SIXTH BIMONTHLY ON THE AUTOMATIC TRANSMITTER PROGRAM

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-03424A000800010031-0
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 19
• Document Creation Date: December 22, 2016
• Document Release Date: February 8, 2012
• Sequence Number: 31
• Publication Date: July 8, 1959
• Content Type: REPORT

L
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
u ~.~~~ ~,,..., ... -- -
EXT E~YNU 6 YRS B~ ~.~ ~
REV OATE 10 ~~o- SY~
ooe -,~- T Trv~ '~~G
ORiG COMP ~~. 0?i _,i~- -
rACfS ,~av c~A:a<
ORIG CLASS ~-- ~p AUTN~ MR U?1
V
JUST r NEXT RE
Sixth Bimonthly Report
Automatic Transmitter Program
CO~V~10~-~T~AL
Prepared by:
8-July- to S-Sept.-1959
25X1
25X1
CO~v~~-,v~.~ ~ ; {,
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
Payee No?
I ?
PurpO 8e ?
1
II.
Abstract
1
III.
Factual Data
2
1. The Transmitter
2
2. Automatic Impedance Matching.
7
IV.
Conclusions.
9
V.
Future Plans
10
VI.
Identification of Key Technical
Personnel.
]1
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 1
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
I. Purpose
See Bimonthly Report No. 1.
During the past reporting period several significant decisions
have been made and conclusions reached concerning the final design of the
transmitter and matching network. In the RF circuitry of the transmitter
itself a broadbandi ng technique has been adopted due to the difficulty of
operating Varicap type tuning devices at~ high power levels. Although
electronic band switching is employed, it is of a considerably simpler
nature than that described in previous reports. Considerable work has been
performed in an effort to obtain a high output power from the special tran-
sistors supplied by the customer. These units have proved somewhat dis-
appointing in that, on the travelling wave oscilloscope, it is possible to
see that in Class B operation they do not have a fast enough turn off time
to operate efficiently at 30 mc.
The electrical design of the antenna matching network has been
essentially completed. Mechanical layout, which is a mayor consideration
in a piece of equipment of this type, has been completed and construction
of the control mechanism started. Modification of a currently available
variable capacitor is being carried out, in order to provide a unit with
the correct maximum and minimum capacitance values.
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
III. Factual Data
A. Oscillator
(i) Specifications
r.
(a) Operation from 3-30 mc.
(b) Two bands: manually switched covering
3-15 mc; 15-30 mc.
(c) Crystals are specified as covering the
range from 3-15 me oscillating in furxlamental mode.
(d) The range from 15-30 me will be derived
by using the specified crystals in the shear or 3rd overtone mode.
(e) The oscillator circuit must ensure oscilla-
tion at either the fundamental or 3rd overtone. Thus the 3-15 band will be
~~
derived entirely from fundamental mode oscillation and the 15-30 band entirely
from 3rd overtone oscillation.
(ii=ircuit Operation
A diagram of the oscillator and its auxiliary cir-
cuitry is shown in Figure 1. T1 is a 2N381t used as the oscillator transistor.
The oscillator has four bands; 3-7 7-15 : 15-18 18-30. The band switch Sl
(double pole - double throw) switches between 3-15 and 15-30, whereas the
switching between 3-7 and 7-15, and 15-18 and 18-30 is accomplished auto-
matically.
In switch position a - a the circuit operates from 3-15, and must
oscillate in the fundamental mode only; i.e., if a 3 me crystal is plugged in
the circuit must not operate at 9 mc. This is accomplished as follows:
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
 ? Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
RECTIFIED DC
AS DERIVED
FROM OUTPUT,
g FILTERED TO PASS
SIGNAL ONLY UP
TO A80UT 18 MC
OSCILLATOR AND ASSOCIATED CIRCUITRY
F 1 GURE 1
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
T3 and T~ are 2N123 transistors which act only as switches. T2 is a
2N2l~7 transistor and amplifies the AC control signal from the emitter of T1.
It should be noted that the output of this circuit is rectified to provide
DC control signals for switches T3 and T~.
The 100 K.pl. resistor and 1 ?f capacitor (Cl) combination at the base
of T3 and collector of T4 is significant. If this capacitor is allowed to
charge to -25 V it will turn T3 ~on~. If, however, a DC signal is applied
to the base of T~ it will turn T~ ~on~ which in turn will short circuit the
1 ?f capacitor and keep T3 ~off~.
If T3 is off the 100.rt - 1 ?h combination is connected to T1,
whereas if T3 is on, the .02 ?f capacitor is grounded thus providing an AC
bypass for the 100,.tt - 1 ?h series path. With the 100.st. - 1 ?h path in
the oscillator circuit the range of oscillation is from 3-7 mc, If this
network is short circuited the range of oscillation is from 7-15 mc.
The sequence of operation in the 3-7 me band is as follawss
(a) A 3 me crystal is plugged in.
(bj The 1 ?f capacitor starts charging with a long time constant.
Before T3 can be turned ~on~ the oscillator oscillates at 3 mc.
(cj It can not oscillate at 9 me because the 100.st - 1 ?h com-
bination is still in the circuit, because T3 is still off.
(dj As soon as oscillations occur T~ is shorted thus discharging
C1 (1 ?f capacitor) and ensuring that T3 remains off.
The sequence of operation for 7-15 me is as follows:
(a) A 9 me crystal is plugged in.
(b) The circuit can not oscillate until T3 is shorted so that the
100 .1t- - 1 ?h combination is bypassed. This will occur when Cl is charged,
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
___ __
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
(c) Cl will remain charged since under this condition there is no
signal at the emitter of T1 and thus there is no DC signal available for appli-
cation to T~.
The transistor T5 is short circuited from 3-7 me and introduces a low
pass filter later in the system. This serves to reduce distortion since the
cutoff frequency is set at about 8 mc. In the 15-30 me band the oscillator
circiri.t must ensure oscillation only on the 3rd overtone of the crystal.
Thus if a 6 me crystal is plugged in, the circuit should oscillate at 18 me
rather than 6 mc. In this case switch S1 is in position b. It should be
noted that the 100.s~. - 1 ?h conbination is short circuited. Furthermore it
should be pointed out that there will be a DC signal at point B (base of T6)
from 15-18 me thus short circuiting the 21~ ??f capacitor in the collector of
T1. The tank circuit in the collector consequently consists of the 2 ?h in-
ductor of the 50 ??f capacitor. This provides a lower frequency cutoff of
about 13 me which is adequate to ensure proper operation. However, this low
cutoff frequency results in marginal operation at 30 me so that it is necessary
to add the 21~ ??f capacitor in series with the 50 ??f. This raises the cutoff
frequency to about 18 me and results in an appreciable increase in output at
30 mc.
The circuit operation in the 15-30 me range is quite straightforward
and may be summarized as follows.
With a 5 me crystal (fundamental frequency) the circuit will oscillate
at 15 me and a signal will be returned to point B, short circuiting the 21~ ??f
capacitor.
With a 7 me crystal the circuit will oscillate at 21 mc, but no DC
signal will be returned to point B since the AC has been filtered. Thus the
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
tank is comprised of 50 ??f in series with 21~ ??f.
B. Amplifier Stages
...~_
The oscillator signal is at a relatively low level and
must be amplified. The exact amount of amplification required depends to a
large extent on the capabilities of the output transistors. However it is
clear that several low level, Class A, linear amplifiers will be needed.
These can be broad-band covering the region from 3-30 mc. The transistor
found most advantageous in this application is the 2N509. Atypical low
level amplifier stage is shown in Figure 2.
The operation of this stage is extremely simple. The feedback net-
work from the collector to the base shapes the frequency response such that
the gain rises slightly as the frequency increases. The output impedance of
this stage is quite low, making the cascading of several stages relatively
simple.
It should be noted, however, that linear operation of this circuit
is only possible for low signal levels (about 1 V rms with a .100 st load).
Since a final output swing of about 30 V rms is required it becomes
clear that a driver stage for the output is essential. Because of the
extreme harmonic distortion specification, a push pull driver is necessary.
A pair of 2N509 transistors appears adequate for this application although
all tests of the final stages have not yet been completed. It may prove
necessary to use a higher power transistor in the driver application.
The output stage will also be push-pull in order to derive maximum
output. Several "pulse-transformers~~ have been tried in the push-pull stages
and broad-bared operation from 3-30 me appears to be feasible. However up to
i Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
5A
LOW LEVEL PREAMPLIFIER
~ FIGURE 2
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
the time of writing it has not been possible to use a transformer in the out-
put of the final amplifier stage because of the power level. Two approaches
are being taken to solve this problem:
(i) Evaluati.on of new transformers supplied by Aladdin Electronics
which may be able to cover the range in two bands.
(ii) The output circuit shown in Figure 3 circumvents the need for
an output transformer.
This circuit has been tested extensively acid shows a great deal of
promise. However it is necessary to tune out the output capacitance in order
to meet power level arad distortion requirements.
The input transformer has a 2:1:1 turns ratio to provide abetter
impedance match for the driver stage. The driver can consist of an identical
circuit, or a conventional class B stage as shown in Figure 1~. An advantage
of this circuit compared with that shown in Figure 3 is that both halves of
the input transformer are at AC grourxl potential, thus reducing shunt capaci-
tance problems. However both circuits perform well.
The amplifier has not been completed at the time of writing although
the various components have been tested rather extensively. It now appears
that the final configuration will consist of a preamplifier comprised of
several common emitter feedback stages. These will be followed by a push-
pull driver and finally a push-pull output stage. The amplifier will be
broad-band except for the output stage which should have a tuned circuit at
the collector. As described elsewher?, it appears that this tuning can be
conveniently accomplished by the automatic antenna matching network. The
power output will be about 1 watt from 3-30 mc.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
 i
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
6A
PUSH PULL OUTPUT STAGE
FIGURE 3
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
6B
CONVENTIONAL CLASS B. STAGE
FIGURE 4
~ Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
It may prove desirable to include in the amplifier a simple AGC in
order to compensate for variations in oscillator output. This AGC would also
accomplish gain equalization with respect to frequency and thus eliminate the
need for carefully designed shaping networks.
The system described thus far is relatively insensitive to tempera-
ture variations: the greatest problem will be encountered in the allowable
power level of the output stage.
2. Automatic Impedance Matching
~ i
The accuracy with which the servo system produces a 500 ohm input
impedance depends on the sensitivity of the servos. Since the magnitude of
the error signals is determined by the RF output power level, the drop of
this level by an order of magnitude with respect to that originally antici-
pated has necessitated further work on the sensing circuitry. It appears
that by modifying the transformer in the phase detector an output of suffi-
ciently high level can be achieved. Further improvement is also obtained by
the optimum transformation of impedance level between the sensing circuitry
and the ring modulator. Work is currently in progress to evaluate the opera-
tion of miniature transformers having various transformation ratios, Since
the miniature transformers have characteristics which differ considerably from
those predicted by "ideal" transformer theory, the desired transformation
ratio is being determined experimentally. A plot of 400 cps output from the
ring modulator as a function of phase angle at 3 me is shown in Figure 5.
(Further modifications may improve this characteristic.) It is expected that
an input to the servo amplifier of about 3.5 rav will activate the servo motor.
This signal level is obtained when the phase angle error is about ].!~ degrees,
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
7A
400 CPS AT e~?
(IN MV )
6
4
2
PHASE ANGLE
l IN DEGREES )
25 -2
0 -1
5 -10
-
5
0 +
5 +1
0 +1
5 +
20 +2
5
_2
4
-
_a
PLOT SHOWING 400 CPS OUTPUT VOLTAGE FROM THE RING
MODULATOR AS A FUNCTION OF PHASE ANGLE AT 3 MC
FIGURE 5
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 1. I
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
which represents a negligible mismatch loss (a 20 degree phase angle repre-
sents only a 2 1/2?~6 mismatch loss).
An accurate evaluation of the system sensitivity can not be made un-
til the torque necessary to turn the variable capacitor and inductor has been
determined. These elements are being built in the model shop.
A pictorial view of the automatic antenna matching network is shown
in Figure 6; the motors, gear boxes, variable coil, and variable capacitor
are shown in their approximate positions within the case.
The work in the RF transistor circuitry has indicated that the RF
power output can be increased by tuning out the capacity associated with the
output stage. It was initially assumed that terminating the output stage in
500 ohms would very nearly yield maximum output power. Since the output
capabilities of the transistors are so limited, in order to maximize the out-
put it may be desirable to modify the sensing circuitry so that it will lead
to the production of 500 ohm load shunted by an inductance which varies with
frequency in a manner which cancels the output capacity of the stage. The
manner in which this could be accomplished is as followsa The phase detector
nulls when the phase of a reference voltage is the same as that of the line
current. Thus, a purely resistive load prod uces a null if the line voltage
is used as the reference voltage. To produce an inductive load, the line
current must lag the :Line voltage, so the reference voltage must also lag the
line w page. This lagging voltage could be obtained by applying the line
voltage to a phase shifting network having the desired characteristic. The
magnitude detector nulls when the ratio between the magnitude of a small
series impedance and the magnitude of the load impedance becomes some fixed
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
PICTORIAL VIEW
AUTOMATIC ANTENNA MATCHING NETWORK
FIGURE 6
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 .~
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
quantity. If the series impedance is a resistor, the magnitude of the load
impedance at null is not a function of frequency. If, however, the magnitude
of the series impedance varies with frequency, the magnitude of the load
impedance also will vary with frequency in a similar manner. It seems rea-
sonable that a phase shifting network and a series impedance which varies
with frequency in the desired manner could be synthesized. farther efforts
along this line will be made when the characteristics of the RF circuitry
have been definitely established.
The transmitter oscillator circuitry is capable of providing an
output at the fundamental frequency of crystals when the manual bandswitch
3.s in the 3-15 me posit3.on. In the 15-30 me range operation is on the third
overtone of crystals, the fundamental frequencies of which are between 5 and
10 mc. Within the two ranges set by the manual bandswitch, electronic
switching takes place automatically in order to provide the correct frequency
at the oscillator output. Preamplifier stages have been constructed but their
gain-frequency characteristic has not been finally determined, awaiting settle-
ment of the output stage problem.
It appears that the maximum output which will be available from the
transmitter will be approximately 1 watt over the 3-30 me range. Although
as oscillators, transistors would be capable of more impressive performance,
considerable difficulty is experienced in driving the units as amplifiers be-
cause, at high currents, the gain of the transistor falls off quite sharply.
Some difficulty was anticipated with the antenna matching network
sensing circuits due to the low power level available from the transmitter,
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
 ~ _.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
particularly in the case of the phase sensing circuit. Redesign of the trans-
former has however increased the sensitivity of this circuit so that it now
appears that satisfactory operation can be obtained..
The assembly for the motors, torque multipliers andvariable react-
ances which form the antenna matching network has been designed as is presently
being constructed. The variable capacitor which is used at the input to the
n network is being built from parts removed from a commercially available
miniaturized variable capacitor.
V. Future Plans
Because of the increased output capability of the special transistors
when used as oscillators, the possibility of designing a high power oscillator
stage will be investigated. Precautions will be necessary to prevent exces-
sive RF crystal currents.
As soon as the output stage design has been optimized, the RF
circuitry will be built in final form. By this time it is anticipated that
the matching network and its associated controls will have been built so that
the complete system may be tested. With a mechanical servo system, tests
made on anything other than that which is intended as the final version are
relatively meaningless as friction plays a key role in determining thresholds
aryl unless great pains are taken in the construction of any preliminary ver-
sions, the amount of friction present would be excessive.
The bui]tiing of the matching network calls for completion of the
variable capacitor currently under construction. Also required will be the
high Q variable inductor which is being built at the present time. Some
experimental work remains to be done in order to provide a very high Q even
Declassified in Part - Sanitized Copy Approved for Release 2012/02/08: CIA-RDP78-03424A000800010031-0
 __l _
Declassified in Part -Sanitized Copy Approved for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0
u CONF1D~~sTfAL
when this coil is in close proximity to the metal case. Ferrite slabs which
are c~srently being pressed, will be placed around the coil in order to provide
a low loss flux path.
VI. Identification of Key Technical Personnel
See previous Bimonthly Report.
Declassified in Part -Sanitized Copy Approved, for Release 2012/02/08 :CIA-RDP78-03424A000800010031-0