Back to the InfoAge HomepageBack to the Press Index
By Jack Mofenson,
Evans Signal Laboratory, Belmar N.J.
Dr. Harold T. Webb (right) adjusts the auxiliary tuning crystal in the lunar receiver while E. K. Stodola looks on. Behind Stodola is the nine-inch type-A indicator which records the echoes.
|THE RECENT EXPERIMENTS performed
by the Signal Corps Engineering Laboratories in receiving radar
echoes from the moon have aroused much comment from engineers,
astronomers and others engaged in technical pursuits. Although
the scientific aspects of sending radio-frequency signals throughthe ionosphere
are certainly of importance, the work done on theproject is better classified
as an engineering achievement. As yet, no long-term systematic
observations have been made. This article is confined, therefore, to a
discussion of the technical characteristics and general description
of the equipment employed.
Briefly, the experiment consisted of transmitting quarter-second pulses of radio-frequency energy at 111.5 mc every four seconds in the direction of the moon, and detecting echo signals approximately 2.5 seconds after transmission. Display of the detected signals was audible as well as visible. Technically, theexperiment utilized well-established radar techniques, but with radicallydifferent constants throughout the system. Considerations of pulsewidth, receiver bandwidth, transmitter power and the precise frequency of the returned signal due to Doppler effect, were such thatcareful attention had to be givento the design of the overall equipment.
After preliminary calculations were made concerning transmitterpower, the reflectivity coefficient of the target, and receiver noise figure, it was apparent that receiving radar echoes from the moon was technically possible. Under the direction of Lt. Col. John H. DeWitt, a project called "project Diana" was set up in September 1945 to develop a radar system capable of transmitting r-f pulses to the moon, and detecting echoes more than 2seconds later. Prior to entering the Signal Corps, Colonel DeWitt, who at that time was chief engineer of Radio Station WSM in Nashville,Tenn., designed and constructed transmitting and receiving equipment for the purpose of receiving echoes from the moon. This equipment employed substantially similar transmitter power and frequencyto that used by the Signal Corps, but the attempt was a failure due to insufficient sensitivity in the receiver. Colonel DeWitt's appreciation of the problem and personalsupervision were the driving forces that made the present experimentsuccessful. Assisting Lt. Colonel DeWitt were: E. K. Stodola, Dr. Harold D. Webb, Herbert P. Kauffman and the writer, all of Evans Signal Laboratory. Credit is also due the members of the Antennaand Mechanical Design Croup, Research Section, Theoretical StudiesGroup and others.
The practical implications of radar contact with the moon are numerous. During the war the Germans used the V2 Rocket whichclimbed some 70 miles above the earth, and the future holds the unhappy prospect of missiles going far higher than this. The matter oftransmission of radio signals to great distances above the earth for detection and control of such weapons becomes a problem of military importance. Further, the use of a reflector far beyond the earth forradio waves makes possible directmeasurement of the ability of radio waves to penetrate the ionosphere.A more complete investigation in this direction is indicated.The possibility of using the moon as the reflector for a part-time long-distance point-to-point communication system is also being considered, as well as using the moon as a target to measure field strength patterns.
Lt. Colonel John H. DeWitt Jr., in charge of the project, a modified version of the SCR-271 early-warning radar used at Pearl Harbor. DeWitt is the former chief engineer of WSM.
The author, Jack Mofenson, adjusts the position of the waveform-monitoring stub. Over this transmission line traveled the 3-kw transmitted pulse and a millionth of a billionth watt echo.
Determination of Requirements
Several of the constants which determine the maximum distance at which a radar set can detect targetsare peak transmitter power, radio-frequency of the transmitted signal, duration of the signal, receivernoise figure, and target echoing area. These constants, among others, are concisely summarized inwhat has been called the free-space radar equation.
This equation has already been derived in Electronics In this equation, r is the radar range at which a signal may be detected, Pt is the transmitter power during the pulse, Go, the transmitting antenna power gain, A. the absorption area of the receiving antenna, o the effective echoing area of the target, and P,the power of a barely discernible signal, on the same basis as Pt. The power gain due to ground reflections (not considered in the freespace equation) at maximum effectiveness increases the range of thesystem by a factor of 2. This is equivalent to a power gain of 12 db.In the case of a target as large as the moon (2160 miles diameter),calculations showed that in orderto receive an echo from the wholehemisphere of the moon at once a pulse width greater than 0.02 seconds was required. This set a lower limit on the transmitter pulse widthwhich corresponds to an optimumbandwidth of 50 cps for the receiver. These requirements eliminate, for the present, the use of themicrowave frequencies, because of considerations of pulse length.
Propagation studies indicated that electromagnetic waves at a frequencyof 110 mc were capable of penetrating the ionosphere, and because ofavailability of equipment, a radar set operating at 111.5 me was chosen for the experiment. The peak power available in this transmitter wasequivalent to 3000 watts for Pt usinga 0.25-second pulse. The transmitter had the added advantage of being crystal controlled, deriving its final radio frequency after a series of frequency multiplication from a 516.2-kc crystal oscillator. The receiverassociated with the transmitter wasof the multi-mixer type (quadruplesuperheterodyne) capable of beating down radio-frequency signals to a final intermediate frequency of 180 cycles per second. Such an arrangement permitted use of an extremely narrow pass band, 57 cps, thus making the receiver highly selective and limiting the noise to a very low value. The extremely narrow-band receiver was an advantage, also, because it permitted tuning the receiver to the exact radio frequency of the returned echo. The importanceof this can best be realized by considering the fact that due to the relative velocities of the earth and themoon, the returned signal may differ from the transmitted signal by as much as 300 cycles, due to the Doppler frequency shift. In using a highly selective receiver whose final mixer is tuned to receive the precalculated frequency of an echo return from the moon, the receiver rejects any signal returned at anyother frequency. To reduce the noise contribution ofthe receiver, a high-gain, low-noise-
figure pre-amplifier was connected between the antenna and the receiverproper. The minimum perceptible received power was P, readily calculated from the formula for noise figure.
In this formula E2/4R is the maximum available signal power at the receiver input terminals in watts, where E is the signal voltage at the antenna terminals, and R is the effective impedance in ohms. KTB is the maximum available noise power at the receiver input, where K is Boltzman's constant, 1.37 X 10-23 joules perdegree Kelvin, T is the temperaturein degrees Kelvin, chosen, at 300 degrees, and B is the noise bandwidthof the receiver in cycles per second.For this receiver B is 57. For a one-to-one ratio Eq. 3 gives signal-powerto noise-power of
1.48 X 10-18 watts, taking the effective noise figure of the receiver as 7 db.
The best antenna available at this frequency was a 32-dipole array utilized by the SCR-271 early-warning radar. Two of these arrays were secured side by side and mounted on a 100-foot tower. Calculations show that the array had a power gain of152 times that of a single halfwave dipole antenna. Since the effectivegain of a single dipole is 1.64 times that of an isotropic radiator, the value of G. is given as 1.64 X 152 or 250.
The absorption area A, of the
receiving antenna is calculated from
Substituting the value of G, previously given, Ao = 522.1 X 10-7 square miles. The remaining constant to be determined before solving Eq. 1 is o theeffective echoing area of the target. Calculations of the reflectivity coefficient made by Walter McAfee of theTheoretical Studies Group, assumingzero conductivity and a dielectricconstant of six for the moon, resultedin the figure 0.1766. The effective echoing area is this figure multiplied by the projected area of the moon,Pi d2/4 where d is the lunar diameter.This gave an effective echoing areaof 0.1766 (2160)2 (3.1416)/4 or 647,000 square miles.
Substitution of these values in the free-space radar equation gave a maximum range of 573,500 miles andindicated that the effective range of the equipment chosen was more thantwice that needed to receive echoesfrom the moon. By adding the powergain due to ground reflection, a further excess of power of 12 db or arange of 1,140,000 miles was indicated, which meant that accordingto calculations, the received signal should be about 20 db above thermal noise. This calculation of the signalstrength of the returned echo checked closely with observations and indicated that no appreciable attenuationoccurs in free space.
Once the determination of constant was completed, the choice ofavailable radar sets was made. Since no attempt was made to design majorcomponents specifically for this experiment, the selection of receiverand transmitter was made from equipment on hand. A crystal-controlled radar transmitter and receiver designed by Major E. H. Armstrongfor another purpose were selectedsince they met the requirements of power and bandwidth. A block diagram of the complete transmitting,receiving and indicating system is shown in Fig. 1.
The transmitter is crystal controlled, deriving its final radio
frequency of 111.5 me after a series of frequency' multiplications
from a fundamental crystal oscillator frequencyof 516.2 kc.
Keying is accomplished by causing a low-level multiplierstage to conduct by driving its cathode negative for the duration of thetransmitted pulse. In the initial setup, keying was performed mechanically by a relay, but this has since been replaced by an electronic keyer withthe pulse width controllable between0.02 to 0.2 seconds. A block diagram of the transmitter is shown in Fig. 2.
From the diagram it is apparent that the transmitter is of a conventional type. The output is fed overa 250-ohm open-wire transmission line to the antenna array. The antenna contains 64 dipoles horizontally polarized. The effective power gainof the array is 250, or 24 db. The antenna, shown in Fig. 3; ismounted on a steel tower 100 feet high and is controllable in azimuthonly. No provision has been made to incline the antenna in elevation.Because of this restriction, the times of observation using the presentequipment were necessarily limitedto moonrise and moonset. That this condition of observation is the worst possible (due to the long path through the atmosphere and the consequent possibility of trapped radiation) has been recognized. But itwas impractical to procure an array of the equatorial type. Aside from propagation deficiencies, a far more serious limitation was the fact that observations were limited to twoshort periods daily.
The beam width of the array is approximately 15 deg at the half-powerpoints, with the first three lobes spaced approximately 3 deg in elevation. Since the diameter of the moon subtends roughly one half degree of arc, most of the power transmitted does not illuminate the target, which constitutes a serious waste of power. The rate of rise of the moon along its ecliptic is 1 degree of are every
4 minutes, which allowed roughly 40 minutes of observation as the moon intercepted the first three lobes of the antenna. Bending effects due to long transmission path through the ionosphere undoubtedly exist, but no precise measurement of this effect has yet been made.
The receiving system is sufficiently different from conventional
design to warrant a more complete description. A block
diagram is shown in Fig. 4. The entire receiver is frequency controlled,
and contains four mixerstages which heterodyne the radio-frequency signal
to a final intermediate frequency of 180 cps. Since the first
three injection frequency voltages as well as the final radio frequency
are derived from multiples of common crystal oscillator, a highdegree
of frequency stability is achieved in the system. This highdegree
of stability is essential to permit tuning the highly selective receiver
to the frequency of the echo signal. This tuning is accomplishedin
the final heterodyne stage.
In tuning it is necessary to take account of the change
in frequency of the returned signal which resultsfrom variations
in the relative velocity of the moon with respect to theearth. The frequency
of the returning echo may differ from the transmitted frequency by as much
as 300cycles per second, since the relativevelocities of the earth and
moon vary from about +900 mph at moonrise to-900 mph at moonset.
At the frequency of the transmitter, a relativevelocity of 3 miles per
hour between antenna and target causes a shift of approximately
1 cycle per second in the received signal. This frequency
shift, due to relative velocities of thetransmitting antenna and target,
ispresent in all radar echoes from moving targets, but is undetected in
conventional receivers because the band-width of the normal receiver is
many times greater than the frequencyshift. In the Diana receiver,
a band-width of 57 cps is achieved in the final IF stages. It
is therefore necessary to predetermine the Doppler frequency shift for
the particular observation being made, and to select the proper
crystal for the final heterodyne mixer. To achieve the highdegree of accuracy
required in thefinal mixer, provision is made to modify the frequency
of the crystal-controlled oscillator by means of ascrewdriver control which
varies theair gap above the crystal. Final adjustment of the oscillator
is made bybeating the crystal oscillator output against a secondary frequency
standard source, and observing the outputon a monitoring oscilloscope.
The output of the final heterodyne mixer is fed into two channels, one audio, the other video. The audiochannel is simply a power amplifier stage with the output connected toa loud speaker. The video outputchannel is fed into a second detectorto recover the envelope of the 180 cps intermediate-frequency signal,and then is amplified by a high-gain video amplifier and connected directly to the vertical deflecting plates of anine-inch cathode-ray tube. The horizontal deflection is a linear 4-secondtype-A sweep. The visible output is the characteristic low-frequencynoise pattern representing a 57-cycle bandwidth centered at 180 cycles. A sudden upward departure from the base line occurs when an echo signal is received from the moon. This isshown clearly in Fig. 5. The audible signal is random noise of 57-cyclebandwidth, superimposed on a fixed-frequency note, at the intermediatefrequency of 180 cycles, when the echo is received.
As stated previously, tuning of the receiver is accomplished in the final mixer. The injection signal frequency must be calculated for each observation to take into account the
relative velocity of target and antenna due to both the rotational velocity of the earth and the orbital velocity of the moon. These data, together with azimuth angle and time, are calculated daily from information given by the Nautical Almanac and Ephemeris. The detection of the frequencydue to Doppler shift is made with a high degree of accuracy by the selective receiver. This in itself is corroboration that the echo signal is from the moon. Also, the echo interval of 2.4 seconds admits of no other explanation.
The pre-amplifier of the receiver consists of a three-stage tuned r-f amplifier employing two grounded-grid stages (654) followed by a 6SH7 tuned amplifier at the transmitter frequency. The overall gain of the pre-amplifier alone is 30 db with an overall noise figure of 3.5 db and a bandwidth of 1 me. The electrical design of the first two stages was suggested by a development of Dr. F.B. Llewellyn. A simplified schematic of the first two stages is shown in Fig. 6. The use of concentric tubing inductances forthe tuned circuits provides automatic r-f filtering on the direct-current and filament leads. The pre-amplifier wasdesigned originally as an improvement kit for the SCR-271 radar, andlike the transmitter and receiver, was chosen for the Diana experiment because it satisfied one of the requirements, that of a very low noise figure receiver. A tuned impedance-matching transformer is used between the
receiver and transmission line to convert the 250-ohm balanced
input to the 50-ohm unbalanced input ofthe pre-amplifier.
The transmit-receive switching system (t/r box) employed in theoriginal experiment was a set of two mechanically-operated shorting barson the transmission line, operating from a multivibrator-controlled relay during the transmitted pulse interval of 0.25 sec. One of the shorting barsserves to short out the receiver input during transmission, and the other shorts out the transmitter during reception.
Keyer and Indicator
The visual indicator used is a nine-inch electrostatic cathode-ray tube, 9EP7, with a long-persistence screen.The electron beam is caused to scan the width of the tube, synchronouslywith the transmitted pulse, in 4 seconds, forming a linear time base.The persistence of the tube is long enough to retain the pattern for atleast two sweeps. The circuit employed to generate this sweep is a direct-coupled transitron sawtooth oscillator, described below. A pulse equivalent in time to the keying pulseis also generated by this circuit and is applied to the cathode of a low-level multiplier stage of the transmitter, causing it to conduct for the pulseduration and to drive the subsequentmultipliers.
The time-base generator consists essentially of a high-gain pentode amplifier with capacitance couplingbetween plate and grid. The schematic is shown in Fig. 7. The capacitance coupled path includes a cathode follower stage, the left hand sectionof V2 For the duration of the conduction cycle, the anode voltage ofthe pentode V1, drops and capacitorC1, begins to discharge through thetube. As the voltage on the platedrops the current flow in C1, drives the grid negative, tending to cut off the plate current. A condition of dynamic equilibrium then exists with the plate voltage dropping at a linear rate determined by R1, and C1, and thegrid being maintained at a constant voltage, since each decrement in plate voltage causes a corresponding dropon the grid which keeps the grid signal and hence the output of the tube substantially constant. The time constant of R1, C1, is chosen to cause C1,to become fully discharged during the cycle.
When the plate voltage drops to the point where electrons from the cathode can no longer flow to it, anincrease in screen current occurs which rapidly decreases the screen
voltage and correspondingly decreases the suppressor voltage. Thisaction, which is cumulative, has the effect of suddenly cutting off theanode current. This causes the cathode current to be retarded by the suppressor grid and made to flow to the screen. A negative pulse appears atthe screen, and C1, begins to chargethrough the cathode follower until a point is reached where theplate begins to draw current and the oscillator is recycled. The screen returns to its original voltage, and the plate voltage begins to fall. By suitable choice of R1, and C1, a range of from about 0.1 to 3 cps is obtained.
Keying-voltage signals are derived from the differentiated output of the negative pulse appearing on the screen of the oscillator. This is used to trigger a multivibrator whose time constant is controllable by a variable 5 meg resistor, varying the output pulse width from 0.02 to 0.25 seconds.
The addition of the cathode follower stage V2,was made to shorten the charge time of C1, by causing it to charge through the grid cathode space of the cathode follower. This reduces the return trace time. Tube V3, serves as a degenerative phase-inverting amplifier to secure push-pull sweep voltage.
The keyer multivibrator is a conventional cathode-coupled flip-flopcircuit with the initiating trigger applied as a positive pulse on the grid of the normally non-conducting section. A positive pulse varying in width from 0.02 to 0.85 seconds is obtained at the plate of the other section. This signal is applied to a normally cut off pentode whose loadimpedance is the cathode of the12.388 me amplifier stage in the transmitter. For the duration of thisapplied signal, the plate of the amplifier is driven negative, taking thecathode of the keying tube down with it, thus causing it to conduct.
The first echoes from the moon were received at moonrise
on January 10, 1946. The indication was of the audible type in
the form of a 180-cycle beat note occurring 2.5 seconds after
transmission. Although numerous observations have been
made, both at moonriseand moonset, echo returns do not occur after
every transmission. Further measurements are needed before precise
scientific conclusions can be drawn.
(1) The Radar Equation, Electronics p. 92, April 1945.
Page updated February 7, 2004 page created July 7, 1998 ?
Back to the InfoAge HomepageBack to the Press Index