Full text of "An approach to the estimation of life cycle costs of a fiber-optic application in military aircraft"

See other formats

Institutional Archive of the Naval Postgraduate School

Calhoun: The NPS Institutional Archive
DSpace Repository

Theses and Dissertations 1. Thesis and Dissertation Collection, all items

1975

An approach to the estimation of life cycle
costs of a fiber-optic application in military aircraft

McGrath, John Michael; Michna, Kenneth Ralph

Monterey, California. Naval Postgraduate School
http://ndl.handle.net/10945/20761

This publication is a work of the U.S. Government as defined in Title 17, United
States Code, Section 101. Copyright protection is not available for this work in the
United States.

Downloaded from NPS Archive: Calhoun

Calhoun is the Naval Postgraduate School's public access digital repository for
(8 DUDLEY research materials and institutional publications created by the NPS community.
«ist sae Calhoun is named for Professor of Mathematics Guy K. Calhoun, NPS'‘s first

INN KNOX appointed — and published -- scholarly author.

| LIBRARY Dudley Knox Library / Naval Postgraduate School

411 Dyer Road / 1 University Circle
Monterey, California USA 93943

http://www.nps.edu/library

=
>
ma
i>
ow
m
3
'
()
“U
~
—
oO
a)
a a

John Michael McGrath

NAVAL POSTGRADUATE SCHOO

Monterey, Galifornia

THESIS

AN APPROACH TO THE ESTIMATION OF
Piste CrChE COSTs OCF°A FIBER-OPTIC
APPLICATION IN MILITARY AIRCRAFT

John Michael McGrath
Kenneth Ralph Michna

September 1975
Thesis Advisor: Catia eee oles

Approved for public release; distribution unlimited.

1171677

SECURITY CLASSIFICATION OF THIS PAGE (When Data Entered)

REPORT DOCUMENTATION PAGE

1. REPORT NUMBER 2. GOVT ACCESSION NO.

4. TITLE (and Subtttle)

An Approach to the Estimation of Life
Cycle Costs of a Fiber-Optic Applicatio

in Military Aircraft
AU THOR(e)
John Michael McGrath
Kenneth Ralph Michna

READ INSTRUCTIONS
BEFORE COMPLETING FORM

3. RECIPIENT'S CATALOG NUMBER

5. TYPE OF REPORT & PERIOD COVERED

Master's Thesis;
September 1975

6. PERFORMING ORG. REPORT NUMBER

CONTRACT OR GRANT NUMBER(2)

10. PROGRAM ELEMENT, PROJECT, TASK
AREA & WORK UNIT NUMBERS

- PERFORMING ORGANIZATION NAME AND AODORESS

Naval Postgraduate School
Monterey, California 93940

. CONTROLLING OFFICE NAME ANDO ADORESS

Naval Postgraduate School
Monterey, California 93940

- MONITORING AGENCY NAME 4 ADORESS (Hf different from Controlling Office)

12. REPORT DATE

September 1975
13. NUMBER OF PAGES

ios

1$8. SECURITY CLASS. (of thle report)

Unclassified

1Sa, OECLASSIFICATION/ DOWNGRADING
SCHEDULE

Naval Postgraduate School
Monterey, California 93940

DISTRIBUTION STATEMENT (of thie Report)

Approved for public release; distribution unlimited.

- DISTRIBUTION STATEMENT (of the ebetrac! entered in Block 20, if different from Report)

- SUPPLEMENTARY NOTES

. KEY WORDS (Continue on reveree eide if neceeeary and identity by biock number)
Optical fibers, fiber optics, optical waveguides, light
emitting diodes, PIN diodes, economic analysis, cost

effectiveness analysis, life cycle cost model

. ABSTRACT (Continue on reveree cide if neceseary end identity by biock number)
As significant technological advances in fiber optics and

optical data transmission methods are being made, it is
necessary to develop appropriate methods for estimating life
cycle costs for alternative coaxial/twisted pair wire and
optical fiber avionics. Measures of effectiveness are
suggested for each alternative system. An approach, which

structures the technological and demand uncertainties of

DD . ‘an 73 1473 EDITION OF 1 NOV 63 Is OBSOLETE

Pa S/N 0102-014- 6601 | =
(Page 1) SECURITY CLASSIFICATION OF THIS PAGE (hen Data Entered)

SECURITY CLASSIFICATION OF THIS PAGE(Hhen Deta Entered:

Vip betract (cont 'd)

fiber optics, is developed through scenarios as a means of
relating cost and effectiveness. It is suggested that Delphi
and experience curve techniques be used in conjunction with
ordered scenarios as a technological forecasting technique

for estimation of life cycle costs of fiber optics. In addition
a review of the historical and technological background of fiber
optics and their application to the Naval Electronics Laboratory
Center (NELC) A-7 Airborne Light Optical Fiber Technology
(ALOFT) Program is included.

DD Form, 1473

GO Be.
S/N 0102-014-6601 ? SECURITY CLASSIFICATION OF THIS PAGE(When Data Entered)

An Approach to the Estimation of Life Cycle
Costs of a Fiber-Optic Application in Military Aircraft

John Michael McGrath
Lieutenant Commander, United States Navy
B.S., United States Naval Academy, 1962

Submitted in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE IN MANAGEMENT
and

Kenneth Ralph Michna
Lieutenant Commander, United States Navy
A.B., Wabash College, 1965

Submitted in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE IN OPERATIONS’ RESEARCH
from the

NAVAL POSTGRADUATE SCHOOL
September 1975

E SCHOOL
Y. CALIFC iA 93940

ABSTRACT

As significant technological advances in fiber optics and
optical data transmission methods are being made, it is
necessary to develop appropriate methods for estimating life
cycle costs for alternative coaxial/twisted pair wire and
optical fiber avionics. Measures of effectiveness are
suggested for each alternative system. An approach, which
structures the technological and demand uncertainties of
fiber optics, is developed through scenarios as a means of
relating cost and effectiveness. It is suggested that Delphi
and experience curve techniques be used in conjunction with
ordered scenarios as a technological forecasting technique for
estimation of life cycle costs of fiber optics. In addition,
a review of the historical and technological background of
fiber optics and their application to the Naval Electronics
Laboratory Center (NELC) A-7 Airborne Light Optical Fiber

Technology (ALOFT) Program is included.

TABLE OF CONTENTS

I. INTRODUCTION -------------------------------------

II. BACKGROUND ---------------------------------------

HISTORICAL BACKGROUND ------------------------
1. Background of Glass Fibers/Fiber Optics --

2. Background of NELC ALOFT Demonstration
Program --------------- 2 eo

3. NELC A-7 ALOFT Demonstration Approach ----

4. A-7 ALOFT Demonstration Management
Organizational Structure -----------------

III. FIBER-OPTIC TECHNOLOGY AND ITS APPLICATION

Po OR a eee
A. GENERAL -------------- 2222-22222 eee ee ee
B. FIBER OPTICS RELATED TECHNOLOGY --------------

i eteee iatten == ———-—----=—-=----------—----===-
Peat asc —=—————— = — = — —
3. Multiplexing/Data Bug ------2-2--2--------

DESCRIPTION OF THE BASIC COMPONENTS OF A
FIBER-OPTIC DATA TRANSFER SYSTEM -------------

1. General ----------------------------------
2. Glass Fiber/Cables -----------------------
3. Connectors/Couplers a = See = = elie

a. Connectors ------co- nooo ooo ooo ome oon

Boeecouprers for Data Bus Application --=-

igi
EO
16

35
35
37

44
44
45
DZ
DZ

4. Light Sources/Signal Drivers -------------
a. Light Emitting Diodes (LEDs) ---------
b. Semiconductor Lasers -----------------
>. Signal Receivers/Detectors ---------------

SYSTEM DESCRIPTION OF FIBER OPTICS AS EMPLOYED
IN THE A-7 ALOFT DEMONSTRATION ---------------

1. System Description -----------------------

IV. AN APPROACH FOR A COST-EFFECTIVENESS STUDY OF
AVIONICS BATA LINK ALTERNATIVES ------------------

ley

GENBRAD -------—-------~-~---~~--~-----=--------
MEASURES/LEVELS OF EFFECTIVENESS -------------
SOST AMAA AS “66283 reer
1. Life Cycle Costing -----------------------
2. Cost Data Collection Effort --------------
COMBINING COST AND EFFECTIVENESS -------------

l. Ordering Uncertainties Through Scenarios -

a. scenario I -- A Neutral Context ------
b. Scenario II -- A Modestly Optimistic
CE O@SR 2S SoS ons 5555S] Sse aS eee

c. Scenario III -- A Modestly
Pessimistic Context ------------------

2. Constructing Cost-Effectiveness Curves
from Scenarios ---------------+ ecco

Pe VOmGoOctinG EECHNIOURS FOR FIBER OPTICS ----------—

Dee

GENERAL -------------------- o-oo ee eee

aN apa CHNMOUR el ===-os2— aoe eee

DY,

C. THE EXPERIENCE CURVE TECHNIQUE --------------- iy

VI. SUMMARY AND CONCLUSIONS See ee ea 134
APPENDIX A, A-/ Navigation Weapons Delivery System

(NWDS) Schematics -------------------------- WSs

APPENDIX B, A-/7 ALOFT Component Requirements ----------- 140

APPENDIX C, A-/7 ALOFT Component Descriptions ----------- 144

~oenwixX D, Faber Optics Cost Data Collection ---------- 148

APPENDIX E, Industry Contacts for Fiber-Optic
Convenes. 2 = aioe ios SaaS 153

APPENDIX F, Assumptions for an Economic Analysis of

eu: NONI SIE) C3 aaa ll LSy
OE LTO PLD aaa ILS,
ie omen LON (oie = >= ---<-- ss Sse sso ce a= - == === 162

Figure
II-1l

er —2
i= 3

II-4

ipl 5

er |

Prt -2
PEL -3
ee —4
ILI
III -6
ek /
III -8
eit —9
II1I-10
JUEIL Salk
eh 2

met — 13

LIST OF FIGURES

Page
Historical progress in low-loss seyeebdas:
lowest attenuation achieved vs. year --------- IN
A-7 ALOFT Demonstration milestones ---------- 29
A-7 ALOFT Economic analysis activity flow ---- 3]
A-7 ALOFT Project Office organization
Structure rrr r ee em re ren rer reece eer enrnren- 52
A-7 ALOFT economic analysis organization
SEXYUCTULE --——— enn He mm nn nm men wn nn nw ww nn nen 38
Attenuation as a function of wavelength in
a recent Corning low-loss optical fiber ------ 37
pe Iieemiadce cysoeeis. -——-—-—9-——-——————————— 39
Data multiplexing <<<<9--<-<<3-c<-<<<-e--------- 41
ivoeeweaireCrare data busvarchitecture =-----~- 43
ip ee lascee LAadehiber: —---=-—-——-—--——=————-=— 46
Large-core, solid-clad fiber ----------------- 47
eee Oden i Deis ae = ee a 47
Graded-index fiber -------------99------------ 48
Fiber-optic bundle ------------9e rrr nner 50
iitelnogcmea lCmCONMCC EONS (= =—-——==—-————— 53
i peroal ier optic slink Leplacement=—-----—-- 56
Stamdaca package = fiber-optic module ----=-<- 56
Pao MmeCmImmmOaehesmeOnInterlaACeS (=a ———sa—— > — 14 —> ye)

Figure
III-14

HiL-15

ny = 1

iy 2

Laser injection diode bonded to fibers -------
Optical circuit on a chip as envisioned

by Texas Instruments -------------------------
Economic analysis process --------------------

Hypothetical cost-effectiveness curves
displaying dominance -------------------------

Hypothetical cost-effectiveness curves -------
Measures of effectiveness --------------------
Life cycle cost elements ---------------------
Rupr OpEtecscable demand, by Scenarios: =---=-=
Scenario/effectiveness/time matrices ---------

Scenario related cost-effectiveness curves ---

Sample Delphi questionnaire ------------------

The 80% experience curve on an arithmetic
grid -------------------------------------- =e

The 80% experience curve on a logarithmetic
grid -------------------------- oo en renner ee

itm@ecrated Cilrcules ExperlLence curve =====-=<->
Pomwiny lenlortde experience curve ---=-=----—-
MECWeameacteri Stile Stable pattern -=--—=—-=—---—-—
A typically unstable pattem -----------------

A characteristic unstable pattern after it
has become stable ----------------------------

Hypothetical example of a fiber-optic
cable experience curve --------cr rere oe

ree:
Te
76
79
102
104

105

ACKNOWLEDGEMENT

The authors wish to express their appreciation to
Associate Professor Carl R. Jones for his encouragement and
assistance in this work.

The authors also wish to express their appreciation to
LCDR J. R. Ellis, Mr. Roger Greenwell, and other personnel
of NELC who unselfishly shared their time and materials to

mesist in this work.

I. INTRODUCTION

Present day avionics in military aircraft utilize twisted
shielded pair wire and/or coaxial cable to transfer signal
data. These data link subsystems reflect post-World War II
state of the art in electronic development. Electronic signal
transmission by this means exposes avionics to potential
operational degradation and damage because of the suscepti-
bility of metallic conductors to electromagnetic interference,
radio-frequency interference, and nuclear-generated electro-
magnetic pulse. Other sources of electronic interference such
as cross-talk, ground looping, reflection, and short-circuit
loading also degrade system operation.

A recent technological breakthrough in the field of fiber
optics has made fiber-optic data link applications technically
feasible, and perhaps desirable, for use in military aircraft
avionics systems. Fiber optics technology does offer several
significant advantages for avionics data link subsystems. The
primary advantages are that it: (1) is not susceptible to
electromagnetic interference (EMI) nor to electromagnetic
pulse (EMP) associated with a nuclear blast; (2) does not
generate EMI; (3) is isolated from ground plane signals; and

(4) is capable of higher data rate transmission.

As a result of feasibility tests and demonstrations
conducted or sponsored by Naval Electronics Laboratory Center
(NELC), San Diego, approval was gained from the Assistant
Secretary of the Navy for Research and Development to implement
a two-year feasibility program to install fiber optics com-
ponents (fiber-optic cables, light sources, light detectors,
and connectors) in place of standard twisted pair wire and
coaxial cable for selected components of the Navigation/

Weapons Delivery System (N/WDS) of an operational A-7E
Corsair IL light jet attack aircraft. The program, called
the A-7 Airborne Light Optical Fiber Technology (ALOFT)
Demonstration, is a feasibility demonstration to determine
the information transfer capability of an aircraft avionics
system through point-to-point applications of fiber optics.

Concurrent with the A-/7 ALOFT Demonstration checkout, test
and evaluation, an economic analysis was desired by NELC for
the two alternative systems; coaxial cabling and fiber-optic
cabling. These two alternatives, together with their associated
components, will hereafter be referred to as "coax" and "fiber
Spetcs.

The basic format of an economic analysis involves the deter-
mination of the cost and effectiveness of competing alternatives.
A life cycle cost model, as defined by NELC and Naval Postgrad-
uate School (NPS) students is used as the costing basis for the

Ivo alternatives.
12

A contractor will perform the costing of the coax alterna-
tive. He will also determine the measures of effectiveness
for both the coax and fiber optics systems. Naval Postgraduate
School students will perform the costing effort of the fiber
optics alternative. The authors of this thesis perform a
preliminary costing effort for the fiber optics alternative by
developing an approach to costing fiber optics as an emerging
technology. NPS students and NELC systems analysts personnel
will coordinate future efforts toward the desired objective
of numerical estimation of fiber optics life cycle costs.

As a baseline thesis for follow-on NPS theses students,
the authors have discussed the historical and technological
background of fiber optics as well as the background of the
A-7 ALOFT Demonstration. A general discussion of a cost-
effectiveness analysis is presented together with possible
measures of effectiveness (MOEs) for data transfer.

Since fiber optics cost data is either non-existent or
available only on a prototype basis, the authors' basic
approach to costing fiber optics is done through scenarios.
Scenarios offer a means of ordering the uncertainties of an
emerging technology. They define the possible futures of the
fiber optics industry and its related technology. Three sample
scenarios developed by the authors provide specific time-

related estimates as to civilian/military demand, growth rates,

standardization and technological development. These repre-
sentative scenarios are meant to provide the basis from which
cost estimates could be made.

Two exploratory techniques, Delphi and experience curves,
are discussed as they pertain to the costing of an emerging
fiber-optic technology. A sample Delphi questionnaire is
developed as a means of soliciting forecasts from a panel of
experts in order to deal with specific uncertainties associated
with fiber optics. (e.g. When will production bases be
established for fiber optics components?) The information
gained from the Delphi survey can be used to refine the
estimates contained in the scenarios as well as minimize the
number of possible scenarios.

Experience curve evidence is discussed as a means for fore-
casting unit cost reduction as the fiber optics experience base
accumulates. The information required for using experience
curves is provided by the scenarios. Experience curves can
then be used as a means of predicting the cost behavior of
components relating to fiber-optic technology.

It is felt by the authors that these techniques; scenario-
writing, Delphi and experience curves, can be combined as a
cost-predictive method to estimate component prices of an
emerging technology such as fiber optics. These techniques

can then provide a means of estimating costs for the life cycle

cost model elements used in a cost-effectiveness study. Not
only will the fiber-optic component procurement costs be
estimated, but the costs to operate and maintain a fiber-optic
system will also be determined through future efforts.

This thesis, then, is the first step in developing a cost-
effectiveness study which could aid in making decisions
concerning the use of coax or fiber optics in the next series
of military aircraft to be designed and built (VAX, VFX, VPX,
etc.).

It is the basic conclusion of the authors that the emerging
fiber-optic technology deserves full and continuing effort and
attention by research and development (R&D) agencies. Even if
the results of initial cost-effectiveness studies are such
that the decision is made to not use fiber optics in the next
generation of aircraft, the authors feel that it would be a
mistake to cut back or reduce fiber optics R&D funding. The
military services are pursuing extremely meaningful and
productive research and development in a field containing
great potential for future benefits to the military services
in general. It is expected that fiber optics will be used in
some future generation of military aircraft and weapons systems.
These future weapons systems would be the beneficiaries of

today's efforts from the development of this emerging technology.

II. BACKGROUND

A. HISTORICAL BACKGROUND
1. Background of Glass Fibers/Fiber Optics

Glass has been used in a multitude of applications
from very early times. The earliest glass objects come from
Egypt and are dated from circa 2500 B.C. The first vessels
of glass were manufactured in Egypt under the 18th dynasty,
particularly from the reign of Amenhotep II (1448-20 B.C.)
onward. The possibility of drawing hot glass into threads
was recognized in the Rhineland during the late Roman empire
as well as in ancient Egypt and such threads were wound around
vessels as a decoration.

In the 18th century, fine threads were prepared from a
heat softened glass rod by using a "spinning wheel" process.
The next development was a mechanized drawing process by
attaching the fiber from the heat-softened rod to the surface
of a large revolving drum. In 1908, G. von Pazsiczky replaced
the rods with a refractory glass-melting chamber that had a
series of holes in the bottom to provide drawing points. A
different method of production was developed in 1929 whereby
the application of centrifical force forced the glass through

radial serrations resulting in a tangled mass of fibers. /16/

It is entirely possible that early Egyptian and Grecian
Glassblowers observed the phenomenon of multiple total internal
reflections in conducting light along transparent glass
cylinders, and in fact, there are a number of unsubstantiated
historical claims. However, the earliest recorded scientific
demonstration of the phenomenon of total internal reflection
was recorded by John Tyndall at the Royal Society in England
in 1870. In his demonstration, he used an illuminated vessel
of water to show that when a stream of water was allowed to
flow through a hole in the side of the vessel, light was
conducted along the curved path of the stream. D. Hondros and
P. Debye followed the work of Tyndall by doing some theoretical
studies on optical wave propagation in fibers in 1910, but
little else was done in the way of experimentation.

The phenomenon described by Tyndall was disregarded
and lay dormant until 1927, when J.L. Baird in England and
C.W. Hansell in the United States considered the possibility
of using uncoated fibers in the field of television to transmit
and scan an image. They were followed closely by H. Lamm of
Germany who used a crude assembly of quartz fibers to demon-
strate the basic image and light transmission properties of
fibers. Activity in this area then all but ceased for two

decades. [25/

Quite unrelated to previous experiments with glass
fibers as light conductors, manufacturing methods for producing
glass fibers were being perfected. For example, in 1938 the
Owens-Illinois Glass Company joined with the Corning Glass
Works to form a new independent glass fiber firm. The company,
the Owens-Corning Fiberglass Corporation, developed large-
scale production methods to produce glass fibers. The spun
glass method allowed continuous threads to be drawn from
bushings provided with 100-400 small orifices. The threads
falling from these orifices were gathered together and passed
over a sizing pad onto a spool on a high-speed winder. The
resulting fiber had a diameter of around 0.00022 in. ne
Material contained in one glass marble 3/4 in. in diameter
would yield about 97 miles of single filament). /15/

A new burst of activity began in the year 1951, when
A.C.S. van Heel in Holland and H.H. Hopkins and N.S. Kapany at
the Imperial College in London independently initiated studies
on the transmission of images along an aligned bundle of
flexible glass fibers. Kapany, B.I. Hirschowitz, and others
then developed optical insulation techniques which solved most
of the previous light-loss problems. The resultant glass-
coated glass fibers were for many years a standard optical
element for use in fiber optics. Kapany continued his work and

in 1956 first applied the term "fiber optics" by defining fiber

optics as "the art of the active and passive guidance of

light (rays and waveguide modes), in the ultra-violet, visible,

and infrared regions of the spectrum, along transparent fibers

through predetermined paths." /25/

During the ten year period from 1957 to 1967, interest

and experimentation increased such that significant develop-

ments and applications were made in the following areas:

ile

Waveguide mode propagation.

Coupling phenomenon in adjacent fibers.

The use of scintillating fibers for tracking
high energy particles.

Skew ray propagation along fibers.

The use of fiber optics as field flatteners,
Focons, and image dissectors in ultra-high-speed
photography.

Extension of the spectral range of fiber optics
in the infrared region.

Combining the field of lasers and fiber optics
in lasing fibers, fiber amplifiers, hair trigger
operation in fiber lasers, and light switching by
waveguide "beating."

Poobicatdon ot fiber optics to various photo-
electronics devices, data processing, and photo-

copying systems. In this field of photoelectronics

ILS,

alone, fiber optics have been applied in multi-
stage image intensifier coupling, high resolution
cathode ray tubes, end window vidicons, and
various forms of scan converters.

Application of fiber optics to the field of
medicine: cardiac catheter assemblies to record
and observe oxygen saturation of the blood;
application of fiber-optic endoscopes for appli-
cation to gastroscopy, bronchoscopy, rectroscopy,
and cystoscopy; hypodermic probes; in vivo cardiac
oximeter; laser coagulator for treatment of remote
tissues using fiberscopes; scintillating fibers
for radiology; endoscopes for the inspection of
the pericardium, thoracic cavity, bone joints,

living fetus and peritoneal cavity. [25/ [36/

However, before 1967, in the field of electronics,

glass fibers were not Seriously considered as a communications
medium for transmission over even moderate distances (about
1 km) because of high attenuation losses associated with glass
Primary emphasis prior to 1968 was on image trans-
mission devices of short length (5m) and illumination devices.
*Attenuation, or loss of light in a glass fiber, is expressed

in terms of decibels per kilometer (dB/km). This subject will
be discussed in more detail in Section III.

The first serious interest for communications was expressed
by K.C. Kao of Standard Telecommunications Laboratories in
England in 1968. At that time, technology was paced by the
ability of industry to draw fibers of long length and low
tess. /32/

In 1967-68, laboratories began development programs
Beomacvelop low-loss fiber optics in response to inquiries from
telecommunications laboratories. An attenuation level of
20 dB/km was set as an acceptable goal (Figure 1-1), since
that level of performance was believed to be compatible with
existing telecommunication systems configurations and would be
sufficient to tip the economic scales in favor of optical

waveguides. /8/

: 10,000

=) e

= | 0 i

i= |

ra) "e

= lO “e,

Oo @

gf otacsteace oe -

,o lOy LEVEL REQUIRED FOR = *

aa EFFECTIVE COMIVUNICATION :

= ° g
tt yp

o IgG? IS6e 1969 1970 IS7TIl 1972 197% 1974 1975

YEAR

Figure II-1 Historical progress in low-loss waveguides:
lowest attenuation achieved vs. year

At that time the fiber optic communication technology,
involving multimode fiber optic bundles and discrete semi-
conductor sources (light-emitting diodes, LEDs) and detectors
(silicon positive intrinsic-negative diodes, PINs), received
great stimulation and impetus from the announcement in November
1970 by Corning Glass Works of glass-fiber waveguides with
20 dB/km attenuation at a wavelength of 820 nanometers (nm).
(Commercial-grade fibers up to that time had about 1000 dB/km
attenuation). In 1971, Bell Laboratories developed liquid-
core low-loss fibers with losses less than 20 dB/km out to
1100 nm. (/35/

- In August, 1972, the Corning Glass Works announced
that they had surpassed the attenuation goals by developing
fibers with an attenuation loss of 4 dB/km at wavelengths of
850 and 1060 nm. Losses between 600 and 900 nm were all below
12 dB/km. /8/ By August 1974, Bell Labs had developed a fiber-
optic cable with an attenuation loss of only 2 dB/km at 1060
om. [6/

The development of low-loss fibers was not the only
obstacle to overcome, however. Even the mundane problems of
making connectors that worked and figuring out ways to repair
a broken fiber in the field looked like serious roadblocks.
There were so many problems as late at 1972 that few expected

fiber-optic systems to find anything but specialized applications

until the 1990s. However, most of the earlier problems have
been under parallel attack in dozens of laboratories around

the world, most notably in Russia, Japan and western European
countries. The stumbling blocks of 18 months ago have virtually
disappeared. ''This is one of the fastest-moving technologies

I've ever seen,"

says Don Alexander, who monitors cable develop-
ments from International Telephone & Telegraph Corp.'s head-
quarters in New York. /26/

"A lot of things have come to pass in one and a half

years instead of five,"

agrees Herbert A. Elion of Arthur D.
Little, Inc., who has been working on optoelectronics since
1968. Elion, who has been working on fiber optics with a group
of 27 clients from four continents -- both companies and govern-
ment agencies -- says that spending for development efforts in
fiber-optic systems topped $100 million in the past year (1974).
He expects it to double in 1975-76. "People argue about the
time scale," he says. ‘Some projects have been advanced from
1979 to 1976." [26/

Technologically, it appeared that it was feasible to
use fiber optics in communication and data link systems. With
unlimited potential for future application and the door already
cracked, it only remained for both industry and the military to

expend time, effort and money in research and development programs

in order to start reaping the benefits offered by fiber optics.

2. Background of NELC A-/7 ALOFT Demonstration Program

Man has employed optical means in military communica-
tions since ancient times. Early writers, such as the Greek
historian Polybius (c. 205-125 B.C.), refer to the employment
of visual signaling, including flags and smoke signals. Flag
and light codes for naval communications were developed by sea
forces during the sixteenth century. In 1875, the U.S. Navy
began experimenting with electric lights for signaling. By
1916, Rankine had patented a voice communicator utilizing a
vibrating mirror to modulate the optical carrier. The Navy
developed a cesium vapor lamp which could be amplitude-modulated
electrically at voice frequencies in 1944. Despite considerable
effort and ingenuity, however, practical systems were limited
to audio bandwidths until about 1961. By 1970, three advances
of potential significance were reported: the development of
the first injection laser which operated continuously at room
temperature, the development of the first continuously operating
dye laser, and the production of the first low-loss fiber optics
transmission lines. These, and other electro-optical advances,
such as light emitting diodes (LEDs), helped set the stage for
fiber-optic communications systems. /33/

While visiting England in 1970, Dr. John M. Hood, a
former student of H.H. Hopkins, recognized the suitability and

timeliness of fiber-optic techniques for naval and military

applications. Upon his return from England, he was instru-
mental in having a Fiber Optics group established in the
Electromagnetics Technology Department at NELC. The group,
funded by internal research and development funds, was dedi-
cated to the development of a practical technology for meeting
the problems arising from the specific uses that fiber optics
offers to the Navy. It was clear that a natural and obvious
application was to improve the internal data links of military
aircraft. It was also recognized that the potential for ship-
board use was just as great. By mid-1971, various agencies
of the Department of Defense (ONR, ARPA, NAVELEX and NAVAIR)
had committed funds for continuing fiber-optic research. In
April 1973, a Fiber Optics Development Plan was promulgated
at NELC, setting forth a program for identifying and meeting
the Navy's needs in the fiber optics field. This plan then
became the official NAVAIR-NAVELEX development plan. It has
since been superseded by the proposed DOD Tri-Service Technical
Application Area Plan for Fiber Optics Communications Technology,
Gated 25> March 1975. At the time of this writing, the plan has
been agreed to at the working level but not yet approved at
the command level. /14/

In January 1973, NELC entered into a contract with the
Federal Systems Division of IBM Corporation under contract

number N00123-73-C-1665 for the design, fabrication and

ILS:

laboratory testing of a high speed, multiplex fiber-optic data
link to interconnect the tactical computer and head-up display
from an A-/ aircraft. The work was performed at the IBM
Electronics Systems Center at Owego, New York, during the

period February to May, 1973. The final report was completed
mine 1973 by H.C. Farrell and RN. Jackson. {14/ In partic-
ular, the tests, made on the link between the ASN-91 computer
and the Head-Up Display (HUD) took the form of performance
comparisons between the fiber-optic link and the original
conventional shielded wire cable, as well as experiments on
special properties of the optical link. The results were
conclusive: in a noise-free environment there was no detectible
difference in performance between the two types of interfaces;
in the presence of an electrical noise generator, however, the
output display was unaffected when the signal was received via
the optical channel, but it incurred serious deterioration

when the shielded cable was used. Part of the laboratory tests
in this contract tested the link through the full requirements
of MIL-STD-461 and MIL-STD-462 (military standard specifications
on Electromagnetic Interference (EMI) and Radio Frequency Inter-
ference (RFI). These tests results were the first quantitative
validation that fiber optics were definitely immune to RFI and

EMI. /14/

The results of the IBM tests were made known to program
review officials in the Navy Department and the Department of
Defense. These officials recognized the need of a major feasi-
bility demonstration to design and implement fiber-optic links
at a full scale system level for test and evaluation. At this
time, NELC made a proposal to Commander, Naval Air Systems
Command (NAVAIR), for a two year program to install fiber optics
in place of standard twisted-pair and coax cabling in the
navigation/weapons delivery system (N/WDS) of an A-7 aircraft
for test demonstration and evaluation purposes. Subsequent to
this request, Dr. Malcolm R. Currie, Director of Defense
Research and Engineering, submitted a memo, dated 6 August 1973,
to the Assistant Secretary of the Navy for Research and Develop-
ment in which he expressed confidence in the role of fiber optics
technology for naval applications and thereby urged prosecution
a pLocram for exploiting it. /12/

This request culminated in approval by the Assistant
secretary of the Navy for Research and Development and subse-
quent funding-go-ahead by OPNAV 982 and AIR360 for the imple-
mentation of the A-7 Airborne Light Optical Fiber Technology
(ALOFT) Demonstration. The project was initially funded in
March 1974 under AIRTASK A360360G/003C/4W41X1-001. /14/

In July 1974, the Chief of Naval Material, assigned

the Naval Air Systems Command lead responsibility through

FY 1976 for the development of the fiber optics technology to
fulfill military systems needs and applications. Commencing
in FY 1977, the Naval Electronics Systems Command is designated
to assume lead responsibility of the fiber optics development
program. /11/
3. NELC A-7 ALOFT Demonstration Approach

As soon as the AIRTASK was received by NELC in March
1974 to initiate the ALOFT Project, NELC managers and engineers
consolidated plans and objectives into a formalized Development
Approach. The project was to consist of a two-year program
with a milestone schedule as outlined in Figure II-2. The
major project phases were as follows:

(1) A six-month system analysis and design effort to
be performed in part under NELC contracts to
define the system performance requirements, to
design the system, and to provide a system
installation plan.

(2) A six-month contractual effort to fabricate and
checkout the demonstration system in the contrac-
tor's system integration laboratory.

(3) A three-month test and evaluation program of the
demonstration system while installed in an A-/

ground simulator.

SQUOJISSTTW UOTIeAISUOWSeG TAOTY /-V Z-II eAansta

Moday vtAow Yaloug
vonviedarg woday puury

OWl9G WaASAS UOISsiTy Bpeso di
UONRNCAQ wWaIsaAs
my sicaf ByPi4 yesany fy

uolessuoiwUag
Wa}SAS 1OJOPNUNS puNngiH
UONENLAQ waysAs wy
SYS9f JOCINUNS punoig LW
Yap aoUadIIIY 2 OWIAG Gey
INOHIBYD -Quonyneosuqey wazsds

7 SBA, Weg) MIEN
Waly, , sisAyeurly SUIdISAS

L¥Od3uy | M31A3Y M31A3Y $}9049U05 337]
TYNid 9O0ud uyGd- fom
oe vi) Vv ae “NS < SMAIAGY /Peayly OD yalosg
Prof w vy Wirfonosvur rnvywWwa4arlaqnos wi SHINOIW AD

TG ee yan Luvs Y3LdV AD

SOUOISOHW (L4AOTY) wOoNeujsuoWAag
Asojouyda| ssqiy jeondo 1YysI7 ausoguly Z-Y

sisAjeuly 1s049 QJIAD ajiy

Oo F Ww

ro)

(4) An eight-month test and evaluation phase of the
demonstration system including aircraft modifica-
tion, ground check, and flight test while installed
im an A-/ test aircrare.

(5) Funds permitting, an economic analysis is to be
performed concurrently with the checkout, test and
evaluation of the demonstration system; the objec-
tive of which will be to analyze the comparative
cost and performance benefits of the fiber-optic
system versus a wire interconnect system.

The possibility of utilizing Naval Postgraduate School
students' theses efforts to conduct independent research and
give complimentary support to the economic analysis was first
discussed by NPS students and NELC (Code 1640) in early 1974.
The resulting proposals of theses investigations in this area
proved desirable to both NELC and NPS. See Figure II-3 for
A-7 ALOFT economic analysis activity flow.

4°, A-7 ALOFT Demonstration Management Organizational
otructure

The A-7 ALOFT project is assigned to NELC under the
Aircraft Internal Communications Project Office, Code 1640.

A project has been established within Code 1640 for the manage-
ment of this project. The basic ALOFT organizational structure

is shown in Figure II-4.

moTZ AATATIOe stTsAyTeue oTWOUODS TAOTV /-V €E-II ensty

-AANLS (0.13 ‘SIVG “SOVIG)
SNE VLVG W3LSAS DINOIAY IWLOL V 3L3 1dWOD

ALIALLISNAS SISAIVNY LIS3SN28

{ ‘wosasat CN Os LYOddNS SON

SL1IS3N39 AYNLNG
ASILNVND ONV SNIWY83L30

SAIAILVNYSLIV ANVY ONV
SLISSN38 GNV LSO9 3LVNIVAS

.SSAILVNYS LIV C3193 7139S ANVY OL G24/0%

3LVNIVAI3SY ONV ASIOOW

SNOILOSFONd LSO9 SOILdO VLVG LSCO S311d0

¥Y398I3 IWLOL SLVWILSS HY39t14 1037109 SOdN

TAGOW SISATVNY

SNOILOSFOd LSOD LOSNNOOYSLNI VLVG LSOD LOSNNOOYSLNI
DIWONOD3 GINS

3YHIM IWLOL SLVWILSS 3Y¥IM 1937109 YOLOVYELNOD

DANJONAAS VUOTAIeZTUeESIO VOTFFO 2O9elOAgG LAOIV /-V b-IT[ oanZs ty

CSie a

SUIYSIT punoury
pue f‘usisog uoijeagajuy
UOT}VI[VISU] 2-V

asa “Wadi

Pl9old pue uoljeotsqey
‘udtsaq waysks

O0bb 9P9D - OTUN

OPSS YP POD = OVICIN

Sjuoulorinboy
Juilesulsug usisaq
SIUMpPIey] TeJUIWUOI AUT

yzoddNs [eotuyso |
yofoug pue juawdoraaag
surayskS a1idO voarg

APO SOL
WATT AVeN

yUuauIas sue; yolorg
EAO1V L-Vv

Under the current program effort in the A-/ ALOFT
project the economic analysis function will be expanded to
include some in-house management along with Naval Postgraduate
support and contractual assistance. This structure is shown
in Figure II-5, which does not present the other organizations

(see NELC-TD 369 for full organization).

~ NELC
CODE 1640

ECONOMIC ANALYSIS
MANAGER
NELC CODE 230

NPGS
SUPPORT CONTRACTOR

Picume ti=5 A-7 ALOFT economic analysis
organization structure

felis FIBERVOPTICS TECHNOLOGY AS APPLIED TO DATA LINK SYSTEMS

A. GENERAL

Recent breakthroughs in fiber-optic technology have made
the application of fiber-optic waveguide systems to military
information transfer entirely possible and feasible. The area
of avionics data transfer will possibly be the first major
application or beneficiary of fiber-optic technology. Several
military utilization applications have been studied and a
number of feasibility demonstrations have been made. These
studies have pointed up dramatic performance and potential
cost advantages for a wide range of system applications.

Engineers at the Naval Electronics Laboratory Center have
summarized the important properties of fiber optics as follows:
[30/

(1) Cross-talk immunity between fibers and fiber cables.

(2) Security from signal leakage and tap-in attempts.

(3) No electrical grounding problems.

(4) No short circuits which could damage terminal equipment.

(5) No ringing problems.

(6) Large bandwidth for size and weight. The increase in

bandwidth, combined with crosstalk/noise immunity,

makes miltiplexing at high data rates possible.

(7)

(8)

(9)
(10)

Gy)
(12)

(13)

Small size, light weight (glass is 1/6 the weight of
copper), and flexibility - thus, ease of installation.
Potential low cost - when considering common factors
such as size, flexibility, equivalent bandwidth, and
manufacturing quantity. The strategic availability
and cost of copper as compared to glass will play a
future role.

High temperature tolerance (500 to HOOO-C) .

Safety in combustible areas and hazardous cargo areas
(i.e., ammunition and fuel storage areas).

No copper (strategic material).

Potential Electromagnetic Pulse (EMP) immunity.
RFI/EMI, noise immunity (glass, a dielectric, does not

pick up nor radiate signal information.

FIBER OPTICS RELATED TECHNOLOGY

Certain principles, components and data link systems should

whole.

be discussed before delving into the actual components used in
the A-7 ALOFT project. This discussion is necessary for a

greater understanding of a multiplexed fiber-optic system as a

Attenuation

Light is attenuated as it moves down an optical fiber.

Light is lost both to absorption and scattering in the fiber.

35)

The absorption is determined primarily from the bulk of the
glass from which the fiber is made. It converts light into
Meat. The scattering is due both to the bulk material and to
fiber manufacturing defects. Radiation losses can also occur
because of bends in the fiber, but losses are not significant
unless bends are below a minimum bending radius.

Attenuation is a primary factor in the economics of
fiber optics communication systems. It determines a system's
repeater spacing, source output and detector sensitivity.

Attenuation can be measured in decibels because of the

exponential nature of light attenuation in a fiber as given by:

- XL
ome bee
where
Eimer o ce ot ercceiving end of fiber
Pi > input power
o = extinction coefficient
L = fiber length

Extinction coefficients are sometimes used but decibels
have become the accepted measure of attenuation.

eet
Perenuatwon (ab) = lO log Py

Pi
a aE) dB/10

The graph in Figure III-1 shows how a low-loss fiber's

attenuation changes with the wavelength used. It was obtained

by Corning Glass researchers using one of their 4 dB/km low-

loss fibers. (3/

Attenuation dB/km
O

|
0.6 0.7 0.8 ORS 1.0 {|
WAVELENGTH (4™)

Figure III-1 Attenuation as a function of
wavelength in a recent Corning low-loss optical fiber

Po Moaulaclon

Light, as a carrier signal, from a light source such as
a Light Emitting Diode (LED) must be modulated in order to carry
data. The traditional modulation techniques of amplitude and
frequency modulation require complex electronic circuitry, sine
wave sub-carriers, etc., and increase costs, as well.

Digital Transmission is the easiest modulation mode to
implement with optics. This results from the approximate

linearity of LEDs in which the light output varies directly to

Sy)

the drive current. The digital signal can be connected to
the LED input port through a driver circuit or by digitally
controlling the bias current to the LED. Where used as a
binary on-off keying device, this technique causes a logic 1
miput to give a logic 1 light output. /2/

Figure III-2 shows a block diagram of the typical
fiber-optic system using a digital signal from a LED source.
The signal could be a linear signal from a laser diode as well.
For high speed operation, one would use wide bandwidth amplifiers.

Modulation rates achievable with LEDs are considerably
lower than those possible with semiconductor lasers because
the rise times in the LED are limited by spontaneous minority
carrier lifetimes, rather than stimulated minority carrier life-
times, as in the laser. Nevertheless, very useful modulation
rates are possible up to a few hundred megahertz (MHz). Assuming
an acceptable fiber loss factor in the range of 50 dB/km, one
finds a 200 MHz limit with fiber optics for a 300 meter length.
This is primarily a function of the electro-optic devices
available. Coax, on the other hand, is limited to 20 MHz for
the same cable size and length, and a twisted pair wire pair to

1 MHz. /28/

|
| pa age
FN
1) sta DRESSSGH
| ome ee ee ee eee

yaw i

LINE
RECEIVER

C. FIBER OPTIC CABLE

Source: NEEE

Figure III-2 Typical interface systems

3. Multiplexing/Data Bus

One of the basic data transmission methods naturally
applicable to fiber optics is multiplexing, a well known
technique which provides efficient use of a transmission
medium. A large number of single-strand wires may be replaced
by a single twisted pair for transmitting information, or
Similarly, a single fiber-optic cable may replace many single-
strand wires or single-strand fiber cables. In short, multi-
plexing is the process of combining several information channels
and transmitting them over a single communications link. The
two primary methods of multiplexing are time-sharing and
frequency separation, as shown schematically in Figures III-3A
and III-3C.

The need to multiplex is becoming more and more evident
as avionics systems become more and more complex. Some systems
engineers at NELC feel that continued use of conventional
approaches might not be capable of providing the information
exchange required by future integrated, computer controlled
multiplexed navigation, fire control, and communications systems.

Multiplexing of an avionics system configuration provides
advantages in several areas: reduced weight, increased flexi-
bility, ease of modification, ease of maintenance, reduced life
cycle costs (attributed to reduced maintenance and modification,

irrespective of investment costs) and a higher survivability

mace. /2/
40

05310 NVS SIAN Zurxetdraqtnu eqeq =€-III ernst y

ozf:i0¥
6) g Y ISNNVWHO

WV Lay | vv 4 ONINWHS AONSNDSYS O 2 TSNNVYHO

sme SSE SENS

oF G 13NNVHO
SX
(saaunes WV>a) vv G TANNWHO
Sapna WIONSTIDAM

Y ISNNVHO

ONINVHS - AWIL OV

AND IB2) Band6

13/0} y-7074—) yaxardi inn
§911d0 er Te — uaXa alsin

yooeoeseoe 7. |

YAaX ATI LINAS

The optimum multiplexing approach is known as "data

' in which a central control computer addresses each of

bus,
several remote units in turn on a programmed basis in the

time division multiplex approach (TDM), or by addressing

remote units individually by suitable filtering in frequency
division multiplexing (FDM). Figure III-4.

Certain avionics systems of the F-15 have been multi-
plexed using the data bus system, with the avionics units tied
directly to the data bus through their own interface units.
Among the avionics units multiplexed on the F-15 are inertial
navigation set, inertial measurement set, navigation control,
radar warning device, radar fire control, air data computer,
heads up display and the altitude heading and reference set.
Total capacity of the system is one megabit. /1/

Multiplexing on the B-1l will be more extensive. The
system is designed to handle over 12,000 electrical signals
which will be multiplexed into a single twisted pair wire cable.
Each of three separate multiplexing systems will have a data
capability rate of approximately one megabit. /13/

Fiber optics do offer certain advantages over coax/
twisted pair cables, such as increased data rate capability
and better RMI/EMP immunity, but fiber optics don’t offer much
extra in terms of multiplexing alone. It is true that most of

the advantages of multiplexing can be gained by using conventional

ainjzoaqtTyore snq ejep AyeroAtTe TeotdAy, y-LII eanstTy

| eee es
pTVNIWESL I YNIWHSL
| VLVO | viva

TVNIWYSL

4
|

SALOVO8

‘Y3TIOYWLNOD
sng
Wid

L YINNVHO VLVG

| TANNWHO IOYLNOO
NOLS N NOD es °

NOILVNIWHSL Bid
3VWVO '378VO

cables. It is also true that there would not be substantial
additional weight savings achieved by replacing the few
remaining twisted pairs in the multiplex system with optical
fibers. Present fiber-optic technology does not permit a
direct one-for-one replacement of twisted pair used in the
data bus concept, since fiber optic cable "T" connectors and
"star'' couplers (multi-terminal connector) are not readily
available outside research laboratories. It would be possible
to install a point-to-point system with fiber-optic cables
running from each remote unit to other remote units and the
central computer. Although this design would not necessarily
mean size and weight savings, it could provide the EMI and EMP
advantages which are even more important in a multiplex system
where increased emphasis is placed on integrity of the signals

being transmitted on a few wires.

C. DESCRIPTION OF THE BASIC COMPONENTS OF A FIBER OPTIC DATA
TRANSFER SYSTEM
l. General
It should be kept in mind that the A-/7 ALOFT Demonstra-
tion is designed to utilize "off-the-shelf" components. The
short term objective is to prove the feasibility of a multi-
plexed electro-optic transmission system for integrated digital

airborne avionics systems utilizing the A~7E as a test bed.

The system, as designed, is a demonstration only and is not
envisioned as being a design prototype for future generation
avionics systems. Future fiber optics avionics systems would
not necessarily be designed to incorporate all or any part of
the present "off-the-shelf" technology of point-to-point
systems, e.g., discrete circuits, and multimode fibers. Rather
it is probable that future systems would be designed to incor-
porate improved LED or laser injected diodes, single mode
fibers, integrated optical circuits and a data bus concept, etc.
2. Glass Fibers/Cables

Light is able to propagate through glass or plastic
fibers because of the well known phenomenon of Total Internal
meelection Chirk). For this phenomenon to hold true it is
meeessanry Lor certain conditions to exist. First, light rays
must hit the entrance end at angles less than the critical
Peeieemesanegle, 0, Of Otherwise be deflected from the desired
course. Figure III-5. Second, the fiber itself must have met
exacting manufacturing standards to prevent surface imperfec-
tions which will cause absorption and scattering of light. In
particular, metal ions such as iron, nickel or cobalt --
normally used to color glass -- should be eliminated because
of their light absorptive characteristics. In addition, the
fiber (or fiber bundles) should be designed so as to prevent

leakage of light from fiber to fiber because of cross-coupling

effects. These effects are associated with the penetration
of light into the surrounding low-density medium. The pene-
tration depth is small, reaching at most 2 A( A=wavelength of
transmitted light). Leakage is therefore significant only in
sufficiently dense fiber bundles. /27/

A light ray undergoes a multitude of reflections even
when propagating along a relatively short fiber. Calculations
show that in a fiber about 50 microns diameter, there are up-
wards of 13,000 reflections per 1 meter of fiber length. ([34/

To prevent leakage of light, fibers are coated with
Special materials which provide a high reflection coefficient.
This material is usually a dielectric coating called the
outer cladding. The outer cladding has an index of refraction
(n) somewhat lower than the glass core. As a result, light
rays are trapped in the core by reflection from the cladding,
as shown in Figure III-5. Note that the zigzag path slows
arrival of some rays.

n, +

| ie EN 2-75 microns
Xe. | Yt Crhousandths
of a meter)
nN, where n,n, )n,

Figure [11-5 Typical glass-clad fiber. The optimum sheath
thickness is approximately equal to the wavelength of trans-
ferred radiation.

absorptive jacket
cladding f
a Tins
LE SC ae
Q microns

Figure III-6 #Large-core, solid-clad fiber. The diameter is
slightly wider than a human hair.

Light rays that graze the cladding at shallow angles
are reflected back into the core resulting in a zigzag path
for some rays while other rays follow essentially straight
lines along the core. Figure III-6. This zigzagging can
create problems in timing for long distance communications by
distorting the on-off digital pulses used for high density
communications. This particular problem, however, would not
be a factor in short distance data link systems in aircraft.

One method of eliminating the delay problem is to make
the central core so small (a few microns) that only a single
ray can pass through it. Figure III-7. Such fibers are called

"single mode'' fibers and must be used with lasers.
absorptive Jacket

cladding
dime aaa ei CO CE eee

Figure III-/7 Single-mode fiber. Core diameters are only a
few microns, typically on the order of the light wavelength.

Transmission of a light pulse down a single fiber
introduces a new set of considerations and limitations not
previously encountered. Present indications are that an
information rate of at least 3X10!9 bits/second should be
attainable in a single fiber guide over lengths of one kilo-
meter. /35/ Present techniques for utilizing single mode
transmission incorporate laser-injection-diodes as a light
source. However, one of the most troublesome problems is
splicing (joining) two fibers together such that the signal
can travel on without distortion or undue attenuation.

Graded-index fibers have an index of refraction which
gradually becomes lower from the center outward. Instead of
travelling in zigzag paths, light rays follow a roller-coaster-
like sinusoidal path. Figure III-8. The gradually changing
refractive index actually speeds up light rays travelling
farther from the central axis. This results in light rays
arriving at nearly the same time, even over long distances,

thus minimizing ''smearing'’ (nodal dispersion) associated with

ABSORPTIVE JACKET
ES Cn at A Bt Sade SW Rte erg Velo FORANEONE |

Figure III-8 Graded-index fiber. Its refractive index de-
crease with increasing rapidity (parabolically) from the center
outward.

other fibers. The most common graded-index fiber, known as
SELFOC (self-focusing), was developed by Nippon Sheet Glass
Company, Ltd., of Japan.

SELFOC offers several advantages over the total internal
reflection fiber including larger bandwidth with no appreciable
wave form distortion, and the capacity for single fiber imaging
and special multiplexing. The disadvantages are a lower
flexibility than the TIR fiber due to a larger diameter and
the difficulty in bundling SELFOC fibers effectively. /1/

SELFOC fibers are possible candidates for an optical
data link because of their major advantage in their capability
to preserve the mode pattern and the fact that the absence of
a core-cladding interface eliminates the potential source of
defects from impurities and scattering centers which may occur
during fiber drawing. However, in the opinion of R.L. Ohlhaber
of IIT Research Institute, the typical high attenuation (approx-
imately 200 dB/km) as well as complex fabrication procedures
and their associated cost all but eliminate SELFOC for long
distance communication at the present time. /35/

Individual fibers may be bundled into a cable (multi-
mode) no thicker than the lead of a pencil as shown in Figure
III-9. Fiber bundles have enormous signal-carrying capacity
for their size. Each fiber in the bundle, carrying signals

as rapid on-off bursts of light, has the capacity for many

thousands or, theoretically, even millions of voice channels.
By comparison, as pointed out in an article by Mr. John Free,
22-guage twisted-pair wire can carry 48 one-way voice channels
while a coaxial cable might carry 5400 one-way channels. /18/
For most applications many fibers must be bundled
together to couple them efficiently to available light sources
and to provide redundancy against broken fibers. For cylindrical
fibers, the closest possible bundling arrangement is hexagonal.
Due to the empty spaces between fibers in a bundle, only a
fraction (the so-called packing fraction) of the total bundle
area is capable of accepting light for transmission. This
fraction must be accounted for in designing applications

requiring a minimum light transmission for detection.

JACKET
FIBERS

20 to !25 mils
€ 500 to 3/00 Am)

Figure III-9 Fiber-optic bundle

Thirteen fiber-optic cables (multi-mode) are to be
used in a point-to-point system application of the A-7 ALOFT
Demonstration. The A-/ ALOFT Fiber-Optic Interface System
Components Requirements call for a cable composed of 367
fibers, each fiber having a diameter of 0.00215 inches. See
Appendix C. The cables are to be covered with a non-metallic
jacket and shield which is non-toxic upon decomposition. Such
a jacket might well be made of an improved dielectric plastic
polymer compound such as "Hytrel."’ Extruded Hytrel tubing,
made by Valtec Corporation, is completely flexible yet exhibits
erush-proof characteristics. /14/ Polyvinylchloride (PVC), an
early candidate for protective cabling, has been eliminated as
a candidate for protective cabling material because of its
toxicity upon burning and its poor mechanical characteristics
at high or low temperatures.

The fiber-optic cables for the A-/ ALOFT program are
of the medium loss category, with a maximum optical attenuation
of 590 dB/km at 910 nanometers wavelength. Cables with such
attenuation characteristics would hardly be suitable for long
distance communication links, but are completely suitable for
relatively short distances aboard ships or aircraft. Cables
with light losses of 350 dB/km means that half of the signal
is lost in less than 10 meters, half of the remaining signal

within the next 10 meters and so on. That's an enormous loss,

pill

but even so, enough light emerges at the end so receivers can
accurately decode the transmitted signals.

Long distance communications would require a lower loss
cable (i.e., less than 20 dB/km) as well as repeaters. For
example, if the one or two dB/km fibers developed by Bell and
Corning Labs were used, repeaters would be spaced every 10
miles. That's better than current wire and coaxial cables,
which require a repeater every few (approximately 4) miles. /18/

Current fiber-optic cables being used in the A-/7 ALOFT
project were supplied by Valtec Corp. Two hundred twenty-four
feet of this fiber optic cable is used on a straight point-to-
point system for ALOFT. It should be noted that transmission
requirements in the ALOFT system configuration could have been
met by 13 coaxial cables utilizing 224 feet of RG-316 coaxial
cable -- but only at the expense of increased EMI/RFI suscepti-
bility and with a slight increase in weight. /14/

3. Connectors/Couplers
a. Connectors
With any fiber-optic system there is always the
problem of connecting the fiber-optic cable at either end to a
light source and a data receiver. The fiber surface at either
end must be rigidly held in position. The ends are polished
and anti-reflection coatings are sometimes added in order to

reduce attenuation. In the case of the connector at the source

ena, tt must be positioned such that a majority of the light
from the source falls within the acceptance angle of the cable.
Figure III-10. In the case of detector coupling, the detector
surface must be large enough to collect the spreading output

light.

rn detector
ZG ‘surfoce

Figure III-10 Multimode cable connectors

A single mode fiber, offering bandwidths up to 10olt
Hertz, requires critical source alignment with a laser source
because of its small size and small numerical aperture. Numer-
ical aperture, NA, is defined as a measure of the light
gathering capacity of the fiber:

NAS aes rn OF

where n, is the refractive index of the material outside the
mIDereandmessas) the incident angle of the light ray.

Multimode fibers offering a bandwidth of 10° Hertz
can be very easily coupled to multimode emitters (e.g., Light

Emitting Diodes), which operate at low power and are more

efficient. They are also less expensive than a laser source.
Low power operation and efficiency are intrinsic properties
of LEDs -- not causes for the lower costs. The basic problem
of LED-to-fiber and fiber-to-detector couplers is to maintain
the proper geometry for efficient coupling. Extremely close
tolerances other than concentricity are not required. Simple
machined housings and epoxy cements are proving adequate.
Sealectro Corporation has supplied NELC with hermetically
sealing connectors of the type shown in Figure III-12.

Simple machined housings are all that is required
of multimode fiber-optic connectors. By comparison, stringent
capacitive and inductive design requirements of electrical
connectors cause housings to be more complicated. Often, parts
must be gold-plated in order to satisfy these design require-
ments’.

The problem is not simple when considering multi-
channel connectors. ITT-Cannon Corporation had to tackle that
problem in order to design and build a 13-channel bulkhead
connector for IBM to mount in the wall of the A-/7 computer.
Five prototypes were sold to IBM. One was delivered to NELC.
The development of this connector is undoubtedly of importance,
as explained by Mr. Anderson of Galileo Corp. when he says,
"The development of this connector could be among the most
important developments of the entire program (ALOFT Demonstra-

mon). 1soy
54

Little information is available on single fiber
connectors. Alignment of the microscopic size core of a fiber
wien a source or detector surface can be critical. Present
methods normally involve imbedding fibers in a substrate
material or using epoxy cement as a binding agent to hold the
fiber in alignment. The Deutsch Company developed a mechanical
single mode connector for Corning in mid-1975, but further
information was not available to the authors.

One of the biggest problems of interconnecting
fiber optics involves the question of just where to make the
connections. Some feel that the LED-fiber interface is the
most obvious interconnection point while others feel that the
critical nature of the optical interface will prevent making the
connection at that point. NELC has considered three basic
Bopeeeehes to the problem, Figure I11-13. They have decided
that for the present, the optical interface has several advan-
tages over other proposed fiber-optic interface methods:

(1) Elimination of contact discontinuity at the "break
point” because of the optical coupling instead of
electrical contact. This eliminates such connector
problems such as oxidized contacts, mechanical
Beolobmetey (bene pints), etc,

(2) Throw-away modularity. The electronic circuitry,

LEDs, etc., could be replaced if either failure

bd
occurs or technology advances necessitate updates.

Nya, FIBER OPTIC
‘ USS) REPLACEMENT CABLE

Paewies LT T-1) Typical fiber-optic link replacement.

HERMATIC
SEAL

METAL SHEATH

a
ee a Hy

BULKHEAD

ENCLOSURE

EEECTRONIES

LEO
(PHOTODIODE)

a BULKHEAD ~ Source: NELC

CONNECTOR

Figure III-12 Standard package -- fiber-optic module

JV8SSOd 131 ON

S3A

S34
OM
ON
SJA

S3A

C~SOVIVILNI WOYLIITI 9

S IN3IS O39

PhLitiowmiv

] Qievase wv wy

SooRejz9qUT 02 sayoeroradde otseg

CI-LII ean3sty

HY YEAS Us

MGEVAL Md Av gPilss te |
DUO Bicis WLUA Sab ye

a a ae OE

(2_eE ee eee SUL

SHIVGOL WIS
SIQGS DLS Sots
|

LAOHUYA SOL ALIHIEOD 5

SLYMOs UY) Waste
Guvanyis HIPAA ATLOFE!O SHQYIG GE KY

Ut S3A | WIVISSOND WHNVHD LHGINIGY OL Apa

ON SjA
ON SUA
Sih
ON SHA
0} S3A
ON STA
ON S34

S3A
S3A

S3h

(HIGG) G3sCH “SIA

L—FOVIUILNE WOIWLIFTT 8

Swag

JOVSUILE! Wolldd

SuUscid (iw3s 073)

2 Pew ee ae 408 Shee

| OUATWAD 29d GUS WHF ise GUE Xo

i GVA NG

Cl Si V4lSSi0 Ls
B NOISHO Lod

aa ad
la

ae) ee oe eee oe ee ee

w VAI Verh eas

| er
| LRIERIGUIAND GEIVS NE LAGS G2 Ghz

sn EES we Cn a er ee Bl

MuyyaAgOS Aeny ar

ALD YINGONS Ave ED

Gull

{

ae?

Lysis ng |
fa

;

WIS O-F/LNIUD GILG A
19444
dy?)

MOLINGNOIIS O-9/ LINDT LY

ee ene

SHIISSY sT8¥I 30 | QUV HL Gad: a3
ase iaee tales ety SANNEDE
ALIABILHOISIO LOVLOI Kid W3idiJat:

ae
Ka

PU UUI Ap tot Jessel
QINOYLIIITI AHsGOM YO SHLSIXE Sson

<j——_ S3HVONddY dIS¥e
Nar NYS YATRA I

a vitvy ==” 2 “mene @

(3) Diode-circuit matching and engineering is no longer
the systems designers’ problem.

(4) Easy to install and replace.
b. Couplers for Data Bus Applications

The concept of the data bus is becoming increasingly
evident in the design of new generation aircraft. A data bus
system is potentially less expensive to install and maintain,
lighter in weight and smaller in size, more reliable, easier to
modify and expand, and less vulnerable to damage than systems
based on point-to-point links.

If fiber optics are to be considered as viable
replacements for electrical lines in data bus systems of the
future, properly designed couplers, junctions, and terminations
must be perfected.

Successful laboratory models of both single-access
Trunk Couplers (T-couplers) and multi-access (star) couplers
have been tested at NELC. It was concluded that star couplers
make it possible to implement a data bus with a large number
of terminals without a repeater. If a system using "'T" access
couplers is used, a repeater is necessary if there are more
than ten terminals. It was concluded that the information flow
requirements of a modern military aircraft can be met using

either access couplers or star couplers. /38/ If the number of

terminals is large, resulting in unacceptable attenuation
levels, a repeater would be required in an access coupler
system.

4. Light Sources/Signal Drivers

Various types of light sources can be coupled to fiber
optics for useful purposes. For instance, typical tungsten
filament lamps, bulbs and other common light sources are used
in connection with fiber optics in market areas which include
TV, stereo and appliance illumination, gas and electric burner
pilot light indication, dashboard and cockpit instrumentation
lighting, medical endoscopes, etc., and monitoring of remote
light sources. However, for communication purposes, only the
semiconductor laser and the light emitting diode appear
attractive for interconnections on aircraft and spacecraft.

The signal driver for the A-7 ALOFT Project utilizes a
discrete circuit driver-amplifier with LEDs, resistors, capaci-
tors, and integrated circuit amplifiers all mounted on a circuit
board. The much more desired hybrid fiber-optic driver is yet
to be delivered to NELC by an impending contract. It will be
delivered at too late a date for consideration in the ALOFT
Eroject.

a. Light Emitting Diodes (LEDs)

Light Emitting Diodes are the most widely used light

source today. They will be used in the A-7 ALOFT Project because

Ne,

of their availability and their operating characteristics

which readily satisfy important characteristics which must

be considered in the selection of a light source for fiber-

optic systems. These characteristics are: /2/

(1) Wavelength of light output within frequency spectrum

detectable by available photo detectors.

(2) Size of light source emitting region is compatible

(3)

(4)

(5)

(6)

with the multi-fiber cables.

The power requirement 1S compatible with the air-
craft electrical system (£ 28vdc). Specifically,
it is TTL compatible (4 5vdc).

Coupling efficiency allows light emission such that
output power is radiated with an angle, 9, for
efficient coupling to a fiber-optic cable. In
addition, power efficiency of LEDs provides sufficient
light to overcome coupling and cable loss and does
not require external cooling.

The response time of LEDs is fast enough so as not
to distort high rate (15-20 Megabits) signals.
LEDs, which have a much longer lifetime than laser
diodes, are believed capable of operational life-
times measured in hundreds to thousands of

continuous hours at 25°C.

A light emitting diode is a semiconductor chip
which contains a P-N junction, mounted in a header and encap-
sulated beneath a transparent window. This semiconductor
basically converts an electrical signal from the aircraft
electrical system into an infrared (~ 90004) light for
transmission through a fiber-optic cable. Light emitting
diodes make use of a P-N junction for light generation in
Pech the same way as injection lasers except that no optical
resonator is used to control the gain in the device.

The intensity of light output from the LEDs is
proportional to the current through it. Thus, the amplifier
output current controls the light intensity. Since LEDs
operate at much lower current densities and optical densities
than semiconductor lasers, they do not suffer unsolvable
degradation and reliability problems.

The amount of information an LED can transmit is
limited by its frequency response -- how fast it can be turned
on and off. At this time LEDs can be modulated up to a few
hundred megahertz. This is suitable for some 50,000 voice
channels, which require 4000 Hz of bandwidth each, or some 30
TV channels, each requiring six MHz of bandwidth. /18/

Driver requirements for LEDs are much less severe
than for semiconductor lasers. In general, the voltages on

the LEDs and semiconductor lasers are approximately 2 to 3

Polts. For some applications, LEDs with one TIL output, can
be driven with currents of approximately 20 mA at frequencies
not exceeding 30 MHz. For these device applications, transistor-
Beansistor-logic (TTL) circuits are convenient drive circuits,
whether they are off-the-shelf items or custom integrated
circuits. /1/ In summary, electronic drive circuits for LEDs
and semiconductor lasers are readily available for requirements
at least to 50 MHz.

b. Semiconductor Lasers

Of all the laser sources, the semiconductor laser
holds the most promise for high data rate fiber-optic systems.
Their characteristics of small size, simplicity of design,
ease of high frequency modulation, and relative high power
conversion efficiency make them ideal as light sources. /1/

A laser diode can be as small as a speck, barely
visible to the eye. They are capable of emitting very narrow
Spectral outputs (spectral widths of less than a nanometer are
possible), which makes them ideal for the microscopic core of
a single mode fiber. Lasers can be pulsed in the gigahertz
range, and thus can transmit far more information than an LED.

Several companies are now working on laser injection
diodes. Corning Glass has developed fibers which have a square
cross section which can be bonded side by side into a flat
ribbon. This ribbon can then be bonded to a laser injection

diode as shown in Figure III-14. /1/
62

2-MIL FIBERS

96 -MIL
LASER DIODE

Figure III-14 Laser injection diode bonded to fibers

A few of the companies involved with laser injec-
tion diodes are: Sperry Rand, IBM, Bell Labs, and Texas
Instruments. Bell Labs revealed in mid-1975 that they have
been able to integrate familiar optical components such as
lenses and prisms on special substrates. Bell has also inte-
grated all the components needed to generate, modulate, deflect,
and detect optical signals onto a single chip. /18/ Most optical
engineers feel that the greatest potential of laser diodes will
be realized when integrated optical circuits (10Cs) are as
common as integrated circuits (ICs) now used in calculators
and other electronic equipment. Instead of transistors on a

button size surface, IOCs will have microscopic lasers, modulators

(to put signals on a laser beam), photodetectors (to convert
light back into electronic signals), and optical switches to
route light into fibers for long distance communication. /18/

Figure III-15.

£LECTRO OPTIC ELECTRO.
MODULATOR SwiTe "a

7 = 2
Ty A
we = {caste
a

pepo
g

GaAs SUSSTPRATE
WAVEGUIDE FGERS sae

Figure IL1i-15 Optical circuit on a chip as envisioned by
Texas Instruments. Bell Labs recently formed such components
on a single chip.

An important consideration of semiconductor lasers
is that they can be modulated at extremely high rates. The
modulation rate is intrinsically limited only by minority
carrier lifetime in the semiconductor crystal. ae life-
time has been determined to be less that 10719 seconds, which
implies a modulation rate capability of ten gigahertz. /22/
The bandwidth available is phenomenally higher. Light wave-
lengths involved translate into some 500,000 gigahertz -- enough
bandwidth, theoretically to carry some 83 million TV signals

simultaneously. The limitation in signal carrying capacity is

how fast light sources can be modulated. /18/
64

5. Signal Receivers/Detectors

For the purpose of this study, detectors are analogous
to a receiver. Optical signals are required to be demodulated
through use of a photodetector which is sensitive to low light
signal levels at the incident wavelength. A photodetector is
a device in which the voltage or current output depends on the
intensity of light falling on the light sensitive region of
the device. The incident photons cause hole electron pair
formation in the junction region which causes current to flow
through the junction to an external load resistor which causes
a voltage drop proportionate to the incident photons striking
bme detector junction. /28/

Detector requirements for fiber-optic applications are
not particularly unique and much of the technology which has
been developed in the past is applicable. However, some very
important characteristics and requirements must be considered
for fiber-optic applications: (/2/

(1) Wavelength of transmitted light must be within the
region of wavelength sensitivity of the receiver.
(2) The size of the light sensitive region of the

receiver must be compatible with the particular

fiber-optic cable for efficient light energy

coupling.

(3) The electrical power system of the aircraft must
be compatible with power required by the detector.

(4) Sensitivity must be such that incident light rays
from the original signal source can be demodulated
with a minimum amount of distortion.

(5) Mechanical constraints, such as simplicity, light
weight, ruggedness, temperature coefficient, etc.,
must be met.

These conditions can be met by using commercially
available positive intrinsic-negative (PIN) diodes with
commercially available amplifiers. PIN diodes are quite
satisfactory for short run applications such as the ALOFT
system, but the avalanche photodiode is preferred in the long
run where greater sensitivity is required in the bandwidth
regime out to 15 megacycles per second. This improvement is
obtained at the cost of more complex biasing networks and less

proven reliability. /13/

ieee oro ee) DESCRIPTION OF FIBER OPTICS AS EMPLOYED IN THE A-/

ALOFT DEMONSTRATION

1. System Description

The original A-7 data communication system as utilized
by the A-7 Navigation Weapons Delivery System (NWDS) is a point-

to-point system which uses twisted pair wire and coaxial cable

mmecriaces in the Navy and Air Force versions of the operational
aircraft. Certain portions of that system, as shown in Figure l,
Appendix A, are being converted to a multiplexed fiber-optic
interface by the A~-/ ALOFT Project. The original wiring will
be left in the aircraft and will be reconnected for use upon
completion of the A-/7 ALOFT Demonstration. Since no change

in the input/output (1/0) design of the avionics (other than

the computer) was authorized, the fiber-optic interface with
the peripheral avionics units has been achieved through

external adapter units which contain all electro-optic and
multiplexing/demultiplexing (MUX/DEMUX) circuitry and which are
connected to the avionics with wire adapter cables.

The data communications encompassed by that portion of
the system shown in Figure 1, Appendix A, which has been con-
verted to a multiplexed fiber-optic interface by the ALOFT
Project, consists of 123 signals. After electronic multiplexing,
these signals are transmitted in the ALOFT Project over only 13
point-to-point fiber-optic cables, as opposed to approximately
300 wires which were required to transmit these same signals
in the original A-7 system configuration. The fiber-optic
configuration of the system is shown in Figure 2, Appendix A.
Figure 3, Appendix A, shows only the electro-optic, MUX/DEMUX

and fiber-optic portion of Figure 2, Appendix A, that is being

installed in the ALOFT Project. The computer shown in Figures
2 and 3, Appendix A, is an internally modified version of the
original A-/7 computer containing all necessary electronic
multiplexing/demultiplexing circuitry to reduce the interface
density required for the transmission of the signals to 13
channels of information flow at a maximum of ‘a 10-megabit data

rate. /14/

IV. AN APPROACH FOR A COST-EFFECTIVENESS STUDY

OF AVIONICS DATA LINK ALTERNATIVES

A. GENERAL

Engineering research and development of fiber-optic
cabling in aircraft has reached the stage where it is approp-
riate to begin assessing the cost and effectiveness of this
emerging technology as a possible replacement for coaxial and
twisted pair wire cabling in avionics data transmission wiring
suites. The general approach for the analysis, as required
by SECNAV INSTRUCTION 7000.14 A, and as desired by NAVAL
ELECTRONICS LABORATORY CENTER, SAN DIEGO, is an economic
analysis. An outline of this process is provided in Figure
IV-l. The basic format of an economic analysis involves the
determination of the cost and effectiveness of each of the
competing alternatives, i.e., fiber optics and conventional
wiring. Once this task is accomplished, the decision maker
should be better able to make a rational choice between the
competing systems.

The framework for cost and effectiveness analyses for any
system usually follows one of two conceptual approaches:

(1) Fixed Effectiveness Approach - For a specified

level of effectiveness to be attained in the

ssoo0id stsATeue ITWOUODY T-AI ean2ty

SSATLYNUSL TY suVdHOo

SLI4INI@ GNY SLSOS
TEVISTINVNO-NON S91YdSzd

JALLYIN ILIV
HOV3 30 1S0)
SATLVLILUVAD SNIEM IS

A Aye
TER

QOHLIW ONTLVWIL
LSOQ dLVIudOddd¥ 3800

JATLVUYILIV HOV2 JO SLISINIG
JAYLVLILNVAD 3NIHY3L30

INSNBUNSV3H LI43N38 YOs
VINBLIND TLVUdONddY ATddY

E
ae

SNOILdWASSY SLY INWUOS

accomplishment of some given objective, the
analysis attempts to determine that alternative
which is likely to achieve the specified level
of effectiveness at the lowest economic cost.
(2) Fixed Resource Expenditure Approach - For a
specified cost level to be used in the attain-
ment of some given objective, the analysis
attempts to determine that alternative which
is likely to produce the highest effectiveness. /17/
While either approach is possible, the fixed effectiveness
approach might be more appropriate for the alternatives being
considered in the case of data link systems. The fixed
resource expenditure approach would apply more to an entire
weapons system purchase, such as fighter aircraft, where a
resource constraint can probably be more easily stated. Further,
fixing resource levels would require extensive and detailed
cost data at a subsystem level which is, in most cases, not
available. Therefore, the authors feel it is appropriate to
fix effectiveness at a desired level for both competing systems
while minimizing costs.
A level of effectiveness as referred to in most cost-
effectiveness publications usually relates to a single measure
of effectiveness and the unit values that may be achieved for

a given unit cost. In fiber optics there exists a myriad of

fit

effectiveness measures that must be evaluated. A level of
effectiveness for fiber optics would therefore consist of the
quantification of all the MOEs.

Each specified level of effectiveness will have a cost
associated with it resulting in the well known cost-effective-
ness curve. Figures IV-2 and IV-3 serve as examples. Figure
IV-2 illustrates the case where one alternative, B, exhibits
"dominance'' over its competitor, A, in every case. When dom-
inance occurs, there is little need to proceed further with
an analysis. Common sense would clearly indicate a choice of
the dominant alternative. Figure IV-3 illustrates a case
where alternative A exhibits dominance over alternative B over
the range of the first four levels of effectiveness. However,
alternative B is dominant at effectiveness level five and
above. This could be of considerable significance if, for
instance, weapons systems designers and decision makers insisted
on acquiring a system capable of operating at effectiveness
level five or above. Both alternatives could reach level five
but at considerable cost differences. The obvious choice in
this case is to choose alternative A for the first four levels
of effectiveness and to choose alternative B if effectiveness

Hevelertive is desired.

6 ALTERNATIVE B
ALTERNATIVE A

taf

ues

LiJ

= 2

lid

a /

ea 2 3 4 S - 6
COSst LEVEL

Figure IV-2 Hypothetical cost effectiveness curves
displaying dominance

ALTERNATIVE B

ALTERNATIVE A

EFFECTIVENESS sever

[ 2 3 4 S 6
COST LEVEL

Figure IV-3 Hypothetical cost-effectiveness curves

B. MEASURES/LEVELS OF EFFECTIVENESS

The actual determination of measures of effectiveness
(MOEs) will be accomplished by a contractor to be determined
by NELC. Possible MOEs can be derived from the physical
characteristics of the equipment such as weight and size as
well as system performance characteristics such as data rate
capacity and mean time between failure. Figure IV-4 lists
several possible MOEs along with suggested measurement scales.
If multiple MOEs are chosen to define effectiveness, a vector
of MOEs will result. Collapsing the vector to a scalar intro-
duces two problems. First, a method must be determined to
combine MOEs measured by different scales. Typical scales to
be considered are ordinal, linear interval, and ratio scales.
Ratio scales, as used in such measurements as weight, volume,
and mean time between failure, are special scales which have
a natural zero point and an arbitrarily defined unit size.
Linear interval scales, as used in measuring degrees centi-
grade, have an arbitrarily defined zero point and an artibrarily
defined unit interval. Ordinal scales are measures of relative-
ness. Examples of ordinal scales include measures of hardness,
measures of deterence, and degree of EMP/EMI immunity. In fact,
many utility indices, such as rankings of cost and/or effective-
ness issues by individual decision makers, are representable by

ordinal scales. The second problem concerns the relative

weights that must be assigned to the components (MOEs) of the
effectiveness vector. Such assignments are necessarily
subjective because they depend totally on the judgment of the
individual making the weighting assignments. Both problems,
combining MOEs measured by different scales and assigning
weights to the MOEs, can be eliminated if the fixed effective-
ness approach is utilized since the competing systems will
have the same effectiveness level.

When the relevant MOEs are determined, actual magnitudes
can be assigned. The right hand side of Figure IV-4 illustrates
five hypothetical assignments. The assignments represent five
different levels of effectiveness that may be required or
desired of the competing systems. Once the costs are determined
for the competing systems at the different levels of effective-
ness, the cost-effectiveness curves as illustrated in Figures
IV-2 and IV-3 can be constructed. These curves then provide
the decision maker with the necessary information to make a

rational decision.

C. COST ANALYSIS
l. Life Cycle Costing
The costing methods as required by NELC will be done
imeemaeon the life cycle costs (LCC) for both a coaxial and
fiber-optic aircraft avionics system configuration as repre-
sented in the A-7 ALOFT Demonstration.

(Tedtzeu_ZOdAYy) sSsUdATIOIFFO JO Saanseop

y-AT 294N3T J

Ot=

000° 0€

O¢
002
SHA
se

Sak

Cee

000 °0€

OT
OOT
OS
ON

Oc-

000‘0Z

O¢
002
OOT
ON

OOT

v7,

OT-

000*ST

OS
OO€
SZ
ON

Scr

(YSTH = ¢ ‘MoT = T)

SSHNHATLLOUSHa dO SHAT

000° OT

OS
OO€
OS
ON

Sot

u (004

Zod SiSTMyz)
I (9 $00130q)
T (9 $90139q)
T (saTo49)
I ($372)
u (wy /aP)
I (ZAASYeZIW)
O (ON/S2X)
I (seq;nor FO #)
O (ON/S9X)
ad (99eF SFqnd)
YW (spunod)
qY (szy “qUTeR)
UY (sry aYSTTT)

(A'IVOS)

LNAWAYNSVAN)

(PANT Te™ OF SISTA]
ues) BOUeTSTORZ ISTA]

(‘due Sutjzezsado
JSOMOT) BOUCRSTSIA pO)

(‘due sutjerzedo
ASSYUSTH) FoueISTSeA Jeoy

(SAn{T Freq O02 SS [SAD
ues) DOUeASTOR uUoTIAeACTA

uotjdumsuos Aamog
uUOTJeENUsAIAIe [eUsTS
yaptmpueq Teusts
AQTunNWWT TWa/ dW
Aouepunpsr s{tqey)

AjTunuWT prezey
9ATF pue “Butrpunorzas *syAzeds

wo}SAS SoTUOTAR
aod ‘oun[oA 20 ‘azts

waj3SAS SoTUOTARe aod WY8TOM
Atedsy oF sWITy ues

SIN[Tey vs9eMjJog oUTT ues,

SSANGALLOG SHA JO SAdMOSVAW
IVOLLAHLOdAH

The task of estimating life cycle costs for wire inter-
connect components (coax/twisted pair, etc.) will be accomplished
by a firm yet to be chosen by NELC. Currently, two NPS students,
CDR R. Johnson and LT E. Knobloch, are developing a life cycle
cost model which, together with the two suggested costing
methods of this thesis, can be used by NELC to prepare cost
estimates for the fiber-optic alternative.

Life cycle cost estimates used for making a particular
decision, such as data link selection, need not be the total
life cycle costs for the system. Costs which would be the same
for each alternative and costs incurred prior to the decision
(sunk costs) should be excluded. Sunk costs are those resources
(money, etc.) which have been expended and which cannot be re-
covered. They are therefore irrelevant and should not influence
future decisions. However, any assets created as a result of
such expenditures are relevant.

Care must be used in the choice of costs to be excluded
lest their omission improperly influence the decisions to be
made. For example, consider one aspect of the present cost
analysis. It has been decided to exclude the electronic equip-
ment (including the MUX/DEMUX components) not incidental to
drivers, connectors, cables, and receivers because those equip-
ment costs appear to be common to both data link alternatives.

Eis restricts the analysis to the trade-off between the costs

and reliability of competing alternatives. One might consider
whether or not the costs of the MUX/DEMUX units are really
common to both alternatives. For instance, electronics manu-
facturers might find it more costly to modify their standard
electronic units to accomodate fiber-optic components (such

as multi-channel bulkhead connectors, etc.) than to use off-
the-shelf units for conventional wiring/cabling. Thus the
assumption on which the MUX/DEMUX equipment costs were excluded
would prove to be invalid.

NELC has undertaken the preliminary definition of
relevant LCC elements for the purposes of defining a LCC model.
The cost elements listed in Figure IV-5 were derived by NELC
after examining over 600 documents on life cycle costing, or
subjects related thereto. The search included local, navy-wide,
DoD, and private industrial firms that have apparent expertise
in electronics and communications equipment and systems life

cycle costs. /29/

2. Gost Data Collection Effort
Cost data collection, one of the first steps of a cost
analysis, provides specific costs for elements of the system
on which a simple price tag can be placed, or for which a
nominal extension can be made of costs experienced in similar
programs. A literature search and telephone survey by the

authors confirm the generally known fact that cost data for

1.0 RDT&E Costs are assumed as sunk costs for this study only.

Z-0

B. 0

Investment Costs

NB BH HS HY PS PO
NAO EWN

Prime Mission Equipment (PME)
Installation

Support and Test Equipment
Initial Supplies

Initial Training

Inventory Management/Support
Initial Transportation

Operations and Support Costs

3.1 Operations

3.1.1 Operations Personnel
Sly Training
Dpuppert

Maintenance Labor

Replenishment Spares and Material
Transportation

Maintenance Training

Support Equipment Maintenance

Lo oO WH W LW
NO DO BD DO ho
in & WN Fr

meedre 1V-5 Life cycle cost elements

elements of a fiber-optic data link system, other than pre-
production prototype costs, are not generally available. It
is true that costs (i.e., price to the user) of fiber-optic
Systems components such as cables, connectors, receivers and
Beavers, can be obtained -- but these prices usually reflect
contract prices on one-time bids. The prices paid today are
not indicative of prices that will be paid tomorrow for com-
ponents produced on either a one-time contract basis or a
full production mode basis.

Most of the available cost information has been obtained
from NELC sources (APPENDICES C, D, and E). The cost informa-
tion provided by NELC has been verified by the authors as being
representative of the wide range of costs generally associated
with components of an emerging technology.

One of the principle reasons for wide cost dispersions
is the lack of standardization of component parts. For instance,
if one needed a fiber-optic system to perform a particular
function, and if this person was to approach several fiber-optic
manufacturers for bids, he would immediately be faced with the
problem of non-comparability of different manufacturers' compon-
ents. The customer would be faced with the problem of defining
perhaps dozens of his own desired design requirements: single
mode, multimode (how many fibers?), desired cabling (will it be

toxic upon decomposition?) , packing fraction, numerical aperture,

index of refraction, attenuation limits, flexibility, diameter,
per cent breakage tolerance, etc. After defining his needs,

he should not be surprised to learn that no two manufacturers
have similar cables to meet his needs, nor are standardized
couplers, drivers and receivers available. The customer would
find, however, that a fiber-optic system could be designed and
built to meet his needs -- but at considerably higher cost
than he might have first anticipated.

A few examples will be given to illustrate the uncer-
tainties involved in gathering data for component costs.

Although module driver/receiver units have cost in the
mamee Of hundreds of dollars, Mr. J. R. Biard, of Spectronics,
Inc., indicates that it would not be unreasonable to expect to
see prices for driver/receiver modules drop to a $10 - $12 range
when in full production. /7/

Galileo's 400 dB/km multimode cable was selling for
peeo0/f£t in 1974. It was selling for $0.75/ft in August 1975.
Mr. Rodney Anderson, of Galileo, indicates that he could reduce
Ehat price by half, or more, with purchase quantities greater
than 100,000 feet. He feels that his 35 mil fiber-optic bundle
could compete with micro-coax cable on a cost-per-foot basis
but, as yet, there is not enough consumer demand to generate
cost savings which in turn, with competition, would lead to the

lowering of prices below $0.75/ft. [4/

Mr. Robert Freiberger, of Corning Glass Works, states
that his company "has been in a full production mode for fiber-
optic bundles for the past 7-8 years." /19/ In 1974, Corning's
19-mode 30 dB/km cable was selling for $17.37/f£t when sold in
less than 5,000 meter quantities. Corning reduced attenuation
from 30 dB/km to 20 dB/km at the 820 nm wavelength while re-
ducing the price by 36 percent in 1975. The price in mid-1975
was $10.97/ft for purchase orders of less than 5,000 meters and
$5.56/ft for purchases greater than 5,000 meters. Corning's
current emphasis, however, is on single-mode cables rather than
multimode bundles. Corning's most important fiber-optic product
is a single-mode low-loss ( < 6dB/km) cable called CORGUIDE.
CORGUIDE presently sells for $13.50 per meter or about $4.11
per foot. This equates to about $.59 per foot for each low-loss
fiber as there are seven individual fibers in CORGUIDE.

"Corning is putting millions of dollars yearly into fiber

" states Mr. Freiberger. One of

optics research and development,
their recent developments, in conjunction with the Deutsch Co.,
has been the development of a hopefully reliable mechanical
fiber-to-fiber connector for single-mode cables. Corning's
efforts are aimed directly at capturing a major portion of the
potentially large market that will result from fiber optics

utilization by American Telephone and Telegraph Co. in the

1980's. Mr. Freiberger sees little chance of lowering prices

for a military market in fiber optics in the near future as

it would take a potential $100 million per year market to in-
duce Corning to drastically lower prices or alter production.
"In a full production mode, with markets above $100 million per
year,'' Mr. Freiberger states, “it would not be unreasonable to
Mook for costs of CORGUIDE to drop from $4.11 per foot to about
$.10 per foot. This equates to a little over l¢ per foot for

low-loss fiber."

Mr. Freiberger makes the interesting pre-
diction that Corning's costs of production for low-loss fibers
will continue to decrease. As this occurs, the currently less
expensive medium-loss multimode fiber-optic bundle (with
hundreds of individual fibers in each bundle) will become more
costly to produce than low-loss cables such as CORGUIDE. [19/
Costs for connectors are not, in general, as uncertain
as other fiber-optic component costs. The exception would be
the 13-channel bulkhead connector developed by ITT Cannon Co.
for NELC/IBM at a price of $500 each for a total of six connec-
tors. /31/ It has subsequently been reported to the authors
that ITT Cannon Co. has sold this same connector to a leading
aircraft manufacturing company at a price of $50.00 each. /14/
Single channel connector costs are nominally low at
$2.50 - $3.50 each. This lower price is generally attributed
to the fact that mechanical connector technology and manufacture

is not new. Connector manufacturing companies already have the

production base necessary to produce fiber-optic connectors
for multimode cables.

The authors were unable to obtain any meaningful esti-
mates of the costs of fiber-optic integrated optical circuits
(I0Cs). According to Mr. Biard of Spectronics, Inc., the
development of integrated optical circuits today is in the
same relative position that integrated circuits were in in
1958 -- a full three to four years before a firm production
base was established. /7/ Mr. Biard makes one clear distinc-
tion, however; in 1958, the electronics industry was receiving
substantial financial assistance from the U.S. Air Force for
the specific purpose of perfecting and developing integrated
circuits. The electronics industry today is not receiving
the funds and support necessary for the same pace of development.
Mr. Biard feels that unless more government funds are made
available for the purpose of IOC research and development,
fecerared optical circuit growth and development will be much
slower than the previous growth of integrated circuits. Ample
statistical data exist in the field of integrated circuits
such that meaningful cost analogies, for the purpose of pre-
dicting costs, could be utilized once a cost data base has been
established for I0Cs.

It is the authors' assessment that a more detailed cost

gathering effort was not warranted at the time of this writing.

Statistical data could not be correlated because, in many
cases, there was no common ground for comparison. It was
generally observed, however, that costs are definitely in a
downward trend. The rate of price decline will continue to
depend on demand and technological development trends but it
would not be unreasonable to expect prototype costs of some
components to be reduced by a factor of 10 within the next

few years.

D. COMBINING COST AND EFFECTIVENESS
1. Ordering Uncertainties Through Scenarios

Technological forecasting is by definition, an area
fraught with uncertainty. It has been seen in earlier dis-
cusSions in this work that technological developments in the
field of fiber optics have many uncertainties -- all of which
should be considered by a decision maker. For example, before
B decision maker can make a final choice of future avionics
data link systems, he must face the overall questions of how,
when and why to implement any given system. In the case of
fiber optics, he must concerm himself with the future technical
composition of the fiber-optic data link. One most certainly
would not choose the one-time application of discrete circuitry
used in the A-7 ALOFT Demonstration. In fact, technological

developments are accruing so rapidly, he might not choose any

of the now existing components. Listed below are several of

the important uncertainties about which a decision maker must

be concerned:

(1) Technological levels of sophistication desired for

(2)

fiber optics: multimode or single mode fiber-optic
cables; low-, medium-, or high-loss cables; point-
to-point or data bus systems; data capability rates
of kilo-, mega-, or giga-bits per second; discrete,
modular, or integrated optical circuits; rugged
strength or small size, light weight cables;
redundant or single path data links; LED drivers

or laser injection diode drivers; standardization

of components to meet military specifications; low-
loss T- and Star couplers; reliable single mode
mechanical connectors; bandwidth -- How much is
“enough,” etc.?

Avionics systems design requirements: Will military
decision makers insist on higher EMP/EMI immunity
standards for future avionics systems; Will data
transfer rate requirements for complex computerized
avionics systems be increased beyond present data
link capacity; Will wiring-path redundacy be required
for increased reliability/survivability; Should

avionics systems be utilized in any one type of

©)

military aircraft -- or all types of military air-
craft; Are fiber-optic data links desirable for all
weapons syStems; etc.?

Timing: When will each of the technological develop-
ments mentioned in (1) above be off-the-shelf available;
When will technological advances level-off enough to
preclude an existing generation of fiber-optic com-
ponents from approaching either apparent or perceived
obsolescence as happened in the case of lst, 2nd, and
3rd generation computers; When will there be sufficient
market potential to induce fiber-optic producers to
mass produce components; Does even the strongest
possibility of a military ''go-ahead" in this area
offer enough incentive for industry to establish a
production base for mass production; What market
potential (measured in millions of dollars and/or
millions of feet of cable) will be sufficient induce-
ment to industry; When will military design require-
ments force military decision makers to utilize fiber
optics in order to meet EMP/EMI immunity requirements;
When (and how much) will government sponsored R&D
funds be made available to industry and/or the
military for continuing research and development;

When will data transfer rates greater than the limits

of coax cable be required; When will the fiber-
optic data link system be technically and/or
economically feasible for avionics suites; What
is the earliest time frame we could expect to

use fiber optics in shipboard use; etc.?

The uncertainties described above can present a con-
fusing situation when taken together. Scenario construction
often offers relief in this area by structuring uncertainties
in a logical sequence of events in order to show how, starting
from the present, or a base year such as the beginning of FY
1977, a future state might evolve, step by step, to a terminal
date, say 1990. The purpose is not to predict the future, but
to refine information on the forseeable "climate" for various
fiber-optic technological advances and system utilizations.
Kahn, in the introductory chapter to his study on scenario
technique, emphasizes that "the scenario is particularly suited
to dealing with several aspects of a problem more or less
simultaneously." /24/

Through the use of a relatively extensive scenario, the
analyst may be able to get a "feel'’ for events and for the
branching points dependent upon critical choices. These branches
can then be explored systematically. The authors have attempted

to structure several events and branches on a representative

basis of (1) a neutral context, (2) a modestly optimistic
context, and (3) a modestly pessimistic context. It should
be emphasized that these are only three of an infinite number
of possible scenarios. An entire study could be made on the
dozens of uncertainty branch points and the resultant event
trees which could develop from each.

Two of the advantages that Kahn points out in his
discussions are: Scenarios are one of the most effective tools
in lessening "carry-over' thinking; scenarios force one "to
plunge into the unfamiliar and rapidly changing world of the
present and of the future by dramatising and illustrating the
possibilities they focus on.'' Secondly, scenarios "force the
analyst to deal with details and dynamics which he might easily
avoid treating if he restricted himself to abstract considera-
tions. Typically, no particular set of the many possible sets
of details and dynamics seems specially worth treating, so none
are treated, even though a detailed investigation of even a
mew Arbitrarily chosen cases can be most helpful." /24/

The analyst should be aware that certain dangers may
arise from the use of scenarios to help guide and facilitate
further thinking and analysis. Specifically, the initial con-
jectures might be assumed erroneously to be sufficiently
correct to lead to scenarios with some content of "reality."

However, as Kahn remarks, "'a specific estimate, conjecture, or

context, even if it is later shown to have serious defects, is
often better than a deliberate blank which tends to stop thought
and research."
a. Scenario I - A Neutral Context

The scenario begins in October, 1976, the beginning
of the 1977 Fiscal Year -- the year that NELC has chosen for
an economic analysis for fiber optics. One million feet of
fiber-optic cable is produced annually. Flight testing of the
A~/ fiber optics demonstration aircraft has been completed as
scheduled. The results of the first Delphi questionnaire have
been received ava refined. These results, together with the
life cycle cost model constructed by NPS and NELC, have been
exercised. The results are such that it has been decided to
expand the model to analyze a particular multiplexed data bus
system utilizing the building block components described in
NELC TD-435. (See Appendix C) By October 1977, with the
expansion of the models, a trade-off analysis is performed for
a data bus concept. The results of the economic analysis indi-
cate a choice of parity between fiber optics and coaxial systems.
There seems to be no question of the technical feasibility.
Test results indicate that desired EMI/EMP immunity can be
obtained. It is decided by military aircraft designers and
decision makers to utilize multimode fiber optics with modular

LED circuits on a limited number of aircraft as a pilot

application. A multiplexed multimode data bus avionics system
will be designed and installed in the Navy's 16 ES3 aircraft
being developed for the Tactical Airborne Surveillance Exploita-
tion System (TASES) program. The aircraft will be built in
1981. It is decided that this will be the only multimode
application. Future aircraft avionics designs will utilize
single mode if the TASES avionics systems work well. These
follow-on military aircraft of the mid-to-late 1980's will

also utilize 10Cs. By mid-19/7/7, fiber optics are still
primarily used in laboratory and test demonstrations. Demand
is small. Any one of the major fiber-optic cable producers is
able to produce enough high-, medium-, or low-loss fiber-optic
cable in a period of only a few weeks to satisfy the market
demand of the entire United States for one year. Fiber-optic
cable production in 1978 totals two million feet. As of 1978,
there is no standardization of components (cable size, connectors,
circuits, etc). Modular type driver and receiver circuits have
been produced in small quantities on contract bases for various
contractors to use in laboratory applications of multimode and
single mode fiber optics for the past two years. Successful
demonstrations of T- and Star-couplers have encouraged the Navy
to conduct further demonstrations of fiber optics feasibility.
Following the ''float-off" sea trials between the Rohr and Bell

Cos., the prototype 2000-ton Surface Effects Ship (2KSES) is

mall

to receive a fiber optics data bus system, utilizing multimode
cables, in 1978. The data bus system utilizes prototype T-
couplers and modular driver and receiver circuits. In 1978 it
is decided to use fiber optics in the avionics package of the
VPX (replacement aircraft for the Navy's P-3). The Navy, now
convinced of the technical feasibility of fiber optics, plans
to use a single mode fiber-optic data bus system in the VPX
when production commences in 1983. It is apparent during the
1978-1979 time period that the military is the primary user of
multimode fiber-optic cable. Cables, connectors, modular
drivers, etc., have been standardized for military application,
but much of this effort will be of questionable value as multi-
mode applications are planned to be phased out, during a five-
year period, in favor of single mode applications. Industry's
efforts are concentrated on technological developments relating
to single mode cabdles in conjunction with integrated optical
circuits. The sale of multimode cables to military consumers
has little financial impact on the producers. They are not
dependent on a military market. The Navy and Air Force have
decided against large scale retrofit programs. However, the
Air Force is retrofitting one B-1l bomber in a program similar
to the ALOFT Demonstration. The Air Force will utilize a single
mode data bus system in the B-l Demonstration. They will use

prototype components developed by industry. By 1980, the U.5.

Army has begun to replace an initial segment of four million
feet of tactical communication lines with single mode fiber-
optic cable. Their plans are to replace a total of 16 million
feet of 26 pair coax cable by 1985. One fiber optics producer
wins the contract, but he is still not dependent on the Army
for continuing profits, etc. Because fiber-optic cable is
relatively simple to make, he is able to stay far ahead of the
Army's need through use of his pilot plant facility, and in
fact can produce the entire 16 million feet of cable with only
a few months of productive effort. Industry is still concen-
trating on the single mode market. Production in 1980 totals
three million feet of fiber-optic cable. In 1980, the United
States is experiencing a rate of inflation of 6-7 percent per
year, but certain materials are considered "strategic" and are
in short supply. Copper is one of these strategic materials.
The price of copper (in terms of constant 1975 dollars) has more
than doubled, while the cost of raw materials for glass (also
in terms of 1975 constant dollars) has remained constant. There
are sufficient raw material reserves for glass in the U.S. to
last for an estimated 100 years. The cost of petroleum base
products has risen in a manner similar to that of copper and
thus has caused the costs of fiber optics protective cabling to
double. Almost all laboratory and test bed demonstrations

utilize single mode cables in conjunction with IO0Cs by 1981.

By 1983, low-loss (<5 dB/km) long distance fiber-optic cables
are a reality. The Corning Glass Co. is in a full production
mode for the production of single mode fiber-optic cable.
American Telephone & Telegraph Co., the principle receiver of
Corning's output, begins replacement of one million feet of
aging coax and twisted pair cabling. Six million feet of
fiber-optic cable is produced in 1983. One million feet of
cable will be replaced during each of the first two years.
This replacement rate will be increased to five million feet
per year in 1985. During the period of the mid-1980's, fiber-
optic applications boom, but the largest users are companies
in the communication industry. In retrospect, it can be Seen
that technology development rates during the late 1970's and
early 1980's were quite significant. However, production
growth rates were almost stagnant by comparison. Industrial
producers utilized their pilot operations to produce only
enough to satisfy occasional customers such as the military,
and experimental laboratories. In the early 1980's, the
military began to design avionics systems for single mode data
bus applications. Twelve million feet of fiber-optic cable
are produced in 1986. By 1987 there is increasing fiber optics
applications by computer companies, electric power companies,
aerospace industries, civil aviation firms, etc. This continuous

demand helps maintain a stable production growth rate of 50 percent

per year. in 1988, 23 million feet of fiber-optic cable are
produced. Component prices start a continuous decline over a
period of time in accordance with the experience curve theory
as explained in this thesis. Total industry output is 46
million feet of cable per year in 1990.
b. Scenario II - A Modestly Optimistic Context

During FY 1977, interest in fiber-optic systems has
increased to the point that other follow-on fiber optics demon-
strations are planned. The successful A-7 ALOFT Demonstration
has proven the technical feasibility of point-to-point multi-
mode applications. The cost models developed by NPS and NELC
are utilized by analysts who conclude that single mode applica-
tions will be used in yet to be determined future military
aircraft. The resounding success of the A-/7 ALOFT Demonstration
has helped pave the way for a similar demonstration with the
Air Force's F-15. Funds have been made available for the Air
Force Avionics Laboratory to replace the conventional coax data
bus system of an operational F-15 with a fiber-optic data bus
System. Prototype T-couplers and modular hybrid cricuits are
used with multimode fiber-optic cables. One million feet of
riber-optic cable is produced in 1977. In early 1978, infrared
light emitting diodes are beginning to be replaced with laser
injection diodes for laboratory applications. Monolithic

integrated LED circuits are introduced as standardized fiber-optic

components in 1978. In early 1979, the prototype 2000-ton
Surface Effects Ship (2KSES) demonstrates the feasibility of

a fiber-optic data bus system using multimode fiber-optic
cables. It is decided that future data bus applications will
utilize a single mode fiber-optic cable in conjunction with
I0Cs. Multimode cables will not be used in operational aircraft
avionics systems. The decision is made in 1979 to utilize a
single mode data bus system in the VPX (replacement for the
P-3). The aircraft is to be built in the mid-1980's. In 1979,
monolithic integrated optical circuits have been perfected and
are available commercially. However, they won't be mass pro-
duced until the American Telephone and Telegraph Co. begins
Msesor Liber optics in 1983. Even though interest is high,
demand for fiber-optic components does not warrant full scale
industrial production. The fiber-optic cable producers can
keep up with demand with only a few production hours each day.
Total production is four million feet of cable per year in 1980.
By 1980, standardization of components has been completed.
Single mode cable connectors have been successfully demonstrated
for three years. By 1981, integrated optical circuits are off-
the-shelf items but supply is limited because they are not full
production items. However, their continuing successful use in
laboratory applications indicate that the real future of fiber

optics continues to point to integrated optical circuits

together with single mode cable as the desired goal of fiber
optics technology. The successful Air Force F-15 Demonstration
in 1981 has further convinced the Navy and Air Force to plan
future avionics systems around the single mode data bus con-
cept. The B-1l bomber demonstration in 1982 was a success. The
application of fiber optics helped reduce total weight by six
hundred pounds yet provided ample EMP/EMI protection. The
Army starts replacing five million feet of tactical communica-
tion line in 1982. Army plans call for a replacement of 16
million feet of 26 pair coax cable by 1985. This will be
followed by the replacement of 25-50 million feet of permanent
long distance communication lines by 1990. In 1982, demand for
low-loss single mode cable by the Army accounts for almost 50
percent of the total U.S. demand (4 million feet per year).

The Army's share of the user market will dwindle to only a

few percent per year after AT&T starts its replacement program.
By 1983, AT&T is using ten million feet of single mode cable
each year. 1983 is considered the base production year with

a total production of 8 million feet per year. Russia, Japan
and European countries are also active in the fiber optics
market. Empirical data can now be gathered to verify growth
rates of approximately 50 percent per year. In 1984, the VPX
aircraft is built. It utilizes single mode fiber optics. In

1985, AT&T has the only data link system capable of handling

the high data rates required for a data bused "wired city"
concept. A few strategically placed buses in a city are
capable of transmitting data at the gigabit level. Other
telecommunication companies prepare to follow AT&T's lead.
By 1986 they are installing their own fiber-optic lines.
After the additional telecommunications companies enter the
market, production growth rates steady at 50 percent per year
as total production is now 20 million feet of fiber-optic cable
per year in 19386. Total production is 45 million feet of cable
per year in 1988, 100 million feet per year in 1990. Newly
constructed military aircraft avionics systems all use the
fiber-optic data bus concept by 1989. New ships also use
fiber-optic data bus. Total military usage is approximately
five million feet per year, or only a fraction of the total
used by the major communications companies.

c. Scenario III - a Modestly Pessimistic Context

Flight testing of the A-/7 ALOFT Demonstration air-

craft has neither proved nor disproved any of the claims of
hopeful proponents of fiber-optic systems. Results from the
Delphi questionnaire were somewhat late in being received.
This, coupled with a less than satisfactory data collection
for the life cycle cost model, has delayed the proposed
economic analysis for a period of several months. The decision

points on whether or not to use fiber optics in the avionics

packages of the proposed LAMPS helicopter, VPX (P-3 replace-
ment), TASES (ES3's), and VPX (F-14 fighter follow-on), pass
without a conclusive economic analysis from the ALOFT Demon-
stration. The decision is made to continue to use known
reliable coaxial systems in the above mentioned aircraft. One
reason given by military planners is that the SALT II Agree-
ments with the Russians, among other things, were so successful
that the needs of EMP/EMI immunity are no longer a driving
= It has also been argued very successfully by newly
formed coaxial cable manufacturers! lobby groups in Washington
that the coax/twisted pair data bus system of the F-15 has
functioned perfectly well for years. '"'Besides,'’ they argue,
"think of the thousands of productive workers who will be
thrown out of work if coax is no longer used."’ Military
planners decide to use protective shielding for EMP/EMI
immunity if and when the need arises. In 1978, because of
constant pressure from Congress to "cut the fat'’ out of the

veut

military budget, R&D funds for fiber optics research are
to the bone."' The planned 2KSES shipboard application is
cancelled. It is clear that the 1980-1987 generation of
military aircraft avionics systems will not utilize fiber
Cmetles, lathe period 1976-1985, NEE@, thewAir Force Avionics

Laboratory, and the Army Electronics Command use their limited

R&D funds to their best advantage in continuing to demonstrate

oh,

Ge tCechmical feasibility of fiber optics applications as

data transfer links. All technological developments, such

as laser injection integrated optical circuits, single mode
connectors, T-couplers, and Star-couplers are slowly standard-
ized -- mainly as a result of the military's Tri-Service
effort in this area. A firm production base is not yet
established but, as components become standardized, and as
more component firms enter the market, prices to consumers
continue to drop. Between 1980 and 1982, production doubles
to two million feet of cable (mostly single mode) as more
potential users are beginning to follow the lead of American
Telephone and Telegraph Co. In 1983, with production at four
million feet of cable per year, AT&T begins to replace its
first million feet of aged long distance communication line.
1983 is considered to be the base year for production. More
telecommunication companies follow AT&T's lead in the late
1980's, mostly because the strategic aspect of copper avail-
ability has driven them to find a substitute cable. Copper
prices have more than tripled since 1975 (constant 1975 dollars).
Raw materials for fiberglass, on the other hand, are not
strategic in nature and are available at about the same relative
cost. The strategic nature of petroleum based products has
tripled the cost of fiber-optic protective cabling. Production

tetals eight million feet of fiber-optic cable in 1986. The

100

mid- to late-1980's see constant growth rates of 30 percent
per year. Total production is put at 22 million feet per

year in 1990. This period in the late 1980's sees the mili-
tary designing operational single mode data bus avionics
systems for aircraft to be built in the early 1990's. The
Army begins to replace the first segment (one million feet)

of tactical communication line in 1987. The Army will replace

all 16 million feet of its coax line by 1995.

The three scenarios presented are the authors' own percep-
tions of how the fiber-optic industry might develop. Some of
the information contained in the scenarios, however, was
obtained from a literature review and conversations with
military and industry contacts. The scenarios provide a
framework for hypothesizing how the fiber-optic industry might
evolve over time. It should be emphasized that many more
scenarios could be developed by varying relevant branch points
or events such as technological advances and military and
civilian demand requirements. To summarize some of the infor-
mation provided in the three scenarios, the following graph
(Figure IV-6) represents the annual production demand quantities
tGerocenarios -, IL and [Li for Liber-optic cabling by pro-

duction year.

soya

PRODUCTION OF FIBER-

ANNUAL

100

.90
50
©
40
=<
i“)
&
is
i 30
&
S
g
a“ A)
Lid
= 20 Os
: g
a Ny @)
: ec &/
Oo. => 2 VO
Oo & OMe
S S
me)
© @
(¢)
O

OXOTO

(y)
I976 77 78 79 80 B81 82 83 84 85 86 87 88
PRODUCTION YEAR

Figure IV-6 Fiber-optic cable demand, by scenarios

W———_6

sicia ) ©,

2. Constructing Cost-Effectiveness Curves from Scenarios
In addition to ordering and structuring uncertainties
existing in the fiber-optic industry, scenarios implicitly pro-
vide the information for cost and effectiveness levels required
for the comparative analysis. To illustrate how scenarios can
be used to construct cost-effectiveness curves, it 1s appro-
priate to summarize the characteristics of scenarios and
related assumptions. The following items are relevant:
(1) Each scenario is a list of assumptions
of how the fiber optics industry and
technology may evolve.
(2) The events listed in the scenarios are
ese in time.
(3) At any time in the scenarios the specified
levels of effectiveness, in most cases,
are possible. If an effectiveness level
cannot be achieved, then the cost-effective-
ness analysis cannot be accomplished and
infinite cost should be assigned to
eliminate the alternative.
(4) The cost of achieving a specified level
of effectiveness are scenario and time
dependent.
With these characteristics and assumptions it is possible to

construct the matrices shown in Figure IV-/.
103

SCENARIO A

EFFECTIVENESS

LEVEL E. E. E, E, E
YEAR - 3 5

ae oe oe
a. * A
i ie

STE)

198¢ a

EFFECTIVENESS

LEVEL E E E E E
YEAR

O76
1977
USTs

roy

1980
Figure IV-7 Scenario/effectiveness/time matrices

The elements of the matrices are the cost, oa where
i specifies the year and j the effectiveness level required.

The rows of either matrix trace out the cost-effectiveness

104

curve for a particular year at five effectiveness levels.
Possible cost-effectiveness curves for 1978 are constructed

for scenario A and B in Figure IV-8.

Es 4 v
5
Re
NS @
3 Sv S
4 S SS
a

Caz —Y c&
Res

EFFECTIVENESS (EEV EE

Ol
|
|
|
|
|

| 2 S 4 be)
COST (1978)

Figure IV-8 scenario related cost-effectiveness curves

To illustrate how scenarios can be used to construct
cost-effectiveness curves for the fiber-optic alternative,
Scenarios A and B are first defined. Assume scenario A depicts
rapid technological advances, standardization of components,
and high demand from the civilian community while scenario B
represents continued research and development outlays, no
civilian demand, and prototype components. If effectiveness
level three is desired in 1978, then the C93 cost element

should be computed for scenarios A and B utilizing the life

105

cycle cost model (Figure IV-5) together with the assumptions
of the respective scenarios. Since the scenarios contain
demand quantities and growth rates for fiber-optic components,
cost estimates can be made for these components by utilizing
the experience curve or other applicable techniques. Experi-
ence curves will be discussed in Section V-C. The component
cost estimates then can be applied to the life cycle cost
model to obtain the Cg3 cost element. Estimating costs for
the five effectiveness levels and both scenarios results in
the cost-effectiveness curves depicted in Figure IV-/.
Scenarios, therefore, provide the basis from which
cost estimates can be made. The selection of a particular
scenario, from the many possible scenarios, should be assessed
in terms of the likelihood of occurrence of particular
scenarios. As a possible method of minimizing or limiting
the number of scenarios and refining estimates contained in
scenarios, the authors suggest using the Delphi technique to

be discussed in Section V-B.

106

Vee ev OP COSTING TECHNIQUES BOR FIBER ORTICS

A. GENERAL

In this section, two techniques for costing fiber optics
are discussed. First, the Delphi technique is discussed as
a method of obtaining estimates identified in and required
for scenarios. These estimates include such items as demand
quantities, growth rates, and technological advances. The
second technique, the experience curve, is then explained.
It permits the estimation of costs to the government based
on cumulative quantity produced. This discussion is followed
by a demonstration showing how the estimates contained in
the scenarios can be applied to the experience curve to

predict the future price behavior of fiber-optic components.

fee lot DELPHE TECHNIQUE

ieciberscptics, data do not existto establish firm
Eechnelogical, price or demand trends. In this case, regres-
sion, sampling, smoothing or other mathematical analyses are
mot applicable as a basis for forecasts. Hence, predictions
must rely on the opinions of experts. Kahn makes the
observation that many books go into considerable detail on
the methodology of forecasting, particularly technological

forecasting. While the methods seem very impressive when

107

viewed in terms of how successful they are, their track
record is not as good as some of their proponents would
suggest. /24/ Kahn, however, expresses great interest in
the Delphi method which he thinks has great potential in
areas involving emerging technologies.

Kahn makes the statement that while Delphi is an excellent
method of technological forecasting, it works best when it
polls the experts who are actually attempting to achieve the
given result. Not only will they have some idea when the
innovation can be expected, but they could also have a large
influence on program outcomes.

Delphi, as a technological forecasting technique, is
generally credited to Olaf Helmer, T. J. Gordon, and N. C.
Dalkey of the RAND Corporation. Initial work was done by
Helmer as early as 1959. Helmer's publication of a "Report
on a Long-Range Forecasting Study" by the RAND Corporation,
in 1964, discussed the Delphi technique in detail. /20/

In his report, he describes his now well known method of
Selemrine forecasts from a panel of experts im order to
deal with specific questions, such as when will a new process
gain widespread acceptance or what new developments will take
place in a given field of study. Instead of the participants

@atherins together to discuss or debate the questions, they

108

are kept apart, usually answering assigned questionnaires
through written or other formal means, such as on-line
eomouters.

The advantages of using Delphi to poll the experts then
are: (1) The Delphi seeks to systematically codify the
opinions of experts while minimizing bias. (2) It polls
the experts who are actually attempting to achieve the given
results. (3) It eliminates typical problems of face-to-face
interactions among members of a panel. Many of these prob-
lems are psychological factors which tend to reduce the
value of methods based on face-to-face interaction (e.g.,
brainstorming sessions). /23/ Some of these psychological
factors are: unwillingness to back down from publicly
announced positions, personal antipathy to or excess respect
for the opinions of a particular individual, skill in verbal
debate, band wagon effects of majority opinion, and
Geasuasion, /5)/

There are also disadvantages to Delphi. Some of these
have been pointed out by Ayres and Cetron in their books
concerning technological forecasting. (/5/ /10/ These dis-
advantages include: (1) It is difficult to allow for the
bias of the pollster. For example, the framers of the

questions can to some degree guide the trend of the answers.

109

(2) Panel members dislike starting with a blank piece of
paper. They also dislike being involved in extensive itera-
tions and evaluating projections outside their areas of
expertise. (3) Extensive iterations can result in a heavy
investment in time and money to the researcher.

The authors have taken the above mentioned advantages and
disadvantages into consideration in deciding to recommend
Delphi as an appropriate technique to use in the costing of
fiber optics. Primary reasons for the Delphi selection are:
(1) Fiber optics is an emerging technology which is fraught
with not only technological uncertainties, but also total
demand uncertainty. In addition, there are component price
uncertainties. (2) The experts in the fiber optics industry
can be easily identified (see Appendix E). (3) Users and
producers alike would benefit from the results of a Delphi
study. It is to their mutual advantage to cooperate in
efforts to realize the potential benefits of this emerging
technology. (4) Improved forecasts or estimates of future
demand quantities, industry growth rates, technological
‘advances and component prices are expected to decrease the
range of the estimates for these variables. As a result,
the number of scenarios to be developed can be fewer since

the range of estimates is smaller. Based on the above

110

considerations, the authors recommend a Delphic approach.
Specifically, the authors make the following recommendations:
(1) The first iteration should provide a firm starting point
for participants by structuring events into sub-categories
of technology, demand quantities, growth rates, and component
prices. (2) The number of iterations should be limited to
two or three unless further refinement is required. (3) The
panelists should be required to self-weight themselves as to
their expertise in evaluating events in any given field or
sub-category of the questionnaire. These weights should
range from 1 (highly qualified) to 5 (not qualified).
(4) Construct the Delphi questionnaire format such that
there is an interweaving of timing, producer feasibility,
and user desirability. Timing is particularly important
because of the heavy time dependence of scenarios. The
feasibility and desirability aspects are particularly impor-
tant to the producers and users respectively as they relate
to their particular areas of expertise. (5) Panelists should
be informed of the study plan and its schedule prior to
involvement. They should also receive the final results of
the study.

The following Delphi questionnaire, as developed by the

authors, is meant to accomplish the above suggestions. It is

iL ileal

an example of how time, desirability and feasibility can be
interwoven to form a simple, yet comprehensive, approach to
establishing refined estimates for scenarios. The events
are representative of the types of questions which should be
included by one who conducts a Delphic study of an emerging
9

technology. Any revision of the suggested Delphi question-
naire, due to information not now available, and its

application would be an aspect of the economic analysis to

be performed by other NPS students and NELC personnel.

WIE

azteuuotjsenb tyudjeq etduesg J[-A 2ANsTY

* CHUMIVN

-NOIZSGNO ITHd1ad V dO YANVA AHL OL
LSAUALNI AO SNOILSHNO YWAHLYNA *°OLG) Gc {( )
sores (7 JUeAD UT se °°990) “SWwORT
J[eus-9yj-JJO se uotT}ZONporad ut o72e
SHOLOANNOD ‘,,2,, “ANQUL THNNVHO HIONIS) 47 |( )
en et Uo Notes 7049) oT qe
-[teAe F[IYS-aeyuA-JJO ete sxtO}ZDIUUOD
adkq ‘*o30 ‘,YVLS, THNNVHO-ILIAW| € |( )
OR Ae 3 Sea ee *(uotTqsenb aAo0ge
Ur Se.‘s099) “e71Ge[TeAe FT[SUS
-dU}-JJO 212 SAVATIIIA VY SADATIAP
JInNsAtTS [eorqdo pazerZaeqQuT OTYRTTOUOW] Z | C )
a *satjodououl on *s790e12
-U0) 10} UOTITJEdUIOD UT Sie SoTU
-edwods om}? ueuq eAOW “sweat JTeys
3UI-JJO Se peteptsuos pue uot Jonpoad
uT o1e ‘°90]90 ‘sapotp odAQ NId ¥ AAT
SULZT[I4N ‘StaATIIeA/SABATAp ATe[Npow| T |] C )
a MOCAOMMOciIatoo | LdOy dda lass —- =
“CAZ1iS LNASddd YlaH
0002} CO66l| S86T 7/T OL GAZIYMNIVINIW FUV LAVEOYIV AuavVL
-ITIW NI SUPLNdWOO SOINOIAV ‘A1dNVXa)}| V | (€)

pasndo0 aADY {IM
juaAd pawou aut jou) Xn Wa

oJ st Ayyigoqoid \? 5
©
aut 1893 NOA YDIUM\YS

Aq 409K ay} haus

Buiw, aqoqosd

(suotjdumsse 3utpnyout) +f
SINHAd OLSVd

da LALTVnO

AITUOOd ‘(S) OL Gala
-I1vnNdD ATHDIH °(T)
LSAYALNI AO

SVauV NI YHMSNV OL
SNOLLVOLALITVNO

NO DNILVY- alas
INVdLOILUVd

Ayiigisoay 49aDnpos4 Ayipiqosisap sasn

Jae:

Pa4Nd50 BAD IIIM
JUSAd PoWOU JU} OU}
oJ S!' Ajyiqoqoid
ay} [aay NOA YdIym
Aq 409A ayy sajuy

Buri} ajqoqosq

Col SO ALGVo
AGOW FIONIS ‘SYOLOANNOD ‘SYAALAOAY
/SYAAING AINGOWN ‘:SHIdNVXE °LSOO

ONINYAQNOD SNOLLSAND WHHLO °° OLA)

Poe e eee eee ee eee eereees F9N909 °F
"°(S,5 GL6T°}#UeISUOD) 4F/907-OT “A
°° (8,5 GL61 FUeRSUOD) IF/90E-02 °9
°° (S,$ GL6T JUeISUOD) I3F/907-OF “A
°° (8,5 GL6T Jue ISUOD) 3Z/90S-07 “V
("oa “sio
-injoeynueuw Sutjeduos [erzsAes ‘apow
uotjzonpord [[nzZ sumssy) :3Sso0d [TTA
S307 3003 000 O01 UE peseyoand °(°3990
‘stsoyq sty jo q xtpuaeddy ut pajze
-OUTTep se) squeuerztTnbexr aouewizo0jized
a[qeo Joow oF poustsep ‘ (wy/gPodsc>)
SSOT-MOT SaTqed oEAdO-A9qTF BspowT ITH,
Sree a MiSteOrotda SOlLdO dddid------

INGA

)

( )

114

juaAd pawou ay} youN
Jd Ss! Ajyiqoqoid
yt |9a} NOA YoIUM
Aq 402K ayy 194uz

, peatsep st

eo @@e@08@ @ @

eoe7#e#ee#eeewte?

eoeoeee@#ee#e#@e#ee # @

eoeeewee#@ee 8# #@#

cee (190490) a
eee) Lk
" ZHW COS < O
° ZHW OOT < ‘ad
°° ZHN OC < OV

YIpTmMpueg s<1OW Yyonw MoH] ZT

029

soqel Aazysuelt

e]ep poseotour A0F AAeSSIao9U 92q
ITEM (WH OO€ 3% ZHW OZ *° 98°F) 24TH
ated pejzstm3/xeoos Aq peptaoid mou
2eYy FO ssaeoxe UT YAPTMpueq TeUsTS) TT

Coma) Ot

"3TFOAQAA e JOU -- swaqshs pzeOg

-~dtus 240} poustssg

“sua SAS YUTT

elep UT suwaqT paepueqjs se sotqdo

zoqty Ao[Tdwa [tM sdtys AreATTT] 6
*swoqSAS YUTT eIep
SOTUOTAG UT SWOT prepueqs se sotqdo
zaqty Ao,Tdwe [{[IM JFyezAoATe AACITTIW Q

eee e @

Teuotjerz9dao

LNA rm

oN TN

( )

115

SPOULITMuMTO UOTIJonpord 196° [e204

eeeee #

e«eee8 @®8@6m6UhOrmhCl OrmhUchOOrmhUCUhOHmhlUm AIUIO "ad

TSIt O00 000 0OL =D
° 22°F 000‘O000'OI “A
eel oO Om OO alae ay

>8q TIT4 eTqeo (wy/_Pocs
-OOT) SSOT.unTpew (szeqTF 00Z<)

oe @ @e@ @ &@ 8 @ ee emhUhUh rh A940 °d

meet OOO OCOOMO0L 2
Bee loot OOO COO OL “a
moe OO OOO O00 La V¥

:8q [ITM eTqeo

oT3do ASqTF Spow aTSsuts (wWy4/gpos>)

SSOT-MOT FO uotTRZONpoOAd °*sS’yn TeIO]
QNVWHG) NOLZLONGOdd SOLLdO wddld---

INWAA

Sul

C )

Oe,

116

C. EXPERIENCE CURVE

This section will develop experience curve theory and
explain how it can be used as a forecasting technique to
help predict the cost behavior of products such as fiber-
optic cables, drivers (LEDs) and receivers. Experience curve
theory should not be confused with the well-known learning
curve theory. Learning curve theory predicts cost reductions
for two cost elements, labor and production inputs (materials),
whereas experience curve theory predicts cost reductions for
all cost elements including labor, development, overhead,
capital, marketing, and administration. Experience curve
theory is much broader a concept that incorporates learning
curve theory. To facilitate the development of experience
curve theory, the subsequent discussion will explain both
theories noting similarities, differences, and the factors
which explain both theories.

Both the experience curve and learning curve theories are
expressed as cost quantity relationships stating that each
time the total quantity of items produced doubles, the cost
per item is reduced to a constant percentage of its previous
cost. For example, if the cost of producing the 200th unit
of an item is 80 percent of the cost of producing the 100th

item, and if the cost of the 400th unit is 80 percent of the

ey,

200th item, and so forth, the production process is said to
follow an 80 percent unit experience or learning curve. Figure
V-2 shows a unit curve for which the reduction in cost is 20
percent (i.e., 80 percent of the original cost) with each
doubling of cumulative output. The arithmetic plot illustrates
that the reduction in cost for each unit is very pronounced
for early units. For example, on the 80 percent curve, cost
decreases to 28 percent of the original value (100) over the
fest OO units. Over the next 50 units, it declines only 5
more percentage points to 23 percent of the first unit cost.
A plot of the same relationship on a log-log scale, as shown
in Figure V-3, makes the relationship linear and reflects the
constant rate of reduction. Log-log plots are used almost
exclusively because the straight-line relationship is easier
to construct and use for predictive purposes.

The mathematical relationship between cost and quantity
for experience curves and learning curves is represented by
the power equation:

a Equation (1)

Cc, = Cyn
where: C, : is the cost of the first unit
C F iS the cose Of Ener nenewnmine
n : is the accumulated units produced (experience)

A : is the rate at which cost declines with
experience (slope of the experience curve).

118

oO oO
© vt

L803 LINA

O
O

400 600 800 1000”

200

CUMULATIVE UNITS

Figure V-2

1000

ene pw =

i) i ae
ee Blak
: i | .
we dfe | :
1, || | a :
; ec
:
= = Sao
ie a Nee : 5
re) om :
- ae ’
E a Pe
‘a wes :
E mae 7) a
w at :
Eula PY
G . Jee go"
; = <<
> Va
a eS
i aq 3h
OU | Lio : BP
‘aaa 5 a
ant =
oD | ;
a
w)
a i
~
; wv
of
oO
oO
os
: oo)
‘ Le
i
:
KH
O °
o

LSO9 LINN

119

The slope of the experience curve, x bears a simple
relationship to the constant percentage to which cost is
reduced as the quantity is doubled. By letting S represent
the fraction to which cost decreases when quantity doubles,

and using Equation (1), then:

s = ©2n & C,(2n)~* Os A
(ee Cc, Ge
_A
See Equation (2)
A= - 1085
em Equation (3)

For example, an experience curve with a slope of \= 0.415
has a constant percentage reduction in cost to 75 percent of
its previous cost each time accumulated quantity doubles. In
order to avoid confusion, subsequent use of the term "slope"
will refer to the constant percentage reduction.

The history of learning curve theory dates back to 1925
when, in the aircraft industry, learning patterns were first
observed by the Commander of Wright-Patterson Air Force Base.
The phenomonon observed was the constant reduction in direct
labor hours required to build airplanes as the number of air-
craft built doubled. /21/ Subsequently, learning curve theory

has been documented and used in many industries to predict

120

cost reductions for direct labor and raw material -- or

production inputs. Typical learning curve slopes have

ranged from 75 to 90 percent. Some of the factors commonly

mentioned that account for direct labor and material cost

reductions are summarized as follows:

(1) Job familiarization by workmen. This results from

(2)

(3)

(4)
(5)

(6)

(7)
(8)

(9)

the repetition of manufacturing operations.

General improvement in tool coordination, shop
organization, and engineering liaison.

Development of more efficiently produced sub-
assemblies.

Development of more efficient tools.

Substitution of cast or forged components for
machined components.

Development of more efficient parts-supply systems.
Improvement in overall management.

Workmen learn to process the raw materials more
efficiently, thereby cutting down spoilage and
reducing the rejection rate.

Management learms to order materials from suppliers
in shapes and sizes that reduce the amount of scrap
that must be shaved and cut to form the final

pPEOaGuICE.

yal

The above list of relevant factors is not considered
complete. It also tends to understate the importance of the
one item usually considered most important -- labor learning.
SO

Experience curve theory dates back to 1965. /9/7* Experi-
ence curve theory is much broader in scope than learning
curve theory. It considers the full range of costs which
include development, capital, administration, marketing,
overhead, as well as labor costs. Raw material cost is not
included in this list. The cost of raw materials usually
depends on factors such as availability of supply. For
example, the price of unprocessed timber fluctuates from
year to year partly as a result of federal policy concerning
the nation's timber reserves. Strictly speaking, correct
measurement of the experience effect therefore requires that
expenditures be calculated net of the cost of raw materials,
i.e., on value added to the product. In general, experience
curves do not apply if major elements of cost, or price, are
determined by patent monopolies, natural material supply, or

government regulation. The experience curves apply to products

* Experience curve theory is primarily credited to Mr. Bruce

Henderson, founder and President of Boston Consulting Group,

Inc., a management consulting firm specializing in developing
corporate strategy.

in industries with multiple producers who interact rivalously
as well as other products in purely and perfectly competitive
industries. Experience curves cost reductions on value added
range from 20 to 30 percent every time total product experi-
ence (accumulated quantity) doubles for an industry as a
Whole, as well as for individual producers. These reduc-
tions represent experience curve slopes of 70 to 80 percent.
Empirical data have been collected which verify these
experience curve slopes of 70 to 80 percent. Many of these
data collection efforts were for products in the chemical
and electronics industries. Reports of the Electronics
Industry Association, the Manufacturing Chemists' Associa-
tion, and the 1965 Statistical Supplement to the Survey of
Current Business by the United States Department of Commerce,
among others, were used in gathering these data, as were
Boston Consulting Group sources within the relevant industries.
Figure V-4 on integrated circuits and Figure V-5 on polyvinyl-
chloride are two examples illustrating the experience curve
Saucer withthe characteristic Cost reductions. To permit
comparability over time, prices were expressed in constant
1958 dollars.

Price and experience (accumulated quantity) follow one
of two characteristic patterns: stable, as shown in Figure V-6,

or unstable, as shown in Figure V-/.

o4
average
ie eae

$40.00.

(qunjsuod $) LINN Y3d 3dIYd

=»

-
‘
oe od .
: e
Ao - &
i
ce ae
ae pee
i
og ob ge )
Os ec
trdee -
‘
‘ ee . /..-
° a5
oe

Ss
e

INDUSTRY TOTAL ACCUMULATED VOLUME

(million units)

Figure V-4

rve

ts experience cu

ireul

Integrated c

( juoisuoo g) GNNOd Yad Jd

'
Rusts
{
{

t
!

|
> t ; } - e ee
a en i ee +

nee,
7

+
pe a —e
. ‘
. FY
: ? { : :
. a 2
ar ete i
‘
t
)
i
¢
4
ae of ne a aera ae
t
4

a og
} .
| .
_— a

40
Oyle)

VOLUME
Source

INDUSTRY TOTAL ACCUMULATED

(million pounds)

, Be

Boston Consulting Group

Figure V-5
Polyvinylchloride experience curve

124

e +
1 . ' 5 ’ ¢
— Sa le bool re Sa oc —-— - tae =~} oe

a a

—- =

BW a wane Wrens oe Ny aerate Osea oe es [~~

ew ee

rf
i
*
- ~© os. - settee o-
ry
tn ee ee Seed co
ry
i . mat
° 1 -— ee
‘
‘

L509 GNV 35diud LINN

TOTAL ACCUMULATED VOLUME

-6

Figure
A characteristic stable pattern

TOTAL ACCUMULATED VOLUME

Figure V-/
A typical unstable pattern

125

A stable pattern exists when the cost and price of a
product maintains a constant quantitative difference over
time. In Figure V-6, price and cost parallel each other
over time thus indicating a stable pattern. Products follow-
ing this type pattern tend to be found in technological
industries which are experiencing rapid growth as well as
being very competitive. Integrated circuits is an example
of such a product. Its straight-line trend relationship is
illustrated in Figure V-4.

When prices do not decline as rapidly as cost, an un-
stable pattern, as shown in Figure V-7, will exist. Prices
are set below cost to establish an initial market. As
volume and experience reduce cost, the prices are maintained,
gradually converting the negative margin to a positive one.
If prices do not decline as fast as costs, then competitors
are attracted to enter the market. At some point, prices do
start to decline faster than costs. The experience curve for
Polyvinylchloride, as shown in Figure V-5, illustrates the
point. Obviously, prices cannot decline faster than costs
indefinitely. At some point, a reverse bend in the price
curve reestablishes a stable relationship between cost and
price. Figure V-8 illustrates an unstable pattern transform-

ing to a stable pattern in different phases.

126

UNIT PRICE AND COST

TOTAL ACCUMULATED VOLUME

a Figure V-8
A characteristic unstable pattern
after it has become stable

In Phase A, costs typically exceed prices. This is
always the case in the very early production stages of a
new product. It covers an extensive period if the future
potential is obvious and competition appears severe in the
very early life of the product.

In Phase B, the market leader is effectively holding a
price umbrella over higher cost producers who are entering
the market and increasing their market share. In effect,

the dominant producer is trading future market share for

eilecent profits.

Ley

Phase C is a shakeout period. This phenomenon is caused
when a producer thinks that his own interests will be better
served by lowering the price faster than industry costs are
declining. The typical slope of the experience curve during
this phase is about 60 percent during the period in which
industry experience doubles. This, in fact, does not
occur unless the cost-price relationship is unstable. An
unstable market is characterized by rapid growth, a large
number of producers, and a large difference between price
and cost for the lowest cost producer. The motivating
factor for the lowest cost producer to lower his price is
to increase his market share. High cost producers must then
either accept lower profit margins or drop out of the
industry.

At the end of the shakeout phase, the stability of the
relationship of cost to price is fully established and
Prase BD, i.e., stability, emerges.

The factors, identified by the Boston Consulting Group,
that cause the experience curve effect include:

(1) The "learning curve effect"

(2) Competition (rivalry) among producers in a
given product market

(3) Economies of scale and specialization; the

Mecale eEneEcE

(4) Investment in capital to reduce cost and
increase productivity.

The learning effect, people learning by doing, has already
been discussed in learning curve theory and is the major fac-
tor which causes reductions in labor costs. The second
factor, competition (rivalry) among producers, forces each
producer to find means of lowering his total average costs
in relation to his competitors. The successful low-cost
producer will then be able to lower his prices and induce
a situation which causes a "shakeout" of those producers
who have been unsuccessful in reducing costs. This will
give the low cost producer an increased market share. With
increased market share, the third factor, economies of
scale, can be realized. With scaled-up volume due to in-
creased market share, it is possible to use more efficient
tools and spread their cost over enough units so that both
labor and overhead costs are reduced. Increased volume
may also make it possible to consider alternative materials
and alternative methods of manufacture and distribution
which are uneconomic on a small scale. The final factor,
investment in capital, is a further attempt to reduce cost
by displacement of less efficient factors of production.
This can be accomplished by automating various stages of

production thus reducing labor costs. This may not be

possible or desirable if the market share is not sufficient
to warrant the investment.

To use the experience curve as a predictive tool the
following elements are required:

(1) Che fbeoem teria (price

(2) Initial experience (accumulated quantity),
represented by Cy

(3) The slope of the experience curve.

With these three elements it is an easy matter to con-
struct an experience curve on a log-log plot. As a hypo-
thetical example for low-loss fiber-optic cable, an initial
price of $4.00 per foot with an initial experience of
100,000 feet is assumed. Experience curve slopes of 70, 80,
and 90 percent are plotted to illustrate a range of cost
reductions possible. Figure V-9 illustrates the three
experience curves with their different slopes. The initial
point (100, $4.) is common to all three experience curves.
A second point for the /70 percent curve is obtained by
multiplying (.70) (4.00) = $2.80. This $2.80 figure is for
the doubled — of 200. Hence the second point (200,
$2.80) is obtained and a straight line is constructed to
complete the curve. The 80 and 90 percent curves are

constructed in like manner.

130

These curves illustrate how prices might decrease as a
function of accumulated quantity but they do not indicate
when these price/quantity relationships will occur. Time
frames can be established if the rate of growth of the
accumulated quantity is known. Use of the standard formula
for annual compound interest is applicable. The formula is:

A= (100 (1 + i> Equation (4)
Where: i : is the annual interest rate

T : is the time in years $1.00 has been invested

A: amount accumulated after T years.
Changing Equation (4) to multiples of accumulated quantity
produced and using growth rate in place of interest rate
yields the new formula:

mA =A (1 + g)yt Equation (5)
where: m : is the desired multiple of any accumulated

quantity produced

A : accumulated quantity produced

g : annual growth rate of the product

T : years required to attain the desired multiple.
Solving Equation (5) for T provides the desired result:

log m

re ee Equation (6)
Hoe (lc) :

For example, the time to double accumulated quantity (m = 2)

having a growth rate of 40 percent per year (g = .4) is

1978

10.00

198]

wre

I980

Igig
|

1
*
eae
‘
‘

. ‘
— =) ee

a :

Oo
( suoisuod $) 1004 Yad 3Okd

- 4
eee — *.. -
° ?
«oc- ---}-- ~ Sa (he
} . v
sees. 6. = - . .
)
. o-

Z
{
'
;
i
\
}
!

-J a ame

Sai

es =| ie

—a ee eye Bo

0.0!

100,000

000

10
ACCUMULATED QUANTITY

1000

300

100

(in thousands of feet )

Figure V-9
Hypothetical example of a fiber-optic

cable experience curve

132

approximately two years. Figure V-9 indicates (hypothetical)
a ten-fold increase (m = 10) in accumulated quantity each
one year interval (T = 1). By solving Equation (6) for
growth rate an annual growth rate of 900 percent is obtained.

Experience curve theory then offers a means of fore-
casting price reductions for well-defined (standardized)
products. Again, to use this technique, estimates of first
unit cost and a production base (initial accumulated quantity)
associated with the first unit cost are required. If time
frames are desired, estimates of growth rate must also be
provided.

Since most of this required information is either non-
existant, or available only on a prototype development basis,
the authors have suggested constructing scenarios of the
fiber-optic industry's alternative futures. The example
scenarios in Section IV were developed in terms of fiber-
optic component evolution and standardization, military and
civilian demand requirements, and possible growth rates that
might occur. Scenarios thus provide the information required
to use the experience curve technique to predict cost of fiber-
optic system components such as cable, drivers (LEDs) and
receivers. These component cost estimates could then be

used as inputs to the life cycle cost model.

-
;

— a — aia «
a” a Ps ee

VI. SUMMARY AND CONCLUSIONS

This thesis contains the results of the initial cost-
effectiveness investigation of the fiber-optic alternative
for an avionics data link system. The study was intended
as an initial approach toward the desired objective of
numerical estimation of fiber optics avionics data link life
cycle costs.

The historical and technological background of fiber
optics as well as the background of the A-/7 ALOFT Demonstra-
tion was discussed. A general discussion of a cost-
effectiveness analysis was presented together with possible
measures of effectiveness for data link systems.

Scenario-writing was discussed as a means of ordering
the uncertainties of this emerging technology. Sample
scenarios were developed by the authors to provide specific
time-related estimates as to civilian/military demand,
growth rates, standardization and technological development
in fiber optics. These representative scenarios are meant
to be examples of scenarios which can be established as a
base for making cost estimates.

Two specific forecasting techniques, Delphi and experience

curves were discussed as relevant to the costing of this

134

emerging fiber optics technology. A Delphi questionnaire
1s proposed as a means of soliciting forecasts from a panel
of experts in order to deal with the specific uncertainties
associated with fiber optics. Experience curves were
suggested as a means of predicting the cost behavior of
products such as fiber-optic components.

It is the basic conclusion of the authors that: (1)
These techniques, scenario-writing, Delphi and experience
curves, can be combined as a cost-predictive method to
estimate component prices in an emerging technology such as
fiber optics. (2) Meaningful component cost predictions
can then provide a means of estimating reliable costs for
the life cycle cost model elements used in a cost-
effectiveness study. (3) At the present time, the uncertain-
ties associated with future cost estimates of fiber-optic
components, uncertainties of demand and production, and
lack of standardization will require careful analytical work
if reasonably accurate life cycle cost estimates are to
result. (4) The emerging fiber optics technology deserves
full and continuing effort and attention by R&D agencies.
Even if the results of initial cost-effectiveness studies

are such that the decision is made to not use fiber optics

in next generation aircraft, it would be a mistake to cut
back or reduce fiber optics R&D funding. Future military
communication and data link systems may well be the bene-

ficiaries of today's development efforts.

136

APPENDIX A

A-7 Navigation Weapons Delivery System (NWDS) Schematics

SoeJAOWUT [VOTAIET® SGM/N L-V “T O4NsTA

' 9$ WIA

Y3au"V LIdMDO9

dOMN

dyV

10V

dOV

|
i
|

1 YOLO3BNNOD

| QV3HN ING

Q3Z1IYNSS3Aud

Bw3aLNdIWOOD
WOLOVL
fOL
LG-NSV/CS6'dD
SOMN

' AWN
Y¥3 1ddO0O0

NOsSv

s(SOINOUT O43)

OIWNd

(SOINOU1I373)

QNH

137

‘uoT}eAnStyuoo weyshs TAOTV /-V “7 eANstTy

| om |
i]
[owe | 3
‘ 43 1dd0Q ‘
|

sa18v9 HaLdVOV OSIM —- aman eae
| §A18V9 01dO H3dIg ee ee

SAV 134
SIW

SAVI13H
Siev

S/d H1ldOV
NWI

YAaldVOV y3ldVOV

| =
Lo
» |

| walavav

O32 | O03 O03
VayV HY nosv a " 7 - WaHV HH
AVS
AVS ;
SOINOIAV HY 0/42
SOINOIAW H7 SSS SS SS ee
ina eee ee —= ene ene —_—asew ee ew ew ewe —Olse IOAN
ny ate, se Ba eee ey, , Of & ‘GQ314!1G0W
| im 'O/a
— —— = mae —— -—
‘SHONOYHLG335 O31V4aS ~~, they , LNINWLYVdiNOD XO7T
; “acs "INSWLUVANOD LN3SWdINO| Glos LAV
3 a INSWLYVdWOD LNIWdINOS-GIW

YjldVOV
O3
yd

-—-- HI

SLOANNOODSIGC O/3 QVAHNING S A

EER eee ee ee 9! PF

LNAWdINOS OI GMs HY

YICOINS
INVYOVNOD
AILLOYHL

1aNvd
— NOLLAWS
/ANOSIAGYV

INd
TOVYULNOD
dNd1V1N8

VANVd Nd “INd
oa iaS ALVNOISAG Vay
LNAWVWYV Pau 1 Wu

|
'
A2044 A2043 A389
FLR COCKPIT AREA RM AVIONICS
E/JO ADAPTER | 2% E/O ADAPTER E/O ADAPTER
>a
eae
ta I LW Le)
LJ
alts ‘ G) pet | 2
LH MID-EQUIP Sie oo
LH AVIONICS - Oo 0 omy.
P3336 Sea Ce ae
A390 SEAT BULKHEAD Doon w wn
Bee ms © ©
S10) at) F13067-144-F/0 FEEDTHRU = es Se & KIS
; iS ~~ Ee KR ORK o. 10
(2) — F T3067-15A-F/O 565 6 8 8 2s
MODIFIED Pace 2 © O,<
cm Lu
NWODOC uw re eto Ol-z
(1)— FT3067-16A-F/O te

N76 oho P2176 |

t
, 1

Lat ad pelo 37 go ales

| J2178 == ep
ET3067-18A-F/O 2178

—FT3067-19A-F/O
FT3067-20A-F/O

FT3067-23A-F/O
FT3067-24A-F/O
FT3067-25A-F/C
FT3067-26A-F/O
FT3067-27A-F/O
FT3067-28A-F/O

y
2

=a oo ieee ED

LH AVIONICS

LH AVIONICS
E/O ADAPTER
A388

—_ a ame ow

E/O ADAPTER
A387

EF reer OPTIC CONNECTORS
FIBER OPTIC CABLE

Prem n= (ALOFT “tiben) Opeie imuerrace Contileuratlom:

Woe,

APPENDIX B

A-7 ALOFT Component Requirements

A. ORIGINAL SYSTEM
(PORTION CONTAINED IN ALOFT CONFIGURATION)

Tota!
Power
Con-
Unit Cost Unit Total sump-
No Required (FY7S) Wt Wt tion
Type of Components Required Footage (S) (gm) (Ib) (W)

. Single Cables/Wiring |

These data to be derived and provided at a later date by LTV and IBM
. Signal Connectors through analysis cf the quantities, cost, weights, and characteristics of
a. Single-channel bulkhead the components that are used in the existing A-7 N/WDS design to

Originally transmit the 115 unmultiplexed ALOFT signals.
b. Single-channel pressure bulkhead

c. Multichannel bulkhead or rack/panel | |

d. Multichannel pressure bulkhead

e. Access couplers

. Signal Drivers
a. LEDs

b. Drive circuitry

. Signal Receivers
a. Photodetectors

b. Amplifier circuitry

Source: NELC TD 435

140

B. TWISTED PAIR — AFTER MULTIPLEXING

Total
Power
Con-
Perform No Required ,| Unit Cost Unit | Total sump-
Type of Components Reg or (footage) (FY75) Wt Wt tion
Required Part No (S$) ($) (gm) (1b) (W)
1. Single Cables/Wiring
2. Signal Connectors
a. Single-channel Determined to be infeasible with presently available twisted-pair components at 10-Mbit data
bulkhead rates and within MIL-E-5406P Class II Aircraft Environment.
b. Single-channel
pressure bulkhead
c. Multichannel
bulkhead or
‘ rack/panel
d. Multichannel
pressure bulkhead
e. Access couplers
3. Signal Drivers
a. LEDs
b. Drive circuitry
4. Signal Receivers
a. Photodetectors
b. Amplifier circuitry |

141

C. COAX — AFTER MULTIPLEXING

Total
Power
Con-
Perform Unit Cost | Total sump-
Type of Components Req or No Required (FY75) Wt tion
Required Part No (qty/footage) ($) (Ib) (W)
1. Single Cables/Wiring .| RG-316 13 (225 ft) 0.30/ft 66.08 WO127 e275 --
ft
2. Signal Connectors
a. Single-channel Sealectro
bulkhead 50-622-9188-31 36 Brea a? 3.371g | 0.21 Ib -
ea
$0-645-4576-31 26 eed ed 85.02 25202717021 16
ea ,
b. Single-channel Sealectro
pressure bulkhead | 50-675-7000-3 | 5 8.34 ea 41.70 4.342 {| 0.05 lb —
ea
c. Multichannel — 0 No multichannel coax connectors considered _
bulkhead or feasible for this application. Single channel
tack/panel utilized instead with printed circuit board
d. Multichannel connectors.
pressure bulkhead 0
. Access couplers — 0 ~ - = - -
. Printed circuit- Sealectro
board 50-65 1-0000 26 4.46 ea 115.96 2.320g | 0.133 Ib -
ea
3. Signal Drivers SN54S 140 Is 6.26 ea olase 0.004 lb | 0.286
a. LEDs — = =e = om = =
b. Drive circuitry — - =_ = = = ae
4. Signal Receivers SN548 132 13 IBi2Sea 172.64 ea 0.585

a. Photodetectors = ae = a = = =

b. Amplifier circuitry — = = = = = =

142

D. FIBER OPTICS — AFTER MULTIPLEXING

Perform
Req or
Spec No
Part No

Type of Components
Required

1. Single Cables/Wiring
See NELC Per-

2. Signal Connectors formance Re-
a. Single-channel quirements
ealkhead Sheets in

appendix A for
b. Single-channel description of
pressure bulkhead required
c. Multichannel components
bulkhead or
rack/panel

d. Multichannel
pressure bulkhead

e. Access couplers

3. Signal Drivers

a. LEDs

b. Drive circuitry

4. Signal Receivers

a. Photodetectors

b. Amplifier circuitry

Total

Unit Cost Cost Total
No Required (FY75) (FY75) Wt
(qty/footage) (S$) ($) (Ib)

13 (225 ft) 2.50/ft

6.94/ft

2.50 ea 13.568

€a

22.192
ea

255.00g
ea

0.389 Ib

3.50 ea

500.00 ea

13 (12 digital &
analog
a. 14*
b. 12 digital
1 analog

* x

80.00 ea

2.50 ea
32.00 ea

1120.00

30.002
32.008

0.085 Ib
WHO2 7a

13 (12 digital
1 analog)
a. 14°

b. 12 digital
1 analog

0.463 Ib
0.036 Ib

Total
Power
Con-
sump-
t10n
(WW)

*The one analog link in the ALOFT design requires two LEDs: one direct signal transmission and one for feedback for
linear compensation. Therefore, the transmission over 13 data channels requiies 14 LEDs and 13 photodetectors.
**The weight figures for driver circuitry and amplifier circuitry include the weights of the LED and photodetector,

respectively.

143

APPENDIX C

A-7 ALOFT Component Descriptions

Description of Components Required as Building Blocks for a Point-to-Point Informa-
tion Transfer System of 115 digital signals (A-7 ALOFT baseline):

A. Coax Interface System Components Requirements (assuming digital transmission only
over 13 data links — no analog). Prices in quantities (F Y75 prices).

1.0 CABLE. Type RG316
Spec: (Description) 502, 0.102” OD, 29.4 pF/ft, loss = 3.8 dB/100 ft @ 10 MHz,
temp =-55 to +200°C ;
Requirement: 225 feet
Price: $293.70 per 1000 feet; $66.08 total system cost
Weight: 0.012 Ib/ft; 2.7 Ib total system

2.0 CONNECTORS

2.1.1 Terminal Connectors. Type Sealectro 50-622-9188-31. All connector prices subject to
10-percent gold surcharge

Spec: Crimp type coax connector — straight plugs for RG316. All connector specs
MIL-C-39012 SMA

Requirement: 2 ea for each cable link (13) 26
2 ea for each bulkhead connector = 10
Total 36each
Price: $3.27 each; $117.72 total system cost
mctent: ~—3.3/ 19 each

2.1.2 Bulkhead Receptacles. Type Sealectro 50-645-4576-3 1
Spec: SMA receptacle MIL-C-39012

Requirement: 26 each (on each adapter unit, 2 for each cable)

2.2 Pressure Bulkhead Connectors. Type Sealectro 50-675-7000-3 1
Spec: M!IL-C-39012 SMA Source: NELC TD-435

Requirement: 5 each
Price: $8.34 each/S41.70 total
Weight: ~4.340g each

2.3 No multichannel connectors. 144

2.4 PC Card to Coax Connector. Type Sealectro 50-65 1-0000
Spec: MIL-C-39012 SMA
Requirement: 26 each
Price: $4.46 each/$115.96 total
Weight: ~2.320 grams

3.0 LINE DRIVERS. Type SN 548140
Spee: Dual line drivers, 509, Schottky for operation at 10 MHz

Requirement: 13 each (assume only one gate used per IC)
Price: $6.26 each/$81.38 total

Power: 22 mW ea/gate; 0.286W total

Weight: ~ 1.973 grains each

orev RECEIVERS. Type SN 545152
Spec: Quad Schmitt trigger

Requirement: 13 each

Ree: owls. 28seach/S172.64 total
Power: 45 mW ea gate; 0.585 W total
Weil, ~ ].973%erams each

B. Twisted-Pair Interface System Components Requirements

Conclusion reached after searching for qualified components that twisted-pair
interface not possible. Components for 1Q-megabit data rate did not readily exist. RG-108
could have been used if constraint of MIL-E-5400P Class 2 environment had not been a
requirement. RG-108 is only good for -40 to +80°C temperature range which is below
Class 2.

C. Fiber Optic Interface System Components Requirements

1.0 FIBER OPTIC CABLE PERFORMANCE REQUIREMENTS
1.1 Number of fibers: 367 — | percent (4)

1.2 Number of broken fibers: four if unterminated; seven if terminated
1.3 Fiber diameter: 0.00215 inch

1.4 Core glass area to total fiber area ratio: 285 percent

1.5 Numerical aperture: between 0.54 and 0.67

1.6 Maximum optical attenuation: 400 dB/km

1.7 Cable jacket and shield to be nonmetallic

1.8 Termination diameter: if terminated, active area diameter to be 0.0455 inch

145

1.9 Termination loss: without lenses or refraction matching, throughput loss to be <2.0 dB

1.10 Environmental range: temperature, temperature shock, vibration, mechanical shock,
and altitude capabilities to conform to MIL-E-5400P Class 2

{1.11 Mechanical requirements: impact, bending, and twisting to conform to MIL-C-13777F
1.12 Tensile strength: 35 Ib

2.0 SIGNAL CONNECTOR PERFORMANCE REQUIREMENTS

2.1 Fiber Optic Cable: Fiber optic cable used with connectors to be as required in perform-
ance requirements for fiber optic cable

2.2 Termination Diameter: Connector termination for fiber optic cable to be 0.0465
(+0.001) inch diameter

2.3 Cable Retention: Connector retention to exceed breaking strength of glass

2.4 Optical Loss: Maximum optical throughput loss to be <2.75 dB measured at 8Q0 to
950 nm

2.5 Environmental Requirements: Temperature, temperature shock, vibration, mechanical
shock, and altitude capabilities to conform to MIL-E-5400P Class 2

2.6 Connector Durability: All requirements met after 1000 cycles of mating and unmating

2.7 Pressurization: Connectors designed for use as pressure bulkhead penetrators to mect
pressurization requirements of MIL-E-5400P Class 2. Also to maintain required gage pressure
of 30 (+5) psi during steps 2 and 12 of MIL-T-5422 for Class 2 operation

2.8 Requirements apply to single-channel and multichannel connectors

3.0 DIGITAL SIGNAL DRIVER PERFORMANCE REQUIREMENTS
3.1 Electrical: Input to ITTL load; power supply to be 5.0 (40.5) Vdc

3.2 Optical Output: Optical half-power points to be 50 nm apart ana within range of 800
to 950 nm

3.3 Power Coupling Ability: 1.25 mW into 45-mil-diameter fiber optic cable

3.4 Logic Code: 1.25 mW into 45-mil cable at application of high TTL input; <0.01W into
45-mil cable at application of low TTL input

3.5 Pulse switching time: <10 ns

3.6 Environmental Characteristics: Operate in all conditions of MIL-E-5400P Class 2
environment

3.7 Operation Lifetime: 10 000 hours continuous at 25°C

146

4.0 DIGITAL SIGNAL RECEIVERS
4.1 Responsitivitv: Platform 600 to 1100 nm
4.2 Power Supply: +5 (40.5) Vde and -5 (40.5) Vde

4.3 Yransfer Characteristics: Convert input optical signals to standard TTL output format
with fanout of 10. See following table: |

Radiant Power Power Supply Electrical
Input (watts) (Vdc) Output
Min Max Min Max
4x 107? Oe (0 27V
50); 5) @ I-mA output
current
2x 1078 0.5 V
-5 (40.5) @ 16-mA output
current

4.4 Electrical Ourput Switching Time: <10 ns
4.5 Environmental Characteristics: Operate in all conditions of MIL-E-5400P Class 2

4.6 Operation Lifetime: 10000 hours continuous at 25°C

147

APPENDIX D

10M

Fiber Optics Cost Data Collect

JIIN ‘2OpeoW *] +:e0an0¢g
S}s09 eTqeg OT3dQ-A9qTYA UO JooUS eIeq AreUTUTTOITg

(840°0

90U9
-195F5 TP
100s. 1a76° 1 6T wug¢g XOpuy WIZE’ [TTWg’*ZT doot
Ort asec ¢ LE wu¢gg SATIOVAF wwi¢y’O “{[TU QT 900T
TOG Gasol ¢ T9 wUdgg -oy) wmigc’o TTWZ°¢EZ qO0t
é é T9 wcg -- wug/'Q [TWH’OE (w4/AP OOT) VOOT
(w45-2<)(-yoob> fw cot >) 0S°0 c8°0 ley aes O10 uozSUTATTd
(AOQTF | - — (si ae
/43/1S) [AP07/35
SIS, 19 9609 10 wyzZ/
145/76 EG (wy /aP OZ)
45 /6T 61 2g
Be 3S l OOOT/aP ¢°9 BUTUIO)
uoTjeUuTuW
-29} YITM) 00Z 20 w¢g/
oAd/T Too wu CZE°T (uy /AGP O19)
-ououl ¢9°7ZS 9¢5°0 838°0 ‘[Tw €¢ 2 \at VA 116 Yoo TeA
- _— SSE TIS eX Oy,
TAUTA/TTOO /AF8TE 40 WZOT/ O8-W
-ouow ¢/°€S 08°0 wl CZ °T (wy4/adP ZLE) Teotjdg
SS 210. ‘TTwW QC at Salo uPoTIoWy

/dPorv

/33997 20 wW/g/

oAd/T Too (peainseolu)
-ouow Q¢°Z$ Gaye 17.0)
uu CCITT (w4/aP LG) 1/e%
99°0 ¢L°O ‘TTw ¢y sy aoG 1 0 OoTTTe9

“Qy/e0TAg = SOQqTYA ~"eTG = WN) SCUOT DRAG
JO “ON A0qTY BuT Ie

148

JIAN ‘t0peay_ *] sao1no0s
S$}S09) GaT uo Jooys eyeq AreCUTWT Teg

WTZ9OWAOY | agexoed

-udu soz dn OF IG
oS OU fgT ao idn QOS +
S-OL poddoug {| On°6s bs [T}U 0zZ STW OC ueFRAoOqUeT LOST
oU0D
oTsue Jey ,¢z ut
00°0ST STTW OG} St atomod Jo Z%C/<
OUdD
| opsue Frey .¢z ut
OS*6Z$| BFP TTW gT/_wEg STF“ O€| st tamod jo zoG<
00°SES Hf S[TW C uUrTIIaquey
° : oT 3ur
yoed #7°€Z71TIL | 00° zs ‘i JICU EZ UT Aw c:
eyed Zi IxitL | 00°2z$| BTP Tu gt/,wucg uULTIADqueT]
SQuUDWUMOy BOT Ad | eoAay VEXD 09 OOPRJANS | SI Pastas soereU5

SuFIITWY | SuTAI~WY oourystq| uazsgzed uoptsstwy

Sa soe 8 C28 Se Pe Ae ee a ee Os pe oes)

cot | su
Sa fu

S°T | sw

GCF MU

i ek fart
XCWAC* T mrt
VA OS | VK O

30 34

Orel S2-1
VoAOW -

a SLL1aS

Et ECSUadS
SOFuorzzosods

OL) 16 eye DK IN

009 Se ec

CQO9 COC toes

Gaae Pe j2ed) Puc

3no Ste oG aie eye y qogeyhy

149

(f+V)
O0O0°ST

00°OT

00°8T
G9°6

OEr9c Ss

(satTun 0¢)
3run/jso9

OSL] TOA
uUMOPYvoag
OSADADY

OIAN ‘A0peeyW *y, :o0an0¢
SJSOD SpOTpojOYg UOOT[IS uo AJosys ejeq AAeUTWITeIg

CTP THY 07/8"

AOS} 970.1 TFL OST < JA¢T | eUg0z Zuo /Mu/VNO* | ETP Tru ¢/g8¢°
A00Z< 947-OL FOG cm Cue > M/V ¢° BTp [Fw o¢
A00Z jdy< | vugz> M/V S°
ASL 9%-OL TFL CST | . 3dy> | eugz> aye Ge
AOOT ST-OL THY OST 3dy> | edogz M/V ST°

ce oe

DOULIIUY
DouLvTpey oF vorzy
DATROV DducjAstg

osvyseg M/V Wil OTG6"
aed oF hy

DOULTpPLCAAT

( pu) rosy
VATIOV

O8 IXTL
IL
Ws

oJULSUON

ST9TXdS

Guac i585
USSG Ses

Siro} 8 66 Sale ro rels

OTSQUN

BT 0102 OW

070m dS
Jog

JOSO0dd

yoo} Our

4070 Nid
qt Nid
Lan

LOZ7-C80S
od) ar

ITA]

150

Quantity of Bundle
Fibers in Diameter Attenuation Cost
Compan Bundle inch dB/km Lt

Galileo High Loss:

300 0.045 (50a Orr
2,000 07125 WSOGE O262
Medium Loss:
05035 400 CATES
05125 400 3.00
Corning High Loss:
0.087 1,000 35
0.120 1,000 . 69
OntZ 9 MOOG oO,
Ol s9 12000 1 as
Low Loss:
0.050 30 0)

A Comparison of Commercially Available
Multimode Fiber-Optic Cable
source: Army REPORT EMCOM-4271 by CW3 Richard D. Parent

Nov. 1974
ITEM DESCRIPTION RRECE
i P.V.C. Jacketed 35 mil dia fiber-optic cable S057) ft
Z End terminator for TO 18 Mating D5 Oe,
Be End terminator for TO 5 Mating 50.00
4 Hybrid, TO packaged, receiver in connector 400.00
5 Hybrid, TO packaged, transmitter in connector SSO
6 Complete 100-foot long Hybrid System (Complete 900.00
fiber-optic data transfer system including:
end connector with LED, end connector with
photodiode and amplifier, and 100 feet of
15 dB/km fiber-optic cable. This does not
include power requirements.)
7 LED mounted in TO 18 size connector 50.00
8 SPD mounted in TO 18 size connector 50.00

Cost Breakdown of a Fiber-Optic Data Transfer System
manufactured by Galileo Electro Optics Corporation,
Galileo Park, Sturbridge, Mass., 0158
Source: Army REPORT EMCOM-4271 by CW3 Richard D. Parent
Nov. 1974 |

Boi

General Cost Information

COMPONENT SOURCE
Fiber-optic 1. The Valtech Corporation re- Telephone con-
cable vealed in August 1975 that they versation with

have developed a commercially ECRR JOnNe Leis.
available 40 dB/km fiber-optic Code 1640, NELC,
cable with 1 to 40 fibers. 9-2-75

Prices range from $2/ft for the

single fiber cable to $12.

2. NELC accepted delivery of Telephone con-
medium-loss (590 dB/km) multi- versation with
mode (367 fibers) fiber-optic LGOReIJOUN ELIS
Cable 2b 4 cose ome eoU) ft.

This is the cabling to be used
in the A-/ ALOFT Demonstration.

3. Galileo's K2K medium-loss (<500 Telephone con-
dB/km) multimode cable is sell- versation with
ing at $0.75/ft. Lower prices Mr. Rodney
would be considered for quantity Anderson,

purchases of 100,000 ft. Galileo, 8-18-75
4. Corning's single mode (7-single Telephone con-
mode fibers per cable) cable, versation with
CORGUEDE as 'selline sommelier” wilice oper &
meter, or about $4.11/ft. Freiberger,

Corning Glass
Works, 8-18-75

Drivers/ It eEDiserete Circuit caivede NELC TD-435
Receivers receivers for the A-/ ALOFT
Demonstration cost approxi-
mately $110-120 each.

2. NELC has awarded a contract Telephone con-
to Sperry Univac for the de- versation with
livery of 60 Hybrid module LODE? J CHa ris

receivers at a cost of $54,000 2 September 1975
Give., S900 each). NEL@ has the

option to obtain an additional

30 receivers at $285 each.

Connectors 1. The ITT Cannon 13-channel bulk- Telephone con-
head connector cost was $500 versation with
each (6 made). Subsequent cost LCDR JOHN ELLIS
has been reported as $50 each
(unconfirmed).

2. Sealectro has provided NELC with NELC TD-435
Single-channel bulkhead connectors
at oe o0=s. SOmcecnmniomEtse in the
A-7 ALOFT Demonstration

Arn

APPENDIX E

Industry Contacts for Fiber Optics Components

NELC contacted the following list of manufacturers by
mail or telephone. The representatives on this list were
considered to have candidate components for the A~/7 ALOFT
demonstration. Some manufacturers do not appear on the
tables in the text because their component seemed unlikely
to exhibit the desired performance. Omission from the
following list means only that there was no response from

the manufacturer to an initial contact by NELC.
Source:
PUBER OPTIC CABLES

American Optical Corp Walt Sigmund
14G Mechanic
Southbridge, MA 01550

Ealing Optics Corp Henry Murphy
2225 Massachusetts Ave
Cambridge, MA 02140

Edmund Scientific Co
101 E. Gloucester Pike
Barrington, NJ 08007

Fiberphotics Bill Zinky
2257 Soquel Dr
Santa Cruz, CA 95060

Galileo E/O Corp Rod Anderson
Galileo Park
Sturbridge, MA 01518

Corning Glass Works Rich Cerney
Corning, NY | Bob Freiberger

sis

Electro-Optical Products
Division

Box 7065

Roanoke, VA 24019

hs) 3:

NELC TD-426

(617)
SSL IL

Col)
491-5870

(609)
547-3488

(408)
475-5242

(617)
347-9191

(607)
974-8788

(703)
963-0371

Valtech Corp
99 Hartwell St
West Boylston, MA 01583

LEDs

Fairchild Microwaves
Optoelectronics Div
4001 Miranda Ave

Palo Alto, CA 94303

General Electric
Corporate R&D

1 River Road
Schenectady, NY 12345

Peseronix, Inc
1900 Homestead Rd
Cupertino, CA 95014

Meret, Inc
1050 Kenter Ave
Los Angeles, CA 90049

Monsanto

Electronic Special Products
3400 Hillview Ave

Palo Alto, CA 94304

Motorola Semiconductors
Box 20912
Phoenix, AZ 85036

RCA Indl Tube Div
Dene. G

New Holland Ave
Lancaster, PA 17604

Spectronies, Inc
541 Sterling Dr
Rachardsom.. EX. 75050

Texas Instruments
Mami Station 12
PO Box 5012
Dalkase EX 75222

Welty Trout

Don Staub
Bruce Cairms

Jack Kingsley

Tony Heinz

Dave Medved

Grant Riddle

Francis Christian

Jim O'Brien

J. RovBrard

Gene Dierschke

154

(617)
835-6082

Cas)
493-3100

(518)
346-8771

(408)
7 fe oO

(213)
828-7496

(408)
257-2140

(602)
962-3186

Chae
597-7 Go

(214)
234-4271

(214)
238-4561

PHOTODIODES

EG&G, Inc
35 Congress St
Salem, MA 01970

Patnent ta Microwave

& Optoelectronics Div
4001 Miranda Ave

Palo Alto, CA 94303

Hewlett Packard
620 Page Mill Road
Palo Alto, CA 94303

Inotech
131 Main St
Norwalk, CT 06851

Monsanto

Electronics Special
Products

3400 Hillview Ave

Palo Alto, CA 94304

Motorola Semiconductors
Box 20912
Phoenix, AZ 85036

Quantrad Corp
2261 G S Carmelina Ave
Los Angeles, CA 90064

RCA Indl Tube Div
Dept G

New Holland Ave
Lancaster, PA 17604

SpeceLonses, Inc
541 Sterling Drive
Richardson, TX 75080

Texas Instruments
POCO oe 1Ez
Dal basa egy 52 22

Ed Danahy

Bruce Cairns

Hans Sorenson
Stan Gage

Ray Pennoyer

Wayne Stewart

Dave Durfee

Frank Ziemba

Jim O'Brien

Jz Rew Brand

Ed Harp

Col
745-3200

(415)
493-3100

(45)
BOD) 8) LAL

(203)
846-2041

(415)
493-3300
(602)

244-4556

(213)
478-0557

ea,
Sesh Fleer

(214)
234-4271

(214)
238-3274

UDT, Inc
Zeaa 50th St

Santa Monica, CA 90905

CONNECTORS

Amphenol Connector
2801 South 25th Ave
Broadview, IL 60153

Deutsch Co

Elect Components Div
Municipal Airport
Banning, CA 92220

ITT Cannon
666 E Dyer Road
Santa Ana, CA 92702

Don Dooley

Don Warenburg

Ted Alsworth

Ron McCartney

156

C7ili3))
SOs)

(312)
345-9000

(714)
849-6701

(714)
557-4700

APPENDIX F

Assumptions for an Economic Analysis
of the ALOFT Project

The following assumptions shall be utilized by NELC,
NPS students, and NELC contractors in the performance of
an economic analysis of the A-7 ALOFT Project:

The external electro-optic adapter units which were required for the ALOFT project
would not be the design approach for the multiplexed fiber optic interface in a point-to-point
data transfer system or data bus of the future, since the MUX/DEMUX and electro-optic
drivers and receivers would physically replace the I/O design presently being utilized in the
electrical interface of the peripheral avionic units.

Assuming that next-generation point-to-point data transfer systems are going to make
increasing use of electronic multiplexing to reduce the interface density, the resulting
increased data rate requires close consideration of the tradeoffs in selecting the interface
medium due to increased susceptibility to electromagnetic compatibility (EMC) problems
encountered at high data rates.

Three generally recognized methodologies exist at the present time as alternatives for
the multiplexed system interface medium which are sufficiently proved to be considered:
coaxial cables, twisted pair, and fiber optics. Millimeter waveguide technology has not suf-
ficiently developed at this time to be a viable alternative.

Once the assumption of multiplexing is made, the transmission of the data over any
one interface channel requires four basic components as the building blocks for any point-to-
point interface system: a driver, a cable, a connector, and a receiver. In a data bus system
only one additional component is required above the basic building blocks: an access coupler.

Due to the various types of connections required to install a data transfer system in a
particular vehicle or platform, there exists a subset of the connector building block consisting
of the different varieties of connectors. These varieties can be generally classified as: single-
channel bulkhead, single-channel pressure bulkhead, multichannel bulkhead (or rack/panel),
and multichannel pressure bulkhead. A splice connector is not an autonomous part of this

SL)

_— = os 7 ;

7 am a i hee oe
— i.) Ss oS ae:

ei’. | or

“subset because the basic design of a single-channel bulkhead connector can be adapted to
‘fulfill the requirement for a splice at no mujor change in cost. A printed circuit card con-
nector tsa possible variety, but is also an adaptation of the single-channel bulkhead

- connector.

The optical signal driver can be considered to be dimensionally the same size as its
electrical counterpart since both can be created from discrete amplifier circuitry or hybrid
or integrated circuits. The same can be basically stated for the signal receiver. The only dif-
ference that may occur in the evolving design of optical drivers and receivers from their elec-
trical counterparts is a modular optical component such as the light source or light detector
that can be inserted into or removed from the discrete or hybrid circuitry. The necessity of
this modularity depends on the reliability that is achieved by the manufacturers of the optical
components. If the optical components become as reliable as the rest of the driver or receiver
circuitry, modularity will not be required. This contingency also has major bearing on the
maintainability that will evolve for optical drivers and receivers.

For installation and handling considerations, the fiber optic cable can be considered
to be susceptible to the same requirements as the coaxial cable.

Preliminary review of the performance capability of twisted-pair components leads to
the preliminary conclusion that twisted pair is not a valid alternative for a multiplexed inter-
face due to inability to handle high data rates without extreme susceptibility to EMC prob-
lems. This preliminary analysis requires further analysis to validate the conclusion. However,
for initial analysis purposes no effort will be made to gather the component data for the
twisted-pair case until this conclusion is refuted by further analysis.

The performance, cost, weight, and power consumption data for the components
required for a coaxial interface are readily available from commercial manufacturers. A
review of the data has led to the selection of candidate coaxial components that could have
been used in a coaxial version of the ALOFT interface. The identity and data for these com-
ponents are included in table 1C. These components shall be utilized as the coaxial interface
baseline in the analysis.

The performance, cost, weight, and power consumption data for the fiber optic com-
ponents required are not readily available from commercial manufacturers. NELC Code 2540
(Fiber Optics Systems Branch, Electro-optics Technology Division) has compiled performance
requirements for each of the basic fiber optic components that are projected to be required
in FY77 to fulfill a fiber optic point-to-point interface system requirement in an aircraft data
transfer system. These performance requirements are constrained by a MIL-E-5400P Class 2
environment. These performance requirements for each building block are attached as
appendix A to this concept report. The economic analysis will gather the cost, weight, power
consumption, and any other required data for the fiber optic components. This data gather-
ing will be an iterative approach based on the Delphi predictive analysis method.

Source: NELC TD-435

158

y — ; ‘
-, 7 sos -
> oiet ts

- —

NO"

ee
2
i:

14.

BIBLIOGRAPHY

Air Force Avionics Laboratory Technical Report AFAL-TR-

73-267, Part 1, Fiber Optics and Related Technology,
by D.D. Matulka and others, p. 1-147, November 1973.

Army Electronics Command Research and Development
Technical Report EMCOM-4271, Application of Fiber

Optic Technology to Army Aircraft Systems, by
CW3 Richard D. Parent, p. 1-53, November 1974.

Army Strategic Communications Command Report AD-754 566,

Optical Fiber Links for Telecommunications, Part l,
by Stephan F. Fulchum, Je.) ewe Joamesm.

Burke, Dr., p. 36, December 1972.

Anderson, Rodney, Galileo E/O Corp., Telephone conversa-
tion with author, 15 August 1975.

Ayres, R. U., Technological Forecasting and Long-Range
Planning, p. 8, McGraw-Hill, 1969.

"Bell Crashes the 2 dB/km Barrier," Electro-Optical
Systems Design, p. 4, August 1974.

Biard, J.R., Spectronics, Inc., Telephone conversation
with author, 15 August 1975.

Bielawski, W.B., ''Low-Loss Optical Waveguides: Current

Status,"' Electro-Optical Systems Design, p. 22-28,
April 1973.

The Boston Consulting Group, Perspective on Experience,
The Boston Consulting Group, Inc., 1970.

Cetron, Marvin J., Technological Forecasting, p. 149,
Technology Forecasting Institute, 1969.

Chief of Naval Operations letter dated 3 July 1974.
Currie, Malcolm R., Dr., memo dated 6 August 1973.
Electronic Design News, p. 12-13, 15 February 1972.

Ellis, John, LCDR USN, NELC (Code 1640), numerous
telephone calls throughout 1974-75.

159

Oy.

ro.

S..

ZO,

Ze.

22,

24.

Zi;

Zor.

PnevyelOpedia Britaniea, Vall0, p. 475, L967.

Encyclopedia of Science and Technology, V. 9, p. 536,
McGraw-Hill, 1964.

Fisher, G.H., Cost Considerations in Systems Analysis,
p. 10, American Elsevier, 1971.

Free, John, "The Coming Age of Fiber Optics: How wires
of glass will revolutionize communications,"

Popular Science, p. 82-85, August 1975.

Freiberger, R., Corning Glass Works, Telephone conversa-
tion with author, 18 August 1975.

Gordon, T.J., and Helmer, Olaf, Report on a Long-Range
Forecasting Study, Report P-2982, The Rand Corpora-
tion, Santa Monica, September 1964.

Hirschman, W.B., "Profit from the Learning Curve,”
Harvard Business Review, p. 125, January-February 1964.

Ikegama, T., and Suematsu, T., "Carrier Lifetime
Measurement of a Junction Laser Using Direct

Modulation,'' IEEE Journal of Quantum Electronics,
Ve OB-Ge =p. L145. spre igGc.

Jantsch, Erich, Technological Forecasting in Perspective,
p. 137, Organization for Economic Co-Operation and

Development, Paris, 1967.

Kahn, Herman, and Bruce-Briggs, B., Things to Come,
p. 186-188, MacMillan, 1972.

Kapany, N.5S., Fiber Optics, Principles and Applications,
p. 1-10, Academic Press, 1967.

"The Light Wave of the Future," Business Week, p. 49,
1 September 1975.

Drs itgare ot Bereyhinskia> bei yeeand Valakh, M. Ya.
Fiber Optics, Israel Program for Scientific
Deanigtaclons, pp. 5o, L972.

Naval Electronics Laboratory Center Report 405-103,
Fiber Optics Communication.

160

- 2 wr |
_)
- > =a

BO.

Say

Syn

a3.

34.

356

50)

Sh.

oor.

Naval Electronics Laboratory Center Technical Note 3000,
Economic Analysis and Life Cycle Cost Model for
Naval Telecommunications Systems, by R.A. Greenwell
and others, p. 18, 30 June 1975.

Naval Electronics Laboratory Center Technical Report
1891, Fiber- and Integrated-Optic Communications
Technology, by W.E. Martin and D.J. Albares,

24 August 1973.

Naval Electronics Laboratory Center Technical Document

435, A-7 ALOFT Economic Analysis Development Concept,
by J2R. Ellis and ReA Greenwell lap ee5.) 7 July 97>:

Naval Research Laboratory Memorandum Report 3004, EOTPO

Major Issue Review, Fiber Optics and Integrated
Opercs. EOLPO Reporter NO c7 Orn p ee pe Diletelgi ee

Naval Electronics Laboratory Center Technical Report
1763, Revision 1, Transfer of Information on Naval

Vessels Via Fiber Optics Transmission Lines, by
Hal, GibeaalOr, Dp. 5, otlayeelou ae,

Novotny, G.V., "Fiber Optics for Electronics Engineers,"
Electronics, p35.) NO. 622 eo.

Onliaberwt, L.. LoneePachwel rancnlcotonmin liber. Opetecns
Electro-Optical Systems Design, p. 18-21, August 19/1.

"Peek-A-Boo, I See You,'' Electro-Optical Systems Design,
pe, June 1973.

Rand Corporation, Military Equipment Cost Analysis,
pee gee Loy lL,

TaysteueeH.'., Caton, W.M.>5 and Lewilse A,L., ‘Data

Bussing with Fiber Optics,'' Naval Research Review,
p. 12-25, February 1975.

ILO

INITIAL DISTRIBUTION LIST
No. Copies

Defense Documentation Center 2
Cameroun ofa Lon
Alexandria, Virginia 22314

Library, Code 0212 2
Naval Postgraduate School
Monterey, California 93940

Department Chairman, Code 55 1
Department of Operations Research
and Administrative Sciences
Naval Postgraduate School
Monterey, California 93940

LCDR J. M. McGrath af
511 South Newland St.
Denver, Colorado 80226

LCDR K. R. Michna af
1374) Jackson Street
North Chicago, Illinois 60064

Dean of Research Administration iE
Code 023

Naval Postgraduate School

Monterey, California 93940

Professor C. R. Jones, Code 55J5S 5
Naval Postgraduate School
Monterey, California 93940

Associate Professor B. R. Pierce, 1
Code 6423

Naval Postgraduate School

Monterey, California 93940

CDR D. Forsgren il
CNO (OP-982E2)
Washanezon, D.C. 20350

Chief of Naval Material 1

NMAT -03422
Washington, D. C. 20360

162

eZ,

kde

14.

AG.

12)

A. D. Klein

Naval Air Systems Command
ATR-360G

Washington, D. C. 20361

CAPT R. N. Winkel

Naval Air Systems Command
ATR<-533

Washington, D. C. 20361

Andrew S. Glista

Naval Air Systems Command
ATR-52022

Washington, D. C. 20361

Ne buble:

Naval Electronic Systems Command
ELEX=304

Washington, D. C. 20360

K. C. Trumble

Air Force Avionics Laboratory
AAM

Wright-Patterson AFB, Ohio 45433

J. Ramage/Gene James
AFFDL/FGA
Wright-Patterson AFB, Ohio 45433

CWO R. Parent, USA

U. S. Army ECOM

Avionics Lab, AMSEL-VL-A

Ft Monmouth, New Jersey 07703

T. R. Coleman

LTV Vought Systems Division
Code 2-14000

P, Ov Box 5907

Dallas, Texas 7/5222

R. Betts

IBM Federal Systems Division
Owega, New York 13827

EGS

20.

Commander
Naval Electronics Laboratory Center
271 Catalina Blvd
san Diego, California 92152
Attn:
Code 1640 (J. Ellis)
Code 235 (R. Greenwell)
Code 220 (D. Williams)
Code 2500 (T. Meador)
Code 4400 (G. Holma)
Code 6/700

164

de ee

Thesis
Mi88475

eC.

McGrath
An approsech to the
estimation of life
cycle costs of a fiber-
optic application in
military aircraft.

approach to the estimation of life cy

iin

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "An approach to the estimation of life cycle costs of a fiber-optic application in military aircraft"

See other formats