-------
repetition  frequency  f   will  be  accurately  reproduced by a system whose
                      m

bandwidth accommodates the  triangle's  primary  Fourier components  at  f  ,
                                                                     m

3f  .  and 5f  .   For example, a VCO  with an input bandwidth of  100  kHz will
  m        m

not pass the third and fifth  harmonics of a 50-kHz  triangle wave,  and  its


resulting FM sweep will  resemble more  a sinusoidal  sweep at 50  kHz.  As


discussed earlier, a  nonlinear sweep like this tends  to  add spurious


sideband components to the  target  display of an FM-CW radar.  Therefore,


it  is imperative  that the sweep  repetition  rate fall  well within  the VCO


input bandwidth.



                Once a clean FM signal  has been generated,  it  needs to  be


sent  to both the  laser-beam intensity  modulator and the  mixer.   If the


connecting cables do  not have a  flat response  over  the sweep  frequency


range, an additional  AM  component  will be introduced  on  the FM  signal.


This  FM-to-AM conversion can  be  a  significant  cause of spurious sidebands


in  the target-spectrum display.  If buffer  amplifiers or power  amplifiers


are used to  interface the FM  signal with  other system components,  the


amplifier response must  also  be  uniform over the sweep region Af.



               The next key modulation component to consider  is the  laser-


beam  intensity modulator.   The most well  developed, off-the-shelf  com-


ponent is a  type of electrooptic crystal  such  as KDP  whose  optical trans-


mission changes with  increasing  voltage applied  to  opposite faces  of the


crystal.   The intensity,  I,  of the output changes with applied voltage,


V ,  according to the following equation:
 S



                        out      J
                       	 = sin

                         in



A judicious  choice of the bias-voltage  value V  and the  range of the argu-
                                               b

ment  [V  + V ] allows operation along  the nearly linear  portion of the

   2
sin X curve, so that the useful modulator function  becomes:
                                   46

-------
                             I
                             out

                            1— = KVs + Tc                         (22)
                              in



where K and T   are constants related  to the  shape  and  center point  of  the


optical transmission characteristic of the modulator assembly  (see  Figure


16).



                The applied voltage changes the  transmission of  the  modu-


lator in a linear, analog fashion.  KDP modulators typically offer  excur-


sions of 100:1  in transmission  at bandwidths up  to 10  MHz, and  even up  to


25 MHz with special driving circuitry.  Once again, the  nonlinearities

                                 2
introduced by the modulator's sin X characteristic and bandwidth can de-


grade the FM signal appearing on the  optical carrier in  such a  way  as  to


introduce spurious sidebands in the final target display.  The  better  the


reproduction of  the original sawtooth frequency  sweep, the closer the


lidar performance is to  that modeled  in Section  IV-A.  It should be clear


from the present discussion that the  design  and  specification of modulator


components is very important in reducing system  sideband noise.  It is


also important  to note here that since the transmission  of the  modulator


assembly is varied about an average value T  , the  average transmitted
                                           v/

beam intensity  is nonzero.  The resulting dc component in a square-law-


detected signal envelope has an important effect on system performance


and must be carefully considered during the  design of  the lidar receiver.





     2.    Laser Transmitter



          The incentive for using an  optical radar wavelength lies  in  the


nature of the intended target.   Since a radar designer's goal is to maxi-


mize the amount of signal power to be detected after reflection or  back-


scattering by the target, the designer should choose an  operating


wavelength smaller than the target dimensions.  For particulate targets


such as the aerosols and molecular constituents of the lower atmosphere,
                                   47

-------
 relatively high backscattering and cross sections can be obtained for



 wavelengths less than 10 ^im,  at which a mix of Mie and Rayleigh scattering



 phenomena  occur.   As the radar wavelength shortens,  the bulk-scattering



 cross  section of the particles rises, but the corresponding path attenua-



 tion increases until in the  far ultraviolet regions  the useful  lidar



 range  may  be too short to be  practicable.   On the other hand,  in the in-



 frared region the backscatterlng cross section is lower by an order of



 magnitude  or more.   A convenient middle ground is the visible region,



 particularly because presently acceptable techniques for judging smoke-



 plume  opacity are visual.  A  visual  technique is one involving  the human



 eye, whose response  to light  peaks at green wavelengths.   Because these



 visual  methods typically involve comparison of plume darkness against



 the scattered blue light from the sky,  a laser wavelength in the green



 is a reasonable  choice for a  CW lidar opacimeter that is to  complement



 or even supplant  existing remote-measurement techniques.   Also  important,



 most photodetectors  are  more  efficient  and  sensitive to green light than



 to longer,  red wavelengths.





           Power output levels from 10 mW to 1  watt can  be supplied by such



 commercially  available CW gas lasers  as  the HeCd and argon types  operating



 at various  wavelengths between  4416 and  5145 A.   The stability,  durability,



ease of handling, and  lower cost  of HeCd and argon lasers  also make  them



 likely prospects  for use in a mobile  or  field-portable  instrument.   CW



 tunable dye  lasers have  recently  appeared on the market  and may prove



suitable in  the present  application should  their prices drop significantly.



They might  also provide  a means  to identify the  chemistry  of the  effluents.



However, for basic plume transmission measurements at a single wavelength,



a tunable CW laser would be more  sophisticated than  is  required.





          The power level actually required from the  laser transmitter



is constrained at the  upper extreme by laser-performance and eye-safety
                                   48

-------
considerations  and  at  the  lower end by  receiver  sensitivity  and back-


ground noise.



          HeCd  lasers  are  now  available  that  can transmit up to 50 mW at


4416 A.  Argon  lasers  with rated power  outputs from  25 mW to 10 watts


can transmit  at  any  of a number of wavelengths:   5145, 4880,  4765, and


4579 A.  The  trend  toward  standardizing  transmissometer  sources to the

                                           o
green region  favors  use of the argon  5145-A line.  Such  factors as tube


life, cooling requirements,  line stability, and  electrical power drain


can limit the choice of such an argon laser to one of  the lower power


models for mobile or field-portable applications.  Also,  the ability of


the electrooptical modulator crystal  to  handle the rejected  beam power


during its attenuating half-cycles limits  the amount of  power that can


be transmitted  by an FM-CW lidar.



          Finally, eye-safety  considerations  indicate  that  the trans-

                                                              2
mitted beam intensity  should not exceed  approximately  1  m\V/cm .  This


number is derived from data showing the  interrelations between continuous


source radiance, retinal irradiance,  retinal  image size,  and thermal


damage (if any)  done before the eye's natural blink  reaction occurs.12


For example,  if  the  lidar  receiver design  shows  that one  watt of power


must be transmitted  to a target in order to get  a detectable return sig-

                                                              3   2
nal, the transmitted beam  s cross-sectional area should  be 10 cm  or


greater.   This  in turn means that the transmit lens  must  be  some 40 cm


in diameter in order for the one-watt beam to be eye-safe even for per-


sonnel looking directly into the transmit  aperture.  Given the receiver's


sensitivity and  the  target's lidar volume  backscatter cross  section, eye-


safety considerations  then lead to the first  step in the design of the


transmitter optics:  choosing  the output-aperture dimension.  The re-


mainder of the  transmitter's optical  design is straightforward and,


since it has  no  peculiar dependence on  the CW technique, will not be


pursued here.


                                   49

-------
          Internal laser noise is one  additional  characteristic  that  can



have a serious effect on CW lidar performance.  Since  the CW  lidar  target



information is contained in the frequency  spectrum,  the spectral  charac-



teristics of noise become important.   Appearing as either spectral  base



line "grass" or as spectral line-broadening  frequency  instability,  the



noise components can interfere with  the desired target signal spectrum.





          Laser amplitude noise can  be caused by  the 60-Hz harmonics



riding on top of the dc gas ionization or  exciter voltage.  Amplitude



noise can also appear in the VHF region of the spectrum at a beat fre-



quency between two laser modes.  Since both  of these AM noise sources



steal carrier power and introduce it as spurious  sideband components,



the designer must either specify components  that reduce the noise level



or choose modulation parameters so that the  spurious noise components



will not fall in the target signal spectrum, or both.





          In a similar fashion laser frequency instability can cause a



kind of FM noise that is converted into AM noise in  the laser itself.



The severity of this internal FM-to-AM noise conversion is difficult to



estimate and even more difficult to measure  in the presence of other



FM-CW system noise sources.   Laser manufacturers claim that laser FM



noise is generated by acoustical disturbances to the laser cavity caused



by the flow of cooling air or water, loud noises, or random motions of



the laser mount.   Since all these phenomena occur in the audiofrequency



range, such internal noise could show  up as  low-level audiofrequency



components on the detected high-frequency modulation envelope.  On  the



other hand, atmospheric effects can also convert the VHF frequency  in-



stabilities directly into VHF AM noise, as discussed later.
                                   50

-------
     3.   Optical Receiver



          The design of an optical receiver  for  incoherent,  square-law


detection of backscattered laser signals  has  been  amply  discussed  by


others.2'13  A few receiver characteristics,  however,  have effects on


performance that are peculiar  to an  FM-AM Ctt'  lidar and should  be examined


here.



          The modulated laser-beam intensity  v^s expressed in  Eq.  (22)  as
                         I
                         o
= I.  JKV  + T                          (22a)
   in|_  S    CJ
Since V  = V  sin[2nf t'j , where f   is  the  FM  sweep  function,  the  average
       S    o        o           o

value of I  KV  is zero.  However,  the  quiescent  modulator  transmission
          in  S

value T  is nonzero, and therefore  an  average nonzero  intensity equal  to
       O

I  T  is continuously transmitted.   Since  the laser modulator input  in-
 in C

tensity I   is very nearly constant, consider I   instead  as
         in                                    o
                         
-------
which can be redefined as two power components
                           P  = P  + P                              (24)
                            R    S    C
          The most sensitive, low-cost light detectors for use  in this


application are square-law detectors such as a photodiode or photomulti-


plier tube (PMT).  The current output of the detector is proportional to


incident power, so that in the case above, output current is:
                                                                    (25)
where i  is the FM sinusoidal function and i  is a constant.   The rela-
       S                                    C

tive value of i /i  is just the ratio of modulator parameters  K T /K V .
               C  S                                            C C  o o

This value typically averages 1.5 and higher for practical electrooptic-


crystal laser-beam modulators.  As a rule, the better the modulator's


reproduction of the input signal, the higher the value of K T  /K V  .  The
                                                           C C S S

resulting effect is that the average dc component, i } could exceed the
                                                    \j

continuous current rating of the detector for a bright lidar target.


Another result for a slightly smaller target return would be that i
                                                                   \s

could drive the detector nearly into saturation, thereby reducing the


effective gain for the smaller, information-bearing current, i .
                                                               O


          Even when both i  and i  are in the dynamic range of the de-
                          S      C

tector, i  creates a steady, additional shot-noise component above that
         Vx

of i .   In short, the relatively larger reflected power component P
    i>                                                              C

contains no information,  yet it can significantly reduce the lidar system's


signal-to-noise ratio (SNR) in the case of a strong target echo.  The dc


component can mask the returning, information-bearing FM sinusoidal sig-


nal, P .   The design of the lidar receiver should therefore take into
      b

account both the modulator parameters K T  and K V  and the maximum ex-
                                       C C      S S

pected target reflectivity.  As a practical matter, since the  lidar



                                   52
-------
operates continuously,  it lends  itself well  to  incorporation  of  an  auto-


matic gain control  (AGO to reduce PMT gain  for bright-target conditions.


An AGC could keep both  i  and  i   in  a range  where  i   can be separated,
                        s      c                  s

undistorted, from i  in signal-processing  circuits.
                   \^r


          Considering the distributed nature of the atmospheric  target


particles, an experienced lidar  designer would  point  out here that  cer-


tainly the backscattered signal  from ranges  very close  to  the receiver


would be decades stronger than those from  distant  targets, and would


therefore mask remote-target information whether there  is  an  AGC or not.


The solution for a pulsed lidar  is,  of course,  to  make  sure that the re-


ceiver's field of view does not  intersect  the transmitted  beam for  some


appropriate distance from the receiver.  This is not  a  complete  solution


for a CW lidar, however.  The continuous average signal effects  discussed


in the foregoing paragraphs combine  in the receiver's field of view to


introduce a spatial integration  effect on  the CW lidar's distributed-


target performance at any range.



          The modulated laser beam,  frozen in an instant of time, would


appear as a column of light with  alternating light and dark sections


spaced at the modulation wavelengths (Figure 17).  For a constant modula-


tion frequency f , the peaks are  spaced by \ = c/f .  A receiver viewing
                o                            o     o

one thin slice of one section will see the full  maximum and minimum value


of the intensity as the column is allowed to proceed  across the  field of


view at the speed of light.   As  the  field of view  is  increased to include


one-half wavelength, the receiver will capture  greater total  power, but


the relative ratio of the total's maximum to its minimum will  fall  be-


cause, just as a dark region is  leaving the  field of  view, a  light  region


is entering.   When the receiver  views a full  wavelength, again the  total


average received power increases, but the minimum value of intensity for


the dark section is exactly  offset by the maximum values being received


from the light section at any instant of time.


                                   53
-------
             INTENSITY I  (X) = K sin  |27T — I + K T
                       O           I   -v  /    c C
                                                                              LASER

                                                                          TRANSMITTER
                                         WAVELENGTH OF  INTENSITY MODULATION
                    RECEIVER
                         (at   SIMPLIFIED ANALYTICAL MODEL
              (b)   A TYPICAL EXAMPLE OF SPATIAL INTEGRATION GEOMETRY
                                                                                  ia



FIGURE  17   SPATIAL  INTEGRATION OF BEAM  INTENSITY  FUNCTION BY RECEIVER
                                          54
-------
                            In  the  expression  for  intensity  of  the  frozen  column of  light


                  as  a  function of  length,  it  is clear  that  every time  length  x increases


                  by  one  wavelength \ ,  one cycle  is  completed:
                                                                                     (26)
                  Of  course,  when the beam is  allowed  to  move  again,  the  length  x  traversed


                  by  a  modulation peak at the  speed  of light is  just  x =  ct,  which,  when


                  substituted in the sine argument  above,  gives  Eq.  (23)  for  I  (t).
                                                                             o


                            The lidar receiver field of view,  Q,  encompasses  at  any  instant


                  some  extent of beam length Ax where  Ax  = 0R  at range R.   The power re-


                  ceived  is  just the integral  of the intensity function over  the spatial


                  extent  AX  that is  being viewed:


                                         .x+Ax
                                     Ki  J
P  = K   J      K V  sin|2TT —I dx + K T  dx          (27)
 R    If       SO    I    \  I       C C
                       P   = K K V
                                   \>
                        R    1  S O  2n
\
/2n
| —
 \
       cos| — xl - cos
K T  Ax    .   (28)
                  The power received at  an instant  of  time  is  clearly  a function of  field


                  of  view,  because  Ax is the  extent of the  beam subtended by the field  of


                  view at  range R.   Also,  since  the value of x is arbitrary,  depending  on


                  where the beam is frozen in space and phase,  Eq.  (28) can be  simplified


                  by  letting x = X  /4 so that:
                                 o
                                            *o     /2n  \
                                DR = KiVo to  cos(r Axj+ KiKc
                                   Ax    .            (29)
                  As  the  field  of  view is  widened,  the  term K K T  Ax increases  linearly.
                                                             J. kj v>

                  However,  the  cosine term periodically peaks at K K V  \ /2n every  time
                                                                  ISO   o

                  AX  reaches  an integral multiple  of  wavelength \ .   The peak value  of  the
                                                                 o


                                                     55
-------
information-bearing cosine signal can be expressed as a fraction P    of


the nonzero average power:
                   rel   K K T <0R)    rel   i
                          ICC               c

It is clear that as receiver field of view increases, the detected sinu-


soidal signal current drops in relation to the average dc current.  Since


the average dc component is a noise source and a limiting factor as dis-


cussed earlier, an increasing field of view acts to degrade system SNR


independent of optical background noise.



          Therefore, in a CW lidar system, the receiver ideally must view


less than one-half wavelength of the modulating frequency f .  As the


field of view widens, the stronger de-induced noise swallows up the fre-


quency component f .
                  o


          Similarly, when f  is a sweep function of time, f (t), in order


for the receiver's response to a distributed target to be uniform for


each frequency in the sweep range Af, the receiver must view at most a


fraction of the wavelength of the highest modulating sweep frequency,


f
 o,max


          As an example, assume that the maximum sweep frequency is 10


MHz, so that Ax < 15 m is the spatial integration constraint.  Assume


also that the target is at 200 m, and that the receiver can be no further


than 5 m from the transmitter [Figure 17(b)].  If the receiver is to view


only Ax, its beam diameter at the target must be
and therefore field of view Q should be smaller than
                                   56
-------
                             D   0.4
                             R =
which is not an unreasonable practical constraint.




     4.   Atmospheric Effects


          Just as pulsed-lidar performance  is affected by atmospheric

turbulence and nonuniform aerosol or particulate  concentration,  the C\V

lidar beam undergoes these atmospheric effects.   Unlike  the pulsed sys-

tem, however, the CW lidar continuously reads out the magnitude  of error

introduced by these effects.  For example,  if,  in the idealized  lidar

of Section IV-B, the backscattering coefficient of  the near clear-air

volume changes while that of the far volume does  not, the continuous

readout of [log P     - log P   ] would vary accordingly about an average
                 near        far

value.  A chart recording of this relative  reading  would show  the magni-

tude of excursions from the average value,  giving a quick and  reliable

measure of system accuracy under changing atmospheric conditions.  An

operator would know in a minute or two whether  conditions were too turbu-

lent for adequate plume-opacity measurements.   Indeed, microwave FM-CVV

radar has been used recently specifically to measure air turbulence,1 "*• lp


          Additional atmospheric phenomena  may  have deleterious  effects

on the beam's modulation envelope.  In this experimental study,  noise

from these peculiar effects did not appear  significantly above other sys-

tem noise levels.  Nevertheless the design  of an  FM-CW lidar should take

them into consideration.


          Particles and molecular constituents  of the atmosphere are not

at all stationary when they scatter light.  For example, molecular oxygen
                                                                     4
at 273° K consists primarily of molecules moving  at speeds between 10
      5
and 10  cm/s.  Recall that a moving target  or scatterer  shifts the fre-

quency of any signal that it reflects along its velocity vector.  For a


                                   57
-------
radar  signal  frequency  f  ,  velocity  v,  and  speed of light c,  the Doppler-


shift  frequency  f   is:
                                14
For  a laser beam  at  f   = 6  X  10  Hz,  where  f   is  the  optical  carrier
                     o                       o

frequency, oncoming  oxygen  molecules  could shift the received  signal fre-


quency by  as much as 2  GHz.   Since  the various  scatterers in a volume  of


clear air  are moving with a continuum  of  velocity  components along the


lidar-beam axis,  the Doppler  shift  in  the color of the reflected  laser


light would be  a  continuous spectrum  from 0  to  2 GHz around  the optical


carrier  frequency.   In  an optical FM-CW radar such as  that shown  in Figure


14,  the  random  Doppler  shift  in the optical  FM  sweep signal would  be on


the  order  of the  sweep  deviation Af obtainable  from the  tunable laser.


Therefore  a change in return-signal frequency (color)  due to Doppler


shift would be  similar  in magnitude to the intended frequency  changes


due  to target-echo delay time.   As a result, Doppler effects from  the


distributed atmospheric scattering target could cause  serious  inaccuracies


in the target range  profile in  a coherent, optical heterodyne  system.



           In an FM-AM-CW lidar,  however,  the range information is  con-


tained not in optical FM but  in the VHP or HF modulation envelope  around


the  optical carrier's intensity (Figure 15).  The  photodiode receiver


throws away the optical-carrier frequency information; changes  of  20 GHz

           14
in a 6 x 10   Hz  laser  frequency would not be detected in such  a receiver.


Only the HF envelope signal is  processed.  Its  highest frequency during


the  sweep  is on the  order of  10 MHz, and  it  too undergoes a Doppler shift


due  to atmospheric particle velocity:





                     f  = — — — - X 107 =  33 Hz

                         3  X  10
                                   58
-------
This shift is small compared to the value of Af = 5 MHz used in a prac-



tical FM-AM-CW lidar, and is also small compared to the resultant beat



frequencies, typically 100 kHz or so for a target range of 200 m.





          A second atmospheric effect, coupled with Doppler shift, could



conceivably add amplitude noise to the Doppler-frequency noise just dis-



cussed.  The severity of this new effect is difficult to calculate;



measuring it was one of the main goals in<\hls recent effort.  It is



another manifestation of the FM-to-AM noise conversion mentioned earlier.



The theory behind this second effect is cumbersome to express mathemati-



cally.  A qualitative explanation will help here.





          Those who have observed an object illuminated by a laser have



undoubtedly noticed a kind of speckle pattern overlaid on the object.



The pattern is caused by constructive and destructive interference of



wavefronts emanating from various point reflectors on the object.  Whether



a particular point will appear to be a bright or dark speckle is deter-



mined for a given beam diameter by (1) spacing between points on the ob-



ject, (2) distance to the observer's eye, (3) observer's pupil size, and



(4) the wavelength of the laser light.  When the observer moves his head



he sees the speckle pattern change, as if the object were glistening.



He is changing the angle between lines drawn from two given object points



to his eye.  As the two path lengths change, the apparent source bright-



ness goes through bright and dark interference cycles.  In effect,



changing the determinants of the speckle pattern causes intensity varia-



tions, or "AM noise" at the observer's eye.





          In a similar fashion, a stationary lidar receiver viewing



moving point scatterers will see a glistening image.  Its intensity



appears to vary as any two scatterers move into the field of view in such



a way as to cause interference fringes to move across the receiver aper-



ture.18  As the aperture is enlarged, the effect decreases because more






                                   59
-------
fringes are being spatially integrated over the aperture dimension, and



any new fringe moving in amounts to a smaller percentage of the whole



pattern.  Similarly, two scatterers in the field of view may have Doppler-



shifting wavelengths as well as shifting positions that can cause speckle



variations.  The glistening speckle effect caused by billions of indivi-



dual scatterers at various speeds, ranges, and angles-off-axis might be



complicated and serious enough to cause significant AM noise in the re-



ceived signal.  Theoretical analysis of this point becomes quite diffi-



cult, and in the absence of experimental data one could speculate that



the AM noise might be dominated by the optical Doppler shifts, which



vvould add noise with bandwidth up to 2 GHz on the return signal.  Or the



AM noise might be caused only over relatively small spatial regions, so



that only a very small receiver aperture or narrow field of view would be



adversely affected.  The effect might be serious for an optical FM-CW



lidar with its need for coherent heterodyning, while not showing up at



all in an FM-AM-CW system.





          As a practical matter, during this experimental study, no



serious Doppler or FM-to-AM conversion noise occurred to degrade system



performance.  If any of these noise phenomena occurred, their contribution



was below that of other readily identified and correctable system sources.



It is not certain, however, that an optical design for an FM-AM-CW system



somewhat different from the one used here would also be free of this



noise.  Nor is it clear that a purely optical FM-CW coherent radar would



escape these effects.  Future CW lidar designs should not ignore the



potential for such Doppler spreading and FM-to-AM noise conversion to



degrade system performance.  However, actual results show that at least



one FM-AM-CW lidar design is largely free of these adverse atmospheric



effects.
                                   60
-------
     5.   Signal Processing





          The foregoing sections have discussed the design of an FM-AM-CW



lidar without specifying the manner in which the received signal frequency



is subtracted from the transmitted signal, nor the method for displaying



and measuring target range and returned signal power.  Some theoretical



considerations of measurement uncertainty were discussed in Section IV-



C-l-a, but. an explanation of the signal-processing functions required in



an FM radar still needs to be presented.





          In Section IV-C-1 the advantages of VHP or HF frequency mixing



as opposed to optical, coherent heterodyning were pointed out.  Indeed,



in the FM-AM-CW technique any of a number of inexpensive, commercially



available HF mixers can be used.  Such mixers contain nonlinear elements



that multiply the transmitted FM signal envelope with the delayed, re-



ceived envelope to generate sinusoidal signals at the sum and difference



of the input frequencies.  Since only the difference component is of



interest the mixer must be followed by a low-pass filter that eliminates



the higher sum-frequency components.





          Practical mixers have limits to the dynamic range over which



their conversion of input-signal to ou,tput-signal amplitude is linear.



Care must be taken to guarantee that the photodetector output does not



exceed the mixer's maximum linear input level.  Otherwise, comparisons



of various frequency-component amplitudes will be inaccurate.  The mixer



will also introduce internal conversion noise around the signal during



the. heterodyning process.  If the input signal is small enough, mixer



noise can overwhelm it.  There is also a conversion loss through the



mixer that typically attenuates the signal power by a factor of five or



ten.  While mixers are functionally simple to use, their detailed input-



output characteristics can be quite important in the design of a CW lidar,



where low-level and wide-dynamic-range signals are expected.






                                   61
-------
          Once the mixer output has been obtained and sum-frequency com-

ponents filtered out, it remains to analyze  the spectrum of  the difference-

frequency signal and noise.  As discussed in the foregoing sections, this

spectrum contains the target range and reflected-power information  that

is of primary interest to  the operator.  Signal processing at this point

is required to perform the following functions:

          (1)  Determine the frequencies of  interest.

          (2)  Extract the target signal spectrum from the noise
               spectrum.
          (3)  Modify and  format those signals in such a way as to
               provide a convenient measurement of target charac-
               teristics.

          Function 1 is simple when the system's modulation parameters

and target characteristics are known.  A narrow bandpass filter tuned to

the known target frequency will eliminate all components except the

target-frequency component.  If there are three targets, three such

narrow-band filters can be used, as in the idealized CW lidar opacimeter

discussed in Section IV-B.  A portable AM radio does much the same thing:

it can be tuned to a predetermined broadcast frequency to process the

voice or music signals found there.

          When the target  spectrum is not known in advance, as in this

research and development effort, it is necessary to scan through the

frequency spectrum until an overall picture of the target range has been

developed in order to find any interesting spectral characteristics.  A

spectrum analyzer is a useful instrument for this purpose.  It sweeps a

narrow-band filter of some bandwidth B from zero frequency to some maxi-

mum frequency, displaying  component amplitude Y versus sweep position X.

While it is certainly beyond the scope of this report to explain spec-

trum analyzer theory, it can be said that the instrument is of great

utility as a development tool.   But since it cannot view three separate


                                   62
-------
frequencies simultaneously, one cannot always use the analyzer to make


instantaneous comparisons of the reflected power levels of FM-CW lidar


targets.  The analyzer takes some time to sweep from the frequency of


the near target to that of the far target.  During that sweep time,  the


target characteristics may have changed, and therefore the analyzer  sweep


rate introduces uncertainties in measurements referenced to  time-varying


signals.  For slowly changing signals, however, the scan rate may be fast


enough to reduce the uncertainty to an acceptable level.



          Functions 2 and 3 are straightforward electronic design prob-


lems and are not peculiar to the FM-AM-CW lidar technique.   The idealized


CW lidar for measuring smoke-plume opacity shown in Figure 7 includes


functional signal-processing blocks that are readily achieved with inex-


pensive commercial components.  In the CW lidar design the need for  compli-


cated signal processing can be reduced by careful design elsewhere in the


system.  A good choice of modulation and frequency parameters would  allow


the target spectrum components to fall away from, or in between, noise-


spectrum components.  For example, the so-called semiconductor 1/f noise


spectrum diminishes in strength at frequencies above 10 kHz, while the


modulator's power-amplifier signals may leak into the processor circuits


as noise in the frequency range above '1 MHz.  By choosing modulation rate


f  and deviation Af carefully, the designer can cause the desired beat-
 m

frequency spectrum for targets between 100 and 300 m to lie  between  50 and


500 kHz, thereby avoiding the given 1/f and leakage-noise spectra.
                                   63
-------
                        V  EXPERIMENTAL PROGRAM








     The objective of the experimental program undertaken here was to



determine the engineering requirements necessary to design, fabricate,



and operate a CW lidar with range and amplitude resolution sufficient



to obtain transmittance measurements on semitransparent targets at ranges



of 500 to 1500 feet.





     After surveying some of the analytical groundwork described  in the



foregoing sections, the intensity-modulation, or FM-AM-C\V, approach was



selected over the less easily implemented coherent optical FM-CW  tech-



nique.  The next task included the design and construction of a laboratory-



model CW lidar with flexibility for doing research over a wide range of



parameters.  With the completion of the lidar, preliminary system-noise



and performance-limitation measurements were made.  These measurements



allowed the specification of optimum operating regions, and in these



regions the first ranging and remote opacity measurements were made using



the FM-AM-CW lidar.  The details of these experimental program activities



and their results are included in the following pages.







A.   Design and Construction of a Laboratory-Model CW Lidar





     The FM-AM-CW lidar technique chosen is depicted in Figure 15, and



the modulator technique is detailed in Figure 16.  For convenient refer-



ence this system has been redrawn in Figure 18, where each component is



specified by function, manufacturer, and model number.





     Item 1, the sweep function generator, is needed to provide very



linear voltage ramps because the required linear frequency sweep f (t),



is patterned after this voltage waveform by the voltage-controlled
                                   65
-------
                                        POWER
                                       AMPLIFIER
                               ©
    TO BEAM-
EXPANDING OPTICS
                                          ARGON  LASER
                                                          ©
                                                                                             BUFFER AMPLIFIER
                                                               VOLTAGE   (?)
                                                             CONTROLLED
                                                             OSCILLATOR
      FROM
 RECEIVER OPTICS
               ©
PHOTOMULTIPLIER
      TUBE
   SWEEP
  FUNCTION
GENERATOR
©
                                                                                           LOW-PASS    ©
                                                                                             FILTER
      ITEM
      Alfred Model 325A tor Wavetek 114) Function Generator
      Exact  Electronics Model 7030 Voltage Controlled Generator
      Coherent Associates Model 30 Modulator Driver Unit
      Spectra-Physics Model 164 00 2-Watt Argon Ion Laser
      Coherent Associates Model 27 Electro-optic Modulator
      RCA Type 7265  (or 6810 A) Photomultiplier Tube
      Relcom Model M1 Mixer. 0.2-500 MHz
      2-MHz Five-pole Low-pass Filter
      Hewlett-Packard Model 8552/85538 Scanning Spectrum Analyzer
      Tektronix 453 Dual-Trace Oscilloscope
                                                             SCANNING   (»)
                                                             SPECTRUM
                                                             ANALYZER
 DUAL-TRACE
OSCILLOSCOPE
                                                                                                                      SA-1979-16
                  FIGURE 18    LABORATORY  MODEL FM-AM-CW  LIDAR EQUIPMENT  BLOCK DIAGRAM
-------
oscillator, Item 2.  The specified VCO  had  a remarkably linear voltage-


to-frequency transfer characteristic.   Its  primary  limitation was its


input bandwidth, which  tended  to round  off  the  sharp turnaround points on


a triangular or sawtooth sweep waveform.  This  input bandwidth of 200 kHz


limited the sweep repetition rate to  f  < 20 kHz in order to maintain
                                      m

sweep linearity and small  relative  turnaround time  t .   The majority of


measurements were made  using a 10-kHz modulation rate.   Buffer amplifiers


were used to reduce the load at  the VCO output  and  thereby preserve sig-


nal purity.



     The electrooptical modulator crystal (Item 5)  has the measured


voltage-to-transmission curve  shown  in  Figure 19.  The crystal presents


a 65-pF capacitive load, requiring  a  rather powerful driver stage to pro-


vide the ±50-volt excursion around  a  bias point at  75 volts that makes


maximum use of  the linear  portion of  the modulator  transfer curve.  The
CO

cr
o

it
o
Q
O
z
o
in
Z
<
a:
v-
             20   30   40
                         50   60  70   80  90  100  110  120  130

                          APPLIED MODULATOR VOLTAGE — V
140  150  160  170
                                                                   SA-1979-17
        FIGURE 19   MEASURED MODULATOR TRANSMISSION CHARACTERISTIC
                                    67
-------
specified power amplifier, Item 3, is designed to drive the crystal at



frequencies between dc and 10 MHz, a convenient range since the maximum



frequency available from the VOO is 11 MHz.  Moreover, spatial-integration



effects in the receiver field of view dictated that the half-wavelength



of the highest intensity-modulation frequency be no shorter than approxi-



mately 30 m to encompass reasonable smoke-plume targets.  This factor



also supports the choice of 10 MHz as an upper limit to the modulating



frequency.





     To achieve research flexibility, a durable, tunable, 2-watt Argon



io>% laser transmitter was specified (Item 4).  The unit provided plenty



of power for most experiments at 5145 A.  No internal laser noise could



be measured save for 60- and 180-Hz ripple variations in the laser output



intensity.  These variations were well below the 60-Hz noise introduced



by other components in the system such as the photomultiplier-tube power



supply (Item 6).  Only by purposely detuning the laser could one generate



a beat frequency between two laser modes, but even then these beat fre-



quencies were greater than 25 MHz, well out of the range of the overall



10-MHz system bandwidth.





     The photomultiplier tube (Item' 6) originally chosen for use as the



laser receiver had an S-ll photocathode that introduced too much dark-



current noise for adequate measurement of low-level backscatter from



clear air.  After switching to an S-20 phototube, it was possible to de-



tect light backscattered from clear-air volumes at 100-m ranges.





     Photomultiplier output current signals were buffered and mixed with



the FM sweep drive signal supplied by the VCO.  The output frequencies



of the mixer consist of the sum and difference of the two input-signal



frequencies.  Since only the difference frequency is desired, a low-pass



filter (Item 8) was used to eliminate the higher sum frequencies and pass
                                   68
-------
the lower beat-frequencies containing target-range information on to the


display devices (Items 9 and 10).



     Figure 20(a) shows the two kinds of data display employed in the


operation of the CW lidar.  The dual-trace oscilloscope display shows


the voltage waveform (bottom trace) that drives the VOO; therefore this


trace is an indicator of the FM sweep waveform, deviation Af> and repeti-


tion rate f .  The top trace is the output of the mixer showing the
           m

Fourier sum of all target-frequency components as well  as the error-


frequency components arising during sweep nonlinearities (note the blur


in the trace corresponding to  the  turnaround points of  the  triangular


sweep waveform).  The second photo in Figure 20(b) is the spectrum-


analyzer display (Item 9) of the frequency components comprising the top


trace of Figure 20(a).  Three  main target frequencies appear, each of


which is 10 dB above its surrounding sidebands.   Indeed, when these photos


were taken the lidar was viewing three  targets downrange, and Af had been


adjusted to reduce each target's sideband envelope so that  the three tar-


gets could be resolved on a display of  the beat-frequency spectrum.


Notice that the differences in signal power returned by various targets


show up as height variations from  component to component.   By measuring


Af, f , and f  from these displays, one can calculate target range from
     m       R

Eq. (7).  By measurement of the heights of frequency components on the


spectrum display, the relative power scattered back from the target ranges


can be determined directly.





B.   Target Display Characteristics



     The goal of the overall effort is  to develop a CW  lidar for the re-


mote measurement of smoke-plume opacity.  The measurement of the distance


to the plume is only incidental to this goal.  Nevertheless, since


analysis of FM radar performance shows that single, discrete targets will
                                   69
-------
      OSCILLOSCOPE TRACE SHOWING MIXER OUTPUT SIGNAL (top) AND VOLTAGE
       WAVEFORM THAT DRIVES THE FM GENERATOR.
                                       —50 kHz
                          BEAT FREQUENCY — kHz
   (b)   SPECTRUM ANALYZER DISPLAY SHOWING  THREE MAIN TARGET FREQUENCIES
       COMPRISING TOPMOST SIGNAL TRACE ABOVE.  NOTICE SIDEBANDS BELOW
       EACH TARGET SIGNAL. LARGE SPIKE IS THE ZERO FREQUENCY MARKER.
                                                               SA-1979-18

FIGURE 20   DATA DISPLAYS USED FOR REMOTE  TARGET MEASUREMENTS
                                   70
-------
generate spurious target components in the beat-frequency spectrum, and



since these spurious targets can introduce inaccuracies in the measure-



ment of the relative power reflected by two targets (Figure 11), it was



important to experiment with discrete targets at various ranges before



tackling the distributed targets presented by plume and clear-air scat-



tering volumes.







     1.    Discrete Targets





          The CW lidar equipment shown in Figure 18 was used for these



initial target measurements.  The modulated argon laser beam was expanded



to a three-inch diameter and transmitted via periscope through  the roof



of the laboratory and out along the target range shown in Figure 21.



Also on the roof and next to the steerable output periscope mirror was



the tripod-mounted lidar receiver, consisting of a 6-inch-diameter front



lens, a collimating lens, and a photomultiplier tube.  The mirror was



adjusted to direct the beams down the parking-lot range, and the receiver



was then aligned with the transmitted beam.  With the laser power output



reduced well below eye-safe levels, an assistant standing 90 m  from the



transmitter held a white card in the beam and began walking away from the



lidar site.  A series of photographs were taken of the spectrum analyzer's



display of the difference-frequency spectrum as the target moved slowly



out in range.





          The spurious sideband components generated around each of the



single-target returns are shown in Figure 22.  The CW lidar's mixer



output-frequency spectrum is displayed on a spectrum analyzer for nine



target positions at successively greater ranges.  Each photo presents



the frequency-component amplitude on a logarithmic vertical scale versus



range (beat frequency) on the horizontal.  This scale tends to over-



emphasize the sideband strength on first impression.   The large signal at
                                   71
-------
   BACK
  TARGET
REFERENCE
SEMI-TRANSPARENT
    TARGET
               STEAM
               PLUME
,•:••::.-.    . BLDG. 1 •.:};'•••»:.•::'.
                                          GLASS
                                          PLATE
                                       REFERENCE
                                                                 'BLDG.
                   OBJECT
                   TARGET
                       GLASS PLATE
                       REFERENCE
                                                                      SA-1979-19
     FIGURE 21
    LIDAR TARGET RANGE  USED FOR REMOTE  RANGE AND  OPACITY
    MEASUREMENTS
  left is the zero-frequency marker generated by  the  spectrum analyzer.

  Since the sawtooth  sweep repetition frequency employed  is f  =10 kHz,
                                                               m
  the sidebands  appear  at multiples of 10 kHz across  the  display.   The

  apparent sideband width is an artifact of the display device.

            Notice in Figure 22(a) that the sidebands  are approximately 20

  dB below the main target component at 50 kHz.   This  is  a result  of the

  fact that a target  located at a range node meets  the minimum-sideband

  condition discussed in Section IV-C.  In Figure 22(b) the target has

  moved to a range halfway between range nodes.   Notice the increase in
                                      72
-------
  (a)
  (b)
                              — 20 kHz

                             SA-1979-20
FIGURE 22   DISCRETE-TARGET SPECTRA
             SHOWING SPURIOUS SIDE-
             BAND  CHANGES  WITH  RANGE
                     73
-------
sideband levels.  As the target moves out in range, alternate photos



show the changes in the sideband power envelope for range-node locations



and nonrange-node locations.




          An interesting feature shown in Figure 22(h) is that just as



the target moves slightly off the range-node position in 22(g),  the first



sideband to the left of the main target component drops below the under-



lying system noise level.  That a particular sideband can be eliminated



in this way may be a useful feature in future lidar opacimeter designs.



When the clear-air signal is brought up above the system noise and the



spurious sideband induced by the target is eliminated in this way, any



remaining sideband amplitude will be contributed solely by energy back-



scattered from clear air near the target.  In a real smoke-plume measure-



ment, however, the target would not be a white card but a volume of dis-



tributed particle scatterers.  A tenuous steam plume near the experimental



target range provided an excellent opportunity for examining such a dis-



tributed target's characteristics when viewed by a CW lidar.







     2.   Steam-Plume Characteristics




          Figure 21 shows the location of this steam plume near  the ex-



perimental target range at SRJ.  Figure 23 is a photograph of the rooftop



lidar site after the transmit mirror and receiver telescope had  been



steered over to view the steam plume.  Notice that the beam's path is



clearly visible due to clear-air scattering all the way through  and beyond



the plume.  The plume itself is very reflective, and its bright  front-



surface reflection causes the plume target spectrum shown in Figure 24(a).



Here the two plume sidebands appear approximately 10 dB below the main



plume return.   Additional weaker sidebands can be seen at higher fre-



quencies, just above the system and background noise spectrum.
                                   74
-------
                                                            SA-1979-21
FIGURE 23   ROOFTOP LIDAR SITE
-------
                                      •50 kHz

                        BEAT FREQUENCY — kHz

STEAM PLUME TARGET RETURN RISES ABOVE NOISE SPECTRUM BY NEARLY
20 dB.  TWO LARGEST SIDEBANDS  ARE 10 dB LOWER. LARGEST SPIKE IS
ZERO-FREQUENCY MARKER.

  
-------
          The steam plume return is much like a discrete target return.


The moving water droplets in the plume do not appear to have caused any


severe Doppler spreading in the target spectrum.  Figure 24(b) is the


spectrum of a discrete target—a wall—near the steam plume.  The two


spectra are quite similar, leading to a preliminary conclusion that


multiple-scattering or Doppler effects on the beam as it passes through


a plume do not make the plume target spectrum significantly different


from a discrete-target display.



          The detection of light scattered from clear-air regions in


front of and behind the seemingly discrete plume target proved more dif-


ficult with the elementary CW lidar system built for these  initial efforts.


Unlike the bright return from the front surface of the plume, the back-


scatter from extended volumes of clear air posed a problem  unique to the


FM-AM-CW technique.





     3.   Clear-Air Targets



          The primary problem in detecting clear air in the presence of


a bright steam plume return was that the plume return masked  the clear-


air return.  Since target energy appears only as spectrally narrow side-


bands at multiples of f  in the beat-frequency spectrum, the  strong
                       m

sideband-energy contributions from the main plume return are  much greater


in amplitude than the contributions by clear-air scattering.  It was


necessary to study the clear-air backscatter first by eliminating any


bright targets other than the clear-air scatterers in the lidar's field


of view.



          This was done by steering the beam shown in Figure  23 off to


the left of the plume.  The receiver was then trained on a  segment of


the beam at about the range of the plume.  The target spectrum of this


return is shown in Figure 25(a) on a linear scale.  The main  target
                                   77
-------
                              (a)   MAIN-TARGET COMPONENT
                                   FROM CLEAR-AIR RETURN.
                                   LARGE SPIKE IS ZERO-
                                   FREQUENCY MARKER.
                                   NOTICE CHANGE IN COMPONENT
                                   AMPLITUDE AS DISPLAY SWEPT
                                   FROM -fR TO +fR
                              (b)   WHEN THE LASER BEAM IS BLOCKED,
                                   THE NOISE LEVEL IS SEEN TO DROP.
                              c)  CLEAR-AIR RETURN SIGNAL SHOWN ON
                                  LOGARITHMIC SCALE APPEARS 10 dB
                                  ABOVE NOISE.  LARGE SPIKE IS ZERO-
                                  FREQUENCY MARKER.
                              (d)  CLEAR-AIR RETURN ON EXPANDED
                                  SCALE, SHOWING THE MAIN COM-
                                  PONENT IS SPECTRALLY NARROW.
                  1.0 kHz
                                                           SA-1979-23
FIGURE  25    CW  LIDAR CLEAR-AIR RETURNS
                           78
-------
spectral component is several times greater than the surrounding noise



spectral amplitude.  When the beam is blocked, the spectrum shown in



Figure 25(b) results.  The noise level is lower by the measure of white



noise generated by PMT detection of the nonzero average intensity of the



reflected beam.





          On an expanded frequency scale (1 kHz/div) it can be seen that



the clear-air component is spectrally narrow  [Figure 25(d)].  This photo



shows a spectrum scan made with a 30-Hz filter bandwidth.  One can con-



clude that use of a similarly narrow filter to monitor the clear-air



sideband center frequency will give a similar SNR better than 6:1,



allowing adequate resolution of the 10-to-50 percent attenuation in the



clear-air return that would be caused by an intervening smoke plume.





          In order to get these clear-air returns it was necessary to



narrow the receiver field of view to reduce spatial-integration effects.



As discussed in Section IV-C-3, a CW lidar receiver viewing more than a



fraction of the intensity modulation wavelength will suffer a loss in



SNR.  The series of spectrum photos in Figure 26 shows the spatial-



integration effect on the SNR of the PMT output for constant field of



view and increasing frequency—i.e., increasing number of modulation



wavelengths in the field of view.  Each photo shows the spectrum of a



single modulation frequency component as detected by the PMT.  As this



frequency was increased, the component was recentered on the display.



A similar series would result if the frequency were left constant while



the receiver field of view was progressively widened.





          The top photo shows a 25-dB SNR obtained from clear air when



the receiver views a length of beam equal to only about 3 percent of the



modulation wavelength.  The following photos show how the SNR drops below



10 dB for a receiver viewing more than a full wavelength.  One can reason-



ably conclude that spatial-integration considerations are indeed important
                                   79
-------
                                                 EACH PHOTO SHOWS THE SPECTRUM OF
                                                 THE FREQUENCY COMPONENT USED TO
                                                 MODULATE  THE  INTENSITY  OF  THE
                                                 LASER BEAM, AS DETECTED BY THE PMT.
                                                 AS THIS FREQUENCY IS INCREASED, IT IS
                                                 RECENTERED  ON  THE SPECTRUM DIS-
                                                 PLAY.   SPATIAL INTEGRATION OF THE
                                                 SIGNAL OVER THE CLEAR-AIR VOLUME
                                                 BEING   VIEWED  RESULTS  IN   LOWER
                                                 SIGNAL-TO-NOISE  RATIO  FOR HIGHER
                                                 FREQUENCIES.   FREQUENCY INCREASES
                                                 FROM PHOTO TO PHOTO, MOVING DOWN
                                                 THE SERIES.
                                                                          SA-1979-24
FIGURE 26    EFFECTS OF  SPATIAL INTEGRATION ON S!GNAL-TO-NOISE  RATIO
              OF CLEAR-AIR RETURN
                                     80
-------
in the design of a CW lidar and in its subsequent performance in



distributed-target situations.








C.   Remote Opacity Measurements





     Having established that a discrete target, a thin-plume target, and



a clear-air volume of limited linear extent all have similar target



spectra when viewed by a CW lidar, it was possible to proceed to the



proof-of-principle experiments in remote measurement of lidar target



opaci ty.





     The lidar beam was again pointed downrange as shown in Figure 21.



An 18-inch glass plate was placed in the beam at  the fifth range-node



position from zero so as to minimize its sideband-amplitude contributions



at other range nodes.  The plate was oriented so  that its specular re-



flection did not return directly to the lidar receiver; instead, the re-



ceiver viewed only the diffuse scatter from the glass and dust particles



on the glass.  Since the glass was 90 percent transparent, most of the



beam continued downrange to the ninth range-node  position, where a ply-



wood board was placed in the beam.  The receiver  viewed the back target



through the front glass.  With both targets at node ranges, their com-



bined sideband levels were relatively low, as shown in the linear display



of spectral-component amplitude versus beat frequency, Figure 27(a).  It



can be seen that the plywood reference target is  about 1.26 times stronger



than the return from the front glass reference target.





     When a second glass is placed at the seventh range-node position,



it adds a third component between the two reference-target components,



as can be seen in Figure 27(b).  The important change from 27(a) to 27(b),



however, is the change in relative amplitude between the front and back



reference targets.  Now their ratio has been reduced to 1.02 from 1.26,



meaning that the light from the back reference—passing twice through
                                   81
-------
                                          (a)  FRONT GLASS REFERENCE  AT FIFTH NODE
                                              RANGE AND BACK REFERENCE TARGET
                                              (plywood) AT NINTH NODE RANGE SHOWN
                                              WITH NO INTERMEDIATE SEMITRANSPARENT
                                              TARGET.
                                      GRANGE
                                          (b)
                                          (c)
                                          (d)
INTERMEDIATE GLASS PLATE PLACED BETWEEN
FIRST TWO TARGETS AT NODE-RANGE SEVEN.
NOTICE THAT BACK-TARGET AMPLITUDE
DROPS BY 20%, MEANING THAT ONE-WAY
ATTENUATION OF GLASS PLATE IS 10%.
ALUMINUM SCREEN REPLACES CENTER GLASS
PLATE. BACK-TARGET LEVEL DROPS 43%
BELOW ORIGINAL, MEANING THAT SCREEN
IS 69% TRANSPARENT.  THE SCREEN ITSELF
DOES NOT APPEAR AS A SPIKE  BECAUSE
ITS REFLECTIVITY WAS MORE THAN 10
TIMES LOWER THAN THAT OF THE GLASS
AND PLYWOOD.
FINER ALUMINUM SCREEN MESH REPLACES
CENTER TARGET.  FURTHER REDUCTION
MEANS ONE-WAY TRANSMISSION OF 58%
FOR THE FINE MESH.
                                                                            SA-1979-25

FIGURE 27   CHANGES IN RELATIVE TARGET  POWER LEVELS CAUSED BY VARIOUS
             TARGET OPACITIES
                                       82
-------
the new center target—has been attenuated to 81 percent of its intensity



when only clear air lay between the references.





     From this one can surmise that the  two-way  transmission  through  the



center glass is 81 percent of the transmission through only the clear-air



path of Figure 27(a).  This corresponds  to a one-way  transmission of



/0.81, or 90 percent, which is remarkably close  to  the 90 percent trans-



mission of the plate-glass target when measured  in  the laboratory.  Such



a remote measurement of the opacity of a semitransparent target at a



range of approximately 160 m was first made with  this CW lidar in April



1973.





     Other semitransparent center targets were measured as well.  Figure



27(c) shows the spectrum for a 1/4-inch  aluminum wire mesh sprayed black



to reduce its reflectivity.  Indeed,  the return  from  the center target  is



not visible on this scale but the telltale drop  in  the amplitude of the



return from the back reference target allows calculation of a 69 percent



one-way mesh transmission.  Figure 27(d) shows still  further  reduction



when a 1/16-inch mesh is substituted  in  the center  target position.   Its



transmission, calculated from the remote spectrum display, appears to be



58 percent.





     More than 50 different measurements were made  of the semitransparent



glass and mesh targets, separately and in combination.  The results of



tests on these thin targets are given in Figure  28.   The chart is labeled



in percent transmission (one-way) and equivalent Ringelmann scale on  the



right.  The error bars represent the  range of opacity values  measured for



each type of target, the small-dots represent the actual remotely measured



value, and the large dots represent the value measured in the laboratory.





     The accuracy of measurements made with this first CW lidar trans-



mi ssometer was limited by a number of factors; the  effects of each factor



can be mitigated or eliminated by suitable refinements in technique and





                                   83
-------
     100
      80
    z
    w


    <
    
-------
subsequent engineering development.  The most important factor was the



nonlinearity of the so-called "linear" spectrum-analyzer display scale



used when making three-target opacity measurements.  The bottommost divi-



sion on this scale can represent anywhere from 0.1 division to 1.5 divi-



sions, depending on the signal gain taken in the spectrum analyzer's



front end.  Some of the extreme data points in Figure 28 result from



photos taken before this equipment limitation was discovered, for which



scale-correction factors were not known.





     Another limiting factor was the physical nature of the reference-



target setup.   High winds during two experiments tended to jostle the



front glass reference plate, steering its specular reflection near the



receiver aperture and causing a corresponding increase in the amplitude



of the front reference target.  Interestingly, an operator watching the



continuously refreshed target spectrum display could easily monitor this



changing reference level and write off any measurements as being subject



to error.





     A third measurement error resulted as predicted in Figure 11.  The



spurious sideband generated at Node 9 by the target at Node 5 added to



the amplitude of the main target component at Node 9 (and vice versa) to



give an inaccurate measurement of the height of one component relative



to the other.





     Finally,  there was the inaccuracy inherent in making centimeter-



stick measurements of traces in polaroid oscilloscope photos.  Earlier



project plans had called for construction of a bank of three narrow-band



intermediate-frequency amplifiers, each tunable to a target frequency,



for continuous electronic measurement of signal amplitudes.  Unfortunately,



other equipment considerations required that the beat-frequency spectrum



be scaled down from the 0-to-l-MHz range to the O-to-100-kHz range.  Since
                                   85
-------
the breadboard IF amplifier could not be scaled down accordingly, use of
this otherwise promising signalvprocessing technique had to be postponed.

D.    Discussion of Experimental Results
     The successive milestones in this experimental study can be listed
as follows:
     (1)  Preliminary analysis of FM-fW lidar.
     (2)  Design of laboratory-modei lidar.
     (3)  Detection and ranging to discrete targets,
     (4)  Detection and ranging to distributed scattering targets
          including clear air.
     (5)  Remote opacity measurement using discrete object and refer-
          ence targets.
     (6)  Proof-of-concept analysis and report.
In reaching each of these milestones, no fundamental obstacle to the use
of the CW lidar was found,  Every significant source of error or per-
formance limitation was traced to an engineering technique or component
that is amenable to improvement, given spfficient attention in a subse-
quent engineering development effort.  In reviewing the results, it is
unfortunate that remaining time and funds were not sufficient to accom-
plish a seventh milestone—namely, remote opacity measurement using the
detected clear-air volumes as reference targets,  Jn order to accomplish
this milestone, an improved, low-noise PMT receiver 4s required, as well
as a more refined signal-spectrum processor based on the three-AM-radio
concept depicted in Figure 7.  More realistic smoke pj,umes than the highly
reflective steam plume used here need to be employed i.n testing a refined
engineering model.
     These are all straightforward tasks for the fixture,  That this first
hardware assembly was not quite up to reaching th$ ideal seventh milestone
-------
is hardly an indictment of the technique, which passed the first six



milestones in a most positive manner.  It is encouraging to have learned



that the atmospheric Doppler and FM-to-AM noise-conversion effects feared



by theorists do not degrade system performance in practice.  It was sur-



prisingly easy to align and operate the eye-safe CW laser system, as



opposed to its pulsed-laser predecessor at an equivalently early stage



in its development.  And it was gratifying-*o be able to make reasonably



accurate remote measurements of target opacity between Ringelmann 0 and 1



even with the rudimentary system implementation possible within the con-



straints of this first-phase effort.





     The principal objective of the program was successfully accomplished:



the concept of CW lidar for remote opacity measurement was demonstrated



along with its capacity for detecting backscatter from remote volumes of



clear air.  The engineering requirements for design of a subsequent engi-



neering research model have been determined and reported.





     It is concluded that the FM-AM-CW lidar technique for detection,



ranging, and transmissometry through atmospheric particulates is feasible,



and that straightforward improvements, primarily in receiver detector and



postdetection signal processing, will allow accurate measurements of smoke



plumes thinner than Ringelmann 1.
                                   87
-------
                              REFERENCES
1.   D. W. Jackson, "Development of a CW Lidar for  the Remote Measurement
     of Smoke-Plume Opacity," Proposal No. ELD 69-165, Stanford Research
     Institute, Menlo Park, California (22 January  1970).

2.   W. E. Evans, "Development of Lidar Stack Effluent Opacity Measuring
     System," Final Report, SRI Project 6529, Stanford Research Institute,
     Menlo Park, California (1967).

3.   W. B. Johnson, "Lidar Applications in Air Pollution Research and
     Control," J. Air Poll. Contr. Assoc., Vol.  19, pp.  176-180 (1969).

4.   W. B. Johnson and E. E. Uthe, "Lidar Study  of  Stack Plumes," Final
     Report, Contract No. PH22-68-33 (NAPCA), SRI Project 7289, Stanford
     Research Institute, Menlo Park, California  (1969).

5.   P. M. Hamilton, "The Use of Lidar in Air Pollution  Studies," Intern.
     J. Air and Water Poll., Vol. 10, pp. 427-434 (1966).

6.   E. W. Barrett and O. Ben-Dov, "Applications of the Lidar to Air
     Pollution Measurements," J. Appl. Meteorol., Vol. 6, pp. 500-515
     (1967).

7.   W. D. Conner and J. R. Hodkinson, "A Study  of  the Optical Properties
     and Visual Effects of Smoke-Stack Plumes,"  Cooperative  Study Project,
     Edison Electric Institute and U.S. Public Health Service.

8.   J. E. Yocom, "Problems in Judging Plume Opacity—A Simple Device for
     Measuring Opacity of Wet Plumes," J. Air Poll, Control  Assoc.,  Vol.
     13, No. 1 (January 1963).

9.   C. S. Cook, G. W. Bethke, and W. D.  Conner, "Remote Measurement of
     Smoke Plume Transmittance Using Lidar," Applied Optics, Vol.  11,
     No. 8 (August 1972).

10.  M. I. Skolnik, Introduction to Radar Systems, Section 3.3 (McGraw-
     Hill Book Company, San Francisco, California, 1962).

11.  T. E. Honeycutt and W. F. Otto, "FM-CW Radar Range Measurement with
     a COg Laser," IEEE J. Quantum Electronics (February 1972).

                                   89
-------
12.   D. H. Sliney and B. C. Freasier, "Evaluation of Optical Radiation
     Hazards," Appl, Optics, Vol. 12, No. 1 (January 1973).

13.   M. Ross, Laser Receivers (John Wiley and Sons, Inc., New York,
     New York, 1966).

14.   E. T. Ebersol, "Radar Sees Gnats and StuffJ" Microwaves, p. 9
     (February 1973).

15.   "FM-CW Radar Detects Atmospheric Pressure," Electronic Design, Vol.
     21, No. 15, p. 24  (19 July 1973).

16.   A. Siegmann, Proc. IEE, pp. 1350-1356  (October 1966).
                                   90
-------
BIBLIOGRAPHIC DATA 1- Report No. Z
SHEET EPA-650/2-73-037
4. Title and Subtitle
FEASIBILITY OF A CW LIDAR TECHNIQUE FOR MEASUREMENT OF PLUME
OPACITY
7. Author(s)
Richard A. Ferguson
9. Performing Organization Name and Address
Stanford Research Institute
333 Ravenswood Avenue
Menlo Park, California .94025
12. Sponsoring Organization Name and Address
EPA, Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
3. Recipient's Accession No.
5. Report Date
November 1973
6.
S. Performing Organization Kept.
No.
10. Project/T«sk/Worlc Unit No.
1 1. Contract/Grant No.
EPA 68-02-0543
13. Type ol Report & Period
Covered
Final Report
14.
15. Supplementary Notes
16. Abstracts
    This  report describes the work  performed  for the Environmental Protection Agency  during the  initial  proof-of-
 concept phase of a  program  to  develop an eyesafe  CW lidar for remote measurement of the  opacity of smoke plumes
 from  industrial smoke stacks.  The analysis, design, construction, and evaluation of a laboratory model CW lidar  were
 performed under SRI Project 1979 from  30 May 1972 to 30 May 1973 under EPA Contract 68-02-0543  to deter-
 mine  the limitations and  potential of the technique.  The proof-of-principle experiments combine what  is called  an
 FM-CW radar technique with an argon laser.  The  technique involves modulating the intensity of the  laser beam at
 a frequency  that changes rapidly and linearly with  time.   A portion of the transmitted signal is mixed electronically
 with the light reflected from the targets in a device similar to  a radio receiver.  Each taiget appears at a particular
 frequency.   By tuning the radar's receiver  to these  target  frequencies,  the  researchers were  able to measure both the
 range and the opacity of semi-transparent  targets over distances of 100 to  200 meters.
17. Key Words and Document Analysis. 17o. Descriptors
   Remote Measurement
   Laser
   Lidar
   Optical  Detection
   Smoke-Plume  Opacity
   Measuring Instruments
   Air  Pollution Monitoring
   Smoke Abatement  Enforcement

17b. Identifiers/Open-Ended Terms
   Air  Pollution
   Lidar
   Clear-Air Scattering
   Measuring Instruments


I7c. COSATI Field/Group  14/02,  20/05, 4/01,  13/02
18. Availability Statement
Release unlimited.
19.. Security Class (This
Report)
UNCLASSIFIED
20. Security Class (This
Page
UNCLASSIFIED
21. No. of Pages
102
22. Price
FORM NTIS-SS (1O-70)
                                                                                              USCOMM-DC «OS20-r»7!
                                                       91
-------
   INSTRUCTIONS  FOR  COMPLETING  FORM  NTIS-35 (10-70) (Bibliographic Date Sheet based on COSATI
  Guidelines to Format Standards for Scientific and Technical Reports Prepared by or for the Federal Government,
  PB-180 600).

   1.  Report Numb«r.  Each report shall carry a unique alphanumeric designation.  Select one of the  following types:  (a)  alpha-
      numeric designation provided by the sponsoring agency, e.g., FAA-RD-68-09;  or, if none has been assigned, (b) alphanu-
      meric designation established by the  performing organization  e.g., FASEB-NS-87; or, if none has been established, (c)
      alphanumeric designation derived from contract or grant  number, e.g., PH-4J-64-932-4.

  2.  Leave blank.

  3.  Recipient's Accession Number. . Reserved for use by each report recipient.

  4-  Title and SubHtla. Title should indicate clearly  and briefly the subject coverage of the report, and be displayed promi-
      nently.  Sec subtitle,  if used, in smaller type  or otherwise subordinate it to main title.  When  a report is prepared in more
      than one volume, repeac the primary title, add volume number and include subtitle for the specific  volume.

  5.  Report Dote.  Eoch report shall carry a date indicating at least month and year.  Indicate the basi* on which  it waa selected
      (e.g., date of issue, date of approval, date 01 preparation.


  6.  Performing Organization Code.  Leave blank.

  7.  Aut! V,..).   Give name(s) in conventional order  (e.g.,  John R. Doe,  or J.Robert Doe).   List author's affiliation if'it differs
      from the performing  organization.

  8-  Performing Organization Report Number-  Insert if performing organization wishes to assign this number.

  9-  Performing Organization Name and Address.  Give name, street, city, state, and ?ip code.  List no more than two levels of
      an organizational hierarchy.  Display the name of the organization exactly as it should appear in Government indexes such
      as USGRDR-I.

  10.  Project/Task/Work  Unit Number.  Use the project, task  and work unit number* under which the report was prepared.

  11.  Contract/Grant Number. Insert contract or grant number under which report was prepared.

  12.  Sponsoring Agoncy Nome and Address.   Include zip code.

  13.  Type of Report and Period Covered.  Indicate interim, final, etc., and, if applicable, dates covered.

  14.  Sponsoring Agency Cod*.  Leave  blank.

  IS.  Supplementary Notes.   Enter information not included elsewhere but useful, such as: Prepared in cooperation with . . .
      Translation of ...  Presented at conference of ...  To be published in ...  Supersedes . . .        Supplements . . .

  16.  Abstract.  Include a brief (200 words or less) factual summary  of the most significant information contained in the report.
      If the 'eport  contains a significant bibliography or literature survey, mention it here.

  17.  Key Words and Document Analysis,  (a).  Descriptors.  Select from the Thesaurus of Engineering and Scientific Terms the
      proper authorized terms that identify the major concept of the research and are sufficiently specific  and precise to be used
      as index entries for cataloging.
      (b).  Identifiers and Open-Ended Terms.  Use  identifiers for project names, code  names, equipment  designators, etc.  Use
      open-ended terms -written in descriptor form for those subjects for which no descriptor exists.
      (c).  COSATI Field/Group.  Field and Group assignments are to be taken from the-1965 COSATI Subject Category List.
      Since the majority of documents are multidisciplinary in nature, the  primary Field/Group assignment(s) will be the specific
      discipline, area of human endeavor, or type of  physical object.  The application(s) will be cross-referenced with secondary
      Field/Group assignments that will follow the primary posting(s).

  18.  Distribution Statement.  Denote releasability to the public or limitation for reasons other than security for   example "Re-
      lease unlimited".  Cite any availability to the public, with address and price.

  19 & 20.  Security Clossificotion.  Do not submit classified reports to the National Technical Information Service.

  21.  Number of Pages.   Insert the total number of pages,  including this  one  and unnumbered pages, but excluding distribution
      list, if any.

  22.   Pric«.  Insert the price set by the National  Technical Information Service or the Government Printing Office, if known.
FORM NTt*-SB (10-70)                                                                                      USCOMM-OC 4O3ZO-P71

                                                             92
-------
ENVIRONMENTAL PROTECTION AGENCY
    Technical Publications Branch
      Office of Administration
  Research Triangle Park. N.C.  27711
     POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
           EPA - 335
        OFFICIAL BUSINESS

 AN EQUAL OPPORTUNITY  EMPLOYER
                                  Return this sheet  if you do NOT wish to receive this material I  I,
                                  or  if change of address is needed f~~l.   (Indicate change, including
                                  ZIP code.)
-------