MIL-HDBK-472 Mantenibilidad y Prediccion PDF

MI L-HDBK-472
- 24 May 1966
MILITARY STANDARDIZATION HANDBOOK
MAINTAINABILITY PREDICTION
DEPARTMENT OF DEFENSE
WASHINGTON D,C. 20360
A?IL- HDBK-472
MA INT.\INABILITY PREDICTION
:4 Jltiy ]966
1. Thls maintainability handbook. V.as developed b} the I)epartn)cnt of

?-~fcnqe ir, :!(,eord~ce \\itb es~abiisheu proce[lur~o
)
-. This publication ivas approted on Z4 nl~v 1966 for printing and
InpIIlc!nn in the mi]itar~ standardization handbook series.
:}. This dwunlcnt profidcs information on current maintainahilit!

prwlicticm procedures. It i~i]] prn~idr valuable information and guidanw
to personnel concerned with the drsikm, de~)clopment, and production of
myipmcni and s~stcrns requiring a high order of maintamability.
4. Ever~ effort has been made to reflect the latest information m

waintqi~~hilit~,
..... . .. . nrdirtion
., procedures. It IS the intent to re~ieu Lhis
l:mcl~wmk pcrlorhcally t.n mswre its conlple Leness and currency. Users of
(111s [Iocllnl(nf n re encouraged to report an] errrms discovered and an~) recoin-
))l(,n(j~tinns for ch:lnges or inclusions to the Comnlander, Naval Air S\stems
(:nmmancl, I)cqmrtment of the N:IVJ, Mxsh!ngton, D. C. ~03GiJ.
MIL-HDBK-472
24 May 1966
FOR EWORD
The purpose of this Maintainability Prediction Handkmok is w familiarize

project managers and design engineers with current maintainability pre-
diction procedures. To achieve this objective, particular care has been
exercised in selecting and including only those procedures which are currently
used in predicting the maintainability of equipment and systems. The hi@-
lights of each maintainability prediction procedure are presented m a clear,
lucid and irttell igible manner and include useful supplementary information
applicable to specific procedures, using the following format.
1.0 GENERAL
Philosophy, Assumptions and Summary
Applicability
Point of Application
Basic Parameters of Measure
Information Required
Data Basis
Correlation Between Predicted and observed Values
2. () ANALYTIC KXJNDATION
3.0 APPLICATION
This type of presentation facilitates reference to all or to only those parts of

a procedure whfch are of particular interest to the user. For example, the
manager may be interested solely in the non-technical aspects, such as:
Point of Application, or Applicability. Conversely, the engineer, may con-
centrate on the technical aspects only, such as Analytic Foundation and
Application. Thus, through the use of this handbook, maintainability
engineers, working with a new development, can select the most applicable
maintainability prediction procedure for a specific equipment or system.
ii
-
MIL-nDBK-472
24 May 1966
Prediction facilitates anearly Assessment of thematurity of the maintainability

design and enables eariy decisions concerning the compatibility of a proposed
&sign with specified requirements or the choice of better alternatives.
The maintainability prediction procedures I and III are applicable solely to

electronic systems and equipments. Proccdurcs II and J\ can be used for
all systems and expnprnemts. in applying procedure II to non-clectromc
equipments the appropriate task times must be estimated.
In conclusion, the use of this hardbok facilitates the desi~, development,

and production of equipment and systems requiring a high order of maintaina-
bility.
.,.
Ill
MIIm HDBK-472
24 May 1966 -
TABLE OF CONTENTS
Page
INTRODUCTION
The Need for blaintainability Prediction

Definition of Maintainability
Basic Assumptions and interpretations
Elements of Maintxiinability Prediction Techniques
summary
MMNTAINABILITY PREIXCTION PROCEDURES
PROCEDURE I 1-1
SECTION 1.@ GENERAL 1-1
1,1 Phi losoph!, Assumptions and Summar.v 1-1

1.2 Appl]cabjlity i-l!
1.3 Point of Application 1-2
l-~
1.4 Basic Parameters of h!easure
1.5 Information Required }-2
1.6 Data Basis 1-3
1.7 Correlation Between Predicted and Observed Values 1-4
SECTION 2,0 AN.+ LYTJC FOUNDATION 1-4
2.1 Structure of Time Elements in Fix of Malfunction 1-4

]-12
2.2 Elemental Activities
2. 2.1 Assumptions Relating to Elemental Activities 1-14
2. 2.2 Major Characteristics of Elemental Activities 1-14
2.3 Synthesis of Time Distributions 1-16
2.3, 1 Modes for Synthesizing Distributions of Time 1-16
2. 3.2 Cumulative Time Distributions and the Monte Carlo 1-17
Method
SECTION 300 APPHCATION
3. I Preliminary Procedure 1-19

3.2 Steps of the Prediction Procedure 1-19
-
MIb~BK-472
24 hhy 1866
TABLE OF CONTENTS (Continued)
Page
ADDENDUM A CALCULATION OF THE PRO134BIXJTIES OF Al-37

OCCURRENCE OF ELEMENTAL ACTMTIES
ADDENDUM B DEFINITIONS OF MAINTAINABILITY TERMS AI-44
PROCEDURE II 2-1
SECTION 1.0 GENERAL 2-1
1.1 Philosophy, Assumptions and Summary 2-2

1.2 Applicability 2-3
1.3 Point of AppU cation 2-3
1.4 Basic Parameters of Measure 2-3
1.4.1 Corrective hlaintenance (Part A) 2-3
1.4.2 Corrective Maintenance (Part B) 2-5
1.4.3 Preventive Maintenance (Part B) 2-5
1.4.4 Active Maintenance (Part B) 2-5
1.4.5 Maintainability Index (M 1) 2-6
1.5 Information Required to Apply This Procedure 2-6
1.6 Data Basis 2-7
1.7 Correlation Between Predicted and Observed Values 2- 7
SECTION 2.0 ANALYTIC FOUNDATION 2-8
2.1 Basic Considerations aad Definitions 2-8

2.1.1 Applicable Maintenance Taska 2-8
2.1, 2 Applicable Flmctional Levele 2-9
2.1.3 Replaceable Item 2-1o
2. z principles of the hiaintahability Prediction Rocedure 2-1o
SECTION 3.0 APPLICATION 2-11
3.1 PART A Corrective Maintenance Rediction Procedure 2-11

3.1.1 Maintainability Prediction Durtng Final Design Stage 2-12
3.1.1.1 Prediction Procedure 2-12
3.1.1.2 Determining Functional Levels of the 2-12
Equipment or &stem
3.1.1.3 IMermin!ng Functional Levels at wMch 2-14
Maintenance Features are Effective
-. v
MI& EDBK-472
M May l@36
TABLE OF CONTENTS (Continti)
Page
3.1.1.4 Prediction 2-15

3.1.1.5 Obtain@ Predicted htntnia.inabili~ 2-26
Pammeters
3.1.1.6 Calculations 2-26
3.2 PART B Active Maintenance Predction Procedure 2-27

3.2.1 Corrective Maintenance PredictIon 2-27
3.2.2 Preventive Maintenance Predtcflon 2-29
. PROCEDURE III 3-1
SECTION 1.0 GENERAL 3-1
1.1 Philosophy, Assumptions and Summary 3-1

1.4 Basic Parameters of Measure 3-2
1.4.1 preventive Maintenance %3
1.4.2 Corrective Maintenance 3-3
1.5 No=&non RuwIred 3-3
1.6 Data Baei~ 3-3
1.7 Correlation Between %edicted and observed Values 3-4
2.1 Determination of Sample Size N 3-5

2.2 Derivation of Equation for Sample Size N 3-6
2.3 Task Sample 3-9
2.3.1 Task &llllpU~ 3-9
2.3.2 Task Sample fk?lectlon Technique 3-11
2.4 Application of Check Llsta 3-11
2.5 L1.nsar Regression Equation for Downtime 3-12
2.6 Maihemittlcd Relationships of Maintenance Times 3-12
3.1 Gene-l Approach 314

3.1.1 Determination of Sample Size 3-14
3.1.2 *p-by-Siep Procedure for Calculating Sample Size 3-14
Vt
TWL-HDBK-472
24 May 1966
Page
3.2 Sb?p-by-Stap Procedure for Calculating Task Sample 3-16

3, 3 Step-by- Step Procedure for Performing Ta 8k Prediction ?-IF
3.4 Calculstio~ of Maintenance Indices 3-30
ADDENDUM A DESIGN CHECK LISTS AND SCORKNG CRITERIA A3-33
CHECK LIST A - SCORKNG PHYSICAL DESIGN FACTORS A3- 33

CHECK LIST B - SCORING DESIGN D] CTATE& FACILITIES A3-47
C7HECK LXST C - SCORKNG DESIGN DICTATES-IWI.NTENANCE A3-!i2
SIULN
PROCEDURE IV 4-1
SECTION 1.0 GENERA L 4-1
1.1 Philosophy, Assumptions and summary 4-1

1.4 Basic Parameters of Measurement 4-2
1.5 Information Required 4-2
1.6 Da* Basis 4-3
1.7 Correlation Between Predicted and Observed Values 4-3
SECTION 2.0 ANALYTIC ,OUNDATION 4-3
2.1 General 4-3

2.2 Theoretical Considerations 4-7
2. 2.1 End Kern Identification 4-7
2.2.2 qeratiunal Ihtnctton 4-7
2.2.3 Coirective Maintenance Action Definition 4-7
2.2.4 Ma Ifurction Detection Analysts 4-8
2.2.5 Prevcntite Maintenan& Task Time Analysis 4-8
2. 2.6 Corrective Mai,,tenance Task Time Analysis 4-9
2.2.7 Tots? Maintenance Task Time Analysis 4-12
.
BIBLIOGRAPHY
vii
MIb HDBK-472
24 May 1!466
LIST OF ILLUSTRATIONS
Table i-ieadmg Page
List of Categories and Elemental Acttvtties of .4ctive 1-13

Repair Ttme
1-2 Fitted Dlstrlbutions for Complet~on Times for Elemental !-!5
Acti\ities
1-3 S}-nthesis of Time Distributions 1-20
1-4 Parameters of Distributions of Elemental Activities i-24
1-5 Proposed Format for Determining the Cumulative Time 1-27
III stribution
1-6 Frohability of Completing Eiemental Activity of Final
Malfunction Test by Designated Time
1-7 Combining Matrix for Active Repair Time S.ynl.hesis
l-b Observed Distributions of System Final Test Time and
System bgistic Time
1-9 Combining Matrix for System Down Time Synthesis 1-34
1-1o Occurrence Probability Functions for the Elemental Activities AI-36
1-11 System Readouts AI-43
2-1 Average Parts Failures per Part per 106 Hours by Part 2-18
Category
Maintenance Task Times 2-20
Interchange Time 2-23
Element Times Based on the Work Factor System 2-24
An Example of the Use of Element Times for Determining 2-25
Interchange Time
3-1 Part Class Failure Distribution and Sample Size 3-17

3-2 Task Prediction, R-7801 3-23
3-3 Maintenance Analysis, R-7801 /FPS-20 3-24
.,.
Vlll

MIL-HI)BK-472
.- 24 May 1966
Figure Figure Heading Page
A Comparison Matrix of hfaintainabilit~ Prediction Procedures 4

1-1 Distribution of Malfunction Active Repair Time for the 1-5
AN; ASB-4
1-2 Distribution of System Down Time for the AN/ASB 4 1-6
1-3 D~stribution of Malfunction Active Repair Time for the 1-7
ANr ARc-34
1-4 Distribution of System Dow~ Time fcr the AN,; ARC- 24 1-8
1-5 Distribution of Malfunction Active Repair Time for 1-9
the MD- 1
1-6 Distribution of System Down Time for the hlD-1 1-1o
1-7 Structure of Time Elements in Fix of Malfimction l-lx
1-8 Distributions of Times Required w Perform Repair Actions 1-18
1-9 Resultants Synthesized From t a, t~andtc 1-18
1-10 Probability of Occurrence of System Logtstic Time 1-33
2-1 Functional Level Diagram of a Typical Communications System 2-13

2-2 Worksheet for Final Design Maintainability PredictIon 2-16
- ~- : ~ui-ktdleei A 2-32
2-4 Worksheet B ?-33
2- 5 Worksheet C 2-34
2-6 Worksheet D 2-35
3-1 Comparison of Population ar. rl Sampling Mean Distributions 3-8

3-2 .Sample Size Nomograph 3-1o
3-3 Maintainability Prediction Form 3-19
3-4 Maintenance Analysis Continuation %eet 3-20
3-5 Maintenance Diagram AN/GRT-3 3-21
3-6 AN/ FPS-20, Transmitting System 3-26
3-7 Amplifier-Mixer, AM-1347 /FP&20 3-27
3-B Expanded Iiew, Plate Power Supply 3-28
3- fl Plate supply, Block Diagram 3-29
3-1o Nomograph - Downtime 3-32
4-1 Procedural Flow Block Diagrnm 4-6

4-2 Related Constraint Matrix 4-14
4-3 End Item/Corrective Maintenance Action Matrix 4-15
4-4 End Item/Preventive Maintenance Action Matrix 4-16
4-5 End !tem/Operational Functional Matrtx 4-17
ix
MIL-HDBK-472
24 My 1966
INTRODUCTION
THE NEED FOR MAINTAINABILITY PREDICTION: The prediction of the eqxxtecl

number of hours that a system or device will be in an inoperative or down state
while it is undergoing maintenance is of vttal importance to the user &cause of
the adverse effect that excessive downtime has on mission success. Therefore,
once the operational requirements of a system are fixed, it is imperative that
a technique be utilized to predict its maintainability in quantitative terms as
early as possible during tic design phase. Thi B prediction sho~d be u~~d
continuously as the desi~ progresses to assure a high probability of compliance
with specified requirements.
A significant advantage of using a maintainabili~ prehction procedure ia that

it highlights for the desi~er, those areas of poor maintainability which justify
product improvement, modification, or a change of design. Another useftd
feature of maintsinabi lity prediction is that it permits the user to make an early
assessment of whether the predicted downtime. the quality, quantity of personnel,
tools and test equipment are adequate and consistent with the needs of system
operational requi rem ents.
.-
DEFINITION OF MAINTAINABILITY: Ml bSTD- 77 ?i wimes maintainability as
foIlows :
$lM~n~n~i~ity is a characteristic of
design and installation which is expressed
as the probability that an item will conform
to specified conditions within a given period
of time when maintenance action is performed
in accordance with prescribed procedures and
resources.
This definition has fostered the development of many maintainability prediction

procedures for providing an assessment of system maintainability. Each of these
uses various quantitative measures to indicate system maintainability. However,
all of these measures have a specific relationship to, or constitute some element
of the distribution of total system downtime. Hence, if a universal method or
technique can be developed to determine the Total System Downtime Distribution
for any type of system, this would facilitate calculating the measures of maintain-
ability currently m use.
.
MIL-HDBK-472
24 May 1966
BASIC ASSUMPTIONS .4ND INTERPRETATIONS: Each maintainability prediction

procedure included in this handbook depends upon the use of recorded reliability
and maintainability data and experience which have been obtained from comparable
systems and components under similar conditions of use and operation. It is also
customary to assume the applicability of the principle of transferability. This
assumes that data which accumulate from one system can be used to predict the
maintainability of a comparable system which is undergoing, design, development,
or study. This procedure is justifiable when the required degree of commonality
between systems can be established. Usually during the early design phase of the
life cycle, commonality can only be inferred on a broad basis. However, as the
design becomes refined, during later phases of the life cycle, commonality is
extendable if a high positive correlation is established relating to equipment
functions, to maintenance task times, anti to leveis of mainte.aance. Alt.ho@
the four techniques contmned in this handbook have been proposed and appear to
fit certain applications, it should be borne in mind that they have not truly ben
tested for generality, for cxmsi stency one to another, or for most other criteria
dealing with broad applicability. It should also be borne in mind, though, that
experience has shown that the advantages greatly outweigh the burden of making
a prediction. For that reason, it is not the purpose of this document to deter
further research or inquiry.
ELEMENTS OF MMNTAIN.ABILITY PR ED! CTJON TECHNIQUES: Each maintain-

.~i II+.. __~.-I{~+i~ +~ohiq,.o ,~+ili?~t mrnn-dvre~ whj~h are ~cifically
EUL A&., ~.bw.-..., ,uti. ti. -- -------- ~..--
designed tO
satisfy its method of application. However, all maintainability prediction methods
are dependent upon at least two basic pararnetera:
(a) Failure rates of components at the specific assembly le~el of interest.
(b) Repair time required at the maintenance level involved.
There are many sources which record the failure rate of parts as a function of
use and environment. This failure rate, is expressed as the number of failures
per unit of time. A typical measure is failures per 106 hours. The major
advantage of using the failure rate in maintainability prediction calculations is
that it provides an estimate of the relative frequency of failure of those components
which are utilized in the design. Similarly, the relative frequency of failure
of components at other m aintafnable levels can be determined by employing
standard re[iabili~ prediction techniques using parts failure rates. Failure
rates can also be utilized in applicable regression equations for calculating the
maintenance action time. Another use of the failure rate is to weight the repair
times for various categories of repair activity, in order to provide an estimate of
its contribution, to the total maintenance time,
hl;LinL7illfil) ilit~, prc(liction, thereforc, is o most usc!fr.tl ]nstrl~n)rnt to both nlnnager
and engineer hccause it provides for improved system ~ffectiveness :md rduws
ar..fnlinlstrat lv( ancl maintenance costs.
T))(I comparison matrix, Figure A, is included to prnvidc a summnry of thr signifl-

r~nt attrihtlte~ of ~neh nlaintiinabilit}, . nreriirtinn rrroce(iurc inrlu(ie(l in t.i}is han(l-
IKmii. A[i(]itinnai details may ix Obtained ~JJ referring to S) W!iflL illainLainabilitJ
pre(lirtinn promvlures of interest.
MIIrHDBK-472
24 May 1966

..

I= I -,
1-
MI~HDBK-472
24 M8y 1966
PROCEDURE 1
1.0 GENERAL
This procedure is used to predict system downtime of airborne electronic and

elecrro-mechanical systems involving modular replacement at the flight-line.
Just as a masonry building depends upon the brick as its basic building block,
the procedure relies on the Elemental Activity as the fundamental element of
downtime from which other mea sures of downtime are developed through a process
of synthesis of time distributions.
The Elemental Activity is a simple maintenance action of short duration and

relatively small variance which does not vary appreciably from one system to
another. An example of a basic elemental activity would be the opening snd
shutting of a door, or opening and closing of a radome on an aircraft. It should
be obvious that the performance time does not depend upon the construction of
the house or aircraft provided that the door or radome are similar. Therefore
if one should record the times required over many trials, to open and close a
door, he should be able to calculate the mean time, ~, and the standard
deviation, C, of this Elemental Activity. This is precisely what has been done
in this prediction procedure for various Elemental Activities which in total
ccm.pr!se Lie basic Categories nf Active Repair Time. These activities are
listed in Table 1-1 and the corresponding recommended values of ~and @
for the Elemental Activities of each category are shown in Table 1-2.
The technique of using the basic building block, namely the Elemental Activt~,
and building the structure step by step to include other measures of downtime is
explained in 2.0 Analytic Foundation and detailed in 3.0 Application.
1.1 Philosophy, Assumptions snd Summary
Figure 1-7 entitled Structure of Time Elements in Fix of Malfunction illustrates

the %uilding block principle. An examination of this figure shows how elemental
activity times are synthesized to produce maintenance categories and that these
in turn combine to produce Malfunction Active Repair Time. The remainder of
this chart showing the complete buildup to total system downtime is self evident.
There are two exceptions to be noted. The first is that the category entitled
Distribution of Final Test Time is not derived from other Elemental Activities
since it has constant parameters and is shown in Table 1-8, The second exception
h that this procedure does not include the techniques for calculating Total System
1-1
.
MIbHDBK-472
24 May 1966
Downtime but rather concludes with the determination of System Downtime. The
reason is that Total System Downtime is the result of combining the distributions
of .System Downtime with Initial DelaJ. The methods of determining Initial Delay
require some further refinement before they will be considered for Inclusion.
In summary, tbe philosophy of the entire prediction procedure is based OD the

principles of synthesis and transferability. As already discussed the synthesis
principle involves a buildup of dowmtlmes, step by at.ep, progressing from the
distribution of downtimes of Elemental Activities through various stages culminati-
ng finally with the Distribution of System Downtime.
Tbe transferability principle embodies the concept that data applicable to one type
of system can be applied to similar systems under like conditicm of use and
environment to predict system maintainability.
1.2 Applicability
This maintainabili@ prediction procedure is applicable to predict flight-ltne

maintenance of airborne electronic and electro-mecha.nical systems involving
modular replacement at tbe flight-line. The procedure may also be used for
maintainability prediction in echelons of maintenance other than flight-line such
as field or depot by extension of formulae through further development work,
= rcqutrcd, += \~~~.~& ~+-her ~!e~emt~l ..- ~fi++~,itte~.
-...
1.3 Point of Appl jcation
The technique can be applied at any time after the design concept has been established,
provided the essential data enumerated in 1.5 entitled Information Required is
available.
1.4 Basic Parameters of Measure
Tbe ultimate measure of maintainability is the distribution of System Downtime.

Intermediate measures include the distribution of times for the various Elemental
Activities, Maintenance Categories, Malfunction Active Repair Time, Malfunction
Repair Time, System Repair Time and System Downtime. (See Figure 1-7. )
1.5 Information Required
In order 10 perform a maintainability prediction the following information must

k available,
1-2
MIbimBK-#a
24 Hay )9W
(b) Number of flight-line replaceable comwnents of each type.
(c) ZJst d fMgbt-Mne nqheeabh mmpments oontabhg adjuu&nenta

or flight-line replaoaable ~.
(d) Number and A8racter of readmte (bfonltortng dwtcen for portlon6

of the ay8tem).
(e) Nmnbar of tgpm of spar- urried.
@ Number of pmmme-rddnlng ommeckm.
(g) Number of teat potnts.
(h) Nmne of apeetal teet equipment (designed specifictiy for the uysbm).
(1) Number of zna.gnetrons.
o) Estimates of durations of verage mleaion.
*) Mimriixig scha&les !0: spesatiorw IU2dmaintenance peraonncd tncludtng

all ahtfta and all breaka for hmch, cof!ee, etc.
(1) Estimates for intervals oecuphd by tmachethled activities such aa

debrteftng.
In the ortgtnd development of the prediction procedure, data wexw employed from
ma Munction repairs on the AN/ASB-4 Bombtng and IUavi&ption Syutem( wed in the
B-52 bomber). In testing and refining the prediction system, data were used from
seven other systems:
AN/APN - 89 AN/APX - 25
AN/ARC - 34 AWARN - 21
AN/ARC - 65 MD-1
A N/AIC - 10
1-3
MIIr HDBK-472
24 May 1966
Thus it is expected that this prediction procedure should be more adaptable to

electronic or electro-mechanical systems similar to those mentioned above.
1.7 Correlation Between Predicted and Observed Values
Figures 1-l to 1-6 show the correlation between observed and predicted values
for the distribution of Malfunction Active Repair Time and System Downtime
for various equipments.
2.0 ANALYTIC FOUNDATION
2.1 Structure of Time Element6 in Fix of Mdfunctton
Using Figure 1-7 as a guide, it is evident that when the latter is viewed from
right to left the following relationships become evident:
(a) Total System Downtime consiBts of: ~~
1. Initial Delay.
2. Sywtem Downtime.
System Dountime comprtses:
i. system @tatic Time.

2. System Repair Time.
3. System Ftnal Test Tfme.
(c) System Repair Time is the product of Malfunction Repair Time and the
number of malfunctions.
(d) Malfunction Repatr Time consists of:
1. Malfunction Active Repair Time.

2. Malfunction Administrative Ttrne.
~/The calculation of Total System Downtime is not shown in this prooedure

because it depends on the distribution of tfmes of Initial May and the
procedure requires more development.
1-4
MIL-XDBK-472
24 hlay 1966
Io.oo
6.00
6.00
S.oo
4.00
3.00
2.00
1.00
0.6
0.6
WI
0.4
0.3
02
alo
De
~6
.05
.04
.03
.02
.01
0.01 0.1 0-s I 5 IO 30 50 70 90 95 9899 99.9 99.99
PROBABILITY OF COMPLETION WITHIN TIME t
Figure 1-1. Distribution of fila)function Active Rfpair Time for the A% /AS13-4
1-5
..
ML-H DBK-472
24 hby 1966 -
4J
0.01 0.1 0.5 I 5 10 30 50 70 9095 wm m, 99.99
PROBABILITY OF COMPLETION WITHIN YIME t
F@ire 1-2. Distribu~iofi of System Down Time for ihe AS,ASB-4
1-6
MWHD13K-472
24 May 1966
Figure 1-3. Distribution of Malfunction Active Repair Time for the AN/ARC-M
1-7
MILHDBK-472
24 May 1966
6.0
5.0
40
3.0
2.0
z
K
=
o
z
4-
I.0
0.8
:
+ 0.6
0.5
0.4
0.3
0.2
0.1
oh 0.1 0:5 I 5 10 30 50 70 90 9s *a 99 99.9 99 99
PROBABILITY OF COMPLETION WITHIN lIME t
Figure 1-4. D~strlbut!on of System Down Time for the AN; ARC-34
1-8
MIGMDBK-472
24 May 1966
100 I 1 r I 1 1 1 I I I 1
8.0 t 1 I I
1 I 1 1 1 i 1 I
b
0i-ii me m- PREDICTED t-
2.0
1.0
Oa
0.6
05
0.4
0.3
0.2
0.I
0.01 0.1 0.3 I s 10 30 50 70 90 95 9699 99.9 99.99
Figure 1-5, Distribution of Malfunction Active Repair Time for the MD-1
1-9
MXL- HDBK-4 ?2
24 May 1966
100.0
80.0
60.0
30D
40.0
30D
20J2
lox)
B.o
6L)
So
40
3.0
2.0
I.0
0.8
0.6
0.4
0.3
0.2
0.1
0.01 0.1 0.s I 5 10 30 50 70 90 95 98 w 99.9 99.99
PROBABILITY Of COMPLETION WITHIN TIME t
Figure 1-6. Distribution of System Down Time for the MB1
1-10
2
!$
k
J ------
1-11
MI L- HDBK-42
24 May 1966 -
(e) hlalfunction Active Repair Time is comprised of:
1. Preparation Tim~.
-. Alalfunction Verification Time,
2. Fault Location Time.
4. Part Prc)curement Time.
L. Repair Time,
G. Final Malfunction Test Time.
(f) Elemental Activities: Each of the maintenance categories of (e)

above consists of a series of Elemental Activities. as shoun in
Iahle 1-1 and discussed in 2.2.
_ Elemental
-. Activities
To facilitate analtsls and minmlize variations in the performance time required,

each malfunction rnwntenance category is subdivided into smaller maintenance
actions, labellcd Elemental Activities. As (Iisrussed It] 1. u these represent
rclati{$clj simple md brief maintenance act]ons, which require a short time to
pc rfo rm. Therefore, in the prediction procedure x origin~ly developed, the
assumption was that Elemental Acti\ritics were normally distributed. However,
after further refinement of the proccdurc three optional distributions are available
ior use in a prediction as follows: @ee Table 1-2. )
(~) The fitted normal distribution.
(b) The fitted lng-norms! distribution.
(c) A corrected time log-normal distribution.
The determination of which Elemental Activities are normall} distributed, and

\vhich Iog normally, is based on the h}~thesis that Elemental Activitieshaving
standard deviations less than an arithmetic mean of one hour, are representative
of activities of a routine nature and are assumed to be normall~ distributed.
This is due to the fact that execution time is not significantly influenced by
changes of personnel, characteristics or surrounding events.
On the other hand, Elemental Activities having standard deviations greater than
the arithmetic mean, or an arithmetic mean greater than one hour, are con-
sidered both as more complex and as containing manv possible subactivities,
all of ~vhich may not need to be performed to define the achvlt}. in such a case
1-12
\
MI L- HDBK-472
24 hfay 1966
I ) TABLE 1-1
I LIST (~F CA TFCORIFS AND FLFMFNTAL ACTIV!TTES OF Aml!rE REPAIR lTMr
I
I
( Categrrry Elemental Activlt~ ~Acl,L!tv~o.
I i
Sjstcrr turr{ on, narrmup, scttlng dIals and couders as ncccssar).
I ActIvIty cl PIUB lime awatt]ng particular component stabilizatlotr.
, l.@enmg and clobmg raoorne.

Gaming access and remstalllrtg covers Iother than radome).
Obtalnmg test equipment and or Tech Or&rs.
Checking matntcnance records.
I , Prnerrrmg components IrI antlclpatlon of need.
!
S44tlng urr lesl equipment.
Obscr}mg IxlI:4tw2ns cd,. J
Uswg lvsi tqutpmenl 10 $erifv maliunchona inherenll) n(,l reproducible on ground. 2

Perform lrzg standarrf test problems or checks 3
Test!nR for pressure Ieaks 4
Attempting 10 obccrvt tlusl$t! or nOn-eSt91ent SymPIOm(IIJ. 5
Using special test equipment rteslmed specifically Ior tiIs equ!pmerrl. 6
Making a visual mtcgrlty chvck, 7
Fault self -ewdent from s.wnptom observation.

Interpreting symptoms by mental analysis only (from knowledge ~exrlence).
Lnterpretmg dIsPlsys M djfltrent srttlngs of controie.
Inler?sretmg meter rcadmga.
Removmg uniI(s)/suburtil (s) snd checking in shop.
Swilchmg and/or substikttmg tmtt(s)ltrubttnlt {s).
wil(hln~ anrl/or suhslitutirrg part(s).
iiemovmg anti ciwchng parsa.
bl~klrrg 3 WISUS) mle~lty check.

Cbcckmg $Ill!ages. cnntmulty. Wawfnrms andinr aipal tracing
ConsulilnR Ttw+ Orders.
CorrlerrtnE wl!h Tech Reps or other mamtenunce personrrel.
Performtrrg standard tom problem (n].
laolattnK pressure leak.
l:sin~ :pcc Ia! tesl cqulpmcnt dcslgncri sm.c~flcally for tht~ qutpmcn:.
CSbtammg replacement component from a[rcraft spares or tool &x I

Part Obtaln]ng replacement(s) from &nch, ahop. or pre-iseue stock. 2
Prcmutwmcnt CMtatrrlrrg replacement crrmporrcnt(sj hy cmmlballzation. 3
Atlemptlng 10 Obtain replacement Crsmpcmrnt(sl Unsvallable. 4
Replacing unlt(rs)/subun! tfs). I

Replacing parts. 2
Correct wtg improper mstallatton or delectl.e plug- kn mxrnactlon(s). 3
Making atdjustmcnls m sircraft. 4
Rcpalr M.!klng adjustments In shop. 5
Baktng mogrretmt. 6
Precautionary repstr activity (Include? S- called fau!t Ioca!lon, parl 7
procurement. and repsir t trees sipetrt when symptom not vprlft~d).
Rcpstrm~ wirtng or eonnerllons. 8
Final Mal- 1
hctton chcckou! folloulrtg completion of repair.
functmn Test I
...
.;
1 l-I::
14flL- HDBK-472
24 May 1966
tbe tendency is for the applicable distribution b be skewed to the right and is
assumed to have log normally distributed completion times.
Table 1-2 shows fitted distributions of completion times for Elemental Activities.
It is recommended that the distribution parameters shown in this fable, which
are denoted by an asterisk, be used in any prediction. Predictions based on
these distributions yield the best empirical fit to reported maintenance time.
in those cases when a prediction of txme time, rather than the time reported by
repair personnel, is desired, the distribution of corrected time is used.
(See columns 7 & 8 of Table 1-2. )
2. 2.1 A.ssumpticmc RelatLng to E!ement~ .4ctivitics
The selection and phrasing of Elemental Activities is based on the following

assumptions:
(a; The mean time required for the performance of an Elemental

Activity is independent of system design and support facilities.
(b) The frequency of occurrence of an Elemental Activity correlates

with some factir of system design or support facilities.
(c) ihe Elemental Activities in any maintenance category are independent

of each other.
)
(d) The total time required in any maintenance category is completely
accounted for by one or more of the Elemental Activities in the
category.
2.2.2 Major Characteristics of Elemectai Activities
TWCIof the characteristics of Elemental Activities are of major concexn. These

are the probability of occurrence and the distribution of time required to complete
an Elemental Activity. Studies indicate that the distribution of the time required
ior the performance of an Elemental Activity is independent of the type and design
of the system involved. For example, large differences would not be expected
between the times required to open and close a radome on two types of aircraft,
or the times required to make a simple electrical adjustment since these do not
depend on the design characteristics of the aircraft. On the other hand, the
probability of occurrence of an Elemental Activity does correlate to some factor
of system design or support facilities.
i-14
r;;i,
MIL-11D13K-j7Z, ,..
$&!:
... . , 24 May 19GG.J!. 1
+ ,.. ..
; . T&OLK 1.2
.,. ,. .,. .,,.
.,,,,4. -..,.
,: ,!, !#,+
G, c . :;.
FIT7EIIDISTRJBUTK)NSFOR COMPL@ON , ,. ,...
,!
.,
TtAfESFOR ELEMENTAL A~ti
Dlotrlbufhm 3
Cormcled Tlmc
(Img Normsl)
Number
P u
P u
.4
Prep2ra- 1 0.102 0.068 0.085 0.606 o.Onz o.ri43
llom 2 0.665* 0. 3R4* 0.Slrl 0. S:!6 o.5s6 0.64s
3 o.23s 0.135 0,204 0.534 0.198 0.643
. 4 O.27S 0.196 0.21% o.6s7 0.200 0.643
s 0.330 0.448 o.195* 1.023 0.190 1.011
6 0.140 0,104* 0.112 0.6GI n.109 0.643
7 0.070* 0.0s0 0.Octl 0.54> 11.058 0.643
., n 0 107 0.127 0.0G9* o.93n 0.008 1.011
Md- 1 0.ltts 0.263 0.0.391,409 0.O.?* 1.011

. Iundon 2 0.693 0.416 0.485 0.63:1 0.469 0,641
Vtrf(i - 3 0.3!)7 (j.
4X 0.2rIl?0.W18 0.?Co 1.011
c8110n 4 0.329 0.605 G.137* 1.215 0.1!33 1.011
5 i.394 1.236 [.047* I.o.wl 1.019 1.011
G 0.309 0.171 ).269 0.520 0.261 0.643
7 0.059 0.i3? ).024 1.345* 0.023 1.011
Fault 1 0.010 G.ool ).010 n.100 0.010 0

LocalIon 2 0.019 o.04U ).008 1,201 0.008 1,011
3 0.333 o.25!* ).2G3 0.686 0.256 0.643
- (\ I
4
5
6
0.141
0.821
0.324
fl.172
o.7BFI
0.346
!.OIW* 0.955
).532 0.808
).221* G.ei:~
0,0S6
0,672
~>:::
1.011
1.011
*
. . nll
. .
7 0.436 0.480 }.293* o.66t* 0.284 1.011
@ 0.181 0.191 ).125 0.865 0.1:2 1o11
~ 0,)40 0.213 ).077 1.095 0,074 1.011
Jl) 0.R07 1.owl ).w? 0.!t65 0.491 1.011
11 0.344 ij.?~l ).240 0.444* 0.233 1.011
12 0. 4G6 ~.420 ).342 0.785 0.333 0.643
I 11 !1. se? O..A. J,321, 0.78!l 0.413 0.643
1.4 O. 6R3 0.960 1.WIG* 1.044 0,364 1.?11
!5 P. :12(I* 0.)62* 1.u15 0.477 O.277 0.643
PW4 I I-1.ilizo.0:1(1 ).012 1.086 0,012 1.011

Proou*- 2 I 0.313* o.z3a* ).249 0.675 0.242 rJ.643
ment 3 0. 31s 0.171 1.277 0.S08 0.269 0.647
., 0.199 0.144 1.161 0.649 0.1s7 ().643
K@ r 1 ;.394 0.s18 t.239 1.002 0.231 1.011

2 0. .&qfi 0.s0: ).203 1.122 0.138 1.011
1
3 o.W41 ().0s5 ).051 O.726 0,049 0.64?
4 0.415 0.541 1.253* n.997* 0.246 1.011
,., 0.997 0.862 1.7s0 0.749 0324 0.643
r, 1.416 0.701 1.%9 l,-
4GFG 1.238 0.643
1 0.7S2 0.747 1 534 0.U29 0.515 1.011
8 0.810 [ 0,461 [ 1.011
J_
9A leg normal diet

rtbkbns,lhe dlstrlbutlon whl+ fWWJ bt ~ Ims &n marked wtth
M Mterld(
( 1-15
MIL-HDBK-472
24 May 1966
For a majority of the Elemental Activities, the probability of occurrence, P,

is predicted for a particular system by solving an equation generated by a
multipk-linear refgramion antdysis. Table 1-10 list~ the occurrence proba-
bility function, p, for a number of Elemental Activities. In these equations,
tbe dependent vsxiatdes are based on observations of comparable systems, and
the independent variables ~e certain qusmtitatfve characteristics.
The occurrence of multiple Elemental Activities is a function of the probabilities

of the Constituent Activities; thus, the probabili~ of conjunctive occurrence of
two independent activttiea is the product of the indtvldual probabilities multiplied
by the probability of non-occurrence of all the remaining activities. For three
independent activities (A, B and C), the probability of joint occurrence of any
*O such as A and B and a non-occurrence of the third C would be:
In this type of notation the bar over a letter indicates the probability of non-
occurrence and a plain capital letter is the probability of occurrence.
2.3 Byntbesis of Time Distributions
There are several techniques which are used in this procedure to synthesize
ttrne distributions. For the most part these depend upon the use of Monte Carlo
techniques which simulate acmai conditions by selecting random samples from
cumulative time distributions of applicable parameters. The various techniques
which are used for combining or synthesizing distrfbution6 of time are explained
in 2.3.1. The details of the specific methods of developing cumulative time
distributions and utilizing Monte Carlo methods are discussed in 2.3.2 and
illustrated under 3.0 Application. A summary is also included in Table 1-3.
2. 3.1 Mode8 for Synthesizing Distributions of Time
There are four synthesis modes, depending on the manner in which a systam CM
fail. These are described as:
(a) Equal Sampling, Adding Variates: When one of the components fails
during a time interval and all other pssible failures occur during
this same period. All repair actions are performed. (Mode 1,
Figure 1-9. ) .
1-16
MILr HD13K-472
24 May 1966
Llnequal Sampling, Adding Iariates: This applies when only onc

failure occurs and the others will probably occur at a rate corresponding
to their occurrence probabilities (the relative frequency, determined
from failure rate data). In this case, various combinations of the
repair action will take place. (310de 2, Figure 1-9. )
Equal Sampllng, Not Adding Variates: This applies when only one
failure can occur at a time and each component ha6 an equal probability
of failure. Only one of the repair actions is possible to correct
system failure. (Mode 3, Figure 1-S. )
Unequal %rnplin~, Not Adding Variates: This is applicable when oniy

one of the components can fail at a time. in the order of their relative
probability of failure, The repair consists of perfomn ing only one
reptir action at a time, since failures wi!l probably occur one at a time,
in accordance with the relative frequency. (Mode 4, Figure 1-9. )
Figures 1-H and 1-9 illustrate the basic principles of synthesis. Note that three
norms! distributions of time are shorn in Figure 1-8. Two of these distributions,
te and tb, overlap to a certain degree, while the distributions to time tc for the
third does not, to any extent.
Figure 1-9 shouts the resulting synthesis for each of the four synthesizing modes.
Note that resultant mode 2 is skewed to the right. This is due to the effect of tc,
wjth its larger mean and its own distribution, intermingling with the sampling of the
other two. This is the mode which can be expected to occur in a complex system.
~. ~. ~ ~~ulative Time D*@rlb~ti~n~ ~d the M~nt~ Carlo Method
The basic technique used in the precbction method is @ develop a cumulative

distribution of times as a means of selecting random time samples for s~thesis
purposes. Figure 1-1 is an example of a cumulative distribution of Malfunction
Active Repair Time. The abscissa represents the probability of completion of
an active repair action within a time, t, (i. e. , in t or less hours. ). The time
of completion is showm on the ordinate. Since the expression within a time t
is used or in t or less hours the distribution is cumulative in nature.
The Monte Carlo method uses cumulative distributions such as Figure 1-1. to
determine parameter values. This involves the selection of times, t, by
r,andomly picking a probability value of the abscissa and reading Lhe corresponding
time, t, cm the ordinate. The random selection of probability is performed by
1-17
----
Mlb HDBK-472
24 May 1966
F we 1.8. Distributions of Times Required to Perform Relulr Act ons
&uQ!-
0
i !/<(02
!,%
Figure 1-9. Res~tan~s %nthesized From to, to and tc
1-18
MJb HDBK-472
24 May 1966
using a table of random numbers, or by utilizing a random number generator when

a computer is employed, to select as many values asneedcd for the prediction
procedure for the desired parameters.
Typical of the mrthod is the s}mthesizing of Elemental Times. In this case

a cumulative d)strlbution is developed from the normal or log norrrlaI distribu-
tions as required, and the selection of random times is nbtalned as explained
in ~Section 3.0 .application,
Table 1-3 summarizes the synthesis method beginning with the synthesis of
E1ementd Activi~ies to detelop the various maintenance categories and corl-
tinuing th process up to ancfincluding the development of System Downtime.
3.0 APPLICATION
?. 1 Prelimina~ Procedure
(a) Def~nc the system in terms of its constituent units and their input and
output boundaries.
fb) Compute the failure rates of the system and of the flight replaceable
components listed in the }~gend ~f Tab!e 1-1o vti!izing accephbie
standard reliability prediction procedures.
(c! Finf~ :he number of system readouts and calculate the readout factor
as outlined in Table 1-11 of Addendum A. This readout factor is used
to multiply each value of time given in the distribution of system final
test time in Table 1-6.
(d) Estin]ate an a~erage flight length (in hours) during which the system is
to be operatccl,
3.2 Steps of the Prediction Procedure
Step (1). Compute tie occurrence probability of each Elemental Activity as

outlined in .4ddenaum A, The act]vit]es marked with an asterisk in Table 1-10
cannot occur with any other activity wilhin the category; therefore they should
be omitted from the calculations in Steps (2), (3), and (4) below.
Step (2). Determine the probability of occurrence of each Elemc#tal Activity

a!one by multiplying i~s computed tolai occurrence probability by the product
.
.. . >--
~. HDBli-4/i!
24 May 1966
t
1
i
i
i
E
I
b
(
(
8
;
.:. ;
i
..
.
. . . . .
--L.
---1
.
1
.
---
1-20
MIhllIX3K-472
24 Hay J 966
of the complements of the probabilities of occurrence

of all the remsining
Elemental Activities wfhin the maintenance categoxy. The fonnuia used in
obtaining the probability of occurraee Of tie events AZ~ A3, . . . Am is:
Step (3). Within each category, &termine occurrence of arty No activltiee

which can occur together. Use the following fomnula for act!vltiesAl , A2, . . . An,
where, Al and AZ are the activities whose joint probability is to be computed:
%obabtlitieg ~ 1% will be considered b be zero.
Step (4). Within each category determtne the probability of occurrence of all
posslb)e tiiple Elemental Activities. Multiply the product of the computed total
occurrence probabilities of alI possfble combinations of three activities by the
product of the complements of the total occurrence probabilltiea calculated for the
remain!~g act!vltteB wttidn the catego~. If the occurrence probability of a
triple combination ts determined to be 1% or lemh cordder it as zero.
~p (s). Within each category add the probabilities calculated in Steps (2), (3),
and (4)to the probabilities calculated in Step (1} for the activllles marked with
an aaterisk ti Table 1-10. The sum obtained
by t.bts addition, thus, will not
contain the probabilities of joint occurrences of four or more activities nor tb06e
computed probab{llties which are negl{gfbiy smaIl (1%or less), therefore the mum
of the probab[llties will be less than unity as shown in tbe following example.
However, the probabilttieo can be made to awn to unity ae Ehown in Step (6).
The following example does not rcpreeent aIl of the combinations possible,
however, it is being presented to show the method used for calculating combinations
of Elemental Activities, (Combinations iesa than 1% omitted as illustrated below).
Example h Step (2), assume in addition that activ!ty A2 was marked with an
asterisk in Table 1-10 which means it cannot occur with any other activity in a
category. We have the following:
MIL- HDBK-472
24 May 1966
P(AI) = .30
P(A2) = .40
P(A3) u .10
P(A4) = .05
= (,0081) ( = 1%., omit)
= (.0021) ( = 1%., omit)
= (.0009) ( = !%. , omit)
.95 Obtained by adding the

first four numbers.
Step (6). Normalize each category by multiplying each of the probabilities summed
almve by the reciprocal of their sum (1. 176). This step ensures that the probability ies
sum to 1 and presupposes that the probabilities of joint occurrences of four or
more activities are negligible.
1
Example Using the values of the previous example we have ~ = 1.176
and (. 30) (1. 176) = .35, normalized P(Al)
(. 40) (1. 176) = .47. normalized P(A21
(.10) (1.176) = . 12, normalized P(A3)
(.05) (1,176) = . 06, normalized P(A4)
rim Total Probabi 1ity
Step (7). Complete Table 1-4 as follo~s:
(a) Insert, in Column 2, the double and triple Elcrnental Activity numbers
whose probabilities of occurrence were calculated as explained in
Steps (3) and (4) above.
.
Ml L- HDBK-472
A May 1966
(b) Determine~ia and Clap Orpia and ala for the

Bzl /38
distribution of the multiple acti~tities from the following equations:
Enter the calculated parameters in Columns 3 ad 4 of Table 1-4. The equations

for the standard deviations reflect the assumption of the statistical independence
of activities or mdtiple activities listed in Table J-4.
(c) M u, > /L I or if ~1 > I hour, enter LN in Column 5.

If cl c ~1 < I hour, enter NTin Column 5. LN means
log normal, and N means normal (see 2. 2).
(d) Complete Column 6 of Table 1-4 by inserting the appropriate normsJized

occurrence probabilities as calculated in Step (6).
(e) For each LN entry in Column 5, compute (and enter in Columns 7 and 8)
p ~ and ~ 2 values from the following equatkns:
H=
The ~ 2 and at values in Column 7 and 8 will be identical tn those in Coiumns

3 and 4 for Elemental Activities designated by an N in Column05. This transforma-
tion of the mean and standard deviation of each log normal distribution to the same
parameters of its nomnal distribution transform is required for the next step.
1-23
.
htIL-~DBK-472
24 bflly 1966
TABL& 1-4{~tlnued)
1 2 I 3 1 4 5 6
1
Bade ckcLlr-
Uemeold Dltiri - mtiti- rewe motrt-
Calegmry Aetivtty butlon butloo Pmb- button
Number Pmrunetera sbuny PsramelerO
PI u, Crz
0w8. ) (hrs. ) fbra. )
Fault 9 0.140 0.213 LN 0.077 1.095

katloo 10 0.60? 1.000 LN 0.507 1.Mt5
(CO* 11 t).344 c.351 LN 0.240 9.644
tirmed) 12 0.466 0,430 N 0.466 0.430
13 0.582 0.541 N 0.582 0.54)
14 0.603 0.960 LN 0.396 1.044
15 0.320 0.162 N 0.320 0.162
--
.
- -.
.
.
.-. 1 0.022 c. $x! G.0L2 1.0M6

2 0.313 O. 238 0.313 0.238
3 0.315 0.171 0.31s 0.171
4* 0.199 0.144 0.199 0.144
--
-..
Rrpatr 1 0.394 0.5 8 LN o 239 1.002

2 9.360 0.603 LN 0.203 1.122
3 0.066 0.055 N 0.066 0.055
4 0.41s 0.541 LN 0.2s3 0.997
5 0.993 0.862 N 0.993 0.862
6 1.416 0.701 LN 1.269 0.468
7 0.752 0.747 N 0.752 0.747
B 0.810 1.114 LN 0.476 1.030
--
I I
---
I t
1
Canoot occur wltb sny other activity.wltiln the cahgvry
J-25
MIL-HDBK-472
24 hhy 1966 -
Step (6). This step tranafoms a normal variable with mean~zand staadard
deviation ~2 to a standard normal variable with a mean of zero and a standard
devlatlon of un{tyo In this fomn the cumulative normal d{atrhution is readily
available tn Tables, or can be easfly calculated on an electronic computer. The
tianaformatlons for normal and log normal variables, respectively are:
!09 ti - log p*
z, = (lor parameters designated LN in Coiumm 5
C2
of Table 1-4).
The values of/L2 and a2 me obtained from Columns 7 and 8 of Table 1-4.
The tl values are gtven in Column 1 of Table 1-5. The cumulative probiabtlitles
of completing an Elemental Activity by a designated time tl are obtained by
entering a table of the cumulative normal distribution and obtatntng each desired
probabili~ opposite the entry Z,. These probabilltiea are then entered in Cohnnn 2
of Table 1-5. ColumxI 3 of Table 1-518 determined by multiplying the values in
Column 2 by the probability of occurrence of the Elemental Activity which is listed
in Coh.urm 6 of Table 1-4. Ibat is, mu!tipl y each cumulative probabi 11ty of complet-
ing an Elemental Activity in a given time ( tl) by the probability of occurrmce of
the activity. Thh3 1s to be done for each maintenance categoxy.
Step (9). For each of the twenty discrete values of time ( tl) listed Ln Column 1
of Table 1-5 sum the corresponding Elemental Activity probabilities shown in
Column 3. For example, amurntng there are 8 Elemental Actlvit{es comprising
the Category of Prepamition each Elemental Activity will have an individual
sheet such as Is illustrated by Table 1-5. Therefore, there will be eight
probability values, one on each ahet?t, for each of the twenty discrete values of
time (tI} listed [n Column 1. Summing each of these eight probabilltie6 (Column 3),
for each time tl will result in a total probability value for ti. Shine there are
twenty t I values, twenty points wil 1 result, which when plotted represent the dis-
tribution of the Category of Preparation Time. Other category distributions
with the exception of Final MaMunction Test Time. will be determined from
their Elemental Activities in a similar manner. Final Malfunction Test Time
haB constant parameters and is gl ven in Table 1-6. Plot the six individual
categury distributions, i, e. , Preparation, Malfunction Verification, Fault
Iacation, Part Procurement, Repair and Final Malfunction Test.
1-26
-
MIL-HDBK-472
24 May 1966
~
z
Z DA
1 [
II
u-
L
MXL-HDBK-472
24 May 1966 TABLE 1-6
-
PROBABILITY OF COMPLETTNC ELEMENTAL ACTTW~
OF FINAL MALFUNCTION TEST BY DESIGNATED TIME
f
~Time (H~ur~) Probability
I
I
0.01 0.009
f
t 0.02 0.022
I
I 0.03 I 0.043
I
\
t
0. 0s I 0.083
I
1
0,07 0, 180
I
0.10 0.250
I
1 1
0.15 0.270
0.20 0.380
0.30 I 0.580
0,40 0.700
0.50 0.790
0.60 0.840
0.80 0.890
0.90 0.905
1.00 0.934
1.50 0.982
\
2.00 0.993
3.00 0.999
5.00 1.000
I 10.00 1.0;0
log Normal
Distribution -==i
1-28
MIL-?KDBK-472
24 May 1966
--
Step (1O), Determbe the d.istrhutlon of malfunction active repair time by
using the combining matrix (Table 1-?), the inetruction8 thereon, and the Bix
category distritmtiona plotted in Step (9), above.
Step (11). From the above distribution (the values are now plotted and smoothed),
select a minimum of 200 random vaiues of rnaifunction active repair time (fitted).
To each time, add a value of adrniniatrative tlrne to be calculated in Step (12)
below. The ref3ult ie mrdfunction repair time.
Step (]2). The following equation ie u~ed to compute administrative time, t:
t=
-Q in [i - F(t)] /S
10
where: F (L )is a randomly selected probability. An acceptable method is to use
a table of random numbers for this purpose.
To obtain Q and ~ use the equations:
P = 0.027 X3 - 0.233 X2 + 0.521 X + 0.230 onci
.
a= 0,310 - 0,064 X
where: X 1s a value of malfunction repair time eelected in Step (11) i
Choose a minimum of 200 Xs and for each value compute the corresponding U
and ~ vaiucs. AISO select the mme number of F ( t ) ~Iues by H~pling by me~e
of a table of mndnm numbers. U6ing the three setg of values, Q , B , and F(t),
we obtain 200 (or more) values of t from the above equation.
Wp (13), Multiply each of the 200 values of active repair time plus administrative
time (i. e. , malfunction repair time) by 0.9501.
where:
1-29
L
Mlb HDBK-472
24 May 1966
TABLE 1-7
COMBINT14G MATRIX FOR ACTIVE REPAIR TIME SYNTHESIS
(X = Occurrence)
Instructlorl:
For each row, uu.m randomly eelectcd samplee of time* frwm tho~e
maintenance categories denoted by an x, ThiB summation will be
executed (M) (P) ttrnes for each row, where M is the desired number of
samples of malfunction active repair tlrne, and P is the indicated probability
of occurrence of each mw.
1 Mai-
Row Prep a- function
Nm.~ber ralton Verifica-
tion
1 x x x x x x 0.170
2 I x x x x x ! 0.022
3 x x x x x 1 0.022
I
4 x x x x x 0.064
5 x x x x x I 0.173
t 6 I I v.. ~ x Y. 1
n
..
nlo
7 ! x x x I x 1 0.012
8 x x x x 0.029
I
9 1 x x x xl 1 0.077
10 x x I x x I o. 0%
11 1 x x x x 0.011
12 x x I x I 0.008
x J 0.011
x x x 0.050
x r
xl 1
I
0.018
I x 0:032
x 1 x 0.011
x 0.008
1 19 1 x I x I I 1 I 0.173 I
20 x
Raw30m vduee of time may be selected from all graphs by using My

published table of random numbers. These random nurnbera represent
probabilities, md, from ihe cumulative distribution graph, the time
which correspond w the ra-dom probability is determined LS a random time
frmm that distribution.
1-30
MILHDBK-472
24 May 1966
~or is the median of the predicted distribution of mal.f unction active repair tlrne;
~1 is the ~ystem f~il~re ra@; ~d N,5 is tie ~ticipated avera& flight len@.
When malfunction repair time is mlItiplied by 0.95N, system repair time 16
obtained. The factor of 0.95 accounts for an overlap time observed during which
two or more malfunctions are king repaired concurrently.
Step 04). Plot the system repdr times obtained jn Step (]3).
Sep (15). ThiB step involves system final test time. Its probabll!ty of occurrence
was observed to be about 0.5. l%e probability di~tribution of final test times as
shown m Table 1-8 was obgerved mainly from AN/ASB-4 data. A system readout
factir, as computed !n Table 1-11, is used as a multiplier of the time completion
probabilities of Table 1-8 when a different system 16 being considered. The kger
the number of readouts, the higher the probability of completion of system ftnal
tegt h a given time.
Step (16), Plot the distribution of system logistic time, given in Table 1-8,
and draw the best fitting line through the plotted points. Determine its
(i.e., a logistic time event) probability of occurrence, N20t from Figure 1-10.
The distribution of systcm logistic time is based on observed data The probabili-
ties of occurrence were observed to be an increasing function of the number of
.-. component types replaceable at the flight-line level.
Now the inputs from S@p (14), (15), (16) are available in the form of cumulative
time distributions for system repair, system final test, md system logistic events.
A new combining matrix (TsMc 1-9) is now developed for the synt.heeis of a system
downtime distribution. Repair time occurs 100% of the time, tist time occurs 50%
of the time, and logistic ttme occurs a t~ariable percenage of the time according
to Figure 1-10. The occurrences are indicated by the Xs in Table 1-9.
The numbers which replace the Xs are drawn at random from the fitted distribu-
tions of repair times, finaI test times, and logistic times. A distribution of syetem
downtimes is obtained from this synthesis. Steps (17) and (18), which follow, describe
the method of completing the combining matrix in Table 1-9,
*p (17). Complete the combining matrix (Table 1-9) by finding the quantity
IOON20. This represents the relative proportion of the total sample of system
dowmtimes which wiIl contain systcm logistic times. Draw IOON20 random
numbers betucen 00 and 99 and determine fl and r2 (the quantities O! those
numbers hefiveen 00 and 49, and betwcn 50 and 99, respectively). Place an x
in f , gpaces &ve line so ad x in r ~ spaces below @nd including) line 30.
1-31
MILr HDBK-472
24 May 1966
TABLE 1-8
OBSERVED DISTRIBU110?W OF SYSTEM FI_NAL TEST

TIME AND SYSTEM LOGISTIC TIME
Systcm Final Test Time System Logistic Time
Time, t,
I
Probability of
in liours Time, t, or Less
0.15 0.020 0.50 0.010

0.20 0.055 0.60 0.020
0.30 0.172 0.80 0.071
0.40 0.31 0.90 0.08
0.50 n,4s 1.00 G. i~
0.60 O. 56 1.50 0.17
0.80 0.728 2.00 0.22
0.90 0.790 3.00 0.28
1.00 0.830 4.00 0.32
1.50 0.945 5.00 0.37
2.00 0.980 6,00 0.40
3.00 1.000 8.00 0.42
10.00 0.43
15.00 0.57
20.00 0.78
40.0 0.82
50.0 0.90
Go. o 0.94
150.0 1.00
1-32
24 May 1966
0.18 -
r
0.15 -
. .
0.1
m!
0.05 -
01
r I 1 I 1 1 1 I 1 1 1 1
L
+ m ~v z
NUMBER OF COMPONENT TYPES REPLACEABLE AT FLIGHT-LINE LEVEL
Figure 1-10. Probability of Occurrence of System Logistic Time

1-33
MILIIDBK-472
24 May 1966 TABLE 1-9
COMBIWNC MATRIX FOR SYSTEM ~W TIME SYNTHESIS

(x = Occurrelme)
r
Syntheota
Number
1
~
system
Repmr
Time
S@em
Final
Text
System
Logictic
Ttmr
Synthesis I
Number /
i fv8~,
TIrxxe
/ SYUm
?::
Syxtem
Lopsllc
Time
Ttme Time
T
I
x x 51 x I
2 x x I I
52 x I
3 x x 53 ! x
4 x x
Ii
54 x i
5 x x 5s x
6 x a 56 x
7 x x 57 ...
I
1 6 x x 5* x
3 x x 59 x
10 x x 60 x
11 x x 61 x.
12 x x 62 x
! 13 x x 63 x
i
14 x x 64 x 1
15 x x 65 x
16 x x 66 x
17 x x 67 x
lb a x 6B x
19 x x 69 x
20 x x 70 x
21 x x 71 x
en
*- Y. x 72 x
23 x x ?3 x
24 I x x. 74 x
25 x x 7s x
26 x x 76 x
27 x x 77 x I
2B x x 7b x
2: x x 79 x
30 x x 130 x
31 x x 81 x
32 I x x 82 x
33 x x 83 x
34 x x 84 x
35 x x 8s x
36 x x 86 x
37 x x 87 x
38 x x RR x
39 i x x 89 x
I 40 x 90 x
x
41 91 x
I x x
42 x x 92 x
I
43 x 93 x
x
I
44
45
x
x
x
x ,
94
9s
x
x
! I
I
I 1
96 x
46 x x
I 97
I
47 x x x
48 x x 98 x
49 x x 99 x
I
I
50
1
x
1
x 100 lx
I I I
1-34
MIb~BK-472
24 May 1966
step (18) . Select a quantity, Q, of an integral number of hundreds (no fewer

than 200) random times from the distribution of system repair times, Step (14).
Select O/2 random times from the distribution of system final test times,
(step 15). %Iect Q Nzo r~~m times from tie digtrib~ion of lo~stic time
plotted in Step (16). Put these times in place of Xs in Table 1-9 and add
them across each row to obtin system dowtimes. Plot the distribution of
system downtimes and draw the best fitting line through the points.
1-35
MIL-HDX-472
24 May 1966
This page intentionally ieft blank.
1-36
MIL- HDBK-472
24 hIay 1966
ADDENDUM A
CALCULATION OF THE PROBABILITIES OF OCCURRENCE OF ELEMENTAL

A CTMTIES
The probabilities of occurrence of Elemental Activities, calculated from the

oc~~rrence probability ~ctions, are given in Table 1-10. These functious
were determined, for the moat part, by subjective selection of those system
chararteriatms wh lch are deemed logically responsible for the occurrence of
the activity. The selected characteristics were then used as the independent
variables in a multiple-linear regression analysis, with tbe dependent variable
being the observed occurrence probability. The foIlowlng statements concerning
the occurrmce probability functions must be complied with:
(a) If P t9 a calculated occurrence probabfl!ty, and if P <0, set P O.
(b) If the denominator of any term O, delete that term and

recalculate P.
For explanation of the symlmls ueed in the occurrence probability functions,

refer to the legend of Table 1-10.
Al- 37
MIL-HDBK-42
24 May 1966
LLEMESTAL AC T2VIll
I Svtim Lurn-oo, -win-@. BWW

dtds d ceurneru aa necesaq
2 Act Ivlty tl PIW ttmr wallt~ Par-

lICUIU eOmpm-74 mabti lu4t0n. I
3 ~nimg md ClOaIU rebmr 4*/A,
4 GeImIU MCem d rctnetd;tq L,t/A,

eowrt (other then rhdomr i.
(l, -A,,)
)8s5-0.022 n, -Q3t9
A,,ol, -11
Mdfunciton 9 1- Chnew:ng Indlrat, m? ofdy

Verlf le..
tlon
JO 2 USIIW Irst rqutpmeni to verlf) D07 ? Uf, I
mdlu4ctt004 Ialwewly U R-
produ:tbl~ on WOWS
1: 3 Perf~mmg AW+ MM pmblem~

or chcckc.
12 4 Teettng [or pressure letit. 0065 V441
~-13 Altcn-@Ing 10 dmrvve tumve ).7!5W4,,l 005 IN,*)

or mo-emateal wpmmbl.
14 USIM CPH42 tern qwpmen! 0997 - O*Y5 Ip,,) - 1.627 (P,l] *,O
drDv& @eelficdtv for thlc
aymrrn
It Mti IW YWJS! !tietrlt) check. 1149 1P,,) - 0.036 {?,.1 I
FmlI
I Ocu,e.
+
IL 1 0.2!0 f~, ) 0670 (P,,) 0.012 lP, I 0030
I
1? 2. 0.228W,1. 0.097W,) - O.OIG(;:) 0623
I
i
lF 3
(%)
1s 4 0.0? N,, I
20 3
2 134 (p, ) - 0473

I
N,,
. CMnm cccur with my IX*? =ttvltv wt!b!n Ibr CKCOIV
.41-3R
MIL-HDBK-472
24 May 1966
STEP
U(, X LEMEYJTAL AC7W2TY PROBAB!LtlW YV14CWN
CYCCURKEY4CE
21 6 a??r rP,J amz + -o o v?) 002!
22 7
M 6
24 9 O.M 1P*) nz?l (*#l - 0.0s7
25 10 Qots W4,1 - O.w 1~~) - 0 0s$ &) 0017
26 11 0.5 lf~ IP,ol
n 12 0.002 (+ - 0.oo1 (MJI 0.096 (~,) 0.024
26 13 {?~ (?,)
B 14 moo4
T!
so 15
[0.,,3 ,PJ a352 ~+ * 0022] M*
8
. 11 1
32 2 I-*,,
33 3 0.02
w 4 1.4 444*O)
xl
s
37
38 onoo3t41 0.0?7 y * 0025
40 QO06 I%J
41 {P,,) {P,J -0.03
42
CM601 occur wtth any ohr =tMyy -ltUa b O*FY
Al-39
MIL-HDBK-472
24 May 1966
TABLE 1-10 LEGEND
( AS to be expressed as failures per hour)
N, - Number of flight-line replaceable components in system.
N2 - NMber of different t-ypes of flight-llnc replaceable components In

system.
N3 - Number of readouts in system, as determined by use of Table 1-11,
N, - Number of system readouts whose function is to evaluate a standard

test problem.
N5 - Number of different types of spares cmried aboard the aircraft.
N~ - Number of connectors (electrical or mechanical) which maintain

dynamic pressure integrity.
N7 - Number of CRTs in the system (excluding built-in test scopes)
Na - Number of circuit parameters monitored by built-in meters.
N9 - Number of test points in the system.
N,. - Is special test equipment, designed specifically for th~s system, used
for flight-line maintenance?
Lf yes, N,. = 1.
Lf no, Nlo = o.
N,, - Does the system contain any information devices, as determined

in Table 1-11.
If yes, N,l = 1.
If no, Nll = O.
N12 - Does the system contain any auditov devices, = determined In T*1c 1-11.
If yes, N12 = 1.
lf no, N,z = O.
(continued)
AI-40
MIL- HDBK-472
24 May 1966
TABLE 1-10- LEGEND (Continued)
N13 - Is special test equ!pment used to reproduce any infl ight conditions not
otherwse reproducible on the ground ~
If yes, N,, = I
If no, N13 = 0,
?4,4 - Does system contain any components which require an unusually long time
to reach positional or thermal steady-state condition? (Time-delay relays
and magnetron warm-up are not considered unusual. )
if yes, Nl~ 10
Lf no, N,4 o.
~15 - Anticipated average flight length (in hours).
td,6 - Are there provisions such as mock-up, go-no-go tester in the shop for
checking flight-line replarcable ~?
If yes, N16 1.
T.fno, N,6 =0,
N}7 - Are there provisions such as mock-up, go-no-go tester in the shop for
checking fl ight-iine replaceable black boxes ?
If yes, N17 s 1,
If no, N17 O.
N,. - lfN8 > 0, then Nl@ s 1.

LfN8 * O, then N18 = 2
N19 - Does system contain one or more -magnetrons?
N20 - Probability of occurrence of lo~stlc t irne, as determined by he usc of

Figure 1-10.
N21 - Readout factor, as calculated in Table 6-1.
A1-41
MIL-HDBK-472
24 May 1966
-
TABLE 1-10- LEGEND (Conttnud)
Pn - Probability calculated in the nt~

step of Table 1-10
Al - Failure rate of the system.
AZ - Summation of failure rates of flight-line replaceable components located

in the aircraft radome.
AS - Surnmatton of faflure rates of flight-line replaceable components for

which spares are carried aboard the aircraft.
A. - Summation of failure r*ee of fltght-hne ~hceabie components whose

operation is reflected by a readout.
15 - Summation of failure rates of those flight-line replaceable components,

whose operation is reflected by a readout and for which a apare is
carried aboard the aircraft.
A~ - Summation of failure rales of flight-llne replaceable components which

contain adjustment(s).
A, - Summation of failure rabes of flight-line replaceable components which

contain either adjustment(s) or flight-line replaceable parts, or both.
A8 - Summation of failure rates of flight-line replaceable parts (excluding

lamps and hardware).
19 - Summation of failure rates of flight-line replaceable parts for which

spares are carried aboard the aircraft.
A12 - Summation of failure rates of components accessible through -- or which

contain -- access covers (other than radome ), removal of which covers
may be necessary for flight-line maintenance.
A13 - Failure rate of that flight-line replaceable component which has the
maximum failure rate.
A1-42
MI L- KDBK-472
24 May 1966
TABLE 1-11
S}STEM READOUTS
(B)
Readouts Able to Reflect (A) Normalized (c)
System Operation Quantity Weight ing Product
Factor (A) (B)

I. Visual
A. Information Devices ---- 0.05 ----
[i. e., posit ion-,

velocity-, and
ang! e- indicating
instrume:lts)
B. Circuit Monitors m * o. 03+ ----

rim 0.003 (n-1)* ----
(i. e. , voltage-current, n m_ ----
power-pressure, and n E_ ----
frequency indicators)
C. CWhd~ Ray IX S@avs ---- 0.14 -----
D. Optics ---- 0.05 ----
Il. Auditory
A. Audio ---- 0.03 ----
B. Transmitter Sidetone ---- 0.03 -----
Total = TcAaJ =
(Number of System Readouts) (Readout Factor)
n is the number of different parameters thsi can be read on each

monitor by a switching arrangement.
x$ This ~cightin~ f=tor must be m~tip]ied by q of t~e ClrCUit
monitors listed in Column (A).
AI-43
MIb HDBK-472
24 May 1966
ADDENDUM B
DEFINXTIONS OF MAINTAINABILITY TERMS
1. Preparation time is the time spent obtaining, setting up, and calibrating
maintenance aids; warming up equipment; etc.
2. Malfunction verfficadon ttme ts the time spent testing the system to observe
previously reported synptoms of malfunction.
3. Fault Iocation time is the time spent arriving at a decision as to which items
caused the system to malfunction. This includes time spent working on
repiacing, attempting to repair, nad adjuating) portions of the system shown
by subsequent interim tests not to have been the cause of the malfunction.
4. Part procurement time is the time spent by the maintenance man in procuring,
or trying to procure, necessary replacement items.
5. fiepair time is the time spent replacing, repairing, or adjusting all items
suspected to have been the cause of the malfunction, except those subsequently
shown by interim test of the system not to have been the cause.
G. ?wffilf~l~tiuil finai kd Lime is ihe time spent com-irm mg that the malfunction
in question has been corrected, after which time no further maintenance
is performed on that malfunction.
i. Malfunction active repajr t!me is the composite of Items 1 through 6, above.
8. Systcm final test time is the time spent confirming that the system is in
satisfactory operating condition (as determined by the maintenance man)
following maintenance. It is possible for a system final test to be performed
after each correction of a malfunction.
Q
. . k~stic time is al! replacement procurement time, except that time when
the ma.intermncc man is engaged in the procurement acliv~tv.
10. I.n]tlal delay time is the time between the moment the equipment becomes
available for maintenance and the moment work is commenced.
11. hlalf unction administrative time is all time beiweer, the beginning .md end
of \\ork on a rnalfunchon, except for logistic or active maintenance time
[or that malfunction.
Al-44
MI LHDEIK-472
24 May 1966
12. System achnirdstrative time Is all system downtime other than active
matnteaance ttme and log fatic time.
13. S@t.em downtime h the time interval between the commencement of work
on a aydem mahnction and the time when tie system hae been repaired
antior checked by the maintenance man, and no further maintenance
activity is executed.
14. TotaI system dmmtime la the ttme intervaJ between the reporting of a system
mahmction and the tlrne when the system haa been reprd red ancVor checked
by the maintenance man, and no further maintenance activtty is executed,
15. Test point is a jack or similar fitting to whtch a test probe is attached
for measuring or observing a circuit pammeter or waveform.
16. Readout is a device butlt into the system which monitors, efther prl.mafily
or incidentally, the operation of some portion of the system.
17. Standard test problem is an evaluation of the performance of a system,

or any part of it, conducted by setting psmmeter~ into the system; the
parameters are operated on and the result obtatned from system readouts.
AI-45
MIL- HD13K-472
24 May 1966
PROCEDURE II
1.0 GENERAL
This maintainability prediction procedure describes the methods and techniques

which are used to predjct Corrective, Preventive and Active Maintenance
parameters.
As appljed in this procedure corrective maintenance time includes only actwd

repair time which js the period when repair work is in progress. Therefore,
it exclude~ such parameters of measure as admtrdstrattve time or log!stlc
time, ek. , which me usually considered in deflnltiorm of Corrective Mtidmance.
Similarly, preventive maintenance time includes ordy the actual active time which
is required for rcp~r during a preventive maintenance shut down. It does not
include preventive ma.tntenance time wldch is expended while equipment is
in operation.
Active rnaintcnance combines both corrective and preventive maintenance

because it includes the time when both of these activities are actually being
performed, The details imd parameters of measure are discussed in 1.4.
There are two methods which are presented for predicting Corrective Maintmance.
Th~ first method described in Part A of this procedure results in a maintainability
prwiict~on expressed in hours becauae it utilizes tabulated maintenance task repair
times, recorded in hours, which have been established from past experience.
Those data are discussed in 1.6 and tabulated in Table 2-2. The second method,
explained in Part B of this procedure, does not use tabulated task times- hstiti
it utilizes estimates of m?n-hours required to perfom a maintenance task which
are Insed OI? past experience )1 ~ analysis of the design with respect to malnten~ce.
The WO different measures, one in terms of hours which is representative of

aetunl elapsea Ezne, and the d.her in man-hours which is a measure of manpower
required to complete a maintenance activity h a given ttrne, have of necee!sity
rcsulteci in the development of a different symbology for each method. These are
mplained in Parts A and B respectively of this procedure and summarized in 1.4.
However, once the repair times have been established either in hours or man-hours
the actual ;~redi~tion uroce(tures for hth Parts A and B are very similar since
each uses LvOrk sneets whicn closely resemble each other.
2-1
~HDBK-472
24 May 1966
Part A 1s solely concerned with corrective maintenance, during the final design
singe of the product development cycle, and describes the method of obtxintng
a prediction of the mean time to repair expressed in hours when the information
&scribed in 1.5 is availnble. ~/
Part B cover6 active maintenance which includes both preventive and corrective
maintenance and details the methods for obtaining an estimate of the mean time
expressed in man-hours for performing botk types of maintenance.
1.1 Philosophy, Assumptions and Summary
The two most important parameters of measure in the field of MaMatnablHty are
the durzttfon of downtime due to maintenance and the number of personnel required.
Each are important measures of matntainabllity and ideally, both should be kept
at a minimum if possible. However, for a critical mission, the number of maMe-
nance man-hours required may not be as important as minimizing the time required
to repair regnrdiess of the number of men involved or the inefficiency of their
utilization. Conversely, when downtime is not of paramount significance, the number
of ma-hours becomes an important pammeter of measure and control. This pre-
diction procedure outlines the methods of predicting both parameters of measure,
the results of which can be utilized for design improvement or other evaluations.
7%1s procedure aasumes that repair times, applicable to certain equipment

types, can be used to predict the maintainability of Hke equipments and devices.
Therefore, ittabtdatee certain repair times which can be used as described tn
2.0 and 3.0 to develop a prediction of corrective maintenance in hours.
bother assumption is that reasonably guod estimates can be made of the

maintenance tasks required for corrective and preventive maintenance ~ analyzing
basic features of the design. Moreover, it is asaumed that the number of hours
or man-hours required to perform these tasks for both corrective and preven-
tive maintenance can be estimated from the information developed.
A fundamental philosophy in any case, is that the magnitude of the repair time,
for a discrete repair, is the sum of the individual maintenance task times which
are required for its completion. Seven such maintenance asks are assumed to
effect the magnitude of maintenance time. These are: Localization, Isolation,
Disassembly, interchange, Reassembly, Alignment and Check Out. The procedure
~/ Adaptations of the ksic prediction method for each of the earlier stages of the
product development cycle arc detailed in MIL-M-23313A and NAtSH7PS 94324.
These include: Early Conceptual, Equipment Planning, Earl: Development and
bite Development Stages.
2-2
w
MIL-HDBK-472
24 hfay 1966
also assumes that each of the maintenance task times are a function of the
method and level of repair. For e=mple, it is postulated that it should take
longer to replace a pan whicJ is wired than is required to repiace a part which
is plugged in. Also that at the part level of repair it takes longer to perform
a repair task than at the subassembly or equipment level because Aesa time
iB required for the discrete steps involved at the latter levels, Accordingly,
nine functional lewe}a of repair are described. These are: System, Subsystem,
Equipment, Croup, Unit, Assembly, Subassembly, Stage and Part.
In summary the assumptions and philosophy of this procedure are applied in

predicting either corrective maintmwme measured in hours, or active maintenance
time consisting of corrective and preventive maintenance measured in man-hours.
The techniques of application are described in 2.0 and 3.0.
1.2 Applicability
This procedure can be applied to predict the corrective maintenance time, as

outlined in Part A, and the active maintenance time, aa described in Part B,
of shipboard and shore electronic equipment and systems. It can also be used
to predict the maintainability of comparable equipments and systems of other
branches of the armed services provided there is similarity of design, we and
application.
This procedure may also be applied to predict the maintainability of mechanical

equipments or systems if the maintenance task times for the applicable functional
levels can be established. ~)
1.3 Point of Application
The maintainability prediction methods described in this procedure are applicable

only during the final design stage.
1.4 Basjc Parameters of MeRsure
The basic parameters of measure which are applicable to this procedure are as
follows :
1.4.1 Corrective Maintenance (Part A)
The basic measure of maintainability for this method is the Equipment Repair
Time (ERT) expressed in hours.
~ For &finitions of matite.~ce tasks and functlond levels, see 2.0 entitled
Analytic Foundation.
2-3
MIL-HDBK-472
24 M8y 1966
The ER T is defined as the median of individual repair times and is expressed

by specific formulations for various statistical distributions as follows:
(a) When repair times follow x normal dmtribution the basic parameter of
measure is the Mean Time to Repair (M7 T R). Since for this distr-
ibution, the median is equal to the mean, the MT T i? is a satisfactory
measure of the true ERT. The ERT is therefore equal to the MT TR
and is expressed as follows:
= MTTFi (2-1)
where: A = average part failure rate in failures per 106 hours.
Rp = repair time required to perform a corrective maintenance action

in hours.
(b) When repair times follow an exponential distribution:
ERT = 0.69 MTTR (2-2)
[c) When repair times follow a l~g-nerma! d!stribut!cn 0! repair times:
MtTR
ERT = (2-3)
ontilog (1.15 t72)
where: ~ is the standard deviation of the logarithms, to the base 10 of repair

times. The average value of C7is approximately O. 55 in which case:
ERT = 0.45 MTTR (2-4)
(d) When repair times follow a log-normal distribution the geometric

mean time to repair (M TTRO) occurs at the median, therefore
it is a measure of ERT .
The geometric mean time to repair (M TTRG) is:
(2-5)
2-4
MXL-HDBK-472
24 hf8y 1966
1,4.2 Corrective Maintenance (Part B)
~e applicable formulation to obtatn the mean corrective maintenance time

(Mc) in man-boure fE:
(2-6)
where: A= average part failure rates tn failures per 10 hours.
Me = the man-hours required to perform a correctl ve maintenance

task, (Corrective Maintenance Time)
1.4.3 Preventive Maintenance (Part B)
The mean preventive maintenance time (Rp) is expressed in termf3 of mean man-
hours and is equated as:
~ (fMp)
Rp : @-7)
Zf
where: Mp = fie man-hours required to perfomn a preventive maintenance

action. (Preventive Maintenance Time)
f = the frequency of occurrence of preventive maintenance actions

per 106 hours.
1. 4.4 Active Maintenance (Part B)
This parameter is expressed as the mean active maintenance time which {E the
sum of the preventive and corrective maintenance man-hours required to
maintain a product for a apec!fied period, divided by the total number of preventive
and corrective maintenance tasks required during that time. Mathematical y
ii is expressed as follows:
(2-8)
2-5
MI~HDBK-472
24 May 1966
where: E = mean actl ve mainLmance time
Wc = mean corrective maintenance time (resulting fmnn time, ti)
Wp = mean preventive maintenance time (during time, tj)
~~ = the mm of the put failures mtes
2 f = the mm of the frequency of occurrences of preventive maintenance

tasks.
Li = operating time during the period tj
tj = calendar time in operating inventory
1. 4.5 Maintainability Index (MI)
The total maintenance effort required to maintain a product in operational

status per unit of operating time is expressed as:
(xx) iict~ + (Zf) Iiptj

MI = (2-9)
i
where: t = operating time
~C, ~p, tit tji X A, Zf have been previously defined.
1.5 Information Required to @ply This Prcmdure
Since this procedure Is applicable to predict maintainability during the final

design stage, pertinent functional level and design detaila must be available.
These will inclu& but are not necesearlly restricted to tbe foliowing:
(a) Packaging: The equipment packaging arrangement muet lM Imown

to the extent that a &tailed breaiwbwn into tbe various equipment
groups, units, assemblies, subaeaembllea, and parts can be
determined,
2-6
MI~HDBK-472
24 May 1966
(b) Diagnostics: The diagnostic procedure that would be followed

in the event of failure of each part m the equipment, including
application of all designed fault location features, must be available.
(c) Repair Methods: The repair methods that will be used to correct
failure of each part must be known.
(d) Parts Listing: A complete listing of all parts categorized by

equipment subdivision is required.
(e) Stresses: Electrical and environmental stresses under which each

part wilI operate.
(f) Mounting: The method by which each individually replaced part, and
each repIaceaMe subassembly), assembly, or unit is mounted must
be known.
(g) Functional Levels: The functional levels at which alignment and

checkout are performed, following the replacement of each part,
must be known.
1.6 Data Basis
Part A utilizes tabulated data compiled as a result of over 300 observations

of maintenance activity in the U. S. fleet. These data in the form of corrective
maintenance task times, as shown in TabIc 2-2, are the basic data used for
predictions of active repair times.
Part B, depends on the establishment of equipment maintainability parameters

based on experience since this method of predicting active maintenance times
does not utilize tabulated data of maintenance task times. Where experience
information is lacking, estimates are required of the minimum permissible
values of equipment maintainability parameters to assure that the operational
requirements can be achieved with:
(a) the planned operating and maintenance equipment
(b) the designated and planned maintenance personnel
1, 7 Correlation BeMecn Predicted and Observed Jalues
A validation study carried out or. the AN URC-32 Transceiver and the AN SRT-16
Transmitter sho~,ed good correlation be~een predicted and observed corrective
2-7
NDkXDBK-472
24 May 1966
maintenance resuIts. These equipments were helng ut[lized in many ship types
from destroyers to submarines.
2.0 ANA LYTIC FOUNDATION
Active Maintenance time consists of two basic components, namely, Corrective

and Preventive Maintenance time. Corrective Maintenance ie the Maintenance
performed to restore an item to a satisfactory condition by providing correction
of a malfunction which has caused degradation of the item below the specified
performance. Preventive Maintenance is the maintenance performed to retain
an item in satisfactory operational condition by providing systematic inspection,
detection, and prevention of lncjp{ent failures. Preventive Maintenance can be
either scheduled CT unscheduled depending upon the requirements of the m,ission.
Active Maintentice Prediction assesses tbe average man-hours of work tn

perform the required corrective and preventive maintenance tasks. These
active maintenance tasks do not consider the effects on elapsed maintenance
time due to logistics problems or administrative procedures. The active repair
time estimate of corrective maintenance predictB the downttme due to active repair
which is the result of a malfunction causing system downtime. The preventive
maintenance time estimate, on the other hand, predicts the downtime duc to
preventive maintenance activities.
2.1 Ba8ic Considerations and Definitions
2. 1.1 Applicable Maintenance TaSkS
The corrective maint.enancc action is divided into the following corrective maintenance
tasks:
(a) Localization - Determining the location of a failure to the extent possible,

without using accessory test equipment.
(b) Isolation - Detemnintng the location of a failure to the extent possible,

by the use of accessory teet equipment.
(c) Di sassemblv - Equipment disassembly to the extent necessary, to

gain access to the item that is to be replaced,
(d) Intercha= - Removing the defective itcm and installing the J eplaccrnent.
2-8
IvIILr HDBK-472
24 May 1966
(e) Reassembly - Closing and reassembly of the equipment after the

replacement has been made.
(f) Alignment - Performing my alignment, minimum tests and, or

adjustment made necessary by the repair action.
(g) Check Out - Performing the minimum checks or tests required t.n
verify that the equipment has been restired to satisfactory performance.
2. 1.2 Applicable Functional Levels
(a) Part - One piece, or two or more pieces joined together, which arc
not normally subject to disassembly without destruction of designed use.
(Examples: electron tube, resistors, mica capacitor, audio transformer).
A part may be a replaceable item,
(W Swe - A combination of two or more parts which form a portion of a

subassembly. A stage ts usually considered as one tube or transistor,
together with its directly associated parts. (Examples: amplifier stage,
detector stage). In certain cases a stage might contain two or more
directly associated tubes or transistors such as in the case of a push-pull
amplifier stage or binary counter stage. A stage is not a replacement
Item itseif. If an entire siage is repiaued, the replaceable item is termJcd
a subassembly, assembly or unit.
(c) Subassembly _- Two or more parts which form a portion of an assembly

and which is replaceable as a whole, but having a part or parts, which
are individually replaceable. (Example: IF Strip, terminal bmrri with
mounted parts).
(d) Assembly - A number of parts, or subassemblies, or any combination

thereof joined together to perform a specific function and replaceable
as a whole. (Example: audio-frequency ampjifier). The distinction
between an assembly and a subassembly is not always exact. An
assembly in one instance might be a subassembly in another where it
forms a portion of an assembly.
(e) Unit - Any combination of parts, subassemblies and assemblies

mounted together, normally capable of independenPoperation in a
variety of situations. (Example: electronic power supply, radio
receiver), A unit is normal!y directly accessible (e. g. , mounted
2-5
MI L- HDBK- 472
24 Ma} 1966
4
on the front panel) and can be removed without prior disassembly

of the equipment or group. A unit may be a replaceable item.
(i) Q$!W - .4 collection of units, assemblies, or subassemblies

which IS a subdivision of or system,
an equipment but which
(set)
]s not capable of performing a complete operational function.

(Example: antenna group, indicator group). A group is not
normally a replaceable item.
k) Equipment (Set) - One or more units and necessary assemblies,

subassemblies, and parts connected or associated together and
inc!udins all necessary cabling and accessories to perform an
operational function. (Example: radio receiving set, sound measur-
ing set, radar set). An equipment is not normaliy a replaceable
item.
(h) Subsvstem - A combination of equipments, groups, etC. , which

perform an operational function within a system. Subsystems form
the major subdivision of a system. (Example: one station of a
communications system). A subsystem is usually located on one
ship, or in one geographical location.
(i) Svstern (Electrical - Electronic) - A combination of two or more

subsystems or equipment, generally physically separated when
in operation, and such other assemblies, subassemblies, and
parts as are necessary to perform an operational function.
(Example: communications system including all stations; fire
control system Including the tracking radar, computer and gun
mount).
2, 1. 3 Replaceable Item
A unit, assembly, subassembly, or part which is normally intended to be replaced

during corrective maintenance upon failure of the item.
.-.6
) Principles of the Maintainability Prediction Procedure
The maintainability prediction method described in this procedure depends upon

utilizing the estimated maintenance task times required for performing the
specific maintenance tasks which, in total, comprise a repair or maintenance
action. These maintenance task times may be obtained from tabulations,
2-1o
IbflLr HDBK-472
24 May 1966
such as Table 2-2, of Part A of this prediction procedure or they may be estimated
times as described in Part B. The sum of the maintenance task times is then
multiplied by tbe failure raw, expressed in failures per 106 hours to obtain an
estimate of the number of maintenance hours required for that specific maintenance
or repair action. The mean value is then detemnined by summing the number of
maintenance hours and dividing by the sum of the failure rates.
A similar procedure iE followed for preventive maintenance except that, in this

case, instead of using failure rate parameters, fre~~cy, (f) is used whi~ is
a measure of the number of preventive maintenance actions for a similar pertod
of 106 hours. Hence the mean preventive maintenance action time is the sum of
the product of frequency and preventive mainteriance time divided by the sum of
the frequencies.
The symbols which are used for corrective and active maintenance are explained
in 3.0 which describes the application of this procedure,
Failure rates for use tn performing a maintainability prediction may be obtained

from, but are not necessarily restricted to the following:
(a) Handbook for Prediction of !%dpboard and %ore Electronic Equipment

Reliability, NAVSHIPS 93820, Naval Ships Systems Command.
(b) Reliability Stress and Failure Rate Data for Electronic Equipment
Military Standardization Handbook, MIL-HDBK-217.
Through the use of this procedure quantitative predictions of maintainability can

be obtained in hours or in man-hours depending on which units of measure are
used for the maintenance action time.
3.0 APPLICATION
The application of the techniques outlined for both corrective and active maintenance
prediction arc described in I%ti A and Part B which follow:
3.1 PART A Corr~ctive Majntena.nce Prediction Procedure
As stipulated in 1,0, General, this procedure is only concerned with the final
design stage phase of the product development cycle. The earlier ~d in~r-
mediary design stages are not treated in this procedure. Tht prediction proce-
dures for these phases of the product development cycle are included in
NAVSHIPS 94324 and M1bSTD-470.
2-11
MILHDBK-472
24 M8J 1966
-
3.1.1 Maintainability Prediction During Final Desdgn Stage
TMS procedure is intended to predict malnbbabfli~ at the flna.1 design stage.

It takes into account all of the pertinent functional level and details available,
and is applicable when the information lfEted tn 1.5 la available.
3.1.1.1 l%ediction Procedure
lWs matntainabtlity prediction pxwcedure involves determining the detailed functional

level breakdom of the equipment, eatablidhg the functional levels at which xnainte-
nance features are effective, and at which maintenance tasks will be performed,
and performing the prediction ?med on equtpment design features and approximate
part failure rates.
3. 1.1.2 Determining Iknctional tivel~ of the Equipment or System
The first step in the procedure is to detemntne the functional level breskdown of
the equipment or system. This is done by divjding the equipment or system into
its various physical subdivisions beginning with tbe highest subdivision and
continuing down to the ftems such as parts, subassemblies, assemblies or units
that wU1 be replaced tn corrective maintenance. The various functional levels
or subdivisions (part, stage, aubaasembl y, assembly, unit, group, equipment,
subsystem and @y@tern) of ccnce!m =re defined in paragraph 2.1.2.
The functional ievel breakdown is most easily established and certain determinations
required during the prediction are mor~ easily made tf a functional level diagram
similar to that shown in Figure 2-1 is prepared. A hypothetical communications
system is tmbdivided Into Its varioug subsystems, equipments, groups, etc. ,
down to the item~ that will be replaced during corrective maintenance. Each branch
of the diagram is terminated with a circle which tndicatea the item or items that
will be replaced to correct malfunctions existtng in that branch,
The connecting lines indicate physical relationships and not electrical or

operational connections. In preparing such a diagram, care must be exercised
tn establishing the appropriate functional levele for the various subdivisions,
especially where an item may have a nomenclature that includes the name of one
of the functional levels (e. g. , Power Amplifier Assembly), in some instances,
the functional level location of an item may not be the same as tic nomenclature
indicates.
2-12
MILPKDBK-472
24 M:ay 1966
L
-----
i I ~1 I I
i i
I I 1 I.ri I I
I i
I I
I I
I
I
I
:
I
I I
I
I I
I I
< .,
I I
I Q) i
i I
-. I I I
I I i
1 I I
1 1 l;-=l
I ! I w
- I I I
I
I
I
1! 1
I
I I
I
,
I
I
I
(
I l
,,,
1, I I /
I 11 I
,
I
I 1 i I
I
I
I
I
I
l-!
-..
2-13
MIL- HDBK-472
24 May 1966
3. 1.1.3 Determming Functional Levels at Which Maintenance Features Are Effective
After the functional level breakdow~ has been established and a functional level
diagran~ prepared, the functional ievels at which focalization, isolation, accees,
alignment, and checkout features are applicable should be determined based on
the overall characteristics of the design.
The functional levels at which features for localization, isoiation, alignment, and
checkout are effective for each replaceable item can be indicated on the functional
level diagram as shown by the symbols A in Figure 2-1. The access functional
level can be determined directly from the functional leveI diagram as indicated
m c below, therefore, a symbol identifying it is not required, The functional
level at which each c)[ these features i~ effechve is determined and shown in the
functional level diagram as follows:
(a) Localization: The functional level to which a failure Cm be Iocamd without

~s~ng accessory test equipment is indicated by ~.
b) Isolation The functional Icvel to which a failure can be located using

accessory test equipment at designed test points is indicated by ~.
If there are no designed test points below the localization functional
level, then isolation is assu.m ed to be effectt ve at the same functional
level as lmm.lizsttiort.
(c) Access: The access functional level for a replaceable item is that
level to which disassembly must be accomplished in order to gain access
to the item that is to be replaced, and from which reassembly must be
accomplished after replacement of the item. This can be determined
directly from the functional level diagram as the functional level of the
first rectangular block abcne the replaceable item. For example, in
Figure 2-1, to replace a part in a modulator power supply access
must be gained h the unit level, and to replace a subassembly of the
TT Demod assembly access must be gained to the assembly
level.
(d) Ali grunent: The functional level from which alignment must begin
following replacement of a variable item such as potentiometers;
or follo~ing replacement of subassemblies, assemblies, or units
containing tuning drives, mechanical couplings or other major
adjustable devices is indicated by A
(e) Checkout: The functional level at which restoration to normaI services is

to be verified using self-test or other testing facilities is indicated by ~,
2-14
MIL-HDBK-472
24 May 1966
3,1. 1,4 Prediction
The actual prediction is perfomed using a worksheet such as that shown in

Figure 2-2. These sheets are prepared according to the instruction below
for the appropriate columns.
step 0). item( Indicate the item being covered by the worksheet. When
individud parts are replaced indicate the item which directly contains the
replaceable part. When modular subassemblies, assemblies, or units are
replaced, indicate the item that is being replaced.
Skp (2). Method of Repair Indicate the t~pe of item that will actually be
replaced to eliminate a malfunction.
step (3). C]rcuit Designation List the circuit desi~ation (e, g. , V101)
of each part in the item designated at the top of the worksheet, with the exception
of the following categories which should not be listed:
Mechanical hardware such as:
gear trains chassis rollers
clutches {mccnanicaij sure-ws
bearings nuts
cams washers
dials and registers tube shields
Electrical hardware such as:
feed-through terminals wiring harnesses
terminal studs sockets for plug- in parts or subassemblies
wiring terminal boards
cables printed wtrlng boards fboard only)
Vlen the worksheet covers a replaceable modular subassembly, assembly, or

unit, begin the list on the second row of the column.
2-15
-
MTL-HDBK-472
24 May 1966
-

Km

u
. Q
._
!
I I , I
I
~,li z
1
I
2-16
hfI&HDBK-472
24 May 1966
step (4). Part Type Llat the identifying names or type designations of all
parts opposite their re8pecti ve circuit &slgnationa. men the worksheet covers
a replaceable modular mkssembly, assembly, or unit, enter the designation
of the replaceable item LOthe top row of the column followed by the designation
of each pm.
Step (5). Failure Rate List the failure mte ~, of each P* opposite tbe
appropriate part designation. The fafiun rate for each pafi and tube tn the
equipment can be determined by use of standard reliability prediction methods
wtth applicable electrical and thermal stress condltiona considered. Failure
rate data should be taken from sources which condder only tie random
catastrophic failures asaocinted with normal equipment operation. %Itable
references for failure rate data were listed in 2.2.
When using these references care should be exercised in not mixtng fathare
rate data of one with the other. Once a selection has been made this same
reference should be used throughout uxdeas other failure rate data not included
theretn iR required. in thlE case other sources may be used if justified and
approved.
IMermi.ne the sum of the ~ values listed and record this value in tbe awn
space at the bottum of the A mlumn.
~p (6). Maintenance Taak Times (Columns 6, 7, 8, 9, 10, 11 and 12).

For individually replaced parts detexnnine the maintenance task times through
the use of Table 2-2, For each part, record theue task times in the appropriate
columns oppos!te the respective part destgnationm For replaceable modular
subassemblies, assemblies, or units, determine a single time interval for each
maintenance task for the overall replaceable item and record each Ibne interval
in the top row of the appropriate maintenance kmk column. Table 2-2 provides
charts for determining maintenance task times for the localization, isolation,
disassemble y, reaasembl y, alignment, and checkout tasks. Interchange time
is determined through the use of work factor data aa shovm in Tables 2-3 and 2-4.
The use of the charts tn Table 2-2 requires determining the specific functional
level within the equipment at which all maintenance features are effective.
As used in this procedure, functional level denotes the physical subdivisions
of electronic equipment at which diagnostic, replacement, and test features
we effective. Tbe functional levels under consideration are: 1) part. 2) stage,
3) subassembly, 4) assembly, 5) unit, 6) group, 7) equipment, 8) subgystem,
and 9) Bywtem.
2-17
MIL-HDBK-472
24 MaY 1966
TABLE 2-1
AVERAGE PARTS FAILURES PER PART PER 106 HOURS BY PART CATEGORY~/
Failures/106 Ftilures/i06
Part Category Part Calegory
Hour~ Hours
Capacltora Rectifier. (Power Selenium) 2.96

FIud
Cermmic, Feed-Thru 2.06 Relays 2.06
Ceramic, (Otbcr Than
Feed Thtu) o. s? Re~latorfi
EkctroIytlc 2.48 Fixed
Mica 0.46 Carbon Composition 0.36
P@er 0.50 Carbon Film 0.39
Metal Film 0.51
Variable Other Film 0.60
Air 0.13 Wi r?would 1.40
Ceramic O. OB7
P18st\c 0.00045 Variable
Vacwm 3.08 Compocltlon 0.42
Wire !Vound O. 84
Connector 0.058
Resolvers 25.47
Counters, Mechanical 4.54
(TaUy Register Type) socket8 0.042
Cryatale, Dmde 2.98 SWilches 0.48

(Signal
Miser)
Detector and
Synchrm
I 1.80
C~stala, Frequency 1.36 Tranafomners 1.16
Hardw~ Trandntoro 1.03

Electrical 0,033
Mechanical 0.092 Ttia, Eiectron
Crt 24.66
Inductors (Deflection, 0.28 Magnetrons 544.54
Focus, Pulse, Filter, ReceiVi~ (including

R. F. (MIs) VR T*s) 9,42
Special Pupose and
Metera Tranmntttitg (in-
Electrical 1.36 c] U&B ATRO TR,
Mechanical 2.19 Klyntron, ThyrUron) 42.88
Motors Vlbratoro (Chopper) 32.18

Blower 2. 8s
Reao)ver 25.47
Synchro 1.80
Other S.84
~/ Reproduced from Handbook for the Prediction of Shipboard and Shore Electronic Equ~pmeM
Reliability NAVSHIPS 93820.
2-18
MIL- HDBK-472
24 May 1966
The charts in Table 2-2 are used In the following manner:
The appropriate chart, TUBES or P.4RTS. is selected depending on the t.vpe of

item to be replaced. The application notes for the chafis define the specified
types of items each chart applies to,
To ohin localization md isolation time. the chart is entered using the FUNCTIONAL
LEVEIS column headed by the type of replacement that will be made. For example,
replacement of a part would indicate Column 1, a stage; Column 2, a subassembly;
Column 3, etc. To obtain disassembly, reassembly, alignment, and checkout
times, the chart is entered using Column I only.
The appropriate corrective maintenance task column is deterrn ined b} the type
of maintenance task being considered.
The ro~v for the maintenance task time is found by determinltig the functional
leveI at which maintenance features are effective, or at which each corrective
maintenance task is performed,
Special limitations, regarding specific application of the information amd data

in Table 2-2 are detailed in the application notes associated with the table.
he average maintenance task times shown in the PARTS chart are not applicable
to n-IeChfJAid device. - SM...
.,. -L -. ~~mw . . . .... ---
~. ~- +7.mime .nrl ~~~~ing~.
!n adriitinn, the interval
between the time a mai.function occurs and the time that an operator or maintenance
technician becomes aware of i~ has been excluded.
Examples
The following examples are presented to demonstrate the use of Table 2-2:
(1) The Modulator Power Supplv. illustrated in the Functions: Level Diagram
of Figure 2-1. is demonstrated in this example because of its apparent
localization and isolation features. From the Functional Level Diagram.
localization for the power supply is determined to be effective at the
Equipment Level. Considering the Power Supply to be a plug-in chassis,
then the Tubes Chart of Table 2-2 is used. Since the Power Supply is
located at the Unit Level where failure can be removed and localization
is effective at the Equipment . Level, tbcn the chart is entered from the
unit Column (5) down to the intersection of the Equipment row, This row
is +&en carried over ta the intersection with the localization column where
the maintenance task time is read as 0.037 hours.
2-19
MIL- HDBK- 472
24 May 1966

r
.
, .
L
.
J-
!mt
1 ,-.

2-20
MIL-HDBK-472
24 May 1966
TABLE 2-2 (Coetl~d)
A.PPLICATYON NOTES AND DE ~ItWTIONS
NO~S (me wernrrlpl \eI1*r* on ske[ 1)
1. Tbe verage t-k times }n tbc TUBES eh-rt Add be ~Pl~ed to front D@ Iuees, eaceemd PIUS-IB mtxkler
subsoeemblt*=, Pluc-tn cvctals. end tubes {eseltii~ ~iel pup- ttifi).
2 The sv?fege teek t!meo fn the PARTS chtrt should be mmtled to IUUD nol w the Iront panel, Irerrwotors,
spec IaJ purpose ttiree lmagnclnmB, klyctrww, lc. ). wl~d In non-waled moduleo, wd ell elerlmnlc
parta (rea Lslors, c~acttora. Ihetors, treosforene?s. etc. ).
The IOLd rcpiaccmen! time hould include dl~ueembly time, $nlercho~e t!me snd reeeaembly flm* However.
stm- ar, vcrege intei+mncr timt cppllceble 10 all 11UUI08B Is dlfflcutt to @Natn, 11 Imo not been Imtuded m
tn? tcbuktied vere#e leek ttme. httemkeage lime muet be dewrmiaed lrom TdAc 1-3-13.
; %eclf~c mcbnlquea are u-cd to Ieolale lube Iatlumc. The dest~r must corrsldcr Ihes? Iechnlques In
ttetemnlnlng R. Ior tubes.
[a) Test polrth sre nonnelIy 001 ueed to Iedete tube laliures.
(b) Tbe mosl common method 0[ iacdeling lube faliures duri~ corrwtlvr malolenence is by ubdltutlon of
koown good ttbe for the au-ted lube. Ahbough lsolat~OtI Ietturm msy be effecthe U 1* urnll or
eeeembly level, tdte iedetlon can be eccoetplkebed at tke tube or s~o ievd tbro@ eubetltutto~ moecte
*%34 SS:S met paste.
(c) Some tube falhms Icqen IIJwneols, bmlteo ovf~ope, etc. ) CM be kvcaed tieudly; boerever, this Ie
true in only ve~ real! pemen!ue of cWt. -
2 Becbuee Ihe tesk durationo were &rived Imm held eka, thie mrdat(on time includee the tube checke. tc. .
lbe electronic lachnlcln uw~ly makes belore -Idlw to we tern Points md Polnt-Io-polr9 checks Tbt8
verage time akeo ~wrds for the POLIA--POM checks wWch -uld ueud) y be ~irtd 10 isOIUi k Idled PU1,
TM ~verue ttme (ntervale in thIo cha~ ~ UII Include dm~rdetrettvt time. A&nlolstreilve time coneleta of lime
ewendcd In Pam procurement ltlme ~ot by malntertame pereonoel 10 ObWnlog ~lecement Ilema) and to rmo-
iechnlcal nhtpboed rmut Ine (!nepectlon, eating, 04her mllltary rbAteo, etc. ).
When quipment matntenenee Ieeturec e-e locallxetlon to Ike futvotmttal level !hrottgh wh!ch Iallure is bel~
=mowd @te t- m of FUNCTIONAL LEVELS coltmm),& mt ttee
thevalue ehmrn In the ISOLATION column at
this functlond level, instead use 0.000 bows.
To determiw the time for the cflsmeembty, twueemt#y, m)lfrmtent ettd ebeckout tacks, the &stgoer ehould mdy
uoc column I of WNCTfONAL LEVELS in the ~roprlate nw M wftkeh tb teak ia per{omned.
To properly Wproxlmste the checkou! time, enter column 1 u the Iuncttond level M which the checkout Is bel~
mad. and multlply by the number of @erU!onel modes effected by the rqleced funetmrul level.
.< stu@ made u ltK Federe! Elcclr,: Cowratlon of 762 tube replacements tn one type of h,phoe.rd !ransmltl?r
!ndmaled Lhu only SO Miums (7%} could Poestbly have been located th~uO vl~u~ mc~nt
2-21
MI L- HDBK-472
24 May 1966
(2) Similarly, isolation for the Power Supply is determined to be effective

at the Un:t Level from the Functional Level Diagram, F!gure 2-1.
Thus entering the Tube Chart of Table 2-2 in the Unit Column (5)
to tie Unit Level row intersects at the top row. Carrying this row
over to the intersection with the Isolation column presents a maintenance
wsk time reading of O. 265 hours.
(3) U an equipment must be disassembled to the unit level to replace an

assembly, the disassembly time will be O. 094 hours. Thm time is
obtained by locating the intersection of the UNTT row of Column 1 of
functional levels, and the DISASSEMBLY column. Note that functioned
level~ Column 1 is uBed in all caees except localization and isolatlon.
Step (7). In@rchange Time This ttme can be obtained from Table 2-3 for
most typesof tubes and replaceable itemg. For items having mounting methods
not conforming to the description given in Table 2-3, the interchange times can be
determined from Table 2-4.
In determining the interchange time from Table 2-4, each deta[led step (unsolder
joint, remove unit, etc. ) tnvolved h removtng a failed ftem and in positioning
and attaching a new item must be accounted for with a specific element time
interval, The same ttrne interval for an attachtng etep can be used as an appNM-
mation for a corre apondtng detaching mep. The interchange tune for an hem is
equal to the sum of all ttme intervals. An example of determining interchange
time for a part {B shown in Table 2-5.
Step (8). Rp
Determine the repair time ( Rp ) rewired for Perform:llg a o~~e
corrective maintenance action in the caee of failure of each replaceable item by
adding the values recorded on the worksheet in Cohm.ns 6 through 12.
For individually replaced ptmts, RP is determti~ ~dependently for ~~ part)

and is recorded In the Rp column of tie worksheet opposite tbe respective pati
designation.
For replaceable modular eubaseembliee, aeeemblies, or uniti, Rp is ~te~~~

once for the overall replaceable item, and is recorded on the worksheet in the
top row of the Rp coluznn.
step (9). ~RP When individual parts are replaced, multiply each recorded
part failure rate { ~ ) by the reepecttve value of Rp, -d record tie pr~ct ~ the
~RP column oppoeite the respective part designation.
2-22
MIb HDBK-472
TABLE 2-3 24 May 1966
INTERCHANGE TIME
TL!13ES AND FL!SES
PART TYPE AVERAGE TIllE (HOURS)
Plug-i n tubes 0.015
Wired tubes (4 wires) 0.149
U:red tubes (more than 4 v.nres) O. 149 + O. 034 per wire over 4
All tubes with shield ADD O. 007

with C~iUllJl ADD O. 027
with C~ , ADD 0.007
Plug-in fuses 0.010
Screw-in fuses 0.015
All fuses with screw cap 0.014

PARTS OTHER THAN TUBES M4D FUSES
PART TYPE AVERAGE TIME (HOURS)

-.
IJzrts with 2 vnrcs or 2 tabs
to be soldered 0.081
Parts wvth more than 2 Wires or

2 tabs to be soldered 0.081 + 0.034 per wire over 2
with clamp ADD 0.027
Parts attached with screws,

nuts, and washers ADD O. 022 for each screw, nut and
washer combination
NOTE: For attachments not conforming to the above type deterrnine the inter-
change time by using work factor dat% Table 2-13
.
MIL-HDBK-472
24 May 1966 TABLE 2-4
ELEMENT TIME9 BMED ON THE WORK FACK)R SYSTEM
ELEhfENT TIWE
ELEMENT DESCRLPTION
{Hours)
1. PLUC-tN9 (includes hmdllw)
PID-type tubes. plug-in parts, etc. o.007s

Tube cw or shield 0.003s
Fume
loseR Intohorizontal holdrr o.00s0

Inseul into vetiicsl bolder 0.0075
2. WRING A- SOLDERING
Wtre - wr8ppIog Md SPllclq

Bue copper wire (1) Ed 0.0150
(2) Ends 0.0237
Jumper wire md csble leads (1) End 0.0134
(2)Ends 0.0265
Put wllh urts.1 leads (Includes part )mtdllrg)
(1) Eod 0.0176
(2) Ends 0.0289
tihkrt~
Per )olnt 0.0058
3, REPLACE M E Nl WITH HARDWARE
Rcplsce scnw lato t-d bole 0.0093

RepIsce screw through clesrsnce hole O.0023
Replsce w8s&r o.Oolfl
IWplsce ad 0.0071
Rtp\ace stop nut 0.0210
Replue 6et ucrmw 0.0075
Apply glyptol Ocrew 0.0018
4. PART HANDLING
Pull up pm and ponltion in chsa~ia for saeembly 0.0025
5. PRINTED CiRCUIT WIRING
Replacc (Inseti) o.oo33/End

Solder O. 0056/End
T& ioterchmge tuk abosrd ship rhea mt Invoive the repetitive (cyclic) factory type of regulsrly
occurriog motlon8: lherefore, these interchange element times are based on non-cycllc and
I rrefpdarly occurring motions.
2-24
hflL-HI)BK-472
24 May 1966
-.
u
tc
w
Ceoo 00 DC-O 0
I ,, 1( I
KKVW *K XXX
2-25
MIL-HDBK-472
24 May1966
When subassemblies, assemblies, or units are replaced, multiply the recorded

sum of ~ 8 (from space 17) by the recorded value of Rp.
Determine the sum o; the ,\Rp values listed and record tie in the sum space
at the bottom of the kRp column,
,,S&!p (1o). log Rp List the logarithms of each recorded value of Rp m the
log Rp column oppdte the respective Rp value. Where values of Rp are less
than one, the logarithms should be expressed as negative numbers. (For example,
log 0.25 = 9.3979 - 10 = -0.6021).
step (l]). ~ log Rp When indltidual parts are replaced, multiply each recorded
part fidlure rate ( ~ ] by the respective value of log Rp for that part, record the
product in the ~ log Rp column opposite the appropriate part &signation.
When subassemblies, assemblies, or units are replaced, multiply the recorded

sum of ~~s (from space ]7) by the recorded value of log Rp.
Determine the sum of the log Rp values listed and record this in the sum space
at tbe bottom of the~ log Ru column.
3.1, 1.5 Obtaining Predicted Maintainability Parameter
After all worksheets are completed, the data should be consolidated to determine:
(0 Equipment failure rate ( z ~). This is tbe total of all recorded sums
of failure mtes (the value recorded in space 17 of the worksheet
ilhmtrated by Figure 2-2).
(b) ~ARp This is the total of all recorded mme of kRp (space 18).
(c) ~~ log Rp This is the total of all recorded sums of~ log Rp (space 19).
3.1.1.6 Calculations
(a) Mean time to repair (MTTR). The predicted MTTR can now be obtained
by the calculation:
~hRp
MTTR =
XA
(b) Geometric mean time to repair (MTTRO]. The geometric mean time to
repatr, which 16 the medfan Equipment Repair Time ( ERT) when the
repair times are log nomnally distributed, is calculated as follows:
2-26
MI L- HDBK-472
24 May 1966
MTTRG = ontllog
x (A log F@
[ v\ 1
L LA J
3.2 PART B Active Ma]ntenmce Prediction Procedure
{ )
,.-. 1 corrective hllin:en~n,:e Prediction
The forms which are used to perform a corrective maintenance prediction are
shown in Figures 2-3 and 2-4 which are labelled Morksheets A and B, respectively.
The step by step procedure is as follows:
(a) Qolumn A, Part Identification: List identifying part numbers. names

and t}-pc designation of each part.
b) Norkshem A - Coiumn B, Circuit Designation- List the circuit designa-

tion of each part in the product except for the following categories for
which no circuit designation shall be Iisted:
1. Mechanical hardware
2. Electrical hardware such as:
~e~~-ilh~l~~h teT7??in21S
Terminai stucis
Wiring
Cables
Wiring harnesses
Sockets for plug-in parts of subassemblies
Terminal boiirds
Printed wiring boards (board only)
Circuit designations of subassemblies or assemblies (e, g. , modular

assemblies) that normaIly will be replaced upon failure of any part
contained therein shaf I be listed separately, followed immediately by
the circuit designations of all parts included in that subassembly or
assembly. Part listings shaIl be grouped on worksheets according
to product subdivisions (e. g. , unit, equipment. set, system, etc).
(c) Column C, Failure Rate: ( ~): The failure rate of each part shall
be determined in accordance with MI L- STD-756. The failure rate for
each part, in part failures per 106 hours, shafl be listed in Column C
2-27
MIbMDBK-4?2
24 May 1966
opposite the respective part fctentlftcatton. For parts with life

characteristics measured in units other than time, an operational
analysis shall be performed and a ~ in part failures per 106 hours
shall be computed. The sum of all part failure rates {Z ~) on each
sheet shall be determined and recorded in the space at the bottom of
Column C.
(d) Column D, Locaiizatton: E, Isolation: F, D{saseemblv: G, Interchange:

H, Reassembly; J, All gnment: and K, Checkout: List the estimated
average time required to perform each of the listed corrective maintenance
tasks corresponding to tbe respective part identlflcation, This may be done
by foIlowing a sequence of steps, similar to, but not necessarily restricted
to the following:
(1) Assuming that each component fails in Its most likely mode, note
the fault localization features and determine the necessary steps
to localize the fault to the module or function. Estimate the
average localization time.
(2) Determine the average isolation time by referring to schematics,

noting the loation of test points and esttrnate the observations
and steps required to tsolate the trouble to tbe failed component,
(3) Note mechanical aasembly details by referring to mechanical or

assembly drawings, as required, to estimate disassembly time.
(4) Observe the method used to attach a failed component to its mounting
surface and perform an analysis to estimate the interchange time.
(5) Refer to mechanical drawtngs and determine the steps required to

reassemble a con:ponent to Its required state. Estimate the
reassembly time.
(6) When alignment or adjustment is requtred, note the characteristics

of the end result and estimate the time required to perform the task.
Estimate the alignrn ent ttme.
(7) Estimate the check-out time by reviewing the minimum established

performance procedures.
2-28
.-
hfIb HDBK-472
24 May 1966
(e) column L, MC: The sum of the values recorded in Column D, E, F, G,

H, J and K 61M.11be detemnined and recorded lD Column L whtch la headed
MC, the man-hours needed to perfomn a corrective maintenance task.
(f) Column M, AMC: The value of each Mc multiplied by the respective value
iTl- a.iiure rate ( x )shEJl be recorded in Column M oppodte the re~c~ve
part identification. The mm of ail {~Mc] values on each sheet shall be
determined and recorded in the space Indicated( x A Mc ) the bottom of
Column M.
w Worksheet B: F@ure 2-4 shows a copy of Worksheet B. This is used to

record the results of the eurnmdfon of ( ~ k)ad ( x ~Mc 1gho~ at ~tbm
of Worksheet A. In essence then, the sum of all the en[ries Ln Column 2
of Worksheet B, is the sum of all the falkre ratea (x ~) and represents
tie fti)ure of the equipment or product and hence is called the Product
Failure Rate. Similarly, the sum of all the( z ~MC)V~UeS appearing
at the bottom of Column 3, Worksheet B is the total corrective maintenance
do~mtime per 106 hours. Hence by substituting in Equation 2-6 the mean
man-hours for corrective maintenance EC is obtatned. This equatjon ia
repeated for immediate reference.
X(AM,)
.- Tip
n
3.2.2 Preventive Maintenance Prediction
Worksheets C and D are used to record the data required to calculate the mean man-
hours of preventive action Hp. The forma are illustrated by Pigure8 2-5 and 2-6.
References such as Column A, Column B, etc. , refer to the respective columns
of Worksheet C. References such as Column 1, Column 2, refer to the respective
columns of Worksheet D.
(a) Column A, Description of Preventive Maintenance Ta8k8: Determine all

of the preventive maintmance tasks that must be perfomned during the
service use of the product. A detafled description of each task shall be
presented in a separate report. Determine those preventive maintenance
tasks which can be performed while the product is opcratlng. Determine
the frequency of occurrence and man-hours required to perform each task.
The frequency of occurrence, man-hour6 required, and a statement that
the tasks can or cannot be performed while the equipment is operating
shall be included in the detaihxi description of each task. A brief descrip-
tion of each task that cannot be perfomed while the product is operating
shall & listed in Column A.
2-29
MIL- HDBK- 472
24 May 1966
b) Column B, Frequency of(kcurrence( f): The frequency of occurrence of

each task, listed in Column A, shall be determined using existing data.
The frequency of occurrence for each task, in events per 106 hours of
calendar time shall be listed in Column B opposite the respective task.
The sum of all frequencies of occurrence on each sheet shall be determined
and recorded in the space at the bottom of Column B.
(c) Column C, Mp~ The man-hours required to perform each task shall be
determined using existing data. The man-hours required for each task
shall be 1i steal in Column C opposite the respective tasks.
[d) Column D, fMp: The value of each frequency of occurrence multiplied

by the respective value of man-hours required shall be recor&d in
Column D opposite the respective task. The sum of the values appearing
in Column D shall be recorded in the space provided at the bttom of
Column D.
(e) Worksheet D: The Worksheet C sheet nurntn?r shall be recorded in

Column 1. The sum of each Column B on each Worksheet C shall be
recorded in Column 2 opposite the respective Worksheet C sheet number.
The sum of each Column D on each Worksheet C shall be recorded tn
Column 3 opposite the respective Worksheet C sheet number. The sum
of all values in Column 2 shall be recorded in the space marked
Frequency of preventive matntensnce tasks at the bottom of the last
Worksheet D. The sum of aIl values in Column 3 shall be recorded tn
the space marked Total preventive maintenance time per 106 hours of
operation at the bottom of the last Worksheet D.
(f) Calculating tip: The mean preventive main%nance time, Xp in man-hours

is calculated by substituting the data recorded in Worksheet D in equation
(2-7) which is repeated for immediate reference, namely:
~ (fMp)
~p=
x,
Q) Calculating w: The mean man-hours of active maintenance which term

includes preventive and corrective maintenance is determined from
equation (2-8) namely:
2-30
ML-H DBK-472
24 May 1966
(b) Calculating M1: The MI (Mal.nta!nabfll@ Index) shall be calculated by

~bstituting in equation (2-9) which lE:
..
2-31
bULEDE#K-472
24 May 1966
. . Y . -
. .
. .
. .
. C9
A
m
.
. 1
i2
u
I
..-
2-32
MnrEDBK-472
24 May 1*6
Workabeet B
Contractor Date
Contract No. Sheet of
Workeheet
Sheet
A
Number [ Total
IIA
Cohnn n C Column
Total
iMc
M
1
Worksheet A
Sheet Number
A
Column
Total
C
~iMc
Column
Total
M
i
I 1
I I
Subtotale I I I
roduct failure rale. zA = Total of column 2 subtotals
Total repeir time per 106 hou~e,~ ~Mc= Total of column 3 subtotals
r-igure 2-4 Worksheet B
2-33
-...
MIL-HD3K-472
24 May 1966
-
-.
2-34
MIL-HDBK-472
24 ?da~ 1966
Worksheel D
Contractor Date
Contract No. Sheet of
I Worksheet C
~ Sheet Number Total ! Total I ! Sheet N~ber ~ Tot~ ! Tot~ 1
I 1 I i
1
[ I {
~ I
r
I I
I
t I
t I
I 1 ,
I I
i
I
. . 1 1
I I
Frequency of preventive maintenance tasks, f = Tot~ of col~n ~ SUb-tOt~S_

Total preventive maintenance time per million hrs, fMp = Total of column 3
sub-totals
Figure 2-6 Worksheet D
2-35
MIL-HDBK-472
24 May 1966
PROCEDURE Ill
1.0 GENERAL
This maintainability procedure describes a method of performing a maintainability

prediction of ground electronic systems and equipment by utilizing the basic
principles of random sampling.
It out] ines the methods used for selecting random samples of replaceable items
from the total complement of components comprising the system; subdividing
this sample into smaller su~samples by discrete classes of items and conducting
a mamtamabllity analysis for every replaceable item in the subsample. Typical
classes of replaceable items are listed as transistor, receiving tubes, trans-
mitting tubes, etc.
The methods of scoring each maintenance task in the repair by replacement

cycle are also detailed. These include the use of design check lists containing
a description of applicable scores and scoring criteria and the substitution of
these scores into a regression equation, the solution of which, results in an
estimate of downtime.
1, ] Philosophy, Assumptions and Summary
The underlying philosophy of this procedure is that system failures are principally
due t n the malfunction of replaceable items and therefore, the time cycle for the
var. -w. steps required to replace these items is a measure of downtime which is
a puameter of system maintainability. The duration of this downtime is assumed
to be a function of specific design parameters which relate to: the physical
configuration of the system; the facilities provided for maintenance by the design;
and the degree of maintenance skills required of personnel charged with the repair
responsibility.
Accordingly, it is assumed that a similar type of maintenance activity is required

to repair a part of the same class when repair by replacement is used because
the same steps are followed. These include the time required for preparation, fault
location, fault correction, adjustment and calibration and final test. On this basis,
it is inferred that the analysis of the times required to perform each listed step
comprise maintenance tasks which can be evaluated in terms of elapsed time.
The various classes of replaceable items are subdivided into categories such
as transistors, receiving tubes, transmitting tubes, resistors, capacitors, etc.
3-1
ML- HDBK-472
24 May 1966
The procedure also assumes that because of a basic uniformity of design, a mdom
selection of replaceable items by class will provide a representative sample of
maintenance tasks whose time of performance can be established by stmuhtion
in a manner representative of system characteristics in actual operation. The
method of selection of this sample is described in detafl in 6ections 2.0 and 3.0.
The ~signment of the times of performance for each of the steps involved
in the maintenance cycle, commonly referred to as maintenance tasks, is
determined by using three types of check lists. These are intended to provide
a uniform method of scoring the various maintenance tasks and are labelled
Check Lists A, B and C respectively. Check List A is used for scoring physical
deEign factors, Check List B scores design dic~tes-faciltttes and Check LJst C
is used to score design dictates-maintenance skiIls. The theory 1s employed that
by using these check lists which include uniform scoring and scoring criteria,
variations due tn indltidual appraisers are minimized and the resulting scores
can then b correlated with actual downtime. A re~SSiOII equation @WdOn 3-7)
is provided for this purpooe which provides a corresponding eattmate of downtime
when the numerical, A, B and C scores are substituted therein.
1.2 Applicability
This maintainability prediction procedure is used to predict the mean and maximum
corrective maintenance downtime for ground electronic systems and equipment.
The procedure is adaptable for perfonntng maintainability predictions during the

Desi@ and Development Stage. Usually prior to completion of design, a gross
estimate is performed as a first step. This is later followed by more detailed
estimates as development progresses.
1.4 Basic Parameters of Measure
The basic parameters of measure are:
nc~ = mean corrective maintenance time
ii@ = mean preventive maintenance time
El = mean downtime
3-2
MIL-HDBK-472
- 24 May 1966
max = maximum corrective maintenance time (95!I Confidence Uvel)
The mathematical equat]ons for each of the parameters of measure are shown
in 2.6.
In this application preventive and corrective maintenance are defined a follows:
1.4, 1 Preventive Maintenance
The maintenance performed in an attempt to retain an item in a specified condition.
1. 4.2 Corrective Maintenance
The maintenance performed, as a result of a failure, to restore an item to a

specified condition.
1.5 Information Required
In order to accomplish the task predictions the evaluator must have detailed
information and have accessibility to, and be familiar with at least the following:
(a) Schematic diagrams

.-
(b) Phys[cal layouts
(c) Familiar witJ the functional operation nf the equipment
(d) Description of tools and test equipment
(e) Maintenance aids to be incorporated in the prime equipment
(f) A description of the operational and maintenance environment
1.6 Data Basis
The data utilized for the development of this prediction procedure were obtained
during the surveillance of three equipments of varying complexity, use, maintenance
and packaging concepts, and the nature of the circuitry. The three equipments were:
(a) AN/ FPS2: Imng range search radar, two channels. Average com-
plexity is 10,976 parts. hlaintensnce is performed at the part level.
3-3
MIL-HDBK-472
24 May 1966
(b) ANi FST-2: T~o-chael data processor which converts analog radar
returns to digital form. Average complexi~ is 114,500 pwts.
hlalntenance is performed at the moduIe level.
(c] .%h/GK4-5: Time-di\fision data link transmitting equipment. Contains

both digital and radio frequency sections. Its average complexig is
44, 520 parts. The dimtal section uses boards and maintenance is
provided by modular replacement, that is replacing the defective
boards at the board level. The radio frequency section is comprised
of individual parts and maintenance is performed by replacing defecttve
parts.
fhe correlation between predicted and observed values can be good provided
that adequate information is available and mature experienced analysts are used.
The follo~ring is a tabulation of some typical results.
PREDICTED VERSUS ACTUAL TIMES
Equipment Type of Data Active Downtime (Min8. )
AN/ FP&6 (Radar) Final Design Prediction 67.26
Actual 94.03
AN/GRT-3/GRR-7 Final Design Prediction 51.90
Actual 63.31
Although it appears that in the case of the Radar AN/ FPS-6, the actual measure
of maintainability does not correlate with the preliminary and final design predic-
tion, it has been verified, that this has been due to a nonrepresentative sample
of maintenance tasks. This was demonstrated by selecting maintenance tasks
for simulation from actual failures in the field, the analysis of which, resulted
in good correlation, The recorded data also show a high degree of correlation
between the actual and predicted values of maintainability for the AN/GR>3/GRR - 7.
3-4
MIL-HDBK-472
24 May 1966
2.0 ANA L}TJC FOUNDATION
The fundamental approach to obtaining a maintainability prediction of a complex

system or equipment is to randomly select, for maintenance analysis and
evaluation, a repreaentati ve sample of replaceable items from the tots! popu-
lation of parts and component which comprise the system. This total sample
is termed sample size, N, beoauee it includes parta and components of all classea
of items in use. The N sample is then subdivided into a number of eubmunples
of size n, termed Task sample~. Each of these task samples represents a
specific class of parts such as; resistor, capacitor, motor, etc. The elze of
each n sample (Task Sample) is determined by considering the relative
frequency of failure for a particular ciam of replaceable ltemw This me~s
that classes of items having a higher failure rate would be represented by a
larger sub- sample than those of classes having a lower failure rate.
The details of the technique for determining the sample sizes, accuracy and
confidence levels are explained in 2.1 and Section 3.0. The use of sampling in
predicting downtime is justified in this procedure, on the basis of uniformity
of design with respect to like categories of replaceable items. Thi ISimplies
that on the average, it should take the same time to correct any resistor, or
cspaci tor failure as 1s required for any other resistor or capacitor since the
methods of mounttng, fault location, adjustment, calibration, and final test
are similar for replaceable parts of the same cAass. l%erefore, these mainbname
actions are referred to as sampies of maintenance taske tn order to provide a
status of universal applicability and the item to which these maintenance actions
iue applied Is called the maintenance Task Sample. On this basis ff sufficient
maintenance task samples are randomly selected from each class of replaoesble
items, these ~hould suffice to provide a prediction of downtime for that specific
category.
The downtime is calculated by performing a meintatnabflity analysis of the

maintenance tasks which entails a utep by step accounting of a logical diagnostic
procedure. This results in numertcal scores which are assigned by followtng
certain seortng crtterta of applicable check liuta. These numerical oeores ~
then translated tito a quantitative measure of downtime in hours by aubatitutig
tbe ~cores in Equation 3-7 which 1s a regression equation developed fmm past
studies and experience with mmparable systems.
2.1 Determination of Sample Size N
The sample size is determined by substituting in the following equation:
3-5
MIPHDBK-472
24 May 1966
-
(3-1 )
+
= confidence level
v= population standard deviation
X = population mean
k= the desired accuracy of the prediction given as a

percent of the mean
2.2 Derivation of Equation for Sample Size N
Equation (3-1 ) can be established from an analysis of the Normal Curve, P,

assuming normality of the population and the Normal Curve, S, of the distri-
bution of means of a sample of size N taken from the population. Figure 3-1
shows this relationship.
The first stipulation made in tbe derivation is that It is desired to know the mean
of the population {Curve P) within prescribed limits, i.e. , ~ t k ~ where k
! z Lb* %ccuracy desired. %: e=mplc, suppose !: were e6timiited that the
population mean is 100, and it is desired to state this value wfth 10x accuracy,
the required interval will be 90 to 110.
When samplee of a specific size are taken from a particular population, it has
been found that the mean calculated for each sample will vary. These variations
depend upon the variance of the parent population and the size of the sample taken,
Moreover, the sample means distribute themselves normally (Curve S), hawing
the same me~ as the parent population (Curve P).
The standard deviation fUY ) of the mean distribution curve (Curve S) {II related
to the parent population by the following expression:
(3-2)
.
where:
WY = standard deviation of samples of size N
3-6
MIL-HDBK-472
L
24 May 1966
r= standard deviation of population
N = sample size
mMI, it IB seen that the standard deviation of Je sample me~s( ~Y ) is directly

proportional to the standard deviat~on of the population ( a ) and inversely
proportional to the square root of the sample size N. Therefore, within practical
ltmits, it IS possible to have Oy take on any value by varying the sample size.
From Figure 3-1, the following relationship should be obvious:
k~ = # W: (3-3)
where:
confidence to be applied to the measurement

(~ = 1.645 for 95% confidence)
accuracy
*betItutlng vaiue of ~W from Equaiiun (3-2) we get;
(3-4)
with stated accuracy requirements ( k) and desired confidence ($), it is necessarY

to find the value of N l~hich satisfies the expressed equality.
%lving for N:
N=
+
[1k~
u
2
(3-5)
u the coefficient of variation, the equation may be

If we label cx=~ ,
wrttten as:
N=C k
[1k
2
(3-6)
3-7
-.
Ml bHDBK-472
24 May 1966 -
Figure 3-1. Comparison of Population and %mpling Mean Distributions
3-8
-
MIL- HDBK-472
24 May 1966
This equation has been solved for several values of k accuracy for 95% confidence.
The results are shown in Figure 3-2, Sample Size Nomography,
The use of the nomograph will be illustrated by the following example:
I&ample 3-1
SUfpXe tie expected populat~on mean and standard deviation are estimated
to be 100 and 50 respectively. 50

Cx+ = 0,5
100
It is estimated that M aoouracy of 10%is desired. Fmm the chart, the sample
size, N , is approximately 67.
It must be recognized that the value of Cx must be approximated in order to be

able to establish the magnitude of the sample size N. Field experience with ground
electronic equipment has ahown that when applied to thi B procedure a good
practical estimate for Cx is 1.07.
2.3 Taak Sample
It is tmportant to emphasize at this point that there ts a clear distinction between

Task Sample and !%imple Size, N, discussed in 2.1. Sample, N, conmins a
rntxhme of the total number of replaceable items that will be scored, which are
randomly selected from the total quantity of parts in an equtpm ent or system.
The Task Sample, however, is a Sub-Sample, or a percentage of N containing
within it a quantity of parta which are representative of a specific category, or,
class of parts.
2. 3.1 Task Sampling
in the tnterest of clarity it should be stated that Task Sampling involves the
sampltng of replaceable parts for evaluation of hypothetical part failures.
The results of the evaluation will, on the average, be repre.sentati ve of the
number of maintenance tasks which can be e xpected to occur, due to ~
faflure, under operational conditions.
Task Sampling becomes a necessi& when the complexity of. equipments now
appearing in the field is considered. For example, the AN/ FPS- 20 radar has
over 10, 000 active electronic parts and hence evaluation of each part with respect
3-9
MIL-HDBK-472
24 May 1966
I000
600
300
Id 100
60
30
10
.I ,2 .3 .4 .5 .6 .7 .8 .9 LO 1.1 1.2
COEFFICIENT OF VARIATION Cx (&#)
Figure 3-2. Sample Size Nomograph
3-10
M!LrHDBK-472
24 May 1966
to tw contribution to maintenance time would be both impractical and unnecessary

This is particularly true, when it !s reallzed that the physical arrangement and
function of many of these parts are similar with reswct ti a maintenance task.
Therefore, the evaluation of randomly selected parts, wltht.n each Taak Sample
will re~ult @ azcurate predictions of maintainability.
2,3,2 Task Smmple Sdectfon Technique
{a)
General .COnslderatlons: Through the conrdderation of vnriou8 f8ctors
associated with the failure and replacement of partB, the maintenance
capab!l!ty of the equtpment can be predicted. The maintenance time
thus derived is an estimate of the aver~e time to accompl!rh a
maintenance task under actual oixwsting conditions.
(b) ..
Pmcees: The procem of tagk ~election is illustrated by Example
3-3 which le ehown in 3.2. Reference to Table 3-1 shows that the
parts are llBted by part class, quantity and average part failure
rati per 106 hours. The expected number of failures per 106 hours
per clam is computed as the product of the quantity of parts and the
average clam fatlure rate. The ratio of the expecbd number of
failures per 10s hours for a clam to the total of the expected number of
failureB for all claases is the percent contribution to fatlure of that
class. The product of the percent contri-button of each ciasu and the
sample size N is the claae sub-uample, n, which giveB the number of
replaceable items to be anaiyzed for that claB9. The last column of
Table 3-1 Bh0W8the sub-sample sizes , n, rounded off to the nearest
whole number.
2.4 Application of Check LteJte
Check litiB are used to evaluate each maintenance step which 1s stmulated for
each of the ~oplicable number of maintenance taBkB randomly selected per class
in the quantity as specified by tie computational method described in 2.3 and 2.3.1.
Each of the check lists are scored frwm O-4 and the resulting ecorea are applied
to a regression eqlation (Equation 3-7) which is used to calculate doumttme.
The re are three general categories of design check IIats which are used for
assessing desi~, support and personnel vartaL1es,
a) Design Check Liste Af38efming the D Variable: The basic types of

check liBtB which are used to aeeem deBign (D) variables are !abclled
as A, B, and C. The prticipal variables which are included in these
check hats zme outlined in the following paragraphs. For greater detail
It is suggested that reference be made to each indiviciu,al check i!%
(see Addendum A).
3-11
MIL-HDBK-472
.-
The A Check List is used b score specific maintenance tasks

that are a function of physical design variables such as
packaging, access characteristics, test points, displays, etc.
The E! Check list is used to score Design Mctate6- Faci]itie8.

These Include such variab)es as: external test equipment,
ccmnectvrs, assistmee (operations personnel), etc.
(3) The C Check List evaluates the personnel requirements relating

to physical, mental and attitude characteristics. Typical of
the criteria are: visual acuity, iogical analysis, memory,
endurance and energy, etc.
2.5 Linear Regression Equation for Downtime
The linear regression equation wh!ch is used to calculate the predicted downtime
1s:
Met s ontilog (3. 54651 - 0. 02512A - 0. 03055B - 0. 01093C) (3-7)
The coefficients of this equation were derived from 101 corrective maintenance
ta8kB appearing on appropriate check Iieti for tbe AN/ FP%20 long range
sealch r M&, AN/ FST- 2 ho channel data processor and AN/GKS S data link
tranmnftting equipment. The symbol Met means corrective maintenance ttme
and tbe letters A, B and C, respectively, represent the scores obtatned by the
use of the respective applicable check lists.
2.6 MathematlcaJ Relationships Gf Maintenance Times
The applicable mathematical expressions for calculating the various matntatn-

ability indices are shown below:
(3-8)
N
where: &t = mean corrective maintenance ttme
N= sample eize of corrective maintenance taaka

Met = corrective maintenance time of individual maintenance tasks
3-12
MIL-HDBK-472
24 May 1966
similarly:
(3- 9)
where:
~pt . mean preventive maintenance time
= preventive maintenance time of individual m ain~nance

MPt
tasks
~mox is expressed as:
Mmo~ = ontilog
[
~ct +1.645C IOQMet
1 (3-1 o)
where: Nt
~ log Mcti
met = 1:1 : moon of log Mc;
NC
NC-l
(3-11)
where:
Kt = mean dowmtime
Fp, Fc = number of preventive and corrective maintenance

tasks per thousand hours
3-13
MXLHDBK-472
24 May 1966
3.0 APPLICATION
3.1 General Approach
The maintainability prediction technique for evaluation of electronic equipment

is accomplished in four steps as foIlows:
(a) Determination of sample size, N
(b) Determination of @k sub-samples, n
(c) Prediction
(d) Calculation of maintenance indices
Briefly the statistical selection of a sample of failed parts/components and the

quantitative evaluation of the contribution that each assumed failure makes to
tbe total equipment maintenance time, permits the calculation of the overall
equipment mean downtime. The justification for, and the steps involved in,
determining the total downtime, are given in detail in the text that follows.
The use of this method permits the calculation of an accurate system maintain-
ability figure, without requiring the time consuming empirical evaluation of
the maintainabi My of each part/component in a complex electronic system.
Detailed explanations for each of those steps, plus illustrative examples of
the process, are contained in the following paragraphs.
3. 1.1 Determination of Sample Size
The sample size to be used in the prediction is dependent upon the statistical
accuracy desired. With stated accuracy requirements (k) and desired confidence
level,
#
(~) , the s~ple size (N) which satisfies these requirements is computed
as shown in Equation (3-l) which is repeated here for convenience: 2
[1
w
N:
kx
3. 1.2 Step-by-Step Procedure for Calculating Sample Size
(a) Decide on the accuracy (k) or precision of prediction desired; that is,
decide how large a confidence interval is adequate for prediction
purposes and express that interval as a percent of the expected MTTR
(the latter is to be expressed in minutes).
3-14
-
MIL-HDBK-472
24 May 1966
(W @tide on be ~=el Of confidence (~) IO be -ociati wjth the

interval determined in (a) above.
(c) Obtain an estimate of the population value of the coefficient of

variation (Cx) . The expression for this statistic is:
+ (3-12)
where:
u = the population standard deviation of MTTR
Y = the population axithmeti c mean
(Note: The estimate of Cx maybe *seal on eqwience ~ti simi~~ SYSteID8

on which adequate measures are available. If satisfactory data are not
available, use Cx = 1. 07).
(d) Substitute these values in the following equation and solve for Sample
Size, N: (Thi B is a variation of Equation (3-1) in which Cx is sub-
- stituted for u
T
(3-13)
where:
c~ = the coefficient of variation
+
= the normal deviate corresponding to the desired confidence
level; and
k = the desired accuracy of the prediction, given as a percent

of the mean.
3-15
MfbHDBK-472
24 May 1966
Example (3-2~
Assume that it is desired to state the MT TR with an accuracy of t25 percent

wjth a confidence of 95 percent. What sample size 1s required? Substituting
in Equation (3-13) we obtain
Figure 3-2 is a nomograph which relates the sample size to Cx for a number
of common accuracy values. To solve umder the same assumptlorm ae in
the precedtng example find the value of Cx (= 1. 07) on the abscissa, follow
up to the curve Iabelled 25%, and across to the ordinate to read N (= 50).
3.2 Step-by-Step Procedure for Calculating Task Sample
(a) Determine the failure rate contribution of each part type in the
system or equipment to the ove~ll failure rate, That is, determine
what percent of the failurws will be attributable to tubes, to resistors,
to transistors, to capacttore, etc.
(b) Ustng the percentages computed shove, calculate how many of the N
items (determined from (a) a&ove) will be apportioned to tubes, to
traneistora, to re8i8tort3, etc. , 1.e. , calculate the class sub-sample
size, n.
{c) Select randomly from each part type enough item8 to meet the require-
ments set up in (b).
Example (3-3~
Assume an equipment with parts as shown !n the first column of Table 3-1 and
that the number of each type of part is aa Ehown in Column 2. Given the part
hilure rates as in Column 3, the numixsr of failures expectad per 10* hours
appear -n Column 4. Expressing entries in Column 4 as percents of tbe
the
sum of @lumn 4 then yields Column 5. This is the permit contribution to
the total expected failuree of each part lt steal tn Glurnn 1. Multiplying tbcwe
percents by the sample size of 50 ytelde Column 6 which shows the num~r of
simulated maintenance tasks per part class or category. The valuew in
Column 6 are then rounded off to yield the values tn Column 7.
3-16
MIL HDBK-472
24 May 1966
>
5
w
N
t-
3-17
MIL-HDBK-472
24 h!ay 1966
-
3. 3 Step-by- Slep Procedure for Pcrforrr. ing Task Prediction
TCIaccomplish the task predictions. the cia,luator should have available

detailed inform atlor. inc!uding schcmat}c diagrams and physical layouts.
He must atso be thoroughly familiar vith the functional operation of the
equipment. (Xher ]nfn~iit]on needed IS a description of the tools and test
equipment to be protlded and the maintenance aids to be incorporated in the
prime equipment. .+ description of the operation and maintenance environment
is ~so e~remely ~;~uab]e. Figure 3-3. hlalntwnability Prediction Form.
has been developed tn facilitate the mahtenance analysis and task scoring.
On this form specific information required to complete the task or part
failure prcciicxlon is iisted along uit.h iacntification data regarding equipment.
evaluator, ew.
hlaintcnance Analysis: Pr]or to task scoring, it is necessar~ that,

for each task a maintenance analysis he performed. This analysis
entails a step-by-step accounting of the rnrmt likely chagnostic
procedure. Beginning uit.h the symptoms of malfunction, each step
requi red in locating the defective part is recorded. Complementary
to each step, notations regarding access problems, test equipment
requirements, and related information which is important to determining
the task scores, are made. Figure 3-3 the Maintenance Analysis
for~t &-Id T~~uiE 3- -! tk ccmth-luatim Shcat, illustrate a f017J-t&t uSd
for this analysis. The form is divided inti ti$o columns, The left
column !abelled Maintenance Steps, is used to record each test
or step that a technician should make, Scoring comments associated
with each step are entered in the column on the right. Completion of
the maintenance analysis provides n firm basis for the scoring. The
full scope of a m,alntmance situation is realized through this process.
S~stem hlaintenance Di~ram: This analysis may be facilitated by the

preparation of a System hl~nt.enance Diagram, which would clearly
detail the system functional block diagram, with main signal paths,
major test points and other diagnostic aids shown. An illustration
of such a diagram is sho~vn in Figure 3-3, Maintenance Dia.gran~
AN/ GRT-3. This diagram assists in the determination of the
malfunction symptoms and in selecting steps to isolate the malfunction
to a functional area. It is necessary to have a schematic diagram for
each block to troubleshoot within a block and to determine the effect
nf an assumed failure on the output (s) of other blocks. The illustrated
diagram is representative of the minimum requirements for such a
diagram and ma! be expanded to [ arying dcg-rees depending cm the
complexi~ of the equipment and the information available.
MI L- HDBK-4?2
24 May 1966
.-
.
Equip. Task No.
Assy D} Date
Primary function failed unit,ipart
Mode of failure
Ma.lfurfc:[on symptoms
Nlaintenance .4nalvsis
!iIai ntenance Steps Scoring Comments
Checklist Scores
Predicted downtime _ Min.
Figure 3-3 Maintainability Prediction Form
:{- ~9
M!L- KDBK-472
24 May 1966
\I.QSTEN.iYCE ASALYSIS COXTISLAnOS SHEET
Equip. Par! Task So.
\lalntcrlancc Steps scoring Comments
Figure 3-4 Ma ntenance Analysis Continuation Sheet
:]-2(,
MX~KDBK-472
24 May 1$66
g ~g____________________
1~
F----- 1
I I
1 I
f +
--
I TW
I
L---- ~
g 1>G------w>--
-v
vi
J-; -----------
i I@
J-
@@e@
____
-------=-- SE w 5
J---m l-------- +---- ---
I I I
I
I
I
I
I
I
I
I
1
, 1 I
I t II I
3-21
MIb HDBK-472
24 May 1966
ExarnPle (3-4)
The use of t-he maintenance analysis procedures described in paragraph (a)

Irill hc illustrated b~ evaluating the time requiremcmts for a specific maintenance
task, Resistor R-7801, appearing in Amplifier Mixer Ahl-1345/FP%20 of the Radar
AN FPS- 20, hzs been selected for this purpose. The following discussion
illustrates the procedure to be followed (see Table 3-2).
Evaluation of Resistor Failure
The resistor failure was assumed to have nccurred by opening. Following

&his assumption, a step-by-step maintenance analysis wm made, drawing
from both general maintenance experience and techntca.1 data. This procedure
is illustra&d in Table 3-2 Maintenance Analysis, R- 7801 /FPS-20. Here
the demiled steps necessary to isolate the defective resistor are listed.
For each step, uomnients rcgdrding availability of test indicators, need for
test equipment, access problems, and related inform ation needed to effectively
score this task were listed. Figure 3-6, AN/FPS-20, Transmitting System,
illustrates functionally the major circuits associated with the transmitter section.
Within the diagram each step has been numerically identified. It will be noted
that the troubleshooting pa~h chosen is one of several possible routes. The
route estahl i shed IV3S based on the importance of the check and the e-e with
witlch ;t crmid be taken: For example, in Step 2, the trigger input which is
vital for the proper operation of three portions of the transmitter, was tested.
The third step was selected because of the ease with which the information
could be secured (built-in metering) . Such choices are generally consistent
with procedures employed by electronic technicians.
Fig_urc :-7 Arnplificr hlixcr AM-1347, FPS-20, illustrates the mechanical

layout of tic section of the equipment wlthln uhich the defective part was
located, ,Is indicated in the maintainability analysis, the plate power supply
sub-chassis had to be removed for further testing. Figure 3-8, Expanded
liclv, Plate Pn\vcr Supply, shows the underside of the power supply. Here,
thr terminal board cm uh]ch resistor R-7801 is located has been identified.
Functionally. the use of reslstnr R-7801 is illustrated in Figure 3-9, Plate
.Sllpply Block I)lagrm The rcsi stor provides continuity between the rectifier
and series dropping electron tubes. Its opening caused loss of plate supply
volta~e 10 the buffer arnplifler, thus preventing operation of the transmitter.
These illustrations, in assoclaticm with other technical material secured from the
applicable technical orclcr, provided the basis for the scoring comments in the
Nla[nlenancc .4nalvsis Cont]nuation Sheet (Table 3-2).
MIL-HDBK-472
T.4SKPRED1~TIOY, R-7901 24 May 1966
31~T~m]LITY PREDICTION FC)R?,l
Equip. .4,,, ~FPS- 20 lnlt ~Pari R-7901 Task \o.
A9sy. Ilv Date
Primary function failed unit/part Series remstor in 620 V. D. C. power supply

output network.
Llm$? of failure Resistor opened
MaJfunct\on symptoms l~o return on radar indicator
Maintenance .4nalys1s
I Maintenance Sleps Scoring Comments
Checklist Scores
Predicted downtime . . . . . . . . . . . . . . . ...1.... . . LO \!in.
TABLE 3-2
h!IL- HDBK-47Q
24 May 1966 TABLE 3-3
M.41STEN.4!!CE ANALYSIS. R-7901, JFPS-20
>faintpnanc~~ An~ysis Continuatlon Sheet
Equipment AX, FPS-20 J>art R-7801j FPS-20 Task No. 1
Slai nlenmce Steps Scoring Comments
1. Equipment malfunction is initially Rudio Freouenc} Ilonitor (I D-446)

indicated t.)}no target returns on normaJl!< connwtcd to system is used
Indicator. Maintenance action iJro- to monitor power output at bidirectional
ceeds to isolate trouble to major coupler (C U-516J. Test equipment
equipment functions. Pover out- S.erves variety of tests for equipment
put check at bidirectional coupler adjustment and repair. Preliminary
(CU-516) isolates trouble to trans- calibratlon or test set-up may be
mittcr function. required. Proper values listed in T.O.
2. Trigger pulse is checked at 1 PA oscilloscope is used to check trigger

Mm.lulalor (jl]D-276) LOIsolate trouble pulse at J-1405 on front panel of
to transmitter or modulator unit of modulalor MD-276 (I PA modulator).
the transmitter section. Presence { Oscilloscope set-up and adjustments
of trigger indicates trouble in R. F. requi red, Proper reading listed in
generating slages (Wdo, Buffer T C).
Amplifier, Mixer, or Power
&nplifler).
3. hleter reading on Am\jl]fier- Jllxer, (-athmic current meter M-77(J2

intermediate Power Amplifier and provides front pane) indication of
Power Amplifier are observed and trouble In AN1-1347. Proper value
checked against required values. listed in T.O.
No cathode current on meter ?$1-7702
of Amplifier- M~xer AM-1347 indi-
cates trouble is in second amplifier
stage or power supply (PP- 1347).
I
3-24
JLI]L- HD13K-472
24 hlay 1966
TABLE 3-3 (Continued)
I \Iaintenance Analysis Continuation Shce~

1
Equl]wnent AN /FPS-20 Part R-7901 /FPS-2O
llai ntenance Steps Scor~ng Comments I
I 4. Power Suppl! PP- 1377 monitors Front panel mounted mctrr \i-7@l
~ arc obsemed (fuse lights and and fuse lights pro~-iclc ra])l(l check
meters) to isolate trouble to the of .4. C. input voltages to r~ctlfiers
mplifier or power supply. No and 11. C. voltage 10 amp]lfler
6201. I). C, noted on meter tuh?s. Pro]wr meter reading listed
(hl-7EiOlj. inT, O.
5. Power Supply PP- 1377 is removed Exlcrmd access requires removal

from front of cabinet and the 620 of Jmwcr supply from c abinct.
ll. D. C. circuit checked. Tube Chassis must be rcmovud partially
check made prior to chassis re- to allow disconnecting of cables lo-
moval from cabinet. Trouble is cated in rear. Power Sup!]]y con-
:-..~bq.q 4A
LZIUICAALU
-n. ..
&u +b,.
vneie+n
-..-.
- tain% hea~y transformers and f]lters
requiring strength and endurance.
Two men required to remove and
place on workbench, Multimeter
and tube checker required to isolate
trouble resistor in G20 V. D. C. sec-
Lion. Spring lock fasienvrs permit
rapid removal of chassis. Resistor
board rnounteci on the underside of
chassis by screws. Resistor is
soldered to terminals. Some delay
to be expected in repair action due
to part location and necessity to use
c=e in part removal to avoid
heat damage or solder ccrntaminatim
to adjacent parts.
-. - -
f- 7
1
I
I
I
I
II I 1
I
If I
1: I
1; I
1
1
I I
I I
I I
I I
I I
I I
I
I
I I
I
I
I
!
I
1
I
I g la
I
I Im
I
I I c
!
I I
le
I
! II- I 1 I !L
I I 0
i
I
I
3-26
MI~ HDBK-472
24 May 1966
u% M7702 M7T03
I IST BuFFER 1!
o FILAMENT AND BIAS PCWER SUPPLY

u
q
J772
Bzmimlm
F7701 F7702 F7703
m J7724
FRON7 VIEW
Figure 3-7. ~plifier-lvlixer, AM- 1347/ FPS-20
3-27
MI~HDBK-472
24 May 1966
R-7801
BOARD IIS
./
Figure :1-x. Ewanderi lIeN, Plate Power Supply
3-28
MIL-HD33K-472
24 May 1966
R-?ou
120 VACJPLATE ~ v78a ~
VOLTA6E RECTIFIER
TRAN9F TUBE
Ac- 7?001 OROPPWG
m- t $ $ 3
TUBE
TUBE
%L\
V7804
VOLTAOE
I ~ ~
V7805
VOLTAOf
REFERENCE
TUBE
5
1
V7@06 %
vOLTAGE
REFERENCE
TUOE
t {
e i
PLATE ?+
vOLTAGE RECTIFIER
TRANSF TUBE DROPPINfl
k T7803 +360 Voc
t b L i J
UIFF
AMP 7{
780
nl[
Wel I w
VOLTA@E vOLTAOI
REFERENCE REFERENC
~
JOE AA
TUBE
r
. 1 -#
Figure 3-9. Plate Supply, Block Diagram
3-29
.
M13.rHD13K-472
24 May 1966
3. -! Calculation of hlalntenance lndlces
(a) Task. Scoring: The cleslgn predlct]on 1s accomplished b) completing

the three design-related check llsts for smnple tasks. Specifically,
~ese rheck lists are: .%, Scoring Ph~sical Design Factors; B,
Scoring I.)eslgr D~c!3tes-. ~~~illtj[>g; and ~, ~oring ~e~i~ f)ict3tes -
hlalntenance Skil!s. These check Ilsts are presented in Addendum A
together wqth ail ]nstructlons necessary for scoring each item.
The scoring for each Item ranges from O In 4. lnterm~dinte v:dues

of 1, 2 and 3 are provided for some question~ hhere lh~~nature of
the char;icterlstlc being assessed ma} Lake on la:i 111x!Jlagnitudes.
Th]s 1s contrasted to the vcs-no situation. The questions have been
framed in a manner that permits Reneral application across equipment
Iincs.
Situations may arise \\hcrc insufficient information is Oiailablc to score

a particular check list question: or, a question is not applicable to a
particular task. lf Insufficient data is available, the n~era~e question
score for the check Ilst should be inserted. For example, if 1-1 questions
were scored and the total score was 42, the average qucsticm score is
th rec. tnsertin~ this \alue for the unevaluated item. the final score is 43.
ror oiiler siha~i(tns, i. c. . nol qq)iicabie. a sc+orc o; ivur- shouid W UWA.
The rvusoning hvre is that if a particular item d~es n~~ aPPl}, it is not
detrimental to maintenance.
To illustrate the scoring process, tic scores obtained for the sample
maintenance analysis tasks are shov:n in Table 3-:1, Task prediction,
R- TR()], The score for each check list question is nbtainecf by referring
to the scoring comments in the maintenance anal!sis and the technical
data available for the equipment. The task was reviewed for items that
pertain to each question and the questions were then scored in accordance
with the criteria, presented in Addendum A. in cases where the quality
of a feature is scored, the uorst condition encountered is used.
To illustrate further hour check list scores are obtained. some of the
specific scores in Table 3-3 tii!l be examined. in Check List A, question 2
receiied a score of 2 (external latches and fasteners meet two of the
criteria that the~ are capti!e. need no special tools, and require only
a fraction of turn for r~!ease) In Check L!st B question 1 received
a score OJ 1 (: or :; p~eces of test equipment are needed). Examination
,;- :](J
MIL-HDBK-472
24 May 1966
of Table 3-2 indicates that an oscilloscope, multimeter, and tube checker

were used to accompl~sh this task. For Check List C, quetiion 5
received a score of 1. This score was assigned because the initial
symptoms gave very Iittle indication as to the cause of malfunction
and because a number of the major units had to be checked to isolate
the trouble to a functional area,
@) ~wntime CalcUlati~n . The last step in the prediction process is to

calculate the predicted downtime ~or each task. This is accomplished
by inserting the total check list scores for each maintenance task
performed in Equation (3- 7). The equation is repeated here for
Immediate reference:
UC, m antilog (3. 54651 - 0. 02512A - 0.03055B - O. O1O93C)
To facilitate this calculation a nomograph was developed for the pre-

diction equation and is shown in Figure 3-10, Nomograph - Downtime.
The use of this nomograph permits the determination of downtime
directiy in real time (instead of log values). All instructions for use of
the nomo~aph to calculate maintenance tndices are contained in
Figure 3-10; and are the mean active corrective maintenance downtime fie t,
and the maximum active corrective maintenance downtime Mmox . The
meti corrective maintenance tire+, wet, is Ubt&iirIadby substituting Lxi
Equation (3-8). The ma.xtmurn corrective maintenance time, MmoX,
la found by substituting in Equation (3-10).
3-31
I
d\
\ /
\
\ /
\l
\!
3-32
hlIb HD131i-472
24 hlay 1966
DESIGN CHECK LJSTS AND SCC)RING CRITERIA:
The three Dcsigx Check Lists A, B, u(I C anG f.heir respective scoring crlterla
follow:
CHECK LIST A - SCORING PHYSICAL DESIGN FACTORS
(1) Access (ExternalJ: Detemnines if the exiernal access is adequate for

visua.! :rtspecbon and manipulative actions. Scoring will apply to external
packaging as related to maintainability design concepts for ease of
maintenance. This item is concerned with the design for external
\isua.1 and man]puiat]ve actions which would precede interns! maintenance
actions, The following scores and scoring criteria will apply:
Scores
(a) Access adequate both for visual and manipulative tasks

(electrical and mechanical) . . . . . , . . . . . . . . . 4
(b) .4ceess adequate for visua!, but no manipulative,

tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
(c) Access adequate for manipulative, but not visual,

tasks, . . . . . . . . . . . . . . . . . . . . . . . . . . 2
fd) .4CCCSS not adequate for visual or manipulative

tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Scoring Criteria
An explanation of the factors pertaining to the above scores is consecutively

shown . This procedure is followed throughout for other scores and scoring
criteria.
(a) To be scored when the external access, while visual and

manipulative acticms are being performed on the exterior
of the subassembly, does not present difficulties because of
obstructions (cables, panels. supports, etc. ).
(b) To be scored tvhcn the c\lernal access is adequate (no delay)

{or visual inspection. h~lt nnt for n3n~p~ll~ti~:e actions.
A:)- 3:{
MI~HDBK-472
24 May 1966
External screws, covers, panels, etc. , can be located visually:

however, external packaging or obstructions hinders mani-
pulative actions (removal, tightening, replacement, etc. ).
(c) To be scored when the extenml access is adequate (no delay)

for manipulative actions, but not for visual inspections.
This applies to the removal of external covers, panels,
screws, cables, etc. , which present no difficulties; however,
their location does not easily permit visual inspection.
(d) To be scored when the external access is inadequate for both

visual and manipulative tasks, External covers, panels, screws,
cables, etc. , cannot be easily removed nor visually inspected
because of external packaging or location.
@) latches and Fasteners (External): Determines if tie ecrews. C1lPS,

latches, or fasteners outside the assembly require special tools,
or if significant time was consumed in the removal of such items.
Scoring will relate external equipment packaging and hardware to
maintainability design concepts. Time consumed with preliminary
external disassembly will be proportional to the type of hardware
and tools needed to release them and wiH be evaluated accordingly.
Scores
(a) External latches and/or fasteners are captive, need no special

tools, and require only a fraction of a turn for release . . . . . . 4
(b) External latches and/or fasteners meet two of the

above three criteria, . . . . . . . . . . . . . . . . . . . . ..2
(c) External latches and/or fasteners meet one or none

of theabove three crtteria . . . . . . . . . . . . . . . . . ...0
Scoring Crlterta
(a) To be scored when external screws, latches, and

fastmers are:
(1) Captive
(2) Do not require special tools
A3-34
MIL- HDBK-472
24 hfay 1966
f:\I Can be released with a fraction of a turn
Releasing a DztSI fastener }vhich requires a 90 de~ee turn using a standard

screw dri~.er is an example of all three conditions.
fit To be scored when external screws, latches, and fasteners

meet two of the three conditions stated in (a) above. .4n
action requiring an Allen urench and several full turns for
release shall be considered as meeting onlv one of the above
requirements.
(c) To he scored when external screws, latches, and fasleners

meet only one or none of the three conditions stated in (a) above.
(2) TUtches :inc! Fasteners (lnterna~: Determines if the internal screws.

clips, fasteners or latches within the unit require special tools,
or if significant time was consumed in the removal of such items.
Scoring will relate internal equipment hardware to maintainabilit-v
ciesign concepts. The types of latches and fasteners in the equ ipmcnt
and standardization of these throughout the equipment shall tend to
affect the task by reducin~ or increasing required time to remove and
replace them. Consider internal latches and fasteners to be !vithin
the inkrior of the asscmbl~.
Scores
(a) Internal latches and.for fasteners are captive, need no special

tools, and require rmly a fraction of a turn for release . . . . 4
(b) lntcrnal latches anrt; or fasteners meet two of the above

three criteria . . . . . . . . . . . . . . . . . . . . . . . . ..2
(c) Intcrnd latches and. or fasteners meet one or none

of theatmve three critcria . . . . . . . . . . . . . . . . . ..~
Scoring Criteria
(a) To be scored when internal screws, latches and fasteners are:
(1) CYtptive
(1) M not require special tools
A3-35
MIL-HDBK-472
24 May 1966
(3) Can be released with a fraction of a turn
Releasing a D2US fastener which requires a 9(1-degree turn using a standard

screw clrlver u)ould be an example of all three conditions.
(b) To be scored when mtcrna] screus, latches, and fasteners

meet NO of the three conditions stated in (a) above. A screw
which is captive can be removed witi a standard or Phi Iiips
screw driver, but requires se~eral full turns for release.
(c) TO km scored urhcn lntcrnal screw, latches. and fasteners

meet onc of three conditions stated in (a) shove. An :Iction
requiring an Allen wrench and several full turns for release
shall he considered as meeting oniy one of the above require-
ments.
(4) Access (lnternaf): I)ctcIm~ ines if the internal access is adequate for
visual inspection and m3nipulativc actions. This item applies to
internal packaging concepts in relation to design ior ease of main -
tellance. Internal is to mean ail wori{ accomplished after gaining
access to some portion of the equipment.
Scores
(a) .4cress adequate both for visual and manipulative tasks

(electrical and mechanical) . . . . . . . . . . . . . 4
(b) Access adequate for visual. hut not manipulative.

tasks . . . . . . . . . . . . . . . . . . . . . . . ...?
(c) Access adequate for manipulative, but not visual,

tasks . . . . . . . . . . . . . . . . . . . . . . . ...2
(d) ACCCSSnot adequate for visual or manipulative

tasks . . . . . . . . . . . . . . . . . . . . . . . . ..()
Scoring Criteria
(a) To be scored when the internal access. while performin~

n~anipulative or visual actions in a sub-as sembl~ or unit,
does not present difficulties because of thr internal construc-
tion or part location.
(b) To be scored when the internaI access is adequate
(no delay) for visual inspection, hut not for manipulative
uctions. Components and parts carI readily be located
~isuall~ during the mmntena.nce task; however, internal
construction or part Iocption hampers manipulative actions
(testing, removal, etc. ).
(c) To be scored when the internal access is adequate for

manipulative actions, but not for \tisuaJ inspections.
Components or parts can be easi IY tested or removed;
however, their physical location does not easill permit
visual inspection.
(d) To be scored when internal access is adequate for both

visual and manipulative tasks, Components or parts cwmo!
be easily tested or identified because of internal construction
or location during the maintenance action.
(5) Packatin& Determines the access (within the subassembly) to components

or parts requiring mechanical disassembly. This question concerns
itself with the internal packaging of parts relative to the maintenance
action. Current design concepts have been concentrated on module
type packaging; however, : .. 11..
even Lhese va4j hi k mwh.ritc~llj
L 14 .
,Ielu ..- A.1
~hlvuL&b,
while others are plug-in type only. This item deals with the mechanical
problems involved in gaining access to failed components or parts.
Scores
(a) Internal access to components and parts can be

made with no mechanical disassembly. . . . . . . 4
(b) Little disassembly required (less than 3 min. ) . . 2
(c) Considerable disassembly is required (more

than3 min. ). . . . . . . . . . . . . . . . . . ...0
Scoring Criteria
(a) To be scored when less than one minute is required

to gain access to the failed components cr part.
A3-37
MIGfiDBK-472
24 May 1966
(b) To be scored when less than three minutes is expended in

gaining access to the failed component or part.
(c) To be scored when more than three minutes is expended

in gaining access to the failed component or part.
(6) Units - Parts (Faiied): Determines the manner m which umts

or parts are removed or replaced during the maintenance action.
Since units and parts are electrically and/or mechanically secured
in equipments in many different ways, the time to remove such items
varies considerably, Mechanically held items include tubes prntected
from vibrations b~ speclai shields or clamps, printed boards clipped
into their sockets, and parts and components held by brackets.
soldered items include resistors, capacitors, etc.
(a) Units or parts of plug-in nature . . . . . . . . . . . . . . . 4
(b) Units or parts of plug-in nature and mechanically held. . . 2
(c) Units ofsolder-innaturc . . . . . . . . . . . . . . . . . . 2
Scoring Criteria
(a) To be scored when units or parts are plug-in types requiring

only to be pulled out. Plug-in t+ype parts such as tubes. some
rela~s, crystals, etc. , would be included in this category.
(b) To be scored when units or parts are plug-in types, but are
mechanically held by clips, shields, clamps, etc. Also applies
to maintenance requiring the removal of a tube having external
grid or plate connections, anti-vibration shields, etc.
(c) ro be scored when units or parts are soldered-in types such as

resistors, capacitors, etc. , when the removal of parts requires
the unsoldering of part terminations.
(d) To be scored when units or parts are soidered-tn mechanically

held t3-pes such as transformers, Jacks, etc. The remo~al or
replacement of parts requires mechanical disassembly and
unsoldering.
A3-3~
MIL- HDBK-472
24 May 1966
(7) Visual Displays: Determines if sufficient vlsua! information pertaining

to the equipment malfunction is displayed within one area or unit.
Cirruit indicators and meters provide, to some extent, symptom
analysis. Therefore, it is important that these indications be disp)ayed
within one area to ensure rapid analysis and action. If several areas
must be consulted before a qualifled estimation of the difficulty can
be made, much time is required
Scores
Sufficient visual information on the equipment ik given within

one display area . . . . . . . . . . . . . . . . . . . . . . . . 4
Two dIsp)3y areas must be wnsulted to obtain sufficient

visua] information . . . . . . . . . . . . . . . . . . . . . . 2
More than two areas must be consulted to obtain sufficient

VI Su:l in formation . . . . . . . . . . . . . . . . . . . . . . 0
Scoring Criteria
-. TCJ h scored whcri visual infol.matitm associated with the fauit

or malfunction is displayed within one area. Applicable if
) diagnosis and repair can be accomplished successfully

s-yrnptoms derived from one display area or subassembly
followlng
of the
system.
TO be scoreti when two display areas must be consulted to provide

visun] in(ormuticm assoriatcd with the fault or malfunction. Two
sepal ~te display areas cm t!:: system (meter panel and fault
indicators) mv%t be consulted t~ diagnose malfunctions successfully.
To be s~ored when more than two areas or subassemblies must be

conslilted to provide visual information associated with the fault or
malfunction. This would be indicative of a least maintainable
condition.
(8) Fault and Operation Indicators (Built-b Test Equipment~: Determines

if an equipment malfunction or fault is clearly discemtble via audible
alarms, indicators, etc. , and that such information is clear!y presented
for rapid maintenance action, The use of indicators is increased as
A3-39
M?LHDBK-472
24 hflly 1966
complexity increases, and equipment availability becomes more

important. Although visual and audio alamns usually indicate that
a problem exists, they do not always determine the exact location
of the malfunction. The more precise the indication, the better
the maintenance condition.
Score6
(a) Fault or malfunction information is provided

clearly and forrapld action . . . . . . . . . . . . . . 4
@) Fauit or malfunction information clearly presented,

but requires operator interpretation . . . . . . . . . 2
(c) Fault or malfunction information requires no

operator interpretation, but is not clearIy presented. 2
(d) Fault or malfunction information not clearly pre-

sented and requires operator Intevretation . . . . 0
Scoring Criteria
(a) To be scored when an equipment fault or malfunction

occurs and is evidenced by alarms, indicators, etc. ,
which provide for rapid diagnosis and maintenance action.
An example of this would be when a power supply failure
occurs because of an open fuse which is pointed out by an
indicator or alarm.
b) To be scored when an equipment fault or malfunction occurs

and is evidenced by alarms, indicators, etc. , but r%quires
further tests for isolation of the fault. lass of output power
is evidenced by an alarm: however, further diagnosis must be
made to &term ine the exact cause of trouble.
(c) To be scored when an equipment fault or ma.l!knction occurs

and is not clearly determined by alarms, indicators, etc. ,
however, provisions for rapid diagnosis and maintenance a&ion
are available. Applies when some preliminary testing might
be required to determine if a fault or malfunction such as the
lose of some voltage, (B +, Bias, etc. ) exists. Once determined,
however, maintenance is expedited, such as in the case of an open
fuse.
A3-40
hfIL-KDBX-472
24 May 1966
(d) To be scored when an equipment fault or malfunction occurs and

is not clearly discernible, and which requires s~ptorn interpre-
tati on. Testing is also necessary to determine the equipment
status and cause of failures.
(9) Test Pointi (Availabili~~: Determine if test points are available for
needed tests pertaining to maintenance action. A test point shall be
considered as aoy test probe receptacle where specific system operation
data can be obtained. This definition eliminates as test points connector
pins, on printed circuit boards, terminals, tube pins, etc. f%e number of
test points available and the amount of hfomnatlon ytelcied wlH affect
de time to establish the cause and location of fault.
scores
(a) Task did not require use of test points . . , . . . . . . . . . , 4
(b) Test points available for all needed tests . . . , . . . . , . . 3
(c) Test points available for most needed tests . . , . . . . . . . 2
(d) Test points not a\*ailable for most needed tests . . . . . , . . 0

.
Scoring Criteria
(a) To be scored when the maintmance action did net require the use
of test points, but when, instead, the malfunction can be diagnosed
and repaired via built-in test equipment.
(b) To be scored when all needed tests were accomplished at test points.
Sufficient information to diagnose and repair the trouble was available
at test points.
(c) To be scored when at least 51% of the required tests were accomplished
at test potnta. Troubleshooting required that several separate tests,
most of which made use of test points, had to be made.
(d) To be scored when the majority of needed tests were not accomplished
at test points. Malfunction diagnosis and repair requtred the making
of tests for which few or no test points were available.
A3-41
~HDBK-472
2$ May 1966
(10) Test Points (Jdentffication]: Determines if all test points required

during the maintenance action are properly Mentffled by circtdt
symboI and pertinent test data. This preoise information provides
diagnostic data to aid tn troubleshooting the malfunction.
8001W
(a) All test pointi are i&ntified with rquirad readtngs

given . . . . . . . . . . . . . . . . . . . . . . ...4
(b) Someare 8uftabIy marked . . . . . . . . . . . . . 2
(c) @tits are not marbd aud test ch.ta are not
given . . . . . . . . . . . . . . . . . . . . . . ...0
Scor @z Criteria
(a) To be aoorgd ~ all test points needed for task &nnpletion

are kkkified (ctiouit symbol), wtth required readings gi%n
(+8VDC. -18VDC, 115VAC, etc.). This is indtcatlve of a best
maintainable condition.
(m To be scored when the majortty of test points required for task

completion are sdltably Identffkl.
(c) To be scored when last points required for ~ oompiettom

are not suitably identified. -Troubleshooting at test points is a
cause for delay because required voltage read&gs, signal
characteristics, etc. , are not specified. This would indibate
that a least maintainable condition exfsts.
(11) Labelltn& Determines if pruds associated with the maintenance actions

mw identified with respect to N rctiit symbol and part identification.
Proper iden~fication of parts can be an important asset to the maintenance
task in that. if part circuit number is emitted from the equipment,
considerable time could be wasted tracing the circuit to identify it.
~imilarly, if information is hidden, requiring rem.ovtd of other parts
to read it. much time u_il\ be consumed.
fiIIL. HDBK-472
24 May 1966
(a) All parts labe!led ~I,Ith full identif!.lng lnform~tion, ~d

all identif:;ing information clearl! .isiblc . . . . . . . . ~
(b) Aii parts iaheiled with full identifying lniorma~ion. but

some inform ation hidden.. . . . . . . . . . . . . . . . . Q
(c) All information visible. but some parts not fully

identified ...,.... . . . . . . . . . . . . . . . ..:
(d) Some information hidden and some parts not fuii~

identified . . . . . . . . . . . . . . . . . . . . . . . . .()
Scoring Criteria
(a) To be scored when all parts associa!mi with the mainlcnantc

act ion are identifier! and this information is clearly visihlc.
To include testing or removing of parts that are clearl}
identified (V401-6RE6) or (R- I 225-401)fl ).
(b) Applies when all parts associated with the n~aintcn.ancc action
are identified, but some of this information is not visihlc.
Applies to testin~ or remo~ing parts thnt arc Iabrllcd, Imt
uhich Information is hidden b} {~i]structlons.
(c) Applicable when alI circuit symbols arc visible, hut some
parts nssocinted with the tasks arc not identified. Parts
required for testing or removal are not identified Ivith
reference to part value, etc.
(d) To he when some parts nssociatecf

scored with the mainten:mre
task contain hidden
circuit symbols and are not fully identified.
Parts required in testing or removal are not identified and infor-
mation is also hidden.
(12) AcJustmcnts: Determines if adjustments such as tuning and ;lllumment

are required, after a maintenance artion, to make the equipment
operate according to specifications. An adjustment will be any action
which resets or changes varinhlc cmmprmcmts such :1s polcntinmtitcrs,
variable capacitors, slug- ttied coils, etc. , whcreb~ the nperatlon of the
A?-l?
MlbfiDBK-472
24 hlay 1966 _,
systcn~ assemb I! or suh:issembiy 15 affected, These actions. depending

upon their critica!in an{! frequency, uill affect the merall maintenance
time.
Scores
(a) 1% ad.iustments or reiilignment are necessary to place

equipment back in operation . . . . . . . . . . . . . . J
h} A few adjustments. but no major realignments are

required . . ...<... . . . . . . . . . . . . . . . .?
(c!) Man! adjustments or major realignments must be

made. . . . . . . . . . . . . . . . . . . . . . . . . . .~
.Scoring Criteria
(a J To be scored when no adjustments are required to bring the

equipment back to normal operating specifications. Appl ics
to repair of the malfunction, if the equipment need only be
turned on.
(l]) To be scored when a few adjustments of a minor nature are

required to place equipment back in operation according to
specifications.
(c) To be scored when many adjustments (time-consuming) or a major

tuning or alignment is required to place equipment back to normal
operating specific attmm.
(13) TestinE (In Circuit): Determines if the defective component or part

can be tested without removal from the circuit. This question is based
on the nature of the equipment and the repair concepts associated
with the particular design.
Scores
(a) Defective part or component can be determined

\vithout removal from the circuit . . . . . . . . . . 4
(b) Testing requires removal . . . . . . . . . . . . . @
A3-44
h!l L- HDBK-472
24 May 1966
Scoring Criteria
(a) Applicable when the component or part can be decisively

determined as being defective without remo~al of any part
from the circuit,
b) To be scored when the component or part must be removed

from the circuit to be decisively determined as defective,
When testing has isolated the trouble to a particular part or
component, however, a definite opinion cannot be made until
such part or component is electrically or physically removed
from the circuit for further testing.
(14) Protective Devices: Encompasses equipment design provisions

for self-protect ion against damage to components or parts after a
malfunction has occurred. If a system has protection devices such
as fuses, circuit breakers, etc. , then the equipment can be protected
from further damage as well as aiding in isolating the malfunction.
If no provisions have been made, further damage and increased repair
time could result.
Scores
(a) Equipment was automatically kept from operating after mal-

function occurred to prevent further damage. (This refers to
mah%nction of such areas as bias supplies, keep- aiive
voltages, etc. )........ . . . . . . . . . . . . . . . . .4
m) Indicators warned that malfunction has occurred . . . . . . . 2
(c) Noprovisions have becnrnade . . . . . . . . . . . . . . . . . 0
Scoring Criteria
(a) To be scored when automatic shut-off devices protected parts

or components from further damage after a malfunction occurred
in a critical area. A typical example of such a malfunction would
be if the Bias supply fails and B+ voltage is automatically cut off
by circuit breakers, fuses, or relay action.
A3-45
MIL- HDBK-472
24 May 1966
(b) To be scored when automatic shut-off-devices do not protect

parts or components from further damage, but when visual
indicators or audible alarms warn personnel of the situation.
(c) To be scored when a critical malfunctionarid parts or

occurs
components are not protected by automatic shut-off devices,
indicators, or alarms. Involves malfunction which damages
parts or components because automatic shut-off devices or
alarms were not provided.
(15) Sale& (Fersonnelj: Determines if the maintenance action requires

personnel to work under hazardous conditions such as close proximity
to high voltage, radiation, movtng parts, high-temperature components,
or on elevated stmctures, etc.
Scores
(a) Task did not require work to be performed in close

proximity to hazardous conditions (high voltage,
radiation, moving parts and/or high temperature
parts) . . . . . . . . . . . . . . . . . . . . . . ...4
!%) Some delay encountered because iIl precau?itinb

taken . . . . . . . . . . . . . . . . . . . . . . ...2
(c) Considerable time consumed because of hazardous

conditions . . . . . . . . . . . . . . . . . . . ...0
Scoring Criteria
(a) To be scored when the maintenance action did not require

personnel to work under hazardous conditions. The maintenance
action did not require precautions t.n be taken, in that the task
was not associated with high voltage, moving parts, etc.
(b) To be scored when precautions were taken because of hazardous

conditions causing slight delays tn the maintenance action.
A typical example would be when a shorting probe must be used
to discharge high-voltage capacitors.
A3-46
-
MIL-HDE3K-472
24 May 1966
{c) Tobe scored when precautions taken because of hazardous

conditions caused a considerable delay to the maintenance action.
Maintenance required that testing be done in close proximit}
to high voltage where extreme caution was necessary, or the
closeness of moving parts (gears, motors. etc. ). caused delay
because of precautions taken.
CHECK LIST B - SCORING DESIGN DICTATES- FACILITIES
The intent of this questionnaire is to determine the need for external facilities.
Facilities, as used here, include material such as test equipment, connectors,
f?tc. , and technical assistance from other maintenance personnel, super ~isor, etc.
(I) External Test Equipment: Determines if external test equipment 1s

required to complete the maintenance action, The type of repair
considered maintsinably ideal would be one which did not reauire
the use of external test-equipment. H follows, then, that a maintenance
task requiring test equipment would involve more task time for set-up
and adjustment and should receive a lower maintenance evaluation score.
Scores
(a) Task accomplishment does not require the use of

external test equipment.. . . . . . . . . . . . . . ~
b) One piece of test equipment is needed . . . . . . . 2
(c) Several pieces (2 or 3) of test equipment are

needed. . . . . . . . . . . . . . . . . . . . ...1
(d) Four or more items are required . . . . . . . . . 0
ScorinK Criteria
(a) To be scored when the maintenance action does not require the
use of external test equipment. Applicable when the cause of
malfunction is easily detected by inspection or built-in test
equipment.
A3-47
-.
MIb~BK-472
24 May 1966
b) To be scored when one piece of test equipment was required

to complete Sufficient information
the maintenance action.
was avaiiable through the use of one piece of external test
equipment for adequate repair of the malfunction.
(c) To be scored when 2 or 3 pieces of external test equipment

are required to complete the maintenance action. Th]s type
malfunction would be complex enough to require testing in a
number of areas with different test equipments.
(d) To be scored when four or more pieces of test equipment are

required to complete the maintenance action. Involves an
extensive testing requirement to locate the malfunction. This
would indicate that a least maintainable condition exists.
(2) Connectors: Determines if supplementary test equipment requires

special fittings, special tools, or adapters to adequate] y perform
tests on the electronic system or sub-system. During troubleshooting
of electronic systems, the minimum need for test equipment adapters
or connectors indicates that a better maintainable condition exists.
Scores
.
(a) Connectors to test equipment require no special tools,

fittings, or adapters . . . . . . . . . . . . . . . ...4
m) Connectors to test equipment require some special

tools, fittings, or adapters (less than two) . . . . . 2
(c) Connectors to test equipment require special

tools, fittings, and adapters (more than two) . . . . . 0
Scoring Criteria
(a) To be scored when special fittings or adapters and special

tools are not required for testing. This would apply to tests
requiring regular test leads (probes or alligator clips) which
can be plugged into or otherwise secured to the test eqvipment
binding post.
MXL-HDBK-472
24 May 1966
(b) Applies when one special fitting, adapter or tool is required for
testing. An example would be if testing had to be accomplished
using a 10 & attenuator pad in series with the test set.
(c) To be scored when more than one special fitting, adapter, or tools
IS required for testing. An example would be when testing requires
the use of an adapter and an RF attenuator.
(3) Jigs or Fixtures: Determines if supplementary materials such as block

and tackle, braces, dollies, iadder, etc. , are required to complete the
maintenance action. The use of such items during maintenance would
indicate the expenditure of a major maintenance time and pinpbint
specific deficiencies in the design for maintainability.
Scores
(a) No supplementary materials are needed to perform

tat3k . . . . . . . . . . . . . . . . . . . . . . . ...4
(b) No more than one piece of supplementary matertal is

needed to perform task.... . . . . . . . . . ...2
(c) TWOor more pieces oi suppiemeniary materiai

are needed . . . . . . . . . . . . . . . . . . . ...0
Scoring Criteria
(a) To be scored when no supplementary materials (block and tackle,

braces, dollies, ladder, etc. ) are required to complete maintenance.
Applies when the maintenance action consists of normal testings
and the removal or replacement of parts or components can be
accomplished by hand, using standard tools.
(b) To be scored when one supplementary material is required to

complete mai.ntenanoe. Applies when testing or when the removal
and replacement of parts requires a step ladder for access or a
dolly for transportation.
(c) To be scored when more than one supplementary material is required

to complete maintenance. Conoerm the maintenance action requirihg
a step ladder and dolly adequately to teat and remove the replaced
parts.
A3-49
hfxb HDBK-472
24 May 1966
-
(4) \isual Contact: Determines if the nature of the equfpment, location,

or maintenance action causes the members of a team to be hidden
from the view of each other at times during the task.
Scores
(a) The activities of each member are always visible

tothe other member . . . . . . . . . . . . . ...4
(b) On at least one occasion, one member can see

the second, but the reveree is not the case . . . 2
(c) The activities of one memlx.?r are hidden from the

view of the other on more than one occasion . . . 0
,Scoring Criteria
(a) Applies when the team members are visible to each other during
the entire maintenance action.
(b) To be scored if one member of the team becomes hidden from

view of the other member or members during the maintenance action.
(c) Applicable if temn members are hidden from view on more than
one occasion.
(5) Assistance (Operations Perscmnel~: Determines whether or not information

or assistance from operations personne~ is required, and if required,
to what extent.
(a) Task did not require consul@tion with operations personnel . . 4
(b)%mec ontactw asrequired . . . . . . . . . . . . . . . ...2
(c) Considerable coordination required . . . . . . . . . . . . . . 0
Scoring Criteria
(a) To be scored when the maintenance action does not require the
assistance of operations personnel. This would apply if physical
or verbal aid to the technical personnel was not required. (Jam
than one minute. )
A3-50
MIL- HDBK-472
24 hfay 1966
(b) To be scored when the maintenance action requires a small

amount of assistance from operations personnel. (One to
five minutes. )
(c) To be scored when the maintenance action requires considerable

assistace from operation personnel in the operation or repair
of the malfunctioning equipmenL (Over five m~u~s. )
(G) Assistance (Technical Personnel\: Determine tke number of technical

personnel required to complete the maintenance action, not including
adrninistmtive or operations type personnel.
Scores
.
{a) Task required only one technician for completion . . . . 4
(b) Two technicians were required . . . . . . . . . . - . . 2
(c) Overtwower euse d... . . . . . . . . . . . . ...0
Scoring Criteria
(a) To he scored When oniv one technician vv&srequired tc complete

the maintenance action.
(b) To be scored when two technicians were required to complete the

m aintenanee action.
(c) To be scored when more than NO technicians were required to

complete the maintenance action.
(7) Assistance (Supervisors or Contractor Personnel): Determines whetier

or not the services of supervisor or contractor personnel (TECH.
REPS. ) were required to complete the maintenance action and the extent
of their participation in the task.
Scores
(a) Task completion did not require consultation with

supervisor or contract personnel . . . . . , . . . . . . 4
A3-51
MIL-HDBK-472
24 May 1966
fb)Some belp needed . . . . . . . . . . . . . ...2
(c) Considerable assistance needed . . . . . . . . 0
Scoring Criteria
(a) To be scored when no supervisor or contractor personnel

are consulted during the maintenance action.
(b) To be scored when a small amount of assistance from

supervisor or contractor personnel is required to complete
the nlaintenance achon.
(c) To be scored when considerable assistance from super~isor

or contractor persomel is required & complete the maintenance
action.
CHECK LIST C - SCORINC DESIGN DICTATES-MAINTENANCE SKTLLS
This check list evaluates the personnel requirements relating to physical, mental,
and attitude characteristics, as imposed by the maintenance task.
Evaiuniion procedure for ibis check iist can best beexpiaineci by way of severai
exampies. Consider first question which deals with arm, leg and back st.rcngth.
Should a particular task require the removal of an equipment drawer weighing
100 pounds, this would im~se a severe requirement on this characteristic.
Hence, in this case the question would be given a low score (Oto 1). Assume
another task which, due to small size and dclicatc construction, required
extremely careful handling. Here question 1 would be given a high score (4),
but the question dealing witi eye-hand coordination and dexterity would be given
a low score. Other questions in the check Ii st relate to various persomel
characteristics important to maintenance task accomplishment. in completing
the check list, the task requirements for each of these characteristics should be
viewed with respect to average technician capabilities,
Scores
Score
1. Arm. kg, and Back Strength

.>
-, Endurance and Ener~
A3-52
.
MIL-IIDBK-472
24 May 1966
score
3. Eye-Htmd Coordination, hlanual Dexterity,
and Neatness
4. Visual Acuity
5, Logical Analysis
6. Memory - Things and Ideas
7. Playfulness and Resourcefulness
8. Alertness, Cautiousness, and Accuracy
9. Concentration, Persistence and Patience
10. Initiative and Incisiveness
Scoring Criteria
Quantitative evaluation of these items range from O to 4 and are defined in

the following manner:
4. The maintenance action requires a minimum effort on the part of

the technician.
3. The maintenance action requires a below average effort on the

part of the technician.
2. The maintenance action requires an average effort on the part of

the technician.
1. The maintenance action requires an above average effort on the

part of the technician.
O. The maintenance action requires a maximum effort on the part

of the technician.
These criteria will be used in scoring the following specific divisions of physical,
mental, and motor requirements.
A3- 53
MIL-HDBK-472
24 May 1966
(1) Arm} leg and back strength : Determines the degree of =m, leg,
and back strength required to complete the maintenance action.
Refers to any effort. no matter how minimal. Varying degrees of
strength are required for various maintenance actions are related
b equipment design.
(2) Endurance and ener~ : Determines the degree of endurance and

ener~ required to complete the maintenance action. Endurance
might be referred tn as the physical counterpart of patience, where
a sustained physical effort is required. Enerm required to complete
the maintenance action when the task requires vigorous activity or
exertion by the technician is also assessed. This applies to the
necessit} of lifting and carrying heavy as~emblies, took, or parts.
(3) Eye-hand coordination, manual dexterity, and neatness: Determines

the degree of eye-hand coordination required to complete the maintenance
action. Refers to any act involving the use of the eves while manipulating
the hands to accomplish the same action. This type of action would be
applicable mostly in testing and measuring actlv~ties: however, it is not
inconceivable that this item would also be appI]caMe m other areas of
the maintenance action. Scoring shaJl be proportional to the degree or
the intensity of the requirements of the task.
Determines the degree of manual dexterity required to complete the

maintenmce action. When the skillful use of the hands is required to
accomplish the task, appropriate degrees of necessity shall be established.
Those type actions invoiving manual dexterity would more naturally apply
to the repair, assembly, or disassembly of equipments rather than the
troubleshooting processes.
Also determines the defrree of neatness required by the maintenance action.

Applies specifically to the requirement of the actual repair where tidiness
is of prime importance to accomplish the task adequately. Since
equipment is designed and constructed in accordance with quality control
specifications, it is importint to consider the care which has to be
exercised during a particular repair.
(4) Visual Acui ty: Determines the degree of visual acui~ required to complete
the maintenance task. When the maintenance action is such that the
visual accuracy of the technician is required to accomplish the task,
A3-54
MIb KDBK-472
---- 24 May ]966
a degree of requirement shall be established. Such actions shaIl

include the need for accurate and prec!se visual activity in finding
indications of trouble, fault~ components, or the visual sensitivity
sometimes necessaw in reading certain oscilloscope presentations.
(5) bgical Analysis: Determines the degree of Iogica! analysis required

to comple~ the maintenance action. Refers to the need for involved
logical analysis or for extensive mental reasoning to determine the
origin of the fault or malfunction. If the problem is such that it requires
orientation on the logical signal sequence, then this shall also be
considered as part of this question.
(6) Memory: Things and ideas: Determines the degree to which the
maintenance action requires a knowledge of the equipment past
history with reference ti) component or part failure, tools to be used,
and sequences to be followed (assembly, disassembly, etc. ).
Also determines the degree to which the maintenance action requires

a previous knowledge of the equipment. Refers to the degree that
the task requires recall of concepts or principles of operation, function
and operation of circuits and parts, or electronic theory and maintenance
- proWlllres.
v) Planfuiness and resourcefulness: Determines the degree of plsmning

required to complete the maintenance action successfully. Refers to
the extent to which the task requires a planned and methodical approach
to assure rapid diagnosis and repair of the equipment fault or malfunction.
Also determines the degree of resourcefulness required @ complete

the maintenance action. Refers to the capabilities necessary in dealing
with a situation or in meeting difficulties pertaining to the diagnosis
and repair of the equipment. Conditions sometimes exist where certain
needed materials such as tools, test equipment, or technical publications
are not avai Iable, although substitution is possible, by some improvised
method, to accomplish the task adequately.
(8) Alertness, cautiousness, and accuracv: Alertness is a readiness or

promptness in comprehending and a keen awareness and knowledge of
all events or factors affecting the maintenance action. Cautiousness is
the exercise of forethou~ht. srI that risks mav be avoided or minimized
A3-55
MI LHDBK-472
24 May 1966
during the maintenance action, (A purveyance of all possible consequence

before making a decision. ) Accuracy is attained by the exercise of care
by showing close attention to the details of the maintenance taak and
cautioutmesa in avoiding errors. The design requirements for these
characteristics are to be assessed.
(9) Concentration, p ersiatence, and patience; Concentration is the clone

mental application or exclusive attention to tbe maintenance task and
tbe direct focusing of tbe mind upon one thing to the exclusion of
everything else, Persistence refers to maintenance taeke with the
implication of being able to carry performance to a successful conclusion.
Patience is tbe quiet persever~ce, calmness in working, and being
undisturbed by obetacles, delays, or failures which might occur during
the maintenance task.
(10) Initiative and incisiveness: Idtiattve is the ener~ or aptitude displayed

in the initiation of action and the ability or power to introduce a new
measure or course of action. Incisiveness is the keenness of mind wd
acuteness of understanding the task at hand,
A3- 56
?QL-HDBK-472
24 May 1966
PROCEDURE ?\
1.0 GENERAL
This procedure is based on the use of historical experience, subjective evaluation,

expert judgement and selective measurement for predicting the downtime of a
system/ equipment. The procedure uses existing data to the extent available.
It provides an orderiy process by which the prediction can be made aod integrates
preventive and corrective maintenance. Task times to perform variou8
maintenance actions are estimated and then combined to predict overalI syetem/
equipment maintainability.
1.1 Philosophy, Assumptions and Summary
This procedure recognizes that throughout a mission, a system/equipment performs

various operational functions and that the maintenance time depends upon the
specific operational function which is in process. To clearly understand tie
meaning of an operational function we wrill discuss the mission of a commercial
airliner as an example.
Prior to take-off the first operational function is to warm up the engines and
perform z preventive mainiemlncx? check o~t. Ihxing t!!!s pertml n failure may
occur (associated with this first operational fimction) which requires corrective
maintenance action. This is the type of failure which is readily detectable
such as a malfunctioning engine, A failure requiring corrective maintenance
may also be detected during the specified scheduled preventive maintenance
routine which involves a detai!ed checking of all instrument readings.
A simiiar type of reasoning is applicable to other operational functions of the

aircraft such as taxiing down the field, in flight and Iandtng. As a general deftii-
tion applicable to all systems either mechanical or electro-mechanical, an operational
function is defined as that particular function which the system is perforrntng at the
specific inwrval of time during which the maintainability analysis is being conducted.
In other words, the procedure requires the development of a mission/maintenance
profile which specifies the various operational functions of the system and the
scheduled preventive maintmance actions required for each operational function.
Another significant assumption is that the estimate of task times can be made best
by a maintenance anaiyst working closely with the design engineer, or by the
design engineer himself. Therefore, it is assumed that the maintenance task times
so estimated are practical, realistic and applicable for performing a maintainability
prediction.
4-1
Mlb HDBK-472
24 May 1966
1.2 Applicability
Because of the nature of the time estimatmg techniques, this procedure is

applicable to al I system s~equipments.
Time ana)ysis can be performed as soon as sufficient system/equipment deft.nltlon

exists. This level of definition for initml estimations is usually available fairly
early in the definition phase, The application of the time estimating procedures
will pemnit the updattng of the equipment desfgn and can, therefore, take place
throughout the design and development phases.
1.4 Basic Parameters of Measurement
The intrinsic maintainability of the syst.emiequipment is predicted under the

asmrnption of optimum utilization of specified support equipment and personnel.
The intrinsic matntatnabtltty 1s given by the following parameters.
Mean Corrective Downtime - MCDT
Mean Preventive Downtime - MPDT
Total Mean Downtime - TMD1
1. S Information Required
The following information concerning the operational and maintenance environment

of the system and subsystems is required to make the initial maintenance task
time analysis. Subsequent schematics, assembly drawings, etc. , will be used
for updating this time estimation as the system design continues.
a. System Block Diagram
b. Functional Flow Diagrams
c. Shbsystem Block Diagrams
d. Subsystem Flow Diagrams
e. End Item List
4-2
-
hl[L- HDBK-472
. 2!4 May 1966
f. End Item Failure Rates
~. hltiintenance Concept
h. lda~ntainabilitj Goals
i. Operational Resources (facilities, personnel, supp{>rt equinment, etc. )
j. A detailed definition of the task being performed
k. Imcatiori at ~hich the task is being performed
1. Environmental Constraints
1.6 Data Basis
This procedure utilizes expert judgment and existing data snurres on maintenance
task time, but the procedure does not rely solely on existing data. The applicability
of the data is decided by the anslyst and is supplemented h~ his expert judgment
in estimating maintenance task time uhcn such information is not available.
1.7 Correlation Between Predicted and Ohsrrved Vtilues
Procedures similar to this one have been applied selectively .md verified to a
high degree of accuracy. This accurac~ is npplicahle only ~vhere specific
maintenance actions and equipment end Items arc being anal}zed, and is
dependent upon the qualifications of the persrmnel performing the evaluation.
2.0 ANALYTIC FOUNDATION
2.1 Gencraf
The analytical foundation of the task analysis procedure integrates the development
of task performance time for preventive and corrective maintenance actions.
A maintenance action is defined as the exclusive maintenance task ~vhich occurs
at a specific location and within a specific set of conceptional and physical
constraints. This maintenance action permits the logical development of elapsed
times, subsystem equipment levels.
The mean corrective maintenance time for the system equipment wil] vaq for
each individual scheduled preventive maintenance action appllcahle to a specific
4-3
MIL- HDBK-472
24 May 1966
operational function. The task zma.lysis procedure permits the evaluation of

these times from the end item, up to the system level. The products of the
procedure are:
a. The elapsed time b perform preventi~e maintenance action, assuming

that no detectable malfunctions exist in the system.
b. The elapsed time to correct malfunctioning end items detected during

each preventive maintenance action of an operational function.
c. The distribution of corrective maintenance times for detectable

malfunctioning end Items for each preventive maintenance action
of an operational function.
d. The mean corrective downtime ( MCDT ) for detectable malfunctioning

end items for each preventive maintenance action of an operatiomd
function.
e. T%e distribution of corrective maintenance task times for the system

and subsystems.
f. The preventive downtime ( PDT) for the system and subsystems for a
Z?ec!ficti caleni?ai time.
g. The total mean corrective downtime ( MCDT ) for the eystem and subsystems
for a specifjed calendar time.
h. The total mean downtime for integrated preventive and corrective

maintenance for the system and subsystems for a specified calendsu time.
These maintenance downtimes relate only to the inherent maintainability of the

equipment, since administrative and other delays are not normaHy definable
during tie deai~ of the equipment. The estimated elapsed time required to
perform matnt.enance on a system will vary as a function of tbe conceptional
and physical conetralnts within which the esttrnation was made. These constraints
consist of the availability of physical resources (i. e. , personneA, spares and
con sum akdes, support equipment, and facilities) and applicable maintenance
and operational concepts (i. e. , testing concept, level of repair, m~sion de-
scriptions, etc, ). The applicabili@ of specific constraints must be documented
if a given time estimate is to be meaningful. Only single elapsed ttmes are
estimated for each maintenance task. This number should approximate the mean
time required to perform the task under actual conditions. The correlation
4-4
MIbHDBK-472
24 hhy 1966
Wwen estimated and verified task times described tn paragraph 1.7 juattfiem
the use of single value~ for the purposes of this procedure.
A series of mission/maintemmce profiles will be established based on the

system operational requirements. The8e profiles *ail specify the echeduies
of operational functions and preventive maintenance actions for a given calendar
time. The mean commtive downtime and preventive downtime for the system
are calculated tn sequence by function, mission/m~nmn~ce profile * -d
complete system. A procedural flow block diagram of the procedure is shown
tn Figure 4-1. lb explanation of each block in the diagram follows:
The end itenm (1) of tbe system are identified and categorized under the appropriate
headings as: system, subsystem, assembiy, etc. , down to the smallest piece of
equipment on which a specific maintenance action will be accomplished. The
failure rate 1s identified for each end item. The preventive maintenance actions
of an operational function (2) to be performed on tbe categorized end items are
cieflned (e. g. , check out, servicing, adjustment, etc.). The physical and
conceptual constraints previously described mu~t be defined and documented for
each function. Ihe corrective maintenance actions (3) to be performed on
appropriate cate~rized end items are deff.ned. Tbeae actions wtll inclu*. tit
are not necwmaril y limited to the maintenance actions of @@t, remo V% ~P~~ei
. adjust, repatr, etc., epec!fled by the applicable constraints. The phystcd nd
conceptual constratnt8 previoudy duacxibed defined s set cf corrective nctton,
which can be undertaken. The detectable end ttem malfunctions for each preventive
matntensnce action of an operational function are defined, (4). Thoee end items
which can be detected as malfunctioned, but which cannot be corrected (within the
Constratnta of the location at which the maintenance is occurring), are grouped.
No troubleshooting witl be cimducted wkhtn these groups of end items since,
by definition, no corrective action can be umiertaken. A task analysis is con~cted
for each preventive maintenance action (5). A distribution of end {tern task times
and a total time for each nperatlonal function iE generated. The total task time
for the operational function is compared to the allocated time to determine if the
maintatnabillt y design of the equipment ts adequate. Xfnot, the dhtrfhtton of
end item task times pennitg McnMfication of critical desi~ points. A tad
analysis k! conducted for corrective matnt.enance (6) associated with eaoh of the
prevent{ ve matntenence of an operational function. Thts ansJysis is corductd b
deriving the t roubleshoottng, repair, and verification time for each end item
previously defined a+ a detectable and repairable malfunction. These times
are described in terms of a distribution of end item corrective maintenance ttmes
versus frequency of occurrence and by a mean corrective downtime ( MCDT)
for the spectfied operational function. As before, the MCDT and distributions are
used to identify critical design points. The preventive nnd corrective maintenance
4-5
MIIJ-HDBK-472
24 May 1966
-1-
~
1%1 lx II
1. 1
1
.- m
I
4-6
MIL-HDBK-472
.- 24 May 1966
times and associated MCDTS are integrated (7) over the previoudy specified
calendar time (e. g, , 6 months) to derive the total preventive downtime, tda.1
mean corrective downtime, and the t.eta! mean downtime, where all times
are related to the inherent maintainability characteristics of the system and
exclude administrative and other delays.
2.2 Theoretical Considerations
A parametric description of the previously described procedure is provided

m the following paragraphs and is referenced to the numbered blocks in
Figure 4-1.
2. 2.1 End Item Identification
Each end item of the system is described by Ii such that 11, 12 . . . . In.l,
In are inclusive of all end items within the system. The failure rate of each
end item is given by ~i , where ~i is assumed to be a constant over the
specified calendar time, and where ~i is the failure rate of end item II, etc.
2.2, 2 Operational Function
The preventive maintenance actions of an operational function associated with

the system are given by Pj such that PI) P2 . . . . Pm -1, and Pm are lnciusive
of all preventive maintenance actions and where each action is defined by type
(e. g. , inspection, servicing, etc. ) and hy physical and conceptual constraints
(e. g. , personnel, spares and consumables, support equipment. facilities,
testin~ concepts, etc. ). A new function must be defined if the type or constraints
are modified.
A subset of Ii is associated with each function pj. The operational functions of

the system are given by Or such that O,, 02 . . . . Or-t, Or are inclusive of all
operational functions, where each function is defined by the end items of the
system being used. An operational function must be defined for each different
subset of Xi utilized durin~ the operation of the system. By definition, all end
items Xl, 12 . . . . In-,, I ~ will appear in at least one subset of Ii. associated
with the preventive maintenance action of an operational function.
2. 2.3 Corrective Maintenance Action Definition
The corrective maintenance actions associated with the system are given by
c,, C2 . . . CQ-l, and Cg are inclusive of all corrective maintenance actions
where each action is a maintenance action taken to correct a detected malfunction
4-7
MJL HDBK-472
24 May 1966 -
indication (e. g. , test, remove, replaCe, adjust, repair, verify, etc. ) within a
specific subset of constraints. The actions are assigned to each system end
items so that the times involved are exclusive to the individual end items,
(e. g. , he removal of M end item, aasuming that all access to that item has been
accomplished).
2.2.4 Malfunction Il?tection Analysis
The concept on which the analytical procedures are based prescribes that only those
system end items (Ii) which can cause identifiable malfunction indications during
tie preventive maintenance, Pm, action or operation~, or , functions will contribute
to the mean corrective downtime for that function, The pxwbability that corrective
maintenance will occur is a function of the~~i ,of the subset of Xl associated with
the specific function. The end items 11, 12 , . , In are assumed to have only one
failure mode for purpose8 of thie di6cuseion. However, in some cases it may be
necessary to specify the various possible failure modes (e. g. , for a resistor -
open, shorted, etc. ) to provide adequate downtime estimation accuracies. The
determination of the necessary level of definition required to achieve specific
accuracies of the task time estimation is beyond tbe scope of this document. Those
end item failures which can be detected but not repaired within the constraints
associa@d with a specific preventive maintenance action or an operational function
will be grouped tito the sm~iem ieolatabie black boxes. The identification
Gf one G: these black boxes as ~ing failed wiii require that the system be
transferred to another fuction (e.g. , the system shifted to a degraded operational
mode or the actual equipment moved to another maintenance area).
2. 2.5 Preventive Maintenance Task Time Analysis
The task times for preventive maintenance actions are given by:
oTm f
= i:,Ti m
where: PDTm = The total preventive maintenance performance time for

action Pm
T 1- = The time to perform the maintenance task on end item I ~ as

required by action Pm
A distribution of the individual task times within each action can be developed
to identify critical design potnts as previously described.
4-8
MiLr HDBK--li~
24 May 1966
2. 2.6 Corrective hlaintenance Task Time Analysis
The end ]terns defined in 2.-.q -t as being detectable during a specific preventive
maintenance action or an operational function serve as a starting point in the
conduct of the correc!~ve maintenance task analysis. The fault isolation concept
for the system under action, pm of m Or function, is def!ned. The trouble-
shooting, repair, and verification time for repairable end items, or the trouble-
shooting time for non- repairable end item groups are derived based on the
defined fault isolation concept. Thus, for action pm:
Item. I : T ~m (~T$l~) + ~cim + V,m
item 2: T2m (~l,am) + TC2m + h2m
Item n: T nrn = (~T~nm ) + TCnm + v.m
Item ith:
L.
where:
Tim The total time required to correct malfunctioning end

item II during action Pm of aJI operational ~ction
w The troubleshooting test times required to isolate end

item xi during action pm
= The time required to remove, replace, adjust, or othenise

Tcim
repair malfunctioning end item li during action pm
Llm The time required to verify that the system is good, given
that II is replaced, repaired, adjusted, etc. , during
action Pm
For function Or:
Item 1: 1,, : (~~slf) + Tel, + G,,
4-9
-.
MIL- HDBK-472
24 May 1966
Item 2: : (~Ts2 ) + Tc2f + ,2

r r
Item n:
Item iw: : (XT,,,) + Tc,, + TV,,
where:
T:*? The total time required tn carrcci tnahnctiming end item

Ii during function Or
S1, The fault isolation test times required to Isolate end item
II during function Or
The time required to remove, replace, adjuat, calibrate,

Tcf
or otherwise correct the malfunctioning end item I i
during function Or
T vi, The time reouired to verify that the system ifi mod. given
that 11 is rep~aced, repaired, adjusted, etc. , during
function Or
In addition, the time to isolate the non-repairable end item groups during
action Pm is gtven by:
Tjm = ~T~j~
where:
T- The total time required to isola~ he )th gTOUP during

Iur
action Pm of an operational function
= The troubleshooting time required to isolate tie jl~ grouP

Tsjm
during action Pm
The time required to isolate the non-repairable end item groups during function
or is given by:
Tjr = &jr
4-1o
Mlb HDBK-472
24 May 1966
= The total time required to isolate the jth group during tbe
function Or
. The troubleshooting time required @ isolate the jt~ !PW
during function Or
The mean-corrective- downttie of the system or identification of the requirement

to shift it m another maintenance or operational function during action Pm is
given by:
where:
MCDTm = The mean- corrective- donttme for the system during

action pm of an operational function
Aim = The failure rate of detectable malfunctioning end item

Ii during action pm
Ai,m = The failure rate of the it~ end item in the jt~ non-repairable
group which can be isolated during action pm
The mean-corrective-dowtime of the system or identification of the requirement

to shift to another maintenance or operational function during function Or
is given by:
where:
MCDTr = The mean-corrective-downtime for the system durtrig

function Or
Aijf = The failure rate of the i ~h end item in the jtb non-repairable
group which can he Jsolated during function Of
4-11
M.Ib HDBK-472
24 May 1966
2.2.7 Total Maintenance Task Ttme AnalyBis
A total maintenance time analysis 1s conducted to define the total time requtmd
to perform preventive maintenance, and tbe total meam-comective-downtime,
for maintenance of the eyetem. Tbe total time for preventive matntfmance
IS given by:
PDTt = Xem PDTm
where:
PDTt = Total praventtv-downtime durtng the 8pectfied

calendar ttme
m = Frequency of occurrence of the mt~ pmventtve maintenance

actton during the peclfied calen~r time
The mean-corrective-downttme for the uystem is dertved from tbe mtnslon/

xnalnterumce profiles.
The mean-comective-downttme for the uyutem is given by the weighted

(normaIt zed fallum mtee) of the MCDT fcr Uadl set!= Pm d G Cr
operational fimcttono Therefore,
wbe re:
MCDT~ = The mean-correctlv-downtime for the system for the

gtven mi8aion/m8intanence profile
Applying the equation to a hypothetical mtsalon/nmintemace profile IIMUlte in:

-. MW HDBK-472
24 May 1966
+
em,
(Ai+Q1]
The total mean-corrective-downtime of the smm for tie mhsicdm~in-~
prufile ie given by:
MCDTL = f(MCDTJ
where :
MCDTt = The total mean-corrective-downtime of tbe system for tbe

mission/maM6ma.nce profile
f = The number of detectable faihmw occurring dur!ng the

cdeack? tlm e
The total mefm-doumtlme of the system with a specified mtssion/maintemnce

prt)fik iB @V8n by:
~am PDTm + MCDTt

P =
where:
Tp = The total mean-downtime of the system with a specified

mi8sion/maint4mance profile for the calendar time
am = The frequency of occurrence of tie action % ~flx tie

calendar period
The use of a mix of mission/maintenance profiles for tbe system gives a total
mean-dowmttrne of:
4-13
MIL-HDBK-472
24 May 1966
where:
t The total mean-downtime of the sys~m for a given mix of
mi6sion/rnaintenance profi le6
o? = The frequency with which the p tfi mission maintenance profile

will occur during the calendar time
3.0 APPIJCATION
The development of the systa!n maintenance times is initiated by establishing

and grouping the physical and conceptual constralnte existing wiWn the
maintenance environment. The aLlowable corretive maintenance actions
(e. g., remove/replace, repair, test--troubleshoot--, adjust, etc. ) are specified
for each end item for each Bet of constraints.
Step 1 - The constraints applicable to each preventive maintenance action of an

operational function, and to tbe corrective maintenance action are related through
the use of the matrix shown in Figure 4-2.
CORRECTIVE MAINTENANCE ACTIONS

k 1
c1 c~ Cq
PI x x
~w
2:*
l-az f2 x
220
ww-
>++
UJZQ
a-d
~;
Pm x x x
Figure 4-2. Related Constraint Matrix
An X at a row/column junction of the matrix Indicates that the applicable con-

istralnts to the actions Pm of an Or wII1 permit the accomplishment of the
corrective maintenance actions (C ~).
Step 2 - An end item corrective maintenance action mstrlx, as shown in Figure

4-3, is used as an aid in tbe comluct of the task time analysie procedun. The
corrective maintenance actions assigned to each end item are described to match
the established physical and conceptual constraint.
4-14
hfI&KDBK-472
24 May 1966
.
A ttme wlue at the matrix junction of an end item row and an action column
indicates that this end item is acted upon or utilized during that action. For
example, in Figure 4-3 end item (1J is acted upon during corrective maintenance
action (cl ).
END iTEM I CORRECTWE MAINTENANCE ACTIONS
1, A, c, c~ c~
C9
I, h, T,, o T 1, TI
9
k~
I
1~ o ?22 72, I
I
I
13 k3 T3, 0 T3, I
I
t
I
I
I
I
i
I
x~ )S~ Tn, ---- ---- ------ Trig
Figure 4-3. End item/Corrective Maintenance Acdon Matrix
undertaken on tbe system end items within the specified physical and conceptual
oonstrainta associated with the preventive maintenance action of an operation
function.
!Xep 3- An end item, preventive maintenance action matrix, as shown in Figure

4-4, is used to oalculate the individual action performance time (PDT) @S
(kmcrihed in Paragraph 2.2.5) and the related m-correctfv-downttm e ( MCDT)
@s described in Paragraph 2.2, 6).
4-J 5
MILHDBK-472
24 May 1966 -
END ITEM I PREvENTIVE MAINTENANCE ACTIONS
P, P, P* P2 Pm Pm
Ii hi Corr.
Prwt.
Prcv. corT. Prav. Corr.
1, AI 1, 0 0
I
I
X2 AZ 12, 12, 0
Is As o 0 T3*
#
I
I
I
I
I i
I
I
I
Tp,
I IJ I
1 I
t I
I I
I
I (in-l)
I
I 1 1
X* A* o TR, Tot 0
I I 1
PDT,
/
MCDT2
PDTm 8 ~Tim
Figure 4-4. End M6m@reventive Matnmce Action ~x
Each weventive maintenance action colxmn is divided tnto two parts; the fird
for the times required to perform the specified action on the affected end
items of the sWem; and the second part for the times required to troubleshoot,
repair, and verify detectable malfunctioning end items.
The end item /corrective maintenance action matrix is used to establish those end
items which can & corrected if malfunctioning within constrain specifjed
4-16
MIL HDBK-472
24 May 1966
for the individual preventive maintenance actions exist. The mean-corrective-

downtime for each preventive maintenance action is calculated by establishing
the troubleshooting path to each detectable end item malfunction, considering the
specified troubleshooting logic and constraints, and the interrelationships of the
various detectable malfunctioning end item6, The time tnm ia placed at the end
item (In) t action (Pm) matrix junction of Figure 4-4. This time is the summation
of the troubleshooting, repair, and verification times for repairable end items.
The time Ttjm to troubleshoot non-repal rable end item group6 is submitted as
required, as shown in Column p2 Corr. of Figure 4-4. The non-repairable grouP ~
this example consist5 of end items 14 through In- 1.
Step 4 - An end item/operational functton matrix as shown in Figure 4-5 is used

to calculate the MCDT ~ for each operational function as described tn Paragraph
2,2. 6.
END ITEM OPERATIONAL FUNCTIONS
Ii Ai o, o~ 03 ----- or
1, A, 1,, 0 0 T,r
1~ X2 Tz, Tat o 0
13 AS o 0 T3a o
I
I
I
I
In An o Tnt Tn, T ~,
MCDT, MCDT2 MCDT s MCOTr
Figure 4-5. End Item? Operational Function Matrix
4-17
. .
MIL- EDBK-472
24 M8J 1966
The MCDT, for each opemtional function is calculated h a manner ldentia.1

to that used for tie NCDfm for preventive matntenanoe ctions.
~ - Tbe W.a.l preventive mabtenance time, tbe total mean-correctlvedmmtlme,

and the total mean downtime for maintenance are cald.ated for each miaai~nWn-
tenance profile a8 described in Pamgm@ 2.2.7,
CwItodhuls: ~fig activity:
Army - MI Navy - AS
Navy - AS Project No. MXSC-0327
Air Force -11
Review activities:
Amy - EL, Ml
Navy - AS, EC, 0S, SH
Air Force -11, 13, 14, 17
4-18
MIL-HDBK-472
24 May 1966
--
BIBLIOGRAPHY
1. Maintainability - Prediction: Theoretical Basis and Practical Approach

of, LSAF Aeronautical Systems Division, 15 Feb 1962.
2. Maintainability - Prediction: Theoretical Basi ~ and practiral Approach

of, USAF Aeronautical Systems Division, 31 Dec 1963.
3. Maintainability - Design Criteria Handbook for Designers of Shipboard

Electronic Equipment, NAVSHIPS 94324, 30 Apr 1962.
4, A Research Plan for Ikveloping Methods of Maintainability Measurement

and Prediction, Rome Air Development Center Tech. Note No, RADC-TN-60- 5,
AD-231 090, 6 Jan 1960.
5. A Review of the Literature on Design of Information Job Performance

Aids. Proj. No. 1710. USAF Airborne Systems Division Report No. ASDTR
61-549, 1961.
6. MSintainability Measurement and Prediction Methods for A. F. Ground

Electronic Equipments, Phase II Report RADCTN-60-221, AD-247 155,
15 -p 1960.
7. Maintainability Measurement and Prediction Methods for A, F. Ground

Electronic Equipments, Phase Ill Report RADC-TN-61-141, AD-262 454,
15 Jun 1961.
8. Maintainability - Prediction Technique, Phase IV Report RADC- TDR-62-156,

AD-277 199, 15 Mar 1962.
9. Maintainability Engineering, Phase 1 Report RADC-TDR-63- 85, Vol. 2,

AD-404 898, 5 Feb 1963.
10. Maintainability Technique Study, Final Technical Report, Phase V Report

RADC-TDR-63-85, Vol. 1, AD-404 899, 5 Feb 1963.
11. Maintainability and Supportability Evaluation Technique, Wright-Air

Development Division, Part I, Part II, Vols. 1 and 2, May and Jun 1860.
5-1 4
f
NUTE nib form my Dot k M-d to rwqll=t mpa of rhcunmta, aor w mquut ~ dm@t30m, or CiUifiitmn of
apDcifkatm nqwwmmb Oa Crumat CQIlhcts.~nu Subdwd a thb form tb ad -*W* or tnlpJy $uLhorUdJon
to mm my portwm of th nferwmd dtxua=atfd or to ~d mmtrutual roquirwmenh.
,. ....
(Fold dons Mb Ufv)
DEPARTMENT OF THE NAVY
111111
OFF)CiAL USINE=
~P4ALT~ FOR RIVATE USE t=
I
~
BUSINESS
CLASS
FfRST
REPLY MAIL tPERM}T NO. $2603 WASHINGTON D C t
& -
POSTAGE WILL aC PAID BY THE OEPAWMEP4T OF THE NAVV
Commander
Naval Air Systems Command (AIR 51 122)
Washington, DC 20361
.-. .. . .
STANDARIW?ATION DOCUMENT IMPROVEMENT PROPOSAL
(* lnsfructiau - Rcverx we)
OOCUUENT MUMOC n 2 00 CUMEWT TITLE
In OTHER (hpuity)
b Ruomm.nO.d WO-dIDW
c. haaonfR*tmulo for Ruomwundstton
!
i
1
I
I
I
I
I
b WORK TELEPHONE NUMOER Afsa
I. NAME of SLOMITTER fL~I, Ftrtf. ,wif - OpImOI
Co&l - optional
(hCk+d#
I
1
MAILING 400RCSS ISIWI CIIY. Slink ZIP Cndr I - Og!,ond 8 0A7E OF SUSMISS!ON (V YMMDD)
1
I
I
ml FORM
2 MAR 1426 REv IOUS EOIYION IS 0f150LETf
MIL-HDBU-472
mTKE 1
12 January 1984
MILITARY HANDBOOK
TO ALL HOLDERS OF MIL-HDBK-472 :
THE FOLLWING PAGES OF MXL-WBK-472 HAVE BEEN REVISED AM) SUPERSEDE THE
kGES L!STED:
NEM PAGE DATE SUPERSEDED PAGE MTE
tlt 12 Jan 1 iii 24 my ?966

tv 24 my i (REPRINTED MITHOUT
CHANGE)
Vfl 12 Jan 1 v!! 24 *Y 1966
VI!! 12 Jari 1 V!tf 24 *Y 1966
ix 12 Jan 1 ix 24 my 1%6
1 12 Jan I 1 24 MY 1%6
2 12 Jan 1 24 hy 1966
3 24 ?kY 1 ( REPRINTED kCNT CHANGE)
?2 Jan 1 4 24 my 1%6
6~1 12 Jan 1 5-1 24 My 1%6
2. THE FOLLOHING PAGES ARE TO BE ADDED:
PAGE DATE
x 12 Jan 1964
xl 12 Jan 1984
xl! 12 Jan 1984
Xt!i 12 Jan 1984
xiv 12 Jan 1984
5 12 Jan 1984
V-1 thru V-33 12 Jan 1984
A-V-1 thru A-V-22 12 Jan 1984
B-V-1 thru 8-V-27 12 Jan 1984
C-V-1 thru C-V-8 12 Jan 1984
D-V-1 thru D-V-9 12 Jan 1984
3. RETAIN THIS NOTICE AND INSERT BEFORE TABLE OF CONTENTS.
4. Holders of UIL-HD8K-472 will verify that page changes and additfons

tndicatwl Iwrcio h8v@ baan tcrcd. This mttce WI 11 be retained as check
sheet. Th{s i ssuanct, together w! th appended pages, ~s a separate
Publication. Each notice Is to be retained by stocking points unti 1 the
Ml 1 I tary Handbook !s cmpletely revfsed or cancel led.
INNTY -79021
.
MIL-HDEUM72
MOTICE 1
Cust@ans: Preparing activity:

Amy - M! Wy - AS
mavy - AS (Project MNm-7902)
Air force- 17
Rtviewacttvtttt$ :
Amy -
my - EC, 0S, SH
A!r force - 11, 13, 14
-
MIL-HDBK-472
Prediction facilitates an early assessment of the maturity of the

maintainability design and enables decisions concerning the compatlbllfty of a
proposed design with specified requirements Or the choice of better
alternatives.
The maintainability prediction procedures I and III are applicable solely to

electronic systems and equipments. Procedures II and IV can be used for all
systems and equipments. In applying procedure 11 to non-electronic equipments
the appropriate task times must be estimated. Procedure V can be used to
predict maintainability parameters of avlonlcs, ground and shipboard
electronics at the organizational, Intermediate and depot levels of
maintenance.
In coocluslon, the use of this handbook facilitates the design, development,

and production of equipment and systems requlrtng a high order of
ma!ntainabil!ty.
1? Jan 1984
xr&HDBK47a
2fJ MAY1966
TABLE OF CONTENTS
Pegs
DWRODUCTION
Tbe Need for MaialalaaMMy RecMcUon 1

Definition of MILbtalndJlll@ 1
BMiC Amumptiom zad Merpretatkms 2
Elements of Mnlntalnabiltty Rediction Tecbxdques 2
-mary 3
MAINTAINABILITYPREDICTION ~OCEDURES
~OCEDURE I 1-1
SECTTON1.0 GENERAL 1-1
1.1 Philosophy, Assumptions and lihunmary 1-1

1.2 AppLkabili& 1-2
1.3 Pointof Application 1-2
1.4 Basic Parameters of Measurw 1-2
1.5 Xnforrnatton Required 1-2
1.6 Data Beeis 1-3
1.7 Correlztlon Between Predicted and @erved Values 1-4
2.1 Structire of Time Elements in Fix of Malfunction 1-4

2.2 Elemental ActiviUes 1-12
2.2. I AsmunptionEI Relating to Elemental Actkltks 1-14
2.2.2 Major Characteristics of Elemental Activttles 1-14
2.3 Synthesis of Time Distributions 1-16
2.3.1 Modes for Synthesizing Distritnationa of Ttme 1-16
2.3.2 Cumulstlve Ttme Dtstritmtions and the Monte Carlo 1-17
Method
SECTION 9.0 APPXJCA170N 1-19
3.1 Preliminary Procedure 1-19

3.2 Steps of tbe Prediction %occdure 1-19
iv
1411-HDBK-472
Page
3-16
3.2 Step-by-Step procedure for Calculating Task Sample
3-18
3.3 Step-by-Step Procedure for Performing Task Prediction
Calculation of Maintenance Indices 3-30
3.4
A3-33
ADDENDUM A DESIGN CHECK LISTS AND SCORING CRITERIA
A3-33
CHECK LIST A - SCORING PHYSICAL DESIGN FACTORS
A3-47
CHECK LIST 8 - SCORING DESIGN DICTATES-FACILITIES
A3-52
CHECK LI~? C - SCORING DESIGN DICTATES-MAINTENANCE SKILLS
4-1
PROCEDURE IV
4-l
SECTION I.O GENERAL
1.1 philosophy, Assumptions and SU~rY 4-1

Point of Application 4-2
1.3
1.4 4-2
Basic Parameters of Measurement
Information Required 4-2
1.5
4-3
1.6 Data Basis
4-3
4-3
SECTION 2.0 ANALYTIC FOUNDATION
Genera 1 4-3
Theoretical Considerations 4-7
H
2.2.1 End Item Identification
4-7
2.2.2 Operational Function
4-7
4-7
2.2.3 Corrective Maintenance Action Definition
4-8
2.2.4 Malfunction Detection Analysis
4-8
2.2.5 Preventive Maintenance Task Time Analysis
4-9
2.2.6 Corrective Maintenance Task Time Analysis
2.2.7 Total Maintenance Task Time Analysis 4-12
SEC71ON 3.0 APPLICATION

4-14
v-1
PROCEDURE V
SECTION 1. GENERAL v-1
1.1 Philosophy and Assumptions v-1

1.2 Point of Application v-1
1.3 Bas!c Parameters v-2
1.4 Information Required v-2
1.4.1 Method A v-2
1.4.2 Method 3 v-2
::
MIL-HDBK-472
Page
SECTION 2. REFERENCED DOCUNENTS v-3
2-1 Issues of documents v-3
SECTION 3. BASIC DEFINITIONS AND MODELS v-3
3.1 MTTR elements v-3

3.2 Basic models v-4
3.2.1 l+eantime to repa~r v-4
3.2.1.1 Mean repair time for the n RI V-6
3.2.1.2 Average repair time for the n RI V-6
3.2.2 Percent isolatlon to a single RI V-6
3.2.3 Percent Isolation to a group of RIs. v-7
3.3 Hean maintenance manhours per repair (=frepalr) v-7
3.3.1 Mean maintenance manhours v-7
3.4 Nean ma~ntenance manhours per maintenance action (~/NA) V-8
3.5 False alarm rates v-9
3.6 Nean maintenance manhours per operating hour (~/ON) v-lo
3.7 Nean maintenance manhours per fl~ght hour (=/FH) v-lo
3.8 Naxtmum corrective maintenance t!me for the (0) v-lo
percentile (Mma. (0)1
3.8.1 Approximate ~.~, ($> v-lo
3.8.2 Accurate k.. (@) V-11
SECTION 4. APPLICATION V-II
4.1 Method A, early prediction procedure V-11

4.1.1 Predlct\on requirements definition V-11
4.1.2 Replacement concept definition V-12
4.1.3 Determlnatton of the prediction parameters V-12
4.1.4 General prediction model and submdel selection V-15
4.1.s MTTR computation V-IS
4.1.s.1 Method of computing SC and S, V-18
4.1.5. 1.1 Nethod 1 V-18
4.1,5.1.2 Method 2 V-19
4.1.5.2 Computation of A, To and Tn V-19
4.1.5.3 Deteradnatlon of MTTR v-21
4.2 Nethod B - detailed prediction procedure v-22
4.2.1 Prediction requirement def!n~t!on v-22
4.2.2 Replacement concept definition v-22
4.2.3 Fault detection and Isolatton output Ident!flcatlm V-23
4.2,4 FD&I outputs and hardware correlation V-24
4.2.5 Prepare maintenance flow diagrams V-26
4.2.6 Ttme line analysts V-31
4.2.7 RI and FD811output correlation V-32
4.2.8 Compute ma!nta!nabllity parameters v-33
12 Jan 1984
MIL-HDBK-472
Page
APPENDIX A TIME STANDARDS A-V-1
APPENDIX B ESTIMATES OF M.,. (0) FOR LOGNORMAL REfAIR

DISTRIBUTIONS B-V-1
APPENDIX C ACCURATE COMPUTER M-.. (Q) ANALYSIS c-v-1
APPENDIX D ACCURATE MAN.UALMm., (0) ANALYSIS D-V-1
BIBLIOGRAPHY 6-1
12 Jan 19B4
MiL-HDBK-472
-
LIST OF ILLUSTRA~IONS
Table Table Heading Page
1-1 List o z Categories and Elemental Activities of Active

Repsr Time 1-13
1.2 ~it~e~ Di;tri~ution~ for ~o~pletion Times for [lementa~
Activities 1-15
l-~o
1-3 Synthesis of Time Distributions
1-4 Parameters of Distributions of Elemental Activites 1-24
1-5 ProDosed Forma for Determining the Cumulative Time
Distribution
1-6 Probablllty of Completing Elemental Activity of final
Malfunction est by Designated Time 1-2B
1-7 Combining Matr x for Active Repair Time Synthesis 1-30
1-8 Observed Distr butions of System Final Test Time and
System Logfstic Time 1-32
1-9 Combining Matrix for System Down Time Synthesis 1-34
1-10 Occurrence Probability Functions for the Elemental
Activities Al-38
1-11 System Readouts Al-43
2-1 Average Parts failures per Part per 106 Hours by part
Category
2-2 Halnterlance T&5k TIRej
2-3 Interchange Time
2-4 Element Times Based on the Nork Factor System
2-5 An Example of the Use of Element Times for Determining
Interchange Time 2-25
3-1 Part Class Failure Distribution and Sample Size 3-17

3-2 Task Prediction, R-7801 3-23
3-3 Maintenance Analysis, R-7801/FpS-20 3-24
V-I MTTR Elements for Prediction Procedure V v-5
A-V-I Elemental Maintenance Actions A-V-2

thru
A-V-5
A-V-II Common Maintenance Tasks A-v-6
B-V-I 60th, 70th and 80th Percentiles of the Lognormal B-V-:

Dlstributlon for Means and Sigmas from .1 to 2.6 thru
B-V-i4
-
MI L-HDBK-472
Table Table Heading Page
B-V-I; 90th, 95th and 99th Percentiles of the Lognorma~ B-V-15

Distribution for Means and Sigmas from .1 to 2.6 thru
B-V-27
c-v-I Llstlng for Computer Program to Compute M.,, (0) Hhen

Elemental Maintenance Activities are Normally
Distributed C-V-6
C-V-II Sample Input/Output Data for M.,., (O) Computer progrm c-v-7
D-V-I Prediction Example Input Data D-v-8

D-V-II Prediction Example Manual Calculation of Mm,. (0)
(0= 0.90) D-V-9
L
MIL-HDBK-472
LIST Of ILLUSTRATIONS
Figure Figure Heading Page
Mnparison Matrix of Maintainability prediction Procedures 4

Distribution of Malfunction Active Repair Time for the
AN/hsB-4 1-5
-2 Di~t~{butlOn of 5ystem Down Time for the AN/ASB 4 1-6
!-3 Distribution of Malfunction Active Repair Time for the
AN/ARC-34 1-7
Distribution of System Down Time for the AN/ARC-34 1-8
Di~trjbution of ~a~f~n~tion Active Repair Time for the MD-1 ~_9
Jistributlofi of System OowriTime for the MD-1 1-!(J
Structure of Time Elements in Fix of Malfunction 1-11
Distributions of Times Required to Perform Repair Actions 1-18
Resultants Synthesized from t., tb and tC 1-!3
Probability of Occurrence of System Logistic Time 1-33
2-1 Functional Level Diagram of a Typical Communications System 2-13

2-2 Horksheet for Final Design Malntainabi]iy Prediction 2-16
2-3 ldorksbeet A 2-32
2-4 Horkshee! B 2-33
2-5 Norksheet C 2-34
2-6 Morksheet D 2-35
3-1 Comparison of Population and Sampling Mean Distributions 3-8

3-2 Sample Size Nomograph 3-1o
Maintainability Prediction Form 3-19
;:: Maintenance Analysis Continuation Sheet 3-20
3-5 Maintenance Diagram AN/GRT-3 3-21
3-6 hNIFPS-20, Tr&nsmltt!ng System 3-26
3-7 Amplifier-Mixer, AM-1347/FPS-20 3-27
3-8 Expanded View, Plate Power Supply 3-28
3-9 Plate Supply, Block Diagram 3-29
3-10 Nomograph - Downtime 3-32
4-1 Procedural Flow Block Diagram 4-6

4-2 Related Constraint Matrix 4-14
4-3 End Item/Corrective Maintenance Action Matrtx 4-15
4-4 End Item/Preventive Maintenance ActIon Matrix 4-16
4-5 End Item/Operational Functional Matrix 4-17
t41L-liDaK-472
.-
LIST OF ILLUSTRATIONS (Continued)
~ Page
Figure gure Head ng
v-1 RI Data Analysis Sheet - A V-13

v-2 RI Data Analysis Sheet - B v-t
4
v-3 MTTR Submodels V-16
Definitions of MTTR Submodels Terms V-17
v-4
v-s Matrix for correlating FD&I Features uith Ris v-25
V-6 Fault Isolation Output and RI Correlation Tree V-27
V-7 Manual Fault Isolation Output and RI Correlation tree
(partial) \-28
v-8 Sample: Maintenance Flow Diagram V-30
v-9 Example Time Synthesis Analysis V-31
v-lo Maintenance Correlation Matrix format V-32
A-V-1 Standard Screws A-V-7

A-V-2 Hex or Allen Set Screws A-V-7
A-V-3 Capt~ve Screws A-v-8
A-V-4 DZUS Fasteners A-v-8
A-V-5 Tridair Fastener A-V-9
A-v-6 Thumbscrews A-V-9
A-V-7 Machine Screws A-V-IO
-. A-V-1O
A-v-8 Nuts or Bolts
A-V-9 Retaining R~ngs A-V-11
A_iJ_i~ Drawhook Latch A,-fl-l!
A-v-1 1 Spring Clip Catch A-V-12
A-V-12 Butterfly Latch A-V-12
A-V-13 ATR Latch A-V-12
A-V-14 Lift and Turn Latch A-V-13
A-V-15 Slide Lock Latch A-V-13
A-V-16 Terminal Posts ConnectIons A-V-13
A-V-I7 Screw Terminal Connections A-V-14
A-V-18 Termipoint Connections A-V-14
A-V-19 H!rewrap Connection A-V-15
A-V-20 Taperpln Connection A-V-15
A-V-21 PCB Connections A-V-15
A-V-22 BNC Connectors A-V-16
A-V-23 Quick Release Coax Connectors A-V-16
A-V-24 Frict!on Locking Connector A-V-16
A-V-25 Frtctlon Locking Connector with Jackscrew A-V-17
A-V-26 Threadlocklng Connector A-V- 17
A-V-27 Slide Locking Connector A-V-17
12 Jan 1984
MIL-HDBK-472
-
L;ST OF ILLUSTRATIONS
f igure Figure Heading Page
A-v-28 Dip ICS A-V-18

A-v-29 Guided CCAS A-V-18
A-V-30 Guided CCAS Hith a Tool A-v-18
A-V-31 tion-GuidedCCAS A-V-19
A-V-32 Modules A-V-19
A-V-33 Crimp Lugs A-V-19
A-V-3.4 Form Leads A-V-20
A-V-35 Soldeing Terminal posts A-V-20
A-V-36 Soldering PCB Connections A-V-20
A-V-37 Resoldering Mitt a Braided Wick A-V-21
A-V-38 Resoldering Using a Vacuum A-V-21
A-V-39 Form Flat Pack Leads A-V-21
A-V-40 Panels, Doors ana Covers A-V-22
A-V-41 Drawers A-V-22
A-V-42 Display Lamps A-V-22
c-v-1 Resulting Repair Time Distribution for a Sample System

Containing Four (4) Repair Types with Different
Normally Distributed Repair Times C-V-8
12 Jan 1984
~iv
M]L-HDBK-4~2
INTRODUCTION
THE NEED FOR WINTAINA6!LITY PREDICTION: The wediction of the expected

number of hours tha: a system Or device will be in an inoperative or down
state while it is undegc;rg maintenance is of kital importance to the user
because of the adverse effect that excessive downtime has on mission success.
Tnerefcre, once the operational requirements of a system are fixed, it iS
imperative that a technique be utilized to predict its maintainability in
quantitative terms as early as possible during the design phase. This
prediction shouid be updated continuously as the design progresses to assure a
high probability of compliance with specified requirements.
A significant advantage of using a maintainability prediction procedure is

that it highl)ghls for the designer those areas of pwr maintainability which
justify product improvement, modiflcatlon, or a change of design. Another
useful feature of maintainabilityprediction is that it permits the user to
make an early assessment of whether the predicted downtime, the quality,
quantity of personnel, tools and test equipment are adequate and consistent
with the needs of system operational requirements.
DEFINITION OF MAINTAINABILITY: MIL-STt)-721cieftnesmaintainability as follows:
Maintainability: The measure of the ability of an

Item k be retained in or restored to specific
conditions when maintenance is performed by personnel
having specified skill levels, using prescribed
procedures and resources at each prescribed level of
maintenance and repair.
This definition has fostered the development of many maintainability

prediction procedures for providing an assessment of system maintainability.
Each of these uses various quantitative measures to indicate system
mainta~nabllity. However, all of these measures have a speciftc relationship
to, or constitute some element of the distribution of total system downtime,
Hence, if a universal method or technique can be developed to determine the
Total System Downtime Distribution for any type of system, this would
facilitate calculat~ng the measures of maintainability currently In use.
HI1-HDBK-472
BASIC ASSUWTIONS AND iNTERPRETAT10N5: Each maintainability prediction

procedure inc?uded in this handbook deDends upon the use of recorded
reliability and maintainability data and experience which have been obtained
from comparable systems and components unoer similar conditions of use and
operation, It is a~so customary to assume the applicability of the principle
of transferability. This assures that data which accumulate from one system
can be used to predict the maintainability of a comparable system which is
undergoing design, development, or study. This procedure is just~fiable when
the required degree of commonality between systems can be established.
Usually during the early design ohase of the life cycle, consnonalitycan only
be inferred on a broad bas~s. However, as the design becomes refined, during
iater phases of the life cycle, commonality i~ expendable if a high positive
correlation is established relating to equipment functions, to maintenancetask
times, and to levels of maintenance. Although the techniques contained in
this handbook have been proposed and appear to fit certain applications, it
should be borne in mind that they have not truly been tested for generality,
for consistency one to another, or for most other criteria dealing with broad
applicab!llty. It should also be borne in mind, though, that experience has
shown that the advantages greatly outweigh the burden of making a prediction.
For that ~eason, it ~s not the Durpose of this document to deter further
research or Inquiry.
ELEMENTS OF MAINTAINABILITY PREDICTION TECHNIQUES: Each mainainabillty

prediction technique utilizes procedures which are specifical y designed to
satisfy Its method of application. However, all maintainabilty prediction
methods are dependent upon at least two basic parameters:
(a) Failure rates of components at the specific assembly level of

interest.
(b) Repair time required at the maintenance level involved.
There are many sources which record the failure rate of parts as a function of
use and environment. This failure rate, is expressed as the number of
failures per unit of time. A typical measure Is failures per 10C hours.
The majoradvantage of us!ng the failure rate in malntafnablity prediction
calculations !s that lt provides an estimate of the relative frequency of
failure of those components whfch are utillzed in the des~gn. Sim\larly, the
relative frequency of failure of components at other maintainable levels can
be determined by employing standard reliability prediction techniques using
parts failure rates. Failure rates can also be utilized in applicable
regression equations for calculating the maintenance actfon time, Another use
of the failure rates is to weight the repair times for various categories of
repair activity, in order to provide an estimate of its contribution, to the
total maintenance time.
12 Jan 1984
MIL-HDBK-472
24 :NY 1966
Repair times are determined from prior experience, simulation of repair tasks,
or past data secured from similar applications. Most procedure break up the
maintmance action, which is a more general expression than repair action,
into a number of basic maintenance tasks whose time of performance is summed to
obtain the total time for the maintenance action.
SUMMARY: It is emphasized that the selection and application of the proper

mahtatibjlity technique results in many economies measured in terms of
man-hours, materiel, and money. These Savingsare attributable to the fact
that maintalnabil!ty prediction is considered to be a tool for des@n enhancement
because it provkkm for the early recopdtion and elimination of areas of poor
maintainability during the early etages of the design life cycle. Othemlse, areas
of poor maintainabili~ would only become apparent during demonstration testing
or actual use, after which time, correction of design defjchnciee would be
cxmtly and uodd y delay schedules end miseions.
Maintainability prediction, therefore, is a most useful instrument to both manager

aod engineer becmse itprovideB for improved system effectiveness and reduces
administratiw end maintenance coots.
.
The comparison Figuru A, (s included to provide ummary of the signifl-
matxix,
oant8ttrfbutesof each m ahtafn@tlityprediction procedure included fn this hand-
book Additional detailo may he obtained by referrfng to specfftc msintatnability
prediction procedure of Merest.
3
MIL-HDBK-472
ML-HDBK-472
,.
w 9
.
12 Jan 1984
MIL-tiDBK-472
PROCEDURE V
1. GENERAL
This procedure can be used to predict maintainability paranete~S of aviOItiCS,

ground and shipboard electronics at the organizational, intermediate and depot
levels of maintenance. It can also be auplied to any application environment
and type of equipment including mechanical equi~ment. Time standards are for
electronics only.
1,] philosophy and assumptions. The maintainability Procedures presented

oere permit analyzing the maintainability of electronic eCIUiCNM?tIt/SyStWtIS
iwliiding direit d[c~n:dbility Zf d!agno;ticjlisolation/test capabilities,
replaceable tefnconstruction,-(a replaceable item (RI) is any of those
physical ent ties normally removed and replaced to effect repair at the
maintenance evel for which the prediction IS being made) Dackaging, and
component fa lure rates, in addition, the following assumptions and
stipulations apply to any predictions made using the procedures given here:
a. Fa lure rates experienced are all in the same proportion to those

predicted.
b. Ony one failure at a time is considered.
c. Mantenance is performed in accordance with established rna!ntenance

procedures.
d. Maintenance is performed by maintainers possessing the appropriate

skills and training.
e. Only active maintenance time is addressed; administrative and

logistlc delays, and clean-up are excluded.
Two separate methods are presented. Method A is an early prediction method

that makes use of estimated design data and can be applied nwch earlier than
Method B In the development of an equipment or system. Method B Is a detailed
prediction method that uses actual detailed design data to predict
maintainability parameters.
The application of the procedures presented here permits the user to monitor
the overall system maintainability throughout the design and development of
that system. The user can identify whether or not the specified
maintainability design requirements will be met before the system Is
complete. Thus, if it appears the maintainab!llty requirements will not be
met, the designers can be informed and the necessary changes can be made
before they become prohibit~vely expensive.
1.2 Point of application. Both of the prediction methods (Method A is the

early prediction and Method B Is the detailed predictim) of this procedure
can be applied at any equipment or system level, at any level of maintenance,
and for any maintenance concept pertinent to avtonics, ground electron~cs, and
shipboard electronics. (ldhllethe prediction methods were developed
MIL-HDBK-472
specif~cal:y foelectronic equipments and systems, there is nothing inherent

in the methods t?ia: shoul~ preVefIt th?fr from being applicable to eiect~o-
mecnancal or me:nanical equipments or s~stems.)
1.] Basic parameters of measure. Mean time to repair (MTTR) ?S the primarj
maintainability Darameter that can be predicted using this procedure. The
otner maintainability parameters that can be D~edictetjusing tnis procedure
arp mailmum ~orre~ti~~ maintenance time at tne o Percentile (M... (Q)),
Dercent of fau!ts solatatle to a single replaceable item (I ; per:ent ~f
faults isolatable to <N replaceable items (1,,),mean mdintenan:e manhours
per repair (EiFi/repai;), mean maintenance mannours per operating hour
(~/0~), mean maintenance manhours per fllght hour (WH/FH), (For details
see DaragraDh 3.2 )
1.4 Information requireq. These data items must be Drovlded as part of the
mainta~~ Dredlction if they are not provided from another source. (See
MIL-STD-756 and MIL-STD-1629.)
1.4.1 Method
. A. To use Method A the following data are necessary:
z. The n[]mhorand contents of either actual or est mated) the pr mary

replaceable it~~;
b. Tne fa)iu e rates, either predictea or estimated associate wth

each replaceable item.
c. The basic fault isolation test strategy of each replaceable Item.
d. The replacement K!ncept, if fault isolation is to a group of

replaceable items.
e. The packag!ng philosophy
f. The fault isolation resolution, either estimated or required (I.e., %

of faults isolated to one replaceable item or the average replaceable item
group size).
1.4.2 Method B. The data necessary to implement Method B are:
a. The replacement concept for each replaceable ~tem or group of items.
b. The fault detection and isolation outputs associated with each

repaceable item.
c. The failure rate of each replaceable item.
d. The maintenance procedure that is followed to remove and replace each

rep aceable item.
.-
1; nan 1 C)GA
MIL-HDBK-472
2. REFERENCED DOCUMENTS
~.1 Issues of documents. The following documents of the issue in effect On

the date of invitation for btd or request for proDosa~, are references In this
Drocedure for information and guidance.
STANDARDS
Militarv
MIL-STD-721 Definitions of Terms for Reliability and

Mairttalnability
MIL-STD-756 Rellabiltty Modeling and Prediction
MI1-STD-1629 Procedures for Performing a Failure Mode,

Effects and Criticality Andlys!s
(Copies of specfflcatlons, standards, drawings, and publ~cat~ons required

by contractors in connection with specific procu~ement functions should be
obtained from the contracting activity or as directed by the contracting
office~.)
3. BASIC DEFINITIONS AND MODLLS
3.1 MTTR elements. Corrective maintenance actions conslstof the following

: %epa~at~on, Fa~lltIsolation, fault Correction (further broken down
tzlsk$
into Disassembly. InterchamJe, Reassembly, Allgnment, Checkout). The time to
R. Hence the task t~mes are
perform each of-these tasks-is an elemen~ of PIT
called MTTR elements.
The definitions for these MTTR elements and the r abbreviations re used in
the prediction models are as follohs.
MTTR ELEMENT (Abbrevlatton) DEFINITION
Preparation (TB ) Time associated with those tasks required

m~ to be performed before fau t Isolation
can be executed.
Fault Isolatlon (Tr, ) Time associated with those tasks required to

n~
isolate the fault to the level at uh~ch
fault correction begins.
Disassembly (Tb ) Time associated with gafning access

;
to the replaceable item or items
identified during the fault Isolation
process.
Interchange (T, ) Time associated with the removal and

fl~ replacement of a faulty replaceable
Item or suspected faulty Item,
Reassembly (T, ) Time associated with clostng up the

nJ equipment after interchange is performed.
-
MIL-HDflK-472
.
TABLE V-I. MTTR Elements for Prediction Procedure V.
T 1 Isolation to I Isolatlon to ! Isolation with Mlblguity T

IIsolation to I Group with Group IGroup with Iterative! (Requires Further I
lSingle RI I Replacement I Replacement I Isolation) I
I I ) I
I I I Secondary ~
~Preparation 1 Preparation ; Preparation { Preparation I Preparation {
I
I I 1 I !
I I secondary
lIsolation I Isolation I Isolation I Isolation I Isolation /
I I i i I I
1 i
lDisassembly ! Disassembly I Disass@nbly I Disassembly \ I
1 I I I I I
I I I
lInterchange [ !ntewhange I Interchange I Interchange ~ ;::~;dary I
I.- I I { I
1 Correction I
preassembly I Reassembly I Reassembly ~ Reassembly I I
1 I I I I I
1- 1 I
ignment I Alignment I Allgnment I Alignment I 1
I I I I
~rheckot I I 1 ! I
,>. ! Checkout ! Checkout ! Checkout ! {
I I I I I I
I
]Start Up I Start Up [ Start Up I Continue I Stati UP I
v-5 12 Jan 1394

MIL-HDBK-472
3.2.:. 1 Mean repair time for the n Ill.

-
J= number of unique FDLI outputs (refer to Daragraph 4.2.3)
A.l = failure rate of those parts of the n RI which would cause the
n RI to be called ou; in the jh FDL? output
R.:= average repair time of the n RI when called out in the j FD~I
output as computed below:
3.2.1.2 Average repair time for the n RI.
Mnj = number of steps to perform corrective maintenance when a failure

occurs in the n RI and results in the j FD&I outputs. Includes all
ma!ritenance elements preparation, $so~ation. ~t al This may include
operations on other Ris called out in the j fault isolation result. (e.g.
isolation to a group w~th iterative replacement).
T. = Average time to perform the mb corrective maintenance step for

n~ the nt RI given the j FD&I output-
3.2.2 Percent isolation to a single RI. The model for calculating the
percent Isolation to a s~ngle RI (1,) is:
K
s A,
k-l
1, = x 100
NJ
Mere. An, = failbre rate of those parts of the nh RI wh!ch wfauld cause
the n RI to be called out in the j FD&I output.
A, = failure rate associated with the k FD&I output which

results in Isolatlon to a single RI.
K= number of FD&I outputs which result in isolation tc a single

F/l
12 Jan 1984 V-6

MIL-HDBK-472
3.2.3 Pe*cen! isojation to a grodp of RIs The model for calculating the
percent ,solaticnto a group of N or less RIs (I.) !s:
P
I h.
x N= p= 1 x 100
NJ
Hhere ; AO = failure rate associated with the p FD&I output which

results in isolation to N or less RIs.
A e) = same as for I,
P. number of FD&I outputs which resdlt in IsolatlufitG ~ Or ~ess

RIs.
Other maintenance parameters that can be predicted using these procedures are
dS f01~OU5.
3.3 Mean
. maintenance manhours per repair (~/repair).
N
~ ?..MMH.
Filil/Repair. n=l
N
E 1.
n=1
Uhere:
N= quant!tyof RIs
1. = fai)ure rate of n RI
m. . mean maintenance manhours required to repair the n RI
3.3.1 Mean maintenance manhours required to repair the n RI.
12 Jan 1984
v-.
7
llIL-HDBK-472
J = quantity of FD&I results
1. = failure rate associated with the j result for the n R]
MMH.. = maintenance mannours requiled to repair the n- RI given the

J FD&J result
(Replace repair times in the apDroprate Method A and Method B procedures with
the maintenance manhcur~ required for each reoair action).
3.4 Mean maintenance manhours per maintenance action (~H/MA).
This is the same as ~/repair except that time spent as a result of sySt.em
failure fal;e alarms must also be fncluded in the maintenance manhours.
Two types of false alarms &i-econsidered:
:) Type 1 false alarm is detected during normal operations but cannot be

repeated during the fault ~solatlon process.
2) Type 2 false alarm is detected and isolated to an RI when the RI does not
have an actual fault.
F -n s frequency of occurrence of type 1 false alarms ~/
F 2n = frequency of occurrence of type 2 false alarms ~1
~1 expressed as a fraction of the nth RI failure rate
WHO = mean maintenance manhours associated with Type 1 false alarms.
M=. = mean maintenance manhours required to repair the n RI
12 Jan 1984
v-(!
MIL-WK-472
\ 3 5 False alarm rates False alarms are dependent on the system type,
oDeratin9 env~ro~mentt maintenance environment, svstem design and fault
detection and isolation implementation. Therefore, a standard set of false
alarm values would be impossible to d@rive. A sample of false alarm rates
e~perienced on tw~ 1978 vintage equipmen! are presented for reference
purpo$es.
$YS.TEM/EQJ?fWENT Type 1 (F,.)2 Type 2 (F2,)~
Heapon Contrcl System
Radar Subsystem .41 .25
Computer Subsystem .63 .65
Control Subsystem 1.32 .31
Power Subsystem .37 .66
Auxillary Subsystem 1.31 .54
Airborne Radar System
RF Unit NA .44
Transmitter NA .31
!?ecelver NA .12
9 Antenna NA .08
Analog Processor NA .07
Digital processor xl NA .65
Digital processor #2 NA .50
Control Unit NA !00
Power & Antenna Servo NA .33
NA - Not Available
~1 The ratio of type 1 false alarms to actual failures

~/ The ratto of type 2 false alarms to actual failures
12 .hn 1%4
v-9
MIL-IiDBK-472
3.6 Mean maintenance manhours per operating hour (~/OH).
This includes the entire manpower that IS required to maintain a system;

corrective maintenance, preventive maintenance, and maintenance caused by
false alarms.
N N PM
2. = ),.expressed in failures per operating hour
F. = frequency of r preventive maintenance
m, = mean maintenance manhours to perform r preventive maintenance type
PM = quantity of unique preventive maintenance types
N = quantity of RIs
F In s frequency of occurrence of type 1 false alarms 4/
F2. . frequency of occurrence of type 2 false alarms ~/
PIMHO = mean maintenance manhours associated with type 1 false alarms
~. = mean ma.lntenance manhours required to repair the nth RI
~1 expressed as a fraction of the n RI failure rate.
3.7 Mean maintenance manhours per flight hour (~/FH). This is the same as
~/OH where k = X. is expressed in failures per fllght hour.
3.8 Maximum corrective maintenance time for the (0) percentile (M.ba
(0)). Two Mma. (0) models are provided. The first yields an
~oximate value and requires that system repair times be lognormally
dist~!buted. The second gives a more accurate value.
3.8.1 Approximate Mm., (0), Appendfx B contains tables of M_a, (0)

values for selected values of system 0, system MTTR (MEAN), and standard
deviation of system repair t:me~ (SIGMA) .- !4TTR may be predicted using Method
A, and the MTTR models in paragraph 3.2.1 above. SIGMA is usually determined
mm data on similar equipments. Approximate Mm., (41)values for values
of 4, MEAN, and SIGMA not covered in Appendix B may be calculated by using
the following equation:
i-o
t4:L-HDBK-472
Mma. ($) = exp [log 14TTR+ OSIGMAJ
N
klhere: SIGMA = r.(log R,,; - [(! logR. ; i NJ
i=l
d
i=i

N-1
3.8.2 Accurate ti~,.(@). Append~~ C presentsa c~puterlzed means Of

pred{ctlng an accurate M.,, @. AppenciizD presents a manual calcu?atlon
method for predicting an accurate M.,. (0).
4. APPLICATION. The application of the early and detailed maintainability

prediction techniques is described in 4.1 and 4.2 respectively.
4.1 Method A - early P redlc~ton procedure. This sectiofiprovides a

step-by-step procedure for performing an early prediction of mean tire?to
repair. The tasks involved in performing the early prediction are:
a. Define the prediction requirements.
b Define the replacement concept.
c. Determine the prediction parameters.
d. Select the appropriate models.
e. Compute the 14TTR(or other parameters).
Descriptions of each of these tasks are provided In the following subsections.
4.1.1 Prediction requirement definition. Th!s step of the prediction is in

some resDects the most imDortant as~ecc since it establishes a common baseline
of understanding the predtcttonpurpose, approach and scope. During this
step, the malntalnablllty parameter(s) to be evaluated Is defined, the
prediction ground rules are established. and the maintenance level for uhlch
the prediction is being made is defined.
Parameter definition Includes the selection (if required) or the parameter(s)

to be evaluated and the establishment of a qualitative and quantitative
definition of each parameter. If the prediction is being performed in
compliance ulth a customer statement of work defining the parameter to be
analyzed, it must be determined if the stated parameter is consistent with an
equivalent parameter contained in this methodology. if not, the prediction
mooels must be changed accordingly, As part of the parameter evaluation, It
must be determined uhlch elemental maintenance tasks (e.g., preparation,
isolation, etc.) are to be included in the analysis and which are to be
exciuoed.
The last aspect of th!s step is to explicitly define the maintenance level for
which the prediction is being made. If the level is defined in terms of a
specific maintenance organization (e.g., direct support unit, depot, etc.),
then the tasks to be performed are readily defined by the maintenanceconcept
as described in the following section If the level is deftned by operating
level or location {e.g., on-site, flight-line. etc.). then this level must be
redefined In terms of the maintenance organization(s) performing maintenance
at the leVe~/10CatiOn.
12 Jan 1984
v-~1
MIL-HDBK-472
4.1.2 Replacement concept definition.

The maintenance concept must be -
estab shed, so that in conjunctionwith a definitionof the prediction
requirements (Daragraph 4.1.1), a basellne is established which defines the
prediction to be performed. Hith respect to the maintatnabillty prediction,
the primary ouput of the maintenance concept is the definition of how a repair
is effected and what the replaceable items are.
As part of this process, a complete set of replaceable items is identified.

If the maintenance concept allows for fault isolatlon to a group of !?1sand
repair by group replacement, then the RI groups can be reclassified as RIs ~f
each of the groups isindependent of other groups.
4.1.3 Determination of the prediction parameters. This step involves:
a. Def\ning theRXs.
b. Determining the failure rate (predicted or est!mated) associated with

each RI.
c. Defining fault lsolat~on test methodology for each RI.
d. Defining the replacement concept.
e. Defining the packaging philosophy.
f. Determining the estimated or required fault Isolation resolution

(i.e., X% to 1 RI or average RI group size).
Forms slmllar to those in Figures V-1 and V-2 should be used for the data
collection process. Data is collected on these forms at the level for which
predictions are performed. For example, If a repair time Is to be computed
for every equipment wlth{n a system, then a seperate data collection form
should be used for each equipment. Data should be tabulatd as follows:
F!rst tabulate al? the primary Rls and their associated failure rates
In tie respective columns of Figure V-1.
b. Next describe all methods (V) for performing each elemental actlvlty
(m) in Figure V-2. (Mote that some maintenance act!ons do not require that
all the maintenance elements be fncluded).
c. Next enter the appropriate number of headings (V.) for each

elemental activity along the top of Flgwe V-1.
d. For each elemental activity, (m,v) synthesize times (T-.) using

times, selected in accordance w~th paragraph 4.2.6 and note them In the
respective column of figure V-2.
e. Next enter the associated failure rate of each RI for the elemental
activity that It pertains to in f~gure V-1.
12 Jan 1984
V-12
\
l:!
+>,
I c
b.- 1
)
.-
Jan 1984
I I
; Tm ; ~m ;
I th v
\ !4TTRElement (m) ~ V ~ Description of the v Method ; v I
I
I
~--- (
o
I I
I
I I
I
(
I I
I ~
I ~ I
,
1
I
I
I
I 1
I I
I (For additional data, see fi12
( of bibliograpl~y.)
I I
i
I I
t
I
I
,
I I
I
( I
1 I I I
I
(
,
I
I 1
( 1 I
I
I
I
1
1
I
$
t I
I I I
I
I I I
I I
.. . . ---- .. .- --- . -_l-
FIGURE Y-2. P.1Oata Analvsis Sheet - B

12 Jan 1984
These completed data sheets provide the bas!s for the early prediction
technioue. Once they are complete the subnmdels can be applied.
4.1.4 General prediction model and submodel selection.
The general form of the prediction mooel is:
Hhere:
70 = Average Preparation lime

7,, = Average fault Isolation Time
T, = Average Disassembly Time

.-
1, = Average Interchange Time
1. = Average Reassembly Time
7. = Average Alignment Time

. 7 c = Average Checkout ?ime
~~, . Average Startup Time
7. = average time of the m element of MTTR.
Variations of the model are limited to deleting the time elements for
elemental activity terms that are not necessary to complete certain
maintenance actions.
The selection of submodels Is dependent on the replacement policy imposed.

The appropriate $ubmodels for computing the average time for the above
elemental activities are given in Figures V-3 and V-4,
The sutmdels presented are of a general form and can generally be applied to
any e~uipment level (i.e.a systgm, subsystem, equipment, etc.). The only
limitation being that if SC or S, are computed, the prediction level must
be consistent with the RI grouping rules presented in paragraph 4.1.5.1.
Otherwise, the elemental activity submodels are applied at the lowest leve~
for which an MTTR prediction is desired.
4.1.5 MTTR computation. The MTTR is computed at the level al which the
average number of RIs contained in a fault isolation resul~ (S) or the
average number of iterations gquireg to correct a fault (S,) is
established. For example, if S, or SC can be estimated for each equipment
within ? system, the~ ~$e lowe$t ~eve! ~h~~ ~h~ U?Tp can be p~ed~cte~ ~j ~;,~
equipment level. Higher level predictions of MTTR, such as system Ievei MT~R,
can be computed by taking a failure rate weighted average of the equipment
MTTRs within the system.
:2 Jan :93{
-
.. .
. *O L
b. .W
- .J 1-
.-1 -T
-
,,
.-,
.L-
-
I
.--11..- --b
z
,t-<
-
--- >
- : .-
----
. I
-- . -. ii .--i
j, >--~ n
.-
G
9--
s +CJ
-. .
- - -,. ,.
.9 -
- . .
t-i
-l -
.+- 1 -.-i
,:1- .
4- . ~ .
e

,
* 8
- *fl
i41L-)@BK-472
th
%w -
urnerqu~re~
toprepare
* system for laul! Iso~atIon using the v mrthud
th
t)me rqu!red to Isolatc a fdl using the v mclhod .
% -
111
ttm~ rqulred 10 pcrlorm d16as$embly urJin& the Y method
Tr) -
v
th
- bmt rqw~d to Interchmge m! RI using the v methcd
1
v
th
7A - nmc required m lign or cslibrste sin i?l using the w method
v
!h
time required to check a rep$tir USIISC the v method
c\
M shrtUPn ~vw~mu~m~tlwv[h method
ST, - time raqtimd
kP, - ~aliure
rate
ofItls
~s~oc{sted
withtheVthrnethd of perforrmng preps mlloo
th
F] - failure mte of flls associated with the v method of performing fault
v i301stion
J Dv
- failure rete of Rlstssocimted
wtth the Vti method of petiormtng dlss~eemb}y
Ln - faflure rate of RIstmroclafecf w-M the v ti methcd of performing rets*embly

r
1 v - failure rate of Ills sa-~ated

wl~htheVthme~trod
of perforMhIK Interchame
hA - failure mtc of with(heVm me-

RIsassociated ofperforming slignment
v
cv - failure mte of Rlsaseocmled with the V* method of performing checkout
th
k STV - tsllure raw of IUSassocmtad with the v melhod d ~tlo) IItlng start-up
P - the number of unique ways to perform preparation
th? number of unlquc wrIys to perform (suit lsolatton

rl
\ - the number of unique wsyri to perform disassembly
u
\
R - the number of unique ways m perform reassembly
vI - the number of unlqus umys to perform Merchsnge
- the number of uniqur ways to perform alignment

A
- the number of untque weys to perform check-out
c
- thenumber of untque ways to perform wrt-up
ST
5C - the vem~e number of M corIt8Mted In hull Ieoblkon resull
E, . the s~rage number of tnbrchs~es rsqulred to correct fsull
A - the number of untque ccweees (A s VD or VR)
7 - the verage number of unique accesses required per fault !sofatlon result
- the (allure rate of We M UtU r~ulre the S w type of Sccess
- the totAl system failure mte
th
- the tl
MC requl red to dlesssemble the B ccess
Dm
.-. - the tkme ~utred to rasnemble the a
th cccesa
TRS
L
Flcuu v-l). ~
L- 1-
t41L-Hi)BK-472
Computation of repair times below the level at which S(I or G) is established

m?y result fr an inaccurate account of repair times. The only exception to
this is jf fault isolation is down to a single RI (< = 1) for the entire
system, equ~pment ...), then the MTTR may be computed at any level since
ambiguities between RIs do not exist. Otherwise, the fcllowing criteria must
be followed:
!n orde; to compute a repair time at a given level, a value for 3 (1 or G)

must be established at that level. After the level at which the repair times
will be computed has been selected, the appropriate models are selected to
compute time for each elemental activity at that level with higher Ieve
repair imes being computed using a failure rate weighted average.
Values for $C, ~,, ~, ~~ or ~~, where required, should be comptited

as detailed in the following subsections.
4.1.5.1 Method of computing 56 and %. Two methods are oresented for

computing the dverage number of RIs in a fault isolation rgsult (SC) an~ the
average number of iterations required to correct a fault (S,), compute S (1
or G) using the specified or design requirements, or compute S (1 or G) by
assessing the approximate fault isolation capabilities of the system or
equipment.
4.1.5.1.1 Method 1. The first method of computing $, or :6 depends upon

how the fau-tion requirements are specified. In the fault Isolation
resolution is specified as follows:
X,% to~ N, RIs
h. RI~ < X2%to ~ Nz RIs
N* RI~ ~ X,% to s N] RIs
and X, + X2 + X, = 100
then,
~=x (N; ]+ X2( N1+N; +:) .X, (N2+N3;1 )
1~
If the fault !solation requ~rements are spec~fied as follows:
x,% to: N, RIs
x,x to~ N2 RIs
1002 to:ti, I?ls
Hhere X,% c X22 c 1001
i~ idli ;?i+~
MIL.-HDBK-472
~hpr,
N,+i \ ( N, + ti, + )
=
x
( 2
+ (X2-X.)
\
]o~
2 ) (100-%2)( +:+)
pre~i::p~ L!TTR u;ing this method of computing ~ is based orIthe assumption

the specified faut isolatlon requirements have been (or will be) met,
resulting pred ction is the inherent MTTR that will be realized by
eving the spec fied requirements. This approach is valuable during the
ea-;y stages of equiDment development for purposes of allocation and
assessment of the requirements facility. This &ppro&ch should tIotbe used
when data is available on the actual fault lsolaticm characteristics.
4.I.5.I.2 Method
2. The second method of computing ~, Or <G Involves an
analysis of the fau~t isolation characteristics of the subject
equipment/system as fJllows:
a. P-epare a simple block diagram depicting the system and how each
major function is related (i.e., show functional interfaces).
b. Group the functions (RIs) into G RI sets such that:

an estimate of the fault tsolatlon (number of RIs) can be

determined for each RI set.
each RI $et is independent of any ~ther RI $et.
each RI set established is the smallest set that can be

estab ished.
c For each RI set (g) estimate the average fau t Isolation resolution
or th~ average number of RIs per fault isolation resu t depending on the
replacement philosophy in question ~~ ,,Q if iterative replacement,
>(6,0 if group replacement).
d. Compute the average $, or ~, for the system or equipment using a

failure rate weighted model.
If the repail times are computed at lower levels, then the overall ~ does not
have to bp computed.
4 1.5.2 Computatj~~Of
.. and ~o, The average ni~rntwr of accessg:
~ ~r~___
(disassembles and reassemblies) requi~ed per fault isolation result (A) can
be comDuted as follows
12 Jdn i98:
MIL-HD8K-472
I 1,
g-l
and,
1
Mere:
~q = the average number of accesses required per fault Isolatlon result

In the g RI set, (G RI sets established the same uay as was
done for S)
F 98 = the probability that the a acces~ will be required for any

random fault Isolation result
AQ = the number of untque accesses in the g RI set
A9 - the fa~lure rate of the RIs located In the g RI set
1 qb = the failure rate of the RIs located in the a access location

of the g RI set
~@ = average number of RIs per fault lsolatlon result for the g~ RI

set
The computation of ~D and ~u is exactly Ilke the method used for ~ wtth
one modification. Each probability !s multiplied by its appropriate
disassembly for reassembly time. The equation for ?D or ~, 1s:
G
ML-HDBK-472
Nhere:
The same equatfons ~lso hold true for reassembly, (~~ )

9
where:
T~w. = the disassembly or reassembly time for the a access of the

~ h RI set.
Note here also that if the RIs are grouped Into Just one set instead of G
sets, then all the subscripts g will fall-out and the failure rate tieighting,
of the g RI sets Is not necessary.
4.1,5.3 Determination of ?4TTR. The MTTR can nou be computed by sunning up

the average times computed from each submdel. Thus, the MTTR !s expressed as
M
MTTR = I 7.
m-l
,,-, . . . . .
HIL-HDBK-472
If the repair time computed is for a lower level then the higher level repair
times are cc+nputedas follows.
)4TTR
MTTR. - mean repsr time of the b lower evel
10 = failure rate of the b lower level
B - quantity of lower level breakdowns.
4,2 Method B - Deta!led prediction procedure. Thts section provides a step

by step ~rocedure for performing a deta! led prediction of mean tfme to repair
(i4TTR).The tasks Involved in performing the prediction are:
a. Def!ne the prediction requirements.
b. Define the replacement concept.
c. Identify the fault detection and Isolation outputs (FIXI outputs).
d. Correlate the FD&X outputs and hardware features.
e. Correlate replaceable items and fault detection and isolatfonoutput$.
f. Prepare a maintenance flow diagram.
~ Prepare time line analyses.
h Compute the maintafnab{llty parameters.
Descr pt ons of each of the tasks are provided in the followlng subsections.
4.2.1 Prediction requirements def!nltion. This step !S similar to that

requ!red for an early predtctlon; refer to paragraph 4.1.1.
4,2.2 Replacement concept definition. This step is sim}lar to that required

for an early pred~ction; refer to paragraph 4.1.2.
v-2?
MIL-1411BK-472
4.2.3 Fault detection and !solation output identification. Th~s step

involves the Identification of all the outputs which are used in the fault
detection and isolation process. Normally, the fault detection and isolation
processes are segregated. However, for purposes of malntalnabillty
prediction, the fault detection methodology is considered as the first step of
fault isolation and is properly included as part of the Isolation capability.
Any time associated ulth fault detection (e.g., mean fault detectton time) is
normall~ excluded from the prediction model, but can be included If desired.
The term fault detection and isolation outputs is defined as those

indications, symptoms, printouts, readouts, or the results of manual
procedures which separately or in combination, Identify to the maintenance
technician the procedure to be followed.
FD&I outputs wtll vary \n form, format, c~]exity and duta content from
system to system and some uIII be more obvious than others. The maintenance
actions taken in response to these outputs may depend upon the system
maintenance environment and the system operating CritlCa]ltV. It is
important, therefore, not only to Identify the FD&I outputs but also to ensure
that the FD&I outputs identified are the ones that wII1 be used In the
intended maintenance environment.
Some of the more common generic FD&I outputs are:
a. Indicator or annunciator.
b. Diagnostic or BIT output.
c. Meter readings.
d. C)rcuit breaker and fuse Indicators.
e. Display presentation.
f. Alarms.
9. ~mprope~-system operation.
h. Improper system response.
,. System operating alerts.
o iDply the oreoiction methodology presented herein, the predictor should

f!rs~ !dentlfv all primary unique outputs upon which the maintenance
technician relies to make decisions on the repair methodology (e.g., perform
adjustment. rer)laceRI, proceed to a different method of fault isolation,
etc.). Secondary outputs should then be identified for those cases where the
primary output yielded a result which dld not correct the problem and further
isolation is reQured.
,-,
---- a
MIL-tiDBK-472
4.2.4 FD&I outputs and hardware correlation. The key to this predictto~
procedure, and by far the most demanding of the prediction tasks, iS the
establishment of a correlation between the FD&I outputs (See paragraDh 4.L.7)
and the hardware for wnlch the prediction is being made. This step jemands a
thorough understanding of the system hardware, software, monitoring and
diagnostic capabll~ties, and of the FD&I features Inherent to the system,
Ft)&Ifeatures are those hardware and software elements, or comb~nat~ons
thereof, which generate or cause to be generated each FO&I output.
his task can be accomplished either from the top down or bottom bp. The top
down app~oach ~nvolves a fault tree techniaue where the top of the tree s
each unique FD&] output; the next tier identifies the FCXI feature(s) which
can yield the subject output: and, the bottom tier Identifies the RIs or
partial RIs which upon fa!~ure would be detected or isolated by the subject
~%: feature. The bottw up approach involves identification of all the
circuitry in terms of RIs associated with each FD&I feature, and the analysts
of how a failure of each RI presents itself in terms of a FD&l output.
Either approach requires the same five steps to be performed:
a. Identify all FD&I features.
b. Ident~fy the circuitry associated with each feature.
c. Identify the fD&I sequencing.
d. Establish the RI failure rate associated with each FD&I featUre.
e. correlate the FDLI features with the FD&I outputs.
FD&I features are those hardware and software elements, or combinations

thereof, which generate or cause to be generated each FD&I output. Typical
features include diagnostic program routines, BIT routines, BITE, performance
monitoring programs. status monitors, and test points.
After the FD&I features are icientlfied,the circuit schematics are analyzed to
ic!entlfythe components tested or verified by each feature. The outputs of
this analysis are then translated into a matrix as shown in Figure V-5. The
matrix Identifies, for each FD&I feature, the RIs and canponents which are
tested by that feature. Also included in the matrix is an identifier which
defines the order in which the FD&I features are utillzed during the iSOlatiOn
process.
12 Jan 1984
v-24
UIL-HDBK-472
The matrix is used to identify the failure rate of each RI associated with
each Fl)&Ifeature. The first FD&I feature Is examined and the failure rate of
each component associated with that feature \s entered in the matrix under
that feature. The second feature Is then examined. etc. If a component is
tested by mre than one feature, the failure rate is assigned to the first
feature which would result in a positive failure indication. If different
tests of the sane component check different failure nxx!es,then the failure
rate Is apportioned to each feature based on the relatjve occurrence of each
failure mode. The failure rates for the components under each RI in each FD&I
feature column are sumned together and entered as the failure rate for the RI
checked by that particular feature. This assumes the feature e!ther checks a
single RI or can check multtple RIs by some sequencing scheme. Those
components which are not included under any F(MI features represent failures
not isolated with the FD&l features. The failure rates of the failures not
isolated by the FD&I features are noted In the manual isolation failure rate
column of the matrix to complete the accountingof the total equlptnentfailure
rate. All manual isolatlon cases must be accounted for,
If those cases where the n failure rate ts known to result in several FD&I
outputs, but the allocated failure rates are not known, the rationale for the
assumed allocation of the failure rates shall be stated.
FIGURE V-5. Matrix for correlating FD&I features with RIs.
12 Jan 1984
MIL-HDBK-472
The next step in the correlation process is to associate the FO&I features
with the FDM outputs. This is accomplished using a fault tree type diagram .
such as the sample shown tn Figures V-6 & V-7. The Top of the tree consists
of all FD&I outputs; the second tier contains the Fil&lfeatures uhlch
separately or jointly result in the given FD&I output; and, the bottom tier
presents the RIs associated with each FM feature and the failure rate
associated w!th that feature. The circles are used to assign numbers to all
unique FD&I outputs. The triangles identify the order in whfch RIs are
replaced when the replacement concept calls for iterative replacement.
4.2.5 Prepare maintenance flow diagrams. Next a maintenance flow diagram

(MFD) is prepared to establish the Rel values for insertion in the
!lalntenance Correlat~on Matrix (Figure V-10). The MFD is prepared to
illustrate the sequence of maintenance required. The symbols used in the t4FD
are:
o StartMi Point (i.e., Failure Occurs and is

Detected) or Ending Point
Activity Block. The top of the block indicates

I I a sDeclfic maintenance activity and the bottom
ind~cates the time assoc~ated with that act!vity.
This is the only symbol that denotes ttme.
FD&I Outputs. Designates the primary or secondary

uniaue FD&I outmt which defines the subsequent
mai~tenance activity to be performed. The-j
associated with the output is entered in the
circle.
Decision Point. Defines a point In the

o maintenance flow at which time the maintenance
technician must make a dectslon on which
subsequent path to take.
0 n,j Path Identlfler. Uniquely identifies each path

by unique RI(n) and FD&I output (j).
Continuation. Designates continuation fromor to

another place In the maintenance flow diagram.
n
12 Jan 1~4 V-26

-=
r
mSEQUENCC
T15TS
3
Q
ANTENNA ANTENM
K)s4TIOM ARD Mln AND
RATES ROLL RATES
m
610
&=7m
RI
A
nj
Fi~ure
hncl.ature
fOr
V-8
Q 1o
A =9,s43
FIGURE V-6. Fault Isolatlon outpu t and RI correlation tree.
V-27
3:L.HI)BK-L72
I
m
#J~~~,iJtj-ii
011 041
A.z.106 A 0.204
QQ::~m{pf
039
A = 4.308
1
041
k 0.204
01 i
A. 0.2s4
001
A. 0.ss9
039
A. 0.120
031
A*13940
I
~. 0204
610
ArOB79
2
Ot 1
A,62b6
03 i
A*077S
I I
J
1 ou:put and RI correlation
. tree (partial) .
FIWRE V-J. ?!anual fault isolation
V.7R
HIL-HD%K-472
\
The HFD (as Illustrated in Ffgure V-8) starts on the left side of the figures
as a Fa\lure Occurs and Is Detected event. If IsolationIs inherentin
fault detection, the neKt Item shown in the MFD is the unique FD&I outouts.
If isolation is not Inherent In detection, the next ftem in the MFD is the
fault detection output. This would be followed by activity blocks which
define the procedure followed to achieve fault isolation. The activtty
b)ock(s) is followed by the unique primary FD&l outputs associatedwith the
maintenanceactions that have been executed.
following the FD&I output symbols are shown the activities required for fault
correction and repair veriflcat~on.
If a FD&I output results fn non-ambiguousmaintenance (i.e., primary isolation

to a s~ngle RI, or group RI replacement),then an End symbol will directly
follow the fault correction and vertftcatfonacttv~tfes. lfeW&J W?pW
results In an ambiguous result, a verification decision block IS shown after
each verificationactlv~ty (except the last). Any actlvlty (e.g., clean-~p)
performed after a positive verificationdecision Is shown In an activity
block(s) between the decision block and the End symbol. Associatedwith each
End symbol Is a path identifierwhich un!quely ident!fleseach path by RI and
FD&I output. For example, the path associated with the second RI and FD&I
Output #12 would be designated as 2, 12.
Care must be exercised to ensure that all possible maintenance actions that
could be followed as a result of a FD&I output observationhave been accounted
for, especially those that result in Manual Fault Isolat\on.
The R.j values are computed by adding the times assoc~atedwith each
activity block from the Failure Occurs and Is Detected event to the
end event for the subject (n, j) pair. Note that only the act!vityblocks
have t~me associated with them. The time entered in the fndiv~dualactivity
blocks fs computed from a time ltne analysis prepared in accordancewith
paragraph 4.2.6. Elemental times entered in the time Ilne analysis are
extracted from the following sources In the order given:
a. Actual times eKperlencedon the subject equ!pment.
b. Standard times from Appendix A.
c. Actual times experiencedon slmllar equipment.
d. Other recognized time sources.
e. Engineering judgement.
In the establishmentof the time line analyses, the number of maintainersmust

be considered, For example, If a given equipment has two technicians
performingmaintenance,one technician may perform disassembly to ach~eve
access to the faulty RI while the second technician simultaneouslyperforms
other work. In thematntenance flow diagram, this would show as a single
maintenance actlvlty w!th the associated time being the elapsed clock time.
If the parameter of Interest was ~loii, Instead ofMTTR, then the the
entered In the actlvlty block would be the combined MMH in lieu of the elapsed
time.
12 Jan 1984
V-29
KIL-EDBX=472
.
0.
-
.
I
s.
b
e
E-
Jan 19B4
141L-HDBK-472
--
4.2.6 Time line analysis. The estimated times used In the two prediction
methodologiesare synthesizedusing a time line analysismethod. A tlw llne
analysls conststsof computtngthe total elapsed time of a maintenanceactton
by accountingfor the t~me required to perform each step. The procedurefor
performinga time line analys~s is as follows:
a. Identifyeach task that comprisesthe maintenanceaction,
b. Determine the t!me required to perform each task by either actual

times, maintenancetime standards,time studies, or englneer!ng judgement.
c. Determinewhich actions can be done sitnultanewslyif more than one

ma!nta ner IS available.
d. Determtne the overall time to perform the maintenanceaction by

smnlng up the times to perform each action.
Figure V-9 is an exampleof how a time is synthesizedfor a simple physical

task. The time associated with each task !s extractedfrom the table of
maintenance time standards shown in Table A-V-I In Appendix A.
RI NAME: NODULE (l/R) ELEMENT MAINTENANCEACTION: INTERCHANGE
DESCRIPTIONOF THE ELEMENTALTASKS TIME/ACTION QTY TOTAL TIME
REIWVE QUICK RELEASE COAX 0.04 4 0.16
RENOVE SLIDE LOCK CONNECTOR 0.09 1 0.09
RE140VEMODULE 0.09 1 0.09
REPLACE!WDULE 0.11 1 0.11
REPIACE SLIDE LOCK CONNECTOR 0.12 1 0.12
REPLACEQUICK RELEASECOAX 0.04 4 0.16
TOTAL TIME 0.73
FIGURE V-9 - EXAMPLE TIME SYNTHESISANALYSIS
12 Jan 1984
WL-HDBK-472
4.2.7 RI and FD&I output correlation. The results of the preceding section
are sunanarizedIn a matrix which shows the relationship annng the RIs for
which the prediction is be!ng performed and the total set of FD&I outputs.
The matrix (Figure V-10) identifies the RIs across the top and the unique FD&I
Outputs down the left column. In reference to the math models (refer to
paragraph 3.2) the RIs are the n paranwters and the FD&l outputs are the j
parameters. Each RI column is further divided into three columns:
15 I I 1 1
FIGURE V-10. Maintenance correlation matrtx format.
Under each RI column, enter the failure rate (A.j) of the RI (obtained
from the FD&I correlationtree) (see Figure V-6) that is associatedwith each
FD&Ioutput. For each untque output which has only one RI associatedwith It,
enter a 1 In the ().jcolumn for that combination. For those outputs which
are associated with 2 or more RIs, the Q, value is determinedby the
replacement concept. If the replacementconcept 1s group RI replacement,
enter under Q.j the number of RIs associatedu~th each output. For example,
if three RIs could contribute to the same FD&I output, then a 3 ts entered
12 Jan 1984
V-32
MI L-HOBK-472
.
in the Q., for each of those RIs. If the replacement concept is Iterative
replacement, then Q., \s assigned based on the order of replacement. That
is, the first RI to be replaced upon recognition of the subject F08J output Is
designated as Q.j = 1, the second Q.j = 2 and sc forth. In cases of
integrative replacement, the values for each Q.j Is based on the relative
failure rates of the RIs, with the highest failure rate RI assigned as the
first replacement item.
4.2.8 Compute maintainability parameters. Once the MFD and Malntmance

Correlation Matrix have been completed, compute the maintainability
parameter(s) using the equations in section 3.
.- .-_.
-
MIL-IUNK-472
APPENOIX A
TM STAKWD5
10. The ti- standards are tabulated in Table A-V-I. The times tabulated in
Table A-V-I have corresponding figures referenced which illustrate uhat each
time represents. Table A-V-II contains coqosite times of ccxmn maintenance
actions that may occur. Colums two and four of Tab?e A-V-II denote which
tires of Table A-V-I were used to synthesize each actlvlty (letters denote
remval (A) and replaceable (B) tires).
20. Other maintenance tasks can easlIy be synthesized by the following method
(for an ex~?e. see Figure v-9, in paragraph 4.2.6).
a. List the actions involved for the maintenance task.
b. Obtain the tiws for each action by using Table A-V-I (tiws that are
not Ilsted should be established either by actual data, time studies, or
engineering judgemnt).
c. Compute the tim by sumlng up each ind~vidual time.
-.
>lIL-}ll)BK-~72
TABLE A-V-X. & mental maintenance actions.

b
1
Standard Times
Time
standard Rem we Replace Interchange Reference \
Number Description (min.) (min.) (min. ) Figure ~
FASTENERS
1 Standard Screws 0.16 0.26 0.42 A-~!_ 1
~ I!exor Allen ~q3e I
Screws 0.17 I 0.43 0.60 A-Y-2
3 , Captive Screws 0.15 0.20 0.35 A-V- 3

:_ .. !.-.. .
4 DZUS {U4 LUHUOCK) 0.08 / 0.05 0.13 A-V-4
I
5 Tridair Fasteners 0.06 i 0.06 0.12 A-V- 5
6 Thumbscrews 0.06 0.08 I 0.14 A-V-b
7 Machine Screws 0.21 0.46 0.67 A-V- 7
8 Nuts or Bolts 0.34 0.4-4 0.78 A-V-H
9 Retaining Rings NA 0.27 NA A-1- 9
LATCHES
10 Drawhook 0.03 \ O*OZ 0.06 A-V-IO
11 Spring Clip 0.04 0.03 0.37 A-IJ- ] !
12 ButWrfly 0.05 0.05 0.10 A-v- I~
13 ATR (spring loaded,

pair) 0.45 0.69 1.14 A-V-13
14 Lift & Turn 0.03 0.04 0.07 A-V-14
15 Slide Lock NA NA NA A-v-. i J
TERMINAL
CONNECTIONS
16 Terminal Posts
(per lead) 0.22 0.64 A-V-10
17 Screw Terminals 0.23 0.45 0.68 A-V- 17

I
18 Termipoint 0.22 0.30 A-V- Ii{

I
19 Wirewrap 0.09 0.24 A-v- 19
0.07 0.07 0.14 A-\l_:f)

20 Taperpin
I
I
12 Jan 1984
TIMLE A-V-I. Element; L maintenance actions (continued). 1
~
Stidard Times I
~
Time I
Remove Replace \ lnterc~ge Reference
Standard
Ik2scription (min. ) I (min. ) I (min. ) ~ Figure
Number
1 I 4
~ TERMINAL I I
CONNECTIONS (cont. ) I t
21 PCB a) Discretes 0.14 io.17 ! I A-V-~)

1
2J b) Flatpacks 0.14 0.13 perl 1%-(-2 i

per lead Iflatpack
I
c) DIP [Cs
8 pin 0.46 0.52 i A-V-2 1
23
24 14&16 i ().90 \ 0.86 I

I A-V-2 1
pin I I
I
CONNECTORS , !
I I \
I ~-1~-z~
~ 0.07 ~ 0.10 0.17
25 BNC (single pin)
0.07 0.12 (). lg A-V-22
26 BNC (multi pin)
I i A-L-23
0.04 , 0.04 0.08
27 Quick Release Coax
NA A-; -2JI
28 Friction bcking NA ! NA i
1
29 Friction Locking with I 0.38 \ A-V-25
one Jack Screw 0.18 ! 0.20 I I
Thread Locking , 0.09 \ 0.17 0.26 I A-v-26

30 I
1
1
,j-~-~?
, Slide Imcking 0.09 { 0.12 0.21 I
31 i
! i
J
PLUG lN I I
MODULES 1
1
32 DIP ICS (into ~-\l-)~
0.21
DIP sockets )
I I
(
CCAS (without tool)

(guided)
A-\-2y I
40 pin NA
I
0.11 ,j-:-~q
33 SO pin , 0.04 ~ 0.07 I
I I
I ,
CCAS (with tool)
(guided) I
40 pm 0.06 0.07 0.13 ,t-,- 12 I
34
I
b 80 pin I 0.09 i 0.08 0.17 I f\-\-3o
35 )
I I
I!
1
I
I
I
1
(
A-[l_) 12 Jan 1984

m!l.-ttDBK-472
TABLE A-V-1. ~lmental mimtenance actions (contiwed).

f w 1
Time I Stad8td Time

Referencx
Number Figure
I
,
PLUG IN
MODULB (cow) II
I
I
I
j CCAS (without tcml) (
; (sot @d@
I
40 pin MA NA NA A-V-3]
i
36 i ao pin 0.04 0.16 0.20 A-v-31
37 0.09 0.11 0.20 A-V- 32
I
I
38 I strip wire 0.10
1- -
39 cut wire of sleeving 0.04
40 ( Dress Wire with
ISieevilqg I 0.21
i-
41
42
crimp hlgs
Form Leads (per
1- 1-
1
<
I
0.27 A-V-33
A-V-34
lead)
h- 0.03
43 Trim Lads (per 1
had) 0.03
0.55 1
, 0.13 0.68
44 Adhesives
45 Conformal Coating 2.20 i 0.23 2.4S

46 soldering A) Terminal I
Posts 0.22 A-V-35
47 B) PCB 0.06 A-V- 36

i-
4a Reflow Soldering 0.25

49 Tinning Flatpacka I
(dipping) i- 0.30
50 Resoldering A) Braid@ 0.16 A-V-37

Wick i-
i
0.09
51 B) Solder I
1
A-V-38
1
Sucker
I
52 FomtI Fla@ack YJEads 0.11 A-v- 39
(Mechsnkslly)
1- 0.29
1
53 Clean Su tie
I -
54 Panels, Doors, & 0.04 I 0.03 0.07 A-v-40
Covers !
12 Jan 1984 A-v-4

}tlL-liDBK-Li2
TABLE A-V-I. E1-enLal maintenance actions (continued).
Stmdard Time
1
Standard ! I Remwe Repiane M&w&m@? Reference
Number Descri@ioa I (min.) (min.) (min.) Figtm
MBCELLANEOUS
(cont. ) I
55 Drawers (Urge) 0.09 0.10 0.19 A-V-4 1
56 Display Ixuxw 0.10 0.11 0.21 k-V-42

I
57 Threaded Connector
Ccnrers 0.11 0.14 !
0.25
I {
I :
I
i
I
I
I I
I
.-L ....._..__. -..- -.

i
_l--- .- -- --- 1- .-
NOTE: Oata obtained from RADC-TR-70-89, Maintainability Prediction and

Oemmstration Techniques
A-V-S 12 Jan 1984

CJ
m
-- . L?
--
m s
m
7
--
w -
m
-
a 1-
0 N-
N L?
m
m G: m
m 7 N
.
.
1-
C-J
m m
-.. . m In
U
m
--- C-J G G <
a E
.
G -
.. . ..-. -
I I
u
.
r) 1+ o 1+
m
I ___ .- -. . ..
12 Jan 1984
A-v-b
.
ST ANOAQO !ica~b%s
THIS TIME ts~OR ALL sTAhDARG THREADEo Fasteners

SUCM AS. SLOTTED HE&O. OI+JLLIPS MEAD, AND ~tLLISTER
l+ Ao
THE TIME GIvEN IS THE TIME QEOUIRCD TCI QEMOVEI
StEPL ACE THE FASTENER FROM THE HOLE AND DISENGAGE:
OR ENGAGE IT L?Y SEVERAL TwISTING MOTIONS OF THE
HAND (A PPROXIMATEL~ S TWISTS I
TOOL ffEQUIUEC) IS STANDARD SCa EWDRIVER (FLAT HEAD.

Pt+t LLIm. OR~Exl
FIGURE A-V- 1. Standard screws.
HEX OR ALLEN SET SCREWS
THIS TIME IS Fo$? HEX OR ALLEN TYPE

SET SCREWS
THE TIME GIVEN IS FOU THC TIME TO ?lGHTEN/
OR LOOSEN A HExAGONAL TVC% sET SC f4EW
US?NG AN ALLEN TYPE wRENCH
TOOLS REOUIREO ARE HEX WRENCHES OR
ALLEN TVPE wRENCHES
I
I
FIGURE A-V- 2. Hex or Allen set screws.
12 Jan 1984
MIL-HDBK-472
CAPTIVE SC$lEWS
TH6 IS THC TIME FOfl STANOARDFASrENERS

THAT ARE C14PTl VE TO THE PA NE L/B RACUEl
TUEV SECURE
THE TIME GIVEN FOR THIS ACTION INCLUDES

THE TIME TO EF4GAGEKIR DISENGAGE THE FASTENER
ov A SERIES OF TwIS?t NG MOTIONS WI Tti TblE HANO
THE TOOL I?EOUIREO IS A ST ANOAfl DSCfl EWORIVER

[F MT+IEAD, PHILLIPS OR HEX I
FIGURE A-V-3 . Captive screws.
DZUS FASTENER
m
THIS TIME IS rOR FAsTENERS THAT REOUIRE ONLY
A 1/4 TuRf4 TO ENGAGE OR OISEWGAGE fFASTENEn
IS cAPTIVE)
THE T4ME 61v EN IS THE TIME REQu IREO TO ENGAGE
On DISENGAGE THE FASTENf Fl BY A 1/4 TWIsT Mo.
TION OF THE HAND
TME TOOL REQUla EO IS A ST ANOARO SCREWDRIVER
IF LATUEAO, PMILLl~ On HEX)
FIGURE A-V-4 . DZUS fasteners.
12 Jan 1984 A-v-8

MIL-HDBK-472
I
/
TRIOAIR
c THIS
FASTENER
FAST ENER Is A QUICK ENGAGING FASTENER . > =

-$3
/
,/,,/
TH4T REQUIRES LESS THAN ONE TuRN
,{
e/
9 TM!S. TIME INCLUDES THE TIME NECESSARV TO
tZNGAGE/OR DISENGAGE THE FASTE!WER USING A
TURN OF TwE wAND
c THE TOCJL REQUIRECI IS AN ALLEN wRENCH
. -.
LJ k
FIGURE A-V-5 . Tridair fastener.
-.
THUMBSCREWS
THIS FASTENER IS A THREADED SCREW WITH A HEAD

THAT CAN SE GRASPED EASILY BY THE HAND
THIS T{ME WCLUOES THE TIME NECESSARY TO DISEN -

GAGE/OR ENGAGE THE FASTENER BV TURNING IT WITH
THE HAND
NO TOOU REQUlaEO
Wc
FIGURE A-V-6. Thumbscrews.
12 Jan 1984
..41 !.
. ..
! 1) !:,,-.:
.
-,
MACHINE SCREWS fWITHNUTl
THIS FASTENER IS ANV THREACWD ~ASTENER 7HA1

DO ES NOT TAP INTO THE ST QUC~URE. INSTEAD IT
E?+ GCtCIES INTO A LOOSE NUT
THIS TIME INCLUDES THE T)ME TO REMOVE/OR POSI
TION THE FASTENER ANO NUT ANO TI+E TIME
REQuIRED TO TIGHTEN THE FASTENER
TOOLs REOU!REO ARE A SC REw DQlv EQ AND A wRENc H
FIGURE A-V-7 . Machine screws.
NUTS OR BOLTS
A?4v FASTENER THAT REQUIRES A WRENCH
TO TIGHTEN IT DOWN
THIS TIME INCLUOES THE TIME NECESSARY TO
POSITtON THE w$7ENCW ANO ENGAGE/OR DISEN.
GAGE THE FAST ENEF!
la I
\
FIGURE A-4-8. Nuts or bolts.
12 Jan 1984 ,1-V- 10

RETAINING RINGS
THIS DEvICE IS A. U SHAPED PIECE OF METAL THAT

RETAINS A UN! T/COMpONENT IN POSITION
THE TIME GIvEN INCLUDES THE TIME NECESSAQV TO
--
tINGAGE/OR - OISENGAGE THIS FASTENER
SPECIAL PLIERS ARE REQUIRED TO REMOVE /REpLACE
b
TH!S FASTENER i
;,2
L
-1-
FIGURE A-V-9. Retaining rinlls.
DRAWHOOK

ANY
SHOWN
LATCH
THE TIME
DISENGAGE
No TOOLS
U$TCH
THAT
lf+CLUOES
IS S\ MILAa
THE TIME
THE LATCH
REQUIRED
TO
TO ENGAGE
COMPLETELY
THE ONE
&
Llli m I Ill
i,.
I
J
FIGURE A-V-1 O. Drawhook latch.
12 Jan 1984
L-v- 1 1
t4JL-III)NK-472
Sw!tw CLIP cnTcll
t ANY LATCN slMILAn TO Tt4E ONE 51 iOWN

TIME INcLUDES T14E TIME FJECESSARV 70
!IWOfNWE~n DtSEM3NiE THE LATcH COM.
PLETFLV
* NO TOOLS nEQUIREO
L 1
FUUKE A-V-11. Spchui clip Utch.
EWTTEnFLV LATC~l

nNv LnTc$4 slMILnn To T14E ONE SIKWJN
TIME W+CLUTWS THE PtECESShnv TIME To cNGAnE/
OR DW~NOAmE THE LnTC\l COMWETELY. Nnn~nL-
LV coN$lsTs oF CIFTINQ TI+E lfbl! hNO TUnN)tVG IT
me,
NO TOOLS REOUta EO
I
FIGURE A-V-12. Butterfly latch.
ATR LATCH
ANY LATCH SIMILAR 70 THE ONE S140WN

. TIME INCLUOES 7*$E TIME Nccessnnv Tn uf4scn Ew/
On SCnEW T$+E Ch*UVEn TNE NIB TO HQOnUEttTQ OIS.
lENOfiOE THE SeCUREO UN1l. TME TIME OIVEM !S FOm
A NR OF ATI! LATCNE$.
MO TOOLS *l! OUtn ED.
FICURE A-V-13. ATR latch,
12 Jan 1984 A-V-1 2

I
.
LIFT ANO TuRNLAICM
ANY LATCH SIMILAR 70 THE ONE S+40WN

TIME INCLUOES T?(E TIME NE CESSAHV TO LIF 1 THE
l+ ANa LE aNO TURN 1? ?0 UNSECURE OR SECURE A
DOOR OfJ PANEL
NO TOOLS REQU!UED
I1 1
FIGURE A-V-IL . Lift and turn latch.
SLIOE LOCK LATCH

ANV LATCH slMILAR 10 THE ONE SHOWN
THE TIME GIVEN IS THE TIME NE CESSARV TO
SLIOE 7HE LOCKING DEvICE TO ENCf+GE)OR
OISENGAGE THE PANEL
NO TOOLS REQuIRED
FIGURE A-v-l 5. Slide lock laLch.
TERMINAL
b

aNv
SHOWN
THIS TIME
REPLAcf
IN CLUOE
NE EOLE
Posrs
TERMINAL
IS THE TIMf
A LEAD
sOLDEOING
NOSE
connection
FROM
REQuIRED
A TEQMINA(
OR oESOLOCRINC)
IJLIEt?SAfl
SIMILAR
TO REMOVE
(n~ES
E REQUIRED
No*
OR
TO
fOR
THE
THIS
ONES
TASK
in
v
o
0
?tIL-HDBK-472
SCREW TEQMINALS
ANY TERMINAL CON NE C710N SIMILAR TO THE
ONE SHOWN
THE TIME GIvEN IS THE TIME REQuIRED TO
Rf MOVE/OR POSITION THE TERMINAL LUG AND
LOOSE N(OR TIGt+TEfw THE SCREW
A SCk EWCIRl Vk8 IS fiEQul RED
FIGURE A-V- 17. Screw terminal connections.
TERM IPOINT cONNECTIONS
ANY TERMIANL CONNECTION S! MIMF! TO THE

ONE SHOWN
THIS TIME IS THE TIME TO REMOVE THE CLIP
WITH A PICK OR TwEEZERS ANO THE TIME TO
REPLACE lHticLIP WITHA TERM IPOINT GUN
TOOU REQUIRED ARE TWEEZERS, OR A PICK,

ANDA TERMIPOINT GuN
FIGURE A-V-18. Termipoint connection.
. ,,%1
YIL-HDBli-4i2
WIREWRAP
1
ANY TERMINAL CO NNEC710N SIMILAR TO THE ONE SHOWN
THE TIME SGIVEN ARE TO REPLAcE THE WI REWQAPWJITH
s WI REWRAP GUN AND TO REMOVE THE Connection WITH
AN uNWRAPPING TOOL J
TOOLS REOLJIQEL) ARE A WI REWRAP GUN ANO AN UNWRAP.
PING TOOL
,,
FIGURE A-V-19. WireWrap connection.
TAPER PIN
ANY TERMINAL CONNECTION SIMILAR TO THE ONE SHOWN

THE TtME GIVEN IS THE TIME REG)u IREO TO UNPLUG MATE
y
OR DEM&TE THE CONNECTOR
o
NO TOOL REQUIREO /-
FIGURE A-V-20. Taperpin connection.
%6

ANY TEHMINAL CONNECTED O! RECTLV TO
THE PRf NTEO CIRCUITI?V OF A CIRCUIT CARD
THE TIME GIVEN IS THE TIME REQUIRED TO REMOVE OR
REpLACE f! LEAD ~NOM THE PCB (NO SOLOER!NG OR RE-
SOLDERING TIME)
THE TOOLS REQUIRED ARE A PICK OR NE EDLENOSE PLIERS
FIGURX A-V-21 , PCB connections.
.- ..-.
E-c cowvEc To$ls
ANv CONWECIOm THn T wA54 BAVONET LOOKING OEV,CE
TIMEs GIvEN ARE FO~ MA TIN G/DE MA TlMt2 THE COIW.4ECTORS

BV A TwISTING MOTION
NO TOOLS REQu IREO
H
FIGURE A-V-22 . BNC connectors.
QUICK RELEASE COAX Connectors

ANY COAK CON UECTOR THAT cNcAGE5 OR DISENGAGES
BV A PMw On PULL MoTION
TIMEs GIvEN ARE FOP DC MAT? NGfWA71NC THK CONNECT
ORS BY A DULLING OR J%ISuf NG h@TIc~N / e
No TOOLS Rf. QUINLU /
FIGURE A-V-23. Quick release coax connectors.
FRIc710N LOCKING CONNECTORS
ANV CON NE CTOQ THAT IS MATED ANO SE CUREO

eV THE Ffllc TION OF THE PINS ANO/0= CONNFC70R
CASE
THE TIME GIvEN 1$ THE lfME NE CC SSA~V TO MATE
OFI DE MAt E THESE CONNECTOR TYPES
NO TOOLS RECMJIf!ED
FIGURE A-V-24 . fric~ion locking connector.
12 Jan 1984 A-Y-16

!$III.-I{ I)9K-L-J
FIGURE A-V-?5. Friction locking connector _wit~]~ackscrew.

.-.
lHQEADLOCKING CONNECTOR
ANV cONIWECTOQ THAT lSSECIJRED8v A

lHUEADED CONNECTOUSMELL
K
1-!S TIME GIVEN 15FOR TME DEMATING/MATING
oF THE cONNECTORAND THE SECURING/UNSECUR
IF4GOF 11 @V A TuHNINGMOTION
NO TCIOLSQEQUIRED
FIGURE A-V-26. Threadlocking connector.
SLIDE LOCKING CONNEC1O9
ar.Jw CONtU(CTO!=T-IAT IS sIcIJfJ[l) RV A Ll~DE LOCK

THF TIME GlVEN15T14f TIME REQul\?fDTOMATE/DEMATE
THE cONNEC70FI ANL! ENGAGE/CllSfNGAGF THE SLIDE LOCK
NO TOOLS RCQUIRED
FIGURE ,1-:-27. Slide lockin~ connector.

,1-V-17 12 Jan 1984

/
U(P fc$ @
THIS INCLUDES ANY DIP IC THAT !5 SECURED IN
a DIP SOCKET
,,.&@
THE TIME GIVEN IS THE TIME REQuIRED TO UNDLUG ;;> / ~
OR PLUG lN THE DIP IC
NO TOOLS REOLMR60
===
F1CX?9E A-V-28. m
GuIDED CCA\
ANY GuIDED CCA THAT IS lNSERTED/REMOVED BY

HAWD
THE TIME GIVEN !S THE TIME TO PULL OUT OR
PUSI-I IN THE CCA
. NU TCKJLS REQUIRED
ylcu~ A-v-29. Guided CCAS.
GuIDED CCAS WITH A TOOL
THIS TIME IS ~swcl AT ED WITH ANY CCA THAT

IS w4sEf+7E0/REMOVED wlTH A CARD ExTRAc T.
lfuG TOOL
THIS TIME INCLUDES THE TIME REOUIREO TO
F9SITION THE TOOL AND REMOVE/REO@CE
TUE CCA
A cARo ExTRAcTING TOOL IS REQUIRED
I
FIGURE A-V-30. Guided CCAS with a tool.
&_,., 10
M[l.-Ho HK-472
NOt+A2UlDE0 CCAt
. TMIS TIME 1S ASSOCIATED WITH PLUG )Nc Alt DS
THAT AQC NOT GUIOEO ID
Tl+tb TIME lNc LLloES THE TIMC NC CESSAF4 < ro

~{LMOVE/RCPLAcC THE ccA r ROM TnL LUG(
COMNt CTOIl IOOC5 NOT I?UCLLM3E 1 ~ME +064 d L ml
FIGURE A-V-31 . Non-~;uided CCAs.
%
.-. #-
./
..
MODULES F:,
. .
9 THIS IS THC TIME ASSOCIATED WfTH REMOVING 0$? POC>I
41
/4 ..-
.d
~--
,
lIONING A MOOULAR AsSEMBLY
-----
THIS TIMC IS THE TIME Nf?CESSARV 10 REMOVE 7+4[ Mf>(). %
w&
ULE Ofi POSITION IT 4N PLACE
=&
NO TOOLS RCOUIRECI !
4 I
)
F1(XJRE A-V-3?. tl[>dules
C[<lhll> LIJC,~
. \. .,
I
?!!!%?
TH!s 1s TH[ TIML Assf)c IA7CD WITH 5fc(Jut Nc n If 0t141NAL
LUG TOA WIRE
$?4[ TIML GIVFN lNCLUO[S 114E lIML 10 WSIII(JN !Iff
W,lll. IN T,4c L uG AN OCOIMP II
A (141 M01NG 100( IS RCOUIIJ( 1) o 0
\
12 Jan 1984
?lI&HDBK-472
/
f
/Air ,1
(1,
FORM LEA12S
L
!1(,
TH!S IS THE TIME ASSOCIATED WIT* ~ORMING A
1,
LEAD ON A COM-NENT pFil OR TO CONNECTING
1? TO A TERMINAL /i:\ \
THE TIME GIVEN 15 7ME TIME NECEsSAa Y 10 GRXP
THE L@ AD W!TH THE PL)ERS ANO FORM IT \
( ~~
NEEDLE NOSE PLIERS ARE REQUla E@
FIGURE A-V-34. Form leak.
/
SOLDERING TERMINAL POSTS
THE TIME AS50CJATED wITH SOLDERING A LEADTC

A TERMINAL POST ..-
-.-
THE TIME GIvEN M THE TIME TO HEAT THE TERMINAL / _=--
IWST AND APPLY THE SOLDER .5
A SOLDERING IRON IS REQuIRED
B
FIGURE A-V-35 . Soldering terminal posts.
SOLDERING =8 CONNECTIONS
THE TIME ASSOCIATED WITH SOLDERING A LEAD

?0 A PCB ETCHING
THE TIME GIvEN Is THE TIME TO HEAT THE ETCHING
PAD fiND aPPLY THE SOLDER
A SOLDERING IRON IS REQuIRED
FIGURE A-V-36 . Soldering PCB connections.
A- V-2(-I
MIL-HDBK-472
RESOLDERING
. USING
ABRAIDEDWICR
THE TIME ASSOCIATE DVVITH RESOLDERING A CONNECTION USING

A BRAIDED COPPER WICK
THE TIME C!VEN IS THE TIME TO REHEAT THE SOLDER AND EXTRACT
u
IT FROM THE PCB OR TERMINAL USING A BQAIDED COPPER wlc~
A sOLDER!NC IQON AND BRA IDFO COPPER WICK ARE REQUIRED
FIGURE A-V-37 . DeSoldering with a braided wick.
DESOLOERING USING A VACUUM

-- . -
THE TIME ASSOCIATECI WITH DE5C)LOERK+4G A CONNECTION

U51NG A VACUUM ASSISTED RESOLDERING lPf ON
THE TIME GIVEN IS TI+E TIME REG)u IREDTo REHEAT
AND SSLJCK. UPB THE SOLDER
A DESOLDECIING IRON IS REQUIF4E0
FIGURE A-V-38 . Resoldering using a vacuum.
FORM FLAT PACK LEAOS

.,+@O:i.-
THE TIME ASSOCIATECI WITH FORMING FLAT PACK LEADS USING A
.:--
MECHANICALLV OPERATEO OIE .
THE TIME GIVEN IS THE TIME REQUIRED TO ~SITION THE FLATPACK
ANO AC TUA7E THE MECHANISM
AM+FCHANICALLV OPERATED DEvICE IS uSED TO 00 THIS
m
I
FIGURE A-V-39 . Form flat pack leads.
12 Jan 1984
ML-HDBK-472
\
PANELS. OOORS, AND COVERS
THE TIME ASSOCIATED WITH OPEN/CLOSING PANELS,

DOORS, AND COVERS
NO TOOG REQuIRED
+
FIGURE A-V-40. Panels, doors and covers.
DRAWERS
THE T$ME ASSOCIATED WITH O~NINGtCLOSING
OF DRAWERS THAT ARE ON A TRACK 1
NO TOOLS aEOUl~ED
I
1
-J
\
FIGURE A-V-4 1. Drawers.
DISPLRV LAMPS
g.a~
THE TIMI? REQuIRED TO REMOVE /aEf%ACE PANEL
INDICATORS THAT POP IN AND OUT
---a/
NO TOOLS REQUIRED
FIGURE A-V-42. DiePlay lamps.
1? ,L3n lQM
MIL-HDBK-472
APPENDIX B
ESTIMATES OF Me,. ($) FOR LOGNORMALREPAIR DISTRIBUTIONS
10. This appendi~ tmovides estimates of K,, ($) for 109normalY

distributed repair times. The estimates for ~m,= ($) are tabulated in
Tables B-V-I and B-V-II. The %.. (0) values are found by:
a. Selectlng the percentile of interest {*I either 60, 10, 80, 90, 95.
or 99 percent.
b. Locate the mean repair t me (MEAN) which most closely approyimate$

the NT R of the equipmentlsystem n question The repair times are ~rovided
from O 1 to 2.6 in steps of 0.1.
c. Locate the corresponding repair times standard deviation (SIGMA)

wttich s estimated for the subject equipment/system. Values are provided from
0.1 to 2.5 in steps ofO.1.
d. Read the value of K,. (0) under the appropr ate percentile
col umn
.-
R\l 1
60thi 70th and 80th percentiles of the Iogrmrnal distribution
for means and si~~as from .1 tc 2.5.
!32 PF.JC:!:T ~~ p~~~~;;~ 13CI?~CE:i7
.! .! .0673]5 .109413 .142493

.1 .2 .06J 674 .086984 .130081
., .3 ,(j&~g$47 .0 7LO 73 .I134G6
.
.1 ,4 GJ /I.- c- .:>:~~ .! 3L9Q2
.i . 5. .C323?3 .C50535 .OES ~9L
d .6 .02 SK 6 .044:!32 .081369
.1 .7 .C23J <2 .0J9899 .Z?74R3
.1 .8 .0208!4 .036212 ,069236
I .9 .0;E756 .3332Z? .!-)64Q2
.i i .G .C1 7J 45 .030657 .060532
.! 1 .1 .@i5775 .02d575 .357273
.1 102 .014614 .026754 .?54290
.1 ] J .013613 .025171 .C51 654
.1 I .4 .C12755 .C23 792 .C4920C
.1 i 5 .0119s6 .(322 552 .047193
4 1.6 .01!330 .021454 .C45295
.1 i .7 .GIC73> .02046S .S43565
.1 1 .6 .CIC222 .C19576 .041985
I 1 .9 .OO!J 721 .018769 .040536
.! 2.0 .009286 .018030 .039201
.1 2.1 .006889 .017353 .037966
.1 2.2 .006526 .016729 .0365?.0
*.1 2.3 .0061s3 .016153 .035753
.1 2,4 .007636 .0156J9 .C34757
.1 2.5 .007602 .015123 .033624
.2 .1 201 Q8 .229170 .26Qj7

2 .2 .174630 .218639 .284966
- .3 .146064 .196038 ~76636
.: .4 .123348 .173968 .2601 ti
.2 .5 .106102 .155384 342835
.2 .6 .0928S5 .140157 .226812
J .7 .082564 .127649 .212556
92 0 .0 743G3 .117259 .2COO04
.2 Q .0675Q .loe518 .166953
02 I :G .061965 .101071 .i791g3
.2 1 .1 .057245 .094652 .170501
2 12 .053212 .089063 .162738
.2 I .3 C)49R7 .0)34151 .155757
.2 I ,4 .0466S4 .C 7S793 .149446
.. 1 .5 .044005 .075914 . I 43709
.; I .6 .04!S27 .072424 .138471
. .133666
L 1.7 .039 5C2 .C 692 7;
6- J P .c3 75?2 .066406 .129244
.2 ] .? .035e55 .063791 .!25156
.2 2.: .03429< .06!394 .121364
.2 2.1 .032654 .059 I 88 .117s37
2 202 .031551 C5?;51 .1;4546
02 2.3 .03034.3 :L1552@ .ll14~i
.2 2 .4 .02?228 .05J>07 .iu65eo
.2 2.5 .02E!?6 .051871 .I05g66
12 Jan 1984
YIL-l IDBK-47:
TABLE .3-;
-1.
60th, 7oth and OOth percentiles of the lo~ncrrnal distribution

for means and sigmas rem
.1 to ?.5 (continued).
IIEAN SIGRA 60 F5?CENT 70 PDICENT 80 PERCENT
.3 1 .308999 .337417 .37401 I

.3 .2 .291067 .34.3064 .4! 5832
.3 .3 .261946 .A26258 .427476
.3 .232534 .305822 .421424
.3 :; .206706 .262530 .407261
.3 .6 .165022 .260952 .390244
.3 .7 .167000 .241768 .372776
.3 .8 .151974 .22A94~ .355961
.3 .9 .AJ9342 .2 i 0236 .J4G2i6
.3 1 .0 .128616 .197.333 ,325667
.3 .119418 .165964 .312286
.3 Ii .111455 .175889 .300006
.3 J .3 .104500 .166909 a266725
.3 1 .4 .096377 .156860 ~78346
.3 1.5 .092948 .151606 .268775
.3 I .6 08810J .145036 .259927
.3 1.7 .083748 .139058 .251727
.3 J .B .079818 .133595 244107
.3 J .9 .076252 .128582 .237007
.3 2.0 ,073001 .123965 J?30376
.3 2.1 .070026 .119698 2241 6E!
,3 2.2 .067292 .115743 .218344
.3 2.3 ,064772 .1!2065 .~!~E67
3 2.4 .062441 .1OG63S .207707
.3 2.5 .060278 .105432 .202636
.4 .413035 .441541 .477410

.4 :4 .403256 .458339 .532433
.4 .3 .379013 .4$4249 .561473
.4 .349261 .437678 .569971
4 :; ,3! 9492 .415575 .565313
.4 .6 .292127 .392076 .553272
.4 .7 .267876 .369238 .537551
.4 .8 .246695 .34?936 .520326
.4 .9 .228265 .328447 .502814
.4 1 .0 .212204 .310768 .485669
.4 .198151 .294770 .469222
.4 lj 165789 .280314 .453624
.4 I .3 .174855 .267207 .438923
.4 I 4 .!65128 .255297 .425113
.4 ! .5 .156428 .244443 .412158
.4 1.6 .148606 .234519 .400008
.4 1.7 .141539 .225415 .388609
.4 1.8 .135125 .217036 .377905
4 1.9 .129279 .209302 .367841
.4 2.0 .123930 .202142 .358367
.4 2.1 .119017 .195493 .349435
.4 2.2 .114490 .169305 .341001
.4 2.3 .110305 .183529 .333027
.4 2.4 .106424 .178126 .325475
.4 265 .1028I 6 .j 73061 .318315
B-V- 7 12 Jan 1984
MIL-H.DBK-L72
f)~tt?, 70th and 80th percentiles of the distribution

lognorrr,al
for means and sigmas from . 1 to 2.5 (continued).
~~ ~~~c~:~i EC ?~=c~t/T
9:> .
,5 .515518 .543941 .579215

5- .511835 ,55~17~ .642030
.
.5 .: .495415 .573439 .6ai72c
~,
- .4 .4665!91 .5,4557 ,705714
<
- .5 .436576 .547097 ,7]2464
.5 ., ,406623 .525254 .708736
.5 .; .378393 .5G1846 .698370
.5 .8 .352557 .478493 .684094
.3 .9 .329253 .45605G .66172C
.5 1.0 .s08369 .434920 .65G407
,5 1.1 .289687 .415246 .632878
. 5. 102 .272960 .397036 .615571
.5 1.3 257351 .380223 .598744
. 5. 1.4 .244440 .364711 .582540
.5 1 .5 .232236 .350393 .567030
.5 !.5 .221173 .337159 .552239
.5 I ,7 21}108 .324903 .538165
,5 1.8 .201917 .313546 .X? 4788
.5 1.9 .193497 .5029g5 .512061
.5 2.9 .185758 .293149 .500010
.5 2.1 .178Q2 .28396~ .488540
22 .172023 .275383 .477534
:; ~ 3 .165904 .267337 .46R51
.5 2,4 .160215 259763 .457376
.5 2.5 .154913 .252677 .447959
.6 .1 .617183 .645504 .680304

.6 .2 .617998 .674833 .748023
.6 .3 .604883 .6g7509 .798650
.6 .4 .562133 .6FC129 .83 I 664
.6 .5 .554030 .674565 .849315
.6 .6 .52389i .656516 .854957
.6 .7 .493830 .634874 .851 8R9
.5 .8 .465067 .611644 .842848
.6 .9 .438191 .5813114 .829908
.6 1 .0 .413412 .565060 .814561
.6 1 .1 .390 ?32 .542917 .797851
.6 12 .370043 .521904 .760489
.6 1 .3 .351190 .502J07 .7Q952
.2 1 4 .334001 .483536 .745556
.6 1 .5 .318306 .466152 .728504
.6 1.6 .303s49 .449895 .71 i923
.6 I .? 490784 .434694 .695887
.6 1.8 a2786g4 .420471 .660436
.6 1 .9 .267533 ,407153 .665585
.5 ~ 2 35Z32 .394667 .651333
.6 2.1 .247692 .302946 .S7669
.6 22 .23683 b .371928 .s24576
6 2.3 .239595 .361556 .S12032
.6 2 .4 .222909 .351778 .600013
,5 ~ 5 .215727 .342541 .588494
.....
?!:: -}l:; !;!:-.,:?
TABLE B-l-I.
50:F,,70:h and 8()t!? pprcentil~s of the Iognormal distribution

for means and slpmas from .1 cv 2.5 iCUIIL~iIUeCI).
rut! SIGMA 60 PERCENT 70 PERCENT 60 PERCENT
.7 .1 ,716373 .746569 .781022

.7 .2 .722570 .779571 .852016
.7 .3 .7JJ344 ,?97??9 ,909022
7 .4 .6953E8 .803159 .950660
.7 ,5 .670232 .7s7639 .977825
.7 .6 ,641427 .784357 .992571
.7 .7 .611206 .765936 .997449
.7 .8 .5BI076 ,744451 .994874
.7 .9 .5513 (J! .7214!6 ,986892
.7 1.0 .524368 .697872 .975127
.7 1.1 .498529 .674502 .9608! 5
.7 ! .2 .47450s .65I 733 .944B?2
.7 i .3 .452267 .&!9817 .927964
.7 1.4 .431717 .608888 .910570
.7 1.5 .412740 .589003 .893030
.7 1.6 .39521 I .570172 .875581
.7 1.7 .379003 .552370 .858387
.7 1 .8 .363998 .535559 .841557
.7 1,9 .350084 .519669 .825164
.7 2.0 .337159 .5i14704 .809251
.7 .325131 .490550 793gQ
.7 .313917 .477170 .77894B
.7 ,303443 .464S!2 .764568
.7 293641 .452526 .750696
.7 .284453 .441165 .73732 I
.8 .1 .819263 .847385 .881526

.0 .2 .826070 .883082 .954821
.8 .3 .S21 164 ,906002 1.016467
.8 .4 .806511 .916678 1.064867
.0 5 .7s4635 .916780 I .099959
.8 .6 .758026 .908498 I .122945
.8 .7 .726?86 .894042 1.1356?1
.8 .69052J .875355 1.139942
:; .9 .668344 .853997 1.137748
8 1.0 .638985 .831150 1 .130Q7
.8 1.1 .610877 .807670 1.119681
.8 12 .584254 .784153 1.106543
.8 I .3 .559210 .761OOI 1.0914[3
.8 1.4 .535753 .738477 1 .O751OI
.6 ! .5 .513838 .7I 6737 I .058066
.8 1.6 .493391 .69513R j .040652
6 1 .? .474321 .675921 1 .023110
I .8 .456531 .656893 1 .0056?8
:: ) .9 .439924 .638774 .986339
.0 2.0 .424408 .&1536 .971336
.8 2.1 ,409895 .605144 .954692
.8 2.2 .39630i .589556 .936445
.8 2.3 .383552 .574732 .922625
.6 2.4 .371578 .5@Q8 .907248
.8 2.5 .360317 .5472C12 .692322
12 Jan 1984
?lIL-t?DBK-A~?
TABLE B-i-l.
-
60th, 70th and 80th percentiles of the logncnnal distribution
for means and sigmas from . 1 co 2.5 (continued).
ME/it; SIGMA 60 PERCENT 70 PERCENT 80 PEI?CEHT
.9 .1 .919954 .947994 .981697

.9 .2 ,928s22 .985773 1.056874
,9 .3 ,92 ~&? 7 I .!212250 1 .! 22034
r .4 .91>629 1 .027526 1 .175669
:; .5 .g972] 7 I .032642 1.217311
.9 .6 .fj752G0 1 .029193 1.247496
,? .7 .845L>5 1.0189Q 1 .267407
. .
9 5 .~i 61.33 ! .C03651 ~ .27e538
.5 .9 .7S5637 .984775 1.202435
-
3 i .0 .755GZ6 .363523 1.280545
.9 1 .1 .72GG89 .940S46 1.274135
.2 1~ .07601 .917466 1 .264272
, ~, 1 .3 .6?03E2 .E93302 ! 25!627
.3 1 .4 .644542 .E170S25 1 .237530
c-9 1 .5 .Qolls .84759) I .22 I 64?.
.> 1.6 .597C96 .825270 i .2 C52E5
9 I*7 .575432 .603671 1.188163
.9 1.6 .555065 .782856 I .170733
.9 1.9 .535921 .762e54 1.153192
.9 2.0 .517925 .743673 1 .135688
.9 2.1 .Solooi .725304 1.116333
.9 2.2 .485072 .707727 1.10[210
.9 2.3 .47C)CJ6Y .69091 7 I .G6438G
.a . 2.4 .455925 .674843 1 .0 G7864
-
Q 2.5 .442572 .659473 1.051753
1 .0 1.020504 1 .048473 1.082180

1 .0 i 1 .03!035 ! .087893 1 .158430
] .0 .3 1.031779 1.117232 1.22=70
1 .0 1.023670 1 .136345 1 L84060
1 .0 :; 1.008139 J .145648 1.331084
1 .0 .6 .986831 1 .146879 1.367449
] .,? .7 .961365 1 .140839 1 .3938?0
1.0 .6 .933162 1.129174 I .411429
1 60 .9 .903469 1.113235 1.421364
1.0 1.0 .673152 1.094194 1 .424928
1 .0 i .1 .842916 I .073026 f .423285
1 .0 12 .813247 1.050508 1 .417471
1 C 1.3 .784468 1 .027244 1.406375
I so I 4 .756787 i .003692 I .396740
I J I .5 .730318 .980191 1 .383179
1.0 1.6 .705114 .956985 1.368189
1 .0 !.7 .6811F4 .934249 I .352170
1 .0 1.8 .656507 .912100 1.335441
1.0 1.9 .637043 .890 6! 5 ! .318256
I .0 2.0 .616739 .869640 1.300815
1.0 2.1 .597526 .649797 I .283274
1 ,0 22 .579374 ,830493 j .265756
1.0 2.3 .562183 .811922 10248354
1.0 2.4 .545?21 ,?94071 1.231142
I .0 2.5 .532510 .77632: ~02lt173
. - . . m,-, n
.1:[ -i{!)f{h-+;:
--
TABLE B- V-; ,
60ch, 70tt~ and 80th percentiles @f the logrtonnal distribution

for means and sigmas from .1 zo 2.5 (continued)..
SIGMA 60 PERCENT 70 PERCENT 60 PERCENT
1 .1 J 1.120953 1.14EE59 1,182403

J ,1 2 I .132652 1.169603 1.259642
1 1 .3 1.135756 1221273 J .329565
1 ,1 .4 1.130316 1.24JW6 t .390700
i .1 .5 1.117617 1.256921 I .442159
1 .1 .6 I .096976 1.262914 I .463793
1 .1 .7 i .075768 j .259956 I .5] 5972
1 I .8 1 .049282 1.251572 1,539436
1 .1 .9 1.020649 J .239213 I .55514!
I 1 1.0 .990198 1.222777 1.564127
1 .! 1.1 .960467 1.203613 I .567420
I .1 12 .930212 ] ]~251~ \ ,565963
1.1 ~ .3 .90044 I 1.160141 I .560676
J .1 I .4 .8?1437 1.136998 I .552250
! .1 1.5 .843389 1.113493 1.541346
1 .1 1.6 .8!6413 I .069934 I .528503
1 .1 1.7 .790573 1 .066554 1 .514168
} .1 1.8 .765892 I .043522 1 .496711
1.1 .742364 1.020961 1.462432
1 .1 ;:: .71996? 998957 1.465578
1 .1 2,1 .698665 .977566 I ,446346
I .1 22 .678412 .956624 1.430696
1 .1 2.3 .6591= .936747 I .413356
J .1 2,4 .6406Q .917343 1.395828
1 .1 2.5 .Q3461 .898tW6 I .378392
1.2 1 1 Z21326 I .249178 1382562

I .2 .2 I .234367 1 0291009 1.360609
i .2 .3 1 .239105 1 ,324622 1.432231
1 .2 .4 1,235996 1 .34966? I 496045
12 .5 1 225864 1 .36 G?134 1.551181
102 .6 I .209767 ].375017 I .597300
12 .7 1.188856 1.376665 1.634524
.8 1.164267 1 .372257 1.663328
E .9 1 .137039 1.3 Q746 1 .684418
12 I .0 1 .1OBO76 1 349129 1.698630
1.2 1.1 1.078124 I .332297 1.706841
1 .2 12 [ .0477S2 1 .313033 1.709913
1.2 1.3 1.017512 1 ,292000 1.708649
I .2 1 4 ,987661 1.269748 1.703777
1.2 1.5 .950477 I .246725 1.695940
i .2 1.6 .930134 1223288 I .665695
1.2 I .7 .902747 1.! 99717 1.673519
12 J .8 .876361 I .176229 I .659815
1 2 .851071 1 .! 52989 1.644922
12 ; :: .826824 1.130121 I .629123
12 2.1 .803Q?9 1.107715 1.612652
12 2.2 .781464 1.085834 1.595703
12 2.3 .760295 1.064523 I .578434
12 2.4 .740066 j .043306 I .5609?8
12 2.5 .720795 1,023 ?03 I .543438
I)-y-j 12 Jan 1984

60th, 70th and 80th percentiles of the Iognormal distribution
for means ant sigmas from .! cc 2.5 (continue(i).
sic!% 60 PZi?CEliT 70 PERCE;T 80 ?ERCCNT
I .3 .1 1.521641 1 .349445 1A82723

1 .3 2 I .335649 1 .392164 I .451395
i .3 .3 i J4i959 i .42742; 1 .5:4JE2
] ,: .< I .3408$0 1 .4547EC 1 .CO0414
1 .5 .5 1 .333C72 ] 47426E 1.658621
1 .3 .5 1.319353 I*&lgEtf i .798555
.
I .3 .I 1 .30CCS7 1.431326 1.750194
1 .3 .8 1 2780e9 1 4902G2 : *78377R
Ij
.* .9 i w2524&I 1 .4633 ?s i .&09618
1 .? I C 1 .224686 1 .4732 G 1 .528966
1 ,: 1.1 i .195475 1 .45E3>5 i gd{p~o
I ,3 1~ 1.IE5347 1 441gco i .S4!) 520
; .3 I .3 1 .1:5097 i .4224; :: i .S524CG
1 .: I .4 10104?16 1.4 C: 473 I ,621290
i ,3 J .5 1 .974?00 I ..3 V3J4 I .S46?96
I .3 1.6 1.045:71 1 ,:5 G~.L? 1 .=9480
1 .3 1- 1 .0163?6 1 .333GfJ5 I ,5?9630
1 .3 i :6 .S~9259 1.30s516 1.218271
1.3 1.9 .9Q455 1 .285963 1 .805166
I .: 2 .C .936615 1.26? 566 1 .79@E27
1 . 2 .~ .911754 1.239469 1 .775517
II
. .- 22 .867B74 :216772 1 .7554G0
1 .3 2.3 .864959 i .194502 J .74284i
1 .: 2.4 .842988 1.172726 1.725818
, .-7 2.5 .821953 1.151464 ] .708521
1 .4 .1 I.42I91O i .4496?2 1.482652

1 .4 2 1.436746 1.493179 1 .5 Q045
1 .4 .3 1 .444417 1 ,529600 I .~61S9
1 .4 4 1.445141 1 .559142 ] .7 C4033
1 4 .5 1 .43s404 1.581134 1.764826
I .4 .6 1.427887 1.595997 1.618044
1 .4 .7 1.411394 1.604187 106635I7
1 ,4 .8 1.390776 1.(3)6317 1.901359
1 .4 .9 1.366672 1 .6030?9 1.931906
1 .4 ] .0 1 .340465 i ,595277 ] .955650
1.4 ! .! 1.312257 1.533585 1.973!83
! .4 1W2 I 382e55 J.569714 1.935141
1 ,4 I .3 1 a252768 1.551290 1.992171
1 .4 1 .4 1222412 1,531872 I .994899
1 .4 1 .5 1.192121 1.510939 I .993911
I .4 I .6 1.IQ153 1 468903 1.989748
I .4 7 1 ,132705 1.466!06 1.962895
1 .4 ; :8 1.103923 I .442833 J .s73764
] .4 I .9 1.075910 1.4193!6 1 .9 Q794
1.4 2.0 1.048736 I .395744 I .950254
1 .4 2.1 1.022445 1.372267 I .936451
I .4 22 .997058 1.349004 1 .g~] ~o
1 .4 2.3 ,977>~4 I .32 C046 I .906001
I .4 2.4 .94SOi 6 1 .3034cfi 1 .t!a9 743
j G 2.5 .92t.~4~ 1281315 ! .873099
12 Jan 1984 R-t-f

for means and ~irma. frnm -1 tn ~ 5 (rnnt+n,,erl)
. . . .. . . ... ..--, ,
MEAN S I GMA 60 PERCENT 70 P5?CENT 80 PERCENT
i.> 1 J .522143 i .549867 I .582955

1.5 .2 1.537697 1 .594G32 1 ,66?589
1 .5 .3 I .546553 i .~184C I .737645
1 .5 .4 I .546661 i .66289Y 1.607068
1.5 .5 1 544995 1.687084 I .870057
1 .5 .6 1.535505 1.704517 1.926090
1.5 .7 1.521060 1 .715526 I .974920
1 .5 .8 1,502394 1.720600 2.016S47
1.5 .9 1.480246 1 i203J6 2.051173
1 5 1.0 1.455333 1.715321 2.079160
1 .5 1 .1 I .4283!3 1.706259 2.100972
1.5 12 I .399773 1.693762 2.117143
J .5 1.3 1+370221 1.678420 2.126232
I 5 1 .4 I .340066 I .660770 2.134800
i .5 1.5 1.309726 ;.641291 2.137391
1 .5 1.6 1 .27942% I .Q0400 2.1365!6
I .5 1.7 J .249417 1.598455 2.132645
1.5 1.8 ! 219870 1 .57576? 2,126207
1.5 1.9 1.190921 1.552576 2 11 7585
1 .5 2.0 1.JQ668 1.529110 2o107119
1 5 2.) J .135180 1.505538 2,095111
1.5 2.2 1.108502 1.462003 2,08162!
! .5 2.3 I .0826Q 1.458619 2.067479
I .5 2.4 I .057671 1 435478 2.052283
1.5 2.5 1.033530 1.412651 2.036404
1 ,6 1.&!2346 1.650036 1.683044

1.6 .2 I .638527 I .694771 1.763052
1.6 .3 1.648426 1.733607 1.838905
1.6 .4 1.652140 1 .766 I 63 i .909642
1.6 *5 1 .fxs958 1.792279 1.974508
1.6 .6 1.642327 1.812004 2 .032s 75
1.6 .7 J .629606 1,825571 2.084745
1.6 .8 I .613022 1.833356 2.129734
1.6 9 1 .592=9 1.835838 2.168042
1.6 1.0 1 .56927) J .633560 2.199918
1.6 1 J 1 543559 1.827069 2.225723
1.6 12 1.516052 1.816995 2245890
1.6 I 3 1.487240 1.803824 2260698
1.6 1.4 1.457577 1.788085 2.271241
1.6 1.5 1.427406 1.770242 2.277412
1.6 I .6 1.397043 1.750710 2379884
1.6 1.7 1.366737 1 ,R9856 2479107
1.6 1.8 1.336689 1.707995 2a275496
1.6 1.9 ] .307059 1.685399 2J269431
1.6 2.0 1277969 1 .6=300 2,261254
1.6 2.1 1.249512 1.638893 2,251273
J .6 22 l&21755 i .615339 2.239762
1.6 2.3 1.194743 1,591773 2.226963
1 .6 2.4 ) .166508 1.568305 2.2130s7
1.6 2.5 1.143065 I .545025 2.19S321
~-!J_c, 12 Jan 1984
?!IL- HD9K-472
TABI.E B-V-I.
60th, 70th and 80:h percentiles of the loRnormal distribution,

for means and sl~as from .1 to 2.5 (continued).
rlE&ii sIGnA 60 Pt??CENT 70 FmcCI/T 62 PERCENT
I .7 .1 1.722525 1.750185 1.783121

I .7 .2 I .739258 1.7954:7 I .e63649
1 .7 .J1 1 .750030 i.M5152 1 .939582
1 .7 .4 I .?55046 i S63022 2.011845
1 .7 .5 1 .7543CI 1 .29s846 2.078329
1 .7 .6 i .743462 1.915C) 2.138912
1 .7 ,7 I .737735 1 .9345a3 2.193264
I .7 ,8 1 .722746 ! .9447s: 2J241242
1 .7 I .704075 1.949850 Q ~P~g67
; .7 1:: I .622292 1.9501 w ;.318298
:07 1.1 1.6572G2 1 .94G2@a 2.347804
! .7 12 1.531W9 1 .YZ85G; 2.371 732
1
. ? 1 3 : .6037;C 1 .327523 2 .39a4r34
1 .7 1 .4 I .574(92 1.9!3795 2.404439
:* 7 1.5 i .544s25 !.ss7716 2.4; 4165
!.7 !.5 !.5!4724 1.879-/07 2.429008
J ,7 i .7 1.484358 j .6&3!3C 2.422-377
1.7 1.8 1.454046 1 .S3S303 2.421668
1.7 1 .9 1.423966 1 .817523 2.418309
1.7 2.0 1.394267 i. 795019 2.412580
1.7 1 1 .36505? 1 .7?200G 2 .4048!0
i .7 Z2 I .336425 I .74R6Q 2 .395279
1.7 2.3 10308435 1.725139 2.334233
1.7 2.4 10281134 1.7o15&~ 2.371898
].7 2.5 1254553 1 .67S039 2.35E!4R
1.8 .1 1.622684 1.850317 I .8531e9

1.6 2 1.839906 1.895387 1 963794
f 6 .3 1.851550 1.936513 2.040913
1 .8 .4 1 .857tM3 1.971545 2.113749
! .5 ,5 1.656359 2 .0008B8 2.161637
1,8 .6 1.853994 2.024500 2.244068
1.8 .7 1 .844939 2 .0424s2 2.300697
1.8 .8 1.631657 2.055053 2.551338
1.8 1.814650 2.0 S252S 2.335950
1.6 1:: 1,794434 2.065284 2.434622
1.6 i .1 1.771521 2.063755 2.467545
1.6 12 1.746400 2.05838G 2 .4s4931
i ,s 1.3 1.?19524 2.049630 2.517292
1.6 1.4 1.Q1306 2 .037s2 6 2.534814
1 ,8 1.5 1.6QI13 2.023694 2.547945
1.8 1.6 1.632267 2.007322 2.557075
1 .6 ] 1 1 .6U2042 1.969166 2 .5e593
1.6 1.6 1.571673 1.969543 2.564670
1.8 1.9 1.541356 : 948757 2.564250
1.6 2.0 1.511252 1 .92704! 2.561OE9
1.6 1046!4?1 1 .9046?2 2.555666
1 .a :2 1.4521~9 1.681 6;2 2.546269
1.6 2,3 1.423394 1 .!35S41 6 2.539152
1 ,8 2.4 1.395202 i .834932 2,528543
1.6 2.5 1.367647 1.611362 2.5i 6649
60th, 70th and 80t11 percentiles of the lo~normal dist:ibutior,
. - . . .. . .- - ..- - - ...- - - - --- . - -- - . . -- - - .--
mti SIGf14 60 PERCENT ?0 PERCENT 80 PERCEti7
I .9 .: 1.922826 1.950434 ! .s8324E

1 .9 .2 1 .940467 I .996493 2.064095
! .9 .3 1 .952856 2. C37721 2./a I725
1 .9 .4 1.959971 2. G73785 2.215407
1 .9 .5 1.961937 2.104485 2.264520
I .9 .6 1 959004 2.12975! 2.340576
~ 9 7 1.951503 2.!49640 2 ,4 C7220
1 .9 .8 1.939830 2,1643!3 2,460237
! .9 .9 1.924423 2.174021 2.507543
1 .9 ! .0 1.905745 2 .I 79083 2.549166
1.9 1 .1 1.884256 2.179867 Z .56%Z36
J 9 I .2 1.660415 2.176766 2.615959
1 .3 1.3 I .834642 2.170166 2.641605
1 .9 I .4 1.807336 2.160530 2.6Q481
1 .9 ! .5 ! .778854 2.148!89 2.678926
i .9 1.6 1.749516 2.133532 2.69i289
1 .9 1 .7 J .719604 2.{ 16905 2 .69992i
1 .9 1.8 J .689359 2.098621 2.705168
1.9 I .656968 2.076969 2.707362
-. ; :: I .fZ?8667 2.058202 2.706621
1.9
1 .9 2.1 I .598540 2.036550 2.703841
i .9 22 1 .~66727 i.0142i Z 2.698697
1 9 2.3 1.539325 1.9913= 2.691643
1s 2.4 1.510409 1.968153 2.6~9J0
I .9 2.5 1.482040 ].944?16 2.6727!C
2.0 .1 2.022953 2.050538 2.083301

2.0 .2 2.041006 2.036945 2.164360
2.0 .3 2.054049 2.138600 2.242437
2.0 .4 2.062071 2.175786 2.316861
2.0 .5 2.065175 2.2 U7704 2.3 B7052
2.0 .6 2.063556 2.234464 2.452541
2.0 .7 2.057500 2.256087 2.512975
2.0 .8 2.047339 2.272690 2.566120
2.0 .9 2.033462 2.284471 2.617855
2.0 1.0 2.01Q76 2L9105 2.6 S2167
2.0 1.1 1 .99=06 2.294676 2.701130
2.0 J.2 J .973662 2 029375B 2.734898
2.0 1 ,3 1 .949045 2 Z!89303 2.763682
2.0 1.4 I .922731 2.2 S1678 2e?B?740
2.0 1 5 I .895065 2.271244 2 .go7363
2.0 1.6 ~ .866363 2.250349 2.822857
2.0 1.7 1.836905 2.243322 2.834540
2.0 1 .B 1 .80s38 20226469 2.g42729
2.0 1 9 J .776675 2.208073 2.847734
2 0 2.0 1.746304 2.188388 2.849855
2.0 2.1 i .715979 2.167646 2.849378
2.0 22 1.685832 2.146052 2.846570
2.0 2.3 ! ,655974 2.!2378P 2. R416RI
2 G 2.4 1.626493 2.101016 2.834942
2.0 2.5 I .5974Q 2.077876 2,826567
~-f!-: 1 12 Jan 1984

?:; i. -Ili)i, f:-..;.
TABLE B-!-I.
61JthJ 70rh and flOtll percentiles of the lognonnal distribution

for means and sigmas from ,1 to 2.5 (continued).
rw N ~]Gm& 70 PERCEUT 80 PERCENT
2,i .1 2.123068 2.! 50633 2.183348

?*! 2 2,141479 2.137352 2 .26459S
9
*.. .J
, 2.!5512C 2.23976? 2.343067
2.1 .4 2.1U973 2.277524 2.41 g144
2,1 .5 2.lwl15 2.310599 2.489288
2.1 .6 2.1677il 2.3367~3 :,556049
?!
L.. .7 2.lf&?913 2.3619:7 2.618079
2.1 8 2.154253 2.380293 2.675133
9
G.! .9 ~*14j9:] 2 .39399s 2.727066
2. I 1.0 2.126089 2.403245 2 .77392?
2.: 1.1 2.107497 2.408305 2.815449
201 102 2.066154 2.409476 2,652039
2.1 I .3 2 OQ 733 2.407061 2.883759
2.4 1 A 2.037457 2.4 C142C 2 .910E23
~7! I .5 2.01C 637 2.392S16 2.953476
2.: 1.6 1 .982730 :.381503 2.>~!905
~.
.. . 1 i .7 I .953643 2.366423 2.966635
2., la
, . 1 .924282 2 .35307C 2.977712
Q,
L*. I .9 1.894268 2.336@19 Z .9 B5X15
2.1 2.0 I .8639S2 2.317521 2,990293
2.i 2.1 I .833619 2.297807 2.992346
Z.i 22 1 .8CU3~ 2.277067 2.99! 326
2.1 2.3 1.773127 2.255547 2.989277
9L.! 2.4 1 .743229 2.233354 2.984622
2.1 2.5 1.713660 2.21 G65G 2.978176
.1 2&!23173 2.250716 2 .2 S339C

2.241SC6 2.2977!9 2.364805
$ 2.256092 2.340642 2.443627
.4 2265704 2 .37g207 2.519265
.5 2270799 2.413216 2.591275
.6 2.271511 2 .44255? 2.659169
.7 2.268038 2.467209 2.722628
.8 2260632 2.467216 2.781400
9 2 .2495B9 ? .502702 2.e35323
1 :C 2.235233 2.513PA7 2.684316
1.1 2.217995 2.520665 2.928364
! .2 2.197953 2.524029 2.967586
J .3 2.17:719 2.52361~ 3.002049
I .4 2.! 51537 2.519932 3.031943
I .5 2.125721 2.513274 3.057473
1.6 2.0 S8~66 2.503943 3.078872
j .7 2.070343 2.492234 3 .09&388
1.6 2.041298 2.4?8426 3.110283
1 .9 2.011650 2.4627e4 3.120817
2.0 1.981597 2.445555 3.128253
2.1 1.951308 2.426966 3.132846
22 1 .920934 2.407227 3.134841
2,3 1.890603 2.366527 3.[34473
2.4 1.8604?4 2.365035 3.i31966
2.5 I .8304S2 2.342909 3.127527
12 Jan 1984
q11.-l{!)i\};-4?~
TABLE B-V-;.
60rtI, 70th and 80th percentiles of the lognomal distribution

for means and siQmas from .1 LO 2.5 [continued).
FIEAN SIGMA 60 PD?CENT 70 PERCENT 80 PERCENT
2.3 .1 2.323268 2.350796 2.363428

2.3 92 2.342296 2.396052 2.464994
2.3 .3 2.356S78 2.441434 ?.544120
2 .3 .4 2.367284 2.480678 2 .(20303
2.3 .s 2.373255 2.515590 Z .693049
2 3 .6 2.374999 2.546055 2.761956
2.3 .7 2.372684 2.572029 2.826701
2.3 .8 2.366529 2.593540 2 .867024*
2.3 .9 2.356792 2.6i0678 2.942753
2.3 1.0 2.343763 2.623588 2.993788
2 3 1.1 2.327740 2.632459 3.040099
2 3 1.2 2.309063 2.657517 3.061721
2 *3 J 3 2268026 2.639012 3.118741
2.3 I .4 2J?64947 2.637211 3.151293
2.3 1,5 2.240124 2.632390 3.179548
2.3 1.6 2213839 2. Q4826 3.203703
2.3 1.7 2.186355 2.614793 3,223978
2.3 1.8 2.157913 2.602556 3.240604
2.3 1.9 2,128733 2.588368 3.253819
2.3 2.0 2.099011 2.572468 3363864
2.3 2*1 2,068923 2.555079 30270978
2.3 22 2 .1338&?4 2.536409 3.275393
~ J 9
b 5 2.008249 2.5! 6545 5.~77334
2.3 2.4 1.977915 2.495964 3277014
2.3 2.5 1,947724 2.474520 3274639
2.4 .1 2.423355 2.450867 2.483462

2.4 2 2.442652 2.496356 2.565163
2.4 ,3 2.457790 2.542156 2.644578
2.4 .4 2.468734 2.582018 2.721218
2.4 .5 2.4755J1 2.6 J7754 2.794641
2,4 ,6 2,476210 2.649245 2.13644~
2.4 .7 2.476974 2.676435 2.930365
2.4 .8 2.471991 2.699334 2.992090
2.4 9 2.463491 2,716006 3 .0494e
2.4 i o 2.451?28 2.732568 3. I023Q
2.4 2.436979 2.743!80 3.150740
2.4 Ii 2.419533 2.750034 3.194600
2.4 I .3 2.399681 2.753352 3234003
2.4 1,4 2.377712 2.753370 3.269049
2.4 I .5 2.353906 2.750339 3.299877
2.4 1.6 2.328533 2.744514 3.326655
2.4 I .7 2.301646 2.736149 3.349573
2.4 1.8 2274079 2.725492 3.366835
2.4 2&?45448 2.712786 3.384657
2.4 ; :: 2.216151 2.698258 3.397259
2 4 2.1 2.186365 2.682127 3.4068~
2.4 22 2.15Q48 2.664595 3 413683
2.4 2.3 2.125940 2.645650 3.417936
2.4 2.095564 2. IQ6066 3.419626
2,4 ; :: 2.065228 2. f335401 3,419551
g_v- ] 3 12 Jan 1984

60th, 70th and 80th percentiles of the lognonnal distribution
for means and sigmas from .1 to 2.5 (continued).
SIGMA 63 PERCSJT 7CI ?iRCZI; T 8C FERCE::I
2.5 .1 2.523436 2.550933 2.5 R34?4

?.5 .2 2.5A29b0 2.532634 2.5 S5J17
2.5 .3 2.55B537 2.64281? 2.744985
2.5 .4 2.5700G7 2.52324: 2 .822043
2.5 .5 2.5775g9 2.713733 2. Z9GC76
2.5 .5 2.5~1174 2.752165 2.966721
2.5 .7 2.560944 2.780475 3.033669
2.5 .8 2.577063 2.304656 3.09 S671
2.5 .9 2.569730 2.524755 3.155540
?.5 ! .0 2.559174 2.s40862 3.210i49
2.5 1 .1 2.545644 2.653!10 3J!6C429
2.5 ! ,2 2.529402 2.861656 3 .30636.3
2.5 ! .3 2.510715 2.866723 3.347984
:* 5 1 .4 2.489852 2.668491 3.385357
2.5 i .5 2.4G7i177 2.867197 3.418622
2.5 ;.6 2.442645 2.863074 3.447890
2.5 1 .7 2.416797 2.856356 3.473333
2. 5 1 .8 2.389760 2.847277 3.495131
> 5 I .9 2.361748 2.83 S064 3.5! 3477
;:5 2.0 2.332954 2.822936 3.52e57i
2. 5 2.1 2.303556 2.808103 3.540615
2.5 2.2 22737]4 2.791764 3.549615
2.5 2.3 2.243571 2.774101 3.556370
2.5 2.4 2213256 2.755259 3.560474
2.5 2.5 2.182879 2.735465 3.5@.319
2.6 .! 2.623510 2.650S93 2.683523

2.6 .2 2.643282 2.698890 2.765458
2.6 .3 2.659225 2.743423 2.845354
2.6 .4 2.671297 2.784367 2.922790
2.6 .5 2.679509 2.821549 2.997375
2.6 .6 2.683918 2.954847 3.068765
2.6 .7 2.684LS28 2.884190 3.136664
2.6 .8 2.681780 2.909560 3.2 COI127
2.6 .9 2.675551 2.930985 3.261065
2.6 1.0 2.666144 2.948537 3.317242
2.6 1.1 2.653?83 2.962325 3.369275
2.6 102 2.638707 2.972491 3.41 7!30
~C6 1.3 2.621161 2.979204 3.460817
2.6 1.4 2.601394 2.982651 3 .5003e9
2.6 i .5 2.579653 2.983036 3.535s30
2.6 1.6 2.556177 2.980571 3.567556
2.6 1.7 2.531199 2.g75471 3.595406
2.6 ! .8 2.504936 2.967955 3.619636
2.6 1.9 2.477599 2.958237 3.640418
2.6 2.0 2.449373 2.946524 3.657931
2.1 2.420436 2.9330!8 3.67236S
;:: 22 2 i390950 Z.9I791O 3.683900
2.6 2.3 2A 61059 2.901381 3.6s2731
2.6 2.4 2.330894 2.683601 3.69? 041
2.6 2.5 2.300571 2.864726 3.703009
*V-14
MI L- Hi)Bti-d7;
TABLE B-V-Ii.
90th, 95th and 99th percentiles of the loRnomal distribution

for means and sigmas from .1 to 2.5.
.! .! ,2~7523 .27?1 12 . 4904?3

.1 .2 .??7300 .36C13f3 .F553A?
.[ .- * ,22!OF3 .Ypscp> 1.C7916&
.! .4 20C7(72 .3F6526 1.217?03
.! .5 1cF214 .3FIFF! 1.:SCC60
1 .6 :IF77723 .!.744C0 1.s6f:*l
.1 .7 17F3F0 .365$44 1.40?67s
. 1 7ioQ4 .3573/6 1.43PCII
.4I 62725 .3APP74 1.45P74fl
.1 5613? .340772 1.473312
.1 502!5 .3?3071 1.4e!57c
.! 44?59 .5257P3 j,4$@r47
.1 3?9e? .31F?CI 1.494154
.1 1.4 .135536 .312402 1.4sJ6277
.1 1.5
13~44p .306265 104C6F54
.1 I .6 .12767F .JOC461 1.496214
.1 1.7 1241RP .2Q4Q6p 1.4s4610
.1 1 .F :120945 .2ps76fi 1 .4c223g
.1 I : . 117?21 .2F4F16 1.uFS253
.1 2.P .115094 .2FOJ 16 1.4F5775
1 2.1 112444 .275641 1.4FI?C2
.
I 2.2 :IOC052 .271375 1.477713
.! 2.3 .107605 .?6730! i.4Y3272
.1 2.4 I053FF .?6:40? ;.46P62C
.1 2.5 .103290 ,25y67a 1.463P30
.
7
. .32770? .3FC063 ,536F23
.? :; .41!046 .536?24 .9P09F5
.-2 .3 ,4499s: .&61649 1.3F6603
.Z .4 .454601 .720766 1.71109P
.? .5 .45!04F .752131 1.s62764
.-
7 .6 .442165 .767367 2*15P353
.? .7 .43!120 .7730F2 2.3!2P36
.? .P 419404 .773053 2.434407
.? Q :40772$ .76?49$ 2.533017
.-
9 I :b .3F6427 .76376? 2.613321
.2 1.1 .SF564Q .7566F4 2.67?2?9
.? 1.2 .375445 .74P7Q! 2,733Q@?
.{.n [ J .365g22 .7404s7 2.77?355
.? I*4 .356760 .731FF? 2.F17359
.2 1.5 .34P227 .723246 2.?4S2S4
.2 !.6 .3401!?F ,714633 2.?76C22
.2 I .7 .33?.60F .706118 2.89F541?
.2 1 .F .32545! .60774p 2.91747S
.2 I C .31F684 .~p$550 2.933357
.2 2.C .312277 .6P1544 2,946625
.2 2. .3C6201 .673739 2.95765P
.7 2.; .3P0431 .666141 2.966739
2.5 .2$4>44 .65F751 2.97414fJ
:: ?*4 .2pa7J8 .6515K7 2.9F@094
.? 2.5 .2f4755 .6445P5 2.9F4762
R-v- 15 12 Jan 1984

MIL-t{
Dr\K-L72
TABLE B-i- Il.

-
?c:h, 95tk, and 99th percentiles nf the loj?, normal discri$utim
for means and si gmas from .1 to 2.5 (continued). -
m-A ! CC PEPCENT
.< .4314?1 .4F54f7 .5055ss

.; .542971 .67679] I,C2! 154
.. 1 .6J656P F.!4J36 1.47}47F

1
. .65740: .949!(I4
1.P?CC45
1
. . ,676336 1.02F166 2.255?11
.3 .6Fl?@l l.PPI 14:. ?.5f5f4c
.
x .6795~7 1.I15F7[ 2.e2Fp7p
.3 .672710 ].137e96 3.(150107
. .x .66324P 1.)51C50 L.2J74ss
.
x .6523?2 1.157S4i 503c7/~~
9.\ .64CP65 1.i6C366 :.533F22
..t .~29j ~6 1.15s579 3.65161C
.3 .6[ 73F5 !. 15646? 3.753622
3 .605F,64 1.1516KF 3.F423$5
. .1 .5$4641 1.145643 3.SIS9FI
. .T .58376P 1.13F73? 3.cPF~~7
. .\ .573274 1.1311P? 4,c47yi2
. .\ .56316F 1.1231?! 4. IC0P53
-
t .553450 1.[14s0s 4.14766(
.< T .544J[I 1.1064?6 4.IPS!6F
.
t 2.1 .535J4@ l,pa7p32 A ,7?f,~:F
* t 2.2 .526522 I,cRs191 4.?.5FF31
.. 2.3 .51F?41 1.rPo551 t.?FPC2C
.; 2.4 .5102E? !.07194s 4.314032
. .; 2.5 .50262? 1.0634!4 4.331?P3
.4 .53202F .fp]p]~ .6FFIOP

.e .65541? .77PI?.6 1.07:646
A .7532?5 ,CK(216 \.%13fp2
.4 .t?zzo?l 1.11244P 1.!6197[
A .e66214 102?2]9C 2.3FFF17
.4 .F91??7 !.3222?? 2.773206
.4 5@4a59 1*3C 1347 3. I17F76
.4 ,~~pzp] !.4415~2 3,422195
.4 .?075?3 1.47FOP4 3.6F??5?
.A .9@2096 1.504?63 3.c2557~
.4 .F?4039 I .5??53@ 4.,1331KI
.4 .pp4330 ~ ,534733 4.316(.6(
.4 .F735F? 1.54??61 a.47935P
.4 ,FC2240 1.546164 G.624@72
.4 .F5057? 1.547240 4.753212
.4 .?3FROP 1.54FIP5 4.F6FEI:
.4 .F27CC? 1.?43?.34 4.972600
.4 .F1545? \ .53P??5 5.cc6@33
.4 .F04037 [C%3~6F7 ~e)~p~fa
.4 7Q2F54 ) %?75p4 s.?266al
.4 :7i1935 ~o%pf17@J 5.2?5752
4 .7712$7 IC513367 5.35p5cp
.4 ,760$4F 1.505634 5.41573F
,4 .750F91 1.49759F 5.~67fc5
.4 .7411?4 !.4??3s7 ~.515312
12 Jar~i98i
II-v-lb
!4LL-I!DDK-L ?2
TABLE B-i- II.
90th, 95th and 99th percentiles of the lopnormal distribution

for means and sigmas from ,1 to 2.5 (continued).
v&t; ?Q PERCENT
c .63194:
0/ 01 .7772]2
5. .2 .76CCP9 1.13755C
.5 .3 .F7262? !.557551
c
. ., .4 .sf1634 2.CC51F2
G
., .5 1.C27C:4 7..45?463

. . . .C 1 .c7JJP?r 2.PFC576
c
. ., .7 1. IC4375 3.27$442
L
. ., .F 1.173C74 3.5451]e
. %
. 1.1329?7 ~*077~~l
c, I :: 4.277744
. . 1.i365r!
c
.. 1.1 1.13534P 4.54FF$C
. 5
. 1.2 i.1J@F31 4.7c36C1
. 5. [ 3 1.123902 5.C14637
.5 1.4 1.1157!5F 5.?14599
.5 1.5 1.105413 5.JQ5P32
.5 I .6 I .cs474f 5.5K0427
. 5
. 1.7 1 .0F35.~C 5.71 C226
c,
., I .F 1.P72POC 5.F46F44
j.Cf~2p6 5.S71K?5
:; N 1.C4P51C 6.6F6017
5 2.1 I .C36761 6ol?oFp2
. 5. 2.2 1.@25102 6.m??73
, . e 1.C13579
L .3 6.375??2
. 7. ~.d 1 .PO?226 6.45771?5

-. 2.5 .>$ ;o~p 6.533301
.,t- .1 .73!6?3 .R69P34

.c .2 .P62F42 1.21119s
.s w.T .:F3 12C I.610469
.6 .4 1.CF5S4? 2.04626?
.6 .5 1.16?OOC 2.496460
.6 .6 1.233137 ~.g42055
.7 1 .2FC7F0 3.5732er
:; .P 1.314FIP 3.7FOOSI
.6 c I .337S$6 4.15SF09
.6 I;6 1.352672 4.511422
.6 1.1 1.36C767 4.F35455
.6 !.2! 1.363F02 5.!33293
.K 1.3 1.:&2965 5.406740
.6 ].4 1.35s175 5.657756
.6 1.5 ].353143 5.8?F292
( 1.6 1.34541? 6.1002!4
:6 1.7 1.336427 6.295254
.-l( I F I .~264aK 6.4?4?Q?
.6 ]Cp 1.315PP2 6.640PF3
.6 ?.r 1.3C47F5 6.7s4200
.6 2.1 I .2$33~(7- 6.S36107
.6 Q-
G.& 1.2F1730 7.067644
.6 2*J I .26spFe 7.IF973F
.K 2.A 1.25F213 7.303220
K tic I ,2A~45F 7,d~FF33
f.
12 Jan 19$4
fj-~-17
TABLE B-V-II,
-
90th, 95th and 99th percentiles of the lognormal distribuclon
for means and sipnas from .1 to 2.5 (continued).
MEAN $] GF# 90 PERCENT 95 PERCENT 99 PERCENT
.7 1 .R3t41? .p754?9 .964527

.7 .2 ,963760 1.067007 1.29!44S
.7 .3 t.oe9006 1.264213 !.672466
.7 .4 1.201145 1.457C22 2.0s3129
,7 .5 \02969P@ !.6S7712 2.5567@4
*7 .6 1 ,375F20 1.F@1616 2.9Q762F
.7 .7 1.43tJ660 1.$467$4 3.43344F
.7 .P I ,4t!7353 2,073190 3.?65316
.7 *C ! .524Q34 2. IR1956 4.2776!5
.7 ! .0 1.550769 2.274?67 4.66724?
.7 !.! !.5694PK 2.3534~0 3.032>27
.7 I .2 1.5F!5!3 2.41?9S5 5.374563
.7 I .5 !.Y?F3Q2 2.475851 5.69~F2~
.7 1.4 1.5sI]02 2.5226FI 5.9FFF42
.7 I ,5 1.59049e 2.561753 6.263945
.7 I .6 1.5?7266 2.3941F9 6.51?5!!0
1.7 1.5P195F 2.620949 6.7571F9
:; I .F 1.5750!5 2.642F47 6.97F172
.7 1.566790 2.660576 7.jE3e55
.7 N 1.55756? 2.674722 7.375472
.7 2.! 1.54757? 2.6F57F4 7.534165
.7 2.2 1.537004 2.6?41F4 7.7209~1
.7 2.3 1.525994 2.?00284 7.@76F75
.7 2.4 1.51467P 2.704594 F.@2272!
.7 2,5 1.503J2R 2.706777 P.i5930?
.E .1 .931162 .974251 1.060533

.F .2 1.064056 !.1636?5 !.376219
.F .3 1.192355 1.36030? 1.7417?4
F .4 1.3toF35 1.5562S2 2.14?292
.8 .5. 1.416089 1.744592 2.5FO19?
F .6 1.5065S0 1.9~0432 3.027765
.8 .7 1.5?2S06 2.0?0917 3.47F6$6
.e .? 1.6441F2 2.224F96 3.923?40
.F .9 1.693677 2.352439 4.356?31
.F I .0 I ,732427 2.4643?? 4.773235
.F 1.1 1,762041 2.561?6! 5.170}79
.f? 1.2 1 .7F3PP5 2.646597 5.546412
F 1 .3 1.79951J3 2.719721 5.901525
.e 1.4 1 .FOP?IF 2.7R2694 6.235752
P 1.5 1.F15931 2.F3676F 6.549739
.F I .6 i.el~402 2.PP3065 6.P44~9@
E! !.7 1.F17977 2.922575 7.12C745
.? 1 .? 1.F15187 2.95616@ 7.379905
.? I .9 1.elo471 2.9F4599 7.6229F0
.8 2.0 1.F04191 3.00F525 7.F51056
F 2.1 1.796645 3.O?.F515 F.065175
.? 2.2 I*7$807F 3.045@60 F.266322
F. 2*3 1.77P694 3.C5F58fI e.455421
F 2.4 1.76F661 3.06?467 8.633332
.6 2.5 1.75FI17 3*07FO19 F.60@P5t
12 Jan 1984
B-V-18
90th, ?5rh and 99t~ percentiles of the lopnormal discrib~tion
for means and sipmas from .1 to 2.5 (continued).
~:~m~ 9tI PERCENT 95 PERCEPT ~g p~Rc~MT
.1 1.030931 ].073265 !.157422

.2 !.J64047 1.2606?7 1.464169
.3 1.294263 1.456252 !.61679?
.4 1.417!1C !.6s3454 2.208218
.5 1.52930! 1.P46392 2.629223
.6 1.62F913 2.@3057s 3.06940S
.7 1.7t5256 2.202172 3.519040
.? 1.7?F57! 2.359954 3.s69632
c 1.P49705 ?.503@oF 4.414433
1:6 I.FPSFSI 2.6S!44! 4.t148279
1.1 l,?40243 2.745FS1 5.267616
1.2 1,972227 2.$47311 5.670136
] S 1.S969P3 2_y3K7c7 6*C34>23
1.4 2.015593 3.0!54F? 6.429201
]Oy 2.02900? 3.@F44?F 6.767!33
! .6 2.C3E047 3.144E6S 7.C95653
1.7 2.043413 ~O!?7562 ?.406342
I .? 2.045703 3.24344e 7.69?939
2.04341F 3.283303 7.977272
;:: 2.0429F6 3.317F20 ?.239211
2.1 2.05e762 3.347610 F.4F6634
2.2 2.033049 3.37!)211 F.7204C7
2..3 2.026100 ~e>o?c!?p IT.g4J369
2.4 2.018129 3.4136F9 S.15C321
Z*: 2.00?314 3.429353 s.34F023
1.($ .1 1.130727 !.)72456 1.254932

I.c * 1.263F!85 !.35t!172 1 .55A424
!.0 .; 1.395320 1.552356 I .F?6169
1.0 .4 1.52121R 1.74??5! 2.275099
1.0 .5 1.63P543 1.945314 2.6F4115
1.0 .6 !.74524S 2.13474P 3.115062
!.0 .7 1.S40237 2.31477? 3.559686
1.0 .F ].?23267 2.4P320? 4.010~64
IOc 1.9046~3 2.G3FP@7 4.460544
I.(I I :: 2.05522P 2.7FII?.O 4.904925
1.0 I*I 2.105864 20?lf)2~7 5.339443
l.r J .2 2.147639 3.02675? 5.76115!
1.C 1.3 2.181603 3.13 !322 6.16F045
1.@ 1.4 2.2@F751 3.224$47 6.558PIJ4
1.0 1.5 2.229?93 3.30!z247 6.9S3015
J*n 1.6 2.246149 3.3F743P 7.290236
!.0 1,7 2.25793? 3.44e30@ 7,63066?
I*O I F 2.269993 3,50665F 7.s54661
1.0 ] .$? 2.270E5C 3.55F270 F.262731
1.@ 2.0 2.2750C3 3.603831 C.55548i?
1.0 2.1 2.272fl!$3 3.643565 F.F336C7
--- I ,0 2*2 2.270696 3,67g237 ?.0977?I
100 2.3 2.?66FF1 3.710151 9,s4?7s4
I*C 2.4 2.261662 3.737157 Q.5F7201
In ~.~ 2.255?S9 3.7606:? ?.8{3F20
_
)!1L-}[UBK-: 7?
TABLE B-V-II. -
90th, 95th and 99th percentiles of the lognor~l distribtiion

for means and sigmas from . I to 2.5 (continued).
FIE4N S 1 GFIh sO PCRCENT
I*I .1 1.230548 1.2717F2 1.35ZI!94

1.1 .2 1.363651 I*4559F7 1.646393
1.1 .3 ].495~61 1.64F741 1.g7P922
!.1 .4 1,623916 1.F45723 2.34676!
1.1 .5 !.745074 2.042640 2.744475
10/ .6 1.?572?1 2.235643 3.165563
I*I .7 1.9592F7 2.4215PS 3.603139
1.1 .? 2.050469 2.59E!125 4.C50524
1.1 .9 2*\30goR 2.763723 4.501666
I*I 1.0 2.2006(?5 2.$17525 4.951377
1.1 !.1 2.260751 3.05!232 5.39541F
l.! I .2 Z.slle!o? 3,1PR965 5.eso477
101 I .3 2,3>472(! 3.307!32 6.254@e4
1.1 1.4 2.390363 3.414330 6.6644F9
lot 1.5 2.41?56s 3.511263 7.060537
1.1 I .6 2.443116 3,5?~6p7 7.44[553
1.1 1.7 2.461714 3.677363 7,807232
1.1 2.415997 3.74R03!l e.t57552
1.1 ::: 2.4F6SS0 3.FI 1420 F.492?04
1.1 Z.@ 2.493F1O ~.F6f/174 F.U13026
}.1 2,49R274 3*9]~917 9.lle965
1.1 ::; 2.500303 3.?64214 ?.41!037
!*! 2.5C322F 4.064576 9*6FW01
1.1 ::: 2,49F33g 4.040479 9.95583P
!.1 2.5 2.494FQ4 4.07233R 10.20?73?
1.2 .1 1.330390 1,371212 1045II?7

1.2 .2 1.4633E5 1.5540?7 1.73S668
!.2 .3 1.5960F4 t.74543F 2.06432P
t.2 .4 1.7256P4 !.94!669 2.42239@
1.2 .5 1.l?49723 2.139120 2.R09670
1.2 .6 1.966252 2.S34377 3.220938
1,2 7 2.073906 2.524504 3.6!50483
1,2 .? 2.171PF4 2.707163 4.o$25~7
1,2 Q 2.259683 2.F80649 4.541647
1.2 J :6 2.337?99 3.043R44 4.992?21
1.2 l.! 2.4066(!6 3.!96]42 S.442162
1.2 1.2 2.466274 3.337344 5.8t?5910
1.2 I .3 2.517682 3.467567 6.321415
1.2 !.4 2,56156(! S.5F7146 6.746575
[.2 1.5 2.598641 3.696569 7.159F52
1.2 1.6 2.62?635 ~.7s6415 7.5601@2
102 1.7 2.655213 3.eF7310 7.S46891
1.2 I.F 2,675992 J.9~9~96 F.31961F
1.2 2,692533 4.044809 F.67F246
1.2 ;:? 2.705344 4.112663 9.022F45
1.2 2*I 2.714F76 4.I?4042 9.353627
1.2 2.2 2.721533 4.2294S1 9.670?10
I.2 2.!! 2.72567! 4.27951? 9.9750e3
1.2 2.4 2.727603 4.324597 10,2665?6
l.? 2.5 2.727KOP 4.365!54 10.5458F9
12 Jan 1984 B-V- 20

L
90th, 95th and 99th ~ercentiles

of che lognormal distribution
for means and sigmas from .1 to 2.5 (continued).
CQ P[p~~LT
.1 1.54976[
.2 1.?ss959
.-1 ?.151?40
,4 2.5@130c
. 5. 2.F75F44
,( 2.2PoP7fI
.7 2.7C2C4Ct
.P 4.13760$
Q 4.W2731
! :6 5.C32?CC
I*I 5.4P41!7
1.2 5,a3?cQf7
] .3 6.376~03
].4 6.?12281
lo~ 7,2Jp73~
I .6 7.654398
:. 7 6.05e2~s
I.F F.44s7@l
]*$ F.F2F452
2.C S.194131
..- 2.! $.546PC4
2.2! 0.PF65Ps
~1
&,- 10,213701
2.4 !O.52~42P
2*5 10.?31!13
i.4 .1 1.530127 1.64F530

1.4 2 1.6C2E3F 1.929C53
!.4 .3 1.7s6007 2.241C4P
1.4 .4 1.~27520 2.5?2P?P
I*4 .5 2.05541$ 2.951994
I*4 .6 2.17F012 3.344931
!.4 .7 2.293961 3.75?761
1.4 F 2.402290 4.l@625@
1.4 Q 2.5~23eP 4.62616F
!.4 1:0 2.5~3g6@ 3.@734@s
1.4 !.1 2.67CQ$3 5.52422!
!.4 I .2 2.751641 5.975256
1*4 I .3 2.PIP272 6.423623
1.4 ]OA 2.P7731$ 6.f166P?5
1.4 1 .5 2.$2!+2P6 7*303C9C
1.4 I .6 2.5747~7 7.730632
1.4 1.7 ~.o14124 8.14F311
1.4 t .? 3.@4pof7 F.55523P
1.4 I Q 3*@77046 ?.950761
1.4 2.C 3.lr153s ?.3344??
!.A 2.1 3.1?1ss0 ?.7C6?21
j.a 2.2 :.13FP II 10.065?$5
I,A 2.3 3.152376 JO.413444
1.4 2.4 3.163(26 10.749i2~
1.4 2.5 5.171070 11.873!10
12 Jan 1984
Y!!. -}{!)D}-L?2
TABLE B-V-II. -
90th, 95th and 99rh percentiles of the lognormal distribution

for means and sisrmas from .1 to 2.5 (continued).
MEAN s] Gmn 90 PERCENT ?~ p~Rc~NT 99 PERCENT
!.5 I !.6300!6 l.66?o32 1.747464

1.5 2 1.762S74 1.84966! 2.024796
1.5 .3 l.eswm 2.037257 2.331636
1.5 ,4 2,02792? 2.230515 2.666732
1.5 5 2.157104 2.4270F6 3.027997
1.5 .6 2,2F1827 2.624627 3.412649
1.5 .7 2.400803 2.F20923 3.t?17390
1.5 .F 2.513045 3,013sP6 4,23P612
1.5 .9 2.617P65 3.202121 4.6?2~9~
1.5 \ ,0 2.7J4F?55 3.3?3$54 5.115612
1.5 J*I 7.eo3e54 3.531?434 5.5643R1
],y I 2 2.FF4901 3.7?4FI? 6.015545
1.5 1 3 20$5@193 3.FF2612 6.466339
1.5 1.4 3.024042 4.0315F7 6.9! 4311
1.5 !.5 3.@F2842 4,171681 7.3573FF
1.5 1.6 3.133C33 4,3029P4 7.795?54
1.5 J.7 3.lFIOF1 4,425703 P.22232?
1.5 1 .F 3.22145? 4.540129 6.64J 727
I*5 1 9 3.256630 4.646606 9.05J225
1.5 2.0 3ei?e7044 4.74551e 9.450227
1.5 2*I 3.313126 4.837270 9.F3F325
1.5 2.2 3.335274 4.9?227? 10.215272
1.5 2.3 3.353e59 5.000934 10.5?0951
1.5 2.4 3.369223 5,07s657 10.935353
1.5 2.5 3.3F16F0 3.140?29 I 1.27F556
1.6 1.729916 ].7gs607 1.846532

1.6 :; ;.P62323 1,94?503 2.121C66
1.6 .3 1.9s5601 2. I?5021 2.423362
1.6 2.12E112 2.327250 2.732A57
I .6 :: 2.25@2B9 2.523141 3.t06610
i .6 ,6 2.3F4711 2.77061? 3*4P35P9
1.6 ,7 2.506163 2.9176F5 3.FF0608
I .6 F 2.62)670 3.112503 4.2~45?4
1.6 a 2.730503 3.2P3456 4.72226?
I .6 I :6 2.P3217F 3.41791P5 5.16039F
1,6 1.1 2.92643C 3.C6F597 5.6C5F07
1.6 I .2 3.013180 3.F4CF65 6.05552$
!.6 I ,3 3.09250p 4.0P5401 6.306833
1.6 1,4 3.164612 4.161F34 6.957372
1.6 1.5 3.22?7gl 4.30?s77 7.40497?
1.6 1.6 3.2FR365 4.44?793 ?.847FFo
1.6 1.7 3.340753 4,501367 F.2f4%3
I .6 I .? 21.3P7355 4.7C4F79 F.713F63
1.6 ] ,$ 3.42FSP5 4.F?C576 9.134743
2.0 3.464P54 4.92R758 9.54646S
;:: 201 3.496560 5.029756 9.94F475
1.6 2.2 3.5240?2 5.J23921 10.3403W
1.6 2.3 3.5477?0 5.211613 I0,721F56
1.6 2.4 J*567cp9 5.2$3195 I I .0c2tT24
I .6 2.5 3.5F5P22 5.SK9C?7! 11.453212
12 Jan 1984
FITL-iiDBK-&;~
TABLE B-V-Ii,
90th, 95th and 99th percentiles of the Iognormal discribucioc

for means and sigmas from .1 to 2.5 tcmcinued).
s] GFP ~@ p~pcENT 95 pFRr~~IT ?s PERCENT
1.7 .1 1.f2?F27 !.F6?3!F 10945710

!.7 .?. l*9520p6 2,047465 2.217773
1.7 .3 2.C$53~7 2.2!?.?F7 2.516C34
1.7 .4 2.Z2F155 2.424214 ~*p3p7jc
].7 . . 2.35s1(2 2.61S34F 3.IF745S
1.7 .6 2.4F6FFS 2.F 16547 3.557354
!.7 .7 2.6!0375 3.Q14P07 3.947076
1.7 .? 2.?2s613 3.210C3F 4.3539FI
1.7 a 2.F40F61 3.403!!7 4.775237
].7 1 :i 2,$465F5 3.5?1S23 5.2C7944
1.7 1.1 S,045443 3.775Z56 5.64?235
].7 1.2 3.13?267 3,952559 6,~$6367
le? I .3 3.222059 4.122?!0 6.54677P
1.7 !.4 3.29sF59 4.2P57Q3 6.9PF136
1.7 1.5 3.370s2? 4.440s26 7.44P35F
1.7 I .6 3.435493 4.5FFs?)3 7.E95625
1.7 1.7 3.4?3?Ff! 4.7??905 F*33F373
1.7 ! F 3.546450 4.F?S71R 6.775282
1.7 1.9 3.5?3544 409p393] 9.205264
1.7 2.0 3.635536 5.l@075s s.627437
1.7 2.1 3.672792 5.210460 lC.041102
1.7 2.2 3.7C5666 5.3!3325 10,445728
],7 2.3 3.734500 5.4C96&0 IG.P40?22
1.7 2.4 3.75?617 5.40?7F2 11.226415
1.7 2.5 3,7F1324 5.5F4013 11.602041
I*P .1 I .92S746 l*9&po59 2.044?F0

1 .F .2 2.061F62 2.146530 2.314P44
1 .F ,3 2.j9507p 2.331130 2.6C9502
I .F 2.32EOS3 2.5?.1$?4 2.92F33~
I.F :: 2.45?63? 2.715733 3.270231
1 F .6 2.5FE!525 2.9125C3 3.6335S7
I.F .7 2.7136F? 3.! 10079 4.016462
I.E F 2.g34zzo 3.3@690? 4.416555
log e 2.S4937F 3.50!566 4.E31407
1*F 1:6 3.C5F602 ~c~027fJ4 5.25F445
I.R !.1 3.161497 3.F79476 5.6s50F5
1 ,F 1 2 3.257F26 4.QK0745 6.13F?06
1.F I .3 3.3474?2 4.235PF0 6.587213
I.F 1.4 3.430512 4.404345 7.C380P0
1.? I .5 3.5@69?? 4.565766 7.4e93et
1 .F 1.6 3.577142 4.719?0s 7.939308
!.F l.? 3.6411F4 A.P66663 P.SF6257
I.F 1 .P 306$g410 5.006017 ?.F2PF65
I.F I .9 3.752129 5.!3R043 9.2C5S5F
1.F 3e7@9662 5.262FFI 9.696557
I.F ::: 3.F42339 5.3F@71? 19.l~9F63
I.P 2.2 3eFF04p6 5*4$17g3 19.535232
I.@ 2.3 3,$]4422 5.5$6327 19.942!61
l.!? 2* 3.?44453 5,6?4622 11.340273
l.F 2.: 3.972F75 507F04Q 1 [.729294
Yii.-l-il)! ?.-+;:
TARLE B-,-
II.
90th, 95th and 99th percentiles of the lognormal distribution
for means and si~mas from .1 to 2.5 (continued).
mAN S 1 GMA 93 PERCEPT 99 PERCENT
]*: 01 2.029672 2.06FF27 2.144327

1.9 .2 2.161652 2.245683 2.412223
I*9 .3 2.294F03 2.42?432 2.7P3645
1.9 .4 2.427?5C 2.616776 3.OIPIOI
1.9 .3 2.559c6F 2.1312304 3.J54667
1.9 .6 2.6P9744 3.oog547 3.712012
1.9 .7 2.F 16297 3.206033 4.C)FE456
I*9 F 2.95t75e 3.403S42 4.462047
1.s a 3.C56402 3.59?144 4.$9C622
I*$ I :6 3.16P655 ~,7$22~4 5.311902
log I*J 3.2750P7 3*51#lj554 5.743564
1.9 I .2 3.375412 4.166206 6. IF3305
1.9 I .3 3.469467 4.345450 6.62E?O0
3.557202 4.51F704 7.07F24$
;:: M 3.65?660 4.6P5530 7.329403
l,? 1.6 3.713962 4.845625 7.9F05F7
!.9 1,7 ?J.7F32P6 4.99QFO0 F.430213
j,~ 1 P 3.C46~57 5.144?69 8.P76P76
1.9 I .9 3.904933 5.2F412S S.31935F
1.9 2.0 3.9577s1 5.416346 s,7566!3
I.g 2.1 4.005720 5.541744 1001t!7760
1.9 2.2 4.049016 5.660490 10061206~
1.9 2.3 4.0F?g72 5.772782 11.02F943
[.? 2.4 4.12287S 5.e7f1845 11.431912
{.9 2.5 4.154005 S.S7F920 11.83F60?
2.0 ,1 2.129604 2.16?618 2.24373S

2.0 2.261455 2.344912 2.509P64
2.0 :: 2.394520 2.527P73 2.79P364
2.0 .4 2.527771 2.716343 3.1oee4P
2.0 .3 2:660141 2.909063 5.440546
2.0 .6 2.7?0639 S.!04712 3.792337
2,P .7 2.918344 3.301960 4,162786
2.0 .F 3.042436 3.499503 4.550199
2,0 3.16220? 3.696105 4.952667
2.0 1:: 3.2770F7 S.F~Q620 5.36F230
2.0 1.1 3.386621 4.0$2055 5.794741
2.0 1.2 3.4904F6 4.269495 6.230124
2.0 I .s 3.58F472 4.452200 6.672324
2.0 1.4 3.6PC474 4.629556 7.119371
2.0 1.5 3.766475 4.F@loFl 7.56S411
2.0 1.6 3.E146535 4.966416 F.020727
2.0 !.7 3,920775 5.125311 8.471755
2.0 I P 3.9P?366 5.277614 e.92]oee
2.0 I .9 4.052515 5.423256 9.3674g3
2*P 2.0 4.110456 5,562241 ?.P09F50
2.0 2.1 4.163439 5e69463~ \oo~47251
2.0 2.2 4.2117?? 5.t??0534 10.67FFe6
2,0 2.J 4.2555P5 5.940099 11.lo40e7
?.4 4.2?5276 6.053505 11.522302
;:: 2!.5 4.35106F 6.160fJ4P 11.933(W7
12 Jan 1984
R-V-7.4
HIL-ni)fJL-i2
TABLE B-V-T[,
9f)t!7 , 95cII and 99th percentiles of the lo~normaj distribuclon

for means and si~as from .] to 2.5 (~~ntinued),
2.1 t 2.22?543 2.268427 2.3432@7

2.1 :; ?,361270 2,4442Fs 2,60772?
9 ,
. . . .3 2.4?42s6 2.G2645P 2.6S35FG
7
~.i .4 2.62754F 2.F14CFI 3.~OC446
9,
Lot .5 2.760195 3.(06CPL L.5276P6
2.1 .6 2,F$12F() 3.?010?! 3.F74346
-<.! .7 3.019946 3*3~7:2) G.25Q196
:.1 P 3.145418 3.5?7509 8.520765
?.1 o S.267@lF 3.7~26JF 5.017397
2*! I :; 3.3F4176 3.9p~242 5.4272>s
2.1 1.1 3.496432 4. IF134C 5.64F5?7
7.1 1.2 3.603436 403710K7 6.2733F7
2.1 ] ,3 3.704?41 4.556676 6.717781
:.1 1.4 3.EP0797 4.737536 7.15194C
?.! 1.5 3.F?os4@ 4.?13135 7.610113
2.1 1.6 3.975376 5.n9sc17Q P.@60654
2.1 1.7 4.054176 5.247043 P.512040
.->
c. 1 1 .F 5*4p4p8p
4.127461 P.962FP3
2.1 1.9 4.]~53a2 5.55636! ?.411?34
2.1 2.0 4.25e16(l 5.70!526 ?.P5FOP0
2.1 ?.1 4.315979 5.FA@355 10.300343
?.1 2.2 4.369076 5.~7ppQ7 10.737F74
2.1 2.3 4.4176$0 6mp>$~59 ! 1.169?4P
?.1 2.4 4.462C60 6.21?57C 11.5S5?4F
291 2.5 4.50242F 6.*>JJo~g 12.015363
2.2 .1 2.3294F6 2.36F253 2.442724

-J*
L.*1 .2 2.46]0?7 2.5435G4 2.70S7~9
2.2 .3 2.59~9?l 2.725115 2.9P?227
2.2 d 2.72?302 2.911974 !I.2927P6
2.2 .5 2.F60160 3.103121 3.61592?
2!.2 .6 2.??1722 3.2?74F3 3.957F44
2.2 .7 3.121191 3.4$s?56 4.31746E
?.2 .? 3.247636 3a6aj4~5 4.693522
2.? c 3.371OOF 3.FFF8fz? 5.0F4546
?.2 I:i 3.49014F 4.0P52F0 5.4FFP51
2.2 J.] 3.60479& 4.27$692 5.$05(!53
2.2 1.2 3.714W3 4.4712P5 6.331]27
2.2 ] J 3.?1S237 4.659322 6.765437
2.2 1.4 3.91e574 4.F4316r5 7.20627F
2.2 1.5 4.0124F6 5.0222R6 7.652004
?.2 I.K 4.100938 5.l?625fI F.101049
1.7 4.1F3956 5.J~4720 8.s51944
1 P ~.261616 5.527446 9.C03333
4.334037 S.6F4257 S.453?76
;:; 4.401371 5.FS504Q S.902754
2!.1 4.4637?4 5.97:762 IC.34F66F
?.? 4.52!501 K.IIF4K5 IC.7SOF35
2. 3 4.574703 6.25]]5] 11.228486
2.4 4.C?36!5 6.>77~~,~ I 1.6C0955
2.5 4.66?4S3 6.49F9~6 12.0F7674
12 Jan 1984
}Ili--l{[!i{ t:-+ .-.-
I-
TA3LE E-I-::. -
9(lth, 95th ad 99th percentiles of :he lo~normal distribution

fOr means and sipmas from .1 :0 Z.5 (conticud)
. .
FIEAN SIGMA ?(I p~Rc~NT 95 PERCENT 99 PERCENT
2.3 .1 2.42943!! 2.46t?094 2.3422t4

2.3 .2 2.560934 2.642971 2.RC)4C17
2.3 .3 2.693754 2.F23F90 S.C8924~
2.3 .4 2.e27041 3.010009 :.3F5?75
2.3 .5 2.96005F 3.2@0404 3,705145
.6 3.0F2001 3.39410! 4.042662
::: ,1 3.22Z147 3.590093 4.s97407
20 3 .? 3.349791 3.7F7374 4.76F26U
2.3 9 3.474S!8 3*>p495g 5.15593?
2 3 1.0 3.5?5!88 4.lf1905 5.S53021
2.3 l.! 3.711939 4.377334 5.964001
2.3 1.2 3.F?4197 4.570438 6.3F5313
2.5 1.3 3.931671 4.760495 6.815366
2 .s 1.4 4.034)47 4.$46873 7,252984
2.3 1.5 4.1314F7 5.129030 7.695425
2,3 1.6 4.223619 5.306513 8.142406
2.s ),7 4.310531 5.47P955 F.502]23
203 1 F 4.3$2259 5.646072 9.043260
2.3 1 9 4.46eeF!3 5.$07653 $.494601
2.3 2.0 4,540519 5.963553 9.945033
2.3 2.1 4.607309 6.I136FF 10.393549
2.3 2.2 4.66S417 6.25F025 \o.g3924e
2.3 2.3 4.72?024 6.JJQ6577 11.2F132?
2.s 2.4 4.7F0320 6.52?3?2 11.7190156
2.3 2.5 4.F29503 6.656553 12.15191S
2.4 .! 2.5293F5 2.56794R 2.641F@0

2.4 2.6607@l 2.742424 2.902393
2.4 :: 2.7?341?5 2,$22754 3.IF1599
2,A .A 2.926770 3.loe173 3.47$336
2*A .5 3.059905 J.2Q~044 3.795221
2.4 .6 3.102169 3.490875 4.12F656
2.4 ,1 3.322871 S.6B6S43 4.4781?41
2.4 .F 3.45136T 3.?F333P 4.P44?96
2,4 S.577066 4.0F092R 5.2253?3
2.4 1:: 3.699445 4.?7F240 5.6!934!
2.4 1.1 3.fft6051 4.474439 6.025312
2.4 I.2 3.932904 4.66F755 6.441@76
2.4 I .3 4.042502 4.8604F6 6.8675!!0
2.4 1.4 4.147611 5.04?OOP 7.300966
2.4 1.5 4.241F26F 5.233777 7.140597
2.4 1.6 4.34S769 5.414S27 8.lf35074
2.4 1.7 4,43426P 5.5S0271 ?.633062
2.4 J e 4.519770 5.761297 9.DF3294
?.4 lea 4.600320 5.927162 9.5345!$s
2.4 2.0 4.6759?ff 6.0F76FF 9.9FW41
2.4 2.1 4.?46918 6.242751 10,43605(!
2.4 2.2 4.F13211 6.3Q22F3 10.S~4324
2.4 2.3 4.e75033 6.53625P 11.329F21
2.4 2.4 4.9S2547 6.6746F9 li.77fe20
2.4 2.5 4.91?3930 6.F07625 12.209677
1? ;ap 19?4 ~,--. f-l

?IIL-HDBK-L72
TABLE E-V-::,
90th, 95th and 99th percentiles of che Lognarrdl distribution

for means and sipmas from .1 to 2.5 (continued).
SI GYA 90 PFPCENT os fJFf?cFNT 09 PERCENT
2.5 .1 2.62934P 20667F]~ 2.74150F

2. . 5 .?. 2.760&~7 2.P41sl~ 3.oo@990
2.5 ..7 2.F:3245 3.P2!6?? 3.27?240
2*> .4 3.P2&4?G 3.?n~455 3.5734C2
2.5 . 5. 3.15:713 3.3c542c 3.FfC(?6@
?.5 f 3.2c?232 3,5F7FC2 4.2156?F
?.5 .7 3.423407 3.7FP.733 4.561619
2.. ~ .r 3.552631 3.>70367 4.522y59
2.5 3.679345 4.176F57 5.29?711
2.5 1:: 3.??3C44 4.374:7F 5.6P774e
?.5 1.1 3.~232p~ 4.571145 6.0FFE4S
2!.5 I .2 4.03?693 4.76K421 6.500724
2.5 1.3 4.151?53 4ep5~53p S.922045
2.5 1.4 4.259F20 5.14?FK0 7.35146F
?.5 1.5 4.36310P 5*336p6$ 7.7f7655
2.5 I .6 4.461690 5.>2@C?6 P.22?2?3
2.5 1.7 4.5554QS 5*~9clll F.675!15
2.5 I .F 4.6444?1 5.F73610 S*123?I0
2.5 ].~ 4,72F69? 6.04331G S.574531
?.fl 2.0 4.P@F16$ 6.20F020 10.02S909
-.
2.5 2.1 4.ps29el 6.3C757@ 10.477052
2!.s 2.2 4*?53p44 6.521862 10.927051
?.fl .-.
*,.5 5.01?OP3 6.670i3S 11.375079
2.5 2.4 5.0F@640 6.F144F0 11.F203FF
2,5 2.5 5.13F070 6a95~Fol 12.262313
2.6 .! 2.72929P 2.7G76FF 2.F41165

2.6 .2 2.F605C2 2.Q4144V ;.pa95zI
2.6 .3 2.gs30j4 3.i207i4 3.375139
2.6 .4 3.126220 3.304P4? 3.667?17
2.6 .5 3.2594Q4 3.493150 3.9775F0
2. 6 .6 3.3$?217 3.6F4F7K 403036FI
2.6 .7 3.5237FR 3.F7?25C 4.64560$
2.6 .? 3.653637 4.0754F4 5.0026@l
2.6 Q 3.7E!!231 4.2727PF 5.373764
?.*K I :6 3.9(i60FF 4.470394 5.75FOFF
2.6 1.1 4.027777 4.66756P 6.154472
2.6 I .2 4.145924 4.F635F4 6.56174g
2.6 I .3 4.?.60213 5.C57F1F 6.97F69$
2.6 ]*A 4.3703FP 5.24$664 7.4040FF
2.6 !.5 4.476249 5.43F590 7.v36670
2.6 1.6 4.57764? 5.624122 fi.27521F
2.6 1.7 4.6744s] 5.FOW51 F.718534
2.6 1 e 4.766724 5*9p3431 s. 165462
2.6 4.P54337 6.156573 SI.6i4901
2.6 ;:: 4.937354 6.325C46 lC.0G5F12
2.6 2.1 5.(15831 6.4FF671 10.517223
2.6 2*2 5.0F?F4!? 6.647317 10.?6F234
2.6 2.3 5.15?507 6.FP@Fs7 11*41FO17
2.6 2.4 5.?24927 ~6~A?~~l 11.B65F17
2.G 2.5 5.2B6241 7.0!26?5 12.310950
-
MIL-WW-472
-.
APPENDIX C
ACCURATE COMPUTER Hm., (4) ANALYSIS
10, This section describes a method for predicting M~a= $ when an

accurate representation of the overall repair time distribution is desired.
The prediction requires that a distribution of time for each maintenance
element (i.e., preparation, fault isolation, etc.) be ancnrnor assumed,
The prediction ts general and can be applied to any definable distribution or

combinations thereof, however, the complexity of computing the overall
distribution Increases proportionately with the complex~ty of the maintenance
element distributions, A simplifying assumption cac be made that all
maintenance elements have normally distributed times. This slmpllfylng
assumption !s reasonable since each maintenance element Is the sum of many
inde~endent task times, e.q. the maintenance task preparation may include
timefor equipment war&up~ acquisition of necessary t&ls, etc. by the
central llmlt theorem In statistics, the distribution of the maintenance
element task time approaches a normal distribution as the sample size of
Individual task times increases. %asedon this assumption the detailed
procedure has been developed and a computer program written for computing the
desired IIL..(0). Programs based on other distributions of maintenance
elements can be stmilarly developed and programed.
-. 10.1 General a~proach. In the general approach, we have a system with total
failure rate X,. and with N x J possible repair tyges with random repair
times Rnj, n-l, .... N.J-1, h.., J where Jis thetotai number of unique
fault detection aud ~solation outputs and N is the total number of repairable
iterns. Let A., be the failure rdte of that portion of the n
repairable !tem which is covered by fault detection and Isolation output :
Further, let f, (t) be the probability density function for R.j. n=l. .... N,
n~
~-i, .... J. It !s assumed that f. is continuous and concentrated on [O,W).
WJ
If T 1s the system repair time, then its density functtcm gt(t) (sfnce the
events T = R.j are mutually excluslve) 1s:

g, (t) = I Pm, f, (t) (1)

n~
where:
12 Jan 1984
r-v-l
t41L-HOBK-472
The mean system repair time is

-
(2?
where
p, ~,= E(R.l) = mean repair time R.j, and the variance of the system
repair time is
(3)

2 z
= z P., (0, + p, ) - p?
n~ n~
where
p. 2 = var\ance of the repair time R.J.

n~
Values of Mma. (4I)are given as solutions to tne equation

Mma (~)
M man (0) gT (t) dt = 1 Pmj
f (t) dt=o (4)
o n.J
which are not, In general, un~que. Sufficient conditions for the existence of
a unique solution are that f.,(t) > 0 for all t > 0, n-l. .... N, ~=1, .... J
and that each f.j(t) be continous, conditions easily met in practice.
Equation (4) can easily be solved, under these sufficient conditions,
by using Iterative means on a computer.
10.2 Assuwtlngnormal densltles for the R.,s. In practice, Rfij,n-1, .... N,

j=l, .... J are sums of several independent repair element times which
are themselves sums of a large number of independent repair task times. An
application of the central limit theorem suggests that the denslttes fR
are approximately normal. -j
Specifically, the density f. will be (approximately)

-l
f.,,(t) =
++-- {-( ~:
)21 -
12 Jan 1994
c-v-2
tI!IL-HDBK-472
~- where
pa and OR 2 are the sums of the elemental repair time

n.
J .
means ano variances, resr)ec:ively, Presumably, u, artdo. z

J *J
wil be such that the normal density is, approximately, concentrated on the
pos t!ve real axis, i.e.,
r
(5)
\ J
1 IJ
J
If we let n(t) = ~2fi
1 f e-x/2
dx, then equation 4 becomes
m
which will have a unique solution for all O where O c 4 < 1. The advantage
here !s that only one dens!ty function need be p?ogriwnnedin order to cacuiate
(4) using a computer.
10.3 Computer program. A computer program listing is provided In Table C-V-I for
Derform~ng the normal case described above. A sample input/output for the program is
shown in Table C-V-II. The resulting distribution for the example is shown in Figure
C-V-I .
The means and variances for each repair element which makes up the individual repair
times R., are inputted. p. and o, .,2 are then computed
IJ
and equation (6) is solved for Mm,, (0) for the given O using the secant
method. The secant method solves equations of the form
f(x) * o
141L-MDBK-472
by forming the sequence (for n-~. ?, ...) -
after choosing xc and xl as startin9 points. The sequence is terminated

after the desired accuracy is reached. Several points concerning the computer
program deserve discussion.
First, no lntegra~lon is per$ormed per se in the calculations of
Instead, the followlng approximation is used
-x2/2
f
n(t) =
+-
. ,_ (&i e->/ [b,A+ b7A2+b3A3+ b. A6+bsA] +,(t)
where
lc(t)i~7.5 x 10 for all t and
A = 1/(1 + 0.2315419t) with the b,z g!ven by:
b, =0.319381530
b, = -C.356563782
b, _ 1.781477937
S. = -1.821255978
b, - 1.330274429
c-v-4
MIL-HOB~-472
second~y, the user must provide two initial guesses to M.,. (0) denoted by
XO and Xl in the computer program. It is essential that XO not equal xl since
this would cause zero divides in the Pro9ram. The be$t way to pick Xo and
xl is to estimate an interval which ~m,. (@) ui1 lie. Then, select XO
and Xl as the enopoint: of the interval.
Final~v, although tne Present discussion deals with double subscripts n and j,
t4e d~~tinction~ indicated by these subscripts and independent of the
calculations performed. Hence, the progran uses the data it?:;ngle
dimensioned arrays of length N x J.
The input data is read in the following order:
XO <Initia esirnation), Xl (Ifiitia! estimaticm;, HI ($), LT (Total

system failure rate
N1 (Number of e ements contributing to firjt R). LAMBDA (1) (Fai~ore rate)
W, SIG2 (mean, variance for first element)
Mu, SIG2 (mean, variance for 2nd element)
. .
. .
. .
W (Number of e!ef!!ents
ccx?trl!?utingte second R!, LAMBDP, !2) (fai!ure rate!
MU, SIG2
. .
. .
. .
The following condition must be met:
I LAMBDA (I) = LT
All I
Sample input/output. and program listing follow.

TABLE C-V-I. Listing for camputer program to cqxste M -(0
vhen eiesoental maintenance activities are normally distributed.
00010 lrrLICITRFAl*8(A-M~, RE~L*9(0-7~

00020 RML*BR@,S16X,W,SX:2,~.L~~M,LT
00030 conRoN W, LAM#M(loo ).M(loo),sIsi*flw).l
TDIA
noo40c IO,X1AM INITIAL GUESSESTO MAX(PN!).LT!STOl&L
,00s0c F41LuR[ KATE oF;fSTEM.
00040 R[hP15, @)XO,X1,FNI,L1
00070c ERR=RaI ERR(ISIII
RMA1(WI)
O*OSO tRR=U.00W0
Ooo?o J.O
00100 30 J,$,l
00110 BEAO(S, *,fNLI=l O)NN,L@lIW(Jl
00120 LhNBDh(J)-%MDDA( J)/LT
00t30 WF(.ll=o.oo
00140 SIG2R(J):o. BO
001 so 8020 S=1,N9
Oolboc NO ISTHEWNSSROF [LEKNTSTOFaLWI.
Oot?oc LMSRA(J)ISTJIE r41LLJREMTE Of TWEREPL6CEARLE ITEff
00100c UNOS[tEPAXRIlnEISMOE U~ OF W U[NH7S UiICfi FflLIOU.
00I?0 RfaD(S, *)Ill J,SIG?
00200c mu Is TNEHff3h, sm rs wf wwaw or[MM ELEFKMT.
0021C MURIJ)=lf~(j)+MU
00210m SIG2R(J)=SIG2R (J)+S162
00230 60 1030
00?40 10 lTOTAL=J-1
002s0 ru=xl
002W xHtl=xo
00??040 :N1=XN-(XN-XNN1 )*f(IN)/ (F~XN)-F(XtWl
ii
002s0 URITE(6,2) Ml
00290c HRAX(FHI) ISPRINTIO ATEACHITERATi ON.
~o:ao ]ff YfiPs(IN1-xffI .Lf. [W) GO TO58
00310 xilMIxxN
003?0 lM=lU1
00330 60 1040
00340W tONl]t!IJ:
003s0 LIKI?t(~,l) PHI, XN1
oo3bo 1 F~MATil A, W~X(, F4.3, ): ,;?.2)
043701 FoRflal(5x, Flo.2)
oo3@o STOP
00390 [m
00400 FuNCTION N(l)
00410 Is4rLIcIT tEAL*B(A-M, *).
tEu=t(O-2)
004?0 C SIMBAMI NOMAl DISTRKW!ON :lUCTSW
00430C rORTK NET140D, SEE 7!4[ tthTIONAL
PLIHAU OF STAi~WR@S
00440c NMDI09UofMTWMTICAL rumzo~s
004s0
oo4bo
00470
004B0
004?0
00s00
003Io
00s?0
00530
00540
08ss0
10540
Oos?o
00ss010
005?0
00610
004)o
. .. .
TABLE C-V-II , Sample input /out Tut data for
y~ax(~) computer program.
7.OIJO 9,0D0 .90110 250.ODG

4 SO.ODO
2.OBO .24D0
z.~~o .21D0
I.eBo 20D0
2.100 .18D0
2 100.ODO
Input Data 2.7~o 15[10
3.ODO :14B0
4 50.ODO
1.SD() .10DO
1 .4D0 .08D0
1,7D() .llLJO
1.9D0 .0900
~ 50 .Ono
1.ODO .05D0
1.3D0 80D0
8.17
Output from Program b.o?
8.10
8.10
MAx(.we)= 6,10
Mil. -fii3lJK-L7?
0 000 2.00 00 600 8.00 1000
FIGURE C-V-I. Resulting repair time distribution for a samplo

svstem containing four (L) repair~pcs with
different normally distributed repair .-..
times.
12 Jan 1984 c-v-8

MIL-HDBK-472
APPENOIX D
AcfJJRATEMANUAL M.,, O ANALYSIS
iO. An alternative to the accurate computer M-., (4) ana]ySiS \S

presented. Tne previous technique, Appendix C uses a computerized, iterative
technique tc solve the equation. A complete development of the proposed
manual M~,. (@) technique is given here. The technique is illustrated
using the prediction example described in paragraph 10.3 of appendix C. Hhere
~ossible, the same terminology and abbreviations used in paragraph 10 of the
computerized technique are also used here. Tuo alternatives are given for
handling large numbers of types of repairs:
a. Reference is made to a progranunablesllde rule calculator program

which speeds up the calcu atlcm procedmes.
b. A method of group ng the data is described and illustrated uifh a

prediction example.
20. The manual H-,, (0) analysts is based on the same two basic premises
of the computerized technique. Before giving these premises, SOfIWkey
terminology is first defined. There are NxJ possible system repair types with
random system repair times R.j, n=l, .... N, j-l, .... J where N 15 the
total number of repairable items and ~ is the total number of unique fault
detection and isolat~on (fD&I) outputs. The system repair times R.j are
themselves the sumof the applicable maintenance element ~/ type times E-
-J
tiherzm +nd!catez the element type and where the element types are defined as
follows:
a. Fau t localization
b. Faut isolatlon
c. Disassembly
1/ This element of repair time has untts of hours and should not be confused
~ith the probability density element introduced later, uh~ch has unit$ Of
probability.
d. Interchange
-
e. Reassembly
f, Alignment
9 Checkout
The first premise is that at least the mean (wC ) and variance (UE 2,
m m
Tf cactiof the i. are kq~wfl. 21 The second prefaise is based on the central
n~
limit thenrem which suggests that the pdf of each of the ind~v~dual RfIJ
approachesa Pomal dlstrlbu?~on with a wan (PP ) f~rmed from the sum of
n~
the respective (PF s).
m
M
prmj- z pi (1)
m
M
where 1 lndlcates the sum Is taken over all appl~cable element types and a
vartance (un 2, formed from the sum of the respective u~ 2s jl

n~ m
(2)
~1 The obvious additional subscript nj is ~~tted on ~E and or z

m m
throughout the dlscuss!on,
jJ In forming the on s as the sum of the respective UC s, the

mJ *
variables are assumed to be uncorrelated. lt !s noted that It !s not
essential to know the dlstrlbutton of the Em times, merely their
12 Jan 1984
D-V-2
MIL-HDBK-472
. 30. Both the computerized and the manual calculation techniques are subject !
to the same inaccuracies inherent in the assumption of normality. Although
the central limit theorem s usually applicable only for large numbers, there
is evidence that it is appIcable (le., its use gives reasonably accurate
-esuits) for tne type of s tuation treated here when as few as six E* s are
m.
.
used and the percentile va ue Is not larger than 95 percent. The accuracy of
P
*ma. for the case just refered to does decrease sign!ficact!y at the
(0)
higher percentiles - i.e.. at 99 percent. Fort~nately M~.. (0)

requirenients are usualiy limited to 90 or 95 percent.
40. Let f, (t) be the pdf for the system repair times Rqj. There are NxJ
n~
pcssit!le different f. it)s. By definition, the fQ (t are non-negative
*; r.;
functions whose integrals, when extended over the entire t axis, are unity.
For a given fn (t), the probability that the njh type of system repair
mj
is completed In the infinitely small interval (t, t+dt) !s f~ (t) dt. The
*)
quantity fR (t)dt Is called }he probability density element at the point t.
~j
The relative probability of the nj type of system repair occurring, P.j
is:
P., WA.,IEA (3)
Mhere Xnj ts the failure rate In failures per million hours associated
with the n.j type of system repair and 1A is the total system failure rate
of repairable items, ~1 P.j is also the relatlve probability of the nj
probability density element, The combined probability of njh type of
system repair occurring and being completed In the Intervals (t, t+dt) is
P.jf~ (t)dt. The combined probability of any of the NxK possible system
n~
repair types occurlng and being completed \n the (nterval (t,t+dt) is
~1 Note that P.j is based on all failures and repair procedures that lead
to system repair nj. A particular P., may result from failure of a
single component, several components in series, several redundant
components, or some other combinations of components. Thus, any system,
whether formed of series, redundant, or some other combination of
components may be analyzed by both the computerized and the manual
calculation techniques providing the X.j values can be determined.
MIL-MDBK-472
(since the R., random events are mutually excluslve) the sum of the relative .
probabilitle$ of each of the NxJ possible probability dens!ty eleIWntS
multlpled by each corresponding probability density element at time t. This

NJ
is lP.jF~n\t)dt, where t = I 1. This expression Is the
n-l j,l
probability system repair, completed in the t~me interval (t, t + dt). lt Is
also the system repair probability density element at time t. lJslngthe

probability density element definlticm given above, lPmjf* (t) 1S the pdf
nJ
of the system repair time T, i.e.,

g,(t) - f.ft) - IP.,f* (t) (4)

nj
50. Before using lP.jfe (t) as the pdfof system repair time T, ltmust
nj
be shown that this function Is a pdf - i.e.. that It !s non-negat!ve and its
integral over the entire t ax!s is unity. Since It Is formed from the sumof
non-negat!ve functions, this function !s also non-negative. The Integral of

m
the function over the entire t axis 1s lP.j~ f, (t)dt. This iS just
o n~
the sum of the relatlve probabllittes of each NxJ poss~ble system repair type
probability density element multiplied by the Integral, over the entire t
axis, of the pdf of each NxJ possible system repair type. As stated In
paragraph ~0, these Integrals of pdfs are by deftnitlon each unity. Each of
the NxJ possible Pjs In the function are thus multiplied by unity and the

Integral, Over the entire t axts, of the function becomes just XP~j. Th!$
1s the sumof the relatlve probabll!t~es of each NxJ possible system repair
type probability density ele~nt Occurring, and th!s sum, as uith the sumof
any complete set of relatlve probabllttes, Is unity. Thus, the Integral of
the funct!onover the enttre t axis fs unity.
12 Jan1984
D-V-4
MIL-HDBK-472
60. Let the cumulative distribution function (calf)of T be G,(t). By

deflnlt{on of the relationship between the pdf and calf:
*t
J g?(t) dt -G,(t) (s)
o
S\nce GT(t) IS the probability of ccmplettng system repair at time t or

less, it follows that:

M ma. (@) m,, (0)
J g,(t)dt = lPnj r f. (t)dt = 0 $1 (6)
o 0 n~
where Mm,. ($) represents the value of t at whtch the probab[llty of

repair is Q.
M **.(O)
70. The integral
o
J
f~ (t)dt Is the probability of completing repair
n~
type nj at time t or less. This tntegral !s also the area under the pdf curve
of repair type nj up to (at) a particular t value. To slmpllfy equation (6),
let
M m..(~)
F,(t) D
J f. (t) dt (7)
o n~
Equatton (6) becomes

Mm,, (@)
J gr(t)dt - W.jFmj(t) =0 (8)
o
~/ This is equation 4 of paragraph ?0.1 of Appendix C.
D-V-5 12 Jan 1984

MIL-HDBK-472
Note that the derivation of equation (81 hold; or any dls:ributlon of R.j.
--
Thvx, !f the form of pdf of the R.j and the parameters necessdry to
cnaracter:ze the pdf distr~but~on of the R.j are known, bcth the
computerized and the Inanua? C?lCUldtiOn tectn;ques can De uses without
assum~ng the adf of each of the ind:~idua! R., is normal.
80. TO so!ve equation (8), :t is necessary only to find thE t val)efor uhcb
the ~um, over all NxJ poss;ble system repair types, of ~,F,Jt~j gives the
desired O. Valtjes of +., are calculated as iod!c~ted by equation (4!.
Values of F.,(t) are found from tables of area under suitable pdf curves.
An Inltal t ts selected and evaluated. Successive t values are then examined
Q Is obtained, to the desired level of accuracy.

until the des~ref.i
90. The p,ed!etion example used here to illustrate the manual calcu ation
technique is the example described in paragraph 10.3 of the computer zed
technique. The individual R.j are assumed to have normal pdfs. Table
D-V-I contains the input data for the E- comprlsing each nj system repair
-J
type, plus the PI and U[ for each E. of each nj system repair
. .
type. Usfng this Input data, the individual nj system repa~r type iJW .
~j
2
ok and P.j are calculated as described above in paragraphs 20 and 40 and

mJ
as indicated in footnoes !/ through Ii of Table D-V-XI. Columns 2 through 4.
respectively, of Table O-V-II present these calculated vaues for the
individual nj repair types. Columns 5 through 8, respect vely, of Tab e
D-V-II present the calculated values of F.j. P.jF.j and O for four
iteratively-selected trial values of t. As shown, the desired @ value of
0.90 Is achieved at 8.10 hours. This is the same value given in paragraph
10.3 of Appendix C for the example computer output.
12 Jan 1984
D-v-6
MI1-HDBK-472
ction example described above indicates the relative simplicity
of tne mtirlual
calculation tecnnique Although applied In the example to a
jjstenicomprised of only four indiv dual repair types, the technique is
applicable to ?ny r,umberof repair ypes. For 10rge number~ of repair types,
it wCLld be Ccnieriertto use a programmable :~lcu!at~r with a norm~l
distribution program.
i~ij, Another practical calculation al~ernative for handling large numbers of
repair types with the manual calculation technique is by grouping the data. It
IS assumed the necessary Caiculatlons have Oeen made to obtain the MU S

~J
and all 2s, or they are given directly. Group the ordered repair type
n~
P.m, s intO 1 time Intervals of length At. For each set of data a mst
suitable number, 1, of classes can be found. Sturges has developed the simple
rule 1 = 1+3.3 log,O N where 1 is the number of classes and N 1s the number
of data. ADDroximate the actual di$trtbutlon of repair type p~mjs !n
each of the 1 time intervals by a normal distribution with a mean (p,)
calculated as th failure rate-we!ghted repair type mean
i 1
p, = Zp. A.jlzl (10)
mJ
i
where Z indicates the sum is taken over the ii interval and a variance
(U,2) calculated as the failure rate-weighted repair type variance
2 i~i
al -run L,lrl (11)

n~
D-V-7
I
1,
\ I
, I-,
I
1 I
(
I
I
I
!
I
I
I
I
I,
I
f
m-
. I e
1
f.
m
v
GO
,.
00
r- 92
. .
--
in 80
N u- 1-
g ed UI
ml 0
.
0 0 0
.
u
o m
. .
m
I
I
12 Jan 1984
z.
.
c
c
c
.
r+
n
w
w
1
N
-. z
E
0
.
N1
-
MIL-HDBK-472
61BL10GRAPHY
1. Maintainability - Prediction: Theoretical Basis and Practical Approach

of, USAF Aeronautical systems Division, 15 FeD 1962.
2, Maintainability - Prediction: Theoretical Basis and Practical Approach

of, (JSAf Aeronautical Systems Division, 31 Dec 1963.
3. Maintainability - Design Crlterla Handbook for Designers of Shipboard

Electronic Equipment, NAVWIPS 94324, 30 Apr 1962.
4. A Research Plan for Developing Methods of Malntalnability Measurement and

Prediction, Rome Air Development Center Tech. Note No. RADC-TN-60-5, AD-231
090, 6 Jan 1960.
5. A Review of the Literature on Design of Information Job Performance Aids.

Proj. No. 1710. USAF Airborne Systems Division Report No. ASDTR 61-549, 1961.
6. Maintainability Measurement and Prediction Methods for A.F. Ground

Electronic Equipments, Phase 11 Report RADC-TN-61-141, AD-262 454, 15 Jun 1961.
7. Maintainability Measurement and Prediction Methods for A. F. Ground

Electronic Equipments, Phase 11 Report RAOC-TN-60-221, AD-247 155. 15 Seo 1961.
8. Maintainability - Prediction Technique, Phase IV Report RADC-TDR-62-156,

AD-277 199, 15 Mar 1962.
9. Malntainabillty Engineering, Phase V Report RADC-TDR-63-85, Vol. 2, AD-404

898. 5 Feb 1963.
10. Malntalnbility Technique 5tudy, Final Technical Report, Phase V Report

RADC-TDR-63-85, Vol. 1 AO-404 899, 5 Feb 1963.
11. Maintainability and Supportabllty Evaluation Technique, Mright-Air

Development Division, Part I, Part II, Vo]s. 1 and 2, May and Jun 1960.
12. Malntainablllty Prediction and Analysts Study, Final Technical Report

RADC-TR-78-169, AD-A059753, Jul 1978.
.-
... ..
..
... .

MIL-HDBK-472 Mantenibilidad y Prediccion PDF

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

MIL-HDBK-472 Mantenibilidad y Prediccion PDF

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

MIL-HDBK-472 Mantenibilidad y Prediccion PDF

Uploaded by

Copyright:

Available Formats

MI L-HDBK-472

MILITARY STANDARDIZATION HANDBOOK

1. Thls maintainability handbook. V.as developed b} the I)epartn)cnt of

:}. This dwunlcnt profidcs information on current maintainahilit!

4. Ever~ effort has been made to reflect the latest information m

The purpose of this Maintainability Prediction Handkmok is w familiarize

Philosophy, Assumptions and Summary

Basic Parameters of Measure

Correlation Between Predicted and observed Values

This type of presentation facilitates reference to all or to only those parts of

Prediction facilitates anearly Assessment of thematurity of the maintainability

The maintainability prediction procedures I and III are applicable solely to

In conclusion, the use of this hardbok facilitates the desi~, development,

The Need for blaintainability Prediction

MMNTAINABILITY PREIXCTION PROCEDURES

SECTION 1.@ GENERAL 1-1

1,1 Phi losoph!, Assumptions and Summar.v 1-1

SECTION 2,0 AN.+ LYTJC FOUNDATION 1-4

2.1 Structure of Time Elements in Fix of Malfunction 1-4

SECTION 300 APPHCATION

3. I Preliminary Procedure 1-19

TABLE OF CONTENTS (Continued)

ADDENDUM A CALCULATION OF THE PRO134BIXJTIES OF Al-37

ADDENDUM B DEFINITIONS OF MAINTAINABILITY TERMS AI-44

SECTION 1.0 GENERAL 2-1

1.1 Philosophy, Assumptions and Summary 2-2

SECTION 2.0 ANALYTIC FOUNDATION 2-8

2.1 Basic Considerations aad Definitions 2-8

SECTION 3.0 APPLICATION 2-11

3.1 PART A Corrective Maintenance Rediction Procedure 2-11

TABLE OF CONTENTS (Continti)

3.1.1.4 Prediction 2-15

3.2 PART B Active Maintenance Predction Procedure 2-27

. PROCEDURE III 3-1

SECTION 1.0 GENERAL 3-1

1.1 Philosophy, Assumptions and Summary 3-1

SECTION 2.0 ANALYTIC FOUNDATION 3-5

2.1 Determination of Sample Size N 3-5

SECTION 3.0 APPLICATION 3-14

3.1 Gene-l Approach 314

TABLE OF CONTENTS (Continued)

3.2 Sb?p-by-Stap Procedure for Calculating Task Sample 3-16

ADDENDUM A DESIGN CHECK LISTS AND SCORKNG CRITERIA A3-33

CHECK LIST A - SCORKNG PHYSICAL DESIGN FACTORS A3- 33

SECTION 1.0 GENERA L 4-1

1.1 Philosophy, Assumptions and summary 4-1

SECTION 2.0 ANALYTIC ,OUNDATION 4-3

2.1 General 4-3

Table i-ieadmg Page

List of Categories and Elemental Acttvtties of .4ctive 1-13

3-1 Part Class Failure Distribution and Sample Size 3-17

Figure Figure Heading Page

A Comparison Matrix of hfaintainabilit~ Prediction Procedures 4

2-1 Functional Level Diagram of a Typical Communications System 2-13

3-1 Comparison of Population ar. rl Sampling Mean Distributions 3-8

4-1 Procedural Flow Block Diagrnm 4-6

THE NEED FOR MAINTAINABILITY PREDICTION: The prediction of the eqxxtecl

A significant advantage of using a maintainabili~ prehction procedure ia that