AD-A280 235

,,-,.-,,-,o,, EI Ig1 11 I •
SENSITIVITY AND SIGNAL TO NOISE RATIO IMPROVEMENT OF A ONE
MICRON LADAR SYSTEM INCOPORATING A NEODYMIUM DOPED
OPTICAL FIBER PREAMPLIFIER 0

LASER RADAR TESTBED

Michael S. Salisbury @
Electro-Optics Techniques Section
WLJAARI-2, Bldg 622
3109 P St.
Wright-Patterson AFB, OH 45433-7700

February 1994

Final Report for Period August 1991 to November 1993

DTIC S
Sl ELECTE

Approved for public release; distribution is unlimited JUN 1 41994D

AVIONICS DIRECTORATE
WRIGHT LABORATORY
AIR FORCE MATERIAL COMMAND
WRIGHT-PATTERSON AIR FORCE BASE, OH 45433-7409

94-18164

94 6 13 057
• 0 0 0 0 0 0 0 *
 NOTICE

When Government drawings, specifications, or other data are used for any purpose other
than in connection with a definitely Government-related procurement, the United States
Government incurs no responsibility of any obligation whatsoever. The fact that the
Government may have formulated or in any way supplied the said drawings, specifications,
or other data, is not to be regarded by implication, or other wise in any manner construed,
as licensing the holder, or any other person or corporation; or as conveying any right or
permission to manufactu, use or sell any patented invention that may in any way be
related thereto.

This report is releasable to the National Technical Information Service (NTIS). At NTIS,
it will be available to the general public, including foreign nations.

This technical report has been reviewed and is approved for publication.

MICHAEL S. SALISBURY, Cont r, T/SSI DONALD L. TOMLINSON, Chief

Electro-Optics Techniques Group, Electro-Optics Techniques Group
Electro-Optics Branch Electro-Optics Branch

PAUL F. McMANAMON, Acting Chief

Mission Avionics Division
Avionics Directorate

If your address has changed, if you wish to be removed from our mailing list, or if the
addressee is on longer employed by your organization, please notify WL/AARI-2, Wright-
Patterson AFB OH 45433-7700, to help maintain a current mailing list.

Copies of this report should not be returned unless return is required by security
considerations, contractual obligations, or notice on a specific document.

4 •0 0 0 • 0 0 0
 Approved
REPORT DOCUMENTATION PAGE Fo

Piulic repowrg n for thi COllectiO of infloriitatoui attumatia to ar=1r I hour per elsponse. ludintg Idie,t1e for rorwing uflwudi[fin. ieefthng emating date soufl.
ad m
gaterig intigte data needed, anfaeeigad wn h olction of infononatiori Sendcamme~readugt" i budn aimaeo ansohrepect of thi
for in ormnation
turden to Washington HNitiquartert Sertte. OuriectOrarte
normation,. including swggetiona for reducing thise
=oet ionOf Operations and Paeponis 1215 Jefferape
DeavsHiqgway. Suit@1204. Arlington. VA 222024302. and to the Office of Management And$udget. Paperwork Ackdction Project (07040 14). Waiingtot. OC20503.

i. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
128 Feb 94 ITechnical Re rt 1 Aug 91 - I Nov 93
4. TITLE AND SUBTITLE Sensitivity and Signal to Noise Ratio im- S. FUNDING NUMBERS
provement of a One Micron Ladar System Incorporating a
Neodymium
! r..1....,-f,-Doped Optical
T.-_, .. Fiber
T.. Preamplifier. Subtitle:
.•. . . PE 61101F
r6. AUTH4OR(SY" PR 2004
TA 08
WUO06
Michael S. Salisbury _U_06

7. PERFORMING ORGANIZATION NAME(S) AND ADORESS(ES) B. PERFORMING ORGANIZATION

Technology/Scientific Services Inc. REPORT NUMBER
P.O. Box 3065 Overlook Branch
Dayton OH 45437-3065
L0

9. SPONSORING /MONITORING AGENCY NAME(S) AND ADORESS(ES) 10. SPONSORING/ MONITORING

AGENCY REPORT NUMBER
Avionics Directorate
Wright Laboratory WL-TR-94-1075
Air Force Materiel Command
Wright Patterson AFB OH 45433-7409
11. SUPPLEMENTARY NOTES
This report was prepared for WL/AARI-2 as a Masters Thesis project for the University
of Dayton.

'12a. DISTRIBUTION/ AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE 0 0

Approved for public release; distribution is unlimited.

13. ABSTRACT (Maximum 200 words) In an effort to increase the signal to noise ratio of a con-
tinuous wave, one micron all solid state ladar system, a rare earth doped optical
fiber amplifier has been investigated as a preamplifier for ladar return signals.
This details the experimental system used and provides a theoretical analysis of the
fiber amplifier's effect on heterodyne and direct detection. The SNR is plotted as
a function of the return signal power, and a SNR threshold defines a minimum detec-
table signal power. The return signals required to attain the SNR threshold are com-
pared for four cases: direct detection with and without the fiber amplifier, and
ýheterodyne detection with and without the fiber:amplifier. For direct detection,
these results predict a sensitivity increase of 21.0 dB, yet for heterodyne detection
the predicted sensitivity increase is only 4.0 dB. These SNR equations are then used
to predict experimental improvements of 42.0 dB and 4.0 dB for the direct detection
and heterodyne detection experiments, respectively. Experimentally measured in-
4creases in SNR are then compared to these predictions. Specifically, for direct de-
tection a SNR increase of 36.5 dB has been measured, and for heterodyne detection the
,experimental work yielded an increase of 8.0 dB.
114. SUBJECT TERMS 15. NUMBER Of PAGES
73 S
16. PRICE CO0E
Ladar. Lidar Fiber SN_ _alifier.
17. SECURITY CLASSIFICATION I 8. SECURITY CLASSIFICATION [19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT
OF REPORT OF THIS PAGE OF ABSTRACT
Unclassified Unclassified Unclassified
NSN 7540-01-280-5500 i Standard Form 298 (Rev. 2-89)
Prescribed by ANSI Std Z39-18
298-102

4 a i mi ii n mi I
 ACKIOWLEDGM•ENS

This Technical Report was prepared by the Electro-Optics Techniques Group, Electro- 0
Optics Branch, Mission Avionics Division, Wright Laboratory, Wright-Patterson Air
Force Base, Ohio.

The following personnel contributed to the formulation and review of this report:

Michael S. Salisbury, T/SSI

This Technical Report was submitted by Michael S. Salisbury in February 1994.

Accesion For 0
NTIS CRA&M
DTIC TAB
Unannounced LI
Justification ....................

Y .. ...
Distribution I
Availability Codes

Dist
Avail and Ior
Special
*
f-/ I

00

iii 0

.0..0 . . . . .. ... .. . 0.. . . . . .. . .. .I . . . . . _ _ 0

 PREFACE

A 1.06 micron ladar test bed has been built by the Electro-Optics Branch of the

Mission Avionics Division at Wright Patterson Air Force Base. The ladar system was

built with in-house research finding to evaluate experimental ladar devices as components

of a ladar system. This system characterization of the devices can reveal undesirable

properties of the devices early in the development phase, properties that might not be seen

in a device characterization.
The Air Force Office for Scientific Research (AFOSR) funded initial work on a
rare-earth doped optical fiber preamplifier by Dr. Richard Miers, Indiana * *
University/Purdue University - Fort Wayne (IUPU-FW). The goal of the project was to

investigate the effect of the fiber amplifier on ladar sensitivity by incorporating it into the

one micron ladar test bed. In-house research monies at Wright Laboratory were used to

provide finding for a University of Dayton graduate research assistant, who was hired

through a contract with Technology/Scientific Services Incorporated (T/SSI) to develop

the fiber preamplifier and incorporate it into the ladar test bed. This thesis is the result of
that work effort.

iv
 TABLE OF CONTENTS

A B STRACT .......................
.....................................

ACKNOWLEDGMvENTS.............................................................................
PREFACE...................................................................................... iv

TABLE OF CONTENTS ..................................................................... v

LIST OF FIGURES AND TABLES.......................................................... vii

LIST OF VARIALBLES ....................................................................... vyiii

CHAPTER
I. INTRODUCTION...................................................................I
11. LADAR SYSTEM.................................................................. 6
2. 1 Direct Detection System .................................................. 6
2.2 Heterodyne Detection System ....................................... 1
2.3 Target Selection............................................................ 13
II. SIGNAL TO NOISE RATIO THEORY FOR
DIRECT DETECTION ..................................................... 15..i

3.1f Direct Detection Without the Fiber Amplifier ............................ 15

3.2 Direct Detection With the Fiber Amplifier ............................... 22

IV. SIGNAL TO NOISE RATIO THEORY FOR HETERODYNE
DETECTION ........................................... ....
29

4.1 Heterodyne Detection Without the Fiber Amplifier ...................... 29

L
 4.2 Heterodyne Detection With the Fiber Amplifier .................................... 33
V. PREDICTION OF SIGNAL TO NOISE RATIO

INCRE ASE S .............................................................................................. 37

5.1 Direct detection SNR .................................................. 37

5.2 Heterodyne detection SNR ............................................................ 42

VI. EXPERIMENTAL DATA ........................................................................ 47

VII. CONCLUSIONS AND RECOMMENDATIONS ..................................... 52

APPE NDIX A .................................................................................................................. 53

B IBLIO GRAPHY ....................................................................................................... 58

VIT A.................................................................. ............................................................... 61

VITA.............................
S.... .... e.. ............. ............ 61...... • ! •

vit
 LIST OF FIGURIES AND TABLES

Figure 1: Direct Detection Ladar System ........................................................................... 7

Figure 2: Geometry of the Rutgers Neodymium Doped Fiber ........................................... 8

Figure 3: Fiber Amplifier Pump Scheme ........................................................................... 9

Figure 4: Heterodyne Detection Ladar System ............................................................. 12

Figure 5: Detection Electronics ................................................................................. . 16

Figure 6: Direct Detection Signal to Noise Ratio vs. Return Signal Power 0
for Detection Without the Fiber Preamplifier ............................................... 20

Figure 7: Direct Detection Signal to Noise Ratio vs. Return Signal Power
for Detection With the Fiber Preamplifier .......................... "................................ 28 * *
Figure 8: Heterodyne Detection Signal to Noise Ratio vs. Return Signal
Power for Detection Without the Fiber Preamplifier .......................................... 32

Figure 9: Heterodyne Detection Signal to Noise Ratio vs. Return Signal

Power for Detection With the Fiber Preamplifier .......................................... 36

Figure 10: Experimental Signal to Noise Ratio Data: Direct Detection

Without the Fiber Preamplifier ................................................................. 47

Figure 11: Experimental Signal to Noise Ratio Data: Direct Detection 0

With the Fiber Preamplifier ........................................................................ 48

Figure 12: Experimental Signal to Noise Ratio Data: Heterodyne

Detection Without the Fiber Preamplifier ................................................... 49

Figure 13: Experimental Signal to Noise Ratio Data: Heterodyne

Detection With the Fiber Preamplifier ........................................................ 50

Table 1: List of Variable Values for use in Signal to Noise Ratio

Equations ................................................................................................... . . 21

vii

' .* 'L
 LIST OF VARIABLES

linear polarization unit vector of the return signal

random polarization unit vector of the spontaneous emission

Be detector electrical bandwidth 0

Bo optical bandwidth of the 4 nm bandpass filter
D heterodyne detection electronics voltage divider effect
E electric field
f optical chopper frequency

ga Analog Modules electronic amplifier voltage gain

gi M~fiteq electronic amplifier voltage gain
G fiber amplifier optical power gain
h Planclks constant

ir signal current from the detector 0

ir-se return signal-spontaneous emission beat noise current

ise.se spontaneous emission-spontaneous emission beat noise current
iIF intermediate frequency signal current
(1.2) mean squared noise current

(j,2) mean squared dark current noise

(•.) mean squared shot noise current
(,,2) mean squared thermal noise current

Id detector dark current

k Boltzmann constant

Vi.t
 0

0
N number of spontaneous emission-spontaneous emission beat noise terms at 0
2 kHz
Nh number of spontaneous emission-spontaneous emission beat noise terms at
200 MHz

Pr optical return signal power incident on the detector

Plo optical local oscillator power incident on the detector
Pse optical spontaneous emission power incident on the detector
Pse, 8v incremental spontaneous emission power

RSA spectrum analyzer input impedance

Rtrans detector package transimpedance
9R detector responsivity S
SNRdir,w/o direct detection signal to noise ratio without the fiber amplifier
SNRdir w direct detection signal to noise ratio with the fiber amplifier
SNRhetw/o heterodyne detection signal to noise ratio without the fiber amplifier S
SNRhet,w heterodyne detection signal to noise ratio with the fiber amplifier
t time
T temperature in Kelvin 0
Vr signal voltage after the detector package
VIF intermediate frequency voltage
Vr,amp signal voltage after the electronic amplifier 0
Vn,amp noise voltage after the electronic amplifier
VIF,amp intermediate frequency voltage after the electronic amplifier
spectrum analyzer resolution bandwidth 4
8v
A IF frequency
ad4 predicted SNR increase for direct detection
A,, predicted SNR increase for heterodyne detection S

ix
 O
exd
peimental SNR increase for direct detection
Ah expimental SNR increase for heterodyne detection
O:f:f random phase of the spontaneous emission power

rrl electrical signal power for direct detection without the fiber amplifier
rr2 electrical signal power for direct detection with the fiber amplifier
rIFa electrical IF power for heterodyne detection without the fiber amplifier
FIpb electrical IF power for heterodyne detection with the fiber amplifier
rNI electrical noise power for direct detection without the fiber amplifier
rN2 electrical noise power for direct detection with the fiber amplifier
rN3 electrical noise power for heterodyne detection without the fiber amplifier
rN4 electrical noise power for heterodyne detection with the fiber amplifier

rn,rse electrical return signal-spontaneous emission beat noise

Fn,lo-se electrical local oscillator-spontaneous emission beat noise
rn,se-se electrical spontaneous emission-spontaneous emission beat noise
Fea electronic amplifier electrical noise power
lqopt,d optical efficiency of the direct detection system between the fiber amplifier
and the detector 0
7lopt,h optical efficiency of the heterodyne detection system between the fiber
amplifier and the detector
v bandpass filter center frequency
o~o angular frequency of the return signal
TH miraidom mixing efficiency terms

x 00

••. ... ..e •...... .. ..... _ .... ..... e ......... .• ... .. . .. • .. .... .. •oO0!
 0

0
S

CHAFPER I
INTRODUCTION

Electromagnetic waves have been used for communications since the first demonstration

of radio was made by Guglielmo Marconi in 1895, shortly after Heinrich Hertz
successfully generated and detected the electromagnetic waves predicted by James Clerk

Maxwell. Hertz and Marconi generated the radio waves by changing the current in a wire

over time, causing the wire to emit electromagnetic radiation. The wavelengths Marconi

used represent the upper end of the electromagnetic spectrum, which can be divided into
radio wavelengths and optical wavelengths. The radio wavelengths range from millimeters

to hundreds of kilometers and are divided into microwave, short-wave radio and long-

wave radio bands; the optical wavelengths range from tens of nanometers to hundreds of

micrometers and are divided into ultraviolet, visible and infrared bands. 1 Radio

technology pushed towards shorter wavelengths which allow larger information

bandwidths, and in 1935 the German Army began to use radio for military communication

systems. Shortly thereafter, radio waves were used detect and track aircraft. This
technology was given the acronym of RADAR, which stands for BAdio Detection And

Ranging. Radar advanced rapidly, fueled by World War I and World War 11.2

A new method of generating electromagnetic radiation was discovered based on

the quantum theory of matter. The orbital states of electrons around the nucleus of atoms

give rise to radiation energy transfer as electrons shift energy levels. In 1950 MASER

(Microwave _Amplification by Stimulated Emission of Radiation) technology was

developed, using the electron energy levels in materials such as Ruby to generate
1

* S *
 0

electromagnetic waves in the microwave band. 3 Ruby was then pumped optically, which
resulted in LASER (Light ,mplification by Stimulated Emission of Radiation) technology.
This gave coherent electromagnetic radiation at optical wavelengths with uniform
wavefronts, making optical detection and ranging possible. Detection and ranging with a
laser was initially designated as LIDAR (Light Detection And Ranging), where it was
understood that the source was a laser. Over time, another acronym for laser sources has
become prominent, LADAR (Lser Detection And Ranging), although both acronyms are 0
still used. 2

Ladar and radar systems have different performance properties, and both have
advantages and disadvantages in certain areas. The primary differences between the two 0

systems are due to the nature of the emitting process and the different wavelengths. The
efficiency of a ladar source is much lower than that of a radar source because of the
inefficiency in obtaining ladar transmitter energy by quantum mechanical means. The 0

wavelength dependence of ladar and radar performance results in a tradeoff between

transmission and resolution properties. The long wavelengths of radar allow high
transmission through the atmosphere, even when heavily cluttered with pollutants, rain, or
other particles. Due to the long wavelengths, however, a radar system has very low
angular resolution. On the contrary, ladar wavelengths suffer from atmospheric
attenuation due to light absorption by the atmosphere and light scattering off the •

particles, 4 but the low wavelengths result in high angular resolution and high information
bandwidths. The combination of a ladar system and a radar system can be used to obtain
an optimal amount of information, but the development of ladar technology is far behind 0
5
the radar technology.
Ladar systems have unique applications due to some of the properties already
0
mentioned. Remote sensing of the atmosphere is an obvious choice because of its ability
to "see" particles in the atmosphere. Ladar can be used to track wind patterns such as
 N0

3 I
weather phenomena and turbulence resulting from aircr traftfic. 6 ,7,8, 9 Pollution

monitoring is another use, resulting from the optical wavelength absorption properties of

particles in the air. Some pollutants will absorb certain wavelengths and reemit a

spectroscopic signature, allowing contaminants to be identified and monitored remotely. 10

The angular resolution and range information capabilities of ladar lend it to a wide variety
of rangefinding, tracking and targeting applications, many of which are in remote military

and space based systems. 11,12,13,14 Military and space based applications create
restrictions on ladar systems, particularly the laser transmitters.

Ladar technology has advanced rapidly over the years, fueled by the continual

development of laser sources. The most advanced sources for ladar use gases as a lasing
medium, but these sources are large and heavy, and they have a limited lifetime as the gas

is slowly consumed as a result of the lasing process. Solid state lasers are more reliable,

more compact and have longer lifetimes than other sources, but are less technologically 0
developed. For the remote applications mentioned previously, these advantages are vitally

important, and a large amount of research is currently underway to bring solid state lasers

to the maturity level of the gas sources. 1 5 , 16 , 17 0

Current solid state laser radar wavelengths include 1.06 pm (Nd:YAG) and
recently 2.09 pm (CrTmHo:YAG), where the parenthetical information indicates the

composition of the respective laser materials.1 8 These solid state systems are more 0

efficient, have longer operating lifetimes, and are more compact and light weight than the

common CO2 ladar, which is representative of mature ladar systems with gas sources.19

One of the most important characteristics of any ladar system is its sensitivity. One

method of increasing the sensitivity of a remote sensing device, such as a ladar system, is
to optically amplify the return signal before detection. Rare earth doped optical fiber
preamplifiers, highly developed by the communication industry, offer a compact and

lightweight optical preamplifier for integration into a solid state ladar system. The primary

0
I4
fiber amplifiers developed by the telecommuication ndustry are constructed using
praseodymium and erbium doped fibers, for amplification at 1.3 microns and 1.55 microns
respectively. 20 ,2 1 Unfortunately, these doped fibers will not amplify the return signal
from a 1.064 or 2.09 micron ladar system. Neodymium doped (Nd 3 +) optical fibers,
however, have been developed for use as law sources.22,23 These fibers are primarily
made into fiber lasers by coating or polishing the fiber ends to create a cavity. By
combining the Nd3 ÷ fibers used for fiber lasers and the concept of a fiber amplifier from
telecommunications, an optical fiber amplifier can be developed and incorporated into a
1.064 micron ladar system, which represents mature solid state ladar technology.
When a fiber amplifier is added to a ladar receiver, the return signal is increased by 0
a power gain factor. However, spontaneous emission from the fiber amplifier adds an
optical noise to the receiver, so the benefit of the amplified return signal must be weighed
against the increased noise. 0
Direct detection ladar systems operating at near infra-red wavelengths are limited
by the noise generated by the detection electronics. In a simple, inexpensive system, these
noises can be quite large. When the noise added by the fiber amplifier spontaneous 0
emission does not have a significant impact on the overall noise of the system, a large
increase in direct detection sensitivity is achieved by adding the fiber amplifier, as we will
show later. 0
For heterodyne detection, a large local oscillator (LO) power is mixed with the
return signal in order to ensure that the detection is LO shot noise limited. In this case,
the spontaneous emission directly increases the shot noise. Also, the beat noise term 0
between the spontaneous emission and the large LO has a strong impact on the noise level.
Overall, the sensitivity of a heterodyne ladar system will be shown not to increase as
dramatically with the addition of a fiber amplifier as the direct detection scheme. S

"1 * ..... n .. nlI 9.....1 i 9..

ml • Al 9 0
 I@ 0
In Chapter 2 a description of the ladar system is given, and the experimental setup 0
used to incorporate the fiber amplifier into the system is discussed. In Chapters 3 and 4
signal to noise ratio (SNR) equations are derived for direct detection and heterodyne
4 detection, respectively. The SNPR equations for detection with and without the fiber 0

amplifier are used in each chapter to plot SNR vs. return siga power to give a measure
of the sensitivity improvement for each case. The SNR equations from Chapters 3 and 4
are also used in Chapter 5 to predict experimental SNR improvements. Experimental
results are presented in Chapter 6, and Chapter 7 contains a summary and proposals for
future experiments.

-0-

0@

... . ... t ...

.. . ......
.. . ... . ....
.. . . . .. .. . . .. ....J ... .... .. . t .. . ... . . .. ...O . .. . ... • . ..... . ... 0.
. ....
 0

CRAFER U
LADAR SYSTEM 0

This chapter provides an overvew of the test bed ladar system, focusing on some aspects
of the design that impacted the experimental work. Figures 1 and 4 show the ladar system
layout for direct detection (Figure 1) and heterodyne detection (Figure 4) with the fiber
amplifier inserted.

2.1 Direct Detection System

For the direct detection system, shown in Figure 1, a Nd:YAG laser, Lightwave
Electronics model 120-03, provides 40 milliwatts of linearly polarized output at 1.064 0 0
microns. The laser output passes through an Electro-Optics Technology model 1845-5
Faraday optical isolator to prevent backscatter from other optical components in the

system from reentering the laser head. The beam then passes through a varable rafto
attenuator (VRA) consisting of a half wave plate in a rotatable stage and a polarizing
beamsplitter cube. As will be described in detail later, this is used to reduce the
transmitted optical power to levels appropriate for the targets we have used. Also note
that as the photodetector we chose to use is AC coupled, the beam is modulated by a
Laser Precision model CTX-534 chopper operating at 2 kHz.
After passing through the VRA, the beam passes through a tranmisraeceiv (T/R)
switch, consisting of a polarizing beamsplitter cube and an Electro-Optics Technology
model 1845-5 Faraday optical rotator used to rotate the outgoing beam polarization by 45

' ' .

S....
. I • 'd- • .... .
 7

degrees. The beam then travels to the target after which the backscattered optical return
signal received from the target is rotated an additional 45 degrees by the optical rotator.
The polarization of the received signal light is thus perpendiclar to the original outgoing
light and is reflected by the polarizing beamsplitter cube into the return signal leg.

Polariing
Optical Variable Ratio Beemspher Optical
Isolator Attenuator Cube Romor

Nd:YAG chopper k I

Dichroic.Mliror
Diode
Fiber Coupler

Bund"=s Detection
22 Meter Spool of Filter Elaics 0
Nd Doped FiWer

Fiber Air Splice

FIGURE 1: Direct Detection Larat System. This figure shows the

1.064 micron ladar in its direct detection configuration. The variable ratio
attenuator is outlined, and the fiber amplifier is shown inserted into the
return signal leg.

The return signal from the target then passes through a dichroic mirror and is
coupled into the core of a Rutgers University Nd3 + doped optical fiber, whose double
cladding configuration is shown in Figure 2,23 while pump light from a laser diode is
simultaneously reflected off the dichroic mirror and coupled into the rectangular inner
cladding of the doped fiber.

. .. . •., .........•.. .,. .. - ........

S..
..•. . ,m .9
V , -- @ . . . . . . .• •ql •
 m0

Outer Claddin
125 microns Low Index Polymer 0
Core: 5 x3.3 microw O A
Inner Claddin-:

Hard Polymer Buffer

FIGURE 2: Geometry of the Rutgers Neodymium Doped Fiber. The

double cladding geometry of the Rutgers fiber is shown here. The unusual
rectangular inner cladding of the fiber is used to couple the pump light
from a laser diode into the fiber.

Figure 3 shows in detail the setup used to couple the return signal and the pump light into
the fiber amplifier.2 4 The laser diode used to pump the fiber is a Laser Diode 0

Incorporated model LDT 26010 laser diode with an 808.4 nanometer, 500 milliwatt CW
output and a built in Thermo-Electric (TE) cooler. The TE cooler acts as a heat pump,
which cools the laser diode emitter and decreases the output wavelength, allowing the •
wavelength to be tuned over a small range. The TE cooler shifts the wavelength of the
laser diode to the peak absorption of the neodymium doped fiber, given to be 805
nanometers by Rutgers University. To maintain constant operating conditions, a
thermistor monitors the temperature and stabilizes the output power and wavelength by
using a feedback loop to vary the current to the TE cooler.
A grin rod lens from a Newport F-GRKI graded-index rod lens kit minimizes the
divergence of the.output beam from the laser diode. As seen in Figure 3, the return signal
and pump light are combined using a dichroic mirror, which transmits the 1.064 micron
return signal and reflects the 805 nanometer pump light. The two combined beams are
coupled into the fiber simultaneously using a Newport F-1015 high precision single mode
fiber coupler. As the pump light travels through the fiber it passes through the core and
creates a population inversion in the rare earth dopant, causing the return signal to be
amplified as it passes through the core. Note that due to typically small return signals, the

. . •
0 0 0 00 0 0
 9S

fiber amplifier is operated in the amush signal regime. Thus, spontaneous emission ftom the
upper lasing level is added to the amplified retun. signal. We will see the effect of this
spontaneous emission power later.

Retium Sip.!
1064 a

GRIN
Len Tranmsit 1064 im

805 rm

Fiber Copler

22 Meter Spool of
Nd 3+ Doped Fiber

ST Comiec0tor

Fiber to Fiber Coupler

with 4 Nanometer Baud
Pan Fibter0

FIGURE 3: Fiber Amplifier Pump Schense. A dichroic mirror is used

to couple the pump light and the return signal light into the doped fiber.
The pump is a laser diode operating at 805n.ometers, with a GRIN lens
used to minimize. the divergence. A four nanometer bandpass filter is
inserted following the fiber amplifier to block excess pump liht and to
mintimie the effects of spontaneous emission. The output ST of Cmlme0
the bandpass
filter is connected either to an evanescent wave coupler for heterodyne
detection (see Figure 4) or directly to the detector for direct detection (see
Figure 1).

FIUEUieApiirPm chm.Adcri9mro sue

To couple the amplified return signal out of the fiber amplifier, there is a single
mode AT&T ST type connector at the end of the fiber amplifier. The double cladding
geometry of the doped fiber, however, resulted in severe mislignment of the single mode
 0

core within the ST connector which is designed for standard circular 125 micron diameter •
fibers. The buffer on the Rutgers fiber is made of a hard polymer, iflte the 125 micron
outer cladding of the fiber, which would fit into the connector snugly, is made of a soft
polymer. (see Figure 2) Several attempts were made to remove the buffer without
stripping off the outer cladding, though all were unsuccessful. With the buffer and outer
cladding removed, only the rectangular inner cladding and core are left for insertion into
the connector, making it very difficult to align the core at the center of the connector.
Thus, in order to couple the light out of the fiber amplifier and into another, non-doped,
single mode fiber, the ST connector on the non-doped fiber was placed into a fixed mount.
The connector on the end of the fiber amplifier was then fixed to a Newport three axis S

positioner, and index matching gel was used to provide good coupling between the fibers
as the positioner was used to manually align the two cores. The resulting coupling ratios
were consistently greater than 75 percent, allowing enough throughput power for S 0
experintal data to be taken.
The power coupled out of the fiber amplifier at this point includes the amplified
return signal, broadband spontaneous emission power and any unabsorbed pump power. S

A four nanometer optical bandpass filter, centered at 1.064 microns, is used to eliminate
the excess pump light (at 850 nm), as well as any spontaneous emission power outside of
the 1.064 ± .002 micron wavelength range.
The total optical loss following the fiber amplifier for direct detection effects the
amplified signal and the spontaneous emission. The two losses in the direct detection
system are the air splice loss and the bandpass filter transmission loss, each contributing I
approximately twenty five percent. The total optical transmission of the direct detection
system is thus nloptd = (1-0.25)2 = 0.56. After the optical filter, the fight is then coupled
into a multimode fiber pigtailed to the InGaAs PIN detector package, from which the

_4.
 0

0
output is amplified and measured with a spectrum analyzr. The amplification and 9
detection electronics are detailed in the next chapter. lit

2.2 Heterodyne Detection Systm 0

For the heterodyne detection scheme shown in Figure 4, the Lightwave Electronics
laser provides both the local oscillator power and the power for the transmitted signal. A
half wave plate and polarizing beamsplitter cube are used as a Vanable Rafto Splitter to 0
vary the amount of power split into the local oscillator leg. The local oscillator is then
frequency shifted 200 MHz by a model AQS-2002AI acousto-optic modulator (AOM)
from Intra-Action Corporation. The unshifted zero order beam out of the AOM is 0
blocked with a beam dump to minrmize stray reflections in the system. The frequency
shifted first order is then coupled into a single mode fiber leading to an evanescent wave
coupler, which couples 10 percent of the local oscillator power'into the output fiber 0 0
leading to the photo-detection circuitry.
Laser power not split into the local oscillator path passes through the T/R switch
and is reflected off the target. The somewhat depolarized return 2 5 from the target passes 0
through the T/R. switch again, which directs a linearly polarized portion into the signal leg.
The light reflected into the signal leg passes through a dichroic mirror and is coupled into
the fiber amplifier along with the pump light from the laser diode, as discussed previously. 0
The air splice is once again used to couple the power from the fiber amplifier into the non-
doped single mode fiber leading to the free space fiber to fiber coupler, where the four
nanometer band pass filter is inserted into the beam path. The optically filtered power is 9

now coupled into the fiber leading to the evanescent wave coupler, where ninety percent is
coupled into the output fiber.

....t _...
. .....
.. ...
. .. .. ......•. . .. . . S..
. ... .I . .. .. . . ..,,• , . .... ... ., .0 . , ,. . . . .. ., , . ... 0.
t . .. . .
 0

12
Opcai
NdYA Locad Ptlibizint
Nd.YA Op"ilaor Bnqiil•er
LAWs kola111r So"imr Cube mvn

Tar"e

.
*

Air Fr Wv CoupCr
Si CeuFiler

FIGURE 4: DFeteion Ladar

Dpeterodyne
System. This figure shows

the 1.064 micron ladar in its heterodyne detection configuration. The local 0 0
oscillator splitter is outlined, and the fiber amplifier is shown inserted into
the return signal leg.

The total optical loss following the fiber amplifier for heterodyne detection also 4
affects the amplified signal and the spontaneous emission; The three losses in the
heterodyne detection system are the air splice loss, the bandpass filter transmission loss
and the coupling loss of the evanescent wave coupler. Again, the splice and filter each
contribute a loss of approximately twenty five percent, while the coupler adds a loss of ten
percent to the heterodyne case. The total optical transmission of the heterodyne detection
system after the fiber amplifier is thus rloPt,h = (I-0.25)2 x 0.9 = 0.51. The combined 0

signal, local oscillator and spontaneous emission powers are then coupled into the
multimode fiber pigtailed to the detection electronics, where an intermediate frequency
(IF) signal at 200 MHz is generated during the photo-detection process. For heterodyne S

detection, the detector and electronics are detailed in Chapter 4.

9. 9. ........... .
 0
13

2.3 Target Selection

With the system configurations thus established, the nex issue became determining
the targets necessary to give measurable return signals for each detection scheme.
Dfferet targets are required for the two detection schemes because of the inherent
differences in the detection techniques. During direct detection the return signal is not
mixed with a large local oscillator as it is in the heterodyne case. In order to generate a

measurable return signal, a mirror / glint target is used as the direct detection target. The 4

variable ratio attenuator (VRA) discussed previously (see Figure 1) is then used to vary

the transmitted, and thus the received, optical power. By using the VRA in this fashion,
the strength of the return signal can be adjusted until it is just visible above the noise, with 0

the fiber amplifier not in operation. When the amplifier is turned on then, the resulting

ratio of signal to noise powers yields a direct measurement of SNR improvement for the

direct detection case. 0

For the heterodyne detection case, the large local oscillator power mixes with the

return signal power to create the IF signal, thus allowing a diffuselspeckle target to be
used. The driver for the acousto-optic modulator, however, radiates an electric field at 0

the desired 200 MHz IF signal frequency. As this field tends to be picked up by the

detection circuitry, resulting in a noise spike on the spectrum analyzer which drowns out

the IF signal at 200 MHz, the diffuse target was mounted on a motorized linear translation 4
stage, thus causing a small Doppler shifting of the IF signal away from the 200 MHz noise

spike. More specifically, the moving diffuse target consisted of a piece of flame sprayed

aluminum, tilted at 45 degrees and translated parallel to the beam path by a variable speed 0

micrometer.
In the prelininary preparations to take measrment verifying the results of these

theoretical comparisons, it is important to realize that the neodymium dopant of the fiber 0

has a four level lasing scheme. Therefore, with the laser diode pump light blocked, the

•_
..9_... .... .. . ... ...
. ........ ... .._.. . ...... ...m ... . ..... . ,,p,, , . ... ... .. . . ,... ... t .. .._0 .. .
 0
14

return signal will pass through the fiber unamplified and with no loss due to signal
absorption. This eliminates any need to physically remove the fiber amplifier from the
system to make the signal to noise ratio comparison measurements. The signal-to-noise
ratio equations in the following two chapters are developed with the goal of comparing 4
detection sensitivity of a ladar system without and with a fiber amplifier.

*0

* 4

* 4

*0 0
 0

CHAPTR MI
SIGNAL TO NOISE RATIO THEORY FOR DIRECT DETECTION

The signal to noise ratio (SNR) is defined to be the ratio of the signal power to the noise
power. This section provides a theoretical analysis of the overall post detection electronic
SNR for the ladar testbcd in a direct detection configuration, considering detection both
with and without a fiber amnplifier. The analysis begins with a discussion of the optical
power incident on the detector and follows with a derivation of the signal and noise
voltages after the detection electronics. 'S•

*
3.1 Direct Detection Without the Fiber Amplifier
For a direct detection scheme, the outgoing 9
continuous wave laser power must be 9~L@
modulated so the signal current can be AC coupled into the post detection electronics.
The optical rhopper used to modulate the beam results in a 2 kHz square wave return
signal. Figure 5 shows the electronics between the detector and the spectrum analyzer,
which is used to measure the SNR. The detector is a Lasertron QDFT-250-301 pinFET
detector package, with a multi-mode fiber pigtail. The photodetector is saturated at 220
rnicrowatts of power and has a bandwidth of 250 ME~z. The package includes an
integrated current to voltage preamplifier with a transimpedance of 5.9 kU). The detector
package is terminated with a 975 Ql load'as a precautionary measure, as recommended by
the manufacturer, and is then AC coupled to an electronic amplifier. For direct detection,
the electronic amplifier is an Analog Modules 324A-3-B voltage amplifier with an input
impedanýce of R.. =R. IMlQf, a voltage gain g g. =1000 and an amplification
 16

bandwidth from 200 Hz to 35 MHz. The amplified voltage is then connected to the 50
Ohm input of a Tektronix 495P spectrum analyzer whose resolution bandwidth has been
set to 10Hz.

Tektrorix
Spectrum
Trlmimpsdbnco Load Electronic An Sr
I
N-R m 'A R *.r69 kOhm B 975 C
~~~ gl Itn
V.
-- Ohms C

input -o
Amplifier
gaio input •
Inpedance rimpdance
r I"R ,,0p I "5
amp

Lasertron Detector Package with

Transimpedance Amplifier

FIGURE 5: Detection Electronics. The detection electronics for the 0

ladar system are shown here. The Lasertron detector package includes the
photodetector and a current to voltage amplifier. A safety load of 975
Ohms is used to terminate the detector package. An electronic amplifier is
used to boost the voltage into the spectrum ,nalyzer used to measure the
signal to noise ratio. •

The signal current frolm the detector, i,, at node A on Figure 5 is given by

i, = 91P,[sqr(2,t)] (1)

where 9 =.704 A/W is the responsivity of the detector package, P, is defined to be the
optical return signal power incidenton the detector, and the square wave function (sqr) in
Equation I represents a 50% duty cycle positive square wave function (i.e. values of only
zero and one). The signal voltage V, just after the integrated preamplifier (at node B on
Figure 5) is then

• t• S.... •n . •• J •0

.... .. "*.. ..
.... . " 9''... " "I l 0m
l fi9m n . .. . .. ... .
 0

17
V,=-R...
= •PR.4..[(2,v?)] (2)

where R,.,, - 5.9 kOhm is the transimpedace of the preamplifier. The voltage is then
0
amplified by the Analog Modules amplifier to yield a voltage at node D of

V, -, , 1gV
= g.9tPR..,qr(2*g)] 0

where due to the large input impedance of the Analog Modules amplifier, there is no
appreciable voltage drop across the 975 Ohm series resistance. The signal power, r,,, 0

seen by the spectrum analyzer centered at 2 kHz is then

where R. = 50'11 is the input impedance ofthe spectrum analyzer.

The total noise from the detector consists of shot noise from the optical return
signal power on the detector, dark current noise and thermal noise. The well known
equations for these noise components, given as mean squared noise currents, are 0
(i02)=2eB.(i,), (i'2)=2eI,, and (i,2)= 4kB.T/I.,, respectively, where e is the basic

electronic charge, B. is the electronic bandwidth of the detector, I is the detector dark

current, k is the Boltzmann constant, T is the temperature in degrees Kelvin and

(i,) = 9RP,/2 is the average signai current from Equation 1.26,27,28
The total mean squared detector noise,(i.2), is then
0

0 .. S. . . .. .... .... ... ..

 thus yielding an average detector noise current at node A of 2 6

- 2BI~e+I 4kB4T (6)

Calculating the amplified voltage, V.,.,, at node D yields

4kBT (7)
=gRe91 +2BI

The electrical noise power seen by the spectrum analyzer is thus

2
V

Rm.

Note~, though~, that the noise power terms used in deriving Equation 8 are white noises in
nature, so the power is equally spread across the fil electrical bandwidth B,. The

I spectrum analyzer noise power of Equation 8 is thus displayed as equally divided among
the bandwidth intervals defined by the spectrum analyzer resolution bandwidth, 5v -10
 0

196
Hz. The actual nonse power layal dpayed by the specuum analym, r, is then

obtained by replacing B. with Sw in EquationS8, yildi~ng

r.J = (,dR )2((o)gtP2() v) T)(9)

The noise power into the spectrum analyzer, however, is equal to the detector noises plus

the excess noise added by the Analog Modules electronic amplifier. This noise was

measured by disconnecting the detector package from the amplifier and measuring the
noise level r on the spectrum analyzer due to the amplifler alone. The total noise power, 0

r., is then found by adding this measured value, r. = -57 dBm (2 nW), to the noise in

Equation 9, giving

(* +••z . 0•

r. =r., +r. = .e,)2(( 1) +2e(5VIJ + )+2.o0XIOW. (10)

Using Equations 4 and 10, the direct detection SNR equation without the fiber amplifier in
place is

SNR.,o = r,,
rT,
= 2(11)
(g.R .)2 (e(6v)9• • , +2e(,6v)I, 4+( ) J+ R 1.2. 0x 10- W

A useful measure of detection sensitivity can be obtained by plotting the SNR from
Equation I I as a finction of the return signal power, P,, as shown in Figure 6. A

summary of the variables used in Equation I I is given by Table 1.

.-An
 20

. .

£ S
U.
4,

2,

6.12 6.14 6.16 6.16 6.26 6.2 6•.24 6.2 6.3 6.36 6.22
P hMan "OsiPorerise
msowsm"

FIGURE 6: Direct Detection Signal to Noise Ratio vs. Return Signal

Power for Detection Without the Fiber Preamplifier. This figure shows
the signal to noise ratio as a function of return signal power for direct 0
detection without the fiber amplifier (see Equation 11). The threshold
SNR of 6 is defined to be the SNRL required to discern a signal above the
noise. A return signal power of 0.263 nanowatts is required to reach the
threshold SNRl

This plot is used to determine the return power necessary to achieve a'specified
threshold SNR, defined to be the minimum SNR at which a return signal can be reliably
discerned from the noise. The threshold SNR has been chosen to give a high probability of
detection for normal values of the probability of false alarm. 2 9 From Figure 6, the
mininmm detectable return signal for direct detection without the fiber amplifier for the
chosen threshold SNR of 6 is 0.263 nanowatts. This will be compared to the value for
direct detection with the fiber amplifier, derived in the following section.

• J . . .. L) . . . .... .q .. . . J .. . SCn . . .
 S

21
TABLE 1: List of Variable Valses for use in Sgal to Noise Ratio
Equatiom. This table describes and quantifies the variables used in S
evaluating the signal to noise ratio equations. All variables common to
both the direct and heterodyne detection equations are in the top section,
while the variables distinct to either direct or heterodyne detection follow
separately.
Common Vxaria

v Optical power center frequency (1.064 microns) - 2.82x 10" Hz

8v - Spectrum analyzer resolution bandwidth = 10 Hz
Bo - Equivalent optical passband of the 4 nm filter = L07 x 102 Hz 0
h - Planck's constant = 6.626xI0"' J sec
G - Fiber amplifier power gain = 158
Rtrans - Integrated amplifier trnsmpedace = 5900 f(l
RSA - Spectnum analyzerminput impedance = 50C1
9t - Detector responsivity (Lasertron Detector) = .704 A/Watt
e - Basic electronic charge = 1.6x10"9 C
Id - Detector dark current (Lasertron Detector) f 2.9 nA
k - Boltzmann's constant = .38x10-2 I/K
T - Temperature 298K
Direct DmetetonVaile

Pr - Optical return signal power incident on the detector = 0.4 nWatts

Pse - Spontaneous emission power, 1 lopt,d G h v Bo = 17.5 pWatts
Pse,8v - Incremental spontaneous emission power, n.t,dG h v 8v = 1.7 x 10-'6 W
'nopt,d - Optical efficiency between fiber amplifier and detector = 0.56
N - Number of se-se beat components at 2 kHz = 1.07x 10"
ga - Analog Modules amplifier voltage gain = 1000
Ra - Analog Modules amplifier input impedmance = 1 MCI

Heterodyne Detection Variables

Pr - Optical return signal power incident on the detector = 8 x 10-1' Watts

Pse - Spontaneous emission power, 1opth G h v Bo = 15.8 pWatts
Pse,8v - Incremental spontaneous emission power, nopthG h v 8v = 1.5 x 1 0-'6 W
Pio - Local oscillator power, (near detector saturation power) - 200 gWatts
Tlopt,h - Optical efficiency between fiber amplifier and detector = 0.51
gm - Miteq amplifier voltage gain = 1122
D - Voltage divider effect = 0.048
Rm - Mxteq amplifier input impedance = 50 Q
INh - Number of se-se beat components at 200 Mz = LO5x 10"

• •S 9.. .. .... . ........... •0 ... .. .. m. .. .. . . ._ _

 @

22

3.2 Diret Detecoa With the Fiber Amplier

For direct detection with the fiber amplifier, the return signal is increased by an
optical power an factor G. This gain factor depends on the amount of pump light
absorbed and the length of the fiber. With the 500 mW laser diode pump, measured gains
of 1 dB per meter have been achieved. For the 22 meter length of fiber used in the
experimental work, the 22 dB gain corresponds to a power gain factor of G - 158. Thus
the electrical signal power for direct detection with the fiber amplifier, (given by Equation 0
4 for direct detection without the fiber amplifier), is

r_2 =.." (12)

The noise terms previously dimmed do not change, with the exception of the shot 0 0
noise term. The shot noise is larger because the signal power is now increased by a factor
G and because there are also spontaneous emission photons from the fiber amplifier
incident on the detector. Modifying Equation 10 accordingly, the detector noise plus 0
electronic amplifier noise now becomes

(S..ý.) 2 2e(Sv)9t{{] + P.) + 2e(,Sv)1, + 4k(v)

I2R R. ) +2.0x10-W ,(13)

where P. is the average spontaneous emission power incident on the detector. 0

In general, the average spontaneous emission power is dependent on the number of

contributing free space modes. For a single mode fiber amplifier, the spontaneous
emission power propagating through the fiber is limited to the fraction of emitted photons 0

coupled into the single propagating mode of the fiber. The amount of spontaneous

---- . . ... .... . . .. • .. . ... .. .. + . . ... .. . ++--IO- .. . . ., .... • . .. ..... . • .. . . .. .. . I . . . .. . .. S

O . . ..
 23

mason power emitted into a single mod*, md tsm couena mW the propaating mode, 0

P., =Gh . , (14)

where h is Plmck's constant, v is the nominal spontaneous emsion frequency, and B. is

the optical bandwidth. For this case, v - 2.82 x 1014 is taken to be the bmndpus filter 0
center frequency, corespondaing to a wavelength of 1.064 ^un, and Bo - 1.07 x 1012 Hz
is the bandwidth of the four nanometer optical bmndpas filter. Note that as previously
mentioned, the spontaneous emission from the fiber amplifier must pass through the 0
connector air splice and the bandpas filter, and thus the spontaneous emission power
incident on the detector is given by
•@0

P. = 1..d,.i= q GiO . ? (1S)

where r, is the fifty six percent optical efficiency of the direct detection system after
the fiber amplifier.
For reasons similar to those discussed prior to Equation 9, the optical bandwidth
of the spontaneous emission is broken into small frequency intervals. For convenience, the 0

smallest measurable frequency increment is defined to be the resolution bandwidth of the

spectrum analyzer. The power in a single fequency component is thus obtained by
replacing B, inEquation 15 with the spectmm analyzer resol•ujon bandwidth 5v to yield

P., = qW.Ghv8v (16)

0
 0

24

In addition to the increased shot noise in Equation 13, there me two additional •

noise term rising as a result of spontaneous emission. The first of these tenms arises due
to beating between the spontaneous emission and the return signal, while the second arises
due to the spontaneous emission beating with itseK both of which result from the square 0
law detection of optical radiation. 3 1

The spontaneous emission-return signal beat noise occurs when a spontaneous

emission component beats with the return signal. This noise is maniest at a frequency 0

equal to the separation between the center frequency of the return signal and the frequency
of the spontaneous emission noise components. The spontaneous emission-spontaneous

emission beat noise occurs between two spontaneous emission components of different 0

frequencies. A detailed analysis of the beat noises due to spontaneous emission can be

found in Appendix A of reference [31].

Only the portion of the spontaneous emission-return signal beat noise contributing 0 0
to the SNR at the signal modulation frequency must be considered. This portion of the

beat noise occurs when spontaneous emission components separated by +/- 2 kHz from

the optical firequency of the return signal beat with the return signal. The electric field

incident on the detector from the return signal and these two spontaneous emission
components is 31

R= fPcos(v.t)a, + jFl +2if

cos((a. - 24,),+ 0_,)6, +cos((+. .,)t,+.(17)

O f and Z.f are the random phases

where f - 2 kHz is the signal modulation frequency,
for the respective spontaneous emission components, P,, is the spontaneous emission

power in one 10 Hz frequency component, a, is the linear polarization of the return signal,
4-f and 5,f are the random polarizations of the two spontaneous emission components,

and oo = 2;r.2.82 x 1014 is the angular frequency of the return signal. Recall now that

• ___.O ... •
0_ • . , e • • .. . • gri
 |0

250
the detector curent due to an electric field is equal to the upOMivity of the detector
multiplied by the squ electric fidd vMe twat

When Equation 17 is substituted into Equation 18, two 2 Hiz beat noise current terms
arise from the beating between the spontaneous emission and the return signal. These
terms ame manipulated wing trigonometric identities, resulting in the return signal-
spontaneous emission beat noise current, i,.,,2,

i,_._ = 9j2GP,P.., [cos(27* + _.,)T'. +cos(-2ft+0,)'F] , (19)

where 'FFf and 'Ff are random efficiency terms (i.e. 0 < Ttf :5- 1) arising due to the * *
polarization mixing of the two fields. It is important to note that the signal bandwidth is
assumed to be less than or equal to 10 Hz for this analysis. If the source has a bandwidth
larger than the assumed 10 Hz, the signal power will be divided into a number of 10 Hz 0
components. Each of these signal components will mix with spontaneous emission
components, resulting in a number of beat noise terms at any frequency (i.e. 2 kHz for
direct detection or 200 MHz for heterodyne detection). The total number of return signal-
spontaneous emission beat noise terms at that frequency will be larger, but the magnitude
of each term will be smaller. The sum of these tems results in a total beat noise term
having the same magnitude as given in Equation 19, which assumed a narrowband source. 0
As for the spontaneous emission-spontaneous emission beat noise, the electric
fields due to all spontaneous emission components must be considered, with any pair of
components separated by 2 kHz contributing to the SNR noise terms at 2 kHz. The 0
spontaneous emission electric field is represented by a summation of components, 3 1 each
 0

26

of which is separated by the resolution bandwidth of the spectrum analyzer, 8v =10 Hz, 9
from its nearest neighboring frequency component. That is,

M
'Cat= X1: 2 Pco[(a. + 2 Wv~+ + (D, (20)

where -M and M designate spontaneous emission frequency components at the edges of

the 4 nanometer (B. = 1.07xl10 2 HHz) bandpass filter. To arrive at the corresponding 0

noise current, the squared R. is multiplied by the responsivity of the detector. When the

summation of Equation 20 is squared, the cross terms give rise to the spontaneous

emission - spontaneous emission beat noise current, though only those beat terms affecting

the SNR at 2 kHz are considered. This current is then written as a summation of those

terms beating at 2 kHz, to yield

* 0

where Oj is the random phase of each beat component, 'Fj is a term which takes into
account the polarization mixing efficiency and N is the number of spontaneous emission
beat terms at 2 kHz. The optical bandwidth, Bo =1.07xI012H•, contains 1.07x10" of

the 10 Hz frequency increments, and there are 200 frequency increments in the 2 kHz

band. The total number of components separated by 2 kHz, and thus the number beating
at that frequency, is obtained by N = 1. 07 x 10" - 200 w1.07 x10".

The beat noise currents given by Equations 19 and 21 are then analyzed through
the transimpedance amplifier, the electronic amplifier and the spectrum analyzer, resulting
in 0

..•€
...€ _....
. . . . ... € .. . .. .. .. . . ... ... .. ..9 .. . . .. . . . ... . . . .a . .. .... . .. .. S.
. ..
 0

270
=Sj.t (22)

and

((g.jt.,u.uM)2) (g. 9tP

AS= = N , (23)

for the return signal-spontaneous mission and spontaneous emission-spontaneous

emission beat noise power terms, respectively. 1 0

Adding the beat noise terms from Equations 22"and 23 to the detector noises from

Equation 13 and including the signal power from Equation 12, the signal to noise ratio,
SN,., for direct detection with the fiber amplifier is

Sr,2., SN~~dWW~~~rr +r, =R 2(g.¶RGp~R. ., (24)

2
(gRWW) A. +Rs .2.0x 10-9 W

where

A=A &)
3 2e(v)r(G + P.)
4k( v)T
+2e(av)1,4 + (25)

+gRGP,P., + (P.p) 2 N 0

Figure 7, a plot of SNR vs. return signal power for Equation 24, shows that the

optical signal power required to reach the threshold SNR of 6 is 2.06 picowatts. (The
values used in plotting Equation 23 can be found by referring to Table 1.)

-....
-A
 0

12 28 0
IS

I M
I I IA IM I2 ' Ii 2A
IJI 1* US1 1AS lASlAS
IJI 2.2 2. 2J6I U
P laws Uip PeW Wskmtl)

FIGURE 7: Direct Detection Signal to Noise Ratio vs. Return Signal

Power for Detection With the Fiber Preamplifier. Equation 23 was
used to plot the SNR as a function of return signal power for direct
detection with the fiber amplifier. A return signal power of 2.06 picowatts * *
is required to reach the threshold SNR.

The value found using Figure 7 is 21.0 dB smaller than the power required to reach the
threshold for direct detection without the fiber amplifier, thus showing a significant 0
increase in sensitivity obtained by adding the fiber amplifier to the direct detection
system. 3 2 Next Chapter 4 compares the sensitivity of the heterodyne detection scheme
with and without the fiber amplifier. 0

0
 0

CHAPTER IV
SIGNAL TO NOISE RATIO THEORY FOR HETERODYNE DETECTION

This chapter provides a theoretical analysis of the SNR for heterodyne detection, both
with and without the fiber preamplifier included. As shown, an additional noise term must
be accounted for due to beating effects between the local oscillator and spontaneous
emission fields. As a result of this new noise term, the sensitivity gains when using the
fiber preamplifier in a heterodyne ladar system are not as dramatic as those achieved for
the direct detection case.

4.1 Heterodyne Detection Without the Fiber Amplifier

The detection electronics are again given in Figure 5. The detector package is the
same as used for direct detection, but a higher frequency electronic amplifier must be used S
to ensure amplification of the 200 M&z IF signal. The amplifier chosen is a Miteq AU-
4A-0150 amplifier, with a 50 Ohm input impedance, a power gain of 61 dB and an
amplification bandwidth from I MHz to 500 MHz.
Recall, in a heterodyne detection scheme, a local oscillator is mixed with the return
signal to improve receiver sensitivity. For the test system being considered, the local
oscillator is frequency shifted by an amount A = 200 MI-Iz. Referring back to Figure 4,
the local oscillator and return signal mix at the photodetector, resulting in two DC current
terms and one beat term oscillating at the 200 MHz intermediate frequency (1F). Without
the fiber amplifier in place, the IF current term, i4,, is written as

29

---- - ----
 4
304

i = 9(2iPIPcs(2xr&)) (26)

After the integrated current to voltage preamplifier, the IF signal at node B of Figure 5

becomes

V, =iR.,. = 2%R.J P',P, cos(2rxA) (27) 0

For this detection scheme, the input impedance of the Miteq amplifier is not large enough

to neglect voltage division with respect to the 975 02 safety load resistance. The actual •

voltage seen by the amplifier at node C is thus reduced by a factor D, given as

50.0
D = 9751"=+50 0.048 (28) •

The gain from the Miteq amplifier is a power gain of 61 dB, corresponding to a voltage
gain of g = g. = 1122. The voltage after the amplifier, Vg.., at node D, is thus 6

Vf,•, = g.DVI, = 2 Dg.,D9%.]

VP,• cos(2rAt) , (29)

while the spectrum, analyzer sees the following power level, r.,., centered at 200 MHz

( , =p2(gD
(30)

R14
R0
 4

31

The derivation of the noise power level for the heterodyne detection case without 0

the fiber amplifier is very similar to the derivation used to obtain Equation 10. The noise
power level for heterodyne detection without the amplifier, r,,, is then

(gDR,.)2 j2e(6v)9t(P, + PA,) + 2e(8v)l +4( +)10-4W ,T(31)

]F3+6.3x0 W,()

where the electronic amplifier noise, measured for the Miteq amplifier by the same method
as for the Analog Modules amplifier, was determined to be 6.3 X 10-" W. In comparison
to Equation 10, a shot noise term due to local oscillator power PLO has been added, and a 0

factor of 2 multiplying the total shot noise term has been included, as the outgoing signal
is no longer chopped (see Equations 4 and 5). Also, the voltage divider effect D has been
included. 0

For the heterodyne detection case without the fiber amplifier, it is convenient to
note the limiting noises of the detection scheme. With a measured local oscillator power
0
of 200 microwatts, the various noise terms can be evaluated to show that the local
oscillator shot noise dominates the other noise fctors by almost 10 dB, thus ensuring
local oscillator shot noise limited detection. The noise power, under this limiting case,
then becomes

r.3= (g.DR.) 2R2e(6v)91P,.

14• (32)

Taking the ratio of Equations 31 and 32 yields the signal to noise ratio, SNAW,,.,.,

for heterodyne detection without the fiber amplifier, given as

.,.. u .
...... _ .....m_O.... l 0

9...... O Ii0II ... 9 ..... ..

 O
6
32
o= _ 2(+g.D9R) PP ( •
r.3 (g.DR.) 2 2e(8v)9tP
(

Figure 8 is the plot of SNR vs. return signal power from Equation 33, again using the
parameters found in Table 1. From Figure 8, 1.54 x 10-17 Watts of return signal power
are required to reach the SNR threshold of 6.

Is

1.48 O.16 UJs i11 1.2 1.41

A 1.6 1ie 2.86 2.23 2.40
P blimmlIPowesr(zeIs

FIGURE 8: Heterodyne Detection Signal to Noise Ratio vs. Return 0

Signal Power for Detection Without the Fiber Preamplifier. The plot
shows the signal to noise ratio as a function of return signal power for
heterodyne detection, without the fiber amplifier, using Equation 33. A
return signal power of 1.54 x I0" Watts is required to reach the threshold
SNR. S
 336
4.2 Heterodyme Detectie- With the Fiber Amplfer 0
For heterodyne detection with the fiber mnplifier in place, the return signal is

increased by the power gain factor G. The total electric field incident on the detector is

composed of the amplified return signal, the local oscillator and the sportaneous emission.

The IF signal power, given by Equation 30 for heterodyne detection without the

fiber amplifier, becomes

r =2(g' ')= ,p,. (34)

The detector and electronic noise power level for detection with the fiber amplifier

is similar to Equation 31. However, the return signal power is increased by G in the
heterodyne shot noise term, and the spontaneous emission power adds another component

to the shot noise term. The electronic and detector noise power level is thus

(g.DR. )2f2 e(8v)9l(GP + 1 + P.) + 2e(v)Id + k )T')

rs4 = ' •-6.3 x10 -"4W .(35)

For the heterodyne case, three spontaneous emission beat noise terms must be
added to Equation 35. For heterodyne detection, the spontaneous emission out of the
fiber amplifier passes through the connector air splice, the bandpass filter and the
evanescent wave coupler. The spontaneous emission power is therefore calculated using
7 lopt,h and Equations 15 and 16, giving 15.8 pWatts and1.5 x 10-6 Watts for Pse and
Pse,8v, respectively. The first beat noise term is the return signal-spontaneous emission

beat noise,. given by

4 .. . ... .. .9 . . . . . .... •'9t 0..

 346
r (91g'De'))2 GPP,,JV
=,. Rm, (36)

where there is a factor of two difference between this equation and the return signal-
spontaneous emission beat noise of Equation 22. This factor arises because the signal
power is not chopped for heterodyne detection as it is for direct detection. The second
term is the local oscillator-spontaneous emission beat noise term, which has the same form
0
as the return signal-spontaneous emission beat noise with the local oscillator power, Pl0 ,
substituted for GPr, giving

O(91g.DR~, )2 0
r.,. - •g ) 2 P 1 P,.. (37)

The third term is the spontaneous emission-spontaneous emission beat noise term, given 0
by

(9lg.DR&.,J 2 N,0
4 - .5v (38)

where Nh represents the number of beat noise terms at 200 MHz. This number is 0
calculated similarly to the discussion proceeding Equation 21, where again there are
1.07 x 10" incremental frequency components in Bo and 2.00 x 10' of the components in
the 200 MHz band. The total number of terms beating at 200 MHz is obtained by 0
1
N=1.07x10" -2.00xl0' 1.07×x10". The total beat noise is thus givenas 3 1

r ...
,-= +r"._. +r•_
o ~~~~(9gg.DR.,., 2 Nj(9
,] ( 9)
.,
,P . . + ,
1P 2N'
R= ( g g[2 G
t P.r , +2 P
* 0=

.I .9_ 9 t9 ....... .. _
 - - -

356

The SNR for heterodyne detection with the fiber amplifier is obý from Equati 34,
35 and 39 to yield

SIJ~TMr4 +T..
2(g.Dl..) 2 GpP
e,) (40)

(g.DR..) 2 A. +Ru .6.3 x 10-' 4 W

where

A. --2e()91(GP,+P + P.)
4k(6v)T
R,. (41) *
+29 :GP,
2 P~z + 29t P.
1(u~ N2p

0
Figure 9 shows the SNR vs. return signal power for heterodyne detection with the
fiber amplifier from Equation 40, using the values found in Table 1. The return signal
power required to reach the threshold SNR is 6.0 x 10'" Watts, giving an increase in the
sensitivity of the heterodyne detection system of only 4.0 dB with the fiber amplifier
added. At this point it is impo Want to carefuily examine this result, as the classical result
of using optical amplification in a heterodyne detection scheme is an increase of one over
the quantum egiciency of the detector. 30 ' This chapter's analysis appears to contradict this

result, before a careful examination of the appropriate assmnptions. In a heterodyne

detection system, a fiber amplifier compensates for any losses after the fiber amplifier.
The classical derivation assumes no losses between the optical amplifier and the detector,
so the only sensitivity increase gained in adding a fiber amplifier is due to detector
 0

iii
360
0

:I

4'
0

2, 0

Lis asn 8n am Mm #A anm to am in 1.11

P flVPdlinrWIsWox

FIGURE 9: Heterodyne Detection Sigmal to Noise Ratio vs. Return

Signal Power for Detection With the Fiber Preamplifler. Equation 40
was used to plot the SNR as a function of return signal power for direct
detection with the fiber amplifier. A return signal power of 6.0 x 10" 0
Watts is required to reach the threshold SNR.

inefficiency. For the ladar testbed system used, however, there is an additional optical loss
after the fiber amplifier of -qopth. Including this inefficiency, the sensitivity increase is 0
lOlog(l/( 7)(7..))=4.0dB. If the optical losses are absent, the resulting increase

approaches the classical result. Therefore, very little increase in sensitivity can be
achieved by the addition of a fiber amplifier to an ideal heterodyne detection scheme. 0

Chapter 5 examines the experimental easurements to be taken, and uses the equations
from Chapters 3 and 4 to derive SNR improvement predictions for the experimental work.
0

_ 0
 0
Lm(()

CHAPTER V
PREDICTION OF SIGNAL TO NOISE RATIO INCREASES

The signal to noise ratio equations derived in Chapters 3 and 4 were used to determine the
sensitivity increase for direct detection and heterodyne detection with the fiber amplifier.
The exeiments done, however, will give a measurement of the mpg to noise ratio
increase, not a sensitivity increase. In this chapter, the equations from the sensitivity
analysis are used to predict the SNR increases for ladar detection with a fiber amplifier
included. The various signal and noise terms are quantified. numerically to give the
predicted signal to noise ratios in dB, using the values for all parameters gien previously *
in Table 1.

0
5.1 Direct Detection SNR
From Chapter 3, the signal to noise ratio for direct detection without the fiber
amplifier, SNRdrw/o, expressed in terms of electrical power is given as

,•VR•.., ",•(42) =
="•+r.r +I.. +r.

where rrl is the electrical signal power, rsN1 is the shot noise power, rdark is the dark
current noise power, ltherm is the thermal noise power and Iea is the electronic amplifier
noise power. S

37

J----• • 0nm
I~mmlmmm.t
mam. .mm ml 9..._... ....... i..
 38
The terms in Equation 42 are related to the various system parameters given in 0

Table 1. Theelectncalsignalpowerisgivenby

,(43) 0
r =( 2 ,~9

where Pr is defined to be the optical return signal power incident upon the detector. This 0

optical power is too small to be measured directly by the power meters available, so for

this chapter, it is calculated by setting Equation 43 equal to the e. e tally measured

electrical signal power value of -45.5 dBm (see Figure 10, which we will discuss in more 0

detail later), or 28 nW, and solving for Pr. This gives

(2R=,l,)'
P, = 4 x 10-'0 W ()

Using this value for Pr the following noise terms from Equation 42 can now be evaluated:

S(g0R..) 2 (e(v9~

m= R = 3.13 x 10-'6 W , (45)

F,,,- 'e(v)I -4. 45 x 10-" W , (46)

r, (g.= (4k,(,)T/lR,) = 1.9 x 10-11 W (47)

where e is the basic electron charge, 8v is the resolution bandwidth of the spectrum

analyzer, Id is the detector dark current, k is the Boltzmann constant and T is the

L..... 9.... ... ) . . . . 9).. ... ..... .. ... . ... . . 0

 39
tmperature in Kelvin. Note that the eectronic amplifier noise power of Equation 42 is a

measured value. To perforn this maurent, the electronic amplifier input was

disconnected from the detector packam xW the output was measured on the spectrum
analyzer, resulting in a measured amplifier noise value of

r. = 2.0 x 10o'w (48)

As seen then from Equations 45 - 48, the measured electronic amplifier noise level of

2 x 10- W or -57 dBm, is much larger than the other noise terms, thus allowing Equation

42 to be evaluated as follows S

,.R,w.--2.8 x 10-4 W = 14.0 (11.5 dB) (49)

2x10-9W

Similar to Equation 42, the signal to noise ratio for direct detection with the fiber

amplifier, SNRdirw, is

rT2 +r..,-. r,+rT,.

SMRr.w = 2 (50)
+r.

where rr2 is the signal power, FSN2 is the shot noise power, rr..se is the return signal-

spontaneous emission beat noise power, rnse.se is the spontaneous emission-spontaneous

emission beat noise power and Fea is the electronic amplifier noise power. Specifically,

these terms are

2
T,2 (1/2Xg.RGP,. R)
= S (51)

9 .,. ) _ .. €....... ... .,J..9 .. ...... . .. . t0

 40

( ((= (52)

=@

(gR.2 (9t

= ga~ ~ "S'~,,)2 N (54)

=2.0 x 10-9 W, (55) 0

where the new terms in these equations are the fiber amplifier gain, G, the total
spontaneous emission power, Pse, the incremental spontaneous emission power, Pse,8v, 0

and N, the number of spontaneous emission terms beating at the 2 kHz chopper frequency

within the equivalent bandwidth of the 4 nm optical band pass filter.

Using the previously determined 1 dB/Mn small signal gain of the Rutgers fiber, the

22 mn length of fiber used in the experiment gives approximately 22 dB of gain, orG =

158. Writh this length of fiber and this gain, the spontaneous emission power incident on

the detector after the air splice and optical bandpass filter (from Equation 15) is

P= = ,.,GhvB = 17.5pW (56)

0

Using the optical signal power, the spontaneous emission power and the fiber amplifier

gain, the shot noise power rsN2 is then calculated from Equation 52 to be

...._e
S.... _.. ...
.. . . . ...t. ... ...O. .. ..... .. ... .. i . .. .. . . . . 1,,,, . . . ..O. . .. .... .. . .. . .. .... .9S . ..
 0

41
r,l2 = 2.78 x 10-" W (57)

The incremental spontaneous emission power, from Equation 16, is calculated to be

P5,., = rl 4 ,,~Ghv~v= 1.7 x 10- (58)

rn,r-se is then determined to be

r,,, = 3.61 x 102 W , (59)

while the spontaneous emission-spontaneous emission beat noise term, nse-se, is found

to be

r..,_. = 1.o5 x 10-W (60)

thus giving a total noise level TN for direct detection with the fiber amplifier of 0

rN = TJ2 + r,_.. + rT.. + T. =3.1x 10-9W (61)

This corresponds to a noise level of -55.1 dBm.

The predicted SNR for direct detection with the fiber amplifier is then found by

substituting Equations 11 and 21 into Equation 10, giving

,2 6.9 x 10-4 =222581 (62)

TN 3.1 x 10-9

..t• ..... . . .....

S.... .. ..... ._t. ... .... ... ... .. ...t ..... .. . .. . _t . .. .. ... ... .O .. .... .. . ...... . . .09
 0

42
which corresponds to a SNR of 53.5 dB. The predicted direct detection increase in signal 0

to noise ratio, &d,can then be obtained by subtracting (in dB) the SNR without the fiber

amplifier from the predicted increase with the fiber amplifier, giving

Ad = 53.5dB -11. 5dB (63)

= 42.OdB

5.2 Heterodyne detection SNR

For heterodyne detection without the fiber amplifier, the SNR equation,

SNRhetw/o, is given by

SNR ,,("
tVIle
r,. 3

where r]Ya is the intermediate frequency (IF) electrical signal power and TSN3 is the

local oscillator shot noise power, where we assume that the local oscillator power has

been increased until the LO shot noise term dominates all other noises. From Chapter 4, 0

these terms are given by

= Pi,P (65)
TIF. = RsA
2(g.,D9?R. )2 ,1.(5

and

(g.DR.)2 2e(8v)91P, (6)

-R14

._. • o . ... ._ +._..

O.... _ ._ . .• . .. • ++ ++ + •, .. . 0• _
 43
The IF electrical signal power for heterodyne detection without the fiber amplifier was

measured (see Figure 12) to be -55 dBm, or 3.2 x 10-9 W. This allows the received
optical power for heterodyne detection, Pr, to be calculated from Equation 65, yielding

R=ur, 8.0 x 10-11 W (67)

2P
P1(g 2
1DR.9t)

This value for Pr is used to predict the signal to noise ratio for heterodyne detection with

the fiber amplifier.

Continuing our analysis for heterodyne detection without the fiber amplifier,

however, the local oscillator shot noise power, rSN3, was calculated from Equation 66 to

be

r.W3 = 9.15 x 10-'3 W (68)

corresponding to a noise level of -90.1 dBm. This value is approximately 10 dB larger

than the electronic noise from the Miteq amplifier, which was measured to be 6.4 x 10-`

W, or -101.9 dBm, thus verifying that our heterodyne detection system was indeed local
oscillator shot noise limited. The predicted SNR for heterodyne detection without the
fiber amplifier is therefore

SNR•,,/orwo=3.2 x 10-9 W
rs,
-= 9.9x 10"W = 3238.8 (35.odB) (69)

We now consider the case of heterodyne detection with the fiber amplifier, where

the signal to noise ratio equation, SNRhet,w, is given by

*... 9e.... 0 • 0 0
•
 -- S- r-.

O
44

+ ~,.
SWRA~W r~..
,,,, +(70)

where FIb is the amplified IF signal power, rsN4 is the shot noise term when the fiber

amplifier is included, lnr.re is the return signal-spontaneous emission beat noise power,
rn,lo.se is the local oscillator-spontaneous emission beat noise power, and rn, -s. is the
spontaneous emission-spontaneous emission beat noise power. From Chapter 4, these

terms are as follows:

2
r 2(g=D R.) GP,P1, (71)
RSA

r•, (g.DR..) 2 2e(Jv)9I(P,. + P.) (72) 0

RM

, (g.DR,.)2 29%2GP,P.6,
(73)

r_.
(g"DR•) 2
2
= 2Y. PJP..V (74)

- (g(5DR,)
RS_ 2)

where Nh is the number of spontaneous emission components beating together at the

heterodyne frequency. (See Table 1)

mA
_
 45 4
Using Equation 73, the return sgiml-spoeteous emhuion noise is calculated to
be

r,,, = 7.66 x 10`9 W (76)

Similarly, using Equation 74, the local oscillator-spontaneous emission beat noise is found
to be

r., = 6.06 x 10-" W , (77)

while the spontaneous emission-spontaneous emission beat noise from Equation 75 is

found to be
0

1"-, = 2.44 x 10-12 W (78)

thus giving a total noise level, rN, of 0

rN = +r,_- +r
rT.+, + r.. = 6.4 x 10" W(-71.9dBm). (79)
o

The predicted signal to noise ratio for heterodyne detection with the fiber amplifier
is then found by substituting Equations 71 and 79 into Equation 70, giving

SNR r,__ 5.06 x10-7 =7906(39.0dg) (80)

4r,+ +r..,_. +rr.. 6.4 x10i-1
 46

The predicted SNR inase for heterodyne detection with the fiber aplifier, !,, is then
obtained by subracting (in dB) the SNR for heterodyne detection without the fiber
amplifier from the SNR for heterodyne detection with the fiber amplifiergimng

Ak = 39.OdB- 35.OdB (81)

= 4. dB

In Chapter 6, the experimental data is given and compared to the results of this analysis.

4 0

4 0 4
 4
O

CHAFPER VI
EXPERIMENTAL DATA

In this section we present the SNR data we have taken and make comparisons between
eO
our measurements and the predicted values found in the previous sections. The fr data
taken was for direct detection without the fiber amplifier. Fqure 10 shows the spectrum
analyzer display for this cae, where we note that the average noise level is -57 dBm,
which is equal to the measured value for the electronic amplifier (Analog Modules
amplifier) noise discussed in Chapter 3.

* 00

.16
.20
* .-300
Power (dim) -46
.56
-60
; 70
-90

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.8 2.1

Frequeony (kNi)

Figure 10 Experimental Signal to Noise Ratio Data : Direct Detection

Without the Fiber Preamplifier. This figure shows data taken from the
spectrum analyzer for direct detection without the fiber amplifier turned on.
The direct detection signal is located at 2 kHz. The noise level is -57.0
dBm and the signal level is -45.5 dBm, giving a SNR of 11.5 dB.
47

* 0 d
 0

48 6
This verifies our lmitmg noise assumption for direct detection without the fiber amplifier.
Also, the signal level in Figure 10 is seen to be -45.5 dBm, which is the value used in

Equation 44 to calculate the optical return signal power. The measured SNR is thus
11.5 dB. 0

Figure 11 is a plot of the signal and noise for direct detection with the fiber
amplifier turned on. The electrical signal power level is seen to be -1.1 dBm, while the

noise level is -49.1 dBm, giving an experimental SNR of 48.0 dB.

16

-20
-30
Power (dIm) -40

.10
.70
-10
-90
1 1.2 1.4 1.6 1.3 2 2.2 2.4 2.1 2.81
Frequsecy (kfz)

Figure 11 Experimental Signal to Noise Ratio Data : Direct Detection 0

With the Fiber Preamplifier. This figure shows data taken from the
spectrum analyzer for direct detection with the fiber amplifier. The direct
detection signal is located at 2 kHz. The noise level is -48.6 dBm and the
signal level is -2.2 dBm, giving a SNR of 46.4 dB. 0

The experimental increase in SNR for direct detection with-the fiber amplifier, A., is then

obtained by subtracting (in dB) the SNR without the fiber amplifier from the SNR with the 0

fiber amplifier

*_A_
 49 6
A = 48.0d-11.dB (82)
= 36.5dB

This measured SNR increase is 5.5 dB smaller than the increase of 42.0 predicted in

Section 5.1. In examining the error, it is seen that there are several errors inherent in the
spectrum analyzer used. From the specifications given in the operators manual, errors in
the Display Dynamic Range Accuracy, RF Atenuator Range Accuracy and IF Gain Range 0

Accuracy are ± 2 dB, ±1 dB and ± 2 dB, respectively, giving a possible cumulative error

of± 5 dB. These errors, combined with small, unavoidable experimental uncertainties, can

account for the 5.5 dB difference in the predicted SNR and experimental SNR. 0

Next, Figure 12 shows the Doppler shifted IF signal, located at 199.9982 MHz, for
heterodyne detection without the fiber amplifier.

-30

.20
.4'
Pewur (dim) -00

-70
-10

I1HIM01i4 110.01601 100.00001 119.00761 188.00801

Figure 12 Experimental Signal to Noise Ratio Data : Heterodyne

Detection Without the Fiber Preamplifier. This figure shows data taken
from the spectrum analyzer for heterodyne detection without the fiber
"amplifier. The heterodyne detection signal is located at 199.9982 MHz.
The noise level is -90.4 dBm and the signal level is -55.0 dBm, giving a
S'-,-.1 of 36 dB.

'i "
 S

50

The electrical signal power is seen to be -55 dBm, as was used to calculate the optical
return signal in Equation 67. The noise level is also seen to be -91 dBm, which is
approximately the calculated local oscillator shot noise power described in Section 3.2.
These values give a SNR of 36 dB for heterodyne detection without the fiber amplifier.
Figure 13 is a plot of the signal and noise for heterodyne detection with the fiber
amplifier turned on. The electrical signal power level is -33 dBm, while the noise level is
-77 dBm, giving an experimental SNR of4 dB.

* 0

-10

Fmqnny aW

Figure 13 Experimental Signal to Noise Ratio Data: Heterodyne

Detection- With the Fiber Preamplifier. This figure shows data taken
from the spectrum analyzer for heterodyne detection with the fiber
amplifier. The heterodyne detection signal is located at 199.9982 MRZ.
The noise level is at -77.0 d•m and the signal level is -33.0 d-n, giving a
SNR of 44dB.
The experimental increase in SNR for heterodyne detection with the fiber amplifier, A,, is

obtained by subtracting (in dB) the SNR without the fiber amplifier from the SNR with the
fiber amplifier to yield
fibe ampifie to iel
 51

36d8 (33)
Ah = 44(-

The measured increase in SNR is 4.0 dB larger than the p"cted increase of 4.0 dB. The 0

± 5 dB uncertainty in the display accuracy of the spectrum analyzer is again sufficient to

account for the difference between the predicted and egerimental SNR increases.
In comparing the direct and heterodyne detection results, it is somewhat 0

disconcerting to see pemvental results smaller than predicted for one case and larger

than predicted in the other. This result is caused by the different power levels used by the

two detection schemes. Specifically, the noise and amplified signal levels for the direct 0

detection case are very near the maximum range of the spectrum analyzer, while the

heterodyne detection noise floors are near the minmunm sensitivity level of the spectrum
0
analyzer. These two exremes represent the difference between ± 5 dB in the accuracy,

and it is reasonable to expect the experimental differences of the two cases to have

opposite signs.

,i0
.........Q .• . ......... •
,, .. .. ..J _ ...: . •.... .. ...... .. ....0Jl .... ... ... .. .. 0
 0

CHRAMFR VII
I
CONCLUSIONS AND RECOMMENDATIONS

A neodymium doped optical fiber amplifier has been incorporated into a solid state ladar
test bed to determine its effect on the sensitivity and signal to noise ratio of the system.
The design incorporating the fiber amplifier into the test bed was described in detail, and
some of the difficulties in 'Aimpleenting this were noted. A theoretical model of the

4 electrical signal and noise powers were developed for direct detection and heterodyne
detection. These equations were then used to predict the sensitivity increases achieved by
adding a fiber amplifier to each detection scheme. These same equations were used to
predict the signal to noise ratio increases that would be seen experimentally by adding the
fiber amplifier to the system. Measurements were then taken with the ladar system to
verify the validity of the theoretical model.
For the direct detection case, the measured SNR increase for detection with the
fiber amplifier was 36.5 dB, which is 5.5 d]B less than the predicted increase of 42.0 dB.
For heterodyne detection, the measured SNR increase is 8.0 dB, which is 4.0 dB larger
than the predicted increase of 4.0 dB.
Further wo~rk in this -area is recommended to examine fiber amplifier performance
in true ladar finctions. such as ranging and target detection. This includes examining the
effect of the fiber amplifier on return signals resulting from a pulsed or chirped output.
The research should also expand into eyesafe wavelengths, such as erbium doped fibers at
1.54 microns, which will take advantage of the advanced technology driven by the
communication industry.

52

-A.
 0

APPENDIX A
THE DESCRIPTION OF PHOTON NOISE THROUGH POISSON STATISTICS

This appendix discusses how Poisson statiscsmay be used to describe shot noise in the
photon detectionprocess. 0

In describing the photon detection process, the detector is assumed to be an ideal

photon detector. This assumes there is no current from the detector in the absence of
incident light and there is no noise from detector imperfections or the detector electronics.
The light incident on the detector is laser light, which has constant power and nearly
monochromatic light of frequency v, adding no noise to the detection process. The noise
from the detection process is considered to arise from a randomness in the emission times
of photoelectrons from the detector.
Since photon detection is a discrete physical process, it can be described by
Poisson statistics. For Poisson statistics the probability distribntion function PT(n) is
defined to be the probability n photons will be detected in a sample time interval of T. An
important quantity in the Poisson probability distribution function is (n), the average

number of photon events in T. This quantity is arrived at by taking a large number of time
samples and finding the average n. The Poisson probability distribution function is given
in terms of n and (n) by Equation A. 1 [1I

,(n)=(n"
n!e(X)" •A.1

53
 0

Figure A.I is a plot ofnversus Pr(n) foran (n) given to be 10. It is used to determine
the probability n photons will be detected in any interval T.

13

'4

U 10
12
0
4

0 0.02 0.04 0.06 0.00 0.1 0.12 0.14

P to)

Figure A.1 A plot of the number of photons versus probability

of that number of photons being detected in a time interval T. The
plot is made using Equation A.1 with an <n> of 10.

Now that the statistics of the detection process have been expressed, it is necessary to give
the definition of noise. The quantity of interest is the variance, o&, defined to be the mean
square noise level. The variance of any process can be calculated from Equation A-2 [2]

o; ( n) A.2

Equation A.2 can be simplified to

= (n2)-(2n(n))+((n)2) A.3

= (n 2) -(n)2

For a Poisson process, Equation A.3 can be used to show the variance is equal to
the mean. This derivation, from Boyd, starts with the algebraic identity

.. ..• .. ... . .9
S. _. .
........ . ..... ..... .. . O .... .... .'.... .t .. . . . .. . . . .. ..... . .. _. .
 0

55
0
n =2n+n(n-i).. A.4

Then 0

(n') = (n"+n(n- 1))

= (n) + (n(n - 1))

Using the statistical definition of (n(n- 1)),

(n(n- 1)) = n(n -1)P,(n) A.6 0

Equation A.5 becomes

(W2) = (n) + n(n -I)P,(n)

4-0 A.7
=()+X,,(nl)(e

The n=0 and n- I terms of the summation in Equation A-7 are equal to zero

- ,. 2'= (n).,
(n)0e-(R) A

and the n(n- 1) cancels with part ofthe n! to give

(.-2 (n)!

where (n)2 has been pulled out of the summation. By making the change of variables
m = n- 2, Equation A.9 becomes

,. _IJ .. L. .,,., .. . I)_.. . .. . ... a .. .. .,,, ,,,O ...:...... ,, .I .. ,,

... ........ . . ..• .. .,..... ... • .. . ,,,
 0

56
_mo
M-0W (in)' •

=(n)+(n) 2 XP T (m) A.10

m-0
=(n) +(n)'0'0

Substituting this into Equation A.3,

c9l =(n 2) -(n)'

= (n') + (ni)- (n2). A.11
=(n)

Therefore the variance of a Poisson process is equal to the mean. 1 1]

Equation A. 11 gives the variance in terms of the number of photons detected, but

the SNR is in terms of mean square current. Equation A. 11 can be used to arrive at the
*0

current variance to be used in the SNR equation. The mean squared shot noise current
from the detector is defined as the variance of the current [3]

(1f2).a..2 ((.())2) A.12 0

where i is the current in any time interval, given by the number of photons in the time
interval multiplied by the charge on an electron, e, and divided by the time,

ne
i='-- A.13
T

In Equation A. 12, (1) is the average current flow per time interval for a large sampling,
 57
(n)
= A.14
T

Using Equations A. 11, A. 12, k, 13 and A. 14, the mean square shot noise current is
S
e-0). A.15

3
Since the sample time is equal to one over twice the effective bandwidth,[1, ]

T =-- A.16
2B,'

the mean square shot noise current becomes

,2 = 2eBI. A.17

This analysis has shown the mean square shot noise is proportional to the number

of photons incident on the detector. This result -ees with the physical observation that
the shot noise from a detector increases as the amount of incident light is increased.

0
 O

BIIBLIOGRAPHY

[1] G. Keiser, Q9fca Fiber Communicatio McGraw-Hill, 1991.

[2] C. G. Bachman, Laser Radar Systems and Techniques Artech House, Inc., 1979.

[3] K. D. Moller, QVjg University Science Books, 1988. 0

[4] R. L. Schwiesow and R. F. Calfee, "Atmospheric Refractive Effects on Coherent

Lidar Performance at 10.6 Wm", Applied Optics, Vol. 18, No. 23, December 1979,
pp 3 9 9 1 .
0
[5] V. E. Derr and C. G. Little, "A Comparison of Remote Sensing of the Clear
Atmosphere by Optical, Radio and Acoustic Radar Techniques", Applied Optics, Vol. 9,
No. 9, September 1970, pp 1976.

[6] T. K Lawrence, R. M. Huffaker, F. F. Hall, Jr., and P. A. Mandics, "Considerations •

of a Satellite-Borne Global Wimd Sensing Coherent Lidar", OSA Annual Meeting,
November 1978.

[7] R. L. Schwiesow, R. E. Cupp, M. J. Post and K F. Calfee, "Coherent Differential

Doppler Measurements of Transverse Velocity at a Remote Point", Applied Optics, Vol. 0
16, No. 5, May 1977, pp 1145.

[8] M. J. Post, "Effects of the Earth's Atmosphere on a Spaceborne IR Doppler Wind-

Sensing System", Applied Optics, Vol. 18, No. 15, August 1979, pp 2645.

[9] M. P. McCormick, "Lidar in Space", OSA Optical Remote Sensing of the Atmosphere
Conference, February 1990, 1990 Technical Digest Series Volume 4, pp 67.

[10] F. F. Hall, Jr., "Remote Sensing of Airborne Hydrocarbons and Toxic Polluta,.-
Present Status and Future Projections", OSA Optical Remote Sensing of the Atmosphere •
Conference, February 1990, 1990 Technical Digest Series Volume 4, pp 320.

[11] T. Flom, "Spaceborne Laser Radar", Applied Optics, Vol. 11, No. 2, February 1979,
pp 291.

58
 59

-,Contnulme&-
K=1
[12] P. Hermet, "Design of a Rangefinder for Military Purposes", Applied Optics,
Vol. 11, No. 2, February 1972, pp 2 7 3 .

[13] C. K Cooke, "Automatic Laser Tracking and Ranging System", Applied Optics,
Vol. 11, No. 2, February 1972, pp 277.

[14] M. Elbaum and Paul Diament, "Estimation of Image Centroid, Size and Orientation
with Laser Radar", Applied Optics, Vol. 16, No. 9, September 1977, pp 2433.

[15] S. W. Henderson, C. P. Hale, P. J. M. Suni, and J. R. Magee, "Solid-State Coherent

Laser Radar Technology at 2 jim: Current Status and Future Prospects", OSA Coherent
Laser Radar Technology and Applications Conference, July 1991, 1991 Technical Digest
Series Vol. 12.

[16] P. F. Moulton, J. Harrison, J. H. Flint, and D. M. Rines, "Solid State Lasers for
Coherent Laser Radar", OSA Coherent Laser Radar Technology and Applications
Conference, July 1991, 1991 Technical Digest Series Vol. 12.

[17] B. T. McGuckin and R. T. Menzies, "Efficient High Peak Power Output from
Tunable Diode-pumped Q-switched 2pm TmHo:YLF laser", OSA Coherent Laser Radar
Technology and Applications Conference, July 1991, 1991 Technical Digest Series Vol.
12.

[18] L. N. Durvasula, "I and 2 Micron Laser Radar Transmitter Development," OSA •
Coherent Laser Radar Technology and Applications Conference, July 1991, Technical
Digest Series, Vol. 12.

[19] R. Barbini, "Italian CO 2 Lidar Work and Laser Development," OSA Coherent Laser
Radar Technology and Applications Conference, July 1991, Technical Digest Series, Vol.
12.

[20] S. F. Carter, D. Szebesta, S. T. Davey, R. Wyatt, M C. Brierley and P. W. France,

"Amplification at 1.3 Win in a Pr3 +-doped single-mode fluorozirconate fiber", Electronic
Letters, vol. 27, 1991, pp 628-629. 0

[21] R. Olshansky, "Noise Figure for Erbium Doped Optical Fibre Amplifiers,"
Electronics Letters, Vol. 24, 1988, pp1363-1365.

(22] P. W. France, Optical Fibre Lasers & Amplifiers, Blackie and Son Ltd., Glasgow
and London, 1991.
n 60

S~60
-Continued-

[23] E. Snitzer, "Rare Earth Doped Fiber Lasers," Optical Fiber Communcations '92
Conference TutorialSessions, Tutorial FE, pp 418-484, February 4-7, 1992.

4 (24] I. E. Mlers "Fiber Laser Preamplifier for Laser Radar Detectors," 1991 USAF-
RDL Summer Faculty Research Program Reports, Volume 5.B, Wright Laboratory
Report 26, Wright Patterson Air Force Base, Ohio, July 1991.

[25] A. V. Jelalian, LaserRadar Systems, Artech House, p. 32 Table 1.1, Boston, 1992

[26] G. E. Keiser, OpticalFiber Communications,Second Edition, McGraw-HiUl, Inc., p

244, New York, 1991.

[27] See, for example, E. L. Dereniak and D. G. Crowe, Optical Radation Detectors,
John Wiley & Sons, New York, 1984 0

[28] See, for example, R. W. Boyd, Radiometry and the Detection of OpticalRadiation,
John Wiley & Sons, New York, 1983

[29] W. L. Wolfe and G. J. Zissis, The Infrared Handbook, table 21-3, Infrared 0
Information and Analysis Center at Environmental Research Institute of Michigan,
compiled for the Office of Naval Research, Department of the Navy, revised 1985.

(30] R- H. Kingston, Detection of Optical and Infrared Radiation, chapt. 8, Springer-

Verlag, New York, 1978. 0

(31] N. A. Olsson, "Lightwave Systems with Optical Amplifiers," Journalof Lightwave

Technology, Vol. 7, No. 7, 1071-1082, July1989.

[32] M. S. Salisbury, P. F. McManamon, and B. D. Duncan, "Signal to Noise Ratio •

Improvement of a One Micron Ladar System Incorporating an Optical Fimer Preamplifier,"
OpticalEngineering,Vol. 32, No. 11, November 1993

- _. .q. .......P. ...... .... ... . ...

.. .t .......
.. ... =a .. ....... . ..• .... . .. ... ..a . .... .. ... . .
 -VITA

July 21, 1968 Born - Macomb, ninois

1987 Physics Division Lab siti,

North Central College, Naperville, Ilinois

1988 Co-op Student Employee,

Argonne Natonal Laboratory,
Argonne Tandem Linear Aceerator System,
ATLAS Operations Croup
4
1990 B.A., North Central College, Naperville, Illinois

1990 Research Assistant, University of Dayton

at Wright Patterson AFB, Dayton, Ohio

1992 Junior Electro-Optic;Engineer,

Technology/Scientific Services, Inc.
at Wright Patterson AFB, Dayton, Ohio

1993 M.S., University of Dayton,.Dayton, Ohio

Juy 1,198 on aob

PUBLICATIONS
- Iln

4Optical Fiber huemplifilersfor LadarDetection andAssociatedSNR

ImprovementMeasurements,(To be submitted December 93 Optical Enghee•_)

Signal to Noise Ratio Improvement of a One Micron LadarSystem Inoroating

an OpticalFiber Pre~amplifier,Qj_ "a m teeri•g, November 1993, Authored.

Signal to Noise Ratio Improvement in i.dor Systems Incorporatin Neodymium

DopedOpticalFiberPreamplifiers, roEeadIn of the National Aerospae
E.lectronics Conference (NAECONr), Dayton, Ohio, May 1992, Authored and
Presented.

61
 62

Coherent 1.06 Micron LaserAaiarfor Fiberhpr pfier Research,Ih

International Society for Optical Entiewin LASERS 92 Los Angels,
Caifornia, February 1992, Co-authored.

Coherent 1.06 Micron System for Liquid Cryst Beam Steerer Emzuafton,
Sea QGrou on Active Systems Silver Spring, Mmyland, October 1991
Co-authored.

Neodynmon DopedFiberAmplifier in LidarDetection, ica Socitvf f

America - OPTCON '91. San Jose California, November 1991, Presented.

4 0