LIBRARY OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 AT URBANA-CHAMPAIGN 
 
 LkCeY 
 OOp.2. 
 
The person charging this material is re- 
 sponsible for its return to the library from 
 which it was withdrawn on or before the 
 Latest Date stamped below. 
 
 for disciplinary act.on and may result in 
 trenTJValr'TeiepHone Center, 333-8400 
 »TY OF .UiKO^UBRARYMURBA^CHA^ 
 
 0-10% 
 
Digitized by the Internet Archive 
 in 2013 
 
 http://archive.org/details/remotepowersuppl410olse 
 
.Report No. 4 10 
 
 
 REMOTE POWER SUPPLY 
 FOR THE APE SYSTEM 
 
 by 
 
 DAVID LYMAN OLSEN 
 
 August, 1970 
 
 JHE LIBRARY OF THE 
 
 
 
 UNIVERSITY OF ILLINOIS 
 AI URBANA-CHAMWUGN, 
 

Report No. 1*10 
 
 REMOTE POWER SUPPLY 
 FOR THE APE SYSTEM* 
 
 by 
 
 DAVID LYMAN OLSEN 
 
 August, 1970 
 
 Department of Computer Science 
 University of Illinois 
 Urbana, Illinois 6l801 
 
 This work was supported in part by Contract No. N000 14-67-A-0305-0007 
 and was submitted in partial fulfillment of the requirements for the 
 degree of Master of Science in Electrical Engineering, August, 1970. 
 
Ill 
 
 ACKNOWLEDGEMENTS 
 
 The author is grateful to Professor W. J. Poppelbaum for suggesting 
 the topic. He is also indebted to Mr. Fred Ore of the Antenna Laboratory. 
 He would like to thank Miss Car la Donaldson for typing the manuscript. 
 Particular thanks go to Mr. Fred Hancock of the Drafting Department and Mr. 
 Dennis Reed of the Offset Department for the prompt and courteous rendering 
 of their services. 
 
IV 
 
 TABLE OF CONTENTS 
 
 1 . INTRODUCTION 
 
 Page 
 
 2. A REMOTE POWER SUPPLY 3 
 
 2.1 The Need for a Remote Supply 3 
 
 2.2 Methods of Providing Power h 
 
 2.3 Requirements of the Power Supply System 
 
 2.k Initial Experimentation 6 
 
 3. THE TRANSMITTER 9 
 
 3.1 The System Proposed 9 
 
 3.2 The 216 MHz Transmitter 9 
 
 3«3 Frequency Multiplication Using Power Varactor Diodes 12 
 
 3.^ The 216 MHz to 1+32 MHz Doubler 17 
 
 3.5 The 1*32 MHz to 1296 MHz Tripler 17 
 
 3.6 The 1296 MHz Cavity Amplifier 17 
 
 k. THE ANTENNA SYSTEM 25 
 
 k.l General Considerations 25 
 
 k.2 The Transmitting Antenna 29 
 
 k.3 The Receiving Antenna 30 
 
 5. THE VOLTAGE REGULATOR 33 
 
 6. PERFORMANCE OF THE SYSTEM 3k 
 
 LIST OF REFERENCES 36 
 
LIST OF FIGURES 
 
 Figure Page 
 
 1. Experimental Setup 7 
 
 2. Block Diagram of the Transmitting System 10 ' 
 
 3. Schematic Diagram of the 216 MHz Transmitter 11 
 
 k. Photograph of the 216 MHz Transmitter 13 
 
 5. Multiplier Block Diagram and Signal Waveforms 15 
 
 6. Schematic Diagram of the 216 MHz to ^32 MHz Doubler 18 
 
 7. Photograph of the 216 MHz to ^32 MHz Doubler 19 
 
 8. Schematic Diagram of the i+32 MHz to 1296 MHz Tripler 20 
 
 9. Photograph of the ^32 MHz to 1296 MHz Tripler 21 
 
 10. Schematic Diagram of the 1296 MHz Cavity Amplifier 22 
 
 11. Photograph of the 1296 MHz Cavity Amplifier 23 
 
 12. Corner Antenna Configuration 26 
 
 13. Radiation Pattern of the Helical Transmitting Antenna .... 31 
 lk. Photograph of the System 35 
 
1. INTRODUCTION 
 
 Professor Poppelbaum's Circuit Research Group in the Computer Science 
 Department at the University of Illinois has in recent years concentrated on 
 uncommon forms of information processing, transfer and display. A major 
 effort has dealt with stochastic processing techniques, i.e. techniques in 
 ■which a continuous variable is represented by the probability of occurrence 
 of a random pulse. ' The basic concept of stochastic computing is that, if 
 one input to an AND gate has a probability P of being a "1" and the other 
 input a probability P of being a "1", then the output has a probability P-,P C3 
 of being a "1" if the inputs are independent. This a very simple way to 
 multiply. The fundamental operations of addition, subtraction, multiplication 
 and division can be performed with simple logic gates. 
 
 A working programmable analog computer, RASCEL (Regular Array of 
 
 (2) (3) 
 Stochastic Computing Element Logic), has been built. This machine 
 
 uses a tree structure of general computational elements, each capable of 
 
 providing an output which is the sum, difference, product or quotient of its 
 
 inputs. By setting each element to perform an appropriate function, the 
 
 final output can be made to represent complicated combinations of the input 
 
 variables. 
 
 Following a suggestion by Professor Poppelbaum, a further endeavor 
 
 into stochastic processing is being made by another member of the Circuit 
 
 Research Group with the APE Machine (APE = Automomous Processing Element). 
 
 This machine extends and generalizes RASCEL by assembling arbitrary numbers 
 
 of general computation elements through radio frequency channels. It is a 
 
 general programmable computer consisting of a set of APEs, a set of sensors 
 
 to provide raw input data for processing, and a program control unit. There 
 
are different types of channels for data flow and control messages. The 
 control channels are permanently established but the data channels which 
 link various parts of the computer can be established remotely into any 
 desired configuration through the program control unit. The distinct features 
 of the APE machine can be outlined as follows : 
 
 1. The computer does not have to be assembled into a physically 
 interconnected network. It is in the form of a set of small 
 pieces arbitrarily located without any interconnections. 
 
 2. The computer has high versatility in adjusting its capacity 
 for individual application. An increase in capacity simply 
 means putting additional elements into the neighborhood of 
 the other elements. 
 
 3. The APEs, which are the major building blocks of this computer, 
 have high potential adaptability to being produced in large 
 quantity by large scale integration technology. 
 
2. A REMOTE POWER SUPPLY 
 
 2.1 The Need for a Remote Supply 
 
 As previously outlined, the APE system is a network which is not 
 physically interconnected: The exact placement of the APEs need not be of 
 concern to the user. When an increase in computing capability is desired, 
 more APEs are ordered and placed in the active region. This "active region" 
 is the area reserved for the placement of the APEs into which the radio 
 signals for control and data flow are directed. No elaborate mounting 
 brackets are needed and no cables need to be used to connect the APEs. The 
 system design is based on flexibility. 
 
 To be consistent with the ideas outlined thus far, the APEs must 
 also derive their operating power without a physical connection to a power 
 source. The available power must be sufficient to: 1) operate the receivers 
 which pick up the data and control information, 2) perform all the logical 
 operations required and 3) transmit an answer to the next APE. 
 
 An obvious solution would be to supply the operating power with a 
 battery, internal to each APE. However, this solution is unacceptable for 
 various reasons. First, a battery necessarily contains only a limited 
 amount of energy. The battery in each APE would have to be replaced 
 periodically. This would be a costly and time-consuming task, especially 
 in a system with many APEs. Second, some type of switching would be necessary 
 to "turn off" each APE when the system is not operating. A monitor circuit 
 could be devised to determine when the system is on and to control accordingly 
 the supply of power to the other sections of the APEs. This would still 
 present a constant drain on the battery. Alternately, a manual on-off switch 
 could be provided on each APE. With many APEs, however, this would be a very 
 
time-consuming task, especially if the system were operated for short and 
 frequent intervals. Third, the physical size of the batteries is incompatible 
 with the very small APEs to be developed in the future. 
 
 It is apparent, then, that the operating power for the APEs should 
 be from a remote source. That is, just as the data and control signals are 
 "beamed" to each APE, the operating power must somehow be beamed into the 
 active region from a source external to the APEs. Each APE by itself is an 
 inert black box. When the system is activated, however, the APEs, together 
 with the program control unit, "come alive" to form an operating computing 
 system. 
 
 2.2 Methods of Providing Power 
 
 An obvious solution to the remote powering problem is the use of 
 solar cells. The active region could be illuminated by a light source; each 
 APE would be equipped with a bank of solar cells. The strength of the light 
 source and the number of solar cells required would be dictated by the power 
 
 requirements of the APEs. The APE system, however, has been called the 
 
 (5) 
 "computer in a bag". It is being designed with flexibility in mind. To 
 
 increase his computing capability, the user need only to order more APEs and, 
 upon delivery, place them in the active region. It was decided that if it 
 were possible, the APEs should work even when left unpacked in their shipping 
 cartons. From this standpoint the solar cell solution was eliminated as it 
 would necessitate the unwrapping of the APEs to allow light to strike the 
 bank of solar cells. 
 
 The next method considered was to transfer the energy to the APEs 
 magnetically. The APEs could be placed in a gap, completing a magnetic 
 circuit. This arrangement would resemble a large transformer with many 
 
windings . A primary winding would be constructed and each APE would contain 
 a secondary winding. The magnetic path would have to be iron to carry an 
 appreciable amount of power. Such a circuit, however, cannot have air gaps 
 if it is to perform properly. Therefore, the AFEs would need to fit snugly 
 in the gap to satisfactorily complete the path. Using the APEs in their 
 "unwrapped" condition would constitute an air gap in the magnetic path. Also, 
 the proximity of all the iron, mostly in the form of the large faceplates of 
 the gap, would surely interfere with radio communication to the AFEs. 
 
 The final method considered involves sending power to the APEs over 
 a radio frequency channel. A transmitter would send a continuous, unmodulated 
 signal to the APEs. Each APE would contain an antenna to receive the 
 signal. The received signal would then be rectified, filtered and regulated 
 to provide the voltage required to run the circuitry inside the APEs. It 
 was concluded that with this method the APEs would not have to be unpacked to 
 operate because the wrapping would not hinder the signal. Also, given a 
 powerful enough transmitter, the transmitting equipment could be located a 
 reasonable distance from the AFEs so that it would not physically interfere 
 with the data and control signals. 
 
 2.3 Requirements of the Power Supply System 
 
 One of the basic requirements of the power supply system is that 
 it be able to provide the total power needed by the APEs. This power is 
 used to carry out the logical functions required, to operate the receivers 
 which obtain data and control information and to transmit data to the control 
 
 (h) 
 
 unit and the other APEs. Investigation of various types of circuitry has 
 indicated that complementary symmetry metal oxide semiconductor circuitry 
 would be most suitable for use in the APEs. This circuitry offers 
 
exceptionally low power consumption, is capable of providing long period 
 memory, and has good reliability. It is estimated that an APE, built with 
 state-of-the-art COS/MOS circuitry, will consume as little as lOOmw of power. 
 This lOOmw, then, is the minimum amount of power which must be provided by 
 the power supply system. 
 
 One additional requirement of the power system is that it should 
 not interfere with the communication channels. It was expected that low 
 efficiencies would be realized between transmitter power and the usable power 
 received by the AFEs . Thus, a relatively high-power transmitter was in order. 
 It was feared that, if the power transmission frequency chosen was below the 
 communications channels, harmonics might interfere with these communications. 
 For that reason, initial investigations were conducted at high frequencies. 
 
 2.h Initial Experimentation 
 
 As mentioned in the previous section, a relatively "high" frequency 
 channel was needed for the power transmission. A Rhode and Schwartz 
 transmitter was available for experimentation. This instrument offers an 
 
 
 operating range of 275-2750 MHz in two segments. An initial frequency of 
 350 MHz was chosen since the power available from the transmitter was a 
 maximum near that frequency. Two simple half -wave length dipole antennas 
 were constructed. The transmitting dipole was connected to the output of 
 the Rhode and Schwartz generator (see Figure 1) with a length of 50-ohm 
 RG-8/u cable. A balun would be required to match the dipole to the cable, 
 but to save time, a balun was not constructed. The receiving dipole was 
 connected to a full wave bridge of 1N82 diodes. The output of the bridge 
 was connected to a Simpson volt-ohm meter to measure DC voltage. With a 
 transmitter power of 1.5 watts and an antenna separation of about five feet, 
 
< o 
 
 -J E 
 
 a. 10 
 
 UJ 
 
 x 
 
 < 
 o 
 o 
 
 2 
 
 Z -I 
 «° 
 HO 
 
 & 
 -p 
 
 0) 
 CO 
 
 H 
 
 •P 
 
 a 
 
 •H 
 
 I 
 
 Q) 
 
 
the meter registered 10 volts. (The power indicated was not exact due to the 
 
 previously mentioned mismatch.) With a greatly reduced antenna separation, 
 
 (6 ) 
 a 6 volt ho mi 1 .1 1 ampere bulb could be powered. 
 
 The results of this initial test were encouraging since it showed 
 that the proposed scheme might be feasible. However, the maximum physical 
 dimension of an APE is visualized as being smaller than the size of a dipole 
 at 350 MHz (16 inches). Therefore, further tests -were conducted at a higher 
 frequency. 1600 MHz was chosen because this, again, was a frequency at which 
 the power available from the transmitter was a maximum. The experimental 
 setup was the same as at 350 MHz except that shorter dipole antennas were 
 used. Similar results were observed. 
 
 Additional tests were conducted to determine which of the available 
 high-frequency diodes was best suited for use in the receiving bridge. 
 Selection was based on the voltage measured across the bridge; those diodes 
 giving the higher voltage were judged better than those giving a lower readin 
 Of the three types tested, the 1N82s worked best. 
 
 i 
 
 i 
 
3. THE TRANSMITTER 
 
 3.1 The System Proposed 
 
 A brief literature search was conducted concerning power generation 
 
 in the gigahertz region. Various publications of the American Radio Relay 
 
 (1 \ ( P>\ 
 League proved to be quite helpful. It was decided that the 
 
 transmitter would consist of a 216 MHz transmitter, two frequency multipliers 
 
 and a cavity amplifier operating at 1296 MHz (see Figure 2). This final 
 
 frequency would require a dipole size of about 4.3 inches which is compatible 
 
 with the size of an APE. 
 
 3.2 The 216 MHz Transmitter 
 
 The construction of this unit, as well as the other three units to 
 
 be described, follows the plans given in publications of the American Radio 
 
 ( 7 ) (R) 
 Relay League ' The transmitter is crystal controlled and runs with up 
 
 to kO watts input to the final stage. Referring to Figure 3 5 an overtone 
 
 oscillator circuit uses one half of a 12AT7 dual triode. The frequency of 
 
 oscillation is 2k MHz, the crystal working on its third overtone. The second 
 
 half of the 12AT7 triples to 72 MHz. This stage has a balanced plate circuit 
 
 so that its output may be capacitively ^oupled to the grids of a second 12A3T, 
 
 working as a push-pull tripler to 216 MHz. The plate circuit of this second 
 
 tripler is inductively coupled to the grid circuit of a 636O dual tetrode 
 
 amplifier, operating at 216 MHz. Similar inductive coupling transfers the 
 
 drive to the grid circuit of the final stage, a 6252 dual tetrode. A small 
 
 trimmer capacitor, C^, bypasses the screen to ground to help stabilize 
 
 operation. A milliammeter and a selector switch were installed to permit a 
 
 check on voltages and currents throughout the circuit. These checks are 
 
10 
 
 IV 
 
 e 
 
 9 y 
 
 oc u> 
 
 * u 3 
 <« o < 
 
 cj 5 >- 
 o 
 
 X ^ UJ 
 S o2 -I 
 
 (M»-tf>?: 
 to 
 
 
 
 CJ 
 
 § 
 •P 
 W 
 
 >S 
 
 CQ 
 
 hD 
 
 ■H 
 -P 
 -P 
 
 1 
 
 
 -P 
 
 o 
 
 •H 
 
 « 
 
 O 
 O 
 H 
 
 OJ 
 
 •H 
 
 P=4 
 
 uj Q 
 
 CJ < u z 
 
 <r w o 
 
11 
 
 UJ 
 _l K1 
 
 2 
 < 
 
 
 ^"T^ 
 
 rr 
 
 ytr 
 
 ■^tf-Ht— 
 
 -l t- 
 a < 
 
 _/"yrv-\_ 
 
 i[]!R» 
 
 i — v^ — • — wv— ii 
 
 — ^npn 1, 
 
 I V\^ 
 
 -^ 
 
 is 
 
 M 
 M 
 
 I! 
 
 O UJ 
 
 IS 
 
 V V 
 
 (> 
 
 ^ 
 
 <D 
 -P 
 -P 
 
 m 
 
 3 
 
 CM 
 
 d) 
 
 -P 
 
 <+h 
 O 
 
 •H 
 Q 
 
 o 
 •H 
 
 -P 
 
 I 
 o 
 
 
 
 •H 
 
12 
 
 especially helpful when tuning up the transmitter. As noted on the schematic 
 drawing (Figure 3)> "the metering circuitry has been deleted for clarity. 
 Figure k shows a photograph of the 216 MHz transmitter. 
 
 3.3 Frequency Multiplication Using Power Varactor Diodes 
 
 Frequency multiplication using power varactor diodes offers a 
 simple means of generating high level microwave power. An oscillator 
 operating at a lower frequency drives the multiplier. The non-linear element 
 in the multiplier is a power varactor diode which operates as a charge- 
 controlled switch. The switching action generates harmonics of the drive 
 signal; the desired output is selected with a band-pass filter. 
 
 The usual high power varactor diode multiplier operates in the 
 shunt mode. That is, the diode is placed in shunt with the signal path. 
 The shunt mode is chosen above the series mode mainly for its greater power- 
 handling capacity. With one end of the diode grounded, a low thermal 
 impedance path can be provided to a ground plane for heat dissipation. 
 Varactor diodes are available in packages with a stud on one end; effective 
 heat sinking and simple mounting are achieved by screwing the stud into the 
 ground plane. Diodes of this type can handle several tens of watts of input 
 power . 
 
 Another reason for using the diode in the shunt mode is the 
 isolation it provides between input and output circuits. The diode acts as 
 a closed switch across the signal path during most of the input cycle. This 
 action prevents harmonics of the drive frequency from feeding back into the 
 input circuit. This inherent isolation makes the design simple, and good 
 overall efficiency can be achieved even for high orders of multiplication. 
 
13 
 
 0) 
 P 
 P 
 •H 
 
 s 
 
 w 
 Fh 
 
 EH 
 
 H 
 CM 
 
 0) 
 
 -P 
 
 Ch 
 O 
 
 ?H 
 
 t»D 
 
 O 
 -P 
 
 o 
 
 -4- 
 
 •H 
 
11+ 
 
 A varactor diode multiplier typically consists of five major parts, 
 as shown in Figure 5. The main purpose of the matching section is to provide 
 a good impedance match between the source and the impulse section for maximum 
 power transfer. It must also provide DC isolation between the source and the 
 diode and high-frequency isolation between the impulse section and the source. 
 The output of this section has the same form as the input, but at a lower 
 impedance level. 
 
 The impulse section is the most crucial part of the multiplier. Its | 
 function is to convert the input sinusoid into a narrow impulse with high 
 harmonic content. In this section the diode shunts the line and switches the 
 current through a series inductor. The voltage transient across the inductor 
 then forms the impulse shown in Figure 5. The charge -controlled switching 
 action of the diode can be explained briefly. During the positive half-cycle 
 of the drive signal, the diode is driven heavily into forward conduction. 
 This results in the storage of charge in the depletion region by minority 
 carriers. This charge is extracted during the reverse half-cycle of the 
 driving voltage, resulting in low diode impedance and large extraction current. 
 When the stored charge is depleted, the diode conduction ceases very abruptly. 
 The drive inductor will try to maintain the collapsing current, giving rise to 
 a large voltage spike. The amplitude of this voltage spike, or impulse, is 
 determined by the signal source amplitude, the value of the drive inductor, 
 the diode transition speed from low to high impedance, and the circuit losses. 
 The shape of the impulse is determined primarily by the reverse bias 
 capacitance of the diode and the circuit loading effects. 
 
 The bias voltage applied to the diode controls the forward current 
 through the diode during the positive half- cycle of the drive signal. This 
 determines the amount of stored charge and hence, the point on the negative 
 
 
15 
 
 
 St J 
 
 (t? 
 
 uj 9 
 
 ^t> 
 
 U. U) 
 
 to 
 
 
 
 
 or 
 
 o z 
 
 fe9 
 
 z -»- 
 
 o <-> 
 
 en uj 
 
 LU CO 
 
 or 
 
 
 
 UJ z 
 
 
 
 CO O 
 
 
 / 
 
 / 
 
 / 
 / 
 
 IMPUL 
 SECTI 
 
 
 
 / 
 
 
 z 
 
 CO 2 
 
 / 
 
 
 f 
 
 i 
 
 i 
 
 < h- 
 
 
 
 ■* 
 
 UJ 
 CO 
 
 \ 
 \ 
 
 
 
 
 
 \ 
 
 
 
 \ 
 \ 
 
 \ 
 
 2 Z 
 
 E o 
 
 
 
 N 
 
 MATCH 
 SECTI 
 
 3 
 CL 
 
 e 
 
 o 
 
 > 
 
 5= 
 
 •H 
 
 CO 
 
 M 
 o5 
 •H 
 Q 
 
 M 
 o 
 
 o 
 
 PQ 
 
 <D 
 •H 
 
 ■4 
 
 •H 
 -P 
 
 rH 
 
 I 
 
 ir\ 
 
 •H 
 
 (in 
 
16 
 
 half-cycle at which switching occurs. This switching point is a factor in 
 determining impulse amplitude and stability with respect to driving power and 
 operating temperature . Since the impulse amplitude determines the strength 
 of the output signal, the bias therefore affects the efficiency of the 
 multiplier. External bias can be injected directly into the impulse section 
 or though the adjacent sections. Alternatively, the slight rectification by 
 the diode can develop self bias by providing a DC current path with a voltage- 
 dropping resistor. Since both rectified current and stored charge are 
 proportional to the amplitude of the drive signal, self bias tends to 
 compensate for changes in drive level over a good dynamic range. The 
 rectification efficiency of the varactor diode is very low; consequently, a 
 large resistance in parallel with the diode will develop the necessary bias. 
 
 The output of the impulse generator excites a resonator tuned to 
 the desired harmonic of the drive frequency. The output of this section is 
 a damped sinusoid which rings at the output frequency. This waveform occurs 
 once per input cycle, and the damping is affected by the loading at both 
 ends of the resonator. 
 
 The output filter is the final section of the multiplier. It 
 consists of a bandpass filter which extracts from the damped waveform a pure 
 sine wave at the desired harmonic of the input frequency. The filter, should 
 provide strong rejection of the drive frequency and all other harmonics up 
 to several times the output frequency. The application will determine the 
 necessary rejection of spurious harmonics. 
 
17 
 
 3.1+ The 216 MHz to ^32 MHz Doubler 
 
 The circuit for the 216 MHz to 432 MHz doubler is shown in Figure 6. 
 The various parts of the circuit can be compared to the operating theory given 
 in the previous section. The trap, L-C , is mostly to simplify tuning. 
 Without it, changing the output capacitor, C , would also change the tuning 
 of the 216 MHz circuit. The double-tuned input and output circuits help 
 establish that the output is on the desired frequency. The circuits 
 themselves are rather low-Q. Figure 7 shows a photograph of the 216 MHz to 
 J+32 MHz doubler. 
 
 3.5 The i+32 MHz to 1296 MHz Tripler 
 
 The circuit of the ^32 MHz to 1296 MHz tripler is given in Figure 8. 
 The theory behind the operation of the tripler, as for the doubler, is given 
 in a previous section. 
 
 At these high frequencies, except for the ^32 MHz circuits, coils 
 and wire-lead capacitors are out of the question. Strip lines are used for 
 coils. Capacitors are fashioned with machine screws; a screw runs through 
 the chassis of the multiplier and the capacitance is determined by its 
 spacing from the strip line. 
 
 Figure 9 shows a photograph of the ^32 MHz to 1296 MHz tripler. 
 
 3.6 The 1296 MHz Cavity Amplifier 
 
 This amplifier uses two 2C39A triodes, operated in parallel, in a 
 cavity assembly. UHF circuits, particularly those involving cavities, do 
 not lend themselves well to conventional schematic presentation. The 
 equivalent diagram, however, is given in Figure 10. This is a grounded-grid 
 amplifier. Referring to Figure 11, the large square box houses the cathode 
 
18 
 
 rvrw 
 
 to 
 
 H 
 
 o 
 
 CM 
 
 m 
 
 o 
 -p 
 
 CM 
 
 <M 
 
 KJUO>U 
 
 O 
 
 <> 
 
 A /\ 
 
 ~3 
 
 < 
 
 o 
 
 o 
 
 I 
 < 
 
 (/) 
 
 < 
 
 o 
 o 
 
 CO 
 
 en 
 < 
 
 O 
 <T 
 O 
 
 O 
 
 rH 
 
 CM 
 0) 
 -P 
 
 O 
 
 05 
 
 •H 
 
 o 
 o 
 
 ■H 
 
 -P 
 
 <a 
 s 
 a) 
 
 o 
 
 CO 
 
 0) 
 
 •H 
 
19 
 
 
 O 
 P 
 
 CM 
 
 O 
 -P 
 
 3 
 CM 
 
 <D 
 
 -P 
 
 Ch 
 O 
 
 o 
 -p 
 
 cu 
 
 •H 
 
esj 
 
 20 
 
 OHHi' 
 
 •H 
 
 OJ 
 H 
 
 O 
 -P 
 
 Hb 
 
 CD 
 
 ro 
 m 
 c»> 
 <fr 
 i 
 < 
 
 (/> 
 
 UJ 
 
 u 
 o 
 
 C/> 
 < 
 
 UJ 
 
 I 
 
 o 
 or 
 o 
 
 or 
 o 
 
 OJ 
 
 0) 
 
 -P 
 
 o 
 
 8P 
 
 •H 
 Q 
 
 o 
 •H 
 
 1 
 
 O 
 
 00 
 
 •H 
 
21 
 
 u 
 a> 
 H 
 ft 
 
 •H 
 EH 
 
 VD 
 ON 
 CM 
 H 
 
 O 
 -P 
 
 CM 
 
 m 
 
 0) 
 43 
 -P 
 
 <4H 
 O 
 
 .3 
 
 Ph 
 a5 
 U 
 
 W 
 
 o 
 -p 
 o 
 
 ON 
 CD 
 
 bfl 
 
 •rl 
 P>H 
 
22 
 
 a. 
 
 i- 
 
 \/ v 
 
 OH 
 
 to 
 
 X 
 
 < 
 
 I 
 
 U> ro 
 CVJ 
 
 rnpr^ — i 
 
 t 
 
 Hi- 
 
 im- 
 
 "*■ § ' 
 
 Hi' Hi 
 
 5f 
 
 o 
 
 /\ /\ 
 
 u 
 
 ■H 
 
 & 
 
 -3 3 
 
 a. 
 
 
23 
 
 •H 
 <H 
 •H 
 
 I 
 
 -P 
 
 •H 
 
 S3 
 
 O 
 
 OJ 
 H 
 
 <D 
 
 +3 
 
 <H 
 O 
 
 bO 
 O 
 •P 
 
 o 
 
 1 
 
 ID 
 
 ■H 
 
2k 
 
 input circuit. Input coupling is capacitive through C , a small glass 
 
 trimmer; an approximate input match is established by adjustment of this 
 
 capacitor. The tuned circuit L-,C is effectively a halfwave line, tuned at 
 
 the end opposite the tubes by C . L, is a metal plate which connects to 
 
 the tube cathodes (also common to one end of the filaments); C is a 
 
 beryllium copper spring finger that is adjusted by a screw running through the 
 
 removable cover plate. The plate circuit L C is also a square cavity, smaller 
 
 in size than the input cavity. This cavity is tuned by means of C , a coaxial 
 
 capacitor constructed with a screw running through the top plate of the bypass 
 
 capacitor, C, . The fixed portion of C is a metal sleeve soldered to the 
 
 bottom of the plate cavity. C, , the previously mentioned bypass capacitor, 
 
 4 
 
 consists of the top cover of the plate cavity, a Teflon sheet, and the top 
 plate which connects to the tube plates. This combination does not act as a 
 pure capacitance because of the large fraction of a wavelength represented by 
 the capacitor plates at 1296 MHz. L~, the output coupling loop, is merely a 
 small strip of sheet metal soldered between the center conductor of the 
 output connector and the bottom of the plate cavity. The entire assembly was 
 silver-plated to assure good RF conduction. A fan was installed to cool the 
 fins on the plates of the tubes. 
 
25 
 
 k. THE ANTENNA SYSTEM 
 
 4.1 General Considerations 
 
 The first experiments conducted were made with simple half- 
 wavelength dipoles. Simplicity of construction was the main reason for 
 this choice. After initial tests, however, it became apparent that it would 
 not be advantageous to use dipoles for both the transmitting and receiving 
 antennas. Since it was desirable to beam the signal in only one direction, 
 the dipole effectively wasted much of the available energy due to transmission 
 in a 360 arc. It was also decided that, if possible, the strength of the 
 signal received by an APE should be independent of the orientation of the APE. 
 This feature would allow the APEs to be more or less randomly placed in the 
 active region. Random placement would free the user from having to worry 
 about the exact position of the APEs. With the use of two dipoles, however, 
 placement could not be random. The strength of the signal received would 
 definitely vary with the orientation of the receiving antenna with respect to 
 the transmitting antenna. Some other system was needed. 
 
 It was proposed by Dr. Poppelbaum that a "corner antenna" be 
 investigated. This configuration (see Figure 12) would involve three dipoles 
 placed at right angles to each other. This setup would make the received 
 signal strength independent of the orientation of the receiving antenna. 
 With proper placement of the transmitting array, the distance from transmitter 
 to receiver would not vary greatly as an APE was moved throughout the active 
 region. This would mean that variations in signal strength would be small, 
 independent of the APEs orientation and position within the active region. 
 It was realized that several problems were inherent with this proposal. The 
 use of a simple dipole was abandoned partly because of its relative 
 
26 
 
 ACTIVE REGION 
 
 Figure 12. Corner Antenna Configuration 
 
27 
 
 non-directionality and resulting waste of power. Using three dipoles would 
 excite fields that give constant signal strength with variation of the 
 receiver position, but the transmitter power would be distributed over all 
 space. Instead of using three dipoles, it would be necessary to use three 
 antennas that were directional enough to contain the transmitter power within, 
 say, an octant of three dimensional space. These three antennas would also 
 require a complex feed arrangement to give the proper phasing and impedance 
 matching with the transmitter. 
 
 At this point some additional thought was given to the receiving 
 antenna. It was decided that the use of a simple dipole would be a good 
 choice. All of the directionality and phasing could be accomplished in the 
 transmitting system, because it needed to be produced only once. The 
 receiving antenna, however, would have to reproduced many times (once for 
 each APE) and should be as simple as possible to construct. A dipole seemed 
 a likely answer. 
 
 An APE, then, would be a rectangular box with a dipole on one side. 
 To shield the circuits in the APE from the strong signals of the power supply 
 system, the box should be a metallic structure. This fact introduces an 
 additional problem. So far, all concern was for the signal received versus 
 the orientation of the receiving antenna. No consideration had been given 
 to the receiver as a unit, including the box that houses the APE. It is 
 quite possible, however, that the APE, when placed randomly, would be 
 oriented in such a way that the receiving dipole would be on a side of the 
 APE opposite from the transmitting array. That is, the metal box which 
 houses the APE itself could end up between the transmitting and receiving 
 antennas. The presence of the box would surely interfere with the reception 
 of the signal, especially at a frequency as high as we have here. This 
 
28 
 
 difficulty could be overcome by altering either the transmitting or the 
 receiving arrangement. For the former, one could imagine setting up 
 additional transmitting arrays so that the signal would be incident on an 
 APE from more than one direction. However, this would involve more antennas 
 and a much more complex feed arrangement. To alter the receiver, dipole 
 elements could be placed on all sides of the APE; for no matter which way the 
 APE was positioned, one of the dipoles could face the transmitting array. 
 Such an arrangement would require, as one possible hook-up, a diode bridge 
 for each dipole. With their outputs connected in parallel, the dipole 
 receiving the strongest signal would dominate and supply power to the APE. 
 
 At this point the author decided to use a single circularly 
 polarized directional element for the transmitting antenna. A dipole facing 
 such an antenna would receive a constant signal as it was rotated in a plane 
 perpendicular to the axis of the antenna. It was also decided to place the 
 transmitting antenna directly above the active region. For this arrangement 
 to work satisfactorily, the side of an APE which ends up facing the antenna 
 must contain a dipole. Rotation of this dipole relative to the antenna does 
 not affect the strength of the signal received. The active region will be a 
 plane, placed perpendicular to and directly below the transmitting antenna. 
 The distance of the antenna from the plane will be adjusted so that the 
 signal strength does not vary too greatly over the area. Thus as long as 
 a dipole on an APE faces up, the placement of the APE in the plane can be 
 random . 
 
29 
 
 k.2 The Transmitting Antenna 
 
 A great deal of help with the circularly polarized transmitting 
 antenna was given by individuals in the Antenna Laboratory of the Electrical 
 Engineering Department. The first antenna tried was of conical equiangular 
 spiral construction. This antenna was tried because it was both readily 
 available and would operate at the required frequency. This design was 
 discarded because the results were not significantly better than with the 
 dipole element already used. 
 
 It was decided that a helical antenna might provide the desired 
 features. From tests previously made on helical antennas, a pattern was 
 found which closely matched the pattern desired for the transmitting antenna. 
 The dimensions of that antenna were scaled to give the same performance as 
 that recorded but at the new operating frequency of 1296 MHz. The new 
 antenna was constructed from a length of polystyrene tubing turned to a 
 diameter of 1.17 inches. Ten cells were wound with No. 22 wire and a pitch 
 of 3-28 inches to give an overall length slightly under three feet. 
 
 Before a matching network could be built to connect the antenna to 
 the cavity amplifier, it was necessary to measure the impedance of the 
 antenna at the feed point. This helical antenna is front fed and presents 
 a balanced load. The transmission line runs along the axis of the antenna 
 and out the back end. Using slotted line techniques, the impedance of the 
 antenna was determined as being approximately 375 ohms. A balun was necessary 
 to match the unbalanced output of the cavity amplifier to the balanced load 
 of the antenna. A split-tube balun was constructed. This type of balun also 
 gives a four-to-one impedance ratio. Using RG-8/u cable between the cavity 
 amplifier and the balun, the output impedance of the balun would be roughly 
 
30 
 
 
 200 ohms. Finally, a quarter-wave matching section was constructed to 
 connect the 200 ohm balun to the 375 ohm antenna. 
 
 When the setup was tried out with the cavity amplifier, the 
 performance was worse than expected. The results were not significantly 
 better than with the dipole antenna. A check with the slotted line showed 
 that the trouble was the result of a high standing wave ratio, about 3*8. 
 The high standing wave ratio meant that too large a percentage of the available 
 power was being reflected back to the cavity amplifier. A quick analysis of 
 the equipment showed that the split-tube balun had been constructed incorrectly 
 A new balun was built and markedly better results were obtained. The standing 
 wave ratio was again measured, and found to be about 1.3- This is a very 
 
 
 
 good figure for an antenna, and represents an almost complete transfer of 
 power to the load. 
 
 A radiation pattern was plotted for the helical antenna. The 
 pattern is shown in Figure 13. The antenna has a beam width of about 70 
 and a very good front-to-back ratio. The magnitude of the back lobe is a 
 function of the length of the antenna. The small back lobe shows that the 
 current in the front-fed antenna has almost disappeared at the back. An even 
 smaller back lobe could be realized by winding more cells, and as a result, 
 give a physically longer antenna. 
 
 4.3 The Receiving Antenna 
 
 It was decided in the beginning that the antenna used on the APEs 
 should be as simple as possible since it would have to be reproduced many 
 times. The transmitting antenna, however, could be made as complex as needed 
 since the physical size was not important and only one unit was required. 
 
31 
 
 Figure 13 . Radiation Pattern of the Helical Transmitting Antenna 
 
32 
 
 A half -wavelength dipole was the logical antenna to use for the receiver. In 
 fact, the frequency used (12^6 MHz) was determined partially by the length of 
 a dipole at that frequency. 
 
 Eventually it became apparent that a slot antenna would have to be 
 used. The APEs would be housed in metal boxes. A conventional dipole placed 
 on the exterior of such a box could hardly be expected to work properly. Also, 
 the presence of other APEs in the vicinity would surely interfere with the 
 pattern of a dipole. A slot antenna, on the other hand, seemed ideally 
 suited for this situation. Its operation depends on the existence of a ground 
 plane. This ground plane is automatically available in the form of the side 
 of the box, into which the slot is cut. Also, adjacent APEs will not 
 interfere with reception since they represent an extension of this ground 
 plane. 
 
 The length of the slot should be about ninety-five percent of a 
 free-space half -wavelength, the same as with a conventional dipole. The 
 width of the slot should be as narrow as possible for the lowest possible 
 impedance. A test receiver was constructed from a 2 x h x 6" brass box. 
 A 1/16" slot was milled into one of the h x 6" faces. Experiments conducted 
 with this box indicated that the antenna did indeed perform satisfactorily. 
 
 When the final version of an APE is constructed, the side of the 
 box into which the antenna is cut should be made as large as practical; for 
 the larger the ground plane of the antenna, the more ideal is the radiation 
 pattern . 
 
33 
 5. THE VOLTAGE REGULATOR 
 
 A regulating circuit is needed to control the voltage which is 
 supplied to the APE. The regulator must absorb energy when a strong signal 
 is received by the APE; yet, it must use very little energy when a weak 
 signal is received. A maximum signal strength will be received when the 
 APE is placed in the middle of the active region, directly under the 
 transmitting antenna. A minimum signal strength will be received when an 
 APE is placed in a corner of the active region. There, it is not only 
 farther from the antenna, but it is also off-axis. A very simple yet 
 effective scheme is to place a zener diode across the output of the bridge. 
 The zener voltage will vary by a few tenths of a volt, depending on the 
 strength of the signal received and the power required by the APE. COS/MOS 
 circuitry, however, works well on any voltage between about six and ten volts 
 so this small variation in supply voltage should cause no problems. 
 
 Due to the inherently high impedance of a slot-dipole antenna 
 (about 500 ohms), no series resistor is needed between the diode bridge and 
 the zener regulator. Filtering will be included in the form of capacitors 
 shunting the zener diode. 
 
3^ 
 6. PERFORMANCE OF THE SYSTEM 
 
 A photograph of the power supply system is shown in Figure Ik. The 
 support structure was made from wood to minimize reflections of the signal 
 and resulting interference patterns . The distance of the transmitter from 
 the active region was made variable to facilitate experimentation. 
 
 The system in its present form works very well. The transmitter 
 has been raised from the position shown in Figure Ik; the active region is 
 about eighteen inches from the end of the antenna. Raising the antenna has 
 the effect of minimizing the difference between the minimum and maximum 
 signal received. The spacing cannot be made too great, however, or the 
 signal will weaken due to the l/R effect. 
 
 Experiments with the previously mentioned slot antenna demonstrate 
 the effect of the circularly polarized signal because the box can be rotated 
 360 with no appreciable change in the signal strength. 
 
 At present the required 100 milliwatts of power cannot be obtained 
 over the entire active region. Improvements must be made in the system to 
 overcome this deficiency. Schottky barrier (hot-carrier) diodes will be 
 tried in the receiving bridge. These diodes offer a lower forward voltage 
 drop than the silicon diodes presently in use, and are able to function at 
 the high speed required. It is also planned to double the power of the 
 transmitter by adding a second pair of triodes in the cavity amplifier. 
 Hopefully the combination of these changes will provide the needed power for 
 the APEs over the entire active region. 
 
35 
 
 Figure Ik. Photograph of the System 
 
36 
 
 LIST OF REFERENCES 
 
 1. Afuso, C, "Analog Computation with Random Pulse Sequences", Report No. 
 
 255, Department of Computer Science, University of Illinois, 
 Urbana, Illinois, February 1968. 
 
 2. Esch, J. W., "RASCEL - A Programmable Analog Computer Based on a Regular 
 
 Array of Stochastic Computing Element Logic", Report No. 332, 
 Department of Computer Science, University of Illinois, Urbana, 
 Illinois, June 1969. 
 
 3. Poppelbaum, W. J., ONR Proposal N000 14-67-A-0305-0007, May I969. 
 
 k. Wo, Y., "APE Machine - A Novel Computer Using Autonomous Processing 
 Elements", Thesis Proposal, Department of Computer Science, 
 University of Illinois, Urbana, Illinois, March 1970. 
 
 5. Riley, W. B. , Electronics , March 2, 1970, pp. 39-40. 
 
 6. Olsen, D. and Wo, Y. , "Quarterly Technical Progress Report, Circuit 
 
 Research Section, Part I", Department of Computer Science, University 
 of Illinois, Urbana, Illinois, October 1969 to December 19&9- 
 
 7. DeMaw, D., The Radio Amateur's Handbook , American Radio Relay League, 
 
 Newington, Conn., I969. 
 
 8. Tilton, P. T., The Radio Amateur's VHF Manual , American Radio Relay 
 
 League, Newington, Conn., 1968. 
 
 9. Olsen, D. and Wo, Y., "Quarterly Technical Progress Report, Circuit 
 
 Research Section, Part I", Department of Computer Science, University j 
 of Illinois, Urbana, Illinois, January 1970 to March 1970. 
 
 10. Ryals, A. and Siegal, B., "Snap Varactor Frequency Multiplier Design for 
 
 High-Power VHF to S-Band Conversion in One Step", Application Note 
 No. k, Microwave Associates, Inc., Sunnyvale, California, 
 February 1969. 
 
 11. Patton, W. T., "The Backfire Bifilar Helical Antenna", Ph.D. Thesis, 
 
 University of Illinois, Urbana, Illinois, 1963 • 
 
 12. Huntoon, J., The A.R.R.L. Antenna Book , American Radio Relay League, 
 
 Newington, Conn., 1968. 
 
 13. Jordan, E. C, and Blamain, K. G., Electromagnetic Waves and Radiating 
 
 Systems, Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1968. 
 
 Ik. Westman, H. P., Reference Data for Radio Engineers , International 
 Telephone and Telegraph Corporation, New York, New York, 1956. 
 
UNCLASSIFIED 
 
 Security Classification 
 
 DOCUMENT CONTROL DATA - R&D 
 
 (Security c laaal llcatlon o/ tltlm. body of abatract and Indexing annetallon null be entered when the overall report la c la* tilled) 
 
 1 ORIGINATIN G ACT|\/| T Y (Corporate author) 
 
 Department of Computer Science 
 University of Illinois 
 
 [T-Vmna, T llinnis f, 1 flOl 
 
 la RI^OKT IICUKITV CLASSIFICATION 
 
 Unclassified 
 
 26 OROUP 
 
 |, REPORT TITLE 
 
 REMOTE POWER SUPPLY FOR THE APE SYSTEM 
 
 t DESCRIPTIVE NOTES (Typa ol report and Incluelve datea) 
 
 Master's Thesis. Technical Report August, J97Q 
 
 |. AUTHORfS) (Laat name. II ret name, Initial) 
 
 Oisen, David L. 
 
 5. REPORT DATE 
 
 August, 1^70 
 
 7a. TOTAL NO. OF Ptait 
 
 39 
 
 7b. NO. OF REFS 
 
 14 
 
 la. CONTRACT OR GRANT NO. 
 
 N000 1U-67-A-0305-0007 
 
 6. PROJECT NO. 
 
 »a. ORIGINATOR'S REPORT NUMDERfS,) 
 
 flb. OTHER REPORT NOfSji (Any other number* that may ba aaetgned 
 thla import) 
 
 10 AVAILABILITY/LIMITATION NOTICES 
 
 II. SUPPLEMENTARY NOTES 
 
 12. SPONSORING MILITARY ACTIVITY 
 
 Office of Naval Research 
 219 South Dearborn Street 
 Chicago, Illinois 6060J+ 
 
 13. abstract r^g ^pg Machine (APE = Autonomous Processing Element) is presently under 
 development. This machine is a general programmable computer based on stochastic 
 processing techniques. No physical connections exist between the APEs and the rest 
 of the system. It is therefore necessary to provide either an energy source inter- 
 nal to each APE or a method of transferring energy to the APEs. This report con- 
 cerns the development of a system to transfer energy to the APEs remotely. The 
 nethod chosen involves the use of a microwave radio frequency transmitter. 
 
 Four basic pieces of equipment make up the transmitter. A ^0 watt trans- 
 mitter operating at 216 MHz connects to a frequency doubler and a frequency tripler 
 Ihe output of the multipliers is then fed into a cavity amplifier operating with 
 about 1^0 watts input at the final frequency of 1296 MHz. A circularly polarized 
 helical antenna connects the transmitter. Each APE contains one or more dipole 
 antennas. Since the strength of the signal received does not depend on the 
 orientation of the dipoles relative to the helix, the APEs may be placed randomly 
 in the receiving area. The output of the dipoles is then rectified, filtered and 
 regulated to supply the DC operating power for each APE. 
 
 The system constructed is a feasibility model. The results of the study 
 show that it can provide the estimated 100 milliwatts of DC power required by each 
 APE, with a transmitter-receiver separation of a couple feet. The use of microwaves 
 is in fact a practical method for the transfer of power in this application. 
 
 ^n form 
 
 J \J 1 JAN 64 
 
 1473 
 
 UNCLASSIFIED 
 
 Cawi.nf,, Pic 
 
 •if 1 /-'afi/\r» 
 
UNCLASSIFIED 
 
 Security Classification 
 
 KEY WORDS 
 
 Stochastic Computing 
 
 Continuous Variable 
 
 Random Pulses 
 
 Numbers Represented by Probabilities 
 
 Autonomous Processing Element 
 
 Programmable Analog Computer 
 
 APE Machine 
 
 Remote power Supply 
 
 LINK A 
 
 LINK B 
 
 ROLE 
 
 WT 
 
 LINK C 
 
 INSTRUCTIONS 
 
 1. ORIGINATING ACTIVITY: Enter the name and address 
 of the contractor, subcontractor, grantee, Department of De- 
 fense activity or other organization (corporate author) issuing 
 the report. 
 
 2a. REPORT SECURITY CLASSIFICATION: Enter the over- 
 all security classification of the report. Indicate whether 
 "Restricted Data" is included, Marking is to be in accord- 
 ance with appropriate security regulations. 
 
 26. GROUP: Automatic downgrading is specified in DoD Di- 
 rective 5200.10 and Armed Forces Industrial Manual. Enter 
 the group number. Also, when applicable, show that optional 
 markings hcve been used for Group 3 and Group 4 as author- 
 ized. 
 
 3. REPORT TITLE: Enter the complete report title in all 
 capital letters. Titles in all cases should be unclassified. 
 If a meaningful title cannot be selected without classifica- 
 tion, show titl»» classification in all capitals in parenthesis 
 immediately following the title. 
 
 4. DESCRIPTIVE NOTES: If appropriate, enter the type of 
 report, e.g., interim, progress, summary, annual, or final. 
 Give the inclusive dates when a specific reporting period is 
 covered. 
 
 5. AUTHOR(S): Enter the name(s) of author(s) as shown on 
 or in the report. Entei last name, first name, middle initial. 
 If vjlitary, show rank and branch of service. The name of 
 the principal o >lhor is an absolute minimum requirement. 
 
 6. REPORT DATf- Enter the date of the report as day, 
 month, year, or month, year. If more than one date appears 
 on the report, use date of publication. 
 
 7a. TOTAL NUMBER OF PAGES: The total page count 
 should follow normal pagination procedures, i.e., enter the 
 number of pages containing information. 
 
 76. NUMBER OF REFERENCES: Enter the total number of 
 references rited in the report. 
 
 8a. CONTRACT OR GRANT NUMBER: If appropriate, enter 
 the applicable number of the contract or grant under which 
 the report was written. 
 
 86, 8c, & 8d. PROJECT NUMBER: Enter the appropriate 
 military department identification, such as project number, 
 subproject number, system numbers, task number, etc 
 
 9a. ORIGINATOR'S REPORT NUMBER(S): Enter the offi- 
 cial report number by which the document will be identified 
 and controlled by the originating activity. This number must 
 be unique to this report. 
 
 96. OTHER REPORT NUMBER(S): If the report has been 
 assigned any other report numbers (either by the originator 
 or by the sponsor), also enter this number(s). 
 
 10. AVAILABILITY/LIMITATION NOTICES: Enter any lim- 
 itations on further dissemination of the report, other than those 
 
 imposed by security classification, using standard statements 
 such as: 
 
 (1) "Qualified requesters may obtain copies of this 
 report from DDC" 
 
 (2) "Foreign announcement and dissemination of this 
 report by DDC is not authorized." 
 
 (3) "U. S. Government agencies may obtain copies of 
 this report directly from DDC. Other qualified DDC 
 users shall request through 
 
 (4) "U. S. military agencies may obtain copies of this 
 report directly from DDC Other qualified users 
 shall request through 
 
 (5) "All distribution of this report is controlled. Qual- 
 ified DDC users shall request through 
 
 If the report has been furnished to the Office of Technical 
 Services, Department of Commerce, for sale to the public, indi- 
 cate this fact and enter the price, if known. 
 
 1L SUPPLEMENTARY NOTES: Use for additional explana- 
 tory notes. 
 
 12. SPONSORING MILITARY ACTIVITY: Enter the name of 
 the departmental project office or laboratory sponsoring (pay 
 ing for) the research and development. Include address. 
 
 13. ABSTRACT: Enter an abstract giving a brief and factual 
 summary of the document indicative of the report, even though 
 it may also appear elsewhere in the body of the technical re- 
 port. If additional space is required, a continuation sheet shall 
 be attached. 
 
 It is highly desirsble that the abstract of classified reporti 
 be unclassified. Each paragraph of the abstract shall end with 
 an indication of the military security classification of the in- 
 formation in the paragraph, represented as ( TS), (S), (C), or (V). 
 
 There is no limitation on the length of the abstract. How- 
 ever, the suggested length is from 150 to 225 words. 
 
 14. KEY WORDS: Key words are technically meaningful termi 
 or short phrases that characterize a report and may be used as 
 index entries for cataloging the report. Key words must be 
 selected so that no security classification is required. Identi- 
 fiers, such as equipment model designation, trade name, miliUr 
 project code name, geographic location, may be used as key 
 words but will be followed by an indication of technical con- 
 text. The assignment of links, rales, and weights is optional. 
 
 DD 
 
 FORM 
 
 i j a n e< 
 
 1473 (BACK) 
 
 UNCLASSIFIED 
 
 c-^,,,;*„ n« 
 
 ; t f i /- o f i r\n