EPA-650/2-73-040
December 1973
Environmental Protection Technology Series
-------
EPA-650/2-73-040
DEVELOPMENT OF RANGE SQUARED
AND OFF-GATING MODIFICATIONS
FOR A LIDAR SYSTEM
by
George W. Bethke
General Electric Company
Space Sciences Laboratory
P. 0. Box 8555
Philadelphia. Pa. 19101
Contract No. 68-02-0570
Program Element No. 1A1010
EPA Project Officer. William D. Conner
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, N. C. 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D. C. 20460
December 1973
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
ii
-------
ABSTRACT
This study was intended to eliminate the dominant source of error in the
remote measurement of smoke - stack plume s using lidar, and to improve the
data presentation of the lidar transmittance measurement.
The use of a special grid-controlled photomultiplier tube and off-gating
system has resulted in a large reduction of photomultiplier tube afterpulsing
which has been the principal problem limiting the accuracy of plume transmit-
tance measurements.
A range-correcting time-squared amplifier has also been developed which
permits much faster and more convenient data reduction than formerly possible.
The above two developments have been adapted to the existing van-mounted
EPA lidar system; and the overall system accuracy has been evaluated using
both opaque and semi-transparent targets of known transmittance (T). These
test results show the lidar system to yield lidar transmittances (TL) which
(a) are precise to ^ 0. 015 average, (b) are accurate to +_ 0. 02 average for T > 0. 5,
and (c) are too large at T < 0. 5 (for example, now TL = 0. 52 at T = 0. 50, and
TL = 0. 28 at T = 0. 20). These results are to be compared with the pre-modifi-
cation accuracies of + . 09, -0 at T > 0. 50, and more severely too large at
T < 0. 5 (for example, formerly TL = 0. 59 at T = 0. 50, and TL = 0. 40 at T =
0. 20).
111
-------
TABLE OF CONTENTS
Page
ABSTRACT iii
1. INTRODUCTION 1
1. 1 Background 1
1. 2 Purpose of the Program 1
1.3 Scope of Work 1
2. TECHNICAL BACKGROUND 2
2. 1 The Lidar Concept 2
2. 2 Measurement of Plume Transmittance by Lidar 3
2. 3 Range Correction 3
2. 4 Photomultipller Tube Afterpulsing 5
2. 5 Existing Lidar System Design 6
3. USE OF GRID-CONTROLLED PM TUBE 9
3. 1 Description of Photomultiplier Tube 9
3. 2 Divider/Gating Circuit Used" 9
3. 3 Detection System Description and Performance 12
3. 3. 1 Gain and Dark Current 12
3. 3. 2 System Gating Characteristics 12
3. 3. 3 Photomultiplier Tube Afterpulsing 14
3.3.4 System Linearity 17
4. TIME-SQUARED AMPLIFIER 21
4. 1 Amplifier Design 21
4. 2 Amplifier Performance 23
5. LIDAR MEASUREMENTS OF TEST TARGETS 28
5. 1 The Test Targets and Their Laboratory-Measured 28
T r an smittanc e s
5. 2 Lidar Transmittance Measurements of Test Targets 29
6. OPERATING INSTRUCTIONS FOR LIDAR MODIFICATIONS 37
6. 1 Use of the Modified Lidar Receiver System 37
IV
-------
TABLE OF CONTENTS (continued)
Page
6. 1. 1 Photomultiplier Tube Operating Voltage Selection 37
6. 1. 2 Obtaining the Proper Off-Gating Pulse Position 39
6. 1. 3 Taking Transmittance Data 39
6. 2 Internal Description and Adjustment of Detection System 41
6. 3 Time-Squared Amplifier Adjustments 43
7. SUMMARY AND CONCLUSIONS 46
8. REFERENCES 47
-------
1. INTRODUCTION
1. 1 Background
A mobile lidar (light detection and ranging) system has previously been
developed 1» 2 for the remote measurement of the transmittance of smoke-stack
plumes. A standard pulsed lidar is used for the measurement where atmospheric
backscatter of the outgoing pulse is observed as a function of distance. The
transmittance of a plume is determined by firing the laser pulse through the
plume and measuring the discontinuity between the atmospheric backscatter of
the pulse just in front of and behind the plume. Extensive testing of the system If 2
has shown that the accuracy of the measurement is affected by the intense pulse
backscattered by the plume and that data reduction is complicated by the inverse
square range dependence of the signal return. Off-gating the photomultiplier
detector of the lidar receiver when the backscatter from the plume is received
has improved but not eliminated the error. This project is designed to eliminate
the lidar error in measuring plume transmittance due to plume backscatter of
the laser radiation and to linearize the backscattered signal display with respect
to range.
1. 2 Purpose of the Program
The study was designed (1) to eliminate the dominant source of error in the
remote measurement of the transmittance of smoke-stack plumes using lidar,
and (2) to improve the data presentation of the lidar transmittance measurement.
1. 3 Scope of Work
The following tasks were to be performed:
1. Develop a time-squared amplifier to range correct the lidar output.
2. Investigate the use of a specially designed grid controlled photomulti-
plier tube and associated external circuitry to eliminate photomulti-
plier after-pulsing which results from smoke plume lidar returns.
3. Adapt the time squared amplifier (# 1 above) and the grid controlled
photomultiplier tube (#2 above) to the existing EPA lidar system.
4. Evaluate the accuracy of the modified system for remote measure-
ment of the transmittance of semi-transparent targets of known trans-
mitt ance.
-------
2 . TECHNICAL BACKGROUND
2. 1 The Lidar Concept
The conventional single ended lidar configuration involves a laser trans-
mitter sending out a short pulse of light (10 - 50 nanoseconds) from a laser
source in conjunction with a receiving system mounted on a parallel or nearly
parallel optical axis which collects backscattered light from the outgoing laser
pulse as it propagates away from the system. This collected light is focused on
a field-limiting aperture and is then collimated before passing through a narrow
band interference filter to a photo-multiplier detector. Thus, backscattered
light intensity at the receiving aperture is linearly transformed to voltage across
the photomultiplier load resistor as a function of time, and thus range.
When the lidar optical path is through the clear atmosphere, the backscat-
tered signal at the laser wavelength is made up of two components. These are
the component due to molecular (Rayleigh) scattering and that due to aerosol
(Mie) scattering. A complete expression for system output voltage as a function
of range is shown as equation 15 in appendix A of reference 1. However, this
expression can be simplified somewhat when considering the relatively short
slant ranges at which the plume opacity lidar obtains data. For these ranges,
attenuation of the beam and atmospheric density changes can be neglected, with
the correct expression for system output voltage being
cAS E ER p -.
v = ^ n(r) . aR(180°) + m(r) . ^(180°) (2. 1)
where
r = ct/2. (2.2)
Here c is the velocity of light, A is the area of the receiving aperture, S is the
photomultiplier sensitivity (amps/watt), Eo is the overall system optical effi-
ciency, E is the laser output energy, R is the load resistance, r is the range,
n(r) is the number density of molecular scatterers, dp (180°) is the Rayleigh
backscattering cross section for air, m(r) is the mass concentration of aerosols
(figm/m3), crM(180°) is the mass normalized aerosol (Mie) backscattering cross
section, and t is the time since lidar time zero.
If n(r) and m(r) are constant with range then the bracketed term of equation
(2. 1) is a constant atmospheric backscattering coefficient, ks, and equations
(2. 1 and 2. 2) become
V = 2 ASE ER k /(ct2), (2. 3)
o s
indicating a basic 1/t dependence of the system signal.
-------
2. 2 Measurement of Plume Transmittance by Lidar
The basic technique used by the EPA lidar system for obtaining the one-
way plume transmittance TL is illustrated in Figure 2. 1 which shows a drawing
of a lidar return with photomultiplier tube signal voltage plotted against time and
thus range. Although the signal voltage is inherently negative, it is shown in-
verted for clarity in this drawing. The laser pulse is initiated in the transmitter
and defines t = 0 for the system. The initiation of this pulse is monitored by a
photodiode looking at the laser cavity providing a triggering signal for the system
electronics. Since the transmitting and receiving system axes are parallel but
not coincident and since the receiving system has a narrow field of view, the
backscattered light from the outgoing laser pulse cannot be seen by the receiving
system for the first 50 meters or so of range. Following this, the angular field
of view of the receiving system combined with the collimated divergence of the
laser pulse allows gradual overlap of the two beams with full overlap occurring
at a range of approximately 130 meters (0. 8 p sec.). From this point onward
the signal return shows a general 1/t^ dependence as shown by equation (2. 3) as
long as the atmospheric scattering coefficient, ks, is a constant.
Of course a smoke plume means a discontinuous change in ks which shows
up as a sharp signal spike at the range of the plume as indicated in Figure 2. 1.
The amplitude of this spike can be up to 40 db above the ambient light scattering
for low transmittance highly reflecting plumes. Light scattered to the receiver
from ranges greater than the plume will have been attenuated by the plume twice,
once going out and once conning back. Therefore, the received signal shows a
discontinuity at the range of the plume. If the ambient air scattering signal from
each side of the plume is extrapolated to a common point in a 1/t^ fashion, as
shown by the dotted lines in Figure 2. 1, the ratio of the two signal levels can be
computed as A/B as shown. The one way lidar-determined plume transmittance
is then given by
TL = (A/B)1/Z (Z.4)
All of this yields an absolute measurement of plume transmittance independent
of laser pulse energy provided that ks is homogeneous in front of and beyond the
plume.
The above measurement is easy in principle, but in practice an accurate
extrapolation of data to a common range is a time consuming nuisance.
Also, a phenomenon called photomultiplier tube afterpulsing has previously
limited the accuracy with which the signal on the far side of the plume can be
readl.
2. 3 Range Correction
Since the lidar signal return, for a homogeneous scattering atmosphere,
behaves inherently as l/r2, it is necessary to transfer the data from a polaroid
-------
LIDAR RETURN FROM
SMOKE PLUME
PLUME
RETURN
AMBIEN
AIR RETURN
LASER
PULSE
t -i sec
r=ctp/2
Figure 2.1 A sample lidar return through a smoke plume showing the single beam, single shot technique
for obtaining the one way lidar determined transmittance, T .
JLi
-------
oscillograxn to a double logarithmic plot to make the necessary linear extrapola-
tions prior to computing the far side to near side signal ratio, A/B. This is a
time consuming process done with dividers and can take 20 to 30 minutes for one
oscillograxn before the results are known. If the signal were not 1/r^ in nature
but constant with range, the data reduction could be done directly on the oscillo-
graxn simply by taking the ratio of the average signal level before and after the
plume.
One method of range-correcting the 1/r lidar return to a constant signal
output, is to run the output from the lidar receiver light detector through a t^
amplifier and then on to the oscilloscope. The insertion of such an external
amplifier into the system may reduce the detection frequency of response and it
will have a maximum dynamic range of gain of about 100:1. However, such a
separate t2 amplifier has the advantages of being relatively easy to make, of not
disturbing an existing detection system or interfering with the method of gating,
and of being easily adjusted for use with different lidar range intervals. This is
the type of range-correction system developed here (see Section 4).
2. 4 Photomultiplier Tube Afterpulsing
In order to make lidar-determined smoke plume transmittance measure-
ments, it is necessary to faithfully record the atmospheric backscattering both
just in front of and just behind the plume. There is, of course, no trouble in
obtaining the near side signal right up to the point in time where the laser pulse
intersects the plume. However, the intense light scattering from the plume it-
self (during the time that the laser pulse is in the plume) complicates the pro-
blem of faithfully recording the clear air light scattering immediately on the far
side of the plume. This backscattering from the plume can be as high as 40 db
above the ambient clear air scattering for the case of a dense white plume (or
for test purposes, a white wall).
As already indicated, an effect called photomultiplier tube (PMT) after-
pulsing increases the apparent amplitude of the far side (of plume) clear air
return, yielding plume transmittance values that are too high. The explanation
for this effect appears to be as follows: The PMT cathode photoelectrons
(created by backscattered light from the intense plume return) leave the cathode
and are accelerated towards the first dynode. At some point along their path
these electrons acquire sufficient energy to ionize the residual gas in the tube or
to eject ions from the tube structure upon impact. These relatively slow ions
then are accelerated back to the photocathode where they cause secondary elec-
tron emission about 0. 2 to 1/J sec. after the primary pulse. The secondary
electrons then are amplified down the dynode chain and wind-up as an afterpulse
signal at the anode. Those electrons resulting from light ion impact at the
photocathode cause the earliest afterpulsing while heavier and slower ions cause
the later afterpulsing. The amount of afterpulsing then depends on the intensity
of the reflected signal from the plume or target.
-------
As discussed in reference 1, attempts to stop the initial flow of plume-
return photoelectrons by applying repelling electric fields to the PMT focus elec-
trode, to an external grid,and to dynodes had not been successful. Some prelim-
inary considerations had indicated that the best solution involves the use of a
PMT that has been designed for off- and on-gating via an internal flat mesh grid
located adjacent to and parallel to a flat photocathode. The PMT selected for
this purpose is a specially designed grid-controlled tube (ITT type F4084) as
described in Section 3.
2. 5 Existing Lidar System Design
A detailed description of the lidar system design before this modification
program, is given in reference 1. Figure 2. 2 shows the lidar block diagram,
while Table 2. 1 lists many of the lidar optical characteristics.
To summarize here, the lidar transmitter consists of a Q-switched water-
cooled ruby laser (with photodiode output monitor) beamed through a divergence-
reducing 5 inch diameter Gallilean type collimating telescope. The lidar re-
ceiver consists of a 6 inch diameter telescope with 4 mrad. field of view and 12A
FWHH interference filter (IF) located in front of the photomultiplier tube (PMT)
detector. The original PMT detector was an Amperex 56TVP wired for an off-
gating capability ^ to block the strong plume return signal. However, PMT after-
pulsing * was still present to a degree sufficient to partially interfere with the
far-side clear air return (see Figure 2. 1), thus limiting the accuracy of the
plume transmittance measurement. Data recording was accomplished with Pola-
roid oscillograms of the direct PMT signals.
This entire system is located in a motor van with flip open top and elevating
lidar optics mount*. Portable motor generators provide electrical power when
not otherwise available.
-------
PMT-PS
PULSE GEN.
(FOR PMT
GATE)
DOUBLE BEAM
SCOPE AND
CAMERA
4-12 X AIMING SCOPE
6 IN. DIAM., f/5
LASER PS
WATER
COOLING
SYSTEM
MOTOR
GEN.
PM TUBE
DETECTOR
IF
FROM LASER MONITOR
RUBY LASER
5 IN. DIAM.,
IF-THERMALLY CONTROLLED
INTERFERENCE FILTER
Figure 2. 2 Block diagram of mobile lidar system
-------
TABLE 2.1 Lidar Optical Characteristics
Laser
Manufacturer and Model:
Type:
Output Wavelength:
Maximum Output Energy:
Pulse Width (FWHH):
Ruby Rod:
Beam Divergence:
Repetition Rate:
Cooling:
Laser Collimation
Collimated Beam Diameter:
Collimated Beam Divergence:
Receiving Optics
Type:
Objective Lens:
Collimating (eye) Lens:
Field of View:
Interference Filter
Manufacturer:
Type:
Diameter:
Transmission Peak Wavelength:
Transmission Center Wavelength:
Thermal Shift:
Bandwidth (FWHH)
Peak Trans mitt ance:
Residual Trans mitt ance:
Filter Rejection:
Filter Tilt in Lidar:
Hadron/TRG Model 200B/104A
Rotating prism Q-switched ruby
6943A
1. 0 joule Q-switched
< 30 nanoseconds
1. 0 cm x 7.6 cm
< 5 mrad. (full angle - 1/2 power)
0-3 ppm
Deionized water
10 cm
~ 0. 5 mrad. (full angle - 1/2 power)
Refracting
6 inch diameter,f/5
2.14 inch focal length
4. 0 mrad. full angle
Infrared Industries
3 Cavity Interference
2 inches
6946. OA (25° C)
6943. 4A (25° C)
+ 0.2 A/°C
12.0 A
0.66
< 0.01%
Far uv to 2 1. 0 micron
0 (perpendicular to radiation)
-------
3. USE OF GRID-CONTROLLED PM TUBE
3. 1 Description of Photomultiplier Tube
An ITT type F4084 grid-controlled photomultiplier tube (PMT) was obtained
for this lidar modification. Some of the manufacturer-supplied characteristics
and test results for this particular PMT are listed in Table 3. 1. It should be
noted that all of the ITT tests were made using their standard (nearly linear)
voltage divider, which is different from the more tapered divider designed for
its use in this lidar system (see Section 3. 2).
The principle feature of interest to us for this type PMT is that it has a
mostly transparent internal flat mesh grid located parallel to and close to the
flat multi-alkali photocathode. According to the tube manufacturer, this internal
mesh grid not only permits off- or on-gating the PMT response to the incident
light, but it is also likely to eliminate afterpulsing which results from any inci-
dent light during the off-period.
The flat photocathode and parallel flat mesh grid design reduces the tend-
ency for charges to redistribute on the photocathode during the gating pulse.
Furthermore, the tube manufacturer claims that the photocathode is of special
low impedance design, thus reducing any charge distribution and current density
problems. Consequently, no tendency was observed for the on-recovery time to
increase as the off-gating period was increased to and beyond values of interest.
3. 2 Divider/Gating Circuit Used
The special "tapered" PMT voltage divider designed for use with the lidar
system, is shown in Figure 3. 1. This divider system is significantly more
tapered than the divider used by ITT because it was here desired to optimize
linearity of PMT response to large pulsed outputs, even at the possible sacrifice
of some PMT gain. To achieve a large pulsed output linearity requires both
(1) increasingly large inter-dynode voltage differentials as the anode end of the
tube is approached so as to reduce the current-limiting effect of space charge,
and also, (2) increasingly large valued low inductance capacitors as the anode
end of the tube is approached so as to maintain constant dynode voltages during
the output pulse. It is at the anode end of the PMT where the amplified electron
avalanche is largest and has the strongest disturbing effects which tend to upset
linearity of response.
It should be noted that although the PMT is rated by the manufacturer to
6000 volts total, the maximum total voltage applied to this divider/gating circuit
(Figure 3. 1) must be limited to <3000 volts, and preferably should not exceed
2500 volts. The 3000 volt limit is dictated by the maximum voltage and dc wat-
tage ratings of many circuit components, while a 2500 volt limit provides a
safety margin for the 3000 volt gating pulse capacitors and also reduces heat
produced in the divider circuit. Off-gate pulses of -5 volts are sufficient as
discussed later.
9
-------
Table 3. 1 Manufacturer - Supplied Characteristics for ITT Type F4084
Grid-Controlled Photomultiplier Tube*
Tube manufacturer and type
Number of dynodes*
Multiplier type
Photocathode characteristics:*
Sensitivity type
Quantum efficiency
Quantum efficiency
Tube response time
Grid gated on /off ratio*
Grid A V for on - off*
Maximum overall voltage
Tube gain and dark current:*
Overall voltage
on tube Tube Gain
1000
1500
2000
2500
3000
4000
t*»l
2 x 10 ,
1.2 x 10"
5 x IO5 >
1.7x 10*
9x IO6
IT&T F4084
8
linear focused
MA-2 (modified S-20)
4.8% at 6940A
21% at 4080A
<5 ns**
43 db
-2.5 volts
6000 volts **
Dark current
(amps)
1. 2x 10
4x 10-9
6 x ID'8
7 x IO-7
5 x 1C'6
-10
Notes:
*
These measurements were made by ITT on the supplied tube
(S/N 077201) while using their standard (nearly linear)
voltage divider.
** Based on the manufacturer's general specification sheet.
10
-------
(white)*
Shield
(Sh)
|MHV
(bl
^^H
^
"HV
INPUT"
.01 _
(3000V)~
(yel]
Notes:
acK)* _.
"7 ^
/ 10M
r^
10
k
_x
0 .605
5. IK
5
Grid /
510K<
4
OW)*1_^5T
1
Interr
contn
grid (I
-Cathode (K)
.001
II
Vy\ 1
5M
5. IK
R
OK I 5M
1 II T .,_
sHo
>
> .01
IL
Ir
(3000V) L
< 51
L f,
G\ ^S
"PULSE
C
E
«., 1
Fl *
5) (13) (6) (12) (7) (11)
1 D2 D3 D4 D5 D6
flOO |100 S100 SlOO >75 |5
|| II II li n
II
500
RJ^ R ^
5M
U1
F2
(2)
Anc
^r
[BNC
"OUTPUT"
INPUT"
1) * = Flying lead.
2) All capacitors are ceramic disc type with 1000V
rating unless stated otherwise.
t-^r,
(3)
"" (9
Ac
i
R
WV
.025.
(500VH
r1
II
.001
wv 1-
J\
)
eel.
R R
.02
II
II
||
II
,.02
II li II
.002 .005 .01
R R R
(15) 1
F4 < >-£5M
(10) (8)
D8 D7
| 10 S24
n
-II
.02 .01
**
-------
3. 3 Detection System Description and Performance
This F4084 photomultiplier tube (PMT) and housing are mounted in place
of the originally used Amperex PMT and housing (see Fig. 2. 2), while still using
the original PMT power supply (PS) and gating pulse generator. A ground glass
diffuser has been located between the PMT and interference filter IF (Fig. 2. 2)
so as to spread illumination of the PMT cathode over most of its area, thus
optimizing its total current emission capability and reducing any effects due to
variation of cathode sensitivity with spacial location.
Various new detection system performance characteristics and adjustments
are presented in the following sub-sections.
3. 3. 1 Gain and Dark Current
Figure 3. 2 shows plots of PMT gain and dark current vs. applied
total voltage through the ranges of interest while using the divider circuit of
Figure 3. 1. These curves are based on room temperature measurements made
at 100 volt intervals, with the dark current measurements being in amperes and
the sensitivity measurements being relative while using a stabilized dc light
source. To avoid dynode voltage dc distortions during these measurements due
to high tube output, the light level was always such that the PMT anode current
never exceeded . 002 of the divider circuit current. The entire sensitivity curve
was then normalized to a theoretical gain calculation of 2. 29 x lO'* at 1500 volts.
The above PMT gain (current amplification) was calculated for this
particular tube as follows: First, a curve of gain/dynode (6) vs. volts/dynode
(A V) was derived from the manufacturer - supplied gain vs. voltage information
(see Table 3. 1) plus the voltage distribution of the divider circuit they used for
those measurements. Then this 6 vs. AV curve plus the voltage distribution
of the divider circuit used here (Figure 3. 1) was used to calculate the current
amplification ("gain") at 1500 volts overall on the PMT. Thus, although the gain
curve's shape and slope are accurately measured, its calculated normalization
is less certain.
3. 3. 2 System Gating Characteristics
PMT gating response time and an approximately 20 db lower limit to
on/off ratio was measured by off-gating the PMT while it was exposed to a
stabilized cw light source. In these cases, a Corning #2412 red filter was used
and the light level was adjusted to yield PMT anode outputs less than 0. 1 of the
divider chain current. The latter precaution insured reasonably linear PMT
response under cw operating conditions. For all of these off-gate response time
adjustments and measurements, a Tektronix #551 oscilloscope with #1A1 pre-
amplifiers was used, this combination having a manufacturer - stated risetime
(Tr) of 13-17 ns depending on scope sensitivity setting. The pulse generator
12
-------
1500
a 000
ITT type F4084
(S/N 077201)
1200
Figure 3. 2
MOO
PMT Volts
Photomultiplier tube (PMT) gain and dark current vs. overall
voltage while using the "tapered" divider circuit of Figure 3. 1
13
-------
provided pulses of about 15 ns rise and fall times. Risetime (Tr) and falltime
(Tf) are defined as the time required for the waveform to change from 10% to
90% of the final amplitude.
Satisfactory off-gating of this new detection system is obtained with
gating pulses of -2 through -10 volts, with -5 volt pulses having been selected
for all final gating investigations and adjustments. Since the optimum grid ad-
justment ("grid adj. " of Fig. 3. 1) is a moderate function of overall PMT applied
voltage, the grid adjustment was optimized for Tr sw Tf at 1750 volts applied.
With all the above conditions and method of measurement, it was
found that the detection system satisfactorily off-gates with an on/off ratio of at
least 20 db, and both Tr and Tf are < 50 ns through the voltage range of at least
1550 - 1900 volts. It was also noted that the on-recovery time (Tr) is indepen-
dent of off-gated period for gating periods to at least 1 microsecond in length.
Figure 3. 3 shows oscillograms illustrating these off-gating measure-
ments and results. PMT anode stray pick-up from the gating pulse produces
the < 4 mV ringing in the PMT anode signal at the beginning and end of each gate
pulse. This rapidly damped pick-up does not interfere with lidar useage of the
detector due to their transient nature and due to the fact that the lidar signal
amplitudes of interest are 10 to 1000 times greater than these pick-up amplitudes.
A more sensitive measurement than possible above of detection
system gating on/off amplitude ratio was made by comparing an ungated but op-
tically attenuated light pulse signal with a similar unattenuated but gated light
pulse signal. This was accomplished by making monitored lidar shots at the
white wall of a building. As illustrated in Figure 3. 4, lidar returns were ob-
tained from the white wall while not gating but with a 4. 32 O. D. neutral density
filter located in front of the detection system (Fig. 3. 4a), and also lidar returns
were obtained from that same white wall with no neutral filter but while off-
gating (Fig. 3. 4b). The neutral density filter is of the absorbing glass type, and
had been calibrated at the laser wavelength. The examples of Figure 3.4 as
well as other more sensitive oscillograms never clearly showed the wall return
while off-gating,with the resultant conclusion being that the gating on/off ratio
is at least 60 db.
Comparing these gating characteristics with those of the originally
employed ("old") detection system (Amperex 56TVP tube with dynode off-gating),
we see that the new detection system has Tr » Tf = 40 - 50 ns vs. ~ 35 ns for
the "old" system*, while the new detection system has an on/off amplitude ratio
of s 60 db vs_. ~ 45 db for the "old" system1.
3. 3. 3 Photomultiplier Tube Afterpulsing
PMT afterpulse investigations were made via white wall lidar returns
similar to the on/off ratio lidar method just discussed above. The primary
14
-------
PMT -
(1750V)
Light off
Light on
PMTX
tmmm
1950V
1750V
1550V
(b)
Figure 3. 3 Double beam oscillograms showing response by the F4084 photo-
multiplier tube {PMT) to off-gating via divider/gating circuit of Fig. 3. 1 in the
presence of a cw light source. P shows the 300 ns long -5 V gating pulses at
5 V/cm, while the upper traces show the PMT response at 5 mV/cm through a
93 ohm cable and load under the conditions indicated, all with the control grid
adjustment optimized for 1750V operation. Scope sweep speed is 100 ns/cm.
15
-------
0-
0 -
PMT (0. 2V/cm)
M
0 -
0 -
PMT (0. 1 V/cm)
M
Figure 3, 4 Lidar returns from a white wall using the new detection system (F4084
PMT with circuit of Fig. 3. 1). In each double beam oscillogram, the sweep speed is
500 ns/cm, M is the positive-going laser energy monitor trace, PMT is the negative-
going photomultiplier tube lidar return (PMT at 1700V), and W is the position of the
wall return. Figure 3. 4a has no gating but does have a 4. 32 O. D. neutral filter in
front of the PMT yielding a wall peak return of -0. 8 cm, while Figure 3. 4b has no
neutral filter but is off-gated through the 500 ns wide interval G with a -5V pulse.
Except for such off-gating ratio measurements as these, much narrower off-gate
periods were normally as illustrated elsewhere.
16
-------
difference from the above measurements is that for the afterpulse measurements
the non-neutral filter lidar shots were made with shorter off-gated periods and
also with no off-gating.
These investigations showed this new PMT (F4084) to have generally
weaker afterpulsing than the old PMT* (56TVP), and that afterpulsing which the
new PMT does have ends much sooner after the primary (wall return) pulse. As
illustrated in Figure 3. 5a, the ungated F4084 after pulsing consists of a primary
afterpulse peaking at At = about ZOO ns after the primary pulse, and a weaker
secondary afterpulse peaking at At just under 400 ns after the primary pulse.
The secondary afterpulse then rapidly decreases at larger At. This ungated
F4084 afterpulsing (see Fig. 3. 5a) can be compared with the old PMT ungated
afterpulsing as shown in Figure Bla of reference 1.
As seen by comparing Figures 3. 5a and 3.5b, when the new detection
system (F4084 PMT plus divider/gating circuit of Fig. 3. 1) is off-gated, the
white wall lidar return shows even less afterpulsing (A and A ) than when not
off-gated. The residual off-gated afterpulsing can be seen at higher oscilloscope
sensitivity in Figure 3.5c where the traces have gating conditions which are
somewhat wider and later-ending than in the case of Figure 3.5b. These new
detection system gated afterpulsing characteristics (Figures 3.5b and c) can be
compared with the old (56TVP) detection system gated afterpulsing which is
shown as Figure 3. 1,reference 1 (Figure Bla of ref. 1 is also good for compari-
son because off-gating the old detection system does not influence its afterpulsing).
Finally, Figure 3.5d shows an off-gated clear air lidar return while
using the new detection system.
Using the lidar method of Section 3. 2. 2, measurements of residual
afterpuling by the off-gated new detection system shows the following results in
terms of primary pulse (wall return)/afterpulsing amplitude ratio in db down vs.
time delay ( At) after the primary pulse:
At (ns) ~ 200 300 ~ 360 500 > 700
ratio (db) 52 56 54 61 > 65
The ~ 200 ns peak is the primary (A) off-gated afterpulse, and the ~ 360 ns peak
is the secondary (A') off-gated afterpulse. Since the afterpulses themselves are
also directly blocked during any off-gated interval, it may be that A actually
peaks at At < 200 ns with a ratio of less than 52 db down.
3.3.4 System Linearity
PMT linearity of anode output can be limited by 4 effects: (1) Pulsed
or dc output can be limited by photocathode charge depletion due to high incident
light levels. (2) The dc or average anode current level can be limited by the
17
-------
oo
0
0 -
o-
o-
PMT
(IV/cm)
PMT
(0. 1/V/cm)
PMT
(IV/cm)
M
(d)
Figure 3. 5 Oscillograms showing lidar returns from a white wall (3. 5a to 3. 5c) and clear air (3. 5d) while
using the new detection system (F4084 PMT with ckt. of Fig. 3. 1). In each oscillogram, the sweep speed is
500 ns/cm, M is the positive-going laser energy monitor {ckt. modified between a & b and d), PMT is the
negative-going photomultiplier tube lidar return, and W is position of the wall return. The primary and
secondary afterpulse peaks are indicated by A and A , respectively. Fig. 3. 5a has no off-gating, while oscil-
lograms b, c, and d have -5 V off-gating pulse widths through the intervals G of 200, 300 and 250ns , respec-
tively. Oscillograms a,b and c have 1700V on the PMT, while d has 1800V.
-------
divider circuit current from which the output (anode) current must be obtained.
(3) Low duty cycle or single pulsed current output can be limited by the charge
stored in the dynode capacitors. (4) Pulsed current outputs can be limited by
space charge effects, usually near the anode end of the dynode chain where inter-
dynode currents are largest. The first through third limits above are due to in-
duced distortions in the overall voltage distributions within the PMT.
The above linearity limits can be avoided or extended as follows:
(1) Photocathode charge depletion can be avoided by limiting the PMT illumina-
tion density to low light levels and/or through the use of a low impedance photo-
cathode design within the PMT. (Z) The dc or average anode current limitation
can be avoided by limiting the average anode output to < 0. 05 of the divider cir-
cuit current. (3) The capacitor charge limit for pulsed use can be avoided by
using large enough low inductance capacitors so they are < 0.05 discharged by
the integrated current (i. e. charge) of the output pulse. (4) The space charge
limit can be made less restrictive by employing larger PMT potentials and also
by suitably arranging ("tapering") the voltage divider so that the inter-dynode
potential differences (AV) increase down the dynode chain as the interdynode
currents increase due to the current gain/dynode (6). The space charge limit
is approximately proportional to ( AV)3/2e
For the intended normal lidar use of this new detection system, the
combination of special low impedance photocathode in the F4084 PMT and the
special tapered voltage divider (Fig. 3. 1) with relatively low overall resistance
and large low-inductance capacitors results in only the space charge effects
(#4 above) limiting the detection system linearity.
Measurements were made of space charge-limited linearity as
follows: A pulsed light source with pulse widths of 1 microsecond (FWHH) and
a repetition rate of 105-125 cps was produced by reflecting a stabilized HeNe
laser beam (6328A) off a rapidly rotating mirror. This sweeping laser light
beam was then intercepted by (1) various calibrated neutral density absorbing
glass (not coated) filters, (2) a wide "slit" to spacially limit the period of light,
(3) a ground glass diffuser to spread illumination uniformly over the PMT photo-
cathode, and (4) a Corning red filter to reduce unwanted background illumination.
The cw signal level plus pulsed light duty cycle were such as to limit even the
maximum average anode output to < 0. 01 of the divider current.
19
-------
This space charge-limited linearity measurement was made, as
detailed above, at both 1500 volts and 1800 volts PMT potential. The space
charge-limited actual anode outputs for 2%, 5% and 20% deviations (decreases)
from a linear response are listed below for these two PMT potentials and as
interpolated for some nearby potentials: The voltage values assume a 93 ohm
load.
PMT volts
1500 (1600) (1700) 1800 (1900)
Percent dev.
from linear.
* 2%
5%
20%
Actual anode
output
max. volts
max. ma
volts
ma
volts
ma
1.8
19
2. 1
23
3.5
37
2.1
22
2.4
26
3.8
41
2.4
26
2.7
29
4. 1
44
2.7
29
3.0
32
4.4
47
3.0
32
3.3
36
4.7
51
20
-------
4. TIME-SQUARED AMPLIFIER
As previously indicated, a time-squared amplifier has been developed to
range-correct (linearize) the lidar return as received from the lidar photomulti-
plier tube detector.
4. 1 Amplifier Design
The time squared amplifier (TSA) circuit is shown as Figure 4. 1. The
basic concept of the circuit is to use a linear ramp plus analog multiplier Ml to
first provide a time-squared ramp (and thus range squared - see Section 2. 1),
and to then combine this t^ output from Ml with the amplified photomultiplier
tube signal (from operational amplifier A521) in a fast analog multiplier M2,
yielding from M2 a signal constant with range for a homogeneous scattering
atmosphere.
Still referring to Figure 4. 1, the detailed operation is as follows: The
linear ramp which feeds Ml is derived from the 0 to + 150 volt sawtooth output
from the same oscilloscope (or any linear ramp generator) which is triggered by
the lidar laser monitor and which can also receive the TSA output if desired.
Because Ml and M2 each have a maximum input of 10 volts, the linear ramp is
reduced in amplitude, and selected by the "Far Limit" potentiometer to reach
10 volts at any desired fraction of the total linear ramp. A protective diode
located near T5 in Figure 4.1 limits the voltage reaching Ml to a maximum of
10 volts to avoid overdriving Ml. Since these multipliers both have outputs
which are XY/10, the linear ramp input to both X and Y of Ml yields a ramp-
squared (t^) output with a maximum of 10 volts. This t^ output ramp from Ml
(at T6) is then multiplied in M2 with the amplified and inverted lidar PMT output
so as to yield the range-corrected output. The variable dc input level applied at
T4 of figure 4. 1 allows fine adjustment of the ramp time-zero so as to optimize
the "flatness" of the final range-corrected output.
The negative-going photomultiplier tube*("PMT") lidar output current
passes through a 93 ohm load resistor (between Tl and T2 of Figure 4. 1) to the
100 Me (gain bandwidth product) inverting operational amplifier A521. Since
such operational amplifiers maintain their input (T2) at a virtual ground potential,
the 93 ohm resistor at Tl results in a proper impedance match for the RG62/U
coaxial cable used from the PMT. The amplification factor of A521 is controlled
by the feedback resistors selected by the 'Gain" switch to be nominally 3X, 5X,
or 10X. The 300 ohm and 4-25 pf RC pair in each feedback loop permits separate
optimization of the A521 fast response at each gain setting. The amount of am-
plifier gain chosen is based on the PMT output voltage (at Tl) at the nearest lidar
range of interest. Because both the A521 maximum output and the M2 maximum
input are each 10 volts, the 3X, 5X, or 10X "Gain" setting will determine
whether the near range limit for t^ operation is at a PMT output voltage (at Tl)
of about 3V, 2V, or IV, respectively. The "Mode" switch merely permits by-
passing the entire TSA while providing the proper 93 ohm load.
21
-------
N
N
"FAR
LIMIT"
P.S.
[SM 100/15)
+ 15V -15V
dc dc
"SIGNAL
FLATNESS"
(0 to + 150V
sawtooth input)
-15V
"OUTPUT"
"MODE" 2
VT
-XI
"PMT" [
25K
"ZERO"
Figure 4. 1 Time-squared amplifier circuit for range correction of lidar return.
-------
To permit the PMT output voltage to exceed the above values of 3V, 2V,
or IV (depending on A521 gain) without overdriving or damaging amplifier A5Z1,
an automatic output limiting feature was added to the A521 circuitry. When
the A521 output reaches 10 volts and therefore approaches saturation, the diode-
containing section of its feedback circuit reduces the amplifier gain so as to
maintain its output at 10 volts. By preventing saturation and even possible
damage, this feature permits immediate recovery (in < 100 ns) from what
would otherwise be too large a voltage at Tl, as is always the case when the
lidar near range of interest is beyond the lidar return maximum. The diodes
from Tl to ground merely provide further protection against really large nega-
tive voltages ( > 10 V) at Tl and against any positive voltages accidentally applied
to Tl.
4. 2 Amplifier Performance
The time-squared amplifier (TSA) was first adjusted and tested in the
laboratory, followed by field tests on the lidar system.
With TSA capacitors Cl, C2, and C3 (Fig. 4. 1) adjusted for the best over-
all performance of amplifier A521, the 10% to 90% risetimes and falltimes (Tr
and Tf, respectively) were measured for both amplifier AS21 and also for the
TSA overall (i. e. for the A521 and M2 in series). For the latter measurements,
a time squaring effect was avoided by applying 4. 5 volts dc at T5 (Fig. 4. 1),
yielding about 2 volts dc at T6. Oscillograms for these measurements are
shown as Figure 4. 2, with the measured Tr = Tf values being s 15 ns, ~ 16ns,
and ~ 30 ns for amplifier A521 at 3X, 5X, and 10X gain, respectively; and Tr =
Tf = ~ 70 ns (about 5 Me bandwidth) at all three gain settings.for the TSA over-
all. The oscilloscope risetime of ~ 13 ns and pulse generator risetime of
" < 13 ns" limited the A521 measurements. The ringing seen at the leading and
trailing edges of the applied square wave (at Tl of Fig. 4. 1) are due to the fact
that the 93 ohm input at Tl of Figure 4. 1 did not match the 50 ohm output and
cable from the pulse generator.
When the time squared amplifier (TSA) is incorporated into the lidar sys-
tem (including the new F4084 PMT), its clear air performance is shown in
Figure 4. 3. Here and elsewhere, a BNC Tee connector at the TSA input plus
a short length of extra cable to the oscilloscope permits simultaneous recording of
both TSA output and TSA input (PMT output). Fig. 4. 3a and b show the clear air
return to be properly range corrected when the "Signal Flatness" (see Figure
4. 1} is correctly adjusted, while Figures 4. 3c and d show the distortions intro-
duced by opposite extremes in mis-adjustment of "Signal Flatness". In each
case, the near range correction is properly initiated (N in Fig. 4. 3) at that
point where the product of PMT output volts times TSA gain setting equals 10
volts. Any nearer range excess PMT output voltage is accepted by the TSA
without impedance matching or recovery problems.
23
-------
(0. 5 V/cm)
TSA input
(0.1 V/cm)
TSA output
(0. IV/cm)
TSA input
fO. IV/cm)
Figure 4. 2 Double beam oscillograms showing time-squared amplifier (TSA)
response. Sweep speeds are all 100 ns/cm. The lower (negative-going) trace of
each oscillogram shows the applied test pulse at the PMT (Tl of Fig. 4. 1). The
upper (positive-going) traces of each oscillogram show the A521 amplifier response
(probe at T3 of Fig. 4. 1) and the overall response (at TSA "output"). In each case,
the upper trace shows the superposed outputs at TSA gain settings of 3X, 5X, and
10X (see Fig. 4. 1). As described in the text, the TSA "output" is here not multi-
plied by time squared.
24
-------
0-
0-
TSA
PMT
TSA
PMT
(a) Gain = 5X, SF = 7
(c) Gain = 5 X, SF = 1
01
0-
TSA
PMT
TSA
PMT
(b)_ Gain = IQX^SF = 7
(d) Gain = 5X, SF = .
Figure 4, 3 Double beam oscillograms showing time-squared amplifier (TSA) lidar returns from relatively
homogeneous clear air. Each oscillogram is externally triggered by the laser monitor (not shown), and has a
sweep speed of 500 ns/cm. In all cases, the negative-going lower trace (at 1 V/cm) is PMT input to the TSA,
while the positive-going upper trace (at 0. 2 V/cm) is TSA output. The TSA output range correction starts at
N and ends at F. Under each oscillogram is listed TSA gain setting and "signal flatness" setting (SF).
Figure 4. 3a and b additionally show zero signal traces.
-------
Figure 4.4 illustrates the general performance of the ISA under three
extremes of lidar operation. Figure 4. 4a shows the TSA response to off-gating
the F4084 PMT (260 ns gate) during a lidar shot into clear air. Figure 4.4b
shows the TSA response (plus PMT response) to a lidar shot at a white wall
located about 800 feet from the lidar system while off-gating the F4084 PMT
(250 ns gate). The zero signal trace shows very little effect due to PMT after-
pulsing or electronics recovery. Finally, Figure 4. 4c shows the TSA response
to a lidar shot at the same white wall (same range) while not off-gating the
F4084 PMT. In this severely fast overvoltage case, we see that the TSA over-
voltage protective mechanisms are not able to prevent slow recovery and ringing.
Here, TSA recovery requires about 800 ns, and even then comparison with the
TSA zero shows some small tendency for signal wander at the right edge of the
trace. The apparent ringing of the PMT return in Figure 4. 4c is due to the
above-discussed severe overvoltage effect on the TSA, resulting in the TSA
temporarily not being able to maintain a proper 93 ohm load for the PMT and
its cable. The PMT itself recovers much faster than shown in Figure 4. 4c
from an ungated white wall return (see Fig. 3. 5a).
26
-------
TSA
(0. IV/cm)
PMT
TSA
(0. 2V/cm)
PMT
TSA
(0. 2V/cm)
PMT
Figure 4. 4 Double beam oscillograms showing time-squared amplifier (TSA)
performance under three extremes of lidar operation. Each oscillogram is exter-
nally triggered by the laser monitor, and has a sweep speed of 500 ns/cm. In all
cases, the negative-going lower trace (at IV/cm) is PMT input to the TSA, while
the positive-going upper trace is TSA output with TSA gain at 5X. Off-gating inter-
vals (250-260 ns) are indicated by G, white wall position by W, and the near and far
limits of range correction by N and F, respectively.
27
-------
5. LIDAR MEASUREMENTS OF TEST TARGETS
After incorporation of the new (F4084) photomultiplier tube (PMT) and the
time-squared amplifier (TSA), the modified mobile lidar system was evaluated
using glass, Plexiglass, screen, and opaque test targets at a range of 216
meters. These lidar measurements and also laboratory re-measurements of
the original test targets were made using methods and locations similar to
those previously reported in reference 1.
5. 1 The Test Targets and Their Laboratory-Measured Transmittances
As discussed in reference 1, the test target holder has a clear circular
aperture of 42 inches diameter, with the test targets themselves being 42-45
inches square. As indicated above, these glass, Plexiglass, aluminum screen,
and opaque targets are from those originally used* for both laboratory and
lidar measurements in 1971. The 1/4 inch thick Plexiglass-G was manufac-
tured by Rohm and Haas, and the 1/4 inch thick Parallel-0-Float glass was
manufactured by Libby-Owens-Ford.
These new "laboratory" transmittance measurements were made by
illuminating the targets from one side with a collimated white light source while
viewing the source from the other side of the target with the lidar receiving sys-
tem. The targets were alternately placed in and out of the beam while the PMT
(original lidar Amperex 56 TVP) output was read directly on a Keithly electrom-
eter. This procedure was repeated with the same light source blocked off to
permit subtraction of the small PMT dark current and target front lighting
signals. Because the lidar receiver bandpass filter is centered at 6943 A, the
laboratory transmittance measurements were made at the same wavelength as
the lidar laser output.
In the course of these new transmittance measurements, it was noted that
the originally used* "collimated" light source was actually very poorly colli-
mated (about 115 mrad full angle). The use of such a poorly collimated source
should have no significant effect on the absorbing target (glass & Plexiglass)
results, but it should lead to screen target transmittance results which are too
large. This is because non-collimated (angled) incident light rays which should
not be accepted by the receiver (itself collimated to 4 mrad full angle), can re-
flect off the screen wires to produce rays which are at such an angle as to be
accepted by the collimated receiver. These rays are in excess over those
which are desired for a correct measurement, and thus yield transmittance
values which are too large.
Due to the above reasoning, the new target transmittance measurements
were made with both the originally used* poorly collimated light source (~ 115
mrad), and also with a different and much better collimated light source ( ^ 2
mrad full angle). The results of both old and new transmittance measurements
28
-------
are shown in Table 5. 1. Since the collimated light sources both had beam diam-
eters of 2 inches, and the target clear diameter for lidar measurements is 42
inches, all measurements were made near the target centers where the lidar
transmittance measurements were normally made. The highly collimated meas-
urements were also made at many locations over the entire clear target area as
a check on target uniformity.
As seen from Table 5. 1, both the old and new 115 mrad collimator-deter-
mined screen transmittances are in relatively good agreement, while the good
collimator-determined screen transmittances are significantly smaller. This
is especially true for the "30%" screen which has a high ratio of wire area to
clear area. As expected, we also see that the quality of source collimation has
little if any effect on the measurement of non-screen targets. The slightly
smaller transmittance recently measured for Plexiglass as compared with the
original measurement, may be due to the effects of aging. Of the various target
materials employed, it would seem that Plexiglass should be the most suscepti-
ble to aging.
5. 2 Lidar Transmittance Measurements of Test Targets
As discussed above and in reference 1, the 42 inch diameter target holder
was placed on the roof of a small building at the Valley Forge Space Center at a
range of 216 meters, in the same manner as was used for the 211 meter tests
of reference 1. The target holder was arranged so that any specular reflection
was off axis, and the lidar system was aligned so that the transmitting and re-
ceiving axes crossed at the target range. This was not really necessary, but
it facilitated aiming the system and was easy to do. A lidar shot through the
center of the empty target holder showed no trace of a target holder return
pulse such as would be obtained from even a fractional percent of the laser beam
energy hitting the inside edge of the target holder.
In the above manner, each target was measured three times using the
PMT gated and often also once with the PMT not gated. In each case, the
oscilloscope was triggered externally by the laser monitor, with each oscillo-
gram recording both t^ amplifier (TSA) output and TSA input (via a BNC tee
connection at the amplifier input plus a short cable to the scope). For all of
these test target measurements, the TSA gain was set at 5X. Thus the oscillo-
grams were similar to those extreme cases shown in Figure 4. 4 as well as
intermediate cases to be shown later.
Since the TSA output is range compensated, the lidar-determined target
transmittance (TjJ is (see equation 2. 4)
TL = (A/B)1/2 = (Vf/Vn)1/2 (5.1)
where Vf and Vn are oscillogram voltages from the far side and the near side,
respectively, of the target. Since the transmittance measurement depends on
homogeneous atmospheric scattering, the results should be most reliable if the
29
-------
Table 5. 1 Summary of Laboratory-Measured Target Transmittances at 6943 A
Measurement
Source collimation
(mrad full angle)
Portion of target
Target:
Plexiglass -G #1
Plexiglass -G #2
Glass
Al insect screen #1
50% Al screen
30% Al screen
Old* (1971)
~ 115
Center
.915
.915
.861
.640
.515
.288
This
~ 115
Center
.644
.523
.292
contract
Center
.903
.858
.604
.486
. 192
2
Overall
. 897-. 908
(uniform) i
.599-. 611
.470-. 500
. 180-. 195
* See reference 1
30
-------
near side and far side measurements are made as close to the target as possible.
This closeness is limited by (a) the PMT off-gate width used, (b) the detection
electronics recovery time, and (c) also by the small afterpulsing which even
this new PMT has.
Although this new PMT (ITT F4084) has weaker and shorter-lived after-
pulsing than does the old PMT (Amperex 56TVP), the Vf measurement should
still preferably be made after the afterpulsing has decayed. Oscillograms from
white target lidar measurements show the major afterpulsing to be decayed
within 0. 3 JJ sec after the target return, with nearly complete decay requiring
0. 6 M sec after the target return. This is further discussed in Section 3. 3. 3
and illustrated in Figure 3. 5.
During the day of these test shots (12/1/72, weather clear, cool and
windy), the atmospheric scattering uniformity often changed very rapidly from
good to bad and back again. Use of the TSA makes the presence of these non-
uniformities very obvious. For the lidar shots made during these target tests,
the atmospheric uniformity was relatively good about 1/2 of the time, fair about
1/3 of the time, and poor (shots not useable) about 1/6 of the time. This prob-
lem of atmospheric uniformity provides an incentive for making the lidar meas-
urements as close to the target (or smoke plume) as is consistent with the sys-
tem performance discussed above.
With all the above considerations in mind, the lidar oscillograms of test
targets were measured as follows:
1. For comparison, Vf was measured during four different time intervals
( At) after the target,
(a) from the average return during At = 0. 25 - 0. 55 /I s
(b) from the average return during At = 0. 65 - 0. 9 JJ s
(c) from the average return during At = 0. 9 - 1. 9 fi s,
(d) from extrapolation to target time of the return after At = 0. 65 M s.
2. In all cases Vn was measured from the average return during 0. 1 -
0. 4 H s before the target.
For each measured range interval, the average lidar transmittances from all
similar lidar shots are listed in Table 5. 2 along with the average deviation of
individual transmittance measurements from the listed averages. For compar-
ison, Table 5.2 also lists the earlier (1971) average lidar transmittance values1
and the recent highly collimated laboratory measurements of target center trans-
mittance as listed in Table 5. 1. Most of the results shown in Table 5. 2 are
plotted in Figure 5. 1.
31
-------
Table 5. 2 Averaged Results of Recent Lidar Measurements of Target Transmittance.
(See Figure 5. 1 for plot)
Target
Avg. deviation
from averages
Clear air
it ii
Plexiglass #1
Plexiglass #2
Glass
ti
Insect screen#l
50% screen
it ti
30% screen
ti M
White paint
it ii
Black felt
ii ii
Gated
off?
yes
yes
no
yes
yes
yes
no
yes
yea
no
yes
no
yes
no
yes
no
1971
Lidar
Transna.1
--
1. 00
0.925
0.861
0.891
0. 672
0. 573
0. 685
0.409
0. 521
0. 247
0. 121
Recent
Lab.
Transm.
--
(1)
(1)
0.903
0.858
0.858
0.604
0.486
0.486
0. 192
0.192
(0).
(0)
(0)
(0)
Recent Average Lidar Transmittance* (modified lidar)
No.
Shots
--
4
4
4
3
1
3
3
1
3
1
4
' 1
3
1
At (Ms)=
.25-. 55
±. 010
1.008
0.999
0.941
0.863
**
0. 644
0. 528
**
0. 324
**
0. 274
**
0.071
**
At(Ms)=
. 65-. 9
±. 015
1. 001
0. 990
0.970
0.856
~. 875
0. 626
0. 517
**
0. 292
*#
0. 143
**
0. 064
0. 136i
At (Ms)=
.9-1.9
±. 016
0. 985
0. 984
0~~992
0. 839
~. 87
0. 611
0. 508
0. 506
0. 277
0. 264
0. 118
0. 078
0. 061
0. 103i
At(/Zs) =
extrap>. 65
± . 019
1. 020
1. 002
0. 927
0.873
~. 88
0. 646
0.543
0.566
0. 327
0. 303
0. 156
0. 164
0. 077
0. 082
Is)
* Vf was measured from avg. return during indicated time interval (At) after target.
Vn was measured from avg. return during 0. 1 - 0. 4Msec before target.
** Oscillograms off scale or not yet settled.
-------
x= Unmodified lidar (1971, ref. 1)
A = Recent meas., At = . 25 - . 55 JJ
o = Recent meas., At = . 65 - . 9 M s
D = Recent meas., At = . 9 - 1.9 H s
l.Or-
Figure 5. 1 Averaged lidar-determined target transmittance (TjJ vs.
laboratory-determined target transmittance (T). Only off-
gated results are plotted. At T = 0, the upper point of each
pair is the white target result and the lower point of each pair
is the black target result. The upper solid curve is for the
recent TL results with At = 0. 65 - 1.9 fl s (o and a ), while
the dashed curve applies to the unmodified lidar results (x).
33
-------
Figure 5. 2 contains some sample oscillograxns of off-gating lidar returns
obtained during the recent test target measurements. Figure 5. 2a shows a
clear air gated lidar shot taken through the empty target holder during a period
of relatively non-uniform atmospheric scattering. By contrast, Figure 4. 4a
shows relatively uniform atmospheric scattering as observed during a similar
lidar shot through the empty target holder and taken only two minutes after the
shot of Figure 5. 2a. The remaining oscillograms of Figure 4. 2 show sample
oscillograxns of similar lidar test shots taken through some of the targets listed
in Tables 5. 1 and 5. 2.
One comment to be made about the recent test target results is that 3 out
of the 4 Plexiglass oscillograms showed a far side (of target) return which in-
creased with range* These same three lidar shots all hit the Plexiglass sheet
at about the same spot (center), while the one, more normal, type shot had the
laser beam moved to an off-center location. These three abnormal lidar returns
(see Figure 5. 2d for example) may be due to a lensing effect which is discussed
in reference 1. For the above reasons, we discount the Plexiglass results in
arriving at subsequent conclusions.
Another observation is that the one non-gated glass target oscillogram
shows large near range variations in atmospheric scattering and thus in Vn.
This made the corresponding transmittance results only approximate, as indi-
cated in Table 5. 2.
Although far fewer non-gated than gated lidar shots were made in the
recent series of target measurements (Table 5. 2), the non-gated results are
generally comparable to the gated results. However, the gated results have the
advantage of being valid at At smaller (Vf closer to the semi-reflecting target)
than do the non-gated results. The non-gated shots yielded longer PMT plus
circuit recovery times which usually limited Vf measurements to At s 0. 9
microsecond.
A comparison of the recent non-Plexiglass gated lidar results (Table 5. 2
and Figure 5. 1) with the recent laboratory transmittance values show the most
consistent agreement to be from At = 0. 65 - 1. 9 M s« For the relatively trans-
parent low reflectance targets, the At = 0. 25 - 0. 55 /I s results are also very
good, but they get worse as transparency decreases and reflectance (and thus
PMT afterpulsing) increases. Thus, we find that, in general, we obtain the
best transmittance results when measuring Vf with At = 0. 65 - 1. 9 H s and the
PMT gated.
Now, a comparison of the recent gated At = 0. 65 - 1. 9 MS transmittance
results with the 1971 lidar results (Table 5. 2 and Figure 5. 1 excluding Plexi-
glass) shows the modified lidar system to give about as good results at Tj_, ^
0. 85, and much better than the original system at TL = < 0. 85.
34
-------
(a) No target, air non-uniform
(c) "50%" screen target
(b) Glass Target
(d) Plexiglass target
Figure 5. 2 Double beam oscillograms showing sample lidar returns during test target measurements while using
the new PMT and time-squared amplifier (TSA). Each oscillogram is externally triggered by the laser monitor
(not shown), and has a sweep speed of 500 ns/cm. In all cases, the negative-going lower trace (at 1 V/cm) is PMT
input to the TSA, while the positive-going upper trace (at 0. 1 V/cm) is TSA output with TSA gain at 5X. Off-gating
intervals (all 260 ns) are indicated by G, target holder position by T, and the near and far limits of range correction
by N and F, respectively.
-------
In conclusion, it thus far appears that (1) the modified lidar system gives
better results overall than the original lidar system; (Z) the modified lidar sys-
tem gives its best results when used (a) gated, (b) measuring Vn as close to
targets as possible, and (c) measuring Vf with At = 0. 6 - 2JJ sec after the target;
and (3) when used as suggested, the modified lidar system transmittance (TL)
results (a) are precise to £ 0. 015 average, (b) are accurate to +_ 0. 02 average
at T > 0. 5, and (c) are too large at T < 0. 5. "~
These results are to be compared (see Figure 5. 1) with the pre-modifica-
tion accuracies1 of + . 09, -0 at T > 0. 50, and more severely too large at T < 0. 5.
36
-------
6. OPERATING INSTRUCTIONS FOR LIDAR MODIFICATIONS
The originally issued instruction manual still applies to the use of this
lidar system except where modifications have been made as discussed in this
report. Thus all of Table 2. 1 and nearly all of Figure 2. 2 of this report still
apply (Table 1. 1 and Figure 1. 4 of reference 3). A detailed discussion of the
new detection system characteristics and circuitry is covered in Section 3 of
this report, while Section 4 similarly covers the range-correcting time-squared
amplifier (TSA). Thus this Section 6 is concerned with how the lidar modifica-
tions influence the overall system hookup, and necessary or potential circuit
adjustments in the detection system and TSA.
Figure 6. 1 shows a detailed block diagram of an externally triggered
method for simultaneously recording the lidar data in both direct and range-
corrected forms. It should also be noted that Figure 6. 1 shows slight modifica-
tions in the laser energy monitor /trigger system as compared with Figure 1. 9
of reference 3. Except for changing the energy monitor calibration curve, this
laser monitor modification results in no alteration of instructions^ for its use.
6. 1 Use of the Modified Lidar Receiver System
6. 1. 1 Photomultiplier Tube Operating Voltage Selection
As discussed in Section 3. 3. 4 there is a space charge linearity
limit on photomultiplier tube (PMT) output current (and thus on voltage across
the 93 ohm load) that depends on the PMT overall operating voltage, increasing
as the voltage is increased, but increasing far slower than PMT gain (Fig. 3. 2)
with respect to applied voltage. Thus, it is possible to select a PMT applied
voltage that will produce a non-linear near range signal. In the extreme it may
be possible to produce a sufficiently large, non-linear, near range signal that
the dynode capacitors are depleted and not only the near range but all greater
range signals will be non-linear.
The optimum applied voltage is that which yields a near range signal
maximum equal to or slightly larger than the space charge linearity limit of 2
to 3 volts (93 ohm load) since this will allow maximum system range. This
applied high voltage is determined empirically by observing the lidar maximum
signal and adjusting the PMT voltage until the condition is met. The voltage re-
quired will vary day to day depending on the ambient aerosol scattering,but a
good trial value is 1700 volts.
For any applied voltage, the PMT dc output should be < 5 percent
of the divider chain current obtained by dividing the applied tube voltage by the
divider chain resistance of about 1. 16 megohms. If it is not, tube operation will
no longer be linear at any range. For 1500 volts applied, the dc output thus
should be < 65 fJ a ( < . 006 V with a 93 ohm load. The use of larger applied
PMT potentials would result in proportionately larger dc output limits.
37
-------
At
oscilloscope
TRIG.
MONITOR
LOAD
RG62/U
l^red filter
-/V>ky^-^~ xz
.03
Hh
}
input
sawtooth
out
+ gate
out
DOUBLE BEAM
OSCILLOSCOPE
upper lower
beam beam
H. P.
4220
(photo -
diode)
100K
-VW 11
VLASER MONITOR
On laser
-^
RG62/U
output
saw
PMT
TIME
SQUARE
AMPLIFIER
RG62/L
s~
H.V.
POWER SUPPLY
RG59/U
RG62/U
trig.
input
PULSE
GEN.
-5V
pulse
out
output
ITT F4084
PMT
pulse input
-HV input
RG
58/U
Figure 6. 1 Lidar data recording system with double beam oscilloscope
presentation of photomultiplier tube (PMT) output, both direct
and range -corrected via the time-squared amplifier. This
figure also shows the laser trigger/energy monitor system
in the trigger mode.
38
-------
For the reasons discussed in Section 3. 2, the total potential applied
to the PMT should not exceed 2500 volts. In the event that the PMT is operated
outside of the 1500 - 2000 volt range (see Section 3. 3. 2), the applied gating pulse
or PMT grid,and perhaps the PMT focus electrodes, should be re-adjusted for
the new voltage range as described in Section 6. 2 below.
6. 1. 2 Obtaining the Proper Off-Gating Pulse Position
Before transmittance measurements can be made, it is necessary to
establish the correct position for the gating pulse. This is done by using the
lidar system as a rangefinder. The lidar is aimed at the smoke plume (or the
top of the stack if it is diffusely reflecting) and is fired with the oscilloscope
recording the PMT output at lOV/cxn. For this purpose, use a simple 93 ohm
load at the oscilloscope or set the TSA "Mode" switch on direct ("Dir", see
Fig. 4. 1). The oscilloscope sweep should initially be set at 0. 5 M sec/cm
(ranges of 2500 ft. or less) and should be decreased if no sharp spike return is
visible. The return from the plume or stack will generally produce a saturated
PMT output of from 10 to 20 volts which will not faithfully reproduce the width
of the plume, but will give a sufficiently good range value so that by comparing
the ranging oscillogram with the variable position of the -5 volt gating pulse, the
gating position can be set.
The -5. 0 volt gate position and gate width can most easily be checked
and adjusted between lidar shots while using the same 93 ohm load or TSA direct
mode setting as used above for determining target range. However, now the
oscilloscope sensitivity should be set at 0. 01 or 0. 02 V/cm and the oscilloscope
trigger should be set in the automatic mode. In this mode the. oscilloscope
triggers at a sweep frequency of 60 Hz. This provides a gate out which triggers
the pulse generator providing a gating pulse to the 50 ohm load in the PMT
housing. The derivative of the gating pulse is weakly coupled to the anode cir-
cuit and thus is available as a negative going and a positive going set of spikes
defining the gating pulse position and duration. It is recommended that an off-
gating pulse width of at least 200 nsec be used because of range uncertainties
combined with the non-rectangular temporal shape of the laser pulse and appre-
ciable thickness of the smoke plume.
6. 1. 3 Taking Transmittance Data
As detailed in Section 4 and indicated in Figure 6. 1, the range-
correcting time-squared amplifier (TSA) must receive an appropriately timed
linear ramp. This is accomplished by connecting the lidar oscilloscope (Tek-
tronix Model 551) 0 to + 150V "sawtooth out" to the TSA "saw" (sawtooth input)
BNC connector (Figure 4. 1). Also, the RG62/U coaxial cable from the PMT
anode "output" (Figure 3. 1) is connected to the TSA "PMT" (PMT input) BNC
connector without the use of any external load resistor. Finally, a short length
of RG62/U coaxial cable is run from the TSA "output" BNC connector to one of
39
-------
the oscilloscope vertical inputs which is normally set at 0. 1 or 0. 2V/cm sensi-
tivity. The oscilloscope other vertical input may be connected to the TSA "PMT"
input via a BNC Tee connector plus a short length of RG62/U cable as shown in
Figure 6. 1. When the oscilloscope is triggered externally as shown in Figure
6. 1, this general method yields oscillograms similar to those shown in Figures
4. 3, 4. 4, and 5. 2. To trigger in this manner, set the monitor on "trigger" and
attach it to the oscilloscope "trigger input" with trigger settings being slope =
"EXT +", Mode = "AC (LF reject)", stability = "preset", and level - slightly
positive. An alternate display option is to connect the second oscilloscope ver-
tical input to the laser energy monitor ("energy" setting) and trigger the oscillo-
scope internally.
Before the first lidar shot using the TSA, preset its controls as follows
(see Figure 4. 1): set "mode" switch to "XT2", set "gain" to "5X" and set "signal
flatness" to 7.0 (range = 1 to 11). Although the TSA will operate almost immed-
iately after being turned on, the screwdriver lockable "zero" adjustment has
been pre-set to an optimum value for the TSA after it has warmed up for at least
10 or 15 minutes. The TSA zero signal level noise output with grounded "PMT"
input is about 8 mV pp (about 1.5 mV rms), with the zero signal flatness being
influenced by the "zero" setting as discussed in Section 6. 3. The screwdriver
lockable "far limit" adjustment has been pre-set for a far limit of range com-
pensation at 0. 9 scope sweep from time zero when the sweep is set at 500 ns/
cm and the TSA "signal flatness" is set at 7. 0. This 0. 9 sweep far limit is
slightly altered by other settings for signal flatness and scope sweep. See Sec-
tion 6. 3 for checking and readjusting the "far limit". (For the lidar oscillo-
grams shown in Figures 4. 3, 4. 4, and 5. 2, the TSA "far limit" (F) had been
set closer to 0. 8 scope sweep).
A sample lidar shot can now be made into clear air alongside the plume
target and the TSA "signal flatness" adjusted if necessary. Figure 4. 3 shows
examples of lidar returns with "signal flatness" correctly adjusted and also at
the extremes of adjustability. As can also be seen from the oscillograms of
Figure 4. 3, the near range limit (N) for range compensation is determined by
that range (time) at which the product of direct PMT volts times TSA "gain"
setting drops to a value of 10. Thus the near range limit (N) can be varied by
an appropriate combination of choices of (a) TSA gain (3X, 5X, or 10X), (b)
PMT high voltage and thus PMT gain (see Fig. 3. 2), and (c) lidar laser output
energy. When selecting the final PMT high voltage and TSA "gain", do not
exceed the PMT linearity limits at the near range point N or the PMT high
voltage limits as indicated in Section 6. 1. 1 and detailed in Section 3. 3.
Because there is a surplus of TSA and PMT gain combinations available,
the laser output need never be pushed to even 0. 5 joule/pulse. In fact, operat-
ing the laser at outputs of only 0. 2-0. 4 joule /shot will contribute to longer
flashlamp life, longer life of optical components within the laser cavity, and
especially will result in somewhat narrower laser outputs ( £ 30 ns FWHH) than
40
-------
might be obtained if the laser is driven hard. A wide or multiple laser output
might require longer off-gating and a resulting greater requirement of atmos-
pheric uniformity for transmittance measurements.
With all of the lidar receiver adjustments now completed, the actual trans-
mittance-measuring lidar shots are made through the plume of interest. If the
lidar return trace shows ringing and slow recovery after the plume (see Fig. 4. 4c)
or is driven to a large off scale value at the plume (rather than dropping to zero),
then the off-gating pulse starts too late or ends too soon and must be adjusted.
6. 2 Internal Description and Adjustment of Detection System
It is not anticipated that the following procedures will be required, but
they are included in the event that the photomultiplier tube (PMT) must be re-
placed or it is desired to change its operating voltage outisde of the 1500-2000
volts range while off-gating. As delivered, all PMT adjustments were made at
an applied potential of 1750 volts while using -5. 0 volt off-gating pulses.
As indicated in Figure 3. 1, this ITT type F4084 PMT fits into a 15 hole
socket which handles all tube connections except for the photocathode and inter-
nal control grid, both of which have flying leads (external wires) running from
the cathode end of the tube. Also, the PMT anode is internally connected to a
General Radio type 874-B coaxial connector which projects from the center of the
PMT base. The anode output is then directly conducted to the PMT housing
"output" BNC connector via a semi-flexible lead which is press-fitted into the
874B connector. In addition, the PMT housing has an inner cylindrical "shield"
which is held at cathode potential via a flying lead soldered to a lug at the cathode
(open) end of the shield tube. This inner cylindrical shield is positioned and in-
sulated from the outer cylindrical housing body and other grounded housing parts
by a specially shaped Teflon spacer ring at each end. Also, because the PMT
fits fairly loosely in its socket, the PMT is braced near its cathode end with a
foam rubber plus Teflon tape spacer between the PMT and the rigidly held shield.
To remove the PMT, first disconnect all external cables to the housing and
then remove the PMT housing from the lidar receiver at the housing flange (3
Phillips head screws). Next, all three flying leads must be unsoldered at points
"K", "1C", and "Sh" from the divider /gating circuit (see Figure 3. 1) which is
located in the gray rectangular box at the base end of the PMT housing. Do not
attempt to remove the flying leads from the PMT. Also, the press-fit anode
wire must be pulled from the 874B connector at the center of the tube base. This
is also accessed from the gray box at the base end of the PMT housing. Once
these disconnections have been made within the divider/gating circuit box, the
black outer (ground) cylindrical tube and inner cylindrical shield can be pulled
off the circuit box flange after unscrewing the three 6-32 Allen head bolts
spaced around the circuit box end of the cylindrical housing. This exposes the
PMT which can then be removed. If it is only desired to expose the PMT but
not to remove it, then only the shield flying lead (white) need be disconnected
41
-------
at the shield or at "Sh", Fig. 3. 1. The inner shield and outer tube are then
removed as above while pressing lightly on the PMT face to make sure it does
not pull out with the shield tube.
Any time the PMT housing is removed from the lidar receiving system or
even if the entire detection system is removed including the interference filter
section (such as for a lidar system alignment check^), first turn off the PMT
high voltage power supply and disconnect all external cables from the PMT base.
This is especially important for the high voltage connector, as the PMT could
be damaged by exposure to excessive light if the high voltage were still on.
All PMT focus electrode and internal grid adjustment potentiometers ("pots")
are located inside of the gray rectangular box at the base end of the PMT housing.
Since all of these adjustments must be made with the back cover of the box re-
moved and the PMT operating at the desired high voltage, care must be taken to
keep fingers and metal tools away, and to make all adjustments only with a non-
conducting screwdriver. Also, avoid directing bright light into the open circuit
box because of small light leaks from the circuit section to the PMT section.
Normal indoor room lighting is OK.
To adjust the focus electrode pots or internal control grid pot (Fig. 3. 1),
it is also necessary to provide a suitably weak dc light source which does not
overdrive the PMT (see Section 3. 3. 4). This could be accomplished through
the entire detection system assembly which includes the interference filter. The
focus electrodes can be adjusted at low PMT output currents as measured with
an electrometer or by using a large value load resistor (10K - 100K) at the
oscilloscope. Adjust the focus electrode pots for maximum PMT output in the
order Fl, F2, F3, F4 and then repeat (see Figure 3. 1). These pots are num-
bered 1, 2, 3, and 4, and are accessible from the open back of the divider/
gating circuit box as indicated above.
The PMT control grid adjustment pot has a brown insulating extension
attached to its shaft, and a piece of transparent tape which locks the pot to its
present setting. To optimize the PMT off-gating for a different applied high
voltage, it is first necessary to remove the locking tape. Be sure the high volt-
age is turned off before removing or replacing the tape. Now the dc light source
must be adjustable and turned up to yield a measurable dc anode output signal
across a 93 ohm load at the oscilloscope while at the same time always keeping
the anode current less than 0. 1 of the divider circuit current. This is accom-
plished by setting the scope sensitivity at 0. 005 V/cm, and always limiting the
dc voltage drop from the PMT anode output across the 93 ohm load to no more
than 2 cm (10 mV) as shown in Figure 3. 3. While the -5. 0 volt off-gating pulse
is repeatedly applied to the PMT housing "pulse input", the control grid pot is
adjusted to give a symmetric anode output gating response such as is shown in
Figure 3. 3a.
42
-------
An alternate method for adjusting off-gate symmetry which avoids opening
the divider /gating circuit box and changing the grid adjustment pot, is to change
the -5 volt gating pulse to a different voltage which again results in symmetric
off-gating. The use of this simpler method may result in slight changes in
gating on/off ratio or in gating response time.
6. 3 Time-Squared Amplifier Adjustments
The most routine front panel adjustments have already been discussed in
Section 6. 1. 3. Except for the front panel "zero" setting, the other adjustments
will seldom if ever need to be performed. However, the procedures are includ-
ed here in the event that it is desired to change the time-squared amplifier (TSA)
operating characteristics, or if TSA components need to be replaced.
In preparation for the taking of lidar data, the TSA front panel "far limit"
control may be checked and altered if desired. On delivery, this lockable screw-
driver adjustment was pre-set for 9 cm of oscilloscope sweep with the sweep set
for 500 ns/cm and the TSA "signal flatness" set at 7. 0. The far limit for range
compensation is readily observed by temporarily applying -0. 1 to -0. 5 volts dc
to the TSA "PMT" input (TSA mode = XT2) and self-triggering the oscilloscope
by setting it in the "automatic" mode. The time-squared ramp will then be
presented to the scope via the TSA "output" connector, with the t2 ramp leveling
off at the far limit for t2 range correction (points F of Figure 6. 2a). As seen
from Figure 6. 2a, the far limit for range correction varies somewhat with the
"signal flatness" (SF) setting. It also varies somewhat with the scope sweep
speed. The upper trace of Figure 6. 2a shows the TSA far limit setting as
delivered. One convenient source for the dc voltage input could be a 1. 5 or 3
volt dry cell in series with a 1000 ohm resistor.
It may occasionally be desirable to check and adjust the TSA output zero
signal shape. This is easily accomplished by looking at the TSA output while
(a) having the PMT on and the lidar receiving system looking at its target (but
no laser shot), (b) setting the oscilloscope at its lidar-use sweep speed (usually
500 ns/cm), (c) automatically triggering the scope at its internal 60 Hz rate,
and (d) increasing the scope sensitivity to about 0. 01 V/cm. The top three
traces of Figure 6. 2b show typical zero signal returns with no signal at the TSA
"PMT" input. The TSA front panel screwdriver adjustable "zero" setting per-
mits optimization of the zero signal shape (flatness). This shape will stabilize
as the TSA becomes fully warmed up.
Analog multiplier Ml (see Figure 4. 1) has-a zero adjustment vernier pot
located inside the circuit box and accessed by removing the top cover. The sys-
tem operation is so insensitive to this adjustment that it should never need to be
corrected. However if the adjustment is checked or the Ml unit changed, that
pot should be adjusted (after full warm up) so test point T6 reads OmV when test
point T5 is grounded (first disconnect the sawtooth input from the oscilloscope).
43
-------
SF. = 7
SF= 11,7, 1
TSA output
(all IV/cm)
TSA output:
. IV/cm
*P .05V/cm
. OlV/cm
Figure 6. 2 Oscillograms showing some time - squared amplifier (TSA) adjust-
ments and characteristics with TSA "gain" at 5X. All oscillogram sweep speeds
are 500 ns/cm. In all cases, the range correction ends at point F. Figure 6. 2a
shows TSA output (all at IV/cm) with -0.4V dc at the "PMT" input and two "far
limit" settings. The lower group of three superposed traces all have the same "far
limit" setting but have the "signal flatness" (SF) varied (left to right) from 11 to 7
to 1. The top three traces of Figure 6. 2b show the TSA output shape and noise
when the "PMT" input is grounded. This zero signal output shape can be varied
and optimized with the "zero" control. The lowest trace of Figure 6. 2b shows the
simultaneous ramp signal at test point T5 of the TSA (see Fig. 4. 1).
44
-------
The gain-control feedback capacitors Cl, C2, and C3 of the TSA opera-
tional amplifier A521 (see Figure 4. 1) have already been adjusted for optimum
A521 response as described in Section 4. 2 and shown in Figure 4. 2a. These
capacitors should never need re-adjustment unless the operational amplifier is
replaced or the feedback system altered. If this re-adjustment procedure should
ever be necessary, apply a fast response square wave of -0. 1 to -0. 5 volts to
the TSA "PMT" input and observe the amplifier output at test point T3 with an
oscilloscope probe. Then adjust the variable capacitors Cl, C2, and C3 (they
are numbered) for optimum operational amplifier performance at gain settings
of 3X, 5X, and 10X, respectively. These capacitors and T3 are accessible by
removing the TSA box top cover.
45
-------
7. SUMMARY AND CONCLUSIONS
This study was intended to eliminate the dominant source of error in the
remote measurement of smoke - stack plume s using lidar, and to improve the
data presentation of the lidar transmittance measurement.
The use of a special grid-controlled photomultiplier tube and off-gating
system has resulted in a large reduction of photomultiplier tube afterpulsing
which has been the principal problem limiting the accuracy of plume transmit-
tance measurements.
A range-correcting time-squared amplifier has also been developed which
permits much faster and more convenient data reduction than formerly possible.
The above two developments have been adapted to the existing van-mounted
EPA lidar system, and the overall system accuracy has been evaluated using
both opaque and semi-transparent targets of known transmittance (T). These
test results show the lidar system to yield lidar transmittances (TjJ which (a)
are precise to +_ 0. 015 average, (b) are accurate to +_ 0. 02 average for T > 0. 5,
and (c) are too large at T < 0. 5 (for example, now TL = 0. 52 at T = 0. 50, and
TL = 0. 28 at T = 0. 20). These results are to be compared with the pre-modifica-
tion accuracies of + . 09, -0 at T > 0. 50, and more severely too large at T < 0. 5
(for example, formerly TL = 0. 59 at T = 0. 50, and TL = 0. 40 at T = 0. 20).
46
-------
8. REFERENCES
1. C. S. Cook and G. W. Bethke, "Design, Construction and Evaluation of a
Mobile Lidar System for the Remote Measurement of Smoke Plume
Opacity", Final Report, EEI Research Project RP39 and EPA Contract
No. 68-02-0093, December 21, 1971.
2. C. S. Cook, G. W. Bethke, and W. D. Conner, "Remote Measurement of
Smoke Plume Transmittance Using Lidar", Appl. Opt. j^, 1742 (1972).
3. "Mobile Lidar System Operating Manual", EEI Research Project RP39 and
EPA Contract No. 68-02-0093, January 1972.
47
-------
BIBLIOGRAPHIC DATA
SHEET
No.
EPA-650/2-73-040
3. Recipient's Accession No.
4. Title and Subtitle
Development of Ran ^e Squared and Off-Gating Modifications
for a Lidar System.
5. Repoit D«te (preparation]
September 1973
6.
7. Authoi(s)
GEORGE W. BETHKE
I* Performing Organization Re pi.
No- NONE
9. Performing Organization Name and Address
General Electric Company
Space Division - Space Sciences Laboratory
P.O. Box 8555 -- Philadelphia, Pennsylvania
10. ProieciXTask/Voik Unit No.
19101
11. Contract/Grant No.
68-02-0570
12 Sponsoring Organization Name and Address
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
II Type of Report & Period
Covered Finai Report
6/27/72 to 9/5/73
14.
IS. Supplementary Notes
16.
This study was intended to eliminate the dominant source of error in the remote
measurement of smoke-stack plumes using lidar, and to improve the data presentation
of the lidar transmittance measurement. The use of a special grid-controlled photo-
multiplier tube and off-gating system has resulted in a large reduction of photomulti-
pher tube afterpulsing which has been the principal problem limiting the accuracy of
plume transmittance measurements. A range-correcting time-squared amplifier has
also been developed which permits much faster and more convenient data reduction
than formerly possible. The above two developments have been adapted to an existing
van-mounted EPA lidar system; and the improved overall system accuracy has been
evaluated using both opaque and semi-transparent targets of known transmittance.
17. Key Uords and Document Analysis. 17o. Descriptors
Air Pollution; Smoke
Transmittance; Opacity
Optical radar; Optical detection
Remote sensing; detection
Photomultiplier tubes
Range gating
Data reduction, data processing
Data processing; data reduction, data acquisition
I7b. Idrntifiers/Open-Ended Terms
Remote measurement of plume transmittance
Photomultiplier tube gating
Lidar range correction
Time squared amplifier
17c. COSATl Field/Group 14B 9E
18. Availability Statement
19. Security Class (This
Report)
UNCLASSIFI
lAWIEft
Class (Thii
20. Security Class (This
"UNCLASSIFIED
21. No. of Page*
51
22. Price
FORM NTIS 15 110-701
48
UlCOMM-DC 40SH-P7I
------- |