Radio Field Strength Measurements at 41 Megacycles Citation Doll, Edward Bushnell (1938) Radio Field Strength Measurements at 41 Megacycles. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/9j13-5y87. https://resolver.caltech.edu/CaltechTHESIS:12022024-233356078 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Electrical Engineering) Degree Grantor: California Institute of Technology Division: Engineering and Applied Science Major Option: Electrical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Mackeown, Samuel Stuart Thesis Committee: Unknown, Unknown Defense Date: 2 January 1938 Record Number: CaltechTHESIS:12022024-233356078 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:12022024-233356078 DOI: 10.7907/9j13-5y87 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16889 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 10 Dec 2024 20:48 Last Modified: 10 Dec 2024 20:54 Thesis Files PDF - Final Version See Usage Policy. 24MB Repository Staff Only: item control page

RADIO FIELD STRENGTH MEASUREEENTS AT 41 MEGACYCLES Thesis by Edward Bushnell Doll In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1938 TABLE OF CONTENTS Page . . . .. .. . .. . . . . . .. . . . . ... . . . . . . . . . . . . . . .. . .Acknowledgment .. .. . . ..... .. . .. . .. . . . . . . . . . .. . . . . . I. Introduction . . . . . .. . . . . . . . . . .. . . ... . .. .. ... Surmnary II. Description of Permanent Equipment .......... III. Description of Field Strength Measurement Equipment 3 4 6 13 Description of Tests ....................... . 29 Field Strength Meter C. . . . . . ... ... . .... Continuous Runs ... . .. .. . . . . ... .. . .. . . . . Mt . Wilson . . . . . . .. . .. .. . .. . . . . . . . . . . . . . D. Eastern Stations A. B. v. 2 . . .. . . . . . . . . . .. . .. . B. Standard Signal _Generator ... . . . . . .. ... . c. Standard Field Generator . . . .. . . .. .. .. . . A. IV. .......................... ........ 1 Palomar Antenna Current 6 27 29 30 32 . . . . . . . .. .. .. . . .. . . . . .. Discussion of Results . . . . .... .. . .. . .. . . . .. . 34 A. Aims, Expectations, and Realizations .... 34 B. Difficulties in Operation .............. 36 C. Inadvisability of Mobile :rteasurements . . . 37 33 Antenna Calculations . . . ... . .... . .. . . . . . .. . . A. ......................• 40 Appendix A. Calculation of Self Impedance of a Straight Antenna of Arbitrary Length........ 41 Appendix B. Calculation of Horizontal Field Pattern of Pasadena Antenna ..........•.•.•• 50 Appendix C. Sample Data Sheet and Reduction to db. above 1 mv./meter ··•···••·••··•·····••O 54 VI. Pasadena Antenna Bibliography 39 I NDEX TO FIGURES No. After Page Title . 6 .................................. ~ .. 8 . ... . .. . . . . . . . ... .. 1 Modified A.v.c. Circuit 2 Calibration Curves (Plate Current vs. Signal Input ) 8 . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Calibration Curves (I) . . . . . . . . . . . . . . . . . . . . Fig. 3 Calibration Curves (II) . . . . . . . . . . . . . . . . . . . . Fig. 4 Field Strength Meter . . . . . . . . . . . . . . . . . . . . . . Fig. 5 17 Pi,ston Attenuator, coaxial coils . . . . . . . . . . . 17 Piston Attenuator, coplanar coils . . . . . . . . . . 9 Oscillator in Inner Shield, Cover Removed ••• 22 10 Inner Shield and Oscillator . . . . . . .. . . .. .. . . 22 11 Rear View of S ignal Generator, showing 3 4 5 6 7 Auxiliary Cabinet . . Shielding ....................•..•.• • • • • • • • • 23 12 Completed Standard Signal Generator . . . . . . .. 23 13 Standard Signal Generator (Circuit) ........ 24 14 Standard Field Generator . . . . . .. . . . . . . . . . . .. 28 15 Typical Morn ing and Afternoon Field Strength }!Ieasurements ....•...........••••••••••••••• 31 16 Typical Afternoon Field Strength Measurements Fig. 15 17 Profile Ivlap ..............•••••••••••••••• • • 34 18 .Antenna Mutual I mpedance Components ••••••••• 3g 1g Calculated Field Pattern of Pas a dena Antenna Array .............................••••••••• 40 1. AC KNOWLEDGMEJ\IT The author wishes to express his appreciation for the suggestions and advice of Dr. s. s. liiackeown, who directed the work described herein. In addition, the Bell Telephone Laboratories made the work possible by the loan of a large portion of the e Quipment used, while the cooperation of the Astrophysics Department enabled test results to be obtained. In particular, the fol- lowing members of the Astrophysics Department were very helpful: Captain C. s. McDowell, Mr. George Lewis, Mr. Ben Traxler, Mr. Byron Hill, and Miss Helen Brown. To :Mr. Bernard Oliver, Mr. Charles Gates, Mr. George Holmes, and Mr. Gardner Wilson goes much of the credit for building the measurement apparatus, and for recording test data. In conclusion, the author wishes to thank Dr. F. W. Maxstadt for his heartfelt interest and enthusiasm, and for his encouragement during periods of downheartedness. 2. SUMMARY A radio receiver was modified, and the necessary calibration instruments constructed, to make reliable field strength measurements at 41 megacycles. The accuracy of the equipment was sufficient to measure field intensities to within 1/2 decibel above an arbitrary reference. Measurements were made of the intensity of tre radio signals as received from :r.~t. Palomar. A compre- hensive series of readings was obtained for two months during the daylight hours. Although tests were made under a variety of weather conditions, it was found that atmospheric disturbances had no appreciable effect on signal intensity. As far as could be determined, the field strength rema ined essentially constant during this period. It was found that the average field intensity was in the neighborhood of 140 microvolts per meter. Tests were conducted at Mt. Wilson, with the conclusion that the effect of the mountains in the Pasadena-Mt. Palomar path was to introduce an attenuation of approximately 15 decibels. Some antenna calculations were mad~, and the complete mathematical solution of the self impedance of a linear antenna was obtained. 3. I. IN'TRO DUCT ION An ultra short wave, single link, radiotelephone circuit was projected early in 1936 as a means for establishing two-way direct telephone communication between the site of the observatory at Palomar Mountain, and the California Institute of Technology in Pasadena. Since the Bell Telephone Laboratories were interested in making a field experiment under topographical and atmospheric conditions such as were expected to be encountered over the 90-mile Pasadena-Palomar path, a large portion of the required radio, and telephone, apparatus was loaned by them. It was understood that this equipment would be available for such experimental work as the Institute wished to perform. A progra~ of research was initiated by the Electrical Engineering Department of the Institute, under the supervision of Dr. s . S. Mackeown. The aims were t wofold: (1) to develop a reliable and accurate method of measuring field strength at ultra high frequencies, and (2) to investigate the effect of weather conditions upon the transmission of ultra short waves, and perhaps to arrive at some conclusion as to the mechanism of their propagation in the absence of an optical path. describes the work done to date. 'l1 his thesis 4. II. DESCRIP'l1 ION OF PERMAKENT E Q,UIPMENT Modified ultra short wave radiotelephone transmitters and receivers of the 5-watt mobile type, together with some standard and special telephone equipment, were supplied by the Bell Telephone Laboratories. Only the transmitters and receivers will be described, since the remainder of the equipment was of little importance in experimental work. Transmitters Two type 21A Western Electric radio transmitters, modified for the particular application, were supplied for service at Pasadena and Palomar Mountain. The 21A transmitter is designed for operation in the frequency band from 30 to 42 megacycles. A Western Electric quartz plate is used to maintain the frequency well within 0.025 per cent in the temperature range from Oto 60 degrees Centigrade. The high frequency stability of these trans- mitters enables them to be used as fre quency standards in all test work. The 21A transmitter is rated at only 5 watts, and it was found that more power was needed to guarantee reliable communication. The Astrophysics Department of the Ins ti tu te constructed power amplifiers which increased the power output of each st a tion to 40 watts. After the 5. installation of the amplifiers no difficulty was encountered in establishing reliable commun ic a ti on. Receivers Two type 19A Western Electric r adio receivers (modified) were supplied for permanen t installation at Pasadena and Palomar Mountain. These receivers were used for regular communication, and although the Pasadena receiver was further modified for field strength measurement, it was used very little for testing . A Western Electric type 18B radio receiver was supplied for test purposes. This set was e quipped for mobile operation, and was suitable for measuremen ts in the field. The receiver was considerably modified, as s hown later, and was used in making practica lly all tests to date. The permanent radio equipment at both Mount Paloma r and Pasadena was installed in individual r a dio rooms. For communic a tion purposes the e qui pment needed no attendants, all operations being carried out by remote control. Of the several frequencies on which the stations were licensed to operate, 41.0 megacycles was chosen for permanent operation since interference from other stations was least at this fre quency. 6. DESCRIPT I ON o :E' F I ELD S'rRENG'l1 H I\;:E.ASURELJ!;"lifT EQUI PMENT III. A. Field Strength Ivieter Numerous methods are suggested in the literature for the measurement of field strength. The most common method makes use of a superheterodyne receiver with an adjustable attenuator in the intermediate freque ncy amplifier for adjusting the gain of the receiver in accurately known amounts. For high frequencies such a method appeared inadvisable because of the difficulty of building a suitable intermediate frequency attenuator. Since the Western Electric model 18B receiver was available, it was decided to modify it to serve as a field strength meter. This necessitated some means of calibration, since the receiver at best could serve only as a compa rison device. The standard signal generator, and the standard field generator, constructed for calibration purposes, will be discussed later. The most obvious method of adapting a superheterodyne receiver to the measurement of signal strength is to measure the plate current of one of the tubes controlled by the automatic volume control circuit. The automatic volume control circuit, after modification, is shown in Fig. l. When a modul ated intermediete frequency voltage is applied to the plate of the detecting diode, T1, a - - • "'Tl Ci) To Grids AVC sool't I [ •J • -v G (\I IO 0 0 ... 0 0 0 BCircuit 400 600 1200 50 g'~ R3 0 0 -_.,---+------+---- IJ .02; ~-- ~ R2 0 0 +~--1 l Modified AVC - V g t i 't~ :l _ 4,oOorr ,,ooo, ooo ~ - - - ~ 100)000 ----A IFJ Output , B+ Codan 7. direct current voltage, as well as an audio fre quency voltage, appears across its load resistor, R1 . These voltages are applied to the triode section of T1 , and the amplified voltages appear across the load resistor, R2, of this tube. When no signal is being received the diode automatic volume control plate of T1 is maintained at a negative potential with respect to the cathode, by biasing resistors in other parts of the receiver circuit. However, when a signal is received the grid becomes more negative, causing the current through R to decrease. 2 This lowers the voltage across R2 , and allows the cathode to become negative with respect to the a.v.c. diode plate. The a.v.c. diode plate then draws current, causing a voltage drop to oocur across R3 . This voltage is used to decrease the gain or the set by increasing the grid bias of the radio frequency amplifier, the modulator, and the intermediate frequency amplifiers. In this way the plate current of T1 is seen to be dependent upon the input signal intensity. It was found that the plate current of T1 was very sensitive to change in signal input over a limited region. Since it was desired to measure a relatively large range of field strength, some means had to be found to increase the sensitive range of the set. The a.v.c. circuit described is of the delayed type, and thus comes into action only 8. after the signal has reached a given value. By introducing additional fixed bias into the a.v.c. circuit the set obviously demands a stronger signal to cause a.v.c. action to take place. This additional grid bias was introduced as shown at VG. In this way the set could be made sensitive at any desired signal strength. A set of curves for dif- ferent values of VG, showing the plate current of T1 as a function of the input signal, is given in Fig. 2. Due to the lack of a suitable vaouum tube voltmeter to indicate the one volt level, the curves are plotted against the attenuation, A, in decibels below an arbitrary reference, as determined by the standard signal generator. If continuous recording of field strength over a limited range is desired, a recording milliammeter need only be inserted into the plate circuit of T . The value 1 of VG may be chosen so that quite reliable measurements may be made over a range of about fifteen decibels. In such an application the stability of the set is of course very important. It was found that the set could be depended upon to maintain its calibration over a period of 24 hours within about 1/2 db. For continuous measure- ments this accuracy appeared to be sufficient. For manual operation a more convenient method was chosen. It is to be noticed that the curves in Fig. 2 9. are very sensitive in the neighborhood of 0.6 ma. This suggested the possibility of adjusting the additional grid bias voltage to bring the plate current to 0.6 ma. In this way a reliable and sensitive indication of input signal intensity was obtained. Using a signal generator in the substitution method, measurements could be relied upon to about 1/4 db. It was found that the set was quite sensitive to small variations in plate supply voltage. For portable use the entire power for the set is supplied from a storage battery in conjunction with a dynamotor. Since the state of charge of the battery had a very pronounced effect upon the plate supply voltage, it was necessary to find some means of controlling the voltage. It was decided that this could best be done by inserting small dry batteries in series with the plate supply. For opera- tion from a.o. lines the set was provided with a rectifier unit, in which the output voltage was easily regulated by means of a potentiometer. The set was modified as outlined above for field strength measurements so that all connections could be made through a single eight-conductor cable. All of the auxiliary equipment, including dry batteries, was assembled in a cabinet which was attached as an integral part of the 10. receiver. The complete wiring diagram of the auxiliary cabinet is shown in Fig. 3. 1. The cabinet includes: A 1 ma. d.o. meter to measure the plate current, Ip, of v1 . 2. A selector switch to change the grid bias voltage in steps of l½ volts. This circuit is supplemented with a potentiometer connected across a 3-volt battery, and connected to a 2-volt meter to interpolate between the fixed l½-volt stages. The voltage of the 1½-volt steps remains constant since no current is drawn by t~e gri~ circuit. The reading of the selector switch gives the curve number, as indicated on the curves in Fig. 4 and Fig. 5, while the reading of the voltmeter, Vg, indicates the additional voltage necessary to bring the plate current to 0.6 ma. when operating upon a given curve. 3. A selector switch to introduce dry batteries in steps of three volts, to increase or decrease the plate supply voltage of the dynamotor. A series rheostat serves to adjust the voltage between steps. Figure 6 is a picture of the complete field strength meter. It is to be noted that the entire set is portable, and quite compact. When using a simple dipole antenna on a short mast, two men can easily operate the equipme~t in the field. u> • G) .,,- 300 .ft 1000 .n i 90 ,oo 01 "'o AUXILIARY ao 02 of4 of f,,o 0 CABINET 0 0 t o 0 o 0 0 0 0 ,--------++---------------'-01 [ t1_ • - - _11~~~~~~~~ .- r+~~~~~#~n 5 Field Strength Meter Fig. 6 11. In order to have all readings consistent, a standard routine of measurement was adopted. The method is outlined below. 1. The plate supply voltage was adjusted to give Ip ::: 0.85 ma. with the antenna disconnected, and no external grid bias applied. This corresponds to a plate supply voltage of 195 volts. The plate current was so sensitive to fluctuations in voltage that it was used for indicatio~ rather than the plate voltmeter. 2. The antenna was connected to the set, and the set tuned by adjusting Ip to a minimum. The grid bias voltage selector switch was then set to the point just below the position which allowed Ip to become greater than o.6 ma. The grid bias potentiometer was then adjusted to bring Ip to 0.6 ma. In general, the tuning was again checked at this point because of the greater sensitivity of the meter reading in this region. The position of the selector switch was recorded as #O, #1, #2, #3, etc. , and was called the "curve number." The reading of the 2-volt voltmeter was recorded as Vg. This completed the adjustment of the receiver. From a set of calibration curves of the type shown in Figs. 4 and 5, the signal could then be referred to some reading, A, of the attenuator on the standard signal 12. generator. This calibration could be relied upon to within 1/2 db. if the set was allowed to warm up for about 1/2 hour before starting operation. Because a calibration seemed advisable about every 24 hours, a more direct method was the introduction of an equal signal from the standard signal generator, after the set had been adjusted as outlined in step 2. The attenuator reading then gave a direct measure of the incoming signal. To avoid the use of a calibration curve, this method was adopted whenever the signal generator was available. 4. The above routine is sufficient to measure the relative values of field strengths when a fixed antenna system is used. As will be shown later, the actual field at the antenna may be measured with the aid of the standard field generator. In this way, the field strength can always be measured in db. above some arbitrary reference, such as l mv./meter, for any antenna system. ~ sample data sheet, as actually recorded, is shown in Appendix C. Also in Appendix C, it is shown how such data were converted to give the field strength in db. above l millivolt per meter, in order to facilitate comparison with other data. 13. B. Standard Signal Generator When measuring signal strength it is always necessary to have some means of introducing a known comparison voltage into the indicating device, or field strength meter. At first, it was intended to use a method due to R. c. Shaw, as described by 1 Schelleng, Burrows, and Ferrell. In this method the calibrations were made by introducing into the antenna itself a known field from a local source, to which the name "standard field generator" has been given. The standard field generator is a small portable oscillator, which is shielded, except for a radiating loop. The current in the loop, from which the field may be calculated as is shown in Appendix C, is measured by means of a thermogalvanometer. Such a system has several disadvantages. Due to the long non-optical transmission path between Pasadena and Mt. Palomar, and to the low power of the transmitter, it was expected that the fields to be measured would be quite weak. This meant that if the local field were to be of the same order of magnitude, an extremely small current would have to flow in the radiating loop. Although small radio-freQuency currents may be measured in the laboratory by means of vacuum thermocouples and sensitive 14. galvanometers, any simple metering method demands currents of nearly 100 ma. Since the loop must be placed close to the antenna (not farther away than one wave length) in order to avoid ground effects and other disturbing reflections, such large currents would give excessive fields. These excessive fields could not, then, be compared directly with the field strength to be measured. In order to calibrate the receiver it would therefore be necessary to have an accurate means of measuring the relative magnitude of the si gnals to be measured and the signal introduced into the antenna by the field generator. Ordinarily this difficulty could be overcome if the intermediate frequency attenuator type of instrument were being used. Nithout such an instrument it becomes necessary to develop a means for calibrating the set to measure an extremely wide range of signals. The field strength generator gives only a relatively narrow range of signal intensity because its attenuation range depends on the meter range, which for a square law meter cannot exceed 4:1 with accuracy. Consequently, it was decided to construct a nstandard signal generator" which could be used to introduce directly into the set a wide range of signals with known attenuation below an arbitrary reference. If an actual rr~asure of the field in volts per meter were desired, it was concluded t hat 15. it would be relatively simple to introduce a strong known si gnal into the antenna, and thus calibrate the signal generator input in volts per meter at the antenna. It was also felt, at first, that the receiver calibra tion would not remain constant, and consequently, that a substitution method of measurement would have to be used. This of course would necessitate turning off the signal to be measured every time a substitution was to be made with the field generator. A standard signal generator , on the other hand, would eliminate this difficulty since it would introduce a known signal directly into the receiver, with the antenna disconnected. A standard signal generator should possess four very desirable characteristics. 1. Frequency stability. 2. Constancy of oscillator output. 3. Nearly perfect shielding. 4. An attenuation device which permits the constant oscillator output to be attenuated accurately in small steps, or continuously, over a wide range. The re quirements of frequency stability and constant output were met quite easily. Since measurements were to be made at but one frequency, and since a standard source was always to be present in the form of the crystalcontrolled transmitters, there was no need to know the 16. frequency of the oscillator with any great accuracy. Practically any of the simple oscillator circuits maintains a given frequency if the power supply is kept constant. This was done by using dry batteries entirely, chiefly as an aid to portability and to simplify shielding. The problem of adequate shielding in signal generators is very difficult. Most of the principles ultimately used in shielding were arrived at by the out and try process. ifhen the final signal generator is discussed t he method of shielding will be described. Commercially available standard signal generators almost invariably use resistance type attenuators. In general they are quite cheap to build, and may be relied upon for the ordinary range of radio frequencies. However, for high fre quencies and high attenuations, stray capacities, and inductive effects become very important. It was felt that since the attenuator would be the ultimate standard of comparison, it would be desirable to seek a superior method. It was finally decided to use a mutual inductance "piston type" attenuator, as described by Harnett and Case 2 . In this type of attenuator the output and input coils are enclosed in a metal tube. The signal voltage delivered to the output terminals is proportional to the 17. field at the pickup coil, which is movable along the axis of the tube. Mr. Harold A. Wheeler, of the Hazeltine Corporation, has developed a mathematical treatment of the piston attenuator by using Byerly's 3 solutions of temperature problems involving circular cylinders and square cylinders. A series of exponential terms results for the solution, in which the first term is the only one of consequence. Thus the field varies exponentially with spacing, and the scale on the piston may be calibrated linearly in decibels. (Mr. Wheeler has not published his treatment of the problem, but supplied the above information upon request.) Figures 7 and 8 show two possible types of piston attenuators. The attenuation formulas as calculated by Mr. Wheeler were taken directly from the paper by -Harnett and Case. It is to be noted that the depth of penetration is to be added to the actual cylinder diameter. At 41 megacycles this depth of penetration is roughly only 5 x 10-4 inches, and is negligible for a cylinder of ordinary dimensions. In the coaxial type of attenuator there are two major sources of error. One is incidental coupling, and the other is coupling due to transverse fields arising from departures from axial symmetry. Both of these fields Depth of penetration P= 2.6 mils (f in me) ff Attenuation= 33.3 decibels per radius l27.3 db. in square cylinder) Piston attenuator, coaxial coils in circular cylinder Fig. 7 □ :E □==========□=====I c=========================::;::::;::======~ 2 It::===! ~ =====:I Iti Attenuation::: 16.0 decibels per radius ll3.64 decibels in square cylinder) Piston attenuator, coplanar coils in circular or square cylinder Fig. 8 18. are .attenuated much less rapidly than the coaxial field, and hence the errors introduced by them increase with increased attenuation, and in fact, at sufficiently high I attentuation they will ultimately become the dominating fields. The coplanar type of attenuator utilizes the transverse field, and since the transverse field is attenuated less rapidly than either the capacitive coupling or the coaxial coupling, the accuracy in this type of attenuator increases with attenuation. Since a large attenuation was desired, this factor could not be neglected. It is to be noted from Figs. ? and 8 that the attenuation in the coplanar type of attenuator is roughly only half as much per radius as for the coaxial type. This suggests immediately that the accuracy of reading the scale in decibels would be considerably greater in the case of a coplanar attenuator, than in a coaxial one of the same diameter. For these reasons, it was decided to build a signal generator using a mutual inductance attenuator with coplanar coils in a square cylinder. The square attenuator was constructed to give an attenuation of five db. per centimeter of axial displacement. The input coil of the attenuator formed the tank circuit of the oscillator. Since a measure of the input 19. current was necessary, a radio-frequency thermomilliammeter was coupled to the input coil by a single turn. The meter could not be put in series with the oscillator coil, since this prevented ample oscillator output. The entire oscil- lator, the batteries, and the input end of the attenuator were mounted inside of an aluminum shield. The first signal generator, as described in the above paragraph, proved inadequate for several reasons. In the first place the rate of attenuation was checked by the following method. Since the output of the attenuator is the product of the input and the attenuation, it follows that if the input current be increased by a certain ratio, say 3 to l (9.54 db.), an increase of attenuation of 9.54, as given by the piston reading, should restore the output to its original value, as indicated by the receiver plate current. It was found that the actual rate of attenuation differed from the predicted value by as much as 25%, and, further, that the rate of attenuation was by no means constant for different piston positions. 1. This was explained in two ways: The input coil necessarily carried a very large current in order to give an indication on the coupled meter. This would naturally cause large currents to flow in the attenuator case. Ordinarily it would be expected that this would be of little consequence since 20. just such currents would be expected to flow in order to give the desired attenuation. The attenuator tube was made from copper plates soldered together, and because of the difficulties in construction, it ap:p:,ared entirely possible that these soldered joints did not provide uniformly good electrical contact. If this were the case, an unsymmetrical distribution of currents would be set up, and could easily cause errors in attenuation. The remedy appeared to be the use of an attenuator made from seamless tubing. 2. The inner half of the attenuator was exposed to all of the fields put out by the unshielded oscillator. There was no reason to assume that the currents thus induced in the attenuator tube would be symmetrically distributed, and so it appeared that this could be a source of trouble. 3. The shielding was so inadequate that large currents flowed in the outer shield, causing strong external fields and pronounced body capacity effects. Such fields naturally preclude any possibility of making accurate measurements. From the experiences with the first signal generator it was decided to build a second signal generator, incorporating the following principles: 21. 1. An attenuator made from seamless copper tubing. This called for a circular tube, and, for the reasons already given, it was decided to use coplanar coils. The tubing chosen gave an attenuation of 12.8 db. per inch, as calculated by the equation in Fig. 8. The actual useful attenuation which could be obtained without loss of accuracy due to too close coupling was at least 80 db. (very conservative). 2. A method of measurement of the current in the attenuator input coil without such large currents as in the first signal generator. This was accomplished by using a more powerful oscillator with a separate tank circuit, to which the attenuator input coil was coupled by induction. The current in the latter circuit was measured by placing a thermocouple directly in series with it. The leads from the thermocouple to the d.c. meter were thoroughly filtered. 3. Entirely different shielding. To obtain adequate shielding was an extremely tedious process. Little reason can be offered for some of the principles involved, except to say that they ultimately gave the best results. Since the experiences with the first attenuator indicated that stray fields from the oscillator, external to the attenuator tube, were a source of error, it was decided to place the entire oscillator inside a second copper 22. shield. It was further decided to filter all power leads, both before and after leaving the inner shield. This filtering prevented large fields from being produced by the battery leads. The filter chokes were enclosed in separate shields to remove them as much as possible from the main field of the oscillator.. In addition, the tank circuit coils, and the attenuator pickup coil, were enclosed in a third shield inside of the inner shield. All of the ground connections within the inner shield were made to one point on the shield. This eliminated the principal paths for circulating currents in the shield, and thus considerably increased the shielding effect. The oscillator and the inner shield are shown in Figs . 9 and 10, both with the cover off and on. The separate tank coil shield and separate filter shields are clearly visible. The dotted circles on the cover indicate the openings through which the only external radio-frequency leads were passed. The attenuator and shielded oscillator were mounted within an outer aluminum shield. A copper chimney was constructed from the inner shield to the input end of the attenuator to give complete shielding for the radiofrequenoy leads. The thermocouple element was mounted inside this chimney, and the d.o. leads were filtered before they were led out to the panel. In the first Oscillator in Inner Shield, CoTer RemoTed Fig. 9 Inner Shield, and Oscillator 23. signal generator the radio-fre quency meter on the panel was the cause of much oi' the stray fields external to the shield. Definite body capacity effects were noticed with the hand within six inches of the meter. This effect was entirely eliminated in the second signal generator by placing the thermocouple inside the chimney. In order to reduce current paths between the inner and outer shields to a minimum, only one connection was desirable between the two shields. The most successful method of making this connection was found to be by the attenuator tube itself, in conjunction with the connecting chimney. The two condenser shafts from the oscillator were passed through the outer shield by bakelite shafts. A back view of the interior of the signal generator is shown in Fig. 11, in which the method of mounting the inner shield and the attenuator are clearly shown. Figure 12 shows a front view of the entire oscillator. Connection to the attenuator output is made by a simple bayonet type shielded connector. The front panel contains a filament switch, a rheostat for adjusting the oscillator output, the main tuning condenser control, and a meter for reading the radio-frequency input current to the attenuator. The attenuator scale is calibrated directly in db., and the numbers increase as the output voltage decreases. Rear View of Signal Generator, Showing Shielding Fig. 11 Completed Standard Signal Generator Fig. 12 24. The attenuator piston is prevented from turning by a steel ball which rides under spring compression in a "veett groove in the piston. 'rhe piston may easily be adjusted to within l/10 db. The complete wiring diagram of the signal generator is shown in Fig. 13. The diagram also shows the single ground point connections, and shows the location of shields. Power is supplied by three 45-volt "B" batteries, and one 3-volt filament battery. The frequency of the signal generator was first tested by a Lecher wire system, to be certain that the fundamental frequency was in the vicinity of 41 megacycles. At first, many spurious responses were obtained when tuning the signal generator to the receiver. These responses were due to parasitic oscillations and intermittent oscillations due to improper choice of grid leak and condenser combinations. These responses were eliminated by proper choice of grid coupling condensers. The actual oscillator circuit is conventional, and presents no points for discussion. Although it was of little importance, it was found that a change in oscillator output of ten to one produced extremely little frequency change. The frequency stability under continuous operation was found to be entirely satisfactory. w • Ci) - .,, ~ t---~ l I L __________ _ I - I I I I I I Ob . GE NERATOR SIGNAL S TANDAR D ohms - 0. 005 .,)t:f I Rl - i;ooo 4 I cc - Mein tuning condenser, 35 ;;J sect ion. I C6-5 :per Verni er tuning condens er _ _J i~~,-1~--3 : "I • I I -_7 ...2ife.!£L._7 I I I lI II IIt~ I I .___.__ I i ~~~~~◄:-+-+-~---_,,,.-+--:-_L---J r I ~ _J I _:=_=====- . .________, c4 I I A+ AB+ e- I I I I I I I l l _______ I 1 , - - - - - - - - - - - - - - - - - - , c1 _ o. oooo5J<-f' 0.1 fat' I I 62C3- 0 . 01 µf 25. The stray fields from the signal generator were tested by connecting a 4-turn coil, by means of a coaxial line, to the receiver. No appreciable pickup was obtained, except at a few cracks in the outer shield, and in the vicinity of the meter. These stray fields were only noticeable within a half inch of the case. It is to be recogn~zed that this was an extremely severe test of shielding, since the receiver was very sensitive. No body capacity effects whatsoever were found. Using the thermocouple as a standard the rate of attenuation was checked by c hanging the attenuator input curr ent over a ratio of three to one. The test data obtained are recorded below. Curve No. A A Exp. I 0=- 40 10120 Db. Theor. Db. 0 126.3 135.8 9.5 9.54 l 132.2 132.'7 9.5 9.54 2 118.4 12'7.9 9.5 9.54 3 112.1 121.6 9.5 9.54 4 103.6 113.2 9.6 9.54 5 91.9 101.4 9.5 9.54 6 '7?.8 8'7. 3 9.5 9.54 '7 65.8 ?5.3 9.5 9.54 8 54.3 63.8 9.5 9.54 9 44.6 54.1 9.5 9.54 26. The test was performed by finding the piston position (indicated by A) to give a fixed response ' for attenuator currents of 40 and 120 ma. done This was ap 10 different points, as given by different grid bias values. The difference between the two values of A corres ponded to the measured db. change. The above data show that the rate of attenuation remained constant over a range of 80 db. The discrepancy in the readings is well within the accuracy of the meter, and cannot necessarily be ascribed to the attenuator. The discrepancy of four parts in a thousand is so small that it can be neglected, and the rate of attenuation has been assumed exact as calculated. It may be concluded that the standard signal generator provides an entirely reliable means of producing a signal which is known over a range of at least 80 db. 27. C. Standard Field Generator It has already been pointed out that the direct measurement of field intensity with a standard field generator was impractical, since the output of such a generator is quite high. However, having a method of accurately measuring a wide range of signals immediately suggested a means of overcoming this difficulty. The method c·onsisted simply of using the calibrated field strength meter to measure the db. difference between the weak unknown signal, and the strong known local signal from a field generator. The absolute value of the unknown signal could then be easily calculated. Since perme.nent antenna structures were to be used most of the time, it was felt that it would be inadvisable to construct a separate field generator. Instead, the standard signal generator was constructed so that it could be modified into a field generator by simply removing the attenuator and installing a radiating loop outside of the case. Such an operation could be effected in less than half an hour. The loop installed was 18 cm. in diameter, and was used to calibrate the antenna system as shown in Appendix C. Burrows 1 has discussed the use of such field generators, and the generator was used in the manner recommended by him. The spacing between the loop and the antenna was 28. chosen at one half wave length to minimize ground effects. Figure 14 shows the completed standard field generator. Standard Field Genera.tor 29. DESCn I PTION OF 'rESTS IV. A. Paloma r Antenna Current In order that t he measurements made at Pasadena could be of any signific ance, it was necessary to know the reliability of the signa l sent out by the transmitter at :rv:t. Paloma r. Since t he antenna structure was a permanent installation it was assumed t hat if the input to the antenna stayed constant, the output stayed constant. With this assumption, t he antenna current as determined at the transmitter appeared to be an indication of the output signal. In order to test this, the Palomar an ten- na current was chan ged, and the change in field stre ng th measured at Pasadena. The predicted and measured c hanges as referred to an antenna curr ent of 1.0 ampere a re g iven in the followi ng table. I Db.above r::.1.0 Meas. The or. .20 -12.2 -14 .40 - 7.1 - 8 • 60 - 4.1 - 4.4 .80 - 1.9 - 1.9 1.00 0 0 1.22 1.5 1.7 Ant. At first glance it may appea r tha t there i s a rat her large discrepancy between the predicted and mea sured values. 30. However, it must be remembered that the type of meter used to indicate the antenna current is not very accura te over a 5 to l range. Assuming e meter of 2% accuracy (conservative}, with a 0.2 ampere reading, the current oould actually be 0.23 ampere. attenuation of 12.8 db. This would give a theoretical The error now is well within the amount by which the transmission characteristics could vary during the test. Hence, it was concluded that if the antenna current were knovm, all measured values of field strength could be referred to an ante nna curren t of 1.0 ampere as a standard, as shown in Appendix c. B. Continuous Runs Originally it was hoped to make continuous measure- ments of the field strength at Pasadena. As will be pointed out later, this was i mpossible, and it was necessary to be satisfied with measurements made during the hours from 9:00 A. M. to 5:00 P. M. Such measurements were conducted over a period of two mont hs during April and Iviay. In the beginning, measurements were made every five minutes, but after a few runs it was found tha t the intensity remained so steady that a reading every ten minutes was ample. At first it was difficult to get 31. reliable ante nna current informa tion, and for this reason, t he record~ of April and May canno t be compared within about t wo decibels, a fter reduction to a standa rd of comparison. Figure 15 shows a typical run of me asureme nts made during the morning and afternoon of two successive days. It is to be noticed that the intensi t y rerr~ ined quite constant throughout the day, and tha t there was no appreciable va ria tion from day to d a y. The same type of data were obta i ned over a period of a month, without more than 4 db. v a riation between the maximum and minimum v alues. During Apr i l it was difficult to make me a surements at any time . but during the afternoon"' .I n lig ht of later events, as indicated in t he previous pa r a graph, this proved to be of no i mportance. Two typical afternoon runs, ten days apart, are shown in Fig . 16. During the month of April no a ppreciable va ria tion in intensity was recorded. The montll:! of April and May included about every type of we a t her to be experie nced in sout hern Ca lifornia. The tests showed qui te conclusively tha t even very marked chan ges in weather cond itions had no appare nt effect upon the f ield i nte nsity. It may be conclude d from these tests, con ducted over a two-mon th peri od, that the field inte nsity remai ned essen tially co nsta nt. -,--- -;~ 'i"i _; i j t !i • ' -·,:...--:--r ·J '.:: :i j''.: ,j :. :r 32. C. Mount Wilson Since signals from Mt. Palomar do not have an optical path to Pasadena (Fig. 17), and since the path between Mt. Wilson and Mt. Palomar is free of any obstructions, a measure of the field intensity at Mt. Wilson appeared desirable. The measurements proved quite unsatisfactory because it was nearly impossible to get a consistent reading. The intensity was found to change by as much as 10 db. in a fifty-foot circle. Such variations made it difficult to obtain an accurate measure of the difference in field intensity between Mt. Wilson and Pasadena, and the only conclusion which could be reached was t hat the diff erence was in the neighborhood of 10-20 decibels. Tests were conducted along the Toll Road from the top of Mt. Wilson to an elevation of 2500 feet, but the surrounding terrain, and the presence of a telephone line which hugged the road for practically the entire passable distance, had so much effect upon the measurements that no useable data were obtained. 33. D. Eastern St a tions During the winter months signals were heard from the east co ast. The signals were quite weak and were of such a nature as to defy accurate measurement. Some ge neral conclusions may be recorded about this long distance reception: 1. S ignals only received during months of October t hrough February, and then only periodically. 2. Signals received only during hours of from 10:00 A. M. to 3:00 F. M., with best reception shortly before noon. 3. S ignals characterized by spasmodic fading. In particular, a station would be a udible quite consisten tly over a period of one or t wo hours wit h only temporary fading interruptions, and then would suddenly fade completely, and could not be received again until the next day. Such interruptions occurred at no set time. The above information is offered only to record the fact that signals were heard over a distance of 2500-3000 miles on 4 1 megacycles. 34. V. A. DISCUSSION OF BESULTS Aims, Expectations, and Realizations A good many years ago it was felt that any wave length in the ultra short range (below 10 meters) could not be received unless an optical path was present. During the last six years much experimental work has been conducted on radio signals in this region, and it has been found that the signals were quite often received when an optical path was not present. A partial list of the publications on the subject appears in the bibliography. Burrows, Schelleng and Ferrell l have shown that the effect of a variable dielectric constant of the atmosphere due to pressure, temperature, and humidity gradients would be such as to bend an electromagnetic wave toward the earth. 'rhe path after this refraction would be the equivalent of the optical path on a sphere 4/3 the diameter of the earth. Figure 17 shows a profile map of the path between :Mr. Palomar and Pasadena. The solid line represents the straight line path, while the dotted line represents the path predicted by the above theory. It is to be noted that in neither case is an unobstructed path presented. It was to be expected, then, that some further information as to the mechanism for the propagation of the signals between these two points could be obtained. ~ - • Ci) - .,, 0 PASADENA /000 .2..000 3000 4000 68 1--- I :Z..5 ~ 30 45 MILES 40 • 50 tp, .!55 I I ' 60 65 'TO I I 75 ,, I I I I I - --+---+------1._ ' PALOMA+2. OB5El2VATOf2Y I _}- ~~~1\17 :\ PASADENA TO PQOFILE· MAP ' ' I 1·I 90 .5 I --r ' PALOMAR !I / 'W ' I ~-L.. __ -t ------ : I 35 . Although the reception of the e astern sta tions indicated that 41 megacycle waves are affected by the ionized layers in the upper atmosphere, it is undoubtedly true that such effects could occur only for angles near grazing incidence, and for long dist ance transmission . It appeared utterly unfeasible that the Heaviside layer could be of any importance to the transmission of ultra short waves over as short a path as 90 miles. It was expected that the transmission could be accounted for by conditions in the lower atmosphere. might include: Such effects 1) effects similar to that predicted by Burrows, but of greater magnitude; 2) effects due to refractions and reflections from clouds constituting a sharp change in water vapor content; 3) effects due to refractions and reflections from abrupt changes in temperature, such as often occur at boundaries between layers of different temperature. The measurements conducted were over what may be considered a representative variety of weather conditions. The intensity stayed relatively constant, regardless of the weather. Tests conducted during rainy weather gave no different results than tests conducted during clear weather. Tests conducted when a dense, cold fog covered all of southern California were no different than tests during a dry, hot day. In short, it may be said that 36. so fe.r as the observations showed, no effects due to weather conditions were present. lilly efforts to measure the elevation of, or to detect the actual presence of, a marked boundary between the region of optical transmission and non-optical transmission,. or a ttshadow effect," were nullified by the effects of the surrounding terrain on the slopes of Mt. Wilson. It could only be concluded that a distinct difference did exist between the two types of transmission, and that the effect of the mountains in the path was to lower the signal strength by about 15 decibels. B. Difficulties in Operation Originally it was planned to make continuous measurements of the field intensity as measured at Pasadena. In this way some idea of the difference in transmission between night and day could be obtained. Such measurements were impossible for two reasons: 1) the Federial Communication Regulation requires that announcements be made every fifteen minutes while a station is in operation, and 2) a licensed operator must be present whenever a station is in operation. It would have been possible to build an automatic announcing system, but the second difficulty was 37. insurmountable. Several licensed operators were employed in various capacities by the Astrophysics Department at lil t. Palomar, but it was out of the question to expect one to be on hand 24 hours a day. It was quite difficult enough to have an operator present so that representative readings could be made during working hours. The cooperation of the Astrophysics Departr.ient was satisfactory in every respect, but it had to be remembered that any tests were a distinct inconvenience to them. C. Inadvisability of Mobile Measurements It was at first planned to make some measurements in the field to determine the relative field intensities at different points. Actual field tests showed that without extremely elaborate and time-consuming preparation consistent results could not be obtained because of the effect of the surrounding terrain. Such inconsistencies could probably have been overcome by the use of very high portable antenna masts. However, much work of this nature has already been done (references in bibliography) and there seemed little reason to duplicate it. Originally it was planned to test some of the antenna calculations, such as those in Appendix B, and perhaps to carry out some intensive antenna investigations. 38. The inconsistency and difficulty of making measurements in the field caused this project to be abandoned. 39 VI. ANTENNA CALCULATIONS In order to correlate data expected to be taken in the field it was felt that the directional characteristics of the antenna systems used by the Palomar and Pasadena transmitters should be determined. These calculations also seemed of value in that they would .allow a direct calculation of the expected fields due to propagation in free space. For these calculations, a method developed by G. H. Brown 4 was used, treating antenna elements as ordinary circuit elements with mutual impedances due to radiation coupling. The values of the mutual i mpedances are obtained by moving the surfaces over which the Poynting vector is ordinarily integrated directly up to the surface of the conductor. This method yields the same radiation resistances as t he classical method of the sphere at infinity. In addition to t he power component, a reactive power component is obtained upon integration. The latter component may be considered as being due to a radiation reactance which is analogous to the radiation resistance. Brown has obtained the values of these mutual reactances and resistances for various antenna spacings. The curves in Fig. 18 were obtained from da ta published in the paper referred to above. ·~· • ' ii :1 ] TTuI ;I _ - s ,~ ' ·:-~j-:--; ' • rrri ' ·· ' l 7 ·• 1·•1 .:.,I' I -- ---+ -; :tIT::],I I: ·_}L: l;· J1- ,_t~ I! l- i: :I ;;'· I Li 1 , --a •.J• ~ ',. _;_, L: :l~~,ii.:.:J : -ii . ' : tit= rt:-: 1 r ., .. ·.:.:.._ , J\fii4~, I,.. ' '...: r:;-':T ' I' 1 : -: : l : : i , . : ·7_ :_._:-u. f- .~'' ,- r , : ~- r ~·t i j_ ; ~: -- - w Fl f' .,, .i ! ;,, -+- - -+i-i: ' ' it :, rj_L 1,~ • I I 1: '. 1· I • 1i I ~- r.~:~ ,-• • · '.....,..! t'. ::' t0~L ·t ' ti ' ',_~; , · ... 1•·•1· - ~- --- - -~ t ~;: -i - -:• J , ~-r ' l ~ :; p: -" ..:...:..:-=t:::_;~- SR ~2 ~--: t ,t:.!::.,; ~i : { I:~ jcrtf;j~' • I •· : t -~ 1 r · j • ~ •rrill ~'. · ,•··\. , .. -:-'.'-T:;-: . 1· -~-+;;f:h~-·J~f-' \ I .\ \ . ·: : :·:-,-: 1.. ;.:t. :~ ~ •• , · · 1· I I -1·'' ' · 1·· , , • ·-- · :tTPI~;- {; I ,I : • --- i},,i! ,~,- j, . ~;: 7 - •---JI _ . I t I I ~ ~ .~ '~ . _I . .-. J. • I i-. ·•_1·..1r•73 ' T r ,' 1 " l ;I I _TJ . ,l '_I _ _l'i( I J . f _:iJ _, ... rr . __ _F_. · - 1~--- , T• :T ., I,: l :.: 1\ •• 1 ·, l' ,:·Tl · Tj ",'T 'r+ j ri '-' l ·±t: j '• l •ji:'-l ' l '' ,' 1-il' ! ·" ' [-' · .. l ~I · .. 1 .1 :-:,1 ,i .;t ± : • :~J ~ -~l ~~-... • • -• - · · ·· i - ,- '. • :.'. 1 ,. •• • I 1 ~--- • ._ _ _ , . " • :: ;; • H- T~' . _ 1 . • _. _ , _ ,. . _, _, : " ; ; :::;, ct' ' " ' -1 '", c . . , . . ... . . _ _ _, . . __ . _ . . , ...-.. . ~-+- 4 ..... , ~ ........ . . _ . j . - 1 - -- -• -• . ; ;1.• . ..... 1 . _ , .~ : - l : · · . -- •' , , H , h " • i •. - - - - --- ; • ;') ::: • j _;t_-C ' p.1..1 - I . . ~ j T ' - . 1 ~;::._: :-·-· 1,:,; !-,· ! ·: ;+it •:l-r~:: : ; "+± .tr:: I .. ; r ·~ 1 -+ +~11+-2~.,, '- ~,----.... 1:'"'"i: +-l:_: __..... !· ' ~ , : ;-, ·--... . ,- ----~1:_1.. ,.. . . ...... i .. j, -- ''r-- ' ·- -- • 7 •! · 1·'--+-I ' ·- -- -·-· _, t· .J ,_ ,'- ! -,-- - ,. _, __ _ , __ '.· .. " ,,, _ FfLt'lPH :? p ~t;J~!:)1t 'L itEJk'.£~ 1~~ :!;•- (~ ,Ifil]fu~.:q_ .c__,+1:JJ + ~- •+ - ·· " ,,lt, .L j ,--,- . ~, ,cc . · ~ ~---- ~I- • -7~ ::,.·,t -:. l i---r ev-+-~~ i, , : < --rt T rl -..! - !·• T: + ;+I-... :.: -- 1-·: . -:-:--._-,-...,-.---.,....._,_____. -~-:-,-- ·:- 1--:-- --'. · - -T I I : ' ' I' I 1 • ··- 1:- :---,- -- ~-- · :· : ·:--. 'i ' ' ' : '. - : - l. I ' .I : ' 40. The mathematical treatment of the self impedance of a single radiating antenna of arbitrary length was carried out in its entirety (Appendix A). After the solution was obtained it was founc that the entire ,7 analysis had been published in 1933. . Hmvever, the work has been presented here because it was an independent investigation. A. Pasadena Antenna The Pasadena antenna consists of two excited half wave radiators driven in phase with equal currents, and so spaced as to eliminate interference from the High Voltage Laboratory. The excited radiators are supplemented by two reflectors placed one-quarter wave length behind them. The complete analysis of the directional characteristics of this antenna system is carried out in Appendix B. The resultant horizontal distribution pattern is shown in Fig. 19, in which unit field is that produced from a single half-wave radiator delivering the same power as the antenna array. This antenna array gives a signal gain of roughly two to one, or 6 db. in the direction of Palomar. This gain, coupled with the elimination of interference, gave much better . reception at Pasadena than an ordinary half-wave antenna. I/ . · \ ,( . " "-·/ . ·,, • .• I~ / ' I 7 J,'j() . 210 150 41. A.PPEl-W IX A The ~alculation of the Self Impedance of a Straight Antenna. of Arbitraey Length The field from the antenna is calculated by the method introduced l 4) b7 Brown. The ourrent in the antenna. may be represented by the expression p where Io is the amplitude of the current at the center of the antenna, and \ • \ 14 (2) I I, • I The antenna current may also be written in the fo:rm I: Io E. where J'wt- J C4S A r, +[) ~ Je,r:: fl . (3) (4) The retarded Teator potential will haTe a component in the z direction only, which, for the upper half of the antenna, is giTen by the expression Jo - [ [i]t-¾ ,1. o A 'I-, Upon substituti?Jg the retarded Talue ofl from l3) the Teator potential for the upper half of the antenna becomes {5) 42. Ma.king the following substitutions "- =- Je (-"Lr3 -~) (7) V=~ {A-J+-;:) ttc. (8) Letting¢ represent the azimuth angle as measured in the plane normal to the antenna, and since ii:. VAA, the component of the magnetic fieJ.d normal to the paper, I-IJI, is ~¢ -:: - 'J Ar~. a;;,:, where the subscript indicates that the upper half only of the antenna is being considered. Equation l8) may be differentiated with respect to x by use of the The Tector potential for the lower half of the antenna may be obtained in a similar manner, and after ta.king its curl as indicated aboTe the 43. From lviaxwell' s equations for free spaoe the following expression is obtained for the electric field I Ji° cat:= v'l'/-1 • -, Ordinarily this equation cannot be solTed for E in terms of H but sinoe sinusoidal ourrents onl1 are inTo1Ted, and since oomplex notation is being used, the Talue of E becomes In the system of coordinates chosen the Tertical component of the electric field is then lll) .By- use of {7) and {9) the expression for ;t" /Ip/takes the simple form ..,,j//J-::.J·Iof.j1.u(lt-r -jlvta -).kA'j _j~A/. ,- r . L. E + ~ -2Ccr.S8E S--..-6 ll2) • After simplification the Talue of~ as determined by equation {11) is l 13 l If J is chosen as the r.m.s. value of the ourrent at the oenter 0 of the antenna, and if the current is chosen as the referenoe Teator in the ordinary electrical engineering use of complex notation, the r.m.s. value of the field is giTen by the expression 44. (14) where the current is measured in amperes and the field is in volts per cm. (4) Brown's paper suggested the possibility of obtaining a mathematical solution for the radiation resistance and reaotance of an (5) antenna which is not a half waYe length long. Pistolkors first showed that the power radiated from an antenna could be obtained from -P = where lr 1-.fo the relation - E I= ff. ,,11 (15) --lo Instantaneous field intensity at the conductor Instantaneous current in the conductor. When sinusoidal quantities a.re involTed /: and I in general will not be in phase, and, in analogy to ordinary circuits, a so-called reactiTe power will be obtained. When complex notation is adopted, and if the r.m.s. values of the current and voltage are used with the current as the reference vector, equation {15) will become +Po -[ zi"·il = I:{R .....j x•• J {16) -lo where ] 0 is the r.m.s. value of the current at the center of the antenna, Roa corresponds to the radiation resistance of the antenna, and,.\' oo may be called the radiation reactance of the antenna. Roo and ~oo are referred to the current at the center of the antenna. Both 45. It is only necessary to evaluate the integral for the upper half of the antenna, p since from symmetry both halves give the 1::. same contribution. The r.m.s. value of the current may be A, obtained from (1), and the value of the /f.~,-= field is obtained from (14). If the radius of the antenna conductor is f , Ir1o-, J~.,. f' ~ - (17) and the vertical component of the field at the antenna becomes (18) If it is assumed that the radius of the conductor is small all of the integrands in the expression for Roo remain finite as r..., O, and after considerable manipulation equation (19) becomes • R.. -= 'J.8 'J.8 '1r,- ~/"¥1-1-- r{(~ ~ 1?,.,.tt<11-:i.J,,l~ t""..f"", • :I (J'f i 0 - (/J (21) </I +-'1 @z[ " ~~fld1 At this point the sine and cosine integral :t'unctions are introduced (6) as defined by Jahnke and Emde Definitions; • cP ~(a.)= - [ a. (A-5 <f) (/) di) (22) (23) By expansion of (22) into a power series the value of G:{t)for small values of f. may be obtained in the form (24) where Y :: Euler's constant• To express R.oo in terms of these functions it is first necessary to examine the integra1 47. Let then or Using (24) this becomes /11: i [1.r; J. 9 .,. Y - c.:. ( :u.9, By similar manipulations the Talue of (21) becomes Ko• ~3o[r1-~•J[1? 19 -t V-G: (4s!}~¥f>i f-18-). ~ (,),8JJ 1251 J. r-4 (:Pf-lo3 H .,. II - c.; (2.s)J This expression giTes the radiation resistance referred to the current at the center of the antenna. If the antenna is longer than a half waTe length it is customary- to refer the radiation resistance to a current node. The Talue of Roo then becomes R.~ =Jo X:...1 8(ft-@';Jf- '? -ttH JI- C.:.('IIJ!l +-2f1/ S.l-18)-2S.:(.,,11[j 1, :J.8 +-- Jt - C..: ( ~ eJJj +--1@'-f for (26) 48. In case the antenna is exactly a half waTe length long the radiation resistance takes the familiar Ta.lue In the detemination of ~00 1t is found that the assumption f~0 causes some of the integrals to diTerge. After manipulation all of these infinite integrals may be reduced to the form .Q. - I -=-[ c) :.n .A "" CAn .A ~ .,_-+ f 1 cl."- vu . .+f2 If it is assumed that f(( l. 0 , • this integral reduces to ./o I; where 4 _,,c.c., +[ eo-:s ~ ~ J..,,_ r '<• ~ This expression may then be reduced to by the use of (22), (23), and {24). After repeated operations the expression for Xoo becomes x•• =-3o('-C3 [ 1?t . - ~-+ 2 c.; (:i.s)- C.: (-feJJ +- ( (t--1)[ £.i.(.fo)-z s..: (>sl1(@2+',) .k (.u>JJ. 1211 49. For a half waTe antenna the reactance becomes i1his result shows that a half waTe antenna., excited at the center, is not a resonant circuit, but that it has an inductiTe reactance of 42.5 ohms compared to a resistance of 73.2 ohms. It was intended to oarry the investigation much further, and to find the antenna length to giTe zero reactance, the nature of the Ta~iation of radiation resistance and reactanoe with length, etc. HoweTer, after completing the aboTe work a similar treatment was found (7) published in Gemnan . . • This made it of no particular Talue to continue the inTestigation. independent inTestigation. The aboTe work is presented since it was an 50. APPENDIX B Calculation of the Horizontal Field Pattern of the Pasadena Antenna 'I'he antenna may be considered as made up of 4 halfwave radiators, spaced as shovm. Elements 1 and are excited by transmission lines, while elements 4 2 3 and act as parasitic reflectors. If we now consider the four antenna elements as four simple circuit elements with self i mpedances and mutual impedances, the following equations are obtained: V1 ::= I 1 Z11 + I2Z21 + I3Z31 + I4Z41 ( 1) =- 11 zl2 -t- I2Z22 -t- I3Z32 -i- I 4 Z42 ( 2) V3 -:. I1Zl3 + I2Z23 + I3Z33 -+ I4Z43 ( 3) V :::. IZ + IZ -t IZ + IZ 4 l 14 2 24 3 34 4 44 ( 4) V2 where the currents are measured at the center of the radiators, ~ the voltages are the exciting voltages applied there, and complex notation is used. symmetry it may be assumed that: Z 11 :::. z = z 22 33 ::: Z = Z 44 00 Z13== Z31 = z24 = Z42 = z2 z14 = Z41 =z32 = Z2 3 = Z3 Z : Z 12 : Z 21 34 = Z : z 43 l From 51 Since the t wo excited antennas are excited from identical feeders it may be assumed that V : V =V 1 2 I ~ I =I 1 2 Since elements 3 and 4 are parasitic reflectors there is no external excitation applied to them, and V -::. V = 0 3 4 Upon solving the above system of equations the results are: I [z1(Z2 -t Z3) - Zo(Z3 + Z2)] I4 = I3 :. - - - - - ~ - - - - - - - - - ( 6) - zl A half wave antenna has a reactance of 42.5 hhms (see App endix A). In practice most antennas are made about ~5% of the theoretical length in order to make this reactance zero. Such a choice gives for the self impedance of each radiator, since the shortening has no appreciable effect upon the radiation resistance. The shortening has little effect upon the mutual impedance values, and Fig . 18 may be used for their determination. 52. Substitution into equations 5 and 6 of the numerical values for these impedances yields = I{?2.? + j 19.3) I 3 ::: I 4 = -I(.51? - j 0.29) V Since the total power radiated is twice that radiated by each excited radiator, I = ( 7) 2 ( ?2.?) and (8) 2( 72. 7) The above expressions give the currents at the center of the balf wave radiator.s of the array. Knowing this, it is a relatively simple matter to calculate the field pattern. It is well known that the vertical field intensity from a half wave antenna in free space in the horizontal plane through the center of the antenna is given by the expression* E -60r I c-jlrr * This expression may be obtained from equa tion (14) in Appendix A by placing 1 _ ~ c) - 4 ' an d IZ.z. = It,:: r • 53. where k = T2 1'1" i\ = wave length in cm. I - current at center of antenna (r.m.s. value) E - field in volts per cm. (r.m.s. value) r =- distance from antenna in cm. The field at a great distance, r, may then be calculated as a function of the azimuth angle, 9, by the use of this expression in conjunction with the values of the currents from equations 7 and 8. Since the radiation resistance of a single half wave antenna is 73.2 ohms, the r.m.s. value of the field from such an antenna delivering PT watts is E - o - 60 ~ r0 1 ~ • The use of this expression and the above work yields a relation between the ratio of the field intensity of the antenna array to that of a simple dipole delivering the same power. E E0 The result is: r;;- j73.2 --f w 72. 7 Where: A B [1 - 0.517 cos {kd 2 sin. 8 ) + 0.29 sin (kd 2 sin9 )] = [o.517 sin (kd 2 sin 8 ) + 0.29 cos (kd 2 sin 8 )] This expression is shown as a function of the azimuth angle in Fig. 19. 54 . APPE1'1" DIX C Sample Data S heet and Reduction to db. above 1 mv./meter The sample data chosen was recorded from 9: 30 A . i\'l . to 10:30 A. M. on May 12. The resultant curve a:9pears in Fig. 15. Using the substitution method the field strength meter was tuned to the incoming signal and the grid bias adjusted to give Ip =- 0.6 ma. The antenna was then disconnected and the signal introduced from the st anda rd signal generator. Without changing the field strength meter the attenuator was adjusted to give Ip =- 0.6, while the signal generator current was held constant at 60 ma. The reading, A, of the attenuator was recorded. Data: May 12 Time A 9:30 109.6 9:40 108.8 The antenna current as 9:50 108.8 announced from Palomar was 10:00 108.'7 constant all morning at 10:10 108.9 1.12 ampere. 10:20 109.1 10:30 108.8 Ultimately the standard field generator was pl a ced a half wave length in front of t he center of the antenna, 55. and the set was found to respond at a grid bias voltage cor!L'esponding to A == 66.6, using the calibration curve s in Fig. 5. The thermogalvanometer in the radiating loop was held at the full scale reading of 115 ma. The field from the radiating loop was calculated 1 by the following equation 2 E ::= 240 1T 3 A I ( 1. 05 ) A where E -= r.m.s. ·value of field in volts per meter A -=- area of loop in square meters I = current in loop in amperes .i'~ wave length in meters . Upon substitution of the numerical values the field strength from the loop is obtained as, E = 18.5 millivolts per meter. But 18.5 mv./meter is 20 log 10 ( 1 ~•5 ) : 25.4 db. above 1 mv./meter, and by subtracting (66.6 + 25.4) -= 92.0 from the values of A the field strength is obtained in db. below one mv. per meter. If the signal strength is desired in db. above 1 mv. per meter it may be obtained from the expression, - (A - 92) Actually, however, the antenna current was 1.12 amperes or, ZO LJ 0 ( 10 1 •t 2 ): 1 db. above the standard of 56. one ampere. Therefore the above data may be referred to the signal with one ampere of antenna current by subtracting 1 db. from the measured values. The final value of the field intensity to be used in compari son with other data may be obtained from the expression - (A - 91) The data for plotting then become: Time Field Str. Db.above 1 mv./m. 1.0 amp.std. 9:30 -18.6 9:40 -17.8 9:50 -17.8 See Fig. 5 for the 10:00 -17.7 curve corresponding to 10:10 -1 '7. 9 these data. 10:20 -18.1 10:30 -17.8 The average field is -17.9 db. above 1 mv./meter, which corresponds to O.139 mv. /meter, or :approximately 140 mic:r_o yolts per meter. BIBLIOGRAPHY l. J.C. Sohelleng, ,C.R. Burrows, and E.B. b'errell, "Ultra.-Short-WaTe Propagation", Proo. I.R.E., Tol. 2i, P•P• 427-463; hla.rch (1933). 2. D.E. Harnett and N.P. ua.se, "The Design and Testing of Multirange ReoeiTers", Proo. I.R.E., Tol. 23, PP• 578-593, June (1935). 3. Byerly, Textbooks on .Bessel fUnctions, Fourier series, etc. 4. G.H. Brown, "Directional Antenna.s 11 , Proc. I.R.E., Tol. 25, PP• 78145; January (1937). A.A. Pistolkors, "The Radiation Resistance of Beam Antennas", Proc. I.R.E., Tol. 17, PP• 562-628; W iareh (1929). 5. 6. E. Jahnke and F. Emde, "Funktionentafeln mit Formeln und Kunen". 7. J. Labus, "Reehnerisohe Ermittl,mg der Impedaza Ton Antennen", "Hochfrequenzteohnik und Elektroakustikn, pg. 17, 1933. 8. R.C. Colwell and A.W. Friend, "Ultra-High-~'requenoy WaTe Propagation Over Plane Earth and Fresh Water 0 , Proo. I.R.E., Tol. 25, PP• 32-37; January (1937). 9. L.F. Jones, "A Stu~ of the Propagation of WaTelengths Between Three and Eight Meters", Proo. r.R.E., Tol. 21, PP• 349-386; March (1933) 10. B. TreTor and P.s. Carter, "Notes on Propagation of WaTes .Below 'l'en lleters in Le:ngth", Proo. I.R.E., TOl. 21, PP• 387-426; March ( 1933). 11. C.R. Englund, A.B. urawford, and w.11. Mumford, "Some llesults of a '8tucq of Uitra-Short-WaTe Transmission Phenomena.", Proo. I.R.E., Tol. 21, PP• 464-492; March (1933). 12. H.O. Peterson and D.R. Goddard, ''Field Strength Obsenations of Transatlantic Signals, 40 to 45 Megacycles", Proc. I.R.E., Tol. 25, PP• 1291-1299; October (1937). 13. P. von Handel and w. Pfister, "Uitra.-Short-Wave Propagation Along the CurYed Earth's Surfac&", Proc. I.R.E., Tol. 25, PP• 346-363; March (1937). ' 14. C.R. Burrows, "The Surface WaTe in Radio Propagation OTer Plane Earth", Proo • .1..R.E., Tol. 25, PP• 219-229; February- (1937). 15. R.H. BeTerage, "Some l'l"otes on Ultra High .!frequency .Propagation", R.C.A. ReTiew, TOl. 1, PP• 76-87, January {1937). 16. C.R. Burrows, A. Decino, and L.E. Hunt, "Ultra.-Short-Wa.Te Propagation oTer Land", Proc. I.R.E., Tol. 23, pp. 1507-1535; December ll935). 17. H.G. Houghton, "Transmission of Visible Light Througl::1 Fog" and "Theoretical InTestigation of the '.i'ransmission of Visible Light through Fog", Publications from the Massachusetts Institute of Technology, .Serial ~g. 74; September, 1931. 18. H.G. Houghton, "The Size and Size Distribution of ~og .l:'artioles", Publications from the Massachusetts Institute of Technology, Serial No. 82; January, 1933.