3 ••.
12 X2 1
T3 W x 3- x2 1 4 a 12-9 =3". "
Tn Xn-Xn.1 1 -
T5 14-12 1
Tb a lb-14 2
T7 a 17-16 1
T 1I
b a 18-17
c. Compute Di for each T, as c. The Di's are

follows;:
D1 a 8 " 2 a 16
Dl - nT1l.-
-D2 7 * 4•- 26 07-

D2 (n-1)T 2
D3 •3- l6
e-3.1
D3 - (n-2)T 5.3. "15-"
3 4D a 5 i3
Dn • Tn
D5 -4 0 Z -8
D6 - 3 * 2 - b
D7 a 2
DS a 1" 1
2-56
2 -5 ":"
3. hethod Example
d. Generate a V statistic by d. Since nu8, there will be
comparing each set ot two ( 2-- 28 comparisons ot

Li'I. D 1 through D18 as follows:
L V11- Ii
LEI V I if Di
D > D.3 for all VI2. "12
0, since D 1Q2 16<28
ij 1,2 ... n, with i<j V a 0, since D 4z3 a lo<18

1313
V .j 0 otherwise V1 4 - 1, since DI<D4 a lo>15
In a similar manner the tollow-

in& V. 's are assigned values
of 1 since D i > D j ..
V V6# Vitt
15, 17, ,23''24*V25'
V2 6 , V2 7 , V2 8 V3 4 , V 3 5 'V3 6 ,V
37
V 3 b V4 5 1 V4 6 9V4 7 'V 4 8 0V5 6 'V 5 7 P

V58sV671V68t V78 !:
V ,V V V 8
e. Generate the test statistic e. From step d
Vn
V Vi Vni a 26
i<j
f. Now enter Table X of the f. Enter Table X of the Appendix

Appendix with Vn and the with nab and an OP.10 level of
desired level ot significance. significance. Followin g across
the row n-8, the closest value
to .lU is .089 which corresponds
to an observed Vn of 6.
Therefore, it V haa been from
U to 6 it could have been
concluded at a .089 level of .
significance that the failure
rate was decreasing.
In this example it is desired to

test for an increasing failure
rate since Vn s 2b.
Table X in the Appendix is

symmetrical; Lherefore, an .089
level of significance corresponds
2-57
"....... ... . ........
. . .
.,.........., .... :..............-

.. :. -...... ,% ,..,.;,..-..--..--...-....,...- ..... / ................. ,., ,, .,.% .. ,,,.,....,.-.,,.,
3. Method Example
f. (Continued)
to Vn = 20. Since V =
26 > 20, it can be concluded
that the failure rate is
* increasing.
For the full derivation of this method refer to "Nonparametric Tests

Against Trend," Henry Mann, Econometrica, Vol. 13, 1945, and "Mathematical
Theory of Reliability," by Barlow and Proschan, John Wiley & Sons, 1964,
pp. 232-233.
For an example of the use of the method on empirical data, refer to

RADC TR-bb-425 "Accelerated Reliability Testing for Nonelectronic Parts,"
September 19b6, AD 803 484.
The table from Mann's paper which is reproduced as Table X in Appen-

dix II covers sample sizes up to n=10. Above n-10, tables or the :..anaard-
ized normal distribution can be used because Mann proved that Vn is
n(n-1)
asymptotically normally distributed with mean 4 and a variance or
34
2n• + 3n 2 -5n
7Z
'2.4.b.2 Test Based on Probability Limits and Weibull Assumptions.
I. When to Use
Wnen a set at part failure times has been generated and it is desired
to test it the underlying distribution of failure times is exponential
(o a Welbull shape parameter - 1). in effect, this becomes a form of . o
goodness-of-fit test for deciding between the exponential and Weibull
distributions.
2. L;onditions for Use
a. '.ne sample size must be greater than 5.
b. 'The P (estimate of Weibull shape parameter) must have been

calculatea from the empirical data by the maximum likelihood method ofi' "
Section 2.4.2.1.Z.
c. Tabie IX in Appendix II is required. 5--
2-58
•...- . '.... . .- . . ....--- --... . . ... . . :.-:- - - : : -. - - - .- -- :... , _i -'

3. Method Example
a. Observe and record part failure a. Given the following L0 failures

times. times
92 325 640 1010

130 420 700 1020
233 43u 710 1280
26C 465 770 1330
320 51b 830 1690
b. Calculate the Weibull shape b. From this samae data and from the
parameter by the maximum liKe- example in Section 2.4.2.1.2 the
lihooo method of Section Weibull shape parameter is 1.62.
2.4.2.1.2. _ .•
c. Deciae on the desired signifx- C. For this example, it is desired

cance for the test for to have 10X significance level.
exponentiality. tor the hypothesis test (an
arbitrary decision).
d. tnter Table JX with the sample d. Enter the table at N-20, yL7.95,
size (N), the upper percentage yU0.05. The tablea values for a
points, and the lower percentage 90A probability interval are 0.690
points of the probability and 1.264.
interval.
e. Compare the probability limits e. Since l.b2 is not contained

with the calcu.ated value of in the interval 0.690 - 1.264,
(), the Weibull shape then the hypothesis that 8 is .• .
parameter. 1±YL < YU, then from a distribution with b I 1I
(exponential) cannot be sup-:.•,'.
it can be stated that = 1 and ( nn b
tported. The alternate hypothe-
hence, hdsis is that the failure times
exponential distribution.
were from a distribution with an . .
jncreasing hazard rate, since
> 1.2b4.
4. For Further Intormation
The method presented here is from a paper by D.K. ihoman, L.J. Bain and
C.E. Antle. The paper was titled "Inferences on the Parameters of the
Weibull Distribution" and was published i.nTechnometrics, Vol. li, No. 3,
August 19b9, pp. 445-4b0.
L.4.7 Outlier Tests
2.4.7.1 .arly Failures
1. When to Use
When an early failure is suspected not to belong to a population of

faijures which fits a particular failure distribution, this method may be
2-59
. . . . ."..................
".".-'..."--...".-
• •,.-'.;
. .. ', ,,; ;-'-'.,-.-; .. • ,-, . --.. .-.- . ./-.. - -•-. , , v ;-• . . -.. -... . . . .. .. .•. . . .
employed to determine if an early failure occurred too early to be included

in the population.
a. There must be engineering justitmcation for suspecting the early failure

to be unrelated to the main group of failures. Such justification might
be based upon a difference in failure modes for example.
b. Failure times must be collected.
c. A sample size ot at least 20 in the main group is necessary.
3. Method Example
a. For an early suspect failure a. Consider a Weibull sample ot
time xo, compute F(xo), the size 20 with parameters, esti-
cumulative distribution func- mated from the 20 failure times
tion, with par&meters estimated by the method of Section
from the main group of failure 2.4.2.1.1.
times only, evaluated at xo.
x_1.
a 2 x 101U cycles. To
examige an early failure
at 10" cycles, compute
F(10 5 ) 1
I - exp -10 ) .U0158 ,,'
b. Compute b. p (1 - .0U158)20 = o.9b9.
p - (1 - F(Xo))n
where
n - sample size of main group

of failures.
c. It p > 0.95, omit the suspect c. Since p 0.9b•9 > U.95, the
failure time. It probably does early failure time does not ..
not belong to the population belong to the population. ".'-"
determining the distribution

of failure times.
4. For Further Informacion .'-
If p < U.95, recalculate estimates for the parameters of FWx) with

Xo included in the group.
2-60
...................................
~~..-
... .. ,,... .......-....... ... ,.. . ..
,,........ ...........-.......-......... ,.,.,,-....'-.-•
Z.4. 7.2 Late Failures
1. when to Use
When a later failure is suspected not to belong to a population of

*failures which fits a particular failure distributions this method may be V
employed to determine if a late failure occurred too late to be included
in the population.
*2. Conditions for Use
There muast be engineering justification for suspecting the late failure .V

*to 5e unrelated to the main group of failures. Such justification might
be based upon a difference in failure modes for example. A sample size of
at least 2u in the main group is needed.
3. Method 6xample .'
a. For a late suspect failure time a. Consider a Weibull sample of size

x., compute F'kxo), the 20 with parameters, estimated
cumulative distribution func- from the 20 failure times by the
tion, with parameters estimated method of Section 2.4.2.1.1,
from the wain group of failure
times only, evaluated at x -1.
OL 2 x 10 1U cycles
To examine a late failure at
2 x 10~ cycles, compute
I 7x1. 5
7 1-2x10
F(2x10 ) -e-xpk - 0.98b
I
I
2x.10 1 5
b. Compute b. q -(0.958)20
q =(F(xo))n 0.794
jc. If q > U.95, omit the suspect c. q =0.794 U.95

failure time. It does not
belong to trne population
determining the distribution
of failure times.
a. It q < U-95, recalculate d. Since q -0.794 < U-95, the fail-

estisuates for the parameters of ure time 2 x l0U
7cycles does
FWx with x.o included in the belong to the population. bo, it
group. is neces~ary to repeat the estima-
tion of the weibuli distribution '~ -
parameters 0 and OLfrom 5ec-%
tion 2.4.2.1.1 with 7n a 21 anc. the
failure time 2 x 10~ cycles ~ ~
I~ 1
included.
2-61 (2-62 BLANK) .
r M'
* .3.0 RELIABILITY SPECIFICATIONS )
3.1 Introduction. The purpose ot a reliability specification is to fix

the reliability component of mission effectiveness by quantifying required
* reliability characteristics for the part, equipment, or system. In order to %
achieve this purpose the reliability specification must be stated in complete
* and unambiguous terms. A reliability demonstration test will verify whether
or not the requirement is satisfied.
Ambiguous requirements may result in an item that passes the demonstra-

tion but is not effective with respect to the requirements, and vice versa.
For example, to state that a part must have a life of 1000 hours is ambig-
uous. The intention of such a specification could be: to require all such
* parts to survive 1000 hours; to require all such parts to have a 1000 hour
MTBF; to require 90% (on the average) of all such parts to have 1000 hours; or
any number of requirements. Horeover, in an unambiguous statement, the true
intention of the specification must be accurately stated. For example, if the
*: reliability requirement is that a part must survive 100 hours with 90%
probability it is often tempting to "convert" this requirement to an MTBF
specification based on exponentially distributed lifetimes. For an
exponential distribution an MTBF of 1000 hours is roughly equivalent to a 100
hour survival probability of .90 (since exp(-lOO/lO00)sexp(-.l)-.90 4 8).
*." Specifying a 1000 hour MTBF woulo be approximately equivalent to the
requirement in the exponential case. However, if lifetimes are actually
"Weibull distributed with survival function given by
exp(-(t/308) 2 ), -
then the mean of this distribution is only 273 hours, and yet the survival
probability for 100 hours is
exp(.(1OO/308)2)_exp(_.105),.90 .'
which satisfies the original requirement. Imposing the 1000 hour mean life
would thus cause overdesign. Also, this part actually meets the reliability
requirement of survival of 100 hours with .90 probability, but would probably *
*. fail the corresponding reliability demonstration test that is designed to a

10.O hour hTBF requirement.
Many of these difficulties are alleviated when the underlying life

distribution is exponential. In this case, it is sufficient to specify mean
life, and knowing the mean determines all other quantifiable reliability
measures associated with the exponential distribution. The exponential
distribution has been the life distribution of choice in the electronics
industry because of compelling evidence (both empirical and theoretical) which A
Smakes it suitable for describing lifetimes for complex electronic units whose
failure modes are not wear-out related. Consequently, much effort has been
"spent in developing &nd updating MIL-STD-781C, "Reliability Design
Qualification and Production Acceptance Tests: Exponential Distribution." '.' '
And, in view of the success of the exponential distribution in the analyses of
section 2 of this notebook concerning its applicability to nonelectronic
ft parts, in many cases specifying mean life will be adequate for nonelectronic
parts as well.
3-1
..............................
....-.....-. ,...- .....
S. .. .. . , ,., , * , - *- .*-',-• - * -,* . .
MLi Downloaded from http://www.everyspec.com
In summary, reliability can be quantified in many ways: mean life,

median life, probability of mission survival, etc. A reliability
specification may be any one or more such quantities depending on the
underlying reliability requirements, and/or life distribution. The
reliability specification must accurately reflect the underlying reliaoility
requirements, and be both semantically and quantitatively unambiguous.
3.; Reliability Specification for the Exponential Distribution. The

survival function (or reliability function) for the exponential distribution
is R(t)-exp(-t/0), t>0, where b is the mean life. The most direct way
to specify any reliability requirement in the context of the exponential
distribution is to convert the requirement to a requirement on mean life. The
specification of mean life or, equivalently, MTBF, is particularly convenient
in subsequent reliability demonstration test design since MIL-STD-781L is
based on MTBF specifications. This, along with the fact that the exponential
aistribution is uniquely determined by the mean life (making MTBF an •. 7•
unambiguous specification) should establish MTBF as the best form of - -
reliability specification. The following table presents formulae for
converting various reliability requirements to MTBF in the exponential case.
RELIABILITY MEASURE CORRESPONDING MTBF
Failure Rate 6 - 1/(Failure Rate)
Probability of survival for b - -t/ln(r)

t hours m r
x(p) - the p quantile life t -x( p)/ln(l-p)

(i.e. x(p) - the life . -
beyond which the part will live .. /,
with probability l-p)
The following examples should clarify these concepts.
Example 1. . :",
The reliability requirement for an equipment is expressed as a failure

rate of 250 failures per million hours. The corresponding MTBF requirement is
(l/150)(l,0(JO,0O0) hours * 4000 hours.
Example 2.
The reliability requirement for an equipment is expressed as a

probability of mission ( - 1000 hours) survival of .99. The corresponding
MTBF requirement is -I100/ln(.99) - 99,499 hours-
Example 3.
The .JO quantile life requirement for an equipment is 1000 hours. The
corresponding MTBF requirement ia -1000/ln(l-.10) = 9,491 hours.
P..
..
3 - 2.,.. " ...
, - ' .- -~~~..
. .... . . .,. .. =. -. -... ° .*.. o .• -° .# #• ) . . . • . • .- - . . .' '.•
...- ."..."'.."---....-.,.......... .. . .-.. •.<.•.•.•........... -. '." .-....-. ".. i ..-.. '.."-"??...:-, :"..">."
In order to use MTBF specifications in a reliability demonstration test

from MIL-STD-781C, it is necessary to specify the upper (acceptable) test MTBF
(o0), the lower (unacceptable) test MTBF (tI), the producer's risk U,
and the consumer's risk 0. Further details may be found in MIL-STD-781C.
J.3 Reliability Specification for the Weibull Distribution. The two

parameter Weibull survival function is given by
R(t) - exp(-(t/b)C), t > 0, b > 0, c > 0.
The parameter b is called the scale parameter (also, characteristic

life), and the parameter c is callea the shape parameter. The following table
lists common reliaoility measures in terms of these parameters. ;'-
RELIABILITY MEASURE FORMULA
Mean Life b F (1+1/c), where I is the

usual Gamma function.
p quantile, x(p), i.e. the b(-ln(l-p))l/c

life beyond which the
equipment will survive with
probability 1-p.
Characteristic Lite, i.e. b

x(.b3Z)
Since the Weibull distribution is a two-parameter distribution, any

unambiguous reliability specification must involve two quantities. For
example, specifying (b,c) directly would be sufficient, although not much %
physical significance can be attached to the parameter c. Other specifi-
cations whicn would also he sufficient would be mean life and the .90 quantile
life; the .50 and .90 quantiles; or characteristic life (b) and .95 quantile
life. These possibilities are summarized in the following table.
RELIABILITY SPECIFICATION CORRESPONDING VALUES OF (b,c)
Mean Life, v, and p quantile Must be found by iteratively

x(p). solving:
0 (1+I/c) *-.
x(p) - b(ln(l/(l-p)))I/c
3-3
-. i --. . - -.-..- . -. . - ..-.. . ....

• ..
: ~~~~~~~~~~~~~~.
. . . . -"---°
. . ...............
- -- ""5" -r-'--.i.''-....."-"""'"" """"
Thra•'kr..i * •M s' '.a • - ... . . ; L • _•• • = Downloaded
• ,%r
:. .• from
• r http://www.everyspec.com
. ... 1 . .• , . , .. . • .- .....
RELIABILITY SPECIFICATION COFREFSPONCJIZ,4- VALUES OF (b,c)

ln{ in~i!/{1-g)/ln(1/{t-pi)]J•
Two quantil#-, n,,x(q). C - nx~q)/x(p)j
b- x(p)
ln(l/ ±l-p) ]l/c
Characteristic Life, b, and the b is the Characteristic Life
quantile x(p).
C In[In(l/i/[-p•'"..
- n4x(pj-
It is not often practical to specify reliability in terms of the Weibull

or any other two parameter distribution because the corresponding reliability
demonstration test procedures are cumbersome. Moveover, since there are two .
parameters involved, there is no "OC curve- as in the case of MIL-STU-781C for
the exponential distribution. Finally, and most importantly, there is no
natural ordering of the parameter pairs (b,c). That is, given that
"acceptable" values have been established for the parameters (b,c), there is
no obvious logical way to assign "unacceptable" values to (b,c). For example,
the pairs (b,c)-(308.6,2) and (b,c)a(212,3) both determine Weibull .'.,.
distributions having 100 hours as the .10 quantile. The respective mean lives
are 273 hours, and 189 hours, so that the pair (b,c)-(308.6,2) appears more
acceptable. However, the .05 quantiles for the pairs (b,c)-(308.6,2) and
(b,c)-(212,3) are 69.9 hours and 78.8 hours, respectively. Thus, from this
point of view, the pair (b,c)-(212,3) is more acceptable since 78.8 hours are
survived with probability .95, whereas for the pair (b,c)-(308.6,2) only 69.9 f_,.,
huurs are survived with probability .95.
One way to alleviate this problem is to fix one parameter for all cases.
This is the same as assuming that one parameter, either b or c, is known
exactly. Since neither b nor c is ever known exactly in practice, this is an %.... ..f
unacceptable solution.
3.4 Reliability Specification Without Respect to a Particular Underlying "

Life Distribution. In many instances, there is no evidence to suggest a
feasible parametric (e.g. exponential, Weibull) life distribution for the
equipment on which reliability is to be specified. Also, it is often the case
that when a two-parameter life distribution is appropriate, the reliability
requirement is only sufficient to determine one o0 the parameters. horeover,
in a two-parameter model, it is not always clear how to specify acceptable
values for the parameters versus unacceptable valuei even when the reliability
requirements are surticient to determine both parameters.
3-4
*. ' .. ...
,-• ~ ~Whenthese st.tuations arise, there is a type of reliability specification ~o.•s'':•

h •.1 in conjunction with reliability
which can be use6 demonstration tests which do
•-= ~not assume any particular underlying parametric life distribution (so-called • _
-. ' ~nonparametric tests). These specifications are life distribution quantiles.,• •
;" This type of specifica ton has already been discussed in the sections on
reliability specification for the exponential distribution. There, however,
• it was suggested that these quantiles be used to compute corresponding MTBFs
S~~to facilitate the use of HIL-STD-781C. The p quantile of the life '.i•. ,
S~distribution (denoted
live with probability
by x(p)" Js the life
a-p.
beyond which the equipment
can be spelibity
,.,ntile x(p) d will
or n
pecificdttly,
•-- indirectly, by specifying the probability of survival for a fixed time

hperiod For example, specifying that with 90 probability the equipment shall
survive
no hours is equivalent to stating that the .i0 quantile of the llfedJ.
a000
odistribution shall be 1000 hours, i.e. x(.10)=1r00 hours. In many cases, it-
is of
the
rliof Ths
l sfeto place some
interest
distribution,
some measure
mean life. on While
e.g.requirement
of does
mean life not tendency
central lend.
itself to nonparametric demonetration tests, median ltfe does, since median
life is x(.50), the .50 quantile of the life distribution. Indeed, when mean
!• ~ ~life
is specified, the reliability engineer is often (mistakenly, in general),"'--:.
thinking of median life, i.e. thae" life beyond which 50% of the equipments
will survive, on the average. nc bft
-The following table summarizes the strategy for specfyingi.'' reliability invf
underlying parametric life in distribution is assumed.-...,.'
cases where
prThese of particular
types no specifications can be used directly the nonparametrc test -. -
dpisar discussed in section 5.h ilona
RELiABILITY SPECIFICATION WHEN TO USE"..-.'.i
x(.50), the mestan life. Requirements concern a measure

tndc ofceta
central of the life
tendancy Ined hnma
distribution.
s
The probability ofth rrviving Requirements concern ain general)
fixed time Td
tha partyichlar mission length T whichs
the equipment must survive with

"T o b ah high probability.
central o-the•lif
tndancy
p(5), <h p < .5. See above. distribution.,
The prbabilty Reuiremets cocern.
ot~urvivng
* ie ie .priuia iso ent hc
the eqime.mutsuvvewt .4
4.0 SPECIAL APPLICATION METHODS FOR RELIABILITY PREDICTION
4.1 Introduction. This section is concerned with the estimation of

reliability of mechanical and electromechanical components in service. We are _ _
not concerned herein with the classical eatimation problems faced by
statisticians which are described in Section 2 but with practical examples
where the engineer is required to create an estimate of reliability from
whatever data he can find. The methods described in this section involve the
use of "engineering experience" or judgment to extend what is usually meager
sampling information to obtain meaningful comparisons of materials, designs,
and environments.
It is only through the application of engineering Judgment to available data
that an estimate of reliability in service can be created. The consequence of
such methods will of necessity be a point estimate rather than an interval
estimate because the probability distribution of the estimator cannot be
obtained.
4.2 Background for Reliability Prediction Model Development.

Nonelectronic components have numerous failure modes as compared to electronic
components. Some of the more basic failure modes which affect this class of N-
components are fatigue, creep, impact, thermal shock, corrosion, oxidation,
fretting-corrosion, elastic deformation, relaxation, lubrication failure,
wear, spalling, erosion, leakage, delamination, buckling, and radiation
damage. Detailed discussion of these failure modes may be found in ASME
(1965). In a proper reliability assessment the dominant mechanisms must be
identified and considered since each mechanism represents a competing failure
risk with its own failure distribution.
Several possible approaches are available for the model development, each
of which has definite merit but is also subject to limiting constraints.
One approach is through the analysis of accelerated life test results.

Thli approach presupposes that a large number of devices has been tested or is
currently being tested in combinations representing the various technolobies,
processes, etc. The results of such controlled tests would provide some
inlication of the characteristics and peculiarities of the devices as a
function of the several configurations, stresses and applications included in
the test design. However, the extrapolation of these accelerated test results
to more normal operating conditions would be open to questions of validity due
Lo the uncertainties regarding the extrapolation algorithm. Further, while U9'L
test dats under controlled accelerated conditions should aid in understanding
the reliability characteristics, it is difficult to obtain data that covers
the wide range of technologies and stress conditions that would be necessaty
in order to place major dependence on this approach alone.
An alternate approach involves the development of a reliability model and

its parameters based on a knowledge of fabricatian techniques and the
* anticipated failure modes. Also requi~red by this spprodch is a thorough
* understAnding of the fundamental physical/metallurgical/chemica]/ilectricaI
degradation mechanisms involved, as well as the proportionare weighting of
these mechanisms in translating to the varioub configurations the component
may assuwe.
4-1
A third approach would be to rely solely on the collection and reduction

of empirical operating data where the pertinent information with respect to
the model parameters would be extracted using suitable statistical
techniques. This approach should provide optimal applicability since the
field data reflect the actual reliability experience of the devices operating
in their use environment. Powever, it requires the collection and reduction
of a large database on the entire range of device configurations and
application environments in order to evaluate each of the critical factors.
In some cases, particularly with new devices, the amount of data needed Co
provide sufficiently accurate results may not b,-, available.
The best approach endeavors to utilize the collective data and knowledge
offered by all three approaches atid subject it to careful, analytical scrutiny
to censor out conflicting and discrepant information. This approach includes
the following tasks: a literature review to define the component, equipment
and environmental attributes which will be considered during model
development, derivation of the preliminary model form, data collection, data
reduction and analysis, development of the model parameters, and model
refinement and verification. ,
Regardless of the approach taken, the derived model should have the
following attributes:.-
k verified accuracy over the total range of all factors considered
0 an uncomplicated approach using easily accessible information on

comFonent characteristics and environmental parameters
* dynamic, flexible expression, easily modified to accommodate new..* .

techniques -,'-
* appropriate discrimination against design and usage attributes which

may degrade reliability
The simultaneous attainment of all of the above objectives is difficult,

if not impossible. Often these goals are contradictory or mutually
exclusive. As an example, some years ago a prediction model was proposed for
* microcircuits which possessed commendable accuracy over the range of
"parameters and was based on sound theoretical considerations. Unfortunately,
to use the model, the engineer was required to input such information as
* metallization area, total diffusion area and other such fabrication/design
information. Since this information was not available on vendor specification
sheets and indeed was often vendor-proprietary, the model proved to be
useless. Modifying the model Lo use more generic but readily available device
parameters degraded the accuracy of the model. The overall utility of the--
model, however, was enhanced. A discussion of modeling and model limitations
can be found in Flint et. al. (1982).
In reviewing the literature it becomes obvious that an abundance of -.

models have been adva'-ed for use in the prediction of nonelectronic compon'. " °'
reliability. Unfortititely, mout have been found to be deficient in onc or
4-2
* - ** .* ~ ~ ~, 1.:v
more of the areas below and as a consequence cannot be included in this
section:
* Not application-oriented
* Not yet "engineer-ready"
* Not verified
o Single-vendor specific
* Requires information not easily available
4.3 Graphical Approaches to Reliability Prediction. The graphical

approach discussed below briefly provides a means of translating meager
information into a reliability estimate.
The information is most likely vendor supplied and thus may be

optimistic, based on limited test data, or in the worst case represents a
design objective. For these reasons, unless prior experience dictates \.
otherwise, assumptions must be conservative in nature. For example, if an

L10 (life time beyond which 90% survive) supplied by the vendor is based on
small sample tests or if the vendor is unwilling to discuss the matter, the
L 1 0 life should be reduced by one-half as a minimum.
The ambiguity regarding life frequently must be clarified. Life may mean
average, median, some period of time/cycles by which some percent will fail on
the average, or the time to first failure among a population.
Weibull probability paper is a most useful tool available when the use of
a graphical solution is necessary. When the shape or slope parameter, B, is
equal to one, the distribution reduces to the exponential; an approximation of
the lognormal distribution results from values of 8 in the range of 1.5 - 3,
i and the normal distribution approximation results from a 8 value of 3.5.
With an estimate of some percentile of life, an assumption of the value of
"",and the use of Weibull paper, the probability of failure before time x,
say F(x), and the reliability, 1 - F(x), for any lifetime can be determined.
In those cases when a failure rate must be used in a prediction, it must

be remembered that only for the special case of the exponential distribution,
i.e., 0 = 1, will the failure rate be constant with all other values of 8
*. greater than one resulting in an increasing failure rate. In the latter case
an average failure rate over a stated time period may be calculated employing
the average cumulative kazard function, H(t). The expression is (assuming
"l-F(x) exp- x) I ):
t et. al (1977).
H ()W-- --- (Wilson,
Alternate notai.ion for H(O) is X(t).
4-3
4.4 General Theory of Interference. When the strength of a

component is less than the stress imposed on it, a failure can be expected to
occur. Strength is the ability of a component to resist failuri when
subjected to stress. Stress or load may be defined as a mechanical load,
dimensional variation, environment, temperature, etc. Since boch strength and
stress are variables, they may be described by probability distributions.
The strength of nominally identical components can be expected to vary

due to variations in materials, dimensions, treatments, surface conditions,
and so on. This variability can be described by a distribution function. • ?•,i
Various approaches to estimating the strength distribution function are given
in Bompass-Smith (1969), Burns (1975), Konno et. al. (1975), ASME (1965),
Nilsson (1975), Thomas et. al. (1975), and Welker et. al. (1975). Most
typically, however, so little will be known that both the form of the
distribution and the variability about the mean will have to be assumed.
Unless experience dictates otherwise, the strength variable is often assumed %.'
to be normally distributed with standard deviation equal to 5 - 15 percent of %.%
.
the normal value (Fulton, 1983). ..
The distribution of operational stresses can only be known to a

reasonable certainty when the response of a reasonable number of prototypes to
the full spectrum of operating conditions has been closely observed. Due to
such constraints as time, cost, etc., the distribution of stress cannot be
established and assumptions must be made. Further information regarding the
estimation of the stress distribution may be found in Kececioglu et. al.
(1964, 1967), and Fiderer (1976).
If the expected distributions of stress and strength can be estimated for

a component then by employing interference theory the probability of failure
of the component can be calculated. The concept is presented in detail in
Kececioglu et. al. (1964), Disney et. al. (1968), Lipson et. al. (1973), Kapur
et. al. (1977), Kececioglu (1972, 1974, 1977, 1968), Kececioglu et. al.
(1974), and Dhillon (1980, 1981).
The mathematical foundations of Interference Theory may be outlined as .".

follows. It is assumed that the stress is a random variable X having .- ..
probability density function f and that the strength is a random variable Y
having probability density fy. It is generally assumed that X and Y are
statistically independent (although this assumption is not strictly
necessary). The probability of failure, p(f), is then
p(f) = P 1>
i.e. the probability of failure is the probability that the stress exceeds the
strength. An expression may be derived for p(f) as follows.
P(f) " P x > Y(f f

f(x) f (v,)d.'.
4-4 '- " .

fy (y) (1 - Fx(y)) dy
where
FX (Y) fXW dx
is t.-e cumulative distribution of the random variable X. Thus, the

probability of failure is the area beneath the curve fy(y) (U - Fx(y)) as
y varies over all real numbers. Intuitively, (l-Fx(y)) fy(y) dy is the
probability that the strength is in the infinitesimal interval (y, y+dy), and 'i
that the stress exceeds y. Integrating (or "summing") over all y then gives
the total probability that stress exceeds strength.
The area under the curve fy(y) (1 - Fx(y)) is often referred to as

the "interference zone.i When the two densities fx and fy coincide (i.e.
are exactly equal), the probability of failure is exactly 1/2, although this *O's
i3 by no means the maximum value possible. In fact, when fX is the normal
density with mean vX ad variance ax and f2 is the normal density with
mean my and variance :I, tarinc and %or
can be shown that
u -u
p(f)
if
)pf/21/2
(1 ,.
(2) p(f) 1 as UX my fixed;
*(3) p(f) -. 0 ,as my + 0 uX fixed;
_.
Thus, p(f) can take any value between zero and one. The explanation of (1) "- .-?i / -:.
*above is that when the mean stress equals the mean strength, it is equally
*likely for stress to exceed strength and vice-versa. In (2) above, when the
*mean stress is very large with respect to mean strength, the probability of
failure is close to one. In (3) above, when the mean stress is very small
Iwith respect to strength, the probability Df failure is very small.
4-5
The concept of interference is shown in Figure 4.4-1 where the

interference zone is given as the shaded area. This illustrates the simple
case where the strength distribution remains unchanged across time, i.e., is
not affected by exposure to the failure causal stress distribution. Figure ,. '
4
.4-2(a) and (b) illustrates the case where the strength distribution degrades".--•.
as a function of time exposure to stress as the result of such failure
mechanisms as fatigue, corrosion, and wear. Whether this time shift must be
considered or may be dismissed depends of course on the rate of change
expected. For example, in a naval environment corrosion is a rapid failure
causal mechanism and the effect on the strength distribution must be
considered; on the other hand, in most military ground fixed applications
corrosion is a weak failure causal mechanism and the effect on the strength
distribution may be ignored.
4-6
." I......£
L.' ~.'o'"•
4-6"
r. - .. - ".. .U
"i I
Iwo "te
to
0.4
0.7- STRESS STRENOTH

PROWABILITY PROBABILITv
"0.6. OENSTY DENSITY
0.5.
0.4,
0.3-
0.2.
0.1,
0.0
-2 -1 01 2 3 4
INTERFERENCE IS SH.4ADEO AREA
.<" %* 4.
Figure 4.4-1. Illustration of the Concept of Interference -
". -N.-."%".
"
N. o.°. " **
.7., .
4-7-
0.5
0.I. -
0.41f
0.3-
W.0.2.~
0.1
0.0
-2 -10 1--- -- -3 4
STRESS/STRENGTH
LEGEND: GRAPH4 .. STRENGTH4 -STRESS

INTERFERENCE IS SHADED AREA. STRENGTH IS AT TIME T.
K Figure 4.4-2(A) Time Varying Strength Density
4-8
Yk
0.7,
0.6.
0.5
0.4,'
0.3
I
0.2,
0.0
-2 -10 12 3 4
LEGEND; GRAPH .. STRENGTH -STRESS
INTERFERENCE IS SHADED AREA, STRENGTH IS AT TIME S> T.
Figure 4.4-2(B) Time Varying Strength Density
4-9
4.5 Applications of Interference Theory to Reliability Prediction

Methodology
Source: RADC-TR-66-710 (March, 1967) by Charles Lipson et. al., entitled

"Reliability Prediction-Mechanical Stress/Strength Interference
Models. "
RADC-TR-68-403 (December, 1968) by Charles Lipson et. al., entitled

"Reliability Prediction-Mechanical Stress/Strength Interference
(Nonferrous)."*
4.5.1 Purpose. The purpose of this method is to obtain a point estimate

of reliability in service for mechanical components subject to fatigue,, •
failure. The method is applicable to;
a) Components subjected to completely reversed cyclic bending from

axial, or torsion loads.
b) Components subjected to a combination of static and cyclic loads.
4.5.2 Description of Method. Given that strength and service stress r-

each have a probability distribution of known type, and that the parameters of
the two distributions are known, the probability that a random observation
from the strength distribution exceeds a random observation from the stress
distribution is equal to the reliability. The term "interference" is used to .
designate an occurrence where stress exceeds strength, so that reliability is '.*
the probability of no interference.
The source documents provide extensive tables giving the probability of

failure as a function of the parameters for the following combinations of
distributions:
Strength Distribution Stress Distribution
Weibul 1 Normal .K
Weibull Weibll
Normal Normal
Largest Extreme Value Normal
SmallesL Extreme Value Normal
These tables are perfectly general, and may be used for other types of failure
than fatigue. However, the two reports tabulate parameters for fatigue
failure only. Nonstandard symbols for the parameters are used in these
tables. Table 4.5.1 shows their relationship with generally accepted symbols.
Several tables have been compiled in the source documents giving fatigue
strength parameters for virtually all of the common ferrous, nonferrous, and
light metal alloys, subjected to completely reversed bending or axial
stresses. Included are the effect of heat treatment, surface finish, stress
concentr.:tors, temperature, and frequency.
4-10
0- 41 01.
TABLE 4.5.1. PARAMETERS OF TABULATED PROBABILITY DISTRIBUTIONS AS USED

IN4 RADC-TR-68-403 (LIPSON4, et. al. 1968)
1. Weibull Distribution:
R(x) - exp X -
0isthe lower bound of x
3 parameters 8 is the characteristic strength
bsthe slope parameter
This compares with the usual 3-parameter Weibull distributioni
2. a - 8-x 0 )b.
NrmalDistribution:
2.
whichgreesDithstributi on ain
3.Extreme Value (DEV.

istribution,
RWx *ep[e 8(x-M) -* < x <-
x is the variable
B8 is the scale parameter

2 parameters -. d
IM is the location parameter
4-11
TABLE 4.5.1. PARAMETERS OF TABULATED PROBABILITY DISTRIBUTIONS AS USED

IN RADC-TR-68-403 (LIPSON, et. al. 1968) (Continued)
This compares with the usual 2-parameter S.E.V. distribution,
R(x)
W exp •ec(x-u
as follows:
M.
For the Largest Extreme Value distribution, the same relationship holds, where
R(x) 1 - exp -e - (x-M) I "a
e,%-p[-ej
Note: in 1), 2) and 3) above, R(x) is the reliability function.
The procedure for estimating reliability using the interference method

takes several forms depending upon the assumptions made regarding the strength
distribution and the stress distribution. The various methods are
illustrated withn numerical examples in the source documents. .'". '
.'""
% .
4-12..
..
. . .- -. o
4.6 Application-of Interference Theory for Normally Distributed Strength

and Normally Distributed Stress
,-
Source: (1) Fiderer, Leo, "Design For Reliability in Hostile
Environment," Microelectronics and Reliability, Vol. 15, .-.- ',.-
Supplement, Pergamon Press, 1976, pp. 75-85.
(2) Lipson. Charles, Statistical Design and Analysis of

Engineering Experiments, McGraw-Hill, 1973.
.' .
4.6.1 Purpose of the Method. To obtain a point estimate of reliability
in service for non-electrolic components. The method may be applied to a
diverse number of failure causal factors.
of interference
The description equally of
4.6.2 Description of the Method. in 4.4 applies to this
distributions given
stress and strength .
-.
variables,
method. The difference lies in the assumption that the random
various distribution forms, may be approximated by a normal
which may take
normal distributions may be
distribution so that in practical calculations
assumed without excessive error.
The steps involved in the procedure are as follows:
(1) Estimate the mean, PL, and standard deviation, OL

S.
of the load, L, and lis and Os of the strength,
Where there Is no information available to estimate oL V _

.05p -
or U a value may be assumed from the interval
,15p, Where the part is critical or a high reliability
requirements exista, a worse-case approach should be taken,
i.e., in the range of .10ujto .1511. For most
conditions a may be taken as .09p.
as a new random
(2) Define the difference between strength and load "
variable D:-"--
D - S - L,-
1
with mean p 1- 11s PL,
and
Cs L --
standard deviation OD-
(3) The probability of failure, P(f), is found by computing:
P(f) p
P 0
D <0 = P S-L < 0
D-S~
-P.. ..--
aD aD
4,-13 "-...
"'I '.- i.
G4A
'.44
u Proabifty
0*.
- ) /•,_•
oo•~~ ( • - • . '-.9-
" . . ..'.-.:..
..
-3 34-
Figure 4.6.2-1.
Normalized Density Function of Excess Strength Over Load.
.. .. . .... . . . . . . . . . . . . . . . . . . .
W- ft .- ~-
4-14
F% Downloaded from http://www.everyspec.com
%*
1 ff exp 't 2
/2 dt2
Thus, P(f) is found by entering a table of the cumulative

distribution function of a standard normal random variable with -'"
the value -uD/aD. See figure 4.6.2-I.
(4) The reliability R is then
R -~) % .
J. ,-,
4.6.3 Examples
4.6.3.1 Shaft and Bushing Reliability. The reliability of a shaft and a

bushing after ten years exposure (in a nonoperating state) to a heavy ''-'"- *'
industrial environment is to be determined, i.e., the ability of the shaft to

rotate without drag.
The dominant failure mechanism is considered to be corrosion.

Information on corrosion rates and corrosion product build-up is available
from the Battelle operated Metals and Ceramics Information Center.
In the analysis it will be assumed that the part dimensional variability

is normally distributed and that the maximum and minimum allowable dimensions
may be taken as the upper and lower three sigma points, respectively.
Part Specifications .'
Shaft Stainless Steel

Dia. .123 - .124 l.' '' loss
Corrosion Rate 3.5 x 0- in/year loss
Corrosion Products 7 x 10- in/year buildup
Net Gain 3.5 x 10-5 in/year
Bushing - Aluminum Chromated

I.D. .125 - .127 in.
Corrosion Rate 6 x 10-5 in/y ar loss
Corrosion Products 7.8 x 10- in/year buildup
Net Gain 1.8 x 10-5 In/year
The maximum and minimum diameter 30 points after ten years based on the
buildup of corrosion products are:
Shaft Max 3o, . .124 + 3.5 x 10-5 x 10 x 2 a .1247

Min 30s - .123 + 3.5 x 10-5 x 10 x 2 - .1237
6os = .1247 - .1237 - .001
.001
as
s 8 - .1.42
4-15
~. .
**~ . K
*.*O.° :w...o.. f'.: ., .
• • •o.'o .'-
Bushing Max 30B - .128 + 1.8 x 10-5 x 10 x 2 - .12764

Min 30B - .125 +1.8 x10-5 x 10x 2-a .12464
6 .12764 - .12464 -. 003
0B
.003
OgE 66X
UB -. 12614
The unreliability or probability of failure, p(f), is evaluated by
establishing a new random variable D:
with mean PD B us .00194
and standard deviation a a 2 + a -2 5.27 x 10. 4.

D B L
Thus, P (f) P D < 0 - P DP)G <-DO
-u/0/ p dt .
From a table of the standard normal cumulative

4
distribution entered at -MD/CD --. 00194/(5.27x10- )
-3.68, it follows that
P(f) a 0.0001 .
and the reliability is:
R - -P(f) - .9999.
4.6.3.2 Lifting Eye Reliability. A lifting eye intended to be used in

lifting shipboard equipment while the ship is at sea has a nominal SLrength of
60,000 lbs. The dead weight load is 12,000 lbs. This is all the information
available to the reliability analyst.
It will be assumed that both the strength and load can be represented by
normal distributions and that the nominal strength and load are the means of
the distributions.
us 60,000 WL 12,000
The variability of the tensile strength should be controllable such that

the standard deviation may be assumed to be eight percent of the mean
strength. Huw'ever, the load due to the dynamics of wave/ship action may be
quite variable. Thus, the load standard deviation will be assumed to be
twenty percent of the mean load.
C .08 w 4,80u aL .2 1 2,400

L L
4-16
%'
The probability of failure is evaluated by forming the new random

variable:
D "S -L "•--""
%5
where " -
P WS2L
8
,0 a +n
25366.
thus,
p~f'JD -j)l
LD/%D
~D 0 D~u - ifexp
where
-PD/o - -48000/5366 - -8.9.

D D
function the
From a table of the standard normal cumulative distribution
value of P(f) is approximately 0.
Thus the reliability is
R 1 - P(f) " 1.
4.6.3.3 Diaphragm Reliability. Calculate the reliability of a diaphragm

Each diaphragm must pass 5 cycles
intended to be used in a one-shot device. The .'*.-."...
of a worst-case pressure-time profile proof test as acceptance criteria.
3 sigma point to be 5 test cycles. -
purchaser has specified the lower
failure which resulted

A small sample size population has been tested to
mean lifetime of 11.44 proof tests. The sample size was
in an estimate of a
distributed.
not large enough to prove the lifetimes to be normally
of the production
it will be assumed that the proof test lifetimes
a mean of 11.44 test times and a
population will be normally distributed with
a of 2.147.
and noting
Denoting by L the lifetime in proof tests of the diaphragm,
the diaphragm needs only to
that this is a one-shot device and therefore is
of failure
survive one test, it follows that the probability
< 1~ P L-I< where~k (I -1.4/

P(f) -P 1L
2.147 - -4.86. Referring to a table of the standard normal

cumulative distribution function, it follows that
P(f) - .000000605
4-17
., . ... . .. . . ...
. •.
.- .• - - - ° , • •, .- . . . - -°,: oo ,- , -, . :., . K-..• o,~*-....°
I ., o •.... .. °. . . . . . ° -
and then
R - - P(f) .999999395
. _______
4.7 Application of the Average Failure Rate Method for Grease Lubricated
Rolling Element Bearings
Source: Wilson, D.S., and Smith, R., "Electric Motor

Reliability Model," RADC-TR-77-408, December, 1977
(AD 050179).
4.7.1 Purpose of Method. To obtain an estimate of the average failure

rate of grease lubricated rolling element bearings over a related time period.
4.7.2 Description of Method. The method consists of an empirical

mathematical model which provides an estimate of the Weibull characteristic
life, t, and was developed through the use of regression analysis and a
large data base. Essential tables are provided.
The average failure rate is obtained by averaging the Weibull cumulative

hazard function over the time period of interest.
A modification has been made in the characteristic life model so that it

is valid for a single bearing rather than for a population of first failures
of pairs of bearings as given in the source. The method employed is given in
Ang (1970) under Suspended Item Tests.
The characteristic life model is based primarily on the effect of

temperature on the lubrication qualities of greaae and such secondary effects :',..
as quality, bore size, speed, grease, and load. The consideration of load as ..
a secondary factor is consistent with good decign practice which limits the
loads on grease-lubricated bearings to 15 percent or less of rated load
capacity.
The models are as follows:
1.878 •
=St "t (1)'
1.241
B + 1
q 4.32 Kg *300(--76 S+ 19+7

10 TIs 0- .+300.
..::*...- ..-,,
4-18
*5.-..:..'.
','.. . . -• ., , " -.
. . -"..-.- .-.
,- -..'-.,.-... ' . . ." ." : • -. -• . .. . - . _ .. .. .- .. .• -. " . --... -. .. . - _ ,-. .' -- -< .. '-%
Ix.
where
t - time for which failure rate is required (in hours)
cB *bearing characteristic life (in hours)

LJ
q -quality factor
DN *bearing bore (mm) x speed (RPM)

-
T *bearing operating temperature (degrees Kelvin)
Kg - grease constant L W
N 1RPM i
SP Specific dynamic capacity at 33 1/3 RPM in pounds
TABLE 4.7.1. QUALITY FACTORS
TABLE 4.7.2. K CREASE CONSTANT
MIL-Spec. Max. T*C Kg

jsource Oil Thickener
Sodium and MIL-G-3278A 170 1.35

1 Diester
Solid Lubricant
Lithium MIL-G-3278A 120 1.55

2 Diester
Lithium 150 1.74

3 Silicone ----
Sodium MIL-G-18709A 150 1.41

4 Mineral
Lithium MIL-L-15719A 177 1.81
5 Silicon
6 Synthetic Non-soap MIL-G-81 32 2 170 1.74

Hydrocarbon
4-19
4.7.3 Examples
Example 1.
Determine the average failure rate of a bearing with the following

", specifications for an operating period of 10,000 hours.
Military quality
Grease 5 '.-,
Bore Dia. 13 mm
Bearing Operating Temp. 30%C
Speed 3600 RPM
Load 1W lbs
Specific Dynamic Capacity 505 lb (from bearing manufacturer's

catalog)
Then:
q (quality factor) = .12 (Table 4.7.1)
"Kg (grease constant) - 1.81 (Table 4.7.2)
DN - Bore D x speed - 13 x 3600 - 46800 .' .-
T - 30 + 273 303@K
Using Eq. 2
1. 241
(33
"[130 0 1." 01+3 "4"60
.- "..1
10 10+0
6.009 + 3.990
10 10 .300
4-
9.8 1.241
-7 -5
9.8 x 10 + 9.928 x 10
* 12378 hrs.
4-20G
. -.. ° -S.
:. ~~*Au~ t-~,~w~~.~*~ .:,~~~~*55

,•.t'.'.':-
o. * P-..- M *.• .. ,
Using Eq. 1
S00 (12378) 2.878 .000054
or 54 failures per million hours.
NOTE: This X is valid only if the bearing is replaced at the end

"of
the 10,000 hours of operation.
"Example 2
Determine that period of operation, t, for the bearing of Example I that
" will ~~~Solve

rsolte Eq. I for t
inteaeag alr rate equal to 20 failures per million hours. ""''- ..
1 ...,.:...
t x cx2.878 1.878
-6 (12378)2.7
= 20 x lO-6x 2.878] 1.878 "'- ''
5886 hrs.' .
4.8 Reliability Prediction Method - Rolling Bearings Oil Lubricated ".-
Source: %
%'
(I) International Organization for Standardization ISO 281/1-1977(E), -
"Rolling Bearings - Dynamic Load Ratings and Rated Life - Part 1:
Calculation Methods."
(2) Marks' Standard Handbook for Mechanical Engineers, 8th ed.,

McGraw-Hill, 1977, pp. 8-136 through 8-142.
4.8.1 Purpose of the Method. To obtain a point estimate of the

reliability in service of rolling bearings.
4.8.2 Description of the Method. Standard formulas have been developed

to predict the ,10 life of a bearing under any given set of conditions.
These formulas are based on an exponential relationship of load to life which
has been established from extensive testing.
L W x 106 cyc. ""-..•-

10 K 0P
4-21
L-'.Vi-
whe re
L = the number of revolutions that 90 percent (on the average)

"10 of a population of bearings will complete or exceed without failure,
"i.e., R - .9.
C 0 basic load rating, lbs.
P a equivalent radial load, lb.
K - 3 for all bearings, 10/3 for roller bearings. - "
To convert to hours of life (L 1 0 ), this formula becomes
16,666.67 C
10 N (2)
where N - rotational speed, rpm.
The basic load rating, C, value is readily obtainable from any bearing
manufacturer's catalog. All bearing loads are converted to an equivalent
radial load, P. Equation 3 is the general expression used for both ball and
roller bearings.
P XR + YT (3)
where
R radial load, lb.
T - thrust (axial) load, lb.
X - radial factor
Y - thrust factor.
The X-Y factors may be calculated using the methods described in Source I or,
with some loss in precision, average values may be selected from Table 3, pp.
8-140 of Source 2.
One further formula is necessary to adjust the 1.10 life for other
levels of reliability and less than optimum operating conditions. -
-n a, a2 a3 L10 (4)
where:
al = life reliability factor
a 2 = material properties factor
a 3 = operating conditions factor.
4-22
-.. ,........ ,... ,... ........ ...-. .-.-. . .. , .- . . ... .:--.:: ::.::::
",. -.- ',:: .. - .,, :: --- : -"- - . .: . -, , ,". . . . ". , . - .. .. - :'. .-... - . :
"The reliability factors are given in Table 4.8-1.
TABLE 4.8-1: LIFE ADJUSTMENT FACTOR OF RELIABILITY, al,

(From Source 1)
Reliability .
_ __ _ _ _Ln. a1
S90 LIO I
95 Ls 0.62
96 L4 0.53
97 L3 0.44
98 L2 0.33
* 99 L1 0.21
The life adjustment factor for material, a2, has not been quantified on
the basis of material characteristics but rather on test results and bearing
applications. In general an a2 value of one applies. However, a value
greater than one may apply to bearings made of steel of low impurity content
"or of special analysis. Values of a2 should be obtained from the bearing
manufacturer or from Zaretsky (1971).
Operating cunditions which remain to be taken into account are the

adequacy of the lubrication (at the operating speed and temperature) and
conditions causing changes in material properties (i.e., high temperature
causing reduced hardness). The influence on bearing life of such conditions
may be considered by the application of a life adjustment factor a3. h
"The calculation of basic dynamic load rating and basic rating life
assumes that bearing life is limited principally by sub-surface fatigue, i.e.,
that the rolling elements and the ring (washer) raceways are sufficiently
separated by a lubricant to make the probability of failures caused by surface e
distress negligible. Where this requirement if fulfilled, a3 11, provided
a lower value does not apply, for example, because of a change in material
properties caused by the operating conditions.
Reduction of a 3 values should be considered whenever the viscosity of

the l1bricant is less than 13 mm2 /s (1 mm2 /s - 1 cST) for ball bearings or
20 mm /s for roller bearings at the operating temperature and/or where the
rotational speed is exceptionally low (revolutions per minute times pitch dia.
in mm less than 10,000). Values of a3 greater than 1 may be considered only
where the lubrication conditions are particularly favorable.
In most cases, discussions with the bearing manufacturer regarding the - -

specifics of the application will help in quantifying a value for a3.
Carter (1972) should also be reviewed for guidance. Zfi
4.9 Reliability Prediction Method - Spur Gear Systems
Source: Savage, M., C.A. Paridon, and J.J. Coy, "Reliability

Model for Planetary Gear Trains," U.S. Army Aviation
Research and Development Command, AVRADCOM TR 82-C-6.
"4-23
. - . .....
4.9.1 Purpose of the Method. To estimate the gear system life which
rill result in a 90 percent probability of survival.
4.9.2 Description of the Method and Example. In the design of a -.

transmission to carry power, high strength alloy steels are normally used in %
%_,-
key elements to help minimize the transmission's size for a given power and
speed rating. AG a result, the endurance limit of soft ductile steels is
replaced by a higher capacity which gradually decreases with the load cycle
count. Key elements, such as bearings and gears are designed on a life basis
in order to keep their size reasonable.
This finite life design for lives greater than 107 load cycles is *•i- 7
common practice in the design of bearings. It is the intent of this approach
to extend the Weibull reliability, life and load theory to the gears as well
as the bearings and to combine the component lives in a consistent fashion in
order to predict the transmission reliability and life as a function of the
applied load and its critical component capabilities.
Although this theory will apply just as well to a simple gear reduction
or any transmission composed of bearings and gears, the presented example will
be for the planetary reduction in the main rotor box of a light helicopter. A
schematic diagram of a three planet reduction with the ring gear fixed and the
sun gear as input is shown in Figure 4.9-1.
In this transmission the overall ratio is 5:1 and the output planetary
spider or arm is to be rotated at 300 RPM. The power transmitted by the
transmission is to be 200 horsepower (150 kw). The input speed of the sun
gear is 1500 RPM. The gea;s are all 206 full depth AGMA toothed gears with a
diametrical pitch of 6 in-" (Ng/Dg) (a module of 4.23 mm). They are all %
made of case hardened A151 9310 vacuum arc remelt steel with a material
constant of 21,000 psi (144 MPa). The sun gear has 24 teeth and a 4 inch
pitch diameter (102 mm). The planet gears have 36 teeth each and a 6 inch
(153 mm) pitch diameter. And the ring gear which is internal has 96 teeth and
a pitch diameter of 16 inches (466 mm). All the gears have a 0.725 inch face
width (18.4 mm).
The planet bearings are the other key elements in the transmission.
These bearings are 75-02 single row cylindrical roller bearings with a 1 inch
width (25 mm) and an outside diameter of 5 1/8 inches (130 mm). These r -
bearings have a nominal basic dynamic capacity of 18,200 pounds (81 kn) each.
Since the transmission is isolated from external side and thrust loads by
outside bearings, these three bearings and five gears comprise the critical
elements in Lhe transmission. It is assumed that their loading is
sufficiently light to prevent early tooth rupture or bearing brinelling. It
is assumed that the life of each component is based on Hertzian stress pitting
fatigue and that the strength in this mode is continually reduced with load
cycles.
In order to combine the reliabilities of the transmission components into

a consistent system reliability, all component load cycles will be reflected
4-24
--- ' ". .. . - "-- - - - -.- -"-- " ill'...-.-. •.

4. '
vs
**'*'
I"
I.
I. '-
Figure 4.9-1. Planetary Gear Reduction
* *1.1
4-25
------------------------------------------------------- . . - .
- .-
into a common counting basis of input sun rotations.

kinematics.
This requires a little m
For a planetary gear train, the number of relative rotations of the plant
gear with respect to the planet spider or arm in terms of input sun rotations
is: K
SPA + R)
ýR 5S
where the subscripts denote the respective gears: S - sun, R ring and P-
planet. The R's represent the respective gear radii. Thus:
-2(8)
PE) (28 e0 -0.533e
The negative sign Indicates rotation in the opposite direction to

GS. Since the speeds are proportional to the number of revolutions
-0.533 (1500) - -800 RPM.

-PA ~~~r
Since the loads on the gears and bearings are stationary with respect to
the planetary spider or armi, we are also Interested In the number of relative
rotations of the suul gear and the ring gear with reupect to the arm in terms ..
of input sun rotations:
S/A RS -fRR2 +88 OS
0 - 0.80 13 e
S/A S
WS/ - 0.8 (1500) -1200 RPM
and
RS2
6
)R/A - -0.2 9S
OR/A ' -0.2 (1500) -- 300 RPM
The forces on the components can be found from the power and input speed.
Power 632 lb- InRPM'
I H
T 200-i
T1 50 (63025) -8403 lb-i
4-26
where T, is the total input torque on the sun gear.
As shown in Figure 4.9-2, this torque produces equal tangential tooth

loads FT and a planet bearing load of twice this value, NO
8403 r
FT a TI
T F ibs 1400.5
FB 2F - 2801 lbs
This assumes equal load sharing among the planets and no dynamic loading in
the gear meshes.
Given these loads, one can determine the L 10 lines and effective
dynamic capacities of the five components in terms of their own load cycle
counts. The two basic relationships for each element are:
In
[n t )
where S is the reliability of the component for 9 load cycles and tlO is
the number of cycles at this load for which the tomponeat has a reliability of
90 percent and E is the Weibull shape parameter for this reliability
distribution. Normally E is taken as 1.2 for roller bearings (it may be as
high as 1.5 for tapered roller bearings) and as 2.5 for gears based on testing
at NASA Lewis Research Center. The second relationship is that for basic
dynamic capacity, or:
'L0 . (C')
where C is the basic dynamic capacity of the component, or the load at which
90 percent of the units will last for 106 load cycles. Here F is the
applied load, X10 is the corresponding 90% reliability life and p is the
load life factor. The exponent p is normally taken as 10/3 for roller
bearings and the NASA Lewis Research Center tests for gears indicates that 4.3
"is an appropriate value for gears. The dynamic capacity equation is often
modified for bearings as
Cl O
•" £10 VF(
ZA
' where a is a factor used to increase the life estimate for improved material
properties due to a reduction in impurities of the roller and race materials.
According to the Roller Elements Committee of the lubrication comittee of the
ASME, life improvements of from 3 to 8 times are not uncommon. This factor
4-27 4.
po
S.'.'."
. ,."'"
. . '.,'".".
. . . .
. ,""-,
. . . "."
. . . -. -. '..',,. '.V....''""."'""
...---- ","-;N "-. -"". '".-.. - .'-.. -"--. ." ,'
.- AI
:V. P.;
2Fb
ipVN
4-2
'a' may also be used to derate the life of the bearing if its speed of
RPM-mm) or extremely low (DN <
•peration is extremely high (DN > 400,000 is inadequate. For this
It,000 RPM-mm) or if the lubrication to the bearing
application the roller speed is
DN miD + 130 22aD 800 RPM
DN • 82,000 RPM * mu
Although helicopter DIN•
as the operating speed should not be a problem.
in their transmissions and
manufacturers use improved steels and lubrication
order of 8 in their calculations, we
often use life improvement factors in the
1/1.5 or 0.67 to indicate
chose, in this example, to use a factor of
on the bearing (this is equivalent to a load
unsureness of the dynamic loading the
factor of 1.13). At best, one can say that this factor is conservative, but
reasonable application,
1972 AFBMA standard recommends a factor of 3 for a factor,
In addition to the life adjustment
disclaimers are also present. value of 1.2 is used since the
there is also a load adjustment factor, V. The
higher stresses than the
counterformal contact on the inner race produces
choice of not using this factor
conformed contact on the outer race does. The
life improvement factor. However
may be justified in the wide range of the
one has the load cycling on the inner
tor two identical bearings for which
cycling on the outer race, the first
race and the second has the same load
higher stress state.
bearing will fail first due to its
'BlO life in bearing
With all this under advisement, the component
load cycles of a single planet bearing is:
B10
...
2'BI = 184.8 X 106 cycles
related to .
its a~ similar fashion, the dynamic capacity of a gear tooth is
ot and i the pagni

lptch';
wnere CTis tedynamic caaiyor
Spoint load on that tooth. From tests on a particular gear material (AISI 9310 .;..-
re ated to*
aaiyo ertohi
Inasmlrfsin h yai
B10 7. 2,1
' ","
.'4-29
. . .. . . .. . . ....... ,....... ,......

... . .. . ..
,* .*.'. . . .'..".._....'.".3 "..."....,.
'.
... ... ,.•.g...•-....°.. ... °. . . .*.- -..- ,..--...-..".
:.'.•',k•.'.,'~~~~~~~.
Vacuum Arc Remelt Steel) at over 9,000 ft/mi pitch line velocity, the dynamic
capacity, CT, is given by:
CT= BI F sin -7.
"R1 R2
where B1 is the material constant determined by test to be 21,000 psi, F i3
the face width of the gear, # is the pitch line pressure angle and R1 and
R2 are the pitch radii of the two gears.
I
For the sun-planet mesh,
(21,000) (.725) sin (20")

S 1. 1
*2.0 3.0,-
CS 6,250 lbs
For the ring-planet mesh,

sin (20")
CR * (21,000)
R1 1..(.725)
3 8.0
C w 25,000 lbs.
R
Unfortunately, the wide range of data available for bearing lives is not Y
matched for gears. Since it is on the basis of this data that the life
adjustment factors are established, corresponding factors do not exist in
gearing. More statistical gear life data is really needed for gearing. The
direct application of the NASA Lewis Research Center gear test data outside
the load cycle range of the tests (1.2-3 X 107 cycles) appears to be
conservative.
I.
For a sun gear tooth and one side of a planet gear tooth
( 6250 )4.3 i0 6 620 X 106 load
For a ring gear tooth and the other side of the planet gear tooth
tLR0 (25000) 106 - 241 X 10' load cycles.
The next step is to reflect these lines and capacities to a common

counting basis, input sun torque and rotation.
k
For a single bearing
L BIO t
S P/A
.4-30
...... ....
L-"-..- '. .'-
'- .- , ,'" ' . " '. ..........
'"'. .'' .......
-. . ,..,-. .
.-....... . ..
, ,.,.. ... -. ..
','.."-'."...."..•.-. .," "-".."-."....'.",....-. ". .*'
LB0 * .• £BIO = (184.8 X 106 ,
SL 10 - 347 X 106 Revolutions
B10 - 3 X 10 - 3856 hours
torque for which this

and its dynamic capacigy as an input torque is the
bearing life equals 10 sun rotations
+R 1/p sei"
where• .' . -
is the modified bearing dynamic capacity

.1..)/3.33,
(.(1)8133 13,430 lb
and
133 13(2)13430). .''"
D3 48,601 lb-in.
teeth. This is
For the sun gear we must combine the lives of the individual
done by the product of probabilities of independent events:
sun
- ST where S and ST are the reliabilities of the
gear and a single sun gear tooth, respectively.
in iN
Thus,
S S( r
and % , .
e°
In - L n (
S [
4-31
%.....
S.o . -.. . • -o . o + o . .
'- .- ,- " .'% i',,'-

•' .%"i%
"° "" " "-.'-.'","
"-.'-." , ,".'".'", -. '-."" " " "" "" '' "
where LS is the life of the individual tooth in terms of its ovn load
cycles, and eG is the Weibull shape parameter. In terms of rotations,
counting contacts with each of n planets per revolutions, this becomes:
n9sl
Ls = (a
S/ L S
or
tS "3(0.8) LS .2.4 LS
so each tooth receives 2.4 load cycles for each sun rotation. In terms of
LIO lives
I l ] I~G 1 T 24LeG "
- [ N- n 1 - L
S nsloJ Slo
or
(sl\ LSIO
lie G/c E.1 - I/"
1~/2.5 620 X 10
6
= " ~2.4 --. "
L SI0 72.5 X 106 Sun Rotations
or
6 rl1
LSI0 -72.5 X 10 1500 X 60 hours
LSlO = 806 hours.
Its dynamic capacity can be found from the tooth capacity in a similar
faRhion:
Ds - L1 (nR"RsCs)
1/2.5(_.3) 0 ]1/4.3
D ,L(- ) (3" 2 6250)
S 24
DS - 22,762 lb-in. •
4-32
r
- . o .. .. + o o ,- + . , *. o o. • •. * . . . .. . . • . . -. . . .
. . .. . . . . . . . .. .. . . .. . . . . . . . . . . . . . . . ..
ZOO.
For the ring gear, the calculations are similar to those for the sun
gear. The reliability of the gear in terms of its teeth reliabilities are:
NR
S -S
R T
*" so
in ( -N In (
and
in (-"o) - NS [
where LR is the life of the individual tooth in terms of its own load
cycles. In terms of sun rotations, counting contacts with each of n planets
per revolution, this becomes:
n R-/-A L,
-T -
""R - LR
S
or
L R 3(0.2) L 0.6 L
R Rt R
so each tooth receives 0.6 load cycles for each sun rotation. In terms of
LI 0 lives:
"n
°n In( ["°4 ] L
1)G
or
10 \ 0.6 11 X 109)
LRI0 - 64,710 X 106 Sun Rotations
or
LRIO -64,110 X 106 L10 60] hours *719,000 hours.
4-33
-il :i
ii-L~l-
.2 "." - " *
.. -? -- -.. . . .- .'. ' .. "".'..v "-..-'.."-...
." -. ".- ".",.'. -- . . .". -" .'...-.. .
Its dynamic capacity can be found from the tooth capacity in a similar
fashion. A
"D I P + R i/P
D - RR 00 S-SCR
R
D) (2)]/. (3(2)(25,000))
For the planet gears, the fact that each tooth is loaded on one side by
the sun gear and on the other by the ring gear changes the calculation
slightly.
The numbers of load cycles that each tooth sees from either sun gear or
planet gear is the number of relative rotations of the planet with respect to
the arm. In terms of sun gear rotations, this is
Lp L
P - P
P/A S
p Lp -0.533 Lp .
S
The reliability of a planet gear is the product of the reliabilities of

its individual tooth faces:
-
s.Sp Np S. p
P PS PR
where Sps is the reliability of a planet tooth face meshing with the sun and
SpR s the reliability of a planet tooth face meshing with the ring. Thus:
In -N In + N... '.-n'
... :-q
or
te. /.533 L eG .533L eG

N
5P-0(S10 p RIO
The single tooth face lives LSlO and 'RIO are the same as those of the
mating teeth on the sun and ring gears. So:
4-34
.. -- " -.. , ".
.
,:, ..-.. . . . . . . . . . . . . . ,..,.. . . .
. ,_ ......-.... . . . .. ,_. ... ... . . . . .. .
..- ..-..
'--• " '".' "- ' .:_,_",',
li 1'lOR ________
1 G
Plo
p L~~~Rio eG+LSoeG
6 9'
Ll
L
l
36l/2.5 [(620
L
X 10 .533J
)(241 X 10 )1 x
(620 X 10 * (241 X 109)2.5
L = 277 X 106 Sun Revolutions

Plo
L0 277 x 106 1500(60) 3078 hours

S-.150(0
The basic dynamic capacity for a single planet gear is the input torque
for which its life equals 106 sun rotations:
D l/e PG RpS G n RSeICS CR
Dp N' e e • ,......
• , .4,- -° -.
L30.J300(8 3
3(2) (6250) (25,000)

X2.5(4.3) 2.5(4.3) l/(2.5)4.3
[(6250) + (25,000)
Dp 0 31,100 lb-in.
At this point all the components are rated for 90% reliability life and
basic dynamic capacity in terms of .un rotations.
Component 106 Sun Lifa Dynamic Capacity
Rotations hrs lb-in
Plant Bearing 347 3,856 48,661
Sun 72.5 806 22,762
Ring 64,710 719,000 110,500
Planet Gear 277 3,078 31,100
The combination of these lives and capacities involves the product of the
probabilities of survival of all the components
4-35
-.. --.. . . . .. .. .• •.-.. °-
" ~~~~~~~~~~..,,....•.-.,,..-•,,..•,,..,
---- '•",*'• - •....
..-. ,.,,.,.:.....,-
.. .- ,. ., . *2'.-.*-.
Ný
S,T "B
nSnSSsP nlS R
or
ST B1 + l
L eG LT e -
Since eG does not equal eB, this relation cannot be directly set to: "
LT e'T .,.
in ( ) - in (s-) L...-. S LS
T ~ TI
However, for values of ST from 0.5 to 0.95, a least squares fit can be made
on Weibull paper to find the values of eT and Lrl0 which best characterize
the system.
For the data of this example;
LT1o - 58.1 X 106 Sun Revolutions L-

L 16- 58.1 X [0 646 hours
and the Weibull slope for the system is
eT - 2.12.
At similar situation exists for the system's dynamic capacity. To find:
t) T T
TI
one can take the equation for Ll0 and vary the input torque over a range of
0.1 D to DT and find the corresponding LT10 lives. A least square fit
of this LTo vs. Ti data on log-log paper will produce a linear curve for
which
Dr 22,605 lb-in c.
and
eT = 4.03.
4-36
...........
...:.. ._- ... - : • . -..." , ,;, - ,.• : =. - ,..,S+ ._

_ - • . = _.• • . .- . • • • . . .. .. S '
"So a system Weibull and load life model for this example is
- In [L 2.12
In In
and
)4.03
TO (IL6.
4.10 Reliability Prediction Method - Minimum Information

The methods of this section are to be employed when there is not
sufficient statistical data nor sufficient structural/analytical information
"concerning the nonelectronic part to allow the use of the other methods of
this notebook to obtain reliability predictions.
The circumstances which lead to the necessity of using this section are " -•
.. partly due to rapid advancement of technology. New parts are constantly being
introduced with hardly enough lifetime history to allow the vendor to set a
warranty or service life (also called useful life). This service life is
usually available from the vendor and can bc used in conjunction with data
from similar devices to provide reliability predictions.
The following examples illustrate the use of this type of information in

determining Weibull parameters and failure rates. In the Weibull case, two
quantities (usually the shape parameter or "slope" as it is often called, and
the vendor supplied service life) are used. In the constant failure rate case
(exponential distribution), the service life is sufficient.
4.10.1 Examples
Example I. The expected service life of a hydraulic motor has been

calculated by the vendor to be 12,413 hours where the service life is defined
"" as the minimum life expected without failure of the sotor section exclusive of
" the bearing section. The bearing section has been calculated to have an L 0 .
life of 50,000 hours in this application, i.e., 90% probability of surviving
50.000 hours. The preceding constitutes the only information available to be
* reliability analyst.
"In the following calculation,the motor section service life estimate

"will be conservatively assumed to be the tenth pecentile of failure, LIo, of
a Weibull distribution. The slope, BH, is assumed to be 2, which is
consistent with the scatter in lives to be expected when fatigue is the
dominant failure mechanism. The calculated bearing section Llo is reduced
to 25,000 hours to account for less than ideal lubrication. The distribution
is assumed to be Weibull with a slope, BB, of 1.5 which is consistent with
practice.
4-37
,...................... .......
S....... .-....--- ,', ,_.--........ -. -- =. . S - .'. ." ." . .- ¢ ...-. *...'.. -,.-.*-. .-- -. -.... ,.
The characteristic lives of the motor section, aM can now be

computed. Since the L1O life of the motor is 12,413, it must be that
.90 - exp -(LlJO/ M)
or
exp - (12413/a 2
.90
so that aM (12,413)/(-ln (.90))1/2 38,241 hours .
Similarly, aB (25,000)/(-ln (.90)1/1.5) - 112,070 hours
Finally, an average failure rate for a use life, t, is calculated using

an average Weibull competing risk cumulative hazard model:
t t-,
where: ,• ..
t - 10,000 hours am - 2
a. - 38,241 8
B - 1.5.
aB - 112,070
110,000 - [100002 + (112070) 9.5
or 9.5 failures per million hours.
Should this failure rate be unacceptably high, a lower failure rate can be
obtained by reducing the in-use life or by redesign of the motor section to -
obtain a greater LlO life rating.
i. Example 2. Ball Screw Reliability Estimation
A ball screw is to be applied in an environment which includes vibration

and salt-laden air. Vendor catalog information provides an L 1 0 life of 20 x
106 inches under the conditions of loading and lubrication of this
* application. Discussions with the vendor's engineering staff suggest that the
given L 1 0 is realistic provided a life correction factor of 0.5 is used to
account for the special environmental conditions. The vendor also states that
test data Indicates that the lives will follow the Weibull distribution with a
8 of 3 when grease lubricated as in this application. In operation there
will be 60 inches of travel per cycle and 13 cycles per hour.
-.
.
~~~4
-38...-'"
L°. . -V .
._..- ..-..--...-.--.-....-.-.-..-..-. . ..-. v..-.-v....-..--...-...-..-.--..'.-.-.".-.-..--.-'.i.-..----.-7-.. ,

0,7
Z.ý-
Estimate the reliability for a service life of 10,000 hours.
Life Lio - 20 x I06 x 0.5 inches 107 inches.
cyc/hr - 13
in/cyc - 60
Weibull Distribution-1-- I B - 3
Service Life (t) , 10,000 hours.
Convert LI 0 inches to LI 0 hours.
7
L hrs (13) (60) = 12,821 hours.
10 .'1.60
Since 8 3, and since the L life satisfies .90- exp

1 3
(L/), it follows that - (12821)/(-ln(.9))I . 27,145
hours. Thus, the reliability for t 10,000 hours is

1 R (10,000) .
exp 1- (l000O/2745)} - 0.951.
If an average failure rate over the 10,000 hours service life is 7

required, it can be calculated as follows:
"= •~~~(t) - -, :-.

*t
2--
1(i0,000) - (10,000) 5.0 x 10-6
3J
(27145)3
or 5 failures per million hours.
Example 3. Estimation of a Constant Failure Rate Based on Service Life.
"As demonstrated in this Notebook, most of the time it can be assumed that
the nonelectronic parts represented in this Notebook have constant failure ..-
rates. In this case, if the only information available is a vendor supplied
"warranty" or "service" life, then a failure rate is easily estimated.
Occasionally, the vendor will indicate at which percentile the warranty

*" is developed. That is, if a warranty or service life of 1000 hours is
specified, the vendor may have done life testing which indicates that 1000
hours is the lifE beyond which 90% of the devices will survive on the
4-39
"'-- .
UN
average. This would make 1000 hours the 10th percentile of the life
distribution. To compute the constant failure rate, simply set 0.90=exp(-x
1000) and solve for X. In this case, X U - ln(O.90)/1000 a 0.000105 or
105 failures per million hours. In general, use the following formula:
failure rate = -ln(l-p)/(warranty time) N
where p is the quantile associated with the warranty time. Usually, p- 0 .10
but occasionally, if the vendor information is suspect, or the vendor will not
*.- say what value p should be, use p-0.05.
VI,
** o*..
. .. . . . . . .. .. .. . . . . .- . ,.-+
4-40
,,"-.71°
5.0 PART FAILURE CRARACTERISTICS
5.1 Introduction. The followng sections of the notebook describe

the analyses of the nonelectronic àrt failure data collected for this
study. Section 5.2 presents the results of fitting the exponential
distribution to the failure data for each part class, part type, and
environment. Sections 5.3 - 5.5 present the results of testing the fit
of the exponential distribution against Weibull alternatives for the
data corresponding to some items. Section 5.6 gives part malfunction
data and frequency of occurrence for some parts. It should be pointed
out that the data used for preparation of this notebook was screened
to exclude secondary failures and failures caused by maintenance personnel.
environment abbreviations in the tables of this section follow SThe

the conventions of MIL-RDBK-217D. For convenience, these abbreviations
are defined in table 5.1.1. More detailed descriptions may be found
in MIL-HDBK-217D.
Finally, an explanation of the confidence intervals presented in
sections 5.2 through 5.5 is necessary. These confidence intervals (for
failure rate in the exponential cases, and for the shape and scale
parameters in the Weibull analyses) have been called 160% confidence
intervals." In the tables, however, the lower and upper bounds are
labeled '80% lower" and *80% upper" bounds. This has been done so that
either one-sided bound can be used by itself to form a one-sided 80%
confidence interval if desired. When the two 80% bounds are combined
to form a tu'-sided interval, the resultant confidence is 60%.
TABLE 5.1.1 DEFINITIONS OF ENVIRONMET ABBREVIATIONS
Abbreviation Environment
AIF Airborne, Inhabited, Fighter

AIT Airborne, Inhabited, Transport
ARW Airborne, Rotary Winged
AUF Airborne, Uninhabited, Fighter L
AUT Airborne, Uninhabited, Transport
GB Ground, Benign
GF Ground, Fixed
GM Ground, Mobile
ML Missile, Launch
NS Naval, Sheltered
NSB Naval, Submarine
-NU Naval, Unsheltered
5.1.1 Use of Constant Failure Rate Analyses. We recommend that the

results presented in Section 5.2 be used as follows. First, find the
* 5-
part class, part type, and environment of interest in the tables listed
in section 5.2. The corresponding table will give a point estimate (also
referred as a "prediction" in many reliability circles) of failure rate
per million hours. A two-sided 60% confidence bound on the failure rate
is also given in order to give the user a feel for the precision of the
estimate. Also included is the number of independent sources (usually
projects) which contributed to the estimates, along with total number
of failures, total part operating hours and an estimate of mean life
(i.e., mean operating time to failure). In a large number of cases,
less than 50,000 hours of operating time were available so that care
should be taken to examine the width of these confidence intervals.
In cases where the total part operating hours shown are less than 1,000
hours, failure rate information is not tabulated. For these cases, the
user should be very cautious in using the information presented for reliabil-
ity purposes. Wherever there is more than one source contributing to
an estimate, the observed significance level of a statistical test of
homogeneity (also called a p-value, see Cox & Hinkley (1974), p. 66,
for further discussion) is given. This test of homogeneity was performed
in order to determine if the sources reporting failures were statistically
different. In general, the observed significance level is between 0
and 1 with values close to 0 indicating evidence to reject homogeneity.
We recommend a threshold of 0.05 for the homogeneity test, i.e.,
homogeneity is rejected if the observed significance level is below 0.05.
5.1.2 Use of Weibull Analyses. Most data collected was restricted

to total operating time and total number of failures. While this approach
is adequate (sufficient, in fact, in the statistical sense) when dealing
with the exponential distribution, it does not provide a means of evaluating
the "fit" or validity of the exponential model. The validity of the
exponential distribution for describing the life distribution for nonelec-
tronic parts was one of the central issues addressed by this study.
In order to address this important issue, data sets that contained

actual part lifetimes for each failed part and total operating time were
collected. In most cases, part lifetime data simply did not exist.
However, for a significant number of part classes, part types, and
environments, these data were available and were used to test the fit
of the exponential distribution against Weibull alternatives. The Weibull
family of distributions is rich enough to approximate virtually any
unimodal life distribution and is therefore applicable for most
nonelectronic parts. Moreover, the Weibull distribution is the resultant
extreme-type distribution for describing lifetimes of nonelectronic
parts which fail in accordance with a "weakest link" scenario. This
technique for testing the fit of a distribution by embedding it in a
parametric family of distributions is called a "smooth goodness-of-fit
test," and is described in further detail in Lawless (1962), p. 438.
Since the lifetime data was not collected under any of the commonly
treated sampling plans, i.e., nonreplacement type I or II censoring,
or complete samples, the goodness-of-fit procedure had to be developed
ad hoc, and the smooth goodness-of-fit approach allowed the use of standard
5-2
likelihood ratio procedures in performing the test. The results of

these analyses are presented in sections 5.3 - 5.5.
If the same part class, part type, and environment is reported in

sections 5.3 - 5.5 a Weibull analysis was performed (this is indicated
in the index to Section 5.2). Use the information in the Weibull table
to decide whether to adopt the Weibull distribution, or retain the
exponential distribution. For each of the part classes, part types and
environments analyzed in sections 5.3 - 5.5, a table summarizing the
results is presented. Each table contains i point estimate of mean life
(i.e. mean time to failure in hours), and point and 60% confidence
interval estimates for the Weibull scale parameter (in hours) and shape
parameter (unitless). Also included are total part operating hours, and
total failures. In cases where there was a predominant failure mode,
this failure mode is given in the comments field, along with the
observed significance level for the test of exponentiality. The
observed significance level for testing exponentiality is used to decide
whether to adopt th- Weibull model, or to retain the exponential model.
As before, we recommend that the threshold value be 0.05, i.e. reject
exponentiality if the observed significance level is less than 0.05, and
retain exponentiality otherwise. However, depending on the particular
circumstances, the analyst using the observed significance level may
wish to base the decision on a different threshold value, e.g. 0.10,
0.005, 0.001, etc. In the majority of cases, the exponential model is
shown to be the best fit based on the 0.05 level of significance.
If the Weibull model is shown to be the better fitting model, it

may still be desirable to approximate the distribution by the
exponential distribution. This approximation is useful if the
nonelectronic part(s) under consideration are part of a large system in
which the other elements exhibit exponentially distributed lifetimes,
and it is necessary to analyze the system as a whole. For exponential
approximations it is recommended that the point estimate of the mean for
the Weibull be used in the exponential model whenever the Weibull shape
parameter is greater than one. When the Weibull shape parameter is less
than one, use the mean life estimate from the appropriate table in
section 5.2. These guidelines will yield conservative results (i.e.
lower bounds) when computing system reliability for the series string in
the case where the Weibull shape parameter is greater than one.
A total of 145 part classes/ types were analyzed in sections 5.3 -

5.5. In 6 of those cases exponentiality would be rejected at the 0.05
level of significance. This is not statistically significant. These
results suggest that the exponential distribution is an adequate life
model for most nonelectronic parts that are operated for time periods
smaller than those analyzed in sections 5.3 - 5.5, i.e., there is a time
period for most nonelectronic parts during which a constant failure rate
model is appropriate. This phenomenon, conjectured in previous editions
of this notebook, is supported by the results of sections 5.3 - 5.5.
5-3
5.2 Constant Failure Rate Analysis
5.2.1 Index to Section 5.2
Part Class Part Type Identification Number*
ACCELEROMETER FORCED BALANCED 1*

ACCELEROMETER PENDULUM, LINEAR 2
ACCELEROMETER PENDULUM, SINGLE AXIS 3
ACCUMULATOR HYDRAULIC 4
ACCUMULATOR HYDRAULIC-PNEUMATIC 5
ACTUATOR ELECTRICAL 6
ACTUATOR ELECTROMECHANICAL (LINEAR) 7*
ACTUATOR ELECTROMECHANICAL (ROTARY) 8 -
ACTUATOR ELECTROMECHANICAL (LINEAR) 9*
ACTUATOR FYDRAULIC-PNEUMATIC 10
ACTUATOR MECHANICAL 11
ACTUATOR ROTARY 12
AIR CONDITIONER COMFORT 13
AIR CONDITIONER GENERAL 14
AIR CONDITIONER PROCESS 15
ANTENNA COMMUNICATION 16*
ANTENNA MICROWAVE (COMMUNICATION) 17
ANTENNA RADAR 18
AXLE GENERAL 19*
AZIMUTH ENCODER OPTICAL 20*
BATTERY RECHARGEABLE 21*
BEARING BALL 22*
BEARING ROLLER 23*
BEARING SLEEVE 24*
BEARING NUT GENERAL 25
BELLOWS GENERAL 26*
BELT GEARED 27
BELT TIMING 28*
BELT V-BELT 29*
BINOCULAR NITROGEN PRESSURIZED 30
BLADE ASSEMBLY GENERAL 31
BLOWERS 6 FANS AXIAL 32*
BLOWERS & FANS CENTRIFUGAL 33*
BOOT (DUST & MOISTURE) GENERAL 34
BRAKE ELECTROMECHANICAL 35*
BRUSHES ELECTRIC MOTOR 36*
BURNER CATALYTIC 37
BUSHINGS GENERAL 38*
CAM GENERAL 39
CAMERA MOTION (TV) 40*
CESIUM BEAM TUBE GENERAL 41
CIRCUIT PROTECTION
DEVICE SPARK GAP 42*
CIRCUIT PROTECTION
*Note: An asterisk "*" indicates that a Weibull analysis is included

in one or more environments in Sections 5.3i - 5.5.
5-4
DEVICE SURGE ARRESTER 43*

CLUTCH FRICTION 44*
CLUTCH GENERAL 45
COMPRESSOR GENERAL 46
COMPRESSOR HIGH PRESSURE 47
COMPRESSOR LOW PRESSURE 48
COMPUTER MASS MEMORY FIXED HEAD DISK 49*
COMPUTER MASS MEMORY MAGNETIC TAPE 50*
COMPUTER MASS MEMORY MOVABLE HEAD DISK 51*
CONTROL TUBE ASSEMBLY GENERAL 52
CORD/CABLE GENERAL 53
COUNTER ANALOG 54
COUNTER DIGITAL 55
COUNTER MECHANICAL 56
COUNTER WATER CLOCK 57
COUPLING FLEXIBLE 58*
COUPLING FLUID 59
COUPLING GENERAL 60
COUPLING RIGID 61*
CRANKSHAFT GENERAL 62*
CROSS HEAD GENERAL 63
DIAPHRAGMS BURST GENERAL 64*
DIFFUSER GENERAL 65
DISC ASSLMBLY GENERAL 66
DISTILLATION UNIT FROM DISTILLING PLANT 67
DRIVE GEAR 68*
DRIVE GENERAL 69
DRIVE VARIABLE PITCH 70*
DRIVE FOR COMPUTER
TAPES & DISCS CAPSTAN MOTOR 71
DRIVE FOR COMPUTER
TAPES & DISCS DISCS 72*
DRIVE FOR COMPUTER
TAPES & DISCS MAGNETIC TAPE TRANSPORT 73*
DRIVE FOR COMPUTER
TAPES & DISCS REEL MOTOR 74
DRIVE ROD GENERAL 75
DRUM GENERAL 76*
DUCT GENERAL 77*
ELECTRIC HEATERS RESISTANCE 78*
ELECTROMECHANICAL
TIMERS GENERAL 79*
ENGINES GENERAL 80
FEEDHORN WAVEGUIDE 81-
FILTER GAS (AIR) 82*
FILTER LIQUID 83*

in one or more environments in Sections 5.3 - 5.5.
5-5
%5-5.
(:.F
PArt Class Part Type Identification Number
FILTER OPTICAL 84
FITTINGS GENERAL 8S
FITTINGS PERMANENT 86*
FITTINGS QUICK DISCONNECT 87*
FITTINGS THREADED 88*
FLASH LAMP GENERAL 89
FUSE HOLDER BLOCK 90*
FUSE HOLDER EXTRACTOR POST 91
FUSE HOLDER PLUG 92*
GAS DRYER DESICATOR MOLECULAR SIEVE 93
GASKETS & SEALS DYNAMIC 94*
GASKETS & SEALS STATIC 95*
GEAR ANTIROTATION 96
GEAR BEVEL 97*
GEAR HELICAL 98*
GEAR HYPOID 99
GEAR SPUR 100
GEAR WORM 101
GEAR BOX MULTIPLIER 102
GEAR BOX REDUCTION 103
GEAR TRAIN BEVEL 104
GENERATOR AC 105
GENERATOR GENERAL (OXYGEN GENERATOR) 106
GLASS (SIGHT GAUGE) GENERAL 107
GROMMET GENERAL 108
GIMBALS GENERAL Ioq
GIMBALS TORQUE 110
GYROSCOPE SINGLE AXIS III
GYROSCOPE TWO AXIS ROTOR 112
HEAT EXCHANGERS COPLATES 113*
HEAT EXCHANGERS GENERAL 114
HEAT EXCHANGERS RADIATOR 115*
HEATER WATER 116
HEATER BLANKETS GENERAL 117
HEATER, FLEX ELEMENT HEATER TAPE 118
HIGH SPEED PRINTER ELECTROSTATIC 119*
HIGH SPEED PRINTER IMPACT 120*
HIGH SPEED PRINTER THERMAL 121
HOSE FLEXIBLE 122
HOSE FLEXIBLE, PROPELLANT 123
HOSE GENERAL 124*
HOUSING GENERAL 125
INCINERATOR FROM SEWAGE TREATMENT 126
INSTRUMENTS AMMETER 127*
INSTRUMENTS FLOW METER 128
INSTRUMENTS HUMIDITY INDICATOR 129
*Note: An asterisk "*" indicates that a Weibull analysis i's included
5-6
INSTRUMENTS INDICATOR 130

INSTRUMENTS INDICATOR (LIGHT) 131
INSTRUMENTS INDICATOR (FLUID LEVEL) 132
INSTRUMENTS PRESSURE GAUGE 133
INSTRUMENTS TIME tZTEt 134
INSTRUMENTS TOTAL TIME METER 135
INSTRUMENTS VOLTMETER 136*
JOINT, MICROWAVE
ROTARY GENERAL 137*
KEYBOARD ELECTROMECHANICAL 138*
KEYBOARD GENERAL 139
KEYBOARD MECHANICAL 140
KHOB GENERAL 141
LAMP XENON 142
LAMP HOLDER GENERAL 143
LENS OPTICAL 144
LOW SPEED PRINTER DOT MATRIX 145*
MANIFOLD GENERAL 146
METAL TUBING GENERAL 147*
MODULES GENERAL 148
MOTOR GENERATOR SET AC 149
MOTOR GENERATOR SET
MOTOR GENERATOR SET
MOTOR, ELECTRIC
MOTOR, ELECTRIC
MOTOR, ELECTRIC
DC
GENERAL
> 1 HORSE POWER, AC
> 10 HORSE POWER, AC
DC
150
151
152*
153
154
i•-
MOTOR, ELECTRIC DC, (4 HORSEPOWER) 155
MOTOR, ELECTRIC HYDRAULIC, DC 156"
MOTOR, ELECTRIC SERVO, DC 157"
MOTOR, ELECTRIC STEPPER 158*
O-RING GENERAL 159
PARTICLE SEPARATOR GENERAL 160 ...
PITCH HORN GENERAL 161,-
PLOTTER ELECTROMECHANICAL 162
POWER CIRCUIT BREAKER CURRENT & VOLTAGE TRIP 163 -
POWER CIRCUIT BREAKER CURRENT TRIP 164*
POWER SWITCH GEAR GENERAL 165*
PRECIPITATOR ELECTROSTATIC 166
PRISM OPTICAL 167
PROPELLER GENERAL (FROM SHIP) 168
PROPORTIONING UNIT FROM DISTILLING PLANT 169
PULLEY GEAR BELT 170*
PULLEY GROOVED 171*
PULLEY V-PULLEY 172*
PUMP CENTRIFUGAL 173
PUMP HYDRAULIC 174*
*Note: An asterisk "*" indicates that a Wetbull analysis is included
5-7
• ......-.-...-. ._.._...,.,?...... .. •• .... .....- _,. .-.--..-...- ,...............,......-..........-.

PUMP HYDRAULIC (ROTARY) 175

PUMP PNEUMATIC 176*
PUMP ROTARY 177
PUMP VACUUM 178*
PUMP VACUUM - LOBE TYPE 179
PUMP VACUUM - RING SEAL TYPE 180
PUR IF IER CENTR IFUGAL 181
QUILL ASSEMBLY GENERAL 182
RADOME MICROWAVE, ANTENNA 183
REFRIGERATION PLANT FROM AIR CONDITIONING PLANT 184
REGULATOR ELECTR ICAL 185
REGULATOR PNEUMATIC (PRESSURE) 186
REGULATOR PNEUMATIC (VACUUM BREAKER) 187
REGULATOR PRESSURE 188
REGULATOR TMPERATURE 189
RESILIENT MOUNT GENERAL 190*
RESILIENT MOUNT SHOCK MOUNTS 191
RETAINING RING GENERAL 192*
SEAL GENERAL 193
SEAL SOLDER 194
SENSORS WATER LEVEL 195
SENSORS/TRANSDUCER/
TRANSMITTER ACOUSTIC (HYDROPIiONES) 196
SENSORS/TRANSDUCER/
TRANSMITTER AIRFLOW 197
SENSORS/TRANSDUCER/
TRANSMITTER FLOW (LIQUID) 198
"I
SENSORS/TRANSDUCER/
TRANSMITTER HUM ID ITY 199
SENSORS/TRANSDUCER/
TRANSMITTER INFRARED 200
SENSORS/TRANSDUCER/
TRANSMITTER MOTION 201
SENSORS/TRANSDUCER/
TRANSMITTER PRESSURE 202*
SENSORS/TRANSDUCER/ -
TRANHSMITTER TEMPERATURE 203*
SHAFT GENERAL 204*
SHOCK ABSORBERS COMBINATION 205*
SHOCK ABSORBERS RESILIENT 206*
SLIP RING-BRUSH POWER & SIGNAL 207
SLIP RINGS GENERAL 208 J65
SOLENOIDS GENERAL 209

SOLENOIDS LINEAR 210*
SOLENOIDS ROTARY 211
SPRING COMPRESSION 212
5-8
o.. o ".
-,oo•o• "o .' . .O~ oo, . •. ~ O. o. . . . . o . , • . o . . o •. . . . . .-.
•. -. --.. " ". , -o • °. " .- . *- q ° . - . ' . . ,. " - . , . o " . • o. , . . . - - , % - .

Part Class Part Type Identification Nuaubr*
SPRING GENERAL 213

.•SPRING TORR IS ION 214
,'S PRO•tET GEN•ERLU 215
STEAMBOILER GENERAL (FROM SHIP) 216
STOW PIN GENERAL 217
SWITCH COAXIAL (ELECTIR0MCHANICAL) 218
SWITCH FLOW (LIQUID) 219"
- SWITCH nIMLOCK 220
SWITCH PRESSURE (AIR FLOW) 221'
SWITCH ROCKER 222*'
SW ITCH THERMOSTATIC 223' K
SWITCH THUMBWEEL 224*
SWITCH WAVE GUIDE 225*
SWITCHBOARD CONTROL FROM OXYGEN GENERATOR 226
*SYNCRO TRANSMITTER 227%
SYNCRO ASSEMBLY GENERAL 228
SYNCRO/RESOLVER LOW SPEED LOW LOAD 229
TACHOMETER GENERAL 230
TANK NON PRESSURIZED 231
TANK PRESSURIZED 232
* TELESCOPE BORES IGHT 233*
TELESCOPE GENERAL 234
TERMINAL BOARDS
THERMOCOUPLE
GENERAL
GENERAL
235
236
.- ..,
- TRACK BALL ELECTROMECHANICAL 237*

TRANSMISS ION GENERAL 238
TRUNNION ASSEMBLY GENERAL 239
VALVE CONTROL-MANUAL 240
SVALVE GAS (AIR-VENT) 241
VLEHYDRAULIC 242*
VALVE PNEUMATIC 243'
VALVE SOLENOID OPERATED 244,
VALVE (ISOLATION) PYROTECHNICALLY ACTUATED VALVE 245
VALVE (RELIEF) PRESSURE ACTUATED 246
VALVE (FILL & DRAIN) HAND OPERATED PLUG VALVE 247
VALVE (BIPROPELLANT-
HIGH THURST) SOLENOID OPERATED 248
VALVE (BIPROPELLANT-
THURST) TORQUE MOTOR OPERATED 249
WASHER FLAT 250
WASHER LOCK 251
WASHER SHERR 252
WASHER SPRING 253
WASHER STAR 254
"WATER DEMINERALIZER MIX-RESIN 255 '.
WINDLASS FROM ANCHOR 256
*Note: An asterisk "*" indicates that a Weibull analysis is included ',
5-9
•+' :•, ::. .. :..".'.,;...:..:::..:. .......

• • ..... '.....,......-.........................,............................-..........._.....
COLE a a
z0:
I LI IL 45I I. 14 0
J..
SL 5 Q I .JW
5.4 I 5.4 5 I 5 5. 5 5 5 5U
'1 5 a
z £~ s a 2 a w w a
wa~
5Z 0.W 1 ~ .a L . - 5 C
~
~~~~~~~ aIE ~ 4 q
--- a 5 I-a j
i4 - 5
Oe~'
U Z It I
5 5 OS5U
ao aP- 11
(h i.
a w M
- ac-.. - .
aM
am 0 5 ao a -.. aina . 0 at ao
aa 00 a J aa a 00 a a
a a
9$A
Cc At
5() S 05.V 5 e s v $.--4 5 *s ca-(A I s
8
wI4 5 - U W- Na L
5 ~ :,c 5 Ufl
z~ a
I~ ~ 5 ~ 4~ 5a 5 55
WaO Wda 5CS5 a ~

I.- 1 0 1 ZO 6- 1 44 a a
w I9 10
X. 1 4 5 0a 1
x a
cc K W
WiZ Z
L S
a
S ~0 5
5
5
~
I
Las~~~~ 5- >U- a
LU
a S
a
5
-- *56
5 5 ~ Lj
L
aa
>--*~
aa
LL.a
a
5-10 ~j
-
ai
t*-a-*I,.-------*
1
CI* C4 a
I to IL
WS. LL a a h. a ZO=
4c ac U) Ii ac to a a
a
a-~~~~- a- a ma a
Z
W aS a' &J a Z W W Ca
-
a
W- S Dil
m
'aat
a a ah Oiaa
~ alaa
09 aa a ~ a hiJ
4 a a
aD 4ca Ia
a~~~~~~~aL aL aa i. a a ~
a a a-L ( a atona a a-
a L
a a aaa
a~ a a A s
-C 0a aL a
A .0~ 0 a
~
a- . 1 0
4~ I~a 0 a a a a
a C a, a -. a.- a ~ 6- -
w a.0 a ar a aA a aaUa , a a
~ a a a a a acao
oia a-.aaa =aa
)0a- * L... :U 2 .j
.ju a ~ aa
a
zaaa
a
0
a
a wa ag awi ag
a a- a a a a aC
Z~ aa a 0 aaa
a.
a~X ~~ ~ ao
agoo.-----a
a aW
Z
O aa
aa ~ .a -
a ~ ~ Ii V
~~~~ ~ ~~ ~~a a a
a a (
aa
---------
'. d-.--a..
a
a a aaa aa a
a
aw i~ a aa ~ a - .
a ~~
~0a auaa ~ ~ ~ ay a ~ a ~ 0
w
~ a~ I I-f
La4c a0 4 m
0a
a5-11- ~ a
*1-4
a ay
aI M aL aL
I* a CLl A p- a- Ia
a a a I a
a a *a a
a "cp :C " a a
* ~~ ~a a
- $0.
wZ a ~~inl
*
a*
P" 00 a -
a I
z a4 ae Ia I
a.. a j a. m ~ I- V
*~U -jaa
I w 0 a a4 aL
L& a
C a-4a
In a 0 ao
w a w aa 4a
a- a a -
a a a6a
a 4-------5- a2
a ~~ ~Sq C I -
a~t 1 N I .
WI U4 (ala C-4Ia a a
*a
a-a N aa
Ia
mC r' a A
a dc a q
a a aa a
a ac a
o i ao
I a a a a ma a
L) a
00a
La na .
o- . ,a .
- I -
-
1 ,4 1 ma
a- a0
w~~~~~~ a4 ~ a m Sa 1. a .J
a~~~ a waa w a a w
II La aL a aL a a a a a
a a aI a a a (# a a a
z -1 1 a a
a
aa a a a (J a a a5a a3
* LZ 1 a 4a aI
de IaL
1
OL I M
a-4
V,
a aA aN
* II IL S a
a -a a maa a aa a a aK
a a m az -
wZ
a-
LA.
ac1 aa a
L"6aaa
E
40 ~ ~*
a
a c (A ac a al aV)ar
2 a 00 a a aia C
m aI
a a a L aa
01a.--a aJ.----
a c a a a a.
L
aI- a-a- aZ a a az Laa.
S4c ac .Ja a a 4 a .J
C) ao af a-4 a0 at a a a a 4C)vl 1a
a-a~o at
a a0a0 At 1
au a 0 a0 a4 a. JdC a, aJ a aac a
*i a" aI a a a Laa
a-~~UL a0 a .n - 6
~
2 ~
a0a ~Ca 0 a a eg
2 at W ha a C
w .M a-aL C ii .hj - a C
*i a a ae
LAS4~lw~~
a a-w.a-~Ca a- a a ILJa aE
wDa-~
n a0
a ~Q a
ra,-
a
--
L
aj aaaa
~ a aa
a
a
a
aa~
*.a ~ 4
4a
U
LOLA~a
LLnf
a j a a4 aC ac a a6a
a u W. a a a a1
ai a a a LJ a a - a L
U a ~ ~- 4 0
a a a a
-
2 ~a ~a aa:
a aaa
aa a a .-
-I. a. aP4 . a a a - aa .
a aL a Z a a ai aI i
4~~
a
a
~~ ~i~~~aa ~ a
a
a0
U~
a L
.--- I
a
a.-a..
aa
4 a
U
..
*.a-
*~~ a a
ma
i
La
~a
x
rZ
(A
a a
) -
a II a-c.~ a a a ~ ac
4i aE @aa ao 4a-a a ' oao
X aWZ a Wa C 0.a-fa a 'a a2a : 0-a 4
C)#~ a :a ~ rn wa ea 5na14 C a
all
zi a 4 m 1 i eI
1 mi I eatI I
Icc (A I I AI 4
OC I f~w
* ..- - s a .
I.
1 - I II I I
IU I .. 4
4 1Z Z e :1
-j I- 1.4 1. 5.
*N ,*5*15
;11 -:A
z o.~n
es aoil
a-- 6Now
o~@-da 1 5,
01S6
.4 ' 5 5 5 5
WI
4c
Ifl 65
J
4sqss
I
*~~~~ z
w
6 Wo 5I 6
S 6 6a
I I 5-6 6
I)ow
IL
3
3H
onIIO
3 3
3
9
33
3 3
3
3 3
C3
3
am
3
3
3 3
ki I LI II- 4 I3
02M3 I I I 32 (A I
I In I~ COP1 03
* *U. W
3 33 ~ 3 ~ ~
:: : 3
: :
I 3 3
:
3
:
I* I 34-33 It
31- 30 34 0 3 S4 *-*
*~401 es n
: Z1494 " :W
W~b3
I dc 1 41 -C 3
3 33 3 1
*1 19 3 41 a
hI5S17
la Ia o, a-d ao -- o I
a CL ai CL I. I a
a = a a aa ca ON
I, La j 1. 04 a a a a
cc " a1 a. a4 a 1- 1
Um O aaa s upO *
I 4 EcCa C S S 66 4 IW *
(a (Aa a
0 10-0.Ja
g~ a 4I
U a a
0
.I a
aa
a a U
Owl* ~ a a a
a a Z a a
a-a
a a a
~w a
a
a a
a
I.
b
a
a a a a
IS
z
Z~a0 w,
w"a .
w
- a 6 C
aa 0 a
S
~ W a-
W a
a 0
C
a~ ~ ~aa ~Cza ~aaaKa ~ a aC2I~ ~ VE aC

a
I ~L W~4-a
a
.
aa saa4-
4a
aa
a~~~~~~~~
a
a a a
~4(l --
S
(l~
5644-.
N
Z
0
a
a
a
C
*
aa
a
Z Wa-
ALa
9L 41C a aa J L.I ~
ax I. a I. - w a a a 5
a~~~~~ a
aI .- aa 5-a Z
aC a9L ~ a a. L) a
*~ a ~ 44 a a -0 a a a a 5-
6 u12
6* ~~~ u u
A I C
M r3 4L
6* * 6 * 6 6 ~AS
*l .4Sw
: :
6i i-- I
6 66 6
43,
Ii S *.u LL
U I L
, 6 ~ 66S 6
hi-J S USh '4- ~ i 6. I *lu* I lau WE Volu
soI i IS O 046 6h 6 JAI

I -- 96 M 6 1 i-C~ 1 0 0 Sh
6 60
i~C- 0e 6 6 1
44- 6 6S
-- 6 K.4-
6c t*6 .- m: 4 z.
6I b.-
43
IV
6~~~a 6 ~ 5 ......
4K
I ~ I ZL K5 O - I45 6
1-
1. 1 1-d-
3 6 u 0 5
66 6 SI
5 6
*0
K z
6 h 6' 66O IL 6 h ~ 6i 6
6 06 6 5 -19 6 S
44 I
L I I
I L
.4I i 1. 1 6. 6
de " 4t I. O ~
ILS A IJ
Ih.- 0. x . *
I-~L1 6 6 -I
.Li6
I W 6 N
I M II 6
*
Z 1W I I
U I
0 z -z; : >:
zrn
391:
4c~ w 4 w w
~ I 5 20
•al fn I I
,:
* ~ ~ I '
~011
t"
0
**
1
w *-40=0 4 *
4
z4
'. . ..
Ju
+ •; --
, MA<+
0"4
*~~ %
.c 44 4 1 a* 4
4 .4 * 40
2~~~-I l
4 94-6 40 4 *J a
eq. Lim
COCO
~~~
i'l ~Ii, •' -
.4
4 4 .. 4
. ....
.
"U.
,44 .<'
C4 of 4
k 4 N-Ott
014h
4 I I se-"_
IN . I . U.
4 . , - -: -::•2
ic
?.. '41 "0'1,On ,A A~
;,,;'.'.
4 04 .4: U0
ZC 14 0
.4 4 4 0
4 m 4~' I I 44 *4
4 .Z1.4
• I'
*~~~~~~~~J *.-n 4 0 4 4
4L -C --
?,0
.I44 4 4 I 4 4- - , . . -oUZI
4* b
4
~
4 .4 4 4 5-21 0 - 4
.4A
JOc
w 0u 0 0
IL I- 0000
I j- 1. - 0" 0 ch~
o o Iw
~'~' : $-AI
I
W
oIl
o
4 0 00
ILi Ij
0w C-CU
- 0~ 0f
U 0 I I U 1-I
IN
O s W
a5-22
W
IA U .a aIa
Ge ,, : 0: I'll.:
1 of: NI va-
a. '- LL Wl 45 I I- Ia
=2
w
asin,:
a6
I
a,
w aL
e I. (A
a011
so I o
a0
l
acý
a
S M- I a
I 49 ac a W
FI e4-wt :a a4
z~I - I.
a ac ina
4c a a aC aI
a Ca .Ja
a a C a a
S a-S .5 >C
a 4L-a-- a aW
-- a~~ 0i4.a
aa1 LaL, i
* ~w~ ~ u .JU L) j a aa~,~
a~ a Z a aZ
a a~ a a 0 a 5-2 a3
cc I n I a a *e *
w S I i LL
US LL .-I
aA aua in f I (g
a Bja I a Iuj * m
aý 09- m P mm,
m -i ý W. m n mj a aA aI
o m am m m a a a Um U
I~ ca
m
a ~~~
m
a a a a m
Z
m
aaLa a a mm m m m*
cc wa---Eaa C a a m m w
.1I
*--. 0in4S m S-i m -J ac m' c.)J
*i ~4mJ
* ~- a1c (na
m aim m
I
Ca
4c
1.m..
m
10 m e
I
w .
in a 1a00 a m m .
a 4 a aa aa 3 a .
09 a a a ( a 4 1-
* a - a m m~ I-
LL a a (ai
Lai p
* a a- a a a a m
4 . - - 3m
az a a (e m
*0 a 0 =
a L a La L a 0 1--. -.- a-
x mI a I
aI o I w a ZVIm a
a U ag a
4K 50 a a
ea ja.
aim ~ m O' m5-2a 4
6 I .4
4
I I LI I W4 I LB f4 IA I
C
I I 4 B BJ I
r4 34 3
IL *1 0 2.
toZ' 2
I* U. 21
B
LU11I U. 3 - U.
4 3 4D : a
44j 44j I- B0
Bn 1 -4 1-
1a W P- 4B
B BIC 4 B IAI ILlI :. a.
wa
B
.4
. . . 4 1 B >0- 1 P4 0-1 P4
C. IW 1 O I W, n I 1).
0 1 B 2 4fB4 3 4 4B 4
in 1 -0 B I I 3
B 91 : a B B . 4.4 Br-a.' I B
3 lU B0 31 U3 B
B oe 44 4to 4 440 4 B
B. 33 4 3 I B
B I to1 B2 0C.)23
~ ~ ~ 4 B ~ B3 (~(~~4
4......'..4.
0 BsI (1 C~
BiCID
*~ ~ ~~ ~4 ~ ~ 3
( :
4
4
AB
4
(4
4
4j~004
4
4 WI 4 4
-S ~~ I
BaN I~ 3
~~~~0~~
NIC4 I1O
C3
~ NB *
Ug 4
' I Bl
~B 4.63(B 4
M
4 4 4 @ 4 0
4 B-
BB ~~~~~ ~C 1
4 4
B 4 I Z I A3 4
4 4 4 4 3
0 ~ ~ ~B ~~~
~uZ***
4
W I
B 4 ~~ 4I 4 3
... z z 4
*.............
0 0* 4 4 14 I1B N 4
B - 1 Co. 0 (4 OI
C94. o . = - LL, *3 w*z4

Z B9 24 4 I 4E I1 B
BD 4I B 4L
I
4 45 -2B5 3 3 4 3
3 B 3 4 4
PIN
'9.9 a a 9 a a m
* 0 0 CDI OD 9 9 -*JUL-
IIA.:
(N.I :CA S
OC2 1 4 4 1 Z z1
"
~~a~~de aý a I
of:L iK
Ia a 9 A t9
1- IL 1
a
40
a a~ 1 0 tsI La
1. V
CNa C
00 ata
a9 9 9 9
- a a ~ a5a-a6
C4I 61- 1-t--1I
40 I I C- UI - , %
at z
Id m I.
*C.
*~~ LL.~ ~ ,
6
5.4~C-SI I I S
I ~~ e eOs ~OLai
I I I IJI hI
.-- IA5. I WI IA
14VIL
~~4~C.
00
I I
cc = , S CO)I II
I soIII
6L. 4 40 4
I0 I I I*4
II I I - 1
I I I I
.j I
* 1 0 0I 0 I tun c . . I
.1 b, I
I 4cfI! *m m
1 * I I-S
Ij I I
I1
Zi 1z I
I .6 1 I.J
.1 * ..
~1I O 6
I ~ WI
*1
I l Ui
..41
W
IfLI
uL
4
4 ~ IX Va
ox I llx
4 4 4141
I A Ii
ISI .41 -SI 54. I-27

I I ).141
I I 5.-5 5.-I 5-I S-I O~a -I.5 -4 fI I

?A I
ow
I a " lo aCI-
a U. i *
t- Ca p
O-o I -= : t-
CL1- 1.a
ca - Lj0 C a
ar . aI Cx
aC aa 00
09 ia
() a a a
~
* w ~ .U - a
IIC aZ .J--a a *
--cc a4 a
a9 > ta a c
aj z
ao aj a A
- a a a8 ~5
a~~% Z aa
8~1 6t 6 6.6
04 6
1 ~
6 ~ ~6M ~~ 6 6 *
I 6k
16 Os r.i 06 i
$-- *h 06w , jlS6
W
I- P4
j 4c
acI" - "I
4c 6 ~ ~
6 60 - t6 1a66 -l I o 461 *
99 N I in Jog 6 6
6 I'a 6 I ( - 0
doi : 0 I
6 40 qv*
me6 M It~f6 0 .6 N
IL-4l 1, 4c 41Ci
I Q)O I zI slz 0 I 06 06
6 0.. I~ 6 6 6 W 6if6 N
6 ~ ~~~ -a 6 6 6
II W 6 6
InN
5-29
Ik1
I~~~0 S S SS S S C4
N
a~~~~C , s a ,. s
owl * 05 V)1 5 "On
'4 55 * S S I
*L S
I I~ j
Z a j uj ,i S,
zl &*
wZ W
I at 04 4S-0S _SI
* UD~. WM S
S Sc
az 5 II 5 S
j S~ 'U UM
a
S
*
~ S
.~lS
S
S ..l S Sa I
NI
S g aZS C
35.
FA Wh
1A
0 5 s ~* a I - -. 5- S--
* 5 5 I a S I
dc a 8
I I S I% S
ci 0 -s9I s 0a) W4 due 5

CO S Sao S WS aI
W 0. 4 00I
* S I 5IA. I
tip Ia
9I'&-..
W-
o******-*-*
goo
I I S
C4Ij
II low-S
-------
aPS
f
z5-30
9L-
I4.
ow
a~-
at an a 2aa
I a a a aI a a a
aaa
in. fISS 1 411.1

a. =(f at4 1 4 at 6-
a aac
~~~ aORaa
'W a. 1- 1. - I 1.
a - - - a I a a a
4c -a- a a -- a-
aa aa a aa In a a a
- a - -xa a- ~~~~~~~~t
- - a -
a~N a a i a aaa a L" a a R -
hi- a Ia ~ a a i~- a~5-3a1
-t~~~~~ a a a .a .- .~ a.- .~ . ~ a.
ILV
"* as I a" 'a a

La a * a- aq a
I cc I- a4 ac
W a&
*4 a0 a 4-46-
Ba. Z-c 2 -10 aLI Ion

*4 a a a0
1- a a a a
Zn a" a 1 of
w ~~I
wc-a a a
at Uzi LL")
~u. w a a%
(A a a a
ac
I ~ a a
a a- ~ a
aa a Z a-a2
I j
1-INC
z U. 0 0 0 4 0 4: 0
CL U0 S0 S
S S6
a4 24
i 1.-
U S I ~0 S
5 I
S I
S S
0 5
Sca
Li- I Ss
I. So an I o f I S
w
Oltw Cal 00l poS 0501 IV* IF*S-N
Z-. INC4 S wS 1 49S 1 51 15-

I S1) Sý S0
o w a , I 4 a C4 a I a I C a SI
Sc S. S I S S S
I~~~~ S oS
f S S
S9 It S S
a a i a a a a a 5 -3a3
* C Lai NOI I I 6
e j I *
* ~ ~~ ~ I6
I
0-ju
61 6, IV
06 I, 0 0 06
Iý8 I( t-I I I WI
- W4 w M
:
-o=I 400
* 4 o 1 In C I 1
WC l I 6 -0 C":~c"
6
U
41 Q
16"1
- 0 64 *
6t I
cc4
Li: 6 .6
:j: 4
i.4
6
w6 I 3 6 3
I
6 6.46
6 6.66 646 56-I 4
0~.4 M0 In 0 0 .0~
w.0I
.0
I
~ M 0I
0 I
0 In 0 0%
00 0 aS t 3c
0 0
IL
.4
$A0 .01 0~
I @04000
00 0 20 30 0
00 0-4 0 w4 w.4 w.0
""a I Io Lo I
0 00 P4 IN
04-
a&A Li
1. % N
0-, 0
zq sw "0I.. pq~I b
' 0
I 00
I LIS 0 00 0 04 4 4
0 cc
00 0 0
00 1 e4
0 0 I
*4 o i 0 I. 0
04 0 NI O .40
0 0 40 CI 0 0 0
00-I 00 .0 @ .0 0
W 0 0 0 (0 0
zJ 0 0 o.. 0 U0 W
u 0 * 0j '
0 0 40 0 04
0 ~"- 0 1 00 00.
0
0jJ~0 .
I 0 I j I I _0 I0i
* 0 0 40 40 4c 4c0 4c 1
0I (A .40
0 Z w
0)1~ 0 0 0i 0 0 0
.I
0 0i 0 0, 0 0
0 I: 0o g o toe0
-K
0o
>
5-35
I -*-a--a
-aI---a---- a-a----a -- a-a---0!
al aa a a a too
C.
* ~1K j ~ ~
a
1
o f
a a a
a a
1
aa
aaa4a
a~ .
Wfl a CA 4a
L I 9LLA1- *
iL I a: Viag inso'in a
a~A a, aI-a~ a w
a aa aA a8 a. b.1 111a a aI,
a = C49 ( A a ac' a a 4 a a) a-
6-8 LL aLA'l aj
0'
~ ~ ~
a ~ ~
a * ~ * ~ a -Ci- ~aw
a a a )
-J ag aý a a& 0 j a a a ac an a :
a - J ac a4 aI - a a () m..
a~~~~10 aa
aO a a a a1
00 a a aa a aa h
ac -a 9
* *-a a Z a a -0 0 S
4aa.a
1
a a a% ~~ a a U.
I-...-a-. 5 C
I I" a a ai a d af o af a
a
a ~
Oa
aj C a --- a--
4 a . a a
oa ---------------
aa
oa 94a aa a Q aa
aa 9- C 9 a a90
a ~ ~a_ ~ ~ ~o. ~ a _ ofa q_j
Oa m
a. a a ea 44 a a c*4aaC4
a ~ a 0.90 a~ aa9 O .-
aazZ S aw zm
aLo*as #a z.
Nwa.j~ a J~-g aa 0a
a~
a~
aw
a
a
ago
tJa
9-.1
~- a
K a m.U a a a *a
fa a
4 OCQa
-....
N a
a
I
4~I a 0 a I aL
a a a z a ~ aj al a I- -z
w~~~ ~ a39aa a a
a a a
cc a ca
laa Coal a a0 aw
a~d aa
a a ~ a a a a a a-a6
I I 0
I I I II
I I
I I I I C
LL.I
4L II I
I
I ~m
0Z I5 cc)II
W 01I W~ I I01 I I 01.
I 1
-1
a I II,
:0 ~~c La-
a I OI
a I a
4I .J
I I I C~4
U I 1I
I I
I I I 0
~
14 Ic ~ I 4
Z ~~ ~ ~~10 I 4aC W
*Ia I I
aa I a * 1
g~4
I W--14
01 I IO 01 I5I-I7
* 0 0 0 0
* 0 S S 0
0 0 0 0 0
* 0 S I I
* S S U I
r- o * 0 5
(q 0 0 0 S I
* 0 WO 0 0
* sro * a
* W305 I 0
* bSI 0
o CS.sSS S 0
W I - I .0 S
U I (0 I SibO 0 I
K * m 0 0 0 0
o *
S S
14 U 050 0 0
0- 0 04 0 0 0
* 0 0
0 I 0 I
14 I 04 0 0 I 0
So. I Z I I S I
14 0 0 0 I
0- I U 0 0 S 0
0 W W 0 41 0 0
SiJ I B.W - I I I S
-- 0 US-1I- S ('II 0 0
* U(/) S O* I
* JS&. W 0 I I S
* 14 0 S 5 0
* ' I I S I
* I I (0 0 S .. ,
O - I S 5- 5 I
* S S Z I S
* W I S W 0
* 5- 0 0 K 5
O ----S -- 0 Z
* IS 0 5
* 0 5 0 0
O W I 5 0 0
* U I 0 0 0 S
* 0 -0
O S S 0
0 S 0 S
*
O
: :
I
*
*S
.-
U0
,*
-,
14 0 * S p-I *CWiI-S
S-I I0 *-O.Z00
. * * 0 00 0 0
J 0 0 0 5- I 0
( 0 0 I 0 0
5- 0 0 5 0 0
I I 0 0 0
U 5 0 I -*--*0--I
5 * * * *
.-. 0-0--0 0 0 0
0 0 0 W 0 0 AZkI
I 0 0
0 W14
0 Z5-Z 0 0 0
0 w . * a - . *
* W 0 0 U I 0
0 I 0 Z 0 I
a. o 0
0 0 0 5 0 0
0 W 0 .J0 00
L
0 0. 0 0 0 0
0 0 I
O *
W
I 5- 030 WU 0 0
0 (A 0 0 0 0
0 14 0 Li ZOo 0
0 IXO &00 I
5 I WI 2 I 0
5 5 I 0 0
U
W 0 0 0 0 0
0 0 I 0 0 0.
S ) 0
hi
0 0
L
5-38
-a--
a a S a S.
cc a
~
* ~aa ~a ~ ~ g~ 0.a ~ NI
- e
P
~ mZ
a a a C
a a a a -a P4 a a
* a
aa & a a a C Cm al a
*
*a-j
S ~ I *
ba
ii S N
m a
oc . o
a *j
a9 ao a a a
a ta - '
a a
a a Ia-
a-aox a,
a a K
a a a a a
Pa- a -
a Ca Ca a
~
O ~ a
aacaa
~ ~~~1
I aw 40
aa
. ao a~
a
a a a a
~ asma a l a4m a,
a
sin44 a a C4 a aa
")~ g
a a I 0 a
a-emi a- a a~ a a aa za a m a..a
W a a- em om a a a a
3 aa an.
ma a 1 a J
a A-o- a aa a aIaj
ava a al Ua al aI IE
w a w
a ~ a ~a" a
wa
a
zJa
a
z
a la
Zz I
zgmo
a aI Za 3mi
I w36L
a aw
a
a ~~ ~ a
~ ~ a ao ~ h
x
a~
A
w
Z
a a
a ~~~ a1a a
a a"
a ~ aL7
aa ~ 4I 0
ab a a a
a aa a- a9
a a a aa 9 4L I
a1 1 C" ~ ~I
-- c 10 : W
09~g:
z
W. CL4La $- .5 co' I 5 5 ILai

5 35
I _~jWa-w
I ac
I qI
&aa
h.Li I- cc
L. J
a. COa a
I -
C.Z
a i a- m aL- f (4 a 10001, C1-- 24 a
a0 1 4W qr 1. Coa: a CL a a
IA -J- a
* I C4 '-~ I AD
'W
CLL1 05 I IJ 5 (5 Z
0.5CJ .41
0.0 40'
00aZ
LUI= Z u09
a I i
*~ W-
I csa
I-A
I
a O.-aa
5
a as
555
a a-~ ~aa 5~
zI5
0- a afl
4a Z 4
w0 U3
.Cl Uz =9 L(6 0 0'J L 5
I5 4
-. 1-I- --- I0 I--in
a pa aI
04 1
I w Ij I
a- 0 ai 4 0 n I In
.- - IL I
* I - - -4
W
'4DI S I deIA S
I I Iku
IL I
La I S 1
Lai W LL, oIll
ki .. ~ -
I NI NI I1 5
a4 -- a a-- --
"ci a L aA aI-aa a
a0 I .i
aC a NaI a
IZW90 a Ca a a4
Oa a
aa a a . a) am a a
aK cc a a
IL : a ~ 1.
a- ma a a aIn ah
> a i a a
Laia
a~~~~~~ a C 44L~aaw
0 (A
a a 1-4 a a aA u
a Z&a a , a z a )a '- a
aa a a Z a 5-4 a2
a %*laaogw a a
I~LI 4 I6 6I
C 6I r 0 O
i
~ *
6
~ 0~
*
a 4N * I I
0 6 NI -41 I O de 4l h
;L m ) Li
I I
*
I
I
l 36
0'
W' I
I
0
Z
1 ZA ZI Z
z4 I I I0
MA. I L6 6 I
I IL I IS
0 C 6 I4
Z WC
6 WI
6 .664 I > >

.6I .6
j~ (I 6i WIW W
.. J_-56 I -AI
I I
I.-I ~ f
$. P 0 I. . 4 b I 6 6
6 Z --- o XI XI Z
L) .g.. S.
6-6 6lI 6
Igi I W
I I
6
6
6
6()
6 ~Li I 6 ( l ( I 6
0
0 ' 0
a, 1- 1
A"
6o II
z IZ vz
5-4
6-Ip -Il P
9~ ~ ~ ~ S aE I
* i -j$- I in I tsl I I WJI .
W I ~ 4 w6~I If
ai a U. I 4L
I-6. 4 1-
-.8 a I9 .. 1 j
t4 (I Joel I ty:aI
Ii.? I I I
* 604 1 aI-
fl rf O
II O fI-
I - I I _ I
I C4
I Z U
.J I w I wi Sa, I a.
x x
w LA)
: ki VA
a. a
I 84.41~ - 1 5g.5I-44
O~
I I-4
I~~~
94I N i- I-I
4 Mf.I
I A
m2
mg~fl * *
.- I f 4c 5
- mi
SM 4Z 1 "I :
" "0hI
I 5 0
SMflS f A
S. qI.
I 54 S I S .s
S5 5- 5-45
-- a~~~~~~~V - a- a - - 7----ov----..
az" a. a a, a a a
us IL 1. s C a
wI ul S ý 1. I a a
* W aL
* a a a a a a a
* a a SPS
a a
SO 08 I~6 4c 0)) S
Ia I
I": : :a
az a aa M aI e7.
~~~ a. %
I-. a a a ~ a a 5a- a6
AI 0. I
ow A
A A
-. 0 3 0
up A AIA
* A A I A A I
a A wl A*A
I UI 5A0N w
: : . .Z : 49
IL A6 a.A.-
a, I A AL do 0A CAA1
A A ca , 1i Cag
J I Aw
w A 4' *) .- A -a im I I. U
c IJ1 wLI .
A P. I.-
C -4 Ow
IL I
La : :
e ~~ w U
p
-. 4I 4UA
CA A(. AW AU
0 A *1 .JU. ~A -~5-47 - IUA *

*L 16 -
I0
u * t- I me
I o"Zt PC- NCI
= 't - I0
Z L
-i I I I
WSB
14 I in
zi IK I m8 V
I I I
Ix "Ba c me CIf IC IL
Ba~ ~ I ~~0
'C 0a9
1-III
I 4P4
IrvI, 7V..w
* Ba5-48
a. o a aPI a,, .0 S o1
* a ow I oll m ~
0 0v "1S
a u
a
P ft 1 I 41 .1
I~~~~ I
Za I I= 00 0 a :
1.4 I S S I 00 0
4L LL In I a 5 a
I.- = I I L.
0 aa
0 a o af a a ..0' - &I a
L) I foim 1- 11 1f V 1 a a ... a
aI.a-.
Loa CD a
a a I . I I:
aJ I II aI0 I
cc. w4 I I .I
a9L-C a taa C
WII91:
w O Z 1.1 OZ
~~ a~' -49Cam
a raaa w a 41 00
ar a n I *I
LL M
0 C CZ I I 0
0. 1 = : :,m
0
ac1 CAL
e4 S
Ia- a 14
L J 0- II 1 4 I. 6j m
04 a .J a JI
IC a-J
.. - I I Ue I Ui I- m8I I
a
I- i LL A I 46I
mi I I -j LL S I I
1 6 J-4 4 1I Il 1 6iI-4 4 S1
uI a Ca I
* a.I S (0 I 4 .. I I t
U. I9 S. 8IC1-C 11 4 I C
e me I -4 4[" 14 cc C41 1
t 40 0 0
-a a
s e - i de 1 1
0A SI C6 w I45 1 I I t I
* a 4 a 1- a CC 4 a 60 C
-. ~W
IW I IC IU. i 5
4* i I IL I I
IL I c IL CIc I I I
a I
I d I UOJ aU 1 6. a w
a w- ~ ~ S 1 I C, SS h ~ e . ~ 0L.J
Uzi 4C Qo
SA 8 Lai "
*-1
I ~~C I
~~~ : .4 a :- 3t
I t a hi
z
~
LL.
U S 5 hi ~
z hi
0 a
:
CdI C 2' - hi~ C hi 5C 50 6 Z i

in Ij 6 us 6 us6 I . I
so 4, 6 0 06 6
in 6 6 6 66
616 6 6 6 6 6 5 36 Ia;
I WI
mi .46 6 6 U)' 06 wP4
6 ~
~~~~ 6 V.
I
@
Il ~ el
~~~m
61
' 0
P4
I Wz4
0 9"60
6- 6 4 1
I~~~~~~~.~A 6 m I 6 5 I
.6 6 6 6 6 >6 >6 >

I WI W
C SI
~II
W.J6-.4
CY
I ~ .6NI ~0 . ~
I
ILa
~ I -
W jI W
LL46
-j- 6 6 . 6 . J6 .JS1
6 66*6 6 6 6 6 WI I UW I
16*.
....46 '. r 6 4 6-6:.
z 6--
64
6
--
~ ~ ~ -
~ --
~ ~ 4" 1 - Of
- 6 .M-CI66 6 6.6 4
3 36
4
1 6 0 6~
I~~~~~~
66 6 U .4 .4 .4 .4
IL0
I~C 46 6 I" I I.
I~ ~ 6K 9 I* @6 I I
. .) .4 .4 .6 .. . .4 .I .6
000~~~~~~ S
I. . . .. J . . .6 .6 .* .I 10
9" I-- Sc
IIL j I
44 - - - 4
UI 0 1. 0
Sc S
I..
4 I S S
14
5 4I I S
14 55 SS
~ S ~ S FL
Z I5-52
S hi
a a0a 4L
COL
C14
aS5 P
" I w 1*
a i W - a
a la
CLa I-
W= ') A wla
a 41 ~
de 4 a O
1j-54 aC
~C: : a a ~ a (nWS
-a~ a0
a
m4 I- a
-- a-41C a a
a a a a
aa
I
5 ~ 4~U.
.ja. w a
a w
aJK
Ii a 5% #i I
a. cci m af
a) a-c
C)KW=a a -
ao a4 a PwC ,
a0 a aa05 OZ
LLI 05. a. w a qa a W4-0a a
w Ca
0 a a
4 a a 0, a
a
Xw a a- ) a a
a
: ax aI 5
C
a--5
-t I go aI
i L
a aa
z : :
a aU
b4 ItIS
00 ,
a a La Ca
0- a ai
Z wa w ~ aa a *.~.a09L
'4 . 1 fa
I.-
a.aCzK
~ au
-
a
.*.'-**
a >a
-4- z--
~ a
aaa
~
a L aZA L894f
a a , a
1
aa aA .a. hJ a
a a a5 -5a
cc :4c I- : 0 IC
I'4 dc
S S I I S
a a a a I W ~ I
Z~~~ IS ~ ~ a3 - 1
I xKa I I S SS I 54
w z0 aIL a aI
0 Sý a -Ijaa
SS S 14IC'
U~L I S
_4to
U.. a I a a S a
6L.I12S
~~~~~I0
,a
S
S
S
~O
a
~- a a a a a U S
z .j ~ 11
-
I
a
(S
(~I
a-
@
1
I
0
C a I -1
a 01 10
-
a~~ S aI a ~a I I ~ I S~
1
F. e
IL
%.j, -
I. Ii
uI Iu Ij "J -4
I.I
u I LI S
a aajI L Ia I i I I I
La a
a-I I5 -5I
IL C4
0~ ~~ I "
uI u I NI 0U UU I (I
WI I. *i A L . I
~5 -41
* I U I
.4 0 4 S .4 .4 SI
Z W I W I CIY O pa '.1 0 .6 1 a. P4 I
I Ii
* a 1 .4 6 0I ~j * mIo I
4L sI 4 IL
I~
II ~~ ~ - ~ I (-II
4,IS
~ ~ . I
WZa
I
11"
0z UI 4n IaS
w Go I
I ~10 ~~ I zI5I
w-,O 1J 11 -*-0
49
vtb I- Uj e4 1. 110,
I *...,
U.3
. .
B M,-I B B..-I''0-
*
4*.* .'. .','

- *., 3 B
I
1BI
I 03 N
* u 3 IsI
A 1
CI 0
t44 in B
~~~~C
. 1
34
o1
~~ ~ Bcm ~ *
I
B
~BB
.
B3
4 in B
3
B B
I
B
B
0 "
4 3~~~~0 3 I0
5 I
s
Z~~~~~~~Q B IN BY
ac. 0-.0
,
%A,
BA._j O ZI 3-• I
BI UziB B
Ih
~- * B , . :.. 43
°.* . ". ° .
I3B I BO LA w
I
3r U I B 3
4 * .~B B *. o.
)B
B O , - - , -
C .,I *
S ~ 5-5U'
N4 6 1a 1
A Itw C4 a a a(D m
L6 W a 6 *aa a U
do
a 10 a s i 1 a gh 0 Iin 4
I aS I.- Ia -4 0
Ui ..h a1 na I Ca 10
a
a
a
a
a
o -aa - a
a
: a a
~ aa a a0a
b-
2 ~ja~ ~ aNaC
a ;
~~~Z a
a I
w~~OLw aZIa ~ s I- ua
~ ~ a a~af I~a a awa
U. 4a w a~~ I a aX a1
ac a ca aa Z~- a w
a aE a) z~6 a .l v~aaa
a L(l az aA aZ 0 a: :a
a -I -S a-
(4)-.-I
~
~~ ~~
~~
*a aa a a
o aA IS 16 I.
a da a a
a ca a
~ a a I W S S a I I
I 0L I a a
wI IL ai I g
W a D a a a w a1 I
I 1. 40 a
a ~ ~ ~ a4 a ~ ~ d a"
(A ac ag a) a c v ai ai
Up L a
i:~I a ~a maa
a
a
aa
a a a a
a-
p~oa a aa a a- a-
a a I a a ae
a a aa a aa aL
Z aA aC
C9 a
.9K 1
.9Z4 I 1-.-
I=
W I S
a ~~1La.~~~~
a a~oa a a
W il.
a aa an
a 4 4I a a
P.-a :PA to a- a a i af
W
a : Wa 1.a al
KW d )I a
dc x i =) Z
KL a a z
a Ln- --
a~ >.
a ~~ ~ O a a4a a
*-
In
Iz 4
a L II m I K
I 5-5.
%.4 09I 4
LLI
S.A CLI . I4 I C L O Ii I
* I 0 u 'o I
I I
I IE
4K
I I= - ~ CL1
ac FI I 41 I lK C4
I 4 I- I.- C4
-C 000
I II 4
A 4~ IL
I
I I I4 I ~ I 4tw-
z 944- (1 4 4 IL
I z I I I I IU I
xI..
IL KSI 4I
I~ 4 1I II
0 .
II .---- II
,--.I
5. 60I
* --- 0-
I S
I. I.
I I ~4 00
44 W6.S
p. 5 1
C4L Ia
WiI I I
I-~Z
4 S
IL La
~ ' ~~j~dc
I- 5 al I"
~
i~a m w * Nd
* .14J I G.W i I *
I 5- 5-61
0* L. I Im 4 C- in 0
L) I B. I 6 l
"W 1 0.4.6 E
I
ac 6O IO
o,6 IL Ia I. .46
.4
6 - I 6nI c I so Iý I I
ac ". P-6 I IOI I 6j Ij I
.j .1 I I I I I Ij U I
~
6.4 ~~ ~ ~ ~ ~ . i Q" ~..~-"
Z L 1 C -I C I
W 6 .W .1 Cl (I 00 .9" 1 to 11
W.ii-4 6 UI 16 WI W I W
I)
~~6.446-I
I Im
I
of(f
1- 11
% I4
am
in I cc I
.. a
I 0
16.
* z I I I I .4M .46
IWW) I (a 03 1 WfLa II l Il
I
I 9L
>. I
I 6 c
6
-C
II
~ I 6
I I ( l (
4
.6 %
I I I z Ii II
w A*I
OI Ix IX 6 00
Ij 6i 6 I I - 6
u II I I I II I I I
ee 6 Z I, I I Z> I 6 I
W~~7 I II I I I'z
~ 6 I I 6-I2
40 P
II 60
6p
*~~~a
6 6 in 6 in 6
o- 66 4
66 i 0 00
*L 60-4
6 n I r I- p
I I
I. cc
La 6 ELm
6. I d 6L
~ at 0 .J A 6 OU
4 (6
or 06
6 I6 0. a: If' I. L..J
4I W. C466
IL 1-4 . *6 6 0 6Z1
. -
w.4 6
66
ui 6 2
I 64 I 6 6 0 6 4
* INC II6I 6 6 ( I
6 - 09 6 6
6.- 6
6. 6 .* 0) 1 W O
, : gN. I. w I 6.4 646
6
*' aZ v 6 so6
ap a 6
=--~
be 6E
~
6 ~ 0 4
6 I6 6 I
I~ IL
.JI
*0
6 - I 4 p- it
0 I. I * 4 -C6
I'~~ 6m
6 6- I N I M i 4. - 6
Z1 6 C.) I
'6 4. N 0
0. 66 . .
z .6 N 6
6 46 fI j~6coo0
6 f
6 -6 4 0 -
W 6 N 6 .-
I 0j6 6 0
660.16 I 6 - I I I * 6-0.' @ .4
*
46 6I 0 6 66 0 e 00 I
z I w
6~~ *
00 *20--- To
I I LI t o 6 6 66
-z
6-64 6- 6 I 1- 0. 6 PI 4
6
6 z4
66
6
4066
6 6 ~' 4
NL 6 6 4
La6
>. >6 z 6K 6Z
g r
~~z
I~661 .- ~66, 66 I ~ No
6 4 46
l a 0
6L
60. 6 6 46I 0.
*5-63,~
I 6 ~
.6 U 6 6 I 6.4 64 16
GC I IL *t cc
90.0= : I~
CO4-0 I Ii1 *j zG
2--40 IC. I "- 5
WS I I 54 5 0 m
z I 6
u j 9L
P-I _j I CL U P- I" I0
I I
U. Si L _ ý WA
*l 66 41C V
a-I
w w0
cc AlS cc
WI I
I S I I - 155 S 5 5 4
I I I IS I I S I S
I I II~~I -0 I
S ei 5. 5 (fI ~
IL. 10 C.,
08 0 1
S cn I j14 5 .55 04 k f1- NI de C1 .06
I-
I lIZ I 6'-' a~.I-e.
9LJx II II bd-LAI
P-1 N
4 0
ooSS 00 e
1.- S. 0 S 1 I. L"SII o"~-. . -. - *

If 1S0 It I a, I4 "0 I -ZC M S SwS
o' (A . I.-I LLI I U) S I I.- I
9L ~LA =C: MWL

0: w :g w:
I -
5-. AC LL, SL- >. L6
I- I Iv . I 01
4L I IL S I L
- Sd (
a.~ ' wi e w
1- z w - 1- WI 9WI : L S"
o S
~~C
W 2 a
S
S
).
WI W
I X C0
2
S
I
S
Ii6 S && I
cc 7 IK z2 Sc I4I zj e I.I 2 co e
5-64
Wol In* I * I
cc so I e 1
I I
U. I I I II I I
I04 1 . I q* I a s 6q
I
It IK (A I
-j #IA I. (6C
WU I L6 % I -
I I tICZ I I-
I doS N£ 41 4uo 1 P I
I. CL I I
I I I a I w I I I I
1 1-O I c" I4 co (A6
D I 46 .-
$~*1-I I I I
Iz .6If~W I6 .6e I
I. J - I I ( 41 em 40U I IS
*-.-.1 6 W
0SII 6 0
MI
IC .I4 96 6 96
xI I 6 &I6
C * I I& 3 4 I I I .. I I I
6~ z- w Z I
K-- 6 a 2 ----.6N a--.m--.i

I
O I6 K I I ~ 5-65 I
I IICILO
UJ -- I CD 4 .
LLI
I.- I at I E
cc I I I LA
* ý m o, I WIJ
Wzc V) I
z a-
I S
L4A
I W
me
w0
-I I
I
-AIin
m
I a z I I Z
I~
*
~ 9- I I
a I
I ~Il.
I~~~ 0 14I D
z 2>LC IE
" "I I I A
I I ~ 5-I6
6 I It ,\1
IOL
C6 0
11
-. 4 In
A.1~I I -. j
*I 0 I I
I eL *
I~ ~~~~~I 1-I 6 666
10 6L
i
I
I ~I I
OK I. I $A.
6.1. Ge 4
wILI. (A
Z~-
Wi
2.W coo I II
m6
I
4
W
W CA
.
I
0 0
16
to16 I
I .J
I
.4 I OCIII I 4
00f~Z I - I U
I~! - I~I
@1
II> Iwe2
I WIa
I I I W..J-.4 CI
I 01
IW.J6-14 0 1
I - I UI4- I I
U.4- I II I I
II I u40i 4 0 I I II
Ih I
I~ 41 I I I Z - 4
NI.1
I-I
I K a- I II -6 I -- I~I
I UJz IWO
I I II I Lai ( 4
6 UI JI I9 I 9 II Wag6
I z . 1
I 1- I I0
I -4I I I I
I I
I . I 9I
6- II 6 I 6-x
I 6 I U I 6
t.
I I. I UZ
UO 41-6
I I
I
Ii
UO
I
I .1 44 I I 6I ~ .%a
6
I I
I
I
06Cl
I .41
0
643
Z I 6
1
M
I
.-
4
0
1
1
1
0
I~
s
* t>6 I 4 II
0 C 0 .U 6 0
It UJOI 0' wU I@
I 0.U I 01 4WQ NI5-67**

I 4..UI W~I 66
*Ow WU--~
00.
-00..
1 C 1 Co 0
I 0 1 1 # I I .
z4uI Ii S4 01 0
C1
u1 :0 :
4NI I Ir I "
- i
09 ". I.--1
- I -J Iý IA S& I
1.-1 11.4
uJOL& 1 IlI C-I 1 l I I .J ~I
4. I4 I I 1. i I Z
ml
I~0 I I I I ( I I 41
C4 C4
Z
I
*
~~ cc I
I ~~ I r-I
U I
4
in I
C" I
10
W 16
W
~I I
I 1 s4
1
fi I I I I
I I
I II I
I I I Wl
C a.I I I I
I IL I I I I Ii I
I~~~~~~0 I 00
() ~ I 4 ~ 1 () F
I k..
6CO~4%6
C2 I " 6
cc kn if .
6 0
cc OA
U.w 6
Ia (4 (6 a
6 N m -Jc NI 6r
I 4;0
*6
64 6 16 6
61. ~ ~ ~ o 0 6~
in- 0 6 ' I 6
o6.
0 06
*L 61. 9L I
X X :51A U
Cl
6 I 6 6 I 56-6 9
'"O oo 4 0 03" , I ..
4L =. ccI a'In I m
04L I - n1 0 1 - *1
I~~ ma
09099a-
a a a a C C2 0a C
a , 49 al a a
* Q2a a
001
9 a . I I ~0bIZ 064
9 ~ "Zs l4
9~.
a- ~l
.. ow OOa
W= 09 090 0
I- a L .aJ &
Z 0~ ~~9 09 ~ 09 9
~I 9 4
99
~ ~ 9
~ - -j -1
-j U.W 9- 4iP NI ('I 09 IS @ 9 La
J .a .. .I
4 aa a a aa a Za Za Za Z
*i i W ~ W
40 0 9 9 1 (A C5 0 - ý G
I9 .9 .9 cc ...
to4 49 9 .IJ9
419 0 1 ) 09 099 @99
1
N Ii X m~ '0
a I ~~~~ a a a
-
0a0
~-9 9
a
:
9NI
I
U.; aI 9 a
- .959 II 9 a I 9
aa a a a m m : 4 k.
a 9 9 9I -C 9
* ~~ ~a--...............,a.-
~ ~ 9I 0
I
0110
9
I
9 I
..io9
a
-.
0a1
a ~W 9
lK LA.I
9 a i a a 9
9 -I 9 I 9 9~.-.JIn9 9 9
CL
aI I c 9 9C A I 9C
9
a- 9 9 I9 z I9 fi I99 ap 9
z I ja .. j AJ9 w5 .99 W. I---a W-t**--
e------
c9 I 9 49 9 49 4 =
x- V)a a a- - - 9I 9.- : aM 9- /9 4 I 4
9X 29 CL- 2go I9 IL a I I
2~ I WI WI WXI x W I~
-*J a. wf
2:) 2LA tQ:
2:z
* ~~w0 9 9 ' : : : 9
WI Wm WI WI) Li. w
I i -K~ I I ( C 0 z
5-709 I 9
I* Co~ ~ b, -- -0' =--I 6i
co 6 I ,
* 6 6 6 I
zi
Z~
Z ,o I 6jLJ
S I &a6 I
6 K i h 6 0 Ii I-II
Z~~~~hi 6r
0 ' II- 1-h -4 6c I 1- = 6 i. AS I-h
hi die 6c I in Ii
-JI " I C :-C I 1
:4SW *~ S 00n
I~0
I .L. h uj
Iw h .- a. 6 6 a4*6
I~~~~
II 6 I- I ICA"
I
* w 'a*
* ISI
I *96
*L
416C0
1~ -. 06 N'A
6 i uj 0
-46
hi I 6 6
K~hISo
rhi I I 6 6
if I I6 - 6 ( II 6 - 6
fI wbI h
*A hi
p65"71
-LN -% - % -. .t.%
Ia I b Io IM I I I I" Ir
ac I in I 4 I) Iý I OW P Iwo
GC CLuI Il fI I
IL .11O1
-(I I I tI ~ I I
49 1 :i~
'A-
I La La1 41
W 9L a I. I
-0 144 I0 I' I' I "
uj JýI* I Ii I
41 II I II I z .I z~ z.
2~~~4
I I W 0 49 61 4IWcI
'I I .~1 I .I>J>
mI-41 61 1i toI I deP IN toI C4 .l

Iz w I I I CL 1. 0
Ic I 0QI ~ 10 01. C
cn1 V2, 21 .
'n4 4c 4 1 4M 6n
0' 010
1.1 14 .4
cc" M% 11 W, 11
31r 1 I. '--.
I I I II II4
P4 I I I II I I I I I
II I N I (1 I m %1
I
14
I
I
I
I
~~Ui I I
I
IZA
I
Z1 I
1.0XI
V
.1 P-
I
.41
I f1
NI
V
N
4 II I I I I 0 I (41IZ
I ~ ~ ~ I~~I 9
I
R
I
-
I I I
I6 Ia IC I IV R WZA I II I=
I I W
-I I I II 5I-I2 I
V% I I I
* I ID I I
a~~~~~ I II I 6 4 *
I I I I
I v I M I I1-
I2 I %I, I I Q
I I 04 I 1 I 4
* r4 1
cc
a ens a g
sc.
m O, .0
I * I *
I I a * *
(5I
fUI II II III IlU
(l11
I S lI
0I
-i N I S I
I I I I
I l NII
X 08. uI I
I- 1 "4 -c -C.
a aa a a N Lo aL
M ac a a
Z 0 ~ a a a ,
0
0
a Z ~
1a - a 1 6.4aa
S .
Z s aOa a Z a C a aa a
ua -it a0
a z) C~a ( A aV) ~cc a a a
I-j.
L, U aý aU .. a Ia
_j a aa
P4a
a a a a a- a m a aa a
aj a j a w. ~ a C w
aK
Oa i; n cc
aa C S
4c.wa
gm a-
~ a
LA
a C
aC ac a a a a
aI< -o aIa.a In a a- ILa
Z aaa
:..I..-. I a
Z .I I a
a la aa
40 a a a a-c
~ a.a u a
~ ~ a~ a Aa..
a~ Z a aa I.j
a 5
a ~ ~a~ ~ A=
~a~L)4 ~a
a-n
C
a~ ~ a
.J ~ ~
aZ
C
~ a
" a Q
i 4-
la-h
a 0
a ~A a ai a n aa a
aw aa aa a z
>a a> n a
C~~~~ a> - a c, aa a i -- - - a
a- w a a r a ý a L a a
*~~~~~~~~~~~~~~ *. a---a---
a-.---a.. Za.
2E.
* a a aj a a a a an a a7
*
an- a a.J a a a an ~5-7aa4
a a a
0 fa II I I
* ~ o I CC0 m e = mm m m WS
w 4-0 C £* e
00.
~
Z L"I -*U.- IA
OC
K "6 1 1 -:1
6 6 z ,I '-
tgz4I I .e I " Ic . O I JC1
z I I-( I 1W 1 OI 0 1 P
z C I I1 0
IL- 'D 0. 1
I a I I I
*.0 ) MI
0e qe .1m
Ik _ I _C _S I I
CCUI
-~~~
* I
e--e--e.-
e----- ~w I S
a I
*OIL CO Q.0 W I I
I W I xI*
. .I IS C'' 4
* I-a
u:: :1 .C z1C 4 e I
I I 0Ze Qf
*
>f ~ S I ml
S
(C
I ~4 Z '
w *6 4c a 11 S 0:
* e- - e---- e I 0 5-75
ak a- a aa a a a a
n a a a, a a a
a a a a)a a
Cc
I.-aaa t-, (dc a
wa 0i
Uaa ~
a ~ aaJ
a a~ a aa a a a a
a Og aK aK a 1 a "a a0
Co aL to a
~~~~~L
I
a L- am aI, aa a
Z a w a a a
a a g ) I aý a a0 ao a aac
a U - a a aa a
a i- It In a a n MI A aa a
a- a a a aa
a-~~~4 a I
2 a J as a a a a aa a a a a
~ CL - a -a (a C m
* a
.- a
~ U
4a
w
ae-
a-
a~
-a
a
a
a a
aA
a
ak
a
a
a
a
a~ ~ aj.a- Ca -a a aa a
----- a0 Go aO 90-La- a0a -a
a~ a a x44ja aa a a a a
aa.. wj aU a
a Ua" a a aa
a. 4 j a a a aM aUZaaI
a a a. CO a a aa a a aW
a a - a a a b- a
5a-a6
a a . a a a a a a a a a
I II Iq
gn 0 1 at I I
14 M Co
--
(A I a a a N
0o
0.4
*' Le
I- a a a'
" Ir 4 *
1
Ip I
a I I I I
I IL I
aa9
I-
I* I
Z aw a Ca I (1 ~l I I I I I
- I0 I >
z
*
.11SI
~I
~ a ~ al ".J-4 C
aIl0
I UI
.II
I
I I W1
I ~4~A a a ~1 I *
a I
a
a W
a a
ah aa
a aj ai a[ 1 a )
aL
I a I I I Z I
a 4 I a
I
*~~~~~ I I I ~v
I 4~ I I I 41
CII
I ~u LaI L I I ILz
IIo I I I In
5-77 I I
aa~k%
a ~ 21,a a 1N~
1
aI -0 Z41o I
j M La I. ~
aa Ia
I U L
a aa
o k:
i a,
*4 a a X
LL)Z
00
O.W
a %ac o a=
a aC
I L...
* ~ ~ a w
D
a. a Uzi
a~% A
a a a-a8
t
0 I
* I
U).
W w
-
~ .4I ~ I
I
0.
0
0'I
8 I .J
II
=0
I~~~~c I to(l II
*~~~~~ I I S I 6 - I ~~
I W
OOC.1.
I L-J I- en m 0 . i. 0
I ~ ~ ~ C) ~ . 4 -4
LIII
* ~ ~ I oo U" I LI I
Ir to, I I W
40 0 t4 0 kn
I I I I ~00 I I I
cc I I
C,-JO C,. 0,
I ' C4, -I -"JaI
I i _j _I _ I _jI C
I IPI I I I- I 02 4 A e I I
w- -- i -. I I 0U IL IA Ir
z z I z I I cmI Ij I
W ICC. CL 0. 9L41I
I
I
I ~ >.
- ~ I
I
1.4
LII
1
1.4
I-
2
-
I 2
-
14
I
2
IS
I01I z
Ln6I
I 4 I I .4
X I W I W 55 C.
I - *2w1 2 1 X J
S - - - - I-- E=*
I I I I I 5 -7I9 I I
I.-~*h .
IA aat aa a
aaac
'coo :0 4
<
1cgI goI IC. a
OJ" W
a- I a a a a ai,
a a 5 I a a s a a a I
a
a
~~
a
a
~a ~~
aa
a a
a
a a
C- a -caa a a
a
ta-C
ai i a a a
OrIa
a , a
a~a aa a a0
44 u4 lo t.1 I
uj I IS I I I Em .
I' 0I . Ij Wto
W -j I... AI111
IJI I
I u "A I
Ij
~
2 I ~ I S ~10 14 ' S I
I I LaiI I L& W5 I
1`4 1.
:0: :0
x CP~ Q. 0
MIA-I W I
U -W I r) ,
4j U. z' 1
LL)I zI~ S 34
S0, CI O
I
*~ ~ ~ ~~1 Z I.'11
I IAJII S I I ( 5I 81
-f 'A
I C
4c
* I:L:
LL 1 *U
LaI
of I
* z I I 0
Z w
o12 --- *s-4. CO
I-Z' II 0 0II
I0 0 o
*~0- 1 CL
4c z In >
*i Z
~ 5-82
I I IiI
UU
COC 06
GO I0Lac to * 0K e .
I.
I I 6 I
I I
CL I- a , LIM VV6 (46
I 40
U.6 6 (
LU
( NI 1- to6 i.4I 06 CI
6 II W WI WI W
I C*,I 0O 41 0 0*6 0 I4 6 I.. 0 j6
I. *
P4 WE0 1 WI 1 406
I
.41 IA 0 I
*iLL I W 0w
* * I In I C 1
K
I id LL.I U.0
S I Ii
on -4 1 Go I 6
Gota C4 4 1 01%
1 4 I I-6 In 00 * 1
zI I 4 1 Iw -
P. 0m 6 01 1
I4 ICIJ
oo _.JCO I I. 9 L M M4 k mI I, I
I I I
ZI& 1 ~ ~~~ CL 0 1 .1-4 C,1
1 0 C" 0 ' , 06 ) I I
2j . 1
- 0 6
I
0*
W 06 .I 1, 4 N 1. I49 so1 64 .J . 1 *
0 0 1I 0 1 I o I WI I YI,
I W I I W
JI * W W
9I 0. I
64* IL C IL
I
0*
6 4 6
6 40
IC I1 I 1 1 "1 : 6 6
A X
* I ui * . I
I I' WI
*6 64 4
* 2 I I 64
*
* I I I I I *I I ~ .6 IJ..
2z6
11 In
I
U .
2
6 I I I ZI
I i 4. .I
I *46 6.46 6.466-83

-. I I I I I I
* I
o fa al
on.a-....--a*a-a*a-a-a-- -- a-a*'-*--a-a-*--
an aý 1a . a a a ao an a0 a a1
"0 aZ
a~
~ aC0
a
a0 con a)
2 a" a
m
ua
a
a
I".
N
a~~~~~s aw aa a a * I a a
aIX a a a C a C
ajI a 0 on acoo a aa a a
0 a a a M a m
aZ h - 4 a m 'a 1a a a (l h h
a z I q a q a aw Ir~ a
Oma wZ)
a Li 00. a P4 a m aI n aI r a6l N4 a~ an aU.;aa a~ a ~

a 0M c aa a aZ
a -j I. I a
2~~~~~L"
a
aW
a
a I
aa
aa
a
a aO~
aja
a Z
~
0~~~~~~~~~ a a a a za~
z-~-a-a.a.a.a-----
a~~~ ~~ aa a a ~~
0 I
a- aa a-a a a a a a a ~ a
a
a
a a a a ~~~~L4
Imaa-
*~~~~~~~~O (JD aa a aa a a (j o~
*~~~~~~~~. a a a aa a 6
L~a
ia a a a a a a aa a a m a m
a a Q a Q
*~~~ an
a0 a,
a a
an a4 aý
0 av w. OP ~a (a a
wa 4 a a ga a- -a1a aY ai Caa.
J a
1 a0a a
*
w
0a
0
80 - a~
accNO
~ C
a W
Ca
a (~ m
1a 4-.a am 4-a a a0 wa
J
Wa
.)a a
leia
a
0 a
a~0*a ao a aa an do~w a >a >
a p-a ~ ~ aa*a ~) ~a-a ~ wIi-a ~ ~ a4
a I aw a W
-- aa ~a 4a a %aa
a a aa a a .Ja a a a
a ~ ~~ ~ ~ ~ ~ . I(&
a0aj4 .Iaj a a a w ~a w
* a a a a a a a a a a U Qa ja
ao a I' a a* 14 a0 a a0 a Z
* a a. aM
a ~~~ ~~~~~~ a1aa
a
A
C
aea
a a
a ~La a
(~
(
)
a ~~ a ~a - a a aI a ~ a ~
a~~
aLma i
~
4--.*-a-a.-a-a-a-Z
La a a j a a av
- Z1 a a a
a
Za
a
a
a ~ C
a a a a
IL
a a ~ J a
a
a~ a
a (a
~(l ~a
n
a A a. a p ac.-aa .
-aa -aa ma ab4 aZ aZ :CaaOna
'(1 a
m~ ~
m ~ U~ m ia a aM ala Na Nap5 m84 C
a (aa a a a a a .a(na a~a a -I qa maea p-

WM
f C4
wsas
- i J 1-$
*% <
a fe al cm I
0 Al @.W ~ 16 IP 6 -*I
a..
UOW a
* W~l~d ~ 00
06~
- S I w -
Lai
6 -
% V6
a a 00 A
I pt CI I
I I C.) I I 4101I Icl I

I cu 3 3 I S S I .
ac I a aa 3 I * C
I. I S I II 3
I1
II 03 I I I I I I I S I
I IU .
11
a:a:Ia In0.
W I Ii IL
ItI
I
U, C, "1 11
-04~: . &W S
1~~~ ~ ~ ~ a*- oI eLa e ) yC
I,1 I I I 00 60 I I 1-. S I I
I
II LA~ I8 of (O I
0a (0 .1 4w C)'I S I 0' I W
3 0
w. 01 0I
NI 0 JI .~ i 0 I
W-.JW
3 3 5 3 I S ~~4WQ I @3 03 N u
I * g * a I 3-B. 1 a ~S CI 4j
IsIAI I I "li IW I, UI I I I I
~3 I I 3 I S
I I " : = :
w X
Z---.--...~--I..--I*I*- I c I: * z
I I I I S I I 86 W 3 a a5
I.-6 6L
6 m 6
c "4 1 4* - *
*44c
1 6 6
6. 6j 6 Lh
-j
6 6 6 6 6 66 6 6 *46
6 6 6 66 ~ i 6 6 ZI%
4L 1-- 4 66 6r:
I r
L0 m 0
66 6 .4o 6 W mi 6 A
64
-j 6 6
6
00
.. j~LL _* I
* ~ 6 6 6 6 6 6 6 6 -'t
C) 09 - 6-i-
II 00. 6
0I I 6l
6 C4 1 16 :
6 wi 6 . 6 6 U 6 L -
Iw W I W W66
6 6 4a4 6
O i" .. *
* 6 6 6 6 6 6 5 6 -8 7
.6 6 . - - -6 -I*.
4 Win 6
p .CO
bL
CL: -a
~ ~ I
at * L 0 -0 SI
I
* B
I IO
*
z *
PI P"*r I I
CI Z 0 0
I 0 0 5 I I0
- -I.- S Ix
S
LU S S~J I 06 5
m
I
*
~ -
~ I
L
I
IA
99i(.*
)C~
S
S
*1 I
* L iU I W S
Oz A 1 1 0 9 er
4w I. IsIL aa 0 LM *r
0j 1 U4 S U U I 4 U U US
Cn ~ ' 0 . -I 4 W I SW0
1 % '
0 ZU Z
I A. w
U4 "& &o I q U 0 S I_S UII IU
Ur U I
S w IS
Q: : W
z. z~.
U I I I W
UK U W U IfSU"I I NJI M WI I WI S WI
U I8 I " (I U -I -j1 I () J
I 40 1
Uj IO I U S0 4c 1
U of 0U I wfI
Uth U ()
S U I Ch. I 4U 41 U on 15
1-~~ 15 t1 -" V . ,
U z . -l SI 4
U CA I I IU U S (A A A
I 4 I 4K K I 4 U C I A U I I
*~ ~ I NI
Z1. NI * ZA 0 ZuI I U I I U U
U I I I I I1 x I -. U N S N I N
IcZ uA IC Io U)
I Z I Li A Z
I I I I I I U 5U
I I I U I9
ItS
~~~~ 0 0~ Of
S ei. L j*
0.4 0 n Uj 0I 0 0 0
0
I& j- 1 -- 0 o f 1 6 UO
04 0 0 0 0 0
cr W C O IAI
0I4 I I 0 .J
I .04 .. 0J I 0 00 * w
* L Z-
S 9= 5 0 1 = I Z
. 0 10 ft e . 4
00 -1 co cc
000
0I1 00 C, 00
1~ 4
<ZI C cc cc00-5 05
f 0 5 I-,. e N
00
uO NO
V-0 NI JI 0 0 5
O 0 W
N S I ~5S 90
0. : " 7
Z)co a-.-a---40
I In 1 -4
a a 1a
V) aa
-I Ca I-- a
I _a00. W C14 *a
ii I " fa I -q
ag In I aI "d
aZ Z a aa
30 a aa cc
~ ~~C
I
4 .J a aI
U~~ a a MA IJa
* a a a a a aL
C6 a aa
a a a a a
a a Z-a a L a1 "4 a
-. .
a~
a
~ ~aaz-a
a
mit
a
a
04&i
a
aIi aca
ao
a aA uua
a a a
4-.-.-a--a-.-5a a1
La
I Z:
nI : 1 en : 1C
.
.I 6 I I I 1 09 1 a
IL Il=*I
CC I a I 3
I KINS<
I I Cal
I "II
0 a
"Z a-a
2 aw w W, 0 f l el~ .0 a 6.I 1
I uj-jI .- P4J S I tn I I 1Ih 40 1S A I W

mi
.J La W IA S La I I
0 0
4
I - S6 w I C, ai i
08 a -Je C" I a $I . p I
: ~ .~: 0: : i ; C04 Ci4 f1 0~

N
S a a a a a I S I I0
(A aN I S' I
I L 0J S I0 o fI -4 I. .1 II I4 S S
Iin 0S I In 1 0 o IA IP INI,
I L S j I1
-I I i I 6-4 I i. a a
WI ZW WI WI Uzi Oro a I
o a I x I S a 8
* I k i IA. L i w u i I
I 0
a - i a a > a 0-
5-92
NL'
6owl*
- * * ** -
. - I * S - - 1 . - . m -
V44L a ma IL a
or 505 I ~
w w o a Oa a,
* a I SIS
*e a a e a * a C
* isa &W I.- aL a a a w I-- m itaS
&I -,U
0-1 1a 1 ilk1.
ac 4K I-
ac 0-.*.4~ I *- *. I -- *
:) a caW I a-
= I I I
j U. L I -O aL ab a a 0 0
I
u .I'
- wa I=, OZ
I S I C.-1.-0-. .
C4 a IL I-.-.
uI--low,
a 0 . 1 -CZ n0 0 m -aZ-".
40-c 1 1 41u4 0 a0 1( a a w 14
a a
I a aa aI
00 0 (4 00I4
I - Ic 0a isJ. I, - 0 2' IJ L&
w a =I I u I I
a I SI z a I I z 4a
a
S
a
~ ag u
a
4
a
a
u
am ao
zaIs a
a
K'
'
z
Lcc Ucc m
aU
I-~' x 0~- -. IL- 1-. a0 -

x W E(
I i1z
.or
.. A.
5-93
4.1-S..
. . . .
oj 40 wa I
104CL B B q COLO I ~
0 0.a B mB n'0
0 Bl
c BjI
B.
B a j I-- B B -a
1. - 6- 1B B.
Or I .
-It t IA
B m .AB BB I%
.j"I L& BB B
B B I B B B B
l~a BI B B14 I B B
B- B B B- B B B B
Z BW I LA B I ac
W W B 1
* I./ B. Bm BI ago. 1 0) J
m - IBc B
I.. q B 1
O 30 m -j cc m 0) O= 2
B I LB ccj Bl Bn I J B B co j 00A B 4 B 9 BC
4~0 B B00.B B
B - - - -- -
I
-gi B
~ B
BJ B
Cc
. ' B
I
Bc
. - I
I B
B
C4 B at
uB
I I
B~~~ B NIuj
B.. B L.B~
4B.0
0
B
BWZ
B B
B 4-c
B
BNBB
I B
~ B
B OB
B --
4- ~
B-
4~
B ~ B
B
B
O~~~I
B
I0
0..B-LB
o 4
..
B ~ ~~ BB
BB
B
Z B
B
OB
B- i
B
0BLB
4 B
B
BAIB(OB C. B .. J~ B P..
4A B B0 B .A BB
B >B 0 B B K B
(DLB [ B IW1L c
Z B B
5-94-BB
B B Bow
0.4 1
aa
a~La
I Z IL IaL- S-
a a* o -C IA
0
w~ I L.-
InI
aE CL Ia
a~~ ~LL)J-~a
a~ ~ .JKa 4.)aa
aZ aJ orZ a
a I a
ac aj z z ISa
a a 49 aa
aa ~ a -- 5-a--
-aago-a-a-a-
a a a a a
w 0. 1
* L WZ a co-
a a.~ a a a .a S .1
a a ~0 a4
%0a~
%
wz m aq a m m
an a a a a a
aua4.aa4
a aa
a :, a.a a a
a0 ~
I~caP. aa a a
a a
a aW o
aZ
a
a a
~ ~~~aa
n a ý CA
a
ar a a
a~ a a a a
az a ae a- a a. a O a
a a a a a a a a
a4c a aa
a~to a0 0
Z a a- Ca
40, a Caa a
0wZi a, a a aa Iqa
Q 4c
a :w: q 0
a" a a
aj wa a aa MW
* a a a a a a
*~~~ aV a a aa ~
a~~~ a~ ~ W.a W a a a a
~~a
a - aa
aa
a-
Za
a
~a
aEa4
a a0
a~~~~ a lopa,
a a C az :Cl0a
Qaa a a- a6
6~eg i I 6 in I
ac 0004I
OK,
z: ....
S.
.o .:o- •, :C • I :
S.. . . .. . .. 4_6
Zo A 66
ot c - 0 wA
-o . .K, I,- 6 I,-o
n LJ_j,,,- I.-'
,,I •
IN 1. "a : 2
#
cc
,, 6 -4 ,61 ,- ,,
06 6
I
6
K
1 I466
S 6 ....... .. a', 6. ,
6 I6 6 W
6 L64LI
Z 6 ..... 6. .. 66 ... 6
4 : 6 6 I 6 I .4L
= 6
S6 6 6 6 •6 6 6 6 * :.
I 4 . w I - wl, 1. 401 ,C
o'141..
0 0
IA. I 6 6 66 IA 6 6 LA %
I.- O LA ""W
....
" 6 .*A 6 "" 6-* 6" " "6" " " " "" " " "" "
. .- " 6- " - . . ' - - - ' ",6,6-6"

Ic 10 a 33 3
3 3 3 I 3 3 3 3
LL. 3 3 3
".4
3
IL
L 3L
3
3y 3L 3
g *3 A. I--
I
0N
3 3
W= U 3
e4:
L6 01
U.3
c- 1 34 3 3
3 3
3o 3 -Cow 340 3 34 Jc 'A3 "
o $. 6:3
3 ofJ3 043
ZC I4 3II 3 -
(L
C32, *
3M 3 3 M 3
3~ 3 ~3 3
5- 3 3-3
CL die3~
w3 4
I " 3 A z3 3 .4 u
* ~ o LA > eu3
LA.4.4 I ~ 3 3
0e * u 3 ai tu *
C) Li : N;
c I :~~U w 3N
~~~~~fl~~~~~~~
I ~ 3 -U 3 3 J 4O ~ 3
4M P %
34 33 3 .Z 3 u 3M 3 3 3-5-98N
a~w a in I S *
~
o ~ ~ t aI
Ia =! ~I- 1a .a41 a
Z w Lmma pC a-a
9L I-
:K : a 5
)::.
W j: ac " 4 I.-a
1-: a
LLaW a a a a Z
0 a--a-a--
a ---- 55
a a
a--- In a) " a a.'4
I*j a 0- CL W v
*- a a w 0a a a a 1.
..-
* .. 9 aa0.
a* .
a --- '--a.-
a C a 1a o a 0 a a
Z I z4 w
wA 5Z . a
Lw w a o r0
Z 5 a 5 I5I-99
a a aA a a
II
-- e ---- ---
I----I I-a-I -a---*--*u-- -- a--
* w Iz L C I
.4 I I II IO I L LLI,
.1.. IS
5 ~ II I .4 4
CI
*L 0.~ ) II aZ a NZI
a~
cc 03 a a a a .4 I
2~& a0 a
Jdc I a Io I IK
OD:
I- 4I I Y 1 46 I
In: I OD: 4 w
ACI 100a a
II a I A& 0 a
a4 so a I aof I td I af
1 O1 *KLi ala
a~ a am
*~9 C Q 4
w0 z
a .-aa
Li I 'I III (1 I-. Oz I
> aC
*L
*~~ aL a ~I- w 0
5-100 ........ :....: . . .

4 :777p
a-. a a.a m e
I. C, m z I u __i
w2;: .44 ama

.4:
I = aI
I4 a.
* U. . ,. .4 1
9L 1a I.-. ~ ,- * aa
~ ~ ~~~.
a~ IL - .
in
U ~ I 1 - 00
U- * a
I a* ac
I I- a - L
CL a 4"
* a a i a 4--.-
a a
m aa a a U I
9U a E jW
ac a a a aa
z i :i Lia
K 2 ~ -a-(m Wa~ K 2
0550101
a a a
L12 1 10 1
(A Sj I ,
LAa a
0 aa az
I. U)
4E4
'C i I Pa I
1.- 0 a0 -J a 1
I a-m m aW 17
0- 1- CL z a a ,
z z ow w a-
la,-I. : .n
I 06
* a I a. l ow.4.
a X.a z I
tI I I -LJ
a Il IS
> . * .
a -a
x LA LZA a o
* IN%
a I~ wZ O.5-1020
04I I I 6
O: 0160
I s I I s
: : a nita
ULI @6 J -II
j0
I .6- I"M 1 D I 6 I
$0 I-
Z C" I 4 4
-j LA 6i 0 I8 E E
I 6 qv) 5 0
IA IN
w
.6 .40 Ij V I C6J -4 M
I1. C4 1 1. CL I
I-to- I
9L~
U I ZS'4
at 46 .66
.I (" IC0 Ion.
6 W-i o 6 Al *I L 6 I W
I 6 I I Z I w- 6 6 U 1.
IL Z 6 I 6I LLI146 .4
60 6 o 1 Con 6 .6 4
C CL 6I & ~ 4
*I , I w IW IA I ZA 6i
Z CO 6 l 6 , I Z CO 6 6
0 C, Z
6 6 6 I 65 103
C4
* ~~
3
a
~ 3343 ~ C4 aa aa 3 ~
* ~ ~~ ~~ a3
I
* j u ***
* -0 A3 a aa Owlaa
LL..
,j I (
Om
0.
0>
j
* Le L4 .
: a
Ia I
3L a ua Z O O -P M
C* 34 3 cc
Xa "3 1. a
M -C cc V3 I a 2 3
LL a Si L I
4 3i .jA 3 4 3~ w
I- L& aL 0Z~I w
a
I j 3ML 30 Q3 Ob 3 2 M M 3 1" 0
3IM aW 3 3 .3 -in M 3 OW - 3 Q
34 4~
3 z I * 3M >3 a ,- : <4w at 0
. I
4 W -0"- -
0)' 1
La.
IZ WM
w I..,
lo wa
.-
43-
4a
3 MI
a~-4
3
Zia 3U
~ w<34-(a C
031
cn aoV o
4n * 0
w 3-w-
MW 3
~~ a a a ~ 3a
m** - -w~- 09(.)
a ., 3. .*). 0 a
3i a 3 3M 3w33 c.I 3
a a a a.j a a3 - a O .J .
1 3 3*
Z3-( 4 a o aL 03- ai N aaZ0a
. .- ..
V4 I II
*9 OILS
ag SI
U . I IS
Uz "'Lw * *
a ~ at 6
(* USS S
Cc
-j & W
* * *
S WE SMAX
* 4..j
~ 00 I S 41
*-s
0
U - a. 4 -1SA '
I *IL
I 5L z I
LAS II S
0.W 5-LS
~' W~h5-p105
a11 aý b, ,a *a a S
a'
a aj a a a w aI Ca C
aw~ v-w aC I aC a aexa I
a
ac aIL a m a S
Ca m U faLl a 0 0 0
~~~~~~~
m a *ma sj a a
Sa10n a .a m .~ s U )
aý GeCa.
~~~qL~~~w a00a m a a
a a aa 4@g
wm a~
kW " 11
Oa
C a. aj a a
a . mi aU a a a aL a ia ua
b~~~~~~ a
a m a C a C
a a a a a a a a a
aL ul a. a& a a C a.a
a lug
aa
a
~~aa0 aZ a ~Qaaa 'a m
a. a.a~ ,
wa Ow
Z wa a a
a M
w~ a a
a
a
(aA
a
(-
.a .a
a a a *f
.a .a
CaO I a l 0a a a W a
~a(lwE
a- aCo
*aM- a" C,
cc
>
a
>
w
1a w
a
400
OMa
1acaa
*-m. a) Iaaa I a n a a. qava
ap
aý a1 a a ( a I4
a a
a~~-
a a %n~ 0
a a~ a a0a ao Va
a a
Ua
l.a
Ua
w
o
a- a a aM 4 -a aa
a a Za Z
1~~~ P 00 ~ L
W
U0 a-4
apI
a-aar
ay N a a 4 a%
a L a4L --- a-*-*--a..
a .. a a aaac
*~~" a
a. a0 aq, I= a IPaaa -o a a
a *
a
a
mo
~ ~aa a-. Z
Ca
0.Lij
Ca
i LaI Wa ~ x a
a Na NaOa
a~~ wZ a ma ~ waeeaA~-a ~ M a
a~~~~~~~~O a *a C* a 4m aa-am a .
a~~~
00 a< Na
La a " a
a
ai
a =zIn~ ~ ~ a q
a a aa a aa ca a
..
j aa a a a a -. Z V a C
4~~ a ~ 0 ~~~ a apia N
mU a5 La aZ a ao a0 a - a a
w a a a aa aa a a r..Li
Z aa a a a a a5 a a1a6
w a aa a a a a a
inI
-. 40 Ln I
Z Oz
a' Awl :~ :'
_z w
.4
II
I L6 in W , I.
'i I J L I I .- i so I
I ago
I JZ=I 1 I 0 0
I -j
I M 4I I $.- I
I-I 4 ob IL I 0
* m 00 , s
I = II I C
do I ..
J* on 1
Z-
IC
: WLI
WII w C I
in w i 0'
U a c' ~ ca
3M I L me -AZ
I L I-
-
I I acuII
I I
. Ii w. Z
*%J ZV I
OL) I
ZI
I.Z I
0'. I I 9- I I IZ 0'-' I -
5-10
I Go I I I0 If I I 0
I -- co II IY I4 ID I
1 * 1 1. ¶ 1 I 4Al (
V)I
zI 0'I 01I I W I
2N'
0 : Z" Z" "
IL w I I
I I I I I : i
" Iu Ij I.
Cc " a* Ic I- PI I I- I I." - I-
*~- I ý IL I La ~
0I 4II ~ 0I I0 I z~
0.W
0101~ ~~ ~ II I- ~ ~ I.1 Z .1.J
w IA ~ .I C CA~-I
W I (A
W
'
0*
0, I 14 Iq I 4I
~
-
wN
0 I 21 2011
I0 6m
Z.K
I,
4"n-- I
W=1 COD
I (l4A
1 W I I I I I I I
I ~ Iw~
I w I IZ I Uz IZ I I IK Iw
I ~ ~ 0I 9 I ILI b(
x x I xI I 00 I I
Ij I I w I W I w I z
> cc >
z
I
I
~I ~ ~0 1 I
zl
II
---
w I
I-I
I
I
I-I-
I I I I I I5 I108
4 o l in i t- I
N oo 0 4 ) I u ~
4 I 4 4 @ .4
ul La a , 0 m
UCN 09 1 i . : 4c We-
: .1 1
w *i
LA . A
4~~~~~~ 4.4 z 4 e~
4I- J ~ 4 3 44 I
4~ 411 4l 4 0611 l 0
40 0
as I 4 aI
0. O..JI II eq -3 6 -V. 04a
. m I oil In I W- I ) (4
* 4 *a a 2e 2
* L. ai h 4 4
* ~ ~ ~ ~ ~~
_j 4 4 44 4
4~~ I wi w W fil I th
aic 4e 4 4W
I4 * 4
z.. x 4i I
11 4.. I. IK
IA 2 :oz:
4 .4 .4. .4 4 4 C5**409
I0
ILj f-
eL La.,
4c I 0
a I
I S
0.4 5 5
I- in SA I4 I. e
-C agDI a
a 30 1 .01 : -1
0-S WI I 0 I
'~ ~ S NI. 1b4-
* 00I
Af L11I-_
I 0 WI-G.
z
P" *---....I.. u0 4
I 0AN
* I Ik22
I ~ ~ Ii.110
- - - - - - - - - - - - - - - " I ? '
a4
,.U3 a
8I I" ~~10 aa
a-a a S S S S S
w1Jaa4 1-a1a .4
in I I cw I bW C I 4
P- I r I
U~ ~, I Ia
Ia c c I ar"
4A SS S
aI-I
I S S I I I S I
S I 5 1 I5 ~I *I%
I
*~
a 5-111 fl
a a a= 60 I in
* S B B1 0 :
a ~~~ j
* CL I. Si 6 a . w . I I B
a a 09)
a- Sa a. U. w I
Sa 0.a aa4~a
5-z a aa a I- a-a
I-- I U a - a
Zj wn ag a I " I of j I BA 6 -6I 6
a. GODSA a S c ti a ~ a .5 B
1U~ a0 S9
a~uo -j" B I.- B - I
a- -. I I6 a-* . : B
a a I a,UU aZ a B UQ
I
ma I I Wi a m at I
*W Z W Z
zU
a-
wn
a .. a a
CA Oz-I5
a
ma
z
B
4
a,
i
S
a
.J
Uz
a
6 : :
a
w
Of~
a-- aa a-.~a m a- -112
~ aa 4B a ~ 0 a a 00
a- 5
B 5 a-U a:
S
***,n
C1 4A I
* g
* g I I I
Iw
La L I- I I
I II I II
I
(A W
I W
00.0
4I I
a a ~~~-40W *
~ I
Id Ll &16
* ~ I
& I 101
I
0 1 49W
00 I -Ja Z-I
e w 0
0 0
I IJ %.b
w '-
I W W 01 0I I W
0. I
La
WII 1
I- " La
OI 0 INI~ II -
*~L W I II - 0Jd
I-I - :- go
I~~a II
W, z
I w
I (.
.1 W
* *
at ) I III
>
w 1
Z.***-I
I Ii I
I
w I W
I I- I I I5-113
S M CODa
ao
: a(A
* laa
*~ ~ a
-aI a L I.-I aL
*1 aA a ~ Ia
a 09 a-
4c ao
1= 0 aso
I
t0 a 1 a a ( a an
aL i aa
I-
a0.IaJ 00 Lai'aa
_~z~ aa * 9f~1
a w-a-~
a .a L a
~4a
-- a a a a
a ~,4 a
(fl a a
a ~~ ~ J X Soaa
* a5a1a4 a-a
I . . I -. 1 -
a c
a I I I ( 6 I
w me 6 0
s, cc
ag IL
*~~~W N 0
Im1
6 *
.
~I6 0
*
= do I
I in I C C 4
I0 (A IA
4K 6 6 6I
6 I 6 ~~~~~z
3 I 3z
ui CL0 $0 6 *... 6 NII 06w a
I I Ii 6 I Z
I 14A I.
4I~~~
6 I 61 10 0 0 ~Q3
(3 66 (3 6 06
6-4~ 6
II .. J 6
I ~
14
d
z
64c ILc I 6 6
146C.
6
1. I~ 6
6 I
6 I I~ jI ~ .I I~I
6&
IP4
2 I I (I I&
E1I 4 06I 6
0 I 1 z6 1
W~~~~~~~~~
II
fIt
~~1 (l ~I t. 6 ~66 ~ W
c (.W I W
4 I W MI W
La.iL I. 6 61
I~~~i 3 6 Z I 2en
A I 46 4 6 I in
6~~1 6q
I I I 6 6 I.- I I 6 6 ( I 1. ~
Iý I I I A (l (A A
I I
I 6 I6 6
Ix
I j 6I I a
.L Z 4m I I
I0 I 5I *
6 46 ~ @6u (Al I 41 0 I .6
(
0
I
0l * 06 06 0QI 66 p6
'A 1 I 6
.64 6 6 6 4 6 061
b.
I.- I I Io I
0
I I I
I I, W z
Ij 6 6 Z~6I U..i 6 6 6 I I
* I I-ACDI 406 46 6 ~ -l I fdI5-1(l10
~ ~~~~~4
6~ A 34
S6N 1 4 6 6
I
I~~~U I II I I I I
IL D I0 Io do C1 Cd
9K I I I.I C
z W Z I1~ .4 I N I I0NJ
i a * 1 P, 40 I11 1 'I
I 00.
cl I GI NI v0 I I j I JI
i LAW ~ I 0 I -I W
I I
IL U.4 .6
I-J Z I- IVI I I I I IfE

O~
1. I
~
MO4
I
I
U
* Ir
I0
I0 I
1' II
IW
I
Is
I
I,
)
I
coCI Ij Icw I4 4 g, 01
001 (NII W I I00,7.J
ui ~~ D ~ A ~ ~ 2 21
W I0 4 I I co I I I 4 .1
II A w I uiI uI orI(-)
I I I z I I w. . . . . 1 1
* JI L a-I I I I I I
* III I IIM IE IC I
W I de Ic I -D
I 5 (1 CI I Cl Z I5 I I116
- L -. 4
I 3I 33 a Nb4 I
*
I3 . 1..
:=
LUJ
I.-
' o
-4 :
o
-
S0
I
a
WLI
~Q
6. 4E
* N 30 I
(a
C OC. 0
3 0ILI
O 0 0 I
W i U. w
a m o I~
4c
00 4* "I S# 0 1
'n 0l cl
a* Ioni I
0 Ir I 0
II
0 0 0-i
0. 0 Q tw-0
4 ~~ ~ ~ 0 m
00 0 0 U0
0-4~ I ~ 40 0 4C I 33 I
L..
z
U :*0 0 00 0
0
~~3-a40-
S .45I
J
3A
I ~ ~3I o3
~ ~ ~1h.W
z
~I 3 oI W
3 3 3 3 3 5 1I7
Ic
:D in a a 1a 1a
I I I S 4 a
LA. ac I .
m~~ a aa (h aaa
El.
IA !NC Sa
9 0o_2 aj
l a. Ca
_J," a I
4
W
a a a 1 2 a Sn
I .4 0 Z-W91- a - -- '1 1
.0 a1a.6 5b -.0 a& aJ a a
a ae
a LA- a . a Za
.4 LL I 4 IS U.1
a x .. - cc
a- I.J .45 WO Ip 1 S 1ja a
a,. a0
-L a. (..1 at- .4 1 a4a16
a - a a - j a' It 51
10 a a
qw a14 ac .4c 1 C4 a a N. I
a~LL qr w V S-S a
a. 4.4 a -araN
I.-I J a: ~ N 4
<z w a X: PAO
- I0
a a ai aago-
I a I
a a[ a S SagS w
a w aa Z a aý
I W- zi aL z aL
a LA w a O S
f C). 5I "AL I SC~
a I I(f
-4 0. a *a 5(fl a a aa5.4118 C
16I 4 a ; s a a C 4 5 A 5
II
S~~ Io I I a Sa a I "I I Oe
CL I - I Io I4
Z *L I , I
I S S a I I I I I U . US. ~
*b of InI~ me ~I
I I I IIII in
e4 #in~ ac
I U S U
o
14S
a _I CC (n
I
~l.40 5 -s
I..
Ie
a.
S~L
I ~ ~ IC4
WD
~ ~ ~ ~ 0 I N. I~ I~
I I I, 1 1 Y I .C 4 4
I SI S (5 h A
W -c
a j - I, S,
a a WI 06C *
an o I ~
WIfndClWIW
*Wa .a C
f sM
~~ Iu
I
IA ul S
Uz I
I
a
I
Z
I
WO
W.
I
I I
~0 IAJ
I IL-i I
I I I
IL I
W La
0l oN * sa N I
S O N N C E IA *c *
aI M CI I I Ia. I5I I119

a a
r ao
aL aU
.- ILa & aj
cca
* a
* ac
I
* ~ -iUco
~ a
~-o
-".41-x
uw =0
ILL f % I
* a a I- a.
a~ -Ca
a aU
* L a a 1
S
:10 41
D:
I WI
77.
Z S a w I z- I aa ~
a a w a W
a L a Z a :
a--a5.1a0
5-.--- -
6- t
I6b a I 3
44 6 *
It 9 *6 U¶'
) N 1Y
C4I M,
09 : 0. 6
11
OZ6
~ 6 *
V
I 6 s,'
6W
41--
L) 4 6 6 66 6 6
6 6 -46 I 66
II. 6 6 K 6 6 6 I U
6 I Q
6 La
6 6 66 6
6 4c*
9
4 -
Z I0
46 * 6
.
09 . 6 46 46
O 6 4 * Oem s~l ~6 C66 6 W 6
(fCXDR0 I6 6 .66 6 4I 6I
66
M6 -Cia 0I I W
IAJi~so 6o
--.. ~ II
6
4 4
6
6
I
4 6 1660
6 IwA
6 W
uI I .- a. w 6 6 6 6 66
I 6 6 6 6 I 6 .)I
66 6 Z
6
C- A d 6 6 4
I -. due
del I I0 6m 614 640 06-
w " 6* I 64 6 w in - 6
* ~~ ~6P4 - I 66 6
4L6 6 6
I I C6 C 6 IL
06 LL.
6~
I 6 I
~
: 40 64
w6 646 4 (3. 6 0
*
*~~~
~ N @646 *-~NI ~ I' 0
6 1.6 6 4~6 @6 45IA121
6 6 6
6-BZ6 6 6 @6 fl -6
09 in4W ex a m
CL :) II I I
043.0 at a .0 I
I acI
W I - a IL 0 Sa
j. L& W I- 01aj a Li Wa I *
< S C
.4 I (lLaXIlUJ
Z~~~zs - yf
I ~ 01
W-II-IS I W-.h-.4 I US I
C j 1 001- a I q Ii #A I I0
'j~I ' 0U I
Lflj
00 10-, I I
.C I X an5 -.-. I po-I 8-O.Z

ZIa Xa.' aýO aW I-.
(n-~ IA. - o-a
I LL W M: I .
II a I II I I I
::ow 0:6-w
Z WO Z WO
0 I ZI
X SOl IX Z 0 Z M
Z > -
>a
515 >LJ I 4 I I >
zL
CL W 1( c 0
S:w: 2 * w 5se122
I~~~ýP
Sm 4 aa 'IsI a
* Ju a a I
* a a aa aa a a a S S S
4, N ain0 flo I ~~~~4

aI ta Ia
a - a
a a a a a a a, a
al ao
I a I a a aoo a
Ca ~ ----
-a -- a- a--aa a a a a1
a+ a W~
IZaa ai a a
aCA
a
a-.
a 2
3~~~~~~~~ a c~
a fa ej xaa a a a a a a -
w ~0.w ~
a-4 a' axi
IC a a~ Cw'5~
z 4 a (4 a W aa a a a a a a r
a-a *a (~~a~aZ *a a a . * a a a a an
a W a Qa 4II- a * aa Zaa
a~~ a: a4: a aa a a. Z .a
a a ~ a4 (4 a a a a a a5a a1a3
UI
~~~
-- sg---.o---- ~ 1 ol-8-'-8-*
8 ~~~ ~~ I
* ~ ~ 84 ~ 8 4 8 8 U8 8
*J in 8 CY 5 I U
,A r-s I.- 08O NOi 8 I
m
O O 4c I- 8. I i I
5- i 8L U.Io~
L
8~ ~ ~~~~ 8 0 fS @ -z -z
W 0gj 0 of "0 1f VI08
8 8 8 o f -Joe C- m
2 IIV wj 01 40 0
N 084
5 ~~4K S8z5 ~ ~ ~8 8 ~ £
88
Z
I
SO
u 8~8
8 .
~ V08
a.
8~
A.
8 ~ ,~
V. 8 4~-- -8-8 -8-- - 8 ZI I Z
Ilk of ::: go :~
IIJZ~~~1 £50 8 @ 8f. @ I-8
6.408 1J
10 8 0 O e~
~ $32~~1
8 841. 8
* * 8 5-.S
~ 0
085L-4
8 8 @
0
~~ 8Z 0 5 0 I @ 4
14 8 f 8
: ~ N
*~1z 5jJJ 8 z
8 8 855184
w'g .0 C
a ~--. A . IL. . g~g

aIO V N
z
aa aa a aa S aa a
a j a9 W. w a e aI
ll
*I a w a a aa a a
a, a ~ aa a
, - * I006 0 a 4L6 w"ca a
aactLl a 3 aS a
5- a.c aat 0 a0 .1 a aI ag P.
14
:
a a 1 a
I Ia ,
0
ac
Ia. bel a Il IL I
-C 00 40 00
* oar
- ~ a a a .- a a a
Z~u ae a.a h i I S m
% . - a * i . a ~
14
a
f~IJZ
5
I
MI
a ~ a ~ a IAJ.
64a~ . -
4c 5 1 U~Z
ah -4c a - If.a1 a5 w 55 5
*--a 4a a§ a *-.- ~145 5 Saa
a 0 ~ 4a ma a ~ 4~ a atoa-
aW a, a IJ aZA aZ a a Zw
a~LL a a ww * h 5 5 5
a ~
a~~ a
a , a)
----
aI
aaj ZW
x .
.. ~~
aus a aa .Ja
a 4a aa aa 1
--
U'
Ia. ma
LZU :~ :Wa
~ an 41 a a a k~ a ') 45 a1a25
Z s a- B~-Oa
a I a a C 0.505 aVC
a . w Ia I.-a a*a
ago-44cf- -. . dc"A -
m'ece0 =,Coc
a1& a LA aj a a
~ a a a
a aa =
z aa
a a
La a w
a c a a Z a
u- a L 0 a- a aL a
Go a0 ao a1 a~ - owaa
l a
a ~ a ~ c 19" -j a. a
C~0 aC a00a 'a
- a aa a -a a 0~wZ a a
a a ~ asaJs.- a , wJa.- a afa
a
w- 6. x a
a
0
a4
a a
a
h .
- I-a
a
a .
a1
a
0 aI
a
a
aW a a i w a ZJ
aL aa -.-.-..-- aI---a*a
aUj aa w reoae
a) aUN a (.. aý aW C 0
aa a w a a a~ a aa a
a -a aat a a a ata
ae a 9 a. 1- a s-
a .a aUw a. U
aX a( S~ c LZa U a~A (aa a
a wi a afl Ow.(.l W a C1ffa
zaa
a dca ~ N 42 a ~ a *
a a ~ 0
sa .. j~ * 0Q a 00 a~a .J~
a a a 5- a a U aa a -1a6
I I--
*~~~0 0
Sj Ie 43
m
ZWe '4 ~
wn 0, en I Q
0 4e I- *c
-j LA~. e * *
S I IL I. I I go I SY I
S I
CL 4 L" I e I I I
fl o (of 46 0 #-e
soS i go
ze
C4 14
0= 0I1 1 Iý " I
CD 40.
7- to-,j
00-44L '
I 4 4 * *
-- I
S NC.
1.,UA
QI~~~ INC
----.-- a-a --- a - ----- a -- t

ai a ama
aa aa a a aa a-
I ~ a aL I IaaaaaI (l
Z~(d aw w. 6-
aL aL *a a. I W o-- a
a01 aL 340.1 *a V I
a LI i ý0. I~
a- a1 a us 64 do. Up Ia
I at "ag
4 a a w a-d a
a ae aa A
ai a a a a . us a a I
., C. aa a:
Z w
a~~~0 a Zaa a
CL- a g I do a a o Ifa" - 194 a0a
a-aL a aa a.
a-CW a a a- aO- - - a aa
w~o 0 w a ba~ OWaw
- a .a
I -- a
ei
aW.a-
aUS44 .
a
a
a
aa
.aa
a- a
0.UL
a a
a
a
a-a
-a0-~-auB
Z-a
.aa.
A
a
a
..
aa~
1n
*~Z a a. w a.wa a aC
a aa a a a * a aa CL
a ~~ - ~ a ~
aa ~ -. ~ a t a aOw j
aj
Ia aW:
~
a ~ a(1a 1
a
a
i aja a
(
a a a a a
a..a
a a a -- a--.
a W
a>
a(9 Uz
a a
wa
: :
aC9 a
i' a ~.az
a ~~ ~ ~ w ~ 4 ~.ala~ WZaW
~ ~
.aLaC ~ ~ O
w a ~a a C ~ a -128 ~ *5
* ~ ~~~I ~~~~~
.. ~ LAi
a Caaa 0 0 a q .~~ ds
" a a0 a in aa
*
Ij -- 4 CA I II
-4 aO I
a.I LL.a aS
aLaiao, a * LaU
an uj L * I n
aL tLiS S L"
LU
-4~0
a a S _ __ a. ~~a I0
M AEC4 a aa ai
a ai a6 w
-JL0
a LU atava
aa a a
a cc
aa ~ a ~ a IS a
a a
S Ia I
4c
aac
a
aa
a------------ aa a a
I- a
~a I ~ Ia ~~aaa 4 at aI-
a a Ka aa0
cia Ob .41
Ca a- ao a
ai a a a 4c-a
Caa ac a - a-G a
aI ; a uaa
a ~
aJ -4
w- a
a ~ a u i~WJu I
VI w
a z
az a
a a,.
A a a aa a ta.a a a L C
a a a it a a
a~~~~~_ aa wa aa Ca
ai a- aa
a IL a
> - -
a a.a
a a ~ ~ aa c a Ia w-...--.- Z a ~ wo
a C.) a a a a a tj a a
a ILB
a a
a
a~~~u Ia
A
L ai a x x
a : ma
_
a aA aj a aj a
aj
a a ~ a a a
aa6Ja
a aa
0- I
a
ai aL
ac
aa ~
~~~~2 alala a C ~
.
aa a- ~ 0 oa a*a
a a~i a ~ at aNa
a ~ Ca -a a~ i qa .a5-129 C
ZIZ
Z
LL. * -- - s a g- aor-a- -
~ SI S I S I I S I L.
w S * aa I
a. 0i M
I i aL .
a-~ a a0 S Sa AL. CA a C4
be s a a a a s w .
0~~~~~~V S S S a S 'low S Q
z a, 4. o- 0 C,
ci Nan
1040S i .~e~
"W 1 a,I de 4 *LC
W~J-I a 0e . In 05 .4 5 I L
S ~ ~ 4 m
a ~i s a a I a 5 p 5 )5
IL s a a a s Z
C S I 5L
I w I I IS I 2i a ss a
Ii a.(
w La W L5 us mi w Z
*j a ----.-- a'.i-i-.-
* s a a a a a s 0 a a s5S13
9 L
IL q - 0 -I.-I--
Z 4c Ac (A I
. 1 a 0.0s
-A45o "I 4WII :16Sn
AH 1 l41 nIwI I I
I S
I S 3
*6 S
* 0,
x I = 0 I
W a I Si W IS A
Sw Ic
1 c I i I Sc
I
5--13
-- -1-- 'IL a-- -- -- s -a - --in -- - -a-. a-

0 a Ca S 5 a5
a 5 5 aa
a a a sS
aa a a a a a a a a a
ac 5 S 5 UO 69
4~~ cca N4 a(
a-IA a s a -
I- 5 L I 5 5 I U.5 a
0. CL P-h.
(A CY IIu~
a* s aL- 0-i - : .s.
U I . 1 4c dc I
hi-55 An 1- 4a 49 W= a 'CO It" d oJ5S 1 4,0 a aJc
4 do i aa a a4 .4 an ac aL so Iic I 5cL
a. 5 50 (4-d
IL 5 C, h (
a~0 -, - a a aa 00
-
5- SjU. a& 5 SI Ua
*P a0 a a C)c 5 5I5 J
a CI aý a a c a& a a a Uaac
a ' ~~~~~Z-
~I N 5-(A-~ f
w 55a pt~
ui 4 a
sle-5Ga m
5~Zi 00o1 fa
a I~I
4
afa
5
~a
G
.5
-.
M
55
a 00
i..0
a005
5 (5
..JUa
IL4
M
(a aW OZ a 1 a1 aW aZ a a a
a Ii w Za
ILi aIa I hi *5I h. I5
S 5 9-
>-O
a
5>
uj
S
aW
M
a~~~~ m
0-
UIL a
O~5
a
:Ca
aa Z4z
94La
ZU
-4
i ja a
~S5
Ia
O
4 all
aa a ~a ha a a aa w54 a 45 h132
IIS SI I
HW 1 96. 4I S 5 S
dc ~ S a GC
. I S a
ua WZI aL 0 w a I... m
I 0- P4 Ij S0
Z Of
a0 I t ~S 4n 1 4c a
I - A WOII
C
-C atmac
goi
1
.Y
1 Ja hi I 1- 0. X 1 I
-! 008 m 0 1.
O- a- al a . ..
..
hi
a
J
C- I
a
j
m
~
a
a C
a
a
m s-
.aa
I
:
a wia
69
a gVJ
'Cm z' L
aw ~ w.i Or L)-
C~wa z 0
wag0 4
I a
I I
IL
a O w aa a i
I I hi - S~.J I S a 13a3 i- a *5
Cie -- *-13-O)
3 L Z I . I a 3 3 I I
at CL I Ia NI
ac3
* ~~ ~ oI m * Q
(~ I 3 I L3I I.
*. A LA I. M e u
Ua
s
I
K
~ 4c.
~ Il
I
(0
c
I
IZ
co I
3 WIs
NI
I L WI
I I I Z I
I- I-3
t 3 I 41 I 1
3M
La. . (A
0- 04c 3 3" 0
> I u Z I I OD I C 3 I a 1 I0
0 I I -I - 4 Icc
4 14C 0 I3 In 3 jx IZ . 00 co
0. 0
-A Go (%l -C La 0 43 Cb
Or3 At
'.
I.- (A I I :, I N
Wa 4 (a0 I IW, I
3 o - I4 I M
La 3 3 ~ J
Mix*
ccn 3 A l @ 0 4 fI I
w: :: u: Cc-
* I 4L IL 4L 14 I I
*~~ 1 a f
LL;'OW 3 N .1 4
z- I I WI> 0
5-134
I I P% I ID I
e ~ ~I
~~~ m .. m
0.
O I 0 I I0 I lo :I . . If
I~~~~- C e C
e I, I
4E a
I I C II m s S
zD IC
e~~~Z I , 01
b£ 0b-I4:0
I ue CI N'I W I W
Z
' 0C N -mine
lI
*~~~~~~~ QI
.c -J ~ I . * l ~
m C L I.-C Ii I
z j * r IA a*S: I WI I I
SJLJ
it. CS N b eJ
l 1-- leI I I I 0i -i -
I
I I I
=- IZ
_I UsI Lc- l e
Te Za -N
I-c
4~~~~L I0 I u.1 0
I o-4 I e eI
I ea IL I.. I, 1 e
0)e~
I-en 0 In I
fe e) 4 1
w ec
e 00 C,14 C,
too am, I It
Z "t
I ~
. 4 Ir 41 t- inI I IL
0. 11 I Il
1oe I
C--a *1
co
w2 ~ feN ~-il I ~o -1M 010 )1 01 11 0
00 00
cI =I
x x:: :2 in
CL I wCA I I I YAP
I1 1 1 I I I
I
I I 11.- I .1 *I * U.~ I5 I I135

*L -0 - a ao I 1 I S
a LO ai Q a w
L6aWI
a a L LL, a- *a a
0a *aa
*aa a m
0 f.
-a, ama amaa~
I aJ a q a. o alu 4w. aI a
Ia c . D a
a -i 6 a a a aa a C
a~ a a aa a a a
a- a-6 Z a- a. aL a - a a x a
a = a a .. ac ac a a !2
ao a a a1 a a18
maa, a
ca
U. a Z a aa fi a
.Ca a_ aO a' a a O
Lai a 1 J
0 a4aC
w ~~ ~a 0 a ~ ca~ C4 L16Ja- a .j
a 4 a Ca L"~ Caa .w
fl Zoa-w 1 *a * a >
a UJJ-b- a 'Con aa w
aU. a w at a a w
a L a a a az
5-136
6 SS OW I
*S
91 8 :
I *L # 24 1
*WM do
UP I SSS
S- S 4
P4 1
r
U.
0-4
S
5 6 * 5
-4SzS5S I - ss
IGe "S I. Sc c w o S I
SJ I1 I S S
.,W w
5-e I M MS
(A p se go s I.- I
5-4-4 S S 54 0 0
I.- UUIS
-Cp
Z S
* A S.
I I
aJ Z
I- I
M 44
I* hJ 5a 5
55 -137 5S
at' wVtwwa a 4F *
'IL IL C4.
Z U5 1 0:t aa *
mL a aa a- an aL aI aa-
ag
at=4a( I. a a
aa
ac
aaA
a t A t-
(
a
.a a
- 0aUa a 0aw
a &~Z ~ a
I.-aa
a a a ~ 4 .
af.
a
a
*
aA
a a BO -
aw
W I ~~ Q a C
aU, a~
a0 aaa
a aa a a a a a&
a i aa a a S S
ao a aaa
. * aaa
a a4 a aj (Ia a a
a
ga
~ ~ a~ a~ ~ . - a w a o
a~Z.c
a- a alaa a w a a C z~ C
a& aa a a C I
a~~~A
a0 ac
a~~~~~u a a 4a dc
a, La. a aw a.a y
a a - a 5- a a = a -a a I
aa a- a a a 5-1a3a8
a a a a a a a . - a - -
* ~~~ - - - - - - .
* a q~
aa a a a * a o f
~
~a~ ~~~~~
~
* .41 aa a a a .
ow
a ~~~~~ 4 a 03:
3I 43
MW 4
4
4S 4
o~~ i4- 4 W:
I I
7 a ZI Z 4 Z7
LA. I4 Z
cc LL I" * Z
or
L1 I WI eI 4l Ina~
4- U. I
I ILIAJ 6 043 4 -
7~ 4 US P.lS I 4
a Aw
41. 4- 1 440 I Z.J
U. .W 4 is 0 a .J
=4=14
Wm I * W> I 4f I
Az4
06" 0I0'4- I.-4 5 (

I de P3 41
4*. 4 0 6 44 4 U 0 a 4
I 6 -0c
I -S - -
LAI 4
43 US
S 4S ~ 4 SMI 4L"S I LA,
63 I 4
ag 6
4~~ 4o 4 C"
0
I44
I
-a "W
~ I 6
0.(
4
4m I
*i to i6 A4 4. U. O j .1
b13A.
w $. CL x4) -4@ 1@
00 3 4 j 0 04 1
L.8 43 4 4 CO 6
4 I 0,,
c~~-4**
..... . ..
1..
4. 40. 1 -(0 01
%P. . 4
3 I
w .- I I 66 U I I 4I .4
4(4 6 74 CL 4
4C 0-
35. z z(-
I~ Wf I
U. I 5.i
I IL I IL
. - a
I. 1 a
M I -.1 *z--*-- I ..B
WJ I I.
I
( In
* 4
4 0- - 6 w
4- 74J UJ
5-1394
bC 67
-g2 -...-- R -- -- a
a a a a a aI
a aa aa
CL en SI 0
0
a a a I I aZ
I. CL aAa aA IL aL aL
l I. ac 4 a
JW
a a aa
- I I I In a a ~ a
- I ~
~4
4 a *j a a a a .j a a a
LI aa a a a a ,.J aLai
Ii w
I L6~( me a a a
a) aC me a Ia a a0
*L 0- CLx 4 a a& a 1IL z a4
W U.a
I
a I I (. I n a a; I
*~~~z
I a a a a
I 4L
Ca 4 IaI ILI~ Ia~-
I aij NLfza Ci
z m
000 I I waLJ w
0. aI I a I a4 W
-I >
a> -- a- I aa a >
--- I-
aL L a Ozl lal 0 I"

zLA I a
I
OZ.0
* a zl-
a- *UP I I I 140f
ZS-5
w 1 0 5 5 5
B4 0 I
W B I B * I 5 5 B I
ZO
a-.
B
s *.
W n
s
O
m s Un" s ~
LBJ *5 N ' 0'a :* I
I . I > >
O ' -l I -0
-- 9 44L0
~~ B Q O 0. : z
WO I@15 .W WI 5 0 JB .%
C I BAZd w 5 A
.4 0 IDA 'B I W - 0 -- qWI I, WI
-A 50
c00c
W I * 5 Ba0 W
a -I. I0
*
B
*i
cc 5
D4oI
Bn
C5
B
B
*
01 0
5B 5 UB
3
5
03
Q .
M%
c1 j
C
* II B * B 4 I B 5 4
0 B0 5 J
,A
4 4
B w
B W5 ~ 5 5 0 0 0 ii..
w~.... -- Z ' ZZO

z
e
BO
31
~
O~~~~~
B I 8 ~ ~0I 0 C 4
I~
~ ~~ I 0I 1 2 n n
O
Ow
O B ~ a X
W , 5
X
W . 5
X
5.5
**
B9L
*~~~ 4 f ~~ff a ~ 4.zw~0 I 0 I 0
Z 0 6 * 0
I W3
one Os Co *6 0 5-In 141 5 *o 0
It
d ~ -- s-s -a------ -- a- a -- s-s--
-~on aaa a on,
-~.4 S 55 * :
10 I
C
5
a
~~~ ~ 1 1, 5
a N aa
55*
*
I * N 5 - 5 I a
*~ 5 ~ . 5 5 AU
U~ 45 a N 5 49
z w 5 C) 55 ~
ao ac a ac aOagg
&- ac i- a- ao W a
a aW
cc ac
a a-- 04 aIn
g.4a a s- a Dat a a ga5 -1
0 00C# 58 Tt 4C J.f w
co Ja. W 5 5 55 111U W
-.
s-a 5 N 1" so i go N. W
4 a. a a I an a% 6
alo aC I on1alo ac
LL aC a a' ae &61 a4
w L a sr a s a aaM.. 5 I ai 5 .a
a- 5 I 5 5 a- ~ a a a a
-aJ a a ------.--
w
a I0
L a 5
a 30 5 Z
a W
CL
5 5 a
5
aa
W5
a
oc
a
5
a .~ a ~ 11a
azl aW 5 0 I Z I
a~~~~A 1 @5V
o~9 O .4 5 C 1 4La
M a
C.) a IA Lw".5
5-1 aa4a2W a a A
: u iI o I S
I *a I CI0 4 I0
'A I do I ao
m III I I 0 I
I 30
I. I.- 5 S SI
I 04 I le
t- II 0 0
D" o I Ir4 Ia
g ~ila...
99 44 I. SE I I
*
C S -8144..
* I U. I S I .J. W SW
Lh.I
LL ataC
Gj~~ I WIA
* -00z I 0
I i S 5- II I
0- I Iw MI IA S C I W I IA
0~L 4 40
~ IS
*
IS
*
a
I
-.-.- I-.'I>eS5
5
aI
5 InI-
-.-..---. C
S--.S S S *~ I--.---.S--- 0
~
S
SW-
~~~
S
4 0 ~ .-C
I
5
SS
05 W
.
£01 , ~4 0
9L1
::'>
. I0 0 C0S
SC
00 a 0 4
aR CI5 I
0- 0-'
.z X a- Z>
CL z iL
w Iaj Z: uI C) CO
I 5 20
0- W~I 5 0 0 2 4a143
a C lo a aa aa
w
9L a aaa
aa
a ea L
z M
z~ a
Za LL a LW w
Zw a -*---- I.
I-
a L I.
a-a
6-4 0 S 0
ai ai
a~ a ~ I- a
a
a ft 4c a- a a 4 a .ja a
Ia a 6- a K
I a aa a I a M S 4 $-S
Q a a 4
w
a
a
a a ww
~ 6-I a
W
K
1-
4
~ aaa -1
a ~a a
a
a~~~~~~
w -h aa aa f~Z a *
I 4 aa a 4a a
La :~ La a aah6 a L4 a a~*.
wZ 0 1 1-- a, aj Ia.-I (A a
am ai aO I I
a a. a K 4a 6.L I.9 a
a. a; a t" a6 61--a*
0--.a--a4 1
a.. J a a a C4 1ac f ae a t a a
a a. a q4 - 4a a S(
wf- a ma K D a taC-M 1 a £
a (A 4- a 1 a CAa-
0 a m ~ .a a
"w I 0 u
WE L-4fla
a ea Ca aa 6 W a Va 0.-CK a C
K~OV a.4 0 00 a a
mi ag 6A 9
a *---a-*6
a a a --- a--
aa a a a I a
.. wz a *aI L. aI
>-~ 4 aa a (6a 6C a 0-
zK 6>64 a %a 4 aC
a 4ja 6L Ca K. a a Z C a
6 SM 6 ~ 6 a SM a44
II ae 'IL z 0
S .4 S 'K
I. cc
z a a w,
I S *a
-J~-
4 -J ~
S" La
* * *
a a. .6
94a
,
dc
I. a r
S de I
" - -JcSt
JW -L
00 ~Sa 00 ~
1 0 :
a-J in S I "**
*& Sc S I =SS
I W
Sf w SS
40. 5
SUA 0. I 03 4A
W - I. x S S
J . . .
a LLA a.S **-S~
4.a a%
0.I a aa a . a
9L-a
.~
ZI a a
I~ )- II I" I
1 S
q V;
I a a 4 a.4 am a a ao a3
La
a
~m
00. a
a
a9aa
a a01
aL
a1
~~
Cam~
a~~cma
*a
naa
a
a
w
4E
a
~a
a
a
aI
~a
a
aa
ag
aa
a~
a
a
a
a0
~~
aw
ar
a
a
-j.
C0~
-m@
~a
O
0aqa
w (A a
aj
a
a
a
a
a
Caa
a
aL
aI a
a
a
a.
a.a
a
a
a
ac
a
a'
>i a 0 1 ao a aO Ia. ac 0 a1 a a I
P4~1 a a a a a 4 a
M4.- aa a. a(~aaA a
a ~ ~-aa go a aa.f-PWa..5Pa aI a.a
**---a~~~~~~~~
aa a aCaqa
a a C Oaaa
a~I a aa aa
-.
q ~ 9L(l
~ ~ ~
a ~ ~ aa ~ ~
.ad .~ 4 qaaa
a a a OaaW aa 4
a
4i ac a a a 4a a
a a zj a L W/ a a a a w.
4I aga
a0 a a, a- a a a a
a ao aw - a M
a~
a~ a us a
a a- aU
~ 1
s a a Z a axaa
-
a u
-Q>-a > LA L
4c -z w za>x
I w aK aj -aJ
a a a 3 a aa a a (31a6
1. g a0 * 100
a K aA I
0 aaa
~ :. 1, 99 ~
4D a-i
La
U)SM - ~ A L a
I a 1. 6.4 L
-4i
.- ~~<W
a a.a :
a WI Ci _j LL
* b I a L6a--
W
~~~--S~~~i
aa Ku-a J ' Ki4 Oh
-Joa:) I qr I a-D a a a
as -01 a " a
a~LA I ~ a Z -Kaja
a
a 0
aM a
a aK a K a K a
a a O
--- a-
OV-a--C4
a~ ai az
,z .
a
a~~~~
~L aw
4
a-i
-o a a w a
- a(
aa i.a a K
ax amOy,
U.
0 W . aa z04 c 04 a C
a a *aLL m C
a
~ w0
a WZi Ca 050 0a - U-a
4K NO a * a Ka C
.. a aa
.J4a .~aC
FNam
4~~~~~~~~~
V a -a
~~L a- a ~5-147a
I a) I a
aa
I.
w I CL a naa
jIwa I,
aa a
In a a ab
aL a 0.- X I
Z a a~aa a a0
~
w a I- a -a, a a -.-
a a
* I
w
ig 1-
4--.----5-1a8
40(
cc In
z~~ 11.Z
U. * 0
:0 SI 0 I I
0:
(I)j
=1
4I I'-0
6- t.4 I. 0 0 I 0 I I t +D
b-I
K L 11 A 0
I M 1 0" 1 go. I I ,C .
1J Z I N 1 01 w 1 l L.- 0 I0 0 C41
w - j I0 0 I 0 0 go
QI I
0 00-
0 .1 1m .0I w1 > 0
I I" 0I W WI
I.- 0 I I
*
0
0
4
0 I
0
I
0 I
010
0
0
0
I 0
Zo
~Zcc 4.
*.. o 0 0 z I 0 0 I 0 0 I 4
044-fI
I~ ~ Lai tu La I b0 S b4
0
0 ~~I W 0 j 0 A 0
o-
W
50 0 I 0
LI 0 I II 0
cc b- w 0 1 0 ZO Uz I I LO 57 I 0
I 0 4 I o I I I 5-149 I 0
* I I I I_
z La 0 a 1. 1A f.
0 D ac a - ,. .
I4 I I I I
Ck
6j I I I
z~aa ~w a -s aJ
o a : I
~0:
01
l
aI
CL to I I
I NI
x~U I )
0I - &I I
I
& 03 1zI I I
* I -I I I, -I S
I I I11WAP-2 Iwa *: ."V

* 3o 414 i I3 n 1 410"
a a a a
3 0 I as3I en I ~W a1 i ~A
Vi de ," a a 3 I NC
4z a 096 1 in s 0 * 1 N 0 WS 3l)
-4J 1 a4 01 1 '3 *3 1a w im 0 3i0
46
CD.
a d O
1
a um
.3
0 S
.
a cc
3 .3a
ig
a a ma
i ~V3)
w ~a ~ ~3('3 ~ 333 ~ I3c0 ~ 3 ~ ~ "a "0
0 0113 a a a a - i 33 *(i m m a
3 L~ I. i n 1 -0 1 o f 1 -1 a 1* i"
I- I iI I4 C' m X
4K O3 I5a a a i i 3 i O
o~ ~ ~ ~ ~ ~ i~ 3C) ' I ) a
UJ
3~~~~~4 i i3 3 3 a 14
iA w i ~
-- a
I ~ a~
0b4i in
~0) 3
q Ia.m---3
013
qa
Mar1a
-a
C,
Ni
3
r3
ac
3 I
40 18 coo 3
z a a al 3r I- V) I Go ',i
ac 31 -Cm 0) 0i 00
04 aa 1 . a :c
a -.- It
* 0 co 3" I I 3 . I I 3. -
a a am dI n 03 l I I
* ~~~1 Ita Io- I
4ia0
"- N3 03 m
I 1ai an 30 30 4(
i-.Z ~ Ca Mi 0
a
3
a0 ~ i
~ ~~~ i
~ acc 1
i ~ V -
0
3
m i a M
a al 0 M I t, f 34 IN I
Ini 4 i" 34 a it. 3 3 3' a
aL 3 a a3 cr.ma
j(D i ~ 3 £ 3
IM-
3 i I 3 I3 .3J wa Ij-. .3 I a, 0 i 3 ,'.
1 (.) 0 C 4 WS-s 3 4
I a 0. OI CL Mi Ci C3L 3 i 3
I I-I Ia ai Xa X.- X.4 Ni ZOf 3 3 3
Z a L,)) Ni
u'I 3 P-a
L 3 3 Xi. 3
3 i 3 3 3 i i
3j 3Z 3Z~ 3 3 z
a 3 m~ ~ ~3 i .3 ~a .~a
5 -15a1
OW
-4
I a a* (Ao S| _~~-Z E . ,-
z a a ae
aZm[• 0. "a -
d• u I-- Pt: z Me '..'..
Z Z
w" a. a " -- . -
aa* *a
a a a - •
• - - - ~ a 0 a a a a . . - , , " ." .- -. ,-.-, .o - " .- " -. % . .* .'• / /
. -a
. . ..
a a"a• ° "
a O Na.
W W - .a
5-152 -2:..i •
------ I----
5 5 5 I 5 5
* s s I *
I C C C I I P-I WI
= 5
5 () n s C
U)
C *
0
4u II 5 0 C s k.
C4~~
a 5 5 5 5
~~~
I 5
5
55I
5I
5
C I SC II OW
s
.5~~~~~> 0 I: 5 frI:IC 5 , 4: >
Cs
, us oz
U. L& U.
i .A
C-C
5 I s 0 4 1C
C I 5
C- I Lai
J 4I~ q I 5 5 ~ Q
~t.
I ()
5 1
.-0 CI1 I
II
* Ci . I I .I
1q 5 , , A " -
I1E!~iI~i!4w 1'a)$! 9L I-. U,
I- L C 4 ey 1. to 1 10 (
N qt 1'-
in I lb I * 5 I I-4 U )s 5
I~~~ 50 5 1 - I
5
Ma I I 5 CCl
5 5
5 5 I W 5 S..5 (a S. 5j
C W5
I 4c. 4I 1 I
mc Ce L ,~
w C No~ I s C~
5 0 ls (i 0
Cof
IA
5 ~I w
I
w w
I
La
C
Z
I 5 5
IO 4= Cz C w
I
Os NI Vs P-C5-153 s s
H CAOL-C,--
C S C
-00=0 w 40
40 i
*~ ~ I x
1
C
40: CCC
C W CC
.4 a
WIC
C
SI
.. a a ~ a a
J m
CX awr C a
0 Zl l-
.AJ a . w~6
~
I.- ~ HC7 ~ I. I a C- 6
1 6 m C & . m
C- CoutH Iw I
I~144b- 0C-I-
CC
0,
I
oo6
=1 H1
H
Cr
a0
I
f
H
NII
I-.. I4-
1 14
Z - H a -H
*~~~0 6-j " IO H C 6'- .

H1 OC .9 = C -f C 2
C i w I .)14
a. a
C
C~~ Hc
*~C H HIC 9
H .1.- C - I H
*~~~ Hj XL a.a (~ LL 11~

H ~~ C) co WI HCOHSfH
IH ca1LA 0.-AC
Cz Ho a W -= 6 0 C
* H 6 C C-. a 6 SC a 5C 154
I IL IZ I La
w I I I I I
L
a 1 -. 4 :) -4 aa a co lt a0 a)
co a aAa C II
z
Zn aa Ia a
*ia a a s
I
UJ
IL
de
I. aA
(~
4
az
I I
~W0 aW
I
I '.4
w
0.O
Cc0I
~
a.1.
a~
I *
.1*
I r. Y,-
cc
Z----a-a -. 0 1S
Q.4LL a 4 S
a-' aa a a- a 1 In4a
~I I I I S . I S I.-I
a-.
a I a a a c ~ a a
a a6 p.- acI-1 aa W la1 0
.Jna dLI VI. I Z I.- V,

4LJb A
-I - L
I.I I I I I -C I
I L I IA - I a L ai L6a
I W 64 I C4 I I x SM I I I I
xl1- 1I -I l Va)~
. J
9I I I -
a I.. I 1l ia ar a) W I4 -. I
I cc a 0 a
I a aj a a W
I LIII NLII I I
U. I i4
Z -
ZIi
a
aa
a mn
" a a. In ~a A a aa
z- , a a
Na
a S~aaO
I S a a I ~ 10 a-L55S
IL
ce A
-jU. LA
ae I coa
**
-a
Ia ca C h
a IZ aWaa
4 .1a z
a-4x a a0 a
lan'
Le w
~-5-15-
:m -lw
II
L0:C- ina 0 F
Q. I-- 1 : Cal a1
I. o aa
f -
M 4c lz &I I
az a a 1
La a0 aW 1 Z
O acZ~-.-.-----aa a~
9-a 0 a a a a a a
*a a go I
Lu Z a1 af
1.
aC
aL
0
I, aI - (A
I :I :
A
u~ aIn a~ cc I In I
aj-j a a aý a a: 0 4,
*. a-
0 1 4W 4'
U, 1a a 1 a' I .. a a
* L Z- a 0-~w
o-
a1
a .a -.
0
Na N~ a a .~ s
* I I as a a aag a aao
ar a.
1. a aL a
fa I 4
I a a. a a a-aaI
a i :1 a a UI : a. 4C1al aZJ
a a a a a5 -1a 7
a a a a a a SO
a a aOI
4a ve G. S
I IL o
*~0 a aa CO 4
z
Lh IL 0 .0 I- -44 4L La a two,0
I I a
ai a i
iq I )a
y a =" 0 it
I( ý d I II 4 I 4
0a Z-.---a-. 2A ~-
*w
: --
Z Z
* 4
IL 1.- I
La -I w I ** 14 I :
Ic I0 a ar 10
00 -JC1 a I I c a4 Pa *1 I I
I ~
~ ~ ~ I I
j W - I IW _I
S C) 6I4 I SiS I S - I i
IC OM I lo:I I W W= Lon SS
Go I 6 4
S O@J S I* 2a
41J: 0 aW0U S PS W a
I I I L I I
= I I I I I Z a a a S
-at 0- Oa..=a
co a 5 a 5- 0 a a-4a, 7
a- =
Za I -
IL I z4 a I I z4
>
CA w -K I aL am VW) I w
m~ I-~ I I10 wis 4 .J "I.-

a g u 0 - B W
cI I
CL 0 I nC I
oa j I BS 1
I I--U0 :u
0- o b I4 B 0
Bc Iq 4 0 *
0 cr NB
(a I 1 I ODI
.5J IZ
W-II--4 ~ s 0 4 I W
-- ~ *
* ~uI mu
B B I
*4.E
I S
O J" I I w I I-I
Or I- l
- I w w I ccB
I ~ ~ O *WO am,-(~I
ZQ I A I 0
cc
2 ~IA 0 x
*
: W
B
O(SB
B
0. --- ~--B--I..
z
IjJ W B S 0 45 11.I B BZ
o-o
aO 40
u Os-u5 -159
a a S I t- a a a a a
am aý aD go a a a a
b
~
a
o
~~
a4
1a0
a 0
a
.~d
a1
aa
a4 a
a
a) a a0
M Ua a a aa0 a C
e~~~~~j~1 a a a aa a
O a a a a a a a a a ga
M a aa N .a C h
a za aa ~ a G ~ a a a a w
a ~
.~a J Caa (la
m (a a a a Ua"
aL6a * * a * a a a a m
m -a a wa0.0 a * r-. a a a~a6
wal -m*4
fi
P-a (
a
1
a
1
-a
a ~a0 a ~ aa
a)
a a a
Ca
(a
a
a~
Za mi . -. aa
~~~~~~ a a
a
aa
a.1
a
aaa
a
a
a :
a1A a 0 a al~ aIa*-a - a- la0 ma a

1-~4 a0
a ac a -0 a a I
(.)
a1
4~
a 1 00 a*
a, a1
a a a 1.~ a a a a C
aa ~
a a a 0 a a a a a y0
LA a a a a a a aa a a a C
a)
a ~I.-w a ~ IN a Cm aaI a a a W
ax *I *aIa a *I ar a
Z- -. a A.i-a4 a
vap4 a C Ga G aIa .a ~ faaaa a a a w1
aa ~4 a(A~a ana ~a a a a a
aI b. a a a a a Naa a a a a a
aa a
. a a a a a a uC.
a 4 a a a a a a Z
a~ ~ a' a wi a a L a aa a a a a a~
a 4-------------------
-------- a K a a a a a a
:j
*
: I I 1
I
24
I
: -0:
I
It
". 10
: 0
s i
49
mi I 0
a,
.
s
* I W I I 41 (4 4
OIL- -I W
at
*
UI
I
4I
N
WI
j1
0
1co5
1I I
I 0
I W
0 J
I
~S,
I
§ 5 5 - Ia US U S U
zt l ~ m i * 04 1 P4
I
(J5 I I II
*j I 01Q I5 I
0 U I Iof&I I Il Ii
I I I I> I 4 4
>4 4
I
I I I LI mi
I I
I C Cc 0 I I I I
6.4K
Ia 14 w
5 ~~ I"
~ ~ 0 51
a
I ii.U~5 0 WI I I I
5 16 -45 @1 Icm
4 5 5
~~ I I I I 00 5 5 I I
I I I
z ' I I -) I II
z
W~~~ I W-C
Sr ICI I
II a. 0 IL
La u5 IIjI
W I 5 5 I 5 . s. I In ,
W
54 161 W f5
5- S e W sCAOs NI @5 05
P4 1
a~~~~~. a
a&6a a a a a a C :a , :
"a# a af O af 0 a
aR a a a a a Ca
6. Ca
U. 6
ui I L I. aA aO a N a a1 1 a: i a i
a 4 1a Wasa Na a
a a II aa u
a an .. a vCa NI Caj a
=4(xa Ca M
a~~~~~p I aP" a a~ a~
I~
a g aN 1.0 aCA a' a5 a1 a 0
Ia ~ t
W I I r II ~ 4La Ca 0
.--- 0aaI a Ia aý a0 06a aI I IA 1

~
a (fl 4 a N a a a a
a w a
a4
W
aJL W
aa
a
~a ~ 4 ~a a
~
a a
~ a
~ ~
aa = c
a
a
a
a()
a Za 2.IL
)
a.a a a a # a a 4 . 4
Mca a4 a4 aog a L-. a a a a .a (
a~~~O
a0 aN aN -1 a
a
a
*-J
a a
aI
z a
" a w a w IP
w a
at
a
I
~ a
a a a
. I
d
az z----a-a.-----a- 21 12 L 77
a CL a R a 60 a a
a ~~ ~aaj X
A a
aI
W
a1 a: X aIJ a
a
a44
a a (a
-a
h
~
a a aa 3c a P.-
a d a a 0 a w
a a aa a a a a5a a1a2
*ZA WC a* I 6-
I LL.p~
ag .6 of. %E 16 N 1 4i
.. '.s
9 6i8 es
~~6Lle I
I6 La 6j j 6 6 \
miILI
ILc 0
-4n
I A 10 6*"
x
0- 2, 51. 14 6d
be .s - 5 NS p. J
IZ -- 14L- I Of-.4 0I
6 65-153 ~
LonL
a,'
I L I" O-
005 Um
LLt "s os a- 0go
o i tO. S nt 0 I ta
o c a Z v
0 Ccj fa
zA o Iy . IL ZI
kv* II
* ~- ~ ~ ' cc0 0 I S
V)S01 c o%
0 IL~
ito tata Ia Oso
J
4: L'A Ul Z
i s U a..~ at ss WI 516
*3 & M I
atSI
* ~ow
I- I o5
.'*v..
LU
I C ai
a IL
I 4cS
*~ L S I
~M4c
WC a
* I - W I O
S-
*- I M= CK1
I-'
~~S.
*
I
S
SS
S I IW
I
* ~ c uI cc
..---
ccS I
* -LI
* S 5 5-0
S i U.j -4 4c
Downloaded from http://www.everyspec.com k.aa4
az a O f t- a a
a-a0000.0 a a a
* ~~ ~ a
ao
~ =
a a
a aa
aa a
uoI " (-o f en C.): I a

-,~- I tý0. I I4 aaI
aI j a1 a a qaa a
2i a4 0 a aa a
I a ac a a A# a
0 a Z - - a I al Y aiz c
4 a 00 a aa
40 a a . .. J
-
*~4 a, .I a1 a 0 V, aa
aa L16 a
a
a' a C, C', r-
* I-a I a a aa a a L .¶i
a-J ai aj _j aI aI I a
W a-a
z
~ ~W t~a
a
a
~ ana
aa C a~
Caa
a a
a
a
a
4 a
a
..
Z: z~ w~ a
a
L w~i-~
...
I.. awa a
Ca ~a
x~
a
(9
..
x..'
a
I=a
(A
a
a aa
aL
a z
a t7 aW a Oa a. a aOa Z
I a. a a an5a a 1 a6
.44
m- _jJ-- M 1 :
.
II ag MC
c P.- •I-
"I I I
"
I.,e, I
I
ac
.4I
IV
*
,
I
Ic I - 1
S.
u.II o ., ,° ,
I~l OW * ~ I
4., .-4.1 I . ,
GAJ~ :
0
.j
:
4- I
z~ 4
I Z" I I'-"
3 I z-.----,-..- -
* I 167
-Z 0 ." . ... . .
aa
CCL m a,!i
z a a
z aZ- a a a C
SL
I a . a c Ia
a or C a LQacO
wJ a -
LL~ to-, ) (a maO
CC a I~ ~
aa * a a a a
~ a a a a ~fna
Z a I 0I a a a a Za
* aI a a a a -00
Z a a a a a Za
o ~ ~ ~ ~ u a~ U.----a--a.--.aa
aI g 1J
a 0 a a4 a ao a a
az ab a1 W, aP &Fa a,
IL~~~d aY a LA *
(DA a a a a a4 : a 6aJ OS
4 I a0
00-J
a a a 4a 4 1 W a9 a u0a ( a
aL I. xa I~ I cn IWa
w -a aw -~ ~ e~aa .. a0
LL 4 a a jaa j
z f~Z a * a aa a >
*--a
I ~
w...Iaa-~
a
KI-4a- -d
a OS
-aa ~QL a aia
a a
ai
..aa&
a ~I a
a ~
~a F I~ Ln
0 a a
a aj
a a
a aq 0- aý am a a&
a " a 0 a a Z
I aI- a am aU) 4
a *a la ab ah ag
0 a LI I j I I aA
a z
aI" a a aa
a a aa
* 4 a a >
a h a CaZ a0
a a ma a mZ 4 a C m
I WZ
Wa 0 ~aa~.a-Ua Ca 168 C
Ik
ce OL of I OL
I ag * N 4
I -. m~ 5 0* - I
L0 ;'- 9
t2 11 00 Ij I 2j 2
I 9 4L
u* " . -i I-
W i.
=3 OK99 (Ai I
_ .tj La 04 .JS .W
U I- I
'>e -C A
*j Uz L.
WI
~ MI
I~I
S cc
WL S W
LUS.
4eS-J 4c Iý I a c =) I
D *j .. * *b 0
-
so. 1 21 . I-C I I.-~C Imp
X 5- 2L S,
4---**)*
I I W
00
00 .
A-.
I.- *0 I
)
LA.I
I L
*S U.
0- fu4 ~ 0
2n Z:w=:ý
S- L i 4e ~ ' *
Ni 4n I m.Jin S g o'
IL II* '
I f 0Z L 00
U. I z
- S I
0 LL a a
*i I 1 4 i I Cj
4 45 C) I
I 4 5 S 5 4169
L"L
* a~ CI I- ow w I
w CL L* w IL w I. CD *
*j-j I- I LO i I- I
ix 0 1 C CK fI a II I . -
* £ I N SI LL ...
0d I cc I In d
0 I Ide %
00.0 i .0 0.0.'400:
do j ClI 1 -40Iuj n1 *i. I I C4 L

I.-I CL x 1I.4 I I. 1- d X
2I1 I0 0
Iz z If I g
L4J I LJ4 jI I I wý40 1 C w I
x I. x IW coIa U
t- - 3E P I
*~ZO - - -- III.--.--I.- z.I
ILL I z- 3.
*j Iy Ib II I I
I. I z I j or I I La W I
1 00 =I
co , (A 0 00=)
CD
~ .
I CLJ 0
xWI,
0
=0 V
I
I
0- .
I I
I10
IA
*I I I I
IK > 6 WI "
5-1.70
ago~
cc I I I 1
IZ j.
I l
be IL M
*d m -ae
IJ
I 6 6 II I
I I I
Ni m :U: Awl
. , * ~ , . , , ,
In
I~~Z I I~
I In
I'
~ Ij 6W
~ ~ ~
~ .J6~j*0hjI-IN ~
W ~~ ~ ~ ~ 6j -I41* ZO cc.
40 -1 I 1 10a I 0
64 fl~W .
5 I I
1: r
00
4
I .66 I
.4
W~i6 I
1
41 ~
*
I.
-II
I J I W.J6- C
Irn 4** IC
I~~~L
I 4 (A U. I
II 6I L mL* I ft
o~~4 I
w-
I~a 'J I L6
'A I I.- icl w 6-r-

I A 4------------ -
* I I U * * - ' r 9L
W 6i .
IL
Sc
I ~~ ~T : 6
:
~ IL6 ~ A
~
6
6-
.
.
(A
.
1
0
0
I0 6
* ~IIA ~ r142a I N
'
S
6 06
I
N
I IL6 QI
Il
KZx
14 I 4
Q I us
14 -4
46 I. S W .I 0~I (I
61*O9.l .~I N
I ~6 .6 406 54-171l
Cl~I6
IA I IM Iw I Ir I %
I aaaI
b* a a a
a --- -*
J
L ----
2~W
I.-
*
-IJ IA
**
I a
I-. " I.- I I aue 1:1 I canI,
=498co :)at(
~4Uzi -. I
156 I
LU S "2 I
4f
.1k.~
w
I I L LU I I i L& L
* I - I I-a
CC IN
I
I9 NI CLUO
G .I-It
4
-.
a
1~ ~ 1
00 a
0a::
i~0
j - ~ * * ~a N
9-p
2 4cIt ¶~%
I I a I A I aa I ir~
Z .Z I (D I LU I III I L S
* ~ u "I- US Ih a
IC N QfUiS
* ) I 554I zi. I I I waI I6

I
*j I I 60S IO Iz uO I IV
a- w5a172
a za
a : U)
a-l (A-
a-a
al ~0 L 0
W-LI S*V-.a
-
a%
a a
a-aa
a ~~I
a na a L . ItQ
U
a.~
~ ~ aa a
~ a ~ aa~
a
. 6-
a
a
a
ug
l
a
* ~~~~~I.
S ai
a
ago a aa s C
a-j a a0 C. a -).a ~
2 Sa a
a i~j al aa Caa a00
Sal I @.IJ
S * a(I Naa J .. a
* z : : :~ :
aC C. a O-D U) ~
* a a a a ~Z a a-s 5.a
ma I~ W~ a
a u : 79: :ZR, "W '
a a-a
a a aa a1 a3
me I UNI
LI 1 -0 UM
1- a-4
o"" L I.
-C1 -4
Ii
I )
6 IZ
00
La I LI I
uw u
I~.I I
0 -. (M
CL x0
LU 0-u
:5:174
. .~ . .I L. . . . . . . . . . . .
SI I ILW
.6* ,
Lon
4:0:0i
2 m z z
0- 5 55I
r44
uiL ui I'l 1. 1 C l I' S UI.

W
441
uj CL4 I. S -'
a~~ I S ~Lji I
cc (A -4: 1 : "I I)A .

*U S I. UUJ1.
G' 0 0 )c
.I I S
-5- I . - -. 1 -.- -.- CI
C~r U "M Sc W I
Z" as a 1 0. al
IA S
S~~0
r4
IW
* -
In
S-n
-U
.
1
SM
s N' as
.
I %MSJ
-25..'.
01 I'
s.J
0-4
I
01
1
I
0U
015I SS 40II S~
W S ~I0)
CC c oo I I S S
I S: S
15 1
I~ 24 I * Sc Sr I
Lo) I.- S1 I I
Lai 4 1 1 4 I I S
*~u S -l 51 6-l
5-
55z4
55
5S 55
6S 5I
I.
0)
S
5 'A' S
NI I
5
S5
~S 'cý
CL
xi 0. QI 1S 0. U IS
x- x x DSV I S
z. :C x O I: : : >z 0' 20'
5-175
'SL
z 0
ZI I
Z L, II I
Z I I6 I I I I
1,1
I
14 : onI O
I.-ZCL I do
Ii I
I - -j I
I de I
* U.S U.S
I II4
I. NI 1Z=
*~( 0 fI co
I~4 n I
I CZ 0IC-O 0
-0.: I 401
I I I I
I-1
89 I>II
I-
u
I I I I
~5-176
I I I 1..41.
~
9 ~~~ 6 ga 6
o6,
mi 6
4L N6% fn I 6 W at9eIn m
---I--I - ,-.'
,-,--.-.e-.e.e
c a j -5 - t .- 1. M* so,
9K W In6 I
9.4 .4 I& 6U 6 I
961
* 6 I I I L46I S C). J I
LL J
Id I
z g I e I IL Ce 0'
= 6) 199 1 o f II IOwn w4 I4 4
e
-~~4 * * Ie I "I i e
.J I 6~~C 6 61.•fe w 9~ a6 me6 m 9 I
40 "a: A oI %I° " °-6 °o
ogi "°°°'log:
4~ 00 ~ cc II 4 II e ~ - t :6
400 -of .) 9 46 6101w e e49 0

6 6 6 6I6A
~ e a 6 6 6 6 a
-L
2 ~ ~ 4 4C
U.1~!
A 6
14.
d 9e 6 6
-_,CW 9-~~~ 9- 6e e eg 40 , 0-17

m X ----.- 6;.1 e-e*e-6- "bill
,e'60
,t_• , .. .. 4 9. 6 . ,0 • . 6 --I - _e . C•-

.. I .. Je I_.~... . .. Je.f . '. I, . - ., .- - -. .,. .
Z * I2 ez 16 W
UZI6 6B I 6
* C e~ 6 e *e *e * a *e ~ * 656 61e7
eg (a a a a a4
* S S I 6 NI
U8 1 a.D~'01 a g (
* a a a aM*
Z Ia C
Z w a a a
La
:: u:a aL a
a CA L a:O
uj I- **
:.. v0
0 1 IK ac
Ia. La I Ma
~ . o:
0 LK
I 6a I $- .3
a~0 I4 a 4 1 I
a h
*~ ~aD a z
ILL-a,
a af
~4
a l Sn
a
a
~- a
a
a aa
5
a ~
a a -4a
* a ~ --- a-a--* ZaL
*L aj _j 0
a 4 a ( si
dx I x I h
a W
a a ~ a aA
a --
. a
a-a- L7:
*~ f I
a~~0 - a aa
* a aa -. ~ a.a6
* a ama I. 2 a a
~m a (,a s-45aqa 78
14 'LLI
09 1 Gou I1.
g
I
a
I. I. I
I 1.- I~j N 0 41 1 I W
: i L j.~ I *
01 C41 4 o I I o1
00-j - I N - 1Im 01 i9Li L I '
o * 010**
I I I I I -J N I 11 m
;. I- 0 -I * e -K a Q
I" ha. Ig w u I I
IA I I I IL a~ t
II W
W II I I II tA
* I- I I K I 5 I179
* ~--*~*I--*I*-.-.-IFE Z a a . g
* C4I
of II
* I *
I II
i 0. I. IU
La 1 00
j1-14
L& W
*1
1
Z
Woc 4w w -0
I I- I I Z I
4---*--I X00
~m * * 0
0 * * U *
I- I IA I I I
U. JI I& I I
Z I 1I 'A
W 4I "
(~ * 5-180
9 a
a
I S : 1I , : -40 GO I
IgII
a
"Swj "o ,*a I- 1 ua 0aP" *a

oO.0~~~~ a 0 .0- a
p..a
a
a
~aI
a4
~~~
ac
a
ac
a * a
m Q w aal
. L a aa
'S~~~~ a aL a~ aima
Iaf a
# a4
a C O a aa
a f a. US 0, a a 'n
I
ae ai
ai I ac aon a ao
I a a aa q
~~~~~
m4 a~ ~ - i' in
aa a~ w
ao a aco w da w- a
*~~~4 an 0.W*aJ_aj
a- a a a ~ a
a 0.
0~E a C
a aL #-
UI~aJK
a-a *a a lal 0 a0
a
a W-l-a~
ooa
a a ma ...
a ~w..a-aa
4- aW
* a 4W.& a I a a ~UU~ a
* Ja a -ia a w;. a a
a- a1 aa a
*I a,
a a a- a -
aa
a i----
a b
aa
_---
m a a ~ aL
a~~~~, aaa a-4: a
j
*~~L a
*
..
a a aa a ~ a a
e
a
*~j a a . . .- -aa
a aa v .. J a
a~~~
a ~ a a
0
a~ a af
0 -
a-~~~ a a a a. .a .j .a . . .a.j a a
C*4
6--i s------ -.-------. -fn
10 IL :5 64 49 1 0& In
I -- = IA I - M = M
44 6go 666
NL I
W: W : 4:z :
o* ~ d~
---- *z-----
LI~ ~~ U 1
4C ..- (9a*
61.~IZ 1. 1. 6
1 K
LL.
96 6L - 1 6It-. 6 =6 W It P6
-. I- 4L m
I- i 66 - ~6L 6
*--I - co- x -
o- toU S
W 6L
s )
SM 66 6 SM 6 6 9
P i , 6-~ 6
~60
6wagwO qg e..m w I
SM ~ ~~ ~-.. ~ i.6 ~ i 6
6 0
-. I.6
de 6 0 u 66
Id10 .- u 6 ~ 66 .
6 6 6 6 6 M 66 66>
z 0 4 I6
UZ 6 16. 6 W 4c
.. 6 6 6 -.-.. 6--6 C~g 5-182- 6
6 6 66I 6 6%
1-- 0 aL
on-a--
w a aa
I ~ aL a a a
LL a L4
a a * a * I or
au1
1
SL Z 0 C4 N 1- a a6
aC $. a *0*aa0
IL 00 INO m
a~A ae aU I
K aa a a a a
C ~
a--a--5 -.-- a1a3
o r %C o a me e.
I 0 0 n CA I I
coA
09L . . a I I dug
Se 00.0 a~~ a t~
SM I, - I aNI:.
I.I
pAI I 0 a, a a (i
zW
z wuj 3
w._j~
I. 1 00 a1. 0 1 a
~.44I
a a * ab
a
a I I
a
.0
a
9K 4 a a a
-ALIu SM I a a II a I W
a aI-aa a a a z
I I.i. a I Z ~ I I I
a ~
a a a aa a I za
J4 A
cc
w
* ~
a a a ~ I a 0a
a a Z . 0 4. .1
a 0 ac 0Za ~ .1t in '
~ a a a a -01 I 1 I
V) 0
wd - I I I I0 .
*1~z 41C f
W _I I _ 45 I I
U*1 a. we aI a
I 99 I I
z I I 2
0 WI 144a WIC a I i
UI Iz -Z
04 .1 j Z.1
I ~1-a
-. ba b-I 0 I 5.18.4l I
I I. CD I
a
w 1 -4 :) I %
m :a
Zr
x l-
uJ a aan' m~W Z
I.- a-
a Us U -
0 4 or
I a r
CL I- a
Ia a in a C
0a
m -o: aa a41
I~~~ -Ii& I hi ~ ~ .
LL5~ hiaW
(A a
I . ( LL. I
4 w
.K a a :
Ua.-sa
I: a
a & a
aw o a1 a (m aJ aa
'Woat aC o m
I Lu z a ac a -Jc
* 4c .-W a a
a ) 0 5 a 1 aS
a1 a a so 0 a a
a a a00 a
a a 4a- a a aa 4 a a00
s. a ma m a a I.
a a r. "-
a a :
ts as a a ak
c a a, -CI
Zc
k
a
waato as
* ~aca~
.j
a a. Ia
~aa-m 000
in.U N' Ew~s Is a m.l w
-
Z__
w
*Li w
* ~ ~~ 0
I
~a 8
*U
a ac ~~ 5a *.W* 0
a aa
- -
5
a
0M
; ..
0
a- a------- CO a) a-
...
a a aL a
* a a I Li 5
*w hi- a .i
ZI a aa w S a...
Z
Laa w A1 5
II amn
a
IA- a aseae a**-
~ ; N; a 85
a 60 6 6 '66 i W
N a , 1-201: 02, , 00
a I 6 i 6I
6 k
-:~4
09L 6
: IiI - I6i
I
6
I
I
I
6 .4 I
ii AK I
6
~ ~ mI II de I mi I
me
6~
* 6 Ifa I6 6 6 (h (d• hi ha. -

2 i 6 , ~~~Z 6 6
i i 6 I i 6
i ":..-•
Q, , ,,o
1,
,61 , a0, a0, WO W
.
6 = 6 . I 6 I I > > >

I _j.- " 6 II CI 1 1 r,- 0• 0 .1 0a lJ
6j I 6II 6| I -I
.. . . .. . (" Lo," ,..," ,. . ...- ' -

aI ii i I u
1...
2
16i66- i i 6i
.. s i ......
t, -
I Z i_ Zw- 16 ....
l
6 Ia. a 'M I 6
I I 6• 6 I ,.-- 6 6 4 I6 I 4 . a. . k. , '
S i iS I
i(.iI C.I i.6 C)
a I 2
S1 16. 5 I 5I
5
~ 6 6
6 5 6.6
6 . I
16 I
6
1
I6
i6
64
6J,
SI I 6 I d I5 - 6I I
O a a a r4 i 1 - a. I I ha :.a -I. . .*.

Sc - - I I o f I P-I I
a 6-a c a I
a 4 I I f - I... I so. I a K aI lo a i N a
a
I. a a a 6 u a N
0 N6
(n W 0% C" I Ch ICA
la i IN
, ,
S, S ,S _jISI , ,_j I ,S
S _j S NI . -" , ...- '
P- Pi P- $- I- @- p- m itn
Ge u I w Iw L" w w
iIL o- IL I o. o. I I
S--o I s• .- .Co "5 .

1~~~~~ 66 L. w .- L -I o-."
5-186 aa
S.. ,P
. . .6..
•'•,- -•,,-•,•. .- , .a.. .
,,•,.. . .. .a. . . .
. .. ..• .. .. .. ... • .
•-•.•- . . 'j . •. ..- S .- •.r_
.,•r• . 5-. . .. . . .. . . .. ..
. . . . -. _
a a I a a4 a1a *
P4 a aa a a C
-
~ a a a a a
wa -m Ia :a-a a: m
C
a Z--0
aD
a-1 umI - s$ :
a" a
- 0.1
C' ab ( a
I-
o l
l a
awn a a a o :
a a
a-
a-C a
aI
a ~~o aar
D-
aoa.
b, I
00Ka
Oa
a * a aa a >
a
--
U.1a
a~~~~~~ ac
t~ ~
~aa Zaa a
h' w
a I I aj a a-CKQ a (
* ~
~ ~ ~ ~ ~ ~a '~ 1 I 1.a
~x.~aaI
Ia
aSlZ
a a
I~~
* ~~a a a 0
*a a a a C) a a a-GaI -l
a tj~ a a a a
a5 187
a. a a a a .a. a
NL Downloaded from http://www.everyspec.com
I I I IA I a I J I
Owl~~C I fI na II
* 14
14 14 ~*
I IL I 1
I I LAD i I l C
j Ui
777 I 7aI7L
l
OCZ C k.- I q1 I~~ CIn .1j .. J jg
= t I I I I I I
-14c
ZC. ZIC ZC
I .4 151 or-
W0 0.
II =
de0I *
I .0 1 C4 DI i I .1m
x I . 1 01
MI r 4 . -
I~a $01 aI II n
G-wZ cn C4 11 CI C
b11
I qi
OI ci -d~m 0' _d C
-~~ of- o 1. 00, 1 I . *.~
I OL I MI 10I: : Ni.
D_4 C4 V, 1
I 0 I IC X * 1
I W I I L I
* I h I .. I ... I .8
*~~ I ~ ~~~
.
z I
I 11
(,E
11
**I
14
Z
i
~
* I I I SI 1.I -I Q~fl I NI M188

ccI -- M ca 6.a
Lb I x a
Pz # ac f a
W a ao I Zw bIa
L 1- 0 aU a CL aj i a ~ a.
=M 1a~ a at
-i aI *4 W -4 a1 4 U. L4 IPSa
a~a. a aa a a~ a 2 a 6W
aaa
*4 4 a
2 lK ~ a-
f
~ I
t-
2 a W W a~
~ o'
O.IaJ
I
I- a a
ar
a a U.W I a sa
0
a
I* a iAJ
Z I M (I.(A InCq aa ~ ~O ~ 1 CL-
- dea(hLJ
-Ia on '..~-
0'
*a aj az
-~~ a ~~a. 'ol a

L) aL
a
a
a a a a S
a 2a a a L a a a 2
aa aw Sa ij a a aS
W aU .) a aa aa
a a a aa a a
aI...---
S~~~ ora
a aa
*X aq a a-Za a
a~C aI a I(
aa
a
O~
~a4* a
a
a
a-aa
a a~
~&aL a
a
ma
a 4.a
CL a
G
*
a ~saa
aJZ G-Uaa a~a' a a G .- #a
a ~ 4a a *3~a a00a w
* i 000 a wa ..la Xaa 00 a -J a a.
Z0 a a a1 a) a I a a ac a5 a
a ~U a a aw aa a
x a a
(A z
--- -- -- a a . -...-- a-. 0 a
a a a~j a a a w a 3
W-
*
...J a
I-CA aa ~ a£
a a *C*
a -
a 0~
.. ~
Q
Z4~ a aa a 4~ I ~a54 a189
arc 091 o
,ti t:w ~*a----.--a--.a --- ~

a-- ~ ~ ~ay 7 M---------a-
I---.
af a& a 1 a
a. aL a a aa " =aa
or a aCieaa a
a a a aa a aZ
Ca wZi aa a A a Lu aI.-
a K a l4za *
a V) ' a aa aIa I0. 0. A N
w a -i a- ~ I - ~
~~~~~~~~ ai aP 4
Ia i L LL ai a. W
p ~~ a a a ~ a~ a a<
~ a4a
aor aa a
() J a a aa a C ~ a ,
a a a La a6
a I
aK LL Z a Z W
a paa OZ.aI a
a
00
.. a*. -j
W~ja-- a
*O
O- wD f
a
CL
CW
a'
a 0
a) Ia
I w
wPbJ
c
z
C a
-
a
I
w='
W
l aL
a
Ln a.
s a
. a a 1---4 a. a a.
z ac a
a 'n a I.-
p~ a aa
aw w- -.
a~ a ~a CaImu ~ w a W a .- 4l
cq ~
00
W a) pe I
a00 'A a .ja a
* ~ a ~ ~ a
a
C
a
4W
a- o a
a O
a)
ej ILJ
a
T W a
I
.. a aa.a a - a a aoz a
a 0 a a 0 --
>- aa aa a
* ~ - I > f
4C a a z
> w aZ ZO aZ >K-a-
, aZ az a -- --
* a a a a aa a1a0
.. g i u
@6
a0 6 06*16 I 06 0 6 0
I .J I OIL I I6 6
at 6 I
4%4 I
6 U 0." Ii NiJ 064 '1 O > > >6~
A wl 0 6A ij I 3a6
ma 6
6 In
I'M I
4I I--6
cc
I 0 w
.4~u
La 6 6 I
--. u
6 I W J I W
il A II z
O I L&~ i 0 6
I t O441 Ip NIC)N
6 C.)6 C 6 Q .. 1
I I 6 I I i I I I
4 4 4 4 I
I I 0) I # I o f
It 0. I 6 6 0
I I I to I 6 6
I 6I 1 4 I4 6 I- 1 c I
I # 'I 6,'. - -
I-I-.J
1. I I I I I
z ZI z
4060 IZ 40 60 0
4~ 0 0 06M. N N
66 6 I
~
(,)~000050~~J40u
6
I
~
6~
~
6
~ ~~I
I
6 4
4I
6 .64wj6
~ z'. 0
46 0
10 6 N
0
~ 60
.1 6 68.6 6 C 4
-aI
w 6 6 6 I
>
> 0 4
6
I I 6 6 I 6 6
* 4 i I I
z 6I I I I z
I 6
I161
~ 6 6 6 6 6 6 6 6 656
6 6
* ~ 66 6 I 6 6 6 I 6
LL,* I S ac a a a
a w a a a
a a m1a W8 1.4 aa a4 a~ aw ac Ccar
4Z ar
00. a 0
ma6 ore aa
a aa
a a a
Z
I ai a a a aj
A a ac a4 a
ISaM)a
I a a a a a
w a
aI aW WI
a
a
MW s a : : :
a a a a
* ~~ zh mi a a aa
a aa a a aa aa a a
a- a a a a a a a a5 a1a2
I8.
0 00 I 0 S 4
0 Uh
0 e L
4c 60 04 00 "a 1- 0 0
= 0
041
I If4c M1 0
0 10 4I
01 O0 Iil Z
I
5-0z
* ui5-4
O4-
&.
i
0
;'0
.00
00
0
00
.4
WWW
0~1~'..
Wi-i
~P45- S
8
05 i
in 0 ~U~# ~0 0 ~4U A 0)I-
0
O .Jj.
a
i 0 0
0
0 ..ia.
w-
0
f
~x 00 w
0
0
.i
n
o 4 0 0 0 ~ 4:R:I
0 0 ~ 0 0 0 0 4 i5 -
193>
ago~
CO1 S4 aW a*a
z Z
a a0
Z &a a a
1- I w4 5- I la45-j 1. 6.
La ai U. w
f _aL&
L) S I S 4 U Ia I
SI
a
I
~x
0 E- :0) as - a
* L I ftISI- I 5J56 4 0 I
*..ILi I641 Q )4-
i
ac a.. I-.I.I X
I
00 0 a00
~%
u)aP" I
aaC a 0 aa
a ow a or 4J a.aa..a..
5-Z. a a . I 9- 49
4 a a 0 5-0
4 S 4a a .-
c.
L6 I .. J a aUj~a aa a a Aw~ a
Lj~Li a aI ua a
SM w ~9*
w- I ai a ao a Lai4 I wag.
a Caa-SO a ma 5-d) a-a
aCa-a ae.
Z~m
S a 4 a a a5 a194
Z4 a *' aa aa.AO
a a a a aa- SA 4
- ~ ~~ a
u a
~ ~-4 9 0 I La i aS S
a~~~=
a a an
£ U b
Uh
:
I A.~ Iu +36
wi a -am a S a
eea : a
- I *-. a dca a
w- I" 4aI
* a ~ a a a ai. a
-
a Ca4
w P4
a-~4 a -
I w aa W a a a Iac w w a C
0 ~ I-
LLV C
a
*A a a a
t&.J-~a aL -- a-4-Deg4a a
a
9 a
a ~ a ~ a I 1.-L
a:
0Ji U-6a 9
vI A a
co a a
a- aI a a a
ae a54 4a a w
a~~~_ a
a ~ a-
O
-aL a
a
aaIOC -a 0 a
L
.. a a i
a0 J 0 a I.-
a~I- IL Ia
00Sa
a0 a a-a a oaa a4-.-a..
a-aIa -
I u a Qia L
a i a a
0 aa a a0ia a h
0
w~~~~~a
aa I l .J a
P4 ~
z we a a
a
II)
I. ai a1 10a
IL- If 4ra
a. :a aa Ca Q~atfla .- W
41 a ~a
a-a 0
* Ia ~~~A a aOa
.. aa
i a a a a a a
a a a 0 a 5a a1a5
0 I a
a~~~~~l a a a a a a a
a to " a a a Vo
aa ma a 0
Ca
m a N
aI ~~~
I
a a j ~ aI ~ ~
~aA a * a a .
a0 a1 af af a4 a a C ag .a C
o ar a1 1 1
ane D ao a a a a a a l
in ~ aC a w a h h h h
w 9L a - j:~ a C
-@@
(h ~ ~ ~ ~ ~ aa a ~ ~ a '
(a (h h h
Id
a~~Z CZ
a a a aa m w
ac oIa c a al a a0 a4 a - 01 a 0 a a
U~~C a4 a .. aa Ca- a Q V4 a
a~~~4 a a a aa a a
a a aJ a an a a c aI C-
"a"IiJ *aIa , a 1 .a '4a a Ca >aL >
-a 4 af Ia a àa 1 a a La1 a ~ .a.

a .. Ja.a a a1. a a a a a w P
Ca Lua aV a-
CL a a ( ( a (
a C a a a a a a ta V
~ ~ ~
a a ~ ~ a ~ a I a 2 a
a a a a Z Zaa aZ aZ aK a a ia a
*
a aa- a a a aa La a
x O x
""i
.. w
a--a-w*aw I---w-Z aa I
a ma a a a i~ a a w~
a ) a a ,a a al
* a * a a a a a (ha (ha a (h (h
a~~~~~~1 a a aa a COO
a Z a 4= a1 (19 a41 az a a: az---a. -a--------a
a a a a a a a a a a a
a~~~~1
a a a a 4
a a a~
aa ~ Na a a an. ~ : a5a19a6
-1-* -u--s--
s--s--s --- -- s- u-- -- sCie-
sc S aIt a aa s s
W :
cc
-- d = ca
(A a aa 5
~s
ps
PI
s
~m
acc a * 5 *
9
* s ~ ~ ~~~~ a 5 O s C
5~~~~U Q Q
W C500. 1- 1.0 Cs:

1-0 .u MNaa
W~~~~~~L
u a s a m S u a a
U~~ 5 5 (A5 I
a a ~a s ZS Zuao
I ..J ag a1 a a a4 a 0 a [4 1 w 01 * *1 o -
a a- doS. ma
aM4A I 50 4 4X 50 5I NA I 1 5 V 1
-i~a. Sal w
-4 4N5I W
N
a0 5 a 54
1 5 a1 40 15 2 1
a- a- a1 I- a a- a aW K s 5 41 1 5'
4--.--.s-.-u-u-a- sWsZ 2.2J aZ aZ
s ss sa a ~ s a ~ ~ a ~ u u
a~~~ a a.~ f f 4
h 4h 4
h
u - 1 I --- 5.--.5--5-
x a
~
- ~ mu~ L I W: I W s s m a N ~
al > I.- a 0 a a . ~ p.
Uz 5 cs 5Z UZ 5= MOaUO aZ 5- 5 a a w N
~ aa a a s a 5aa * £ u a197
a in
wa .- "~
* U a a
0 az a a0 a
clIa
IM a a
0 Z~-.- a m-a
0o a1aa2
aa a,
a L 4 aa
a ~a a a
* a a a
~ a a
I5 19a
* ~~ ~. W e ote
b3 I Is
LL W Ij n I 4N
- f I
3I 0 410 1
W
IS I I >
I
I 1. - 41 q IL Ii1 1I 09)
a
I C
Or I I0 N I dI I 0 I I I
(-I
oo it 0' L I- M Ic
so
4cf g . (41 o
No 0.I LI
LLD IL I
4~4 I 21 I
* IL I L
IU 6 f WI I IAl
I'
a I 21 3
I 9L R IL
I X I UL I 1 0 OZ I
I 3E
IiiI I I I I I (fl 0 I
IL 019 41 CI ~ Z I5 I1I9
aa a a a
* aa aa owan
Ni -4 M Cc I aw S 4m 0 4.
a 0S 09I aI WE a *
a 0za
~ aaCaa
000
*
a~
a ~ 0
000
~ a
maaZ
*
I.- . n aa a w a . .~ i
a SZ
a4 a 0a
ac aC 14 "- I I.
a Ki L a ai a . a Z Ua
2
ow-l4 a 1.-f-(A w 2 a C a c a.
aie: Y a a 04 ac 4 4h a P- a r
I a cw 11 so M, a w ca
aL a &o a I- a. . a4
CV, i * a- 4zI
am a qr a W Ia I WO4 I ~ I
Sj I
a4---* a a a 4 ---- a- a a,
a 4 - a a- a a aa
* A a w
a~ a a-C~ qc.
*~~I aag I
w al I---------"ag a---------
* aa a a a I I a
a
ao
*
~~ ~ ~~ 4 ~ ~ ~ Ia OU)a
-w ~ a aa
a
~ 4a ws
a
~acW9 Ua I
az ao a a,
Lo44 ao a a
a E 9 I 0.-a aa wE aCa 0I-00
I=~
a ~a ~
OO~
a
~~
a)
a
a
aI0
4~aa
I-0.Z
a
~aaa m
4 aa
aa
a
a
l0Z
0
a a L a a .azaa UX- o
& a
14 > > I II I
a Ua a aC W ae z
a 1a-
A a a
~j a.
m
Q a I ~ aI Wa4
a-
a 1
a
h
0
a a a5(200
a a ~ a aa a I aA
Z I SL S S
LI I-- Sj
S S S WI
W S
b-
~4 40~ 3
I
a
4 ~ _ In 49J W S4
5- I-- I S S
* 2 5 w ~ia S S4
~~,.a4S-
' S 5
z~ us
s Ii- 4 I S
S
-
~.S
4
S 1A
5
IX
I- S
31 SL S 2 5
* S5S201 Ba
5.3 Weibull Analyses -- Project 1. This section contains the

results of fitting the Weibull distribution to nonelectronic part
lifetime data collected on a ground mobile mortar locating radar. For
each part type and class represented, a table is presented which gives
the part identification number, a point estimate of mean life in
hours, a point and 60% confidence interval estimate of the Weibull
scale parameter in hours, and a point and 60% confidence interval
estimate of the Weibull shape parameter. The form of the Weibull
survival function considered here is given by:
Pr(survive t) - exp[-(t/b)C],
where b>O is the scale parameter, c>0 is the shape parameter, and
t>0 is measured in hours. This convention is followed throughout
sections 5.3 - 5.5.
In addition to these quantities, the total part operating hours

are given, along with the total number of failures. The "comments"
field contains the predominant failure mode observed (i.e. the failure
mode which occurred most often in the sample) and the observed
significance level for testing exponentiality. We recommend that
exponentiality be rejected if the observed significance level is below
0.05, although other thresholds may be used according to the
particular application.
In some cases, less than two failures are reported. However,

because there are multiple systems reporting (53 in this case) and one
or more parts per system, the systems which have no failures were also
used as "data" so that the two parameters of interest could actually
be estimated. This applies to section 5.4 also.
In only one instance in project 1 is exponentiality rejected at

the 0.05 significance level, namely for electrostatic high speed
printers (identification number 1-2-36). The Weibull shape parameter
estimate in this case was 0.748. The corresponding Weibull
distribution would, in this case, have a decreasing failure rate and
its probability density function would be shaped somewhat like that of
the exponential distribution. One interpretation for the shape
parameter being less than one is that infant failures were still
taking place in the printers after they were installed. If this is
true, it would be indicative of poor vendor quality control.
5-202
5.3.1 Weibull Analyses Summaries. Following is an index of the

nonelectronic parts analyzed in this section. Each part is identified by a
number of the form "x-y-z". The prefix "x" identifies the project from which
the data was collected ("x" is 1 in this section). The number "y", being
either 1,2, or 3 indicates the sampling and censoring scheme (in the
statistical sense) used in collection the data for that part. These schemes
are described in the Final Technical Report of this study. The number 'z" is
the sequence number as listed in the index that follows.
The column under "Best Fit" indicates which distribution is the better
fitting distribution (assuming 0.05 significance level) with E-exponential,
and W-Weibull. The format for this index is used in sections 5.4 and 5.5.
also.
-U
Index to Project 1 Weibull Analyses

Sequence Best
Part Hame Part Tye Environaent Number
ACCELEROMETER FORCED BALANCE GM 1 E
ACTUATOR LINEAR GM 2 E
AXLE GENERAL GH 3 £
AZIMUTH ENCODER OPTICAL GM 4 E
BATTERY UECW.AGEABLE Gm 5 E
BEARING BALL GM 6 E
. BEARING ROLLER GM 7 E
- BEARING SLEEVE 8 E
BE~LLOWS GENERAL GM 9 E
*BELT TIMING GM 10
" BLOWERS & FANS AXIAL GM 11 E
BRAKES ELECTROMECHANICAL GM 12 E '*. •'
0 ELECTRICAL MOTOR GM 13 E
B? USHES
CIRCUIT PROTECTION DEVICS SPARK GAP GM 14 E
CIRCUIT PROTECTION DEVICE SURGE ARRETER Gm 15 E
CLUTCH FRICTION GM 16 E
COMPUTER MASS MEMORY MAGNETIC TAPE GM 17 E
COUPLING FLEXIBLE GM 18 E
- COUPLING RIGID GM 19 E V
CRANK SHAFT GENERAL GM 20 E
DRIVE FOR COMPUTER TAPES/DISCS MAGNETIC TAPE DRIVE GM 21 E
DRIVES GEAR GM 22 E
. DRIVES VARIA.BLE PITCH GM 23 E
" DRUM WEAPON LOCATION UNIT GM 24 E
"DUCT GENERAL GM 25 E
ELECTRIC HEATERS RESISTANCE GM 26 E
FILTERS AIR (GAS) GM 27 E
FITTINGS PERMANENT GM 28 E
GASKETS & SEALS DYNAMIC GM 29 E
GEAR STATIC GM 30 E
GEAR BEVEL Gm 31 E
GEAR HELICAL GM 32 E
GEAR SPUR GM 33 E
GEAR BOX REDUCTION GM 34 E
HEAT EXCHANGERS COPLATES GM 35 E
I1GH SPEED PRINTERS ELECTROSTATIC GM 36 W
*HOSES GENERAL GM 37 E
INSTRUMENTS AMMETER GM 38 E
INSTRUMENTS VOLTMETER GH 39 E
JOINT, MICROWAVE ROTARY GENER kL GM 40 E
KEYBOARD ELECTROMECHANICAL GM 41 E
MOTOR, ELECTRIC SERVO, DC GM 42 E
MOTOR, ELECTRIC STEPPER GM 43 E
POWER SWITCH GEAR GM 44 E
PULLEY GROOVED GM 45 E
5-204
/,.. . - .... ...-. .,- ,~'-***,'*: .- . - .* .'* .- .* .'.* .. -..- .. *•.. - .* -. •.. , - .. • .. . . . ., . -
,".'". . . """-:'-""""" - *- .
."-' .- "'*'*-". **'- *."-'"--""
"" ""-*.'.,.,.**.i ' ' """" """ ." . •' ••• :
Index to Project 1 Wejtbull AlvTeS

Sequence Best
Part Name Part TEnvironment Nuaber Fit
RETAINING RING GENERAL GM 46 1

SHAFT GEUUAL GM 47 E
SOLENOIDS LINAR GN 48 £
SWITCH PZSSUU (AIR FLOW) GM 49 R
SWITCH THUMOSTATIC GM 50 E
VALVES PIhUMATIC GM 51 E
iL
5- 2
• %
i •d~w
5-20
1-41~ -
z La z I.-z u I-
ILZI~~~~~3
)IZ 00q'I LI
0 Of.- it I a
pq I I U I 0 I A I II 4
604 . 1 6
+1 4..4
Z 0 cis~-
Ii 4 UA 0 I 4
0 O
00.- I I e iI 0
0 O& :-0Q0: OI 2 9L
IS 01 0 Xej
0 G.Z 0
0 I WYI. I O.7Z I oFt
04a. 1 060 ac Com
04.4 0 a
AWu I IL XI
WUI. U)I-IW
4c I Ice
O D 0I-s I I o -C .MI I
I ?SC-O 00 CsoI 1 I 0W00.0

in
u I I A.
I~~- 4 I
WI 0j (J 4II I ~ ; #
I cc))- I0 40014 1 1f 40 I ' q4
Sf0
hic - o
too-
: : I w
6
006: I
40 II
00 OW
*~dt.t.~ ~d 0-
..
-JI t,-l0 W G-~ ~ . *
* ~~ 4 ~ .0
~ c
1 4~I
ui.~
~ I
eI
0 ~ hX 000
.
.~
4 0
-
w- 6 1 ej .-.--. w..
9 I I
* ~ % 0- 0 I oI Iw
> i 0 .1 W 0 I0 W I h 00
L) I- 0 0
We -W
~5-206
*1ý
0*56o I C
P4I I I
.
hi DI 1 40
-I 0 4JO z 1 0 to
40- c .4
* U -- I --0 0 ol.. 0 &A

. I
04~~i 0 0 la I
I OI0
ja WI
5*q ~ ~ 4 0 4
#I I .,I
0 0 0 . W~I I S
gpo a I I o- 1 4w 1 ~ 0
* ( f m 0 f
05 I dc aof
at O 0. 0
: 1
I 0
1
O I 0 PL-
hi I 4 I w I 0i W 1 N I I S - S0 %--
* I o 0 0 I 40 0 *
O~~~U I0 I 3
IOU)
z*-C z. z 54 .
thJ~a II I 0 L I 0
woe - I "Je I
* 0 6 0 6 44 I
%. 40
00 0 0 . 0* @6 0 ,a
> i 0
Z4
fC h 0 W I
0
Cl
e
h
: *
~: * 5 : -207
---- -- -
a U a qW
u*~a aa a ma * A L a aa *
mal
40C a
40= I ~ aMI *a a a 03 aa
Z*~
ILz 00 a
ZCaw
l- a
upa a-
a-
LA a
= faa
a
* 0:
W
I :z
a
-
wa4
M~
()a 0-*- a -a a a -o -
I~a aIL n a Z
4 : Ni: aL
a a a" a aI a aa
a O. ao N a- a' ajz _ma a, a a

-' ~ ~ a
a mai
01
a *
aa04m
m waa aa
a
w
a La (aa f' Z aa >
a a abea a:~
aI. aL a' a 1aaa
a o - (00 u a -- - -a 00
a u
a'W
aa a- S--Ga a a- a-a
a
a
I.
a 10
a
Z wa
)-j
a'
a- a
#a aJ ma
.
a Z w
IA.
~ a
a aI c4a ta a a Wa a a 0"~
a a
z~a *
aZ .ala m aag .-
cao a a a.
jj
8a
an a(z a
I~ a
cc~ a ca ui~ w"
(A a 5-2a8 '. '

* I 69L I
-. I. I- C,n
~~ ~:z :
Oh.O f
M I 6 -
-j II I
* I 06 I I n :SI J
S SI WI 6& a6 6 W
P- OILW
0 1 40. 1.~ *0=(D
I
I0 w s 1o
a a 1 ~ 1
w 40 in~
II I I b6 1,4
_z 1 4-4 Z" I -A I Z" I
0 C4
0aI I.- W fo I 061-
J I I I- 46 SI S 1J
a CO 1 I 00V10 .J S
1 N~
*~~~~~ 6 J I I in 1 Ih.c 6
Iin i xU
PL Ccw
*~C II I II I
c
* C 6 6 > 4
I
u: I
4W WZ I6
5-209
CI
CL) 0 0 a 00
a
Ia 40= 0 I qv0 00= .4 aC4 a
aa 00. Ir w II 0 00.
Z I I Ia Ia
0 0 0 0 : 9a o" I"a p a.
Ia P4 W L 02 a a
I.- II Il a - a
.4: ~4
0.4 .4 a
: az Wa a 0
140. a c a0 a .4 4 a 00 4
II C', a OW"6 a a-
0u O f f~J - I4
a wa
Oa
*-*-a-*-0
.-.--- *
a1 aA- aS.a Ia ol a.-0 v

a~~
aa ~ ~~~ w a ýa1 I a w
*~F 0'
o0 a
1 ~ 0 ~ 0 1.-l
00
wM 0 N .0a w a 0 a 4
-aaz .. 0 - 4 a bU
w4 0 0 w 0~ 4
.. JZ 00 0~ .4 0 (AZ 0 .40 a
o W ) 0 L0 .. w.. 0
ae a il ac I
> >
LWJ Za Uz Z-L
4 a ~a
a ~.m005 a200
I a, n
In. r s-
6i w-666
m.a 6 - %4 66 6a
46 - f46 .
w4 6 U
o6 t6 6 W 6 6 6 W~
6 N II 0
z * ý
$0zI* 6 6 6
*6 0.
LZ * '6 6 16
* , ~I w9, 66 4
0 : .6 ""-
" " I
2 WmZ 6- ~ 9S:~
U4 I ago II W
0. 61 00.0 6 j
6 6 0.
0 5
oom I I W
W 6
*~~34 1
0. I0J (A An- 0
0 .
0-J
In W, I 0 66 6 IL 0. J
o1 fto ac6 W
I 46i ob (j
W"o@ W-~~I JoelZ
I~~ti >C,1.vt 6 o6 >
*I Ic
0% - 94 Z1 66
6 Wi- 4 W 6 Vol6 w
6 ~ 6 0 ~ LA.fl
I 0 0.
ILO- 60, ~n- * n

*~
9 W ~ ~6 . 6 0 6
9W
I ~% \~0
6 6 IS 6 5 6 6
tin 6 0-~ 5 6 6 I.0

0 0.
*~W 9i 00w6l e f, W * .6 .g 9 9
I I a 4W 6 46 406 O
0-.~'I 00 2%r
6 00
.46 I06 @6 .- 0~6 '
* ~J
6 600 I 151 g 6*~6U NI 0261
.4~~~~P6 :
de
6 6L ~~ 6 0- .6 0z
6 ~~ 6 I 4c I6
L6 LL.6 ~ I
I. 6 I. I.-I W a'ea 1-a*

z 449=) *6 0 0H 6 -W6I 4nl2
W6 0
64
~0 *
6
:
6 ac
w6
0
.- W S 66 I I 6 W 6 6 X
4 6 a ' 6411 4 I 6 6 ILL
6Z IUJ I 'l6 16
>W 1 z 6 0
p t3
IL ~ * (06 3 a "r a6 >
z 6
w-II
w I W 1-646 W 6 46 6-66
I0.- I c I~
.- "Z6
z-- -:Z z
66 6L 64 64 4L In 0
~~~:4c : : 0 Z":~
6 610 "4 I Ow Q I - I8 P.
a" 20 a 24
C- N 6W 19 649 I 24 I
00 a I- a.6.. 00
W UW, -. 6 -
WAJ 60 6 -I= 6 -6 1- WJ
W ~ 0u) 6 (AI I WDI co 6 6
O~~~ 606 6 W
I Z 945-212I
- UL .,.-.- . .... ,.
ow 00 W1
aa Z a a6 1 6 Z amaa
6C6 0C 1.
a- I6 I 1
* SI I 6 1 I I 6 4
j.0 :
.46 s-
w.,
988 6 4
14
b-I CL= 6 6 $.
Z I
*~~~m
OIL ae
W &I
I
6I
Qd0s II
InCO
~a *A6 0
6-4 6 I I P-
I PO _i l W"-j 6 "
-" t I -
w U 1 0
h II I 45 I 4 1 C46
-i 05
5~ ~~~ L 6 6 56
%ca 6 6W6O 6 0 a
*~~~~WI
b Z- 6 0 Ii I
9L 6 0 1466 -: i4 6
6.1 -~ - -
Id w 6ZJ
La I 6 2
-K-: 5 616
&*a a at a a *
"4
OIU P
I.- aXI IOaa
Ca
LL. 4 .w a a
aL )t (I
wa cawa a1
~~ aO
O ~ ~ ~ a5- ~ ~ ~
a*.~aQ ~aI a ~~a f
a0 a aI a aa a
0ao
*---...-a.* a1
a L a a. af >j. a1 aL a a >
a L a Ca a U aL m a a L
a IL aj a Ia a CCa a 0
a~a -j ) a Lil
a"L a a .1 Laa.
a. a a Pa Wz Ia ý a
a _WI Z~
a~4 a 0a.aa 1 C3~
a ,1 b"-a a 10 a claa UaQ.4-4 aI
4 a (~ (
a aAL
a a ~fl0.tA a
aa
a a
u a- a:aa *
( L0(la
Ja
0. a ma 4a.a * a. a a 4agoa
1-- a *.J~ a~ 00 a ao a - l aI~ 1 00 a a

w aa a a - aa a a
1 a a a1. a, aa
eia'-***a.* *401.---a-----a
a I a- a j I OU a a a1 a1- a -aj 'J Aa

a~~
aC
~Oa-
x~w
4a
a n
cW 4 .
a -
a a 0- 4
1Z1n
(Aa(~i ~ a
aj a a
1.
V)J
P4 w araa
o
a a x~( 0 a fl W0 a a
a SZV a Ia a A A-
a m a a
I-- 04 "z-----a
a-*.-a.a 0( a ai
a >-
(0~~~
a aZ a a m
a o 10 a a a a a
aa ~a ~ (l a a 0 a5-21a4
*W a I aa a
* I I 8 WI I IS~
P-IL 080 IL 8 W .
L I -- I -8 S i
I II *
IL 001
w 0 1 1 'a
a I
0
4 -18CID 1
*I W'
>8 1 ~0Go 0
cj a5 I
18
I C= 08 14 CL m IGo Qa
a-04 :
C. I~E) I 4Q* () I L %4a
* .45I I I. S 8 2
II CALA 08( WC I
8 0 1 08 1 08 00
I
ILI e
S I ar s - a F.- S I 8
I
IIw I of IJt Sal t I. l I ft mIS o
08M t 841 cwO t- I W : .48 Sa O
I.- I 'I 1 %6
8
_~ P,1
I
0"I-
I - I l-j
W8 1
II
5.~~ aI ~
*U (A A
(AI V()8I ca I
8 4 A.I8 8 24L a
I I I ~ u I 8 I 8I 8
40 18 Or I 8 . I u
0 8O 8
5-215
ot O T C04W! OW9
Z 'e
I4 DL: 0v L= R
aI aCL a aI
Int aO aI I
a Z a - a a U I to aI -
Z 1 ac Im a- a4 a- -O P 6
0 X011- 0Z ~.4 a- *
0.S
0 a Wa a aa~
4 aa aa
a Ka
S la a a
Z W im a aa a a wa-j
w a a wa 0~ 0, .a ~ : LIaZA
a a m..jm
Ca a a-a a ~ a Caa a.
.4~- aa a .4 aa
aa a L aa
a Ca *-- ---- *a*
a UL a) ai a0 IL Sa a. a
a a~ a a
a CL ai
a
a
~ *aa
~aP. 1
Waa
~ ~ a0aa
OCa
~ 0
a
aaZJ
* a ajUJ
wzm a a w a a K
aZ aa waa $. a aa aWa
aZ a a Z a a- i a a 1 Z
AL "Za ag C, a" C
a wz ai cca w Z Ji=a
aPw-4 0
a. .az
a a
ao N~
o-
00
~a4 .4
aL Wa4
a 4
0
o
a~
a0 In~
F a,
.
a1a
Oa4
0 1a
j W
I.- I-- U. _
Z I-,W a a a1. ao a i a a a" a.
0 a L a 0 a
a~~~~~
-a a.-
LL a a
a .
a
a -
a
-
a
aL -a
ac2a aA
-.-
~~~OC
. a
aaa M
o
.4 *
a. a
4 CLa 0aO - aa m .- fa m
*~Z a- aa
* Z a a
*
..a a-aa w2 a
*
p-
~
~ 't m
ac w a 4~ aa a 4~ d
a a aj am
-a
-- - 0-----* aT-O-a*a 3 -----
a a
w a a 0 a I %w
0 aa .aa w a a a~ 4a aI
>za:
a- az *a - a a i..
>z.
2 a 04 a
u0 Ia
a aal
.4~~~~~ 0.:
Laa a Iaa a 4 L,a
* a iJ.4 ma u. a a a
5a21a6 .4 a .-
0 0 0 "1 : I
0La W0 , I 0-
41*
NO0 4 OI 0I
co0 8i SD I I
*n ~
O0 2 a Li0. ~ 0I 3
444. 00 (A
0.0 I- 4c"" 0 -0 1 U.
0I. )b0 0- I h 0 0I
Z I 0
4L 11 'A 1 WI
I I Of "Z
I
00 -1It
: ~ : zJ in
dc1:0
C.J
JOCB :
I0 004 . j 00 1
CLo
Loii
d " : -
W:
ZC 00
14
o0.C~- Sc. 0
40J I 0 o 0 K LA. )4
I.
I4 Ib Yj I,
0~c * o (A
4~ ~~01 0 0 4 . .....
0 001 0j
0 0 5 0 I217
IM
I,0
*LZ -- a a-- -Iw----.--
I---.e---e
an aa 6 a 00 06ý
i
10 1 0- 1116 0" 0-CL0 ILe
I. x 0- w I6 x4
We
ve at S
UI *i Oa-06 0 .05
h.6.
1 6 in 6 06 1 a
CNIj
I N I= w5 IL = I* h
OC 0 e -eja 0.. gi w
Z 06 6 LJ 66
UJ60 - ( )
co-6
@6Z0
qt 546
m W-
17 05
:c -- ": c
L46.4 I in I
Uot I p-
I ~ ~ ~ ~ U I6O-fe660S(e W
I .6 ge 1. 6n Z0 40 =.I
01 -eo1
C 4cW06 -46
It6i 00 a 4*
We a co 0-. 00 6 -6
=~ ~ ~~~ a .6 a a - JIIL
R
Z
W -
1~~ ~~~a a a (0 aa
6h in I h
w I l Ie w 6 66 Sl 6 0
We 6 w I 66 L~L I
I . I 6 - . - 6- 6 0
-~ZV 5 6 666
66 S~ 6 -- 6- .-- 6 h 6 6lie
0- a 6 6 0- 6 6 6
6 6 0 6 Z>
641J o fS~
06 KL) 5-218 ~
I-. 6 a 4
5.46 W~a- 6 6@4'Now
I W I aImmS ~e
0 MC
000 a *aa e
a~Uz
00
a-
ai
aa
e
a.
a
I
z : Ogg
a- a
~~~~~~c
nz a 0 a:
I
I
1.-m
w
i~~~aZ,.ilaa
10a a-U
uj- a>
a1
aJ m a
0:LS-:
a Z
w a
a
aa
: 2
aI. a 41Za ai.
Z aI m a Um 2m a 024
a j aI ~ a aa~
aoý
I ., Ij a a ) 0
00 a vaa Cl.S
a WeoU a aa h aa a aaa a
a a * a. L1"A 1a 01:~i ~ W
a~4 ar I
a ~~ ~ ~.jaa
U ~ ~ ~ ~aa~.a.a
a c z
4
dcI a mi'aaz
1 I Atm a a
at 0 Mm a Ca aL 0l4m a
1 i w a 1.
an a a-
a0 an a00a I ~ a
aI4aK o hb4 a q .~
aI .a~m 0 a w~ a .JZK a e 0 a I4L
a Iaa-so~ I. C~ I 11 I @16 a-U ap-
z
.b aa aa U a -
-~00 ac a~ Cmln a C'4.C f 0 a 1aa a
a i aNI-- -a - ac w --- m..-. .A
a A M aI: :
Il aM a a t a-
- -sar a= a1.-4a6e -C
a aa a 41
hil aIa .. m a m
aj aL aL
a~~~c aa a a
aWaa a aI
a-u-----------
D. a.---a.- ----
aa a aa a
* a a a a~ ~ a ba a
*~~~~~~~~
hi ao 'a h
* a a- a a ..J a ~ a m-mainI
a6 (0 aLA aZ I Ua f * e j*~0
I hiaN2
a--t a5m-219
~ a a aa a 2 4 a
a~J a a% a
ow 11 0 OwlI@9 Im9I 4
N N t0I6
P-I
z UI
I I I
I~~0
IIWO I I -
£L OASI £ I I w
IUI Co gO
WI11
>41
IL : I n s >4I
Lai6
1- 4 I oI I I 1I
w
0.Z I 1W LaiZ 6 4(C 6 IWI
0.0 :
1: .40 0 "
ca~ ~ oo -Ju 0* . -
:y 00 1 1: 01 1 in6
1Lon I-- I 4 I (on
W IW I 409 1- I."1.II
* I I G9 I 11 I9 .4
I £ I 1
I~ ~~~~~~~ 6.WII
WZ IZIIEII
~ Z
I
I9c
'II -W
>
I U JI$0I I 4
WI owlj~
V& I 1-4Z1
I I I 0. 6 II :z5
I 1I W I I 0-
x - I £4 - I £4
I WZ- 9Z I W0 C4 wI- I 1II
U 5-220
h S I SC a I I
I I I -j LA I @
.4~~~~6 I l N I. .I~ 1
: Z 0
~: * : V: : WMa
I US I S
xS
I IUP ~ I
II 5-. I I~ o f
1 all o-I f I W I 41--f I -II J
II I S S ol I W4
I I- I I S-S I I S II S-
Sol
I- aS- I S P4
E
*w I
I.Am
0- S-, am:u 64
e ac
I 00 1 IOJo
-jA
Z1 I c Ie I
II
I I I 1
~:
= I-
89a41.s *i
*we%
~,,
I ~ Z IASI U
CM SI II S
w Ci I .
w IU I
S-
41 a Wk I S I S- I-221
V.a%
II
0"I'LL
"0 40 oza, 1
a a
a Z a 40 > Z a 16 >waa
a ac11
* ~ p I i
a ei a0Ji 1 I 01 1 a1
* a ma Z ai~ aa a z a h,
ILE 6 6I ~ ZI L I-
ina ~
a" a a a0 a z
a ba aa a1 aO
JZ aZ -a 1-- a 24- a --. I -a
a ~ aL i 4(wo
0a
aJC : -a
a 0.
Wa
10..
~ Na
46:
0
-
-
AC:
3Ia
s a
6
.W~
0089
-10.
aI
a~~~. a () aa a I I ow 0-a
*O ao a0 a1 I. I
a0 1
aI
I a a* aa 6
0. a aa a w a a a
U. a0 a---*LOW--
* wIag waag
aca
5-222
g.* a.a . a .ID

a " ha W, 1 a, W" Qa a
04 1 4 a a6 S6 0
WZ gig g l
"a J 6 Ca -
a 6 a -a a a s6
NO 6
V).
m a
w
-- a a
a
-,
,
a
a
~~ja
.. - a
a
-aa
a
.a
waa
~ ~ WW 14 a
94 a -aa
WO, .a 9 'IL
wU
Y1 a a W
mg. a IV-
aa a 6 a a
a a! ~~~~~A
a:
00- aC
a - a 1 a4 14aa
a0 at 1a 000 1
AL ~ ~ :: .a
aa pa a iWa a a W
I a
: ~ .~:: ~ a a1.
a (.0
IL W 6 0.4
I -t w zja
6~~~$-
6I'6 WS-J W 6a w-u a *
I(.306.-C
*aa iZ~ 6 a
(66 at wa-aa1a..Z
66 30- a
a qw&O~~s
4.46-a a -a Z I
a a
a
a
~C
EfO.fl
A q
aa
0
a a ---- a.*-* a w a a --- a~--
41C a aaa
a a a a 65-223
LIV)
* IL I-.. a m
*L I I I a Z Ia ~ II C~
*IV
mX aaa
% b,
IL I ~ -
Lid
a
4
aa
~~ CL-
~ ~ ~4w I . LillIIlS
a 9
a
X
w
z "04
a ~~ 9I ~ ~ 3I ~~J ~ ~ I 4J I- :
aca
9 : I C2
-
La 1-- a a) R ). I as Ii I
~ Wo S (A ILn I WI
00. I 0 I-
In i I 0 0 1 4LI- C
I
III a 5 C4
4c I
so j. 1 00 1a00 ao
C K ~ aa3 a m.
aq 0"
0P I 1 09 L& IK 9 I .
2~ ~~ I I8S I ~
I I z~ Il W I99
a~i ~I ac aA aa aa I a *3 Ac~
0a.
n4 a
I '' Z
I~
(4n
- I
'e a~ ~ S
S
9193
9
~ Cgo
S I la-
ac >SIII~
*a a 4 6 aa a a 2-2a4
0.I--..-.-a-.-IWO.
a
Ve 0 000
0. too n: I o
I j I us I -
IA 4c I
6. 1 .4 2 2 0 1 II 0 @03
m0
1 v0uzZ 0 es $-- 02
w
o Zt
1 0 0 I
0 0o O
O
6-. 0 -,0 S 0 o -- S -
0 IO 0
0-
I:C1 1 co
N:D
M4 0L
# 1
0 - I0 IJ 0 0A I
I.- I ~
.sS 390WI I W
0i 5Z I .
z~
6-
43 0 0 3 U~b
(0 0 .J3 00 C~ 0 J
I I Ao.
*4c s1 wZ 0l w 53:-
ve IV 0 aj 3 0 I0 0
I-. 000 0- 40-401 0

In 410 0 . 0
.- I I z- ~
N. w-~ 1. ZOSo JZ * S 2-I
%
* 5 0
00 3 -.-
I~d I(~ 00 toe-'
Z41C
*d f
O 0
40- 5 * 0 180 40-I
3 j 30 .0
33 000 I0
CL 0 050 4
WZ
.. ~* 3: 0 ~ :4:
C~JC 0 0 0 4 C.S S D
O 00 3
o 0 0I 50 5 0
Z
O CIL
W0 0 00 3 W0- 3 C C~o
Z
z L0- 3i )0 S I W
Z0 3
* We
U) U) 3 S 3 ) 3 oo5-)235
I a aa a a In
a 1 a 4 a1 aw Ca a a
.N gl a 4aa a ' aa aa i
1a '?:0.: Waa *
'-4, a *
*r
-4. 0 ~ :
am
m a cc
-4 a
Ws.
~
a.
a -aa
aa
C
m
I a * a ~ a I aa a
a aR S-at
W4w
a
q
a ~o - a
a
soa ~
aa x
OD- - - a ao a o 40 - C1 -a
. a I- a
ama
~
0-
a : ar a me a (
-: a
>a atw a a 1 aA a0
a c~a00.0 a a .. 00
Da0a0 a 0a
aa
a~~ a 1U i~ a ELaL,
a a a-~ Waa a W
a a aW a a wig
a w a. WO W N a a
I
ai a -
aaa aOaW Z
ao 4a-4a.4
a a .s IU~
ac a -a 0a.-.
a.-a*
a" -. - -CL a aa a -
a so ao a a-jaag 00 a 0 4Z
a-aaI.- a a - a
Oa c I-.Z 4c
is..A0 aK Ll.a
00- a ar a .. m a Cc .0
ar aa
o~~ ~
a>a ~ aTV) a 0
aIWO
aO a0 a at- aa a
.5 a aa a Ic
aa a----**a--a -- a-.-a A -- -- -- **--*-*6-
ES- Ca W a aa, KS- a a W a a

aa a a a
*--*a**-aa aa a a
J a -- - -
a .. J a 0. a a
>z a> )> a
a a . a.. .a. ..O

ub I _ I I *U
I I NZ I 6 6S
W I I I 1 .4
z 4 IC Ia I 6 1
S I WI
I 7'1 Ln f 0.3 jv 4 CVi
I~
4101 ()I 1 I I 114 a W
N
5'Iz .6I 0"u

A04 '.0-
I 6a 14 I
-: Z WI
146 Is
Ac S " I:
~ 51*' 6 W
I 0..5 I Q6I f-S 0 I 00.J5I0I I I0

I 4
II
IDj IZ II1 00 1 -466.3
I 6 00( 0S 4
(-0,1
wWt cc , I&
I
I4
I I
I
3 1
6 ~ 0
V co6(A1 InJ IIca6
CL~ I CLZ
I--
II I £3
f=4 U :16I 6 44 5 .
6 050 . S 0546 W
w
6 ~ 6
6 WE
4£
J~I 06
I
.1 6 C
.
..
45
~ 6
4
6 C
> 1*1 C 0~I I 4U 4
00000..5~~~~~~
II 00I 00.S0 6 00
-. .4Cn1@6I
0. I ... I I I5-2267
bta . a a Ia a
09 I I % II N
-4 1 I I
Com a 1d 1 1
a a S-4 a a a -
IWW
LL. ~ . ~ d aC
cc in In:
:
W
I so4 In a a 5- 0 j Io I .
8-4 I a 1
I ILl 8~
ew I AI -C.U~ nIZ
a Q c 1I W
O~ I I C W I 01I W I
a a I I ZZa a ~Z
a~9 a21 a ac0~
NI Z
ZZ 5-, 124
_i4 I 0.fl
G5161I I _s-a- ai Z 0a-4
c)I aq 0) ( a . w a of" m1 I
V, 'L'A I0.t 14 C,
A z aJK=
0 Ewa V I I I -Joe=) 1 01
Cui a1 00 a .41 I om I WIt 00 I *
1-.61.81.. J
a 1 a 00 0 aC C a
1. a a c
Ki
I W" I C
0) C IA M a =LMa-
ws I- w 2-1.- 1:
C
a a
0.ZVII
a
0Iwo ac I
I I 0
9LLLI
oo aua 0
wa
a a a 3. 1 a a
I Q a-228
9?* . 10 Ca 5 C! I1L 5-- S S 1!

40-4 o -4 i
m4 I0 aa SI 1.
Ss w
5I 5c I. I
U. . LL I
1 0' :- uiI IJ
CM 5i. S I *..
an -j 50 I I I a I WI
COI NI
I z I II Lo(~ -
5 4 CL IUl D5-5 I S
5-~~ 5 I2 5II I
co1 I5.4iL I % V IA.
u* I.- t- 5e *---I..-Iaoj I - - -
in I i I- 5. CA 1 1 C"
01 4.4
oc I5 . I t "41 I *
If 44I allILn
0.1s C .i-0h w
I SaZ
A IjK .~ 14 055 W I Ca .J oc I
wi ~ w % I
w S ID, 4
a I Z4, Si I I. 2
W 1 . L-L.* I9 .'I * * Cc I I
s(lI
aI. 4 I= 1 010
* ) c- Cc U)I I, I A 40Ja S
5 XI 6-- 5 5 I
a' I~ Ge I :j
-4 1-
.. ...
. .. . .. . .. . ... . . . . . .
. . .. .
w.. . . . . . . . . o* w I*.I
. .. . . . .. . .. . . . .- . 229 . . .
in I e
ow
-- a - a -a -- A -- - -am----
: a,:
a .aaa
IL .
W a
00 -1co 01 *
U.~
zi aA W a ZaI O'
IL MZ WI
0 IS ai *Ia
a aa1 00.0 paa
* ~a* Mm * ac C.)
000
4
a aa M en -a
.j
:~ - z a- a a
-izI
4a " a
: a
. a
..Z "
0
_n:
A t 0-
: Ia : a
a>
I a
a a a a a a a a
S u ck 'V-) a o-a IE Z a aaa - . a aC
k
*~A a ~ - a U-) a ~ -
a La a Oa1
1~a ce i.~ Ln1a I ma
ILn .I W a cn INa
0 a-
-co cn M a 00 aC
40a S ( I oo Ia 0 ajQ I"
II aX
a a
Z Na Ca- a ma a ao 0,a a a
* a c3 e X I.- S uI I ýaa
-i Ia LL aj , ~ I ~ .
--- a-- Z ar a~ 3 C
a00. a a G,- .- -a 0
X0 a a ar
a w a -- In a---a a w
a . a aa aa U 0
a
a~
a
a 1- .J
a ~0 a
a aaA
~> a >
.J
a
a
a u al SM a aa Lla a aS~
a aa ~ aa ktJ Ua O a
a 1-41 I- Wk S a a5 -230-a ~
a f.~ 7-7-
IS m
-. ~~ 5~ La
.. 4n S
LZ-
M~ ~
I
'
I I -'II Z , A4 :
Z I .. 5-I I 2
w4 I
: 0) C. (p
5- CL 0M 1 N I
Zj Z 11 f
~i~
o
M
I
I I
Ca
I 5
5-
a
I i CID 00
-- I I C
z t-
I I 041
5.0 0 1 I .0
* @o~mas * a
Lai W a r. .
.Ls,
I
I 005- I5-2'
5.4 Weibull. Analyses -- Project 2. This section contains the results of

fitting the Weibuil distribution to nonelectronic part lifetime data collected
from a ground mobile artillery locating radar. For a description of the
tables in this section, refer to section 5.3.
In only three cases is exponentiality rejected at the 0.05 level of

significance. These cases are listed below.
Part Class T Shape Parameter Estimate

Accelerometer Forced Balanced 0.297
Azimuth Encoder Optical 0.641
Crank Shaft General 0.413,.-'
Note that in each of these cases, the corresponding Weibull distribution

has a decreasing failure rate. This is perhaps indicative of the occurrence
of infant failures which were not properly screened by the vendor, or, as with
the Accelerometer nd Crank Shaft, a very small sample of failures.
5.4.1 Weibull Analyses Sumuaries. Following is an index of the

nonelectronic parts analyzed in this section. See section 5.3.1 for a
description of the entries of this index.
¢.. .>.
S'.-o
.21bQ
5-232
Index to ProJect 2 Weibull AnalLmes
Sequence Best
Name PrtaPart Type Environment Number Fit
ACCELEROMETER FORCED BALANCED GM 1 w

ACTUATOR MECHANICAL GM 2 2
"ANTENNA COMKUNICATION Gm 3 R
Ajax G AL GM 4 _
. AZIMUTH ENCODER OPTICAL GM 5 W
BATTERY RECHARGEABLE GM 6 1
*BEARING BALL, GM 7 E
* BE.ARING ROLLER GM 8 E
BEARING SLEEVE GM 9 .
BELLOWS GENERAL GM 10 '
BELT TIMING GM 11. E
BELT V-BELT GM 12 E
BLOWERS & FANS AXIAL GM 13 E
BLOWERS & FANS CENTRIFUGAL GM 14 .
BRAKES ELECTROMECHANICAL GM 15 E
BRUSHES ELECTRIC MOTOR GM 16 E
• US-,NGS PROTCSTON DEVICE
MCICUIT
GENEAL
SPARK GAP
M 1.7 E
GM i8 E
* CIRCUIT PROTECTION DEVICE SURGE ARRESTER GM 19 E
"CLUTCH FRICTION GM 20 E
COMPUTER MASS MEMORY MAGNETIC TAPE GM 21 9 "'-
CRANK SHAFT GENERAL GM 22 ii
DIAP.RAGMS BURST GENERAL GM 23 E
"DRIVE F)R COMPUTER TAPES/DISCS MAGNETIC TAPE TRANS. GM 24 E
, DRIVES GEAR GM 25 E
DRIVES VARIABLE PITCH GM 26 -
* DRUM WEAPON LOCATING UNIT GM 27 ,
DUCT
ELECTRIC HEATS GENERAL
RESISTANCE GM
GM
28
29 E
E I,
ELECTROMECHANICILL TIMERS GENERAL GM 30 E
FILTERS AIR GM 31 E
FILTERS LIQUID GM 32 "
FITTINGS PERMANENT GM 33 E
FITTINGS QUICK DISCONNECT GM 34 E
FITI"STREDE M 35
FUSE HOLDE BLOCK GM 36 E
LUSE
F HOLDER PLUG GM 37 E
GASKETS & SEALS DYNAMIC GM 38 E
"GASKETS & SEALS STATIC GM 39 E
"GEAR BEVEL GM 40 E f
GEAR HELTCAL GM 41 E L
GEAR SPUR GM 42 E
GEAR BOX REDUCTION GM 43 E
HFAT EXCHANGERS RADIATOR GM 44 .
HIGH SPEED PRINTERS ELECTROSTATIC GM 45 ,
ir HOSES GENERAL GM 46 .
5-433
Index to ProJect 2 Weibull Analyses

Sequence Best
Part Name Part Type Enviroument Number Fit
INSTRUMENTS AMMETER GM 47 E
INSTRUMENTS VOLTMETER GM 48 E
JOINT, MICROWAVE ROTARY GENERAL GM 49 E
KEYBOARD ELECTROMECHANICAL GM 50 E
METAL TUBING GENERAL GM 51 E
MOTOR, ELECTRIC > 1 HORSE POWER, AC GM 52 E
MOTOR, ELECTRIC SERVO, DC GM 53 E
MOTOR, ELECTRIC STEPPER GM 54 E
POWER CIRCUIT BREAKER CURRENT TRIP GM 55 E
PULLEY GROOVED GM 56 E
PULLEY V-PULLEY GM 57 E
PUMP HYDRAULIC GM 58 E
PUMP PNEUMATIC GM 59 E
RESILIENT MOUNT GENERAL GM 60 E
RETAINING RING GENERAL GM 61 E
SHOCK ABSORBERS COMBINATION GM 62 E
SHOCK ABSORBERS RESILIENT GM 63 E
SWITCH LIQUID FLOW GM 64 E
SWITCH PRESSURE (AIR FLOW) GM 65 E
SWITCH THERMOSTATIC GM 66 E
SWITCH WAVE GUIDE GM 67 E
TELESCOPE BORE SIGHT GM 68 E
TRACK BALL ELECTROMECHANICAL GM 69 E
VALVES HYDRAULIC GM 70 E
r-234
6~r 6 66 66
* ~~ ~6 ~ ~ ~ - 0I.
6 I 0. -
*~ 6 ~ .66I I6~ z I
z IZ I-I z 4c " 1 6 I
LU I x J I.-. -- w I- .66 .
b41 6 I m I I 46
uj Z
26
V) 2 6 W-o
OH 4-nI
Z WI-
4 ~ ~ ~ " 6 4 I6 r4 I66
:
0 1L O I I.I j C6 I -j
cc I .- : ~
> 0 ~ V*6. * Q
0w I w-
O0.J QI6 -
t-6 6I I- I . I Z
00 4.."4 4m LMC
6 ~~ ~ C I- cn .L6
-C. 14.
6" 1.- B- 1 %2
-i (06 I U. 2--
--
-I lz(W IL Il I. at GL I-
C.)- 0 r- I L'Z I
Z~LW 64
In 0 0II (6 4M IA*
4~~ 6 06 A N 0 .- f6 I.-fo 6 .
I LI (0 . *4 i
-w .v~W
* I I 0 6 6 ICA.
0 I I6 I IJ I I I I.
I- I - 6 I-----*-*.00- .
IW I 6II
Ii.. 6o 60 Ii 6 cc6
a-.
I.- (n 06- 6 MI con0.4
I nI 4
.~.2.>...
W I I WIL
I 00 I ~ 0L6I
6
.. 0I
66
6- 6
6
7
.6
06
~~ f
Iz
L~l 6-
641 W~ 06 5-235 B- I 6-6
No 0O Z0 0.
UO 0 a C
w ~~ a- ~ ~ ILa
-
a aa
~ . 0.w
-
a aI II
fi
a
a~4 aa a aa
Z LL, aa a Z a W-C a Qaa Z
LL
09a0aa0 ~ LL~Ia
2 a
0ZZ
aa4
I
a N
~S W
l4 a4 an - A. W.
La- ai
-10
a
a. a1
aWI
aO
IW ~ 0
a 0a'aa
U-* a a -a a a U a - ac
Ia- aa a a (~a a- a C0
a- n a aa La-a a9 0 aa aW
0 c a a-a a I C a-
a aa
6 ac a
-- a-a a -..-.---- a-- a C
a a a z W.a
a 1-. a am ~
a i a aa
1a~ W aa w
a~~~ acaac I- aa -. a aa V
Q a
a gama- aEw 0
a
1.-lCL xaa a
0 ai
*
a1 a
gag a
a'a.
-C 0 I .V
V a0 P-j aa, 1aa.
waoK ..-.--- %
0a--
:
a-a
-a
a aCa
---- a
a
I Ca-
aC.~ a
.a-~a
a
a
a
~ a
aa
a
a
aa ~ a
~ --
a
a
a
a
-
a
a
a (a
a-C~
01-W
a
a
a
N
.x a a a- a a a I' OC4
x I--0 a W. go; C a :
a " L wCut -
La aW
>E
-- z - -a
a.
-
~0. a
a
a
ai
a------
a
-
a >
C.
C
a
aa a a
a a..
a---------------a
a uj a5---236a
4~~ 21 ~ 0
44 ow ot'-~ I4 I ~~
4
CotI I, : I : -
U 0 z 4
(0 4 44 W 44
4
* -w I CL 4
CL W1 - 4" :0-4 I-
0 to I I*l 4 W
z 4c" (44
4
I4 IIArI
* ZA I
1-4
00.44. a.
4 14
''
1 I 0
1(hI 1.
W
W I
z 1 0
*~ W 4 . (no4
Liii 044 I j I ago4:

(~~4 -. 4 I Z 'I 3
I A(I I IWI
1Lii 44 I
-I4 I U.4
I 1
I4
(3 I .(fl:
I If,4i
L'i w I--
Lai 4*4~
W I 44
-. 4
in (9 J I. I I I-I
4L x
0.401 I
0, ac
01 4LS N'6
404O 0d L.4 '
C4- 14 L4
9 I 0
14L1
005
. 5-23
* %
. .I
. .
I OaL I
ey 00. co -I*3 00.

o a=- 3
IL I III IIw
I L I IL 3 II~ P"
=z I I- * (0 I l- -4 I
R I' s
9L QI
I.- LL x Ll , W
I- N Lai - I I : : S-I
I A
0- I4 WI 14.
I II 4 .- 3 £0 1-W
coI 04 10.-1 1.I-)
:CLO:). CL -. 1- I 494
3 001co nI I I 00- I NI I
I~L I4 II I I=
>m 3M 414 3 .
6 W "I IU I
JI I
(a
m
.. SU)
4
I 03
-(7U-
X" NIz
FI
0
:
3
II
a:
.
) 34
01Pl
1=0
P
NI
I
L)
1- 0:
1W0 NI00-
I C.) I 5-c. I t)
I r'I 9: o L0-
0. IA NI
3 I-m I I Wx I I 1h
4 in A I In.,
I 0- CLII M I I u IC
4- I- 3 0 3I 9
I0 GIJ I0 I I I Io I 0 I I-
I J I at I UI 3 - I at
II. I U I $.- I m I 3
Ul (n I I co cn I
0. w 0 12
3 1- I . 3 I 1 I . 3 .c
I) I0 U) cn i
CTo I
Zz 3..3 **-I-* z--
* I I I I 11 I I0
I~~~ ~ >
5-238
4. - M**FJ'*01; V
IP IL M IL 6 65
-4 z 1-041 w 6 . I $O.Z I I I CLI
N mo. 0 .4 OZ =01.- I -I
&ACL (A ICO 1 I
L IIA-I IX Ii x4
W~~ 6I -I 1 W II .J
W.I
I-w e I V I
Q a*Z~ Ii InO a 1 00.11 1 0a W
Z WjA I q*II ZI> L

(CL Z -c.- ,
W-fIl
Om 1-4 gL 14 O I j114
1~-4 Iz I 4 I z I
6 I K I
w.MI I0 "" IA I W
$.- I mZ 6 I.-I >
0 io- I: P-I I WIg

I~~ I. .41 @4I
- 0
.12~ - ~ C-4.1-6
W6 0 of CI IA1 -6-I1
Z
46- 6- I f J c aw I oN f
a00. On
*~~~~I
II. oO-0 I 4c1 -JC
I<- Ix a- I I OSI O I
w IL uj A L
6A I II I. L6 _
46. I Ii I' 45 I"
*( 4 ~ wI3 CO (A IWI
I w I I I 6
Il IW6 it towL
14 I )I 6I (C 06- I .CODC
0 Ij. I .
W6-I Im..
Es-~MU6I II I XI - I
0f
I4
-% ý
*z 0 0 I I0 C
I LI I .
L6 I I6 W I6 0z I0. z
. 1 u
6 ca
L c-I0.I Ij .JI 00 I .6I I- I .6 0
6 (I 6 ~ Q 6 6 (CI-23I9
*~~~~~~~~~.
16 6J - 6 64 .1 - I A
.4 la 2 wa a t. 4 a UI a (""4 a.
14 0 Q 14a -4 a a-a CM~ a*e=
'a IK~ I I3 Ia jN I
aj a a u aa Ca Ia
a aL a IXa
a I a ax 4
Zh A z a- a Wa
a i a 0si aa aa C
aJ..a a a Z Ia. a a aGo
a~w N a1 ~ 03a 00a :
si
C~~ID a I _j
aa - a W COD - Iaa
Z I
a L4Aafl
I 1a a(a L4Laf
wi Ci a a wa O~a M, >
a-io
I ~~
a
aa
aCC:
~ ~ O-( h
Iaa
aa
O..a C -
LL.4
1ao 0 a aO a.aD
I a ~ r-.~ ~ a wZ . 3a"6a(~aa
a 00"
JZ aj aIaa 0- ta
a c. (0 a0 0"aaa o~0 a a
a (3al wia
aWK aZ .- a"aaZ la
a w*z a= a00 au Za 1a a - ac qv
q~~~ a n a 4 aa a
a~U a -- a I-u -. I aL -3 W a
a0 a0" a
a~z - a~
u.aaa
~ (3
- a
~ ai
.
f
a4 w a aa a w dcaa a W 4
a & a- a a aw~ - a Z
a ~ ao lu '"o.- Caa W- - 0h
a~ a I-- ai a ~ w -
aa- a ~ a < a -.-. a z
(li i 0aLl A ,
a aL a aD a I aI
aa
ca I oua a wZ 05 Jw 0a
4w~a 001a a4~ a a 4a a *
*~~ -0 I ~ ~ ~~~~ a aa.- la3
a 00 a a a a 5-2a 40 a a 0 a~ .
I 1 asI
-- *l ~ I W II mI - W
wt. I Jl
I*.a" .-. I
*~C W a ~ f/I
&j I ) IC
de
b-4~
z LI IVI I I Z
0 co I O Il
0 WI
In I Ia
*I
%A1 I
toI 1. -1 1 Lai5I
ag J Ciea I .115
L: Z >'
b-ZWI0
I I I W
O~l I WI.
II I~ 4II
w 1 1I)
4'- Sz Io W
I I II
=0 II - I - U f
4* 6 0 :.
W C-44 -. .- ~
I ~C) I I. I. (A I
IS g 40 401 I 0 4C K
~
* * I 00 I IlI 1-4 1
I 4 00
-. X I--
~~~I - a VI 1 co- W
u-
I ~ ~~ 1II ~ ~(Sa'II ~ ~ ~ k 61
Il
04
Z -C
I (7 ~
4c
I SI W- 01j a
Iw
u00I
i 06
~~ )
*I-a
~ ~~~
1
I II
-
U
a CmI
1
3
C)I )
1QI~
- I II W l 11
I~a uLs Ik II I
I I II I
cc 4 I II LJ 4 1
IW . x 0
0. 59
II I
IU
*I
I
II
I ~ I
Wj1 ~
II II 0>.I
I .I IW
Il IL coj
I ..JI1- I U) I .1I I 5-241 I C ) .I

10* 0 a , al C. L a
ar a - aa aa e
"iI
aa ..
I doI .I P4 CAC6(A 1
IwI 11 I. aA
41J
w a
~ a a a a a K a aa .4
o~baOai a 1 .
I (t a4 _a 1 4 f 1a40I
.4 I Ox0 a A
Oa .i
Q.. a4aA a.: ±1 1 Z
wa O* cj rII0m a W Cc a I . w
.4~LL
~-Jz
a~ aLl I
a
-aa
Otaaaa
0'a
I " -C
-
maa%
--
a - az La -.... zja O
a~~C Laa a
a jz a >1 _Aa
I a1II S 9.4
>
a )0. I -a aA a) 0. a 0a
E
a ac Z~a a, aa zW I
cc a* ai aD 00 a .4-a w 0 Ca a a a- 1 -4.4
aa a Z a (lU.a aa C Z
ai w - 04 0S A4
IW I.W - a l a W
a) Qj a wf
~
*--
a h aIQa
4.4 W e 544 a a Z
a~~~~~ ~~~ 14;aa- aua a C
a
a~~~~~ ~~~ c).
- - aaW
uas a C z-
*--- - -
a>* a aaa
aO
ai 0~ aI 5a
.4 ~ Z~u ~ ~ a
a00
c- ac
aowaal
a9
-Q.. .. a -
m
-
4)
Z ~aj :- x 0.-a
a
(D
a I aa
aC
a COaS.4 a
454
.- a E
ca aL ao a~h a aa a
w aa a a w a aa
.. J a 9a a .. a5a a2a2
a J 4 a aq- a. a a o a 0- a ' ao a1
-.4 I 0 I .1 -4 I l 1 too I P-
-e
I
0
C
0.~
. I
0)1
'
1 i 9L- ~C
I O I I 00.0 II W
*i ~ _j I I I ~
z IU I.--III I -
I I I I a-41 I N I I ~
%nI I Z I1 z I c I" to II
I~~ i
~I I
I
II
I
Il
II
. I
I
t
hi
I- I
I
I II
II
I 26
I. LoL- o O* fc..110I
Li- I I ui 0 h La
*~~1-6 . I
CL=I4 IL 11 III 14 II 015 14
GOM C1 . It O(i CID a

1-C4 I I I I 1- I II I
"" I I
I z w I II
Iu ~ 0.
-1I 0(0
I I- Z
W
I
1
1-
O I
I
r..I
II
I
I
I
W
WI
J1i
I
I NII
II
0~
Q W III
I ~J 0
0 1I
IZ
W
I0
I I-C""(3 I zoo
CL 5)O C ( I I I 4
:
I I I.4 II I It" I -i
I 210 I OIL'I I -I 21
I
I W.-0
tw
I I
II
I 'A
-cm ~ 2
( II
.
01-IJoe
M IL
I
.
W-4
Z.
lugs
I- IL
I 0)1
: 1
I..V
21.61
i : w
'0 4c: : AM0
4c cci I o 0I 0~) co0 00 -1
NIP ac I - X o:
h. I I I I L16.
I ~I0~ 111 ~ 0
V % I Ul I I
I - I I I&. UJ 0 I&..
ow I I
I I I as
I ~~L I II I
I 4 C ~ 0.I
IO W 0.
: 1I 0 m0
(3 I
I -4 hi1
I (3
31 I~
I~I I
I
0-- II
II
I -iWII
I 1-1 I 141W£1
W
II
~ z 3I- W I
5-243
* ~ ~~I 00.
w Z
OM I oi
CI I
OCL O c0 'L a*
1 C MM . .-
CW1 W'N
rw~0 N
coSI 4
La I. S La-Ii I SI 1
-Ccj. SI I I W I S JI
I.- 1 ~
w5 i~ 001.wZ ' i N I q
-OS
1
D CL W OI m WLI
I OILO V- Ij Cgt I -I
z '" <
(D WI ;I , Z
I IL
*
-. I I " t
* z I.-I
bI0uý
0. 1
ý : -
J
Z, i -- z -:"
U 0 IS 4-14 I N
0 W L...
(1- 0 wv M P. 4/ .-.-
0~~ 1--. CL X S - I-.L
00m00 a..'e!.J 00S 1
K-00 La$0
II
(0 Z ~ 4 4:
,, L . .
UII- :1 *:x0-- .: (
o~~C IIW WII
I-: I Z
I :-J O L, : 4
I X-
a- ..- -- *-
I I
5-244
-~~4b - --- --- w

I LII
CL Ik 9L I I
2 WI I I~ II ZI a Lj-4 C~41:~
a= COD L'I L~lI 02 I .0I W
x I I.- I I--Is
W I
CI NI ILZLS I CL 1 4 WIA
I V.. L V) I 1 ) CL (A) 0 CL.
II b-- I I xII-
U. 40.-. o 0 IV'0I
IIc In .
LI 'L lii 0 WI L
cc IA
W* -j a::: I
A 3A ,
(4 I
C3 CL m 0C L Im I I zl U
I cc1 2o IýC a LL)0 A
I~ W I ir
o~~ ,a .IQ %i
0 L I IP Ir Ii
I. I . I
1 9~ :
-i I IZ
". (f .4 I I
I
4 6. Cl
~0.0 h I
Cc r"I I
14 1 "0
1. 0I Iqr - II -44 I '5
(A C'I0(L( I
(A W I I
0 CL xA I n no I
*I Ico Ii Wo 00 It I 4c t do -0 i 1 00
-C-
I SJ CCLL I I 1.4 I .5 I IJ
rLL
I I Ii E - sI '(,A- 12 In W0-I
I w I X: I- I -
---
-~ - -- C C
Go t- in , I ,, ".
- ~ (I.a I III
Im- 00
DO'
--.
I
---.-
I
w
I 13 - =(A
(3wi
'A
I (I
~ u
W
:
(
--
1 ( S Li I J
-
i ~ .
5-24 5
C1 II I
N Go= I C14 0N 09I
z I I z Z 4 C 0-4
(A j ".
I
:0.
I
0.
w z ZI . Z
I -OKI w
40 _ Go I
IJ I -
.0 I -*- 9 -. 0 -.
I
COI I WI &( W
Z1 N Sol-~'I
I' Z. It- I a WI-
_ .zl I A I. 1Q
9 N I Zr ~
0 2, 91
4A~~ C0Z GIM(
. ghci I0 9 6 t
~
L.A 9 - CKW
-i.4 P- cc-4m0
u I 'r 'A. I I. ' L
waJ I :-0. I eI- .

V) 0 CI./ I *-----------?
L W LaJ I
* ~ - 0I 1i
~' j 0 IIl. b
00
LaL
0. I
C4
1 *onto I 00=c)3 t~
co : a a -, - : 0 - 1.
z I IL I. z,-D z a I
, 0- -- W aZ cm I uis
m aa aa 4 (A a a w 4
o 8Z.t n a a a 0
m lb- a
I W I I am _I I
C.)Aj I -W I . I -. II .*-
a Z>a Ii2 . >
OS CLM I
0aral -a 444La I an a I-a
I-I LaO 0 X~a I a- a0 OILa a 1Z
W~ ~ a a 4O a j O~~ 1 a W
I W
If (Da
ow :
_ Iz ch1 f I W NI a
4c""I a a W
01 I 4.~
oII WI -4I W ~.~ -
z aw (o w I ptagm 0
2~4Z ~ ~ 4"DI
a -lI up - Iw I 3,
a
3
4A X1 I-- -l JZ
Cf0.J)a 4la I (A.U
aD a I
a ( 1Cn
au
a ~~ ~ I ~ . ~ Cc a
Iu ft
u
MLI.
'A a 1 l -O Y
I C 66 0 6I S&JZ 51 .J
(06 w 5I-a47
a 5) (AJ(A (A a 0
,saA I a wa 0 -7
I MS 4C 00.0 I
"
z- 6- 1 n 4 1-4
w 4I 4 S
I - I -
I.- * z w PI I. W.S
4x (A.,OLS
(OS I La
04 4-0
4 WI go
I mm toGMM t
w WIl Lg Z
uI Mt
z*
0 0 11-4
4- I I - S0ýWILO4 I Sn I~ O
CO).4
Z I I
08 o f vi~*~ I scw -C O o l
" : "I W .
0
*
cc I W S .4
S-I
tL-
14" -4 :U
I I S S 4
I ~
4.-
--
J ---J1. LlIn
I1 S-
Z
ca -
w ".Il
> > I J I > I I

Uzi1 CK3 IZ 05 w ( W I
5-24 8
£ II O C I I
I 00. 0.
CI -,~ I -, 1 CIO 8 I-I
sk~ i N I I IA 11 8-
I to *.Z I I QzU I 1-4
x*. I m I K 0..J1
OZ"I : I
80 iI 0. P ILI
z08 I08 0" ag

8
I
I- I (A cn I $- I 118 8 PI8
4~ 0 0 IL I We IL 0 t!.*
M'4:
1: 0 1 (08 8-
8-48
I
U61 L6
0'' Owl
OZ 4
to
cc M
IId
I I ~ II
'II
I I
I
LO.
08
8~c
I
I ~ ~~~ .J 488 II)- " .J8
I - 1z" I W I Ij MI W
I ) P M 1 0 1-- 1 .4 I6 9
4I M C1 ICL I I : 4 I
0 0I- Ij a.4 I Q.I I(0 CI =8-
I .00
-1 MI -C I 008- 00 414 -4 -
IL I j j I I I6 . I
* Z ~~~ 8 8 0 1
(n ý48 1.-W -44

I 0.-I V) N
84 II I4w 0 ixI 4L8 (0
8 8 I I I I
IL I D Iw 88 Z I ww
Ii in VI I
f I Io w- I 0
(AI
88-?& :* a.
24 I I5-2I49
0. ~84 8-
I I -- 8 I- -. 8 . 8- . N.
W I I4 > I ) - 1 5I~s
I--I,
COM S I -aa
CL.I
4"
, 1a~ 1
(D
~ 1
I : z040,U. I IWe
Cpq"
ON 0
0 I W-
oý 1 MN . - 9 :
I I 1.-I La O
V9 0 In3 643 3 11 3
w W' Z rý I , . I
I~( 3 3 33 1-
3-~~~q I 3 W 3 3a
IA I X
3. I a- IL 3 j Ii I3 O23
O 0 4n : ::~
3 0 W10 3 0
0'0' -- La-- 3 - f 0'
*~~~
iI W I $ . co. I WIc *
o 3 I3 I 1J & g
3
(
2'
i
r- < S W3 :S Y
0OU t I. I X 3 3 0 F3 A"
-aj 00 a I Mf
-I I00s
I~0 233 I 34 74.-I
3 ~~ 03 2 * i%3 !, -0
L2 %
00-
z 3 0 3 3I 03- s 0
I II 14 I &L6
MW # IId
I 3 -3C (z I
W 33 I II IL i 4-
3i 004
cci I 0-O 3 * 0
(n 3 ' 3 o V)d ~
Ip I
W1-3 ~ 3J 103 Li.
I i3 I-aI
A,
L
I
,n q C1 ON)I
04LI (I 01CLO
I I I
*M IA - I - I Z I uji- 'I
L I6 ILr4
1k§- ("I 1-
U I
I I. Wx I~ W
4 aC Laji I A I II
1.e 0 - -- .
10I
W- 6 6
LL~~~L I I. I
6-4 06I ~ IIl .4 ~ d I C%&I I b
1IIK ZO >. I w I rI I >

ICL I -A
Ln s L. I C. W IL
Uz Iwj IN W
i40.I I 4 : 06
0I :s I l 1I- ..
*c Z I I I Ia I
IIL~9 I. 1SI 01 0.- 1 i
I.- IZ 22I'c=( i Z 0 jg=1 0
z l o 100 I 1 1 I O -10 I1 1
I4 0
L .. .. .. I I)
I -- 1 K-0 L I I LA" . o f a
1 01 A I I W
6 1 ca4 II I 1
*L 00
I
N-
00 ('di
I 6 N b"'
Us~~C IIii I ~ I s
Co Ixug I I~ I
(A I x 0
11- I =(I
0' I I Z
5- 251
LV'..-
Z I -SI O .
co IjJ S SlS
Z 1. 6 0* 4"64
w I.- 5 I-L
z w0
0 b4
8-S837
5. 0 wXl 1) 1 Xi
S -1 SO S S -I (A S VS
U i * i ~
-- 5 -e
--- S -. L5
S~L S 0 4 I5
OL
9L=
s
450 : IAL
I 04s-j1--.
: m 0oi
~
1 co S0 to I~. I S I S
I LIC
I t c .-- . I !i .
* C4~S I 54-C0 WOI M S

:~ I Z :5
_iw 9: _.jz m C
I Cal -. U::
i :l
I- ~I
wo a.~
"Iw L
II
I i -4 1 wS.I II w h
(A IC" Con 'A 0. SI.. S
* I i c" -
I.'~ Ii
or :' C. S -.-.--.- 'I.-.S
P%
nor
3xj I.%5
-
0. S
S 5-o 3c 10
S 5 5 h 5- 5 1 0( m
C/I 01
W 31 X0 1
0~n I 5 5S =(A5
CA- S S SI
5... 5 I
> i i
35.l
:: '
5-252
.........................
LU
Q 6 6 L CLZ w0
Z -CPi 40 Z 4c " P1 .4r
mý g*
us a =00.- g
ag % 6 -4 46-1 0" 1 4m6
w: W6-0 OS Z ai 41
o : OB- ~ 6 Os Z i- Sa0.
w I 6 e L 66 ~ I
OCO I I 2' 0a
I-~~ 0 - -W I cmJ
f - a
6 UO I 6 6 L6 -.
*~A I'- ' l6 I.
a *CO I V)4 CL V) 1
16 4
6K A 6
0.6 - - a _--cm-M-I I w - CC-6
a 0 1.- CL 6 u. 6 0 C 6 1 0) a
* q Isi 6) W 6w
a I- 6 6 6j~~~.. 6 - %
10 0"ZC 40 0614C4 6
I 6i
m Wn 6 .- IW I
I a 9 1.- 6 I
6,. -. ,l.l.
LILL~:
Z I w 6 6 6
I. I6 I *- -- 6 *--6 t..cu --- a- --
w a '1 -1 ~
1T 6 14 3 0K~
Z (A
~ f 6Z
I.-------------
5-253
cni K 1**~* r----------D --
CI 1 000
N 00a~
m co£ 00. to * i
9K
6 -5I
~
0 4c I I -
OLZI I 1
DI Iz I .Z vIW '
I El
L~~~~I& ( Z
-j0 jc I 4II
0 Il
* L Z r. a Z ->
CL D- 00. I : I
C3 CL 0 m 6
00 L
*~0 J %
6~~~~ý I I 4-4-4 I a II,14-
I A I '- Uz I I I w I IIM I
I z I- aI I t I
I j I U CVf M(
*i j z I. I "
L) Op.- I I5 C 6
6j Ir s w I COI
1 I1
03 -fI GoI I I I I I
N 0 I 0IQZ 3S OM I LOI
d 00l m I
~ ~ co -1I 03
~1III- ~ 6 ~
02 #
6 ~ ~ 14 ~ 5
.
I US LS6
0
ClI 6
V,
'6
* II f6 4 (
LA)I
I' < I I I <U 6 l %l-rI0

I LL64 I IY I
cc1 I IL I4
1 CcL
6n I- I wEd- m I Ed
L~j
X::AJ-: : m X.
Cd2j A1 "o-
w I cc I wa
h 5 5 6 I I I I *-
* 5 6 6 7 I > 5 6 O 4 I*.M
Sz 5
X Io I L I L
a w sA s (D a ... 0 U
(n > . /~
5-254
I. .
OIL
S C
CID3 4~
NL I I
uLs I. IsI
a . I- X I
g9L4CA (l a Z ~ - Zu I
09-- a1 ado~
3 uiuja 4ui
Q aa1 a -
U 41a
a*
a i a
u. a a 6
a a 21aa
a ~I a'Lon,4 I ~ ( 33
~ ~ aa ~
a-
LL'
~ Z
03~ ~ 3-aw ~ AS-Z
wj 'a~
I
OCL
3
- C
a 0 L-(0
z C91J0 0~ a -
-AZa
a a c ac Z
a
9o w
a0
aae
4c a I- e I
a u .
ccI a:
IA
Ia
-a
I I
.
Z
L
:) a1
O
1
a, Ln 0 a . Z
aa,4~a a 1- @ U .3~ )3~
0.
LL
a~~~~ a
c0o Ua 0 7 - a n I.
-: 10%~31
a avZ
-6 a
40-4 a a 3 434 a
' 0,0 0
>
75-1)
ad CI
a 4~3a cog a
5-255W
Q. Ii 84 0. m I.
09LO I
r4
1 * CO I GOIm O
*j _j I Ij I
w z =I .I . z
2w I, 1 21 2 WI1* I W
VIC I E
::
z PZ qc z Ib A I4- I l*.I
o Mot- 0z 0 MOB--
1.4IU?0.A 0P1 I I
I~w I 1. I l
WI 1 W
I.-:"~OW I >
z- a In Z
a
I
IL 0, Qo
0 w
I -j- ca I -0-i-.-I ca * I.-1
, 0 I Owl w a IcwI
* .IZ I CUI 01 z I~ I4
4 c WI 9 4L= I ca 1 9 LWI
* 00.
4
a*:
* CLl
ro
tco
M
I ~ II.0
I
0_j~O
It0 3 I o :
I
I
I
U.jI
I
I I Z I I Ic
IZ- _j 1."1 z. 1
w 2 .. L..II4 wI IL
C?.V) I
I ~.- I II U00
f04.(A
I ('I
01'
I
I
I
)
Ll
I
r - - ~ W-
1
1.. .-. -.....
%
S 0 I I
"
1.- I *I 0I
~Z"
l
1(-I I .0
I I C .- h NI I C 01-1
I" OQO
8:0 1-.1 .1 0 1-0.2 Ufl
F I 00J4 3-j0 o11- I 0
00 I 1. 0 II 00
014
OZ cI 0ZI
WI-.4 I I ~L, I I WII I I el w
Il fI I I i I fD I Io
I I9 ~ Z
14,--
I- .II.
I~~a -
I- 0
II $---
c
CL' ,
I:~u I ? I 1 U? W
I2 b-I Idj I. (.1 :O ta I
ID 0 l* IL I V)
It IO Ic 39 O=
W z0 O xW
5-256
w
44 Ck M 9 C .999
0?.
ca* - 00.
a OIL
(~J
I 9 @0~0 9 9 (S 9 O~0 9 49a
W -i9 )*
9 9 9 1 9 9 9 9 l.I
Z I. 9i 9 Z us $- z Z
W . z
*~ ~~~ 91
9
-9
4
I :1
I
a)999
9s
.
4
Z~ ~ . 29W.
Z = W2
$A1. 9L V)ILI>I
IaJZ~
~~~ 1-I1 11 9
z 0 z AM.J I9C-i
w o 9. 1 0 9 9
00 Q , W9 a c9 1 9
9 I 9 la9 9I 9 CI-.
* 00.09 9~009 I J 9 0. * 0
00. go: -i
I : 090
M Go W 99.9 am co cc 4W
U
0 Lai
I~~L
9~~~~~- 9i'
~ cc(.)'I
9~~I 1-
919 04C9
W C. -4La
I- 'A 2 - 2 P,
Wl9
9 -1 9 .J- 9 9Z (f00
c ý4.4-4 1 .9 ' (l 4-<J 4 9 90 3 9 (l
W 9A 9
I -9
fa
I. -z
P I l-J=: -.-. # 01 1-Ja 1 -
0 08Z : W : < Z 9j 0) -tw o
0 0 00 (D 1-9dL to I.- & = In- I
(0 ~~~ "-j41.0
Z 90.909
La 00.
W~ 9 IA
OJO ~ cn 00
Cc 4
NIJ w- I9I09
__~~ I 9 2
9- 0( * 0(..
9a I9 z~ 9 9 I
>: >
5-257
P,
aA CLM (nf at"= 0

.4I $0. I C 4 ?0.CLI II C
00.0
I I I MI I I 00.0
m
I >1I IU
o j - I I I I I. ai
I.-I
I-I
I. IIZ
I I.
I.- W
I
I
LII
a -3 0I cc -Cc 1I I0
U.. I I I
w 0 I I WI IL 6'W
W
I~z '~ >1 I >1
I* :) I I e m I m I
I ~ Id
003.
MI
I * o mc
I 0
11 0- I- - * :JII
-z I
U NO Uzi I Lai *
I N I IJZ La , IZ
u I u oZ II .4
0' 0.~CA 0. CAI I 03 0.-ll N
I- I3 I - I I I I I-
o Lb. m 1 011 0 I' II 1 &IIU I.
0)
<L'1 1 W ga
%i cc 0 L I- I.- Z ' - -
W LIS
Z-~iJ~
I
LI--.----..WLK : "0: 3Mý
I='-~ 5-25 -
I aS:
n- 9 fg: :
S I
909
S~0
I "; - I
101j
%
z4-
:
z Lu z ZS Z I 9 ZI
moo 0 0 TW
th CLI (A
w
09W
nJ9
I I E
0 9 9
1
w 08 = I 0
1.
co
w
-oLI..
C--ID - .~S
9 49 S 0
-I I
9 99 _u
I 9
LuZ 039p
994155
> 9
S 4.
W
CL Mu
9
I N
99 9 ~ I
9
w Z 00 0 -j
I 9 90 0 9 .. 99 9-w
9049 9L
99 JC*
9~~~~ 090 LJ4
Zi W
w .L '
I ~~~V ) 9 a-
t
49 Z 5
45 C 5 99 4 -4 1
9
0.-(hS 0 .S9l
cc9J=
09
La Z C 09 (O 3 Ln ~ -cci 0 3 * *
- , . t
0 0 5 ', j -I 0 %9
.x 9 -JW 9 p% ..
9 9 9
Lu I 59 I 9
u2 5
z 4 90 4 '~ 0-'
9 ~
19
'CO
09
L.w
W0
ia'1-
9 9s
9 (.
*
L
WI
I.
2E
9~D
9
9 -9 W n 'N
0. 9
CLS~
0 W .
CO ac0 9j 499
70 3:0
QI 09 Lu0
Z i ~
5-259
r- % a ac a cm a N
4Z Ge W
I IWa z qt Uzi
0 70-- - 0 aaaa
0~ - I - 80s
aIP aL aA CL a OC A 0 aC
- a L aI.-a 1 iai
a a a ax a a
a a aa ac
a-
a 00 ( a a- acoI-i aa X
U-: ac a
aa a
a ~ aa a -c
aO co a3
a CaI
a~L -a ao 4
1O - - -.
LL : an
a (l
a 0 a- S ..J~ a 0'a a-
aA a- an aOaa
j a a a
-tm aý a-Z A.aa " - aa a.
In*
a 418D 1- I < 0J >J a a:a >1
a 00a (- -0 a .0 00. aj 00
a j a - Ia a aa o I a
a I.- a
I (J -
U- J - a -a ( j
aI0- an a 4c
a ' a- a *a a Zia. - a - I i
Ua4 a (0~ au WI -aa -
a~~ ~Co.a. a ma0
a
aa C
a a -- a-(
:l --- . I-(
a
aL
a
4 aaa a 41a*
aa Ca a 0 aI C Cc.- (l a-
a~L ao a 4)Iaa
a
a ~ a Ca ~ .Jka
'a.0Z
C aI
M
n a Z a Ca
a m
-0 4
a
ac
a
~ a a
a
T1-O
0
-aa
aa 0w.. a m ~ a
a
a
a
.U
0
a a
... .4a a a 1 a
LN ~ a iaaa
2 4X Ia aa
a >
aL 3E a a
a: L a ( 1 ua0 a . a 1a
a a~:an a 5-26a0
a n a
rr ?S fIW'
WiP a ~'AE M~
S I- z
La I.- : I
'M 0'& I
& to
m I Olic
D- S- I U
Z 0a)
4c" I S
: l 03 : ;'
l -I S .
to Sn' .41 z- 0 0C 4
ccam a
Am aj
Z
Ij >6. I I M3IiS
t O~CL * t en I
WE E . M I
I, MI
a--.WICD I' I (D
SI )
L II
aL * Ln
sI -- Int
- Q
*~~~ -1Z
OIL", 2 £j
r~~~~~~.
i N I Q awcO
44 "- n LinC
5
a~~~~L
5 2 '
1-CL &on
.41 cc C. t.-& . .4.
43 0
a 0 00 *
00 '
I 8 'S LW.
Id - IU S ia.J S
I IJ
*~~~1-
fl61. W0.4
I IIC Iz b
X IA I S
*M Io (AII
Lai I I
!
Z I
ZA - z m~:
I I
Sj I
: 39K:
I I 2
do
(A L) 1 - 260
to"Iu
* a - -
. . . .
* .j - a.. z
r f-q
A.~r
rt' -. t-r-.w~w-.rsr-a
I~ W 1
00.0 I:0 0.
-gu I -- a -J
0 mo- 0 ,
W CL W 1 0- a I Ia 4
WC
AA 0.
t w * II~
00 -. 5 I- co -. .
--- - 2 -- - oI f2 .-
La ac 0 ar 40
Z IL * Z C I >
* 00.0 a 0.0 a
OO ~~I Il ) f
*~ a n i z iI Z
M"aaa
a L* 17 : I %I
:A b i i4 -r a.
z 0-Ii I- 0 %
*~- *I CJ1. r- I owlZ22
S I- ,
LX.0 4L
*~~~~~~L a--. --- *1a -
cc)- 6 Ij P - I
ccI
a: " :a- 0 - i w a-
I a a am ata
CL&
1W0
a) aO
a
9ý La2
0
OLx
X-
W
(n
I
W. UmD,
0(f
~' IjA
5-262C
a~~~~~~~ a I I- I1a aati W W

a ~~ .f .I . ~ . a .A . . a a
I, COM o to
0~~~~~.J I (La I II S
Ii( -4 m- 1 W CL ) N' CL
NC
4i-' I. I
.I
r 4E
0- I 1 Zm01 1~-
m mI "Z'__ _
0: 0.
&/ Z./ S
t* p- mao Go
U.U a . n 2
L6i
SS itO I n I 1 44-I t. (CI I , - I7 7 7
orW Z mn. I: Z2 '' I >
CL n in '-4
cc so on- : 00,
I I4.
4* * 3 C
I- 0' 6 i
:~
LLJ
ma'-.W
' -
: ..
:
-a a
S-
z
< it
SI-
CL, 0 : ,0 - in
0
Li : M Sj I Li OZ -: in - c
:#. Go_; 00 Na
:%*-41C 00J
I IS
00c
I.i
U. _ I I
I4 I n r in65
w 9i . . . . . %.i~ ... - ."

a. a....- M
Go OD.---. 1. 0.-
0l c- CL 0
WIA I -q
;c -- I '
CL k-I I I i k z 3j I
z. :
-i" a c"
(A 9L (A 0 0- 0 I
I. 131 3 3 1
I- /903/
Z. .f0U)z
to I)
-~~~ I) I
C
I
C,
IO
R
3
I~a I
I I ~ I
R
1I '933 1-0
Ii WZ
4w I I-
03Iw i
2 ui 0 > A > f
('I II 0. () I MWI
I
-Z 0yI R
I I w IO l I a.9 &I.
I 22 IL 1
I"
13.1-
L" (D :3 I WI- I IIL 0.D
MIA.( "VI.
I I I I z I Lq
I o 4- Iof WO I Go I
0 - (L 09 -1 013: -LO
1-0.1 103
00 tn 00 I0 < j cc 1 00 m
ton IL : I L IC1 I
. I .- I I P. I I
W1" IU. II WP. I oI 0. 3L
~~z 4
fl) 'WI
20:1 :i
w!~ II
I 20
39 ZO
IJ
D UI IA
a.
Z >
9
I I 9
~l II
I I
II t
I
5-2 64
Nis.
CA I k
I ILoll,
5
I01 * OIL:> I
mca N O LAII I (~
u* I. I II
2'J e ~ SI I IjI I w
I-- 0 La. -W I 0Liz 0 . P..i .

I X S I- I SM X
I I IIS ~ I be
*al I I (. a'
*S I I CD' S-lift-(
~~~A.~~~ I U. 5I C, I2
*~aCO
oI Oui 'i
I. **11-
-
:
a C :~ :: i
I
L.1
1 I:
Z I. ci 9 I -Z '-, I (0'"
CaS Z II MCAIS1~-~~
Sc W 0I agoI I
-Jdc "*,
S~~~~~' 4 I' I 1 II I 1"S2
00.0O I
Ca 5l1
s000
SM
4I
:
L.....~.. 00
1
4-4
CS L& I Os 5
Cow In m co
zz c
Ca~E lSaw-I
I-- CL -- - C )'
f1
I a
e- > I cc wI C - .s
5 ~ ~~4 DO 0 5 I6 0 * I 10
* Z a fl 1- I I
m 6 j I' Q IU WI7
I SM I a
I I%
* a. I ~5 0.I65I
fa CI ( II
- e IL =0 0. 0 C? OI = I (:* 0
4 I 0.0 I II 4 I 00 Cc
0 I
OCLw
,~~ ~I.-w I X
U~~~~~
I4 I * C)
I &a P4 . 001 (A II
I I
0.4 0.
LI
*
-- 0
OA
Go 9 ':
:ca
(I
w
-
I z
0
t4I
a0 : z::
to
La Z) >fZ 4 AZ 4f 1 >2
CL
:) w4 W CL M PI I
N: I:
4L I Ic
_ 4z
I ~ 4 t 4
JZ
- I I 9-(
Z
DO
" Z
ILI V: N :' CL:

I 4 *I I 40 Q 0 cci
WI- ff91
JgK 0)9 I WIIZ 1 0 4 fj
-Je I 4
Z 0Zb NI l 60 I 06= I I -CLaW
0 t
La C 0 Iq P- 9L o 091- CLX .
InI
4m Ii 0
I i 2 0 " I Z< Ir" 0 4c
I 0. 1 4 1 1 W 4n 0. do I , &-
I - " -. 6'- J Iw
4 x I. w
0~ -
(19~~~I
IC
Ii z V -
~~ 4 ~
1.- I >
C12 I. i:l I (~ w a
0f IE
4
(x
1 1
0 ~ 0
,4* 4
5-266
.... .. .. . . . . . .
Aw Z
IL $ a I I 5
L) (LZr I pil
ZlI
00
K0o
0 Zj
(A CL (A
z Z
5
05
5
AC0-
0o0
S
5
0.0
U
us X u >
09 ,- (D I
z
u 5
00 c'0 I 1 -il W I
Q
5 .(sIs s r' Z Z 4I~ s 006Z
5011 cc:a
LLs Z CD s~
w Zc
W, . -,. %.
Oso -I
WI0 0 LAW
OSCW
S ~ I I
IL
-
a*~~~
I~ I
MOo!s(3
III
t
u..
J
Z0
S O~l...qc
I I L)I I - 5 W
LIJZqs I 's Z 01
IL 0$' ' 1I W
.7
-j
4-
~ Z ZI
DO
i I _*z 0
1I
Z")
1
Uop w u5-s Z-s

L CL0 I)
L0 t
01'4X 0 wX
IL 1 . X0
Ii I
Si
0. I -
V) M o 0. w :
ma:I LI
f 00 0 ~*5
i I= IA I2 09
In O
z Sz
>. >
-I--- I - - --- - -.----
5-267
C,,
Il z I l0'
Ii
ZUujI
coZO. 2
ujIS 54 - SI ~ Sa a.
NL ZZ40 Si00.0 z (uIjI
W) aL a CLS )C ) 0 11 O
r4~ t IJ La-4 I(z
Ca- :1 a-7 a 0 1 al
I a I I . -J S m oo a -sa
ULa
a
ai
InO
4
CL ua
0 " 1a ji l
O
(4
~ z4 (l
*L l
0 a. u 0. %n I.
a0 z: : ckz
0I CICa
I" 'A I-
C, V)
I- L X Ln I. CL In %
a-0-----) 0. V I1-----Go -o 00.
aj - - - - Laia- ~ -
40.a a
LW- W" 5
Ix A
Go V) CA ( oV
- 3. -0 a - 0
* 0
a
~~~~~u)
cc u
j
.
N
3~ Wa a ~a a
Lo
0j aL LaOa
ma 5 a I Ia
*x . 0 x 0
a-
- - - --------- - - w.. - - - - a)
x -- -- - OU - -- -
Q-26
ui.
. . . . . . . . . . . . . . . . . . . . . . . . . .
t- I a-4
I a* 0- MA- ~~ \
-- - " '
I 10 L 1 a~l
-' z
zooa--
. 0 Ia " I b I%A I
a
xo4-Z -a
aj 040a)o-a
- -
I
a V
Q
(A0~
.
L;w go - co
5-4 C,'
0a
0 -0
aI LM
I a a aI a
0 , 1 ta
3r~a
i
'a ~ Z IJa 104
a a5 ~
I a a5 C
U, WZ 4C0 5 >
X 0'
&I CL (A C) 1. CL. a
u 00. 0c a~ cc00a a4.a
* i w ala
Ch -ac:- Q m0
XLW In a hIa (a
0 0a CL x Ln a WL
I go i co N 2' p -j
aa 0
V-4a
*~0 IC~ I
a (f 0
La
---L)
aaaaa
j a -
. 4 ~ aa
w 4w, 6#.. a i *.a

10-
I~ a
L~a ~
~~I
*- I C~-4
~ ~ a WI.
~014 U.
0-I
a Q5
2
*L- 4 ~ a 14I~ 9 aI~ .
V -,
cc In W 4
V ~ ~ I ~ _ a OV ~ I ~ a - ' 0
00-' wa : " %00
i .. >
z qa 14.. aa z >=
w 44
5) w
- a a Ia5-269
Tx.
I
5.5 Weibull Analyses -- Project 3. This section presents the results of

fitting the Weibull distribution to nonelectronic part lifetimes collected
from an air defense. ground fixed system. For a description of the tables
presented in this section, refer to section 5.3.
The part classes/types for which the observed significance level was
below 0.05 (supporting the Weibull over the exponential at the 0.05 level of " '
significance) are listed below.
Significance Shape Parameter
Part Class Level Estimate
Drive for Magnetic Tape 0.028 4.649
Computer Tapes Drive
Track Ball Electromechanical < 0.0005 2.319 -
5.5.1 Weibull Analyses Summaries. Following is an index of the

nonelectronic parts analyzed in this section. See Section 5.3.1 for a
description of the entries of this index.
-A
.. . .. .. . .. . . . .. . .. . . . . . . . . . . . . . .
5-2-70o*
L-. -' .
Index to Project 3 Weibull Analyses •

Sequence Bea t
?art Name Part Trype Eavironmerit Seumerc Fit
BLOWERS & PANS AXIAL OF 2 E

BLOWES & FANS CENTRIFUGAL GF 3 E
COMP R MASS MEMORY IXED HEAD DISK MEMORY GF 5 E

COMPUTER MASS MEMORY MAGNETIC TAPE GF 6 E
COMPUTER MASS MEMORY MOVABLE HEAD DISK GF 7 E !
DLIVE FOR COMPUTER TAPES/DISCS DISCS GF 8 E
DRIVE FOR COMPUTER TAPES/DISCS MAGNETIC TAPE DRIVE GF 9 W
FILTERS AIR GF 10 E
HIGH SPEED PINTERS IMPACT GF 11 E
KEYBOARD ELECTROMECHANICA.L GF 12 E
*LOW SPEED PRINTERS DOT MATR.IX GF 13 E
MOTOR, ELECTLIC SERVO, DC GF 14 E
PULLEY GEAR, BELT GF 15 E
PUMP PNEUMATIC GF 16 E
PUMP VACUUM GF 17 -
ZÊNSOL/TRANSDUCER/TRANSMITTER PRESURE GF 18 E
S ENSORITRANSDUCER/TRANSMITTER TEMPERATURE GF 19 E
SWITCH PRESSURE GF 20 E
SWITCH ROCKER GF 21 E
SWITCH THERMOSTATIC GF 22 E
SWITCH THU1MBWHEL GF 23 E
TRACK BALL ELECTROMECHANICAL GF 24 W
o- -- . , ,
....................................
..... .... ....- ....
. - . . ,
.. -...-.....
. . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5-2 71
gsa~N- a,,a
61 :
1:AL -: 1a CL I
0I.
-s *0. : a I CCa
00. i
co al
I 1*a:)c 1
P I C wo-J
X-II-I W Ua U aL
SL (AI
a
a C
C aia
L
a aI (
:,La @ W
m: a- A. L
aA I a I Ia V WI I
WI a I- a *a:
a18 z.Z 0.
w- I z z
IL I- to I .
O M. aCAJ 4a- I Wc1- aI .

0.x-l
* a 0*a
aD
0 1-0.-(l
00
IL 00
1 ~ j cca S 00S4a
* Z s .~ a w a 0 JXO
a~~~~~~~ a W' -. wL a A
%
a oO a 000
C a-.±a
a (2 5 -0.1.r0
Ca 00 aL 0 a W QWa a
Z ~~ ~ Na
C - a .1 .3 aa .aa a
a-a a aI ac
m a I
ae a. a
~~~~~c
' a a 6
a
a a
I L a a. Ia ý a a
a
ar
(A
w
a
aA
~ w a
acoo "s
aIa
I
~" i.~
..
I--
MCaC
a
* ~~~ :9K a -' a

a'ILKaa
ac I
gs
a
~ Z~
aa
9L a Z
IL
a a~
Ia-
a
*a
a
~IJ a
S
0a
a aa Oha-
a-. 00
a
a
a a
>z
a
. a
~~~ a a>4a
to La a a a a a I
a a
w ------- *- a aj a-272
'ILL In1. 6 CL I
*n 40 m C1 0I oo I S 40
at P4
0 6 66
4c 8I SM
"I
00.0 ~I S
I
6
PI I-
66
CL U I I. I (A CL- w .
14 6 5 I L I x
41I-i I I) 6 46
W2 I 1 6 1 -461 -
kc
1V m 4 406: I 0
W
WI IA Oj .16 WI
Go 0. CI lI
* 0' ~ ~I
V I VII 0
~0 eI II ~ CA w I
D- e I #.IWut I $0OI W N W
s COMM I in I -'I 00~ 0 a w I
I 10 (0 W&ni I 0 I (1 IL. s
I I C.6 6 WW0
b-II ~~ ~ ~Ul -i40,0626

nI 6 -W 1 00
I- - I LA 6 I P I I I >w- 11
I P- 1I . o- 0 6
S MI W 6.41 jZ 1 0 1 ) 6 I"
<"P 00I C4
*L 41f 00.
64 'A4
In 6L I" 6 W
cr -, AM0 I N I 1 0
*~~1 go - "
U- 08
~0 0 I.5-0
001 4I
4
N ~ 0 CI
6
IL
41 0
_Ix 4 ~~6 u..O

o'
(I 0 6
O 41
600 0.( 0 6 N
6 I.- I I . I I6
406 z~I:: t :4
".4 1. "- I. 6 4
so 4 I
Lc 6I ~ $I"- WW~40
El2w W I 66 IC0e
I W I N
CLI I I
W LU I
I W WQ .: a.u
4C WI I C Vd
z# I IA I 6 6 LL I
co I- *-.*--'
w-- ------
5-27 3
* g2. .- ~ ~ * * * ** **- . . -*.

F Aq~VW~b'~Y~ ~SZP..C. U~PF9
a1 a0 aW aI
a L ac
faWI aL Zea
) Na
an ~~ ~ ~
ri am ~o 0 Goo aa
co:~ : * ~ :a *
(~ $.~
a 0 .0
a
~ S Ca ~ ~ a 0
a m a 0-C a a m
LaD a II . I.- - Q . .
z * 0.
82 6
a
QC
Ma
Cw
S
f
a
0- a
zWe I a C I- In I
::~
Sa so -A aa 0. 1f 0a.-1 ~
I4 I
ao a aa IO cs I In
I- N aIZa - -aa *.
CA %V*
a a a a a a a r t
* a L ai a0 w *
~~i
a a LOUI.J
'aa
* ~ ~a00. a1
. a a: ..J a I=Q a a
a~~9
a a 4a
0
a
aaJ- -I
a C
~ ~ f
a- ~ -. ~aa 4
aa
- 0a
a C a aCade
a c ao NaIaa~ Z a CI
a0 -JK M a Za
a 00.0 a *aa aa a0
a a w
. a1 00-C-/C Ca 4 aa a 0C-/a
- Qa
*~~ - W
w a
Is w
wn&- a CW&aaa
a uw a w a aa w a /a w a a
a- a~J
a *a a a- *a5-27a4m
~
0,-a me *-C Z~ aa Na a7
1'ý0 I,
-1.
0. Oa
OIa-- 5 -- -----. S ----

C, S SSS
it S. aa CSaa a
I
aL a maa _a
C aa C
S3 S E S w I 5aW S C
.- aa a - t--a taa a
aaJ 00i 0 aa a -a
a a
(~I S I IW i
LSa
Z a 1 a ) Sa a.
*z a c 0. a a--a a
~~ ~ ~ ~ ~ (d4
1
ka . a 2- aa4
* aa
C4 I w a aa~
*~u a 4 us a aa
r a a aw
:
a ~ a
:a a a&a6
at 10
E :~ :a
a aI
l~a 0 ~ a
a Z
-
a a
4 aa a a ~wa 4 aa aa gj
LL
-- a a a Iaa
-- a
a Ot L a aL a1 it IL Ma a
aa 0&. a a a0aa aa 3
w aa aO O~ a WaO aI O a *aa
1 ZP - a .. JZ a 0 S5 .4
1a
Pa a ChIaIna a a --- a
5
C
CA aL aA (Aaa a
aICea a
a aaa aa~ 1- J
a4 mO 5 a * &0-a0 a
a a (o awZ
040 aa >
a~~~1 40-1 $-na
sivawea w
S~6 a49 000a _j- u
aj
a fl~0a~a
a aa
a
w
aa
a a in~u a~a
a PSa
aI
a aaaa a
aoo Z aa E
a, *a 01 u ae
'A C4: a a I
w aA .- -
CL 5 S5 LA I 5 5
0J de aj WE Na .J a *
a a094 ww a
5-0-5a27a
0.4 5
lift
I~~~ w ac
3 3 323 a= cc3
3LA -4 6-4
w I x 31
L2' C" Z Owl-

a 13.Z
-3 ( 3.0 3 4
M C4 I I M
3-41 3
IA.
1o
: *%
.
1 1
:: ~ 1
&, :
_j I- 4b
00CO 3 M oCo I V: I...
3z I I. I I ~ z
3 j z I en I jz 00 10-4
I
3 f3W
*~e I I::~:~
3~- - "- -) U3 I 3 U) L
3M 034 3 0- 1 00 3 3
3
1 W3 VY I- ~ C433
3M 3 o I3 -03 3 0 o- I NI
>~~~~Z I3 004*
0. *
Ice "
0 w- P4 -I ccU 3
$.-, 40 1y1 4
9-9L
cI I a0 11
a: >zz
35-276
V4 CL -4 15
0
z0 z en I
00. as IL Z5
~ a I:
z 4c"" -e
~ Z 1141-
:LI..
g ui
0 Mo. I 0 W
1..4 I
fI I I
0I O.JO -I IIL ~ U 5s- Q -
i rsIi $-.
La. I @55 w I
_* en
4 I I O li zl 6.05 I
K" I '
6n U I ovI
U I Go I- 1 60 5 I ZIS
(AS
CL(A-a~ 5
flI
I &W4
ago qc 09 1 n51 4 c1 '

ISO...M
z U)I I
5 I0 I4 ui Z 0 -JcC s-
*I
0 0 IC L
- I
I,
. I
I
s-9 I0IS
I0 -j ca
- IL
1
C')
1
0. ) I en I W
I~LL I I i I I
I I II OS I I n i
x I.-%
WI
1-4
>
0
I
Z6.
z0
I ?.I
3It
KJ1 W
u54 I
I II
0
I W-
Z-
>
5
I
45
WI
I ~
wag
a
za
6-I U la. I w z LLL
5-27 7
*4 ~ ISd I I OILIS
.4 -j I -, j
I. : Iý I I.-
I C- Z I 9 -- -Ii
x I.- I i
W*xI
Ii oc I
" I"s- " .AI
I.-I I. I IS I I
* 0. Z- - ---.- 4 *43
*~~~ I 05 1 Q
0 0
I Lait
Id 4 I II N
La Z L2 'm
be A. M of I W:L 0 9L C4S
I 9L 0 S ZI. e O - I22
I 'a .I I Z I-
U* o I Sn I Iu w . I I I I
Z ZI5
)"
Q.4 L o. (A 5-j" 05 .- fS O
* )( 001-
ac41 I
C14a i S ioc S Iy
M* i 00
m 1 0 1~ v Ii do 1 00 1%W
_S If
S4C of Ua.L I W 9IK

Ij 0 1 ( I C-1
W u(f LI-I I
U) F- : COO~X-
Z.0c:
I S
5e ZI L I IM I
A. co ca &U
0-: 9L
5~a 1 0
0 I
IIAI S * Si Z II
a. . 5 LL!5
* (0 I II I I I I I
5-278
C.) I 0 Ia0 C. fn.d I3 6 c a a (
a. cci03a~
I2 Z
I IL Z-
Z
0 of- 03": 0-.4..
lot-3
I.-5 x ai >I aW32

Ic I~ ai
IU4U,
wow
a- O C-L-.-a..
C* 0 >
-*-*
1
a.4 10. LJ . S. 3 -.. .I I" :1. ..g
0 & 1- s
OZ Z
4 a
a o It w a
a c"
0 C1 wAz 3"4z"0 0
a n ae " a a
I.- Iaz aI _a ac LA. a :a -1 9

3j -= . N0 .4 I. ILI i -& I o
Ma ao I 1 00 Wi 0 40-,w0 a.
* j I . - a
3v -. a-
aCD
I **4 -*---aa : zc a : t 20"
- - - I n.
- .. . .
UJ U,3U
W (01 a 3i
~0
:--------- r a~ W
wP
z 5-279
p C6-I
MW j.
... .- 2.
;. . d . * '
6 h 6 64 Owl46 6
be SL. 6 WII 99L
6 IL W 66 CL 6
C a '- a6 *
dcZ 6 0 6 6 E I W
I.-I *j z 6 6 I *
q6 :: I6- i6 z 66
0 aX ~ $.-- IA I
* 4 6 6 Lai 4c 16 I6 W
* 04 6 a *66 4 .4 I64
*~Ui I -4 > 6 Q
6 WELA I 66 W I
q666 >
w 6 -
L16 I I WI 6
I- .i Z"~ 6 z I .
46A u1'CO 66
1 Wn
61 I. 9I IL a.,6
I 4z 4 1
Owl z I E )
6c 6M4c 4w
K~ .W 00aU S-
6~ em 66 too6-
1 04 wo II
6 66 I W I P6 w
I ~~ ~ Z 46166-* *
6ZP.z-6 ,o
*I 4 .
(A6 I Z a @
* ~~IZ ~ ~ m. ~ mu6ij L
*6w
- >::
zmu
6 6
I ~ fl*Lfl6 ~ Lfl a NI-280

U II 8I
I L
I ~~~4 0
~II~M Was8
CI I d
-wow
i -u I- s
0 0 I i- I
*Il. I g Wl1 9 W lbWI-
0
I. WW U. 0U
I.-05I
L6. C4V
oa~Z
4uj
.W
40e o -1G3 06I
dow
*I
0
1
I IA
I- Z
~:
i;
Iý w4 I
~$ " " ZZ ~ In:
I
440'I Z $- .: IV
I
*
I n IL so I IA
K
IA I
I
.1
1.
c
C(
I:;
1.
J2
1-jaIc
0
I I 0' IN ta. nI P
S-9
691 -1CA I~I I- i n
CI
0 -t
00 '
4 IL. Ij
8-0.
I- )
I II
I I
e
zC' I
0 "111I
09* -W-
I
4w e4 * L**e IW W
2 6
N II 8 ~a.O
8II -~ L.I1I 8 I 8 o f*
m '' W I I ui---. 8 -- 4I
%.
. a' .L
~ : ~:5- 28 14
------- * -- -*--a--a
a-- --- -~-
LOL1
e!I.O
A
oo 0CL
COO: ma
t
6a S A
mw a aL S *
ze m
0 S
Z . - J ~ - a - a.
LA I Iý 6 I 4 66 6I W N
LI.- S
S ~ - ~ 6 ~ 5- - -
I Sa I cc aa 1 a a
a~" maJ I a# m
* a -g* >0 : -,: 6 >
*~ mz ~ ~ S m Z
* I 016 O I 6n ~ I4W (A
*1 Q U 0* ~
Lai I W a (al wi 6 U
In I
00 154
w IL maa wjS a - i-
s cl 4 0------ i i 0 "a qa-----

6- 06 w
(A a h W I
Ca 0 -. a C
mm.j a a 0 aa a m.J&l a 00aA-
hi 6 azo - * C
IZ
~ a *
az
ma
a
m
hAaa.%.
a a aa .>
6 6Z 6 a a a a
m
-.-.-- .--- *--*--*--.5- 28------2 -
I~1 II*6 IWO aa

C4 ar I m I I e4 a 2 . .1 10C
* 00.
0' I= 6
I C4I 0I 6 6
I bI 0. II I".. 6I
I E
A Ia=C aZpI iaI C4m I

--.
W a I -. 6 I~ 3p
jI W I --. I .6I .J
W II 6 a 6 64
UI0.U I Na
4 II I. m I (A 0.U Ip 646 6 .
C.) 6 .6 . I.) *-
I w . I
NL I - I I6
6-~~ I -. II N I I
Ba 6 I WI W I 6 I WI:
I 0 I I.,I66
a
I~
a Ca I~ ~ C III
Of1
06I - I
.J046*
o *s >W I w z I Il I Z > I
I w
CL =Z II c.6 W ve a- z1 a oaa
I CL a 0)6 I j OCL I (6)6 I~

I Io Io cc~) I 6 I
I~ 6 ~ 1 a U.6a
j
a0
-a a--. -a
:- I- ") " Ia
6-a I (26 I 1-I W a4

I I I. Ia-a paa
I6 g I0 t4I
a (fO.Cf ZO 1
61 )0.L6 I r) IA
I L Ia.
a
Ij W 0 6 : a Ica
w oZ I ol -o= I V6 I W I6 z I NrI 4-4 =6 Ic
0~ 0 6 .- & 06 6 .6 I.- CL X
c 06 0 0h~ W 6 B
ac a 4 I rZ 1 06 U 1J (6)
0- 1 I 6L 6- I4I._
a Io I
om a ae aý a A aU I
1 I.-
6 L a a I I £06 w
4 I IV : -. I(
w a a - 6 6w a 6 - 6
aC am 10aa-a
9
U-ul_ U-i
O I 64
~
a 6-
* U)
~6w6
I WI1 (CkL
a
U
- 666
W
W
I > I. > I 6 4 6 > >
I I U.I zL I cc I m 6 I 6L
5-283
5.6 Part Malfunction Data. Table ,.6.1 gives the malfunction data
and frequency of occurrence for each part name based only on the information
which was available when the part failure data was collected. The malfunction. _
data foZ each part name are accumulated over all use environments and
part types for the particular part name.
Not all malfunctions reported are mutually exclusive. For example,

"improper adjustment* and "improperly installed" may overlap. We leave
it to the reader to combine the malfunction data into categories, as
needed, using the information presented.
I.',. )•
,\ %
.. . . . . . - . -. . . '... ...
." . o,.
"Table 5.6.1. PART MALFUNCTION DATA

FREQUENCY
PART NAMEMALFUNCTION
PART MANE -TCCURFINCE
ACCELEROMETER
DEFECTIVE PARTS INSIDE 100.0
ACTUATOR
BEARLINGIN
CABL LSULATION FRAYED
& BRAKE RUSTED 6.7
6.*7
CABLE SLEEVE NEEDS FIXING 6.7

IMPROPER CONFIGURATION SHOULD BE -2 6.7
IMPROPER CONNECTOR INSTALLED 6.7
I R.EQUIRES ADJUSTMENT OF TM 6.7
REQUIRES OVERHAUL 6.7
SAFETY WIRE BRACKET BROKEN 6.7
THERMAL SWITCH FOUND TO BE DEFECTIVE 6.7
"UNOKNOWN 40 .0
ANTENNA
"UNKNOWN 100.0
AXLE
DAMAGED 50.0
UNKNOWN 50.0
AZIMUTH ENCODER
ANTENNA, WON'T MOVE 32.3
CASING ROTATES 3.2
-CRACKED CLASS DISC 3.2
ENCODER MARKING SHOULD BE REMOVED 3.2
INCORRECT ANTENNA ROTATION 3.2
LAMP DESIGN DEFECTIVE 3.2
NO MOVEMENT BETWEEN DWELLS 3.2
OPTICAL ASSEMBLY DEFECTIVE 3.2
RESISTER IS DEFECTIVE 3.2
UWKNOWN 41.9
BATTERY
CONNECTOR PANEL DEFECTIVE 20.0
"CONNECTOR PINS SHORTED 20.0
CONNECTOR SHORTED 20.0
K-i MISWIRED 20.0
SHORTED VR1-3 TO CdASSIS 20.0
5-2 85
. . . ..
••.\K.-.~
.
o
.. % . . ....
..
.................
•. . .•. ..... ............... "..•%,.'.." -.. , -..-. ..- ,
• o°• .- ..•.".'.,...o.'.,.. .. '...... .. ".".". .... ..... .. ,....•..",."..-.".
"~~~~~~~~~~~~~~~~~~~.........-..'.....
..................... *,
Table 5.6.1, continued. PAr.T IUi.FUNCTION DATA
PART NAME 1AALFUNCTION FRQUECY
BEARING
AZIMUTH DRIVE INOPERATIVE 4.2

BLOWER BEARINGS WORN OUT 4.2
BLOWER INOPERATIVE 4.2
DEFECTIVE 4.2
EXCESSIVE PLAY 12.5 .%
LUBRICATION DRIED OUT 4.2

PRINTER INOPERATIVE 4.2
REQUIRES OVERHAUL 12.5
UNKNOWN. 45.8
WLW HAS EXCESSIVE WEIGHT 4.2
BELLOWS
CRACKED 50.0
UNKNOWN 50.0
BELT
BROKEN 10.0
DUE TO EXCESSIVE USE 60.0
WORN OUT 30.0
BLOWERS FANS
BEARINGS WORN OUT 15.8

DEFECTIVE SENSOR 18.4
"DEFECTIVE SWITCH 2.6
EXCESSIVE CURRENT HAS SHORTED 2.6
p EXCESSIVE VIBRATIONS & BEARINGS LOOSE 5.3
HAS EXCESSIVE VIBRATIONS & MOUNT IS
LOOSE 5.3
"IMPROPER INSTALLATION 2.6

MOTOR DAMAGED 2.6
NOISY DUE TO DEFECTIVE BEARINGS 5.3
REVERSE WIRING 2.6
SHORTED 5.3
SWITCH IS LOOSE 2.6
SWITCH NOT PROPERLY INSTALLED 2.6
"SWITCH NOT WORKING 2.6
UNKNOWN 23.7
5-286** .
•. . . . . . . . . . . . . . . . . . . . . .
• . .° . %°. .. .
Table 5.6.1, continued. PART MALFUNCTION DATA
PART NAME 11ALFUNCTIOwT TEGUENCY

BRAKES QccuemcgI
ASSY SCREW TOO TIGHT 2.3
BRAKE CABLE TOO SHORT 2.3
BRAKE DISCS WORN OUT 2.3
BRAKE PAD BUSHINGS SCREW LOOSE 2.3

BRAKES CORRODED 4.7
EXCESSIVE GAP 18.6
IMPROPER ADJUSTMET 7.0.
IMPROPER POSITIONING 4.7 07
IIPPOPERLY INSTALLED 2.3
LOCK SCREW NEEDS REPLACEMENT 2.3
NEEDS ADJUSTMENT 18.6
PARTS ARE WORN OUT 2.3
PARTS BROKEN 2.3
RECRIMP TERMIAL, DIC REQUIRED 2.3
SCRE LOOSE 2.3 ýj
UNKNOWN 20.9
WORN OUT 2.3
BRUSHES
NEEDS ADJUSTMENT 66.7 " "-",

SHORTED 33.3
BUSHINGS..."
UNKNOWN 16.7 *:-
WORN OUT 83.3

CAMERA
DEFECTIVE POWER SUPPLY 40.0

PICTURE BECOMES WEAK & BREAKS UP 20.0
UNKNOWN 40.0
CIRCUIT PROTECTION DEVICE
PAINT STENCILED ON ARRESTER SHORTED 14.3

UNKNOWN 85. 7
CLUTCH
ANTENNA FAILS TO ROTATE 62.5
ANTENNA MOVES SLOWLY 12.5
COMPUTER MASS MEMORY
5-287 ,...,
. . . . .. .. .. .. . . . . . .
S•-'--'-- ' .'-'...- .. "." ."..'." . ." . ."." . .".".- ." ". -.-. . .-..... .. :-.*
-.-...- *..-.- --..-.-. 4'
• .... ~~~. .. .
.......... o....... ... -o ... . °...-. o.. %o
Table 5.6.1, continued. PA__TIA.LFVNCTION DATA
PART NAME MALFUNCTION F QF
BAD REEL HUB ASSEMBLY 1.0

BAD SOLDER JOINTS ON CONNECTOR PINS 1.0
CABLE WIRES SWITCHED 1.0
CAPSTAN MOTOR DEFECTIVE 1.0

CAPSTAN MOTOR JAMMED 1.0
CONNECTOR PINS HAVE BROKEN 1.0
DEFECTIVE CAPACITOR 2.0
DEFECTIVE CCA 1.0
DEFECTIVE CHECK VALVE 1.0
DEFECTIVE CONTROL SYSTEM 1.0
DEFECTIVE DISC 1.0
DEFECTIVE IC 6.9
DEFECTIVE MOTOR ASSEMBLY 1.0
DEFECTIVE POWER SUPPLY 1.0
DEFECTIVE SOLDERING ON IC. 2.0
DEFECTIVE TRANSISTOR 3.0
DEFECTIVE TRANSISTOR IN SERVO ASSEMBLY 1.0
DEFECTIVE VACUUM CHAMBER SENSOR 1.0
FLOPPY DRIVE ONE IS DEFECTIVE 1.0
FUSE BLOWN 1.0
HAS LOW PRESSURE 2.0
MTC DEFECTIVE
PINS 1 & 3 ARE DAMAGED 1.0
1.0"'" -
POWER SUPPL DEFECTIVE 1.0

QI & Q2 IMPROPERLY ORIENTED 1.0
READ/WRITE HEAD IS DEFECTIVE 1.0
LIBBON CABLE BROKEN 1.0
SERVO AMP. DEFECTIVE 1.0
TWO PINS SHORTED ON CONNEC' 1OR 1.0
UNKNOWN 8.9
WON'T ACCEPT CERTAIN PROGRAMS 1.0
WON'T ACCEPT VRS-131 1.0
WON'T LOAD 19.8
WON'T RECORD 4.0
WON'T REWIND 9.9
WRONG FUSE WAS INSTALLED 1.0
COUPLING
UNKNOWN 100.0
CRANK SHAFT
BRACKET BROKEN 40.0
5-288
:-••n•
.. " z"---
-- ^-t '- •-."
-- -". - . . . . ... ... .. . " " ""- . . ... .... . ' . . . .

PART NAME MALFUNCTTO OF
OCCURRENCE
UNKNOWN ou W-ý5
DIAPHRAGMS BURST
UNKNOWN 100.0
DRIVE FOR COMPUTER TAPZS/DISCS

DATA LOST FROM TAPES
DEFECTIVE
2.3
25.6
ql
DEFECTIVE IC 2.3
LOOSE CONNECTION 4.7
UNLNOWN 65.1
DRIVES (GARA) K.•'.'
IMPROPER AWDUSTMENT 30.0 Va
IMPRPER INSTALLATIONS 10.0
NEWDS R1PIACEMENT 10.0
U)MOWN 30.0
WORN OUT 20.0
* DRUM
BAD ROTATION AFTER DROP TEST 4.3

BRACKET BROKEN 4.3
CONNECTOR P•4 DEFECTIVE 4.3
DAMAGED DUE TO OVERSPEED & MANY IMPACTS 4.3
- DRUM HAS EXCESSIVE WEIGHT BEARING DAMAGE 4.3
DRUM OUT OF ALIGNMENT 8.7
DRUM STICKS 30.4
HAS OPEN TACH WINDING 4.3
NOT PROPERLY ALIGNED 4.3
SERVO AMP POUND TO BE DEFECTIVE 4.3 "-'''
UNKNOWN 26.1
DUCT
CRACKED 30.0
IMPROPERLY INSTALLED 30.0
SUNKNOWN 20.0
USED UP. NEEDS REPLACEMENT 40.0
ELECTRIC HEATERS
5-289
VJ,.
oV
Table 5.6.1, continued. PART MIALFUNCTION DATA
PART NAME AkLFUNCTION FREQUENCY OF 'I

OCCURPENCE %
SHORTED 4.0_
UNKNOWN 60.0
ELECTR•OMECHANICAL TIMERS
UNKNOWN 66.7
FILTER
DAMAGED 16.7 p
LEAKING 38.9
LINE FILTER PIN-D OPEN 5.6
NEEDS REPLACEMENT 11.1
UNKNOWN 27.8
FITTINGS
IMPROPER ADJUSTMENT 7.1

LEAKING 28.6
NEEDS CLEANING 7.1
UNKNOWN 57.1 ••
FUSE HOLDER
DAMAGED 100.0
GASKETS & SEALS
BAD INTERNAL SEAL IN GEAR DRIVE 6.7

IMPROPER INSTALLATION 6.7
LEAKING 6.7
POPS UP DURING ANTENNA MOVEMENT 6.7
UNKNOWN 26.7
GEAR
UNKNOWN 60.0
WORN OUT 40.0
GEAR BOX
NEEDS OVERHAUL 33.3

NO OIL IN GEAR BOX 66.7
5-290
Table 5.6.1, continued. PART M."ALFUNCTION DATA

PART NjAME MALFUNCTION FRECUENCY OF
HEAT EXCHANGERS
NOT PROPERLY FABRICATED BY VENDOR 20.0
UNKNOWN 40.0
"HIGH SPEED PRINTERS

AC INPUT SHORTED 0.9
AC INPUT SHORTED UNDER LOAD 0.9
SEARINGS WORN OUT. REQUIRES REPLACEUENT 0.9
BELT IS SLIPPING 0.9
DEFECTIVE IC 10.5
DEFECTIVE POTENTIOMETER 0.9
4 DEFECTIVE TRANSISTOR 3.5
i FEEDING MECHANISM NEEDS REPAIR 0.9
HAMM4ER BLOWER NOISY 0.9
HAMMER DRIVER DAMAGED 0.9 ..*
"HAMIMER HEAD BROKEN 0.9
HAS RIBBON SKEWING PROBLEM 0.9 .v,
LOOSE PULLEY 2.6
MOTOR & ROLLER ARE DEFECTIVE 0.9
MOTOR DFF•CTIVE 1.8
MOTOR JAMMED 0.9
MOTOR NOT WORKING PROPERLY 1. 8
"MOTOR SHORTED 0.9
NEEDS OVERHAUL 0.9
NEEDS WILE EPILACEMENT 0.9
NO POWER AT - 32C 0.9
OVER CURRENT, SHORTED 1.8
PAPER FEED NOT WORKING 1.8
PAPER FEEDING MECHANISM JAMMED 0.9
PAPER SPINDLE ROTATION SLOW 0.9
PAPER SPINDLE TENSION LOW 1.8
PARTS KISSING 3.5
PRINT FINGERS BENT 1.8
PROBLEM WITH TAKE UP 0.9
PULLEY BROKEN 0.9
RESISTOR R-35, IS OPEN 0.9
RIBBON MOTOR INOPERATIVE 0.9
RIBBON WORN OUT 1.8
RIBBON WORN OUT & BAD SWITCH 0.9
ROLLER PRESSURE IS LOW-NEEDS ADJUSTMENT 0.9
ROLLER PRESSURE NEEDS ADJUSTMENT 0.9
*
5-291
°,
.......
-. .......-
.,:........ .......... .:.. .,......
.. ,.-.......,-...... ., . ..... ... ,.:....,:.. ...- : .. .-....-...-.- : . ..- :.: ... : .:.
Table 5.6.1, continued. PART IALFUNCTION DATA
PART NAME MALFUNCTION FREQUENCY OF

,•, ~OCCUW•.ENCE % •
STEPPER MOTOR DEFECTIVE
STEPPER MOTOR IS INOPERATIVE 0.9
TAKE UP REEL LATCH IS INOPERATIVE 1.8
TIMING BELT
TIMING & PULLEY WORN OUT
BELT BROKEN 0.9
1.8
i
TIMING BELT DAMAGED & WORN OUT 0.9--
; ~TIMIG BELT WORN OUT 1.8•.'-
";" TOP ROLLER PRESSURE NEEDS ADJUSTMENT 0.9 "
"TRANSFORMER HAS OPEN LEAD 0.9

UNKNOWN 32.5
HOSES
DEFECTIVE 28.6
HOSE HAS CRACKS DUE TO ANTENNA MOVEMENT 42.9
UNKNOWN 14.3
WORN OUT 14.3
=. -
INSTRUMENTS
SEALING DAMAGED 20.0

UNKNOWN 80.0
"-p JOINT, MICROWAVE ROTARY
DEFECTIVE 40.0
UNKNOWN 60.0
XKEYBOAkD
CABLE(GP554)DEFECTIVE 4.3
IMPROPER CONNECTIONS 4.3
LED DISPLAY DEFECTIVE 4.3
LOCKED UP 13.0
U-32 DEFECTIVE 4.3
UNKNOWN 69.6
*
LOW SPEED PRINTERS

I DEFECTIVE CA.PACITOR 8.3
DEFECTIVE IC 16.7
DEFECTIVE RELAY 8.3 h-"
DEFECTIVE SWITCH 8.3
FUSE BLOWN OUT 33.3
TEAR BAR BROKEN 8.3
5-292
.. . . ..-... . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
*.... .......... . . . . . . .
I
"Table 5.6.1, continued. PART 1,ALVNCTION DATA
PART N FREQUENCY OF
TERMINAL WILL NOT LOBE FEED8.'1

% TRANSFORIMER SHORTED 8.3
METAL TUBING UNNOWN 100.0
MOTOR, ELECTRIC
HAS OPEN TACH-WINDING 7.7
PAPER TAKE UP UNEVEN 7.7
SWITCH DEFECTIVE
UNKNOWN 15.4
61.5 !
WLU MAP DRUM OSCILL-ATES. DEMAGNETIZED 7.7
POWER CIRCUIT BRFAKER
COONhECTOR DEFECTIVE 12.5

IMPROPER CONNECTIONS 25 .0
UNKNOWN 62.5
POWER SWITCH GEAR
UNKNOWN 100.0
* PULLEY
DUE TO EXCESSIVE USE 33.3

PRINTER INOPERATIVE 22.2 '"
WORN OUT 44.4
p.... PUMP /
CIRCUIT BREAKER TPS 7.1 "

DEFECTIVE CHECK VALVE 7.1
DEFECTIVE COMPONENTS 21.4
HAS HIGH ION CURRENT 7.1
LOW PRESSURE 14.3
OUTPUT PRESSURE VERY LOW 7.1
UNKNOWN 28.6
RESILIENT MOUNT
IMPROPER MOUNT 40.0
5-293
p.'o
j..
.. . ..... ... . -_. p . ..... . ....
. . _._'_-.
-.- ..- ._._o•. ..-. " '_'," " ' ,
PART NAME WREQUENCY OF

OCCURPENCT-•
SCREWS MISSING 20.0
UNKNOWN 40.0
RETAINING RING

UNKNOWN 33.3
SENSOR/TRANSDUCER/TRANSMITTER
NO TAPE LOADING 30.0

PRINTER INOPERATIVE 30.0
RELAY DEFECTIVE 20.0
VACUUM COLUMN NOISY 20.0
SHAFT
HINGE REDESIGN REQUIRED 11.1

RUST UNDER GUSSET 11.1
UNIT HAD GREASE WHICH HAS FROZEN 11.1
UNKNOWN 66.7
SHOCK ABSORBERS
IMPROPER MOUNT 50.0

REQUIRES REFURBISHING 25.0
UNKNOWN 25.0
SOLENOIDS
UNKNOWN 100.0
SWITCH1
BRAKE INTERLOCK SYSTEM STUCK 1.6

CONNECTION LOOSE 1.6
CONNECTOR BASE PULLED OUT 1.6
DAMAGED 4.8
DEFECTIVE WIRING 4.8
IMPROPER CONNECTION 3.2
IMPROPER INSTALLATION 6.5
IMPROPER WIRING 9.7
IMPROPERLY BONDED 1.6
LOOSE DUE TO EXCESSIVE VIBRATIONS 1.6
5-294
S,.,... ,.,., . . ~~~~~~~~.........•...

...... '...,..'.../.. ".......... ,.. _...... . ................ ,._..
Table 5.6.1, continued. PART MLFUNCTION DATA
MAALFUNCTION nELNY OF
PART NAME
"LOOSE INSTALLATION
LOW PRESSURE 1.6
NW-!' REPOSITIONING 4.8
NEEDS REWIRING 1.6
OPE 1.6
Q-3 LEADS SHORTED 1.6
3.EVUSD LEADS 1.6
SHORTED 3.2
SWITCH Is •ENT 1.6
UNKNOWN 43.5
TELESCOPE
DAMAGED 50.0
UNKNOWN 50.0 ,
TRACK BALL
CONSOLE CANNOT ENTER MODE BITE 3.2
67.7
DEFECTIVE COMPONENT
DEFECTIVE DIODE 3.2
DEFECTIVE IC 3.2
DEFECTIVE LAMP 9.72to
IMPROPER SOLDERING INSIDE 3.2
LOOSE CONNECTOR 3.2
UNKNOWN 6.5
VALVES
v %
CRACKS ON BODY 12.5

DUE TO LOW PRESSURE 12.5
IMPROPER SEALING 12.5
SEAL WORN OUT 37.5
UNKNOWN 25.0
. .
.S us o •..
/#° %" .z
"5-295(5-296)BLANK
. . ..... ........
**. •*o.4.4...... °.....--o. e "
• ,. .. ° . " • "°" • ". .""
. . . . . . .• ..
"e" * %
'o°" ° " '.•, °'•°*' °
°o% . • . . . ..
- . .-..-
.
.
°.
. *. . ...-. . . . .
...
.'." .. .'..'* '.,.*• . .- , .- 4 , .- 4
-. , -...-.. ,
6.0 RELIABILITY DEMONSTRATION TESTS
6.1 Introduction. The purpose of a reliability demonstration test is to

the
decide if additional reliability design effort is necessary to achieve
specified reliability for the nonelectronic Item when it is operated In the
The reliability specification (see section 3.0) identifies
field environment.
the parameter(s) and the values to be nominally and minimally acceptable.
which lead to one
Reliability demonstrations are statistical hypothesis tests
of two mutually exclusive decisions;
(a) The reliability parameter(s) of the component is (are) acceptable %
and no additional design effort is required under the contract;
(b) The realiability parameter(s) of the component is (are)

unacceptable and additional design effort is required.
The demonstration is designed to have a high probability that the decision %%.-
reached is correct. When the decision is (a), the consumer runs the risk that
the decision is incorrect, i.e. that (b) is true but there were an unusually
low number of failures during the test. The probability of this type of
incorrect decision is called the consumer's risk (s). I4-
Similarly, the probability that the decision is (b) when (a) is true
(i.e. the component has an unusually large number of failures during the test)
is called the producer's risk (a).
It is important that the demonstration simulate the field environment or

that there be a known relationship between the field environment and the test
environment. For example, if the component actuation rate in the field is low
and the effect of actuation rate is known, it would save test time to raise
the actuation rate and lower the acceptable reliability values accordingly.
6.1.1 Statistical Characteristics of a Reliability Demonstration Test.

There are six essential characteristics of a reliability demonstration test.
(1) The reliability parameter(s) in the specification. If the

distribution of the number of failures in the period of time (0,T] is
parameters will be
available in a mathematical expression then the reliability
related to the distribution parameters.
(2) The acceptable values for the reliability parameter(s). For

example, the upper test MTBF (00) in MIL-STD-781C is the smallest desired
value of HrBF.
(3) The unacceptable values for the reliability parameter(s). For

example, the lower test XTBF (e 1 ) in MIL-STD-781C is the largest
unacceptable value for KTBF.
(4) The producer's risk, a.
(5) The consumer's risk, 0.
6-1
•*:.- *.* :
t.O RELIABILITY DEMONSTRTION TESTS
6.1 Introduction. The purpose of a reliability demonstration test is to

decide if additional reliability design effort is necessary to achieve the
specified reliability for the nonelectronic item when it is operated in the
field environment. The reliability specification (see section 3.0) identifies
the parameter(s) and the values to be nominally and minimally acceptable.
Reliability demonstrations are statistical hypothesis tests which lead to one
of two mutually exclusive decisions:
(a) The reliability parameter(s) of the component is (are) acceptable

and no additional design effort is required under the contract;
(b) The realiability parameter(s) of the component is (are)

unacceptable and additional design effort is required.
The demonstration is designed to have a high probability that the decision

reached is correct. When the decision is (a), the consumer runs the risk that
the decision is incorrect, i.e. that (b) is true but there were an unusually
low number of failures during the test. The probability of this type of
ino.orrect decision is called the consumer's risk ($).
Similarly, the probability that the decision is (b) when (a) is true
(i.e. the component has an unusually large number of failures during the test)
is called the producer's risk (a).
It is important that the demonstration simulate the field environment or

that there be a known relationship between the field environment and the test
environment. For example, if the component actuation rate in the field is low
and the effect of actuation rate is known, it would save test time to raise
the actuation rate and lower the acceptable reliability values accordingly.
6.1.1 Statistical Characteristics of a Reliability Demonstration Test.

There are six essential characteristics of a reliability demonstration test.
(1) The reliability parameter(s) in the specification. If the

distribution of the number of failures in the period of time 10,T] is
available in a mathematical expression then the reliability parameters will be
related to the distribution parameters.
(2) The acceptable values for the reliability parameter(s). For

example, the upper test MTBF (G0) in MIL-STD-781C is the smallest desired
value of MTBF.
(3) The unacceptable values for the reliability parameter(s). For

example, Lhe lower test MTBF (01) in MIL-STD-781C is the largest
value for MTBF.
(.t
(4) The producer's risk, a.
"consumer's risk, 3.
6- 1
s) Tthe sampling pitet which defines how and what parameter* will be
obuerved, and the criterion for endinE rhe demonsitration test and reaching a
decisiotn.
In what follows, R will denote a generic reliability parameter (e.g.

MTBF, probability of survival ot a prespecified time period, etc.) larger
values of which are preferred. The smallest acceptable value of R in (2)
above is denoted by RU, and the largest unacceptable value of R is denoted
by R1 . R0 must be strictly greater than R1 . The values of R which lie
between R, and R. are called indifference values. The ratio Ro/R1 is
called the discrimination ratio.
The statistical characteristics of a demonstration are summarized in the

operating characteristic (OC) relationship between R and the probability of
passing the test when R is the "true" reliability, P(R). Characteristics (4)
and (5) give two points in the OC relationship, namely P(Ro)-(l-U), and
6.1.2 Cost of Demonstration. The cost of demonstration is determined by

the number of samples collected and the calendar time the test facility is
occupied. Cost efficient demonstrations require the smallest number of
samples and the least calendar time to meet the risk objectives of the
demonstration.
If the reliability parameters are related to the parameters of a life

distribution, then the most cost efficient demonstration is to measure the
time of failure (a variable) for each sample. The demonstration is then
called a variables test and the sample of failure times provides the maximum
information on the reliability of the design of the component. If the
reliability parameter is the fraction of components in a lot which will live
beyond some time T, the demonstration will attribute success or failure to
each sample according to whether or not the component is operational at time T
(attributes testing).
The rule for terminating the demonstration also affects the cost. Given
the values for the characteristics in 6.1.1 (l)-(5), it is possible to design
a fixed sample size (or fixed time) or sequential demonstration test. In a
fixed sample dL-onstration test termination occurs when (i) all the sample
values are observed or (ii) when enough failures have been observed to decide
the reliability is unacceptable. Similarly, a fixed time demonstration
terminates when (i) the fixed time limit is reached or (ii) when enough
failures are observed to decide the reliability is unacceptable. During a
sequential demonstration, a sequence of decision points for both acceptable
and unacceptable reliability are formulated and a decision is reached the
tirst time one of these decision points is reached. On the ave.:age a
sequential demonstraticn will require a smaller sample than a fixed sample
t I- I r thor Samine RO, RI, u, b. SequentLia demonstrations are
pw;iHjhJe for both variables and attributes testing.
. . :;tj) Sie Ijml._tditmions. In general sample size (N) (or average,

i ! ;t4;•quint ia 1. d(1cm riir;trat io0 tests) is the dependent vartiable ')t
a reliabilitj demonstration (i.e. the demonstration characteristics in 6.1.1

(1)-(5) are allowed to determine N). N increases as either the discrimination
ratio, a, or 6 decrease. There is a maximum value for N (NMAX) in all
practical situations. If Lhe N required by the characteristics of 6.1.1
(l)-(5) exceeds NMAX, One of the parameters3 (usually a) must. be changed
(&O, a, or B can be increased, or R, can be decreased).
6.1.4 Summary. The remainder of this section presents step-by-step

instructions on the use of various types of reliability demonstration test
plans. The section is arranged to present first test plans based on
attributes, followed by variables test plans.
6.2 Attributes Demonstration Tests
6.2.1 Attributes Plans for Small Lots
1. When to use
When testing parts from a small lot where the accept/reject decision
for the lot is based on attributes, the hypergeometric distribution is
applicable. Attributes tests should be used when the parameter of
interest is the fraction of components in a lot which possess a certain
reliability attribute.
The example demonstrating the method is based on a small lot and

small sample size. This situation frequently characterizes the
demonstration test problem associated with nonelectronic parts. The
sample size limits the discriminatory power of the demonstration test
plan but frequently cost and time constraints force us into larger than
desired risks.
The attribute of intezest may be that a part survives at least t

hours. A "success" for a component tested would be that it survives t
hours. The parameter to be evaluated then is the fraction of the parts
in the lot whose lives would exceed t hours.The estimation of the
parameter would be based on a fixed sample size and testing without
replaceeint. The selection of the criteria for success (t hours) can be
derived from a requirement (such as mission length, for example). If the
lot size is 30 or more, then the Poisson approximation may be used to
make the calculation simpler. (See Section 6.2.3).
3. Method Example
a. Define criterion for a. A part that lasts 1CO or more

success/failure, i.e. hours on a given life test is
define the attribute, considered a success. Parts
failing before 100 hours are
considered failures.
6-3
3. Method Example
b. Define acceptable lot b. Lots in which (l-po) - 90% of

quality level (l-po). of the parts will survive 100
hours are to be accepted by this
demonstration test plan with high
probability.
c. Specify producer's c. Let a - .2. This decision is

risk (a) (i.e, the an engineering one based on
probability that the consequences of allowing
acceptable lots be defective lots to be accepted
rejected). and based on the time and dollar
constraints associated with
inspecting the lot.
d. Define unacceptable d. Lots in which only 1-p

quality level (l-pl). - 20% of the parts will survive
100 hours will be accepted by the
demonstrations test plan with low
probability.
e. Specify the consumer's e. Let 8 - .022 (Taken for

risk (8) (i.e., the convenience in
probability that calculations).
unacceptable quality lots
will pass the
demonstration test).
f. Now that Ct, 8, f. Given: lot size N-lO

l-po, and 1-pl
have been specified l-Po = .9
the following steps l-pa W .2
describe the a - .2
calculations required 8 Ta .022
to determine the
sample size and accept/
reject criteria which
will satisfy the
stated risks.
g. The process consists of a g. The calculations are as

trial and error solution follows: If N - 10 and it
of the hypergeometric is assumed that samples
equation using N, 1-po, are taken from a lot with
1-pl and various sample 1-po - .9 then that lot
sizes until the condi- contains 9 good parts and I
tions of a and a are defective part. 4.s the
kN
3. Method Exampl e
g• met. The equation used is first step in the trial and

error procedure assume a
(N-r) sample size of 2-. The
Pr(x) x n-x possible outcomes are
either 0, 1 or 2 good
parts. The probability of
each outcome using the
hypergeometric formula is
x-max(O,n-N+r), 1, 2
min(n,r)
Pr(2) ( - .8
where x *number of 767 .
successes ,..
in sample Pr(l) a .2
r -number of Pr(O) - 0
successes in
lot
N - lot size
n - sample size
- r; The same calculations for
x x;(r--x) 1-pl .2 results in
Pr(2) - .022
Pr(l) - .356
Pr(O) - .622
h. Find the number of h. From these 2 sets of .,-

successes which satisfies results it can be seen that
C and 8 in the if a sample size of 2 is
calculations involving specified, then a and
1-Po and l-pl. will
w be satisfied if the -,'',
decision rule is made that if 2
successes are observed in the
sample the lot is accepted and for
all other outcomes the lot is
rejected.
If 1-po - .9, then Pr(2) - .8, '*'

therefore 1-.8 a .2- a If 1-p1
- .2, then Pr(2) - .022 = 8.
NOTE: A different sample size can

be traded off1 against different -
as, 6, l-po, P*l,
6-5
..........................................
3. Method Example
i. The demonstration test i. The test procedure is as

is then specified. follows:
1. Test a random sample

of 2 parts from a lot
of 10 parts for 100
hours.
2. If both parts survive

100 hours, accept the
lot.
3. If only 0 or 1 parts
survive 100 hours
reject the lot.
4. For Further Information.
There are "Tables of the Hypergeometric Distribution" by G. J.

Lieberman and D. B. Owen, Stanford University Press, Stanford,
California, 1961 to perform the mathematical calculations of Step g.
Also if N becomes large (say 30 or more) then the binomial or the Poisson
distribution can be used as an approximation for the hypergeometric
distribution.
6.2.2 Attributes Plans for Large Lots (Binomial)
1. When to Use
When testing parts from a large lot where the accept/reject decision
for the lot is based on attributes, the binomial distribution is
applicable. Strictly speaking, all reliability attributes testing should
follow the hypergeometric distribution as long as individual parts are
placed on test and tested to failure without replacement. However, when
the lot size is large, the binomial distribution is a good approximation
for the hypergeometric and therefore the example presented in this
section covers the use of the binomial. Attributes test should be used
when the parameter of interest is the fraction of components in a lot
which possess a certain reliability attribute.
The attribute of interest may be that a part survives for at least t

hours. A "success" for a component tested would be that it survives t
hours. The parameter to be evaluated then is the fraction of the parts
in the lot that would survive t hours. The estimation of the parameter
would be based on a fixed sample size and testing without replacement.
The selection of the criteria for success (t hours) can be derived from a
requirement (such as a mission length, for example).
6-6
3. Method Example
a. Define criterion for a. A part that lasts 100 or
success/failure, i.e. more hours on a given life
define the attribute, test is considered a success.
Parts failing before 100 hours are
b. Define acceptable lot b. Lots in which 1-po - .9

quality level (1-po). (i.e., the life of 90% of the
parts will exceedl0O hours) are
to be accepted by this
probability.
c. Specify producer's risk c. let a = .01.
(a) (i.e., the proba-
bility that acceptable
lots will be rejected).
d. Define unacceptable lot d. Lots with only a true

quality level (l-pl). fraction of acceptable parts
l-p 1 - .5 are to be accepted by
this demonstration test plan with
low probability.
e. Specify consumer's e. let - .17 (selected for

risk (B). (i.e., ease of calculation).
the probability that
lots of unacceptable
quality level
will be accepted).
f. Now that a, B, f. Given: lot size N - large
i-po, and 1-pl have say > 30
been specified, the
following steps describe i-po = .9
the calculations required l-pl = .5
to determine the sample c = .01
size and accept/reject B = .17
criteria which will
satisfy the stated risks.
g. The process now consists g. Assume a random sample of

of a trial and error size n = 10 is taken from a
solution of the binomial lot whose true fraction of
equation using l-Po, good parts is .9. Solve
l-Pl and various sample the binomial equation for
sizes until at a given the total number of
decision point,the condi- consecutive outcomes whose
tions of a and 6 are summed probabilities equal
6-7
3. Method Exampl
g. satisfied. The binomial a starting at 0
equation Is: successes. The
calculations for this
Pr(x) ( l.p X decision point are:
where xn x- saple 41 Pr(l0) -m10)(.9)l0(.l)0

1 - .3486
observed suc-
cesses in sample Pr(9) - .387
A-p lot fraction Pr(8) zw .1935
defective Pr(7) - .0574
Pr( 7 or more) = .9865
Then
Pr(6 or less) - 1-Pr (7 or

morp)
1.0 - .9865
Z .01 (which
satis- "'
1a.
the a (•
risk).
Perform the same type of
calculations assuming the true
fraction defective is .5. In this
instance sum the probabilities
starting at 10 successes until
succeeding consecutive
probabilities sum to the value of
0. This yields the following
results:,-.
Pr(l0) .10!.510 (.5) - .001 .

Pr(9) - .010.
Pr(8) - .044
Pr(7) - .117
Pr( 7 or more) z.17 (which
satisfies
the '
risk)
h. The demonstration test h. The test procedure is as

is then specified, follows:
1. Test a random sample of ':•.

10 parts for 100 hours.
6-8
. -.. -.. . . . . . . . .. . .. -,.• .- -

3. MethodExml
h. 2. If 7 or more parts
survive 100 hours,
accept the lot.
3. If 6 or less successes
are observed, reject
the lot.
There are several published tables for use in determining binomial

probabilities in the event that the sample size makes calculations too
lengthy. One of these is Tables of the Binomial Probability
Distribution, National Bureau of Standards, Applied nathematics Series 6,
Washington, D.C., 1950. It gives individual terms and the distribution
function for p n .01 to p a .50 in graduations of .01 and n - 2 to n ...
- 49 in graduations of 1. ,. %
6.2.3 Attributes Demonstration Test Plans for Large Lots (The Poisson
Approximation Method)
1. When to Use
In attributes demonstration test plans if the lot size gets much

above .100 the calculations required to generate a demonstration test plan
become very time consuming. The Poisson distribution can be used as an
approximation of both the hypergeometric and the binomial distributions
if the lot size is large and if the fraction defective in the lot is
small. This method can therefore be used in lieu of the previous two
methods in many cases.
"If the lot size is large and the fraction defective is small, this
method is applicable. Its use is initiated by specifying a desired
producer's risk, consumer's risk, acceptable lot fraction defective and %
unacceptable lot fraction defective. As before, it is also necessary to
specify the characteristics that constitute a defective part since this
is an attributes type test.
3. Method Example
a. Define criterion for a. A part that lasts 100 or more

success/failure. hours on a given life test is
considered a success. Parts
failing before 100 hours are
6-9
°.,..l--.. .
-••• . . . . -. -. ~ ~ .~~~~.
.,*.*.-. ° ... , ..-,. . ,*. . • . .' - -.- • . -- , . ,
3. Method Example
b. Define acceptable lot b. Lots in which 1-po a .9

quality level ( 1 -pc,. (the life of 90% of the parts in
the lot will exceed 100 hours) are
to be accepted by this
probability.
c. Specify the producer's c. Select a = .05.

risk (a) (i.e., the
probability that accepta-
ble lots will be rejected).
d. Define unacceptable lot d. Lots with only a true fraction

quality level ( 1 -Pl). of acceptable parts 1-p 1 a
.75 are to be accepted by
this demonstration test plan
vith low probability.
e. Specify the Consumer's

risk 0 (i.e. the.
e. Select 0 " .02.
probability that lots of
unacceptable quality
level will be accepted by *'
this plan).
f. Now that a, 8, f. Given: lot size N-1000

1-po, and I-p1 have l-PO - .90
been specified, the 1-p, - .75
accept/reject criteria a - .05
are determined by the $ - .02 K.
following formulas:
c (npO) exp(-npo) 6'
c (np 1 ) Xexp(-np 1 )
°T
x;..
g. The solution now consists g. Assume n - (sample size)

of trying various values - 100.
of n in the above formulas Then,
until they are approx- npo - 100 (.10) - 10
Imatel, satisfied. npl a 100 (.25) - 25.
Using a digital computer to
*" 6-10
S~~~~~~~~~~~~~.................... ..... ... ,....-.....,..,..-.-......-..,,- ..-..-......-...-..-... ?,......-•. ..... N+...'

3. Method Example
compute the formulas in (f) above

leads to cl 5 , and
* .049
_ * .022.
The decision criterion is now
specified as c-15 or less
failures.
h. The demonstration is then The demonstration test procedure

fully specified. is as follows:
I) Take a random sample of 100 parts
from the lot of size 1000 and
test each part for 100 hours.
2) If 15 or less fail to survive

100 hours, accept the lot.
If more than 15 parts fail to
survive 100 hours, reject
the lot.
4. For additional examples using this method, refer to E.B. Grant,

Statistical Quality Control, McGraw Hill, 1964.
6.2.4 Attributes Sampling Using MIL-STD-105D V
P 1. When to Use
When the accept/reject criteria for a part is based on attributes

decisions MIL-STD-105D is a useful tool. These sampling plans are keyed
to fixed AQL's (Acceptable Quality Level) and are expressed in lot size,
sample size, AQL and acceptance number. Plans are available for single
sampling, double sampling and multiple sampling. The decision as to
which type to use is based on a trade-off between the average amount of
inspection, the administrative cost and the information yielded regarding
lot quality. For example, single sampling usually results in the
greatest amount of inspection, but this can be offset by the fact that it
requires
it gives less training
a greater of personnel,
amount and record
of information keeping
regarding is simpler,
the lot being and
it
sampled. The main difference between MIL-STD-105D plans and the previous
plans is that the unacceptable quality level need not be specified.
2. Conditions for Use:
The user of a MIL-STD-105D sampling plan must have items a and b

below. MIL-STD-105D will determine items c, d, and e below, for a given
type of sampling type (i.e. single, double, multiple, etc.):
a. Lot Size '.'.

b. Acceptable Quality Level
c. Sample Size
6-11 *
7.--f
d. Acceptance Number
e. Criteria for Acceptance or Rejection.
The specification of the AQL is an engineering decision based on the

fraction defective that a user of parts considers acceptable. Lots with
this percent defective will be accepted a high fraction of the time.
Operating characteristic curves are supplied vith each sampling plan and
these can be used to evaluate the protection afforded by the plan for
various quality levels.
MIL-STD-105D also contains plans for normal, tightened and reduced
inspection plans vhich can be invoked if the fraction defective of lots
seems to be varying or trending.
3. Metho___d -Example•"
a Determine lot size and a Given a lot containing 100

pacify AQL and type of parts and an AQL Is specified
sampling. at 6.5% with single sampling
specified.
b. Enter the table with lot b. From Table I (Sample Size Code
size and select the Letters) on page 9,
sample size code letter. MIL-STD-105D, find the sample
size code letter for a lot of
size 100. For this example
and for normal sampling, the
specified code number to P
(General inspection level II
is the default). r
c. Enter the single sampling c. Enter Table I1-A (Single

plan table for normal Sampling Plans for Normal
inspection with the code Inspection) page 10 with code
number from Step b. letter F. Under the column 10
titled Sample Size, find the
number 20 in the same row as
the letter F. This is the
number of parts to be randomly
selected and inspected.
d. Enter the same table in d. Find the column in

the proper column for the Table II-A page 10
specified AQL. corresponding to an AQL,
of 6.5%.
6-12 -
WM •
3. Method Example
e. Proceed horizontally e. At the intersection of row F

along the Sample Size and column 6.5%, the
Code Number row until it acceptance nu~mber is 3 ana
intersects with the AQL the rejection number is 4.
column to obtain the
acceptance number.
f. The Single Sampling Plan f. For the single sampling plan

from MIL-STD-105D is to N 100, AQL - 6.5%, select a
1
select a random sample of random sample of size n - 20
size n from a lot of size and inspect it for attributes
N, inspect it and accept criteria. If 3 or less
the lot if the number of f. defectives are found in the
defectives in the lot is sample accept the lot. If 4
equal to or less than the or more defectives are found
Acceptance Number. If the in the sample reject the lot.
observed number of defects
is equal to or greater
than the rejection number,
the lot is rejected.
In addition to the example discussed above, MIL-STD-105D contains

other plans for any lot size and for selected AQL's from .01 to 1000%
(AQL's over 10% are defects per hundred units, rather than percent of
defective units). MIL-STD-105D also presents operating characteristic
curves for each sampling plan.
6.2.5 Sequential Binomial Test Plans
1. When to Use
When the accept/reject criterion for the parts on test is based on

attributes, and when the exact test time available and sample size to be
used are not known or specified then this type of test plan is useful.
The test procedure consists of testing parts one at a time and
classifying the tested parts as good or defective. After each part is
tested, calculations are made based on the test data generated to that
point and the decision is made either that the test has been passed,
failed, or that another observation should.be made. A sequential test
will result in a shorter average number of parts tested than either
failure truncated or time truncated tests when the lot tested has a
fraction defective at or close to pO or Pl.
,-. The parts subjected to test will be classified as either good or

delefective. In other words, testing will be by attributes.
6-13
b. The acceptable fraction defective in the lot Pot the

unacceptable fraction defective Pl, the producer's risk a,
and consumer's risk 8 must be specified.
c. The test procedure will be to test one part at a time. After the
part fails or its test time is sufficient to classify it as a
success, the decision to accept, reject or continue testing the
lot will be made.
d. The part lot size must be large (greater than 100).
3. Method Example
a. Specify PO, Pl, a, a. Given a lot of parts to be

8 tested by attributes. Lots
having only PO - .04
fraction defective parts are
to be accepted by the
demonstration test plan 95%
of the time (i.e., a "
.05). Lots having P, "
to be accepted 10% of the
time (i.e., 8 - .10).

b. Calculate decision points b. The decision points are:
with the following formula
8and 4L
16 a .05
.L
1-a
.10
f a- .105a.is
c. As each part is tested, c. In this eaample, if

classify it as a failure (.10/.04)1 (.90/.96)8 is:
or a success and evaluate 1) > 18, reject the lot.
the expression: 2) < .105, accept the lot;
3) between .105 and 18, the
(p /Po)f( (l-P 1 )/(1_-Po)) test is continued.
where f w total number of

facilures
s - total number of
successes.
If at some point, this
expression exceeds (1-0)/
a reject the lot. If at >
some point, this expression ,..
is less than $l(1-a)

accept the lot. Continue
6-14
*.-A.Oo..
* * * a I. *I* t *~***.*. . . .--- -. ,

3. Method Example
c. sampling as long as neither

of these conditions arises.
d. The operating character- d. The five points on the OC

istic curve (i.e. the pro- curve are as follows: '
bability of acceptance as
a function true fraction Prob. of Accept
defective) can be roughlyV.0 r.O
sketched from the follow- .04 .95
1.00 0.00
.063 .56
Probability of The last point above is

? Acceptance calculated as follows:
2
0 1 ln(.94)/(ln(.94)-ln( .5))
PO 1- - .063;
1 0 ln(18)/(ln(18)-ln(.105))
0.56
where:
"
* ln((l-p )/(1-Po))
1
SP ln((l-p )fl-p
0 ))-ln(p o)p-)
)-cin(-
l-n)/(
Pa - ln((l-8)/Q)-ln(B/(l-a))
6.3 Variables Demonstration Tests
6.3.1 Introduction. Reliability demonstration tests conducted in

industrial applications are virtually always constrained by time. It is
almost never the case that a demonstration test is carried out by placing n
items on test, and waiting until all (in the complete sample case) or r t n
(in the failure censored case) items have failed and recording their
respective lifetimes. In practice, such sampling schemes are not used because
the time necessary to complete the test is random, making it impossible for
management to allocate the correct amount of time and resources to conduct the
test. Instead, a time truncated test is appropriate (and often easier to
administer) because an upper bound on the time to complete the test is known
4 .
in advance of testing. Such tests were developed in MIL-STD-781C for the
exponential distribution, and have been used almost exclusively in industry
for electronic equipment.
6-15
S......~~~..................... • =•..o........ ._=._:-...'.=.•'"-.- .-.. -. ........ ••-_.

Another aspect of sampling for reliability demonstration tests is

replacement versus nonreplacement tests. That is, if n items are placed on
life test initially, should failed items be replaced (or repaired to new
U .
working order) or not. Just as in the failure truncated case discussed above,
the replacement life test presents a problem with respect to planning, since
the ultimate number of items needed to complete the test is random and thus
impossible to plan exactly in advance. Moreover, except in the exponential
case, replacement tests are mathematically extremely difficult to levelop in ______
the time truncated case. Replacement tests are appropriate, however, when the
item under test is a complete system, and "replacement" signifies
"repair/restore to new working condition." Indeed, the HIL-STD-781C time
truncated tests are replacement tests. Whenever the item to be tested is a
complete, complex system in which the predominant failure modes are due to
electronic (or other constant failure rate) equipments, then the MIL-STD-781C ......
time truncated tests can be used. However, if the system is primarily
composed of nonelectronic parts having increasing failure rates and the
predominant failure modes are associated with these parts, then a replacement
(by repair to new working order) test is out of the question, since in order
to restore the system to new working order at each failure, each wear-out
related part would have to be replaced with a new part whether failed or not.
In summary, when the exponential distribution Is assumed, the time
truncated tests presented in MIL-STD-781C are recommended in the replacement
case. In the nonreplacement case, MIL-HDBK-108 (H 108) contains time
truncated test plans for the exponential case. In view of the applicability
of the exponential distribution to most nonelectronic parts in section 2 of
this notebook, these documents should be adequate most of the time. When the
exponential distribution is not justified, then a time truncated,
nonreplacement demonstration test is recommended. Although they possess
interesting statistical properties and are mathematically tractable, failure
truncated demonstration tests (which include the complete sample case) are not
desirable when time must be limited, and are not recommended here. For
information concerning statistical inference for various life distributions
under failure truncated sampling, refer to section 3 of this notebook, -_._.... _.
MIL-HDBK-108, or to Mann, et.al. (1974) or to Lawless (1982).
6.3.2 Time Truncated Demonstration Test Plans
6.3.2.1 Nonparametric Reliability Demonstration Test
1. When to use
This type of test is applicable to any situation in which reliability
(i.e. probability of survival for a preselected time period), median --.
life, or any quantile of the underlying life distribution is specified. " '
This test procedure is valid no matter what form the underlying life
distribution assumes (i.e. exponential, Weibull, Gamma, Lognormal, etc.)
as long as it is of the continuous type.
2. Conditions for use
The user of this type of test plan must specify (or select from the
table of test plans) the producer's risk (a), the consumer's risk
6-16
- *. - -
-.. -.---. ,-..,

(t), the acceptable reliability (RO), the unacceptable reliability

(RI), and the time (T) corresponding to the reliability values (i.e.
R0 is the acceptable probability of surviving the time T. while R1 is
the unacceptable probability of surviving the time T).
"• ~equipments
The teston entails
test forplacing a of
T units predetermined fixed number
time, and recording the of partsitems
number or
that fail before time T. Failed items are not replaced. The
demonstration test is passed (i.e. items are judged to have the
acceptable reliability if c orif less
c+l items
or morefail before
the demonstration test F0)
is failed items fail time T. and
before time
T. The value of c is predetermined by the user's specifications.
"3. Method Example
a. Specify RO, R , T, n, a. An axial blower must survive

"oor select them from T-100 hours of continuous use
the table of test plans, with high probability. The
table 6.3.2.1. acceptable reliability is
O- .95
and unacceptable reliability

is . . -' -'
R .85.
A producer's risk of no more

than .10 and a consumer's
.'4 risk of no more than .10 are
"acceptable.
b. Determine sample size n, b. From table 6.3.2.1, test plan
and pass/fail number c as 9A is appropriate. The sample
follows: size is n-60, and the test is
passed if 5 or less failures
Choose the smallest c and occur before time T, and the
the smallest n which test is failed if 6 or more
satisfy the failures occur before time T.
inequalities: n)C(1l~-Ro
Ro-
> (1-R-R
This value of n is the

sample aize, and c is the
decision criterion; that
6-17
3. Method Example
b. is, the test is passed

for c or less failures
before time T, and the
test is failed if c+l or
more failures occur before
time T.
Alternatively, table 6.3.2.1

can be used to identify
n and c. ,
c. Because the binomial c. From table 6.3.2.1, test

distribution is discrete, plan 9A, the exact risks
the planned risks cannot are:
be achieved exactly. The
exact producer's and * .079
consumer's risks are .097
given by one minus the
first summation in b
above, and the second
summation in b above,
respectively. Alterna-
tively, if table 6.3.2.1
is used, the exact pro-
ducer's and consumer's
risks are given there.
The test plans LA-13A '""
are based on planned values
of .10 for both risks, and
the test plans 1B-13B are
based on planned values of
.20 for both risks.
"d. Once a test plan is d. Figure 6.3.2.9A is the

defined, it is often operating characteristic
necessary to know what curve for test plan 9A.
the probability of As expected, when true
passing the test is as a reliability is RO - .95,
function ot true relia- the probability of
bility, that is, the acceptance is 1- .079- .921, L%
operating characteristic and when true reliability
curve is needed. This is Rl- .85, the probability
curve gives the probabi- of acceptance is .097. If,
lity of passing the test for example, true reliability
(i.e. the probability of is .90, then the probability -.
accepting the parts or of acceptance is about .45.
equipments) for the .. "
entire range of possible
values of the reliability,
. .> ..... :.. 6 l. ..

. .
3. Method Example
d. not just at R0 and RI.
The operating character-
istic curve is defined by: %
P {acceptance I reliability-R}
S- ~( 1-R)k~'
k-O-o
Figures 6.3.2.•A-6.3.2.13A
and 6.3.2.1B-6.3.2.13B are

the operating characteris-
tic curves for test plans
1A-13A and 1B-13B,
respectively.
TABLE 6.3.2.1. NONPARAMETRIC RELIABILITY DEMONSTRATION TEST PLANS
The planned producer's and consumer's risks are .10 for test plans
1A-13A, and .20 for test plans IB-13B.
ACCEPT REJECT,.
Plan alpha beta RO Ri n equal or lest al or more
IA .095 .096 .50 .40 168 92 93

2A .091 .099 .80 .70 127 31 32
3A .087 .099 .85 .75 109 21 22
4A .086 .099 .90 .80 86 12 13
5A .096 .094 .91 .81 79 10 11
6A .087 .093 .92 .82 77 9 10
7A .095 .098 .93 .83 67 7 8
8A .088 .096 .94 •,4 64 6 7
9A .079 .097 .95 .85 60 5 6
10A .073 .092 .96 .86 56 4 5
11A .063 .096 .97 .87 50 3 4
12A .055 .097 .98 .88 43 2 3
13A .045 .099 .99 .89 34 1 2
lB .181 .194 .50 .40 77 42 43
2B .197 .190 .80 .70 55 13 14
3B .191 .183 .85 .75 49 9 10
4B .190 .180 .90 .80 39 5 6
5B .187 .199 .91 .81 34 4 5
6B .157 .199 .92 .82 36 4 5
7B .183 .183 .93 .83 32 3 4
8B .145 .184 .94 .84 34 3 4
9B .163 .187 .95 .85 28 2 3
lOB .117 .189 .96 .86 30 2 3
SliB .151 .180 .97 .87 23 1 2
12B .083 .199 .98 .88 24 1 2
13B .131 .196 .99 .89 14 0 1
6-19
. . .- ..
". . . "".
1.00
0.95 :
0.90V
P 0.85
6N 0.80
a 0.75
A 0.70
1 0.45
L0.40/
AT 0.35
C 0.30 /
0.40
A0.305--
C0.30
0.00
0.35 0.40 0.45 0.50 0.511

TRUE RELIABILITY
Figure 6.3.2.1A. Operating Characteristic Curve for Test Plan 1A .
A 0.70
A0.67 .
L0.60
IT 0.55
V0.50
0 0.45
F 0.40
A 0.35
C 0J.30
C
E 0.25
P 0.20
T 0.15
0.10
0.60 0.65 0.70 0.75 0.60 0.85

TRUE RELIABILITY
Figure 6.3.2.2A. Operating Characteristic Curve for Test Plan 2A
6-20
1.00
0.95 'I
0.90
P 0.65
R
0 0.80
a 0.75
A 07
a
1 0.65
L 06
T 0.55
V 0.50
o 0.45
F 0.40
A 0.35
C 0.30 (.1.
E 0.26
P 0.20
T
0.15
0.10
0.00
0.65 0.70 0.75 0.60 0.66 0.90

TRUE RELIABILITY
1.00
0.35
P 0.3085..
R 0.65
0 0.10
T 05
Y 0.10 I..
A 0.00 -~.
0.7 0.75 0.008 .009

ERERLAULT,
0.3204A
FPur Oprtn-hrceitcCrefrTs ln4
T 0.6-21
........ ~.................0.....
. .**.* . . . .. . ..
. . . .. . . . . . . . . . . . . . . . . . . . . .
0.. .0.0... .0..0.0...

.0..0. ... .
.
TRUE.. .. . . . .
RELIABILIT . ...
IE
1.00 -4
0.95 VA
0.90
P 0.85
R
a 0.75
0,60 S0.65
Ir 0.50
V0150
S0.70
o 0.45
0.40
A 0.35
IC C
E
T
0.30
0.25
o.,o
P 0.20
0.15
005.
OASb
0.00
.....................................................
I" r,..... I......... v ........... ••
0.75 0.80 0.85 0.90 0.95

TRUE AELIAOILITY b
0.90
P 0.05
S0.75
0,70
0.80
0.55
LTO.S5
y 0,50 S0.35
o0.45
F 0.40
0.30
C 0.25
E
P 0.20
0.15
0.10
0.05
0.00 -- •
0.75 0.80 0.85 0.g0 0.95 1.00 -"•

TRUE RELIABILITY
6-22
...... _.................,......."...'.'.... .... .......... ,...'.....-..'........'..• ........ '..- ... . -'. . '.. -. :4,•'.'
. ....... ' ... .- .. . . - ' -. .
:..., '.'•
.' .' '._
.. ..':.'::":.
•-''..,' "'.'.
:: : -,::,:'., :":'-, :.",':,::. '." -" "., '.L" ,_'" .• ",:: :, :•• " ';':.":.',,'-" .;" '.
PPd
1.00 -d
0.95
"0.90
P 0.85
.0 0.60
8 0.75
S0.65
0.60
.s• T 0.55
0 0.45
0.
0,40
C 0.30
E 0.25
P 0.20-
0.15
o~.o
4- ~~0.10 r'..
0.05
0.00
0.75 0.60 0.55 o.s0 0.96 1.00

TRUE RELIASILITY
Figure 6,3.2.7A. Operating Characteristic Curve for Test Plan 7A
1.00 ,...
0.85
a 0.75
A0.70
L 0.60 .000099
T 0.55 ,.
Y
R 0.50,,"
0.20 .
0.1
0.0.45
0.40 . .,,
A 0-35
0.00
C 0.25 ' ,.•
P 0.20 - ", ,
T 0.16 't'
0.10 "
0.05
0.00
I .%*,•
0.75 0.00 0.85 0.90 0.95 2.00 ,,,
TRUE RELIABIL ITY •''

. . . . . . . . . . . . . . . . . . . . . .
6- 2 3 -
'p
1.00
0.95
0.90
P 0.85
R
0.60
B 0.75
A 0.70
a
1 0.65
"0.60
T 0.SS
0.50
o 00.45
0.40
A0.35
0.30
C
E 0.2S
P 0.20
T
0.1
T
0.10
0.05
0.00.65 0.000 0.90 0.95 1.00
TRUE RELIABILITY
M'U
1.00
0.95
0.90
p 0.85
R 0.80
A
00.75
0.70 I
"L
L. 0.65
(.60
IT .55
-- " J
y 0.50
,. J.45 -
0.40
A 0.36
C 0.30
E 0.25
P 0.20
T 0.15
0.10
0.05
0.00
0.80 0.93 0.90 0.9s 1.00

TRUE RELIABILITY
6-24
S.
, .. .. . . .-, ,-. ... '_, ,, ,-. . -. '• '. '. ,• ,", ',, ,- -' ,", --. .- , .- ., -• .•.. ', ". • , _. _:., _M '
1.00
0.95
0.90
P 0.65 t
0 0.$0
S 0.75
A 0.70
I 0.65
"L 0.60
T 0.55
0.50
o 0.45
F 0.40
A 0.35
C 0.30
E 0.25
p 0.20
TRUE
Figure 6.3.2.11A. Operating Characteristics Curve for Test Plan 11A
1.00
0.1s
0.90
p 100
0.85
R 0.80
a 0.95
A p 0.65
0.70
L- 0.60 ,, .
T 0.SO5,,, .
0 0.45
F 0.40
A 0.35
C 0.30,"-"-
C
E 0.25
P 0.20
T 0.15 '''
0.05
0.00
.............................. -1......... "' ...... ••..................
0.60 0.65 0.90 0.95 1.00

TRUE RELIABILITY
6-25 ..¶'...
. .. .. . . -... .... -i " .... .... ...." ... ..... .'| * ... . .....
1.00 -
0.95
0.90
P 0.65
o1 0.60
* 0.7-
A 0.70
I 0.65
L 06
IT 0.55
y0.50
o 0.45
0.40
A 0.35 I'.
C 0.30
E 0.25
. 0.20
0.15
0.10
0.05
0.00
0.60 0.55 0.90 0.to 1.00

"TRUE RELIABILITY
w*Yg
r.. 1
,,,
tj.
6-26
i. . d10 00 A d
bOlo
I
0.90
A
S0.0
0 0.71
.,. 0.11 +
V 0,10
0.@40
0.41
P 0.40
0.$0
C 0,30
0,10
0.01
0.00
.............. !"l................... !%I......... !........ li......... •........~ *
0,30 0.3 0.40 041 040 0.1 0.40

TRiUC 9I6IAOIILITV
Figurs 63.2.1B. Operating Charzoefistlc Curve for Tom Plan 18
0,90
0.75
0.00
S0,30 ,*',.
@0,40
0.o,0
T 0.20
040
*A 0,210.IA
0.03
0.00 ""....'I .. .... I ......... I.... +I -- lT- ...... - - ....... ''p
*0.00 0.61 0.70 0.71 0.30 0.81 0.90 ,',"

TRUI[ II1LIAISILITY *
Xp.
AFiwe 6.3.2.25. Operating Characteristic Curve for Test Plan 2B

I.P
4,-27
Ib
.... , •.'... -••
_;•
..•:. . . *•..-1;•*•
"1•. _"-'-"-'-•-*- . ,; . ..•_._
-..•
1.00
'4* 0.00
o.eo
A S0.10
60.76
0.66
T 0.66
00.46
F 0.40
A 0.36
00 C 0.30
j
*026
0.16
* 0,011
0.00 -t
0.60 0.66 0.70 0.76 0.60 0.65 0.90 0.95
1.0 Figure 6.3.2.35. Operating Characteristic Curve*frTsPln3
51.06
0190
0.66
0.60
O0.46
C 0.00
* C
£0.26
S0.20
0.16
0.061
0.0
V 0,1 "•
0.70 0.76 o 0.80 o.,'F- 0.66 0.90 0.o6 1.00
Tu" RELIABILITY
Figue 6.32.4S. Operating Characteristic Curve for Test Plan 413
6-28
1.00 P
0.95
0.90
P 0.35
R 0.g0
0 0.75
8A 0.70
S
1 0.65
L 0.60
IT 0.55
V 0.60
0 0.45
F 0.40
A 0.35
C 0.30
c 0.25
P0.20
0.10
0.05
0.08
0.70 0.75 0.80 0.35 0.90 0.95 1.00

TRU9 RELIABILITY
Figure 6.3.2.5B. Operating Characteristic Curve for Test Plan 5B
1.00
0.95
II
9
0.90
P 0.85
T 0.200
a0.15
0.10
a, 0.05
0.600
F,0.6 .0 07 .0 05 .0 09
*~~~~ 0.35ELAILTY
0... Fiur peaigCarceitc uv o0Ts ln6

6-2
. ......
E.2.......................................#
0 ,. .
1.00
0.95
0.90
P 0.85
R 0.80 .
0 0.75
aA 0.70
01 0.65
L 0.6
IT 0.55
y 0.50
0 0.45
F 0.40
A 0.35
0.2
C 0.30
P 0.20 a
T 0.15
0.10 I
0.05
0.00
....................................................................... .................
0.65 0.70 0.7S 0.80 0.85 0.90 0.95 1.00
1.00
I 0.95
a 0.70
A 0.75
1 0.25
p 0.20P
T OS
Y0.50
F0.10
0.7 0.50.35.5 .00.510
C 0.30
IC
1.00
0.905
0.90
P 0.115
ft 0.00
0.7
A 0.70
L 0.00
IT 0.56
T 0.50
0.45
F 0.40
A 0.3S
C 0.30
c 0.25
* 0.20
*~ 0.15 l
0.10
0.05
0.00 ..............................
"" 1....... ...... 8.... " 9 "" .... 9 ..... " 1
0.70 0.7S 0.30

TRUE RELIABILITY
0.o5 0.90 0.95 1.00
Ni
:.oo~Q -- "
1.00
0.95 -,1
0.90
0.0
S0.75
A 0.70
S0.60
IT o.5ss -• x .
0 0.4S5r •.
F 0.40
A 0.35
C 0.30 ,..P
C
E 0.25
P 0.20
T oi
0.10
0.05
0.00
0.75 0.C.0 8.S5 0.00 0.95 1.00
TRUE RELIABILITY
6-31
. .. .. .. .. .. . . . . . . . . . . . . .
...- . .*: _ ........ "

- ....... . ...... -. •.•...•....... ... ..• -..... - __
.. ,
. . . .... . . .. . rTr
.--
............... rT 'rT.....
1.00
0.95
p 0.90
0.65
R 0.60go
0
a 0.75
A 0.70
0.65
L 0.60
T 0.53
S0.30
0 0.45
F 0.40
A 0.35
C 0.30
£ 0.25
P 0.201
T 0.is -
<-.-.
0.10
0.05
0.00
0.75 0.90 0.85 0.90 0.95 1.00

S•rVI RELIABILITY .. *•'
Figure 6.3.2.11B. Operating Characteristic Curve for Test Plain liB

11
A ~~1.00 4P
-f .. "
0,95 A
0.90
R 0.85
0 ,080-"•
A 0.75
* 8 0.70
0.65 -4
L
1 0.60-I
T 0.55 -
o 0.45-
0.40
A 0.35 -.
c 0.30.
C
P 0.20 I
0.15 -' -
0.10 -, "". " -
0.05 -
0.00
0.75 0.80 0.85 0.90 0.95 1.00
TRUE RELIABILITY
Figure 6.3.2.12B. Operating Charac:cristi,. Curve for Test Plan 12B
b-32
.7....
" " " """' '""" '"" " " .. .3 . . . ." ' ' •-•• * • . ;'-•.-r- ... -• ...
1.0o -
og$g -• / .
P 0.90.--
R 0.85
a 0.80 /
A 0.75-
/ ,6
a 0.70
L 0.65 -*
**. o.,o
0.60 -1 , Jh 06
0.O0-1
0.40 1 -"
0.35 -1 .-
C
C 0.30
0 0.25.
" 0.20 1
0.10 -
0.05 J-
0.00 -, ........ ___....... - ......... L
0.75 0.,0 0.65 0.90 O.A 1.00
TRUE RIELIASILITY *
.47................
.:: ,.:. ..
. •'°'S "b "-"
6-33 (6-34 BLANK)

I
2
%
I p.*. * h
4.
.' J
APFENDI I
REFERENCES MID I IBLIOGRAPHY
;-< ..
% 9p*
P
.4
* * * *
APPENDIX I REFERENCES AND BIBLIOGRAPHY
ASCE (1982). Fatigue reliability: Introduction. ASCE, Vol. 108, No. ST1,
June 1982, pp. 3-23.
ASME (1965). Mechanical reliability concepts. ASME design engineering

conference, New York.
Abramowitz, M. and Stegun, I. A. Eds. (1965). Handbook of Mathematical
FrtictJons. Nev York: Dover.
Ang, A.H.S. (1970). Extended reliability basis of structural design under

uncertainties. 1970 Annals of reliability and maintainability conference.
New York: IEEE., pp. 642-649.
Atchison, J., and Brown, J.A.C. (1963). The iognormal distribution. Cambridge
University Press, New York.
Austin, W. H. (1967). Development of improved gust load criteria for USAF

aircraft. 1967 Annals of Reliability and Maintainability Conference, New
York: IEEE, pp. 68-74.
Bain, L. J. (1972). Inference based on censored sampling from the Weibull or

extreme value distribution. Technometr~cs, 14, 693-702.
Bain, L. J. (1978). Statistical Analysis of Reliability and Life-Testing
Models. New York: Dekker.
Barlow, R. E. and Proschan, F. (1964). Mathematical Theory of Reliability.

New York: Wiley.
Bartholomew, D. J. (1963). The sampling distribution of an estimate arising

in life testing. Technometrics, 5, 361-374.
Bazovsky, I. (1971). Reliability of rifles, machine guns and other small

arms. Proceedings Annual Symposium on Reliability, IEEE, New York, pp. 71-73.
Bell, R., and Mioduski, R. (1976). Extension of life of U.S. Army trucks.
1976 Annual Reliability and Maintainability Symposium, IEEE, New York, pp.
200-205.
Bennet, S. B., Ross, A. L., and Zemanick, P. Z., editors (1977). Failure
prevention and reliability. Society of Mechanical Engineers, New York.
Bhattacharyya, G. K., and Johnson, R. A. (1975). Stress-strength models for

system reliability, reliability and fault tree analysis. Philadelphia: SIAM,
pp. 509-532.
BLllet, R. B. (1964). Reliability of hydraulic controls in space vehicles.

National Symposium on Reliability, IEEE, New York, pp. 69-84.
rFI[fman, B. Antle, C. and Bain, L. J. (1972). Statistical Inferences from

.. n.,ored Weibuli sample. Technometrics, 14, 831-840.
T-i
Birnbaum, Z. W., and Saunders, S. C. (May 1968). A probabilistic

interpretation of Miner's rule. SIAM Journal on Applied Mathematics, Vol. 16, .:
No.3, pp. 637-652.
Birnbaum, Z. W., and Saunders, S. C. (October 1969). A new family of life
distributions. J. Appl. Probl., pp. 319-327.
Bompass-Smith, J. H. (1969). The determination of distributions that describe

the failures of mechanical components. Annals of Reliability and
Maintainability Conference, IEEE, New York, pp. 343-356.
Bompass-Smith, J.aH. (1973). Mechanical Survival: The Use of Reliability
Bratt, M. J., Truscott, H. A. and Weber, G.W. (1968). Probabilistic strength

mapping reliability vs life prediction tool. 1968 Annals of Reliability and
Bratt, M. J., Reethoff, G., and Weber, G. W. (1969). A model for time varying..
and interfering stress/strength probability density distributions with
consideration for failure incidence and property degradation. Proceedings 3rd
Annual Aerospace Reliability and Maintainability Conference, IEEE, New York,
566-575.-----
Brewer, J. W. (1971). Interim scale reliability statements consistent with

conventional materials strength criteria. 1971 Annals of Reliability and
Buckland, W. R. (1964). Statistical Assessment of the Life Characteristics.

London: Griffin. ,
Burns, J. J. (1975). Reliability of nuclear mechanical systems. Annual

Reliability and Maintainability Symposium, IEEE, New York, pp. 163-169.
Carter, A.D.S. (1972); Mechanical Reliability. London: Wiley.
Chestei', L. B., Nolf, C. F., and Kececioglu, D. (1975). Com•bined

bending-torsion fatigue reliability - 111. 1975 Annaal Reliability and
Maintainability Symposium, IEEE, New York, pp. 511-518.
Cohen, A. C. (1965). Maximum likelihood estimation in the Weibull

distribution. Technometrics, 17, 347-351.
Collins, J. A., Hagan, G. T., and Bratt, H. M. (1975). Helicopter failure
modes and corrective actions. 1975 Annual Reliability and Maintainability
Symposium, IEEE, New York, pp. 504-510.
Crow, L. H. (1974). Reliability Analysis for complex, repairable systems. In

Reliability and Biometry, F. Proschan and R. J. Serfling, Eds. Philadelphia:
SIAM.
D'Agostino, R. B. and Stephens, M., Eds. (1984). Goodness-of-Fit Techniques.
New York: Dekker.
1-2
-:~:--i.--~-%-
. - - - - . - - -*' %-~' .. . :-.. . . .. ". .-
°-. -. -_ .-. -..-. - "........................................................................................................................................................... %° %
Davidson, J. R. (1973). Reliability after inspection. Proceedings of the

ASTM Symposium on the Fatigue of Composite Materials, Bal Harbor, Florida.
Davis, D. J. (1952). An analysis of some failure data. J. Amer. Stat.

Assoc., pp. 113-150.
Dehnrdt, J. H., and McLaughlin, H. D. (1966). Using Bayesian Method* to

Select a Design with Knoron Reliability Without a Confidence Coefficient.
1966 Annals of Reliability and Maintainability Conference. New York: IEEE,
•' pp. 611-617.•
Dennis, N. G. (1977). PMR, NDE, design practices; present and future. 1977
Z-(:4
"Annual
16 - 170Reliability
.
and Maintainability Symposium. New York: IEEE, pp.
Dhillon, B. S., and Singh, C. (1981). Engineering Reliability - New

Techniques and Applications. John Wiley.
Dhlllon, B. S. (1980). Mechanical reliability: Interference theory models. ""-

Proceedings Annual Reliability and Maintainability Symposium, pp. 462-467.
Dhillon, B. S. (1981). Interference theory application to mechanical

systems. Proceedings Annual Reliability and Maintainability Symposium, pp. 'j
24-27. , -
Dillin, A. L. (1971). Reliability assessment of army weapons and weapon

systems. Proceedings of the Annual Symposium on Reliability, IEEE, New York,
pp. 74-76.
Disney, R. L., and Sheth, N. J. (1968). The determination of the probability...-" 2'
of failure by stress/strength interference theory. 1968 Annual Symposium on
Reliability, IEEE, New York, pp. 417-422.
Disney, R. L., Lipson, C., and Sheth, N. J. (January 1968). The determination
of the probability of failure by stress/strength interference theory..*...
Proceedings of the Symposium on Reliability, pp. 417-422.
Dixon, W. J. and Brown, m. B. (1977). BHDP-77: Biomedical Computer Programs

P Series. Berkeley: University of California Press.
Dumonceau, R., and Antle, C. E. (November 1973). "Discrimination between the
log normal and Weibull distributions." Technometrics, Vol. 15, No. 4, pp.
923-926.
Engelhardt, M. (1975). On simple
extreme value distribution. of 369-374.
estimation17,
Technometrics, parameters of the Weibull or ,-'"
Engelhardt, M. and Bain, L. J. (1973). Some complete and censored sampling

results for the Weibull or extreme value distribution. Technometrics, 15,
542-549.
1-3
•. , . * *. --., . -. . - - ,. ° , . .- . . .° * . .,*.. -. •. .-.. . . . . . -. ..o . ~ , .. . ' o O" .I o

• ,,• .,•,,,• •,,.•
.. , ,-•,•..,. • ,. • ,.• • •: . • .. _ _• .• •. • .-.- • .. " '...-', ", ', .,.•.-. , . - .. '. " . .. '...
Engelhardt, M. and Bain, L. J. (1974). Some results on point estimation for

the two-parameter Weibull or extreme value distribution. Technometrics, 17,
353-356.
Engelhardt, M. and Bain, L. J. (1977). Simplified statistical procedures for

the Weibull or extreme value distribution. Technometrics, 19, 323-331.
Epstein, B. and Sobel, M. (1953). Life Testing. Journal of the American

Statistical Association, Vol. 48.
Feller, W. (1968). An Introduction to Probability Theory and Its .9

Applications, Vol. 1, 3rd ed. New York: Wiley. • \,,
Fiderer, L. (1976). Design for reliability in hostile environment.

Microelectronics and Reliability, Vol. 15 Supplement, pp. 75-85.f '
Fisher, G. K. (June 1981). Advanced power transmission technology. NASA L*. .' .
Conf. Pub. 2210, AVRADCOM TR-82-C-16.
Fisher, R. A. and Tippett, L. M. C. (1928). Limiting forms of the frequency

distribution of the largest or smallest member of a sample. Cambridge Phil.
Soc., 24, 180-190.r--,-".....,.,.
Flint, S. J., and Steinkirchner, J. J. (1982). Reliability models - practical

constraints. NATO ASl Electronic Systems Effectiveness and Life Cycle Costing
Conference. New York: Springer Verlag.
Forrestor, E. R., and Thevenow, V. H. (1968). Designing for expected fatigue

life. 1968 Annals of Reliability and Maintainability, IEEE, New York, pp.
511-519.
Furth, J. A., Upton, C. and Kimball, A. W. (1959). Late pathologic effects of

atomic detonation and their pathogenesis. Radia. Res. Suppl. 1, 243-264.
Chane, P. M. (1970). Quality and safety factors in reliability. 1970 Annals

of Reliability and Maintainability, IEEE, New York, pp. 637-641.
Gnedenko, B. V. (1943). Sur la distribution limits du terme maximum d'une

series aleatolre. Ann. Math., 44, 423-453.
Grow, A. J., and Clark, V. A., Survival Distribution: Reliability

Applications in the Biomedical Sciences. Wiley, New York, 1975. . .
Grubbs, F. E. (1971). Fiducial bounds on reliability for the two-parameter

negative exponential distribution. Technometrics, 13, 873-876. I TAM
Gumbel, E. J. (1958). Statistics of Extremes. Columbia University Press, New

York.
Gupta, S., and Groll, P. (December 1961). Gamma distribution in acceptance

sampling based on life tests. J. Amer. Stat. Assoc., IEEE, New York, pp.
942-970.
1-4
. . .. .
Hahn, G. J., Nelson, W. B. and Cillay, C. (1975). STATSYSTEM- A

user-oriented interactive system for statistic&! data analysis in 1975
Statistical Computing Section Proceedings of the American Statistical
Association, Washington, D. C.: American Statistical Association.
Hald, A. (1952). Statistical Theory with Engineering Applications. New
York: Wiley.
Hassanein, K. M. (1972). Simultaneous estimation of the parameters of the

extreme value distribution by sample quantiles. Technometrics, 17, 63-70.
Haugen, E. G. (1964). Statistical methods for structural reliability
analysis. National Symposium on Reliability, IEEE, New York, pp. 97-121.
Havilan, R. P. (1964). Engineering reliability and longlife design.
Princeton: Van Nostrand.
Heller, R. A., and Heller, H. S. (1973). Analysis of early failures in

unequal. 1973 Annual Reliability and Maintainability Symposium, IEEE, New
York, pp. 198-200.
Heller, R. A., and Shinozuka, M. (1970). State-of-the-Art: Reliability.
techniques in materials and structures. 1970 Annals of Reliability and
Ingram, G. E., Herrmann, C. R., and Welker, E. L. (1968). Designing for.

reliability based on probabilistic modeling using remote access computer
system. 1968 Annals of Reliability and Maintainability Conference, IEEE, New
York, pp. 492-500.
Ingram, G. E., Haugen, E. B., Dicks, C., and Wilson, S. (1965). Panel
discussion - structural reliability. National Symposium on Reliability, IEEE,
New York, pp. 154-169.
Inoue, K., Daito, H. (1977). Safety analysis of automobile brake systems by

fault tree methods. Proceedings HOPE International JSME Symposium, Japan
Society of Mechanical Engineers, Tokyo, pp. 213-220.
Japan Society of Mechanical Engineers (1979). Proceedings of the HOPE

international symposium.
Johnson, C. W., and Maxwell, R. E. (1976). "Reliability analysis of
structures - a new approach. 1976 Annual Reliability and Maintainability
Johnson, W. S., Heller, R. A., and Yang, J. N. (1977). Flight inspection data
and crack initiation times. 1977 Annual Reliability and Maintainability
Johnson, N. L. and Kotz, S. (1970). Continuous Univariate Distributions.

Vols. 1 and 2. Boston, Massachusetts: Houghton Mifflin.
I-5
.......
. . . . . . .. . . *-.
" '- -'. ' '.•" .. i :"-.- ---.'.-:' "'.-.. ''- .'. i. -. ".- .- .- " .-" -" ."i"i.-.' .. .'-. .''-i-i-"," -" -"-"-. -.-..-.-- .'
=. -,, • • • 4• - " " " " ' " ' .. " - . . -... . ..• : . . . ." . " .. •...-..:-... ' i.•
Jones, L. G., and Thompson, D. (1971). Essential elements of analysis from e

army test and field data. Proceedings Annual Symposium on Reliability, IEEE,
Kalivoda, F. E., and Yun, K. W. (1976). Mooeling mechanical system

accelerated life tests. 1976 Annual ieliability and Maintainability
Symposium, IEEE, New York, 206-212.
Kao, J.H.K. (1958). Computer methods for estimating Weibull parameters in

reliability studies. IRE Transactions on Reliability and Quality Control.
Kao, J.h.K. (1960). A summary of some new techniques for failure analysis.
Proceedings Sixth National Symposium on Reliability, IEEE, New York, p. 191. . -
Kao, J.h.K. (1959). A graphical estimation of mixed Weibull parameters in

life-testing of electron tubes. Technometrics 1, pp. 389-407.
Kao, J.h.K. (1965). Statistical models in mechanical reliability. National

Symposium on Reliability, IEEE, New York, pp. 240-247.
Kapur, K. C., and Lamberson, L. R. (1977). Reliability in Entineering

Design. New York: John Wiley.
Karnopp, D. (1971). Structural reliability predictions using finite element

programs. 1971 Annals of Reliability and Maintainability Conference, IEEE,
Kececioglu, D., McKinley, J. W., ano Saroni, M. (Jan. 1967). A probabilistic

method of designing a specified reliability into mechanical component. with
time dependent stress and strength distributions. The University of Arizona,
Tucson, Arizona, (NASA report under contract NGR 03-002-044).
Kececloglu, D., and Cormier, D. (1964). Designing a specified reliability

directly into a component. Proceedings 3rd Annual Conference on Aerospace
Reliability and Maintainability. Washington DC, IEEE, New York, pp. 546-565.
Kececioglu, D. (1968). Why design by reliability? 1968 Annals of Reliability

and Maintainability Conference, IEEE, New York, p. 491.
Kececioglu, D. (1972). Reliability analysis of mechanical components and

.aystems. Nucl. Eng. Design, pp. 259-290.
I... .,'.- .,.
Kececioglu, D. (1974). Fundamentals of mechanical reliability theory and

applications to vibroacoustic failures. Proceedings of Reliability Design for -.
Vibroacoustic Environments. ASME, New York, pp. 1-38.
Kececioglu, D. (1977). Probabilistic design methods for reliability and their

data and research requirements. Failure Prevention and Reliability Conference ,
Proceedings, ASNE, New York.
Kececioglu, D., Chester, L. B. and Dodge, T. H. (1974). Alternating bending -

steady torque fatigue reliability. Annual Reliability and Maintainability
Symposium, IEEE, New York, pp. 15-73'3.
1-6
. . . . .. . . . . . .. . .'.. . . . .. . . . . . . . o"
. . .. ...
. .
Kececioglu, D., and Koharcheck, A. (1977). Wear reliability of aircraft

splines. 1977 Annual Reliability and Maintainability Symposium, IEEE, New
York, pp. 155-163.
Kececioglu, D., and Haugen, E. B. (1968). A unified look at design safety

factors, safety margins and measures of reliability. 1968 Annals of
Reliability and Maintainability Conference, IEEE, New York, pp. 520-530.
Kececioglu, D. B., Smith, R. E. and Felsted, E. A. (1970). Distribution of

strength in simple fatigue and the associated reliabilities. Annals of
reliability and maintainability conference, IEEE, New York, pp. 659-672.
Kendall, M.G. (1938). A new measure of rank correlation. Biometrika, Vol. 30.
King, J. R. (1971). Probability Charts for Decision Making. New York:

Industrial Press.
Kirkpatrick, I. (1960). Predicting reliability of electromechanical devices.

Sixth National Symposium on Reliability and Quality Control, IEEE,
New York,
pp. 272-281.
Konno, K., K. Nakano, and Yoshimura, Y. (1975). A new accelerated fatigue

test. The effective random peak method. Annual Reliability and
Maintainability Symnposium, IEEE, New York, pp. 263-268.
Kullman, L. W., and Phillips, G. W. (1964). Reliability engineering

disciplines applied to commercial weapon systems - guns and ammunition.
Kurtz, P. H. (1975). A strength-stress interference model for time-varying

probability of failure and its application to an electromechanical
system.
Naval Sea Systems Command, TM-75-122, AD A016299, NTIS.
Kurtz, P. H. (1971). Reliability study of a hydraulic control system using

the hybrid computer. Annals of Reliability and Maintainability Conference,
IEEE, New York, pp. 81-85.
Lambert, R. G. (1978). Mechanical reliability for low cycle fatigue. 1978

Annual Reliability and Maintainability Symposium, IEEE, New York,
pp. 179-183.
Lawless, J. F. (1982). Statistical Models and Methods for Lifetime Data. New
York: Wiley.
Lemaitre, J., and Plumtree, A. (July 1979). Application of damage concepts to

predict creep-fatigue failures. Journal of Engineering Materials and
Technology, Vol. 101, ASME, pp. 284-292.
Lemon, G. H., and Manning, S. D. (October 1974). Literature survey on

structural reliability. IEEE Trans. Reliab. R-23.
Lieb]ein, J., and Zelen, M. (1956). Statistical investigation of the fatigue

life of deep-groove ball bearings. J. Res. Nat. Bur. Stand., 5 (Res. Paper
2719), pp. 273-316.
1-7
Lieblein, L. and Zelen, M. (1956). Statistical investigation of the fatigue

life of deep groove ball bearings. J. Res. Nat. Bur. Stand., 57, 273-316.
Liebowitz, H. (1967). Navy reliability research. 1967 Annals of Reliability

and Maintainability Conference, IEEE, New York, pp. 33-53.
Linhart, H. (1965). Approximate confidence limits for the coefficient of

variation of gamma distributions. Biometrics, 21, 733-738.
Lipson, C., and Sheth, N. J. (1973). Statistical Design and Analysis of

Engineering Experiment. McGraw-Hill, pp. 311-343.
Lipson, C., Sheth, N. J., and Disney R. L.(March 1977). Reliability

prediction - mechanical stress/strength interference. RADC-TR-66-710.
Lipson, C., Sheth, N. J., Disney, R. L., and Altun, M. (December 1968).
Reliability prediction - mechanical stress/strength interference
(nonferrous). RADC-TR-68-403.
Lloyd, G. J., and Wareing, J. (August 1981). Life-prediction methods for

combined creep-fatigue endurance. Metr!s Te-hnology, pp. 297-305.
Lloyds, D., and Lipow, M. (1961). Reliability: Management, Methods and

Mathematics. Englewood Cliffs, NJ: Prentice-Hall.
Mann, H. (1945). Nonparametric tests against trend. Econometrica, Vol. 13.
Mann, N. R., Shafer, R. E. and Singpurwalla, N. D. (1974). Methods for

Statistical Analysis of Reliability and Life Data, Wiley, New York.
Mann, N. R. (1 9 67a). Results on location and scale parameter estimation with

application to the extreme value distribution. Wright-Patterson Air Force
Base, Ohio, ARL Technical Report ARL 67-0023.
Mann, N. R. (1967b). Tables for obtaining the best linear invariant estimates
of parameters of the Weibull distribution. Technometrics, 9, 629-645.
Mann, N. R. (1968) Point and interval estimation procedures for the

two-parameter Weibull and extreme value distributions. Technometrics, 10,
231-256.
Mann, N. R. (1977). An F approximation for two parameter Weibull and

log-normal tolerance bounds based on possibly censored data. Nav. Res. Log.
Q. 24, 187-196.
Mann, N. R. and Fertig, K. W. (1973). Tables for obtaining confidence bounds

and tolerance bounds based on best linear invariant estimates of parameters of
tie extreme value distribution. Technometrics, 15, 87-101.
Manr, N. R. and Fertig, K. W. (1 9 7 5a). Simplified efficient point and

t, rvaIl erstimators for Weibull parameters. Technometrics, 17, 237-245.
T-9
Mann, N. R. and Fertig, K. W. (1975b). A goodness-of-fit test for the two •

parameter vs. three-parameter Weibull: confidence bounds for the threshold.
Technomer.rics, 17,.237-245.
SMann,Efficient unbiased quantile estimators
N. R. and Fertig, K. W. (1977).
for moderate-size complete samples from extreme value and Weibull
distributions: confidence bounds and tolerance and prediction intervals.
Technometrics, 19, 87-93. •.
Manning, S. D., and Lemon, G. H. (1974). Plan for developing structural

criteria for composite airframes. 1974 Annual Reliability and Maintainability
Symposium. IEEE, New York, pp. 155-162.
Marble, Q. G. (1965). Improving mechanical reliability of digital computers.

National Symposium on Reliability and Quality Control, IEEE, New York, pp.
136-143.
Martin, P. (1976). Reliability in mechanical design and production.

Proceedings Generic Techniques in Systems Reliability Assessment.
Noordhoff-Leyden, Amsterdam, pp. 267-271. V
Matney, V. D. (1974). Reliability, pollutants and aluminum raw. Annual _

Mesloh, R. (1966). Reliability design criteria for mechanical creep. 1966

Annals of Reliability and Maintainability Conference, IEEE, New York, pp.
590-597.
Moreno, F. J. (1973). Reliability estimate of a space deployable antenna.

1973 Annual Reliability and Maintainability Symposium, IEEE, New York, pp.
182-185.
Mukkayee, B. (January 1979). Residual fatigue life determination of anchor

bolts. Journal of Endineering Materials and Technology, Vol. 101, ASME, pp.
47-52.
Naresky, J. J. (1966). Reliability and maintainability research in the USAF.

Reliability and Maintainability Conference, IEEE, New York, pp. 769-787.
Nelson W. and Hahn, G. J. (1972). Linear estimation of a regression I

relationship from censored data. Part I - Simple methods and their
applications. Technometrics, 14, pp. 247-269.
Nelson, W. (1982). Applied Life Data Analysis. New York: Wiley.
Nilsson, S. 0. (1975). Reliability data on automotive components. Annual .

Niyogi, P. K. (1968). Application of statistical methods and information

theory to structural reliability estimates. PhD Dissertation, Department of
Civil Engineering, University of Pennsylvania, Philadelphia, PA.
1-9
Niyogi, P. K., Shah, H. C., Doshi, K. D., and Tang, W. (April 1969).
Statistical evaluation of load factors for concrete bridge design. Chicago.
Odeh, R. E. and Oven, D. B. (1980). Tables for Normal Tolerance Limits,

Sampling Plans, and Screening. New York' Dekker.
Persson, T. and Rootzen, T. (1977). Simple and highly efficient estimators

for a Type-I censored normal sample. Biometrika, 64, 123-128.
Pettitt, A. N. (1976). Cramer-von Mises statistics for testing normality with

censored samples. Biometrika, 63, 475-481.
Pettitt, A. N. (1977). Tests for the exponential distribution with censored
data using Cramer-von Mises statistics. Biometrika, 64, 629-632.
Proschan, F (1963). Theoretical explanation of observed decreasing failure

rate. Technometrics, 5, 541-554. v
Quality assurance: Reliability handbook. (October 1968). Headquarters, U.S. .. ..

Army Material Command, (AMCP 702-C, AD 702 936), NTIS, Springfield, VA.
Redler, W. M. (1966). Mechanical reliability research in the NASA, 1966 . .

Annals of Reliability and Maintainability Conference, IEEE, New York, pp. -.
763-768.
Reethof, C. (1973). Session organizer's report. Annual Reliability and '." .
Maintainability Symposium, IEEE, New York, p. 181.
Reethof, G. (1971). State-of-the-art; Mechanical and structural

reliability. 1971 Annals of Reliability and Maintainability Conference, IEEE, INX
New York, p. 62.
Roberts, N. (1964). Mathematical Methods in Reliability Engineering. New

York: McGraw-Hill.
Sarhan, A. E. and Greenberg, B. G. (1962). Contributions to Order

Statistics. New York: Wiley.
S•
".. -.:•
Schatz, R., Shooman M. L., and Shaw, L. (January 1974). Application of time
dependent stress-strength models of non-electrical and electrical systems.
Proceedings of the Reliability and Maintainability Symposium, IEEE, New York,
pp. 540-547.
Schupbach, H. and Husler, J. (1983). Simple estimators for the parameters of

the extreme-value distribution based on censored data. Technometrics, 25,
189-192.
Shah, H. C. (1966). Use of maximum entropy In estimating the damage ....-

distribution of a single degree of freedom syotem subjected to random
loading. 1966 Annals of Reliability and Maintainability Conference, IEEE, New ' .
York, pp. 598-604.
'.o , .- " " '•4

1-10
"C,. '•' '
* . .. .. - .-
AV.'.. '.."
Shah, H. C., and Tang, W. H. (1970). Statistical evaluation of load factors

in structural design. 1970 Annals of Reliability and Maintainability
Conference, IEEE, New York, pp. 650-658.
Shaw, L., Shooman, M. and Schatz, R. (1973). Time-dependent stress-strength

models for non-electrical and electronic systems. 1973 Annual Reliability and
Maintainability Symposium, IEEE, New York, pp. 186-197.
Shinozuka, H., and Hanai, M. (1967). Structural reliability of a simple rigid

frame. 1967 Annals of Reliability and Maintainability Conference, IEEE, Rew
York, pp. 63-67.
Shinozuka, M., and Itagaki, H. (1966). On the reliability of redundant

structures. 1966 Annals of Reliability and Maintainability Conference, IEEE,
Shooman, H. L. (1968). Probabilistic Reliability: An Engineering Approach.

New York: McGraw-Hill.
Singpurvalla, N. D. Statistical fatigue models: A survey, IEEE Trans.

Reliab., R-20, 1971, pp. 185-189.
Spoormaker, J. L. (1977). Reliability prediction of hairpin type springs. .
1977 Annual Reliability and Maintainability Symposium. IEEE, New York, pp.
142-147.
Spoormaker, J. L. (1975). Design of reliable plastic assemblies. 1975 Annual

Stephens, M. A. (1974). EDF statistics for goodness of fit and some .4h
comparisons. J. Amer. Statist. Assoc., 69, 730-737.
Stephens, M. A. (1977). Goodness of fit for the extreme value distribution.

Biometrika, 64, 583-588.
Stiles, E. M. (1964). Reliability in mass-produced consumer products.

Tallian, T.E. (.7ily 1982). A unified model for rolling contact life
prediction. Journal of lubrication technology, Vol. 104, pp. 336-346.
Taraman, S. I., and Kapur, K. C. (1975). Optimization considerations in

design reliability by streos-strength interference theory. IEEE Trans.
Reliability, 24, pp. 136-138.
Thomas, J. H., Hanagnd, S. and Hawk, J. D. (1975). Decision theory in

structural reliability. 1975 Annual Reliability and Maintainability
Thoman, D. R., Bain, L. J. and Antle, C. E. (1969). Inferences on the

parameters of the Weibull distributioa. Technometrics, 11, 445-460.
I-il
.- 4' . .........
........ •.. . .... . . . .. .\.. :'_ - . ..- .•.•.... ,_••....
b.+
Thoman, D. R., Bain, L. J. and Antle, C. E. (1970). Reliability and toler.nvce

limits in the Weibull distribution. Technometrics, 12, 363-371.
.,I Tumolillo, T. A. (1974). Methods for calcu.sting thi. -: "ity function fir
systems subjected to rarnm stresses. IEEE Trans. Reliability, 23, pp.
256-262.
Weaver, L., and Scarlett, T. (1965). Reliability and failure distributions of

inertial sensors. National Symposium on Reliability and Quality Control,
IEEE, Nev York, p. 144-153.
Weibull, W. (1951). A statistical distribution function of wide

applicability. J. Appl. Mech., 18, pp. 293-297.
WeLker, D. R., and Buchanan, H. N. (1975). Safety-availability study methods

applied to BART. 1975 Annual Reliability and Maintainability Symposium, IEEE,
Whittemore, A., Altechuler, B. (1976). Lung cancer incidence in cigarette

smokers: further analysis of Doll and Hill's data for British physicians.v.
Biometrics, 32, 805-816.
Wilson, D. S., and Smith, R. (1977). Electric motor reliablity model. Rome
Air Development Center, RADC-TR-77-408 dated Dec 77, AD 050179, NTIS.
Yadav, R. P. S. (1973). A reliability model for stress vs strength problem.

Microelectron. Reliability, 12, pp. 119-123.
Yadav, R. P. S. (1974). Component reliability under environmental stress.

Microelectron. Reliability, 13, pp. 473-475.
Yao, J. T. P., and Yeh, H. Y. (1967). Safety analysis of statically
indeterminate trusses. 1967 Annals of Reliability and Maintainability

Conference, IEEE, New York, pp. 54-62.
Zaretsky, E. V. (1971). Life adjustment factors for ball and roller

bearings. American Society of Mechanical Engineers.
.......................................................
•" •.
† † ††... † † † † † † † †
1-12
APPENDIX II
STATISTICAL TABLE S
TABLE I. MEDIAN RANKS
sample size = n
failure rank j
j 1 2 3 4 5 6 7 8 9 10
1 .5000 .2929 .2063 .1591 .1294 .1091 .0943 .0830 .0741 .0670
2 .7071 .5000 .3864 .3147 .2655 .2295 .2021 .1806 .1632
3 .7937 .6136 .5000 .4218 .3648 .3213 .2871 .2594
4 .b409 .6853 .5782 .5000 .4404 .3935 .3557
5 .8706 .7345 .6352 .5596 .5000 .4519
6 .8909 .7705 .b787 .6065 .5481
7 .9057 .7979 .7129 .6443
8 .9170 .8194 .7406
9 .9259 .8368
10 .9330
sample size = n
"failure size = j
3 11 12 13 14 15 16 17 18 19 20
1 .0611 .0561 .0519 .G483 .0452 .0424 .0400 .0378 .0358 .0341
2 .1489 .13b8 .1266 .1788 .1101 .1034 .0975 .0922 .0874 .0831
3 .2366 .2175 .2013 .1873 .1751 .1644 .1550 .1465 .1390 .1322
4 .3244 .2982 .2760 .2568 .2401 .2254 .2125 .2009 .1905 .1812
5 .4122 .3789 .3506 .3263 .3051 .2865 .2700 .2553 .2421 .2302
6 .5000 .4596 .4253 .3958 .3700 .3475 .3275 .3097 .2937 .2793
7 .5878 .5404 .5000 .4653 .4350 .4085 .3850 .3641 .3453 .3283
6.b75b .6211 .5747 .5347 .5000 .4695 .4425 .4184 .3968 .3774
9 .7634 .7018 .6494 .b042 .5650 .5305 .5000 .4728 .4484 .4264
10 .b511 .7825 .7240 .6737 .6300 .5915 .5575 .5272 .5000 .4755
11 .9389 .b632 .7987 .7432 .6949 .6525 .6150 .5816 .5516 .5245
12 .9439 .o734 .8127 .7599 .7135 .6725 .b359 .6032 .5736
13 .9481 .8822 .8249 .7746 .7300 .6903 .6547 .6226
14 .9517 .8899 .8356 .7875 .7447 .7063 .6717
15 .9548 .8966 .8450 .7991 .7579 .7207
16 .9576 .9U25 .6535 .8095 .7696
17 .9600 .9078 .8610 .8188
16 .9622 .9126 .8678
19 .9642 .91b9
.9659
L[- t
TABLE II. TABLE of 5% RANKS k4 4

sample size a n "•-
j 1 2 3 4 5 6 7 8 9 10
1 .0500 .0253 .0170 .0127 .0102 .0085 .0074 .0065 .0057 .0051
2 .2236 .1354 .0976 .0764 .0629 .0534 .0468 .0410 .0368
3 .3684 .2486 .1893 .1532 .1287 .1111 .0978 .0873
4 .4729 .3426 .2713 .2253 .1929 .1688 .1500
5 .5493 .4182 .3413 .2892 .2514 .2224
6 .6070 .4793 .4003 .3449 .3035
7 .6518 .5293 .4504 .3934
8 .6877 .5709 .4931
9 .7169 .6058 :.....
10 .7411
sample size n
j 11 12 13 14 15 16 17 18 19 20
1 .0047 .0043 .0040 .0037 .0034 .0032 .0030 .0029 .0028 .0026
2 .0333 .0307 .0281 .0263 .0245 .0227 .0216 .0205 .0194 .0183
3 .0800 .0719 .0665 .0611 .0574 .0536 .0499 .0476 .0452 .0429
4 .1363 .1245 .1127 .1047 .0967 .0910 .0854 .0797 .0761 .0725 .- .
5 .2007 .1824 .1671 .1527 .1424 .1321 .1247 .1173 .1099 .1051z-
6 .2713 .2465 .2255 .2082 .1909 .1786 .1664 .1575 .1485 .1396
7 .3498 .3152 .2883 .2652 .2459 .2267 .2128 .1990 .1887 .1735
8 .4356 .3909 .3548 .3263 .3016 .2805 .2601 .2449 .2298 .2183
9 .5299 .4727 .4274 .3904 .3608 .3350 .3131 .2912 .2749 .2586
10 .6356 .5619 .5054 .4600 .4226 .3922 .3542 .3429 .3201 .3029
11 .7616 .6613 .5899 .5343 .4893 .4517 .4208 .3937 .3703 .3469
12 .7791 .6837 .6416 .5602 .5156 .4781 .4460 .4196 .3957
13 .7942 .7033 .6366 .5834 .5395 .5022 .4711 .4434
14 .8074 .7206 .6562 .6044 .5611 .5242 .4932
15 .8190 .7360 .6738 .6233 .5809 .5444
16 .8274 .7475 .6871 .6379 .5964
17 .8358 .7589 .7005 .6525
18 .8441 .7704 .7138
19 .8525 .7818
.8609
U ~ ~'7. 7' •# .. °,
11-2
TABLE III. TABLE OF 95% RANKS
sample size - n
1
j 2 3 4 5 6 7 8 9 10
1 .9500 .776b .6316 .5271 .4507 .3930 .3482 .3123 .2831 .2589
2 .9747 .8646 .7514 .6574 .5818 .5207 .4707 .4291 .3942
3 .9830 .9024 .8107 .7287 .6587 .5997 .5496 .5069
4 .9873 .9236 .8468 .7747 .7108 .6551 .6056
5 .9898 .9371 .8713 .8071 .7486 .6965
6 .9915 .9466 .8889 .8312 .7776
7 .9926 .9532 .9032 .8500
8 .9935 .9590 .9127
9 .9943 .9632
10 .9949
sample size n
j 11 12 13 14 15 16 17 18 19 20
1 .2384 .2209 .2058 .1926 .1810 .1726 .1642 .1559 .1475 .1391
2 .3644 .3387 .3163 .2967 .2794 .2640 .2525 .2411 .2296 .2182
3 .4701 .4381 .4101 .3854 .3634 .3438 .3262 .3129 .2995 .2862 -.
4 .5644 .5273 .4946 .4657 .4398 .4166 .3956 .3767 .3621 .3475
5 .6502 .6091 .5726 .5400 .5107 .4844 .4605 .4389 .4191 .4036
6 .7287 .6848 .6452 .6096 .5774 .5483 .5219 .4978 .4758 .4556 -"-
7 .7993 .7535 .7117 .6737 .6392 .6078 .5792 .5540 .5289 .5068
8 .8637 .8176 .7745 .7348 .6984 .6650 .6458 .6063 .5804 .5566
9 .9200 .8755 .8329 .7918 .7541 .7195 .6869 .6571 .6297 .6043
10 .9667 .9281 .8873 .8473 .8091 .7733 .7399 .7088 .6799 .6531
11 .9953 .9693 .9335 .8953 .8576 .8214 .7872 .7551 .7251 .6971 "
12 .9957 .9719 .9389 .9033 .8679 .8336 .8010 .7702 .7413
13 .9960 .9737 .9426 .9090 .8753 .8425 .8113 .7818
14 .9963 .9755 .9464 .9146 .8827 .8525 .8215
15 .9966 .9773 .9501 .9203 .8901 .8604
16 .9968 .9784 .9534 .9239 .8949
17 .9970 .9795 .9548 .9275
18 .9971 .9806 W971 ...
19 .9972 .9817
20 .9974
,.* .. .,
!I-3. .•
• ° . °.
~ * °,f-.'•*
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION
S/1 .5th 1st 2.5th 5th
.01 .974522 .976956 .980542 .983637

.02 .949602 .954342 .96137 .967447
.03 .925235 .932184 .942484 .951434
.04 .901413 .91045 .923886 .9356
.05 .87813 .889146 .905574 .919946 -.. :*
.06 .855381 .86827 .887551 .904476
.07 .833158 .847817 .869815 .889191
.08 .811453 .827785 .852368 .874093
.09 .790261 .808169 .835208 .859184
.1 .769574 .788966 .818334 .844465
.11 .749384 .77017 .801747 .829938
.12 .729683 .751778 .785445 .815604
,3 .710465 .733785 .769428 .801464
.14 .69172 .716186 .753693 .787519
.15 .673442 .698976 .73824 .773769 AL
.16 .655622 .68215 .723067 .760215
.17 .638252 .665703 .708172 .746857
.18 .621324 .649629 .693553 .733697
.19 .60483 .6333924 .679208 .720732
.2 .588761 .618581 .665135 .707965
.21 .57311 .603596 .651331 .695394
.22 .557868 .588961 .637794 .683019
.23 .543026 .574672 .624522 .670839
.24 .528577 .560723 .611511 .658855
.25 .514512 .547108 .598759 .647065 .. .
.26 .500823 .533821 .586263 .635468
.27 .487502 .520856 .57402 .624063
.28 .474541 .508207 .562026 .61285
.29 .461931 .495868 .550279 .601827
.3 .449665 .483834 .538775 .590992
.31 .437735 .472097 .527512 .580345
.32 .426132 .460653 .516484 .569884
.33 .414849 .449495 .50569 .559607
.34 .403879 .438618 .495125 .549513
.35 .393213 .428015 .484787 .539599
.36 .382844 .417681 .47467 .529865
.37 .372764 .407609 .464773 .520308
.38 .362967 .39779e .455091 .510926
.39 .353445 .388232 .445621 .501718
.4 .344191 .378914 .436358 .49268
.41 .335198 .369837 .4273 .483812
11-4
• '
- .-
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued)
S .5th lat 2.5th 5th
.42 .326459 .360994 .418442 .475111

.43 .317968 .35238 .409782 .466575
.44 .309717 .34399 .401315 .458201
.45 .301701 .335818 .393037 .449987
.46 .293912 .327858 .384946 .441932
.47 .286346 .320107 .377036 .434032
.48 .278994 .312558 .369306 .426286
.49 .271853 .305208 .361751 .418691 .
.5 .264915 .298049 .354368 .411244 . -
C-
=. *
I....... :
' ~...S.•
j .-. *..:
.. .. i~'2.
"~%
11-5
S/K 10th 20th 30th 40th
.01 .987217 .991569 .99472 .997419

.02 .974502 .983113 .989369 .994747
.03 .961858 .9764634 .983951 .991983
.04 .94929 .966135 .978468 .98913
.05 .936799 .957621 .972921 .986198
.06 .92439 .949092 .967313 .983159
.07 .912066 .940554 .961647 .980044
.08 .899829 .932009 .955925 .976846
.09 .887682 .92346 .95015 .973566
.1 .87563 .914911 .944325 .970205
.11 .863673 .906364 .938451 .966766
.12 .851815 .897823 .932531 .96325
.13 .840058 .889289 .926569 .959659
.14 .828405 .880768 .920567 .955995
.15 .816858 .87226 .914527 .95226
.16 .805419 .86377 .908451 .948456
.17 .794089 .8553 .902344 .944586
.18 .782872 .846852 .896207 .940651
.19 .771768 .83843 .890042 .936653
.2 .760779 .830035 .883853 .932594
.21 .749907 .821671 .877642 .928478
.22 .739154 .81334 .871412 .924305
.23 .728519 .805044 .865164 .920079
.24 .718006 .796786 .858903 .915801
.25 .707614 .788568 .852629 .911473 ..-..
.26 .697344 .780391 .846345 .907099
.27 .687198 .772259 .840055 .902679
.28 .677176 .764173 .83376 .898217
.29 .667278 .756135 .827462 .893714
.3 .657506 .748147 .821164 .889173 -
.31 .647859 .740211 .814868 .884596 %%
.32 .638337 .732328 .808576 .879986
.33 .628942 .7245 .80229 .875343
.34 .619673 .716728 .796013 .870671
,35 .61053 .709015 .789746 .865972
.36 .601512 .70136 .783492 .861247
.37 .59262 .693767 .777251 .856499
.38 .583854 .686235 .771027 .851729
.39 .575213 .678766 .76482 .846041
.4 .566696 .671362 .758633 .842,35r-..:-.-
.41 .558304 .664022 °752467 .837314 .
11-6
. . 11-6-. ." -. " .- .

TABLE IV. PERtCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued)
S/I .5th let 2.5th 5th
.42 .550034 .656748 .746324 .83248

.43 .541888 .649541 .740205 .827634
.44 .533864 .642401 .734112 .822779
.45 .525961 .635329 .728046 .817916
.46 .518179 .628327 .722009 .813046
.47 .510517 .621393 .716002 .808173
.48 .502973 .61453 .710026 .803296
.49 .495547 .607736 .704083 .798418
.5 .488238 .601014 .698173 .793541
U~ -Ai
1. -7
TAbLE IV. PERCENTILES OF TKM LOG-NURHA. DISTRIBUTION (Continued)
6I• 60th 70Oh OUrh 90th
.01 1.U0249 1.00521 1.0084 1.01285

.02 I.00488 1.01034 1.01677 1.02575
.03 1.00717 !.0154 1.0251 1.03872
-.04 1.00937 1.02037 1.0334 1.05174
.05 1.01148 1.02527 1.4165 1.0648
M06 1.01348 1.03008 1.04986 1.07791
.07 1.01539 1.03481 1.05802 1.09107
s08 1.01719 1.03945 1.06613 1.10425
.09 1.0189 1.04401 1.07418 1.1174b
.1 1.0205 1.04847 1.08218 1.13073
.11. 1.02201 1.05285 1.09012 1.1,44..i '"
.12 L.02341 1.05713 1.09799 1.1573 , ." .'. .
.13 1.024/2 1.06131 1.1058 1.17061
.14 1.02592 1.0654 1.11355 1.18393
.15 1.02702 1.069/4 1.12122 1.19726
. 1.02b03 1.0733 1.12882 1.2106
.7 1.02893 1.0771 1.13634 1.22394 *-•" -."
.1b 1.02973 1.0808 1.14378 1.23726
.19 1.03043 1.08439 1.15115 1.25058
.2 1.03104 1.087b9 1.15843 1.26389
.21 1.0U154 1.09129 1.10563 1.07717
.2;1 1.03195 1.0945b 1.17274 1.29U44
.23 1.03226 1.09776 1.1797b 1,368
.24 1.03247 1.L0087 1•18b69 1.31689
.25 1.03259 1.10385 1.19353 1.33007
.2o 1.03261 1.10673 1.20027 1.34321 9
.27 1.03254 1.10951 1.20692 1.5631
1.03238 1.11219 1.21347 1.36936
.29 1.03212 1.11476 1.21992 237 ....
.3 L.U3178 .L1•723 1.22627 1.39532
.31 1.03135 1.1196 1.23252 1.4062L
.32 1.03083 1.1216o 1.23467 .2185-"
.33 1.04022 1.12402 1.24471 1.4338-
.34 1.02953 1.12608 1,25065 1.44653
.35 1.02875 1.12804 1.25649 1.45917
.36 1.02789 1.1299 1.2b222 1.47174 . -..
.37 l.U2695 1.13L66 1,26784 1.46423
.3b 1.02594 1.133•32 1.27335 1.49b6b4
.39 l.U2484 1.1348b 1.2787b 1.5w$97
.4 1.02367 1.13634 1..240o 1.52122
.41 1.02242 1.1377t 1,28925 1.536 ...
.°. -.-.
I -,.:i ;-
j, .. • ,.
S/7 60th 70th 80th 90th
.42 1.02111 1.13898 1.29433 1.54545

.43 1.01972 1.14016 1.29931 1.55743
.44 1.01826 1.14124 1.30417 1.56931.
.45 1.1673 1.14224 1.30893 1.5811

.46 1.01514 1.14314 1.31358 1.5928
.47 1.01348 1.14395 1.31811 1.60439
.48 1.01176 1.14467 1.32215 1.61588
.49 1.00998 1.1453 1.32687 1.62726
.5 1.00814 1.14585 1.33108 1.63854
,.'... .%e
K±e ..
'4?4o
,
P J~o
*e
. ...-. .,
. . 4. 4 4. . . .4.
S .°° * .~.. .. o
. . ..... g..4..
TABLE IV. P6RCENTILES OF TH.L LOG-NORMAL DISTRIBUTION (Continued)
S/X 95th 97.5th 99th 99.5th
.01 1.01653 1.10974 1.02348 1•02604 ,

.02 1.03323 1.03977 1.04741 1.05265
.03 1.0501 1.06007 1.07178 1.07983
.04 1.06713 1.08066 1.0966 1.1076
.05 1.08431 1.10152 1.12187 1.13594
.Ob 1.10165 1.12265 1.14758 1.16488 " .
.07 1.11913 1.14406 1.17375 1.1944
.Ob 1.13677 1.16574 1.20036 1.22452
.09 1.15454 1.18769 1.22742 1.25524
.1 1.1724b 1.20989 1.25493 1.26055
.11 I.J905 1.23236 1.28289 1.3184b
.12 1.20868 1.25509 1.3113 1.351
.13 1.22698 1.27807 1.34015 1.38414
.14 1.2454 1.30129 1.36944 1.41788
.05 1.26594 1.32476 1.39918 1.45223
.lb 1.28258 1.34848 1.42936 1.4872
.17 1.30133 1.37242 1.45998 1.52277
.16 1.32019 1.3966 1.49103 L.55695
.19 1.33913 1.421 1.52251 1.59575
.2 1.35817 1.44563 1.55442 1.63315
.21 1.37729 1.47047 1.58676 1.67117
.22 1.3965 1.49552 1.61952 1.70978
.23 1.41577 1.52077 1.65269 1.74901 '"".
.24 1.43512 1.54623 1.68628 1.7•683
.25 1.45453 1.57188 1.72027 1.82926
.26 1.474 1.59771 1.75467 1.87028
.27 1.49352 1.62373 1.78946 1.91189
.28 1.51309 1.64992 1.82465 1.9541
.29 1.53271 1.67628 1.86022 1.99688
.3 1.55236 1.702b1 1.89617 2.04025
.31 1.57204 1.72949 1.93249 2.08419
.32 1.59175 1.75632 1.96918 2.12871
.33 1.61148 1.78329 2.00624 2.171799-
.34 1.63122 1.61041 2.04364 2.21942,-,.,'. . .
.35 1.65098 1.83765 2.08139 2.26561
.36 1.67074 1.86502 2.11946 2.31235
.37 1.69051 1.8925 2.15791 2.35962
.48 1.71027 1.9201 2.19666 2.40743
.39 1.73002 1.947b 2.23572 ý.45577
.4 1.74975 1.9756 2.1751 2.50462
.41 1.76947 2.00349 2.31478 2.5398
.l-lo
. ..
• , ., • . . -. -. - - .- ° -o. °%. °/•.°. , . " , .,. ° few.'°

.,, .. . , . . - , " .
SX 95th 97 .5th 99th 99.5th
.42 1.78916 2.03146 2.35475 2.60385 ~

.43 1.80883 2.05952 2.395 2.6542
1.82846 2.08764 2.43554 2.70505
.44
.45 1.84805 2.11583 2.47634 2.75637
.46 1.86761 2.14408 2.51741 2.80816
.47 1.88711 2.17238 2.55873 2.86041
.48 1.90657 2.20073 2.60029 2.91311
.49 1.92597 2.22912 2.64209 2.96626
1.94531 2.25754 2.68412 3.01983
.5
sp MAW. *
TA$Lk. IV. PERGELNTILES OF ThE LO'G-NOR11AL DISTRIBUTION~ (Continued)
SIX9 .5th let 2.5th 5th
.5 .264915 .296049 .354368 .411244

.51 .258175 .291079 .347152 .403944
.52 .251628 .284292 .340102 .396788
.53 .245268 .277683 .333213 .389774
.54 .239089 .271248 .326481 .382898 -
.55 .233086 .264983 .319904 .37616

.56 .227256 .258882 .313478 .369557
.57 .221.591 .252942 .3072 .363085 ..
.58 .216088 .247159 .301066 .356743

.59 .,210742 .241527 .295074 .350529
.6 .205548 .236044 .2d922 .44
.61 .200502 .230706 .283501 .338474
.62 .195599 .225507 .277914 .332629
.63 .190836 .220445 .27245b .326901
.64 .186207 .215517 .267124 .32129
.65 .18171 .210717 .261916 .315793
.66 .17734 .206043 .256827 .310407
.67 .173093 .201492 .251856 .305131
.66 .lb8966 .19706 .247 .299962
.69 .164956 .192743 .242256 .294898
.7 .1b1057 .188539 .237621 .289937
.71 .157268 .184444 .233G,92 .285077
.72 .153585 .180456 .228668 .280316
.73 .150005 .176571 .224346 .275652
.74 .146525 .172787 .220123 .271083
.75 .143141 .169101 .215997 .266606
.76 .139851 .16551 .211963 .2b2221
.77 .136653 .1b2011 .208026 .257924
.78 .133542 .158603 .204176 .253715
.79 .130517 .155281 .2U0414 .249591
.6.127575 .152045 .196738 .245551
.81 .124713 .148b91 .193146 .241593
.82 .12193 .145817 .169634 .237714
.83 .119222 .142821 .186203 .233914
.84 .116588 .139901 .182849 .230191 -
.65 .114025 .137055 .17957 .226542
.86 .111531 .134281 .176365 .222967
.87 .109104 .131576 .173232 .219464
.88 .106742 .128939 .17017 .216031
.89 W14443i .126368 .167175 .212666
.9 .10z205 .12386 .1b4248 .209369
11-12
11 Eli"
TASLE IV. PERCENTILESi OF THE LOG-NORMAL DISTRIBUTION (Continued)
S/X .5th Ist 2.5th 5th
.91 .100027 .121415 .161385 .206138

.92 .09790', .11903 .158586 .20297
.93 .095842 .116705 .155849 .199866
.94 .093831 .114436 .153172 .196823
.95 .091873 .112223 .150554 .19384 ~
.9b .089966 .110063 .147994 .190916
.97 .088108 .107957 .145489 .18805
.9b .086299 .105901 .143039 .18524 -
.99 .06453% .103895 .140642 .182485
1. .082819 .101938 .1.8297 .179783
".. ",
11-13
TABLý. IV. PERCENTiLEb OF THE LOG-NORM4AL DISTRIBUTION (Continued)
S10 toch 20th 30th 40th
.5 .488238 .601014 .696173 .793541

.51 .481044 .594362 •b92297 .788666
.52 .473965 .587781 .686457 .783794
3 .467 .581272 80654 .776927 .
.54 .4o0146 .574834 .674888 .774066
.55 .453404 .568468 .bb9161 .769213 *- .,
.56 .44t771 .5b2174 .663473 .764369 " "...-.

.57 .440247 .555951 b657824 .759535
.56 .43383 .5498 .652217 .754712
.•9 .427519 .54372 .64665 .749902
.6 .421313 .537712 .641126 .745106
.61 .41521 .331775 .635644 .740324
.62 .409209 .525908 .630205 .735559 '..
-.
.63 .403308 .520113 .b24809 .73081 "
.64 .397506 .514387 .619457 .726078
.65 .391803 .508732 .61415 .721366
.66 .381.96 .503147 .608807 .716673
.67 .360O63 .497631 .o0366ý, .712
.65 .375265 .492183 .598496 .707348
.69 .369939 .486805 .593368 .702719
.7 .3b4703 .481494 .586287 .698112
.71 .359557 .476251 .583251 .693528
.72 .354499 .471075 .578261 .68b968
.73 .349528 .4b5965 .•73317 .684433
.74 .344642 .460922 .56b419 .679923
•75 .33984 .455943 .563567 .075438 , *J.
.76 .J3512 .45103 .558762 .670979
.77 .33C482 .446181 .554003 .066548
.78 .325924 .441395 .549289 .662143 ...- *
•7ý .321444 .436673 .544622 .o57765

b .317U41 .432013 .540001 .653416 .%.. %
.81 .J12715 .427415 .535426 .b49094
.82 .308462 .422877 .530897 .644802
.83 .304284 .418401 .526413 .640537
84 .300177 .413984 .521975 .636303
.85 .296141 .409626 .517582 .632097
.86 .292174 .405327 .513234 .627921
s87 .288276 .401085 .508931 .b23775
.8b .264,.45 .396901 .504672 .619658
.69 .28066 .392773 .500458 .615572
.9 .276979 .388701 .496289 .611516
Li*°"
.'. °-1
11-14
s/X 10th 20th 30th 40th
.91 .273342 .384684 .492163 .607491 -:'," i

.92 .269767 .380721 .48808 .603495
.93 .266254 .376812 .484042 .599531
.94 .2628 .372956 .480046 .595597
.95 .259406 .369152 .476093 .591694
.96 .256069 .3654 .472182 .587821
.97 .25279 .361699 .468313 .583979
.98 .249566 .358048 .464487 .580168
.99 .246397 .354447 .460701 .576387
1. .243282 .350895 .456957 .572638
. ... "-t. i .m*
.- i. ;
.. •° .. .- °
l"'" 5
. . . . . . . . .
T1ABLE IV. PERCENTILES OF THE LUG-NORMAL DISTRIBUTION (Continued)
S60th 70th 80th 90th
.5 1.09814 1.14585 1.33108 l.b3854

.51 1.00b24 1.14631 1.33519 1.64972
.52 1.00429 1.14669 1.33919 1.b6078
.53 1.00228 1.14699 1.34309 1.67174
.54 1.00022 1.1472 1.34688 1.68258
.55 .998103 1.14734 1.35057 1.69331 .. -
.56 .99594 1.1474 1.35415 1.70393
.57 .993731 1.14738 1.35763 1.71443
.58 .9Y1475 1.14729 1.361 1.72482
.59 .989174 1.14712 1.36428 1.73509
. .986631 1.14688 1.30745 1.74524 ,.
.61 .98444b 1.4657 1.37052 1.7552b
.b2 .982021 1.14619 1.3735 1.1652 .
03 .979557 1.14574 1.37638 1.775
.64 .977057 1.14523 1.37916 1.7846o
.65 .974522 1.14405 1.38184 1.79424
b .97195s 1.14401 1.38443 1.80368
.67 .969351 1.14J31 1.J8693 1.813
.68 .9bo716 1.14254 1.36933 1.d222
.b9 .964056 1.14172 1.39165 1.83128
.7 .9b1365 1.14084 1.39387 1.S4024
.71 .958647 1.1399 1.39601 1.8490b
.72 .955904 1.13691 1.39605 1.8578"--'
.73 .953136 1.13787 1.40001 1.8664 ,.- -
.74 .950345 1.13677 1.40189 1.874886.
.75 .947532 1.13562 1.40368 1.88324 Ps,.
.76 .944699 1.13443 1.40539 1.8914a

. 77 .941846 1.13318 1.40702 1.8996
.78 .938974 1.13189 1.40b57 1.90761
.79 .936086 1.13055 1.41004 1.9155
.6 .933181 1.12917 1.41143 1.,92327
.b1 .930261 1.12775 1.41274 1.93092 "
.82 .927328 1.12629 1.41398 1.93b46
.83 .924381 1.12478 1.41515 1.94588
.•4 .921422 1.12324 1.41625 1.95319
.85 .918452 1.12166 1.41727 1.96U39
.86 .;3.472 1.12004 1.41622 1.96747
.87 .912483 1.11639 1.41911 1.97444
.66 .9U94b5 1.11671 1.41993 1.9813
.89 .90648 1.11499 1.42068 1.98805
.9 .90346b8 1.11321 1.42136 1.99468
. .. ...-.
. . . . .. . . . .. . .. . . . .
... . .
SIX 60th 70th 80th 90th
.91 .900451 1.11145 1.42199 2.00121

.92 .897428 1.10964 1.42255 2.00763
.93 .894401 1.1078 1.42305 2.01395
.94 .891371 1.10593 1.42349 2.02015
.95 .888337 1.10404 1.42387 2.02626
.96 .885302 1.10212 1.42419 2.03226 ~ ~
.97 .882265 1.10017 1.42446 2,03815
.98 .879227 1.0982 1.42467 2.04394.
.99 .876189 1.09621 1.42482 2.04964
1. .873151 1.09419 1.42493 2.05523 .,A.¶
-ALL.
11-17
SI 95th 97.5th 99th 99.5th
c1.94531 2.25754 2.68412 3.01983

.51 1.9646 2.28599 2.72636 3.07383
.52 1.98381 2.31446 2.76882 3.12824 Aw
.53 2.00296 2.34295 2.81148 3.18305 ;.\'*-'
.54 2.02203 2.37145 2.85433 3.23826
.55 2.04103 2.39995 2.89737 3.29385
.56 2.05994 2.42845 2.94059 3.34982
.57 2.07878 2.45694 2.98397 3.40615
.58 2.09752 2.48542 3.02752 3.46284
.59 2.11618 2.51389 3.07122 3.51987
.6 2.13475 2.54233 3.11506 3.57723
.61 2.15322 2.57074 3.15904 3.63492
.62 2.17159 2.59912 3.20315 3.69292
.63 2.18986 2.62747 3.24738 3.75123
.64 2.20803 2.65577 3.29172 3.80983
.65 2.2261 2.68402 3.33616 3.86872
.66 2.24406 2.71222 3.3807 3.92788
.67 2.26191 2.74036 3.42533 3.98731
.e2.27964 2.76845 3.47005 4.04699
.69 2.29727 2.79647 3.53.483 4.10692-
.7 2.31478 2.82442 3.55969 4.16708
.71 2.33217 2.85229 3.6046 4.22747
.72 2.34944 2.88009 3.64956 4.28807
.73 2.3666 2.90781 3.b9457 4.34889 ~ .
.74 2.38363 2.93545 3.73962 4.40989
.75 2.40054 2.963 3.78471 4.47109
.76 2.41733 2.99046 3.82981 4.53247
.77 2.43399 3.01782 3.87494 4.59401
.78 2.45052 3.04509 3.92008 4.65572
.79 2.46693 3.07226 3.96522 4.71758
.8 2.48321 3.09932 4.01036 4.77958
.81 2.49936 3.12628 4.0555 4.841712
.82 2.51538 3.15313 4.10063 4.90398
.83 2.53127 3.17987 4.14574 4.96636
.84 2.54703 3.20649 4.19082 5.02885
.85 2.56266 3.233 4.23588 5.09144
.86 2.57815 3.25939 4.2809 5.15412
.87 2.'9351 3.28566 4.32588 5.21689
.o2.60874 3.31181 4.37082 5.27973
.89 2.62384 3.33783 4,41571 5.34265
.9 2.63881 3.36372 4.46054 5.40562
11-18
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued) ,.
S/X 95th 97.5th 99th 99.5th ..
.91 2.65364 3.38949 4.50532 546865. .

".92 2.66834 3.41513 4.55003 5.53173 .y"
.93 2.6829 3.44064 4.59468 5.59485 r
.94 2.69733 3.46601 4.63925 5.658
.95 2.71163 3.49125 4.68374 5.72118
".96 2.72579 3.51636 4.72816 5.78438
.97 2.73982 3.54132 4.77248 5.84759
.98 2.75372 3.56615 4.81672 5.9108
.99 2.76749 3.59084 4.86087 5.97402
1. 2.78112 3.61539 4.90493 6.03723 -
.*.-.-
.'-. '~*..-.... ., .
... "--~~. ".'. ..
i .~~~~...' -'....
:: ,•.t..1'-
-11
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued) h
S/X .5th 1st 2.5th 5th
1. .082319 .101938 .138297 .179783

1.01 .081145 .100027 .136003 .177134
1.02 .079515 .098162 .133758 .174537
1.03 .077925 .096342 .131561 .17199
1.04 .076376 .094564 .129412 .169492
1.05 .074866 .092829 .127308 .167043
1.06 .073393 .091135 .125249 .16464
1.07 .071958 .08948 .123233 .162283
1.08 .070558 .087864 .12126 .159972
1.09 .0691.93 .086286 .119329 .157704
1.1 .067862 .084744 .117438 .15548
1.11 .066563 .083238 .115586 .153297
1.12 .065296 .081767 .113773 .151156
1.13 .06406 .080329 .111998 .149055
1.14 .062855 .078924 .110259 .146993 -.
1.15 .061678 .077551 .108556 .14497
1.16 .06053 .076209 .106889 .142985
1.17 .059409 .074898 .105255 .141036
1.18 .058315 .073616 .103654 .139124
1.19 .057247 .072363 .102086 .137247
1.2 .056204 .071137 .10055 .135404
1.21 .055186 .069939 .099045 .133595
1.22 .054192 .068768 .097569 .131819
1.23 .053221 .067622 .096124 .130076
1.24 .052272 .066501 .094707 .128364
1.25 .051346 .065405 .093318 .126683
1.26 .050441 .064332 .091957 .125032
1.27 .049557 .063283 .090622 .123411
1.28 .048693 .062257 .089314 .121818
1.29 .047849 .061252 .088031 .120255
1.3 .047024 .060269 .086773 .118718
1.31 .046217 .059307 .085539 .11721
1.32 .045429 .058366 .08433 .115727 A
1.33 .044658 .057444 .083143 .114271
1.34 .043905 .056542 .08198 .11284
1.35 .043168 .055658 .080838 .111434
1.36 .042447 .054793 .079718 .110053
1.37 .041743 .053946 .07862 .108695
1.38 .041053 .053116 .077542 .107361
1.39 .040379 .052304 .076484 .106049
1.4 .03972 .051508 .075446 .10476
11-20
"TABLE IV. PERCENTILES OF THEL LOG-NORMAL DISTRIBUTION (Continued)
S/ .5th 1st 2.5th 5th 1. .. ..
"1.41 .039074 .050728 .074427 .103493 .. ." '.

1.42 .038443 .049965 .073428 .102248
1.43 .037825 .049216 .072446 .101024
S1.44 .03722 .048483 .071483 .09982
"1.45 .036628 .047764 .070537 .098636
"1.46 .036049 .04706 .069609 .097472
"1.47 .035482 .04637 .068698 .096328
1.48 .034926 .045694 .067803 .095203
S1.49 .034383 .04503 .066924 .094096
1.5 .03385 .01438 .066061 .093007
AI
* . * - ... -
it
11-21
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued)" ,
SIX 10th 20th 30th 40th

, .% . -
1. .243282 .350895 .456957 .572638

1.01 .240219 .34739 .453253 .568918 . '
1.03
1.02 .234248
.237208 .343934
.340525 .44959
.445967 .56523
.561572
. 1.04 .231338 .337161 .442383 .557944 " , "
1.05 .228476 .333844 .438839 .554347
1.06 .225663 .330571 .435334 .55078
1.07 .222896 .327343 .431867 .547243
1.08 .220176 .324158 .428438 .543737
1.09 .217501 .321017 .425047 .54026
1.1 .21487 .317918 .421694 .536813
1.11 .212283 .314861 .418377 .533396
1.12 .209739 .311845 .415097 .530008 . -.
1.13 .207237 .30887 .411854 .52665 " " "" "
1.14 .204776 .305935 .408646 .52332
1.15 .202355 .303039 .405474 .52002
1.16 .199974 .300182 .402336 .516749
1.17 .197632 .297363 .399234 .513507
1.18 .195327 .294582 .396166 .510293
1.19 .193061 .291839 .393131 .507107
1.2 .19083 .289131 .390131 .50395
1.21 1.88636 .28646 .387163 .500821
1.22 .186478 .283825 .384229 .497719
1.23 .184353 .281224 .381326 .494645
1.24 .1842263 .278658 .378456 .491599
1.25 .180206 .276125 .375618 .48858 ' -
1.26 .178182 .273626 .37281 .485587 .f6
1.27 .17619 .27116 .370034 .482622
1.28 .174229 .268726 .367288 .479683
1.29 .1723 .266325 .364573 .47677
1.3 .1704 .263954 .361887 .473884
1.31 .168531 .261615 .359231 .471023 i
1.32 .16669 .259306 .356604 .468188
1.33 .164879 .257027 .354006 .465379
1.34 .163095 .254777 .351436 .462595
1.35 .161339 .252557 .348894 .459836
1.36 .15961 .250365 .34638 .457102
1.37 .157907 .248201 .343893 .454393 - " -
1.38 .15623 .246065 .341434 .451708
1.39 .15458 .243957 .339001 .449047
1.4 .152954 .241875 .336595 .44641
1-'. -.
-..- ".-..22.
S.
. .-
j
"TABLE IV. PERCENTILES OF THE LOG-NORKAL DISTRIBUTION (Continued)
•/x 10th 20th 30th 40th

...... •,•,
1.41 .151353 .23982 .334214 .443797 ,. •

1.42 .149776 .237791 .331859 .441208
1.43 .148222 .235788 .32953 .438642
"1.44 .146692 .23381 .327226 .436098 .
1.45 .145185 .231857 .324946 .433578
1.46 .143701 .229929 .322691 .431081
1.47 .142238 .228024 .320461 .428606
"1.48 .140797 .226144 .318254 .426153
1.49 .139377 .224286 .31607 .423722

1.5 .137979 .222452 .31391 .421313.
-1.--.-.
___.__-".__ _"__"___ _'
.... . .. *-. . . . . .. . . . . . . .
• . .o
.,. . .... .-. .. .. .

.. . ... .......
.... .- . . . - . .. . . . .... . . . . . .: . .•:.::. .:.
. . . . ..:.. . •..-:._ . .:.:::_.."
. .•_ ::. -:•: :/....
--.. -
"TitBL6. IV. PERCE&NTILES OF THE LOG-NORMAL DISTRIBUTION (Continued) -'-,_"-
SiX 6Uth 70th 80th 90th ,,.',•
1. .073151 •09419
1. 1.42493 2.05523 ,,,,%',
•I ...
1.01 .670115 1,09216 1.42498 2.06072 4.-.
1.0o2 .600 S:::*.. .0901

1o 1.4•2498 2.0o6611 _•••
:
1.03 a64046 1.•08803 1.42493 2.07141".:.-- ;
I.04
.05
.8•61016
.857989
08593
1. 1.4248
1.42469
2.U7661
2.08172..
.:',.-.---,
-08382
1.06 .654965 1.08169 1.4245 2.0Odb73.,-,-..-"
1.0U .6•5194b 1.07955 1.42426 2.09165-"" "" ":,':-"
i.0d .844093 1.077.,9 1.423]98 2.09648 . .
l.Ug .04592 1.07521 1.42365 2. 10122. . .
1.1 .642916b 1.07303 1.4232b 2.•10586 " ' ".
1.. .8-.1 1.22b .104 .008
1.14 .830959 1.06414 1.42141 2.12358 :t_
1.0 .827987 1.06189 1.42084 2.012779
1.16 .825022 1.05963 1.42024 2.13192
1.17 .822065 1.05737 1.4196 2.13597
1.0I4 .819117 1.05509 1.41892 2.13994
1.19 .816177 1.0528 1.41821 2.14383
1.2 .813246 1.05051 1.41747 2.14764
1.21 .810324 1.04821 1.4167 2.15137
1.22 .80419 1.0779 1.4138 2.09648
1.23 .804502 1.04358 1.41505 2.1586 4
1.24 .801614 1.04123 1.41418 2.16251
1.125 .79973 1.03894 1.41328 2.16553
1.26 .785657 1.03661 1.41236 2.11889
1.27 .792994 1.03427 1.4114 2.17217
1.2 .8790141 1.03193 1.41042 2.17539
I.19 .787299 1.02959 1.40941 2.17953
1.3 .704468 1.02724 1.4U838 2.1816
i.31 .781646 1.0249 1.40731 2.18461
1.32 .77b839 1.02255 1.40623 2.13755
i.19 .771042 1,02019 1.4U512 2.19442
1.34 .•77254 1.01754 1.41399 2.19423
1.35 .177043 1.01548 1.40283 2.19597
i.32 .70718 1.01312 1.40166 2.19865 . . ,
1.257 .7b4967 1.01077 1.40046 2.21627
1.36 .7oZ227 1.03641 1.39924 .2082" -
1.39 .759529 1.02607 1.398 2.20632
1.4 .750184 1.00369 1.39074 2.12075"9
•-, :. %.-...,..•
_.:i.,..3 .:" "
.. "•" ..... "'-,..." " :""-"1n.02724n' "1.4u" "83"8'-":: 2""".1816;1' ::; 1:: - '
11-24
1.49 1.02959 1.4941 2.1an5

SIrx 60th 70th 80th 90th -
1.41 .754083 1.00134 1.39546 2.21113

1.42 .751393 .998978 1.39416 2.21344
1.43 .748715 .996623 1.39285 2.2157
1.44 .746049 .99427 1,39152 2.21791
1.45 .743396 .991917 1.39017 2.22006
1.46 .740755 .989567 1.3888 2.22215
1.47 .738126 .987219 1.38742 2.22419 .
1.48 .735511 .984874 1.38602 2.22618

1.49 .732907 .982531 1.38461 2.22811
1.5 .730317 .98009 1.38318 2.22999
11-25
..................................
..................
TABLE IV. PERChNTILES OF THE LOG-NORMAL DISTRIBUTION (Continued)
S/x 95rh 97.5th 99th 99.5th
1. 2.76112 3.61539 4.90493 6.03723

1.01 2.79462 3.6398 4.94888 6.10043
1.02 2.80799 3.66407 4.99273 6.16361
1.03 2.82123 3.68819 5.03647 6.22677
1.04 2.83434 3.71217 5.0801 6.2899
1.05 2.84731 3.736 5.12363 6.35299
1.06 2.8b016 3.75969 5.16703 6.41605
1.07 2.87288 3.78324 5.21032 6.47906
I.08 2.bb547 3.80664 5.25348 6.54201 [-I
1.09 2.89793 3.82989 5.29653 6.60492
1.1 2.91027 3.65299 5.33944 6.b6776
1.11 2.92247 3.87595 5.38223 6.73053
1.12 £.93456 3.8987b 5.42489 6.79324
1.13 2.94651 3.92143 5.46741 6.85587
1.14 2.95834 3.94394 5.50979 6.91843
1.15 2.97005 3.96631 5.55204 6.9809
1.16 2.98163 3.98851 5.5V415 7.03426
"1.17 2.9931 4.0106 5.63612 7.1055a
1.16 3.U0443 4.03252 5.07794 7.16778
1.19 3.01565 4.0543 5.71962 7.22988
1.2 3.02b75 4.07593 5.7b1i5 7.29187
1.21 3.03773 4.09741 5.80253 7.35376
1.22 3.04859 4.11874 5.84376 7.41554
1.23 3.05934 4.13992 5.88484 7.47721
1.24 3.06996 4.16096 5.92577 7.53876
i.25 3.08047 4.18185 5.96654 7.6002
1.26 3.09087 4.20259 6.00716 7.66151
1.27 3.10115 4.22319 6.04762 7.72269
1.28 3.11132 4.24364 6.O8792 7.76375
1.29 3.12138 4.26395 6.12606 7.84467
1.3 3. 13132 4.28411 6.lb0805 7.90546 - ..
1.31 3.14116 4.30413 6.320787 7.9662
"1.32 3.15088 4.324 6.24753 8.26732
1.33 3.91605 4.34373 6.28702 8.08701
1.34 3. 17001 4.3633 1 t.3263b b.14724..."
1.3.5 3.17941 4.38276 6.36552 8.20732 A.
1.3b 3. 1887 4.40206 6.40453 8. 2672 5 ":',
1.37 3.•19789 4.42122 6.44337 8.32704 -- "
1.3b J.Z2069b 4.44023 6.46204 8.36067
1.39 3.21596 4.45911 6.52054 8.44615
1.4 3.22485 4.47785 6.55688 8.50547 "'.
11-26
7•- 7-'
-". .. .... .......- ,.-.-.-...-." . .'.- -.. '.... ..' '.._-.. . •, .... ._ . -'. 'J' •,:-''"" .'' A '"
TAbLk, IV. PERKCbILES OF THE LOG-NORMAL DISTRIBUTION (Continued)
S/x 95th 97.5th 99th 99.5th
1.41 3.23363 4.49645 6.59705 8.56464

1.42 3.24231 4.51491 6.63506 8.62364
1.43 3.25089 4.53323 6.67289 8.68248
1.44 3.25937 4.55142 6.71056 8.74116
1.45 3.20775 4.56947 b.74806 8.7996b
"1.46 3.27604 4.58738 6.78539 8.85802
1.47 3.2b423 4.6051b 6.82255 8.9162
1.48 3.29233 4.6228 6.85954 8.97422 -
1.49 3.30034 4.64032 6.69636 9.03206
1.5 3.30825 4.65769 6.93302 9.08973
|11-2
,.''
. S **. ,
S **S~* .- •°.
. -SS * - S .
SS/-X .5th lst 2.5th 5th
1.5 .03385 .04438 .066061 .093007

1.51 .033329 .043743 .065213 .091936
1.52 .032818 .043118 .06438 .090883
1.53 .032318 .042505 .063562 .089847
1.54 .031828 .041904 o062759 .088828
1.55 .031347 .041314 .06197 .087825
"1.56 .030877 .040736 .061194 .086838
"1.57 .030416 .040169 .060432 .085867
1.58 .029964 .039612 .059684 .084911 L4
1.59 .029522 .039066 .058948 .083971
1.6 .029088 .03853 .058225 .083046 "
1.61 .028662 .038005 .057514 .082135
1.62 .028246 .037489 .056816 .081238
1.63 .027837 .036982 .056129 .080356
1.64 .027436 .036485 .055454 .079487
1.65 .027043 .035998 .05479 .078632
1.67 .026279 .035048 .053496 .076961

1.68 .025909 .034587 .052865 .076145
1.69 .025545 .034134 .052245 .075341
1.7 .025188 .033688 .051635 .074549
"•1.66
1.71
.026650
.024838
.035179
.033251
.054187
.051034
.0707679•.:'
.073769
1.72 .024495 .032822 .050444 .073002
1.73 .024158 .0324 .049863 .072245
1.74 .023827 .031986 .049292 .0715 ".
"1.75 .023503 .031579 .04873 .070767
1.76 .023184 .031179 .048177 .070044
1.77 .022871 .030787 .047633 .069331
1.78 .022565 .030401 .047098 .06863
1.79 .022263 .030021 0046571 .067938
"1.8 .021967 .029649 .046053 .067257
1.81 .021677 .029282 .045542 .066586
"1.82 .021392 .028922 .04504 .065924
j 1.83 .021112 .028568 .044546 .065272
1.84 .020837 .02822 .044059 .064629
1.85 .020566 .027878 .04358 .063996
1.86 .020301 .027542 .043108 .063372
1.87 .02004 .027212 .042644 .062756
1.88 .019784 .026886 .042187 .062149
1.89 .019532 .026567 .041737 .061551
1.9 .019285 .026252 .041293 .060962
!1-28
,L. ao-..:..
TABLE IV. PERCENTILES OF THE LOG-NORMIAL DISTRIBUTION (Continued)
SIX .5th let 2.5th 5thk
1.91. .019042 .025943 .040856 .06038

1.92 .018803 .025639 .040426 .059807
1.93 .018569 .02534 .040003 .0.59242
1.94 .018338 .025045 .039586 .058684
1.95 .016111 .024756 .039175 .058134
1.96 .017889 .024471 .03877 .057592
1.97 .017669 .024191 .038371 .057057
1.98 .017454 .023915 .037978 .05653
1.99 .017242 .023643 .03759 .056009
2. .017034 .023376 .037209 .055496 ~ ~*
AW
Y~lft
r 74W
11-29.
TABLE IV. FERCEMTILLS OF ThE LOG-NORKAL DISTRIBUTION (Continuea)
10th
1/X 20th 30th 40th h.
1.5 .137979 .222452 .31391 .421313
mu
1.51 .1366 .220641 .311773 .418925
1.52 .135242 .218852 .309658 .416559
1.53 .133903 .217085 .307566 .414214
S1.54 .132584 .21534 .305496 .41189
1.55 .131284 .213616 .303448 .4u9586
1.56 .130003 .Z11913 .301421 .407304
1.57 .12874 .210231 .299416 .405041
1.58 .127495 .20857 .297431 .402799
1.59 .126267 .206928 .295468 .400576
1.6 .125058 .205307 .293524 .398373
1.61 .123865 .203704 .291601 .39619
1.62 .122689 .202121 .289698 .394026
1.3 .121529 .200557 .287815 .391881
1.64 .120396 .199012 .285951 .389755
1.65 .119259 .197485 .284106 .38748.
1.66 .118147 .195976 .2b228 .385559
1.07 .11705 .194485 .280473 .383489
1.66 .115969 .193011 .276684 .381437 .-
1.09 .114903 .191555 .276913 .379402
1.7 .113851 .190116 .275161 .377386
1.71 .112813 .188693 .273426 .375387
1.72 .11179 .187287 .271709 .373405 '• '
1.74 .1097b4 .184524 .266327 .309493
1.75 .108802 .183167 .2b6661 .367563 •€-
1.7b .107832 • 181ts24 •.265012 .3b564 9
1.77 .106876 .180598 .263379 .363752
1.78 l5932 .17918b .261762 .361871,
1.79 .105001 .177889 .260162 .360006

1.8 .104081 .176607 .258577 .356157
i1.t$
1.81
1.83
.103174
.104279
.101396
.17534
.174086b
.172847
.257009
.255455
.253917
.356324
.354506
.352704
L1;
1.84 .100524 .171622 .252394 .350918
1.85 .099663 .17041 .25088o .349146
1.86 .098814 .169212 .249392 .34739
1.67 .097975 .168027 .247913 .345648
1.88 U097147 .166855 .246448 .343921
1.b9 .09633 .16569o .244998 .342208
1.9 .095523 .16455 .243561 .34051 --
5...
11-30 "''
, - ° •.-
~~~~~~~~~~~."........
. o-......... . . . -. . ... .. oo.......................'.%...... . o.- ..... .o... -... ,.o.*°.°....,.'°....-. - °.' -°. .o.'
"..'...'-..
'-....-,"......'... ".. "; "...':,~~~.... .'.." ......
"•..•-% "... ;....." ."....-. .. .... " °=...... ... •*, ..... ,
., 10th 20th 30th 40th
li91 .094726 .163416 .242138 .338826 iA. '

1.92 .093939 .162295 .240729 .337156
.1.94 1.93 .093163
.092395 .161186
. 160089 .239334
.237951 .3355
,333858"
1.95 .091638 .159004 .236582 .33223

1.96 .09089 .15793 .235226 .330614
1.97 .090151 .156868 .233882 .329013
1.98 .089421 .155817 .232551 .327424
1.99 .088701 .154778 .231233 .325849
2.- .087989 .153749 .229927 .324286
.' .". . .". .
;.: ,•..11
. .
,-~. . . . . . .-...
~16
TABLE IV. PERCENTILES OF THE LOG-NORMAL DIST,18UTION (Continued)
S/X 6Uth 70Lh 80ch 90th
1.5 .730317 .98019 1.38318 2.22999

1.51 .727739 .977853 !.38:-'4 Z.23182
1.52 .7V5174 .975519 t. 34028 2.2336l
•1..5.3 .722b22 .973188 1.37881 225ý
1.54 .720082 .970861 1.37733 2.23702
1.55 .717556 .968538 1.37584 2.23866
""1.56 .715042 .966219 1.37433 2.24025 "":
1.57 .71254 .963904 1.37281 2.4241711
1.5b .71U052 .961593 1.37128 2.24329
"1.59 .707576 .959287 1.36974 2.24474
l.6 .705113 .956985 1.36819 2.24615 . - ,
1.01 .702663 .954688 1.36663 2.24751 ,-.
1.62 .700226 .952396 1.36506 2.24883
1.03 .697801 .950108 1.36348 2.25011 • '
1.64 .695389 .947826 1.36189 2.25135
i.b5 .69299 .94555 1.36029 2.25255
1.66 .690603 .943278 1.35868 2.2537
""1.7 .688229 .941012 1.35706 2.25482
"1.68 .685868 .938752 1.35544 2.2559 %
"1..69 .683519 .936497 1.35381 2.25694
1.7 .681183 .934249 1.35217 2.25794
1.71 .b7886 .932006 1.35052 2.2589
1.72 .676549 .929769 1.34887 2.25983 • " "
"1.77 .67425 .927538 1.34721 2.26072
1.74 .671964 .925314 1.34555 2.26158
1.75 .66969 .923095 1.34388 2.2624
1.76 .5o7429 .920883 1.3422 2.26539
1.77 .(65168 .918678 1.34052 2.26394
r-" 1. 78 .662943 .916478 1.33883 2.26465 "
"1.79 .660719 .914286 1.33714 2.26534
1.8 .658506 .91201 1.33544 2.26599
1.81 .656306 .90992 1.33374 2.26661
1.82
1i. 3 .654118
.651942 .907486 1.33203 2.2672"
.905582 1. 33032 2.26776 77 7
71 ~1.84 .64*977b •.903423 1.•32861 2. 26829 .. . :.
1.6•5 .647626 .9012 71 1.•32689 2.•26879 . ..
1.86 .645485 .899126 1.32517 2.26926 ::"-:.:"
Sl~~. 87 .643357 .896987 1.•32345 2.2697 :° ':'-":
1.6 b4.24 .89485b l.2172 2.2701.
1.69 .639135 b892732 1.31999 2.8705
1.9 .637042 .890615 1.31826 2.27086 ,
. •. S - .
.. ....-............
%.
" I1-32~~~ .- •".,..
....-.. . . .......... °. .. -..... . . .. °'°.-o*

4.0*
I%" *.% -,.
'ABL. LV. PERCENITILES OF THt. LOG-NORMAL UASTRI8UTIOh (Continue4)
6/
0th 70th 80th 90th
K:1.91 .64496 .888505 1.31652 2.27119

1.92 .63289 .886402 1.31478 2.27149
1.93 .b30832 .884306 1.31304 2.27177
1.94 .628784 .882218 1.3113 2.27202
1.95 .626749 .880137 1.30956 2.27224
1.9b .624724 .878063 1.30781 2.27244
1.97 .622711 .a75996 1.30606 2.L7262
S1.99 1.98 .620709
,bI6718 .873936
.871884 1.30432
1.302.57 2.27277
2,L729
2. .616738 .869889 1.30081 2.273
Li
• . "-:-.:-
. .o
11-33
S/! 95th 97.5th 99th 99.5th .'

1.5 3.30825 4.65769 6.93302 9.08973
1.51 3.31607 4.67494 6.9695 9.14723

1.52 3.32379 4.69205 7.00581 9.20455
1.53 3.33143 4.70904 7.04196 9.2617
1.54 3.33898 4.72589 7.07793 9.31867
1.55 3.34644 4.74262 7.11374 9.37547
1.56 3.35381 4.75922 7.14938 9.43209
1.57 3.36109 4.77569 7.18484 9.48853
1.58 3.36829 4.79203 7.22014 9.54479
1.59 3.37541 4.80825 7.25527 9.60087
1.6 3.38244 4.82434 7.29024 9.65677
1.61 3.38939 4.84031 7.32503 9.71249
1.62 3.39625 4.85616 7.35966 9.76802
1.63 3.40303 4.87188 7.39412 9.82338
1.64 3.40973 4.88748 7.42841 9.87855
1.65 3.41636 4.90296 7.46253 9.93353
1.66 3.4229 4.91832 7.49649 9.98834
1.67 3.42937 4.93356 7.53028 10.043
1.68 3.43575 4.94869 7.56391 10.0974
1.69 3.44206 4.96369 7.59737 10.1516
1.7 3.4483 4.97858 7.63067 10.2057
!.71 3.45446 4.99335 7.6638 10.2596
1.72 3.46055 5.00801 7.69677 10.3133
1.73 3.46656 5.02255 7.72957 10.3668
1.74 3.4725 5.03698 7.76221 10.4203
1.75 3.47837 5.0513 7.79469 10.4732
1.76 3.48417 5.06551 7.82701 10.5262
1.77 3.48989 5.0796 7.85916 10.5789
1.78 3.49555 5.09359 7.89115 10.6315
1.79 3.50114 5.10746 7.92299 10.6839
1.8 3.50666 5.12123 7.95466 10.7361
1.81 3.51211 5.13489 7.98618 10.7881
1.82 3.5175 5.14844 8.01753 10.8399
1.83 3.52282 5.16188 8.04873 10.8915
1.84 3.52807 5.17522 8.07977 10.943 "
1.85 3.53326 5.18846 8.11065 10.9942
1.86 3.53839 5.20159 8.14138 11.0453
1.87 3.54345 5.21462 8.17195 11.0962
1.88 3.54845 5.22755 8.20236 11.1469
1.89 3.55339 5.24037 8.23262 11.1974
1.9 3.55827 5.2531 8.26273 11.2477
." ... -
~.. .".
11-34 . "o.
TABLE IV. PERCENTILES OF THE LOG-HORM&L DISTRIBUTION (Continued)
S/x 95th 97.5th 99th 99.5th
1.91 3.56309 5.26573 8.29268 11.2979 b.

1.92 3.56785 5.27825 8.32249 11.3478
1.93 3.57254 5.29068 8.35214 11.3976
1.94 3.57718 5.30301 8.38163 11.4472
1.95 3.58177 5.31525 8.41098 11.4966
1.96 3.58629 5.32738 8.44018 11.5458
1.97 3.59076 5.33943 8.46923 11.5948 V.
1.98 3.59517 5.35138 8.49813 11.6437
1.99 3.59953 5.36323 8.52688 11.6923
2. 3.60383 5.37499 8.55549 11.7408
1'. -". %*•!•,,•
.g[..-,,
........ •..:
17
i '5
., °.o....-o
**- ...
11-35
""X..
TABLE IV. PERCINTILES OF THE LOG-NORMAL DISTRIBUTION (Continued)
S/X .5th 1st 2.5th 5th

p
2. .017034 .023376 .037209 .055496
2.01 .016829 .023113 .036833 .05499
2.02 .016628 .022855 .036462 .05449
2.0J .01643 .0226 .036097 .053997
2.04 .01b235 .022349 .035737 .053511
2.05 .016044 .022103 .035382 .053031
2.06 .015855 .02186 .035033 .052558
2.07 .01567 .02162 .034688 .05209
2.06 .015487 .021385 .034348 .•%1629
2.•9 .v15308 .021153 .034013 .051174
i.1 .015131 .020924 .033683 .0U50725
2.l1 .014957 .0207 .033357 .050282
"2.12 .01478b .020478 .033036 .U49o44
2.13 .014618 .0202b .032719 .049413
2.14 .014452 .020045 .032407 .04*986
2.15 .014289 .019833 .032099 .048565
"2.16 .014129 .019625 .031795 .04815
2.17 .013971 .019419 .031496 U04774
2.10 •013815 .019217 .0312 .047335
"2.19 .013662 .019017 .030909 .046935
2.2 .014511 .018821 .030621 .04654
2.21 .013363 ,018627 .030338 .04615
"2.22 .013217 .018436 .030058 .045765
2.23 .013073 .018248 .029782 .045384
2.24 .012931 .018063 .02951 .045009
"2.25 .012791 .01788 .029241 .044638 "
"2.26 .012654 .0177 .028976 .044272
2.27 .012518 .017523 .028714 .04391
2.28 .012385 .017346 .028456 .043552
2.29 .012254 .017175 .028201 .043199
2.3 .012124 .017005 .027949 .04285
2.31 .011996 .016838 .027701 .042506
2.32 .011671 .016672 .027456 .U42165
2.43 .011747 .016509 .027214 .041829
Z.34 .011625 .016348 .02U6975 .041496
2.35 .011505 .01619 .026739 .041168
2.36 .011386 .016034 .026507 .040843
2.J7 .011269 .015879 .026277 .040523
2.36 .011154 •015727 .02605 .04020b
2.39 .01104 .015577 .025826 .039b92
2.4 .010929 015429 .025605 .039583
. . . . .. . . . .. . . . .. . . . .. . . . . . . . . .
•11..-....
. . . . . . .. '. - . . .r.
TikBLhr IV. PLRCENTLLES OF THE LUG-NOR14AL DISTRIBUTION (Continuea)
S/31 .5th lot 2.5th 5th -
2.41 .0108b1 .0152b3 .025386 .0M9277

2.42 .010709 .015139 .025171 .038974
2.43 .010602 .014997 .024958 .036676
2.44 .010497 .014857 .024747 .03838
2.45 .010392 .014719 .02454 .038088 ,..
2.46 .01029 .014582 .024334 .037799
2.47 .010188 .0144468 .024132 .037514 •..-
2.48 .0l008s .014315 .023931 .037232
2.49 .00999 .014184 .02J734 U36953
2.5 .009892 .014054 .023536 .036677
.- .•.-..'.. c'.:
,'.............-...
" .. . . .
11-37
' o , o.
TABLt. IV. F(eCLNTIL Of THt. LO-NORBAL DISTRIBUTION

0, ((.ontinued)
SIX 10th 20th 30th 4Uth i- "
2. .087989 .153749 .229927 .324286

2.01 .087286 .152732 .228633 .322736
2.02 .U86591 .151725 .227351 .321199
2.03 .085905 .150729 .226081 .319674
2.04 .085227 .149743 .224823 .318162
2.05 .084557 .148767 .223577 .316662 ,
2.0b .08389b .147802 .222342 .315174
Y.07 .083242 .146847 .221118 .313698
2.08 .U8259 .145901 .219906 .312234
2.09 .081957 .144966 .218704 .310782
i.1 ,0b1327 .14404 .217514 .309341
2.11 .060703 .143123 .216334 .307912
1.12 .080087 .142216 .215165 .306494
2.13 .079478 .141318 .21400b .305088
2.14 .078877 .140429 .212858 .303692
2.15 .0782t2 .139549 .211721 .302306
2.1b .077694 .138678 .210593 .300934
2.17 U077113 .137816 .209475 .299572
2.18 U07b539 .136962 .2u8368 .29822
2.19 .075971 .136117 .20727 .296879
2.2 .075409 .135281 .206182 .295548 -
2.21 .074854 .134452 .205103 .294227
2.22 .074306 .133632 .204034 .292917
2.23 .073763 .13282 .202974 .291o17 , .
2.24 .073227 .132016 .201924 .290327
2.25 .072697 .13122 .2008b3 .28904b
2.26 .072172 .130431 .19985 .28777b
2.27 U071654 .12965 .198827 .286516
2.2b .071141 .128877 .197812 .265265
2.29 .070634 .128112 .196807 .284023
2.3 .070132 .127353 .19581 .282791
2.31 .069636 J126602 .194821 .281569
2.32 .0b9145 .125858 .193841 .2o0355
2.33 .06866 .125121 .192869 .279151
Z.44 .06b18 .124392 .191905 .27795b
2.35 .o67705 .123669 .19095 .27677
2.36 .067236 .122953 .190002 .275593 .' .
Z.37 .066771 122244 .189063 .274424 . ..
2.36 .0bb31i .121541 .lb131 .27J265
2.39 .065857 .120845 .187207 .272114
2.4 .065407 .120155 .186291 .270971
. .......... ,
. .. .. .. . . . . . . . . . . . . . .
. .. .. °°
.- . . .
TABLE IV. PERCENTILES OF THE LOG-NORMAL DISTRIBUTION (Continued) -
S/x 10th 20th 30th 40th y'•
2.41 .064962 .119472 .185383 .269837

2.42 .064521 .118796 .184482 .268711
2.43 .064085 .118125 .183588 .267594
2.44 .063654 .117461 • .182702 .266484
2.45 .063228 .116802 .181823 .265383
2.46 .062885 .11615 .180952 .26429 .'-.*-*.',$
2.47 .062388 .115504 .180087 .263205
2.48 .061974 .114864 .179229 .262128
2.49 .061565 .114229 .178379 .261058 -
2.5 .06116 .1136 .177535 .259996
.. '. .
.. ,.-,-.....
- ... ,. .
""* *,.v- .< ,' *.. -. - -- .- ,
. . . . . . . .. . . . . . . . . . . .
IW Downloaded from http://www.everyspec.com
S/X 60th 70th 80th 90th
2. .616738 .869839 1.30081 2.273

2.01 .614769 .867802 1.29906 2.27308
2.02 .612811 .865772 1.29731 2.27314
2.03 .610865 .863749 1.29556 2.27318
2.04 .608928 .861734 1.2938 2.27319
2.05 .607003 .859726 1.29205 2.27318
2.06 .605088 .857725 1.29029 2.27316
2.07 .603184 .855732 1.28854 2.27311
2.08 .601291 .853746 1.28678 2.27303
2.09 .599408 .851768 1.28503 2.27294
2.1 .597536 .849796 1.28327 2.27283
2.11 .595674 .847833 1.28152 2.2727
2.12 .593822 .845877 1.27977 2.27255
2.13 .591981 .843928 1.27801 2.27238
2.14 .59015 .841986 1.27626 2.2722
2.15 .588329 .840052 1.27451 2.27199
2.16 .586518 .838126 1.27275 2.27177
2.17 .584717 .836206 1.271 2.27153
2.18 .582926 .834294 1.2(325 2.27127 P.
2.19 .581145 .83239 1.2675 2.27099 -01.
2.2 .579373 .830492 1.26576 2.2707
2.21 .577612 .828603 1.26401 2.27039
2.22 .57556 .82672 1.26226 2.27006 :
2.23 .574118 .824845 1.26052 2.26972
2.24 .572386 .822977 1.25878 2.2693616
2.25 .570663 .821116 1.25704 2.26898
2.25 .568949 .819263 1.2553 2.26859
I2.27
2.28
.567245
.56555
.817417
.815578
1.25356
1.25182
2.26819
2.26777
2.29 .563865 .813746 1.25009 2.26734
2.3 .562189 .811922 1.24835 2.26689
2.31 .560522 .810105 1.24662 2.26642
2.32 .558864 .808295 1.24489 2.26595
2.33 .557215 .8064.2 1.24317 2.26546
2.34 .555575 .804696 1.24144 2.26495
2.35 .553944 .802907 1.23972 2.26444
2.36 .552322 .801126 1.238 2.26391
2.37 .550708 .799352 1.23628 2.26336
2.38 .549104 .797584 1.23457 2.26281
2.39 .547508 .795824 1.23285 2.26224
2.4 .54592 .794071 1.23114 2.26166
11-40
- ~~~~.
.. .. ......................... -. -. . ..-..-. '.'.--'-.-.-'......'. .--. "'-..
o...-.-.-..;--.-..----.....-..-.--..................... .......-...... ,-.. .• ...... .:. .,....,
________- - _-__-" __"____"__*-
-" " "•_" ,_____ -" . - "- • - . .. .. . ,.i. . . . . . - -" -- .. . . . ...
TABLE IV. PERCERTILES OF THE LOG-NORMEAL DISTRIBUTION (Continued)
S/,x 60th 70th both 90th
ý..41 . 44342 .792324 1.22943 2.26107

2.42 .542772 .790585 1.22773 2.26047
2.43 .54121 .788853 1.22602 2.26985
2.44 .539657 .787127 1.22432 2.25923
2.45
2.46
2.47
.53b112
.b36575
.535046
.785409
.783697
.781993
1.22262
1.22093
1.21924
2.25859
2.25794
2.25723
n
64. .
2.4b .533526 .780295 1.21754 2.25661
2.49 .532014 .778604 1.21586 2.25593
2.5 .53051 .776919 1.21417 2.25524
.- ...-...
II
* .S ,°
* . ' "
'.-..-
11-41
-o. .-
... ' .'.•.... -. .. -... .,-,. .. -.. - _.........-.. ..-. '.'-." .-- ,- ...- .. '..'-.- .-- ,.-
TABLh- IV. PERCENTILES OF THIL LOG-NORMAL DISTRIBUTIOh (Continued)
S/X 95th 97.5th 99th 99.5th
2. 3.b0383 5.37499 8.55549 11.7408

2.01 3.60808 5.38666 8.58395 11.7891
2.u2 3.61227 5.39824 8.61226 11.8372
2.03 3.61641 5.40973 8.64843 11.38529
2.04 3.6203 5.42093 8.66845 11.9329
*2.05 3.62454 5.43244 8.69633 11.9805
2.06 3.b2853 5.44366 8.72407 12.0279
2.07 3.63246 5.45479 8.75167 12.0751
2.08 3.63634 5.46583 8.77912 12.1221 L
2.09 3.b4018 5.47679 8.80643 12.169
2 .1 3.•64397 5.48767 8. 83361 12.21]57 -""'.
2.11 3.6477 5.49846 8.86064 12.2622
2.12 3.0139 5.50986 8.88754 12.3085
2.13 3.65503 5.51978 8.91429 12.3546

2.14 3.b5863 5.53032 8.94091 12.4006
2.15 3.66218 5.54077 8.9674 12.4464
2.1b 3.6b568 5.55114 8.99374 12.492
2.17 3.66914 5.56144 9.01996 12.5374
2.1w 3.67255 5.57165 9.04604 12.5827
2.19 3.67591 5.58178 9.07198 12.6278
2.2 3.67924 5.59184 9.09779 12.6727
2.21 3.68252 5.60181 9.12347 12.7174
2.22 3.60575 5.61171 9.14902 12.762
2.23 3.68895 5.62153 9.17444 12.8064
2.24 3.6921 5.63127 9.19972 12.6506
2.25 3.69521 5 64094 9.22488 12.8947
2.26 3.09828 5.65053 9.24991 12.9386
2.27 3.70131 5.66005 9.27481 12.9823
2.28 3.70429 5.6695 9.29959 L3.0258
2.29 3.70724 5.67887 9.32423 13.0692
2.3 3.71015 5.68817 9.34875 13.1124
2.31 3.71302 5.69739 9.37315 13.1554
2.32 3.71585 5.70655 9.39742 13.1983
2.33 3.71]864 5.•71563 9.421]57 13.241Z•_
2.34 3.7214 5.72464 9.4456 13.2836
2.35 3.72412 5.73359 9.4695 13.326
2.36 3.7268 5.74246 9.49328 13.3682
2.37 3.72944 5.75127 9.51694 13.4102
2.38 3.73205 5.76 9.5404b 13.452L
2.J9 3.73462 5.76867 9.56J9 13.4938
2.4 3.73716 5.77728 9.5872 13.5354
11-42
TABLE 1V. PECEEhTILES OF THE .LOG-NOP1AL DISTIBUTIOI (Continued)
1 S/X 95th 97.5th 99th 99.5th
I' 2.41 3.73966 5.78581 9.61.038 13.5768 -

2.42 3.74213 5.79428 9.63345 13.6185
2.44 3.74696 5.81103 9.67923 13,7001
2.45 3.74932 5.0193 9.70195 13.7408
"-2.46 3.•75165 5 .82752 9.,72455 13.7815
2.47 3.75395 5.83567 9.74704 13.8219
2.4 3.75622 5.84375 9.76941 13.8622
2.49 3.75845 5.5176 9.79167 13.9023
2.5 3.76066 5.85974 9.81382 13.9423
1.
11-43
, ..
.I..
.
S-'~~~~~~~~~~~~. . . .
.•-.•-.-.-...- . .-.- -........-..........-...... -... , .. . ,
*TABLE V. PERCENTAGE POINTS, by, SUCH THAT P[(/c < by] u y
N 0.02 0.05 0.10 0.25 0.40 0.50 0.60
5 0.604 0.683 0.766 0.951 1.116 1.238 1.378

6 0.623 0.697 0.778 0.937 1.080 1.188 1.304
7 0.639 0.709 0.785 0.930 1.059 1.155 1.256
8 0.653 0.720 0.792 0.926 1.045 1.131 1.223
9 0.665 0.729 0.797 0.925 1.035 1.114 1.198
10 0.676 0.738 0.802 0.924 1.028 1.101 1.179
11 0.686 0.745 0.807 0.924 1.022 1.090 1.163
12 0.695 0.752 0.811 0.924 1.017 1.082 1.151 '-
13 0.703 0.759 0.815 0.924 1.014 1.075 1.140
14 0.710 0.764 0.819 0.925 1.011 1.069 1.132
15 0.716 0.770 0.823 0.925 1.008 1.064 1.124
16 0.723 0.775 0.826 0.926 1.006 1.059 1.117
17 0.728 0.779 0.829 0.927 1.004 1.056 1.111
18 0.734 0.784 0.832 0.927 1.003 1.052 1.106
19 0.739 0.788 0.835 0.928 1.001 1.049 1.101
20 0.743 0.791 0.838 0.929 1.000 1.047 1.097
22 0.752 0.798 0.843 0.930 0.998 1.042 1.090
24 0.759 0.805 0.848 0.932 0.997 1.038 1.084
26 0.766 0.810 0.852 0.933 0.995 1.035 1.079
28 0.772 0.815 0.856 0.934 0.994 1.033 1.074
30 0.778 0.820 0.860 0.935 0.993 1.030 1.070
32 0.783 0.824 0.863 0.937 0.993 1.028 1.067
34 0.788 0.828 0.866 0.938 0.992 1.027 1.064
36 0.793 0.832 0.869 0.939 0.992 1.025' 1.061
38 0.797 0.835 0.872 0.940 0.991 1.024 1.059
40 0.801 0.839 0.875 0.940 0.991 1.023 1.056
42 0.804 0.842 0.877 0.941 0.990 1.022 1.054
44 0.808 0.845 0.880 0.942 0.990 1.021 1.052
46 0.811 0.847 0.882 0.943 0.990 1.020 1.051
48 0.814 0.850 0.884 0.944 0.990 1.019 1.049
50 0.817 0.852 0.886 0.944 0.989 1.018 1.048
52 0.820 0.854 0.888 0.945 0.989 1.017 1.046
54 0.822 0.857 0.890 0.946 0.989 1.017 1.045
56 0.825 0.859 0.891 0.946 0.989 1.016 1.044
58 0.827 0.861 0.893 0.947 0.989 1.015 1.043
60 0.830 0.863 0.894 0.948 0.989 1.015 1.041
62 0.832 0.864 0.896 0.948 0.989 1.014 1.040
64 0.834 0.866 0.897 0.949 0.989 1.014 1.040
*Reproduced from "Inferences on the Parameters of the Weibull Distribution,"

by Darrel R. Thoman, Lee J. Bain, and Charles E. Antle, Technometrics,
Vol. 11 No. 3, (1969), pp. 445-460.
11-44
......................... ... . . ......... o. .. .. % °•.

.o.o
o ,"~ o. o o o .o°- . o. o°*. , .o °o. . .o- .< o. Z.o .- oo ° o o -O- .a o *t °o o Oo o o o - r- - . °-
TABLE V. PERCENTAGE POINTS, b-,, SUCH THAT
S~P[E'/c < byf] -Y• (Continued)
N 0.70 0.75 0.80 0.85 0.90 0.95 0.98

66 0.836 0.868 0.899 0.949 01988 1.014 1.039
68 0.838 0.869 0.900 0.950 0.988 1.013 1.038
70 0.840 0.871 0.901 0.950 0.988 1.013 1.037
72 0.841 0.872 0.903 0.951 0.988 1.012 1.036
"74 0.843 0.874 0.904 0.951 0.988 1.012 1.036
76 0.845 0.875 0.905 0.952 0.988 1.012 1.035
78 0.846 0.876 0.906 0.952 0.988 1.011 1.034
80 0.848 0.878 0.907 0.952 0.988 1.011 1.034
85 0.852 0.881 0.910 0.953 0.988 1.011 1.032
90 0.855 0.883 0.912 0.954 0.988 1.010 1.031 V
95 0.858 0.886 0.914 0.955 0.988 1.009 1.030
100 0.861 0.888 0.916 0.956 0.988 1.009 1.029
* 110 0.866 0.893 0.920 0.958 0.988 1.008 1.027
120 0.871 0.897 0.923 0.959 0.988 1.007 1.025
"1..4..5
-o-4
.* .. ~*
**.*-.v. * °.---"
r.* * .-. ' * .. *. **

* * .** ... *-. ** . ..
ltb-L V. PERCENTAGE POINTS, by* SUCH THAT

PITIC < byl = (Continued)
Y
N 0.70 0.75 0.80 o.65 0.90 0.95 0.98
5 1.557 1.671 1.b12 2.001 2.277 2.779 3.518
b 1.453 1.543 1.662 1.812 2.030 2.436 3.067
7 1.386 1.461 1.561 1.688 1.861 2.183 2.640
b 1.338 1.404 1.491 1.602 1.747 2.015 2.377
9 1.303 1.3b1 1.439 1.538 1.b65 1.89b 2.199
..lU 1.275 1.328 1.399 1.489 1.602 1.807 2.070
11 1.253 1.302 1.367 1.450 1.553 1.738 1.972
14 1.234 1.281 1.341 1.418 1.513 1.682 1.894
•i 1.219 1.263 1.319 1.391 1.480 1.636 1.830
14 1.206 1.248 1.300 1.369 1.452 1.597 1.777
15 1.195 1.234 1.284 1.349 1.427 1.564 1.732
16 1.185 1.223 1.270 1.332 1.406 1.535 1.693
17 1.176 1.213 1.258 1.317 1.388 1.510 1.660
1 1.168 1.204 1.247 1.303 1.371 1.487 1.630
19 1.1b2 1.196 1.237 1.291 1.356 1.467 1.603
20 1.155 1.188 1.228 1.281 1.343 1.449 1.579
22 1.144 1.176 1.213 1.2b2 1.320 1.418 1.536
24 1.135 1.165 1.200 1.246 1.301 1.392 1.504
2b 1.128 1.15b 1.189 1.232 1.2h4 1.370 1.475
28 1.121 1.148 1.180 1.220 1.269 1.351 1.450
30 1.115 1.141 1.171 1.210 1.257 1.334 1.429
"32 1.110 1.135 1.164 1.201 1.246 1.319 1.409
44 1.105 1.129 1.157 1.193 1.23b 1.306 1.392
3b 1.101 1.125 1.151 1.186 1.227 1.294 1.377
3b 1.097 1.Lu 1.146 1.179 1.219 1.2b3 1.363
40 1.094 1.116 1.141 1.173 1.211 1.273 1.351
4. 1.091 1.112 1.137 1.167 1.204 1.265 1.339
44 1.088 1.109 1.132 1.162 1.198 1.256 1.329
46 1.085 1.106 1.129 1.158 1.192 1.249 1.319
48 1.083 1.103 1.125 1.153 1.187 1.242 1.310
50 1.081 1.100 1.122 1.149 1.182 1.235 1.301
52 1.078 1.090 1.119 1.145 1.177 1.229 1.294
54 1.076 1.095 1.116 1.142 1.173 1.224 1.286
56 1.075 1.093 1.113 1.139 1.169 1.218 1.280
58 1.073 1.091 1.111 1.135 1.165 1.213 1.273
bO 1.071 1.089 1.108 1.133 1.162 1.208 1.267
b2 1.070 1.087 1.106 1.130 1.158 1.204 1.262
64 1.068 1.086 1.104 1.127 1.155 1.200 1.256
66 1.067 1.084 1.102 1.125 1.152 1.196 1.251
68 1.066 1.083 1.100 1.122 1.149 1.192 1.246
70 1.064 1.081 1.09b 1.120 1.146 1.188 1.242
72 1.063 1.080 1.097 1.118 1.144 1.185 1.237
11-46
TABLE V. PERCENTAGE POINTS, b., SUCH THAT

Pte/c < b y (Continued)
N 0.70 0.75 0.80 0.85 0.90 0.95 0.98
74 1.062 1.078 1.095 1.116 1.141 1.182 1.233

76 1.061 1.077 1.093 1.114 1.139 1.179 1.229
78 1.060 1.076 1.092 1.112 1.136 1.176 1.225
80 1.059 1.075 1.090 1.110 1.134 1.173 1.222
85 1.057 1.072 1.087 1.106 1.129 1.166 1.213
90 1.055 1.069 1.084 1.102 1.124 1.160 1.206
95 1.053 1.067 1.081 1.099 1.120 1.155 1.199
100 1.051 1.065 1.079 1.096 1.116 1.150 1.192
110 1.048 1.061 1.074 1.090 1.110 1.141 1.181
120 1.046 1.058 1.070 1.086 1.104 1.133 1.171
•""'"-' ",
2.4p..
M.:
p•., .
11-47
.................................................... °•
.1o-4
a
'. ".Oo
*TABLE Vt. PERCENTAGE POINTS, Zy, SUCH THAT

P [Z in (1/b) < tyl -
N 0.02 0.05 0.10 0.25 0.40 0.50 0.60

5 -1.631 -1.247 -0.888 -0.444 -0.241 -0.056 0.085
6 -1.396 -1.007 -0.740 -0.385 -0.194 -0.045 0.079
7 -1.196 -0.874 -0.652 -0.344 -0.168 -0.038 0.074
8 -1.056 -0.784 -0.591 -0.313 -0.150 -0.032 0.070
9 -0.954 -0.717 -0.544 -0.289 -0.137 -0.029 0.067
10 -0.876 -0.665 -0.507 -0.269 -0.126 -0.026 0.065
11 -0.813 -0.622 -0.477 -0.253 -0.118 -0.023 0.062
12 -0.762 -0.587 -0.451 -0.239 -0.111 -0.021 0.061
13 -0.719 -0.557 -0.429 -0.228 -0.106 -0.019 0.059 _
14 -0.683 -0.532 -0.410 -0.217 -0.100 -0.018 0.057
15 -0.651 -0.509 -0.393 -0.208 -0.096 -0.016 0.056
16 -0.624 -0.489 -0.379 -0.200 -0.092 -0.015 0.054
17 -0.599 -0.471 -0.365 -0.193 -0.089 -0.014 0.053
18 -0.578 -0.455 -0.353 -0.187 -0.085 -0.013 0.052
19 -0.558 -0.441 -0.342 -0.181 -0.083 -0.013 0.051
20 -0.540 -0.428 -0.332 -0.175 -0.080 -0.012 0.050
22 -0.509 -0.404 -0.314 -0.166 -0.075 -0.011 0.048 ..
24 -0.483 -0.384 -0.299 -0.158 -0.071 -0.009 0.047
26 -0.460 -0.367 -0.286 -0.150 -0.068 -0.009 0.046
28 -0.441 -0.352 -0.274 -0.144 -0.065 -0.008 0.044
30 -0.423 -0.338 -0.264 -0.139 -0.062 -0.007 0.043
32 -0.408 -0.326 -0.254 -0.134 -0.059 -0.006 0.042
34 -0.394 -0.315 -0.246 -0.129 -0.057 -0.006 0.041
36 -0.382 -0.305 -0.238 -0.125 -0.055 -0.005 0.040
38 -0.370 -0.296 -0.231 -0.121 -0.053 -0.005 0.040
40 -0.360 -0.288 -0.224 -0.118 -0.052 -0.004 0.039
42 -0.350 -0.280 -0.218 -0.115 -0.050 -0.004 0.038
44 -0.341 -0.273 -0.213 -0.112 -0.048 -0.004 0.037
46 -0.333 -0.266 -0.208 -0.109 -0.047 -0.003 0.037
48 -0.325 -0.260 -0.203 -0.106 -0.046 -0.003 0.036
50 -0.318 -0.254 -0.198 -0.104 -0.045 -0.003 0.036
52 -0.312 -0.249 -0.194 -0.102 -0.043 -0.003 0.035
54 -0.305 -0.244 -0.190 -0.100 -0.042 -0.002 0.035
56 -0.299 -0.239 -0.186 -0.098 -0.041 -0.002 0.034
58 -0.294 -0.234 -0.183 -0.096 -0.040 -0.002 0.034
60 -0.289 -0.230 -0.179 -0.094 -0.039 -0.002 0.033
62 -0.284 -0.226 -0.176 -0.092 -0.039 -0.002 0.033
64 -0.279 -0.222 -0.173 -0.091 -0.038 -0.001 0.032
*Reproduced from "Inferences on the Parameters of the Weibull Distribution,-

by Darrel R. Thoman, Lee J. Bain, and Charles E. Antle, Technometrics,
Vol. 11, No. 3, (1969), pp. 445-460.
11-48
... °..'.-o.
_. '.- .-.- ,~.'
....'• %.... .... . .. '...-.....-.......".-.'.........-... ".. .'.. "..."..
TABLE VI. PERCENTAGE POINTS, IL.. SUCH THAT

P (- in C(/b) < tl - y (Continued)
Y
N
66
68
70
0.02
-0.274
-0.270
-0.266
0.05
-0.218
-0.215
-0.211
0.10
-0.170
-0.167
-0.165
0.25
-0.089
-0.088
-0.086
0.40
-0.037
-0.036
-0.035
0.50
-0.001
-0.001
-0.001
0.60
0.032
0.032
0.031
E l"
72 -0.262 -0.208 -0.162 -0.085 -0.035 -0.001 0.031
74 -0.259 -0.205 -0.160 -0.084 -0.034 -0.001 0.031
76 -0.255 -0.202 -0.158 -0.083 -0.033 -0.001 0.030
78 -0.252 -0.199 -0.155 -0.081 -0.033 -0.001 0.030
80 -0.248 -0.197 -0.153 -0.080 -0.032 -0.000 0.030
85 -0.241 -0.190 -0.148 -0.078 -0.031 -0.000 0.029
90 -0.234 -0.184 -0.144 -0.075 -0.030 0.000 0.028
95 -0.227 -0.179 -0.139 -0.073 -0.028 0.000 0.028
100 -0.221 -0.174 -0.136 -0.071 -0.027 0.000 0.027
110 -0.211 -0.165 -0.129 -0.067 -0.025 0.001 0.026
120 -0.202 -0.158 -0.123 -0.064 -0.024 0.001 0.025
-.- '. ~..-.
,q
.. . . . . ....
.. ,:.... '....
. . . . .. . . . .. . . .. . ..' . . . . . . . . . . . . . .
11-49 .. S
I
TABLE VI. PERCENTAGE POINTS, ty, SUCH THAT

P [[-Zn (tb/b) < Ly] - y (Continued)
y
N 0.70 0.75 0.80 0.85 0.90 0.95 0.98
5 0.254 0.349 0.452 0.587 0.772 0.107 1.582

6 0.221 0.302 0.404 0.516 0.666 0.939 1.291
7 0.200 0.272 0.362 0.465 0.598 0.829 1.120
8 0.185 0.251 0.331 0.427 0.547 0.751 1.003
9 0.174 0.235 0.307 0.397 0.507 0.691 0.917
10 0.165 0.222 0.288 0.372 0.475 0.644 0.851
ii 0.157 0.211 0.273 0.351 0.448 0.605 0.797
12 0.150 0.202 0.260 0.334 0.425 0.572 0.752
13 0.145 0.194 0.249 0.319 0.406 0.544 0.714
14 0.140 0.187 0.239 0.306 0.389 0.520 0.681
15 0.135 0.180 0.230 0.294 0.374 0.499 0.653
16 0.131 0.175 0.223 0.284 0.360 0.480 0.627
17 0.128 0.170 0.216 0.274 0.348 0.463 0.605 F
18 0.124 0.165 0.209 0.266 0.338 0.447 0.584
19 0.121 0.161 0.204 0.258 0.328 0.433 0.566
20 0.118 0,157 0.199 0.251 0.318 0.421 0.549
22 0.113 0.150 0.189 0.239 0.302 0.398 0.519
24 0.109 0.144 0.181 0.228 0.288 0.379 0.494
26 0.105 0.138 0.174 0.219 0.276 0.362 0.472
28 0.102 0.134 0.168 0.210 0.265 0.347. 0.453
30 0.098 0.129 0.163 0.203 0.256 0.334 0.435
32 0.095 0.125 0.158 0.197 0.247 0.323 0.420
34 0.093 0.122 0.153 0.191 0.239 0.312 0.406
36 0.090 0.118 0.149 0.185 0.232 0.302 0.393
38 0.088 0.115 0.145 0.180 0.226 0.293 0.382
40 0.086 0.113 0.142 0.175 0.220 0.285 0.371
42 0.084 0.110 0.139 0.171 0.214 0.278 0.361
44 0.082 0.108 0.136 0.167 0.209 0.271 0.352
46 0.080 0.105 0.133 0.164 0.204 0.264 0.344
48 0.079 0.103 0.130 0.160 0.199 0.258 0.336
50 0.077 0.101 0.128 0.157 0.195 0.253 0.328 .. .,
52 0.076 0.099 0.126 0.154 0.191 0.247 0.321
54 0.074 0.097 0.123 0.151 0.187 0.243 0.315
j6 0.073 0.096 0.121 0.148 0.184 0.238 0.309
58 0.072 0.094 0.119 0.146 0.181 0.233 0.303
60 0.071 0.092 0.117 0.143 0.177 0.229 0.297
62 0.070 0.091 0.116 0.141 0.174 0.225 0.292
64 0.068 0.089 0.114 0.139 0.171 0.221 0.287
66 0.067 0.088 0.112 0.137 0.169 0.218 0.282
68 0.066 0.087 0.111 0.135 0.166 0.214 0.278
11-50
TABLE VI. PERCENTAGE POINTS, ty, SUCH THAT

P (- in (6/b) -< i-] -y (Continued)
y
N 0.70 0.75 0.80 0.85 0.90 0.95 0.98
70 0.065 0.085 0.109 0.133 0.164 0.211 0.274

72 0.064 0.084 0.108 0.131 0.161 0.208 0.269
74 0.064 0.083 0.107 0.129 0.159 0.205 0.266
76 0.063 0.082 0.105 0.128 0.157 0.202 0.262 !
78
80
0.062
0.061
0.081
0.080
0.104
0.103
0.126
0.125
0.155
0.153
0.199
0.197
0.258
0.255
• k. _
85 0.059 0.077 0.100 0.121 0.148 0.190 0.246 :"
90 0.057 0.075 0.097 0.118 0.143 0.185 0.239
95 0.056 0.073 0.095 0.115 0.139 0.179 0.232
100 0.054 0.071 0.093 0.112 0.136 0.175 0.226
110 0.051 0.067 0.089 0.107 0.129 0.166 0.215
120 0.049 0.064 0.085 0.103 0.123 0.159 0.205
115. 1
*..o°•o..!o
n.° -.4.v....•
'-.,,
'.-~.
~ . ~
...................
. . . . . .
.. . . ~
. . ~ .*.-*.. ~ ~
.. . .* . .. ~ ... ~ .*.. ~ . ~ .
.. . . . . . . . .. . . .. .. .
.. ........... ...... ..... ..... .....
TABLE VII. TABLE OF COEFFICIENTS FOR COMUTING Za A Z

n 10
p-r/ n 1 7/10 5/10 2/10

Sa iai bi ab aj bi
1 .03958 - .07107- .00255 - .12531 - .08169 - .18478- .67530- .46452

2 .05108 - .07334 .01280 - .12299 - .05458 - .17371 1.67530 .46452
3 .06219 - .07177 .03094 - .11279 - .01793 - .14970 •-
4 .07342 - .06675 .05228 - .09525 .02859 - .11335
5 .08512 - .03797 .07763 - .06949 1.12561 .62153
6 .09767 - .04452 .10833 - .03336
7 .11155 - .02459 .72057 .55919 Aft.
8 .12761 .00548
9 .14751 .05431
* .35022•
10 .20427 3 -,j . - %-
*Reproduced from "An Exact Asymptotically Efficient Confidence Bound for

Reliability In the Case of the Weibull Distribution," by M. V. Johns, Jr. and
G. J. Lieberman, Technometrics, Vol. 8, No. 1, (1966), pp. 135-175.
11-52
.W ! '7
[ABLE VII. TABLE OF COEFFICIENTS FOR COMPUTING Za AND Zb (CoLlrnued)
n" 15
p-r/n 1 1t1/15 7/15 3/15 I.*' .
al bi ai bi ai bi ai 5i
1 .02383 - .04b19 .000b5 - .07776 - .07407 - .13306 - .47594 - .31963

2 .02932 - .U4830 .00683 - .07905 - .06265 - .13055 - .41146 - .29297
3 .03452 - .04890 .01367 - .07754 - .04717 - .12274 1.88740 .61260
4 .03961 - .04832 .02120 - .07379 - .02811 - .11056
5 .04471 - .04b67 .02950 - .06793 - .00541 - .09416
6 .0U4V90 - 04392 .03871 - .05989 .02125 - .07332 "'"".
7 .05525 - .0399b .04899 - .04941 1.19616 .66438
8 .UOb4 - .03461 .Ub059 - .03610
9 .uo677 - .02757 .07386- .01931 .. .*-
i0 .07315 - .01635 .0o930 .00194
11 .0801o - .00619 .b1672 .53884
12 .0o808 .01025
13 .U9740 .03359
14 .I0911 .0b982
15 .14734 .29532
n 20
p-r/n 1 15/20 10/20 5/20

i a bi ai bi ai bi ai bi
L .01682 - .U3402 .00120 - .05593 - .04527 - .09198 - .24498 - .19315

2 .02007 - .03564 .00456 - .05745 - .04032 - .09230 - .22587 - .18644 '
3 .U2312 - .03645 .00b16 - .05754 - .03371 - .09010 - .19843 - .17383
4 .02b07 - .03667 .01201 - .05657 - .02574 - .08597 - .16426 - .15659
5 .02898 - .03b39 .01612 - .05468 - .01650 - .08013 1.83354 - .71001
6 .03189 - .03563 .02053 - .05190 - .0059b - .07264 .
7 .0342 - .03441 .0U526 - .04822 .00595 - .06345
8 .0.s78u - .03269 .03037 - .04361 .01935 - .0524b
9 .040bo - .03045 .U3591 - .03798 .03444 - .03948
l1 .04401 - .02762 .04196 - .03119 1.10777 - .66851
11 .04729 - .02411 .04862 - .02306
12 .U5074 - .01980 .05600 - .01336
13 .0543J9 - .01451 .0Q428 - .00172
14 .05830 - .uU7Y8 .73b9 .01235
15 .06257 .00016 .56132 .52087
lb .ub730 .01054
17 .07268 .02419
"18 .07906 .04312
19 .08714 .07197
20 .11608 .25640
II-53
-. . ..... . .. ~. .. . .. . . .. .............. . ,...,.

TABLE VII. TABLE OF COEFFICIENTS FOR COMPUTING Za AND Zb (Continued)

n a 30
p-r/n 1 22/30 15/30 7/30
i ai bi i bi ai bi ai bi
1 .01046 - .02215 - .00105 - .03780 - .03092 - .06072 - .18830 - .13857

2 .01201 - .02314 .00042 - .03893 - .02916 - .06164 - .18110 - .13688
3 .01345 - .02379 .00198 - .03944 - .02675 - .06151 - .17012 - .13259
4 .01483 - .02419 .00363 - .03951 - .02385 - .06063 - .15631 - .12611
5 .01618 - .02440 .00536 - .03923 - .02052 - .05914 - .14008 - .11862
6 .01751 - .02445 .00718 - .03863 - .01679 - .05710 - .12169 - .10937
.01882 - .02434 .00910 - .03775 - .01267 - .05453 1.95751 .76245
8 .02013 - .02409 .01111 - .03658 - .00816 - .05146
9 .02145 - .02369 .01323 - .03513 - .00325 - .04788
10 .02277 - .02315 .01546 - .03339 .00209 - .04379 .. ,.,
11 .02411 - 02246 01781 - .03137 .00787 - .0391500
12 .02546 02162
5 02030 ,02903 .01413 ,03193
13 .02684 - .02062
302811 .02294 - .02636 .02092 -
14 .02824 - .01944 .02574 - .02334 .02827 - .02161
15 .02969 - .01807 .02873 - .01992 1.09870 - .68119 :
16 701607 .03192 -
.03117 - .01648
17 .03270 - .01466 .03535 - .01173
18 .03428 - .01257 .03904 - .00683
19 .03593 - .01017 .04305 .00129
20 .03767 - .00740 .04743 .00500
21 .03949 - .00420 .05225 .01218
22 .04143 - .00047 .56902 .52514
"23 .04350 .00391
24 .04574 .00913
25 .04821 .01545
"82 .05096 .02331
S27 .05412 .03343
.28 05789 .04719 ' "" •'"
•5. ~29 .06270 .06779 •....
'"30 .08227 .20535
. ~~n =50 ,•-'

p-r/n 1 37/50 25/50 12/50 ."-
i ai bi ai bi ai bi ai bi f'..
1 .00588 - .01296 - .00067 - .02197 - .01880 - .03594 - .10798 - .08046 •'
2 .00648 - .01344 - .00016 - .02259 - .01837 - .03662 - .10666 - .08072.,-
3 •00705 - .01380 .00038 - .02299 - .01772 - .03694 - .10414 - .08013 •.,
•-4 .00758 - .01407 .00093 - .02325 - .01691 - .03701 - .10078 - .07896"-','
S5 .00810 - .01428 .00150 - .02340 -,01599 -,03689 -,09675 -,07731 ""
S6 .00861 - .01444 .00209 - .02345 - .01496 - .03661 - .09213 - .07527 •..
-- 7 .00911 - 01456 .00270 - .02342 - .01382 - .03619 - .08699 - .07287
S8 .00960 -. 01463 .00333 - .02332 - .01260 - .03563 - .08135 - .07013'• "l
11-54 i•i T
TABLE VII. TABLE OF COEFFICIENTS FOR COMPUTING Za AND Zb (Continued)
n - 50
p-r/n 1 37/50 25/50 12/50
aI bi ai bi ai bi aj bi
9 .01008 - .01467 .00398 - .02315 - .01129 - .03495 - .07523 - .06709
i#1
10 .01057 - .01467 .00464 - .02291 - .00989 - .03415 - .06865 - .06373
11 .01105 - .01463 .00532 - .02261 - .00840 - .03324 - .06162 - .06008
12 .01152 - .01457 .00603 - .02224 - .00683 - .03221 1.98227 .80675
13 .01200 - .01447 .00675 - .02182 - .00517 - .03107
14 .01248 - .01434 .00750 - .02133 - .00342 - .02982
15 .01296 - .01418 .00827 - .02079 .00158 - .02846
16 .01344 - .01399 .00907 - .02018 .00035 - .02698
17 .01393 - .01376 .00989 - .01951 .00238 - .02538
18 .01442 - .01351 .01073 - .01877 .00451 - .02366
19 .01491 - .01321 .01161 - .01796 .00675 - .02181
20 .01540 - .01289 .01252 - .01709 .00910 - .01984 V
21 .01591 - .01252 .01345 - .01614 .01156 - .01772
22 .01642 - .01212 .01443 - .01512 .01415 - .01546
23 .01693 - .01168 .01544 - .01401 .01687 - .01305
24 .01745 - .01119 .01648 - .01282 .01973 - .01047
25 .01798 - .01067 .01757 - .01154 1.09036 .69008
26 .01852 - .01009 .01871 - .01017
27 .01907 - .00946 .01989 - .00869
28 .01964 - .00877 .02113 - .00709
29 .02021 - .00803 .02243 - .00538
30 .02080 - .00722 .02379 - .00354
31 .02140 - .00634 .02521 - .00154
32 .02202 - .00538 .02672 - .00060
33 .02266 - .00434 .02831 - .00203
34 .02332 - .00319 .02999 - .00545
35 .02401 - .00194 .03178 .00820
36 .02471 - .00057 .03369 .01119
37 .02545 .00095 .53455 .51043
38 .02623 .00263
39 .02704 .00450
40 .02789 .00659
41 .02880 .00894
42 .02977 .01163
43 .03082 .01473
44 .03196 .01834
45 .03322 .02266
46 .03464 .02793 .. .
47 .03627 .03462
48 .03823 .04360
49 .04074 .05685
50 .05269 .15062
11-55
TABLE V11. TABLE OF COEFFICIENTS FOR COMPUTING 2i AND Ib (Continued)
pUr/n 1 75/100 0 50/10 25/100

iax bl a ,, bi ai bi a ,. b i
1 .00277 - .00632 - .00020 - .01056 - .00944 - .01768 - .04999 - .03820

2 .0294 - .00648 - .00013 - .01079- .00940 - .01797 - .05003 - .03854
3 .U310 - .00662 - .00000 - .01096 - .00931 - .01817 - .04981 - .03868
4 .00325 - .00673 .00013 - .01109 - .00918 - .01831 - .04249 - .03869
5 .00339 - .00682 .00027 - .01121 - .00902 - .01841 - .04889 - .03861
6 .00353 - .00690 .00040 - .01130 - .00885 - .01847 - .04824 - .03844
7 .00367 - .0069b .00054 - .01137 - .00865 - .01850 - .04750 - .03820
0 .00380 - .00704 .0006b - .01143 - .00844 - .u1851 - .04668 - .03789
9 .00393 - .00710 .00082 - .11477 - .00821 - .01848 - .04577 - .03754
10 .0040b - .00715 .00097 - .01150 - .00797 - .01844 - .04479 - .03713
11 .00419 - .00719 .00111 - .01152 - .00771 - .01838 - .04375 - .03667 -. :,,
12 .00432 - .00723 .00126 - .01153 - .00745 - .01829 - .04264 - .03616
13 .00445 - .00726 .00141 - .01153 - .0071b - .01819 - .04146 - .03562
14 .00457 - .00728 .00157 - .01152 - .O0687 - .01808 - .04023 - .03503
15 .00469 - .00730 .00172 - .01150 - .00657 - .01794 - .03894 - .03440
16 .00481 - .00732 .00188 - .01147 - .00625 - .01779 - .03759 - .03374
17 .00494 - .00733 .00204 - .01143 - .00592 - .01762 - .03619 - .03303
1b .00506 - .00733 .00220 - .01138 - .00558 - .01744 - .03473 - .03229
19 .00519 - .00733 .00236 - .01133 - .00523 - .01724 - .03322 - .03151
Yo .00531. - .00733 .00253 - .01127 - .00487 - .01703 - .03165 - .03069
21 .00543 - .00732 .00270 - .01120 - .00450 - .01681 - .03003 - .02984
22 .00555 - .00731 .00287 - .01112 - .0041; - .0165b - .02836 - u02895
23 .00567 - .00730 .00304 - .01103 - .00372 - .01631 - .02663 - .02803
24 .00579 - .03728 .00322 - .01094 - .00332 - .01604 - .02485 - .02707
25 .00591 - .0072b .00389 - .01084 - .00290 - .01576 1.97141 .83495
26 .00u03 - .00723 .00357 .01073 - - .00246 - .01546
27 .00615 - .00726 .00376 - .01049 - .00204 - .01515
27 .00627 - .00716 .00394 - .01049 - .00159 - .01482
29 .00640 - .00712 .U0413 - .01036 - .00113 - .01448
u .U0652 - .o0708 .00432 - .01022 - .00065 - .01412
31 ,00b64 - .00703 .00452 - .01007 - .00017 - .01375
32 .0676 .00698 .00472 - .00992 .00033 - .01337
33 .00688 - .00692 .00492 - .00976 .00084 - .01297
34 .00701 - .006b6 .00512 - .00959 .00137 - .01255
35 .00713 - .00680 .00533 - .00941 .00190 - .01212
36 .00725 - .00673 .00544 - .00922 .00245 - .01167
*7 .00738 - .00665 .00575 - .00903 .00301 - .01121
40
3b
39
.00750 - .00b58
.00763 - .00650
.00775 - .00611
.00597 - .00882
.00619 - .00861
.00642 - .00839
.00359
.00418 - .01023
.00479
-
-
.01073
.00971
i
... .......... ................. . .........
........ ° °....... o°•.'

- *'/v- "° . . *o. . .
*.
. *'
. *.*'o'o
. *
.. . -'.°."
-
.
. "..
. . . *•-.
..,, -o.
. °.
. - .°
*.. - ...-.o . . . ° -
TABLL VII. TABLL OF (.0EFFMIENTh FOR COKPUTING Za AND Zb ((ontinued)
10-00
p-rtn 1 75/100 50/100 25/100
"i ai bi a.- bi ai bi ai b
41 .00788 - .00632 .00664 - .00816 .00541. - .00918

42 .00801 - .00622 .00688 - .00792 .00695 - .00863
43 .00813 - .00612 .00711 - .00767 .00670 - .00806
44 .00826 - .00601 .0073b - .00741 .00737 - .00747
45 .00839 - .00590 .00760 - .00714 .00805 - .00687
46 .00852 - .00579 .00785 - .00686 .00875 - .00624
"47 .O0866 - .0056b6 .00811 - .00657 .00948 - .00559
4. .00879 - .00554 .00837 - .00627 .01022 - .00492
49 .00892 - .0540 .00863 - .00595 .0o097 - .00422
50 .00906 - .00526 .00890 - .00563 1.06321 .69595
31 u0919 - .0U512 .00918 - 00529
52 .00933 - .00497 .00946 - .00494
53 .00947 - .00481 .00975 - .00458
54 .00961 - .00464 U01005 - .00420
55 .00975 - 00Q447 .01035 - .00381
50
7
ji .00990
.01004 -- .00429
.0041O .01,065 --
•01097 .00340
.00298
5b .01019 - .00390 .01129 - .00255
59 .01034 - .00370 .01162 - .00209
60 .01049 - •0034b .01196 - .00162
61 .01064 - .00326 .01231 - .00113
"62 .01080 - .00303 .01266 - .00062
63 .01095 - .00278 .01303 - .00009
b4 .013.3 - .00253 .01341 .00046
65 .01127 - •00226 .01379 .00104
bb .01144 - .00199 .01419 .00163
67 .01160 - .00170 .01.460 .00226
"68 .01177 - .00139 .01502 .00291
_9 .01195 - .00107 .01546 .00359
"70 .01212 - .00074 031591 .00429
71 .01230 - .00039 .01637 .00504
72 .01240 - .00002 .01685 .00583.
73 .01267 .00037 .01735 .00663
74 .01287 .00077 .01786 .00748
75 .01306 .00120 .0v752 .49223
"70 .01326 .00165
77 .01347 .00213
.11-57
-"..........
= o-..
.... .° . o. o.j. . .................
, ° . . .... . . . ... ...-
TVABLE VII. TASLE OF COEFFICIENTS FOR COMPUTING Za AND Zb (continued)
n n 100
pmr/n 1 75/100 10 80/0 25/100
I ai bi ai biai bi ib
78 .01368 .00264
79 .01390 .00317
80 .01412 .00374
81 .01436 .U0435
82 .01459 .00500
83 .01484 .00569
84 .01510 .00643
85 .01537 .00723
86 .01565 .00810
87 .01594 .00904
88 .01625 .01006
89 .01657 .01119
90 .01692 .01243
91 .01728 .01382
92 .01768 .01538
93 .01810 .01716
94 .01857 .01922
95 .01908 .02164
96 .01.967 .02458
97 .02034 .02826
9b .02115 .03314
99 .02220 .01027 hL
WuO .02829 .09493
IL
11-56
p
TABLkt VIII. IABL6 OF EXACT LOWER COhFIDENCE BOUNDS FOR R(tO)*
n - 10
p r/n 2/lU
y - 0 .50 y - 0. 7 5 y- 0 . 90 y - 0.95 Y - 0. 9 9
za/Zb L*(ZI/Z ) L*(Z/Z ) L*(Za/Z ) L*(Z /Z ) L*(Z /Z/)
.0 .726 .543 .319 .154 .010

.1 .734 .564 .357 .193 .022
.2 .742 .585 .381 .224 .045
.3 .749 .599 .409 .274 .072
.4 .755 .b14 .439 .314 .104
.5 .763 .626 .469 .345 .135
.6 .773 .647 .493 .385 .169
.7 .780 .660 .520 .423 .224
.8 .789 .677 .539 .450 .256
.9 .797 .690 .560 .477 .304
1.0 .806 .702 .585 .508 .338
1.1 .184 .717 .601 .538 .412
1.2 .822 .725 .614 .556 .428
1.3 .828 .738 .634 .578 .467
1.4 .835 .751 .648 .594 .484
1.5 .842 .759 .661 .609 .501
.b .d48 .769 .672 .618 .514 ""
.,*"
1.7 .855 .77b .685 .626 .526
1.8 .860 .765 .694 .639 .541
1.9 .865 .793 .708 .645 .545
2.0 .873 .600 .717 .652 .554
2.1 .878 .806 .722 .659 .562 -
2.2 .883 bl81 .731 .670 .569
2.3 .888 .815 .738 .675 .572
2.4 .892 .821 .743 .686 .578
2.5 .897 .829 .746 .688 .578
2.6 .900 .833 .751 .691 .586
2.7 .903 .838 .757 .694 .589
2.8 .907 .842 .760 .700 .595
2.9 .912 .849 .767 .707 .596
3.0 .915 .653 .773 .711 .597 •ui
3.1 .918 .855 .779 .714 .600
3.2 .921 .859 .784 .717 .605
3.3 .924 .863 .786 .721 .609
3.4 .926 b865 .791 .724 .b09 "
3.5 .928 .869 .796 .728 .609
3.6 .930 .870 .798 .732 .b09
3.7 .932 .873 .800 .736 .609 LJ
L*(Z /Zb) is the exact lower confidence bound for R(to)
a b0
•*Reproduced from "An ExacL Asymptotically Etficient Confidence bound for
Keliabi•.ity in the L4ase ot the Weibull Distribution," by H. V. Johns, Jr.
and G. J. Lieberman, Technometrics, Vol. 8, No. 1, (1966), pp. 135-175.
%-
11-59
-~~~~.-..°.
-. °.....'. ....
-. • . •* .... •-. •.-. ..... •,..."....o - . . .-. •,... ... "°. •. ..
"TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued)
n- 10
p - r/n - 2/10 9 99
,," 0.50 y-0.75 y a 0.90 y-. 5 y- 0.
" Za/Zb L*(Za/Zb) L*(Zs/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb)
3.8 .934 .875 .804 .740 .609

3.9 .937 .879 .807 .745 .611
4.0 .939 .880 .809 .748 .613
4.1 .941 .883 .810 .752 .616
4.2 .944 .887 .813 .753 .617
"4.3 .946 .890 .816 .754 .617
4.4 .948 .892 .818 .755 .618
4.5 .950 .894 .820 .756 .622
4.6 .951 .896 .823 .759 .623
4.7 .953 .899 .825 .762 .623
4.8 .955 .900 .826 .764 .623
4.9 .957 .901 .830 .766 .626
5.0 .959 .902 .833 .769 .628
,.;.. ..
L*(Z /Z) iii the ,xact lover confidence bound for R(t0)
a b 0
11-60
.-........ ..... *. "." • .* "'. .' ......'* ' ' ' '.' '
.......... "...'....**...
..... ... " . " .."."."
. . "... ..., w..". -',".... " *...• "'.,' 4 .* ,".".".
S.. ... . .' ..-.T- I~r . . .
•I"T . •T"•
. T" -T . . .
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued) RN
a - 10
p - rli - 5/10
y- 0,50 y" 0.75 y- 0.90 y- 0.95 y- 0.99
Z/b L*(ZI 7Zb L*(Za/Zb L*(Z /Z) L*(Z /Z ) L*(Z /Z)
aa a b) a b) ab
.0 .464 .341 .244 .193 .095
.1 .490 .374 .281 .234 .126
.2 .516 .403 .315 .270 .153
.3 .543 .434 .350 .304 .180
.4 .568 .466 .381 .334 .217
.5 .593 .493 .410 .357 .236
.6 .617 .518 .436 .386 .269
.7 .640 .542 .462 .413 .300
.8 .663 .568 .484 .430 .327
.9 .684 .589 .508 .455 .352
1.0 .703 .611 .531 .475 .378
1.1 .720 .633 .543 .495 .404
1.2 .738 .653 .564 .513 .426
1.3 .766 .674 .581 .526 .440
1.4 .771 .692 .599 .540 .453
1.5 .789 .707 .616 .557 .467
1.6 .805 .719 .631 .573 .486
1.7 .817 .731 .643 .589 .495 "
1.8 .829 .745 .657 .603 .505
"1.9 .840 .759 .670 .618 .518
2.0 .850 .773 .683 .630 .530
"2.1 .859 .784 .696 .642 .538
2.2 .868 .794 .707 .651 .551
2.3 .877 .804 .720 .662 .561
I 2.4 '.885 .815 .733 .671 .568
.576
2.5 .892 .824 .742 .678
2.6 .899 .832 .750 .687 .592
2.7 .906 .843 .760 .696 .606
2.8 .912 .850 .771 .707 .616 ,,,m,
S2.9 .919 .858 .780 .718 .627
3.0 .924 .864 .789 .724 .634
3.1 .929 .870 .797 .730 .640
3.2 .934 .878 .806 .736 .646
3.3 .939 .885 .813 .743 .652
"3.4 .942 .891 .820 .751 .657
3.5 .946 .896 .826 .756 .663
3.6 .950 .901 .834 .762 .669
3.7 .953 .906 .841 .767 .675
3.8 .956 .911 .846 .775 .681 Op
* L*(Z /Z ) is the exact lower confidence bound for 1(t0 )
; : ' •"2 '-'-,"i

2.-2'•. b.
S . . . . . .. . . . . . .. '.
-,:. - -- -?i? -:.-. i"i i 22 i 1:"-'i:i: .'-"i i i ':. ;
, "-" ---
, ." " "6..1"
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(to) (Continued)
p = tin - 5/10
y - 0.50 y - 0.75 y - 0.90 y - 0.95 y - 0.99
za /Zb L*(Za/Zb) L*(Za/Zb) L*(Za/Zb) L*(Zd/Zb) L*(Zd/Zb)
3.9 .960 .916 .850 .783 .687

4.0 .963 .921 .855 .789 .693
4.1 .965 .925 .862 .795 .699
4.2 .968 .929 .867 .801 .704
4.3 .970 .932 .872 .806 .710
4.4 .972 .935 .877 .812 .716
4.5 .974 .939 .883 .818 .721
4.6 .975 .942 .887 .823 .727
4.7 .977 .944 .891 .828 .732 r.
4.8 .978 .947 .896 .834 .738
4.9 .980 .950 .899 .839 .744
5.0 .981 .953 .903 .843 .750
1..,.,,V
.\.
.o* * . - o
,>*
.; .-..
L*(Z /Zb) is the exact lover confidence bound ffor R(tO)
11-62 -."• i'
. . . . - . .-..
. .. .. ... -i . .: • .. • -.:
.. .-. -_- . . _.
. . .',-..•.: ...:.-.-.
..: .• . • • ,. . ,• ••.,• • •,-,-- t-
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued) " -
n- 10
p - r/n - 7/10
y 0.50 y- 0.75 y- 0.90 yu 0.95 y 0.99
Z/Zb L*(Za/Zb ) L*(Z /Zb) L*(Za/Zb) L*(Z,/Zb) L*(Z /Z b)
.0 .415 .332 .260 .215 .141

.1 .448 .362 .290 .245 .163
.2 .480 .396 .323 .275 .191 ,
.3 .509 .423 .351 .302 .214
.4 .537 .453 .380 .328 .243 "- "'
.5 .567 .482 .402 .357 .271
.6 .595 .509 .426 .387 .304
.7 .620 .535 .452 .408 .322
.8 .645 .560 .475 .425 .340
.9 .668 .582 .493 .447 .359
1.0 .691 .607 .514 .469 .377 ' "
1.1 .712 .629 .537 .491 .402
1.2 .730 .650 .555 .508 .420
1.3 .747 .669 .578 .522 .435
1.4 .767 .687 .595 .537 .453
1.5 .784 .704 .612 .554 .467
1.6 .800 .722 .630 .570 .485
1.7 .814 .736 .645 .583 .500
1.8 .827 .752 .659 .599 .514
1.9 .840 .765 .672 .615 .529
2.0 .852 .779 .687 .628 .543
* 2.1 .862 .791 .702 .645 .556
2.2 .873 .804 .714 .660 .567
2.3 .883 .813 .727 .674 .576 ,• -
2.4 .892 .824 .741 .684 .583
2.5 .900 .835 .752 .697 .590
2.6 .908 .845 .763 .710 .597
2.7 .915 .855 .773 .717 .604
2.8 .921 .863 .782 .727 .615
2.9 .927 .871 .792 .738 .626
3.0 .932 .879 .802 .748 .634
3.1 .937 .886 .811 .758 .640
3.2 .942 .892 .819 .768 .651
3.3 .947 .899 .830 .778 .662
3.4 .951 .905 .837 .787 .674
3.5 .955 .910 .845 .795 .684
3.6 .958 .916 .853 .804 .698
3.7 .962 .921 .860 .813 .702
3.8 .965 .926 .866 .821 .709
L*(Za/Zb) is the exact lower confidence bound for R(tO)
11-63
..................
. ** . ~.°°
* * * - ..- \-....----
i
,mo
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(t 0 ) (Continue*)
n 10
p r/n- 7/10
y 75
0.50 y 0. 1 0.90 -y 0.95 y" 0.99
Z /Z L*(Z /Z) L*(Z /Z) L*(Z /Z) L*(Z /Z) L*(Z /Z)
Ma b a b a b a b a b a b
3.9 .968 .930 .873 .827 .715
4.0 .970 .934 .878 .836 .722
4.1 .973 .939 .883 .b45 .730
4.2 .975 .943 .888 .851 .737
4.3 .977 .946 .893 .856 .744 .\ •
4.4 .979 .949 .898 .861 .750
4.5 .980 .952 .904 .866 .755
4.6 .982 .955 .908 .870 .759
4.7 .983 .958 .911 .875 .764
4.6 .985 .960 .915 .880 .770
4.9 .986 .963 .919 .884 .778
5.0 .987 .965 .924 .889 .785
L*(Z /Z ) is the exact lower confidence bound for R(t )
11-64 ... . . .
TABLE VIII. TABL OF

0 YXACT LOWER COHFIDENC& BOUNDS FOR R(tO) (Continued)
'p a. / 1', -
0.50 y - 0.75 y a 0.90 y 0.95 y - 0.99
I/Zb L*(Z /Z ) L*(Z /Z ) L*(Z /Zb L*(Z /Z b L*(Z./Zb)

.0 .393 .315 .255 .212 .153
.1 .428 .345 .283 .238 .176
".2 .458 .376 .311 .264 .202
.3 .492 .409 .336 .291 .226
.4 .522 .442 .363 .319 .247 -. ".-:
.5 .552 .474 .389 .345 .273
.6 .581 .502 .416 .371 .298
.7 .607 .527 .439 .396 .317
.8 .634 .553 .464 .421 .334 'r
.9 .658 .579 .490 .442 .352
1.0 .682 .605 .512 .403 .369 .'. .-
1.1
1.2 .705
.726 .627
t)49 .535
.559 .485
.505 •395
.421 •-•
1.3 .747 .070 .582 .525 .445
1.4 .7686 .605 .546 .462 .90
1.5 .786 .707 .621 .567 .478
1.b .800 .725 .644 .587 .494
1.7 .816 .742 .662 .603 .506
1.8 .830 .758 .680 .b22 .522
1.9 .843 .772 .697 .636 .540
2.0 .855 .787 .710 • 54 .559
"2.1 .868 .801 .727 .668 .573
2.2 .879 •814 .743 .684 .587
2. 4 .889 .826 .758 .700 .600
2.4 .898 .837 .772 .713 .613
2.5 .907 .847 .783 .725 .626
2.6 .915 .856 .795 .738 .b39
2.7 .921 .865 .806 .751 .651 ""
"2.8 .928 .874 .816 .7b2 .665
"2.9 .934 .883 .826 .775 .677
3.0 .939 .890 .835 .787 .689
3.1 .944 .898 .846 .800 .698
3.2 .949 .904 .853 .811 .708
3.3 .954 .911 .861 .820 .719
3.4 .957 .917 .868 .827 .729
3.5 .961 .922 .875 .837 .739
3.6 .964 .927 .883 .846 .748
3.7 .968 .933 .889 .854 .757 1
3.8 .970 .937 .895 .862 .76b
.*i4a/Zb) is the exact lower confidence bound for R(t
11-65
. 4-....
• --::~~~~~~~~~~~~.
.. :..-..:- . .-............... ....... •... , ..-....... ....
n - 10
p - r/n - I
y - 0.50 y - 0. 75 y - 0.90 y - 0. 9 5 y - 0.99
Za/Zb L*(Za/Zb) L*(Za/Zb) L*(Z a/Z L*(Za/Zb) L*(Z a/Z b

3.9 .973 .941 .901 .869 .775
*4.0 .975 .945 .906 .875 .784
*4.1 .977 .949 .912 .880 .792
4.2 .979 .952 .917 .885 .800
4.3 .981 .956 .922 .892 .808
4.4 .982 .959 .926 .896 .815
4.5 .984 ý962 .931 .903 .822
4.6 .985 .964 .935 .907 .827
4.7 .987 .967 .939 .912 .833
4.8 .988 .969 .943 .916 .838
4.9 .989 .972 .946 .921 .843
5.0 .990 .974 .950 .925 .848
I•
.. . .. . . . . . . . . . . . . . . . . .
/Zb
L a( .. .. .. confidence
is the exact lover . . . . .. .. .. .bound
. . ..-
. .for
. .R(y
. . . .. . . .. . . . . . .. . .. .. . . . . . . . . . . . . . . .. o
. . .. . . .. . . . .. . . . . . . . . . . . . . . . . . . . . .
11-66
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(t 0 ) (Continued)
p n- 15 "
p - r/n - 3/15
y m 0.50 y - 0.75 y - 0.90 y - 0.95 y m 0.99
z/Zb L*(Za/Zb) L*(Za/Zb) L*(Za/Zb) L*(Z/Zb L*(Z/ Zb)

.0 .646 .443 .210 .090 .001
.1 .660 .475 .242 .118 .006
.2 .673 .508 .276 .160 .017
.3 .688 .534 .318 .194 .037
.4 .702 .561 .357 .240 .064
.5 .716 .583 .392 .285 .101
.6 .732 .607 .433 .331 .147
".7 .742 .627 .478 .382 .202 .*
.8 .755 .647 .520 4428 .257
.9 .766 .665 .553 .465 .300
' 1.0 .776 .681 .584 .495 .348 .-.
1.1 .787 .698 .610 .533 .395
1.2 .799 .715 .631 .559 .451
1.3 .812 .733 .651 .591 .491 6
"
1.4 .821 .749 .671 .619 .522
1.5 .832 .762 .691 .642 .547
' 1.6 .840 .773 .707 .658 .566
1.7 .849 .785 .725 .673 .587
1.8 .857 .797 .736 .684 .604
""•2.0 1.9 .866
.873 .806
.814 .744
.751 .697
.707 .609
.620'-"
"."2.1 .879 2.1• .823 .761 .718 .642 '•
2.2 .886 .829 .768 .729 .654
2.3 .893 .839 .774 .739 .661
2.4 .900 .846 .782 .749 .676
2.5 .905 .853 .791 .757 .688
2.6 .91] .859 .798 .763 .699
2.7 .914 .864 .805 .772 .704
2.8 .919 .869 .814 .776 .707
2.9 .923 .874 .819 .784 .709
3.0 .927 .879 .824 .790 .711
3.1 .932 .883 .829 .794 .717
3.2 .936 .887 .832 .799 .720
3.3 .939 .891 .836 .802 .723
3.4 .942 .896 .842 .803 .726
3.5 .945 .899 .849 .810 .733
3.6 .948 .902 .852 .819 .735
3.7 .951 .906 .857 .822 .736
3.6 .953 .910 .860 .825 .738
L*(Za/Zb) is the exact lower confidence bound for R(tO)
11-67
A'
-~....
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR Reto) (Continued)
n -15
p - r/n - 3/15
y - 0.50 y - 0.75 y - 0.90 y- 0.95 y - 0.99
Za/Zb L*(Za/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb)

3.9 .955 .914 .863 .828 .741
4.0 .957 .917 .865 .832 .743
4.1 .960 .919 .868 .835 .745
4.2 .962 .921 .871 .837 .745
4.3 .964 .924 .875 .841 .746
4.4 .966 .927 .878 .843 .748
4.5 .968 .929 .880 .846 .749
4.6
4.7 .969
.971 .932 .884 .849 .752
.934 .888 .853 .757
4.8 .972 .936 .890 .857 .763
4.9 .974 .938 .892 .860 .768
5.0 .975 .939 895 .862 .769
11-68
o . .** • -. o°- -.
••.. ,. .. .. ° / .. ... . a, - .- .
,-...- . . ..
. " ." ."
. . * . -, ".* ".
- . .- .. ".
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued)
" 15
p - r/n - 7/15
y - 0.50 y - 0.75 y - 0.90 y - 0.95 y - 0.99
Z /Zb L*(Z&/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za Zb) L*(Za/Zb)
.0 .443 .343 .247 .205 .111

.1 .471 .373 .284 .245 .149
.2 .502 .404 .320 .280 .197
.3 .530 .438 .356 .317 .243
.4 .560 .472 .391 .351 .285
.5 .586 .500 .427 .385 .318
.6 .615 .528 .457 .417 .344
.7 .639 .556 .486 .446 .381
.8 .661 .582 .514 .470 .403
.9 .684 .608 .542 .496 .421
1.0 .705 .630 .562 .521 .444
1.1 .725 .653 .587 .548 .466
1.2 .743 .672 .606 .567 .487
1.3 .760 .690 .626 .585 .502
1.4 .776 .710 .646 .604 .516
1.5 .794 .729 .664 .624 .530
1.6 .809 .743 .677 .642 .543
1.7 .823 .759 .690 .658 .564
1.8 .835 .773 .706 .673 .579
1.9 .846 .786 .719 .689 .592
2.0 .657 .799 .732 .702 .610
2.1 .868 .810 .744 .715 .627
2.2 .877 .822 .757 .725 .642
2.3 .886 .832 .768 .735 .658
2.4 .895 .841 .779 .744 .670
2.5 .902 .851 .788 .755 .680
2.6 .910 .859 .799 .764 .689
2.7 .916 .866 .810 .774 .697
2.8 .922 .873 .819 .783 .707
2.9 .928 .881 .827 .791 .719
3.0 .933 .889 .836 .798 .731
3.1 .938 .895 .644 .806 .743
3.2 .943 .901 .851 .815 .754
3.3 .948 .906 .857 .822 .760
3.4 .952 .911 .864 .828 .765
3.5 .955 .916 .870 .836 .770
3.6 .958 .920 .876 .843 .775
3.7 .962 .925 .882 .851 .780
3.8 .965 .929 .888 .857 .784
L*(Z /Z ) is the exact lower confidence bound for R(tO)
ab0
11-69
'.-4
S..
.....-... __ •. $_....$ _..,
. .-.• . ,_.•
. .-. -. ,,."....-..... . .. ..- •- .- .. ,..- .. '....... . . . . . . . "..--.-. .--....... ... "..-..-.....
n - 15
p a r/n - 7/15
75 9
y-0.50 y - 0. y - 0. 0 y o 0.95 y - 0.99
Z /Zb L*(Z /Zb) L*(Z/Zb) L*(Z /Zb) L*(Za/Zb) L*(ZC/Zb)
3.9 .967 .933 .894 .862 .789

4.0 .969 .937 .898 .867 .795
4.1 .972 .940 .902 .872 .801
4.2 .974 .944 .907 .876 .808
4.3 .976 .947 .911 .882 .814
4.4 .978 .949 .915 .886 .821
I4.5 .979 .952 .919 .891 .827 %.
4.6 .981 .955 .922 .896 .834
4.7 .982 .958 .926 .900 .840 I 4
4.8 .984 .960 .929 .905 .845 /% %
4.9 .985 .963 .932 .908 .850
5.0 .986 .965 .935 .912 .852
It bP 4-
tI*,...
.. . . . .. *
L*(ZaIZ b is the exact lower confidence bound for R(t0
11-70 .
TABLE VIII. TABLE OF EXACT LOWER CONIPIDENCE BOUNDS FOR R(tO) (Continued)
n - 15
p - r/n - 11/15 V.
50 9 99
- a 0. y-a 0.75 "y a 0.90 y-0. 5 " 0.
Z /Zb L*(Z /Zb) L*(Za/Zb) L*(Z&/Zb) L*(Z&/Zb)

ab
.0 .401 .328 .269 .233 .179
.1 .433 .362 .304 .262 .207
.2 .467 .395 .335 .297 .233
.3 .499 .427 .369 .325 .264
.4 .530 .458 .400 .355 .295
.5 .559 .488 .430 .383 .318
.6 .587 .516 .457 .412 .340
.7 .614 .544 .485 .440 .363
.8 .641 .573 .512 .468 .388
.9 .664 .598 .535 .492 .413
1.0 .688 .624 .557 .518 .434
1.1 .712 .646 .578 .539 .450
1.2 .732 .667 .599 .559 .469 .,,,
1.3 .752 .685 .622 .580 .490
1.4 .772 .705 .642 .601 .510
*1.5 .790 .724 .659 .620 .527
1.6 .807 .742 .676 .639 .539 ,
1.7 .823 .759 .691 .657 .560
1.8 .837 .775 .707 .672 .580. .
1.9 .851 .790 .721 .687 .599
2.0 .862 .805 .736 .700 .618
2.1 .873 .818 .752 .716 .634
2.2 .883 .830 .765 .729 .645 \"
2.3 .892 .842 .779 .742 .663
2.4 .901 .852 .792 .755 .675
2.5 .909 .861 .803 .768 .687
2.6 .917 .870 .813 .779 .702
2.7 .924 .879 .823 .790 .712
2.8 .930 .887 .832 .801 .722
2.9 .936 .896 .842 .812 .734
3.0 .941 .903 .852 .822 .746
3.1 .946 .910 .861 .829 .758
3.2 .951 .916 .869 .837 .769
3.3 .955 .922 .877 .845 .779
3.4 .958 .927 .883 .853 .789
3.5 .962 .932 .890 .861 .798
3.6 .965 .936 .896 .869 .806
3.7 .968 .941 .901 .875 .813
3.8 .971
bI
.945 .907
L*(Z a/Zb) is the exact lower confidence bound for R(tO)

.882 .819
.. '
11-71
MEN.2.. %° . . . . . . .
rim
n- 15
p - r/n - 11/15
y 0.50 y-0. 7 5 y" 0. 9 0 y- 0. 9 5 y 0.99
za/Z L*IZ /Z L*(Z//Zb) L*(Z//Zb L*(Z /Zb)
b a bbb
3.9 .973 .949 .912 .888 .825
4.0 .976 .953 .917 .893 .831
4.1 .978 .956 .922 .898 .837
4.2 .980 .959 .927 .903 .846
4.3 .981 .962 .932 .907 .854
4.4 .983 .965 .936 .911 .862
4.5 .984 .967 .939 .916 .870
4.6 .985 .969 .942 .920 .877
4.7 .987 .971 .946 .923 .882
4.8 .988 .973 .949 .928 .887
4.9 .989 .975 .952 .932 .892
5.0 .990 .977 .955 .936 .896
* t.
I -.
.. . .. . . .. . . . . . . . . . .
........................-............
. . . . . . . . . m
.. . . . . . . . . . . .
L*(ZaIZb is the exact lower confidence bc'nnd for R(co)
11-72
TABLE VIII. TABLE OF EXACT LOVER CONFIDENCE BOUNDS FOR R(tO) (Continued)
n - 15
p - rin - 1
y= 0.50 y a 0.75 yn 0.90 y a 0.95 Y 0.99
Za/Zb L*(Za/Zb) L*(Ze/Zb) L*(Za/Zb) L*(Za/7-) L*(ZA/Zb)
.0 .388 .321 .271 .234 .180

.1 .422 .355 .302 .261 .207
.2 .455 .388 .334 .286 .235
.3 .488 .419 .365 .317 .267
.4 .521 .452 .393 .345 .291
•5 .552 .482 .422 .376 .314
.6 .580 .513 .450 .403 .336
.7 .610 .540 .478 .430 .362
.8 .638 .568 .505 .461 .390
.9 .664 .595 .531 .489 .411 A *e..*
* 1.0 .688 .620 .556 .515 .428

1.1 .712 .645 .581 .539 .454
1.2 .734 .668 .602 .565 .471
* 1.3 .754 .690 .623 .588 .495
S1.4. .773 .711 .643 .612 .518,-"--
1.5 .791 .731 .663 .631 .539 KZ
1.6 .808 .748 .683 .653 .555
1.7 .823 .765 ,701 .670 .571 •..
1.8 .836 .781 .720 .687 .587
• 1.4 .851 .797 .738 .704 .605
"2.0 .864 .813 .755 .719 .623
2.1 .875 .826 .770 .732 .640
"2.2 .885 .839 .784 .747 .658
2.3 .895 .850 .798 .762 .673
2.4 .904 .861 .809 .776 .688
2.5 .912 .871 .823 .787 .703
2.6 .920 .880 .835 .799 .718
2.7 .927 .889 .845 .810 .732
2.8 .933 .897 .855 .821 .745
2.9 .939 .906 .865 .831 .758
3.0 .945 .912 .873 .840 .770
3.1 .950 .919 .882 .849 .782
3.2 .954 .925 .890 .857 .793
3.3 .958 .931 .897 .865 .802
3.4 .962 .936 .904 .874 .812
"3.5 .966 .940 .910 .881 .822
3.6 .968 .945 .916 .888 .830
3.7 .971 .949 .922 .895 .836
3.8 .974 .953 .927 .900 .844
L*(Za/Zb) is the exact lower confidence bound for R(to)
11-73
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(t0 ) (Continued)
n 15
p na
r/ *I
y 0.50 y-O .7 5 y aO. 9
0 y-O0.95 y-O0.9 9
Z /Zb L*(Za/Z) *(Z L/Z) / L* Za/Zb) L*(Z/Z

a b b aZb) & ba/b
3.9 .976 .957 .932 .906 .854
40.978 .960 .936 .911. .863
4.1 .980 .963 .941 .916 .871
*4.2 .982 .966 .945 .921 .818
4.3 .984 .969 .949 .926 .884
4.4 .985 .971 .952 .931 .887
4.5 .986 .973 .956 .935 .890
4.6 .988 .975 .959 .939 .893
4.7 .989 .977 .962 .942 .896
* L*(Z /Z )is the exact lower confidence bound for R(t )

a b0
11-74
* 2.ABLE VIII. TIAbLE OF EXACT LOWER CONFIDENCE bOUNDS FOk R(.t0 (Continued)
n - 20
p ftr/n - 5/20
y 0 .5 0 y -0.75 '- 0.9 0 y a0.95 y0.9
L /Zb L-(Z b) 1*(Za/Zb) L*(.Z/Zb) L*(Za/Zb) L*(Za/Zb)
.0 .523 .350 .188 .105 .008

.1. .544 .387 .228 .140 .020
.2 .569 .423 .271 .1.86 .037
.3 .596 .459 .305 .231 .068
.4 b•20 .491 .343 .274 .118 ...
S.5 .642 .524 .365 .31.7 .164
.• .663 .552 .427 .360 .229 I
.7 .683 .582 .465 .401 .287
b, .7U4 b 10 .501 .439 .331
".9 .720 .640 .539 .477 .380
1.0 .735 .bb5 .570 .519 .430
1.1. .754 .687 .601 .552 .464
1.2 .768 .706 b31 .583 .484
1.3 .783 .724 .656 .61•4 .519
1.4 .796 .742 .679 .037 .538
1.5 .809 .758 .698 .659 .55V
1.o .824 .775 .718 .674 .587
1.7 .835 .790 .735 .697 .b09
- 1.b .84b .602 .746 .713 .628
1.9 .358 .813 .761 .728 .644
2.0 .668 .823 .773 .741 .665
2.1 .878 .834 .785 .753 .077
2.2 .- 67 .842 .797 .762 .687
2.3 .895 .850 .807 .773 .695
2.4 .903 .857 .807 .791 .703
2.5 .909 .865 .824 .791 .713
".b .916 .873 •b31 .798 .720
2.7 .922 .879 .837 .803 .726
2.8 .928 .885 .b43 .809 .733
2.9 .933 .891 .851 .815 .745
3.U .938 .•96 .b57 •b23 .752
* 3.1 .942 .902 .865 .826 .755
- 3.2 .946 .907 .870 .633 .763
3.3 .950 .912 .875 .830 .768
3.4 .953 .917 .bO .ts44 .773 "
3.5 .956 .921 .•885 .850 .779
3.b•.959 .925 .189 .b54 .784
3.7 .962 .929 .893 .b5b .786
3.8 .9b5 .933 .698 bb62 .793
L*(C /Z.) is the exact ower confidence bouna for R(t )

a b 0
11-75
- ,. . . . . . . . . ..
% TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued)
n - 20
p - r/n - 5/20
y- 0.50 y- 0.75 y - 0.90 y- 0.95 y- 0.99
Za/Zb L*(Za/Zb) L*(Z /Z ) L*(Z /Z) L*(Za/Z ) L*(Za/Z )
3.9 .967 .936 .902 .866 .797

4.0 .969 .939 .905 .870 .802
4.1 .971 .942 .909 .873 .806
4.2 ,973 .945 .912 .878 .810
4.3 .975 .948 .915 .880 .815
4.4 .977 .951 .919 .883 .819
4.5 .978 .954 .922 .886 .824
4.6 .980 .957 .924 .889 .828"-
4.7 .981 .959 .927 .893 .832
4.8 .983 .962 .931 .898 .836
4.9 .984 .963 .933 .902 .839 4*""
5.0 .985 .965 .936 .904 .843
. .. ....................................
. . .. ....
. .. . . . . . . . . . . .. .. . . .. .. . .. . . .. . .. .. . . . . .
. .. . . . . . .. . . . . . .. .. . . . ..
L*(Z /Z )is the exact ower confidence bound for R(t )

abb
11-76
dA
an- 20
p - r/n - 10/20
y ",0.50 y a 0.75 y " 0.90 y - 0.95 y - 0.99
Z/Zab L*CZB/Zb) h*(Z al L*(Zal Zb) L*(Za/Zb) L*(ZalZb
.0 .415 .333 .257 .224 .141

"1. .0446 .369 .294 .263 .180
.2 .479 .405 .332 .297 .222
-. 3 .509 .441 .365 .332 .264
-. 4 .542 .472 .402 .367 .290
1.5 .573 .502 .436 .401 .336
1.6 .604 .530 .468 .432 .370
1.7 .632 .559 .497 .461 .404
"".8 .657 .588 .530 .492 .432
1.9 .681 .614 .559 .520 .453
1.0 .704 .637 .584 .546 .474
2.1 .725 .659 .608 .571 .495
1.2 .746 .680 .633 .591 .516
"1.3 .764 .701 .652 .614 .534
.14 .780 .721 .672 .637 .546
1.5 .796 .740 .691 .657 .564
1.6 .811 .758 .707 .673 .588
2.7 .924 .877 .821 .815 .073
2.0 .862 .817 .764 .737 .660

2.1 .872 .829 .777 .750 .676 1-
2.2 .882 .840 .789 .761 .681
2.8
3.3 .839
.892 .850
.790 .801
.837 .772
.8210 .692
.628
2.4
2.9 .9350
.900 .804
.860 .851
.812 .8230
.782 .749
.702
"325 .909 .869 .823 .793 .712
'-2.7 "2.6 .916
.922 .878
.887 .832
.842 .8
.815 .80
722
.731
"2.8 .929 .894 .851 .821 .740

2.9 .935 .901 .859 .830 .749 -.
3.0 .940 .907 .868 .839 .758
S3.1 .945 .913 .875 .848 .766
3.2 .950 .918 .883 .855 .774
"'•3.3 .954 .924 .890 .862 ,782
L*(Z /Zb) is the exact lower confidence bound for R(t.)
"."
3.4.958 99 .9 .869..792
11-77
n -20
p = r/n - 10/20
y - 0.50 y- 0. 7 5 y- 0.90 =0.95 y - 0. 9 9
Z /Zb L*(Za/Zb) L*(Z /Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb)
3.9 .973 .949 .923 .898 .833

4.0 .975 .953 .928 .903 .840
4.1 .977 .956 .932 .908 .848
4.2 .979 .959 .936 .912 .854
4.3 .981 .961 .939 .916 .860
4.4 .982 .964 .943 .920 .864
4.5 .984 .967 .946 .925 .868
4.6 .985 .969 .949 .929 .871
4.7 .987 .971 .952 .933 .875
4.8 .988 .973 .955 .936 .879
4.9 .989 .975 .957 .940 .882
5.0 .990 .976 .960 .943 .886
L*(Za/Zb) is the exact lower confidence bound for R(to)
11-78
S.... . -- . .. . . ..- . .--- "."--'-.. ..-.---... . "-.-- - -'.-..- --.- "" '-:- - -" "'. - ."- .
• °. ° ° . . .- .. . . . . . . .° -. .... -. . .. .... .-.. .. .-. "- .-. .° .... " .. .° -. ° .. °° .° -. ° .- " '. ' % "
FOR R(t0 ) (continued)

TABLE VIII. TABLE OF EXACT LOWER CON4FIDEN4CE BOUNDS
n 20
p r/ af 15/20
y -0.75 y a 0.90 y-Y 0.95 y -0.99
-0.50
L(aZ)L*(Za/ZL) L*r(Z aIZ b) L*(Z a/Z b) L*(Za/Zb)

z /Z b
.278 .240 .192AI
.0 .390 .332
.1.424 .367 .310 .272 .221.- -h-
.342 .306 .253
.2 .457 .400
.374 .338 .279
.3 .492 .433
.408 .373 .307
.4 .524 .466
.5 .558 .496 .438.4832
.6 .589 .523 46.39.387.
.497 .466 .8
.7 .617 .550 .412
.575 .528 .489
.8.645
.9 .671 .603 .554 .516 .43659.*.
.629 .8 50.5
1.0 .697
.605 .565 .482
1.1 .720 .655
.629 .589 .503
1.2 .741 .676
.719 .668 .634 .544.
1.4 .780
.687 .654 .564
1.5 .797 .73
.705 .672 .584
1.6 .812 .756
.723 .690 .603 ,'...J
1.7 .827 .772

.788 .739 .707 .621
1.8 .840 .639
.803 .755 .723
1.9 .853 .656
.816 .770 .738
2.0 .866 .670 ~...*.
.830 .784 .753
2.1. .877 .765 .688
2.2 .887 .842 .797
.854 .810 .778 .704
2.3 .896 .719
.864 .821 .792
2.4 .905 .733
.875 .832 .805
2.5 .913 .746
.883 .842 .817
2.6 .920
.851 .826 .757
2.7 .927 .892
.899 .860 .835 .769
2.8 .933
.869 .845 .780
2.9 .939 .9o7
.877 .856 .790
3.0 .944 .914
.884 .863 .800
3.1. .949 .920
.891 .870 .809
3.2 .953 .925
.898 .877 .818
3.3 .958 .930
.905 .884 .827
3.4 .961 .936
.911 .890 .835
3.5 .964 .940
.916 .895 .843
3.6 .967 .945
.921. .902 .851
3.7 .970 .949 :
.927 .908 .858
3.8 .973 .953
L.(Z/Z)is the exact lower confidence bound for R(t)

L*Za/Zb0
11-79................................
IIIEI*~IE
III *hi~~iiism*~
n - 20
p - r/n - 15/20
y " 0.50 7 " 0.75 y a 0.90 y a 0.95 y = 0.99
3.9 .975 .956 .932 .914 .865

4.0 .977 .960 .936 .919 .872
4.1 .979 .963 .939 .923 .878
4.2 .981 .966 .943 .928 .884,..-
4.3 .983 .968 .947 .931 .890
4.4 .985 .971 .950 .935 .895
4.5 .986 .973 .954 .939 .900
4.6 .987 .975 .957 .942 904
4.7 .988 .977 .959 .945 .907
4.8 .989 .979 .962 .948 .910-
4.9 .990 .980 .964 .951 .915
5.0 .991 .982 .967 .954 .919
p'ii
L*(Z /Z )is the exact lower confidence bound for R(t,.'

ab
'1o'
11-80
. ... . . . . . ....--.
....
n - 20
p r/nl -
y- 0.50 y- 0.75 y- 0.90 y 0.95 y 0.99
ZI/Zb L*(Za/Zb) L*(Za/Z) L*(Z /Z) L*(Z /Zb) L*(Za )
aaZb) a b a b a/~Z
.0 .380 .328 .278 .249 .202
.1 .416 .363 .309 .283 .228
.2 .451 .395 .338 .313 .256
.3 .487 .429 .370 .348 .281
.4 .519 .462 .403 .379 .312
.5 ,553 .491 .433 .406 .338
.6 .586 .521 .464 .434 .364
.7 .613 .548 .494 .461 .388
.8 .642 .575 .525 .490 .413
.9 .668 .603 .552 .514 .436
1.0 .693 .629 .579 .541 .462
1.1 .7.-7 .654 .605 .568 .486
1.2 .740 .679 .628 .593 .510
1.3 .761 .700 .649 .615 .533
1.4 .781 .722 .670 .636 .557
1.5 .799 .742 .692 .657 .579
1.6 .816 .761 .713 .680 .600
1.7 .831 .779 .732 .700 .621
1.8 .846 .795 .748 .721 .640
1.9 .859 .810 .763 .737 .661
2.0 .871 .824 .778 .755 .681
2.1 .883 .837 .792 .768 .700 -
2.2 .893 .849 .806 .785 .719
2.3 .902 .861 .819 .798 .732
2.4 .911 .872 .830 .811 .745
2.5 .919 .881 .842 .823 .758
2.6 .926 .890 .852 .834 .772
2.7 .933 .898 .863 .844 .785
2.8 .939 .906 .873 .854 .796
2.9 .944 .914 .881 .862 .807
3.0 .949 .920 .889 .872 .817
3.1 .954 .926 .896 .881 .827
3.2 .958 .932 .904 .888 .836 LA
3.3 .962 .938 .911 .895 .844
3.4 .965 .943 .917 .901 .853
3.5 .969 .947 .923 .907 .861
3.6 .971 .9r1 .928 .914 .869
3.7 .974 .955 .933 .920 .877
3.8 .976 .959 .938 .925 .8R5
L*(Z /Zb) is the exact lover confidence bound for a(,)

a b
11'o-
-,-....-........-.....,...-.....-..- , _.. . .. . .. - .. . . . . - - - . . . . . . . .....-

-,,-<~~............ ..... ...-........-
. ...-.-.-.
' . ....
: ...-. °.. .... ... . .. .- -- .: ...- : .-. -:..- -, .- .- : ..- , -
• _.,• •;,• :.:
.- -:.::.--.-::
... .. ........ ::-.:, - : : .:- ,. ::- . -:. :.-: :. :. : -:-" ::.:: " - -:.--.. :- ::.: .:::. . - .... .:
II
n- 20
p - i/n - 1
y 0.50
0 y - 0.75 y - 0.90 y a 0.95 y - 0.99
I(
Za/Zb L*(Z/Zb) L*(Z/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb)
a. a/ .96a/ a9 .93a|
3.9 .979 .962 .942 .930 .893
4.0 .981 .965 .947 .935 .898
4.1 .982 .968 .951 .939 .903
4.2 .984 .971 .954 .943 .908-
4.3 .985 .973 .957 .947 .913
4.4 .987 .975 .960 .950 .918
4.5 .988 .977 .963 .953 .923
4.6 .989 .979 .966 .956 .927
4.7 .990 .981 .968 .959 .931
4.8 .991 .982 .970 .962 .935
4.9 .992 .984 .972 .964 .938
5.0 .993 .985 .974 .967 .942
4.4
ri' . 5' -. ,
o.S..
*L*(Z a/Z b Is the exact lower confidence bound for R(t ) '. y~p
.................................................................................
11-82
n - 30
p - r/n "7/30
, 0.50 y a 0.75 - 0.90 y - 0.95 y - 0.99
Z /Z/b L*(Z a/Z ) L*(Z /Z ) L*(Z a/Z ) L*(Z /Z ) L*(Z /Zb)
.0 .485 .335 .198 .104 .015

".1 .512 .372 .238 .136 .033
".2 .542 .408 .282 .179 .060
.3 .567 .442 .323 .220 .098
.4 .592 .477 .367 .268 .143
.5 .615 .513 .408 .322 .188
.6 .637 .543 .449 .361 .237
.7 .66l .576 .486 .409 .289
.8 .b83 .b04 .520 .456 .346
.9 .705 .630 .555 .501 .397
".723 .bSb .590 .544 .442
1.. .742 .689 .619 .577 .482
.. 2 .759 .704 .643 .604 .515
1.3
1 .,4 .776
•.793 .726
•.745 .669
.688 .631
.0.5 .552
.589 •
1.5 .809 .761 .711 .676 .608
"1. .823 .777 .729 .694 .033
1.7 .836 .794 .746 .716 .659
1.6 .847 .807 .761. .735 .669
"1.9 .658 .819 .774 .752 b691.
"" 2.0 .869 .831 .767 .765 .712
"2.1 .878 .841 .799 .777 .732
"2.2 .87 .b51 .610 .787 .741
2.3 .895 .860 .820 .79S .750
2.4 .903 .869 .629 .809 .762
2.5 .911 .877 .839 .820 .773
2.6 .918 .885 .646 .829 .783
2.7 .925 .892 .855 .835 .791
2.6 .930 .b98 .8b2 .b43 .b00
2.9 .935 .903 .868 .851 .809
3.0 .940 .909 .074 .857 .816
3.1
3.2
3.3
.945
.94
.952
.914
.920
.925
.881
.- 86
.892
.864
.869
.874
.824
.831.
.838
f
3.4 .95o .929 .896 .878 .842
3.5 .959 .933 .901 .862 .846
K 3.o .9b2 .938 .907 .886 .849
* 3.7 .965 .941 .911 .891 .855
.968 .945 .916 .895 .b60
L*(.ka/L. ) is the exact iower coniidence bound ftr R(tO)

b 0
** , .. - -
'.; .... ..-. .-........-.

..............-... "... -.. ......-..-.-..-. ,...'..., -..-. ,.
S. ... .......
. . . .. . .. ,,-,..,, ., ,, . - • _ .. - .- .. . . .. . .. ..,'.. .' . .- . ..... .. . . -. -. .... .
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(c.O) (Continuea)
n m 30
p u r/n - 7/30
0
y - . 50 y - 0. 75
y - 0. 9 0 y - 0. 9
5 y = 0.99
z /Z b L*(Z /Z ) L*(Z /Z') L*(Za/4)b L*(Z I/Z) L*(Z /• )
3.y .970 .948 .920 .898 .865

4.0 .972 .951 .923 .902 .870
4.1. .974 .954 .926 .906 .875
4.2 .977 .957 .929 .910 .879
4.3 .978 .960 .932 .914 b882
4.4 .980 .962 .935 .918 .885
4.5 .981 .964 .939 .920 .886
4.b .983 .966 .942 .923 .888
4.7 .984 .968 .944 .925 .890
4.6 .985 .970 .946 .928 .892
4.9 .986 .972 .949 .931 .894
5.0 .988 .974 .951 .933 .895
•*..,-. .,.,•
11-84 ;'.--'.
S-....o. .,
ab 0 .• . ". *
4 .. .. .4 . . . - . .4 . 4- .4 ..
1. -4 .
. 4",
t41
I,
"TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Continued)
n- 30
p m r/n - 15/30
y 0.50 y -0.75 y - 0.90 y - 0.95 y - 0.99
Z/Z L*(Z/Z, L*(ZZa/Z' ) L*(Z /Z) L*(Z /Z ) L*(Z /Zb)
aD a/Z)ab a b abb
.0 .399 .329 .269 .230 .169
-. .1 .432 .369 .303 .270 .212
-• .2 .464 .402 .341 .309 .252
.3 .496 .437 .375 .348 .292
.4 .527 .470 .412 .384 .330
.5 .560 .501 .445 .421 .368 '- .• -
.6 .591 .534 .481 .456 .403
.7 .619 .563 .513 .487 .434
"" .8 .646 .593 .541 .514 .463 ý4:
.9 .673 .618 .571 .542 .490 , .
1.0 .696 .643 .597 .567 .518
1.1 .719 .667 .620 .595 .540
1.2 .740 .689 .645 .619 .562
1.3 .760 .712 .664 .640 .582
1.4 .778 .730 .685 .660 .603
1.5 .796 .748 .706 .677 .624
1.6 .812 .766 .723 .693 .643
1.7 .826 .782 .740 .711 .660 ..- '
.840 .798 .756 .729 .676
1.8
1.9 .853 .811 .771
.784 .746
.759 .688
.701 ,'z~x.e
• ,2 .0 .865 .825
Al 2.1 .877 .837 .797 .773 .721 •.,
2.2 .886 .849 .810 .787 .735
2.3 .895 .859 .820 .800 .749
2.4 .903 .871 .832 .811 .760
2.5 .910 .879 .841 .824 .769
2.6 .918 .887 .652 .834 .779
2.7 .925 .895 .861 .845 .789
2.8 .931 .903 .870 .853 .798
2.9 .936 .910 .878 .861 .805
3.0 .941 .917 .885 .869 .813 CA
3.1 .946 .922 .891 .875 .821

3.2 .950 .928 .898 .883 .827
3.3 .955 .933 .904 .889 .834 ." .'
3.4 .959 .938 .910 .895 .840
3.5 .962 .942 .916 .900 .847
3.6 .965 .947 .922 .906 .854
3.7 .968 .951 .927 .911 .861 MI"
"3.8 .971 .954 .932 .917 .868
L*(Z a/Z b) ia the exact lower confidence bound for R(t 0 )

slim
• .. ° . - ...... "
0
n-3
p a r/n a 15/30
y- 0.50 y- 0.75 y- 0.90 y- 0.95 y- 0.99
Za/Z b L*(Za/Zb L*(Za/Zb) L*(Z/Zb) L*(Za/Zb) a L*(Z /Zb)
3.9 .974 .958 .936 .921 .873

4.0 .976 .961 .940 .926 .877
4.1 .978 .964 .943 .930 .883
4.2 .980 .966 .947 .934 .889
4.3 .981 .969 .950 .938 .895
4.4 .983 .971 .953 .942 .899
4.5 .984 .973 .956 .945 .904
4.6 .986 .975 .959 .949 .908
4.7 .987 .977 .961 .952 .912
4.8 .988 .979 .964 .954 .916
4.9 .989 .980 .966 .957 .920
5.0 .990 .982 .968 .959 .924
*% *
L*(Z /Z )is the exact lower confidence bound for R~t

a b
*
11-86
P..-
.t" *. *- °"
.,*
• ... ,.* ,,.......-......-. .,,-- '• .. . ." ,.,--.... ..... . . . .*.,....... -,... .. _ ,.-,". .... '..- ..
TABLE VIII. TABLE OF EXACT LOWER GUNF1DEVCE BOUNDS FOR R(tO) (Continued)
n- 30
p , r/n - 22/30
7 90 9 9
yO0.50 yO0. 5 - 0. y 0. 5 y-0. 9
a /Zb L*(Za/Zb L*(Za/Zb L*(Zb/Zb L*aZa/Zb) L*(Z /Z b
.0 .382 .333 .281 .257 .206 . Ir
.1• .416 .366 .318 .290 .243

.2 .453 .400 .353 .325 .277
.3 .489 .433 .385 .359 .307
.4 .521 .405 .419 .392 .341,
.5 .553 .497 .449 .424 .373
. .5b3 .528 .479 .455 .406
.7 .b12 .560 .512 .486 .438
.8 .b4U .589 .540 .517 .465
.9 .666 .615 .566 .544 .486
1.u .g91. .b4l .593 .569 .511.
1.1 .71o .665 .618 .592 .534
1.2 .737 .bb7 .642 .615 .556
1.3 .758 .709 .664 .637 .577
1.4 .777 .?.9 .685 .059 ,b02
1.5 .795 .748 .705 .677 .623
.0 .dl2 .7bb .722 .b97 .641
1.7 .827 .782 .741 .717 .656
1.6 .842 .798 .758 .734 .675
1.9 .855 .813 .774 .749 .692
2.u .s67 .826 .7d8 .765 .706
2.1. .878 .840 .603 .778 .718
2.2 .686 .652 .816 .793 .732
2.3 .897 .864 .828 .807 .746
2.4 .906 .d75 .839 .bid .759
2.5 .914 .884 .851 .829 .771"
2.b .921 .894 .862 .839 .783
2.7 .928 .901. .871 .850 .795
2.d .935 .909 .880 .859 .806
2.9 .940 .916 .888 .867 .81 b
3.0 .945 .923 .896 .676 .826
3.1I .950 .929 .902 .884 .836
3.2 .955 .934 .909 .691 .645
3.3 .95b .940 .915 .898 .853
3.4 .9022.44 .921 .905 .861
3.5 .966 .949 .927 .911 .868
3.o .909 .953 .932 .917 .874
3.7 .971 .956 .937 .922 .881
3 .6 .974 .960 .941 .927 .58•
L*( a /Z)b is the exact lower confidence bound for R(t 0
I T-M7
--a-•.,,•,.,,•=,•.,.............,.........,......................-.."..-.. . ....-... ,...................-.......................-.......,.'..........-..................."

S-30
p r/n - 22/30
9
y- 0.50 y- 0.7 5 y- 0. 9 0 ya 0.95 y -0.9
3.9 .976 .963 .945 .931 .892

4.0 .979 .966 .948 .935 .897
4.1 .981 .969 .952 .939 .902
4.2 .982 .971 .956 .943 .907
4.3 .984 .973 .959 .947 .912
4.4 .985 .975 .961 .950 .916
4.5 .987 .977 .964 .953 .921
4.6 .988 .979 .967 .956 .926
4.7 .989 .981 .969 .959 .930 "
4.8 .990 .982 .971 .962 .934 94ý
4.9 .991 .984 .973 .964 .938
5.0 .992 .985 .975 .966 .942
til

a b
11-88
.........................................................
TABLE VIII. TABLE OF EXACT LOWER CONFIDENCE BOUNDS FOR R(tO) (Coatinued)
a a 30
p - rn -30
y - o.so - 0.75 ,- 0.90 y - 0.95 0.99
za/ZD L*(Za/Zb) L*(Za/Z b) L*(Za/Z b) L*(Z I/Z ) *Z/b

.0 .376 .326 .283 .265 .222
.1 .413 .361 .317 .297 .255
.2 .450 .394 .349 .328 .286
.3 .483 .429 .381 .362 .317
.4 .516 .462 .413 .394 .350
.5 .550 .494 .446 .426 .380
.6 .380 .526 .478 .455 .409
.7 .610 .556 .507 .485 .436
.8 .638 .585 .536 .515 .463
.9 .665 .613 .565 .544 .490
"1.0 .691 .639 .593 .570 .515
1.1 .715 .664 .618 .593 .543
1.2 .738 .688 .641 .618 .566
1.3 .759 .710 .665 .639 .587
1.4 .780 .730 .6d7 .662 .608
1.5 .798 .751 .708 .683 .629
1.6 .815 .770 .728 .703 .648
"1.7 .830 .787 .746 .721 .667
1.8 .845 .803 .764 .739 .685
1.9 .858 .818 .781 .755 .706
2.0 .871 .832 .797 .772 .724
2.1 .882 .844 .811 .788 .741
2.2 .892 .857 .824 .802 .758
2.3 .902 .868 .836 .815 .773 ,• -:"
2.4 .910 .879 .849 .827 .786 "
2.5 .919 .888 .860 .839 .797
"2.6 .926 .897 .870 .851 .808
2.7 .933 .906 .879 .861 .819
2.8 .939 .913 .888 .870 .829
2.9 .944 .920 .897 .879 .838
3.0 .949 .926 .904 .887 .847
3.1 .954 .932 .911 .895 .858
3.2 .958 .938 .917 .903 .864
3.3 .962 .943 .923 .910 .873
3.4 .965 .948 .929 .916 .880
3.5 .968 .952 .934 .922 .888
3.6 .971 .956 .939 .927 .895
3.7 .974 .960 .943 .933 .902
3.8 .976 .963 .948 .938 .908
L*(Z /Zb) is the exact lower confiderce bound for R(t 0 )
11-89
*. .-.. -......-... . . . . .
. . . . . . . . . ". . . . ..
77-77
. , ,,, -
•• - % .- . . .. . . . . . .. . .
2 ~Lj~ t .'aC'. _. ,•aa.
. .°° -..
n-3 0
p . r/n - 1
y - 0.5 0
y - 0.75 y a 0. 9 0 y - 0. 9 5 y - 0.99
Za/Zb Zb L*(Za/Zb) L*(Za/Zb) L*(Za/Zb) L*(Za/Zb)

3.9 .978 .966 .952 .942 .914
4.0 .980 .969 .955 .946 .920
4.1 .982 .971 .958 .950 .925
4.2 .984 .974 .961 .954 .929
4.3 .985 .976 .964 .957 .934
4.4 .987 .978 .967 .960 .938
4.5 .988 .980 .969 .963 .942
4.6 .989 .981 .972 .966 .945
4.7 .990 .983 .974 .968 .949
4.8 .991 .984 .976 .971 .952
4.9 .992 .986 .978 .973 .955
5.0 .992 .987 .979 .975 .958
r,...-°
. .. . .. .- °
L*(Z /Z )is the exact low.er confidence bound for R(t0 ) . . .,o . o .
a b .......................
11-90
TABLE VIII. TABLE OF EXACT AND ASYMPTOTIC LOWER CONFIDENCE

BOUNDS FOR R(%0) (Continued)
n 50
p a r/n - 12/50
9 99
* 0.50 y - 0.75 y - 0.90 y - 0. 5 y"0.
L*( ±) L * L* LA L* LA*4 L* LAY) L* ) LA*
.0 .441 .368 .320 .252 .207 .15: .163 .111 .080 .048
.1 .47 .405 .359 .292 .249 .197 .203 .147 .119 .073
.2 .504 .441 .397 .334 .292 .240 .246 .188 .155 .105
3 .5.2 .477
.. .433 .375 .337 .284 .291 .231 .196 .144
.4 .560 .512 .467 .417 .380 .329 .335 .277 .240 .187 S;
5 •5b ..
•545 .503 .45b .427 .375 .379 .324 .291 .233 '. ',,
.6 .614 .578 .537 .497 .469 .419 .419 .372 .350 .282
.7 •b3b .b09 .567 .535 .508 .463 .463 .418 .397 .332
.8 .b64 .638 .600 .571 .545 .504 .505 .462 .454 .381
.9 .bo9 .65b .632 .605 .580 .543 .543 .505 .494 .428
1.0 .711 .092 .661 .636 .614 .580 .577 .544 .536 .47G
1.1 .734 .717 .687 .666 .642 .614 .609 .581 .571 .514 • ' .'
"1.2 .754 .740 .710 ."93 .669 .645 .637 .614 .600 .552
j..J .774 .761 .733 .718 .693 .674 .665 .645 .626 .586
1.4 .792 .761 .753 .741 .715 .700 .689 .673 .650 .617
* 1.5 .608 .d00 .772 .762 .736 .723 .710 .098 .674 .644
1.6 .823 .817 .790 .781 .754 .744 .727 .720 .696 .669
* 1.7 .o37 .633 .805 .799 .770 .764 .746 .740 .711 .691
1.8 .649 .84b .619 .815 .787 .781 .762 .758 .724 .710 '
1.v .bal .601 .s32 .s30 .603 .797 .777 .775 .738 .727 '-,
2.0 .673 .873 .844 .44 .816 .611 .791 .789 .750 .742 . .
2.1 .bo3 .•85 .855 .b56 .828 .825 .805 .603 .7b2 .756
2.2 .•92 .895 .865 •bb7 .839 .837 .817 .815 .777 .769
2.3 .9Ou .9U5 .875 .b7d .b50 .848 .828 .b27 .787 .780 ..
S2.4 .906 .913 .683 .887 .860 .858 .540 .837 .796 .791 •,,',.,,
y .,5 .915 .921 .891 .o96 .868 .867 .851 .647 .805 .b00
2.0 .922 ,928 -698 .904 .876 .876 .860 .856 .813 .809
2.7 .929 .935 .905 .912 .863 .684 .866 .804 .823 .817
* 2.5 .935 .941 .912 .919 .891 .892 .874 .872 .- 31 .825
2.Y .940 .946 .916 .925 .897 .699 .881 .b79 .840 .632
3.0 .945 .951 .924 .931 .903 .905 .888 .686 .848 .839
S3.1 .949 .95b .929 .936 .908 .911 .894 .692 .656 .846
3.2 .954 .960 .934 .941 .914 .917 .900 .898 .863 .852
3.3 .957 .964 .938 .946 .919 .922 .905 .904 .868 .668
L*(Z /Z ) is the exact lower confidence bound for R(t )

a b 0
L *(2. /4) is the asymptotic lower confi.dence bound for R(0)"
A a b 0
11-91 ..................................................
".....................................".-...--...'.-.-'.'--..........................................................-.,---....,-...._-.....,...-.,-.-..

BOUNDS FOR R(to) (Continued)
n - 50
p - r/n - 12/50
_z
a L
y - 0.50
La- A L*
A,(/Z / z / /
y - 0.75
LA L*
y -
LA*
0.90 y = 0.95
L.t L*
Y - 0.99
Zb
b \ZLAA\Zb/\VZ-b/A\b
72 J b LAz kZb2 LAZb
bb
LZA\b
.942 .950 .923 .927 .909 .909 .873 .863 IN,

3.4 .961 .967
3.5 .964 .970 .946 .954 .928 .932 .914 .914 .878 .868
3.6 .967 .973 .950 .957 .932 .936 .919 .919 .883 .873
3.7 .970 .976 .953 .961 .936 .940 .922 .923 .888 .878
3.8 .972 .978 .957 .964 .940 .944 .925 .927 .892 .882
3.9 .975 .980 .960 .967 .943 .947 .930 .931 .898 .887 ,'-. -
4.0 .977 .982 .962 .969 .946 .951 .934 .935 .903 .891
4.1 .979 .984 .965 .972 .949 .954 .937 .938 .908 .895
4.2 .981 .985 .967 .974 .952 .957 .940 .942 .912 .899
4.3 .982 .987 .969 .976 .955 .959 .944 .945 .917 .902
4.4 .984 .988 .971 .978 .957 .962 .947 .948 .921 .906
4.5 .985 .989 .973 .980 .960 .964 .949 .951 .924 .909
4.6 .986 .990 .975 .981 .962 .967 .951 .953 .926 .913
4.7 .987 .991 .977 .983 .965 .969 .954 .956 .928 .916
4.8 .988 .992 .979 .984 .967 .971 .957 .958 .931 .919
4.9 .989 .993 .980 .985 .969 .973 .959 .960 .934 .922
5.0 .990 .993 .982 .988 .971 .974 .961 .962 .937 .924
L*...
Z)is the exact lower confidence bound for R(c,.)

LA *Za /Zb) is the asymptotic lower confidence bound for R( t0
. . . . W,¢
L*(Za/•b) is .
he...a.............den......d.....R...........-.
.-*r
11-92
4 * .* .. *°
- - . ____-.•_-_o____ ._____________.____•"
Y 0.50 v - 0.75 y - 0.90 y - 0.95 y - 0.99
L L(bA ) L* La* L LA* L* L

.1~~b
L a A* --IA(
bb ZT 4 )\Zb/bA\Z
.0 .383 .368 .333 .313 .286 .264 .258' .236 .209 16
.1 .419 .405 .368 .351 .325 .303 .297 .275 .249 .224
.2 .456 .441 .406 .389 .363 .342 .336 .315 .288 .263
.3 .488 .477 .444 .427 .399 .382 .377 .354 .328 .303
.4 .522 .512 .477 .464 .436 .420 .414 .393 .368 .342
.5 .555 .545 .509 .499 .471 .457 .451 .430 .408 .381
.6 .586 .578 .543 .534 .507 .492 .483 .466 .441 .417
.7 .615 .600 .573 .566 .540 .526 .515 .500 .475 .452
.8 .. 643 .638 .603 .597 .569 .557 .546 .533 .502 .484
.9 .670 .666 .630 .626 .596 .587 .575 .563 .534 .515
1.0 .695 .692 .658 .653 .622 .615 .601 .591 .560 .544
1.1 .718 .717 .682 .679 .647 .641 .626 .618 .584 .570
1.2 .741 .740 .705 .703 .672 .666 .651 .642 .605 .595 " .. •.
1.3 .761 .761 .727 .725 .693 .689 .673 .666 .627 .618
1.4 .780 .781 .746 .746 .714 .710 .692 .687 .649 .640
1.5 .798 .800 .766 .766 .732 .730 .711 .707 .669 .660
1.6 .816 .817 .783 .784 .750 .749 .726 .726 .6P8 .679
1.7 .830 .833 .799 .800 .767 .767 .745 .744 .706 .697
1.8 .845 .848 .815 .816 .784 .783 .761 .761 .724 .714
1.9 .857 .801 .829 .831 .800 .798 .776 .76 .741 .729
2.0 .869 .873 .842 .844 .813 .812 .789 .791 .757 .744
2.1 .880 .885 .854 .856 .827 .825 .804 .804 .770 .758 .. ,
2.2 .890 .895 .865 .868 .839 .838 .816 .817 .781 .771
2.3 .900 .905 .875 .878 .850 .849 .828 .829 .793 .784
2.4 .908 .913 .885 .888 .861 .860 .839 .840 .803 .796
2.5 .916 .921 .893 .897 .871 .870 .849 .850 .814 .807
2.6 .923 .928 .901 .906 .879 .879 .359 .860 .823 .817
2.7 .930 .935 .909 .913 .888 .888 .867 .869 .834 .827
2.8 .936 .941 .916 .920 .896 .896 .875 .878 .843 .836
2.9 .942 .946 .922 .927 .903 .903 .884 .886 .851 .845
3.0 .947 .951 .928 .933 .910 .910 .891 .893 .859 .854
3.1 .951 .956 .934 .938 .916 .917 .898 .900 .867 .862
3.2 .956 .960 .939 .943 .922 .923 .905 .907 .874 .869
3.3 .960 .964 .944 .948 .927 .928 .911 .913 .880 .876
"L*(Z /Zb) is the exact lower confidence bound for (t(0 )

"L*(Z /Zb) is the asymptotic lower confidence bound for R(tO) .
A a b 0
11-93 " ""
". " ". " . "... . " " " • r'" .. • ,• •,,, '•"•'• •' '•-•.,,••r... .. **"
'**0•,

BOUNDS FOR R( 0 ) (Continued)
n 50
p - r/n - 25/50
y - 0.50 y * 0.75 y - 0.90 v a 0.95 y - 0.99
a) A
zb
-(\z/
SL*(a) L f(. L A*( L*(Z L \Zb/'a) L L Cb \7/b
( LA*
.. 4 .9b3 .967 .948 .952 .932 .933 .917 .919 .887 .883
3.. .966 .970 .952 .956 .937 .938 .923 .924 .894 .889
3.6 .969 .973 .956 .960 .941 .943 .928 .929 .901 .895
3.7 .972 .97b .959 .963 .946 .947 .933 .934 .907 .901
3.8 .975 .978 .963 .966 .949 .951 .937 .938 .913 .907
3.9 .977 .980 .966 .969 .953 .954 .941. .942 .918 .912
4.0 .979 .962 .969 .972 .956 .958 .945 .946 .922 .917
4.1 .961 .984 .971 .974 .959 .961 .949 .950 .927 .921 4
4.2 .983 .985 .973 .976 .962 .964 .953 .953 .931 .925
4..• .964 .987 .975 .978 .965 .966 .956 .956 .935 .930
4.4 .986 .988 .977 .980 .967 .969 .959 .959 .940 .933
4.5 .967 .989 .979 .962 .969 .971 .962 .962 .943 .937
4.6 .988 .990 .961 .ý83 .972 .973 .964 .965 .947 .941.
4.7 .90 .991 .962 .985 .973 .975 .967 .967 .950 .944 , ,,
4.8 .990 .992 .984 .986 .975 .977 .969 .969 .953 .947
4.9 .991 .993 .985 .987 .977 .979 .971 .971 .955 .950 *r
5.0 .992 .993 .986 .988 .979 .980 .973 .973 .958 .953
L*(. /Z ) is the exact lower confidence bound for R(t

a b 0o
L *k/-a/••) is the asymnptotic lower confidence bound for R(r)
A a b 0
11-94
. .. . . .. .. .. . . . . . . . . . . . . . . .-. .-.-.-
......... .-... .. ..
.. . .. . -. .... . . . . . . . . .
"~~~~~~~~~~~~~.
. ......
" ". "-. . . ... . "... -"- .. •" .".-.".".'.'.'..'."...'.-.'..,--. -'.
TABLE VIII. TABLE OF EXACT AND ASYMPTOTIC LOWER CONFIDENCE 4 V.

BOUNDS FOR R(to) (Continued) i.. " -
• =50
p rin =3/0
y a 0.50 y - 0.75 -Y 0.90 y - 0.95 Y - 0.99
L* LAa* L LA*a L * L
L& LA*
Zb Zib) b ý Cb Tbb b 7) Z )
.0 .377 .368 .335 .327 .304 .291 .283 .270 .252 .231
.1 .413 .405 .371 .364 .339 .327 .317 .305 .291 .265
.2 .448 .441 .406 .400 .373 .363 .352 .341 .328 .300 --
.3 .483 .477 .442 .436 .409 .399 .387 .376 .358 .334
.4 .517 .512 .477 .471 .441 .434 .422 .411 .389 .368
.5 .549 .545 .510 .505 .475 .467 .457 .444 .422 .401 e.",•
.6 .580 .578 .542 .538 .506 .500 .489 .477 .452 .432
.7 .610 .609 .572 .569 .537 .531 .521 .508 .478 .463
.8 .639 .638 .601 .599 .567 .561 .548 .538 .506 .492 VI.
.9 .666 .666 .630 .627 .596 .590 .574 .566 .533 .520
1.0 .692 .692 .657 .654 .621 .617 .601 .593 .555 .546
1.1 .717 .717 .681 .679 .645 .642 .626 .619 .575 .572
1.2 .789 .740 .704 .703 .669 .666 .649 .643 .599 .596
1.3 .760 .761 .726 .725 .692 .689 .671 .666 .625 .619
1.4 .780 .781 .746 .746 .713 .711 .691 .688 .645 .641
1.5 .798 .800 .765 .766 .733 .731 .712 .708 .665 .661
1.6 .815 .817 .783 .784 .751 .750 .731 .727 .687 .681
1.7 .831 .833 .799 .801 .768 '.768 .748 .746 .708 .699
1.8 .845 .848 .816 .817 .785 .784 .765 .763 .728 .717
1.9 .858 .861 .830 .831 .800 .800 .780 .779 .743 .734
2.0 .870 .873 .844 .845 .815 .814 .795 .794 .755 .750
2.1 .882 .885 .857 .857 .828 .828 .809 .808 .768 .764
2.2 .892 .895 .868 .869 .840 .841 .822 .821 .779 .779
2.3 .901 .905 .879 .880 .852 .852 .834 .833 .792 .792 ,-..
2.4 .910 .9-3 .889 .890 .863 .863 .846 .845 .805 .804
2.5 .918 .921 .898 .899 .873 .874 .857 .856 .817 .816
2.6 .926 .9:8 .906 .907 .883 .883 .867 .866 .829 .828 r
2.7 .932 .935 .913 .915 .892 .892 .877 .875 .840 .838
2.8 .938 .941 .921 .922 .899 .900 .885 .884 .850 .848
2.9 .944 .946 .927 .929 .907 .908 .893 .892 .860 .857
3.0 .949 .951 .933 .935 .914 .915 .901 .900 .869 .866
3.1 .954 .956 .939 .940 .920 .921 .908 .907 .877 .875
3.2 .958 .963 .944 .945 .927 .927 .915 .914 .885 .882 A,
3.3 .961 .964 .948 .950 .932 .933 .921 .920 .893 .890
L*(Z /Zb) is the exact lower confidence bound for R(t)

a b0
L *(Z /Z ) is the asymptotic lower confidence bound for R(t 0 )
A a b0
11-95 ..
* * . .. . * .

BOUNDS FOR R(t 0) (Continued)
n = 50
p = r/n = 37/50
S= 0.5U y = 0.7j y = 0.90 y = 0.95 y = 0.99
L*(Zb) LA*( b) A*± bA()

L b()LA*(.) L b()LA,*(b) b*~)LA ( b)
3.4 .965 .967 .952 .954 .937 .938 .927 .926 .900 .897
3.5 .9t8 .970 .95b .958 .942 .943 .932 .931 .905 .903
3.6 .971 .973 .960 .962 .946 .947 .937 .936 .911 .909
3.7 .974 .976 .963 .965 .950 .951 .942 .941 .916 .915
3.8 .976 .978 .966 .968 .954 .955 .946 .945 .921 .920
3.9 .978 .980 .969 .971 .957 .958 .950 .949 .926 .925
4.0 .980 .982 .972 .973 .961 .962 .954 .953 .930 .930
4.1 .982 .984 .974 .975 .963 .965 .957 .956 .934 .934
4.2 .984 .985 .976 .977 .966 .967 .960 .959 .938 .939
4.3 .965 .987 .978 .979 .969 .970 .963 .962 .942 .942
4.4 .987 .988 .980 .981 .971 .972 .966 .9b5 .945 .946
4.5 .988 .989 .982 .983 .973 .974 .968 .9b8 .949 .950
4.6 .989 .990 .983 .984 .975 .976 .971 .970 .952 .953
4.7 .990 .991 .985 .98b .977 .978 .973 .972 .955 .956
4.b .991 .992 .986 .987 .979 .980 .975 .974 .957 .959
4.9 .992 .993 .987 .988 .981 .981 .977 .976 .960 .961
5.0 .992 .993 .988 .989 .982 .983 .979 .978 .953 .964
L*(~/. 1 ) is the exact lower confidence bound for Rkt0)

d b 0
S(I, l./1 is tne asymptotic lower confidence bound for R(t0)
A a-b 0

BOUNDS FOR R(tO) (Continued)
n - 50 U
p - rnf .1
y a 0.50 y - 0.75 y - 0.90 y - 0.95 y 0.99
-- a*- LL*(Z
Z Z Z Iz \ /Z&\\ ZaI aZ
•*-Z
z L* LA* ) L*() L a A*(:b& L* A
LA* L* A~b
LA*
.0 .372 .368 .336 .331 .304 .298 .284 .279 .263 .243
.1 .408 .405 .372 .367 .339 .333 .320 .312 .294 .275
.2 .444 .441 .408 .402 .374 .367 .356 .346 .325 .307
.3 .480 .477 .443 .438 .407 .402 .390 .380 .356 .339 M-L
.4 .514 .312 .477 .472 .442 .435 .422 .413 .388 .371
".5 .547 .545 .510 .506 .474 .468 .456 .446 .417 .403 ..-.
.6 .579 .578 .542 .538 .506 .501 .487 .478 .448 .433
".7 .610 .609 .572 .569 .537 .532 .517 .508 .476 .464
.8 .639 .638 .603 .599 .586 .561 .548 .538 .507 ,!93
.9 .667 .666 .630 .627 .595 .590 .577 .567 .535 .521
" 1.0 .693 .692 .657 .654 .621 .617 .603 .594 .562 .548
1.1 .718 .717 .683 .680 .647 .643 .628 .620 .590 .574
1.2 .741 .740 .707 .704 .671 .668 .652 .645 .617 .599
. 1.3 .762 .761 .728 .726 .695 .691 .675 .669 .641 .623
1.4 .782 .781 .749 .747 .718 .713 .698 .691 .663 646
1.5 .800 .800 .769 .767 .738 .734 .719 .712 .o82 .668
1.6 .818 .817 .787 .786 .757 .753 .738 .732 .70] .689
1.7 .834 .833 .804 .803 .775 .772 .757 .751 .720 .708
1.8 .848 .848 .819 .819 .791 .789 .774 .769 .739 .727 .. .,.
1.9 .861 .861 .834 .833 .806 .804 .791 .785 .756 .7,44
2.0 .874 .873 .848 .847 .821 .819 .807 .801 .773 .761 5.
2.1 .885 .885 .860 .860 .834 .833 .822 .815 .789 .776
S2.2 .895 .895 .872 .871 .847 .846 .835 .828 A803 .791
2.3 .905 .905 .882 .882 .859 .858 .847 .841 .817 .805
2.4 .914 .913 .892 .892 .870 .869 .858 .853 .829 .818
2.5 .S22 .921 .901 .901 .880 .879 .869 .864 .840 .830
2.6 .C,29 .928 .910 .910 .890 .889 .878 .874 .851 .842
2.7 .935 .935 .917 .917 .899 .897 .887 .883 .860 .852
2.8 .941 .941 .924 .924 .907 .905 .396 .892 .870 .862
2.9 .946 .946 .931 .931 .914 .913 .904 .900 .878 .872
3.0 .951 .951 .937 .937 .921 .920 .911 .908 .886 .881
3.1 .956 .956 .942 .942 .927 .926 .918 .915 .894 .889
3.2 .960 .960 .947 .947 .933 .932 .924 .921 .901 .897
3.3 .964 .964 .952 .952 .939 .938 .930 .927 .908 .904
LA(Z /Z ) is the exact lower confidence bound for R(t )

LA*(Za/Zb) is the asymptotic lower confidence bound for R(-
11-97
TABLE VIII. TABLE OF EXACT A'D ASYMPTOTIC LU0,QR CONFIDENCE

BOUN.DS FOKnR(t 0 ) (Continued)
n - 50
p - rln -1
0 50 9
1 . y - 0.75 y a 0.90 y - 0.95 y 0. 9
L*
* a
L L* L a L * L*
b A (Z)
Tb LA4± La)
A (_) Zb AA'I
A b) \f.~'
3.4 .967 .967 .956 .956 .944 .943 .935 .933 .914 .910 -
3.5 .Y70 .970 .960 .960 .948 .947 .940 .938 .920 .917
3.6 .973 .973 .963 .963 .952 .952 .945 .943 .925 .922
3.7 .97 .976 .966 .966 .956 .956 .949 .947 .930 .928
3.8 .978 .978 .969 .969 .960 .959 .953 .951 .935 .933
3.9 .980 .980 .972 .972 .963 .962 .956 .955 .940 .938
"4.0 .982 .962 .974 .975 .966 .965 .960 .959 .944 .942 V
4.1 .9•63 .964 .977 .977 .969 .968 .903 .962 .948 .946
4.2 .985 .965 .979 .979 .971 .971 .966 .965 .951 .950
4.3 .9bb .987 .981 .981 .974 .973 .909 .968 .955 .954
4.4 .9a8 .988 .982 .96;4 .976 .975 .971 .970 .958 .957
4. .V69 .9b9 .984 .984 .978 .977 .973 .972 .960 .960
4.6 .990 .990 .985 .985 .980 .979 .976 .975 .963 .963
4.7 .991 .991 .980 .9a7 .981 .981 .976 .977 .965 .9o5 " .-
%,"
4.8 .992 .992 .988 .988 .983 .983 .979 .978 .968 .966
4.9 .93 .993 .989 .989 .984 .984 .981 .980 .970 .970
5.0 .993 .993 .990 .990 .986 .985 .982 .982 .972 .972
b'
L*(/a Lb) is the exact lower cornidence bound for R() "
LA*k a/)
1 s c: asymptotic lower confiaence bound for J((0)
11--98__% .
ITBLE Viii. TABLE OF EXACT AND ASYMPTOTIC LOWER CONFIDENCE

BOUNDS FOK R(t ) (Continued)
n a 100
p a r/n = 25/100
y = 0.50 y - 0.75 - y a 0.90 y - 0.95 y " 0.99
za , L* LA* L* L ( L* .
A.a & L* *
kZb) A&~ (1(b
*. A~) *~ A() (b ýZi- LA~b.
.0 .402 .308 .315 .288 .243 .220 .193 .182 .Ag.9 .119
.1 .437 .405 .353 .328 .282 .261 .234 .222 .157 .156 .
.2 .469 .441 .390 .368 .325 .303 .280 .205 .198 .197
.3 .501 .4,77 .428 .409 .368 .346 .321. .309 .240 .241•
.4 .534 .512 .466 .448 .408 .389 .863 .353 .284 .287 .. ' ,
.5 .5b4 .545 .503 .487 .449 .431. .408 .397 .333 .3b4 -.
.6 .593 .578 .538 .524 .487 .472 .451 .441 .384 .380 . "'
.7 .022 .bo9 .573 .559 .526 .512 .492 .483 .432 .42b
.6 .648 .b38 .605 .593 .562 .550 .533 .522 .484 .4b9
.9 .675 .b36 .635 .625 .595 .585 .569 .560 .525 .511
1.0 .699 .692 .b63 .b55 .626 .618 .605 .595 .566 .550
1.1 ,723 .717 b690 .682 .655 .649 .b37 .628 .603 .585 ,
1.2 .745 .740 .713 .708 .683 .677 .668 .b57 .635 .61.
1.3 .765 .761 .737 .732 .708 .703 .695 .bb5 .662 .b48
1.4 .784 .781 .758 .754 .730 .727 .717 .710 .683 .675
1.5 .o01 .b00 .777 .774 .752 .749 .740 .732 .704 .699
1.6 .81.8 .817 .796 .793 .771 .768 .758 .753 .724 .721.
1.7 .b33 .b33 .812 .810 .790 .787 .77b .772 .746 .741.
1.8 .848 .848 .b27 .825 .806 .803 .791 .789 .764 .759
)..) .0o1 .d•1 .84u .040 .821 .81.8 .806 .604 .777 .775
2.0 .873 .873 .653 .b53 .834 .832 .620 .618 .791 .790
2.L .ts4 .885 .865 .6b.) .84b .845 .832 .b3l .805 .803
2.2 .b94 .b93 .875 .876 .857 .856 .843 .843 .819 .81.5
2.3 .902 .905 .885 .886 .b67 .867 .854 .854 .832 .827
2.4 .91.1 .913 .894 .896 .676 .877 .863 .864 .843 .637 -...
2.5 .';19 .,s21 .903 .904 .885 .886 .872 .874 .653 .847 ,
2.6 .926 .928 .91.0 .912 .894 .894 .880 .882 .862 .65)6
2.7 .932 .935 .917 .919 .901 .902 .688 .b90 .672 .864
2.8 .438 .941, .923 .926 .908 .909 .895 .698 .881 .b72
2.' .4 .946 .929 .932 .914 .91b .902 .905 .866 .879
3.0 .949 .951 .934 .938 .920 .922 .909 .911 .895 .886
3.1 .9.,3 .956 .94'j .943 .926 .926 .915 .91.7 .901 .892
3.2 .957 .9b0 .944 .948 .931 .933 .921 .923 .907 .b99
S.. .9(6L .9b4 .949 .952 .935 .938 .926 .928 .913 .904
L*k2, /4,b is the exact lower contidence bound for (t ).

a b 0
*(2-..at) is the asymptotic lower contidence bound iOL R(. )
b 0
11-99
TABLE VIII. TABLE OF EXACT AND ASYMPTOTIC LONER CONFIDENCE

n - 100 -"1
p - rn - 25/100-
y - 0.50 y - 0.75 y - 0.90 y - 0.95 y * 0.99
z
/Z z /Z \ IZz. IZ Z
kb
3.4
Zb
.965
LAZ
.967
L*LLA
.953 .956
L*\ L
ZZb
-
.940
LT*-
A
.942
L L*
bb LA()LZb
.931 .933 .918
LA*
.909
3.5 .968 .970 .956 .959 .944 .947 .935 .937 .923 .914
3.6 .971 .973 .960 .963 .948 .950 .939 .941 .928 .919
3.7 .973 .976 .963 .966 .951 .954 .944 .945 .932 .924
3.8 .976 .978 .965 .969 .955 .957 .947 .949 .936 .928
3.9 .978 .980 .908 .971 .958 .960 .951 .952 .940 .932
4.0 .980 .982 .97. .974 .961 .963 .954 .955 .943 .935
4.1 .982 .984 .973 .976 .964 .966 .957 .958 .947 .939
4.2 .983 .985 .976 .978 .966 .968 .960 .961 .950 .942
4.3 .985 .987 .977 .980 .968 .971 .963 .964 .953 .945
4.4 .986 .988 .979 .981 .971 .973 .965 .966 .956 .948
4.5 .987 .989 .981 .983 .973 .975 .967 .968 .958 .951
4.6 .989 .990 .982 .984 .975 .977 .969 .970 .961 .954
4.7 .990 .991 .984 .986 .976 .978 .971 .972 .963 .956
4.8 .991 .992 .985 .987 .978 .980 .973 .974 .985 .959
4.9 .991 .993 .986 .988 .979 .981 .975 .976 .967 .961
5.0 .992 .993 .987 .989 .981 .983 .977 .978 .969 .963
f.... •.. ....
a
(Z) is the exact lower confidence bound for R(t 0 )
A. a(/Z is the asymptotic lower confidence bound for R(t)
11-100

n - 100
p - r/n a 50/100
y 0.50 y 0.75 y - 0.90 y- 0.95 y 0.99
z
a L*a
z jAz\/z\
LA*a!) L*k
/Z\
LA*a L*
/Z\
LA*
/z\
a\ L*
z
1) LA**.!) L*
z
L*
z
.0 .377 .368 .339 .329 .309 .294 .287 .273 .244 .236
.1 .413 .405 .377 .367 .348 .333 .323 .312 .284 .275
.2 .448 .441 .413 .405 .386 .371 .361 .352 .318 .315
.3 .483 .477 .450 .442 .422 .410 .400 .390 .359 .354
.4 .516 .512 .486 .478 .458 .447 .439 .428 .399 .393
.5 .549 .545 .522 .513 .493 .483 .476 .465 .441 .430 .. .
.6 .581 .578 .555 .547 .525 .518 .510 .500 .477 .466
.7 .612 .609 .586 .579 .557 .551 .542 .534 .511 .500
.8 .640 .638 .615 .609 .587 .582 .571 .565 .540 .533 ,- -
.9 .668 .666 .643 .638 .617 .611 .598 .595 .567 .563
1.0 .694 .692 .669 .665 .644 .639 .626 .623 .595 .591
1.1 .718 .717 .695 .690 .669 .665 .652 .649 .621 .618
1.2 .741 .740 .717 .714 .691 .689 .676 .673 .644 .642
1.3 .761 .761 .738 .736 .713 .712 .698 .696 .667 .666
1.4 .781 .781 .759 .757 .734 .733 .720 .718 .687 .687 ".""
1.5 .798 .800 .777 .776 .753 .753 .739 .738 .707 .707
1.6 .816 .81/ .794 .794 .771 .771 .756 .756 .726 .726
1.7 .031 .833 .810 .811 .788 .788 .773 .774 .745 .744 ,
1.8 .846 .848 .825 .826 .804 .804 .790 .790 .762 .761
1.9 .859 .861 .840 .840 .818 .819 .804 .805 .776 .776 ,.
2.0 .872 .873 .852 .853 .832 .832 .818 .819 .792 .791
2.1 .883 .885 .864 .865 .845 .845 .832 .832 .805 .804
2.2 .893 .895 .875 .876 .857 .857 .844 .844 .818 .817
2.3 .903 .905 .885 .887 .867 .868 .855 .855 .830 .829
2.4 .912 .913 .894 .896 .877 .878 .866 .866 .842 .840
2.5 .919 .921 .903 .905 .887 .887 .875 .876 .851 .850
2.6 .927 .928 .911 .913 .895 .896 .884 .885 .861 .860
2.7 .933 .935 .918 .q20 .903 .904 .892 .893 .870 .869
2.8 .939 .941 .925 .927 .910 .912 .900 .901 .879 .878
2.9 .944 .946 .931 .933 .917 .919 .907 .908 .887 .886
3.0 .949 .951 .937 .939 .923 .925 .914 .915 .894 .893 *K.
3.1 .954 .956 .942 .944 .929 .931 .919 .921 .902 .900
3.2 .958 .960 .947 .949 .934 .P36 .925 .927 .908 .907
3.3 .962 .964 .952 .953 .940 .941 .931. .933 .915 .913 , .
L*(Z /Z ) is the exact lower confidence bound for R(to)

a b 0
L *(Za/Zb) is the asymptotic lower confidence bound for R(cO) .,.
A a/b 0
11-101

BOUNDS FOR R(tO) (Continuei)
n - 100
p - r/n - 50/100
y - 0.50 y 0.75 y a 0.90 y 0.95 y " 0.99
SL*a) L Z& L*Z I L*Z IL LA L La L*

ZD Zb/ D;
DAZbI b/AZ
3.4 .966 .967 .956 .957 .944 .946 .936 .938 .921 .919
3.5 .969 .970 .959 .961 .949 .950 .941 .942 .926 .924 - .*.,
3.6 .972 .973 .963 .964 .953 .954 .945 .947 .932 .929 ,
3.7 .974 .976 .966 .967 .956 .958 .949 .951 .937 .934
3.8 .977 .978 .968 .970 .960 .961 .953 .954 .941 .938
3.9 .979 .980 .971 .973 .963 .964 .957 .958 .945 .942
4.0 .981 .982 .974 .975 .966 .967 .960 .961 .948 .946
4.1 .982 .984 .976 .977 .968 .970 .963 .964 .952 .950
4.2 .984 .985 .978 .979 .971 .972 .966 .966 .955 .953
4.3 .986 .987 .980 .981 .973 .974 .968 .969 .958 .956
4.4 .987 .988 .981 .983 .975 .976 .971 .971 .961 .959
4.5 .988 .989 .983 .984 .977 .978 .973 .973 .964 .962
4.6 .989 .990 .984 .986 .979 .960 .975 .975 .966 .965
4.7 .990 .991 .986 .987 .981 .981 .977 .977 .968 .967
4.8 .991 .992 .987 .988 .982 .983 .979 .979 .971 .969 r "
4.9 .992 .993 .988 .989 .983 .984 .980 .981 .973 .971
5.0 .993 .993 .989 .990 .985 .986 .982 .982 .975 .973
(Z~a is the exact lower confidence bound for R(t0 )

LA*(Za/Zb) is the asymptotic lower confidence bound for 1V(tO)
11-102
TABLE VIII. TAbLE OF EXACT AND ASYMPTOTIC LOWER CONFIDENCE

]OUNDS FOR R(t 0 ) (Continued)
n - 100
p - r/n 75/100
9 5
I 0.50
U -y=0.75 y "0.90 y - 0. y J.99
L*
L.a L* L* L *1 L* 1A* L* % L * L* LA*)
.0 .374 .368 .347 .339 .320 .314 .305 .298 .273 .270
.1 .410 .405 .383 .376 .357 .350 .341 .335 .307 .306 .
.2 .445 .441 .419 .412 .393 .386 .377 .371 .343 .342 •
.460 .477 .455 .448 .427 .422 .414 .406 .383 .377
.5+.15 .512 .489 .483. .461 .457 .448 .441 .421 .411 ,....s .
.5 .548 .54. .523 :517 .494 .491 .481 .475 .452 .445
.6 .579 .578 .555 .550 .526 .523 .513 .508 .482 .477"
•.7 .609 .609 .585 .581 .557 .555 .543 .539 .513 .508
".8 .o39 .•38 .613 .611 .587 .585 .572 .569 .541 .538 .
.9 .bo6 .666 .641 .639 .615 .613 .599 .597 .568 .566
1.0 .092 .692 .668 .665 .642 .640 .625 .624 .594 .593 16
1.1 .716b .717 .692 .691 .666 .665 .650 .650 .619 .619
1.2 .739 .740 .715 .714 .690 .689 .675 .674 .642 .043
1.3 .781 .761 .737 .736 .712 .712 .695 .696 .666 .066
1.4 .780 .781 .758 .757 .732 .733 .718 .718 .688 .688
1.5 .796 .800 .777 .776 .753 .753 .738 .738 .709 .708
l.b .816 .817 .794 .794 .771 .772 .757 .757 .728 .727
1.7 .o31 .833 .811 .611 .788 .789 .775 .775 .747 .746
1.8
1.o
.846
.859
2.0 .672
.848
.b61
.873
.826
.840
.853
.626
.841
.854
.804
V819
.833
.805
.620
.834
.790
.607
.821
.791
.06
.821
.764
.779
.793
.763
.779
.794
V Y.
2.1. .663 .885 .8b5 .a66 .846 .847 .834 .834 .808 .808i'
2.2 .893 .895 .876 .877 .858 .859 .847 .847 A821 .821
2.3 .9U03 .905 .887 .888 .809 .870 .858 .858 .835 .834
2.4 .911 .913 .896 .897 .880 .880 .868 .869 .846 .845
2.i .919 .921 .905 .906 .889 .890 .878 .079 .856 .856
2.6 .927 .928 .913 .914 .898 .899 .887 .888 .885 .866
2.7 .933 .935 .920 .921 .906 .907 .896 .897 .875 .876
2.8 .939 .941 .927 .928 .913 .914 .904 .905 .885 .884
2.9 .94J .946 .933 .934 .920 .921 .911 .912 .894 .893
.J.0 .950 .951 .939 .940 .927 .928 .918 .91.9 .902 .900
3.1 .956 .956 .944 .945 .932 .934 .924 .925 .910 .907
J..2 .959 .96u .949 .950 .938 .939 .929 .931 .916 .914
3.3 .962 .964 .953 .954 .943 .944 .935 .937 .922 .920
L*(Z /Zb) is the exact lower confidence bound for R(t

a b 0
L,*(4 /Z ) is the asymptotic lower confidence Dound for R(tO)
SA. ab 0
1-103
TABLE VIII. TABLE OF &XACT AND ASYMPTOTIC LOWER COMFDEKCE
-BOUNDS FOR R(C0 ) (Continued)-

n- 100
p -r/n - 75/100
y 0.50 y- 0.75 y- 0.90 y' 0.95 y 0.99
4
b*-)L*T- *Z\L*Z)L1a LA*(/A) L*Z LA* '
'U, \bl\Z
*3.4 .966 .967 .957 .958 .947 .949 .940 .942 .928 .926
3.5 .9b9 970 .961 .962 .952 .953 .944 .946 .933 .931
3. .972 .973 .964 .965 .955 .957 .949 .950 .938 .936
3.7 .975 .976 .967 .968 .959 .960 .953 .954 .942 .941
3.6 .9?7 .978 .970 .971 .962 .963 .957 .958 .946 .945
3.9 .979 .980 .973 .974 .965 .966 .960 .961 .950 .949 .
4.0 .981 .982 .975 .976 .968 .969 .963 .964 .954 .953
4.1 .9b3 .984 .977 .978 .971. .972 .966 .967 .957 .956 ..
4.2 .984 .985 .979 .980 .973 .974 .969 .970 .960 .959
4.3 .9b6 .987 .981 .962 .975 .976 .971 .972 .963 .962
4.4 .987 .988 .983 .983 .977 .978 .978 .974 .966 .965
4.5 V9b8 .989 .984 .985 .979 .980 .975 .976 .96b .968
4.6 .989 .990 .986 .986 .981 .982 .977 .978 .971 .970
4.7 .990 .991 .987 .987 .983 .983 .979 .980 .973 .972
4.8 .991 .992 .988 .989 .984 .985 .981 .982 .975 .974
4.9 .992 .993 .989 .990 .985 .986 .982 .983 .976 .976
5.0 .993 .993 .990 .991 .987 .937 .984 .984 .978 .978
L*(~ /Z)is the exact lower confi~dence bound for R(t.)

a b0
L *(Z /Z ) 8~ the asymptotic lower con~fidence bound for R(to
A a b0
11-104
*-..-.. a *

n- 100
p- r/ 1
y - 0.50 y- 0.75 y- 0.90 y , 0.95 y , 0.99
z jz\ /z' 1 L( /Z /Z /Z\ZL

Z- L*• a L*(ya.) L*Z LA*( L* a) LA L* LA** L* LA*
Z Zb) Zbb Zb
.0 .371 .368 .347 .342 .325 .318 .304 .304 .282 .279
.1 .408 .405 .382 .378 .361 .354 .340 .339 .314 .312
.2 .445 .441 .419 .414 .394 .389 .377 .374 .348 .346
.3 .480 .477 .455 .449 .428 .424 .412 .409 .381 .380
.4 .514 .512 .489 .481 .461 .458 .447 .443 .415 .413
.5 .547 .545 .5221 .517 .494 .492 .479 .476 .448 .446
.6 .579 .578 .554 .550 .526 .524 .512 .508 .481 .478
.7 .610 .609 .585 .581 .558 .555 .544 .539 .510 .508
.8 .639 .638 .614 .611 .587 .585 .572 .569 .543 .538
.9 .666 .666 .642 .639 .615 .613 .601 .598 .573 .567
1.0 .693 .692 .669 .666 .642 .641 .629 .625 .602 .594
1.1 .717 .717 .694 .691 .667 .666 .653 .651 .626 .620
1.2 .741 .740 .718 .715 .692 .690 .677 .675 .650 .645
1.3 .762 .761 .740 .737 .715 .713 .700 .698 .672 .669
1..4 .782 .781 .760 .758 .736 .735 .721 .720 .693 .691
1.5 .800 .800 .779 .777 .756 .755 .742 .741 .713 .712
1.6 .817 .817 .797 .795 .774 .774 .761 .760 .732 .732
1.7 .833 .833 .814 .812 .792 .791 .779 .778 .750 .751
1.8 .848 .848 .829 .828 .808 .808 .795 .795 .767 .769
1.9 .861 .861 .843 .842 .823 .823 .810 .810 .784 . 785 ••' .. w•4m
2.0 .873 .873 .856 .855 .837 .837 .824 .825 .799 .801
2.1 .885 .885 .868 .868 .850 .850 .837 .839 .813 .815
2.2 .S95 .895 .879 .879 .862 .862 .850 .851 .827 .828
2.3 .905 .905 .890 .889 .873 .873 .862 .863 .840 .841
2.4 .913 .913 .899 .899 .884 .884 .873 .874 .852 .853
2.5 .921 .921 .908 .907 .893 .893 .883 .884 .862 .864
2.6 .929 .928 .916 .916 .902 .902 .892 .893 .871 .874
2.7 .935 .935 .923 .923 .910 .910 .901 .902 .881 .883
2.8 .941 .941 .930 .930 .918 .918 .909 .909 .889 .892
2.9 .946 .946 .936 .936 .925 .924 .916 .917 .898 .900
3.0 .951 .951 .942 .941 .931 .931 .923 .924 .905 .908
3.1 .956 .956 .947 .947 ,937
.937 .929 .930 .912 .915
3,2 .960 .960 .952 .951 .942 .942 .935 .935 .919 .921
3.3 .964 .964 .956 .956 .947 .947 .941 .941 .925 .927
L*(Z a /Zb)b is the exact lower confidence bound for R(t)0•i/.:

L *(Z /Zb) is the asymptotic lower confidence bound for R(to)
A a b0
11-10S ,o°* .*. " '.

TABLE VIII. TABLE OF EXACT AND ASYMPTOTIC LOWER CONFIDE1.CE

n - 100
p - r/n=
y - 0.50 y 0.75 y a 0.90 y - 0.95 y 0.99
j a (z ~ aA( Z/A L*z.z ZaL Za\

- L* a LA*() \ L*Z()L*(* *Zb/ \bI Lo(#) LA* h
3.4 .967 .967 .960 .960 .951 .951 .945 .946 .931 .933
3.5 .970 .970 .963 .963 .955 .955 .950 .950 .936 .938
3.6 •973 .973 .966 .966 .959 .959 .954 .954 .941 .943
3.7 .976 .976 .969 .969 .963 .963 .958 .958 .945 .947
3.8 .978 .978 .972 .972 .966 .966 .961 .961 .950 .951
43.9 .980 .980 .975 .975 .969 .969 .964 .965 .953 .955-..
4.0 .982 .982 .977 .977 .972 .971 .967 .967 .957 .959
4.1 .984 .984 .979 .979 .974 .974 .970 .970 .960 .962
4.2 .985 .985 .981 .981 .976 .976 .972 .973 .963 .965
4.3 .987 .987 .983 .983 .978 .978 .974 .975 .966 .968 1
4.4 .988 .988 .984 .984 .980 .980 .977 .977 .969 .970
4.5 .989 .989 .986 .986 .982 .982 .979 .979 .971 .9'12
4.6 .990 .990 .987 .987 .983 .983 .980 .981 .973 .75
4.7 .991 .991 .988 .988 .985 .985 .982 .982 .975 .977 ..
4.8 .992 .992 .989 .989 .986 .986 .983 .984 .977 .978 ' .J ,
4.9 .993 .993 .990 .990 .987 .987 .985 .985 .979 .980
5.0 .993 .993 .991 .991 .988 .988 .986 .986 .981 .982 I.
L*(Z /Zb) is the exact lower confidence bound for R(to) "
LA*(Za/Zb) is the asymptotic lower conficnce bound for R(t)

t.'.' .
11-106
TABLE IX. FACTORS FOR TESTS FOR INCREASING HAZARD RATESk%£I•••-"J

0.10 0.25 0.40 •
N11 0.02 0.05
1.052 o.896 A
-:!•":•
...
5 1.656 1.464 1.305
1.285 1.067 0.926 --
6 1.605 1.435
1050.944 -". %?...
'"
7 1.565 1.410 1.274
1.080 0.957 'ZA,••4
8 1.531 1.389 1.263
1.•255 1.081 0. 966 k',/,
9 1. 504 1.•372 _
1.355 1.247 1.082 0.973
10 1.479
1.233 1.082 0.983 ,
12 1.439 1.330 -','..'.-"-J
1.422 1.318 1.227 1.082 0.986
13 q
1.309 1.221 1.081 0.989
14 1.408
1.081 0.00 ".'-92'•

1.2195
15 1.397 1.299
1.211 1.080 0.994'>-• -
16 1.3830 1.290 1.2474 1.075 10.94 • i
78 1.2565 1.4108 1.206 1.079 0.996•
241.181.2842 0.24i :• • .
17 1.374
1.32951.2276 1.16027 1.0.7 79 .o
.997
18
1.264 1. 193 1.076 1.000..'"....-"[..
20 1.346
1.253 1.186 1.0675 1.002
22 1.330 1.2331 1.082 10098 . ".•
12 1.439 1.330
1.242 1.179 1.073 1.003
24 1.318 1.005 •[
1.234 1.174 1.072
1.070RATES 1.009AZA7D i-.
26 4A0 E 1.305
I. 1A8TT
F1 F . 13 •-
00.02 0.
1.2297!.0.10168 1.071
1.0.2 1.0096=
0.400
28 1.295
1.526 1.464 1.155 1.05210.9-6
34 1.008
1.435 1.351 1.065 .9
36 1.260 1.1761 1.1202 .7
58 1.320
1.123 1.00 1.0.09
48 1.238 1.139
1.25 1.081 1.0109
92 1.244 1.378
1:216 1.082 1.010
10 1.238 1.318 1.01.
1.235 1.32 1.1234 1.082
54
13 1.22 1.184 1.122 1.058 1.011
1.085 0.98 1,. --
12 1.420 1.13 1 1.122
1.199 1.124 1.058 1.011
65 1.239 0.994
1.212 1.129 1.211 1.058
6
16 i.379 1.284 1.206 1.057 1.011 ..
19 1.3 19 .526 1.118 1.057 1.019

1.009
60
80 1.346
1.248 1.26i
1.1392 1.102 1.093 1.060
1.056 1.000
1.302 1.153 1.186
1.109 1.075
1.063 1.002
1.010
62 421.274 1.188 1.010
441289130.3
1.189 1.142 1.117 1.054
.6 1.010
42 1.012
1.124 1.112 1.052
26 1.196
1.171
1.227 1.092
1.168 1.057
1.058461 1.012
520
68 1.123
1.220
1.012
1.185 1.220 1.160
1.022 1.050
1.054 1.011
70
5611 1.2125 1.164 1.012
1.147 1.105 1.056
32 1.179 -"...
1.186 1.208 1.106 1.066 1.012
08
34 1.056 1.012
36 1.61 1.120 1.105
1.192 1.107 1.064 1.012
38 1.255
1.19
68 1~I51-'111 1.53 1.01
1.115 1.083 1.043 1,012

120
70.". 11.148
>'-1
.190>. . .148.-..--.-1-.110.-... ,-.053. 1•.
.012.
11-1o0,7. .•,..
- -.
Downloaded from http://www.everyspec.com '• •
TABLE IX. FACTORS FOR TESTS FOR INCREASING HAZARD RATES (Continued)
N/y 0.50 0.60 0.70 0.75 0.80
50.808 0.726 0.642 0.480.0
0.808 0.726 0.642 0.598 0.552
6 0.842 0.767 0.688 0.648 0.602
7 0.3866 0.796 0.722 0.684 0.641
8 0.884 0.818 0.747 0.712 0.671
9 0.94 0.898 0.375 0.735 0.3695
10 0.908 0.848 0.784 0.753 0.715
17 0.917 0.860 0.798 0.768 0.732
12 0.924 0.869 0.810 0.781 0.746
13 0.930 0.877 0.820 0.792 0.758
14 0.935 0.883 0.829 0.801 0.769
22 0.940 0.890 0.837 0.810 0.779
16 0.944 0.895 0.844 0.818 0.787
17 0.947 0.900 0.850 0.824 0.8795
18 0.950 0.904 0.856 0.830 0.802
19 0.953 0.908 0.860 0.836 0.808
20 0.955 0.912 0.866 0.842 0.814 .
22 0.960 0.917 0.874 0.850 0.824
24 0.963 0.922 0.881 0.958 0.833
26 0.966 0.927 0.886 0.865 0.841
28 0.968 0.931 0.892 0.871 0.847
30 0.971 0.934 0.897 0.876 0.854
32 0.973 0.937 0.901 0.881 0.859
34 0.974 0.940 0.905 0.886 0.864
36 0.976 0.942 0.908 0.889 0.869
40 0.978 0.947 0.914 0.896 0.876

62 0.978 0.949 0.916 0.899 0.880
44 0.979 0.950 0.919 0.902 0.883
46
34 0.983
0.980 0.957
0.951 0.912
0.922 0.8913
0.904 0,896
0.886
48 0.981 0.953 0.923 0.907 0.889
50 0.982 0.954 0.925 0.909 0.891
54 0.983 0.957 0.929 0.913 0.896

56 0.984 0.958 0.930 0.915 0.898
58 0.985 0.959 0.932 0.916 0.900
58
60 0.987
0.985 0.9563
0.961 0.938
0.934 0.923 0.909
0.918 0.902
62 0.986 0.962 0.934 0.920 0.904
64 0.986 0.962 0.936 0.921 0.906
66 0.986 0.962 0.937 0.922 0.907
68 0.987 0.963 0. 938 .923 0.909
70 0.987 0.964 0.940 0.925 0.911
74 0.988 0.965 0.942 0.928 0.913 • •
76 0.988 0. 966 0.94 2 0.928 0.915 •Y' ''•-o
78 0.989 0.967 0.943 0.929 0.916" """ "
80 0.989 0.967 0.944 0.930 0.917....-.," -"
90 0.990 0.970 0.948 0.935 0.922 :- -
100 0.991 0.972 0.951 0,939 0.927
11I0 0.992 0. 974 0. 954 0.942 0.931--•
120 0.993 0.976 0.956 0.945 0.934
11-108
4 --..-. ..
TABLE IX. FACTORS FOR TESTS FOR INCREASING HAZARD RATES (Continued)
NIy0.85 0.90 0.95 0.98
5 0.500 0.439 0.360 0.284

0.552 0.493 0.410 0.326
6
7 0.592 0.537 0.458 0.379
8 0.624 0.572 0.496 0.421
0.650 0.601 0.527 0.455
9
0.672 0.624 0.553 0.483
1.0 0.507
11 0.690 0.644 0.575
12 0.705 0.661 0.594 0.528
13 0.719 0.676 0.611 0.546
14 0.730 0.689 0.626 0.563
,
15 0.741 0.701 0.639057 ...
16 0.751 0.711 0.651 0.591

17 0.759 0.720 0.662 0.602 .-
0.767 0.729 0.672 0.613

18
0.774 0.737 0.682 0.624
19
0.781. 0.745 0.690 0.633
20 0,650
22 0.792 0.758 0.705
0.802 0.769 0.718 0.665 - -
24 0.678 AL
26 0.812 0.779 0.730
0.820 0.788 0.740 0.690
28
0.826 0.796 0.750 0.700
30
0.833 0.802 0.758 0.710
32
34 0.838 0.809 0.766 0.718
36 0.843 0.815 0.773 0.726
0.848 0.820 0.779 0.734
38 0.786 0.740
40 0.852 0.826
0.857 0.830 0.790 0.747 ~
42
4z. 0.860 0.835 0.796 0.752
46 0.864 0.839 0.801 0.758
0.867 0.842 0.805 0.763
48
so 0.870 0.846 0.810 0.769
0.873 0.850 0.814 0.773
52 0.778
54 0.876 0.852 0.817
0880.855 0.821 0.781
56
58 0.881 0.858 0.824 0.786
0.883 0.860 0.828 0.789
60
0.885 0.864 0.830 0.792
62
640.887 0.866 0.833 0.796
0.889 0.868 0.836 0.799
66
68 0.891 0.870 0.839 0.802
70 0.893 0.873 0.842 0.805
72 0.894 0.874 0.844 0.808
74 0.896 0.876 0.846 0.811
0.898 0.878 0.848 0.814
76
78 0.899 0.880 0.850 0.816
80 0.901 0.882 0.852 0.818
90 0.907 0.890 0.862 0.829
100 091 0.896 0.870 0.839
100970910.876 0.847
0.921 0.906 0.883 0.854
120
11-109
N. 'k
AW.
%0..
-~ 0% - 04
z 0
P-1 4 0
- 0
I~I 0 '~.-'ON
go p.M
U.. C40 0
co N0
f. e
ICI
N
co I C.7
4 0 .0 p. .
0 . 0 0% N '
6%
0N (%0 No 0o 0
cz U"C % 04 -i 'to
co0 0 c (0% '0 49' <
LA U'% 0. 0y 0-L-
ao 0O '0 0 0 0 0 u 0
co -7
0% 00
a% -4~ m
LM -4 0 0 0 -..
0 0
N% kn' No 0 0o 0 0 ' -
- 0 0 0% '4 -4.0
:Ad go ' 0 0 o -
0N
CO -%U" 04 ('o LM -4 Q 0 N 80
Fn (14.2 0 0 0 0E"
co 0)0
. i' - 4 0 - 0. h'.
0 0. '.. i. 0 Q 4w' '
C42 -4 0. 0 (D 0 0
-4 NO 0- I0
~ ~~L0
0 4 GO - 0I4d4
09 '0 7ý 0n 0 0, 00 0ý Of
-- 0 0 0 0 0 1
W Ia' *
TABLE XI. UNBIASING FACTORS FOR THE M.L.E. OF C
n 5 b 7 6 9 10 11 12 13 14 15 16
8(n) .669 .752 .792 .820 .842 .859 .872 .683 .893 .901 .908 .914
n 18 20 22 24 26 28 30 32 34 36 38 40
b(n) .V23 .931 .938 .9L3 .947 .951 .955 .958 .9b0 .962 .964 .966
n 42 44 46 48 50 52 54 56 58 60 62 64
S(n) .9b8 .970 .971 .972 .973 .974 .975 .976 .977 .978 .979 .980
n b6 68 70 72 74 76 78 80 85 90 100 120
B(n) .980 .981 .981 .9b2 .982 .983 .983 .964 .985 .986 .987 .990
This table is reproduced from "Inferences on the Parameters of the Weibull

Distribution" by D. R. Thoman, L. J. Bain, and C. E. Antle, Techometrics,
Vol. 11, No. 3, August 1969, pp. 445-460.
S1Best Available Copy

MISION
Rome Air Development Ce#~r.,.
Ccmmuncaticrm and lmte&W~.c C1 :~i ~ :

and enq-(neetVing zuppoa'rlt wt~hin a.',eoa o' technc,:z ~~c;
isA p~uyuided to ESi) P/ztog'Lan OýJice (PO,ý2
efC1Jent6. Vie piccipaZ tchnica.C -J
c-cmmrnu;aI-tj'ono ctocqntz
ve,ittayiciz o, qtound and aorc: ~ ~ ~
7
r1,,tzCýc
ov rti-or, Aotd &ýmte. v'6
ci
Best Available Copy

Radc TR 85 194

Uploaded by

Copyright:

Available Formats

Radc TR 85 194

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Radc TR 85 194

Uploaded by

Copyright:

Available Formats

What is the document about?

What is the document about?

What are some of the sections covered in the document?

What are some of the sections covered in the document?

Downloaded from http://www.everyspec.

RADC'MONE ECTROIVIUC RELIABILIT

* Analysis Cer~ter .OHTRI)

t7''J FIV37 P '. ý IEASE, DISTRIBUTIONl UNLIMITED

Best Available Copy

b as been reviewed bv Lhc, ",, T ", .

RAkDC-TR--85-194 has been r-vicwed and is provod ,

V. S. TUTdILL, Cclonel, UTSA,

Best Available Copy

SEURITY CASiNiAION OF THIS PAGE

REPORT DOCUMENTATION PAGE

This notebook is divided into six sections: Introduction,'--Applicable Statistical Methods

20 I)ISTRIBLJTION/AVAILABILITY OF ABSiRACT 21 ABSTRACT SECURITY CLASSIFICAIION

Best Available Copy

This Notebook is the result of research conducted at Hughes Aircraft

Several individuals made significant technical contributions to this *.

This document replaces RADC-TR-75-22, Nonelectronic Reliability Notebook.

Justiticatio~i Ave '-''a' .'-'_

SECTION 1.0 - INTRODUCTION . . . . . . . . . ........... . .. 1-1

SECTION 2.0 - APPLICABLE STATISTICAL METHODS FOR NONELECTRONIC

2.1 Statistical Failure Models ........ ................ . 2-1

TABLE OF CONTENTS (Continued)

2.4.5 The Extreme Value Distribution............. . . . . . 2-49

2.4.7 Outlier Tests ................. ........................ . 2-59

TABLE OF CONTENTS (Continued)

4.6.3.3 Diaphragm Reliability. ....................................... 4-17

4.8.1 Purpose of the Method. ............................ ............4-21Zv

6.. StatistLical Characteristics of a Reliabili.ty D~emonstration

*6.1.4 Summary .. ....................................................6-3

2.4.2.3.2.1 Graphical Method for Interval Estimation of Reliability %'.)<-e%6

for the Weibull Distribution 2-35

2.4.3.2 Graphical Point Estimation for the Normal Distribution 2-44

4.4-1 Illustration of the Concept of Interference 4-7

4.4-2(A) Time Varying Strength Density 4-8

4.4-2(B) Time Varying Strength Density 4-9

4.9-1 Planetary Gear Reduction .- 25

4.9-2 Planet Gear Forces 4-28-

6.3.2.1A Operating Characteristic Curve for Test Plan IA 6-20

6.3.2.2A Operating Characteristic Curve for Test Plan 2A 6-20

6.3.2.3A Operating Characteristic Curve for Test Plan 3A 6-21

6.3.2.4A operating Characteristic Curve for Test Plan 4A 6-21

6.3.2.5A Operating Characteristic Curve for Test Plan SA 6-22

"6.3.2.6A Operating Characteristic Curve for Test Plan 6A 6-22

6.3.2.7A Operating Characteristic Curve for Test Plan 7A 6-23

6.3.2.8A Operating Characteristic Curve for Test Plan 8A 6-23

6.3.2.9A Operating Characteristic Curve for Test Plan 9A 6-24

"6.3.2.12A Operating Characteristic Curve for Test Plan 12A 6-25

iLlS'l 0t ILLU61RATION.) kon~nueu)

0).. b Qperatxng Characteristic Curve for Test Plan jI -2

t...b uperating Characteristic Curve for Test Plan 2B 6-27

b...b Uperatxng Characteristic Cuive for Test Plan 35 5-28

opera:in6 Cnar~cterisric C;urve 'or Test Pian 5o 6-29

0. 4.0 perz%,n6 (ýnaracteristi.o Lurve for TesL Plan t~b 6-29

7 ) Operatc.r,6 tCnaraC~erIStLI Curve tor Test Plan 78 6-30

0u)2o perat3.ng Characterist2.c c.urve for Test Plan db 6-30