vvEPA
United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-79-023
February 1979
Research and Development
Continuous Reading
Lidar Technique for
Measuring Plume
Opacity
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-023
February 1979
CONTINUOUS READING LIDAR TECHNIQUE
FOR MEASURING PLUME OPACITY
by
Dilip G. Saraf
SRI International
Menlo Park, California 94025
Contract 68-02-1291
Project Officer
William D. Conner
Emission Measurement and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
ii
-------
PREFACE
Remote measurement of the density of smoke plume is an important
capability in the field of air pollution research and control. The
single-beam lidar (laser radar) transmission technique has been judged
useful and appropriate for remote measurement of stack-effluent opacity.
The lidar comparison technique of optical-wavelength power backscattered
from "clear-air" regions surrounding a plume is similar to other existing
techniques that use the transmission of visible light to measure opacity.
Lidar developments to date, however, have used pulsed lasers in
large systems, operating at about one pulse per second or slower. These
pulsed lidars have certain drawbacks resulting from the transient and
high-peak-power characteristics common to pulsed radars. A relatively
expensive pulsed laser with high transmitted power is needed to obtain
the necessary signal levels from the backscatter phenomenon. Further,
the powerful pulsed beam presents a potential eye hazard, and precautions
must be taken in the design of the lidar to prevent operators or by-
standers from looking into the beam. Looking into the bean can result
in a damaging dose of light energy to the retina because the laser pulse
is much faster than the protective blink reaction of the eye.
SRI researchers proposed eliminating the transient, high-peak-power
drawbacks of pulsed lidars by applying CW (continuous wave) radar tech-
niques.2'3 Since radar theory states that the information obtained from
a target is a function of average radiated power, a 1 W CW laser could
hypothetically make measurements equivalent to those made by a 50-mW
pulsed laser transmitting one 20-ns pulse per second. A CW lidar holds
promise of greater eye safety, lower cost, and easier operation. A
dearth of applicable published data on CW lidar experiments led to the
conclusion that a laboratory model should be built to determine the
requirements and potential for designing a field-portable CW lidar for
remote measurement of smoke-plume opacity. The EPA made funds available
for the initial work to be performed by SRI beginning in May 1972.
This initial work for the EPA included building a laboratory model
of the CW lidar. The lidar consisted of a 600-mW argon laser radiating
at 514.5 nm, with its beam amplitude modulated at a frequency that varied
linearly from 1 to 10 MHz at a 2 kHz repetition rate. A small portion of
the output of the transmitter was electronically mixed with the detected
target signal collected by a 15.2-cm (6-inch)-diameter front element in
the optical receiver. The beat frequencies generated by the mixer were
filtered to pass only the difference frequencies, and then were displayed
on a spectrum analyzer. The amplitude of the target signals at spacings
2 kHz apart on the spectrum analyzer display is a measure of the light
backscatter from targets along the transmitted laser beam. The location
iii
-------
of each target is uniquely related to the frequency of the signals on the
spectrum analyzer display. Proof-of-principle experiments were conducted
in 1973 by using discrete targets such as plate glass screen mesh; their
opacities were successfully measured by the remote lidar instruments.
Equipment limitations, however, prevented more extensive research with
the breadboard CW lidar.
This report describes more recent work with an improved research
model CW lidar_system. The report also describes lidar development work
on a new high pulse rate lidar technique. The work was performed by SRI
from May 1974 to December 1976.
iv
-------
ABSTRACT
The development of a laser radar (lidar) instrument for remote mea-
surement of the opacity of smoke-stack plumes is described. The work was
conducted within a number of constraints that were placed on the develop-
ment to make the instrument useful for routine enforcement activities.
The constraints required the lidar instrument to be field-portable, eye
safe, relatively low in cost, and simple to operate. Two lidar measure-
ment methods were studied for the instrument: continuous wave (CW) lidar
and high pulse rate lidar.
A research model CW lidar was constructed, and its performance was
evaluated by conducting comparative tests with calibrated semitransparent
targets and plumes from a smoke inspector training school smoke generator.
The results showed that the CW lidar could remotely measure the opacities
of the screen targets or smoke generator plumes to within 3% opacity at a
distance of approximately 80 meters, if the opacities were below 50% and
the measurements were made at night. Environmental light interference
prevented operation of the lidar during daytime.
Proof-of-principle experiments were performed to demonstrate the fea-
sibility of using a high pulse rate lidar to overcome the effects of back-
ground radiation, which limited the use of the CW lidar to nighttime only.
A breadboard lidar was fabricated and evaluated by conducting two series
of comparative tests with the calibrated semitransparent targets. The re-
sults of the first series of tests averaged 66% too low, showed that sys-
tem improvements were necessary, and indicated in particular that the pulse
length of the laser was too long. After modification of the lidar trans-
mitter to reduce the pulse length, it was observed that the loss in power
that accompanied the pulse length reduction prevented measurements of tar-
get opacities by observing clear air scatter signals behind them as with
previous tests and as required for field measurements of plume opacities.
Consequently, for the second series of tests, it was necessary to enhance
the "clear air" scatter signals by moving the targets in close (40 m)
and by placing a small artificial scattering target in the atmosphere
behind them. With this artificial signal enhancement, results showed
that the high pulse rate lidar could remotely measure the opacities of
the screen targets to within 2% opacity during daytime or nighttime
operation.
This report was submitted in fulfillment of Contract No. 68-02-1291
by SRI International (formerly Stanford Research Institute) under the
sponsorship of the U.S. Environmental Protection Agency. The report covers
technical work performed in two phases between May 1974 and December 1976.
-------
CONTENTS
Preface ..... iii
Abstract v
Figures v1i
Tables IX
Acknowledgment x
1. Introduction • 1
2. Conclusions 2
CW lidar . .' 2
High pulse rate lidar 3
3. Recommendations 5
4. CW Lidar Developments . 7
The research model CW lidar 7
Transmitter 10
Laser receiver 12
Signal processing 15
Spectrum analyzers and display ... 16
Field measurements with the research model
I FM-CW lidar 18
Daytime measurements 19
Nighttime measurements 22
Smoke plume measurements 24
5. Rapid Pulse Rate Lidar Developments 28
Pulsed laser selection 28
The high pulse rate lidar design 30
Laser transmitter 30
Receiver 35
Oscilloscope display 35
Boxcar averager ..... 36
Field measurements with the high pulse rate lidar ... 36
First series 37
Second series 38
References 43
vii
-------
FIGURES
Number Page
1 Basic FM-CW radar performance 8
2 Block diagram of the research model CW lidar system 9
3 Measured modulator transmission characteristic 11
4 Laser transmitter 13
5 Optical receiver 14
6 Frequency response of the 16-channel commutating
filter 17
7 Research model CW lidar 20
8 Lidar target range 21
9 Smoke plume generator used with CW lidar 25
10 Extraction of pulsed signals from noise by
coherent integration 31
11 High pulse rate lidar block diagram 32
12 High pulse rate lidar receiver, transmitter, and
signal processor 33
13 High pulse rate lidar, signal processor equipment
on top of laser power supply 34
14 Pulse signal returned from glass target 35
15 Return signal from semitransparent target 40
vm
-------
TABLES
Number Page
1 Discrete Target Opacities Measured by the CW Lidar 23
2 Plume Opacities Measured by the CW Lidar 26
3 Survey of Commercial Rapid-Pulsed Lasers 29
4 First Series: Discrete Target Opacities Measured
by a High Pulse Repetition Frequency Lidar 38
5 Performance Specifications of the Breadboard High
Pulse Rate Frequency Lidar 39
6 Second Seriesj: Measurement of Target Opacities Using
High Pulse Ra±e GaAs Lidar 41
ix
-------
ACKNOWLEDGMENT
The author would like to thank his associates R. C. Gumming,
R. A. Ferguson, M. E. Hird, D. W. Jackson, J. 0. Knotts, J. W. Shaffer,
and C. J. Shoens for their valuable assistance throughout the program.
In addition, the support, guidance, and encouragement from Mr. William
D. Conner of the Emission Measurement and Characterization Division of
ESRL/EPA is greatly appreciated. The author is grateful, too, for the
use of a mobile, calibrated smoke plume generator generously provided
to the project staff by the San Francisco Bay Area Air Pollution Control
District.
x
-------
SECTION 1
INTRODUCTION
This report describes further CW lidar (continuous wave laser radar)
field experiments to demonstrate the feasibility of a CW lidar operating
at 514.5 nm. Earlier work documenting the proof-of-principle experiments
together with the theory of operation of the CW lidar was described in a
previous report.1
The work described in this report was conducted under Contract No.
68-02-1291 with the Environmental Protection Agency (EPA) for twenty-
nine months ending September 1976. This report also presents the analysis
and results of limited field tests to demonstrate the feasibility of a
high pulse rate lidar operating at 900 nm.
During the course of developments of the CW lidar, it became apparent
that the CW technique would be limited to night use for measuring opacity.
Excessive daylight background radiation at the wavelength of interest
(514.5 nm) degrades the signal-to-noise ratio needed to make reliable
measurements. In addition, the transmitter and the receiver may not be
colocated, so receiver alignment poses a problem for making the necessary
daytime backscatter measurements. Because of these limitations, the
remaining effort of the program was directed toward exploring alternate
lidar techniques that use low peak-power, commercially available lasers.
The SRI study and analysis indicated that a commercially available
high pulse rate GaAs laser would be an ideal candidate for a laser trans-
mitter. This report incorporates that SRI analysis of the high pulse
rate GaAs lidar, along with results of the proof-of-principle experiments.
The experimental work reported here provides the results of investigations
into actual system performance, operating characteristics, and remote
transmission measurement capability using components available at reason-
able cost within the state of the art.
-------
SECTION 2
CONCLUSIONS
Two laser radar techniques were investigated in this effort and
the following conclusions may be drawn.
CW LIDAR
A field-portable lidar based on the tested CW lidar principles was
used for verifying its operation in comparison with targets of known
opacities. Glass plates and mesh screens of known opacities were
measured first. The field experiments were concluded with measurements
of smoke plumes of known opacities. These experiments, conducted during
nighttime, indicated an average discrepancy of 3% of opacity at distances
up to 100 m.
The CW lidar was limited to nighttime use. Because of excessive
background noise levels during daytime, the signal-to-noise ratio for
making daytime opacity measurements is considerably degraded when com-
pared with the nighttime performance, despite the use of both a 0.1-nm
spectral filter and a well-designed spatial filter that admitted only
the radiation from the immediate region of interest near the target
smoke plume.
The daytime operation of the CW lidar was further limited by the
difficulty encountered in aligning the receiver along the path of the
transmitter beam for making backscatter measurements. At night, the
continuous transmitter beam is faintly visible to the eye because of
the backscatter along the beam path, which permits receiver alignment.
However, the receiver field of view is restricted to small and specific
regions along the beam, so visual receiver alignment is critical. During
the day, the operator could not see the beam even through a filtered
viewfinder, so the alignment of the receiver along the beam path was
virtually impossible in the presence of the excessive ambient light.
As a nighttime instrument, the CW lidar has the following distinct
advantages over the pulsed lidar:
(1) The CW lidar does not require the safety precautions of the
conventional high-power lidar that provides measurements on
a single pulse, which could cause retinal injury.
(2) Since the lidar operates at 514.5 nm, near the middle of the
visible region of the spectrum, the plume opacity measurements
are likely to be more closely correlated to observe evaluations
than if the lidar operated at a more distant wavelength.
-------
(3) The CW lidar is continuous reading. This is an advantage
over pulsed lidar systems, for measuring real-time variations
in the opacity of a plume or the ambient aerosol background.
Although there are distinct advantages to the CW lidar, its draw-
backs cannot be overlooked. Perhaps in the near future, advances in
laser and filter technology will obviate both the range limitation and
the daytime receiver alignment problem. In this event, an instrument
utilizing the CW lidar technique would offer an enforcement tool for
making plume opacity measurements rapidly and at relatively low cost.
HIGH PULSE RATE LIDAR
Through the use of a high pulse rate GaAs laser, the technique
of coherent integration extracted laser signals that are otherwise
buried in noise, thus making continuous opacity measurements. The appli-
cation of the technique as developed and demonstrated at SRI was neces-
sarily restricted to a very low average power (15 mW) laser having a
pulse width four times as wide and having a beam divergence at least
six times as wide (20 mrad) as that feasible within the present technology.
Coherent integration was demonstrated for specially fabricated semitrans-
parent targets located at 100 m during both daytime and nighttime con-
ditions. Ambient illumination does not matter in the operation of the
lidar because the receiver electronics performance is thermal noise
limited.
The high pulse rate lidar technique offers unique advantages over
either the CW lidar or the high peak power, single pulse lidar. Since
the power output per pulse is low (less than 1 kW), the transmitter
size is small. Opacity measurements can be integrated for a large num-
ber of pulses per second; thus, a time average of the opacity measure-
ments is obtained. Moreover, the measurements are "continuous" in the
sense that the plume characteristics do not change significantly during
the integration period. Operated outside the visible wavelengths (e.g.,
900 nm) and at peak power less than 1 kW, high pulse rate lidar is eye-
safe, a very desirable feature when making measurements near populated
areas. Operation of a high pulse rate lidar at 900 nm can also offer
definite cost advantages because semiconductor lasers are relatively
inexpensive.
In its investigation of feasibility of a high pulse rate lidar,
SRI analyzed the technique and determined relevant engineering require-
ments. A breadboard lidar was fabricated and field tested to demonstrate
its utility. Limited field tests were performed to measure opacities of
semitransparent targets. Severe performance limitations were imposed
by the rather crude breadboard lidar components available, and the high
pulse rate measurements proved inconclusive. Modifications to the sys-
tem to obtain the necessary performance were not successful and data
taken under conditions that would permit extrapolation of breadboard
-------
performance to the likely performance of a field instrument were not
obtained. However, the data indicate feasibility of the high pulse
rate lidar technique to remote measurement of opacity, and indicate
directions for continued development.
-------
SECTION 3
RECOMMENDATIONS
Because of the background-noise-limited performance of the low-power
CW lidar, as well as the practical alignment problems created by the need
to separate the transmitter and receiver, further development of the low-
power CW lidar technique as a candidate for portable opacity measurement
is not recommended.
The low-power CW lidar described in this report has the advantage
of continuous reading capability, but has background limiting performance
which restricts its use to nighttime. A high peak power single pulse
lidar, on the other hand, has both reduced measurement continuity and
the attendant eye-safety problems. There is also a middle ground: a
high pulse rate (continuous pulsing), low average power laser would pro-
vide the necessary opacity measurement with the attendant advantages of
both eye safety and continuous reading capability.
; It is recommended that a research model high pulse rate lidar be
constructed and tested for additional design development purposes. The
high pulse rate lidar utilizing GaAs laser shows greater promise than
the CW lidar as a portable, eye safe, easy-to-use instrument for the
measurement of smoke plume opacity. This lidar has the potential to
provide opacity measurement for enforcement activities. The research
lidar should incorporate the following features, which should obviate
the limitations of the breadboard model used in this effort:
(1) Coaxial optics to facilitate ease of alignment and portability.
(2) A GaAs transmitter with the following features:
• 300-W per pulse output
• 30-ns pulse width (FWHP)
• 3-mrad beam divergence
• Up to 5-kHz pulse rate.
(3) A receiver incorporating a high-sensitivity silicon avalanche
photodiode detector.
These characteristics provide: (1) a high average power (45 mW) for the
transmitter, thus giving the lidar longer operating range (up to 500 m);
(2) better range resolution, because of shorter pulse width; and (3)
reasonably accurate characterization of smaller physical targets, because
-------
of the smaller beam divergence. This last recommendation (number 3)
should be examined further because the lidar operating performance will
now depend upon the ambient illumination (daytime versus nighttime).
Silicon avalanche photodiode detectors have an internal noiseless
gain mechanism that tends to make its performance background noise limited,
with the system constraints defined by the operational lidar. In general,
the sensitivity of a silicon avalanche photodiode detector is about a
factor of 100 better than a silicon photodiode detector (both operating
under "dark" conditions), which gives more detection range for the lidar.
It has been SRI's experience, however, that the silicon avalanche photo-
detectors are not without their own problems. Consequently, a careful
evaluation of a commercially available device may be essential before
incorporating such a device in the system design.
The above mentioned parameters for a high pulse rate lidar are well
within the present state of the art for GaAs laser transmitters. A lidar
fabricated in accordance with these parameters will provide a unique field
portable tool for measurement of smoke plume opacities. The lidar can
then be compared with other measurement methods to obtain research data
on the applicability and design requirements for a high pulse rate lidar
for enforcement activities.
-------
SECTION 4
CW LIDAR DEVELOPMENTS
The design of the research model CW lidar described in this section
is based on the earlier work performed to demonstrate the feasibility of
such a lidar for remote measurement of target opacities. The basic
theoretical block diagram and system performance are first reviewed.
Then the complete system block diagram, together with design criteria
for each component within the block diagram, is discussed to describe
the design of the research model CW lidar.
The basic FM-CW radar system consists of a transmitter that trans-
mits a continuous wave, carrier signal whose frequency is changing with
time.* The receiver, which is colocated with the transmitter, captures
some of the power reflected and scattered back from a target illuminated
by the transmitter. Since the transmitted carrier modulating frequency
is changing linearly with time, at any instant, the frequency of the de-
layed and reflected signal at the receiver input differs from the trans-
mitter output frequency. The difference in the two frequencies is
proportional to the range to the target. If a beat note equal to this
frequency difference is displayed on a CRT screen, its frequency would
be proportional to the distance to the target and its amplitude to the
target reflectivity. A block diagram of such a radar is shown in
Figure 1 along with the target spectrum display.
If the carrier frequency for an FM-CW radar is at an optical fre-
quency, a high frequency (HF) subcarrier can be imposed on it by an
electrooptical modulator. Adding FM to the HF subcarrier transforms the
light beam into the basic FM-CW radar beam at optical frequencies. The
short optical wavelength increases the backscatter target cross-section
so that an adequate signal return is obtained from volumes of clear-air
molecules and ambient particles.
THE RESEARCH MODEL CW LIDAR
A block diagram of the FM-CW lidar is shown in Figure 2. Though
this diagram appears considerably more complex when compared with the
In principal, the carrier can be either the optical wave itself or a
radio frequency sinusoid that modulates the intensity of the optical
wave. The present discussion is general and applies to either method.
-------
;
FREQUENCY
MODULATION
MIXER
(a) BASIC FM-CW RADAR
CONTINUOUS
CARRIER
SOURCE
BEAT
FREQUENCY
SPECTRUM
ANALYZER
o
LU
D
a
<
oc
TRANSMITTED
SIGNAL
/ ^DELAYED
I I ECHO SIGNAL
h-tR—I
TIME
(b)
O
z
LU
a
LU
DC
LU
CQ
I—to—4
TIME
(c)
cc
LU
I
Q.
O
LU
0.
INCREASE IN SPECTRAL POWER
DUE TO INCREASE IN POWER
REFLECTED FROM TARGET —
S = SWEEP RATE =
Afp
At
R
*R =
TARGET RANGE
2RS
^™*" <-
c
BEAT FREQUENCY, fR
(d) TARGET SPECTRUM DISPLAY
Figure 1. Basic FM-CW radar performance
SA--1979-1R
-------
VOLTAGE-
CONTROLLED
GENERATOR
FM SAMPLE
208 V 3 0 I
POWER 1
SUPPLY |
TTTTTT
JjJJjJL
ARGON
LASER
EXPANDER OPTICS
POLARIZER
=£2 OUTPUT BEAMfcE^f-
TRANSMITTER
RECEIVER
F^ RECEIVED LIGHT tr
Figure 2. Block diagram of the research model CW lidar system.
-------
conceptual block diagram of Figure 1, the basic components of the system
are the same. Design considerations for some of the components in the
block diagram in Figure 2 have been described in an earlier report. A
brief description, however, is included here for the sake of completeness.
Also included in the discussion is the choice of design parameters for
the research model CW lidar.
Transmitter
Crystal Clock—
A 2-MHz clock provides a stable time base to run the timing circuits
for both the transmitter and the receiver-processor. The 2-MHz square
wave is down-counted to 2 kHz to generate a linear sawtooth waveform at
that frequency. A 2-kHz sawtooth provides a little less than 500 us of
a linear ramp followed by a fast reset. The 2-kHz clock is also supplied
to the comb filter, which will be described later.
Ramp Generator—
The ramp generator consists of a digital-to-analog converter (DAC),
which gets its input from the divider logic. The divider logic generates
an 8-bit binary digital word that is upcounted and reset to zero every
500 ys. When this word is sent to a DAC, it produces a voltage that is
linearly rising with time. This ramp is highly linear (< 1/2% non-
linearity) .
Voltage Controlled Generator—
The voltage controlled generator (VCG) is a sine-wave oscillator
whose output frequency is proportional to the input voltage. A com-
mercially available function generator such as the Exact Model 165 pro-
vides such a function. The voltage to its input is derived from the
ramp generator described above.
Compensation—
The output sine wave of the VCG is frequency modulated to drive the
electrooptical modulator driver, which in turn drives the modulator. The
frequency characteristics of the driver-modulator combination is such
that at the upper end of the frequency deviation, its amplitude response
begins to roll off. In the present case, the modulated light output
from the modulator is less by about 1/3 dB at 10 MHz, compared to its
response at low frequencies. This loss in the amplitude response of the
modulated light output at higher frequencies creates additional sideband
frequencies on the target display because of amplitude modulation of the
FM waveform. This can be corrected by compensating for the frequency roll
off of the driver-modulator combination. A circuit was designed to
furnish this compensation, but its net effect on the modulated light was
minimal; hence, this circuit was not incorporated in the final system.
Buffer Amplifier—
This amplifier buffers the output of the VCG to drive the modulator-
driver.
10
-------
Modulator Driver—
The modulator driver takes the buffered input from the VGG and pro-
vides a differential signal to drive the two ports of the electrooptical
modulator. The modulator requires a differential signal of 400 V to
amplitude-modulate the laser light. These characteristics are provided
by the Coherent Associates Model 30 modulator driver unit.
Electrooptical Modulator—
The Coherent Associates Model 27 electrooptical modulator, together
with the output polarizer, provides the amplitude modulation for the
laser beam. The adjustment of the bias point to achieve maximum line-
arity and minimum unmodulated output power is provided by the bias control
circuit of the modulator driver. Figure 3 shows the electrooptical
modulator measured applied voltage as a function of the transmission
curve. A bias voltage of 75 V and an excursion of about ±30 V provide
the optimum modulation characteristics.
TOO
X
o
DC
I
o
o
u.
O
z
o
5
01
Z
<
X
8 20 —
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
APPLIED MODULATOR VOLTAGE—V
SA-1979-17
Figure 3. Measured modulator transmission characteristic.
Laser Transmitter—
The laser transmitter is built around a Spectra Physics Model 164
Argon Laser. The laser has a typical output of 1 W of continuous power
at any of 4 wavelengths: 514.5 nm, 488.0 nm, 476.5 nm, and 457.9 nm.
11
-------
Since ,the in-stack transmissometers are being standardized for visible
wavelengths, particularly in the green region of the spectrum, the FM-CW
lidar was operated at 514.5 nm, which is a green line.
The laser beam from the laser cavity is 1.5 nm in diameter
points), and has a beam divergence of 0.5 mrad. The beam is passed
through the electrooptical modulator, described above, and directed by
two mirrors into a beam-expanding telescope. A short focal length nega-
tive lens and the 15-cm f/8 achromat objective lens expand the laser
beam, which reduces the beam's power density to eyesafe levels. A silicon
photodiode detector mounted inside the transmitter housing samples a
small portion of the laser beam to provide the local mixer input signal.
A few percent of the transmitted power is reflected by a beam-sampling
glass plate to provide an adequate signal for the photodiode.
The laser transmitter is mounted on an aluminum plate 147 cm x 41
cm x 0.8 cm that is reinforced with 7.6 cm (3 in.) aluminum channels to
prevent bending of the plate because of the weight of the laser. The
entire transmitter is then mounted on a heavy-duty tripod. The mounting
plate and its dust-cover were painted black to reduce internal reflec-
tions of the laser light. The transmitter weighs 90 kg (199 Ib) and
requires two persons to move it in the field.
The crystal clock, ramp generator, voltage controlled oscillator,
buffer amplifier, and the modulator drive are housed together with the
laser power supply (Spectra Physics Model 265) in a second unit, a
mobile rack that weighs about 70 kg (154 Ib) and can be rolled easily on
its four casters. A 2-m umbilical cord connects the laser transmitter
to the mobile power supply rack and its electronics (see Figure 4) .
Laser Receiver
The laser receiver consists of an astronomical telescope, spectral
and spatial filters, and a photomultipler tube (PMT) (see Figure 5).
The front element of the telescope is a 150-nm diameter f/8 lens. A
field stop located at the focus of this lens is used to control the
receiver's field of view. Since the focal length of the front element
is 1.22 m, a 1-mm change in the diameter of the field stop at this
point changes the circular field of view by 0.8 mrad. The field stop
assembly may be removed to interchange apertures of various shapes,
thereby minimizing detected background noise when viewing a target. A
lens behind the field stop collimates the light and directs it into a
spectral filter. The filter has a very narrow, 0.1 nm, spectral band-
width, and is thermally biased and controlled to maintain a stable
filter center frequency. The filter eliminates most of the spectrally
undesirable background radiation, and represents more than a 10-dB
improvement over the 1973 lidar performance.
A shutter placed behind the filter blocks the light incident on the
photo-detector. This shutter enables the operator to measure the photo-
multiplier tube (PMT) dark current and to protect the PMT from saturation
12
-------
Figure 4. Laser transmitter.
13
-------
Figure 5. Optical receiver.
14
-------
or damage when pointed at a bright target. The receiver incorporates a
reflex viewfinder; the mirror can be snapped in and out of the light
path to provide means for sighting a target through the receiver, with-
out parallax. The eye piece is mounted on the receiver, perpendicular
to the optical axis, to reduce the chance of accidental misalignment of
the receiver during a long series of visual sightings.
The optical components of the receiver are housed in a black anodized
aluminum tube 17.5-cm in diameter and 152-cm long. The receiver is
mounted on a tripod with both aximuth and elevation control cranks. The
PMT and its housing are mounted behind the tube and help to balance the
weight of the receiver about the mounting point on the tripod (see Figure
The PMT (EMI Model 9840A) converts the received backscattered laser
radiation into an analog electrical signal. It consists of a 50-mm
diameter bialkali photocathode having a quantum efficiency of 16% at
514.5 nm. The PMT has 10 stages of amplification, giving an overall gain
of^10b for a supply voltage of about 1200 V. The string of 10 dynodes,
which provides the electron multiplication, is biased at voltages derived
from a linear resistive voltage divider. The first dynode, however, is
biased at a voltage of 300 V derived from a zener diode to stabilize the
multiplication gain at that dynode. Once the first dynode gain is stabil-
ized, the effects of the power supply voltage variations on the overall
gain of the PMT are considerably reduced.
The PMT anode current drives a 50-Q termination to generate a
signal voltage proportional to the radiant power incident on the PMT.
The 50-fi terminating resistance at the electronics portion of the re-
ceiver generates a voltage proportional to the incident radiation at
514.5 nm. A low-noise, low-level video amplifier (Pacific Photometric
Instruments Model 2A44) provides uniform voltage amplification for the
PMT signals up to 5 MHz. The amplifier has a noninverting gain of 100
(40 dB), an equivalent broadband input noise voltage of 40 yv EMS, and
a peak voltage swing of 1 V. Its input and output impedances match 50-Q
coaxial cables
Signal Processing
The electronic package for the receiver also incorporates an
amplifier to boost the signal from the beam-sampling photodiode in the
transmitter. This signal drives the local oscillator (LO) input port
of the double balanced mixer: the RF port of the mixer is driven by the
PMT signal amplified by the 2A44 amplifier, described above.
The double balanced mixer (Relcom Model Ml) generates the beat
frequencies equal to the sum and difference frequencies of the two signals
at its input ports. The mixer is a transformer-coupled bridge of hot
carrier diodes that switch the signal at the RF port in a nonlinear
fashion, generating sum and difference frequency components at the inter-
mediate frequency (IF) port. The device has a dynamic range of nearly
90 dB, with a 6-dB noise figure.
15
-------
The output of the mixer is passed through a low-pass filter to
eliminate the higher sum frequency components at the IF port of the mixer.
The filter has a 5-pole butterworth low-pass configuration to provide
sharp cutoff and good phase response. The output of the low-pass filter
has frequency components from dc to 50 kHz, corresponding to a target
range of zero to 625 m.
Spectrum Analyzers and Display
The output of the low-pass filter consists of bursts of beat
frequency components; the bursts repeat at the frequency of the sawtooth
signal which sweeps the voltage controlled generator. The sawtooth ramp
is repeated at a rate of 1953 Hz, which is the 2-MHz clock rate counted
down by a 10-bit binary counter. There are no beat frequency components
at any frequencies other than multiples of the sawtooth frequency. One
method of extracting the amplitude information occurring at periodic
frequencies without processing noise components at other frequencies is
to use a commutating filter clocked at the sawtooth frequency.
A commutating filter has an input resistance connected to a node
that has a large number of capacitors, which are switched sequentially
to ground through high-speed transistors. The gate, or the base of each.
transistor, is actuated sequentially by a commutator timing network
driven by a stable clock. The commutation frequency is some multiple
of the fundamental resonance-frequency of the filter, depending upon the
number of the capacitors being commutated. For example, for an 8-section
filter, a clock frequency of 15 kHz gives the resonant filter response
frequencies of 2, 4, 6, 8 kHz, and so on. If the commutator is driven
by the same master clock controlling the sweep generator waveform, the
comb filter frequencies would be precisely the frequencies at which the
information bearing beat-frequency components are generated. With such
a filter, the beat frequency components at specific frequencies of
interest could be extracted and processed to yield target characteristics.
Investigation of the performance of a channel commutative filter
driven by a 10 kHz clock disclosed problems that seriously limited its
use in the present application. First, the filter itself is inherently
noisy. The typical response of such a filter, with no input to the
filter is shown in Figure 6. All of the output frequency components are
noise, since there is no input to the filter. Most of this noise comes
from the switching transients. If the noise components had been con-
stant in amplitude, the processor could simply have subtracted their
known values. Thus if a sinusoidal data input is applied at 2 kHz, the
amplitude of the. measured component at 2 kHz would increase above the
noise to reflect the input signal, with the amplitude of other components
unchanged. However, because of the time variant nature of the original
internal noise component at that frequency, the amplitude of the data
signal could not be extracted from the measured spectral component ampli-
tude. Accurate amplitude comparisons are absolutely required to measure
opacity, so the noisy commutative filter technique was not used in the
final system.
16
-------
LI
RELATIVE RESPONSE
p-
OP
C
fB
3
O
i-!
fD
CO
T3
O
3
en
fD
O
Hi
I
r>
3
3
fD
C
ri-
ps-
3
TO
t-h
fD
i-i
O>
00
iis
O)
00
t
••H^HI-
•*-^HH
"H^Hh-
•o
C
o
O
I
-I
-------
An alternate method for selective measurement of signal amplitude
at specific frequencies is to use a scanning filter of the type used in
laboratory spectrum analyzers. Recall that the output of the low pass
filter consists of signal components at discrete frequencies approxi-
mately 2 kHz apart. The measurement of target opacity requires an
accurate comparison of the amplitude of the backscattered signal from
clear volume behind the plume, first with and then without the plume in
the path. The ratio of the two amplitudes at the frequency corresponding
to the location of clear air volume behind the plume yields the opacity
of the plume (opacity = 1 - transmission). To make this measurement in
the field, the receiver is aimed at the plume with the modulated laser
beam shining through it. The receiver field of view is restricted so
that only the backscattered radiation from clear air volumes falls on
the detector. Moreover, to avoid the separate and important problem of
spatial integration,* the field of view must include no more than 15 m of
path behind the plume (corresponding to subcarrier frequency deviation
of 10 MHz).
To make the measurement of the plume opacity, the frequency of the
spectral component corresponding to the range of the clear air volume
behind the plume is first identified. The spectrum analyzer scanning
filter is then locked to this frequency and the spectrum analyzer output
is displayed on a conventional A-scan display. (The slowly-changing
voltage corresponding to this range cell's signal level is available
from the spectrum analyzer backpanel.) The temporal variation of the
backscattered signal from the clear air volumes then appears. An average
value of this signal voltage is stored at the input of one port of the
analog divider. The laser beam is next redirected to one side of the
smoke plume and the receiver is aligned once again to view that portion
of the beam at a range equal to that behind the smoke plume. If the
spectrum analyzer is left undisturbed, the amplitude of the signal and
the dc output voltage represent the backscatter signal from clear air
volume without attenuation by the plume. The ratio of these two voltages
appears at the output of the analog divider, and is in fact the one-way
transmission through the plume. This ratio is displayed on a digital
readout.
FIELD EXPERIMENTS WITH THE RESEARCH MODEL FM-CW LIDAR
The objective of the experimental program was to determine the
effectiveness of the research model FM-CW lidar in measuring the opacities
of targets with known characteristics. In the first set of outdoor field
tests, the FM-CW lidar was used to measure the opacity of semitransparent
targets such as glass plates and mesh screens of different densities.
These tests were conducted during both day and night under a variety of
ambient conditions, such as meteorological visibility, temperature, and
Page 53, Reference 1.
18
-------
wind, to assess the basic FM-CW technique as implemented in the research
model. The final series of tests compared the performance of the labora-
tory model lidar to that of a transmissometer mounted in the stack of a
calibrated smoke generator on loan from an inspector training school.
The details of these experimental program activities and their results
are included in the following sections.
Daytime Measurements
The research model FM-CW lidar shown in Figure 7 was first used to
measure the opacity of discrete targets, by using a clear air volume
behind the targets as references. The measurement of opacities of dis-
crete targets using the earlier laboratory model CW lidar with discrete
targets instead of clear air volumes as references has been already de-
scribed in Reference 1 (pp. 71, 72). The objective of the present tests
was to demonstrate that the refined research model lidar could indeed
measure target opacity by using clear air volumes alone as a reference.
This performance more closely resembles the actual operation intended
for the CW lidar as an enforcement tool.
The targets themselves included plate glass mounted on a frame 50 x
50 cm, as well as several densities of mesh screens that were placed over
the framed glass to increase the opacity of the target. The geometry of
the target range is shown in Figure 8.
The glass frame was clamped on top of a step ladder so that the
experiments could be performed above the dust and traffic. Furthermore,
with both transmitter and target about 2 m above the ground, the trans-
mitter beam was parallel to ground, facilitating receiver alignment.
After these experiments, the targets were located at elevated heights in
system geometries more closely resembling those encountered in actual
measurements of plumes from industrial smokestacks.
Attempts to measure opacities of discrete targets during the daytime
were vitiated by the effects of excessive background radiance, which
severely degraded the signal-to-noise ratio at the output of the PMT.
The optical receiver had been designed to transmit within 0.1 nm around
the wavelength of interest (514.5 nm); the field of view was limited to
about 2 mrad to cover the regions of interest behind the target. In
spite of these precautions, the effects of background radiance in de-
grading the system signal-to-noise ratio could not be overcome. Prelimi-
nary calculations indicated that daytime operation may be feasible;
however, the background radiance can easily change by about two orders of
magnitude in a certain spectral window depending on a variety of condi-
tions such as the nature of the sky (clear, cloudy), and other objects
within the receiver field of view (e.g., buildings, green trees, and
specular objects). The excessive background radiance forces the operation
of the PMT at very low voltages to limit its maximum dc current to under
100 uA. Because of the background radiance, the gain also had to be
turned down to 5 x 10^ with about a 2-mrad field-of-view aperture. At
these gain levels, the thermal noise of the preamplifier that follows the
19
-------
IS5
O
Figure 7. Research model CW lidar.
-------
LIDAR !
SITE I
••:•'•.:'•: BLDG. 320
GLASS j
PLATE |
TARGET I
PLAN VIEW
;vW':' 8LDG
TARGETj
Figure 8. Lidar target range.
-------
PMT dominates the optical background noise and causes the noise spectral
density at the output of the PMT to increase considerably. The shot
noise from the PMT and the thermal noise from the preamplifier are equal
when the PMT gain is 1.5 x 105; below this gain, the amplifier thermal
noise predominates. No useful purpose is served by merely increasing
the preamplifier gain at this point. The only parameters to manipulate
are either to reduce the effects of background by further narrowing the
width of the spectral filter as well as spatial filter, or to increase
the transmitter output power. The receiver aperture (diameter of the
front element) can also be increased. However, the effects of these
parameters on signal-to-noise ratio improvement would be marginal, and
they include possibly jeopardizing system portability, design simplicity,
eye safety, or ease of operation in the field.
Nighttime Measurements ,
The laboratory model FM-CW lidar measured opacities of several dis-
crete targets. All the measurements were performed at night to minimize
the effects of background radiation. The opacities of the targets were
calculated from the spectrum analyzer display, by taking the ratio of
the amplitudes of the signals corresponding to the location of the clear
air volumes behind the target, both with and without the target in the
path of the beam. Table 1 shows the results of the measurements.
The readings indicated in Table 1 were obtained by measuring the out-
put of the HP Model 3581 audio frequency spectrum analyzer on a digital
.voltmeter. The spectral component corresponding to the location of the
'reference target (clear air volumes) was determined both from the spectrum
analyzer display as well as from the knowledge of the sweep rate and the
range to the target according to the relationship f = 2SR/C, where S is
the sweep rate (Hz/s), R is the range to the target (m), and C is the
velocity of light (m/s). In all cases, these two frequencies were in
very good agreement with each other (always to within 1% of each other).
The HP 3581 spectrum analyzer can be manually tuned to select any specific
frequency component. The temporal variation of the frequency component
can then be displayed on the spectrum analyzer. The Y axis output port
of the spectrum analyzer indicates a voltage equal to the amplitude of
the displayed signal. The temporal variations of the signal can then be
observed on a voltmeter. The readings indicated in Table 1 are the
voltages measured on a digital voltmeter (DVM). A 1-s time constant
filter was inserted between the output port of the spectrum analyzer and
the DVM to minimize the effects of noise on the DVM readings.
The results obtained in the field tests are in agreement with the
measurements performed in the laboratory. The measurements do not agree
in the case where the target consists of a glass plate to which 2 each
No. 2 mesh screens and 1 No. 16 mesh screens were added. The opacity of
this combination of targets proved to be 91% when measured in the labora-
tory. However, the lidar measurements give an aggregate opacity of only
67.3%. This can be explained in terms of the degradation of the S/N
ratio as the denser target attenuates light returning from behind the
target. This weak-signal limitation is obviated by placing the lidar
22
-------
TABLE 1. DISCRETE TARGET OPACITIES MEASURED BY THE CW LIDAR
Number
1
2
3
4
5
6
7
Target
Clear air volumes
behind target
Clear plate glass
Two glass plates
Plate glass + No.
16 mesh screen
Plate glass + No.
2 mesh screen
Plate glass + 2
each No. 2 mesh
screen
Plate glass + 2
each No. 2 mesh +
No. 16 mesh screen
Amplitude of Return Signal
With Target
(units) "A"
N.A.
50
45
27
39
28
18
Without Target
(units) "B"
55
55
55
55
55
55
55
Difference
Units
B - A
N/A
5
9
28
16
2-7
37
Ratio
A/B
0.909
0.818
0.490
0.709
0.509
0.327
Measured Opacity
(percent)
Lidar
N/A
9.1
18.2
51.0
29.1
49.1
67.3
Laboratory
9.45
18.90
52.50
30.10
49.60
91.00
GJ
-------
closer to the target (100 m). In reality, high accuracy for target
opacities in excess of 50% is not required because opacities in excess
of 50% exceed all opacity emission standards.
Smoke Plume Measurements
After the FM-CW lidar had successfully measured the opacity of cali-
brated discrete targets as indicated in Table 1, its performance against
calibrated smoke plumes with various opacities was quantitatively assessed.
A smoke generator was leased from the Bay Area Air Pollution Control
District (BAAPCD). The characteristics of the smoke generator were
accurately known. The opacities of the smoke plumes generated by the
smoke generator could be varied at will. The actual opacities were con-
tinuously measured with an in-stack transmissometer built into the smoke
generator, which provided a continuous record of the smoke characteristics
as they varied during the experiments.
The smoke generator was parked 80 m from the lidar, and all the
measurements were performed at night (as before). A photograph of the
actual setup is shown in Figure 9. The results of the test are shown in
Table 2. The results indicated in Table 2 once again show good corre-
lation between the smoke plume opacity measurements performed by the FM-CW
lidar and those obtained directly from the instack transmissometer built
into the smoke generator. For thinner plumes with up to" 50% opacity, the
maximum disagreement was less than 3%. At higher opacities, the error
increases. For example, when the smoke plume opacity was registered at
70%, the FM-CW lidar measured the opacity at 64%. This again can be
explained in terms of the degradation in the signal-to-noise ratio because
of the great attenuation of the laser light by the denser plume.
Only two readings were obtained on black smoke generated by the
benzene-fired generator. The generator had mechanical difficulties at
first, and when they were resolved, the prevailing winds caused measure-
ment problems. The winds started to blow the smoke into the "clear air"
volume behind the smoke generator, causing the amplitude of the spectral
line at 125 m to jump significantly. This situation yielded meaningless
opacity measurements (negative values for opacity). No additional attempts
were made to gather data on black smoke plumes generated by the smoke
generator since there was adequate prior data on the performance of the
lidar on both white and black smoke plumes.
One drawback of the FM-CW lidar is its limited usefulness in the
daytime. The sunlit surroundings during the day generate enough
background-induced shotnoise in the optical receiver that the signal-to-
noise ratio degrades, causing unreliable opacity measurements. This is
because the spectral radiance of the sunlit surroundings at the wave-
length of interest (514.5 nm) is so high that even a 01.-nm spectral
filter does not provide adequate discrimination. Another difficulty,
which also relates to the high background radiance, arises from the
operating set up of the FM-CW lidar. The transmitter and the receiver
are not colocated, but view the beam at an angle to minimize spatial
24
-------
Figure 9. Smoke plume generator used with CW lidar.
25
-------
TABLE 2. PLUME OPACITIES MEASURED BY THE CW LIDAR
Number
1
2
3
4
5
6
7
Target
Clear air
White smoke
White smoke
White smoke
White smoke
Black smoke
Black smoke
Amplitude of Return Signal
With Target
(units) "A"
—
29
23
19
14
28
23
Without Target
(units) "B"
39
39
39
39
39
39
39
Difference
Units
A - B
10
16
20
25
11
16
Ratio
A/B
0.744
0.589
0.487
0.359
0.718
0.589
Measured Opacity
(percent)
Lidar
26
41
51
64
28
41
Laboratory*
29
40
50
70
30
—
NJ h
Date: 7-29-75
Time: 2030 to 0030 hr
Temperature: 21°C
Humidity: 30%
Visibility: 26 km
Winds: Ousting up to 30 km/hr
Target Range: 100 m
Frequency Deviation: 1.01 to 6.85 MHz
Laboratory opacities are as indicated by the transmissometer of the smoke generator.
-------
integration as described on page 53 of Reference 1. In short, the best
signal-to-noise ratio is obtained by the receiver when its field of view
intercepts only a small segment of the modulated laser beam path in the
vicinity of the target plume area. This is achieved by separating the
receiver and the transmitter by a few meters. This requirement, however,
imposes another restriction on the system setup procedure: the operator
attempting to intercept the laser beam path should visually be able to
sight the beam through the receiver optics. The beam path is faintly
visible when the surroundings are dark. Certain amounts of dust particles
and aerosols are always present in the atmosphere, and they scatter
enough light along the beam path to make this possible. In a bright
surrounding, however, they are difficult to discern despite a narrow
spectral filter.
27
-------
SECTION 5
RAPID PULSE RATE LIDAR DEVELOPMENTS
The difficulties of operating the FM-CW lidar during the daytime
reduced its utility as an enforcement tool for monitoring violations of
emission standards to being effective only for short-range nighttime
observations. This limitation was considered serious enough by the EPA
to warrant exploration of alternate lidar techniques to measure smoke
plume opacity without having to compromise any of the basic design fea-
tures that provided the impetus for the development of the FM-CW lidar:
eye safety, portability, potential low cost (in quantities), ease of
operation, and continuous reading. Any further development, however,^
would also need to provide both day and nighttime measurement capability.
PULSED LASER SELECTION
Eye safety and equipment portability immediately require the use of
a low peak power laser (a few hundred watts per pulse, depending on the
laser) for this application. Commercially available lasers were surveyed
to evaluate the most promising candidate laser for the application. The
following guidelines helped to narrow the field:
(1) Wavelength in the visible or near visible region of the
spectrum.
(2) Efficient to operate without the requirements of multiphase
power circuits and elaborate cooling.
(3) High pulse rate to provide pulse-to-pulse integration
capability.
(4) Small, portable, and easy to operate.
(5) Long operating life.
(6) Rugged.
(7) Low cost in quantity.
Table 3 is a summary of some of the various commercially available lasers
surveyed during the course of this study.
The laser that meets most of the above requirements is a commercially
available semiconductor (such as Gallium Arsenide, GaAs) laser. Perhaps
the only potential weakness is the wavelength of the emitted radiation,
which is in the near infrared region of the spectrum (900 nm). Compared
with the competing candidate lasers, the inherent advantages of the GaAs
laser far outweigh the limitation imposed by its wavelength.
28
-------
TABLE 3. SURVEY OF COMMERCIAL RAPID-PULSED LASERS
1
Laser Type
Gas Lasers
Argon
(GW)
\ s
Helium Neon
(CW)
Carbon dioxide
Carbon Dioxide
Carbon Monoxide
Nitrogen
Solid-State Lasers
Neodytnium Yag
(CW)
Neodymium Yag
(pulsed)
Neodymium Yag
Tunable lasers
(dye lasers)
Semiconductor
lasers
Wavelength
(nm)
351 to 514
633
10600
(nominal)
10600
(nominal)
5000 to 6000
337
1060
1060
530
410 to 800
900
(nominal)
Typical
Output Power
(W)
To 15
10-1
3 to 700
To 4 X 109
8 X 103
106
To 1000
To 107
106
To 106
103
(nominal)
Comments
Low efficience, relatively
bulky, medium price range
($20K)
Output power too low
Far IR, relatively bulky,
higher price for higher
outputs ($40 K) , no suit-
able detector
Relatively bulky, high price
($90 K) , no suitable de-
tector, PRF too low (1/60
PPS)
Pulsewidth too long (1 s) ,
no suitable detector
Up to 100 pps, medium price,
range ($23K) , relatively
bulky
Relatively expensive ($50K)
Medium price range (to
A O (\V7 \
$3 OK)
Medium price range (to j
$50K) , relatively low PRF
(30 pps) , relatively bulky
Relatively bulky, low PRF
(1 to 10 pps) , relatively
expensive ($50K)
(Small size, good PRF (to
1000 pps) , small pusewidth
(50 ns) , moderate price
($10K); (a lidar at this
wavelength gives the most
efficient detector)
29
-------
THE HIGH PULSE RATE LIDAR DESIGN
The high pulse rate lidar is based on the well-known radar theory
that target detection depends upon average power, not peak power. This
concept can be applied to optical radars and, in principle, the same mea-
surements made with a high-peak power laser system can be made with a
high pulse rate laser of the same average power. For example, a pulsed
ruby lidar operating at 1 Joule per pulse (30 ns) and one pulse per
second provides the same signal-to-noise ratio on a single pulse for a
target return as a 300 watt per pulse (30 ns) lidar operating at 5 kHz
during its 20 s of operation. The signal-to-noise ratio enhancement
is obtained by integrating a large number of temporally coherent pulses.
Since noise in an electrical system is random, the signal-to-noise ratio
increases linearly with the number of pulses integrated.
The lidar return signal from clear-air backscatter behind a plume
can stand well above noise, if the power transmitted per pulse is pro-
gressively reduced, the signal-to-noise ratio will proportionally dimin-
ish for returns behind the target. If as the power per pulse is reduced,
the pulse rate is simultaneously increased and the receiver is built to
perform coherent pulse-to-pulse integration on each pulse, the signal-to-
noise ratio at the receiver will remain the same for the two cases as
long as the average power remains the same. Thus, signal amplitudes may
be measured even when the signal itself is buried in noise.
As an example of enhancing the signal-to-noise ratio by coherent
integration, Figure 10 displays a pulse from a commercially available
pulse generator set up for a 600-ns wide rectangular pulse. The pulse
rate is 1 kHz. The top trace in the photograph (Figure lOa) shows the
peak value of the pulse height as measured by a commercially available
Boxcar Averager (Princeton Applied Research Model 162). The boxcar
device averages the signal amplitude at a desired temporal location with
respect to an external trigger, and outputs a dc voltage level equal to
the average signal amplitude (top trace). A dramatic representation of
the true signal-averaging capability of the boxcar averager is shown in
Figure lOb. This figure shows the same pulse from the pulse generator
as before in the presence of external noise from a white noise source.
The boxcar again averages the composite signal and outputs the same dc
voltage level as if the noise were not present. This signal-averaging
capability is the basis for the high pulse rate lidar technique. A block
diagram of such a lidar is shown in Figure 11, and photos of the lidar
are shown in Figures 12 and 13. A brief description of each block in
the figure is presented below.
Laser Transmitter
To meet the requirements of high pulse rate and low peak power
per pulse (from an eye safety standpoint) a transmitter built around a
GaAs semiconductor laser is most suitable. In addition to high pulse
rate and low peak power per pulse, the GaAs transmitters are compact,
30
-------
Figure 10. Extraction of pulsed signals from noise by the
technique of coherent integration.
31
-------
GaAs TRANSMITTER
300 W/PULSE
1-5 kHz
RECEIVER
OSCILLOSCOPE
BOXCAR
AVERAGER
SMOKE PLUME
Figure 11. High pulse rate lidar block diagram.
32
-------
(a) RECEIVER, TRANSMITTER, AND SIGNAL PROCESSOR (R to L)
Figure 12. High pulse rate lidar receiver, transmitter,
and signal,,, processor (R to L) . ^
-------
Figure 13. High pulse rate lidar signal processor
equipment atop laser power supply.
-------
reliable, efficient, and low cost (under $10,000). Because of the trans-
mitter wavelength (900 nm), the effect of background radiation on the
lidar system performance is less deterimental than at visible wavelengths.
Receiver
The lidar receiver uses a Fresnel lens (30 cm) as a collecting
element. A spectral bandpass filter rejects radiation outside the 900-nm
passband. A silicon photodiode detector placed behind the filter effi-
ciently converts the incident radiation into electrical signals because
the sensitivity of a typical silicon photodiode peaks at about 900 nm.
Since the photodiode has no internal gain mechanism (such as in an ava-
lanche photodiode detector or a PMT), the receiver performance is
electronic-noise limited—even in direct sunlight. This indicates that
the daytime operation of the receiver will not degrade system performance.
Oscilloscope Display
The oscilloscope shown in the block diagram of Figure 11 displays
the lidar return. With a dual trace oscilloscope, the lidar return is
displayed on one trace, and the sampling gate of the boxcar averager
is positioned with respect to the strong signal from the target (such
as a smoke plume), as shown in Figure 14. The boxcar averager gate can
be positioned at will on the display. The boxcar averager makes the
signal amplitude measurements (even when the signal is buried in noise)
at the temporal location of the sampling gate.
Figure 14. Pulse signal returned from glass target.
35
-------
Boxcar Averager
A boxcar averager is a device that averages a repetitive wave form
by examining it one piece at a time. One way to understand boxcar
averagers is to compare them to synchronized strobe lighting. At pre-
cisely the proper moment in time, the strobe lamp fires, illuminating
a rotating object for a split second and making it appear to freeze in
its motion. The boxcar averager operates in the same manner. A repeti-
tive waveform and a synchronous trigger are presented to the boxcar
averager. At precisely the selected moment, an electronic gate opens
for a very short selected time period and then closes. The balance of
the waveform is ignored. Since many repetitions of the waveform are
sampled, the output of the boxcar averager will be proportional to the
average level of the input signal during sampling. However, the noise
that accompanies the waveform is attenuated since the average value of
random noise is zero. If more than one point on the waveform is to be
examined, or if the entire waveform is of interest, the gate is scanned
across the entire waveform.
Commercial boxcar averager offer a range of sampling gatewidths
with the minimum width as small as 100 ps. Selection of the optimum
sampling time for a given application is determined by the trade-off
between the signal-to-noise ratio improvement required and the time
inefficiency that can be tolerated. The narrower the gatewidth, the
greater the resolution, but also the greater the number of repetitions
required for a given output signal-to-noise ratio.
The output of the boxcar averager is a dc voltage that is propor-
tional to the signal being measured. This voltage is displayed on the
digital indicator on the boxcar averager front panel.
FIELD MEASUREMENTS WITH THE HIGH PULSE RATE LIDAR
To make opacity measurements using the high PRF lidar, the follow-
ing procedure was followed:
(1) Direct the transmitter at the smoke plume, and align the
receiver (if necessary) so the receiver field of view covers
the smoke plume.
(2) Start pulsing the transmitter. The oscilloscope should dis-
play the lidar return as shown in Figure 14.
(3) Position the sampling gate just behind the temporal location
of the plume return on the oscilloscope. Set the gate width
to 50 ns. This means that the boxcar is averaging an 8-m
path behind the plume to provide the backscatter signal from
clear air volumes behind the smoke plume.
(4) Swing the transmitter and the receiver away from the plume,
and, leaving the position of the sampling gate unchanged,
make the backscatter measurements from clear air "target."
36
-------
This gives the amplitude of the backscattered signals from
clear air volumes in space behind the plume as if the plume
did not exist. The ratio of these two measurements is the
two-way transmission through the plume from which plume opacity
can be easily derived and displayed.
First Series
The high pulse rate GaAs lidar was used in making limited field
measurements that demonstrated the principle involved. The field tests
included measurements of the opacities of targets specifically constructed
for this purpose. The unusually large beam divergence (20 mrad) of the
GaAs illuminator used required that the targets be large. Though this
beam divergence is typical of GaAs illuminators, conventional laser
transmitters such as the one used in the FM-CW lidar have a beam diver-
gence of 1 mrad or less. Accurate measurements of the target opacity
require that the beam width at the target be smaller than the target
width. A 20-mrad beam is useful at 100-m range against targets at least
2-m across.
An experimental smoke generator capable of creating a 2-m plume
could not be acquired in this effort, although actual industrial plumes
are often this large. Instead, a 2-m square wooden frame was fabricated
and layers of cheese cloth were nailed on to the frame to simulate semi-
transparent targets of different opacities. These targets were placed
at a distance of 100 m from the lidar to perform opacity measurements.
To perform the actual measurements, the lidar was pointed at the target,
and the return from the target was displayed on the "A-scope" display.
The boxcar was triggered from the laser transmitter trigger to obtain
the time reference.
The first set of measurements were performed using a wide-pulsewidth
laser. Since the minimum pulsewidth available from the GaAs laser was
180 ns (FWHP), such a wide pulsewidth tends to introduce measurements
errors in opacity readings. Since the target return from a volume tar-
get is the convolution of the transmitted laser pulse and the target,
the return signal from the target is at least as wide as the transmitted
laser pulse. The signal return from "clear air" volume targets behind
the discrete target (e.g., glass plate or semitransparent cloth mounted
on a frame) is at least 30-dB below the peak return from the discrete
target itself. The actual pulse width at -30 dB points could be as high
as 500 ns. The return pulse range must be selected so that it contains
none of the energy from the target reflection itself. Backscatter sig-
nal returned from clear-air volume targets located at least 250 ns behind
the peak return from the target itself will therefore need to be used to
make target opacity measurements using purely clear-air volume scatter
signals as references. In practice, a temporal location of about 400 ns
would be more appropriate. This implies that the clear air volume "tar-
gets" used as reference for making target opacity measurements are
located about 60 m behind the target. This may not be as meaningful as
37
-------
using clear air volume targets immediately behind the target because the
more distant signals are necessarily weaker. The results shown in Table 4
are indicative of this wide-pulsewidth laser limitation.
The opacity measurements performed with the high pulse rate lidar,
as indicated in Table 4, show results below actual opacities measured in
the laboratory. This discrepancy can be partly explained in terms of
the effects of the long pulse width previously described. As described
in the preceding paragraph, the transmitter pulse width when the measure-
ments indicated in the table were performed was 180 ns (FWHF). To make
measurements of backscatter return signal from clear air volume targets,
the sampling gate of the boxcar integrator has to be temporarily located
more than 180 ns behind the peak of the target return signal. Backscatter
signal measurements at distances so far removed from the target obviously
lead to poor signal-to-noise ratios. Moreover, the reference (clear air
volume) "targets" are relatively far removed from the target, causing
additional measurement discrepancies.
Second Series
The field measurements indicated in Table 4 help to demonstrate the
performance of the high pulse rate lidar. The basic concept of coherent
integration using a commercially available boxcar integrator has been
demonstrated with these field measurements. The discrepancy in the meas-
ured opacities of the target are primarily attributable to the long
pulse width of the laser transmitter. If the pulse width is made con-
siderably shorter, the opacity readings can be made more definitive, and
any other discrepancies observed can then be attributed to other system
parameters. An extra effort was therefore made to reduce the transmitter
pulse width and make additional measurements to gather more definitive
data using the high pulse rate lidar.
TABLE 4. FIRST SERIES: DISCRETE TARGET OPACITIES MEASURED
BY A HIGH PULSE REPETITION FREQUENCY LIDAR
Obser-
vation
1
2
3
4
Amplitude of Return Signal
With Target
"A"
10.2
10.1
10.3
10.6
Without Target
"B"
12.1
11.9
11.9
11.0
Ratio R
B/A
0.84
0.85
0.87
0.96
Opacity = (1 - v/R)100
With Lidar
9
8
7
2
In Lab
19
19
19
19
38
-------
The transmitter pulse width of the GaAs laser was reduced by
redesigning its power supply and pulse-shaping circuitry. New high speed
silicon control rectifiers (SCR) were selected to obtain the fastest rise
time when the SCR was turned on. The pulse-shaping circuitry in the
laser head was redesigned to achieve high-speed gate pulses to drive the
SCR. The original laser transmitter consisted of two parallel banks of
GaAs diode arrays independently fired. Coincident output from the two
banks was practically impossible because of the differential delays in
their driving and SCR circuitry. The two banks were therefore connected
in series, and their charging and driving circuitry redesigned. These
modifications resulted in a reduction in pulse width from 180 ns (FWHP)
to 120 ns (FWHP). Further reduction in the pulse width could be obtained
by resorting to hybrid packaging techniques, specially developed SCR
circuits, and carefully tailored charging and discharging circuits. These
steps could not be taken in the present contract effort.
The improved transmitter and the receiver were used to perform
additional field measurements. The performance specifications of the
improved lidar are shown in Table 5. It can be observed that as a result
of the modifications to the transmitter, the output power was reduced.
TABLE 5. PERFORMANCE SPECIFICATIONS OF THE BREADBOARD
HIGH PULSE RATE FREQUENCY LIDAR
Laser Transmitter
Transmitted wavelength
Spectral width
Output pulse width
Pulse rate
Output power (per pulse)
Average power
Output beam diameter
Beam divergence
Laser Receiver
Front element
Spectral filter
Field of view
Detector
Post-detection bandwidth
Minimum detectable signal
890 nm
±3 nm
120 ns (FWHP)
1000/s
125 W
15 X 10~3 W
20 cm
20 mrad
30 cm/f 1.2 Fresnel lens
±5 nm
110 mrad
Silicon Photodiode
(EG&G SGD 444)
5 MHz
4 X 10"
W (-74 dBw)
39
-------
These field measurements were performed on the three different targets
used before. Since the transmitter power had been reduced as a result
of the design changes, the targets were placed 20 m closer to the lidar.
During the course of measurements, the atmospheric visibility conditions
were exceptionally good (visibility always in excess of 40 km). This
resulted in very poor signal-to-noise ratio for signals from the air
behind the targets since the clear air volumes back-scatter is inversely
proportional to the visibility. To obviate this difficulty, a small
target was placed about 25 m behind the target. The target cross sec-
tion was small enough that it appeared as a very small hump on the noisy
baseline behind the main target. This situation is illustrated in
Figure 15. If the amplitude of the signal from the reference target can
be measured in the presence of the noise apparent on the baseline, then
the opacity of the target can be measured using the high pulse rate lidar
technique, in conjunction with the pulse-to-pulse integration by a com-
mercially available boxcar averager.
The measurements of target opacity were performed as before. The
amplitudes of the backscatter signal return from the reference target
were measured with and without the semitransparent target at 80 m from
the lidar. A Princeton Applied Research Model 164 boxcar averager was
used to perform the measurements of the signal amplitudes. Table 6
indicates the results of the measurements.
Figure 15. Target return signal from the semitransparent target.
The target return from the reference target is buried
in the baseline noise. (Bottom trace is the sampling
gate position.)
-------
TABLE 6. SECOND SERIES: MEASUREMENT OF TARGET OPACITIES
USING HIGH PULSE RATE GaAs LIDAR
Observation
1
2
3
4
5
I
2
3
4
5
I
2
3
4
5
I
2
3
4
5
Signal Amplitude of
Reference Target
(mV)
With Target
170
168
171
170
169
161
159
160
162
164
210
207
208
210
209
205
205
203
202
205
Without Target
408
403
410
410
407
401
397
399
399
401
341
340
341
340
339
336
335
332
335
336
Difference
A - B
(mV)
238
235
233
240
238
240
238
239
237
237
131
133
133
130
130
131
130
129
133
131
Ratio
A/B
0.416
0.416
0.417
0.414
0.415
0.401
0.401
0.401
0.406
0.409
0.616
0.608
0.610
0.617
0.616
0.610
0.612
0.611
0.603
0.610
Opacity
ft)
100 (1 -/S/B)
35.5
35.5
35.4
35.6
L_2:LL__
36.7
3617
36.7
36.3
36.1
21.5
22.0
21.9
21.5
21.5
21.9
21.7
21.8
22.3
21.9
Measured in
Laboratory
36
36
36
36
36 |
36
36
36
36
36
20.4
20.4
20.4
20.4
20.4
20.4
20.4
20.4
20.4
20.4
Comments
Daytime
measurement
Nighttime
measurement
Daytime
measurement
Nighttime
measurement
-------
From the field measurements indicated in Table 6, the validity of
the principle of rapidly integrating pulsed signals, otherwise buried
in noise to enhance the signal-to-noise ratio, is well demonstrated.
This demonstration emphasizes the concept of using low average power
level, eye-safe lasers operating at high pulse rates for measuring
opacities of semitransparent targets at practical distances.
Although it was hoped that the target opacities would be measured
using clear air volume "targets" as references and represent real field
measurement requirements, the combinations of low average power avail-
able from the breadboard laser (only 15 mW average power) as well as
exceptionally unfavorable atmospheric conditions (very high visibility)
the clear air references did not in reality provide enough backscatter
signal. As can be seen in the photographs of the lidar return signal
from the semitransparent target out of the lidar path, the return signal
from the artificial reference is barely discernible and for all prac-
tical purposes is buried in the baseline noise. It is remarkable that
with a reference target having such a small lidar cross section and
whose return signal is buried in noise, the opacities of the semitrans-
parent targets were in such close agreement with the measurements per-
formed in the laboratory: the maximum error in opacity reading was
less than 2%.
42
-------
REFERENCES
1. Ferguson, R- A- » Feasibility of a j:w Lidar Technique for Mea'sure-
ment of Plume Opacity. Prepare 4 for EPA by StanTord '' "Research In;- tv
tute under Contract number ka-U2-0543. Publication Number gPA^
650/2-73-Q37. Research Triable Park, N.C. , NovemtSef 'T973T 9t) "pp.
2. Evans ^jjj, ^^ J^evejopmenf; Q| Stack Opacity Effluent Measuring System.
Prepare^ for Edison Eleatr^p Institute by Stanford Resea-rctri^siitut
under_^^j?p4£LcJ^-.65^4r--fffi«-i'tablication Number PB-233-135/AS.
Springfield,, Virginia, July 1967. 96 pp.
* * . TflS*1
* w v. ' "V * *"
3. Jackson, D, \l. , Development of a CW Lidar for Remote Measurement of
Smoke Pjiwme Opacity, Proposal No. ELD-69-165. Stanford Research
Institute, Menlo Park, California, January 1970.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-023
. RECIPIEN1
-v. TITLE AND SUBTITLE
CONTINUOUS READING LIDAR TECHNIQUE FOR MEASURING PLUME
OPACITY
REPORT DATE
February 1979
PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Dilip G. Saraf
. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
SRI International
333 Ravenswood Avenue
Menlo Park, California 94025
10. PROGRAM ELEMENT NO.
1AD605_ BA-63 (FY-76)
TTTCONTRACI /GRANT"fto.
68-02-1291
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory-RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 6/74 - 9/76
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTFiACT _ ,
The development of a laser radar (lidar) instrument for remote measurement of the opa-
city of smoke-stack plumes is described. The work was conducted_withina number of con-
straints The constraints required the lidar instrument to be field-portable, eye safe,
relatively low in cost, and simple to operate. Two.lidar measurement methods were
studied for the instrument: continuous wave (CW) lidar and high pulse rate lidar.
A research model CW lidar was constructed and evaluated. The evaluation showed that the
CW lidar could remotely measure the opacities of screen targets, or smoke generator
plumes at night to within 3% opacity at a distance of approximately 80 meters. Environ-
mental light interference prevented operation of the lidar during daytime.
Proof-of-principle experiments were performed to demonstrate the feasibility of using a
high pulse rate lidar for plume opacity measurements. The evaluation showed that the
lidar1s laser did not have enough power to make measurements under field conditions.
However, the lidar was capable of measuring the opacity of the screen targets at close
range (40m) and by placing a small artificial scattering target in the atmosphere
behind them. With this artificial signal enhancement, results showed that the high
pulse rate lidar could remotely measure the opacities of the screen targets to within
2% opacity during daytime or nighttime operation.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
* Air pollution
* Plumes
* Opacity
* Measuring instruments
* Optical radar
Development
Evaluation
13B
21B
14B
17H
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCI.ASSTFTFn
21. NO. OF PAGES
54
apiSECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
44
------- |