■MCtfJliN^ CORNELL UNIVERSITY LIBRARY c« GIFT OF Prof. Guy E. Grantham Cornell University Library 0840 Modern theory and practice in radio comm 3 1924 031 257 672 olin.anx The original of tiiis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924031257672 Modern Theory and Practice in Radio Communication Modern Theory and Practice in Radio Communication A Text Book Prepared for the Use of Midshipmen at the United States Naval Academy By GOEDON D. EOBINSON M. E., M. S., S. M. Associate Member Institute of Radio Engineers Assistant Professor, Department of Electrical Engineering and Physics U. S. Naval Academy AND PAUL L. HOLLAND Lieutenant Oommandek TJ. S. Navy Instructor, Department of Electrical Engineering and Physics U. 8. Naval Academy ANNAPOLIS, MD. THE UNITED STATES NAVAL INSTITUTE 1920 MODERN THEORY AND PRACTICE IN RADIO COMMUNICATION Price $2.25, postpaid COPYEIGHT. 1920, BY J. W. CONROY Trustee for U. S. Naval Institute Annapolis, Md. BALTUIOBE, US., U. 8. A. PEEFACE. A need having arisen for the instruction of Midshipmen in Radio, beyond that previously provided, the authors have undertaken the production of a text suitable for that purpose. Our endeavor has been to provide a book containing all the information necessary for the student who does not intend to specialize on Eadio, and at the same time to pro- duce a book which will serve as a basis for further study for the student who does wish to specialize. The book is not intended, however, for an instruction manual for a radio operator who has little interest in the how and why of his apparatus. Eequirements of time and space have made it necessary to take for granted various statements which would be derived mathematically, or explained in detail, in a text dealing purely with fundamentals. These omissions are of such a nature that they may profitably be studied only by a student specializing on radio subjects. To understand the work as presented, the student should have a knowledge of elementary electricity and physics. Credit is due Lieutenant Commander P. L. Holland for Chapters I, II, III, IV, VI and XII. The remaining chapters are to be credited to G. D. Eobinson. Depaetmknt of Electbicai. Engineebinq AND Physics. United States Naval Acadbmt, July, 1919. LIST OP ABBEEVIATIOXS. — Capacitance. E — Voltage or Potential. E. M. F. — Electromotive force. 1 — Current. i^Instantaneous current. L — Inductance. M — Mutual inductance. R — Resistance. X — Reactance. Z — Impedance. / — Frequency in cycles per second. a — -2 TT times frequency, mf. — microfarad, mmt. — micromicrofarad. mh. — millihenry. Ij.h — microhenry. Q— Quantity of electricity. CONTENTS. CHAPTEB PAGE I. Electron Theory. Electric and Magnetic Strains 9 II. Capacitance and Inductance. Energy of Charges. Condensers 16 III. Oscillating Circuits. Damped Oscillations. Buzzers. Wave Meters. Detectors. Resonance. Frequency. Decrement. Wave Propagation. Wave Length .... 34 IV. Coupled Circuits. Transfer of Energy. Mechanical Analogues 50 V. Vacuum Tubes. Theory of Operation. Use as Ampli- fier and Generator 63 VI. Spark Transmitters. Gaps. Tuning of Circuits 81 VII. Continuous Wave Transmitters. Arcs. H. F. Alter- nators. Radio Frequency Spark. Application of Three-Element Vacuum Tube as Generator 98 VIII. Detectors 117 IX. Receiving Circuits 131 X. Radio Telephone 164 XI. Radio Compass 184 XII. Measurements. Formulae 199 Appendices. Standardization Rules. Codes. Extracts from Laws Governing Radio Communication 219 Index 247 MODERN THEOEY AND PEACTICE IN RADIO COMMUNICATION. CHAPTEE I. ELECTRON THEORY. ELECTRIC AND MAGNETIC STRAINS. The phenomena of sound waves are familiar to the student of elementary physics. The production of disturbances in a medium as tangible as air, and the effects of the resulting waves, as .sound, upon the human ear seem quite simple. Likewise light and heat waves, by whatever source pro- duced, can be received or detected by our physical senses. These waves are the result of disturbances in a medium called ether, which pervades all space, and are somewhat analogous to water waves. In this medium can be pro- duced waves of various lengths which manifest themselves in vastly different ways. Heat waves vary in length from 1/10000 to 1/100000 cm.; light waves from SBxlO-" to 80 X 10~° cm. Shorter than any waves which we can detect as light are ultra-violet rays. These can be detected by their effect on chemicals or photographic plates. Still shorter are X-rays. Waves in this medium, varying in length from approxi- mately 100 meters to 20,000 meters are the means of con- veying the energy in radio communication. The problem of the radio engineer is the efHcient produc- tion and reception of these extremely long ether waves. In order that the student may more easily understand certain phenomena connected with radio, a slight digres- sion is deemed necessary. 10 Modern Theory and Practice in Electron Theory. Modern science teaches that the atom, which is the smallest subdivision of matter with which the chemist has to deal, is in itself a very complex structure. Each atom is composed of a great many smaller particles, called elec- trons, which are in rapid orbital motion around a central portion or nucleus. These electrons have been proven to be charges of negative electricity, and the nucleus is con- ceived to be a positive charge of electricity. The experi- ments from which the electron theory has been evolved do more than prove the existence of electrons — the weight and size of each electron and the amount of negative charge carried by each has been computed to be as follows : mass, 8.8x10-2^ gram; radius, 1x10"" cm.; charge, 1.6x10"" coulomb. Following this conception of matter, the difference .be- tween atoms of different materials is a difference in num- ber and arrangement and motion of electrons. In some atoms the arrangement seems to be stable, in others the motion of electrons is irregular, and some electrons are constantly leaving one atom and attaching themselves to some other atom. If the negative charges, or electrons, are held by the positive nucleus in a more or less stable arrangement the atom is neutral. If there are electrons in excess, or free electrons as they are called, attached to an atom, the atom is said to be "ftegatively charged. II by any means the number of electrons is reduced below the number necessary for neutral state, the atom is positively charged. What is true of atoms is true of large bodies which are atoms taken collectively. Under this theory we readily understand the charging by friction of bodies such as amber, rubber, fur, etc. Any Eadio Communication. 11 body, when rubbed, has its molecular or atomic structure disturbed. If under this disturbance electrons detach themselves from atoms of one body and attach themselves to atoms of another body, the second body is negatively charged, and the first, with reduced number of electrons, is positively charged. Thus, the potentials of the two bodies are changed in the same degree, and for every posi- tive charge induced there is an equal negative charge in- duced on another body. This condition is unstable — on one body are free electrons, on another there is a deficiency of electrons. If these two bodies are connected by a material through which the electrons can pass the free, or excess, electrons move from the negatively charged to the positively charged body, resulting in what is familiarly known to us as an electric current. This tendency of electrons to move from the negatively charged to the poisitively charged body is manifested in what we are accustomed to call attraction between unlike charges. A material which can serve as a path for the electrons is called a conductor. The ease with which this transfer of electrons takes place is a measure of the con- ductance of the material. A good conductor offers very little resistance. An insulator, or dielectric, permits prac- tically no passage of electrons. Eecent research has led to the theory that this transfer of electrons through a con- ductor is not a continuous passage of numbers of electrons, like a current of air, but is a spontaneous movement from one atom to another, much like the effect noticed when the end ball in a bowling alley ball rack is struck a sharp blow. No apparent motion is noticed in the intermediate ■ balls, but the ball at the far end moves off suddenly to a considerable distance. 13 Modern Theory and Practice in That there is an attraction between two oppositely- charged bodies and a repulsion between two bodies of like charge is proven by many simple experiments as outlined O-CCOOCO Fig. 101. in any text book on Electricity and Magnetism. Of more interest to radio students is the nature of the .stress by means of which this pull or force is exerted. We know from our study of mechanics that there must be some medium under stress linking the two bodies in order that this force may be exerted. This medium is not air, as can be shown by performing the experiments in an exhausted chamber. Innumerable experiments convince us that this force is exerted by means of a strain in the ether. This condition of strain we represent by so-called lines of elec- tric force. These lines are conceived as emanating from positive charges and terminating in negative charges of equal magnitude. Their direction is assumed to be that in which a positive charge if free would move. For simplicity we assume that at one end of each line is a free electron and at the other end is an atom from which an elec- tron has been removed. There is a tendency of the lines Pig. 102. — Electric Field Between Oppositely Charged Spheres. to shorten themselves and at the same time a repulsion between lines in parallel directions which causes the curves as shown in Fig. 103. The cause of this repulsion is as yet Eadio Communication. 13 unexplained. It may be due to arrangement of the mole- cules of the ether under strain. The attraction which exists between a charged body and a neutral body is explained as follows. We have assumed that at the positive end of each strain line there is an atom which has been robbed of one or more electrons, and at the negative end is a free electron. If such electrons exist at the ends of the strain lines on the neutral body they must have been taken from the atoms of that body. As long as these electrons are held on the side of the neutral body nearest to the charged body, there must exist on the other side of the neutral ^ody atoms from which electrons have been removed, or poisitively charged atoms. The effect is the same as though a positive charge had flowed to this side. The negative charge on the near side of the neutral body is called an induced charge. Fig. 103. — Induced Charge. Magnetic Lines of Force. There is another form of strain set up in ether which is called magnetic strain, and represented by lines of mag- netic force. This strain exists in the vicinity of an ordi- nary bar magnet, a piece of iron surrounded by a coil of wire carrying current, which is temporarily given the prop ■ erties of a magnet, also around any wire carrying current. The direction of these lines can readily be shown by sprin- kling iron filings on a piece of paper placed in the field of the lines. 14 MODEBN ThBOEY AND PRACTICE IN For a more complete treatise on magnetic lines the student is referred to any text-book on Magnetism. The point of interest here is the existence of these strains — electric and magnetic — and the fact that the strains are Fig. 104. — Magnetic Fields. set up at right angles to each other. This is made clear by reference to Pig. 105. Suppose positive and negative charges to exist as shown. As long as these charges exist +- + H- -f- + + ''II ' ' / I t I I I I / 1 1 I ; I /> 1 M \ M 1 ; ; i \ \ \ \ \ ^, W \ \ \ I I II 1,1 11 1 1 11/ \ II Electric Strain. Fig. 105.- \ -(' "■■. i' \ r \ f ;-. V 'i !►' .;» y. •' , .■i { -,; Magnetic Strain. -Electric Strain. electric lines of force exist in the space between the plates. When the switch is closed there is a flow of current and a magnetic field is set up around the conductor joining the plates. The electric field is at its maximum before the Radio Communication. 15 switch is closed, and the magnetic field reaches its maxi- mum approximately when the electric field is at its mini- mum, or when no potential difference exists between the plates. These fields may co-exist, temporarily, at less than maximum strength, one increasing and the other decreasing. This phenomenon is of supreme importance in the propa- gation of electric waves from an antenna as will be ex- plained later. Energy of Strain. In order that a spring or a piece of rubber may be put in a state of compression, or strain, work must be done on it. If the compressed spring is released it returns to its original form, and the force of the spring acting through a certain distance represents work. The work done by the spring in returning to its original form is equal to the work done in compressing the spring. The spring in its compressed .state possesses potential energy by virtue of its strained condition. A strain in any medium or elastic body represents potential energy. The strains in the ether which we have just studied represent potential energy. Work must be done, by friction or otherwise, to charge a body. The equivalent of this work, in potential energy, is represented by the strain in the surrounding ether. Numerical expressions for this energy of strain will be given under the heading of condensers. 16 Modern Theoky and Peactice in CHAPTEE II. CAPACITANCE AND INDUCTANCE. ENERGY OF CHARGES. CONDENSERS. A thorough understanding of the underlying principles of capacitive and inductive effects is necessary before the student can progress in the study of radio telegraphy. Capacitance. We have already seen that a body can be charged, nega- tively or positively, and that there is an electric field sur- rounding such a body. The question arises, '' How much charge can be put on a body of a given isize ? " The amount of gas we can put in a gas tank depends upon the size of the tank and upon the pressure of gas in the tank. Like- wise, the quantity of electricity that can be placed on a body depends upon the size, or electrical capacity, and upon the electric pressure or potential. (We shall see later that the physical size of a body is not an absolute measure of its electrical capacity.) The qu^mtity of gas in the tank is proportional to the cubical capacity of the tank times the pressure. The quantity of electricity is equal to capacity times potential. The capacity of a tank is fixed and is measured by the product of its linear dimensions. As we shall see, the electrical capacity of a body is determined not only by its size but by its proximity to other bodies and by the nature of the medium surrounding it. Suppose we charge a body by connecting it to a source of E. M. F., then remove the connection to the source. A certain charge remains on the body. The quantity of Radio Communication. 1'?' electricity is equal to capacity times potential, or Q equals CE. Now connect to the body a gold leaf electroscope as shown in Fig. 201. The gold leaves immediately become charged with like charges and, due to the repulsion, they stand apart as shown. The amount of charge on each leaf, and hence the force of repulsion, depends upon the potential of the body to which they are connected. Now bring near the charged body a neutral body. The leaves immediately approach each other, indicating that the potential of the body has been lowered. This must be the case since the charges on the leaves are proportional to the potential of the body and no part of the charge has been removed. Fig. 201. — Charged Sphere, with Electroscope Attached. Q=EG, or E—Q/C. Since E is less, C must be greater. Next insert between the two bodies a sheet of glass, mica, rubber, or other dielectric. A still further increase in capacity is noted. The increase in capacity due to the presence of another body is partially explained by the concentration of lines of force. When the body is isolated the lines radiate in all directions, terminating in induced charges of opposite polarity on the walls or ceiling of the room or on inter- mediate objects. When the second body is brought nearer, the lines are concentrated within the space between the 2 18 MoDEEN Theory and Pkactice in two bodies. If the two bodies are flat plates the field is as shown in Fig. 202. This concentration of the field into a small portion of the immediate surrounding medium increases the capacity. The eifect of tlie dielectric is to still further increase the number of lines in the intervening space. This also in- creases the capacity. This increase varies with the mate- rial used as a dielectric. The ratio of the capacity with the dielectric in use to the capacity with air between the bodies is called the specific inductive capacity, or dielectric constant of the material. -1- 4- 4- -I- ''7' I M I I 1 I I I I I li\V\ "'III 'I I 1 I I!!' 1 N'. > 1 M . I I I I I I I I I I ' ' W Fig. 202. — Distribution of Lines of Force Between Oppositely Charged Plates, or Between Charged and Neutral Plates. The increase of strain lines takes place in the ether but is accompanied by a mechanical strain in the dielectric. The unit of capacity is the Farad. This represents the capacity of a body which is charged to a potential of one volt by one coulomb of electricity. As this unit is large for practical purposes we use a unit one-millionth as large, called the microfarad. Condensers. A condenser is an arrangement of conductors and insula- tors to store up charges of electricity. In its usual form it consists of a number of metal plates separated by glass, oil, mica, air, paper, or some other insulator. In most of the Radio Communication. 19 condensers used in receiving in radio work the dielectric is air. By arranging two sets of plates, one set sliding into the space between the plates of the other set, both surfaces of all plates, except the outside ones, are made effective. Pig. 203 shows a condenser of this type. Various arrangements and shapes of plates are used. In the type most used in radio work the plates are so mounted that one set revolves with reference to the other set, thus varying the effective area of plates. The shape may be so designed that the effective area, hence capacity, varies as the angular displacement, or as the square of the Fig. 203.— Sliding Plate Condenser. angular displacement. The capacity of any plate con- denser is found by the equation p ■ J! _ Area of plates in cm.^ X dielectric constant ' ™ ™ ■" 4 X 3.1416 X 900000 X distance in cm. between plates. From this equation it is readily seen that by properly designing the shape of the plates the capacity can be made any function of the angular displacement between sets of plates. A few shapes in common use in radio work are shown in Fig. 304. Fixed capacity condensers are made up in various shapes and of quite large capacity. The student can readily see from the foregoing that there must exist, between any two conductors near each other, a certain amount of electrostatic capacity. The amount depends upon the effective area of the conductors 20 MoDBEN Thboey and Pkactice in and the distance apart. In the case of a coil of wire of many turns and spaced close together this capacitance may be appreciable. The effective capacitance of a coil is a capaci- tance of such size that it will have the same amount of energy stored in it that is stored in the dielectric surround- ing the coil, if equal A. C. voltages are applied to each. The amount of energy stored in the dielectric around the coil is, of course, a summation of 1/2 GE^ (see " Energy of Charge in Condenser," page 26) where G is the capacitance between various parts of the coil and E is the Capacity varies as square of angular dis- placement. Kolster type. FIXED PLATES -MOVABLE PLATES — *1 Semi-circular ^ plates. Capacity Percentage change per varies as angular division is constant, displacement. FiG. 204. voltage applied to that capacitance. It is convenient to con- sider the partial capacitances of the coil under two headings. First the capacitance between mechanically adjacent turns. Second the capacitance between turns which are not mechanically adjacent. In the case of any coil the capaci- tance between mechanically adjacent turns is larger than the capacitance between one turn and any other turn ; but it does not at all follow that the energy stored in the dielectric of the coil is stored principally between adjacent turns. This may be seen by considering that as a rough approxi- mation the voltage applied to the whole coil is divided equally between all the turns of the coil, and that Eadio Communication. 21 W= 1/3 CE^. Now if we compare two adjacent turns in a single layer solenoid with two turns separated, say, by nine turns, in the first case Wj^ = l/2 C^E-y^, and in the second case Fi = 1/2 C^^E^^^- G-,^ is nearly half of G^.* On the above assumed case it is seen that J^m^ = 100 E.^^, while Cio is about half of Ci. The result is that the energy stored in the dielectric between adjacent wires is only about 2 per cent of that between the wires separated by nine turns. It is seen from this that iji a single layer 0- AXI8 OF COIL Fig. 204a. solenoid the effective capacitance is due principally to the capacitance of wires which are separated from each other by some distance. The stresses produced in the dielectric by the voltages between the various turns must be similar to Fig. 204a. This .shows that displacement currents flow principally from the neighborhood of one end of the coil to * See A. E. Kennelly, The Application of Hyperbolic Func- tions to Electrical Engineering Problems, p. 210, Fig. 87. The spacing between the copper of the wires has been estimated, in the case of double cotton covered wires of moderate size, as about 30 per cent of the diameter of the wire, and the capaci- tance taken from the above curves, which apply strictly only to straight wires. 22 Modern Theoet and Practice in the neighborhood of the other end. This occurs regardless of the number of turns in the coil (as long as turns are distributed along the entire length of the coil) so that the number of turns and the spacing between turns have rela- tively small effect on the effective capacitance of a single layer coil of given mechanical dimensions. The capacitance in this case, however, does vary approximately as the cir- cumference of the turns. If a multilayer coil is wound in the simplest manner, that is, with one layer directly over the preceding one, the ends of the two layers will come adjacent to each other; o o o o o o o o o 000 bJ I ©®@®@ 3©r ©d)®®® |©©®©@©® Fig. 204b. Pig. 204c. and if there are only a few layers, the voltage between these adjacent turns will be a large fraction of the whole voltage, and the effective capacitance of the coil will be large. The effective capacitance of a multilayer coil may be made rela- tively small in either one of two general ways. One is to have a large number of layers so that the voltage between adjacent layers is only a small part of the total. Coils of the approximate proportions of Fig. 204b are used in some types of radio apparatus. The second way of reduc- ing the effective capacitance of a multilayer coil makes use of a special order of winding which puts mechanically adjacent turns only a few turns, electrically, from each other. The order of turns for a two-layer and for a four- layer winding is indicated in Fig. 204c. These are known Eadio Communicatiok. 23 as " banked windings." It is seen from the preceding that in a single layer coil the effective capacitance is due prin- cipally to the electrostatic capacity between turns which are not adjacent, while in a multilayer coil the effective Fio. 204d. — Variable Air Condenser Semicircular Plates. Leyden Jar. capacitance is due principally to the electrostatic capacity between turns which are mechanically adjacent.* * It should be seen from the preceding that where it is desired to reduce ihe effective capacitance of a coll system, it is much more important to reduce the fraction of the total voltage ap- plied to the coil which appears between mechanically adjacent turns or parts, than to reduce the electrostatic capacity between those parts. 34 MODEKN ThBOBT AND PkACTICE IN The effective capacitance is equivalent to a condenser con- nected across the terminals of the coil. There is also some capacitance between the coil and ground and from the coil to adjacent conductors. This capacity effect makes itself felt in radio compass work where a coil has some tendency to act as a simple antenna, also in the unipolar detector connections mentioned elsewhere. Current Flow in Condenser Circuits, Before going further let us study the flow of current into and out of a condenser. The most nearly correct analogy of an electric current is the flow of water in a system of pipes. A IIBI[II]!Ii!|i[|!|lj|i| Mm II Mm mnm i|i]iiliii||i]i|N|]iittao, jJiiljk|jiiiliiiiii4|]TK ' lillii!lil!lil!iiiii!li' We know from experience that an insulator of any kind .stops the flow of a direct current, while a condenser if of large capacity offers a very low impedance to the flow of an alternating current. Consider an arrangement as shown in Fig. 305, in which B represents any form of pump which tends to force water around the circuit as shown by arrows. 4 is a reservoir with an elastic diaphragm stretched across the middle. When the pump is started the Eadio Communication. 35 pressure on the left side of diaphragm is increased, some water flows into reservoir, and diaphragm is deflected to the right. The quantity flowing into the reservoir is pro- portional to the pressure caused by the pump and to the size of the reservoir. The water is assumed to be incom- pressible and the deflection of diaphragm is proportional to the pressure. As water flows into left side of reservoir an equal amount must flow out of the right side. If the pump is stopped the diaphragm contracts and the flow is reversed until the pressures are equal on the sides of the diaphragm. Pig. 206. If pump is reversed the same operation is repeated, but in the opposite direction. No water actually passes through the reservoir, but there is a flow into it and an equal flow out of it. Now consider the analogous electric circuit containing a condenser and a source of electric pressure or E. M. F. See Pig. 206. When the key K is closed current flows for an instant as shown by arrows. Plate P is positively charged and plate P' is negatively charged. The dielectric serves the same purpose as the elastic diaphragm. A dielectric is conceived to consist of atoms containing no free electrons. The electrons present are so tightly held in the atomic structure that they cannot be detached, but can be slightly strained or displaced from their normal 36 MoDEEN Theory and Pkactice in positions. Their displacement depends upon the potential of the battery and is analogous to the deflection of the diaphragm in the water reservoir. In such a circuit cur- rent flows until the potential difference across the con- denser is equal to the voltage of the battery. The, quantity of electricity flowing into the condenser depends directly upon its capacity and upon the voltage of the battery. For each unit that flows into P an equal amount flows out of P'. This flow is called a displacement current, and is made possible by the slight displacement of the electrons in the dielectric. In the case of the water reservoir the deflection of the diaphragm increases with the pressure until rupture takes place. In the condenser the displacement of elec- trons, and the mechanical strain in the dielectric, increases with the voltage until the dielectric breaks down. The volt- age at which the breakdown takes place is called the dielec- tric strength of the material. Energy of Charge in Condenser. Work is required to produce the dielectric strain in a condenser, and when the condenser is charged this strain represents potential energy. In pumping water into a tank or standpipe the work done is equal to the weight of water times the average height to which it is pumped. Height in this case is a measure of pressure. Expressed as an equation, Work = |^ WH. In charging a condenser to a maximum potential of E, the average potential, or electric pressure, is J E. Work done in charging is -J QE ergs. Since Q = CE, this expression can be written W = ^CE^, in which W is in ergs, C in farads, E in volts. Eadio Communication. 37 Inductance. In Chapter I attention was called to magnetic strain in the ether surrounding a wire carrying an electric current. This strain is represented by flux lines, the shape and dis- tribution of which is strikingly illustrated by the iron filings sprinkled on paper held in such a magnetic field. The positive direction of these lines, or the direction in which a free positive pole would move, can be determined by any of the rules familiar to the student. Perhaps the most con- venient of these rules is the Eight-hand Eule first enun- ciated by Ampere : Grasp the conductor with the right hand, thumb extended along the wire in the direction of the current, the fingers will then indicate the positive direc- tion of the circular lines of flux or magnetic strain. This .strain in the ether, in the field surrounding a per- manent magnet, is conceived to be due to the peculiar symmetrical or uniform arrangement of the molecules or atoms in the magnet. In a wire carrying current it is caused in some manner by the passage of electrons along the conductor. In either case the strained medium repre- sents a certain amount of potential energy. The strain, or the number of fiux lines used to represent it, is directly proportional to the current which produces it. In its re- active effect upon the current which produces it, this strain is analogous to the inertia or momentum of a body. An increase or decrease of current is resisted by the strain just as a moving body resists any force applied to change its speed or direction of motion. This inertia effect described above is called self-induction. Just as the inertia of a moving body differs from friction, so does self-induction differ from resistance. Eesistanee in a circuit or con- ductor is a constant factor tending to obstruct the passage of electrons. Inductive effects are present only when there g8 Modern Theory and Practice in is a change in current or a change in flux. A change in the numher of flux lines embracing or linking with the turns of a conductor, that is, a change in the strain in the ether surrounding the conductor, causes a displacement of elec- trons along that conductor. This is the principle involved in all generators of electro- motive force. A conductor is caused to move through a field of magnetic strain, or the field is caused to move with reference to the conductor. Suppose a conductor to be moved through a field caused by two magnets as shown in Fig. 207. Fig. 207.' The conductor is said to cut lines of force or the interlinkage is changed. By change of interlinkage we mean that the number of lines threading through the coil, or circuit, is changed. Interlinkage of flux lines is best shown in the case of the fields of one or more coils or solenoids. The fields sur- rounding the elements or turns of a coil add up to give a magnetic field as shown in Pig. 308. If a second coil B is placed near to A, so that the flux of A passes through the turns of B, the flux is said to interlink with B. A change in the number of lines, that is a change in the field of A, causes an electron displacement in B as well as in A. Kadio Communication. 29 In any case of a changing flux, or interlinkage, the electron displacement tends to cause a current, that is, it sets up an E. M. F., in the opposite direction to that which produces the change in flux. Take for example a straight wire carrying current. An increase in current causes an increase in flux; this increase in flux or strain ^ets up an E. M. F. which opposes the change in current. This is called counter E. IL F. In the case of a decreasing cur- > / / / M I K \ \ \ Fig. 208.— Magnetic Field of a Solenoid. rent, hence a decreasing flux, the induced E. M. F. acts in the same direction as the original current. In the ease of a coil where the flux lines embrace more than one wire, or turn, the induced E. M. F.'s in each turn add up, so that the counter E. M. F. may become very great. The value of this induced E. M. F. depends upon the rate of change of the flux. A sudden change in magnetic strain produces a great displacement of electrons. This is a point to be remembered by radio students. In the study of alternating current the student has been shown that the 30 Modern Theory and Practice in reactance of a coil increases with the frequency. A coil that would permit considerable ilow of current at any fre- quency met with in alternating current work might have such a relatively high counter E. M. F. induced, when used with the extremely high frequencies of oscillating currents that no appreciable current would flow. Such coils are called choke coils, and are used to choke off high frequency currents and to pass current of low frequency when this is desirable. The unit of self-induction is called the Henry. A coil or wire is said to have an inductance, or coefficient or self- induction, of one henry when a change of current of one ampere per second causes a change of flux, which in turn induces an E. M. F. of one volt. This unit is very large, so for most practical purposes we use some subdivision of the unit: millihenry, 1/1000 of a henry; microhenry, 1/1000000 of a henry. Energy Stored in Magnetic Field. Magnetic strain represents potential energy. If the current in an inductive circuit is increasing, part of the energy being put into the circuit is being used to over- come the inertia efEects of self-induction. This energy is being stored in the magnetic field, and if after an instant the current begins to decrease the energy is returned to t]ie circuit, tending to prevent the decrease. In the case of an alternating current the energy stored in the magnetic field during one quarter of a cycle is returned to the circuit during the next quarter. The numerical value of this stored up energy is given by the expression 1/2 LP joules, where L is given in henries and I in amperes. The mathe- matical derivation of this expression will be found in Chapter XII. Kadio Communication. 31 Permeability. In the study of electric strain we noted that the number of strain lines might be increased by the use of suitable dielectric material. Likewise we can increase the number of jnagnetic lines of force, or flux lines, by establishing the field in magnetic material, such as iron. By magnetic material we mean a material that is attracted by a magnet, or that can be temporarily given the properties of a a magnet. Iron, cobalt, nickel, and manganese are ex- amples of magnetic materials. The ratio of the number of flux lines, due to a certain current or magnetizing force, in iron, to the number that would exist in air, due to the same magnetizing force, is called the permeability of the iron. Tlie permeability of some grades of iron may be as high as 60,000. The number of lines of force, hence, the inductive effects noted, may be increased by causing the flux path to lie partly or entirely through iron. This increase of flux, due to iron core, becomes less as the fre- quency increases. Mutual Induction. When two circuits are so placed that the flux of each cuts the turns of the other, a change of flux in either cir- cuit produces an E. M. F. in each circuit. The two cir- cuits are then said to have mutual inductance. The unit of measurement is the same as in self-inductance, i. e., the henry. Two circuits have a mutual inductance of one henry when a change of current in one circuit of one ampere per second produces in the other circuit an E. M. F. of one volt. Circuits showing a few applications of the inductive effects noted above will be given. The interlinkage of flux lines determines the amount of inductance of a wire or coil. 33 MODBEN ThEOEY AND PRACTICE IN SO various forms of coils are used. With low frequency currents the coil may be wound in several layers on an iron core. By this means it can readily be seen that the num- ber of lines and the interlinkage is increased. An induction coil, as shown in Pig. 209, is used to pro- duce a high E. M. F., and was formerly used to a great extent in spark radio sets. Its operation is as follows: Current from a battery is supplied to the primary winding, which is mounted on an iron core. This produces a strong SECONDARY WINDING - PRIMARY WINDING- 0^ c Hi I I 1} BATTERY Fig. 209.— Induction Coil. flux which links the primary and also the secondary wind- ing, which is mounted concentrically with, but outside the primary. The iron core is magnetized and attracts the armature A. As soon as A moves to the left the primary circuit is broken, the flux lines suddenly collapse, and this sudden change acting upon the .secondary, which consists of a large number of turns, induces in the secondary a very high E. M. E. As soon as the field collapses the core is demagnetized and A is drawn by a spring to its original position, the circuit is completed and operation is repeated. In modern radio sets, instead of supplying battery cur- rent to a circuit containing a make-and-break arrangement Radio Communication. 33 as described above, alternating current is supplied directly to the terminals of the primary winding. The secondary then supplies current of the same frequency and at a voltage which is approximately in direct proportion to the number of turns in the respective windings. See Fig. 310. In such coils as above the primary, or low voltage, wind- ing consists of a few turns of heavy wire, while the sec- ondary consists of a great many turns of fine wire. The secondary is made up in several sections, and as each sec- tion is put in place the end of wire is secured to the pre- ceding section so that all sections are in series. This FLUX PATH, DOTTED LINE Fig. 210. — Closed Core Transformer. arrangement in sections is to facilitate repairs and insula- tion. In case of a burned out coil or broken wire the injured section only need be rewound. One of the simplest applications of the magnetizing ef- fects of a conductor wound on an iron core is found in the high frequency buzzer as used for testing purposes in radio. The operation is the same as that described under induc- tion coils, except that there is no secondary winding. The speed of the make-and-break is controlled by the tension of the spring supporting the armature and, in manufac- ture, by varying the weight of the armature. The use of the buzzer to supply intermittent current will be explaiued under Oscillating Circuits. 3 34 ModeejST Theoey and Peactice in CHAPTEE III. OSCILIATING CIRCUITS. FREaUENCY. DECRE- MENT. WAVE PROPAGATION. WAVE METERS. DETECTORS. Oscillating Circuits. To Prof. Henry, of Princeton University, is due the dis- covery that under certain conditions the discharge of a condenser, througli an inductive circuit, is oscillatory in character. Later, Lord Kelvin demonstrated matheniati- A B m^^^ ^^M Fig. 301. cally the relations between the inductance, resistance, and capacity that must exist in order that the discharge shall be oscillatory. This relation depends upon the condition that if all the energy stored in the condenser is used up, in PR losses and radiation to adjacent circuits, during one discharge of the condenser, the circuit cannot oscillate. On the other hand, if the energy is not all used up in a single discharge, the circuit will oscillate. Radio Communication. 35 The phenomenon of oscillating circuits is analogous to the flow of water between two tanks connected by a pipe between the bases. See Fig. 301. If the pipe connecting the tanks A and B is of large cross-section, when the valve C is opened the water in A rushes through to B and, due to its inertia, rises to a height almost equal to its original level in A, then rushes back to A, reaching a level slightly less than level reached in B. Several such surges take place before the water comes to rest with a common level in A and B. If the valve C is just cracked so that the passage is restricted, no surges take place, but the water level in A falls slowly until height in A and B is the same. Fig. 301a. The discharge of a condenser whose plates are connected through inductance, as shown in Fig. 301a, is analogous. Suppose the plates to be charged as indicated. If the resistance of the . connecting wire is less than V4Z;/C (Z/ = inductance in henries, C = capacity of condenser in farads) the circuit oscillates, i. e., when the key is closed the charge on A sends a current through the wire to B, the potential of A falls and that of B rises. When the poten- tials of the two plates are equal the current does not stop, but, due to the inductance, or inertia effect, continues to flow for an instant, charging the plate B positively and the 36 Modeejst Theory and Practice in plate A negatively. The current stops when the inertia effect is overcome by the accumulated E. M. F. on the con- denser. The potential to which B is raised is not quite as high as the original potential of A, due to the fact that part of the energy is used up in sending the current through the resistance of the connecting wire. In the case of the water tanks, part of the potential energy of the water in tank A is used up in overcoming the friction in the pjpes so the highest level reached in B is not quite as great as the original level in A. As soon as the inertia effects in the electric circuit have ceased, plate B being positively charged, the current reverses and the same operation is repeated in the opposite direction. This surging back and forth continues until the energy originally stored in the condenser has been dissipated in heat losses or otherwise. The surging of the water is due to its inertia. The surg- ing of electric current is due to its electrical inertia or in- ductance. With a given capacity and resistance, the greater the inductance the more times the current will surge back and forth. If the resistance of the circuit is greater than V4L/6', the charge on A acts like the water in the tank when the valve is just cracked. It leaks out until the potentials of the two plates are equal. We have seen that the number of oscillations or surges depends upon the resistance, induc- tance, and capacity. The frequency of the oscillations de- pends upon the capacity and inductance and is determined approximately by the formula /= ^——^ , The oscillating circuit in a spark transmitting set is as .shown in Fig. 302. The condenser is charged by an indue- Eadio Commuxication. 37 tion coil, or by other suitable means, to a very high voltage. The spark gap in the circuit prevents any flow of current other than into the condenser until the voltage across the condenser — and across the spark gap — has reached a value sufficient to break down the air insulation of the gap. When this takes place the space between the electrodes of the gap becomes ionized, its resistance becomes very low, and oscil- lations take place. It is well to study carefully what takes place in an oscillating circuit. To start with, the plates are charged as shown and all the energy in the circuit is stored in the condenser in the form of dielectric strain. When the gap TO SOURCE OF POWER Fig. 302. — Oscillating Circuit of a Spark Transmitting Set. breaks down and discharge begins, current through gap and coil builds up, and after a certain period no potential difference across condenser exists. The current is now a maximum and the energy is now in the form of magnetic strain in the ether surrounding the conductor. The col- lapse of these magnetic strain lines on the coil produces the inertia effect, resulting in a continuation of the current for an instant, thus charging the condenser with opposite polarity and establishing again the electric strain. This operation is repeated several times, but with each reversal of current the potential difference across the condenser is less and the succeeding current is less. The decrease is due to heat losses and to radiation or transfer. The current 38 Modern Theory and Practice in flowing in the circuit may be represented by a curve, as shown in Fig. 303. In this curve current values are plotted as ordinates against time as abscissae. Such a current is called a damped oscillating current or, simply, damped oscillations. The / Fig. 303. ratio of the maximum ordinate of one alternation to the corresponding ordinate of the succeeding alternation in the same direction is called the damping ratio. In Fig. 304, -4/5 = 5/C = damping ratio. This ratio in a pure, damped oscillation is constant and depends largely upon the resis- tance of the circuit. The logarithm of this ratio to the y Fig. 304. base " e " is called the log decrement. If the decrement * is large the current is said to be highly damped. If the resis- tance is low and the inductance fairly large the decrease in amplitude per oscillation is less and the current is said * " Log decrement " is usually abbreviated to " decrement." EaDIO COMMUlSriCATION. 39 to be slightly damped. Knowing the decrement, we can compute mathematically the number of oscillations that will take place before the current dies to any given fraction of its original value. Suppose the oscillating circuit containing a spark gap to be changed in form as shown in Fig. 305. The plates still form a condenser but, if of the same size, the capacity is much less, due to the increase in distance between them. This circuit can be made to oscillate as before change in form. Let us study the strains produced in the ether by a damped oscillating current in this circuit. At the instant ■ill, M I I tC^ I I I I iiilliiili 1 \ I//; / i \ I \ \\ \\ \\ I I ! ! I I Fig. 305. Fig. 306. before discharge begins the energy is all electrostatic, that is, stored as .strain in the ether between the plates. This strain is represented by lines as shown in Tig. 306. When the voltage is sufficiently high the gap breaks down, current begins to flow, and a magnetic strain is built up around the conductor. At the same time the positive and negative charges represented at the ends of the strain lines rush to meet each other and the electric field collapses. Due in part to the good conductivity of the metal plates and the conductor joining them, the ends of the long strain lines meet and the opposite charges are neutralized before the strain in the ether, represented by these lines, has been relieved. This results in closed loops of electric strain as 40 MoDEEN Theory and Practice in shown in Fig. 307. This is the condition at the end of the first quarter of a cycle. The magnetic field surrounding the conductor is near its maximum value and, due to its inertia effect, is beginning to charge the plates with reverse polarity. Electric strain lines of opposite direction are Z^-' Fig. 307. 1 ifO I ,' FiQ. 308. beginning to form and push these strain loops out. At the end of a half cycle the condition is as shown in Fig. 308. At the end of the third quarter cycle a loop of electric strain of opposite direction has been formed. These loops of electric strain are pushed outward from the oscillating circuit with a velocity assumed to be that of light.. The disturbances in the ether resulting from one oscillatory discharge form what is called a wave train. \M Pig. 309. The radiation from a grounded antenna is slightly dif- ferent. One end of such an antenna is grounded, hence the potential of that end is zero and the electric field is as shown in sketch of single wire antenna. Fig. 309. Eadio Communication. 41 Consider what happens when the discharge begins in the grounded antenna. The strain lines tend to shorten and form loops, as in the case of the oscillating circuit pre- viously described, but due to the resistance of the earth, the earthed ends move more slowly than the top ends, so the top ends of the loops reach the bottom of the antenna before the earthed ends do. The result is a system of loops of elec- tric strain as shown in Figs. 310 and 311. / I I I >> r if^A k\ Fig. 310. /,-, II I I I II I I i _Llj liIl Hi ilUlf \r .■(INl \.\ \\ V, 1 I All Fig. 311. Pig. 311 represents the condition before the current reverses and while the antenna is being charged negatively. The iirst loop is being pushed out by the reversed strain lines being formed. The propagation of these strain loops through space admits at present of no rigid scientific explanation. From our conception of strain lines — their similarity to elastic bands — we assume that the loops have a tendency to shorten and collapse. Clerk Maxwell has shown that, so far as con- cerns the magnetic strain which accompanies the electric strain, that a decreasing electric strain is equivalent to a current in a direction opposite to the positive direction of the strain lines, and that an increasing electric strain is 43 MoDEHN Theory and Practice in equivalent to a current in the same direction as the strain lines. There must exist, then, in the dielectric, air, a dis- placement current in direction as stated above. These dis- placement currents set up magnetic strains in the surround- ing ether ; the variation of the magnetic strain in turn setsup electric strains in the adjacent dielectric. Since new loops of strain are constantly being formed, tending to push out those already formed, the strains resulting from the dis- DIRECTION OF PROPAGATION ,' — ^ // ,' -~\ ' / '1 ' ,' 1 ' ' ' / // //,'/ //// //// // /'// /"/ '/// 11 Hi m M a 1 1 III! I II 1 1 III nil III 1 1 I I I I I ' ! I I I 1 1 ' I tl I III nil //// 6URFA/ii;' OF EARTH A B Fig. 312. placement currents appear to loop over outward, i. e., in the direction of propagation, forming loops of increasing diameter. Pig. 313 represents the approximate shape of these waves, at some distance from the antenna. The waves are radiated in all directions from the antenna, so this sketch represents a radial section of the disturbance. The electric and magnetic strains are propagated at a velocity approximately that of light, 3 X 10'* meters per second. The velocity of propagation in any wave motion is equal to frequency times wave length. In Pig. 312 the Radio Commui^^ication. 43 wave length is represented by A B, the distance in the line of propagation between like strains or between particles of the disturbed medium which are in the same phase. The feet of these loops represent points of different potential, hence current flows along the surface of the earth between these points. This flow of current through resistance represents a loss of energy. Since there is a considerable difference in resistance of dry earth, wet earth and sea water this accounts for the difference in range obtainable over land and sea. This same resistance of the / -- — I / I I I I I I ( I I I / / I a J_L I I 11 I, I \ \ \ ^ I I I I I Fig. 313. earth tends to retard the travel of the lower ends of the strain loops, hence they move more slowly than the upper ends, even though the upper ends extend into the rare atmosphere which is more or less, conductive. For this reason the loops which start out inclined slightly towards the antenna soon become inclined forward, i. e., in the direction of propagation, as shown in the figure. In practice we seldom use an antenna in the form of a straight wire except in airplane sets. Usually the antenna is in the form of an inverted L or T shape. The distribu- tion of strain lines around an inverted L antenna is as shown in Fig. 313. 44 Modern Theory and Practice in Effect on Receiving Circuits. Let us study the effect of these waves on a receiving antenna placed in their path as in Pig. 314. An antenna placed at B will be subjected to a changing electric strain and to a magnetic strain at right angles to it. Hence a difference of potential will be induced between the two ends of the antenna. A moment later the field at B is reversed and the direction of the E. M. F. in the antenna is reversed. The result of this E. M. P. is an oscillating current in the antenna of the same frequency as the source of the oscillations. '<^Vi /^-^ /^''-^^ ti mi ill t , I If 1 III A B Fig. 314. We have assumed the electric strain lines as nearly ver- tical. At great distances the strain lines may be inclined forward to a greater . degree, and the more necessary it becomes to have part at least of the antenna horizontal. Under certain conditions and at great distances signals may be received on an antenna only a few feet from the ground and placed horizontally. At short and intermediate ranges the higher the antenna (single wire) the more length of wire is subjected to the strain and the more energy it will pick up. Eadio Communication. 45 Wave Meter. As previously explained, the frequency of oscillations in a circuit consisting of inductance and capacity in series depends upon the values of the inductance and capacity, and is determined by the formula / = l/27rV-Z>C'. Such a cir- cuit, in which the capacity or inductance is variable, when calibrated and marked to read frequency or wave length, is called a wave meter. Both capacity and inductance may be variable, one in steps and the other continuously. The wave length or frequency may be marked on a scale attached to the variable element or the scale may be marked in divisions or degrees and the wave length found by reference to a curve of wave length plotted against degrees or divi- FiG. 315. Simple Wave Meter. i^> ^ / Fig. 316. sions. In case one element is variable by steps a separate curve is required for each step of variation. Fig. 316 shows a wave meter with the usual detecting device, safety spark gap, and buzzer and battery as used to produce, in the circuit, damped oscillations. The use of these devices will be explained in detail, partly here and later under the head of Detectors. The action of the buzzer and battery is as follows : When key is closed, con- necting battery and buzzer in series across terminals of the condenser, current begins to flow from positive side of battery, through buzzer winding, inductance of wave meter. 46 MoDEEN Theoky and Practice in back to negative side of battery. This cairreiit through buzzer winding, which is in the form of a solenoid, causes the armature to be attracted away from the contact indi- cated by arrow head, thus breaking the circuit. Due to the inductance of the wave meter circuit, this current which is flowing when circuit is broken, continues to flow for an ■ instant, charging one side of the condenser positively and the other side negatively. This is the same condition pre- viously described tmder oscillating circuits. Current surges back and forth until the energy is dissipated in heat losses or otherwise. As soon as current ceases to flow through buzzer the solenoid is no longer energized and Fig. 317. armature is returned by a spring to its former position, again making contact. However, since the make-and-break of the buzzer is very slow in comparison with the frequency of oscillations, the energy stored in the magnetic field after one break is dissipated and current ceases before the next make. Hence we get a series of wave trains as shown in Fig. 317, one wave train for each break. Detectors will be described in full in a later chapter, but a few words here are necessary in order that the student may understand their use in this and the following chapter. The detector here shown is a crystal of some metal or oxide in contact with a fine wire. Such a contact offers a very low resistance to the flow of current in one direction and a very high resistance to flow in the opposite direction. Eadio Communication. 47 Its action is analogous to that of a check valve. When the key is closed on the detector-phone circuit the E. M. F. of the condenser is impressed upon the detector. Due to the high resistance in one direction and low resistance in the other, the current which flows through the detector may be represented by a curve as shown below. The diaphragm of the phone is subjected to a series of pulls in the same direction. These .pulls come at very high frequency, but the result is the same as a longer pull but of less strength. It is readily understood that the inertia of any telephone diaphragm is too great to permit it to respond to the indi- vidual pulses of current. Even if the diaphragm could respond to the individual pulses of current the frequency Fig. 318. would be too great for detection by the human ear. With the detector connected, as shown in Pig. 316, the frequency of clicks in the phone is equal to frequency of wave trains, which is same as " breaks " per second of buzzer. This frequency is within the range of audibility. As a matter of convenience, frequencies below 10,000 cycles per second are called audio frequencies; above that number, radio fre- quencies. It is possible for some human ears to detect frequencies much higher than 10,000 per second. The spark gap placed around the condenser of the wave- meter serves to protect the instrument from very high voltages which might be induced in it when brought near oscillating power circuits. The gap is set to break down 48 MoDEEN Theory and Peactice in at a voltage much lower than would be required to damage the instrument. The hot-wire ammeter placed in series in the circuit may- be calibrated to read current directly or the divisions on the scale may be proportional to current squared, in which case the readings are proportional to the energy in the circuit. Uses of the Wave Meter. The wave meter is an indispensable piece of apparatus in all radio stations and laboratories. In addition to its use to produce damped oscillations of known frequency, as just described, it is used in tuning various circuits, in measuring capacity, inductance, and decrement. To measure the wave length, or frequency, of any oscil- lation the wave meter is so placed that part of the electro- magnetic lines of force, or flux, of the oscillating circuit cut or link with the inductive portion of the wave-meter cir- cuit. This changing flux induces an E. M. F., and an oscillating current flows in the wave-meter circuit with a frequency equal to that of the source. As we vary the capacity, or inductance, of the wave meter we find one value at which the current in the wave meter, as measured by the sound in the phones, is a maximum. In this condition the two circuits are said to be in resonance. The frequency and the wave length are the same in the two circuits. This wave length can be read directly from the scale attached to ,the wave meter or from a curve of wave length plotted against condenser setting. The phenomenon of resonance can perhaps be better understood by reference to the equa- tion familiar to students of alternating current electricity. E ^R^+(2.fL-^^y Radio Communication. 49 It is readily seen that for a certain value of the frequency, f, the value of the expression in parentheses becomes zero. When this is the case 7 is a maximum and is equal to E/R. The circuit is said to be resonant to that frequency. A circuit is brought into resonance with, or tuned to, another circuit frequency f, by varying C oi L until 2irfL=l/2irfC. 50 Modern Theory and Practice in CHAPTEE IV. COUPLING. TRANSFER OF ENERGY. MECHANICAL ANALOGIES. Coupling. Under the subject of mutual inductance and in con- nection with the use of the wave meter, mention was made of the fact that when the flux lines produced by a current in one circuit thread through, or cut, the turns of wire in another circuit, a change of current in the first circuit produces an E. M. P. in the second circuit. If the second Fig. 401. circuit is closed conductively, current flows in it with a fre- quency equal to the frequency of periodic changes of cur- rent in the first circuit. If the second circuit is closed through capacity we may have, in addition to the forced oscillation of above frequency, free oscillations of a fre- quency depending upon the inductance and capacity of the second circuit. Suppose an alternating current to flow in the circuit A, Fig. 401. Part of the flux produced by the current in A links with the turns of B. Each change of current in A produces an E. M. P. in B. A transformer is a practical Eadio Communication. 51 application of this principle. Any two circuits so placed that a change of current in one produces an E. M. P. in the other are said to be " coupled." This means simply that each circuit is wholly or in part in the field of the other. The field may be electrostatic or electromagnetic. Thus we have several forms of coupling: Direct, in which the two circuits have some part in common; inductive, when the transfer of energy is effected by electromagnetic lines of force ; capacitive, when electrostatic lines are used to convey the energy. i-T* A o B INDUCTIVE CAPACITIVE COUPLING CONDENSERS CAPACITIVE Fig. 402. — Forms of Coupling. It can readily be seen that the greatest B. M. P. will be induced in the secondary when the greatest number of flux lines of the primary are cut by the turns of the secondary. The circuits are then said to be close coupled. The current in the secondary will be a maximum when the natural fre- quency of the circuit as determined by the formula f=l/2TrVLCj is the same as the frequency of the source; in other words when the circuits are in resonance. The condition for maximum E. M. F., that is, close coupling, is desirable for some purposes, such as trans- formers in power circuits. In certain radio circuits, in 53 Modern Theoet and Peactice in which clamped oscillations are produced, close coupling has disadvantages, to be studied later, which make it desirable to " loosen " the coupling, that is, arrange the circuits so that only a small part of the flux of the first circuit links with the turns of the second circuit. This may be accom- plished in direct coupling by reducing the portion that is common to both circuits, and in inductive coupling by increasing the distance between the coils or by turning one coil relative to the other so that the axis of one coil is nearly at right angles to that of the other. This decreases Fig. 403 the interlinkage. To loosen the coupling in capacitively coupled circuits we usually decrease the capacity of the coupling condensers. Before studying in detail the transfer of energy be- tween coupled circuits, let us consider a simple mechanical analogy. Suspend two bobs of equal weight, by strings of equal length, from a flexible support such as shown in figure. The periods of oscillation of the two pendu- lums, swinging independently of each other, are the same. Start one bob to swinging by drawing it aside about 30 degrees and releasing it. After a few swings the other bob will take up the motion, and in a few moments Radio Communication. 53 the first bob will come to rest. At this instant the second bob will be swinging at a maximum amplitude. The operation now reverses. The first bob again takes up the motion, while the second gradually comes to rest. The energy originally imparted to the first bob is transferred back and forth until dissipated by the friction of the air, etc. The agent by which the transfer is effected is the flexible support, in this case a horizontal string. At each swing of either pendulum the horizontal string is deflected, at the point at which the bob is supported, in the direction in which that bob is swinging. This deflection tends to displace other points along the string, including the point of support of the other bob. Now displacing the point of support of any pendulum has the same effect, so far as producing oscillations is concerned, as drawing the bob to one side, as we did in the first case. Next increase the distance between the points of support of the two pendu- lums and repeat the experiment. It will be found that less energy is transferred — amplitude attained by second bob is less — and that the rate of transfer is slower. If we count the oscillations per second of each bob we will find that the bob that is transferring energy to the other is oscillating at a rate .slightly slower, and the one receiving energy, at a rate slightly faster than that at which either would oscil-' late when swinging independently, or from a rigid support. In other words, neither pendulum oscillates exactly with its natural frequency. The flexible support serves as a coupling betwen the pen- dulums. The distance between the points of support is a measure of the degree of coupling. The transfer of energy between coupled electric circuits is very similar to the phenomenon just explained. 54 MoDEEN Theory and Peacticb in Suppose we tune the circuits A and B, ¥\g. 404, inde- pendently to the same frequency and couple them as shown. Produce in A damped oscillations by connecting to ter- minals of condenser a source of E. M. P. The oscillating current in A will induce in B .similar currents, but neither circuit will oscillate with the frequency to which it was tuned independently. The changing magnetic flux in A produces an E. M. P. in B. The resulting current in B causes a flux which reacts on A, with or against the flux of A. This in turn effects the frequency of oscillations in A. There is a transfer of energy back and forth as with the pendulums. Pig. 404. The curents flowing in these circuits may be repre- sented by the curves of Pig. 405. As, in the pendulum analogy, the maximum amplitude of one bob was attained when the other was practically at rest, so, in the electric circuit, the maximum amplitude of current in one circuit is reached when the current in the other has died practically to zero. To start with, the energy is all in one circuit. After a few oscillations the energy in the first circuit is dissipated by heat losses and by transfer to the other circuit and the oscillations cease. At this instant the second circuit, with maximum current flowing, begins to transfer energy back to the first. All current in the first circuit represented by that part of the curve to the right of the point A is produced Eadio Communication. 55 by a re-transfer of energy. Due to the PR losses and to slight radiation the total energy is constantly decreased until oscillations in both circuits cease. As was mentioned above, coupling two circuits which are tuned to the same wave length tends to cause two differ- ent wave lengths to be produced in this system whenever one of them is excited. One of these waves will be longer Fig. 405. — Current Curves of Coupled Circuits. and the other shorter than the wave length to which the individual parts of the circuit are tuned. In Fig. 401, circuit A, the actual distribution of flux around a coil carrying current is shown. As a matter of convenience, when two circuits are to be inductively coupled, the actual flux may be thought of as being split up into two parts, which together are equivalent to the actual flux. These two fluxes would be as in Fig. 406 (a). It is, of course, flux linking with a coil which causes the inductance 56 MoDEEN Theory and Practice in of the coil, the inductance being numerically equal to N^/I, where N is the number of turns in the coil, $ the flux shown in Fig. 406(a) and I the current which is re- quired flowing through the circuit to cause this flux to be set up. If the coil of Fig. 406 (a) is connected in series with another similar coil and the two coils 80 placed that no flux from one links with the other, the total flux through each coil is unchanged, the inductance of each coil is unchanged, and the total inductance will be found to be simply the C ■^sssms^ J /^" (a) D ^*^^gfeM^ C ) Fig. 406. D arithmetical sum of the inductances of the separate coils. Consider the effects of bringing the coils close together as shown in Fig.'406 (&) . The flux from the right-hand coil is shown by dotted lines, that from the left-hand coil by solid lines. If the coils are so placed that all of the fluxes are going in the same direction, the total flux through each coil is increased, and the effective inductance of each coil is increased by an amount due to the increase of flux through that coil, the total inductance thereby being increased by twice this amount. This added inductance in each coil due to flux from the other coil is known as mutual induc- tance (symbol M). The total inductance is now (La+M) Radio Communication. 57 + (Lb + M) or La + LT, + 2M. If the connections to one coil be reversed, the arrangement of the component fluxes will be the same' except that the ones corresponding to M are now in opposite directions in each coil. The total flux in each coil is therefore decreased, the effective inductance of the combination is decreased, and is now equal to (La-M) + (Lb-M) or La + Li-2M. If instead of connecting the two coils in series with each other each one is connected in series with a condenser to form an oscillating circuit, a current started in one will cause a current to be produced in the other (assuming both Fig. 407. circuits tuned to the same frequency) . Both coils will then produce fluxes and the fluxes corresponding to M will act to alter the effective inductances of the coils. Notice that as the coils are no longer in series, the phase relations be- tween the currents in the two are not fixed, and the fluxes produced by the currents may add or subtract or be out of phase with each other,* thus producing the possibility of more than one frequency in the circuits. To illustrate the existence of two frequencies in coupled circuits a different type of mechanical analogue will be presented. Fig. 407 shows two pans containing weights, * Generally the currents will not be of the same size either. 58 Modern Theory and Practice in each pan suspended by a spring from a fixed point. The weights and springs must be so adjusted that these two devices will oscillate up and down at exactly the same rate, when entirely separate.* The cross beam and center pan shown are made of such light weight that the efEect of their weight is negligible. In this system the motion of the weights corresponds to the flow of current in an oscillating circuit, and the forces correspond to the voltages. The weights, due to their kinetic energy, store energy in the same manner that an inductance does when a current flows. The spring aided by the force of gravity stores energy as the condenser does. The coupling between these two oscil- lating circuits is provided by placing a weight on the center pan, thus corresponding to inductive coupling of electrical circuits. To make this system act as the circuits of Fig. 406 (&) do, take, say 50 gm. from the left-hand pan and 50 gm. from the right-hand pan each to represent M. Take also 100 gm. from some outside source to represent the 2M gained by the coupling. Place this 200 gm. on the center pan. The capacitance of this .system, as represented by the springs has not changed, but if all the weights are started moving up or down simultaneously, it will be seen that the weight carried by each spring is one-half of the total, or the original weight -t-50 gm. (corresponding to L+M). The motion in this case presents the lower frequency at which this system can oscillate. If the left and right-hand weights are started in opposite directions with equal velocity, so that one goes up while the other goes down, the 200 gm. carried in the center will not move. The weight moving with each spring will be the original — 50 gm. (correspond- * To simp-lify the explanation the weights will he taken exactly equal and the springs exactly alike. Eadio Communication. 59 ing to L — M). The frequency will therefore be higher than the original. This demonstrates that a system such as this can oscillate at either of two frequencies separately. If one weight is started into motion the beam acting on the central mass as a fulcrum starts the other weight. The first will stop, and then the energy will be re-transferred, just as in the preceding p'endulum case. The presence of the two types of motion, corresponding to the two different fre- quencies, can be seen by watching the system during this action. Fig. 408. — Radio Frequency Circuits of a Spark Transmitter. If the coupled circuits considered are the closed, or oscillating, and the open, or radiating, circuits of a spark transmitting set, as shown in the figure, the effects noted are, for some reasons, very objectionable. The object in a transmitting set is to radiate as much energy as possible at a fixed wave length. We shall study in detail the effects on energy radiation and wave length of varying the coupling between the closed circuit and the radiating circuit, G and in the figure. If the hot gases 60 MoDBBN Theory and Peactice in remain in the gap so that the resistance of the gap is low throughout the wave train, the curves previously shown, Pig. 405, apply to these circuits. The decrement of a cir- cuit is a function of the resistance, inductance, and capaci- tance, and varies directly as the capacitance. Since the capacitance of the closed circuit is much greater than that of the open circuit, its decrement is also much greater. The ohmic resistance of the open circuit is usually very low and the principal source of damping is the loss of energy due to the radiation of electromagnetic waves and loss due to imperfect dielectric. It is readily seen that the closer we couple the circuits C and the more energy is transferred to the open circuit, but at the same time energy is more rapidly returned from to 0, due to flux of reacting on C. This results in increased damping of the open circuit which is undesirable. Also that part of the energy returned to the closed circuit which is dissipated in that circuit as heat and light, represents a distinct loss. Most efficient operation is attained when a maximum amount of energy is transferred to the open circuit and dissipated in that circuit as electromagnetic radiation. The radiation from the open circuit can be studied as follows. Couple circuits G and closely and couple loosely to a wave meter equipped with a current-squared meter or wattmeter, or other device whose reading is pro- portional to the energy in the circuit. Set wave meter to any desired wave length and read current-squared meter. Take several readings above and below the wave length to which the circuits C and are tuned. Plot the readings of the wattmeter or current-squared meter as ordinates against the wave-meter setting as abscissse. The result will be a curve as shown in Pig. 409, called a resonance curve. This curve shows that energy is being radiated principally at two wave Eadio Communication. 61 T-- fmTnT : : : : \\ ■- :: :: :: ;: .. r: : : a : ::: :: ... : : : : -■■ 7^ = 1" :: :: r l> r ■ :■:: I:::: :::■.: ::: :::: j ::::::::::: ::^ : :; ; : \i : ■ ■ -■' i ; : : :: - ■ ■ ■ : : l::: ::: = ; :;;:: :::: : : : : x3 ,-T :: :: ■s :': :: ::: :: :::. _ ^; :i : : : : \\\ : : J|.. k \:\[-. :: ± -:::::::::::::±:: :: ::: :l :: i ^ mm of- a •a > I o a 9 " a a^ ^ ® r3 tn ra o I a3i3w a3.avn&s iNsaano 62 MoDEBN Theokt and Practice in lengths as indicated by the points A and B. The wave meter circuit picks up some energy when adjusted to any wave length between the limits of the curve, but the points A and B are decided maxima. Suppose the height of ordinates of the line CD represents sufficient energy to operate the receiving device of a station. Then that station can read the signals when adjusted to any wave length between 418 and 538 meters. Hence we say a broad wave is being sent out. This condition is desirable when the message is to be received by all stations, as when distress signals are being sent out. However, when the signal is intended for a par- ticular station, tuned to a certain wave length, such a wave has the disadvantage of interfering with all stations tuned to wave lengths slightly different, and is not as easily re- ceived by the called station as a pure or sharp wave. Now suppose we loosen the coupling and measure the radiated wave length as before. The peaks approach each other as the coupling is decreased until with very loose coupling the energy is radiated at a single Avave length as shown in curve 2. This wave length will be found to be that to which the two circuits are tuned. Various means have been devised to obtain this pure wave and at the same time use closer coupling, with increased energy transfer. Per- haps the best results have been obtained with the iSO-called " quenched gap," which will be described in chapter on Spark Transmitting Sets. Coefficient of Coupling. The coefficient of coupling is a measure of the inter- action between coupled circuits and in inductive coupling is given by the formula M/VL-^L^, where M is the mutual inductance of the circuits and L^ and L^ are the inductances of the individual circuits. Eadio Communication. 63 CHAPTEE V. VACUUM TUBES. As has been explained, every metal is believed to contain large numbers of electrons which are free to move about within the metal. As long as the average velocity of these electrons is relatively low (say less than 100 km. per sec- ond) any electron which happens to come very close to the surface of the metal will be acted upon by a force exactly similar to the surface tension in a liquid, and will not be able to escape from the metal. If the average velocity of the electrons is increased, some of them will reach velocities W?? hIiIiIiIi I I 1 I I Fig. 501, high enough to enable them to pass out of the metal. This increase in the average velocity of the electrons is produced by heating the metal, and is accompanied by an escape of electrons which increases very rapidly with the rise in temperature. Consider an incandescent lamp with two independent filaments. If one of the filaments is heated it will give ofl! electrons. Now if a battery is connected, as shown in Pig. 501, so that the cold filament is charged positively it will attract electrons. As electrons are being given ofE steadily by the hot filament there will be a steady stream 64 MdDEKN ThEOEY and PkAi'TICE IN of them across the space between the tihmienls. It is to be noted tliat as the charges of electricity passing from tlie liot filament to the cold one are nef/olivc tbc current is thought of as passing from tlic cold filament to the hot one. If tlic cold lilament liad been gi\en a negative charge it would Fig. oOIa. — Three-Element Vacu- um Tube; Glass Bulb Removed. repel the electrons given off l\y the hot iilanient, and tliere being no electrons given oif by the cold filament there would be no flow of current. As this seems to completely explain the action wlien the cold filament is negative (if there is only a negligible amount of gas in the tube), the following discussion will be devoted entirely to the action when the KaDIO COMMUXICATION. 65 cold element is cliarged positively. In the practical vacuum tube the cold filament is replaced by a cold metal plate which has the advantage of having much more area than the filament. The circuit from the filament to the plate through the battery (or generator), which keeps the plate positively charged, is known as the plate circuit. Let us consider the forces acting upon a single electron, in the tube described above, at a time when the electron is Pio. 50lB. — Types of Audion Bulbs. part way from the filament to the plate. These forces are due to three things: (1) The action of the plate on the electron; (2) the action of the filament on the electron; (3) the action of the other electrons upon the one being considered. The action of the plate when positively charged is to attract the electron. The action of the fila- ment is considered as being due to the loss of some of its electrons, which results in a positive charge and a conse- quent attraction for the electron being considered. The 66 Modern Theory and Practice in action of the other electrons is in all cases a repulsion, as like charges always repel each other. Its direction in the tube depends upon the position of the other electrons, being away from the plate for electrons between the one con- sidered and the plate, and toward the plate for all others. It should be noted that the force on one electron due to the action of the others changes as the electron passes from the filament to the plate. The nearer to the plate it gets, the smaller is the number of electrons ahead of it, and the larger the number behind it. Ordinarily, at some point between the filament and the plate the force acting on the E Fig. 502. — Two-Element Vacuum Tube Characteristics. electron to move it toward the plate or filament will be zero. If the initial velocity of the electron carries it be- yond this point it will continue toward the plate with a rapidly increasing velocity. If the initial velocity does not carry it beyond this point the forces acting in the tube will carry it back to the filament. The current flowing through this tube will increase with increasing plate voltage until all of the electrons given off by the filament go to the plate. After this any further increase in the plate voltage will not increase the current. The change of current in the plate circuit with changing plate voltage is shown in Fig. 503 for two different temperatures of the filament. We see from these that the current is controlled directly only by plate voltage. Kadio Communication. 67 The tube described above is known as the Fleming Valve, or " two-element vacuum tube." Its check valve action may be utilized at radio frequencies as a detector. However, due to the great superiority of one of its modifi- cations, the simple valve is seldom used for a detector. The idea of introducing another means of controlling the motion of the electrons in a vacuum tube was originated by DeForest. He introduced a " grid " of conducting material between the plate and the filament, thus producing the " three-element vacuum tube," or " audion." The introduction of this conducting grid makes it possible to control the motion of the electrons in the tube independently of the plate voltage. It will readily be seen that charging this grid negatively will act on the electrons as they are given off by the filament and will drive back to the filament some of those which would otherwise have gone to the plate. This is the direct cause of a decrease in the plate current. Similarly, it will be seen that a positive charge on the grid will aid the electrons in getting away from the filament. This will cause some electrons to pass to the plate which would otherwise have returned to the filament. This in turn is the direct cause of an increase in the plate current. It is to be particularly noticed here that although the grid is positive and should apparently draw the electrons direct to itself, only a small percentage of the electrons go to the grid, while the rest pass on to the plate, because normally the plate voltage is much higher than the highest voltage reached by the grid. The few electrons which go to the grid when it is positive constitute a current flowing from grid to filament. The circuit outside of the tube, connect- ing the filament to the grid is known as the grid circuit. Due to the fact that the grid is always placed closer to the 68 Modern Theory and Practice in filament than the plate is, a change in the voltage applied to the grid will have more effect on the motion of the elec- trons in the neighborhood of the filament than an equal change in the plate voltage would have. This means that a change in the voltage applied to the grid will have more efifect on the plate current than an equal change in voltage applied to the plate would have. Stated in another way, it may be said that a small change in the plate current may be produced in either one of two ways, either by changing the plate voltage or by changing the grid voltage. To produce ■'BV' METER Fig. 503. the change in plate current by means of a change in plate voltage will require a larger change in voltage than to produce the same change by means of a variation of the grid voltage. Usually it will require from five to forty times as much change in plate voltage as in grid voltage to produce the same effect on plate current. This ratio is a constant which depends principally upon the construction of the bulb. It may be measured readily by means of the connections shown in Fig. 503. Using values of voltage suitable for the particular tube, the plate current is first changed somewhat by a definite variation of the plate volt- age Eb. The plate current Ib is then returned to its origi- Eadio Communication. 69 nal value by means of an opposite change in the grid volt- age Ec. It will be seen that the current need not be known in value as long as a means is provided to enable one to return to the original value. For this reason a galva- nometer may be used as a current measuring device without knowing what value of current is indicated by the scale readings. The change in plate voltage divided by the change in grid voltage in the above case is a measure of the effectiveness of the control of the electrons by the grid.* Typical characteristic curves showing the variation of plate current and of grid current with variations of grid voltage are shown in Pig. 504 for two widely different values of plate voltage. The plate current is the cur- rent which flows from the plate to the filament through the tube. The grid current is the current which flows from the grid to the filament through the tube. It is to be understood that while tak- ing one of these curves the plate voltage is kept constant at the value marked on the curve. It will be seen that in the case of the plate current an increase in the plate voltage shifts the curve over to the left without changing its shape appre- ciably. (This holds only over a certain range, and is not even approximately true on low voltages.) This shift causes any particular value of plate current to correspond to a lower value of grid voltage than before. In the case of the grid current there is very little change in the point at ♦This ratio, denoted by Mo, is also the maximum value of voltage amplification obtainable from the tube. 70 Modern Theory and Practice in which the curve starts from zero, but for higher plate volt- ages the grid current is smaller, due to fewer electrons escaping the increased attraction of the plate. The values of grid voltage shown above are measured from the negative terminal of the filament. In order to picture the action of the tube in terms of the elementary ideas of Ohm's law, it is convenient to con- sider the resistance of the path through the tube from plate to filament. It is seen from the preceding explanations and from the curves of Fig. 504 that with a constant plate volt- age the resistance of the tube from plate to filament varies as the grid voltage varies, being nearly infinite for negative grid voltages up to a certain point. From here on, as plus values of grid voltage are approached the resistance drops, until it finally approaches a relatively small value. This variation of the internal resistance of the tube may then be thought of ps the feature which controls the flow of the plate current. The controlling action of the grid may also be thought of as simply that of a valve in the power line from the high voltage battery in the plate circuit. The valve closes gradually as negative voltages are applied, and opens with positive voltages. A very small amount of power is sufficient to work this valve. The power controlled hy the valve may be many times as great as the power neces- sary to operate the valve. Also, if the action of the tube is limited to that portion of the Is curve (see Fig. 504) that is approximately straight, the current controlled by the valve varies directly as the voltage applied to the grid. From the preceding paragraph it follows that the vacuum tube may be used as a relay. A simple circuit for use in this manner is shown in Fig. 505. If a varying voltage Eadio Communication. 71 from any source is applied between the grid and filament, the variations which take place in the plate current will be similar to the applied voltage. Such a case is presented in Fig. 506. In this figure Ec is the voltage applied between the grid and filament. Is is the corresponding flow of cur- rent in the plate circuit, and 7^ is the current flowing from the secondary of the transformer to the load. 7^ is due, of course, to the E. II. F. induced in the secondary of the transformer by the variations of flux through this secondary when the primary current varies. It is to be noticed that although, due to the nature of the tube, 7^ varies but never ' ? as l-OAD Fig. 506. reverses, /^ is an alternating current similar to the voltage applied to the grid. As ordinarily operated, the energy available from the variations in the plate current is many times as great as the energy necessary to produce the variations in the grid voltage. When used to utilize this reproduction of a current with increased energy in the out- put, the tube is called an amplifier. The small amount of energy that is used to produce the voltage applied to the grid is used up in the tube and the circuits connected with it. The energy which is available in the plate circuit is controlled hy the grid but comes entirely from the high voltage battery in the plate circuit. 73 Modern Theory and Practice in This control of a relatively large amount of energy by a small amount of energy is made use of in receiving appara- tus when the signal to be received is too weak to produce reasonable signals in the telephones. In this case the energy which would otherwise go to produce a sound in the telephones is used to vary the grid voltage of a three-element vacuum tube. The changes in the grid voltage control a relatively large amount of energy from the high voltage battery in the plate circuit of the tube, and a telephone placed in this plate circuit will have signals reproduced in it appreciably louder than the signals in a telephone con- nected directly to the receiving apparatus. If the signal is not loud enough after amplification by one tube, the out- put of the first amplifier tube may be used to vary the grid voltage of a second three-element vacuum tube, the output of the .second tube may be used to vary the grid voltage of a third, and so on. There are practical limits to this which will be discussed later. The simplest way to transfer the energy from a detector to an amplifier tube, or from one amplifier tube to another, is probably by means of a transformer. The primary of this transformer carries the current from the source of the weak signal. The changes in this current cause changes in the flux through the winding, and this changing flux cutting the secondary winding produces a voltage in it similar to the changes in the primary. As the secondary winding has several times as many turns as the primary, the voltage produced in it is larger than the primary change of voltage. This larger voltage is applied directly between the grid and filament of the amplifier tube to control its output. Eadio Communication. 73 TO SOURCE /"* 1 OF VARYfNG-J (^O I VOLTAGE XJir- f^^ J Fig. 507. Another way to transfer the energy from one tube to the next makes use of an impedance * in the plate circuit of the first tube. Assuming the plate current to be constant, f consider a circuit such as that shown in Fig. 507. As the voltage applied to the grid of the first tube varies, the resistance of this tube varies. With a constant current this variation in re- sistance of the tube causes a corresponding variation in the voltage between the plate and filament of the first tube. This variation in voltage is handed on to the grid of the next tube through the condenser G and the direct connection between the filaments. The condenser C permits the passage of the variations in voltage and prevents undesirable D. C. volt- ages from reaching the grid of the second tube. Amplifiers have been discussed above in connection with weak radio signals. Of course, their use is not necessarily confined to telephone signals, nor is their use confined to audio frequencies. The principles outlined above may be applied directly to radio frequency, amplification, either to amplify the received radio signals before applying them to the detector, or to amplify signals which are to be transmitted. * This impedance is frequently a choke coil, but may be a simple resistance or any combination of resistance, inductance, and capacitance which will permit the direct current to pass but will tend to prevent the current from changing at the frequency or frequencies which are to be amplified. t This Is never quite true, but the same principle applies as long as the current is not allowed to change freely. 74 Modern Theory and Practice in Consider the circuit of Fig. 508 where AC is an alter- nating current generator which is supplying power to the input circuit (between grid and filament) of an amplifying tube. Assume that the power supplied to the tube is 10 micro- watts. The output available at the secondary of the trans- former T might be 1000 micro- watts. Assuming this to be the case, apply to the secondary of piq. 508. the transformer a load which will take 990 microwatts. Now it should be evident that if it is possible to take the 10 microwatts that are left over and supply them to the input circuit of the tube in place of the generator, the generator may be removed, and the tube will continue to generate alternating current without the assistance of an outside alternator. In order to do this it is found that with all ordinary apparatus the following requirements must be met: (1) The output of the tube must be greater than the input, that is, the tube must be an amplifier. (8) There must be an oscillating circuit. (This controls the frequency of the current generated and the phase of the voltage applied to the grid.) (3) Variations occurring in the plate current must transfer part of their energy to the oscillating circuit. (4) Oscillations occurring in the oscillating circuit must transfer part of their energy to the input circuit of the tube, with the proper phase relation, in such a way as to maintain the variations in the plate current. Note. — A valuable cheek on whether a circuit should satisfy the last part of requirement (4) is the fact that if the circuit is correctly arranged, an oscillation occurring in the oscillating circuit will tend to make the voltage of the grid negative with respect to the filament at the time that it tends to make the Eadio Communication. 75 The number of combinations differing at least slightly from one another, that ma;y be used to satisfy requirements (2), (3), and (4) are almost unlimited in number. It is found, however, that the elementary means of satisfying the separate requirements are very few in number, so that an understanding of this small number of possibilities will make it possible to trace out the action of more complicated circuits. As has already been explained, the elementary oscillating circuit consists of inductance and capacitance in series. The inductance may be made up of several parts connected together, and the capacitance may be made up of several condensers, these parts being in series, in parallel, or in any other combination. Another simple variation of the oscillating circuit is of the type shown in Fig. 509. Here inductance and capacitance alternate instead of being put all together. They still form a simple series circuit, and form but a slight modification of the simplest form of oscillating circuit. In order to satisfy condition (3) the oscil- lating circuit need not be one of the above types. It may consist of two or more simple oscillating circuits coupled to each other by any one of the systems of coupling explained in Chapter IV. In the case of several circuits coupled together, it has been explained that more than one frequency of oscillation may appear. Ordinarily only one frequency will be gener- ated. This will be the one which tends to transfer the greatest per cent of the output back to the input with a phase relation such as to maintain the output. Three simple means of satisfying requirement (3) will be considered. The first, Fig. 510a, shows the oscillating 76 Modern Theoey and Peactice in circuit connected directly into the plate circuit, so that all of the plate current passes through the oscillating circuit. Fig. 510b differs from 510a only in that the direct current component of the plate current does not pass through the oscillating circuit. The choke coil permits the passage of the direct current but presents a very high impedance to the oscillations, so that the oscillations pass through the oscillating circuit on their way to the filament. Condenser C is used to prevent a short-circuit of the B battery. Fig. 510c shows no direct connection between the plate circuit CHOKE COIL OR TELEPHOfiE Fig. 510a. Method (3a). Fig. 510b. Method (3b). Fig. 510o. Method (3c). and the oscillating circuit, the transfer of energy taking place entirely by transformer action, due to the mutual inductance of the coils. Neglecting the direct current component of the grid current, to satisfy requirement (4) we have the two arrange- ments shown in Figs. 511a and 511b. Fig. 511a shows the case of direct connection of the oscillating circuit to the grid. This results in a voltage being applied to the grid equal to the voltage produced in the oscillating circuit by the oscillations. Fig. 511b shows the oscillating circuit supplying energy to the input circuit of the tube by trans- former action. Eadio Communication. 7r It may be seen by referring to the characteristic curves of the three-element vacuum tube (see Fig. 504) that when- ever the grid becomes positive with respect to the filament a current flows in the grid circuit. This current consists of electrons, which are attracted to the grid whenever it is positive. If a condenser is placed in series with the grid circuit, these electrons which are attracted to the grid will be unable to leave the grid and will therefore form a nega- tive charge on the grid. Frequently, in a vacuum tube generating circuit, the accumulation of electrons on the grid- will produce a negative charge on the grid great Fig. 511a. Method (4a). Pig. 511b. Method (4b). enough to practically stop the plate current. This, in turn, will stop the generating action of the circuit, for if there is HO input there can be no output. In this case it becomes necessary to provide a path by which the electrons may escape from the grid. This path may be any sort of an impedance which will pass direct current and offer high impedance to the frequency being generated. It consists frequently of a pure resistance of high value. This path is known as a "grid leak." Leakage which is unavoidably present * between the grid and filament assists the action * Some electrons will be carried from the grid to the filament through the tube by the molecules of gas which are left in the tube, due to the impossibility of obtaining an absolutely perfect vacuum. 78 MoDEEN Theory and Practice in of the grid leak, and sometimes malces the use of a grid leak, as a separate piece of apparatus, unnecessary. A few typical combinations will serve to show the applica- tion of the preceding to actual generating circuits. Fig. 512 shows a combination of (3c) and (4a). Fig. 513 shows Pig. 512. l&d CHOKE COIL OR TELEPHONE c. a similar combination of (3a) and (4b). These two are so readily seen that their make-up does not require expla- nation. Fig. 514 shows a com- bination of (3b) and (4a). In this case it becomes essential to use a grid condenser Cg to keep the high voltage D. 0. away from the grid. The condenser Ci completes the grid to filament circuit of (4a). It is seen that there cannot be a metallic con- nection at this point, as this would cause a short-circuit of the B battery. It should bo noticed thfit the condenser C^ at this point is by-passing part of the oscillations around the oscillating circuit so that if it is made large the oscillations produced in the oscillating circuit will not be strong enough to maintain the generating action. If the capacitance at 6\ becomes very small it practically disconnects the grid from the fila- FiG. 514. Radio Commuxication. 79 ment. If the capacitance at C^ is made large it will inter- fere with the action of the circuit by preventing the voltage of the grid from becoming appreciably different from that of the filament. If C, is made very small it tends to stop the generating action of the circuit by not permitting the transfer of energy from the plate circuit to the oscillatmg circuit. Due to the transfer of the energy from one part of the circuit to another through condensers, this circuit may be said to have capacitive coupling. It is probably well to point out here that the capacitance at Cj and at C, is not all due to the condensers located at those points. The capacitances between the elements, inside the tube, and that between the connecting wires, outside the tube, frequently are not negligible. If the decrement of the oscillating cir- cuit in the preceding circuit is low (that is to say, if thc- power losses in it are relatively small) the circuit of Fig. 514: will have sufficient capacitance at Cj and C, to generate without having condensers placed at these points. If only the condenser C. is omitted, the circuit is Derorest'.< " Ultraudion " circuit. Fig. 515 shows a combination of (3a) and (4a), both slightly modi- fied from their original form. In this case the inductance of the oscil- lating circuit is split into two parts as is frequently found convenient. There is direct connection of the plate circuit to one of these parts, and direct connection of the grid „ g,r circuit to the other.* The capaci- tance C may be only that between the elements of the tube * This circuit places the B battery in the oscillating circuit. The effects of the resistance of this battery may to a large extent be avoided by shunting the radio frequency current around the battery, through a condenser of several microfarads capacitance. 80 MoDEEN Theoet and Peactice in and connections to them. Such a circuit as this latter is found in some amplifiers. In this case the constants of the circuit may be such as to give rise to audio frequency. If the circuit acts as a generator when intended for use as an amplifier, it becomes useless as an amplifier, due to the loud " howl " in the telephones. The plate circuit induc- tance in this case may be that of the telephones themselves. This circuit may also be thought of as having capacitive coupling due to the fact that energy applied to Lp passes to Lg through C Eadio Communication. 81 CHAPTEE VI. SPARK TRANSMITTING SETS. In Chapter III we studied the phenomena of oscillating circuits and the theory of wave propagation through the ether. It now remains to study in detail the necessary apparatus and the functions of circuits required to produce in an antenna oscillations of sufficient power to transmit signals to great distances. So far as practicable the de- scriptions cover apparatus used in present or proposed naval installations. Fig. 601. — Spark Transmitter. The end sought in a transmitter is the efficient produc- tion in the antenna of high power oscillations. In the earlier sets the spark gap was introduced directly into the antenna and the leads from the secondary of a high voltage induction coil connected to the two electrodes of the gap. This arrangement had many disadvantages and was soon replaced by the two-circuit system of modern sets. In this 6 83 Modern Theory and Practice in later arrangement oscillations are produced in the closed, or sparkj circuit, and the antenna is caused to oscillate, with the same, or approximately the same, frequency, by induc- tive connection. This spark circuit, as previously described in Chapter III, consists of condenser, inductance and spark gap. It is well for the student to bear in mind that the spark gap is not an essential part of an oscillating circuit. Its purpose as used here is to prevent any flow of current other than into the condenser until the condenser has been charged to the desired potential. Practically all naval installations use A. C. to charge the condenser, but D. C. may be used, and a brief description of a set of this kind, as used by the Marconi Company in transatlantic work, will be given later. " Tone " transmission, that is, trans- mission with a fixed and regular number of wave trains per second is more easily obtained with A. C. than with D. C. supply. Fig. 601 is a line diagram of a modern spark trans- mitting set using an alternating current power supply. The A. C. generator, G, is usually a 500-cycle, 110 to 300-volt machine, driven by A. C. or D. C. motor or by turbine. The reason for using 500 cycles will presently be shown. B is an iron cored inductance used as a reactance regu- lator, to regulate the power supplied by the generator. P S is the power transformer, iron cored, either open or closed core type, with primary winding of few turns of heavy wire and secondary of many turns of fine wire, made up in sections. The ratio of primary to secondary voltage is 110 volts to 30,000 volts, or greater. The condenser, C, consists of Leyden jars in series — parallel arrangement ; or, in the latest sets, of a mica eon- denser of Dubilier or similar design. Eadio Communication. 83 S G is the spark gap, open, rotary or quenched type. The relative advantages of the different types will be studied. L and P' are spark circuit inductances. S' and L' are antenna circuit inductances. All induc- tances in the oscillating circuits are air cored and variable. They may be wound on cylindrical frames or in the shape of expanding spirals. In this latter form they are called " pan cake " coils. P' and S' form the oscillation transformer, the inductive connection between the closed oscillating circuit and the antenna circuit. 4 is a hot-wire ammeter connected in the ground lead of the antenna. K is the sending key placed in the low voltage circuit. The circuits to be considered are as follows : Primary power circuit, consisting of the armature of the generator, reactance regulator, sending key, primary of the power transformer, and the necessary leads. Secondary power circuit, consisting of the .secondary winding of the power transformer, the condenser and leads. Spark circuit, or closed oscillating circuit, made up of the condenser, spark gap, two or more inductance coils in series, with leads, made of braided wire of low high-frequency resistance, and as short as possible. Antenna circuit, consisting of the antenna, loading coil, L', coupling coil, S\ hot-wire ammeter and ground connection. Briefly, the operation of the set is as follows : When send- ing key K is closed the generator sends low voltage current. 84 MoDEEN Theory and Practice in at 500 cycles, to the power transfoiiner. The condenser, which is connected across the secondary of this transformer, has impressed upon it a voltage depending upon the ratio of turns of primary and secondary windings. In modern naval sets this voltage is usually 13,500. The voltage curve of the condenser is approximately sinusoidal and the spark gap is adjusted to break down only at the highest voltage of each alternation. Hence for each cycle of the generator we get two oscillatory discharges of the condenser through the spark circuit, or 1000 wave trains sent out per second. It has been found by experiment that this frequency is most suitable for reception in the case of spark, signals, due in V^^ Fig. 602. part to more satisfactory operation of the telephone dia- phragm, and in part to the ease of detection by the ear of a note of this frequency. Five hundred cycle generators are now standard equipment for radio sets on board' ship. Due to the high inductance of the secondary of the power transformer, and, in high-power sets, to additional choke coils placed between the condenser and the transformer, all radio frequency current is choked off from the power circuit and the spark circuit oscillates, when the gap breaks down, as though entirely disconnected from the power circuit. The voltage of the condenser during charge and when oscil- lations are taking place may be represented, roughly, by curve in Fig. 602. This curve does not apply, as will be explained, when a resonance transformer is used. By Eadio Communication. 85 " resonance transformer " we mean that the primary and secondary power circuits together with the armature wind- ings are tuned to the generator frequency. The two circuits of any transformer may be reduced to an equivalent simple, or series, circuit. If in this case, in the reduced circuit, the capacitive reactance, at the generator frequency, is made equal to the inductive reactance, they balance each other and the impedance becomes equal to the resistance alone. It is readily seen that with the circuits thus tuned to resonance the current from the generator is in phase with the gen- erated voltage, and is increased, due to less impedance, and the voltage on the condenser is correspondingly increased. Several advantages accrue from the use of a resonant or nearly resonant circuit, in addition to that above men- tioned. In any spark circuit when the gap breaks down it becomes conductive, and for an instant practically short- circuits the secondary of the transformer, thus drawing a heavy current from the generator. With the circuits tuned to resonance any change in inductance or capacity, such as that resulting from short-circuiting the secondary, greatly increases the total impedance, thus tending to prevent this sudden rush of current. Another advantage is that the voltage obtainable on the-condenser can be increased with a correspondingly lower spark frequency. Suppose the length of the gap to be increased until the breakdown voltage is slightly above that obtainable with one spark per alterna- tion. The voltage across the condenser will rise with each successive alternation, that is, build up, as shown in the figure, until the gap breaks down. By proper adjustment we can obtain one spark for each two, three, or four alter- nations as desired. The voltage obtained may be several times what it would be with one spark each alternation. 86 MODEKN ThEOEY AND PRACTICE IN It is to be noted that with resonant circuits the maximum voltage on the condenser is reached at the instant that the generated voltage and the current are both at zero value Fig. 603. — Voltage Curve with Resonance Transformer. (see Fig. 604). The time of one complete oscillatory dis- charge of the condenser is small compared to the charging interval of the condenser, hence the voltage across the eon- denser after one or more reversals of the oscillating current VOLTAGE ACROSS CONDENSER GENERATED VOLTAGE CURRENT PiQ. 604. is practically that due to the oscillating current alone and is little affected by the charging current from the gener- ator. This action may assist to a certain extent in the Radio Communication. 8? quenching of the spark. It can be proved mathematically that with one spark per alternation, the voltage across the condenser, when a resonance transformer is used, may be as much as 1.5 times the product of the generated voltage multiplied by the transformer ratio. When using a quenched gap it has been found advan- tageous to detune the audio frequency circuits slightly from the resonant setting. This is accomplished by using a capacity in the secondary circuit about 15 per cent greater than required for resonance. The antenna circuit is inductively connected to the spark circuit, hence has induced in it oscillations of the same, or approximately the same, frequency. Under coupled cir- cuits. Chapter IV, we learned that under certain conditions the wave produced in the secondary of two coupled circuits might be a composite, or broad wave, that is, made up of two waves, one of slightly longer and the other of slightly shorter length than that to which the circuits are tuned. As the coupling is decreased these two waves approach each other in length until finally with extremely loose coup- ling a single or pure wave is produced whose length is the same as that to which the two circuits are tuned. As the coupling is decreased the energy that is transferred to the open, or radiating, circuit is decreased and also the transfer takes place less rapidly. In order to secure a pure wave and at the same time use closer coupling, with increased energy transfer, the Marconi Company devised a rotary spark gap which prevented to a large extent the re-transfer of energy from the antenna circuit back to the closed circuit. In this form of gap the spark takes place between the ends of radial spokes on a revolving wheel, and a sector mounted as shown in the figure. The distance between the sector 88 Modern Theory and Practice in and the ends of the spokes is adjustable so that the spark length can be varied. In the non-synchronous type tlie speed of the wheel is independent of the speed of the gener- ator. In the synchronous type the wheel is mounted on an extension of the generator .shaft and so placed that at or near the instant of maximum E. M. P. a spoke is opposite a sparking sector. The re-transfer of energy is prevented c^^^s::^ Fig. 605. — Quenched Spark Gap. by the lengthening of the gap, with increase of resistance, as the spoke passes the sparking surface. This form of gap has been replaced to a large extent by the quenched gap invented by Wein (see Pig. 605). In the quenched gap we have, instead of a single long spark, a number of shorter sparks in series. The details of construction are shown in the iigure. The object of the radial, fin-shaped extension to the plate is to provide a greater cooling sur- Eadio Communication. 89 face. To assist in cooling a current of air from a small fan is sometimes played on the gap while in operation. The length of gap in use may be varied by tapping ofi across the desired number of plates, or if all plates are in series, by short-circuiting one or more plates by metal clips. "fc> Fig. 605a. — Rotary Spark Gap. The surface of each plate is accurately machined so that the distance between sparking surfaces when plates are clamped together is uniform over the entire area. The length of gap remains constant because if erosion takes place HOLDER - MICA RiNGS ^^ Fig. 605b. — Quenched Spark Gap (Wein). in one spot the spark will jump next time between non- eroded points which are nearer together. The sparking surfaces are usually silvered. The distance pieces or washers, between the plates are made of best quality mica of uniform thickness, and when compressed by the adjust- 90 Modern Theory and Practice in ing bolts at the ends, form air-tight chambers in which the spark takes place. The advantage of this form of gap, the prevention of a re-transfer of energy from the open to the closed circuit, is due to the more rapid de-ionization of the air or gases in the gap, thus increasing very greatly the resistance as the current dies out. As soon as the current reaches zero value the resistance is so great that the reaction from the open circuit is insufficient to again break down the gap. This leaves the open circuit free to oscillate in its own natural period. By the use of this form of gap close Fig. 606. — Current Curves of Closed and Radiating Circuits Using Quenched Spark Gap. coupling can be used, and a larger percentage of the energy in the closed circuit is transferred to the radiating circuit, with no re-transfer and consequent lo.ss. The percentage of energy thus transferred to the open circuit is seldom, in any set, more than 50 per cent of that delivered by the generator. These unavoidable losses take place in the trans- former, condensers and leads, and as light, heat and noise in the spark gap. The energy that is transferred to the open circuit is expended, part, in overcoming the ohmic resistance of the antenna, part, in brush and corona dis- charges, part, as loss due to imperfect dielectric and the remainder in the form of electromagnetic radiation. Eadio Communication. 91 The energy radiated from a circuit, such as an antenna, is proportional to the square of the current flowing in the circuit and is thus analogous to the energy dissipation as heat in a conductor. Eadiation, therefore, increases the effective or equivalent resistance by a certain amount and this increase of resistance is conveniently called " radiation resistance." It is large enough to be appreciable only in cir- cuits of the open or radiating type or in closed circuits of large area and very high frequency. It can be computed for an antenna in terms of the effective height and the length of radiated wave. Direct Current Spark Transmitters. As previously noted, D. C. may be used to charge the condensers of the oscillating circuit. The connections are Fig. 607. — Spark Transmitter, with D. C. Supply. very simple, as shown in Fig. 607. The terminals of the condenser are connected to the D. C. generator, or storage battery, through resistance, R, and inductance, L. The inductance is usually air cored. If iron is used it is desirable to keep the magnetization well above the satura- tion point. The resistance in the supply line should prefer- ably have a rapidly rising characteristic. Metal filament lamps may be used for this purpose. 92 MoDEEN Theory and Practice in If the gap length and the supply current are properly adjusted instead of an arc resulting across the spark gap we get an oscillatory discharge. The D. C. generator or storage battery cliarges the condenser until the potential is sufficiently high to break down the gap. The energy stored in the condenser is dissipated by heat and transfer to the radiating circuit, and potential of condenser falls practically to zero. During this oscillation the supply current is maintained practically constant by the induc- tance of the supply line, and due to this same inductance / TO D.C. eOURCE \_ > -p-VWWA ^ V 't"t**^ '^'^^ RESISTANCE L mmm^ -WW — ipsum^ DRIVINQ_ MOTOR REVOLVING COMMUTATOR Fig. 608. — Line Diagram of a Motor Buzzer Set. the high frequency currents are choked off from the D. C. line. Since the time of one oscillatory discharge is small compared to the time required to charge the condenser to the break-down potential the discharge current is almost a pure damped oscillation. Spark sets of this nature have been used by the Marconi Company in transatlantic com- munication between Clifden, Ireland, and Glace Bay. Some of our latest ships are equipped with a modification of this type, called " motor buzzers," which are small power sets and so designed that the oscillating circuit is completed mechanically by a revolving commutator, and Eadio Communication. 93 at the same instant the D. C. supply is shunted from the condenser. In this type the frequency of the discharge is regulated by the speed of the revolving commutator. Antennae and Fittings. The type of antenna depends upon the location and pur- pose of the station. In ship installations the inverted L or T-shaped antennae are used almost exclusively. Long-range shore stations use the umbrella' type or some fiat top, large capacity, type. A few of the types in use are shown in Fig. 609. The relative advantages of the various types need not be considered here. INVERTED L T-SHAPED UMBRELLA TYPE Fig. 609. The wire used is generally stranded phosphor bronze, of low high-frequency resistance. The free or high poten- tial end must be well insulated from the masts or towers. Bare wire is used on account of added weight that would result from insulation. The "rat tail," or lead-in, from the antenna to the sending set may be insulated wire and, where it passes through walls or decks, must be well insu- lated to prevent short-circuits or brush discharges. An antenna radiates best at a frequency determined by the natural inductance and capacity of the aerial and the lead-in. Since it is impracticable to construct an antenna 94 Modern Theory and Practice in of sufficient size for other than short wave lengths, addi- tional inductance coils, besides the coupling coil, must be added in series to tune the antenna to the desired wave length. The wave length to which the antenna is tuned, by adding inductance, should not, for efficient operation, exceed four or five times the fundamental or natural wave length. The radiating circuit consists of the aerial, lead-in, ground connection, ground, and capacitive connection be- tween ground and aerial, thus forming a series circuit. It is readily seen that a good ground connection is essential in order to keep the resistance of the circuit low. Dry earth is a poor conductor, hence the ground connection must be extended far enough below the surface to insure intimate contact with moist earth. In the case of ship sets the hull furnishes an excellent current path to the con- ductive sea water. On shore the ground connection is often in the form of numerous wires placed well below the surface of the earth and radiating in all directions from the transmitter. Instead of a direct connection to earth, we may use a " counterpoise," i. e., a network of wires con- nected to the lower end of the antenna and parallel to, but insulated from the earth. The aerial and the counter- poise respectively form the two plates of a condenser, with the electrostatic field between. This is the nature of the antenna circuit in an airplane transmitter. The counter- poise may be attached to the upper surface of the wings and the antenna trailed below. The use of a counterpoise in connection with a fiat top antenna tends to give a more uniform distribution of the electric strain lines. Eadio Communication. 95 Notes on Installation and Tuning of Spark Transmitters. The efficiency of a transmitter is materially affected by the manner in which the circuits are laid out. A changing magnetic flux induces a current in any circuit or piece of metal lying wholly or in part in the field of the flux. Such induced currents in other than the regular circuits of the set represent lost energy. Eadio frequency circuits of a transmitter should be so placed with reference to bulkheads, receiver circuits, etc., as to reduce to a minimum these stray currents. Crossing leads are to be avoided. The receiver should be placed as far as practicable from the sending apparatus, and all circuits should be mounted on insulating material. Much loss of energy can be avoided by keeping all parts of the set perfectly clean. For safety all high voltage circuits, transformer ter- minals, etc., should be protected from accidental contact. Assuming the set to be installed, the following procedure is required to put it in shape for transmission. If a rotary gap is being used, the power circuits may be adjusted to resonance to the generator frequency. With a quenched gap the natural frequency of the power circuits should be slightly lower than the operating frequency. Eesonance curves can easily be plotted, and should be plotted for all sets when installed. An easy method of plotting a reso- nance curve for the power circuits is as follows : Insert an ammeter in the generator armature circuit, and, keeping the frequency constant, plot amperes in primary circuit against capacitance in the secondary circuit. The point of maximum current in the primary circuit determines the resonant .setting. A variable reactance is usually inserted in the primary circuit. This reactance regulator, as it is called, may be variable by steps or continuously, by means 96 Modern Theory and Practice in of a sliding iron core. These adjustments are usually made by the manufacturer. In ,some cases a double throw switch is provided, connecting to two taps on the reactance regu- lator, one contact giving a condition of resonance for full power, the other giving increased impedance for low power. Having made or checked the adjustments of the power circuits, the next step is to tune the oscillating or spark circuit to the desired wave lengths. To do this disconnect the antenna circuit entirely, close sending key. Examine the circuit carefully to see that there is no " sparking over," that contacts are good and that gap is operating properly. Bring a wave meter near the circuit and measure the wave length of the oscillation produced. Eelease key, vary the inductance and measure wave length again. A few trials will suffice to adjust to desired wave length. The inductance may thus be calibrated for several wave lengths and marked so that to change the wave length it is only necessary to tap off at the proper point. Next connect up the antenna circuit, with all or a part of the coupling coils in series. Press key and observe the reading of the hot-wire ammeter in antenna circuit. Vary the inductance of the antenna circuit until the reading of the ammeter is a maxi- mum. The two circuits are now tuned to the same wave length. Vary the degree of coupling between the two cir- cuits and observe the effect on current in the antenna. It will be found that there is one best degree of coupling ; any change, either increase or decrease, will decrease the current in the antenna. This degree of coupling can be obtained by varying the distance between the coupling coils or by vary- ing the number of turns of coupling coil in use. In the latter method any change in turns of coupling coils necessitates an inverse change in the loading coil to keep Eadio Communication. 97 the circuits tuned to same wave length. Antenna must be retuned by experiment aiter each change in coupling coil. Having determined the best degree of coupling and retuned the circuits to exact resonance the set is ready for operation. The foregoing method is sufficiently accurate for practical purposes. For more exact results, couple wave meter loosely to antenna circuit and measure the wave length of the radiated waves. If the wave meter indicates resonance at only one setting a pure wave is being radiated. If two points of partial resonance are found, the wave is broad. The coupling is then varied until a pure wave is secured. The readings of the wave meter indicating device, either wattmeter or current-squared meter, in the case of a broad wave, give a ratio of the energy radiation at the two wave lengths. It is not necessary to secure an absolutely pure wave. The U. S. statutes require that where a transmitter radiates energy at more than one wave length the energy radiation at the lesser of the two waves shall not exceed 10 per cent of that of the greater. Such a wave is very selective and easily received. For convenience in shifting from one wave length to another, the coupling coils and loading coils are tapped and the leads brought out to sliding contacts. Thus it is pos- sible by operating a single lever or handwheel to vary the primary inductance, the coupling, and the secondary load- ing coil inductance at the same time. The distance between the coupling coils remains fixed .and the coupling is varied by changing the number of turns in use. This is called common point coupling. Once the coils are calibrated and the leads brought to proper contacts, the entire range of wave lengths is instantly available without further adjustment. 7 98 MoDEEN Theory and Peactice in CHAPTEE VII. CONTINUOUS WAVE TRANSMITTERS. The following devices for producing approximately pure undamped oscillations will be explained. First, the radio frequency alternator; second, the radio frequency spark; third, the arc; fourth, the vacuum tube. Por mechanical reasons the inductor type of alternator 1;= used for high frequencies. There are two principal mechanical advantages of this type. One is that the re- volving parts do not carry windings. The other is that for a given number of poles on the revolving part the fre- quency produced at any speed is Just twice that which would be produced by the ordinary construction. It should be known that the inductor type of alternator is by defini- tion " an alternator in which both field and armature wind- ings are stationary, and in which masses of iron or induc- tors, by moving past the coils, alter the magnetic flux through them." This type of alternator or A. C. generator has the magnetic flux through all parts of the. magnetic circuit constantly in the same direction. The part of the flux which passes through the armature winding is varied by changing the reluctance of the magnetic circuit through this winding. This change of reluctance is, in turn, due to changing the length of an air gap in this magnetic cir- cuit. The change in the length of air gap being produced by passing the soft iron through the air gap. It will be seen that as one inductor approaches an armature coil, the flux through the coil will rise from its minimum value to a maximum, and as the inductor recedes from the coil Radio Communication. 99 the flux through it will decrease to its minimum value again. The increasing flux through the coil produces a voltage in the coil in one direction, the decreasing flux produces a voltage in the opposite direction. The result is that the passage of one inductor in front of a coil pro- duces a complete cycle of A. C. E. M. F. in that coil. In the Alexanderson alternator, which is a representative type of radio frequency alternator, the inductors are spokes on the edge of a steel disc. The steel disc and spokes in this case have a mechanical construction very nearly the same as that of the DeLaval turbine, the disc being thickened near the shaft with the object of producing a uniform distribution of the stresses and consequent maximum strength of the whole. The peripheral speed is extremely high, that is, in the neighborhood of 13 miles per minute. This very high speed in connection with a relatively large number- of poles permits the direct production of radio frequencies corresponding to the longer wave lengths. The very great rate of change of position of the parts of this alternator permits the production of a reasonable voltage with only one turn of wire per armature coil and not a very large change of flux through the coil. This alternator may be connected either directly in series with the antenna, or it may be coupled to it by means of a transformer. In either case the antenna system will be tuned to the wave length produced by the generator. Transmission of signals may be accomplished in a number of ways. A very simple way short-circuits some of the turns of the antenna loading coil. This throws the antenna out of resonance with the applied E. M. F. and thereby greatly decreases the current in the antenna. Changing the field excitation of the alter- nator is another possibility. See Chapter X for other 100 MoDEEN Theory and Peactice in methods of controlling the output of a radio frequency alternator. The production of undamped waves by means of sparks occurring at radio frequencies may be divided into three classes according to the method of controlling the time between sparks. In the first class are those types of appa- ratus in which there is no accurate timing of the sparks. In the second class are those types of apparatus in which the timing of the sparks is produced mechanically. In the third class are those types of apparatus in which the timing of the sparks is produced electrically. Consider the circuit of Fig. 701. This is a type of circuit which is J WWyTn!!n-|-oo-^ ^ used for the production of ordinary damped oscillations, using D. C. supply of power. The time between sparks with this circuit depends Fig. 701. upon the time taken to charge the condenser to a voltage at which the spark jumps the gap. As the voltage required to jump the gap varies with the degree of ionization of the gas in the gap,, the time between sparks will not be perfectly uniform. As the supply current to this circuit is increased the time between sparks becomes shorter, and with suitable gaps the current may be increased until the sparks occur at such short intervals that the current in the antenna never has a chance to die down to zero value. This produces an approximation to an undamped oscillation. Due to the slight irregularity in the time of occurrence of the sparks and the fact that there is no special relation between the number of sparks and the frequency of the undamped wave, the current in the oscillating circuit not only varies be- d± -Tfinr^ Eadio Communication. 101 tween sparks, but varies erratically, as the voltage produced in the antenna by any one spark has no fixed relation to the current already flowing in the antenna. It may assist the current, it may oppose it, or it may be out of phase with the current by any random amount. An undamped wave such as this is satisfactory for the transmission of radio telephone signals, if the sparks occur at a frequency above audibility, but it is not entirely satisfactory for autodyne (or heterodyne) reception of telegraph signals due to the fact that the erratic variations of the phase of the approxi- mately undamped wave cause a poor tone in the receiving apparatus. W ^ j>__/T!iyTO>J -^7!W!r^ Fig. 702. Evidently, any means which makes it possible to time the sparks with accuracy will make it possible to adjust the time of the sparks so that the new spark will produce a voltage always in a direction to assist the current already flowing in the antenna system, thus avoiding erratic varia- tions. Certain types of rotary gap so arranged as to permit the discharge of sparks at high frequencies accomplish this. A simplification of the idea used in the Marconi " Timed Spark " isystem is shown in Fig. 702. To produce the large number of sparks per second and utilize the full wave train produced by each, it is found desirable to use several pri- mary circuits in which sparks are produced at different 103 MoDEHN Theory and Practice ix times by rotary gaps all mounted on the same drive shaft. This difference of time is illustrated in the figure by the fact that the two rotary gaps are displaced 45 degrees from each other. The antenna current is, of course, a summation of the currents which would be produced in the antenna by the separate primary circuits. As the new damped waves in the primaries are timed so as to always assist the oscillation already existing in the antenna, this should be the sum * of the separate damped waves which the primaries tend to produce in the antenna. It is possible to arrange by this system that a new damped wave shall be started in some one of the primary circuits at the beginning of each cycle of current in the antenna sys- tem, at the beginning of every other cycle, or at any other regular sequence. The final resultant current in the antenna is a very close approximation to a pure un- damped wave, if the sparks are not spaced too far apart. Heterodyne or autodyne reception of the wave produced by this system gives a good note. Transmission of signals would seem to be most logically performed either by detun- ing the antenna or by interrupting the D. C. supply. A number of electrical means of controlling the time of the spark are available. The action of the Chaffee gap will be explained to illustrate one of these. A Chaffee gap for use on D. C. consists of two plane parallel metal faces of a few square centimeters area, one of aluminum and one of copper, separated by a few mils and operated in an atmosphere of moist hydrogen. f Al- * Not the simple arithmetical sum, but rather a square root of the sum of the squares of the separate currents, so that the energy represented would be the arithmetical sum of the energies from the separate sources. t Alcohol vapor is also suitable. Eadio CommujStication. 103 though not called a quenched spark gap, the Chaffee gap may be thought of as being the most extreme thing in' the way of a quenching gap. Due to the materials of the gap and the conditions of operation, a single spark discharge across the gap produces only one alternation or impulse of current; that is, this gap puts the spark out the first time that the instantaneous value of current reaches zero. The excitation produced by the action of a current such as this is known as impulse excitation. It corresponds roughly to putting a bell into vibration by hitting it with a hammer. One result of this impulse excitation is that the energy is transferred from the primary to the secondary in the time of one cycle or less of the current in the secondary. Another effect of this almost instantaneous quenching of the spark is that no matter how closely the secondary is coupled to the primary only the one wave length to which the secondary is tuned appears in the secondary. Furthermore, as the pri- mary does not continue to oscillate so that it might get out of step with the current which it starts in the secon- dary, the primary circuit does not need to be tuned to the secondary.* The circuit shown in Fig. 701 is suitable for the oper- ation of a Chaffee gap, the actual circuit differing from the one discussed before only in the use of the Chaffee gap and very close coiipling between the primary and secoiidary circuits. When D. C. is supplied to the condenser it charges until a voltage in the neighborhood of 500 is reached, the gap then breaks down and the discharge of the condenser produces the pulse of current in the primary. The gap * The wave length adjustment of the primary for maximum results in the secondary is very broad, but maximum results have been obtained with the primary tuned to a wave length about 70 per cent greater than the secondary wave length. 104 MoDBHN Theoet and Practice in then becomes non-conducting and the condenser starts to recharge. The voltage now acting to tend to break down the gap consists of two parts, that due directly to the quan- tity of electricity which has gone to charge the condenser, and that induced in the primary by the action of the oscil- lating current in the secondary. Naturally the gap will break down at some point of the antenna current cycle when these two voltages are acting in the same direction. If the current supplied is large enough to charge the con- denser in the time necessary for, say, not over six cycles of the antenna current, with any given supply current the gap will break down at the time the secondary current reaches a certain- definite point in its cycle. This gives a regular cycle of events in the antenna circuit, instead of the erratic action due to a circuit such as the first one explained above under sparks at radio frequency. This definite point on the. cycle at which the gap breaks down may not be reached until after from one to six (or possibly more) cycles have passed without a breakdown of the gap occurring. The number of cycles occurring between dis- charges across the gap is controlled by the rate at which the condenser is charged. The point on the secondary current cycle at which the gap breaks down is determined principally by the voltage induced in the primary by the secondary. It may be seen from the preceding that the approximate control of the time between sparks is due to the variable resistance in the supply circuit, while the accu- rate control of the time of the spark relative to the secon- dary current is controlled automatically by the induced E. M. P. Apparently, if a new impulse is given to the antenna system every other cycle, or every third cycle, the current flowing in the antenna system cannot depart appre- ciably from a pure undamped wave; Eadio Communication. 105 In classifying methods of producing oscillations it is difficult to draw any sharp line between arcs and radio frequent sparks. For this reason, classifications of these two methods usually overlap to some extent. Three cycles of operation of an arc used for the generation of oscillations of radio frequency are usually recognized. The first may be said to be purely arc, as it is the type in which the cur- rent through the arc is always in the same direction and never drops to zero value. The second differs from the first in that the arc current does drop to zero value. In the third the current not only drops to zero value but also reverses. This third type of operation is identically the same /as the first cycle of operation explained under radio frequency sparks above. The operation of the first of the above cycles will not be explained as it appears to be incapable of producing large powers at radio frequencies. The second type of cycle is the one on which all common arc transmitters operate. Fig. 703 shows an ele- mentary diagram of a circuit for the productipn of un- damped oscillations. The coil Lj is an inductance of large value (many henries). This coil produces a moder- ately strong magnetic field in the gap between the electrodes of the arc and also acts as a choke coil, thereby keeping the current supplied by the generator almost constant.* Consider that the arc has * Special care is taken in designing this coil to keep the effective value of the distributed capacitance low so that it may act as a radio frequency choke. ismJ Fig. 703. 106 Modern Theory and Practice ix been in operation so that the conditions are those corre- sponding to steady operation. If by any means the arc is extinguished, the current supply all goes to the relatively small capacitance Cj the voltage across the gap rises very rapidly and the point is soon reached where the voltage is liigh enough to jump the gap between the electrodes. The supply current now transfers itself from the condenser circuit to the circuit through the arc, and the condenser starts to discharge through the arc at a rate determined principally by tlie small inductance L., (a few millihenries). The condenser current adds to the supply current flowing through the arc thus producing an instantaneous value of current through the arc considerably greater than the sup- ply current. Due to the inertia effect of L.^, the current through the circuit CL^ does not stop when the condenser is discharged. It continues until the condenser is re- charged in tile reverse direction almost as highly as the original charge, and then starts to discliarge in the oppo- site direction. It must be noticed that at this point in the cycle the current through the arc is the difference between the supply current and the current in the oscillating cir- cuit formed by C and L,. For this reason as the current in the oscillating circuit rises toward its maximum value, the current acro^ss the arc decreases. When the arc current approaches zero the resistance of the arc becomes so high that the arc goes out entirely. This completes one cycle of the generation of the undamped wave.* For ordinary * The action given in this elementary explanation is supple- mented by the change of resistance of the arc with change of current during the time when the current is flowing. The resistance of the arc being much higher for small currents than it is for large currents. Eadio Communication. 107 operation the time during which the arc is out is only a small portion of the total time for the cycle. The approxi- mate nature of the current through the arc, and through the oscillating circuit is shown in Figs. 704a and 704b respectively. The point A in Pigs. 704a and 704b is the point at which the voltage becomes high enough to jump between the elec- trodes. During the time from A to B the supply current is being transferred from i, to the arc. From 5 to C the con- denser is discharging through the arc ivith the supply cur- C D' E Fig. 704a. rent. At C the current through the oscillating circuit reverses, and from C to D the effect is that of the supply cur- rent and oscillating circuit current trying to pass through the arc in opposite directions. During this period the supply current is transferred from the arc to L.,. At D the arc cur- rent reaches zero and tends to reverse, but is prevented from reversing by the arc going out. Prom D to E there is no current in the arc, the entire supply current going directly through L^ to charge condenser C. At E the voltage between the electrodes becomes high enough to jump the gap, and the preceding cycle starts again. The dotted line in Pig. 704b shows the current which would flow if the first oscil- ]08 Modern Theory and Practice in lation continued without interruption. Evidently the fre- quency of this approximately undamped wave will be at least slightly lower than the frequency determined by the L and C of the oscillating circuit. As this difference is due to the time during which the arc is out, and this time may vary with length of arc, etc., the frequency generated by such a system may not be^ perfectly constant. To enable the arc to follow at radio frequencies the cycle of operation outlined above, several special provisions are necessary. The electrodes used for the arc consist of an ordinary solid carbon for the negative, and a water-cooled copper piece for the positive. The arc is .surrounded by an atmosphere consisting largely of cool hydrogen,* and is operated in a magnetic field. The cool copper electrode and hydrogen atmosphere both make the resistance of the are very high for low currents (and susceptible to rapid changes) and consequently tend to put the arc out alto- gether whenever the current reaches a low value. The mag- netic field is placed so as to deflect the arc from the straight line between the electrodes thus increasing the length and resistance of the arc, and at the same time tending to blow it out altogether. Hydrogen is also more susceptible to very rapid changes of ionization (upon which the con- ductivity of the arc depends) than any other gas. The combination of these features permits the construction of * The chamber in which the arc is enclosed is water jacketed, thus cooling the gas in which the arc operates, and to a certain extent cooling the carbon electrode. The gas is usually ob- tained by vaporizing alcohol or kerosene. Kerosene when de- composed by heat gives principally hydrogen and soot. Alcohol when decomposed by heat gives hydrogen, some carbon dioxide, and small quantities of soot. Eadio Communication. 109 arcs which will generate nearly pure undamped oscillations at frequencies at least as high as half a million cycles per second. The arc as described above is known as the Poulsen arc. In order to minimize the variations of fre- quency due to irregularities in the action of the arc the carbon electrode is rotated about its axis at a very low speed (less than one R. P. M.), so that it burns evenly. As the wear on the copper electrode is extremely slow no such precaution is needed for it. Fig. 705 shows the Poulsen arc applied to an antenna to produce undamped oscilla- tions. This circuit is the same as that of f . ■, Fig. 703, except for substituting the capaci- 2 °™o tance of the antenna to ground for the con- denser C. Signalling with an undamped system such as this is accomplished in either one of two general ways. That is, either by changing the wave length transmitted, or by actually stop- ping and reigniting the arc for each dot and dash. The simplest way to change the wave length transmitted is to change the amount of inductance in the loading coil in the antenna. This in turn is most simply accomplished by short-circuiting a few turns of the loading coil. The effective inductance of the loading coil is, of course, de- creased by this. The transmitted wave length is then cor- respondingly decreased. For high power sets the amount of current and voltage to be handled at the contacts which short-circuit the turns is so very high that another means of obtaining the same result is to be preferred. A closed loop placed anywhere in the magnetic field of the loading coil, so that some of the flux from the loading S 110 MoDERX Theoky and Pkactice in coil links with it, will have currents producefl in it which tend to prevent the setting up of the magnetic field. This will reduce the total flux cutting the loading coil and will, therefore, reduce its effective inductance. A loop such as this may be made of such size that the current and voltage in it are not excessive for one pair of contacts to handle. More separate loops, each having its own contacts, can then be added until the desired decrease in the effective value of the inductance of the loading coil is produced. This produces the same effects on the antenna as short-circuiting the turns of the loading coil, but has the advantage of dividing the current and voltage to be handled among a large number of contacts. A large number of contacts has also another advantage. If out of, say, 50 independent sets of contacts two or three fail to operate, the change in the results on the transmission does not amount to much, and these contacts not being connected to the main circuit may be repaired without stopping transmission. In any case, the contacts would be operated by means of solenoids which would be either directly or indirectly controlled by a telegraph key. Pig. ^06 shows a circuit used for extin- guishing and reigniting the arc. The direction of the magnetic field in this figure is such as to drive the arc toward the lower right-hand corner of the figure. The resistance B is of such value that when connected in series with the large inductance the combination will draw a cur- rent equal to the normal cur- FiG. 706. Radio Communication. Ill rent for the arc. When it is desired to stop the arc the copper contact marked " Auxiliary " is slid in until it strikes the positive electrode, thus putting the resistance E in parallel with the arc. To start the arc again within a short time, all that is necessary is to pull out the auxiliary contact. The instant that the auxiliary contact leaves the main electrode an arc is formed between the two. The magnetic field blows this arc toward the lower right of the figure and against the carbon. If the auxiliary gap is not too short, the arc is again established between car]5on and copper, and normal operation of the arc is resumed, with no current in the auxiliary circuit. The motion of this aux- iliary electrode is confroUed by means of a solenoid which is controlled by a telegraph key. This system is probably desirable only on low power arcs. To start a Poulsen arc it is necessary to put a resistance in series with the arc and choke coil to prevent excessive current when the arc is struck. The negative electrode is then moved into direct contact with the positive electrode. In this condition the flow of current is limited by the added resistance and the resistance of the choke coil (which latter is usually large enough so that no other resistance is needed during normal operation). When the electrodes are drawn apart an arc will be established if a suitable sepa- ration of the electrodes is used. The extra resistance can then be short-circuited leaving the arc in normal operation. The best length of arc is judged by the current produced in the antenna system. After a steady operating condition has been established, the arc can be extinguished and reignited by means of the auxiliary electrode, if this is provided. 112 MoDEEN Theory akd Phactice in As the generation of undamped oscillations by means of vacuum tube circuits has been discussed in some detail, the discussion at this point will be limited to the applica- tion of the generating circuits to antenna systems. The applications of vacuum tube generating circuits to antenna systems may for convenience be divided into three classes. In the first class are those types of circuit in Vifhich the antenna is connected in place of one of the condensers in an oscillating circuit of the generating system. This may be "thought of as direct connection. In the second class are those types of circuit in which the antenna is coupled to an oscillating circuit of the generating system. This is usually connection by transformer action. In the third class are those types of circuit in which the output of the generating circuit is passed through one or more steps of radio frequency amplification before being applied to the antenna. In considering these various types it should be kept in mind that the resistance of any particular antenna at a given wave length has a fixed value which cannot be changed without changing the antenna circuit. If a gener- ating circuit is capable of supplying a certain power P to an antenna, the current to be produced in the antenna, or the voltage to be applied to the antenna, must be capable of determination from the expressions P — PR = E''/R, where / is the current in the antenna, E is the voltage applied to the antenna, and B is the total effective resistance of the antenna (including radiation resistance). From these ex- pressions it is seen that under the given conditions, I and E have but one value, I—VPJB and E=^/PR. If it is Eadio Communication. 113 Fig. 707. attempted to make E or I different from these values, the power transferred to the antenna will be less than P. In using direct connection to antenna, to get the proper volt- age applied to the antenna, it is customary to make use of an autotransformer action as illus- trated in Fig. 707. The amount of inductance connected in series with the antenna depends, of course, upon the wave length to which the antenna is to be tuned. If the A. C. voltage produced by the generating circuit is greater than required to give the maximum results in the antenna the connection would be as in Fig. 707. If the antenna happened to need more voltage than the tube circuit pro- duced, the tap from the plate would be moved to the other side of the antenna tap, thereby giving a step-up action to the autotransformer (as may easily be checked by noting the number of turns in the antenna circuit, and in the plate circuit respectively). The tap from the secondary to the grid is also shown variable so that the most suitable A. C. voltage may be applied to the grid. In using a transformer connection, a circuit such as Fig. 708 might be utilized. Evidently, with transformer action the voltage applied to the antenna is very easily adjusted, as it may be changed not only by changing the turns in the coils, but also by changing the coupling between the coils. Fig. 709 presents a simple circuit for the use of a radio frequency amplifier to apply the output of a generating 8 114 Modern Theoey and Practice in circuit to an antenna. The tube at the right, in this case, would be one of much larger power capacity than the gen- erating tube. Circuits of the first type described above have the ad- vantage of simplicity, and having a minimum amount of conductors in which current is to be set up, tend to have a minimum amount of losses and consequently maximum efficiency. If anything happens to the antenna to change its constants, the power output of the .system will not be greatly affected, but the wave length will be. Fig. 708. Fig. 709. The second type is almost equivalent to the first, but is slightly more complicated, and tends to be correspondingly less efficient. The third type of circuit has as its principal peculiarity that the tuning of the antenna has no appreciable effect upon the wave length transmitted. If the antenna is not tuned for the wave length to be transmitted, the power out- put is decreased, but this does not affect the wave length. This may be of importance, particularly under military conditions, where many slightly different wave lengths are being used simultaneously. Poor work in setting up the Eadio Communication. 11 j antenna and tuning it, or an accident to the antenna after setting it up will not prevent communication with a receiv- ing station waiting to receive on the wave length to which the generating circuit is tuned (unless, of course, the dis- tance approaches the maximum that the set can transmit) . In wireless telephone apparatus this set may have another type of advantage. A wireless telephone circuit usually uses one bulb in a generating circuit and another bulb of the same size to control the output of the generating circuit. If large output is desired a small generating bulb and small control bulb may be used and then a radio frequency ampli- fier with one large tube (instead of using two large tubes, one for the generator and the other to control the output of the generator). A grid condenser and grid leak, such as shown in Fig. 707, are usually used with a vacuum tube generating circuit to enable the grid to be kept at a desirable negative poten- tial. For whenever the grid becomes positive, current flows in the tube from the grid to filament causing a loss of power. The insertion of the grid condenser and a grid leak of suitable value automatically keeps the grid sufficiently negative that only the top tips of the positive alternations of A. C. applied to the grid make the grid positive. When the circuit starts to generate the grid is at the potential of the terminal of the filament to which the grid leak is con- nected. The A. C. applied to the grid at first makes the grid highly positive for each positive alternation. This causes current to flow away from the grid through the tube, thus making the average voltage of the grid negative. The average grid voltage will continue to become more negative until a balance is reached between the current 116 MoDBEN Theory and Peaoticb in flowing away from the grid through the tube and the cur- rent flowing to the grid through thfe grid leak. Frequently, if the grid leak is disconnected, the grid will become so highly negative that the oscillations will stop altogether. The negative charge may then leak away slowly, with the result that the tube starts to generate again, builds up a high negative charge on the grid and then stops again, thus producing intermittent oscillations.* To use any of these vacuum tube generating circuits for undamped wave telegraphy, the transmitted wave length may be changed by making a change in the oscillating circuit of the generating tube. It is usually more convenient to stop the oscillations altogether, by interrupting the D. C. sup- ply, by opening the grid circuit, or sometimes by opening the grid leak circuit. For the use of these circuits with radio telephones see Chapter X. * For a more detailed description of the effects of the average value of grid voltage, see Figs. 1015 and 1016, and accompany- ing explanation. Eadio Communication. 117' CHAPTEE VIII. DETECTORS. In order to receive signals sent out by radio apparatus, it is necessary to cause the currents flowing in the receiving apparatus to operate some device which will translate the radio frequency energy into a form which can be perceived by one of man's five senses. The ordinary system translates this radio frequency energy into an audible sound in a tele- phone receiver. The audible sound in the telephone re- ceiver is produced by motion of the telephone diaphragm at an audio frequency. Motion of the diaphragm at the frequency of the radio signal wave would be utterly useless, for such a frequency does not act upon the ear to produce a sound. The device which translates the radio frequency energy from radio frequency to a frequency suitable for the production of audible sounds in a telephone is called a " detector." In general, a detector acts so that the pull on the tele- phone diaphragm increases (or decreases)* as the radio frequency power applied to the detector varies. All com- mon detectors accomplish this by .slight variations of one simple principle. This principle is that of the check valve. A perfect check valve would be excellent ; such action, how- ever, is never fully attained in a detector. The detector always acts as a check valve which is more or less leaky. * It should be kept in mind that the diaphragm has a steady pull on it due to the permanent magnets in the receiver, and that this pull is much stronger than the change due to received signals. 118 Modern Theory and Practice in (This " leak " reduces the efficiency of the detector and is tolerated only because a better cycle of action has not been produced.) A theoretical curve showing the variation of current through a detector of the above type, for the simplest case, E E Fig. 801. Fig. 802. is shown in Pig. 801. It is seen here that a voltage applied in one direction will produce a much greater flow of current than an equal voltage applied in the opposite direction. A practical curve of a similar type is shown in Fig. 803. Consider a detector of the above type connected at D in Fig. 803. AC is a generator of radio frequency alter- A0{ W T Fig. 803. OD -(c) -(d) Fig. 804. nating current, and T a transformer. Assume the voltage supplied by the generator to be that shown in Fig. 804(a), then the current which flows through the detector will be similar to that shown in Fig. 804(6). The small reverse current is that due to the leakage. Due to the inductance Eadio Communication. 119 of a telephone winding, a current such as 804(6) could be passed through a telephone only with great difficulty, and as explained before, would accomplish no useful purpose (due to the fact that it has a radio frequency, not audible to the ear) even if it could pass through the winding. The result of connecting a telephone directly in series with D would be a current through the telephone of the nature shown in 804 (c) . This is a practically steady unidirectional current much smaller than the average value of 804(6). The current is smaller because the telephone offers a high impedan'ce to the A. C. going to D. The current is nearly ■steady because the inductance of the phones has a great tendency to keep the current from varying and is assisted in this steadying action by the stray capacitance, in parallel with the phones, between the wires of the circuit. This capacitance, although very small, is sufficient to store the greater part of the peaks of the pulses of current long enough to discharge them through the telephone while the detector current is .small, or reversed. Fig. 805 shows a practical variation of Fig. 803. Here the r condenser C serves two purposes. ^^L In the first place it provides a V o( path around the phones which has o o ^^ low impedance for the radio fre- „ „., quency. This permits the A. C. to get to the detector readily. In the second place it supplements the stray capacitance of the circuit in its action of storing the pulses of current. With this connection the current would be steady, as shown in Fig. 804 ((Z), and would have a value approximately equal to the average of 804(6). 120 MoDEEN Theory and Practice in From the preceding paragraph it is seen that an unvary- ing amount of radio frequency power supplied to a detector circuit can produce a steady unidirectional (D. C.) current through the phones. Although this steady power supply and steady current will not produce a sound in the phones, it is a relatively simple matter to arrange some part of the transmitting or receiving system so that the radio fre- quency power applied to the detector will vary at an audio frequency. The unidirectional current will then vary in a manner similar to the variations of power, and a sound I 1 I ^■ . -0 E B ^H ^ J Fig. 806a. Fig. 806b. will be produced in the phones by the varying pull on the diaphragm. The only modification of the simple check valve action explained above that it will be necessary to consider, is that due to a device which gives a current flow of the type shown in Pig. 806a. It is immediately seen from this curve that an A. C. voltage of maximum value less than OB en- counters only a high resistance with no check valve action. The electrical action with this small voltage is simply that of a high resistance. Considering a mechanical analogue, it may be said that this device acts as a leaky check valve with a spring tending to hold it shut. The action of this Eadio Communication. 121 device may be made equivalent to that of Fig. 801 by the simple expedient of applying a steady D. C. voltage (equal to OB) just large enough to balance the tendency of the " spring " to hold the valve shut. A simple method of applying this E. M. F. by means of a potentiometer is shown in Fig. 806b. If, after applying this necessary D. C. voltage, a small A. C. voltage is applied to the circuit, the flow of current caused by the A. C. will be exactly that of Fig. 804b. The total current flow, however, will be greater than this by an amount (equal to BC) due to the steady D. C. voltage OB. The total will be a current such as shown Pig. 807. Fig. 808. in Fig. 807. Curve 806 is a theoretical one which is not exactly obtained in practice. The practical curve corre- sponding to this is one such as Fig. 808, the difference being that the practical curve does not have a sharp corner or bend. The actual check valve action for the practical case is similar to that just explained. The steady value of D. C. applied, for the best results, must be enough to bring the steady values just to the point of the curve where it bends most sharply. This will be approximately at + . The simplest device which will give the results described above makes use of changes at the point of contact between two (more or less conducting) materials. This device is 122 MoDERx Theory and Practice in commonly called a crystal detector. The contact is made between two pieces of mineral or crystal, or between one of the preceding and a metal contact. Examples of such de- tectors are a bhmt metal point in contact with a carborun- dum crystal, a isharp metal point in contact with a piece of silicon, or a piece of zincite (ZnO) in contact with a piece of chalcopyrite (FeS-l-CuS). It is a peculiarity of crystal detectors that some points of contact are much more sensi- tive than others, so that it is necessary to find a sensitive contact hy trial, while receiving electrical impulses (from a local buzzer, if necessary). The chief advantage of this type of detector is simplicity. Probably the most serious disadvantage is the ease with which the crystal detector may be caused to lose its sensitive adjustment. Either a mechanical jar, or a strong electrical impulse is apt to destroy the sensitiveness of the point of contact so that no signals will be received until a new sensitive point of con- tact is found by trial. It will be seen by referring to the curves of Fig. 502 that a two-element vacuum tube may be used as a detector acting in the manner described above. This tube, known as the Fleming Valve, has been used as a detector, but since the invention of the three-element vacuum tube it has dropped out of use. Fig. 504 shows that in the three-element vacuum tube the variation of current with variation of voltage is similar to Fig. 808, so that it is evident that this tube may be used as a detector in the way explained above. As the currents in the plate and in the grid circuits both vary in about the same manner, the action in either one may be made use of. As a detector, the three-element vacuum tube has several advantages. In the first place, its action is very stable. Eadio Communication. 123 No ordinary mechanical or electrical shock affects the action of the tube as a detector, after the shock has passed. In the second place, by utilizing the amplifying action of the tube the energy available to produce signals may be increased, thus giving louder signals, and at the same time aiding sharp tuning of the oscillating circuits by taking but very little energy from them. A further advan- tage is due to the ability of the tube to generate radio fre- quency alternating current (undamped oscillations). This advantage will be explained later. Practically, the connections used for this tube as a detector are always such as to utilize the amplifying action of the tube to some extent. From this it follows that the incoming signaL= are always applied to the tube so as to produce a change in the voltage between the grid and filament, this being the input connection for the tube used as an amplifier. The phones are always connected to the plate circuit. The circuit shown in Fig. 809 makes use of the possibility of detector action in the grid cir- cuit. The condenser Cg pre- sents any free flow of electrons between the filament and grid, outside of the tube. The high resistance grid leak R provides a current path which tends to bring the grid, more or less gradually, to the same potential as the point to which R is connected. In some tubes the leakage which takes place inside of the tube is large enough so that it is not necessary to provide the leak R on the outside of the tube. This is particularly apt to be tho case with tubes which have traces of gas left in them. Pig. 809. 124 Modern Theory and Practice in If an undamped oscillation is supplied to the grid through the condenser Cg, due to the check valve action in the grid circuit, current will flow through the tube away from the grid. As the current flows away from the grid without a free supply of current to the grid, the potential of the grid will fall, making it negative with respect to the filament. As it becomes negative the flow of current through the leak increases, and very soon a grid voltage is reached such that the current flowing to the grid through the leak is just equal to the current flowing away from the grid through the tube. As long as the radio frequency power supplied re- mains constant the balance will be maintained at this par- ticular grid voltage. If the power supply increases a greater current will flow away from the grid, through the tube, and the grid voltage will drop to a lower value for its new con- dition of equilibrium. If the power supply decreases, or stops, the grid potential will go back toward its original value, so that the negative voltage on the grid changes in proportion to the radio frequency power supplied to the tube. This change in the average value of the grid voltage causes a corresponding change in the plate current, and, consequently, a change in the pull on the telephone dia- phragm. If the power supplied changes at an audio fre- quency an audible .sound is produced in the phones. The action of this circuit in the particular case of receiving a damped wave sent out by a spark set is shown in Fig. 810. Fig. 810(a) represents the voltage received. Fig. 810 (&) represents the corresponding grid potential. Fig. 810(c) represents the plate current, and Fig. 810 (cZ) represents the smoothed out current which flows through the phones. It is seen that the detecting action takes place in the grid cir- cuit, and that the resulting audio frequency changes are Eadio Communication. 125 repeated in the plate circuit in an amplified form. Notice in particular that with this action the plate current de- creases its average value when a signal is received, and that contrary to the usual case, the leaky action of the check valve system is absolutely essential. The absence of a suffi- ciently large leakage will cause the grid voltage to become negative and remain negative with but little fluctuation in value as the signals change. This has two bad effects: First, it prevents appreciable changes in the average value \(\f^' ' ^J\j\y^ \f\/^ W « iA/^ - - — lA/^- W ^^^^ Ad) Fig. 810. of the grid voltage, and consequently prevents changes in the value of plate current. Second, if the signal is fairly strong, the grid will become so highly negative that the plate current will be reduced practically to zero. Fig. 811 shows a circmit which makes use of the de- tecting action in the plate cir- cuit of the tube. The battery C in connection with the resistance r provides a variable E. M. F. which may be adjusted so that the .steady values of the current and volt- FiG. 811. 136 Modern Theoey and Practice in age in the circuit correspond to the point of sharpest cur- vature of the plate current curve. This is the steady value of D. C. necessary in the second of the preceding elementary cases (see Pig. 806a). When a constant radio frequency voltage is applied to the grid the plate current varies ap- proximately as shown in Pig. 807. The increase of current above the steady value being much greater than the decrease below the steady value, the average value of the plate cur- rent increases. This increase is proportional to the radio frequency power supplied to the circuit. As before, changes in the supplied radio frequency power will cause corre- sponding changes in the telephone current, and sounds will be produced in the phones by audio frequency changes of the power supplied. The action of this circuit in the par- ticular case of receiving a damped wave sent out by a spark set is shown in Pig. 813. Pig. 813(a) shows the received voltage. Pig. 812(6) pre- Jfl) sents the corresponding values of plate current, and Fig. 813(c) shows J ^j^/v— ^/\/V^ \J\^ — the current through the telephones. In the circuit of Fig. 811 the amplify- ing action of the tube is utilized before the detect- ing action takes place. Aside from the fact that the detecting action takes place in the grid circuit of one, and in the plate circuit of the other, there is an impor- tant difference between the circuits of Pigs. 810 and 813. The circuit of Fig. 810, with the grid condenser, gives a certain amount of cumulative effect. That is, the charge l\[V - ifV- V" (6) i£) Fig. 812. Radio Commuxication. 137 accumulated in the grid condenser by one cycle of radio frequency does not all leak away until a number of cycles have passed. The result of this is that at any instant the accumulated charge in the grid condenser (and consequently the potential of the grid) is due to the cumulative effect of several preceding cycles of the radio frequency. This causes the changes of plate current obtained (for a given change in radio frequency input) to be greater than the changes of plate current obtained in circuits not using the grid con- denser. The final result is that circuits vrith the grid con- denser are usually more sensitive detectors than those without. — . As the preceding detectors produce sounds in the phones only when the power applied varies at an audio frequency, they will not give a signal with any of the undamped wave systems unless .some device is provided at the receiving station to vary the input to the detector. Devices which make this change by interrupting one of the receiving cir- cuits, by changing the resistance of some part of the receiv- ing circuit (as by a sliding contact) or by changing the tuning of some part of the receiving circuit are of the simplest type. These changes must, of course, be made at an audio frequency. Another method of producing variations of the radio fre- quency applied to the detector makes use of a low-power, un- damped wave generating set at the receiving station. The steady undamped oscillations produced by the local source are combined with the incoming oscillations, and the resul- tant is applied to the detector. If the frequency of the local oscillation is made .slightly different from that of the incom- ing signal, at one instant the two oscillations will oppose each other, and an instant later they will add to each other. 138 Modern Theoey and Practice in The result is that the radio frequency power contained in the resultant varies. If N represent the frequency of the incoming wave, and the local frequency is made equal to iV + 1000, the locally generated wave will overtake the incoming wave 1000 times per second. The power applied to the detector will vary at a frequency of 1000 per second, and a sound will be produced in the phones with this fre- quency. If the local frequency is made equal to iV— 1000, vAAAAAAAj '" JWU -(c) J\^ W) =(e) Fig. 813. the incoming wave will overtake the local one 1000 times per second, and the result will be the same as before. It will be seen from this that the sound produced in the phones has a frequency equal to the difference between the received frequency and the local frequency. This method of receiving undamped wave signals is called heterodyne reception. It is also referred to as " beat reception," due to its depending upon the beats between two slightly different radio frequencies. Kadio Communication. 129 Fig. 813 shows the action taking place in a heterodyne circuit with a simple crystal detector. Fig. 813(a) is the received radio frequency. Fig. 813 (&) is the locally gener- ated radio frequency. Fig. 813(c) is the result of adding Fig. 813(a) to Fig. 813(6). Fig. 813(d) is the detector current, and Fig. 813(e) is the current through the phones. If a three-element vacuum tube which is connected so as to generate oscillations of a certain fi'equency has a slightly different frequency from another source applied to its grid, while it is generating, the total radio frequency volt- age applied to the grid circuit of the tube will be similar to Fig. 813(c). A heterodyne receiving action will then be obtained. This action in which the tube provides its own heterodyne is called an autodyne. In general, heterodyne methods of receiving give louder signals than other methods with the same strength of in- coming wave. One reason for this is that energy to prO: duce the signals comes from the local supply as well as from the distant station, thereby increasing the total energy available to operate the detector. Heterodyne reception of undamped waves gives a musical tone, which may be regu- lated to any desired pitch by adjusting the frequency of the local oscillations. Heterodyne reception may also be ap- plied to damped waves but does not produce a musical note in this case, as the " beats " with the damped waves come at random intervals instead of with musical regularity. Probably the determining factor which causes the use of the autodyne detector to be much more common than the use of the simple heterodyne is simplicity of apparatus. For heterodyne reception a complete and separate generat- ing circuit for radio frequencies is necessary in addition to the detector. The autodyne receiver also has the advantage 130 MoDEEN Theory and Practice in of producing so-called " regenerative amplification " in addition to its other actions. This regenerative action consists of taking part of the output of the detector and feeding it back into the input with the received signal. This increases * the strength of the audible signal produced. Its action will be discussed more fully in the following chapter. One disadvantage of the autodyne is, that since it is acting as a generator and a detector at the same time, it cannot act to the best of its ability as either. Another disadvantage is due to the necessity of producing " beats." To produce these beats the generator circuit of the detector must be tuned to a wave length different from the one being received. Under this condition, the generator circuit is not in resonance with the received wave. As the generator circuit of the autodyne is also the detector circuit, the detector loses somewhat by being operated slightly out of resonance with the received wave. This disadvantage is of no importance with short waves, but becomes appreciable with long waves. An autodyne detector used in connection with a radio frequency amplifier also presents special diffi- culties. These disadvantages may be avoided by the use of heterodyne reception. * The increase due to regeneration Is found to be very small compared to that due to the heterodyne action. Eadio Communication. 131 CHAPTEE IX. EECEIVING CIRCUITS. A receiving circuit consists primarily of a device to be acted upon by the electromagnetic waves, a detector, and suitable connections between these two. The part upon which the electromagnetic waves act is ordinarily called an antenna, or aerial. The antenna or aerial consists of one or more wires elevated above the ground and insulated from it at all points except where it is connected to ground through the receiving circuits. Wires buried a short dis- tance under ground, or under water, or wires without a ground * may be used for this purpose. The buried wires are thoroughly insulated. A closed loop or coil may also be used to pick up the effects of the electromagnetic waves. This application will be explained in Chapter XI. So far as the production of oscillations is concerned, the antenna acts in the same manner as the closed oscillating circuit. The inductance of the antenna circuit is partly that of the antenna itself (due to the flux set up around the wire by the current flowing in it) and partly that of one or more coils connected in series with the antenna. The capacitance of the circuit in the .simplest case is entirely due to the antenna. This antenna capacitance consists of the electrostatic capacity between the wires of the antenna on one side and the ground on the other side. The advancing electromagnetic waves produce in the antenna a voltage similar to the voltage at the transmitting station in all respects except size. The magnitude of this received voltage depends upon the receiving antenna and * The type without a ground connection usually has the receiving apparatus near the middle. 133 Modern Theoky and Peactice in the amount of power picked up by the receiving antenna. In any case, both the power and voltage at the receiving end are extremely small fractions of the corresponding quantities at the transmitting station. With a given antenna, in order to pick up the maximum amount of power from the electromagnetic waves, the antenna must be tuned to the wave length being received. The variation in the effects in the antenna circuit when its constants are varied are exactly the same as the variation in a closed oscillating circuit when the same changes are made,* the maximum effects being obtained when the antenna circuit is so tuned that it tends to oscillate at the same wave length as the incoming wave. The simplest tuned receiving circuit is shown in Fig. 901. Here the inductance L is made just large enough to bring the antenna to resonance with the wave length being received. A circuit such as this responds to received signals of equal strength but of different wave length with a strength which depends upon the decre- ment of the antenna circuit and of the received .signal, and upon the difference between the received wave length and the wave length to which the antenna is V. nni resonant. If the sum of the decrements Fig. 901. of the antenna system and of the received wave is small, the strength of the signal will fall off very rapidly as the resonant wave length of the antenna is made different from that of the incoming wave. On the other * This neglects the fact that any circuit having all or part of its inductance and capacitance distributed has a series of natural frequencies, or frequencies of resonance. The preced- ing statement applies to the lowest frequency, which is the only one used. Eadio Communication. 133 hand, if the sum of the decrement of the antenna and that of the incoming wave is large the .strength of the received signal will fall off slowly as the resonant wave length is made different from the received wave. This is the same effect explained under closed oscillating circuits. As the decrement of a simple open oscillating circuit is usually relatively high, this circuit (of Fig. 901) will respond appreciably to .signals quite different in wave length from that to which the antenna is tuned. Where there are a number of signals which may be received at the same time, this is a very decided disadvantage, as any signal which is not of a very different wave length from the one to be received will be heard, and is apt to be loud enough to prevent reading the desired signal. This same property of picking up any wave length within quite a range on each side of the one to which it is tuned is an advantage when it is desired to find a signal of unknown wave length. A cir- cuit used for this purpose is called a pick-up circuit or a stand-by circuit. The decrement of the circuit shown in Fig. 901 may be decreased by connecting in series with it another circuit of very low decrement consisting of an inductance coil and condenser (see Fig. 903). Here L™ is the added coil and C^ the added con- denser. If Lj and Cg are of such size that when connected to form a closed oscillat- ing circuit they would oscillate at the same wave length as the circuit of Fig. 901, putting them in series with this circuit, as in Fig. 903, will not change the wave length to which the antenna system is, tuned. It may readily be seen that the decrement will be decreased, as the amount ~ fiq_ 902. 134 Modern Theory and Practice in of energy stored in the antenna circuit with a given cur- rent is higher than before, while the losses (due to the resistance) are not appreciably changed, for the added resistance is very small. The action here is very much like starting a loaded swing into motion. If the load is light, a few pushes will bring the swing to its maximum motion, and it will not make very much difference whether the pushes are timed just right or not. This corresponds to the simple antenna. If additional load is added to the swing the rate at which it swings back and forth does not change, for the force tending to return it to its vertical posi- tion increases just exactly as fast as the mass to be moved increases. However, with the additional load, it will take a larger number of pushes to get the swing going with the ■same motion that it had in the preceding case. Also, the pushes will now have to be timed just about right or but little result will be produced on the motion of the swing. This decrease of decrement could also have been predicted from the formula for decrement in terms of the inductance and resistance. h — irR/ui L. As L has been greatly increased while R has been increased only slightly, a decrease in 8 necessarily follows, (w was kept constant in this case.) Decreasing the decrement of the antenna circuit serves the purpose of making the tuning sharper, so that wave lengths materially different from the one being received will not interfere with receiving. The nature of the change can be seen from the resonance curves of Fig. 903. Here the abscissas represent the wave lengths being received, and the ordi- nates represent the corresponding in- ^ tensity of signal produced. (The Pig. 903. strength of the electromagnetic waves Radio Communication-. 135 remaining constant.) The maximum intensity of signal occurs when the wave length received is the same as the wave length at which the antenna tends to oscillate. In order to get the loudest possible signal from a given antenna, a given detector, and an electromagnetic wave of constant strength, the energy received by the circuit should be divided about equally between the detector and the rest of the circuit.* This division may be controlled by insert- ing a transformer between the antenna circuit and a sepa- rate circuit containing the detector. By means of this transformer the voltage impressed on the detector may be varied (as compared with the voltage existing in the an- tenna) imtil the best results are obtained. In order to get sharp tuning the power used by the entire receiving system should be as small as possible. In this case the power used by the detector must be kept as low as the necessary strength of signal will permit. A trans- former between the antenna circuit and a separate circuit containing the detector may be so arranged as to accom- plish this also. It is seen from the two preceding paragraphs that the requirements for obtaining two desirable conditions in receiving apparatus are exactly opposed to each other, but both of these requirements are met by the same device (witli different adjustments, of course). The result is that this device (a transformer or its equivalent, technically called a coupler) is practically always used and that a compromise is then made between the conflicting requirements. If interference is not present, use may be made of the maxi- mum signal obtainable, with the resultant lack of sharp tuning. On the other hand, in the presence of inter- * See Zenneck " Wireless Telegraphy," Par. 172. 136 Modern Theory and Practice in Fig. 904. ference, the energy taken by the detector will be decreased to decrease the decrement, thus giving sharper tuning. Naturally, to be received in the presence of interference, a signal has to be stronger than would be required if the interference were not present. Pig. 904 shows the simplest circuit con- taining a transformer as described in the preceding paragraph. This is called an untuned circuit. This circuit in the un- tuned form has about the same advan- tages and disadvantages as the circuit of Pig. 901. To utilize the secondary circuit to the best advantage it is necesary to tune the secondary as well as the primary. The circuit then becomes that of Pig. 905. A variable condenser is usually used for Ca- L2 also is made variable either by cutting out part of the coil, or by using a variometer. It may be well to remember that there is always the capacitance of the coil and some stray capacitance between the connections to the coil in addi- tion to the capacitance of the ccm- denser. For this reason the wave length of the secondary system does not approach zero as the capacitance of the variable condenser does. That is, with any particu- lar inductance, there is some wave length (somewhat greater than the fundamental wave length of the coil) below which it is impossible to fame. Pig. 905. IiADIO L'oilML'XICATION'. 137 lu order to fiinl a signal of imknowu wave length with a circuit sucli as that ot Fig. OO.j, in general, two steps arc necessary. First, the tuning of the seeoiulary circuit is made very hroad Ijy using close coupling, large secondai'v inductance, and small secondary capacitance. Second, tin' Fro. y0.5A. — Commercial Type Receiver; Inductive Coupling. antenna inductance is tlien varied until the desired signal is picked up. If the signal happens to be one which is only just strong enough to be read with the best adjustments of the circuit, it may l:ie necessary to run over the entire range of primary adjustments a number of times using a diifer- ent value of secondary inductance each time. (In this case the signal is heard only after the secondary has by chance 138 MODEEN ThEOEY AND PeACTICE IX been tuned to a wave length near that to be received.) After getting an audible signal the secondary will be tuned, the coupling adjusted to give best results, and finally fine adjustment made on the primary (antenna). The best value of coupling, as indicated above, is usually determined as a compromise between some medium value (which will give the loudest signal) and very loose coupling, wliich tends to eliminate interference, but at the same time weakens the signal. A slight modification of the cir- cuit of Fig. 905 is shown in Fig. 906. Here condenser C^ is connected around the added inductance of the antenna circuit. This condenser is practically in parallel with the capacitance of the antenna so that increasing its value increases the wave length to which the antenna is tuned. Unless inductance L^ is small (say less than four times the inductance of the antenna) the wave length of the antenna circuit may be computed assuming the con- denser Ci to be in parallel with the capacitance of the antenna. The advantage of this circuit is that it is not necessary to make the inductance L^ continuously variable or variable by small steps. L^ will be made variable only by a few large steps, all of the fine variations of wave length being made by the condenser Cj. A further advan- tage in circuits for long waves is that the use of C^ reduces the amount of inductance required to tune for a given wave length. The disadvantage of using C^ is that the addition of the condenser increases the decrement of the circuit. This increase in decrement is due to the additional Fig. 906. Radio Communication. 139 Fig. 907. losses in the coil caused by current which circulates around through ijCi. As long as C^ is not much greater than the capacitance of the antenna this increase in decrement is not serious, the increased ease of manipulation of the cir- cuit offsetting the increase in decrement. In the two preceding circuits it has been assumed that the wave length to be received is longer than the funda- mental wave length of the antenna. If this is not true, a modification of the antenna circuit, such as that shown in Fig. 907, becomes necessary. Here the condenser G^ is inserted in series with the antenna circuit, and is, therefore, in series with the antenna capacitance. The result is that the total effective capacitance of the antenna circuit in this case is less than that of the antenna alone. This decrease in the effective capacitance of the antenna reduces the wave length to which the circuit is resonant. A practical limit to this reduction is reached when the added capacitance becomes so small that the antenna tends to oscillate as though disconnected from the ground entirely. The wave length in this case is about half the fundamental. In any case, when the added condenser is not very small, the wave length may be computed by using the effective capacitance computed as the result of the added condenser in series with the capacitance of the antenna. This effective capaci- tance and the total inductance of the antenna circuit sub- stituted in the wave length formula will approximate the value actually obtained. Condenser Cj in Fig. 907 is known 140 MoDEEN Theory and Peactice ix as a short wave condenser. In sets making use of the circuit of Pig. 906, for long waves, this same condenser is used in series with the antenna for short waves. It should be noticed that it is always necessary to have some added inductance in these circuits to form the primary of the coupler. This variable series condenser may also be used when its presence is not required on account of short wave lengths. As its insertion necessitates the addition of more inductance to tune for a given wave length, its action is to decrease the decrement of the antenna circuit in much the same manner as the circuit of Fig. 902. The advantage of this use is that it retains the ease of adjustment of the variable condenser, makes it unnecessary to have an inductance variable by small steps, and decreases the decrement of the antenna circuit. It is pointed out in the preceding paragraphs that the inductances in a coupler may be changed by cutting out por- tions of the coils. The sim- plest arrangement for giving approximately continuous va- riation of inductance within the limits of the coil is shown in Fig. 908. Here it is seen that the connections are made to the coil by means of contacts which slide on two sets of taps from the winding. One set of taps usually changes the winding by steps of one turn. The other by steps of about 10 turns. Evidently, by using this combina- tion any number of turns from zero up to the full coil can be put into the circuit. An arrangement such as this is Eadio Communication. 141 used in the primary of a coupler except wlien a variable condenser is used, as in Fig. 906. (In Fig. 906 the small steps of variation are not required.) As secondary circuits of couplers almost invariably have a variable condenser for tuning, the small steps of inductance are also unnecessary in the secondary. Where only a portion of the winding of a coupler is in use, the remaining portion of the winding forms a closed oscillating circuit (due to the electrostatic capacity of the coil, which is equivalent to a condenser connected across the terminals of the coil). This closed oscillating circuit is coupled to the active portion of the winding, and the trans- fer of energy between these two parts depends upon the same things as the transfer of energy between any other inductively coupled circuits. This transfer of energy is small as long as the wave length to which the " dead end " is resonant is much different from the wave length being used in the active portion of the coil. If the wave length being used approaches the resonant value of the dead end, the dead end will have a relatively large current produced in it and will absorb a large portion of the energy which would normally go to the detector circuit. This results in very inefficient operation of the set and correspondingly weak signals. The only absolute cure for this trouble would be to disconnect and remove the part of the winding not in use. The losses due to the dead ends may be made quite small by other means which are more readily applied. The method usually applied is to disconnect the dead end from the active part of the winding, and then divide the dead end up into several parts if necessary. This reduces the reso- nant wave length for any one part to a value much below any wave length to be used (if properly done). This cut- 142 MoDEEN Theory and Peactice in ting up of the winding is accomplished by switches, called " dead end switches," which are usually operated auto- matically by the motion of the sliding contact which changes the portion of the winding in use. Such a circuit is shown in Fig. pjq 999. 909. Another way to prevent the dead end from becoming resonant to the wave lengths in use is to short-circuit it. This is seldom done unless the dead end is a large percentage of the whole coil. In the best sets tests would be made during construction of the first set to determine whether cutting the winding into sections or short-circuiting it gives the best results. The dead end eflects discussed above make it evident that very careful design is necessary to produce a receiving set which will operate efficiently over a wide range of wave lengths. Fig. 910 shows a circuit making use of capacitive coupling. In some ways this circuit is exactly the same as that of Fig. 905. In each case there is the antenna circuit, which is tuned to the desired wave length by adding induc- tance, or inductance and ca- pacitance, and in each case the secondary circuit consists of a closed oscillating circuit. However, the inductance in the secondary circuit is so placed that the flux from the primary winding does not produce a voltage in it. The result is that no energy is transferred to the secondary by means of the magnetic field, that is, there is no inductive coupling. The necessary transfer of energy from the primary to the C -#■ ro® ^ * 55 Fig. 910. Eadio Communication. 143 secondary takes place through the condensers C" and 0" . This transfer of energy, and consequently the coupling, is controlled by the capacitance of these condensers. (These condensers are frequently mounted on the same shaft so that they vary simultaneously.) Close coupling is obtained by using large capacitance in C" and C" . This may be seen by considering the effect of making these condensers very large. They might readily be made so large that their effect upon the radio frequency currents would be negligi- ble. Fig. 910 would then become equivalent to Fig. 911. This shows direct metallic connection of the primary and secondary circuits. It appears from this circuit that with close coupling the voltage applied to the detector will not exceed the voltage existing in the primary inductance. If the detector used is of such construction that it needs power supplied to it with relatively high voltage and small current, this may be a disadvan- tage. (This disadvantage is not found with the inductively coupled set.) primary and secondary tuned exactly nrowjnnRnn -xF ^0-tHI ^ Fig. 911. If, with both the to the incoming wave length, the condensers C and C" are set for small capacitance, the voltage existing between the ends of the primary coil is split up into three parts, the parts across the coupling condensers and that across the secondary con- denser.* This voltage across the secondary being only a * With the preceding conditions, the parallel circuit will act approximately as a non-inductive resistance, so that the voltage drop across the coupling condensers will be 90° out of phase with that across the secondary circuit. 144 Modern Theoey and Practice in fraction of that across the iDrimary, the energy is transferred from primary to secondary less rapidly, which means that the coupling is looser. Other arrangements of circuits may be provided to transfer the energy from the primary to the secondary with the aid of condensers. A simple variation of Fig. 910 would short circuit condenser C". One general advantage of capacitive coupled sets is that they may be made more compact than an inductively coupled set. Fig. 912. A representative receiving circuit, including all of the essential adjustments outlined in the preceding paragraph is shown in Fig. 912 (for an inductively coupled set). The switch marked Z* is a dead end switch, and to get the best results must be opened when short waves are being received. The inductance marked " loading coil " is used only for long waves, which can be tuned to only by the use of more induc- tance than can conveniently be placed in the primary of the coupler. Eadio Communication. 145 In practical receiving, interference from other signals than the one desired is a matter that is frequently encoun- tered, and as the number of " other signals " increases with greater use of radio communication, ways of decreasing this interference become increasingly important. The simplest ways of decreasing interference make use of resonance, and in general operate to reduce interference only when the interfering signal has a wave length at least slightly differ- ent from the one being received. Resonance is made use of in practically all of the circuits which have been explained in this chapter, and in each case the circuit would respond more readily to one particular wave length than to any other. The addition of series inductance and capacitance to decrease the decrement (and therefore the interference) was explained in connection with Fig. 902. This was ap- plied to a circuit having no secondary, but it may also be applied to the type of circuit having a secondary. All of the preceding circuits make use of resonance in circuits where the resonant L and C are in series for the current which is ultimately to produce the signals. Circuits such as that of Pig. 913 may also be made use of. In Fig. 913 the main current divides between L and C, and at the same time produces a current which circu lates around through LC. The magnitude of this circulating cur- rent for a given received signal depends upon the difference be- tween the received wave length and the wave length to which LG is resonant, and also upon the ef- fective resistance of the closed cir- cuit (LC) . This circulating cur- rent sets up an E. M. F. across the 10 Fig. 913. l-±6 Modern Theory and Practice in terminals AB^ which opposes the flow of current in the main circuit. For such a circuit to be useful its resistance must be very low, and also the must be large and the L small, as compared with the effective values in the main circuit. Under these conditions the parallel circuit has practically no effect upon the main circuit so long as the wave length in the main circuit is different from the wave length to which L and G are resonant. On the other hand, if current of the wave length to which L and C are resonant attempts to pass through the main circuit, a relatively very large current (maybe as much as 200 times the main cur- rent) is set up in the closed circuit, this circulating current sets up a counter E. M. P. almost equal to the received voltage, and only an extremely small amount of current of this particular wave length succeeds in getting through the main circuit. The application of this circuit to the elimi- nation of one particular wave length should immediately be evident. If in the circuit of Fig. 913 the circuit LC is tuned to the wave length of an interfering signal, and then the antenna system as a whole is tuned to the wave length it is desired to receive, inter- ference from the one particular wave length to which LC is tuned will be greatly reduced. Another method of applying the principle of parallel reso- nance used above is shown in Fig. 914. In this case the cir- cuit LC, instead of being in series with the primary of the coupler, is placed in parallel with it. The action in the LC circuit is the same as before, but the effect upon the receiving circuit is much different. Fig. 914. Eadio Communication. 147 Thus it is seen that the coupler will have no appreciable voltage across its primary for any wave length except the one to which LC is resonant, for LG practically short-cir- cuits the primary of the coupler for other wave lengths. The result of this is that all wave lengths except this one are greatly reduced in strength in the primary of the coupler, whereas before only the one wave length was reduced. The preceding paragraph points out the elementary com- binations of resonant circuits, which may be used to de- crease interference. These circuits may also be used in combination. The principal objection to such circuits as these is the complication of apparatus and of tuning. Another objection is that although these circuits may reduce the strength of the interference greatly, they always * somewhat reduce the desired signals at the same time, due to the additional losses in these circuits. This latter, objection is not as serious as it may appear, as the loss m strength of the desired signal may be made up for by the use of the vacuum tube as an amplifier. The complication of tuning also becomes of little weight as a disadvantage in the special case when the receiving set is permanently adjusted for one particular wave length, and it is not neces- sary to receive any other wave length. Fig. 915 shows the simple receiving circuit applied to the audion detector for the reception of spark signals. This circuit is suitable for the reception of any radio signal except undamped wave telegraph. To get the maximum * This assumes that the receiving circuits were at their hest adiustment before adding the extra circuit, and that no regen- erative action of an audion detector is introduced by the added circuit. 148 MODEEN ThEOEY AND PeACTICE IN results from a circuit of this type a " C " battery with potentiometer would have to be inserted as in Fig. 811. This serves to regulate the steady value of D. C. voltage ap- plied to the grid, so that the plate current will vary about a point on the characteristic curve of the plate current where the increases in current are much greater than the decreases. Either a potentiometer or a switch to change the number of cells in the plate circuit battery may also be provided for changing the voltage applied to the plate. (The change of plate voltage also affects the point on the plate current H^^ Lh|i|i|i|i|i|i1i| Fig. 915. characteristic about which the tube works.) It should be recalled at this point that the circuit of Fig. 915 is the one in which the vacuum tube amplifies the radio frequency and then detects the signals by the check valve action in the plate circuit. Fig. 916 shows the circuit for utilizing the other type of audion detector action (for same field of use as Fig. 915). In this case, as explained in connection with Fig. 809, the detector action takes place in the grid circuit, the audio frequency result being amplified after detec- Eadio Communication. 149 tion. Circuits of this type (having a condenser of small capacitance in the grid circuit) are used more frequently than those similar to Fig. 915, due to the fact explained before that the usual vacuum tube gives stronger signals with the condenser in the grid circuit than without it. Eeferring to Figs. 810 and 812, where the action of the audion for the preceding cases was explained in detail, it will be noticed that the plate current contains not only the audio frequency variations which operate the phones, but also a repetition of the radio frequency which caused the Fig. 916. signal. In the simple audion detector circuit, no further use is made of the radio frequency energy after it acts on the check valve member of the tube, the power represented by the radio frequency variations of the plate current being thrown away. A simple modification of either Fig. 915 or Fig. 916 will permit a portion of this waste energy to be utilized. The simplest variation is shown in Fig. 917. Here the " tickler coil " in the plate circuit forms the primary of a transformer, which passes part of the radio frequency energy from the plate circuit back to the secon- 150 MoDEEN Theory and Peactice in dary of the coupler. As the telephones tend to prevent radio frequency variations from occurring in the plate (.'Ur- rent, a small " bridging condenser," C^, is usually placed in parallel with the phones to shunt these variations around them. If the tickler coil is connected with the proper polarity, this energy will add to that of the incoming signal and thereby increase the strength of the audible signal. This is known as regenerative amplification. With circuits such as that of Fig. 917 the loudness of the received signals increases with increase of the percentage of the radio fre- FiG. 917. quency energy passed from the plate circuit back to the input of the tube, up to the point where the " feed back " action becomes great enough to cause the tube to generate undamped oscillations. At this point the action of the receiver changes to autodyne receiving, but regenerative amplification is still present. (It should be recalled that in receiving spark signals with a heterodyne or autodyne, the irregularity of the beat action breaks up the audible tone of the received signal.) One result of regenerative amplification is that with spark signals the number of Eadio Communicatiok. 151 oscillations in a wave train (that is, the length of the wave train) is increased much beyond that of the wave train striking the antenna. The reason for this may be seen by considering the case of a clock (with a pendulum) which has stopped on account of running down. If a few gentle pushes are given to start the pendulum swinging, it will continue to swing for a length of time out of all proportion to the pushes given to start it. The reason that the pendu- lum continues to swing for some time is that the clockwork mechanism is helping to keep the pendulum going, but is not pushing quite hard enough to prevent its finally stopping. This action is very closely akin to that which a regenerative amplifier has on a damped wave train. One of the results of this stretching out of the wave trains is that with high spark frequency, or with long waves, the wave trains are apt to over-lap appreciably. It may be seen that the over- lapping of the wave trains tends to cut down the variation of radio frequency power passing through the electrical check valve, so that the amplification of sound produced will not increase as rapidly as would be expected from the simple regenerative theory. Any vacuum tube generating circuit may have its component parts so arranged that the- amount of energy fed from the output back to the input is not quite sufficient to produce continuous oscillations. The circuit is then the basis of a regenerative receiving cir- cuit. Likewise, any regenerative receiving circuit may be made to generate undamped oscillations and act as an auto- dyne detector, unless the tube is poor, or the losses in the circuit are high. A simple vacuum tube circuit for the reception of un- damped wave telegraph signals is shown in Fig. 918. This 152 Modern Theory and Practice in circuit makes use of the generating action of the circuit of Fig. 513. The small capacitance condenser in the grid circuit causes the detector action to take place in the grid circuit. The sensitiveness of a circuit, such as this, is afEected to a marked extent by the amount of radio fre- quency fed from the plate circuit back to the grid circuit; that is, by the coupling between these two circuits. This coupling aifects not only the generating action of the tube, but also the detector action. Close coupling, although pro- FiG. 918. ducing a strong heterodyne action, causes the tube to be worked so hard as a generator that it is no longer sensitive as a detector. The best results are usually obtained with coupling not quite loose enough to stop generation of un- damped oscillations. If the condenser in the grid circuit of Fig. 918 is omitted the detector action of the tube is changed to that of the circuit of Fig. 811 (see explanation accompany- ing Fig. 811), while the heterodyne action remains essen- tially unchanged. Eadio Communication. 153 Fig. 919. The application of another type of vacuum tube circuit, of radically different appearance, to autodyne reception is shown in Pig. 919. The gen- erating circuit is that of Fig. 514, the parts of the figure having been rearranged for convenience. The condensers Cj and C, control the amount of energy being fed from the plate circuit back to the grid circuit. (See explanation ac- companying Fig. 514.) It is possible to construct these two condensers so that they may be varied simultaneously, increas- ing one and decreasing the other, so that the effective capacitance of the two in series remains constant. Under this Condition the adjustment of these condensers does not change the wave length to which the circuit is tuned. It should be noticed that with this circuit it is not possible to omit the grid condenser Cg and have the circuit operate. This is due to the fact that it is necessary to have the grid condenser to prevent the high voltage D. C. from getting to the grid. The use of an amplifier to increase the strength of signals from radio receiving apparatus has become so common that the amplifier may well be considered as a part of certain types of receiving apparatus. For radio communication, amplifiers may be divided into three classes according to the frequencies upon which they operate efficiently. The first class is designed for the amplification of audio fre- quencies and gives little or no amplification of radio fre- 154 MoDEEN Theoet and Peactice in quencies. The second class is designed for the amplifica- tion of radio frequencies and gives no amplification of audio frequencies. The third class includes both audio and radio frequencies, giving amplification over an enormous range of frequencies. This third class .is the resistance coupled type (in which the impedance in the plate circuit, see Fig. 507, is principally resistance) and gives less amplifi- cation per tube than the other types. The audio frequency amplifier takes the output from the detector and amplifies it through one or more tubes before it reaches the phones. The radio frequency amplifier takes the radio frequency from the secondary of the coupler and amplifies it through one or more tubes before it reaches the detector. It is evident that both radio frequency and audio frequency amplification may be made use of in the same receiving system. In either of the first two classes of amplifiers there is more or less limited range of fre- quencies over which the amplifier will give good results. Outside of this range the amplification falls off. In radio communication the range of audio frequencies encountered is approximately from 300 to 2000 cycles per second. One ,set of apparatus may be designed to cover this range with satisfactory results. The range of frequencies encountered in " radio frequency " is very wide, ranging from 10,000 to 3,000,000 cycles per second. The second class of ampli- fier, with a particular set of apparatus gives good results over only a small portion of this total range. This limited range is, however, not by any means entirely a disadvan- tage. Using an audio frequency amplifier any audio fre- quency disturbance that gets into the amplifier system (such as "other signals," "static," and disturbances from electrical machinery) is amplified at the same time that the Radio Communication. 155 desired signal is amplified, and at times appears to be amplified more than the signal is. Using the radio fre- quency amplifier, only radio frequency disturhances are amplified, and even radio frequencies, if widely different from the frequency for which the apparatus was designed, are not amplified. The result is that when using a high degree of amplification an audio frequency amplifier will produce more undesirable (stray) sound than a radio fre- quency amplifier. As stray noises tend to drown out the desired signal, this is an advantage in favor of the radio frequency amplifier. The circuits of, the radio frequency amplifier,, are essentially the same as those of the audio frequency amplifier, but the proportions of the apparatus are very different. This difference' is due primarily to the difference in action of inductance and capacitance at the different frequencies. It should be recalled that the diffi- culty that an alternating current encounters in getting through an inductance varies directly with the frequency, while the difficulty that the same current encounters in getting through capacitance varies inversely as the fre- ■quency. As a numerical example, the secondary of the transformer in an audio frequency amplifier might have as much as 100 henries inductance, while a radio frequency amplifier using the same type of circuit would have a rela- tively few millihenries inductance in the secondary of its transformer. At radio frequencies the effects of the capaci- tance of the coils, and other small capacitances in the cir- cuit, are very noticeable, special care being necessary to prevent the inductances being practically short-circuited by the capacitance between the various turns of the induc- tance coils. The capacitance between the plate and grid inside of the tube (and their connections outside of the 156 MoDEEN Theory and Peactice in tube) couple the plate and grid circuits (see Fig. 515 and accompanying explanation) and tend to feed enough of the output of the tube back to its input to cause the circuit to act as a generator of undamped oscillations. If the ampli- fying circuit acts as a generator it is no longer usable as an ordinary amplifier. N'aturally, the greater the amplifica- tion the smaller is the percentage of the output that can stray back to the input without causing this trouble. Due to these effects of capacitance a radio frequency amplifier is not designed for as high amplification per tube as the audio frequency amplifier. Contradictory as it may seem, this does not necessarily lead to a weaker signal in the phones when using a fixed number of steps of amplifica- tion. The reason for this is that with ordinary (non- regenerative) audion circuits the audibility of the signal, or the amplitude of the audio frequency variations produced in the plate current, varies as the square of the radio fre- quency voltage applied to the grid of the tube to produce this change. This means that, for instance, an amplifica- tion of radio frequency voltage of four before application to the detector will produce the same increase in signal strength as an audio frequency amplification of 16 after the signal is detected, or that an amplification of N times by radio frequency amplifiers produces as great an increase in signal as an amplification of N^ by audio frequency amplifiers. This relation does not apply to the autodyne type of detector, for in this case the audibility of the output varies only directly as the first power of the radio frequency input, so that radio frequency or audio frequency amplifi- cation is equally effective in producing increases in the strength of the signal. Radio Commuxicatxon. 157 Figs. 920 to 933 .show the application of amplifiers in a number of typical circuits. Fig. 920 shows the applica- tion of a one-step amplifier to a .simple spark receiving set. This is the type of amplifier circuit in which an impedance or choke coil is placed in the plate circuit of the detector tube to keep the plate current (outside the tube) approxi- mately constant. When an incoming signal tends to change the plate current through the tube the voltage between the plate and filament of the detector tube changes. This FiQ. 920. variation of voltage is passed on to the grid of the amplifier tube through the condenser C and causes a larger change in the plate circuit of the amplifier tube. (See Fig. 507 and accompanying explanation.) The grid leak R^ is usually used in this circuit to enable the average voltage applied to the grid to be maintained at a desirable value for operation of the tube as an amplifier. If R^ is omitted, the electrons which are caught by the grid each time that it becomes positive will accumulate on the grid causing its average voltage to become lower when signals are received. 158 Modern Theoey and Peactice in In this case a strong signal is apt to cause so large an accumulation of electrons on the grid that the plate current will be stopped by this negative charge on the grid, and the tube will temporarily cease to operate. Fig. 921 differs from Fig. 930 in two respects. First, the transfer of energy from the detector tube to the ampli- fier tube takes place through a transformer. Second, the receiving set to which the amplifier is attached has a tickler coil which provides a means for feeding part of the radio A. F. TRANSFORMER \J9 n rn Fig. 921. frequency output of the detector tube back into the input of the same tube. This feed back may be utilized to pro- duce regenerative amplification of spark signals (inde- pendent of the amplification produced by the amplifier tube), or to cause the detector tube to generate undamped oscillations and act as an autodyne detector. The bridging condenser provides a path for the radio frequency around the primary of the transformer. This is necessary as the impedance of the transformer winding for radio frequency, due almost exclusively to the distributed capacitance of the Radio Communication. 159 winding in parallel with the inductance of the winding, is apt to be too high to permit proper operation of the tickler coil. In this circuit the audio frequency variations, which occur in the plate current, pass through the primary of the transformer and produce similar changes in voltage across the secondary of the transformer. These changes of voltage are in turn applied to the grid of the amplifier tube and produce greater changes in the plate current of the amplifier tube. TRANSFORMER Fig. 922. Fig. 922 shows a one-step radio frequency amplifier (in which the radio frequency is amplified before reaching the detector). This circuit is essentially the same as that of Fig. 505. Fig. 923 shows a detector and three-step amplifier con- nected to a spark receiving set. This circuit is of the type that uses transformers to couple the various tubes. The resistance units, marked R, in the filament circuits of the amplifier tubes are of such value as to produce a small 160 MoDEEN Theory and Practice in voltage drop (say one volt) between the filament and the connection to the grid circuit when the normal filament current is flowing. This value of resistance is picked out to give a drop of voltage which will hold the average value of the grid voltage at a desirable point for the efficient oper- ation of the amplifier. As shown, it will hold the average value of the grid voltage at a small negative value (with respect to the filament). The rheostats in the filament circuits are not peculiar to any type of amplifier, and are Pig. 923. frequently used as shown (simply to regulate the filament currents to the desired value). In the majority of the preceding cases it has been as- sumed that where a high voltage D. C. was required it would be taken from a battery. The objection to using a supply directly from a D. C. generator is that, although such a supply is substantially constant as measured by ordinary means, it is sufficiently irregular to produce much undesirable noise in a system which uses it. The prime source of this irregularity is the commutator of the gener- 11 Radio Communication. 161 ator, and any other machines which happen to be connected to the generator. To decrease the effects of these irregu- larities a device called a " filter " is provided. Fig. 924 shows a o— ''OCMWQ^- filter circuit. This filter unit f„om " cons'ists of two coils of large in- '=^''""or ductance (say one henry or o-^umZ3 V Pig. 1108. ciple of the apparatus used to eliminate grinders. This consists of two exactly similar loop ' antenna? in the same plane, placed some distance apart, aiid connected so that the fluxes cutting both coils simultaneously produce voltages in opposition to each other. As the electromagnetic waves causing grinders arrive at both coils simultaneously, the 198 MoDEHN Theory and Peacticb in voltages produced by these waves reach their maximum values at the same time and exactly oppose each other. If the plane of the coils is not at right angles to the direction of travel of the signal waves, the voltages produced by the signals do not reach their maximum values at the same time in the two coils so that there is some resultant signal left over (depending upon the distance that the coils are apart).* Other systems based upon causing the voltages pro- duced in two parts of the system by static to oppose and cancel, while not entirely canceling the signal, have proven capable of considerable reduction of static. A particularly valuable discussion of a number of systems of this type is contained in an article by Lieutenant-Commander A. Hoyt Taylor, U. S. N. E. P., in the December, 1919, "Proceed- ings of the Institute of Eadio Engineers." * If the dimensions of tlie coils and distances between them are both small compared to half a wave length, the voltage In one coll is the difference between two equal and nearly opposite voltages, and the final resultant is the difference between two equal and nearly opposite resultant voltages from the coils, thua giving an enormous reduction in the strength of signal avail- able. Either enormous amplification, or dimensions approach- ing half a wave length, or more practicably a compromise, is necessary to obtain signal audibilities comparable to those directly from an ordinary antenna. Eadio Communication. 199 CHAPTEE XII. MEASUREMENTS AND FORMULA. 1. Energy Relations in an Inductive Circuit. — If i rep- resents the instantaneous value of the current in an induc- tive circuit at any instant and E the impressed E. M. F. and R the resistance of the circuit, then Ei will be the instantaneous value of power expended. A portion of E is expended in sending a current i through the resistance B and is equal to Ri. The remaining part of E is expended in overcoming the E. M. E. of self-inductance and is equal to L di/dt. Therefore E = Ri + Ldi/dt. Multiply through by idt. Eidt = Ri^dt + Lidi. Hence, [ * Eidt = W = f ' Ri^dt + [ ' Lidi ( 1 ) Jo Jo jo The term Ri^dt represents the energy spent in heat, Jo !' . . LP and Lid%= — =— represents the energy stored up in the magnetic iield while the current increases from to I. If the current is gradually decreased, the energy 1/2 LP is returned to the circuit, due to its action in tending to pre- vent the decrease, but if the circuit is suddenly broken a very high induced E. M. F. is created and the en^srgy is expended in sparking across the contacts and volatilizing the metal. 2. Frequency of Oscillation in a Series Circuit. — The equation of frequency in an oscillating circuit as originally derived by Lord Kelvin is based on the law, enunciated by Kirchhoff, that in a circuit in which current is flowing the 200 Modern Theory and Practice in sum of the instantaneous values of E. M. F.'.s is equal to zero. In an oscillating circuit, then, Ri + L di/dt + q/C = 0. Since i=dq/dt and di/dt = d^q/dt^ the equation above becomes d^ R dq q _ dP^ L dt^ LC ""• The solution of this differential equation gives us a value for q, hence also for i. It can be shown by mathematical deduction from the above equation that if the value of R is greater than the ^f^LjC, the charge on the condenser dies away gradually, as time increases, in such a manner that the current is always in the same direction. Thus no oscillations take place. If the value of R is less than above- named value, the differential of q has positive and negative values, or the current in the circuit is reversed periodically, as time increases. The equation for current is of the form below, %= — = • e 2L sm V^-B' {(>E^)'} C in which U is the initial value of the B. M. F. across the condenser, in volts, L is the inductance in henries, C the capacity in farads, R the resistance in ohms. This is the general form of an equation for an alternating current of decreasing amplitude, i = p sin wt where Prom this , = 2nf= yjj^. R' Radio Communication. 201 When the value of R^/'^L' is small compared to l/LC, the equation becomes This is the equation used to determine the frequency of oscillations in all radio circuits. It is seen from the above that the resistance not only limits the conditions for oscil- lations, hut, within those limits, affects the frequency to a certain extent. In the derivation of the above formulae no account was taken of the fact that the resistance of a conductor is not the same for high frequency currents as for continuous or low frequency current. Hence the formulae are not rigidly accurate. 3. Measurement of High Frequency Resistance. — ^There are four methods of measuring high frequency resistance, (1) calorimeter method, (2) substitution method, (3) resistance variation method, (4) reactance variation method. All of these methods involve the use of some accurate high frequency ammeter. The first named is simple and can be used to measure the resistance of a portion, or of all, of the circuit. The coil or circuit or other apparatus whose resistance is to be measured is placed in some form of calorimeter, which may be a simple air chamber, oil bath, etc. The current is measured by the ammeter in series and the resistance computed from the observed cur- rent I and the power, or rate of heat production, as deter- mined from the calorimeter, Power =i?P, 202 Modern Theohy and Practice in The power may be measured calorimetrically or, as is usually done in practice, a known resistance is substituted for R and placed in the calorimeter and direct current sent through, and adjusted until the final temperature is the same as with the unknown resistance. By comparison of the results the unknown resistance is determined. The substitution method is convenient and rapid and is suitable for rough measurements. It involves the use of a resistance standard, and due to the fact that when substi- tutions are made in a circuit, stray B. M. F.'s are elec- trostatically induced, this method is not as accurate as some others. The procedure under the third and fourth methods is practically the same, but since a variable reactance, such as a calibrated condenser, is more often available than a resistance standard, the latter method will be described in full. This method is sometimes called the decrement method, but the name is no more applicable to this than to the other methods, since all measure decrement in the same sense that this does. That the method primarily measures resistance rather than decrement is seen from the fact that in its simplest and most accurate form it utilizes undamped current, which has no decrement. An ammeter is connected in series with the circuit whose resistance is to be measured and the circuit loosely coupled to a source of undamped oscillations. The circuit is tuned to resonance and the current measured by ammeter. The reactance, which is zero at resonance, is now changed to some other value X^. The change can be effected by vary- ing the inductance or the capacity. Calling the currents at resonance and with added reactance Ir and I^ respectively, Eadio Communication. 203 we get the following equations, assuming the applied E. M. F., E, to remain constant, T ■>- -^' From these it follows that If the reactance is varied by such an amount as to make the quantity under the radical sign equal to unity, the equation reduces to R^X-,. When the reactance is varied by changing the setting of a variable condenser, equation (1) becomes (2) given below. This reduction is effected by substituting for X^ its value, in terms of capacity, or — p=- — — -=p where Cr and Ci represent capacity at resonance and with changed reactance respectively. For variation of inductance (1) becomes, by similar reduction (3), R^±^{L,-Lr)y\j4tj^,. (3) The reactance may be varied by changing the frequency of the source but care must be taken that, with the change of 204 Modern Theory and Pbactice m frequency, the B. M. F. is maintained constant. See equation (4). If we observe equal values of current on the two sides of the resonant value Ir and call the corresponding values of cajjacity C^ and C^, the equation becomes 1 C,_-G, I 7,^ The above formulae are rigorous for undamped E. M. F. and are sufficiently accurate for damped B. M. F. when the damping is small. 4. Decrement Calculation. — By Kirchhoff's law the sum of the instantaneous values of the E. M. F. around a com- plete circuit in which free oscillations are taking place must be equal to zero. Hence, in a circuit containing inductance L, capacity C and resistance R, 'idt L^^+Ri + ^=0. The solution of this equation gives, when the resistance is not very large, an equation for instantaneous current as follows: i=Iof- 2Zr sin wt, in which /„ = initial current, t = time elapsed since beginning of wave train, w = Sw times the frequency, T^y- = damping factor, usually indicated by . a, which by definition is the product of frequency times log decrement. The relation between the damping factor and the log decrement can be shown mathematically as fol- lows : If the time of one complete oscillation is represented by T = 2ir/a), the ratio of successive current maxima (sin Eadio Communication. 205 oit being unity) in same direction, as determined from cur- rent equation above, is I £-1^ . 27ra J -g(f + y) , which simplified equals f.^^^e^oT. The ISTapierian logarithm of this ratio is the log decrement, 9 which consequently equals—^. Since a = R/2L, 0) 2Z/ mL X It is seen that the decrement is tt times the ratio be- tween the resistance and the inductive reactance, or ir times the ratio between the resistance and the capacitive reactance, since wL = —^ . From these equations we see that, with known reactance, a measurement of resistance is essentially a measure of decrement. The measurement of resistance as explained in last sec- tion can, with slight changes in formulae, be made directly into a measurement of decrement. The formulae for decre- ment are obtained by application of the relation, just de- rived, between 8 and R. It has been proved that with a slightly damped source the formula for the sum of the decrements of the primary and secondary circuits is the same as the formula for the decrement of the secondary alone when the primary, or source, is undamped. Hence the decrement formulae become g_ ±(6V-(7J FTT^ .,.._ ±(L,-Lr) I I,- Gap, Quenching QQQOO Gap, Spark -• •- Gap, Synchronous •X- Ground Inductance Appendices. 233 — njimM> — Inductance, Iron cored. -^wm^ Inductance, Variable fs> — OR — onnnspw, Key -^. Receiver, Telephone -. Resistance -AA/VW^ Resistance, Variable... VyVW AAAMAA/i Transformer, Iron cored J G~ - Transmitter, telephone ^ . 234 Appendices. Voltmeter Cvh Wattmeter /Tn Wave Meter ■€>■ APPENDIX C. INTERNATIONAL MORSE CODE AND CONVENTIONAL SIGNALS. 1. A dash is equal to three dots. 2. The space between parts of the same letter is equal to one dot. 3. The space between two letters is equal to three dots. 4. The space between two words is equal to five dots. A ■ B - C - D - E . F . G - H I . J • K ■ L . M N ■ - P • Q - R . S • T - U . V • W X ■ Y - Z - 1 2 3 4 5 6 7 8 9 Period Semicolon Comma Colon Interrogation Exclamation point Apostrophe Hyphen Bar indicating fraction .... Parenthesis Inverted commas Underline Double dash Distress call Attention call to precede every transmission General inquiry call From (de) Invitation to transmit (go ahead) Warning — high power Question (please repeat after ) — interrupting long messages Wait Break (Bk.) (double dash) . Understand Error Received (O. K.) Position report (to precede all position messages) . . . End of each message (cross) Transmission finished (end of work) (conclusion of correspondence) 236 Appendices, APPENDIX D. EXTRACT FROM LAWS GOVERNING RADIO COMMUNICATION [Public — No. 264.] [S. 6412.] An Act to regulate radio communication. Be it enacted by the Senate and House of Representatives of the United States of America in Congress assembled, That a person, company or corporation within the jurisdiction of the United States shall not use or operate any apparatus for radio communication as a means of commercial intercourse among the several states, or with foreign nations, or upon any vessel of the United States engaged in interstate or foreign commerce, or for the transmission of radiograms or signals the effect of which extends beyond the jurisdiction of the State or Terri- tory in which the same are made, or where interference would be caused thereby with the receipt of messages or signals from beyond the jurisdiction of the said State or Territory, except under and in accordance with a license, revocable for cause, in that behalf granted by the Secretary of Commerce and Labor upon application therefor; but nothing In this Act shall be construed to apply to the transmission and exchange of radio- grams or signals between points situated in the same State; Provided, That the effect thereof shall not extend beyond the jurisdiction of the said State or interfere with the reception of radiograms or signals from beyond said jurisdiction; and a license shall not be required for the transmission or exchange of radiograms or signals by or on behalf of the Government of the United States, but every Government station on land or sea shall have special call letters designated and published in the list of radio stations of the United States by the Department of Commerce and Labor. Any person, company, or corporation that shall use or operate any apparatus for radio communication in violation of this section, or knowingly aid or abet another Appendices. 237 person, company, or corporation in so doing, shall be deemed guilty of a misdemeanor, and on conviction thereof shall be punished by a fine not exceeding five hundred dollars, and the apparatus or device so unlawfully used and operated may be adjudged forfeited to the United States. Sec. 2. That every such license shall be in such form as the Secretary of Commerce and Labor shall determine and shall contain the restrictions, pursuant to this Act, on and subject to which the license is granted; that every such license shall be issued only to citizens of the United States or Porto Rico or to a company incorporated under the laws of some State or Territory or of the United States or Porto Rico, and shall specify the ownership and location of the station in which said apparatus shall be used and other par- ticulars for its identification and to enable its range to be esti- mated; shall state the purpose of the station, and, in case of a station in actual operation at the date of passage of this Act, shall contain the statement that satisfactory proof has been furnished that it was actually operating on the above-men- tioned date; shall state the wave length or the wave lengths authorized for use by the station for the prevention of inter- ference and the hours for which the station is licensed for work; and shall not be construed to authorize the use of any apparatus for radio communication in any other station than that specified. Every such license shall be subject to the regu- lations contained herein, and such regulations as may be estab- lished from time to time by authority of this Act or subsequent Acts and treaties of the United States. Every such license shall provide that the President of the United States in time of war or public peril or disaster may cause the closing of any station for radio-communication and the removal therefrom of all radio apparatus, or may authorize the use or control of any such station or apparatus by any department of the Gov- ernment, upon just compen,sation to the owners. Sec. 3. That every such apparatus shall at all times while in use and operation as aforesaid be in charge or under the super- vision of a person or persons licensed for that purpose by the Secretary of Commerce and Labor. Every person so licensed 238 Appendices. who in the operation of any radio apparatus shall fail to ob- serve and obey regulations contained in or made pursuant to this Act or subsequent acts or treaties of the United States, or any one of them, or who shall fail to enforce obedience thereto by an unlicensed person while serving under his supervision, in addition to the punishments and penalties herein prescribed, may suffer the suspension of the said license for a period to be fixed by the Secretary of Commerce and Labor not exceeding one year. It shall be unlawful to employ any unlicensed person or for any unlicensed person to serve in charge or in super- vision of the use and operation of such apparatus, and any person violating this provision shall be guilty of a misdemeanor, and on conviction thereof shall be punished by a fine of not more than one hundred dollars or imprisonment for not more than two months, or both, in the discretion of the court for each and every such offense: Provided, That in case of emergency the Secretary of Commerce and Labor may authorize a col- lector of customs to Issue a temporary permit, in lieu of a license, to the operator on a vessel subject to the radio ship Act of June twenty-fourth, nineteen hundred and ten. 4. That for the purpose of preventing or minimizing inter- ference with communication between stations in which such apparatus is operated, to facilitate radio communication, and to further the prompt receipt of distress signals, said private and commercial stations shall be subject to the regulations of this section. These regulations shall be enforced by the Secre- tary of Commerce and Labor through the collectors of customs and other officers of the Government as other regulations herein provide for. The Secretary of Commerce and Labor may, in his discretion, waive the provisions of any or all of these regulations when no interference of the character above mentioned can ensue. The Secretary of Commerce and Labor may grant special temporary licenses to stations actually engaged in conducting experiments for the development of the science of radio com- munication, or the apparatus pertaining thereto, to carry on special tests, using any amount of power or any wave lengths, at such hours and under such conditions as will insure the Appendices. 239 least interference with the sending or receipt of commercial or Government radiograms, of distress signals and radiograms, or with the work of other stations. In these regulations the naval and military stations shall be understood to be stations on land. Regulations, nobmal wave length. First. Every station shall be required to designate a certain definite wave length as the normal sending and receiving wave length of the station. This wave length shall not exceed six hun- dred meters or it shall exceed one thousand six hundred meters. Every coastal station open to general public service shall at all times be ready to receive messages of such wave lengths as are required by the Berlin convention. Every ship station, except as hereinafter provided, and every coast station open to general public service shall be prepared to use two sending wave lengths, one of three hundred meters and one of six hun- dred meters, as required by the International convention in force: Provided, That the Secretary of Commerce and Labor may, in his discretion, change the limit of wave length reser- vation made by regulations first and second to accord with any international agreement to which the United States Is a party. OTHEE WAVE LENGTHS. Second. In addition to the normal sending wave length all stations, except as provided hereinafter in these regulations, may use other sending wave lengths: Provided, That they do not exceed six hundred meters or that they do exceed one thousand six hundred meters; Provided further, That the char- acter of the waves emitted conforms to the requirements of regulations third and fourth following. USE OF A " PUEE WAVE." Third. At all stations if the sending apparatus, to be re- ferred to hereinafter as the " transmitter," is of such a char- acter that the energy is radiated in two or more wave lengths, 840 Appendices. more or lesS sharply defined, as indicated by a sensitive wave meter, the energy in no one of the lesser waves shall exceed ten per centum of that in the greatest. USE OF A " SHAKP WAVE." Fourth. At all stations the logarithmic decrement per com- plete oscillation in the wave trains emitted by the transmitter shall not exceed two-tenths except when sending distress sig- nals or signals and messages relating thereto. USE OF " STANDARD DISTRESS WAVE." Fifth. Every station on shipboard shall be prepared to send distress calls on the normal wave length designated by the international convention in force, except on vessels of small tonnage unable to have plants insuring that wave length. SIGNAL OF DISTRESS. Sixth. The distress call used shall be the international sig- nal of distress # • • ^B B WM • • • USE OF " BROAD INTEKFERING WAVE " FOB DISTRESS SIGNALS. Seventh. When sending distress signals, the transmitter of a station on shipboard may be tuned in such a manner as to create a maximum of interference with a maximum of radiation. DISTANCE REQUIREMENT FOB DISTRESS SIGNALS. Eighth. Every station on shipboard, wherever practicable, shall be prepared to send distress signals of the character specified in regulations fifth and sixth with sufiicient power to enable them to be received by day over sea a distance of one hundred nautical miles by a shipboard station equipped with apparatus for both sending and receiving equal in all essential particulars to that of the station first mentioned. Appendices. 241 "eight of way " FOE DISTKBSS SI6NAXS. Ninth. All stations are required to give absolute priority to signals and radiograms relating to ships in distress; to cease all sending on hearing a distress signal; and, except when engaged in answering or aiding the ship in distress, to refrain from sending until all signals and radiograms relating thereto are completed. REDUCED POWER FOB SHIPS NEAR A GOVERNMENT STATION. Tenth. No station on shipboard, when within fifteen nau- tical miles of a naval or military station, shall use a trans- former input exceeding one kilowatt, nor, when within five nautical miles of such a station, a transformer input exceeding one-half kilowatt, except for sending signals of distress, or signals or radiograms relating thereto. INTEBCOMMUNICATION. Eleventh. Each shore station open to general public service between the coast and vessels at sea shall be bound to exchange radiograms with any similar shore station and with any ship station without distinction of the radio systems adopted by such stations, respectively, and each station on shipboard shall be bound to exchange radiograms" with any other station on shipboard without distinction of the radio systems adopted by each station, respectively. It shall be the duty of each such shore station, during the hours it is in operation, to listen in at intervals of not less than fifteen minutes and for a period not less than two minutes with the receiver tuned to receive messages of three hundred meter wave lengths. DIVISION OF TIME. Twelfth. At important seaports and at all other places where naval or military and private or commercial shore stations operate in such close proximity that interference with the work of naval and military stations cannot be avoided by the enforce- ment of the regulations contained in the foregoing regulations 16 243 Appendices. concerning wave lengths and character of signals emitted, such private or commercial shore stations as do Interfere with the reception of signals by the naval and military stations con- cerned shall not use their transmitters during the first fifteen minutes of each hour, local standard time. The Secretary of Commerce and Labor may, on the recommendation of the de- partment concerned, designate the station or stations which may be required to observe this division of time. GOVERNMENT STATIONS TO OBSEEVE DIVISION OF TIME. Thirteenth. The naval or military stations for which the above-mentioned division of time may be established shall transmit signals or radiograms only during the first fifteen minutes of each hour, local standard time, except In case of signals or radiograms relating to vessels in distress, as herein- before provided. USE OP UNNECESSAKY POWEE. Fourteenth. In all circumstances, except in case of signals or radiograms relating to vessels in distress, all stations shall use the minimum amount of energy necessary to carry out any communication desired. GENEEAL EESTEICTIONS ON PRIVATE STATIONS. Fifteenth. No private or commercial station not engaged In the transaction of bona fide commercial business by radio communication or In experimentation In connection with the development and manufacture of radio apparatus for com- mercial purposes shall use a transmitting wave length exceed- ing two hundred meters, or a transformer input exceeding one kilowatt, except by special authority of the Secretary of Com- merce and Labor contained in the license of the station: Pro- vided, That the owner or operator of a station of the character mentioned In this regulation shall not be liable for a violation of the requirements of the third or fourth regulations to the penalties of one hundred dollars or twenty-five dollars, respec- tively, provided in this section unless the person maintaining Appendices. 243 or operating such station shall liave been notified in writing that the said transmitter has been found, upon tests conducted by the Government, to be so adjusted as to violate the said third and fourth regulations, and opportunity has been given to said owner or operator to adjust said transmitter in con- formity with said regulations. SPECIAL RESTRICTIONS IN THE VICINITIES OF GOVERNMENT STATIONS. Sixteenth. No station of the character mentioned in regula- tion fifteenth situated within five nautical miles of a naval or military station shall use a transmitting wave length exceed- ing two hundred meters or a transformer input exceeding one-half kilowatt. SHIP STATIONS TO COMMUNICATE WITH NEAREST SHORE STATIONS. Seventeenth. In general, the shipboard stations shall trans- mit their radiograms to the nearest shore station. A sender on board a vessel shall, however, have the right to designate the shore station through which he desires to have his radio- grams transmitted. If this" cannot be done, the wishes of the sender are to be complied with only if the transmission can be effected without interfering with the service of other stations. LIMITATIONS FOR FUTURE INSTALLATIONS IN VICINITIES OF GOVERNMENT STATIONS. Eighteenth. No station on shore not in actual operation at the date of the passage of this Act shall be licensed for the transaction of commercial business by radio communication within fifteen nautical miles of the following naval or military stations, to wit: Arlington, Virginia; Key West, Florida; San Juan, Porto Rico; North Head and Tatoosh Island, Wash- ington; San Diego, California; and those established or which may be established in Alaska and in the Canal Zone; and the head of the department having control of such Government stations shall, so far as is consistent with the transaction of governmental business, arrange for the transmision and receipt of commercial radiograms under the provisions of the Berlin 244 Appendices. convention of nineteen hundred and six and future inter- national conventions or treaties to which the United States may he a party, at each of the stations ahove referred to, and shall fix the rates therefor, subject to control of such rates by Con- gress. At such stations and wherever and whenever shore sta- tions open for general public business between the coast and vessels at sea under the provisions of the Berlin convention of nineteen hundred and six and future international conventions and treaties to which the United States may be a party shall not be so established as to insure a constant service day and night without interruption, and in all localities wherever or whenever such service shall not be maintained by a commercial shore station within one hundred nautical miles of a naval radio station, the Secretary of the Navy shall, so far as is con- sistent with the transaction of governmental business, open naval radio stations to the general public business described above, and shall fix rates for such service, subject to control of such rates by Congress. The receipts from such radiograms shall be covered into the Treasury as miscellaneous receipts. SECRECY OF MESSAGES. Nineteenth. No person or persons engaged in or having knowledge o£ the operation of any station or stations, shall divulge or publish the contents of any messages transmitted or received by such station, except to the person or persons to whom the same may be directed, or their authorized agent, or to another station employed to forward such message to its destination, unless legally required so to do by the court of competent jurisdiction or other competent authority. Any person guilty of divulging or publishing any message, except as herein provided, shall, on conviction thereof, be punishable by a fine of not more than two hundred and fifty dollars or imprisonment for a period of not exceeding three months, or both fine and imprisonment, in the discretion of the court. PENALTIES. For violation of any of these regulations, subject to which a license under sections one and two of this Act may be issued. Appendices. 345 the owner of the apparatus shall be liable to a penalty of one hundred dollars, which may be reduced or remitted by the Secretary of Commerce and Labor, and for repeated violations of any of such regulations, the license may be revoked. For violation of any of these regulations, except as provided in regulation nineteenth, subject to which a license under sec- tion three of this Act may be issued, the operator shall be sub- ject to a penalty of twenty-five dollars which may be reduced or remitted by the Secretary of Commerce and Labor, and for repeated violations of any such regulations, the license shall be suspended or revoked. Sec. 5. That every license granted under the provisions of this Act for the operation or use of apparatus for radio com- munication shall prescribe that the operator thereof shall not wilfully or maliciously interfere with any other radio com- munication. Such interference shall be deemed a misdemeanor, and upon conviction thereof the owner or operator, or both, shall be punishable by a fine of not to exceed five hundred dollars or imprisonment for not to exceed one year, or both. Sec. 6. That the expression " radio communication " as used in this Act means any system of electrical communication by telegraphy or telephony without the aid of any wire connecting the points from and at which the radiograms, signals, or other communications are sent or received. Sec. 7. That a person, company, or corporation within the jurisdiction of the United States shall not knowingly utter or transmit, or cause to be uttered or transmitted, any false or fraudulent distress signal or call or false or fraudulent signal, call, or other radiogram of any kind. The penalty for so uttering or transmitting a false or fraudulent distress signal or call shall be a fine of not more than two thousand five hun- dred dollars or imprisonment for not more than five years, or both, in the discretion of the court, for each and every such offense, and the penalty for so uttering or transmitting, or caus- ing to be uttered or transmitted, any other false or fraudulent signal, call, or other radiogram shall be a fine of not more than one thousand dollars or imprisonment for not more than two 246 Appendices. years, or both, in the discretion of the court, for each and every such offense. Sec. 8. That a person, company, or corporation shall not use or operate any apparatus for radio communication on a foreign ship in territorial waters of the United States otherwise than in accordance with the provisions of sections four and seven of this Act and so much of section five as iriiposes a penalty for interference. Save as aforesaid, nothing in this Act shall apply to apparatus for radio communication on any foreign ship. Sec. 9. That the trial of any ofEense under this Act shall be in the district in which it is committed, or if the offense is committed upon the high seas or out of the jurisdiction of any particular State or district the trial shall be in the district where the offender may be found or into which he shall be first brought. Sec. 10. That this Act shall not apply to the Philippine Islands. Sec. 11. That this Act shall take effect and be in force on and after four months from its passage. Approved, August 13, 1912. Index. 247 INDEX. PAGE Abbreviations 6 Absorption, atmospheric 219 Aerial, see Antenna. Alternator, radio frequency 98, 99 Amplification, regenerative 150, 151 Amplifier 71, 72, 73, 80, 153-160, 219 circuits, typical 157-160 classification of 153, 154 magnetic 169 vacuum tube, see Vacuum tube. Amplitude, decrease of, in oscillating current 38 Antenna 131, 219, 220 of airplane 43, 94 circuit, with decreased decrement 133, 134 constants of 210, 211 directional characteristics of 184 resistance 220 series condenser in 139, 140 types of ._ 93, 94, 219, 220 Arc oscillator 105-111 transmitter 105-111 Audibility 221 Audio frequency 47, 226 Audion, see Vacuum tube. bulb, see Vacuum tube. click method of determining resonanse 213-218 control box 162, 163 Autodyne circuits, typical 152, 153 reception 129, 130 Bridging condenser 150, 158 Brush discharge, loss due to 90, 221 348 Index. PAGE Buzzer 33, 45, 46 motor 92. 93 Capacitance of antenna 211 of condenser 18, 19, 208 of a coil 20-24 unit of 18 Capacity, see Capacitance. specific inductive 18, 209 Chaffee gap 102-104 Circuits, antenna 83 coupled ; • • • : 50-62 oscillating 34-39, 75 radiating 39-43 spark 83 Code 235 Coils, choke 30 induction 32 loading 83, 144 Condensers, bridging 150, 158 capacity of 208 discharge of 35, 36 types of 19, 20 Conductor, electrical 11 Continuous waves, production by arcs 105-111 production by r. f. alternator 98,99 production by sparks 100-104 Counterpoise 94,223 Coupling 51-62 close 51 coefficient of ; 62 loose 52 Damped oscillatiqns , 38 Damping, see Decrement. Dead end 141, 142 switch 142 Decrement, calculation of • 204-207 of oscillating circuits 60, 224 see also Log decrement. Index. 349 PAGE Decremeter 206, 207, 225 Detector, crystal 121, 122 definition 117, 225 principles of the simple 117-122 three-element vacuum tube as a 122-127, 129 Dielectric 25 constant 18, 209 strain in 18 strength of 209 Direction finder, see Radio compass. Electricity, nature of 10, 11 Electric field 12 strain 42 Electron, forces acting on, in vacuum tube 65-67 theory . . .• 10-13 Energy of strain 15 Ether 9 strain in 12, 13, 15, 27, 44 waves, length of, see Wave length. velocity of 42, 218 Farad 18 Fields, magnetic 29,30 Filter circuit 161, 162 Fleming valve 67 Frequency, audio 47, 226 radio 47, 226 Grid, see Vacuum tube. leak 77, 115, 116 Ground connection 94 Heat waves 9 Henry 30 Heterodyne reception 127-130 Inductance, unit of 30 variable 140 variable, iron cored 168-171 Induction coils 32 mutual 31 self 27 350 Index. PAGE Inductive coupling 51,52 Inertia, electric 27 Insulators 11 Interference prevention 133-135, 145-147 Ionization oi: upper atmosphere, eilects of 194 Laws relating to radio communication 236-246 Lines of force, electric 12,14 magnetic 13,14 Loading coil 144 Log decrement 38 Measurement, see Chap. XII. of capacity 208,209 of decrement 204-206 of high frequency resistance 201-204 of inductance ' 209, 210 of wave length 218 Microphone 165 Modulation, definition of 167 Modulators 168-170 classification of action of 171 Oscillating circuits, mechanical analogy of 52, 53, 55-59 see also Circuits. Oscillations, damped 38 frequency of 36, 45 undamped, generation of 74, 98-111 Permeability of iron 31 Pick-up circuit 133 Plate, see Vacuum tube. Poulsen arc, see Arc oscillator. Quenched spark gap 88-90, 102-104 Radiation from antenna 40-42 resistance 91,228 Radio compass 188 bilateral characteristics 190 definition of a 222 sources of errors in the 191-194 unilateral characteristics 190 frequency 41, 226 telephone, see Chap. X 164 Index. 251 PAGE Receiving transformer, see Coupler. Relay, use of vacuum tube as 70 Resistance, critical value of, in oscillating circuits 34, 36 Resonance 48,49, 229 curve 60, 61, 224 determination of, by audion click 213-218 transformer 85, 87 Rotary spark gap 87-89 Self-induction 27 Sending apparatus, see name of type. Short wave condenser 139, 140 Spark gaps, function of 37,82 resistance of 37 types of 88-90 Specific inductive capacity, table of 209 Stand-by circuit, see Pick-up circuit. Static, and reduction of its effects 195-198 Strain, see Ether. Symbols 6 Telephone, radio, see Chap. X 164 Tickler coil 149 Transmission formulae 211, 212 Transmitting station, determination of location of 189 Tuning 230 Ultraudion circuit 79 TJndaniped oscillations, generation of, by vacuum tubes .... 74 wave generators, see Continuous waves. Vacuum tube, amplification factor of 68, 69 as amplifier 71-73 use of, as a detector 122-127, 129 as generator of alternating current 74 grid applied to 67 grid circuit of 67 internal (D. C.) resistance of 70 plate circuit of 65 as relay 70, 71 theory of 63 seq. three-element, characteristics of 69 two-element, characteristics of 66 253 Index. PAGE Vacuum tube generating circuits, application of, to antenna, 112-116 grid leak with 115, 116 requirements of 74 typical 78, 79 Wave, broad 62 length 42, 43, 48, 218, 230 in coupled circuits 55-59 meter 45,47,48 propagation 39-43 velocity of 42 trains 46, 230 Winding, banked 22, 23