THE RELATION BETWEEN AVERAGE PHOTOGRAPHIC DENSITY AND TRANSMITTANCE FOR FOUR CASES OF INTEREST.

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04747A000200010030-6
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 27
• Document Creation Date: December 28, 2016
• Document Release Date: March 22, 2001
• Sequence Number: 30
• Publication Date: October 8, 1964
• Content Type: REPORT

Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
STATINTL
October 8, 1964
HBH:bjs-431
To:
From:
Subject:
STATINTL
The Relation between Average Photographic Density STATINTL
and Transmittance for Four Cases of Interest.
Introduction
The analysis of density traces of photographic transparencies
raises a special problem when it is desired to determine the "average
density". The true average density, which is the average level of the
densitometer trace, is in general not equal to the value of density that
is approached when the area of the scanning aperture is increased. This
latter density is determined only by the average transmittance.
It is the purpose of this memorandum to find the relation be-
tween the average density (p) and the average transmission (T) for
four cases of interest:
(1) Square wave
(2) Sawtooth in transmittance
(3) Sine-wave in transmittance
(4) Noise due to photographic grain
1%/
The first three cases will be treated by determining the
spatial averages:
D(x)a(
(1)
where: D( X) = -Zoy T(x-)
For these three cases, the transmittance functions are periodic and the
range X will be chosen as one period. For the last case (4), a simple
statistical model will be used, which consists of a gamma d~stribution
for the probability density function of photographic density.
It will be understood that "transmittance" will always mean intensity
transmittance.
DeclaVMMewrlielbWCD4/02 : CIA-RDP78B04747A000200010030-6
 STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
October 8, 1964
HBH:bjs-431
Low Contrast Approximation
Before examining the four cases mentioned above, it will
be worth while to compare h with T when the density fluctuations
from the mean are small. The density can be written:
D(x) = D i f (k,)
where f(x) has zero mean. Eqs. (1) and (2) yield:
1O o/X to (z)CZX
X
I%W
For small f'(x) the integrand can be expanded:
As the fluctuations become zero, Eq. (4) becomes:
D= - ZogT
(2)
(3)
T = 1Q-~~ f- (ln f0) 1'zi F (ln /O)2 t Z(z) J (4)
(5)
Since {(x) = 0 it is necessary to retain the f 2(z) term to obtain the
next higher order approximation. The average f )) is commonly known
as the variance (~o ),
T = 10 ~~1 t r7 9LJJ~cro~~
Z
I~k- Taking the log of both sides:
lr~ 10 l l Z
1.15
(6)
(7)
where E (= p F Zoo T ; is a measure of the error within
which 7 and - LoyT can be interchanged.
Square Wave
The square wave is defined for one period (X) to be:
T(z)
(8)
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 STATINTL
I%W
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
Evaluation of E from Eqs. (1) and (8) yields:
(AD =
For low contrast (small .6 D ) Eq. (9) gives:
E_ (1Z-8- ~0)(4D~z"' 0.288(4D),
f /
E= z dD - loyz 4D-0.301
October 8, 1964
HBH:bjs-431
Tj
dog Tz I%
At high contrast ( A D becomes large) an expansion of Eq. (9) yields
an asymptote for e
The dependence of C on r D (Eqs. (9) and (11)) is shown in
Figure (1).
The sawtooth wave is defined for one, period (X) to be:
7(x, ) 7? X +T~
Combining Eqs. (1) and (12) yields:
e(>+ / I
where: = /0
(9)
(10)
(13)
For low contrast Eq. (13) gives:
E / Zy >o l ~~ t ) - 6) r? 959 ('11 D) (14)
At high contrast the value of r- from Eq. (13) approaches a constant:
E= Loy(`) - 0133
(15)
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
(o < X j (12)
 Approved For ReI se 2001,404/02: CIA-RDPMB04747A000200010030-6
-AL 11 ~ L-1
o-XO
r 'i nr vgd ~i 8~ 2Q,Q /ml" CIA P 747 , 0001 030-6
 STATI NIL
A roved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
-5- October 8, 1964
HBH:bjs-431
The dependence of c' on JD (Eqs. (13) and (15)) is shown in
Figure (1).
A sinusoidal transmission function can be defined for one
period (X) to be:
T(x)= T (I cos`rrx~
` X
Combining Eqs. (1) and (16) yields:
E = la z
9 f+ 9 C6 z)
At low contrast, Eq. (17) gives:
Cl ~1U)(LID)2 0.288 (AD)2 (18)
I%W
which is the same as the square wave. As the contrast is increased, E
approaches a constant:
6 = Locy 2 = 0. 301
(0s cx /) (16)
(17)
(Q D = peak density difference)
(19)
The dependence of E on t10 (Eqs. (17 and (19)) is shown in Figure (1).
Density Fluctuations due to Grain
The determination of values of U and T for a noisy
densitometer trace will not be carried out as spatial averages since
the noise is the result of a random process. The values will be de-
termined, however, by assuming a probability density function (for
either D(z) 0 7-(x) ) and evaluating the following integrals:
T P (T)dT
(20)
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
-6- October 8, 1964
HBH:bjs:-431
where PT and P1 are the probability density distributions for
transmission and density respectively. For the lack of a better dis-
tribution, the mathematically convenient "gamma distribution" # is
assumed for To (o)':
Po (D) -_
A(AD)
T(r)
(O < D < oo) (21)
where and r are two parameters of the distribution. It is now
necessary to find Pr ( T) for the above distribution.
The differential probability ( d P ) can be written:
d P = Po(D)d0= PT(T)c/T (22)
%MW
Combining Eqs. (20) and (22) yields:
T= /10 "PP(D)cdD
(23)
Employing the gamma distribution of Eq. (21) and integrating:
( a~ 1nJ0)
(24)
Therefore, the value of E is:
E = r A cy(A+l,-; 1n
* Not to be confused with the "gamma function" ( r
(25)
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 STATINTL
A ase 2001/04/02 : CIA-RDP78B04747A000200010030-6
ON" -7- October 8, 1964
HBH:bjs:[431
l:
where
Eq. (25) can be written in perhaps a more convenient form:
where: I _ (Lr7 10
(26)
The dependence of F on O (commonly known as granularity for
fluctuations arising from grain noise) from Eq. (26) is shown in Figure
(2) for different values of %`' , along with the low contract approxi-
mation of Eq. (7). As Z becomes large the quadratic curve (Eq. (7))
is approached for all values of cr
STATINTL
Ext. 562
`/
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Approved For Re~leasp 2Q
1/04/02 CIA-RD047A000200010030-6
Q
1AdOO~ i IAdOO.
Ot13X 1 10213X!
4 Prave r + J se3001 04/J 2 : c~4-R1 p7j~, g4777A000200010030-6
AdO:)
O_Li3X
 STATI NTL
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
-9- October 8, 1964
HBH:bjs-431
STATINTL (1) "Variance of Transmittance as Obtained from a
Gamma Distribution of Density Fluctuations" =memoran-
dum ET:bb:271 (15 June 1964)
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Faw 1074 (a... 1-59) Approved For
^ LABORATORY VISITOR
? TRIP REPORT
^ MISCELLANEOUS REPORT
[J TELEPHONE CALL
SUBJECT:
REPORTED BY;.__
TALKED TOi.-._
STATINTL CONTACT REPORT
- - -------- - ---
C)TI-ERS ATTENDING CONFERENCE!---.. -_ _
MM:bjs-450
020001 ITL
FILE-------- -- -
PROJECT NO -.97 -1.12_
DEPT- ---- -.-7 2 - -- --.- - - -- --- -
TITLE: ._---
DATE OF CALL-- ._ L4
oi per, rman_
PURPOSE OF CALL:__ To_~btain additional data on !!Mic_.r_osp
~~ or S11R~t~1ARIZ.E RESISL7 OF CAL!- OR VISR-9r Bkt::
FoR AM
-STATINTL _--.
hic edge produced at for project Microcap
ra
h
t
h
p
e p
o
T
og
was scanned using the Microanalyzer ;with the Microspot
Aperture to obtain the modulation transfer function of the instrument.
was made because the, data
This second visit to the
obtained during the previous trip (8-20-64 - 8-21-64) indiVATINTL
unexpectedly poor result for. the Microspot system.
n this visit did yield a considerably better
d
b
i
o
ta
ne
The data o
response curve for the MicrosPot system than that obtained previously
but it did not indicate any significant difference between the standard
slit aperture configuration and the Microspot configuration.
STATINTL
SIGNATURE.: _
Approved For Release 2001/04/02 : CIA-RD ?I
STATINTL'i
 F. 104 at, mss) Approved For
13 LABORATORY VISITOR
TRIP REPORT
0 MISCELLANEOUS REPORT
0 TELEPHONE
ATINTL
CONTACT REPORT
,97-112
OTHERS ATTENDING CONFERENCE:___. 111, PURPOSE OF CALL, IO eva uate Waco ass an "Class 1111
micro en.sLtometers for project Microcap
FOR ATrsKnor+ OF SUMMAR212.L RESULT OF CALL OR VISIT-BE BRIEF
1Th it models of tire-"Class i" ana ",_, Lass TINTL
mi crnrtensitometers were made available to as at the
^STATINTL
L
STATINTL
STAmPNTL
Ap
STATINTL
Special Products Plant in. on LLLUL)U1 1J, '?TAT~NTL
The microdensitometer evaluation es s developed for p ojec,
Microcap were conducted on the "Class I" instrument. They
could not be performed properly on the "Class III" instrument
because the basic instrument is not equipped :z~ith a strip chart
recorder.
The optical system of the "Class I" instrument is
essentially the same as that of the old Model 4 instrument
with. the exception that the hyperplane oculars have been re-
placed with designed projective eyepieces. This should,
and from the response indicated by the sine wave test charts,
does improve the performance of the microdensitometer. The
sensitivity of the instrument has been improved some chat
through some modification of the electronic circuitry.
There is still some doubt as to the ability of the in-
strument to achieve the quoted accuracy. Interferometric
techniques were used to determine the accuracy of the screws
and of the ways and although accuracy greater than that quoted
by was apparently achieved, no measurements were
made on the stage itself which rests some 10 to 12 inches above
the guiding ways. This could lead to a significant degradation
of the accuracy of the instrument but wou4d not affect the
precision.
The "Class III" instrument appeared to be a very versatile,
conviently operated instrument for routine analysis of large amounts
of data where precise linear measurements or resolution greater
than about 100 lines per millimeter are not required. The viewing
system of the "Class III" instrument was especially useful. An
eight times (8x) enlargement of the entire sample is displayed at
all times including during the scan. A smaller screen is used to
provide a view of the sample through the analytical optics and is
used for initial focusing and alignment. STATINTL
proved For Release 2001/04/02 : CIAR
MM:bjs-449
0020001 Q(3TL
flI F.
 Approved For.
0030-6
STATINTL
October 30, 1964
M3M:bjs-458
STATINTL
STATINTL
TRIP REPORT
Subject: Trip to STATINTL
NTL
Purpose: To evaluate Microdensitometer and fifil
Color Microdensitometer
Reported by:
Talked to:
STATINTL Others Attending:
STATINTL
STAY NTL
ST NTL
On Monday, October 26, 1964 we visited the _
STATINTL a_
~,.:_n.lor rnicrodensitnmPter. they had lust completed
th
e
see
for The instrument is basically the
o e e addition of two more photo-
multiplier tubes, two more amplifiers and another two pen recorder. The
after passing through the analyzing aperture is separated into three
light
,
non-overlapping spectral bands (specified by~by dichroic r6iFA TL
and filters. The outputs from each of the three channels (blue, green,
STATINTL red) are recorded on - recorders and can also be multiplexed onto
STATINTL magnetic tape. The instrument may also be used as a "black & white"
microdensitometer.
STATINTL
STNTL
On Tuesday, October 27, 1964 we visited the
to test their rnicrodensitometer. a very RIFINMTL
firm which started by producing microphotometers a few years ago,
currently manufactures two models of microdensitometers. The newer
model differs from their first instrument in that the light source is
separately monitored to eliminate the effects of intensity fluctuations and
to allow the photomultiplier tube to operate at a high average intensity
level which lessens the effect of "dark current. " A "dual beam" instrument
of this type was not available at this time. Therefore, tests were conducted
on the single beam version of the instrument. The standard logarithmic
STATINTL amplifier used with the instrument to provide an output linear with
density was also not availa they had "borrowed" a different log-
arithmic amplifier to provide us with the density output. The "borrowed"
amplifier's response time was much poorer than the standard amplifier's
and this may have affected the edge trace data we obtained using the in-
strument.
Described in Tri-p-ff eport dated 17 July 1964, MJM:bb:335 jg
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
TRIP REPORT October 30, 1964
MJM:bjs-458
The instrument resembles somewhat a
scaled down version oft a instrument. The stage is close to
the guiding ways and moves through about 2 inches along one axis.
The lead screw is not directly attached to the stage. The lead screw
drives a lever arm which in turn drives the stage. The accuracy of
the stage travel has been tested interferometrically and found to be on
the order of + 1 micron under specified environmental conditions.
Film sampl ~s inn which surrounds the glass area of the stage. m supplied
to an annula g
Both fixed scan speeds, or, as included on the
instrument tested, continuously variable scan speeds are available.
Selsyns are used with the continuously variable scan speed unit to
synchronize the recorder drive and stage drive to provide a constant
scale ratio (which can be altered by selecting various gear ratios) of ven
stage motion tart paper motion. The chart
he stage
correspond to the dire ct onrin can whc hdri
forwards or backwards
is moving if desired.
pochromatic objectives are used in the
instrument. The sample may be viewed directly by deflecting
the beam to a focusing eyepiece using a mirror which may be flipped
into the beam. A dichroic mirror can be permanently placed in position
to allow for viewing while scanning, but at the expense of sensitivity.
The instruments range in price from $10, 000
to $25, 000 depending upon t e model and the accessories ordered. The
instrument was considered particularly convenient to operate and appears
to be an excellent tool for photographic research where scans of 2 inches
or less are required.
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved F %WT%
TL
October 27, 1964
JG:bjs-448
STATINTL To:
From:
STATINTL
Subject: Safe Laser Powers for Microdensitometers
References: 1. Manual of Physical Properties of-Aerial aStJkMir
JSTAtNtL
STATINTL 2. on the Theory of Bessel
Functions, Cambridge, Y., 1962
STATINTL
STATINTL
3. Microdensitometer Sources and Detectors,
Memo No. JG:bjs-453
1. The purpose of this memo is to show that much more
radiation than would be necessary for use in a microdensitometer can
be applied to film without causing excessive heating. Excessive heating
can cause warping of the base or distortion of the emulsion by means of
stress formation within it.
Section II gives the assumptions necessary and justification
for them. In Section III the temperature rise within the irradiated area is
determined, and in Section IV the temperature rise in the surrounding area
is found. In Section V a typical case is discussed.
II. Assumptions.
The film is assumed to be held between two ring-shaped pieces
of metal, which provide an infinite heat sink. Later calculations will show
that since most of the heat is lost from the surface of the film, the heat sink is
not critical. Because one does not want Newton's rings, a sufficiently thick
layer of air will be allowed to cling to the film, even if it is held between sheets
of gla- s or plastic, that it may be considered to be in air for purposes of heat
loss.
It is further supposed that the film is heated uniformly over a
small circular area in its center. Preliminary calculations show that the
heat loss due to radiation is small compared to the surface losses.
oil immersion microdensitometers are not considered in this memo.
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Approved Fo 04/02 : CIA-RDP78BO4747A0002000120302 -9
To: -2- Octo er , 1964
JG:bjs-448
Estar film bases undergo a change of phase at about 80?C
No definite temperature is given for cellulose ester film bases?but
about 100?C is typical. 100?C is also the point for steam formation
within the emulsion. Because the emulsion will be heated more than the
base, we may take the maximum temperature as 100?C. Assuming
ambient temperature of 20?C, we have a temperature difference of 80? C
available.
If the film emulsion has absorption properties uniform
through its thickness, an exponential law of absorption will apply, with most
of the heat being absorbed near the illuminated surface. To reduce the
problem to two dimensions, the emulsion layer is replaced with a thinner
one having uniform heat absorption, and a volume rate of heat absorption
equal to or greater than the maximum rate of absorption of the real emulsion.
This will cause the heat conduction rate to be underestimated, which is safe.
For an emulsion with uniform properties, the thickness of the equivalent layer
is the point at which all but f /e of the radiation has been absorbed. This
can be found by dividing the emulsion thickness by 2. 3 times the diffuse
density, the factor of 2. 3 being the conversion from common to natural log-
a3kithms. For an emulsion developed to less than completion, the maximum
density is less and will create less temperature rise.
Illumination has been assumed to be from the emulsion side.
This system has the advantage that most of the heat is released near the air
surface, and does not need to be conducted through the emulsion layer.
This system also gives better definition when a small illuminating spot is
used.
III Heated Region
The temperature rise in the heated region may be found by
integrating the temperature gradients from the center to the edge. This
temperature rise is to be added to the temperature rise in the surrounding
region to obtain the total temperature rise.
r = distance from center of spot
k = thermal conductivity
L = equivalent thickness
t = temperature above ambient
P = power delivered to film
J = mechanical equivalent to heat
R = radius of heated area
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 STATINTL
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
To: -3- October 27, 1964
JG:bjs-448
At equilibrium, the thermal gradient at any distance ( r )
from the center of the heated spot must be sufficient to conduct away
the heat absorbed within the distance, t- , of the center of the spot.
The heat absorbed is
and the cross section which it must be conducted through is
IT 1- L.
with conductivity k, thus
The negative sign applies because temperature decreases as distance
increases.
The above equation may be integrated to give the temperature
rise from the edge to the center of the heated area, or
t Utii ` _ L
7r
IV. Cooled Region
Consider a ring with inner radius r and outer radius r--
The heat conducted in is il'rh,' L nd that conducted out is i s' .
where tpe symbols have the same meaning as in the previous section,
and means evaluated at 4A;- . The heat lost by convection
is -. 77 a e) where h. = convection coefficient. Setting heat
lost equal to heat gained we have
%f ' ". 7 - Tr C' .a 1-1 /
Passing to the limit 0 , we obtain the differential equation
Letting
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
To: -4- October 27, 1964
JG:bjs-448
This equation has the solution
where and are Bessel functions of order zero and
pure imaginary argument p 77), and C 1 and C are arbitrary
constants to be evaluated by satisfying the boundary conditions. Two
boundary conditions are: (1) the temperature gradient at the inside edge
of the cooled area must be sufficient to conduct the heat away from the
heated area, and (2) the temperature at the outside edge, where the film
is clamped between metal blocks, is equal to ambient, or zero. In order
for t to approach zero at large X , where
Thus, for small values of X where 4' T- .1 f f ~i
and since ~: e. if we are
interested in small values ( v ) it does not matter how far away the
heat sink is from the heated spot, as long as it is far away, and compared
to the other dimensions of the problem, it is far away.
For small y
and
?~ . biz.
(1
T
From the previous section, when
Thus,
and
7r#.L
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
STATINTL To:
2;7'Tr
From the previous section
or
which can be solved to
...-t ~ _ ...
October 27, 1964
JG:bjs-448
V. For example, take plus X , thin reconnaissance
STATINTL base, _ with a one micron spo . e appropriate values aTATINTL
80C?
4. 185 joules/cal
5. 38 x 10 -4 cal/sec. cm ?C
5x10'Scjn
1.3 x 10- cal/sec
6. 82 x 10-5 cm (developed to a diffuse density of 4. 85,
P = 1.2x10-5
12 microwatts
This power is compared to that available from various light sources
in Reference 3.
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
STATINTL
October .30, 1964
JG:bjs-453
mo for the Record
STATINTL To:
From:
Subject:
%woe
STATINTL
STATINTL
STATINTL
STATINTL
STATINTL
Microdensitometer Sources and Detectors
References: 1. Reference Data for Radio Engineers.
International Telephone & Telegraph Co.
New York, 1956
2. R.C.A. Tube Handbook, Vol. VII
currently possible can be satisfied by increasing the sensitivity of the
detector, increasing the illumination on the sample, or both.
Spangenberg, K. R., Vacuum Tubes,
McGraw-Hill New York, 1948
Safe Laser Powers for
Microdensitometers.
Memo No. JG:bjs-448
Intensity Stability of Laser
Sources,
Memo No. WCT:bb: 357
The need to. measure the density of photographic film
with smaller effective apertures or greater scanning speeds than is
2. Detectors
The usual microdensitometer detectors are multiplier
phototubes.' These devices have sufficient gain in the multiplier section
to insure that only a negligible amount of noise is introduced into the
channel at later stages in the electronics. Power supply fluctuations,
leakage, field-thermal and secondary emission, and shot noise are
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved for Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
JG.:bjs-453
claimed to cause.. spurious response from these detectors (Ref. 's 1, 3).
2. 1 Power supply fluctuations.
The gain of the multiplier section of a multiplier photo-
tube is a function of supply voltage. A small percentage change in
voltage causes a small change in the gain of each stage. However, when
all of the individual stage gains are multiplied together to obtain the over-
all gain, a large change results. According to Ref. 1, p 410, a small
change of P percent in supply voltage will cause a change of ~r < P per-
cent in, the output current, where 2i is the number of stages and
0. 5< x < 0. 7 . A change of A X / percent in current will be interpreted
as a change of )7 1.~ P percent in transmittance, or as a change of
A 91 P
in density where
and ;~7 = 7 , a common situation, then solving the above we have
f 0. 4 percent
Computation from the current versus voltage curves of reference 2 for
a 931A phototube at 1000 V yielded essentially the same result.
Power supply fluctuations cause spurious density readings regardless of
signal level.
The above applies to systems in which the voltage is held,
constant and the phototube current is measured and indicates the need
for closely regulated power supplies in such instruments.
The more common system, however, is the constant
current system, in which the phototube voltage is varied by a feedback
circuit in such a manner as to hold the phototube current constant, and
the phototube voltage is measured and converted to density. This system
is almost invulnerable to line voltage fluctuations, assuming well regulated
reference voltages for the feedback circuit, because the large current
change from a small voltage change of the phototube acts to increase the
gain of the feedback circuit.
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78B04747A00020.0010030-6
STATINTL To: October 30, 1964
SG.bjs-453
2.2 Leakage resistance
Leakage between electrodes contributes to the dark
current. But this resistance is in parallel with much lower resistances
and should cause a slight shift in operating point, but no noise at all.
2. 3 Field - Thermal and secondary emission.
At low signal levels, the primary noise source is the
fluctuation in the dark current due to field-thermal and secondary
emission, the secondary emission being caused by positive ions and
electrons arising from bombardment of gasses in the vacuum tube. Al-
though thermal emission by itself is negligible for most photosensitive
surfaces it is aggravated by the strong electric fields within the tube,
particularly if the cathode or dynodes have sharp corners or burrs.
2. 4 Shot noise
The noise voltage depends upon the amount of smoothing
done. For the usual case of a chart recorder, smoothing is certainly
provided by the inertia and friction of the pen assembly if not elsewhere.
If, however, a magnetic tape output with a high sampling rate is used,
less smoothing can be done. In such a situation, the effective aperture
of the system must be increased. Consequently, one could arrive at a
situation in which the shot noise,,which increases with effective aperture,
other things held constant became large compared to the noise due to
field thermal and ionic emission, Letting (1)
SZ.
2 1?
= mean square noise
current due to field
hermal and secondary
emission
= mean square noise
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 STATINTL
SW
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
To:
October 30, 1964
JG:bjs- 453
6 = sensitivity
Al = noise equivalent input
= bandwidth
--~ = charge on the electron and
luminous i-nut,
the inequality 12` 4 11 will be satisfied if
G
I 2 (-
S' 14/
where P = current signal to noise ratio.
For a median 931A phototube, we have
S = 30 x 10 6 amp/lumen and
= 9. 5 x 10-13 lumen secs.
Also, _C = 1. 6 c 10-19 amp sec, and for an equivalent density error of . 01
= 43. Thus, for ( < 4 cycles/sec, the case for strip recorders,
phototube dark current noise predominates, but7for tape recording.shot
noise predominates.
2.5 Refrigeration
Refrigeration decreases the noise of the phototube under
most conditions. According to reference Z, refrigeration of a 931A photo-
tube to -75? C (approximately the sublimation point of dry ice) increases
its detectivity (reduces the noise equivalent input) by a factor of 20. But,
film should be assesed in an environment with proper relative humidity.
For 20?C and 50% relative humidity, a window approximately 3 1/8" thick
would be required to prevent condensation of moisture. Thus, special
optics would be required to pre-correct for the effects of the refrigeration
apparatus.
2. 6 Remarks on Multiplier Phototubes
The characteristics of multiplier phototubes of the same
type and make vary widely. Selection is common practice. Therefore,
not much validity can be attached to the procedure of measurement of the
sensitivity of one instrument of a make and type and taking this to be the
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6'4-
 ST.ATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
To: October 30, 1964
JG:bjs-453
sensitivity of all instruments of that make and type. Finally, the tubes do
fail, and after the first tube fails, the sensitivity of an instrument will
depend on the detectivity of the next phototube selected.
Mercury arc and laser light sources can increase the
illumination of the sample, but create special problems of their own.
Comparison of mercury arc and tungsten light sources
is straight forward and universal. It is done on the basis of brightness
alone. Tungsten at 3360?K, taken as a standard, has a brightness of
about 3, 095 candles/cm2. The brightest mercury arcs have a brightness
of 140, 000 candles/cm2. Also, for a device with S-4 response (such as
the 931A phototube) the efficiency of mercury light is Z. 23 times that of
tungsten light. Thus, the mercury light gives about 100 times as much
effective illumination as the tungsten.
The comparison for laser light is not as straight forward,
because the laser illumination depends on the size of the illuminated area,
whereas (for reasonable sample sizes) the thermal source illumination does
not. For a circular spot with the smallest N. A. permitted by the
diffraction limit the tungsten gives 7. 4 microlumens, and for an 80:1
rectangle, 750 microlumens. A 12 microwatt laser (4) will give 1. 9
millilumens. The phototube sensitivity to the laser light is less than to
tungsten, and the 1. 9 millilumens, are equivalent to 420 microlumens of
tungsten light.
3. 1 Stability
The major problem with the mercury arc source is the
lack of stability, which will require compensation. An investigation of
the stability of laser sources may be found in Reference (5).
4. Conclusion & Recommendation
It is concluded that for small effective apertures either
mercury or laser sources will provide an increase in instrument sensitivity.
If the two orders of magnitude available from mercury are sufficient,
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6 -5
 Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
To: October 30, 1964
JG:bjs-453
either a mercury or a laser source may be used. The 12 microwatts
of reference 4 are safe for steady state. At the risk of damaging the
film should the scanning stop, one could use much more laser power, so
that if the 2 orders of magnitude of the mercury are insufficient, laser
power is recommended.
For larger effective apertures the mercury is the most
effective source, the crossover point being at about 1000 square microns
for a numerical aperture of 0. 4 and at larger areas for smaller numerical
apertures.
It is further recommended that source possibilities be
exhaused before attempts are made to increase the detectivity of the
detector because of the selection problems associated with phototubes
and the optical and supply problems of refrigeration.
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved WATTNTL
October 28, 1964
MJM:bjs-446
STATINTL To:
From:
Subject: Some considerations in the design of an improved
microdensitometer system
STATINTL
STATINTL
A s a result of the survey of microdensitometers conducted
for project Microcap much information was obtained on a variety of
microdensitometer systems. While several of the instruments surveyed
do reflect the current state-of-the-art of microdensitometer design it is
felt that an instrument of improved performance could be produced at
this time by incorporating the best features of each of these instruments
into a single system.
A brief discussion of the features of each instrument which
are considered the best follows.
The basic components of a microdensitometer are the
mechanical system (stage drive, guiding ways, lead screws), the optical
system and the electronic system.
The best mechanical system appears to be that developed by
the Their long experience in the production of
precision comparators has enabled them to develop a microdensitometer
stage drive system with micron, and possibly sub-micron accuracy over
approximately 10 inches of stage travel in either the x or y direction.
Many present and future uses of microdensitometers (such as the present
moon map project of ACIC) will require micron or sub-micron accuracy
for linear measurements. STATINTL
The optical system employed by- including the viewing
system used on their "Class I" instrument, appears to be superior to other
optical systems based on the modulation transfer functions obtained from
the edge traces and sine wave test pattern traces. Further improvement
STATINTL over the-optics should be possible by using an illuminating objective
with a numerical aperture approximately 0. 8 th4t of the analytical objective
STATINTL as was determined, theoretically, by The viewing system
should be modified somewhat to provide for, when desired, direct viewing
STATINTL
STATINTL memo by RK:bb:406) 31 August 1964
STATINTL and emo by (RK:bb:283) 22 June 1964, Revised 9 July 1964
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
STATINTL To: _2_ October 28, 1964
MJM:bjs-446
of the sample image instead of displaying it on a ground glass screen. The
granular structure of such a screen obscures small detail making focusing
of such detail difficult. The use of a lens to distributethESTA tlNTVr
the photomultiplier surface (used by M1 on their "Class I" it p tft
STATINTL and on the nstruments is also considered necessary
for the optimum per ormance of the instrument. Both bilaterally adjust-
able and fixed scanning and illuminating apertures should be incorporated
STATINTL in the system. STATINTL
The amplifiers and associated. electronic circuitry, including
STATINTL a recorder and the end window photomultiplif,NTL
STATINTL suc as Rat used by the in their icroanalyzer
provide the most versatile (logarithmic or linear amp t ication), sensitive
(approximately 0-4 in density with less than 1 2 area and stable electronic
system for a microdensitometer. The electrical system
for the illuminating lamp incorporates such desired features as adequate
isolation and heavy soldered connections throughout the circuit to prevent
STAINTL intensity fluctuations due to line voltage changes. This type of system should
be used in any future microdensitometers. To reduce the effects of lamp
intensity fluctuations further a form of the double beam system (utilized by
in which lamp intensity fluctuations are detected by a separate
STATINTL photomu tiplier system, could be used to compensate the output of the main
amplifier. For high speed data aquisition to digital recording system must
be employed and it is suggested that a system such as that produced by the
be adopted.
STATINTL
adding alphanumeric data, provides for density clipping, and can record
data at a rate of 3000 cycles per second.
This system uses magnetic tape, allows for programming scan patterns,
Various auxiliary features could be added to increase the
versatility and/or performance of the microdensitometers A gas bearing
platen such as that developed by the provides an excel-
lent means of keeping the sample firm y against t e supporting A W. L
and its use is strongly suggested. In the event that focus is changing be-
cause of emulsion characteristics (it should not change due to non lanar
stage motions in a well designed system), a device similar to TL
_-1-7 t..- __,_-A A _.........~ ......U:a
would be necessary, however, since the device seriously decreases
instrument sensitivity and response time. The use of a special recording
unit, a raster scanning system, and a density level coder woulc8TTL
the instrument to plot isodensity contours. Such a system has been
developed by the for use with the -
_ instrument which is especially suited for such an application. An
isodensity system could, however, be adapted to any microdensitometer
without undue difficulty. STATINTL
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78BO4747A000200010030-6
 Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6
STATINTL To: -3- October 28, 1964
MJM:bjs-446
The various features described above could be incorporated
into a single instrument since there is no problem regarding-the compatability
of the separate components discussed. Based on the cost information ob-
tained from the various microdensitometer manufacturers, it is estimated
that a system incorporating all of the above features would entail a de-
velopment cost of approximately STATINTL
%AW
STATINTL
Approved For Release 2001/04/02 : CIA-RDP78B04747A000200010030-6