UC-NRLF B ^ as7 DD3 GIFT OF ROBISON'S MANUAL OF RADIO TELEGRAPHY AND TELEPHONY 1918 ROBISON'S MANUAL OF RADIO TELEGRAPHY AND TELEPHONY FOR THE USE OF NAVAL ELECTRICIANS BY CAPTAIN S. S. ROBISON, U. S. NAVY REVISED BY CAPTAIN D. W. TODD, U. S. NAVY Director Naval Communications AND LIEUT. COMMANDER S. C. HOOPER, U. S. NAVY In Charge of Radio Division, Bureau of Steam Engineering 4th revised EDITION ANNAPOLIS, MD. THE UNITED STATES NAVAL INSTITUTE 1918 ROBISON'S MANUAL OF RADIO TELEGRAPHY AND TELEPHONY Price, $1.50, postpaid Copyright, 1911 BY PHILIP R. ALGER Secretary and Treasurer, U. S. Naval Institute Copyright, 1912 BY RALPH EARLE Secretary and Treasurer, U. S. Naval Institute Copyright, 1913 by E. J. KING Secretary and Treasurer, U. S. Naval Institute oitt Copyright, 1918 BY J. W. CONROY Trustee for U. S. Naval Institute Annapolis, Md. c Z2)t £orb (^aftmore (pr«»g BALTIMORE, MD., U. S. A. PREFACE This manual, first published in January, 1907, was revised in 1909 by L. W. Austin, Ph. D. The present (2d) revision contains the results of some of Dr. Austin's later researches as well as more detailed instructions relative to installation, care and operation, also additional appendices, containing extracts from Service Eegulations adopted at the International Wireless Telegraph Convention in Berlin, October, 1906, and the U. S. Statute of 1910, relative to wireless telegraph apparatus on merchant vessels. The author is also indebted to Mr. J. Martin, of the Navy Yard, New York; Mr. Geo. F. Hanscom, of the Navy Yard, Mare Island; Mr. Geo. H. Clark and others for figures, illustrations and suggestions. July, 1911. PREFACE TO THIRD REVISION Advantage has been taken of the necessity for a new edition to regroup subjects; to add articles and figures relating chiefly to the use of un- damped oscillations; to note the results of some of Dr. Austin's recent re- searches; to alter appendices to include the London Convention of 1912 and the U. S. Statute of 1912, and, through the kindness of Lieutenant S. C. Hooper, U. S. N., to extend the chapter on care and operation. August, 191S. A new edition being necessary, minor corrections and additions to the third revision have been made. December, 19H. PREFACE TO FOURTH REVISION The copies of previous editions having become exhausted, certain changes have become necessary to bring the Manual up-to-date. Owing to the many responsibilities placed upon Captain Eobison, due to the war, the revision has been made by Captain Todd and Lieutenant Com- mander Hooper, who offer the Fourth Edition as Eobison's Manual of Radio Telegraphy and Telephony. Credit is due to Lieutenant H. P. LeClair, U. S. Navy, in charge of the Eadio Division, Bureau of Steam Engineering, and Mr. Charles J. Pannill, Expert Eadio Aide, Office of Director Naval Communications, for many valuable suggestions and their work in connection with the revision of this Manual. January, 1918. it) CONTENTS CHAPTER I. General Review of Facts Relating to High-Frequency Currents. Electricity; Magnetism; Electro-magnetism; Electro-magnetic induction; Relation of positive direction of lines of force to positive direction of current; Methods of producing currents by electro-magnetic induction; Production of electric and magnetic fields stresses and strains in the ether; Electric capacity; Electric and magnetic induction; Electric condensers; Discharge of con- densers; Ether waves; Reflection of ether waves; Refraction; Diffraction. CHAPTER II. Production, Radiation and Detection of Ether Waves. Mutual induction and coupling; Transfer of energy between coupled circuits; The quenched gap; Methods of producing electric waves; Radiation of electric waves; Damped oscillations; Undamped oscillations; Decrease of amplitude with distance from source; Detection of electric waves; Receiving circuits. CHAPTER III. Electric Units and Their Relations to Each Other. Volt; Ampere; Ohm; Watt; Coulomb; Farad; Henry. CHAPTER IV. Capacity and Self-Induction. Fundamental Equation of Wireless Telegraphy; Self-induction; Capacity; Condensers and Inductances in series and in parallel; Combination of self- induction and capacity in oscillating circuits; Capacity and self-induction of straight wires; Time constants of condensers and inductive circuits; Difference between D. C. and A. C. due to self-inductance and capacity; Skin effect of high frequency A. C. ; Measurements of inductance and capacity in oscillating circuits. CHAPTER V. Power ExrENniTURE and Efficiency of Sending and Receiving Apparatus. Mechanical work done in making dots and dashes of the telegraph code; Efficiency of sending apparatus; Losses in condensers; Losses in closed and opened circuits; Relations between height of aerial, Oscillating current. Wave length and distance of transmission; Efficiency of receiving apparatus; Increase of efficiency due to a high spark frequency; Comparison of efficiency using damped and undamped waves; Losses in receiving circuits. 8 . , r - ■ CONTENTS. CHAPTER VI. Sending Apparatus. Generators; Considerations governing frequency of generators; Trans- formers; Regulation of A. C. sending apparatus; Sending keys; Closed circuits (inductance, condenser, spark gap); Condensers; Dielectric strength of air; Spark gaps; Use of the arc for producing undamped oscillations; The Federal-Poulsen System of Radio Telegraphy; The antenna; The ground; The helix; Wave-changing switches, etc.; Signals and Keys; The arc; Receiving; Wireless Telephone transmitters; Limitations on wave lengths; Open circuit (Aerial, inductance, ground); Open circuit inductance; Aerial accessories; Grounds and ground connections. CHAPTER VII. Receiving Apparatus. Navy receiving sets type A; Variable condensers; Condensers in receiving circuit; Inductances in receiving circuit; Detectors; Electrolytic detectors; Rectifying detectors; Vacuum tube detectors; The oscillating audion; Mag- netic detector; Slipping contact detector; Coherers and Lodge-Muirhead detector; Testing buzzers; Receiving telephones; Relays or ampliphones; Recording apparatus; Direction finders; Belini-Tosi wireless compass; Port- able and auxiliary sets; Airplane radio transmitter. CHAPTER VIII. Installations, Adjustments and Measurements. Installations; Protective devices; Adjustments; Measurements, wave meters and their use; Measurements of wave lengths; Instructions for using the Pierce wave meter of the Massachusetts Wireless Equipment Company; (a) Calibration of sending station; (b) Calibration of receiving station; (c) Precaution and care of the instrument; Tuning curves; Resonance and audi- bility curve; Measurement of damping; Measurement of sending current; The shunted telephone method of measuring the intensity (loudness) of signals; Measurement of inductance and capacity and total resistance; The measure- ment of logarithmic decrement. CHAPTER IX. Care and Operation. Calling; Sending; Duplex operation; High-speed operation; To send a message; Receiving; Interference; Static; Codes; International Morse code signals; Abbreviations; Commercial operation by United States Naval Com- munication Service; Sources of information. APPENDICES. ROBISON'S MANUAL OF RADIO TELEGRAPHY AND TELEPHONY Chapter L GENERAL REVIEW OF FACTS RELATING TO HIGH . FREQUENCY CURRENTS. ELECTRICITY. 1. If amber is rubbed with silk a change in the condition of the amber and of the silk is produced which can be detected in various ways. This change in condition is described by saying that the amber and the silk are electrified or charged with electricity by friction. Both of these terms are derived from the Greek word " elektron," meaning amber. The silk and amber thus electrified attract each other and bodies in their vicinity, but the silk will repel another piece of silk similarly electrified and the amber will repel another piece of amber similarly electrified. Since amber and silk have no effect on each other when not electrified, the qualities of attraction and repulsion are said to reside in the electric charges, and the fact is expressed by the statement that like charges repel, unlike charges attract each other. The silk is said to be positively, the amber negatively, electrified or charged. Positive and negative charges are indicated by plus ( + ) and minus ( — ) signs. The charges are said to consist of static or frictional electricity. Bodies thus charged when not brought into contact with each other or with what are called conductors remain in an electrified condition for some time. Bringing oppositely charged bodies in contact generally removes all evidences of electrification. The charges are said to unite and, being of opposite signs, to neutralize each other, and the bodies are said to be discharged. Sparks accompanied by a sharp crackling sound are produced between highly electrified bodies when brought very near each other. After the spark has passed the bodies are found to be discharged. Charged bodies which can be discharged by sparking at greater dis- tances than others are said to be charged to a higher potential. All bodies, whatever their nature, are capable of being electrified. The presence of static charges of electricity can be shown by what are called electroscopes. One of the most sensitive, the gold-leaf electro- 10 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. scope, consists of two small pieces of gold leaf, which, becoming charged in the same sense (i. e., positively or negatively), by touching a charged body, repel each other, and diverge, and show by their divergence the presence of electric charges. 2. Certain bodies, notably metals, have the quality of transmitting or carrying electric charges through themselves and are called conductors. Bodies lacking in this quality, or possessing it to a very limited degree, are called nonconductors, or insulators, or dielectrics, according to the purpose for which they are used. 3. When pieces of zinc and carbon are immersed in a conducting liquid (fig, 1) the combination is called a prima/ry cell. If a wire is connected to the zinc and one to the carbon and the free ends of the two wires brought near each other, these ends are found to be electrified ; the end of the wire connected to the carbon electrified like the silk ( + ) and the end of that connected to the zinc like the amber ( — ). The carbon is called the negative element or positive pole of the cell and the FiQ, 1. Fig. 2. zinc the positive element or negative pole. A number of cells together is called a battery. The liquid in which the elements are immersed is called the battery solution. If the free ends of the wires are brought together an electric current is established, of which the positive direction is said to be from the carbon to the zinc, through the wires ; from the zinc to the carbon, through the liquid. (See fig. 2, and note 1, appendix.) The current is said to be caused by a difference of potential between the carbon and the zinc. It is supposed to be made up of small electric charges transmitted through the wire in quick succession, the charges being produced by chemical or electric action between the carbon and the zinc in the liquid. The force which causes the movement of the electric charges which make up the current is called the electro-motive force and is usually written E. M. F. If the free ends of the wire in fig. 2 instead of being directly con- nected are immersed in another conducting liquid, as in fig. 3, the cur- rent will flow through this liquid. The immersed ends are called electrodes. The one at which the current enters is called the positive and the one at which it emerges the negative electrode. These are also MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 11 called the anode and the catliode, respectively. The conducting liquid in this cell is called the electrolyte. 4. If the anode and cathode in fig, 3 are made of lead (or prepara- tions of lead) plates, and the electrolyte is a solution of sulphuric acid in water, the combination is called a secondary or storage cell or accumvr lator and a number of such cells is called a storage battery. The Fig. 3. Fig. 4. anode is called the positive plate and the cathode the negative plate. If, after a current has been forced through such a cell for a time, the wires from the primary cells are disconnected and the positive and negative plates connected by a wire (fig. 4) outside of the electrolyte, a current will flow, the positive direction of which will be from the positive to the negative plate in the wire, and from the negative to the positive plate in the electrolyte. 5. For convenience, a battery of primary or secondary (storage) cells is indicated as in fig. 5, the elements forming positive poles by the light r'^ 1-. +i Fig. 5. — Cells in Series. Fig. 5a. — Cells in Parallel, lines and the elements forming negative poles by the shorter, heavy lines. Cells connected as in fig. 5 are said to be in series; connected as in fig. 5a, in parallel. MAGNETISM. 6. A magnet situated at a distance from other magnets and pivoted 80 that it is free to move, will point toward the north magnetic pole of the earth, which in some localities coincides with the north star in 13 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. direction. That end of the magnet wliich points in the direction of the north star is called the north-seeking pole, or simply the north pole of the magnet. The other end is called the south pole. Similar magnetic poles, like similarly charged bodies, repel each other. Dissimilar magnetic poles, like oppositely charged bodies, attract each other — i. e., two north poles or two south poles repel each other: a north and a south pole attract each other. The north pole is sometimes called the positive pole and the south pole the negative pole of the magnet. Wrought or soft iron can be magnetized but only retains its magnet- ism while under the influence of the magnetizing force; steel or hard iron once magnetized retains its magnetization permanently, and special means to demagnetize it are required. All bodies can be electrified, but all bodies can not be magnetized. 7. If a sheet of paper is held over a powerful magnet and iron filings sprinkled on the sheet, the filings will assume positions approximately Fig. 6. Fig. 7. as sliown in fig. 6. Some force connected with the magnet must make the filings assume these positions, which are different from what they would be if the magnet was not under the paper; and from the way the filings are arranged, this force must act in the space surrounding the magnet. This space is called the field of magnetic force, or simply the field of force, and the lines in which the filings tend to arrange them- selves are called the lines of force, and we shall see in chapter II that this conception is used as a basis for electric measurements. The direction of the lines of force at any point indicates the direction of the magnetic force at that point, and their number in any area, the strength of the field in that area. It is found that a small magnetic needle, pivoted so that it is free to move and brought near the large magnet, will lie parallel to the direetion of the lines of force at any point at which it may be placed in the field, MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 13 and that the north pole of the needle always points along the lines of force in the direction leading to the south pole of the magnet. The direction in which the north pole of the needle points is called the positive direction of the lines of magnetic force, and the direction in which the south pole points, the negative direction of the lines of magnetic force. Lines of magnetic force are said to run from the north pole of the magnet to the south pole through the air, and back to the north pole through the steel (fig. 7). ELECTRO-MAGNETISM. 8. If the wire in fig. 1 is coiled into a spiral, as in fig. 8, with the positive direction of the electric current as shown by the arrows and Fig. 8. the battery connections, a field of magnetic force which can be explored by a small magnetic needle, or outlined by iron filings, as in fig. 6, will be found to exist around the spiral, and the direction of the lines of force will be found the same as those around the magnet in fig. 7. If the current is reversed, the lines of force are reversed in direction. Such a spiral, when traversed by a current, is found to have all the properties of a magnet, and is called an electro-magnet or solenoid. The strength of the magnetic field around an electro-magnet rises and falls with the rise and fall of the current, and its polarity depends on the direction of the current. The positive direction of the lines of magnetic force which surround a solenoid is from the north to the south pole outside of the spiral, and from the south to the north pole inside of it, just as the positive direction 14 MANUAL OF RADIO TELKGRAPHY AND TELEPHONY. of an electric current is from the positive pole to the negative pole out- side of a battery and from the negative to the positive pole inside of it. If the number of turns of the spiral is reduced to one it does not lose its magnetic character. The lines of force then form circles around the wire, tlieir positive direction being shown in fig. 9, the upper side being a north pole and the under side a south pole. If the turn is straightened out, as in fig. 10, the lines of force still form circles around the wire, and the north pole of the exploring needle points in the positive direction of those lines. This direction is found to be always at right angles to the wire. 9. It appears from the foregoing that what is called the positive direction of motion of electric currents, or charges, is related to what is called the positive direction of the lines of magnetic force, in the manner shown by the arrows in figs. 8, 9, and 10, and, further, that the terms positive and negative as applied to electric and magnetic effects. Fig. 9. Fio. 10. and so largely used in connection with them, are purely conventional. (See note 2, appendix.) 10. Eeturning now to the statement in article 8 that the strength of the magnetic field around a solenoid rises and falls with the strength of the current, and its polarity (i. e., the direction of the lines of mag- netic force produced) depends on the direction of the current^ it can be further stated that a magnetic field exists around every wire carrying an electric current (fig. 10). The direction of the lines of force in this field depends on the direction of the current. These lines of force always enclose circles in planes at right angles to the wire. 11. Since a current is conceived to he made up of a quick succession of moving electric charges (art 3), the above facts may be stated in another way, viz., that moving electric charges produce magnetic fields in which the lines of magnetic force enclose circles in planes at right angles to the direction of motion of the moving charges. This has been proved to be true for single static charges.* ELECTRO-MAGNETIC INDUCTION. 12. Fig. 11 represents a primary battery, with the two poles of the battery connected by a conducting wire, broken at K. A straight portion * By Professor Rowland, Johns Hopkins University. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 15 A B of this wire is parallel to, and at a distance from another conducting wire C D. When the break at K is closed, a current flows in the circuit, and a field of force is created around the wire. Let ys consider the straight portion A B in which the direction of the current is shown by the arrows, and the direction of the lines of force by the circles (shown as ellipses), at right angles to A B. Several of these lines of force are shown embracing the parallel wire C D. If gold-leaf electroscopes (art. 1) are attached to the ends C and D of the wire C D, and if the current started in A B when the break is closed is sufficiently powerful, the gold leaves will be observed to diverge, momentarily, whenever the circuit is made or broken at K. The stronger the current in A B, and consequently the stronger the magnetic field produced, the more pro- nounced the indications of the electroscope will be. This shows that the ends C and D of the wire C D are electrified when the current is made or broken in A B. When the current is made the end D is negatively charged like the amber and like the wire attached to the zinc element in fig. 1, the end C positively, like the silk and like the wire attached to the carbon element in fig. 1. When the circuit is broken at K the electrification of C D is reversed, C becoming negatively and D positively electrified. A sudden increase or decrease of the current in A B produces the same effect as when the current is made or broken. It is to be noted that the electrification of C D is only momentary. As soon as the causes producing it are removed, the electric charges unite and neutralize each other through the body of the conductor. We know that when the current in A B is made, a magnetic field is created around A B which extends to and beyond C D, and that when the current in A B is broken, the magnetic field disappears, and that the only thing common to A B and C D is this magnetic field, the lines of force in which surround them both, and since we see that one kind of electrification is produced in C D when the lines of force are being created, and the opposite kind when they are being dissipated, we con- clude that the movement or creation of these lines creates the electric charges that we observe in C D. 13. In art. 11 it is stated that moving electric charges create magnetic lines of force. N"ow, we see the truth of the converse, viz., that moving magnetic lines of force create electric charges. These two facts are of general application and are the basis of all electro-magnetic calculations. 16 MANUAL OF RADIO TELEGRAPHY AND TELEPHONT. 14. It is of great importance to keep clearly in mind the fact that electrification in C D only takes place when the current is made or broken or changed in A B. When there is no current in A B there are no magnetic lines of force, and consequently there is no electrification in C D. When there is a constant current in A B the magnetic field is constant and there is no electrification in C D. It is while the current in A B is rising or falling, and the lines of force expanding from or contracting toward A B and cutting through C D as they pass, that C D is affected. A movement of the lines of force is required to electrify C D, and this movement is produced by changes in the current in A B. If the ends C and D are joined to form a complete circuit, a momen- tary current will flow when changes in the magnetic field around C D take place. We have just seen that a moving magnetic field in the vicinity of C D creates electric charges in C D. We would also find that moving C D in a magnetic field has the same effect. The change of current in A B is said to induce the current in C D, and the action is called electro- magnetic induction. The preceding facts can be stated as follows: When magnetic lines of force cut or are cut by a conductor, electric charges (i. e., a tendency to current flow) are induced in the conductor, and currents flow if the conductor forms a closed circuit, the direction of the induced currents depending on the direction of cutting. 15. When the current in A B is rising, tlie magnetic lines of force are expanding, and cutting C D in the direction from left to right, the direction of the momentary current in C D being as shown in fig. 11a. Fig. 11a. — Current in A-B Rising. B D Fig. 11b.— Current in A-B Falling. When the current in A B is falling, the magnetic lines of force are contracting, and cutting C D in the direction from right to left, the direction of momentary current in C D being shown in fig. lib. These momentary currents or movements of electric charges in C D themselvee MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 17 produce momentary magnetic fields around C D, the direction of the lines of force of which are shown by the arrows in figs. 11a and lib. It will be seen that the lines of force around C D, when the current in A B is rising, are opposite in direction to those created when the current in A B is falling. The field of force created around C D reacts upon A B, tending to create in A B a current in the opposite direction to that already in A B, i. e., to stop it. In other words, the change of primary current in A B induces a secondary current in C D. The latter current in turn induces a tertiary current, which is in A B. This influence of two currents on each other is called their mutual induction. 16. The electric charges produced by friction (art. 1), by chemical action (art. 3), and by the movement of lines of magnetic force are all identical in their properties, and the magnetic fields produced by the movement of these charges are also identical in their properties. It is therefore evident that a very close relation exists between electricity and magnetism. Pig. 12. v'' (^V^Z-^^nV' '/, Fig. 12a. Fig. 12c. Fig. 12. — Electric Field Charges of Opposite Sign. — Attraction. Fig. 12a. — Magnetic Field between Unlike Poles. — Attraction. Fig. 12b. — Electric Field Charges of Same Sign. — Repulsion. Fig. 12c. — Magnetic Field between Like Poles. — Repulsion. 17. We have seen that the field of magnetic force around a wire carrying a current or around a magnet can be mapped out by iron filings. In a similar manner the field of electric force around charged bodies can be shown by the use of a dielectric, such as powdered mica. Figs. 12 and 12b show the electric field between unlike and like charges, respectively. Figs. 12a and 12c show the magnetic field between unlike and like poles, respectively. The electric field between two 2 18 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. charged bodies is found to resemble very closely the magnetic field be- tween magnet poles.* In all figures it can be seen that in electric as well as magnetic fields each line of force appears to repel its neighbor, and that they have their ends on points of opposite electrification or magnet- ization. If these lines tend to shorten in the direction of their length this tendency will cause the attraction between the bodies from which they proceed. 18. It may be asked, — what are these lines of force which are not visible and which can not be physically grasped? The only reply is that we believe all electric and magnetic phenomena to be the results of the disintegration of the atoms of matter or the rearrangement of their constituent parts (see note 2, appendix), the movements of which produce stresses and consequent movement or strains in what is called the ether, an almost infinitely elastic, infinitely tenuous substance which surrounds and permeates all matter and all space. The earth is immersed in an illimitable ocean of ether, just as fishes are in water. We move about in a sea of it. "What we call electric and magnetic fields are places where ether move- ments and ether stresses can be detected by the phenomena which they produce, and which are being described. An electric field is a state of strain (stretch or compression) in the ether; it can be removed between any two points by connecting them with a conductor. The release of the strain starts movements of electric charges in the conductor. Movements of these charges produce another state of strain in the ether at right angles to the first. We call this a magnetic field. We have seen that movement of either field creates the other, and that the lines of force in the two fields, when they are thus produced, are in planes at right angles to each other. When equilibrium is restored one field or the other has disappeared, though they can coexist in a transitory state. 19. It lia^ been proved thai light and heat are forms of ether motion also, and that all movements (electric and magnetic) in the ether are propagated with the velocity of light. It has also been proved that electric movements progress along straight wires at practically the same speed that magnetic movements progress at right angles to them — i. e., with the speed of light. This velocity has been measured many times and found to be 186,000 miles, or approximately SOOfiOOjOOO meters per second. * The direction of the electric lines of force at any point in the field can be determined by suspending in it a small piece of a dielectric, such as a glass fibre. The dielectric will lie parallel to the direction of the lines of force at the point of suspension. MANUAL OF RADIO TELEGRAniY AND TELEPHONY. 19 We must learn therefore to think of light movements and of electric and magnetic actions not as being instantaneous, but as being restricted to a definite measurable speed. It takes time for electric and magnetic effects to be propagated in the ether, time for them to be propagated along a wire. The wire guides or strikes out the line of maximum disturbance. 20. Let us now return to fig. 11. Before connection at K is made, the field of magnetic force does not exist, but the wires are electrified by means of action between the zinc and carbon in the battery solution. When the break at K is closed, a magnetic field is established; when the connection at K is broken, the magnetic field disappears. The question arises, — how is this magnetic field created? How is it dissipated? The reply is: It is created by movements of electric charges in A B which disturb the ether and this disturbance is propagated through the ether at right angles to A B, with the speed of light, i. e., at the rate of 186- 000 miles or 300,000,000 meters per second. This disturbance is of such a nature as to produce a state of strain in the ether which may be compared to that produced in a piece of rubber by compression or tension. The strain is relaxed as soon as its cause (i. e., the movement of the electric charges) is removed. The amount of strain (i. e., the strength of the magnetic field) decreases as the distance from the moving charges increases. It spreads in all directions, but except with very delicate instruments can not be detected at any great distance from A B. The creation and dissipation of this state of unstable equilibrium in the ether, which must be brought about by some kind of movement in it, produce electrical movement in C D, or, as it is perhaps better to say, produce electric charges in C D. CD stands in the way of and is disturbed by an advancing or receding wave of movement in the ether, originated at A B. CD is, like all other conductors, an obstacle in the path which creates an eddy, so to speak, in the ether wave and reacts, however minutely, on A B, because the movement of the electric charges produced in C D also creates an ether movement, but in the opposite direction to that proceeding from A B. 21. We have now reached a point where the electric and magnetic actions under discussion are directly applicable to wireless telegraphy, but, before proceeding with this subject, it is desirable to consider more fully the action of A B on C D ; because the creation of electric currents by moving or varying magnetic fields, and vice versa, is the basis of industrial electric power — of that used in wireless telegraphy as well as in other branches of electricity; and other facts or developments of this fundamental fact will appear which will lead to a clearer compre- hension of it. 20 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig 22. In fig. 11, C D is shown parallel to A B. ^x If C D is slowly revolved around its own center ^5-r^^ ^^^^ as an axis the effect on it of making, breaking, '','?'/' crtffC-^^i^^^f^mF^, or changing the current in A B will be found to decrease until C D is at right angles to A B, ,|^ when it will disappear altogether. The lines of force are circles at right angles to A B ; they do not cut C D when it is at right angles to A B because it is parallel to them, and consequently no effect is produced. The induced effects in C D will be found to increase as it is brought nearer A B and to decrease as it is removed from A B. The field near A B is stronger, and more lines of force are created there or dissipated there than at a greater distance from A B — i. e., a greater disturbance in the ether takes place. 23. If the two ends of C D (fig. 11), are brought close together, but without touching, and if the current made or broken in A B is very strong, a spark will pass between the ends of C D at each make and break. If C D is separated from A B by an opaque, nonmetallic screen and the makes or breaks in A B are made to represent the characters of a code, messages sent in this code from A B can be received at C D when each is invisible from the other. By the addition of a battery to C D, similar to that producing current in A B, replies can be sent, and thus a crude wireless telegraphy produced. 24. If A B is coiled into a spiral or helix and C D into a similar spiral or helix (fig. 13), the effect of making, breaking, or changing the current in A B is much greater than when both wires are straight; for the disturbance created in the ether — that is, the number of lines of force produced by the moving charges in A B — is equal, for equal lengths of the wire, and since a greater length is concentrated in the same space, the number of lines of force in that space, assuming the current in the spiral to be the same as that in the straight wire, is correspondingly greater. This stronger field would produce an increased effect on a straight wire; but the length of C D is also concentrated. Therefore the effect is increased still more. 25. We know that A B when coiled as in fig. 13 and traversed by a current forms a solenoid (art. 8, fig. 8). The space inside the coil is called the core, and it has been assumed that the surrounding substance (excluding the ether, which is present both in the interior and on the exterior of all bodies) is air. It is found, however, that if the core of the solenoid is iron, as in fig. 14, instead of air, the effect on C D is very much more powerful — i. e., the number of lines of force created with the same current is very greatly increased. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 21 This shows that it is easier to create lines of force in iron than in air; or, to state the fact differently, lines of force permeate iron more easily than they do air. The relative ease with which magnetic lines of force are created in a substance is expressed in figures and called its magnetic permeability. The permeability of air at atmospheric pressure is called CUKf^ENT IN A-B RISING^ =g^ unity, and on that basis the permeability of the purest wrought iron is 3,000. In other words, within limits the same current will produce 3,000 times as many lines of force in iron as in a body of air of the same length and area of cross section. CURRENT -< c ->- D Fig. 14. 26. If the iron of fig. 14 is extended to include C D, as in fig. 14a, the effect of changes in A B is increased still more, because in fig. 14 the lines of force are partly in iron and partly in air, while in fig. 14a they have an iron path throughout, and are consequently much greater in number. C D can also be placed inside of A B or outside of it, with or without an iron core (figs. 14b and 14c). 22 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 14a. — Closed-Core Transformer (current in A B falling). <"- ^ 1 Fig. 14b. — Open-Core, Step-Down Transformer or Induction Coil (current in A B rising). C Fig. 14c. — Air-Core, Step-Up Transformer (current in A B rising). -< A C "B D »- Fig. 14d. — Auto Step-Down Transformer (current in A B rising). MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 23 27. Since the tendency to current flow in C D is produced by lines of magnetic force cutting C D, and since on making or breaking cur- rent in A B each line of force cuts C D once, for each turn in C D, if the turns in C D are decreased or increased, as in figs. 14b and 14c, the tendency to current flow — i. e., the electro-motive force — is raised or lowered. From this fact, and from the fact that the current in C D is opposite in direction to that in A B, the arrangements in figs. 14a, 14b, and 14c are called transformers. Fig. 14a is called a closed-core trans- former; fig. 14b an open-core transformer or induction coil; fig. 14c an air-core transformer. Transformers are called step-up or step-down with reference to whether the turns in the coil C D are greater or less than those in A B. Fig. 14b is a step-down; fig. 14c a step-up transformer. The coil A B is called the primary and the coil C D the secondary winding, and where A B and C D have some turns common to both, as in fig. 14d, the arrangement is called an auto-transformer. =^ Fig. 13. 28. Keferring again to fig. 13 : When the break at K is closed, a current is started, which progresses upward through the coil, the mov- ing charges composing it creating a magnetic field around the wire. The lines of force, as they expand from the current in the first turn of the spiral, cut the second turn of A B in the same way that they cut C D a little later. They induce a current in the second turn opposite in direction to that in the first turn — i. e., tending to stop it. The same effect is produced in the third and succeeding turns. In other words, the different parts of the coil A B react on each other just as the coil C D reacts on A B. This reactive effect of the turns on each other makes the rise in current slower than in a straight wire, and is greater when the core of the coil is of iron than when it is of air, because of the greater number of lines of force produced. 24 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 29. We find that a stronger current is produced by the same battery in a short wire than in a long wire of the same size and material, and in a thick wire than in a thin wire of the same length, and we say that this is due to the greater resistance of the long wire and of the thin wire as compared with the short or with the thick wire. To establish the same current in the longer or the thinner wire as in the shorter or thicker wire requires a larger battery — that is, greater E. M. F. 30. Now, we find that when the wire is coiled into a spiral and a change in the current is taking place, the turns react on each other and resist the change of the current. This resistance does not depend on the size nor the material of the wire, but only on the amount and quick- ness of the change in the current, and is therefore of a different character from the resistance referred to above. The resistance of a wire to changes in current established in it is called its reactance, and during the change the total effect of the true resistance and the reactance is called the impedance of the wire or circuit. In circuits having reactance tlie production or progression of electrical effects is retarded. It takes longer to create a given current than in the same length of straight wire. It may be said, therefore, that coiling a wire increases its electrical length — i. e., increases the time it takes an electrical movement created at one end of it to reach the other. The currents in C D are said to be produced by the induction of A B on C D. The retarding effect of the coils in A B to the rise and fall of current in A B is said to be due to the self-induction of A B. It has been shown that the amount of both kinds of induction depends on the shape and arrangement of both circuits and the material (iron or air) in and around them. RELATION OF POSITIVE DIRECTION OF LINES OF FORCE TO POSITIVE DIRECTION OF CURRENT. 31. The currents under discussion have been illustrated as being pro- duced by batteries of primary cells, and for many purposes these are very valuable, but for the production of very powerful electrical effects advantage is taken of the fact, stated in art. 14, that when magnetic lines of force cut or are cut by a conductor, electric currents flow in the conductor, if the latter forms a closed circuit. The direction of current flow can be determined by the following rules :* (a) Fig. 15. An increase in the number of lines of force embraced by a circuit induces a current in the opposite direction to that in which the hands of a watch move, while a decrease in the number of lines of force induces a current in the same direction as that in which the hands of a watch move, the line of sight being in both cases along the positive direc- • Rule (a) from Fiske's "Electricity and Electrical Engineering." MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 25 tion of the liues of force. (Art. 7 and fig. 7.) (b) The positive direc- tion of the lines of force is with the hands of a watch when the current is flowing away from the observer. Fig. 15a. (c) The currents induced by moving lines of force always tend to prevent change in the inducing current. Induced currents are, therefore, in the opposite direction when the inducing current is rising and in the same direction as the inducing current when the latter is falling. (Art. 15.) From rules (b) and (c) can be deduced the following: Let fig. 15b represent a conducting wire C D below a line N S, which represents a field of force and its direction. When in the relative positions shown, movement of either wire or line of force toward the other creates a current to the rear, moving either one away from the other creates a current to the front. / : N £. JT^ 7 =j D I r D D Fig. 15a. Fia. 15b. Fig. 15. It will be seen that the field N S can be revolved through any angle around the wire C D as an axis so as to be to the right, left, helow or in any intermediate position without changing the truth of the above statement. These three rules show the relation between what we call the positive direction of the lines of magnetic force and what we call the positive direction of electric current. METHOD OF PRODUCING CURRENTS BY ELECTRO-MAGNETIC INDUCTION. 32. Now let the wire C D in fig. 11 be bent until it forms a rectangle, and let it be placed in the magnetic field between the north and south poles of a powerful electro-magnet having an iron core. By bending the core into the shape shown in fig. 16, the north and south poles are oppo- site each other and more lines of force are produced, because the distance 36 MANUAL OF RADIO TELKGRAPHY AXD TELEPHONY, they have to travel through the air is very much shortened as compared with fig. 14. Exploration of the field in fig. 16 by means of iron filings or by means of a small magnetic needle will show that the lines of force extend directly from a point in the north pole to the opposite point in the south pole. In other words, that they are straight and parallel to each other, and they are so shown in fig. 16. The field is also found to be of uniform intensity, which indicates that the lines of force are equally dis- tributed throughout its area. Now, if C D is moved up or down in the magnetic field, no indica- tion of a current can be perceived, and it appears that the statement in art. 14 (that when magnetic lines of force cut or are cut by a conductor ^kMb Fig. 16. electric currents flow in the conductor if the latter forms a closed circuit) is in error, but when we consider that when C D is moved upward (the field being of unifonn intensity) as many lines of force are cut" by the bottom half as by the top half of C D, the currents induced in the two halves must therefore be equal, and since both flow to the rear we see that they neutralize each other, and the result is zero. Another way to explain this is to consider that portion of the field inclosed by C D as containing a certain number of lines of force. Those coming in when C D is moved induce a current in one direction, those going out induce a current in the opposite direction, and if as many come in as go out no effect is produced. 33. If C D were straight, electric charges would be produced on its ends and would be maintained there as long as the cutting of the lines of force continued, but bending it into a closed circuit changes conditions MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 27 to the extent that cutting of lines is going on all around the circuit, Boine inducing charges in one direction, some in the other, and it is only when there is a preponderance of cutting in one direction that a current actually flows. This would occur if C D were moved from a point where the field is weak to where it is stronger, or vice versa, but the field under discussion is supposed to be uniform. (See rule a, art. 31.) If C D is rotated around one of its diameters as an axis (say the hori- zontal diameter at right angles to the lines of force) when it is hori- zontal, as in fig. 17, the lines of force included will be zero, and when vertical, as in fig. 17a, the lines of force included will be the maximum number possible in that field, so that a revolution of 90° will make an entire change in the number of lines of force passing through the rectangle. For instance, if the revolution is in the direction of the hands of a clock — i. e., if the top of C D moves to the right (see fig. 17a) — the upper half of C D is cutting lines of force in the direction which induces movements of electric charges to the front, while the lower half is cutting Fig. 17. Fig. 17a, lines of force in the direction which induces movements of electric charges to the rear, so that an electric current is established in C D in the direction shown, the number of lines of force included in C D is decreasing, and looking from N to S, the current moves with the hands of a watch. If C D's rate of revolution is constant, a little consideration will show that when it has revolved through 90° and its plane is horizontal it is then moving at right angles to the lines of force, and consequently cutting them faster than when, its plane being vertical, it moves parallel to the lines of force for an instant and is not cutting any; also that the increase in the rate of cutting is progressive from one position to the other. It will therefore be seen that the electric current produced is a maximum when C D is horizontal, and that it is zero for an instant when C D is vertical because during that instant it moves parallel to the lines of force and therefore it cuts none. (No change in number included.) It is also evident that the increase of the current from zero to a maximum is progressive during the first 90° of revolution, that it then progressively decreases until C D has revolved through 180°, and IB again moving parallel to the lines of force when it falls to zero. 28 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. As the revolution continues, that half of C D which during the first half revolution was cutting lines of force in such a manner as to induce a current to the front, now cuts them in such a manner as to induce a current to the rear, its former place being taken by what was originally the lower half, so that the direction of current in C D is reversed. (Eule c.) Another maximum rate of cutting lines of force and consequent maxi- mum of current is produced when C D has revolved through 270°, The current progressively increases from 180° to 270° and then decreases until when the original conditions are restored by the completion of one revolution the current has again fallen to zero. From the above and from an inspection of fig. 17a it will be seen that current is always flowing to the front in that half of C D which is going down to the right and to the rear in the half going up on the left, and that each half revolution the current changes in direction. Such a cur- rent is called an alternating current. Fig. 18. 34. This can be shown graphically in fig. 18, where the rate of cut- ting and therefore the rate of change of number of lines included in the circuit at different equidistant points in one revolution is represented by equidistant vertical lines proportional to the cutting rate, and conse- quently to the current strength. Vertical lines above the horizontal line represent current strength in one direction and, below it, current strength in the opposite direction. A regular curve is produced by joining the tops of these lines. This curve is the curve of sines, because the rate of cutting and the strength of the induced current are proportional to the sine of the angle of revolution.* * Since the lines of force are horizontal, the number cut during the revolu- tion of C D through any angle is proportional to the vertical movement of the extremity of the radius of C D which generates the angle. The amount of this vertical movement is the sine of the angle, and therefore the induced current is proportional to the sine of the angle; also and for the same reason the induced E M F which produces the current is proportional to the sine of the angle of revolution. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 29 35. If C D, instead of forming a closed circuit entirely in the mag- netic field has its ends connected to two rings which revolve with it and touching these rings are the ends of a coiled wire (E F, fig. 19), the currents induced in C D also flow through E F and make of it a sole- noid whose strength varies with the strength of the current and whose polarity reverses with the reversal of the current. If a small magnetic needle were pivoted in E F, its direction would tend to change with each reversal of the current, and it can thus be made to indicate both the direction and the amount of current flowing through the coil E F. Such an instrument is called a galvanometer. The currents in the coil E F are supplied from C D, and they are induced in C D by its movement in a magnetic field. C D has become a source of electricity like the battery in A B. E F corresponds to the coil A B in fig. 13, and the rise and fall of current in E F will produce a rise and fall of current in another coil near it, just as the make or break at K in fig. 13 induces a momentary current in C D. The currents in C D, fig. 13, were induced by interrupted current. Those induced by E F in coils near it are induced by alternate current. Interrupted current was used almost entirely in wireless telegraphy in its earlier development. It has now been replaced by alternate current. 36. It only remains now to make C D produce the magnetic field in which it revolves, and we can dispense entirely with the primary battery in A B. This can be done as follows : In fig. 20, instead of having each end of C D connected to a ring of conducting material, as in fig. 19, one ring is removed and the other split into two equal parts and an end of C D connected to each part, the ends of E F being adjusted so that, as the split ring revolves with C D, one end of E F is always connected Fia. 20. 30 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. through the split ring with that half of C D in which the current is flow- ing to the front and the other end to that half in which the current is flowing to the rear. This arrangement makes the current in E F always flow in the same direction. It rises and falls with the current in C D, but does not reverse, because just as the current reverses in C D, E F changes ends, so to speak, by breaking connection with one half of the split ring and making connection with the other. The current in E F is now said to be a pulsating instead of an alternating current, and the change can be graphically represented by transferring the part of the curve below the line in fig. 18 to a corresponding position above it, as in fig. 18a. Fig. 18a. The alternating current in C D is said to be rectified into a direct current in E F. The split ring by means of which it is rectified is called a commutator, and the entire apparatus (either with or without a com- mutator), a dynamo. 37. With a single coil, C D, rotating in the magnetic field, the current in E F can be made to flow always in the same direction, but in order to make it constant a large number of coils, equally spaced, must be used, so that one of them is passing through the position (horizontal) in which maximum current is produced practically all the time. If there were 10 such coils, each connected to its own split ring (fig. 21), and all connected to E F, the currents in each would overlap, so that the resultant current in E F (to another scale) might be indicated by a line joining the highest point of each (fig. 18b). In other words the current in E F is practically constant. Fig. 18b. The revolving coils are held in place on a cylindrical drum or ring and the whole is called an armature. If this ring is made of iron, the strength of the magnetic field is much increased, because the iron affords a path for the lines of force from one pole to the other and thereby lessens the distance through which they have to pass in the air. (See art. 25.) MANUAL OF GADIO TELEGRAPHY AND TELEPHONY. 31 The tendency to current flow in C D created by cutting lines of force is called the electro-motive force in C D (see art, 3), and is found to depend on the number of lines cut in a given time, so that the faster C D revolves, and the stronger the magnetic field, the greater the electro- motive force and the greater the current produced in any given circuit. Now, if the current induced in C D, instead of all flowing through E F, is divided, so that part of it flows around the core of the electro-magnet (fig. 21), this current can take the place of that produced by the battery in A B and the battery can be dispensed with. Fig. 21. 38. In art. 6 it is stated that wrought or soft iron can be magnetized, but only retains its magnetism while under the influence of the magnet- izing force. Steel or hard iron once magnetized retains its magnetiza- tion permanently and special means to demagnetize it are required. It is found that electro-magnets with soft-iron cores can be made more powerful (i. e., will give a stronger field) than if the cores are of steel, and that electro-magnets with either kind of core can be made to give much stronger fields than any permanent magnet. Also, that soft-iron cores retain a very small part of their magnetism and polarity when the current is broken, so that, if the magnet poles between which C D revolves are made of the most efficient material (wrought iron or mild steel containing no phosphorus), when C D stops they still retain their polarity in a slight degree. When C D starts to revolve again, the weak field generates a small cur- rent in C D, which sends this current through the wire around the poles; this current increases the strength of the poles and consequently of the field, which increases the current in C D and so on. This is called generating or building up, and continues until the limit of the power moving C D in the continually strengthening field is reached, or until the iron core is saturated, in which condition no increase of current will increase the field. 32 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 39. When alternating current is desired, a dynamo, in order to be self-exciting, i. e., to produce its own field, must have part of its cur- rent rectified by means of a commutator. It is more usual, however, to drive a small, direct-current dynamo by means of the same power which drives the larger one, the current from the small dynamo being used to create the magnetic field in the larger one. Such a machine is called an exciter. 40. The fact that magnet poles of unlike polarity attract each other (art. 6) applies to electro-magnets, with or without iron cores, as well as to permanent magnets. Hence two electro-magnets placed as in fig. 13 will attract or repel each other according to their polarity. Each line of force apparently tends to contract in the direction of its length, and by so doing exerts a mechanical pull on the conductors which it sur- rounds. The same effect is observed between a magnet and a wire carrying a current (which, as we know, has a magnetic field around it) and between two wires, each carrying a current. They actually pull or push each other according to the quality of their magnetism, which is determined by the direction of the current. 41. If in fig. 21 the armature instead of being revolved to the right by some outside agency, is supplied with a current flowing through it in the same direction as the current it generates, it will revolve to the left. The current flowing to the front in the winding of the right half of the armature and to the rear in the winding of the left half, makes of the armature an electro-magnet with a north pole at the bottom and a south pole at the top. The revolution is caused by the attraction of the north pole of the armature by the south pole of the field m.agnet, and its repul- sion by the north pole of the field magnet. This action is reversed in the south pole of the armature. The movement will be continuous, because, as the top of the arma- ture moves toward the north pole of the field magnet, the commutator acts to maintain the flow of current as before, and the consequent arma- ture poles are always at the top and bottom halfway between the field magnets. The armature thus creates a current when made to revolve, and revolves when supplied with current. In the first instance we have seen tliat the entire machine is called a dynamo; in the second it is called a motor. Every dynamo will run as a motor if supplied with current. Every motor will act as a generator or dynamo if made to revolve in its own field. The motor can be made to drive another armature in another field. Such a machine is called a motor-generator. It can be run with direct or alternating currents and made to srenerate direct or alternating cur- MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 33 rents of a higher or lower E. M. F. For this reason it is sometimes called a rotary transformer, as distinguished from the stationary trans- formers already described. 42. Electricity produced by friction (art. 1) is sometimes called fric- tional electricity; by primary batteries, voltaic electricity; by electro- magnetic induction, dynamic electricity. But however produced and transformed, all kinds of electricity are identical, and the same is true of all kinds of magnetism. PRODUCTION OF ELECTRIC AND MAGNETIC FIELDS STRESSES AND STRAINS IN THE ETHER. 43. Wherever there is an electric charge, stationary or moving, emanat- ing from the charge are electric lines of force which end at other electric charges and which form electric fields. Wherever there are moving electric charges (currents) there are magnetic lines of force also, which form magnetic fields and these magnetic lines of force are always at right angles to the direction of the motion of the electric charges and to the electric lines of force proceeding from them. Motion, or state of strain in the ether, which these lines of force repre- sent, travels with the speed of light, and the fields of force, while more pronounced and therefore more easily detected near the moving charges, are really all pervasive. They have no limits. Imagine a disturbance — say an expansion of a gas — to take place in the center of an immense rubber ball. A wave of tension, which becomes less as its distance from the center increases, progresses outward through the rubber to the farthest confines of the ball. When the gas contracts, a wave of contraction, also starting from the center, and decreasing with its distance from the center, progresses outward through the rubber to the farthest confines of the ball. If expansion and contraction are equal the ball's former state of equilibrium is restored. In this way it can be imagined that starting a current produces a stress which strains the ether or stretches it in one direction; stopping it releases the strain. Action in both cases starts at the point where the current is produced and progresses outward with the speed of light, and a little consideration will show that it can have no limit, though it soon ceases to be perceptible except under certain conditions, to be later described. The function of wireless telegraphy is to produce these ether move- ments at will. ELECTRIC CAPACITY. 44. We can produce momentary currents in conductors, whether open or closed, by the cutting of lines of force, and the evidences of electrifi- cation are most pronounced at the ends of an open conductor, but these 3 34 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. disappear as soon as the cutting of lines of force ceases. We find, how- ever, that electrification of amber, glass, silk, and other bodies remains after the rubbing ceases. We can produce static electricity on conduc- tors by suitably insulating them. For instance, if two metal plates separated by a piece of glass are connected, one to the positive, and the other to the negative pole of a source of E. M. F. and then simultaneously separated from it, they will be found to be electrically charged. When two plates oppositely charged (art. 1) are connected through wires lead- ing to a galvanometer, the amount of deflection of the galvanometer needle (caused by the magnetic field of the momentary current created as the charges unite and neutralize each other) is a measure of the quan- tity of electricity on each plate. In testing plates of different sizes, shapes, and materials, charged to the same potential by being connected to the poles of the same source of electricity, it is found that different values of the throw of the gal- vanometer needle are produced. Other conditions being equal, plates having the greatest amount of surface are found to have the largest capacity/. Plates of the same capacity will give a larger throw of the galvanometer when charged from a source of high than a source of low potential, so that the amount of electricity stored in an electrified body depends on its potential as well as on its capacity. - fl + H|l|l|lh Fig. 22. Fio. 22a. 45. If two plates, oppositely charged by being connected to the poles of a battery, as in fig. 22, or to the terminals of a dynamo or transfonner are discharged by being connected through a galvanometer, the throw of the galvanometer will not be as great as if the same plates, charged to the same potential by the same battery as in fig. 22a, are discharged through the same galvanometer. By being brought closer together the plates seem to have their capacity increased. It takes a greater amount of electricity to bring them to the same potential than when farther apart. If two plates, charged at a distance from each other, as in fig. 22, and then disconnected from the battery are brought to the position shown in fig. 22a, their potential, as measured by an electroscope, is found to be lowered. The electricity is said to be condensed by the approach of the plates, and such an arrangement is termed a condenser, a somewhat misleading term, but one generally used. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 35 This is analogous to the increased strength of magnetic field produced by shortening the magnetic circuit while retaining tlie same magnetizing force. In both cases the field of force represents stored energy which can be made to reappear in the discharge of the condenser or the dissipation of the field. ELECTRIC AND MAGNETIC INDUCTION. 46. Electric lines of force permeate a nonconductor — i. e., electric induction takes place through it, — in a way analogous to that in which magnetic induction takes place through iron or air. (See note, p. 18.) The permeability of air for magnetic induction is taken as a standard and called unity. (See art. 25.) Its permeability for electric induction is also taken as a standard and called unity, and as we find that iron, nickel, cobalt, and oxygen have a greater magnetic permeability than air, so we find that glass, beeswax, paraffin, nearly all kinds of oil, and indeed most bodies we call insulators, have a greater electric permeability than air. The quality of a body as compared with air in this respect is called its specific inductive capacity, and bodies when considered with reference to electric induction through them are called dielectrics. (Art. 2.) It is found that the best quality of glass has nine times the specific inductive capacity of air. This means that when subjected to the same potential, the electric field, when this glass is the dielectric, is nine times as strong as that created when the medium intervening between the charges is air, it requires nine times as much work to create it, and its discharge can do nine times as much work. 47. Bodies such as iron or nickel through which magnetic induction is taking place are found to change very slightly in shape, and sudden changes in the induction or lines of force permeating them produce slight sounds. The action is also accompanied by the production of heat, but as the magnetizing force (magneto-motive force) increases, the lines of force tend to read a maximum which no increase of mag- netizing force will increase. When in this condition the magnetized body is said to be saturated. There is, however, apparently no limit to the magnetization of air. In the same way bodies (dielectrics) through which electric induction is taking place are found to change (enlarge) slightly in shape, but increase of electro-motive force (in this case potential) does not appear to tend to a maximum of electric induction. The physical strain on the dielectric, however, continues to increase and finally reaches a point where it pierces or ruptures the dielectric, the action being accompanied by a sharp crackling sound and by the production of light and heat, 36 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. which we call an electric spark. If the dielectric is air or a liquid, the rupture is immediately repaired by the action of the surrounding sub- stance on that heated by the passage of the spark ; but if the dielectric is a solid the rupture is permanent. Magnetization is limited by satura- tion. The limit of electrification is marked by rupture. The electric charges are found to have been dissipated after the spark has passed. The condenser is said to be discharged. If the oppositely charged plates are discharged without sparking, a slight sound is produced if the dielectric is glass. This is analogous to the minute sounds given out by magnets when magnetized o^ demagnetized suddenly. Magnetization or electrification seems to consist of forcing to point in the same direction, like magnetic or electric polarities of the molecules of a substance. ELECTRIC CONDENSERS. 48. We have seen that the capacity of an electrified body depends on the area of its electrified surface, on the nearness of its charge to charges of opposite sign, and on the material of the dielectric — i. e,, the sub- stance intervening between the charges. Bodies capable of being electrified and arranged so as to present a large capacity in a small space are frequently called simply capacities, but this term is misleading, and though the term condenser is not entirely satisfactory it will be used. The total charge in a condenser depends on its potential as well as its capacity. Its potential depends on the potential of the source of electricity only, but its capacity, as stated above, depends on its size, material, and arrangement. Condenser capacities may be said to be related to each other in the bame way as rubber bags inflated by gas, A large bag charged to a given pressure contains more gas than a small bag charged to the same pressure. The gas in the large bag is making no greater effort to escape per square inch (i, e., has no higher potential) than the gas in the small bag; but it requires a longer time and more gas to charge the large bag than the small one. So, when connected to the same source of electricity it requires a longer time to charge a condenser of large capacity to a given potential than it does to charge a small one to the same potential, and its power to do work is correspondingly greater. In the same way, it requires a longer time to create the magnetic field of a large electro-magnet than that of a small one, and a stronger mag- netic field (within limits) is created by a large current than by a small one under the same conditions, and the energy stored in the strong field and its power to do work is correspondingly greater. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 37 49. It is evident that a close analogy can be drawn between the electric field in a condenser and the magnetic field around an electrojimagnet. We have seen that any movement of either field creates the other; that they can exist independently only in a static condition ; that, though they have no limits, the center of effort, the point of greatest activity in each, is at the body which we consider electrified or magnetized; that bodies differ in their qualities in these respects; that an actual physical change takes place in the dielectric when electrified and in the iron or nickel when magnetized, and, finally, that both electric and magnetic fields represent stored energy in an infinitely elastic medium, and we shall see that this medium, on account of its elasticity, vibrates and oscillates when either an electric or a magnetic field is suddenly created or destroyed in it. 50. The most common and best known form of condenser is the Leyden jar, which consists of an inner and outer coating or film of tin foil or copper on a glass jar, the glass being the dielectric. Electric induction takes place through the glass and the energy is stored in the electric field, the tin foil merely serving to increase the area over which electric induction takes place, and hence the rapacity of the condenser. Fjaed Condenser Variable Condenser Fig. 23. Fig. 23a. Fig. 23b. Condensers are often made up of a large number of interlaced plates or films of conducting material, having between them for a dielectric larger pieces of glass, mica, or oiled paper, alternate plates being simi- larly charged. Condensers are represented either as in fig. 23 or fig. 23a, They will be represented in this book as in fig. 23. Condensers are also made in which the relative position of the plates, and therefore the capacity, can be varied at will. These are called variable condensers, and will be represented as in fig. 23b. In variable condensers the dielectric may be glass, air, oil, mica, or paper. They are usually made of metal plates with air dielectric. DISCHARGE OF CONDENSERS. 51. If, after being charged by connecting the inner coating to one pole of a source of electricity and the outer coating to the other, the two coatings are connected by means of a conducting wire the charges neutralize each other and the condenser is said to be discharged. The discharge of a condenser, being a movement of electricity, creates a cur- rent and consequently a magnetic field around the wire through which the discharge takes place. 38 MANUAL OF RADIO TELEGRAniY AND TELEPHONY. If the potential is high enough, the condenser can be discharged with- out acti|,ally connecting the two coatings, for, when the opposite ends of wires connected to them are brought within a certain distance of each other, sparks will pass, and the condenser will be found to be discharged, the same as if the wires were actually connected. The charges unite by rupturing the air dielectric. Tlie energy stored in the electric field appears as sound, light, heat, and other invisible ether vibrations. This spark discharge is found when analyzed to consist usually of several sparks, passing first in one direction, then in the other. Each condenser coating is charged positively and negatively in rapid succes- sion, each charge being somewhat less than the preceding until the entire energy of the original charge is dissipated. This form of con- denser discharge is oscillating. The released charge acts like a released musical string which vibrates until its energy is dissipated, and as the same string gives out the same note, whether stretched strongly or only a little, so a condenser when discharged through the same wire always vibrates or oscillates in the same period, regardless of its potential. Just as the note given out by the string depends on its material and length, so the rate of vibration of a condenser depends on its capacity, which, as we have seen, depends on its material and arrangement. 52. Another illustration of oscillatory condenser action can be given: Let fig. 24 represent two glass vessels connected by a U tube with a stopcock at the bottom of the tube. One vessel is filled with water and the other empty. If the U tube is large enough to permit free passage of the water, when the stopcock is opened quickly the pressure in the filled vessel will cause a sudden rush of water up the other side of the tube into the empty vessel, which will continue until it has reached nearly the same height as before (fig. 24a) . It will then rush back into the first vessel, and so on, reaching a little lower level each time until equilibrium is reached at the same level in both vessels (fig. 24b). The only action which prevents the oscillation from being continuous is friction of the water on the walls of the tube and internal friction between its molecules. Eeleased condenser charges would also continue to oscillate indefi- nitely if it were not for the resistance in the discharging wires and in the dielectric and the sound and light produced by the spark. These absorb the energy of the charge, and, being relatively large, a position of equilibrium is reached after a few oscillations. If the U tube in fig. 24 is very small or the stopcock only slightly opened, the water will gradually rise on the other side and will finally reach a position of equilibrium without any oscillation, and it is found that if the condenser discharge takes place through a long thin wire, instead of a thick one, the condenser is slowly discharged through it without any oscillation. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY, 39 53. The oscillation of the water in fig. 24 is due to its inertia. Inertia is a property of all bodies and is in amount proportional to their weight. It is represented by their resistance to change of condition, either of motion or of rest. The water in the first vessel falls by the action of gravity/. Once in motion its inertia (resistance to change of condition) causes it to rise on the opposite side against the action of gravity. When gravity has overcome its inertia, it falls again by gravity and is carried on by inertia. S ^ Fig. 24. Fia. 24a. Fia. 24b. It continues to overshoot the mark, so to speak, until friction, internal and external, brings it to rest. Though the electric charges on condenser coatings appear to be inde- pendent of gravity, they do possess inertia, as is shown by their resist- ance to change of direction and by their oscillatory movements. 54. Let us consider a charged condenser (fig. 25) discharged through a thick wire connecting the coatings. A break in the wire prevents the discharge until the potential is high enough to cause sparks to cross the break. One condenser coating before discharge is at a certain positive potential, the other at an equal negative potential. Both discharge through the wire in the same time, and when they have reached zero potential the electric field has been dissipated, but the moving charges 40 MANBAL OF RADIO TELEGRAPHY AND TELEPHONY. in the wires have induced a magnetic field around the wire. The strength of this magnetic field depends on the amount of the moving charges, i. e., the strength of the current, and on the self-induction (art. 30) of the wire which, as we know, depends on its shape and the material (air or iron) in which the magnetic field is created. All the energy (except that lost by friction) which was stored in the electric field is now in the magnetic field (fig. 25a). The magnetic field, having no continuous source of magneto-motive force (current) to maintain it, collapses on the wire, producing movements of the electric charges into the condenser coatings, which now become charged in the opposite sense (fig. 35b). The electric field is again set up, containing all the remain- ing energy, and the magnetic field disappears until the charges again move toward each other. OscjLiATiNa Condenser Discharge at start energy all electric. Fig. 25. end of quarter cycle energy all magnetic. Fig. 25a. + 1 — ^ ; \ , —±» , tND OF HALr CYcte energy all electric reversed. Fig. 25b. ^r« • — ^ fTEE-au/WTtfT Cl €kd or TM energy all magnetic reversed. Fig. 25c. - 1 + t+ - End Of (Tycle energy all electric less in amount. Fig. 25d. The attraction of the unlike charges for each other is analogous to the attraction of gravity for the water in fig. 24, and the magnetic field caused by the self-induction of the moving charges is analogous to the inertia of the water, which makes it rise in the second vessel, because the collapse of this magnetic field charges the condenser in the opposite sense, and for this reason self-induction is sometimes called electro- magnetic inertia. From the foregoing illustration of what appears to take place during the oscillating discharge of a condenser, we see that the energy before an oscillation begins is all electric. At the end of the first quarter of a cycle it is all magnetic. At the end of a half cycle it is all electric, but in the opposite sense. At the end of three-quarters of a cycle it is all magnetic, but with the direction of the lines of force reversed. At the end of a complete cycle or oscillation the energy is all electric again MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 41 (figs. 25a, 25b, 25c, 25d) and in the original sense, but less in amount on account of the losses which have taken place during the transforma- tions and which are shown by the lieating of the condenser and the wires (and the sound and light produced by the spark if the oscillations take place through a spark gap). At all intermediate points of a cycle the energy is partly electric and partly magnetic. 55. A complete oscillation or cycle is made up of two alternations. The highest potential reached during an oscillation is called the ampli- tude of the oscillation. The difference between the amplitude of two successive oscillations is called the damping and is a measure of the losses. The interval in time between two successive oscillations is called the period. 56. Since every body has electric capacity in proportion to its surface (art. 44), and since movements of electric charges, without which 9, body can not be electrified, always produce magnetic fields, every body must have self-induction, and therefore electro-magnetic oscillations can take place in it. We know that every body vibrates in its own period mechanically, and we find that every body vibrates in its own period electrically, and further that the number of vibrations or oscillations per second depends entirely on the capacity and self-induction of the body. It will be seen that while a closed circuit is necessary for the flow of a continuous or direct current, for oscillating currents a straiglit wire is sufficient. A circuit containing a condenser which would completely obstruct a direct current has no effect on an alternating current other than to change its sign. 57. We must be careful to distinguish between the capacity of a con- denser and the total charge in it, and between the self-induction of a wire and the total induction caused by the current in it. The capacity, it may be repeated again, depends on the material and arrangement of the charged body. The total charge — that is, the total electric induction — depends on the capacity and the potential. In like manner the self- induction depends on the arrangement of the conductor and the sur- rounding material (whether iron or air). The total magnetic induction depends on the self-induction and the current. 58. We can see in a general way that the period of an oscillating circuit depends on the capacity and self-induction of the circuit, and not on the total electric or total magnetic induction, because the capacity and self- induction are determined by the material and arrangement of the circuit, which qualities determine the mechanical period of a body. It takes longer to discharge a condenser of large capacity than one of small capacity, and it takes longer to create a given current in a circuit of large than in one of small self-induction. Increasing the potential gives 43 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. more work to be done during a discharge, but also gives power to do it in the same ratio, so that increase of potential does not change the period, though it may change the amplitude of the oscillations. 59. It was stated (art. 29) that coiling a wire increases its self- induction and enables a strong magnetic field to be created around it, and that this increases the electrical length of the wire — i. e., it takes an electrical disturbance, started at one end of it, longer to reach the other end when the wire is coiled than when the same wire is straight. Now we see that the electrical length of a wire depends on its capacity and self-induction and that its period in seconds — i. e., the time of one complete oscillation (the time required for an electrical impulse started at one end to reach the other and be reflected back) — must be twice its electrical length divided by the distance electricity travels in a second, which we know to be the same as light (300,000,000 meters). The capacity and inductance of a straight wire long in proportion to its thickness are so related that its electrical length is equal to its natural length. From the above the period or time of one complete electrical oscilla- tion of a straight wire one meter long is si^-uuioo-^ second, and it therefore oscillates 150,000,000 times per second. The number of oscillations or cycles made by an alternating current per second is called its frequency. 60. We know that by coiling a wire its self-induction can be greatly increased, and its period thereby lengthened. By adding capacity to the wire in the shape of condensers its period can be lengthened still more, so that by suitable arrangements a circuit having small mechanical length, but comparatively great electrical length, can be made up in a small space.* Pig. 26. -o o- Fio. 26a. Such a circuit is shown in fig. 26. It is made up of a condenser con- nected to a coiled wire, and will be called in this book an oscillating circuit. * It must not be forgotten that every wire possesses capacity by virtue of Its surface, and self-induction by virtue of the fact that an electric current can flow in it. Even condensers have a certain amount of self-induction. MANUAL OF RADIO TELEGRArilY AND TELEPHONY. 43 The oscillating circuit in fig. 26 may have a break or gap in it, as in fig. 26a. If the potential of the condenser is sufficient to rupture the air or other dielectric in the gap, the circuit does not lose its oscillating character. The presence of the gap does, however, decrease the number of oscillations for one charge and prevents the complete discharge of the condenser, because the oscillations cease as soon as the potential falls below a certain value. The greater the loss or damping in each oscilla- tion the smaller the number of oscillations that will take place before the potential falls so low that the spark ceases. 61. As stated in art. 48, the term condenser is not satisfactory, and the word capacity is often used to mean condenser, especially in con- nection with such an oscillating circuit, the condenser being spoken of as a capacitij and the coiled wire as an inductance, which means a con- ducting wire arranged so as to have large self-induction. -vQQQQQ^ -AAAA/^- Fig. 27. Fig. 27a. Fig. 27. — Inductive Resistance. Fig. 27a. — Noninductive Resistance. Fig, 27 represents an inductive resistance, or simply an inductance, since it is assumed that all wires have resistance. Fig. 27a represents a noninductive resistance, or simply a resistance — it represents a coil so wound that the currents in adjacent turns are in opposite directions and the coil has therefore no self-induction. 62. An oscillating circuit whose electrical length can be varied at will is represented in fig. 28. It consists of a variable condenser in connection with a fixed inductance (fig. 28), or it may consist of a fixed condenser and a variable inductance (fig. 28a), or both capacity and inductance Fig. 28. Fig. 28a. may be variable, the arrow in fig. 28a being meant to show that any number of turns of the coil can be included at will. 63. Two circuits having the same electrical length are said to oscillate. in resonance; their periods are equal, though the inductance and capacity may not be the same in each. 44 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. For instance, suppose the oscillating circuit (26a.) is adjacent to a wire, as in fig. 28b, having the same electrical length, we know that for oscillating currents (see art. 56) a closed circuit is not necessary, we also know that by reason of their mutual induction (art. 15) the closed oscillating circuit, which we can call A B, will induce currents in the wire, which we can call C D. Since their periods are equal the induced oscillating current in C D will be suitably timed to the natural period of C D and the two circuits will oscillate in resonance. C D can be called the open circuit as distinguished from A B, the closed circuit. QroOnd Fig. 28b. Oscillating circuits now used in wireless telegraphy have electrical lengths varying from 100 to 5000 meters, giving from 1,500,000 to 30,000 oscillations per second. Those first used by Marconi had electrical lengths of about 6 centimeters and oscillated approximately 2,500,000,000 times per second. ETHER WAVES. 64. As stated in art. 55, a cycle is made up of two alternations or move- ments in opposite directions and can be represented as in fig. 18. Such a curve also represents the crest, hollow, and slope of regular waves on the surface of the ocean or other body of water. The distance from crest to crest or from hollow to hollow of a water wave is called a wave length, and this distance is equal to that of two alternations. Since electro- magnetic (ether) disturbances spread in all directions with the speed of light, and when sent out by an oscillating current succeed each other at equal intervals of time, and since the magnetic and electric forces pro- duced by oscillating currents change direction during each alternation, just as the particles of water rise to the crest or fall to the hollow of a wave, their positive and negative amplitudes may represent the crests and hollows of waves separated by half periods or half wave lengths, an MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 45 oscillating current may be called a wave producer, and the oscillations considered as movements of the ether may be called ether waves. 65. Hertz (in 1886 at Carlsruhe, Germany) was the first to show that oscillating electric currents really do produce ether waves — like those of light, only longer and subject to all the laws governing light waves. For this reason, wireless is sometimes called Hertzian wave telegraphy. 66. The vibrations of particles producing sound waves, as in air, con- sist of to-and-fro movements parallel to the direction of the waves, the latter consisting of alternating conditions of compression and rarefaction of the air. The movement of the particles in ether waves is at right angles to the direction of propagation of the wave, and the electric and magnetic stresses are also at right angles to each other at any point in the wave front. This is called transversal vibration, as distinguished from the longitudinal vibration of the particles in sound waves. Fig. 18. When one particle of a substance is displaced or made to vibrate, it induces its neighbors to follow it, and starts them to vibrating in the same periods but in different phases, each particle starting to vibrate (passing the word, so to speak) at a definite interval of time after the one next to it has started. The vibrations may be longitudinal or trans- verse, as described above, or they may be circular or elliptical, but if they are regular the waves produced are regular. The amplitude of the wave (art. 55) depends on the extreme limits from its normal position of the vibration of oacli individual particle. The wave length depends on the time of one complete vibration of each particle and the velocity with which the displacement or vibration is propagated from one particle to another of the substance. Tt is found that this velocity is equal to the square root of the elasticity of the body divided by its density. We know that this velocity in the ether is 300,000,000 meters per second, and we conclude that the ether must have very great elasticity combined with very small density. 46 MANUAL OF KADIO TELEGKAPIIY AND TLLEPPIONY. It has been stated that electric charges or electrons are the only things which have a grip on the ether, and that when they are vibrating the ether vibrates with them. When a particle is subject to several forces at the same time, it? resultant movement depends on the resultant of the forces and will vary as the foices vary, so that a body can, in effect, vibrate in more than one way at the same time, and can produce complex waves where vibrations are superimposed on each other. This is shown every day at sea by the small waves or ripples on the slopes of large ones, or the short waves from local winds superimposed and propagated in the same or different directions from the long swells due to distant storms. 67. The vibrations producing ether waves, and consequently the wave lengths and frequencies, are of an almost infinite range, for instance: Ether vibrations from 430 to 740 trillions per second (a little less than one octave) are visible to the ej'^e and are called UgJit f Between 870 to 1500 trillions of vibrations per second we have the ( ultraviolet and X-rays, and from 430 down to 300 trillions of vibrations j per second what are called infrarouge rays. '*\ Below 300 and down to 20 trillions of vibrations per second we detect ; ether vibrations by our sense of feeling or by the thermometer, and they f are called 7iea^. > ..'^■' ', ' ' ' ' "« V Twenty-five octaves lower on the same scale are the ether vibrations which we call electric waves and which are used in wireless telegraphy. The shortest of these yet measured is 0.2 of an inch in length; the longest, over 1,000,000 miles. Marconi, in his first experiments, used a pair of small spark balls which gave out waves about 12 centimeters in length. 68. Ether waves of all lengths are subject to reflection, refraction, diffraction, and absorption, and bodies, such as insulators of certain kinds, which are opaque to the short waves we call light, are transparent to the long electric waves used in wireless telegraphy. Practically all conductors are opaque to electric waves. Generally speaking, insulators are transparent to electric waves, but in transmitting the wave they absorb some of its energy. Conductors, being opaque to electric waves, partially reflect and par- tially absorb the wave energy. A simple case of wave reflection is seen when a rope hanging vertically is given a quick jerk and then held taut in the hand. A wave can be seen traveling up the rope till it reaches the top, where it is reflected, travels down the rope to the hand, is reflected there and starts up again to the top, and so continues until its energy is damped out. If a number of equally timed jerks are given, a succession of waves at equal intervals is sent up the rope. When reflected back they meet MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 47 others coming up whose lengtlis are equal to those coming down. At 6ome points the rope tends to move a certain distance in one direction with the direct wave, and the same distance in the opposite direction with the reflected wave; the result is that it does not move at all. These points are found along the rope one-half wave length apart; at all other points the rope moves or vibrates in the resultant direction of the direct and reflected wave impulse, and what are called stationary waves are set up. The points at which there is no movement are called nodes, and points at which there is maximum movement are called loops. This is shown graphically in fig. 18c. J^ V^fVC LEMSTM 3 jt WAVC LtN«TH — -4*-!^ WAVE LWeHHj Fig. 18c. Stationar}^ ether waves can be set up around conducting wires by suit- ably timed electrical impulses applied to the ends of the wires. 69. It will be observed that the point of support of the rope, where it can not move, must, in every case, be a node. So in a conducting wire, the end of the wire away from that receiving the impulses must be a current node, because no current can flow there. It can, however, and a little consideration will show that it must, be a potential loop, for while there is no movement at the point of support, the greatest pressure or tendency to move is there. -fbrcNTiAi. Loop JRRENT Node R)TENTIAL Loop- CURRCNT Nooe- FiQ. 18d. Since the electrical impulses consist of variations of current and potential, which succeed each other regularly, and since at a given point we find a loop of potential and a node of current, we must, at a quarter- wave length distant, find a node of potential and a loop of current. This is shown graphically in fig. 18d, which represents the relative positions of current and potential nodes and loops in stationary electric waves, and illustrates the statements made in art. 54 (figs. 25a, etc.), relative to the alternations of electric and magnetic fields in oscillating condenser discharges. 48 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 70. If an oscillating current be set up in a free wire (fig. 18e) by a neighboring discharging circuit in resonance with it, the free wire will be found by measurement with a micrometer spark gap to have an alter- nating potential in it, varying from nothing at the middle point, C, to a maximum at either end somewhat similar to the full curve EOF. If at the same time the current in the free wire could be measured, it would be found to have a maximum value at C and a minimum at the ends similar to the dotted curve A D B. If the wire A C B is not too far from the discharging resonant circuit and the wire be cut at C and E an incandescent lamp L be connected to the two halves as shown in the figure, the lamp Avill glow. REFLECTION OF ETHER WAVES. 71. If ether waves impinge on a reflecting surface not normal to their direction, they are reflected at an angle equal to that which the reflecting surface makes with their original direction (the angle of incidence is equal to the angle of reflection), so that directed waves may be detected at points not in the line of direction by the interposition of a reflector. Air at atmospheric pressure (about 760 millimeters of mercury) is an insulator. Its density decreases with distance above the earth's sur- face, and its insulating qualities decrease with the decrease of density. At a height of approximately 45 miles above the earth's surface its pres- sure is about 1 millimeter of mercury. At the density corresponding tc this pressure it is a good conductor, and though still transparent to short ether waves like tliose of light, it partly reflects and partly absorbs long ether waves. In the intermediate distance it is at first transparent, then partially transparent, absorbent, and reflecting, simultaneously. It is known that ether waves are guided by conducting surfaces to a certain extent (for instance, by wires), as well as reflected by them, and that otherwise they travel in straight lines. Fig. ISf shows the approxi- mate path of an ether wave started from the earth's surface and reflected from the upper atmosphere. It will be seen that even if the earth's surface did not guide the waves they might be detected at points below the horizon. Other causes of reflection may exist, such as large bodies of electrified air, or heavily charged clouds, which would cause interference between MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 49 direct and reflected waves and make electrical shadows at certain places, i. e., points at which, owing to conditions outlined above, either the waves are so attenuated that they can not be detected or they are com- pletely neutralized. Fig. 18p. REFRACTION OF ETHER WAVES. 72. When ether waves impinge on transparent bodies at any angle other than the normal, if their velocity in the transparent body, on account of its elasticity or density, is different from that at which they were previously moving, that part of the wave first entering the body will move either faster or slower than it did before. The part outside will therefore either fall behind or gain on the first part. This action will affect each portion of the wave front as it enters the body, and the result will be that its direction of movement will be changed. The effect is to bend the wave out of its original path, and the action is called refraction. Ether waves passing through the atmosphere, whose density varies at different points, are subject to this bending action. The bending due to refraction tends to keep the wave in the denser atmosphere; i. e., it is bent towards the earth's surface. DIFFRACTION OF ETHER WAVES. 73. When waves meet a body in their path (for instance, when the comparatively long waves used in wireless telegraphy impinge on a high island or mountain range) at the points where the wave front cuts tlie extreme width of the island and along the crest or summit, new cen- ters of disturbance are created, which radiate some of the wave energy to points behind the island or mountain. It has the effect of bending the waves around the object. This action of waves is called diffraction. In amount it depends on the wave length. From the new centers of disturb- ance waves are sent out, which interfere with each other, not being propa- gated in the same directions. The result is that for a distance, depending on the width and height of the obstacle and on the wave length, a shadow exists beyond it. Partial reflection of the waves toward their source takes place on the side of the obstacle nearest the source. An attempt to show this graphi- 4 50 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. cally is made in fig. 18g, but the best illustration is given by the motion of water around a rock on a windy day. The small back waves on the windward side are reflected to windward. The waves circling or bend- ing around the rock are diffracted. The still water in the lee of the rock is the shadow, in which no action exists. At a distance depending on the size of the rock and the wave length the zones of interference disap- pear, the regular waves from the two sides of the rock unite, and there is no evidence of its existence at points beyond, though it has decreased the total strength of the waves. For the above reasons, high land between two wireless telegraph stations has the effect of decreasing the strength of signals at each station, and, if close to either station, may entirely prevent that station from receiving. (It may be in the shadow or be subject to interference from reflection.) The effects of reflection and diffraction on waves passing over irregular country are very pronounced. The effects of reflection, refraction, and absorption in the atmosphere are equally pronounced, the qualities of the atmosphere in all three respects varying greatly from day to day and between day and night. An ether wave traveling from one wireless-telegraph station to another over rough country and through an atmosphere of varying density, work- ing its way around and over mountains, being balloted from thunder clouds at one point and absorbed by semiconducting gases at another, may be said to pursue an adventurous journey. Chapter II. PKODUCTION, EADIATION AND DETECTION OF ETHER WAVES. 74. We have now seen how to produce electric and magnetic fields, how to utilize magnetic fields for the production of electric currents in dynamos, how to increase the potential of these currents by means of step-up transformers, and how by means of this high potential current to force large charges into electric accumulators or condensers and by discharging these condensers in oscillating circuits to produce what we call electric or ether waves. These operations can be represented graphi- cally or diagrammatically, as in fig. 29, which shows a separately excited A. C. dynamo in circuit with the primary winding of a step-up trans- former, whose secondary charges the condenser of an oscillating circvii containing a sparJc gap. Fig. 29. The secondary winding of tlie transformer is of many turns, in order to give a high potential. The transformer also has an iron core. The great number of turns of the secondary winding, added to the effect produced by the iron core, gives the circuit containing the secondary winding and the condenser a very large self-induction, and consequently a very long period. The circuit composed of the condenser, self-induc- tion, and sparh gap has a very much shorter period, and when the spark gap is ruptured this circuit oscillates as if it were entirely disconnected from the secondary, usually completing its oscillations and coming to rest in a fraction of the period of the circuit formed ])y the secondary and condenser. The oscillating circuit (condenser, spark gap, and inductance) is shown in fig. 29 near a conducting wire, having a few turns of inductance close to those of the oscillating circuit. In this circuit we can consider 52 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Pig. 11. ^LABY ARCO TO AEfflAU i the condenser as representing the source of current, like the battery in fig. 11, art. 13; the spark gap as the break K, the turns of inductance in the oscillating circuit as A B, and the open circuit with one end grounded as C D. The oscillating currents in A B produce like cur- rents, but in the opposite direction in C D (art. 12), and C D becomes a source of ether waves. 75. The production of ether waves and their ^^ detection at a distance from the source constitute \^^^___^fe-= ;£=;==- -" wireless telegraphy. 1 ^ ^ U C D is usually called the open or radiating circuit or aerial circuit. A B the closed or oscillating circuit. The two inductances in A B and C D form the primary and secondary, respectively, of an air-core oscillation transformer (art. 27). When arranged as in fig. 29, A B and C D are said to be inductively connected. Or C D may have part of its inductance common to A B. The arrange- ment in this case acts as an auto-trans- former, and the circuits are said to be direct connected (fig. 29a). If the oscillating and radiating cir- cuits have the same period, they oscillate ( " Or or vibrate in resonance. The radiating J 9 ^ks circuit in such a case receives the in- ductive impulses from the oscillating circuit at the proper time, and the am- plitude of its oscillations is thereby in- creased. The adjustment of A B and C D to any given period and their adjustment to each other's periods is called tuning. It will be noted that tlie oscillating circuit has concentrated capacity, while the capacity of the radiating circuit is distributed. 76. The fundamental principle of wireless telegraphy is that all bodies vibrate electrically as well as mechanically; that their periods of electrical vibration depend solely on the capacity and self-induction of the vibrat- ing body; that these electrical vibrations produce ether waves which are propagated with the speed of light, and which can be detected at great distances from their source by means of instruments specially designed for the purpose. 77. Let us consider a little more closely the circuits in Fig. 29. First, we have the armature windings, the leads to the transformer and the primary winding of the transformer. These form one circuit called the primary circuit. Fig. 29a. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 53 Second, we have the secondary winding of the transformer, the leads to the condenser and the condenser. These form a second circuit, the secondary circuit. Third, we have the condenser and leads, the spark gap and leads and the closed-circnit inductance. These form a third circuit called the clotzd circuit. Fourth, we have the open circuit inductance and the leads to air and ground, forming a fourth circuit (for oscillating currents), the open circuit. There is also a fifth circuit, consisting of the secondary winding of the transformer, the closed circuit inductance, spark gap and leads. This circuit is not given any particular name and only becomes of interest when the spark gap remains conducting, after having once broken down. 78. As has been stated in art. 74, the secondary circuit has a very long period, and, for the same reasons, the primary circuit has likewise a very long period, compared with those of the closed and open circuits. If the design is such that the primary and secondary circuits have the same period they are found to operate more efficiently. 79. While sending, the spark gap appears to be sparking continuously, but it is really sparking a very small percentage of the time, completing (as has been stated) its oscillations and coming to rest in a fraction of the period of the secondary. If the spark gap were in operation all the time the fifth circuit men- tioned above would short circuit the transformer through the gap and the closed circuit inductance and there would be no oscillation and no electric waves produced. When this condition obtains, what is called an arc is produced. An arc may be described as an electrical discharge which produces light, heat and some sound and is a continuous rupture of the dielectric, as com- pared with a spark which is an electrical discharge producing light, heat and considerable sound and which is intermittent in its rupture of the dielectric. MUTUAL INDUCTION AND COUPLING. 80. Let us now endeavor to get an idea of how energy is transferred from one circuit to another until it reaches the open circuit and is radiated as electric waves. Referring to Fig. 29, A B and C D are coupled together by virtue of their mutual induction (art. 15). The induced current in C D represents a transfer of energy from one circuit to the other. If their mutual induction is large, the circuits are said to have close or tight coupling; if small, the coupling is said to be loose. It is evident that the mutual induction between two circuits depends on the self-induction of each, that is, the strength of field produced by vary- ing the current in each circuit. Also, that it depends on the distance apart 54 MANUAL OP RADIO TKLEGRAPHY AND TELEPHONY. of the two circuits, their position relative to each other (art. 22) and the material (iron or air) intervening. Mutual induction will be a maximum when all the lines of force created by the current in either circuit cut the other. In this case the coupling is said to be perfect. If the two circuits are moved in relation to each other so that only part of the magnetic field of each cuts the other circuit, their mutual induc- tion is decreased. The mutual induction between two oscillating circuits alters the effec- tive self-induction of each, making it apparently larger or smaller as one circuit is receiving energy from or transferring energy to the other. 81. Since the natural period of a circuit depends on its self-induction, if the effective self-induction is varied, the period of the circuit is varied. Therefore, coupled circuits having the same or nearly the same natural periods are found to have two periods of oscillation, one faster and the other slower than the natural period of each. The open radiating circuit generally sends out electric waves of two lengths, one longer and one shorter than the natural wave length of the circuit. The closer the coup- ling the greater the difference in length of these two waves. This differ- ence divided by the natural wave length of the circuits is called the percentage of coupling. For instance, if an open circuit, having a natural wave length of 400 meters, sends, when coupled to a closed circuit of the same natural length, two waves, one of 445, the other 365 meters, the , . T 445-365 oA«/ percentage of couplmg= — ~— r — =20%. 82. If the circuits have loose coupling, i. e., are moved farther apart, the mutual induction is less and the difference in the wave lengths radiated is less. This distance can be increased until the two waves practically coincide with the natural wave length of the circuit. This is very loose coupling, but, since without mutual induction, no energy can be trans- ferred, the two can never be the same. Most of the energy is found to be in the longer wave and until recently that in the short wave was practically wasted. The method now used of generating but one wave length will be described in art. 83. TRANSFER OF ENERGY BETWEEN COUPLED CIRCUITS. 83. The transfer of energy between coupled circuits having the same natural period is well illustrated by the mutual action of two similar pendulums connected by a flexible support. If, one being at rest, the other is pulled aside and released, the swinging pendulum gives properly timed impulses to the other through the flexible connection and starts it to swinging also, gradually decreasing its owe swings while the other increases, until the first one stops; at which time the second has reached MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 55 an amplitude nearly as great as that of the first swing of the one pulled aside. In other words, all of the energy has been transferred to the eecond pendulum. The first one then starts again and increases its swings while the second gradually slows down and comes to rest at which time the first is again at its maximum. All the energy has been returned by the second pendulum to the first. The swings are slowly damped by air friction until the system comes to rest. If the periods of the two pendulums are not equal, or nearly so, the impulses are out of step (resonance) and no transfer of energy takes place — the pendulum first started keeps on swinging and the second remains at rest. 84. If the points of support by the flexible connection are a foot or more apart (loose coupling) the second pendulum picks up the swing rather slowly and both pendulums make a large number of vibrations before the second has received all the energy from the first and the latter has come to rest. If the points of support are close together (close coupling) the second pendulum reaches its maximum and the first comes to rest in a few vibra- tions, the transfer of energy is more rapid, and the damping greater. The ball of energy, so to speak, is tossed back and forth between them more rapidly than when they are farther apart — more loosely coupled. Professor Pierce * has photographed the sparks in a short gap in the open circuit when oscillating in connection with the closed circuit and shows that they occur in groups. This particular circuit showed groups of four. In other words, four vibrations sufficed to transfer all the energy from one circuit to the other. THE QUENCHED GAP.f 85. Having once transferred all of the energy to the open circuit, we wish to radiate it and not have any retransfer to the closed circuit. What is called a " quenched gap "f has the advantage of stopping the oscillations of the closed circuit and leaving the open circuit free to vibrate in its own natural period and it, therefore, radiates waves of but one length. 86. What we call the closed circuit is only closed when the spark gap is conducting and its period in that condition is the one measured either when we take the time interval between sparks or determine it by a wave meter. It has a different period when the spark gap is not conducting because its capacity with reference to being charged from the open circuit is less and it is therefore out of step (art. 83) with the open • G. W. Pierce, Principles of Wireless Telegraphy, 1910, p. 248. t Discovered by N. Wien in the course of an investigation on electrical discharges between metal surfaces placed very close to each other, and pub- lished by him in Oct., 1906. 56 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. circuit and the latter does not transfer any energy to it. The effect of the method of construction of the quenched gap seems to be to restore the nonconducting character of the gap the first time the closed circuit comes to rest, and thus leave the open circuit free to radiate. It would be inter- esting to take photographs in both circuits to determine whether this really is the case. 87. Eeferring to art. 80 on mutual induction: The open circuit is first set to oscillating in either the period longer or shorter than its natural period and has reached its maximum when the closed circuit has Fig. 18h. — Oscillations from Quenched Spark-gap. stopped and opened. Thereafter the open circuit is free to vibrate in its own period, and that it changes to that period can be shown by wave meter readings, but in building up it is sending out waves of a different period. The first maximum reached in the open circuit is the highest maximum and, since no further loss by retransfer to the closed circuit takes place, the quenched gap is consequently the most efficient. It will also con- duce to efficiency to make the building up period of the aerial (when it is radiating waves of a different length) as short as possible. In other words, close coupling, but close coupling increases the induced E. M. F. in the condenser circuit. Therefore, there is a possibility with very close coupling of retransfer of energy by breaking down the gap and again closing that circuit. 88. We can, therefore, conceive of a wave train (art. 102) from an ordi- nary open circuit as made up of a series of waves whose amplitude rises and MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 57 falls during the transfer and retransfer of energy from one circuit to the other; the rate of dying away depending on the coupling and being partly natural (due to heating and radiated energy in the shape of electric waves), partly artificial (due to retransfer of energy to the closed circuit) . A wave train from tlie open circuit of a quenched gap can be repre- sented, as in fig. 62, by a building up at a certain frequency (depending on the coupling) to a maximum depending on the radiation or other losses per oscillation, and then oscillations in the natural period of the open circuit, with damping dependent on the radiation and resistance of the open circuit only. The closed circuit starting at a maximum and transferring all the energy to the open circuit in a few oscillations as shown in the upper part of fig. 18h; there being no retransfer of energy from the open to the closed circuit and vice versa as occurs with the pendulums discussed in arts. 83 and 84. METHODS OF PRODUCING ELECTRIC WAVES. 89. In fig. 29a, A B and C D have been given some turns in common, forming an air core-auto-transformer, but, whether directly or inductively connected, these two circuits — the closed and open circuits — must have equal natural periods in order to produce and radiate electric waves efficiently. 90. We can omit the closed circuit, in fig, 29, and excite C D directly from the transformer by putting a spark gap in C D, as shown in fig. 29b. Since C D has some capacity, by virtue of its surface, we can store some energy in it and when the spark gap breaks down, this energy will oscillate in C D as an electric current and will produce electric waves. This is one of the earliest methods of producing electric waves for wireless telegraphy and was usually called plain aerial, C D being known as the air wire now called aerial or antenna. 91. Arrangements for producing electric waves may be simplified still more by dispensing with the transformer and connecting the alternator terminals directly across the gap. The transformer is only used to increase the potential of the condenser or aerial and, therefore, store up more energy between discharges and produce oscillations of greater amplitude. If we could conveniently generate a high potential directly in the alter- nator, the transformer would not be used. 92. Many attempts have been made to do this, the most successful being when the oscillations are generated directly in the dynamo and C D is con- nected directly to its terminals, as in fig. 29c. In this case the armature of the alternator is stationary and consists of one or more turns of wire around each of a large number of stationary poles, past which the field poles are moved at very high speeds and the natural frequency of the aerial (CD), combined with the armature winding, is the same as the alternator frequency. 58 MANUAL OF RADIO TELEGiiAPHY AND TELEPHONY. SOO TO lOOO VOl_TS DC FiQ. 29e. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 59 It will be seen that the oscillations thus produced and, therefore, the waves radiated, are continuous and not intermittent as when a spark gap is used. C D may have but a small capacity, but it receives a charge at everj oscillation, instead of being intermittently charged and allowed to dis- charge gradually. However, in order to produce wave lengths, say of 3000 meters, the field revolutions and numbers of poles must be such as to produce a frequency of 100,000 cycles per second. This involves great mechanical difficulties. E. A. Fessenden and E. F. W. Alexanderson have designed such genera- tors, turning at more than 300 revolutions per second, but they have not yet become commercially successful. Machines are now built to produce high frequency with lower mechani- cal speeds by special windings of fields and armatures or transformers, but such apparatus is not yet in general use.* 93. As has been indicated in figs. 11a and lib, oscillations can be pro- duced by suddenly making and breaking direct current, with or without an induction coil (transformer) intervening, and some very large inter- rupted current sets have been built and are, or have been, used for trans- atlantic wireless telegraphy by Mr. Marconi. 94. Even before wireless telegraphy existed, it was known that an arc produced by direct current would produce electrical oscillations if in circuit with an inductance and a condenser, and that these oscillations were produced continuously if the circuit was adjusted properly. This, known as the arc, as contrasted with the spark method, of produc- ing electric waves, is illustrated in figs. 29d and 29e. It will be observed that the only difference between fig. 29d and fig. 29 is that we have direct instead of alternating current and the arc instead of the spark gap. It is found also that the arc can be made to produce oscillations, if mounted as in fig. 29e, i. e., directly in the aerial, and that it operates better if placed in a very powerful magnetic field, as shown in fig. 29e, and oper- ated at a high potential, in a heat conducting atmosphere, such as a gas largely composed of hydrogen. 95. We see, therefore, that we can and do produce both intermittent and continuous electric oscillations, with alternating current and also with direct current. Continuous or undamped oscillations are, generally speaking, produced without noise and radiated at» lower potentials than intermittent ones. Other advantages of continuous oscillations will appear later. As yet they are not in general use and are somewhat more difficult to regulate than intermittent oscillations. * Professor Rudolph Goldschmidt and Count Arco have designed and con- Btructed such machines. 60 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. RADIATION OF ELECTRIC WAVES. 96. It has been stated that every oscillating circuit must contain inductance and capacity. This is true even though the circuit consists of straight wires, for these have distributed inductance and capacity. If the circuit is formed as in fig. 26a with a coil of wire and a condenser, the inductance and capacity are said to be concentrated or lumped. There is also a certain amount of distributed inductance and capacity, but in general this will be small compared with the concentrated portions. In the case of a linear oscillator (fig. 30), when the oscillations are taking place and the charges are most widely separated, we may imagine lines of electric force to be connecting each unit of positive electricity on one end to a unit of negative electricity on the other. For clearness of conception we may picture these lines of force as having a real exist- , » > > i 1 1 1 ' 1 ■ I I " I 9.' ; •' ! I j Fig, 26a. \ \ \ I I I I / Fig. 30. Fig. 31. Fig. 26a. — Non-radiating Circuit. Fig. 30. — Radiating Circuit. Fig. 31. — Electric Wave Leaving Oscillator. ence and exerting an elastic pull between the positive and negative units, tending to draw them together, while at the same time, provided they are running in the same direction, they tend to repel each other. These lines of force, in the case of a linear oscillator, on account of their repulsion away from the oscillator, form wide loops which tend to snap off and travel away into space when the charges again rush back through the spark gap, thus forming electrical waves or radiation as shown in fig. 31. In the case of the circuit shown' in fig, 26a, where the principal capacity lies in the condenser, the lines of force are concentrated between the condenser plates. They do not loop out to any extent, and hence such a circuit radiates very feebly. On account of these differences an open circuit oscillator (fig. 30) is often called a radiating circuit, while MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 61 a closed circuit (fig. 26a) is called non-radiating, although all high frequency circuits radiate in some degree. 97. Let fig. 32 represent a closed circuit inductively connected to a vertical, grounded, open circuit or aerial, and suppose the spark gap to break down at the point of maximum potential of the charging current. At this instant there is no current in the closed circuit and, therefore, no current in the open circuit. The energy is all electro-static, all in the closed circuit and practically all in the electro-static field between the condenser plates, the capacity of the spark points and other parts of the circuit being very small. As soon as discharge through the spark gap commences, the field of the current in the closed circuit inductance induces movements of electric + u- J 1 Fig. 32. Fig. 32a. charges in the open circuit, the starting point of the disturbance being the open circuit inductance. As the charges in the open circuit separate, they are connected by electro-static lines of force and surrounded by magnetic lines of force, both moving outward at the same rate that the charges move in a straight wire. The electro-static field becomes a maximum when the charge reaches the top of the wire. At this time the magnetic field is a minimum. At the expiration of a half period, when the charges meet again, the mag- netic field is a maximum, but reversed in direction. The electro-static field reverses as the charges separate again. If they can be represented as meeting in the open circuit inductance, the electro-static field just after the end of a half period can be represented as in fig. 32a, where the mutual repulsion of the electro-static lines of force outside the wire has kept them from returning as fast as the charges travel towards each 62 MANUAL OF RADIO TELKGRAPIIY AND TELEPHONY. other. As the charges meet, the ends of the lines of electric force unite and become closed circuits, or electric whorls shaped like smoke rings which, owing to the mutual repulsion of all their parts, expand outward, upward, and downward. It is in some such manner that we can conceive energy to be detached and sent out into space from wires forming oscillating circuits. The expanding rings touch the earth and are guided by it as by any other conductor, thus resembling, near the earth, expanding hemispherical shells. These may be called earthed waves to distinguish them from the free waves which exist momentarily in the vicinity of an ungrounded oscillator (fig. 31). 98. If the point of connection with the closed circuit is considered as at the earth, earthed waves only are generated and detached from the aerial and no free waves exist at any time. The production of earthed electric waves under these conditions is illustrated in fig. 33. •':S>>/'\ .-/>\ ''\; Fig. 33. — Earthed Electric Waves. We know that earthed waves are guided by conducting surfaces; we know that light waves are not; we do not know where the dividing line is between waves that are radiated in straight lines and those that are guided by conductors. 99. For simplicity, we have described the process of radiation in terms of electro-static lines of force, but it must not be forgotten that a moving electro-static field always produces a magnetic field at right angles to itself and at right angles to the direction of movement, so that we have electro-static lines perpendicular to the surface of the earth (at least near the aerial), and magnetic lines in circles surrounding the aerial. Both the electro-static and the electro-magnetic fields reverse their directions every half wave length. The process of radiation withdraws energy from the circuit just as is the case when a resistance is placed in the circuit; hence radiation resistance is an expression often used, meaning the resistance which under the given conditions would use up the same amount of energy as that removed from the circuit by radiation. This radiation resistance depends MANUAL OF RADIO TELEGKAl'IIY AND TELEPHONY. 63 only on the form and dimensions of the aerial and on the frequency of the oscillations, increasing rapidly as the frequency increases. It is independent of the intensity of the oscillations and of other sources of lost energy in the circuit. Eadiation resistance might be called the radiation coefficient. Accurate means of measuring it are not yet in general use. DAMPED OSCILLATIONS. 100. It has been explained (art. 54) that when a circuit consisting of a condenser, inductance, and spark gap is charged by a transformer to a potential so great that a spark passes across the gap, the electricity stored up in the condenser discharges itself through the spark gap, and by its inertia charges the condenser in the opposite sense, only at the next instant to again discharge itself, and so on. All this takes place during the time of one spark, and in fact this surging of electricity is what keeps the spark in existence after the first discharge. This surging back and forth would continue indefinitely were it not for the energy used up in Fia. 34. — Damped Oscillations. Energy Supplied at Beginning of Wave-train. tlie heat of the spark and in the resistance and other losses in the rest of the circuit. But as no new energy can be introduced into the circuit until the condenser is recharged, the electrical surgings decrease in intensity and finally cease. If we represent time by the horizontal axis and the amplitude of the oscillations by the vertical axis, fig. 34 will show graphically the course of the phenomenon. It is exactly analogous to a light pendulum which is set swinging and which is brought to rest after a limited number of swings by the friction of the air. Gradually decreasing oscillations of this kind are called damped oscil- lations and obey the law that each succeeding amplitude is a given fraction of the one before it. UNDAMPED OSCILLATIONS. 101. It has been seen in the last article that the cause of the dying out of a train of oscillations in a spark circuit is the using up of energy in the circuit together with the fact that no energ}' can be brought in from outside to compensate this loss. If means can be found for keeping 64 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. up a constant supply of energy, such as an alternator directly connected to the aerial or an arc transmitter, our oscillations can be made to continue indefinitely and with equal amplitude as in fig. 35. Fia. 35. — Undamped Oscillations. Energy Constantly Supplied. 102. The electric waves radiated during one set of oscillations are called a wave train. If more than one, the wave trains radiated during one-half cycle of the charging current are called a group of wave trains. The duration of a wave train is the time of one oscillation multiplied by the number of oscillations in the train. It is found that the duration of a wave train is much less when the oscillating circuit (A, B, fig. 29) is connected to an aerial with one end free and the other earthed, like C D, than when it oscillates without any other electrical connection. The energy is radiated more rapidly, the vibrations more quickly damped. It is for this reason that the circuit formed by the condenser, spark gap, and inductance is called the closed or oscillating circuit; that formed by the aerial, inductance and ground, the open or radiating circuit. (See art. 75.) 103. Considering the series of expanding hemispherical shells referred to in art. 97, and shown in fig. 33, if there is but one wave train per alternation of the condenser charging current, the thickness of one of these series is equal to the wave length multiplied by the number of oscil- lations per train. Suppose this to be 10 and the wave lengths 500 meters, then the depth of a wave train is 5000 meters, or a little more than three miles. If the frequency of the alternating current is 60 cycles, or 120 alternations per second, we have 120 wave trains per second, and since they travel at the rate of 186,000 miles per second the wave trains have intervals of 1550 miles between them, so that M'Orking at ordinary dis- tances and at this frequency, each wave train has passed the receiving station before its successor has left the sending station. If the alternator frequency is 500, the wave trains are only 186 miles apart, or about the distance of ordinary daylight communication between ships. 104. When the spark gap is set to break down at the maximum charg- ing potential, the condenser absorbs and stores all the energy that can be transferred by the charging transformer during an alternation. When it discharges, it transfers part* of the energy to the open circuit to be ♦ Experimentation has proven that from 80 to 90 per cent of the energy delivered to the transformer is transferred to the spark circuit. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 65 radiated as electric waves. Since its period of discharge is very short as compared with that of the cliarging current the latter current does not appreciably change during the time the condenser is discharging. This current immediately begins to again charge the condenser, but the poten- tial of the latter does not rise high enough to cross the gap so that the con- denser soon begins to return energy to the charging circuit. It does this until its potential and the charging potential (and current if they are in phase) falls to zero. It then begins to absorb energy again with the reverse potential, and on reaching the maximum again discharges across the gap. Fig. 36 is an attempt to illustrate this action graphically. The area included by the curve on the left of the zig-zag line indicates the work Fig. 36. done on the condenser during the first half of an alternation; the zig- zag line indicates the number and amplitude of vibrations made by the closed circuit in transferring the energy to the radiating circuit. The area included by the curve on the right of the zig-zag line repre- sents the work done during the second half of the alternation in recharg- ing the condenser. This work is all returned to the charginor circuit. DECREASE OF AMPLITUDE WITH DISTANCE FROM SOURCE. 105. From the discussion in art. 103 on the thickness of the hemi- spherical shell enclosing a train of ether waves, if we assume this thick- ness to remain constant and that part of the shell near the earth to be represented by an expanding cylinder, it is increasing in size by one dimension only, viz., circumference, and therefore the energy in any part of this shell will vary inversely as the distance, instead of inversely as the 66 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. square or cube of the distance from the source, as would be the case if expansion were taking place in two or in three directions. But it appears that expansion takes place in two directions; since Messrs. W. Duddell and J. W. Taylor, in experiments made for the English Navy in 1905, proved that (at least for distances up to 60 miles) the received ciirrent as stated above varies inversely as the distance from the sending station, and the received energy varies inversely as the square of the distance. But additional experiments by Dr. L. W. Austin show that this law does not hold except for very short distances, and that the amplitude is lessened from other causes than those due to distance alone. We loiow that the energy is absorbed in the atmosphere more by day- light than by night — more at high (summer) than at low (winter) tem- peratures. The amount of absorption as between one day and another probably depends also on the electric condition of the atmosphere. Long waves suffer less absorption than short ones. Irregular country produces large absorption. The absorption over some soils is for com- paratively long distances, 30 times as great as over sea water. Trans- mission over salt water is the best. 106. As illustrating the difference in absorption between short and long waves, the greater efficiency of short waves for short distances, and the rapid falling off at distances above 100 miles. Dr. Austin finds : Strength of received signals at 20 miles, using 300 meter waves, 5 times as great as with 1500 meter waves; at 100 miles, 4 times as great as with 1500 meter waves; at 400 miles, 1.6 times as great as with 1500 meter waves; at 800 miles, signals from 300 meter waves weaker than from 1500 meter waves. Using 300 meter waves, he finds, strength of signals at 200 miles, 0.3 of that at 100 miles; at 400 miles, 0.053 of that at 100 miles; at 800 miles, 0.0036 of that at 100 miles. (See table 11, appendix A.) DETECTION OF ELECTRIC WAVES. 107. We believe that, generally, the direction of the magnetic lines of force at any point in a wave near the earth is parallel to the earth's sur- face and at right angles to a line joining the point with the source of radia- tion; and that the direction of the electro-static lines of force at any point near the earth is perpendicular to the earth's surface. An iron wire placed horizontally and parallel to the lines of magnetic force will be magnetized by a passing electric wave just as iron wires held in the magnetic meridian become magnetized ; pointed in the direction of the station the magnetic effect would be zero. It has been proposed to utilize this fact, both as a detector of electric waves and of their direction, Any conducting wire held perpendicular to the earth will be cut at right angles by the magnetic lines of force and will have electric charges MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 67 induced in it which will create currents, and it is by means of the cur- rents induced in vertical conductors that electric waves are usually detected. A vertical wire thus situated also has a difference of potential created in its ends since it joins two points of the advancing wave whose electric potentials differ. This is also the case in a horizontal wire, if in the line joining its posi- tion with the source of radiation, A very long horizontal wire so placed might have stationary waves like those of fig. 18c set up in it. The total electric is equal to the total magnetic energy in an advancing vrave. If two horizontal conducting plates forming a condenser are in the path of the wave, they will have electro-static charges of different poten- tials induced in them. This potential difference will vary with their vertical distance apart. If these plates are joined by a conductor, electric currents will be produced in it. We see, therefore, that there should be at least three ways of detecting electric waves: (a) By placing conductors at right angles to the mag- netic field; (b) By placing conductors parallel to the electric field; (c) By adding to conductors at right angles to the magnetic field, conducting planes forming condensers at right angles to the electric field. It would seem that by the last method we should be able to abstract the greatest amount of energy from an electric wave and, therefore, be able to detect it at the greatest distance from its source. Methods (a) and (b) coincide since vertical conductors are at the same time at right angles to the magnetic field and parallel to the electric field. From the good receiving results obtained by the use of long horizontal conductors, with a comparatively limited vertical portion, it seems that the parts of the earthed waves near the earth (see fig. 33) travel at a slower rate than the higher parts and the waves are bent over as if fig. 33 were reversed, cutting the horizontal wires at something less than a right angle. However, this subject awaits more complete investigation. 108. It will be readily seen that the induction of currents in another aerial, however great the distance from the inducing aerial, is not greatly different from the inductive actions of the wires A B and C D on each other, which were discussed in the preceding chapter. It was there pointed out that inductive actions caused by ether move- ments could have no limits, however small they might be at great dis- tances. In other words, every change of current sends out some non- returnable energy. Oscillating circuits of high frequency send out more non-returnable energy and radiate better than those of low frequency. Open oscillating circuits radiate faster than closed oscillating circuits. 68 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. RECEIVING CIRCUITS. 109. In practically all cases, except at a few large stations, the same aerial wire is used for both sending and receiving. The advancing waves of electric and magnetic force from the sending aerial cut the receiving aerial and induce in it oscillating currents. If the receiving circuit has the same period as that of the passing waves, the induced oscillating currents in the aerial will increase until the energy dissipated per oscillation, by re-radiation, resistance, and transfer to other parts of the receiving circuit, is equal to that received per wave. If the receiving aerial circuit is directly or inductively connected to a closed oscillating circuit to which part of the energy received per wave is transferred during each oscillation instead of being re-radiated, this closed oscillating circuit will absorb energy, and if its period is equal to that of the arriving waves the oscillations will increase in amplitude with each half period, since a closed circuit radiates slowly. If a detector is placed in either the open or closed circuit so that the oscillating currents produce differences of potential at its terminals and the maximum ampli- tude of the oscillation set up is sufficient to make it function, the passing of groups of wave trains separated into dots and dashes at the sending station can be detected at the receiving station. At the sending station the closed circuit furnishes energy to the radiating circuit, which sends it out in the shape of electric waves. At the receiving station this radiating circuit absorbs energy from the passing waves and transfers to the closed circuit part of what it absorbs. It is evident that no spark gap is required in the closed receiving cir- cuit and that, since no high potentials nor heavy currents need be provided for, it is not necessary that the receiving inductances and condensers should have the same dimensions or arrangement as those in the sending circuits. But in all other features receiving circuits are the exact analogue of sending circuits and the detector could occupy the place of the spark gap. Chapter III. ELECTEIC UNITS AND THEIR RELATION TO EACH OTHER. 110. Fleming says " exact measurement is the very life and soul of all technical applications of science.'' Our attention has thus far been concentrated on the quality rather tlan the quantity of the electro-magnetic actions under discussion. Be- fore proceeding further it is necessary to consider the standards of measurement adopted and their relation to each other. 111. Electric and magnetic actions being forms of energy, and being mutually convertible, as we have seen, are subject to all the laws govern- ing transformations of energy. Wurlc is done when conductors are moved in magnetic fields, the re- sistance to movement and the amount of movement determining the amount of work done. The unit of mechanical work is a foot-pound, by which name we designate the work done in lifting 1 pound 1 foot against the action or force of gravity. Force, by which we mean the cause 6i action or movement (pulling or pushing ability), is measured in pounds, and force multiplied by the distance through which it acts is wor]c. Lifting 10 pounds 10 feet=100 foot-pounds. Exerting a push of 10 pounds for 100 feet = 1000 foot- pounds. The amount of work done in a given time — that is, the rate of doing work — is called power. The unit of mechanical power we call a horse- power, and it represents a rate of doing work equal to 33,000 foot-pounds per minute, or 550 foot-pounds per second. In the above definitions of work and power the units of distance, weight (or mass), and time are the foot, pound, and minute, all of which are defined by law and are called fundamental units. 112. Another system of units, proposed by the British Association for the Advancement of Science and now generally used in electrical measurements, is based on the centimeter, gram, and second, and is usually called the c. g. s. system. The use of this system is authorized bv law and is universal in scientific work. 70 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. The following relations exist between the two sets of units : 1 foot = 30.48 centimeters, approximately. 1 pound =453,59 grams, approximately. 1 minute = 60 seconds.* The units of length and weight in the United States are kept at the Bureau of Standards in "Washington, and the unit of time is determined by the Naval Observatory in the same city. The unit of force in the c. g. s. system is that force which, acting on a gram mass for 1 second gives it a velocity of 1 centimeter per second. This force is called a dyne. The force of gravity acting on a gram mass for 1 second will give it a velocity of 32.2 feet per second = approximately 981 centimeters per second; therefore the force of gravity is equal to 981 dynes and the pull of a dyne represented as a weight is equal to -^\^ of a gram. The pull of a pound, which equals 453.59 grams, must be equal to that of 453.59x981 = approximately 445,000 dynes. The unit of worTc in the c. g. s. system is the work done in overcoming the force of 1 dyne through 1 centimeter, and is called an erg. In other words, an erg is the work done in lifting ^\^ of a gram 1 centimeter. An erg by definition is a dyne overcome through a centimeter, and we see that a foot-pound is 445,000 dynes overcome through 30.48 centi- meters; therefore a foot-pound equals 445,000x30.48 = approximately 13,570,000 ergs, and a horse-power, which equals 550 foot-pounds, per second=13,570,000x 550 = approximately 7,460,000,000 ergs per second. 113. The c. g. s. units of length (centimeter), time (second), force (dyne), and work (erg) are employed to define the absolute units used in electrical measurements. These are electro-motive force, current, and resistance. (Art. 3, art. 29.) From these are derived the so-called practical units in daily use — volt, ampere, and olim. On account of the fact that the names adopted for the practical electro- magnetic units are all names of noted scientists and not related to nor in any way descriptive of the qualities they are used to designate, their acquirement must be entirely a feat of memory. They can be more easily remembered by associating them with the names of the theoretical or absolute units. As, for instance, we say an E. M. F. of 125 volts, a current of 40 amperes, a resistance of 10 ohms. * The unit of time is based on a fundamental unit, being a fraction of tlie cime of a revolution of the earth, and this unit is common to both systems. The foot and the pound are really arbitrary units. The centimeter is a fraction of a fundamental unit, namely, of the distance from the equator to che north pole on a certain meridian. The gram is the weight of a cubic centimeter of distilled water. It is an arbitrary unit. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 71 By agreement among electricians, electro-motive force is represented by the letter E; electric current by the letter I; resistance to the flow of electricity by the letter R; time by the letter T ; work by the letter W ; power by the letter P. The object now is to determine the relation of these quantities to each other, 114. We know that it requires work to move conductors in magnetic fields, or one magnet in the vicinity of another, and the movement generates an E. M. F. in the conductor, and also a current, if the conductor forms a closed circuit. And we also know that the amount of current produced by a given E, M. F. depends on the resistance of the conductor (art. 29). We say that the E, M. F. and current are produced in the circuit because it cuts the lines of force of the magnetic field. We must, therefore, have a definite idea or agreement as to exactly what is meant by the term, lines of magnetic force, and how they are con- nected with E. M. F. and current. The physical basis for the term is the action of iron filings in the field of a magnet (art. 7, fig. 6), but to make a definite basis for measurement it has been agreed, first, that a unit magnet pole shall be one that when placed at a distance of 1 centimeter in air from a like pole of equal strength, is repelled by a force of 1 dyne. Second, if a unit pole, as defined above, is placed in a field of force of such strength that it is acted upon (attracted or repelled) by a force of 1 dyne, such a field is a unit field and shall be held to contain one line of force per square centimeter. It is further agreed, third, that unit E. M. F. shall be that generated by moving a conductor across unit field, so that it cuts 1 square centimeter (1 line of force) per second. Fourth, if this conductor forms part of a closed circuit, and if the cur- rent generated by this unit E. M. F. is such as to cause the movement of the conductor to be resisted by a force of 1 dyne, it is agreed that the con- ductor has unit resistance and that the current produced is unit current. The work done is 1 erg per second, equal to a dyne (-g^ gram), lifted 1 centimeter. Since power is rate of doing work, we can say it requires unit power to produce unit E. M. F. or unit current in a circuit of unit resistance. 115. Let fig. 15 represent unit magnetic field between two magnet poles N and S. Let C D represent a conductor one centimeter in length mov- ing at right angles to this field at the rate of one centimeter per second, and making sliding connections at its ends with a very heavy conductor whose resistance, as compared with C D, is so small that it can be neg- lected and the resistance of the circuit considered as concentrated in C D. Then, if it requires a pull of 1 dyne (1/981 gram) to keep C D moving 72 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. at the rate of one centimeter per second, C D has unit resistance, unit current flows, and, by definition, unit E. M. F. is generated. 116. If the speed of C D is doubled, the E. M. F. is doubled and the cur- . rent (as shown by the effects) is also doubled, we can express this by say- ing : (a) Current varies directly as E. M. F. i i l^i 1 / Fig. 15. If the size of C D is doubled (the speed and, therefore, E. M. F. remain- ing the same) the resistance is reduced to one-half, and we find that the current is doubled as before; we say: (b) Current varies inversely as F M F resistance. Combining (a) and (b) we can say current^ — "^-r^ — ~ or I ° ^ ' ^ ^ •' resistance = -(1). Equation (1) is the fundamental electrical equation and states in mathematical form what is known as Ohm's law, viz. : " The current in any circuit varies directly as the electro-motive force, and inversely as the resistance in the circuit.'^ 117. We also find that doubling the current doubles the opposition to movement and, other things remaining the same, doubles the work pei second, or the power. Power, therefore, varies directly as the current. Doubling the speed of movement doubles the electro-motive force and also the current, but it quadruples the power or work done per second since it represents a pull of 2 dynes through 2 centimeters in 1 second. Power therefore, varies directly as the E. M. F., as well as directly with the cur- rent, and we say that it varies as their product, or P = I E (2) or from (1) R andP^C'R. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 73 The foregoing are physical facts determined by observation and experi- ment. 118. Since magnetic fields containing 20,000 lines of force per square centimeter can be obtained, a rate of cutting of one line per second gives too small a unit of E. M. F. for practical use. On the other hand, the current necessary to produce a resistance of 1 dyne to this slow movement in unit field is somewhat large, therefore to replace the theoretical or absolute units, as defined in art. 114, the so-called practical units have been adopted. VOLT. The practical unit of E. M. F. is the volt and is the E. M. F. generated when lines of force are cut at the rate of 100,000,000 per second. AMPERE. The practical unit of current is the ampere and is one-tenth of the theoretical or absolute unit. OHM. E In order to maintain the truth of the equation /= -^ (1), the prac- tical unit of resistance, which is the ohm, is taken as 1,000,000,000 times the theoretical or absolute unit. Ohm's law then still remains true. 7= „ or amperes = -; R ^ ohms, because this equation in terms of the absolute units is — (amperes) = E X 100,000,000 (volts) ^ • , . ,, r E „,, . . -n^-, r^.^r nnr! nnn / 1, \ > which IS thc samc as i = -„" . 1 lie size of B X 1,000,000,000 (ohms) ' R the units has been changed, but the proportion between them is the same as before. WATT. The practical unit of power is the watt, which is the ergs of work done per second when 1 ampere is made to flow with an E. M. F. of 1 volt. This in ergs (see equation (2)) equals unit E. M. F.x 100,000,000 X ^ , or 10,000,000 ergs per second. Therefore 1 watt equals 10,000,000 ergs per second. The power expended in any circuit in watts equals the product of the volts and amperes in the circuit, or P=IE (2). Ten million ergs of work is called a joule. Therefore a watt=l joule per second. We have seen that 1 H. P. = 7,460,000,000 ergs per second. There- fore 1 H. P.= 746 watts. 1 watt = approximately 0.737 foot-pounds per second. 74 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY, 119. After having selected the practical units, it became necessary, for the purpose of comparison and for everyday use, to represent them in practical form, because the accurate measurement of dynes and ergs is a very difficult matter practically, but it can be done in accordance with definitions given in art. 112. Art. 114 indicates how to measure the strength of magnetic fields and how to determine and compare E. 11. Fs. and currents by the ergs of work done in creating them. A volt or an ampere can thus be definitely created. The reproduction of standards of measurement is assisted by the following facts: (a) The resistance of a conductor kept at a constant temperature is found to depend only on its length and area of cross section, so that standards of resistance are easily reproduced. (b) The current from certain primary batteries is found to be constant when their terminals are connected by the same wire : Since current and resistance are constant, the voltage of such cells must be constant, and this voltage once determined by comparison with absolute volts as determined above, we have at once a practical concrete standard of E. M. F. (c) It is found that the decomposition of an electrolyte (art. 1), by an electric current, always results in the separation or deposit of exactly equal quantities of the constituents of the electrolyte for equal quantities of current. The deposit in a certain time, being weighed, serves as a very accurate measurement of the amount of electricity which passes in that time, and consequently affords a very accurate means of comparing electric currents. When 1 ampere determined as above is passed through a given electrolyte, the weight of material deposited gives us at once a practical standard of current. E 120. On account of the relation 1= ^ (1) between amperes, volts, and ohms in a circuit, if any two of them are known the other is also known, so that only two measurements of concrete units are required. The question of which two should be selected and the exact form that each should take has been the subject for discussion at a number of inter- national conferences, the latest of which has recommended that only two electrical units shall be chosen as fundamental units, viz., the inter- national ohm defined by the resistance of a column of mercury, and the international ampere defined by the deposition of silver. The volt to be defined as the E. M. F. which produces an electric cur- rent of 1 ampere in a conductor whose resistance is 1 ohm. Different methods of measurements produce slight differences in the values of the standards, but the values recognized by law in the United States are as follows : MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 75 The standard (international) ohm is the resistance offered to an un- varying electric current by a column of mercury at the temperature of melting ice — 14.4521 grams in mass — of a constant cross-sectional area, and of a length of 106.3 centimeters. The standard (international) ampere is the unvarying current which, when passed through a solution of nitrate of silver in water in accordance with certain specifications, deposits silver at the rate of 0.001118 of a gram per second. As previously stated, a volt is the E. ]\I. F. which if steadily applied to a conductor whose resistance is 1 ohm will produce a current of 1 ampere; but a concrete standard for the volt is also recognized by law, it being specified : That the electrical pressure at a temperature of 15° centigrade between the poles or electrodes of the voltaic cell known as Glade's cell, prepared in accordance with certain specifications, may be taken as not differing from a pressure of 1.434 volts by more than 1 part in 1000. The latest international conference has recommended the adoption of the Weston cadmium cell as preferable to the Clark for a standard cell. The Weston cell has an E. M. F. of 1.018 volts at 20° C. Standard resistance wires having a known ratio to the legal ohm are made, and these, with standard cells, are used for calibrating volt meters and ammeters, which are the names given to the galvanometers for indi- cating automatically the E. M. F. and cuiTent in any circuit. In this way electrical values are made uniform throughout the country. 121. In addition to the volt, the ampere, the ohm, the watt, and the joule other practical units have been adopted, the most important of which, for our purposes, are : COULOMB. The unit of quantity, the coulomb, which is the amount of electricity passing any point in a second when 1 ampere is flowing in the circuit. FARAD. The unit of capacity, the farad. A condenser is said to have a capacity of 1 farad when 1 coulomb of electricity will charge it to a potential of 1 volt. (Potential and E. M. F. are in some senses identical, one being the passive and the other the active state. An E. M. P. is the result of difference of potential.) If this definition were in terms of the absolute units, that for capacity would read : A condenser is said to have unit capacity when one unit of electricity will charge it to unit potential. 76 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Since by definition a condenser has a capacity of one farad when one- tenth of the absolute unit of electricity charges it to a potential of 100,000,000, a farad must equal — , x roo~00(U)f)n ~^^'^ absolute units of capacity.* HENRY. 122. The unit of self-induction, the henry. A circuit is said to have a self-induction of 1 henry when, if the current in it is varied at the rate of 1 ampere per second, the induced E. M. F. — that is, the counter or reacting E. M. F. — tending to oppose the change is 1 volt. The definition of self-induction in terms of the absolute units would be : A circuit is said to have unit self-induction when, if the current in it is varied at the rate of one unit per second, the E. M. F. of self-induc- tion is unity. Since by definition a circuit has a self-induction of one henry, when, if the current is varied at the rate of one-tenth of unit current per second, the absolute E. M. F. of self-induction is 100,000,000 such a circuit would have an absolute E. M. F. of self-induction 10 times as great, or 1,000,000,000, if the current instead of being varied at the rate of one- tenth unit per second were varied at the rate of one unit per second. Therefore the unit of self-induction, the henry, is equal to 1,000,000,000 = 10® absolute units of self-induction. By agreement among electricians self-induction is represented by the letter L; capacity, by the letter C. Self-induction, when expressed in terms of the fundamental units of length, mass, and time, has the dimensions of a length, and the prac- tical unit of self-induction was formerly called a quadrant on account of the fact that in the metric system, an earth quadrant — i. e., the dis- tance from the equator to the north pole = 1,000,000,000 centimeters, and since the henry = 1,000,000,000 absolute units of self-inductance, it was said to =1,000,000,000 centimeters. In this notation a millihenry = 1,000,000 centimeters. (See art. 125.) 123. The units which have been considered in this chapter have been derived from the relations between electric currents and magnetic fields and are called electro-magnetic units. Another S3'stem of units, also based on the centimeter, gram, and second, called electrostatic units, is in use. The relation between the absolute units of quantity in the two systems is the velocity of light in centimeters per second. This velocity is 30,000,000,000, or 3x10^° centimeters per second, and the electro- magnetic unit of quantity = 3 xlO^" electro-static units. * When quantities are dealt with having a large number of ciphers either before or following the significant figures it is convenient to express them as multiplied by some power of ten, i e.. 10 = lOS 100 = 10', ^ig=lO-\^ij= 10-*. etc. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 77 The coulomb, being one-tenth of the absolute unit, =3x10® electro- static units. The electro-magnetic unit of potential is -g^ of the electro-static unit. In both systems a condenser is said to have unit capacity when unit quantity of electricity charges it to unit potential. From the definition of a farad, given in art. 121, we see that the quantity of electricity in a condenser equals in coulombs the potential Q in volts multiplied by the capacity in farads, or Q = EC, .'.C= -p . Sub- stituting for Q and E their unit values in electro-static units given 3X10* above, C = ^ = 9x10", or the practical electro-magnetic unit of capacity is 9 x 10" times as large as the electro-static unit. The capacity of spherical bodies is found to vary as their radii, and in the electro-static system a sphere of 1 centimeter radius has unit capacity; therefore the capacity of a sphere may be expressed by its radius in centimeters, and capacities are still expressed by some writers and manufacturers by the radius in centimeters of the equivalent sphere. A condenser having a capacity of 1 farad has a capacity equal to that of a sphere having a radius of 9x10" centimeters. A microfarad (see art. 125) =10"'^ farads, is equal to a capacity 9x 10"XlO-" = 9xlO% or 900,000 centimeters in electro-static units. The earth's radius is approximately 65x10' centimeters; its capacity should be approximately 700 microfarads. 124. This difference in nomenclature is very confusing, but it exists particularly with reference to the two qualities of self-induction and capac- ity with which wireless telegraphy is intimately concerned. Microfarads and millihenrys will be used in this book, and where centimeters are found as in some catalogues and some books on electricity, the relations here given — 1 millihenry = 1,000,000 centimeters electro-magnetic units; 1 microfarad = 900,000 centimeters electro-static units — will enable one set of units to be converted into the other. The entire system of units used in electrical measurements is a monu- ment to the ingenuity of scientists, but productive of many difficulties to students. Careful study is, therefore, necessary. 125. While the volt, the ampere, and the ohm are really practical units, the farad and henry are too large for practical use. It would take a very large condenser to have a capacity of 1 farad and a coil of many turns to have a self-induction of 1 henry. Sub- divisions of the farad and henry are in practical use. Multiples and subdivisions of the other units are also frequently used, and for this purpose the prefLxes kilo, meaning 1000; mega, meaning 78 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 1,000,000 ; milli, meaning j^ , and micro, meaning ^ ^^^ ^^^ , are added to the units, and such terms as — kilowatt kilo volt megohm = 1,000 watts, = 1,000 volts. = 1,000,000 ohms. millivolt = i,Joo ^°^*' 1 milliampere= . ^^ ampere, millihenry = ^^qq ^^enry, microfarad = i^oO^OO ^"''^' microsecond=^-^^^-^^^ second, are in common use. The most common practical units of capacity and self-induction (the qualities of electric circuits with which wireless telegraphy is principally concerned, because they determine the period of vibration) are the microfarad and the millihenry, but even these are too large for convenience. The terms mil, meaning inch; micron, meaning ^-^ — meter; circular mil, meaning area of cross section of a wire having a diameter of ■.^„„ inch, are also in general use among electricians. 126. The E. M. F. (volts) in any circuit connected with a dynamo depends only on the rate of cutting of lines of force (strength of field and rate of revolution). The resistance (ohms) in any circuit depends only on the material, cross section, and length of the conductor forming the circuit. The current (amperes) in any circuit depends only on the E. M. F. and the resistance in the circuit. The power (watts) in any circuit depends only on the E. M. F. and current in the circuit. The self-induction (henries) in any circuit depends only on the shape and length of the circuit, on the magnetic permeability (art. 25) of the material surrounding and inclosed by the circuit, on the amount of this material and on the position of the circuit relative to other circuits. The capacity (farads) in any circuit depends only on the shape and area of its surface, on the electric permeability (art, 46) of the material surrounding the circuit, on the amount and location of this material (the dielectric), and on the position of the circuit relative to other conductors. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 79 (Straight wires are said to have distributed inductance and capacity, coiled wires have concentrated inductance, and condensers have con- centrated capacity.) The coulombs in a charged condenser or circuit depend only on the capacity and potential of the condenser or circuit. 127. From the foregoing we can make up a table of values as follows :— A volt= 100,000,000 = 10« absolute units of E. M. P. An ohm =1,000,000,000 = 10"* absolute units of resistance. An ampere = y\ =10"^ absolute units of current. A watt=a volt X an amp. = 10^X10-^ = 10^ absolute units of work per 8econd=l Joule per second= y^ H. P. = 0.737 foot-pounds pei second, A horse power=746 watts. A kilowatt=1000 watts. ^ ^^'^^= 1,000,000,000 =^^" '^^^^^^ "°^*^ '^ ^^P^^^*y- A farad in electro-static units = 9xl0" centimeters. A microfarad = aaa nAr> farad=10-^'' absolute units of capacity. A microlarad in electro-static units = 900,000 centimeters. A henry =1,000,000,000=10" absolute units of self-induction. A millihenry = :.^. henry = 10' absolute units of self-induction. A millihenry in electro-magnetic units = 1,000,000 centimeters. A standard Leyden jar or plate having a capacity of .002 microfarad has been adopted for naval use. In electro-static notation 1 standard jar has a capacity of 1800 centimeters. Chapter IV. CAPACITY AND SELF-INDUCTION. FUNDAMENTAL EQUATION OF WIRELESS TELEGRAPHY. 128. It was stated in art. 56 that the period of electrical vibration of any circuit depends onl}' on the capacity and self-induction of the circuit. Lord Kelvin proved many years ago that when the ratio of the resistance to the self-induction of a circuit is small, the circuit will vibrate in a certain period, which is found to be equal in seconds to 'lirV LC (3) when L is measured in henries, C is measured in farads, 7r = 3.1416. This is called the fundamental equation of wireless teleg- raphy. (See table 7, appendix A.) If R is greater than ^ J ^ the circuit will not vibrate at all. For instance, when a condenser is discliarged through a wire of great resist- ance the charge leaks out slowly witliout any oscillation. A nonoscillatory condenser discbarge, as compared with an oscillatory discharge, is like the flow of molasses into a jar as compared with a large and sudden flow of water into a similar jar. One takes up a position of equilibrium slowly but surely, while the otber vibrates and splashes and only settles down after a considerable period. Equation (3) shows that a circuit having a self-induction of 1 henry and a capacity of 1 farad would have a period of 27r= 6.2832 seconds. Its wave length would be 1,168,000 miles. The standard wave length originally adopted for naval wireless tele- graph circuits was 320 meters; the period was approximately -g-iyi/oiFTF second, that is, they made approximately 900,000 complete vibrations per second. The usual capacity in these circuits was 0.014 microfarad (seven 0.002 microfarad jars in parallel). Therefore the self-induction must have been 0.0022 millihenry. The standard wave length was in- creased first to 425, then to 600 and recently to 750 meters for large ships. It will be noted that the period of a circuit varies as the square root of the product of the inductance and capacity, so that doubling either of these increases the period by V2, i. e., to 1.414 times its former value. Doubling both inductance and capacity doubles the period. SELF-INDUCTION. 129. We see that all conductors must have self-induction, because we know that all currents are surrounded by magnetic fields produced by MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 81 the currents. The production of the field creates an E. M. F. in the circuit opposite in direction to the E. M. F. causing tlie current and tending to stop it, so that self-induction has been defined in a qualitative manner as the inherent quality of electric currents which tends to impede the introduction, variation, or extinction of an electric current passing through an electric circuit. It has also been expressed in quantity as the number of lines of force induced in a circuit by the establishment of unit current in it. It bears the same relation to electricity as inertia does to matter; it represents its resistance to change of condition, and it might be defined as the work necessary to create imit current in a circuit. Suppose we wish to determine the work done in creating a current of value 7 in a circuit of self-induction L in a time T. Since L=the counter E. M. F. of self-induction when the current is varied at the rate of 1 ampere per second, the counter E. M. F. when it is varied at the rate of -=- amperes per second = ~rn - If ^^^ rise in current is uniform, the counter E. M. F. is uniform and the total work done (which equals the product of the E. M. F., current, and time) would be equal to -^ xl xT = LP, were it not for the fact that the current rises uniformly from zero to / and its mean value is -^ and LP therefore the work done=: ~^- (4). Since the factor of time does not appear in the result it shows that it requires the same amount of work to create a given current in a circuit of given self-induction whether it is created slowly or quickly, and that this work is equal in joules to one-half the product of the self-induction in henries by the square of the current in amperes. Therefore in a circuit whose self-induction is 2 henries the work done in creating a steady current of 10 amperes is equal to ^ =100 joules =73.7 foot-pounds. These 73.7 foot-pounds represent the energy stored in the magnetic field; it is the work done by the circuit in creating its own field. If it is in the neighborhood of other circuits, the momentary current created in them during the rise of current reacts on the field and makes the amount of work required still greater. When the current is broken the collapse of the field restores this energy to the circuit, thus tending to prolong the current. In alternating currents, where the rise and fall is continuous, the magnetic field is continually absorbing or giving out energy. In oscil- lating circuits the energy is constantly changing from magnetic to electric and vice versa. 83 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. CAPACITY. 130. Now suppose we wish to determine the work done in charging a condenser of capacity C to a voltage or potential £' in a time T. The potential of the condenser is zero before charging begins and increases as the charge increases, so that the resistance to charging also increases with the charge; therefore it must take more work to add a coulomb of electricity to a condenser of high than to one of low potential. The total quantity of electricity in coulombs in the condenser is Q = E C, and assuming that the condenser is charged at a uniform rate, CE the coulombs per second flowing into it= _, , and this must equal the amperes in the charging circuit. The condenser being charged at a uniform rate, its potential will rise uniformly from zero to E and the total work done during the time T must equal the average potential 4- Xrate of chargextime= ^X^^ XT = ^ (5). Since the factor of time disappears, this shows that it requires the same amount of work to charge a given condenser to a given potential whether it is charged slowly or quickly, and that this work is equal in joules to one-half of the product of the capacity in farads by the square of the potential in volts. Therefore, in a circuit whose capacity is 2 farads, the work done in charging it to a potential of 10 volts = -*-^^^':^ =100 joules = 73.7 foot- pounds. We see that it takes the same amount of work to charge a condenser whose capacity is 2 farads to a potential of 10 volts as it does to create a current of 10 amperes in a circuit whose self-induction is 2 henries. If the capacity of the condenser is 2 microfarads instead of 2 farads, the required work is one-millionth of 73.7 foot-pounds = 0.0000737 foot- pounds. These 73.7 foot-pounds represent the energy stored in the electric field, just as the 73.7 foot-pounds in art. 129 represented the energy stored in the magnetic field. Disregarding losses it is the amount of work the condenser can do on discharge. CONDENSERS AND INDUCTANCES IN SERIES AND IN PARALLEL. 131. When two or more condensers are placed in parallel (fig. 28c), their total capacity C is equal to the sum of their capacities taken singly ; i. e., C = (7i-fC2-f-etc. When two equal condensers are placed in series MANUAL OF KADIO TELEGRAPHY AND TELEPHONY. 83 (fig. 28d), the resulting capacity is one-half of that of each taken singly, or in general i = 'k + i + "''■ For instance, 32 jars connected in 2 groups, with 18 jars in parallel in each group, would, if the two groups were placed in series, have a capacity equal to only 8 jars in parallel. Inductances, however, follow the law of resistances. Two equal induc- tances in parallel have a total inductance equal to half that of each taken singly. Or, when in parallel, -y- = ^ +-^ , etc., whereas in series L=Li-f L2, etc. ; that is, 2 or more inductances, in series, have a total inductance equal to the sum of the individual inductances. Hh s rHHS n 1 3 'A Fig. 28c. Fig. 28u. Condensers, which will be ruptured if used alone, can be used in series, dividing the voltage between them. For instance, a 30,000-volt trans- former can be used with jars, which will stand but 20,000 volts by placing 2 groups in series, as in fig. 28c. Then each jar will have to stand but 15,000 volts. 132. We know that by coiling a wire we can increase its self-induction and, therefore, its electrical length, without any increase in its physical length. So we add to the self-induction of circuits and, consequently, to their periods of oscillation, by adding coils of wire in series, and this is done in practice for both sending and receiving. We can decrease their capacity and, consequently, their periods of oscillation, by adding condensers in series; but this is usually done in practice for receiving only. If a straight wire is broken in the middle, the oscillation period of each half would be half of the original period were it not for the fact that the adjacent ends of the wire and the air between them form a small condenser, which has the effect of slightly increasing the capacity of each half, thus giving it a period slightly longer than half of the original period. It appears, therefore, that we can shorten the electrical length of an aerial (fig. 29) by putting a condenser in series with it, but we cannot shorten it to less than one-half of its original period. 84 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. COMBINATION OF SELF-INDUCTION AND CAPACITY IN OSCILLATING CIRCUITS. 133. In an oscillating circuit, when the condenser is discharged — i. e., when the coatings are at zero potential — the electric energy has been transformed into magnetic energy. If there were no losses in the con- denser due to heating, radiation, etc., the conversion would be perfect, the work in the magnetic field of the circuit referred to in art. 129 would equal 73.7 foot-pounds, and this, in turn, would be again trans- formed into electric energy when the condenser recharges. (See art. 54.) A magnetic field can not be maintained steadily except by a current, but a condenser can be charged and kept in that condition for some time. However, condensers used in wireless telegraphy are always dis- charged immediately, and the energy stored in them before discharge is the stock in trade, so to speak, of the sending apparatus; it represents the work it can do on the ether. This is true, whether the capacity is con- centrated, as in a cundenser, or distributed, as in an aerial. CAPACITY AND SELF-INDUCTION OF STRAIGHT WIRES. 134. The capacity and self-induction of all but very simple forms of circuits are very difficult to calculate, and in general they are deter- mined by comparison with known values. The capacity of a straight, vertical wire of length I and diameter d, well above the earth and away from other conductors, is C = r^yr 4.1454/0^(1] values being given in centimeters. Fleming states that a wire 111 feet long and diameter 0.085 inch, suspended verticalh^ was found to have a capacity of 0.000205 micro- farad, or approximately one-tenth of one standard Leydeu jar. Four wires of the above size and length, being 6 feet apart, were found to have a capacity of 0.000583 microfarad, or about three times as much as one wire. One hundred and sixty such wires in the shape of an inverted cone, 2 feet apart at the top and in contact at the bottom, had a capacity of only about thirteen times that of a single wire. It will be seen that doubling the wire in an aerial does not double its capacity. For wires about 2 feet apart the capacity increases approx- imately as the square root of the number of wires — that is, IG wires would give four times the capacity of 1 wire. The self-induction of a straight wire of length I, diameter d, and cir- cular cross section, at a distance from other conductors is 2 I (2.3026 log. —=--1), values being given in centimeters. The self-induction MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 85 of two parallel wires varies as the distance between them, decreasing with the distance, so that adding straight wire to an aerial does not add to its self-induction in the same proportion. The relation between tlie inductance and capacity of a straight wire of circular section and diameter small in comparison with its length is such that its electrical length is equal to its natural length, and its wave length is therefore twice its natural length. A vertical straight wire, well grounded and of small diameter, has an apparent electrical length approximately equal to twice its natural length ; its wave length is approximately four times its natural length. Pierce states that a single wire 100 feet long and ^ inch diameter, when alone in space, has as much capacity as an isolated flat metallic ditc 16 feet in diameter.* TIME CONSTANTS OF CONDENSERS AND INDUCTIVE CIRCUITS. 135. Every capacity and inductance has what is called its time constant. The time constant of a condenser is equal to C R — i. e., the product of its capacity and the resistance through which it is charged. If C is measured in microfarads, R must be measured in megohms, and their product will then be in seconds. The greater the time constant of a condenser the longer time it will take for it to arrive at a given fraction of the charging potential. For any usual transformer charging frequency this effect is inappre- ciable. The time constant of an inductive circuit = ^^ . The greater the time K constant of a circuit the longer it takes to establish a current of a given strength in it (art. 30). DIFFERENCE BETWEEN DIRECT AND ALTERNATING CURRENTS DUE TO SELF- INDUCTION AND CAPACITY. E 136. The fundamental electric equation 7= ^^s derived from meas- xt urements of the relations existing between electric current and a con- stant E. M. F. in a circuit of constant resistance. Self-induction only affects a current when it is being started or stopped. It increases the time it takes for the current to rise to its steady value and the time it takes to fall to zero. For continually changing currents both in strength and direction it impedes both rise and fall, and therefore acts as a resistance, so that the resistance of a circuit for alternating currents is not the same as for steady or direct currents, but is a combination of the ohmic resistance and the induc- • Principles of Wireless Telegraphy, by G. W. Pierce, A.M., Ph.D. (1910). 86 MANUAL OF RADIO TELEGRAPHY AND TELEPHGNT. tive resistance or reactance (art. 30). Reactance is not a true ohmic resistance, which appears as heat, but is rather a counter or opposing E. M. F. The action is still further complicated in circuits having capacity, as wireless telegraph circuits have, since capacity is found to assist both the rise and fall of current, and therefore to act in an opposite direction to the self-induction and to decrease the total resistance or impedance. In alternating circuits we have /= ~ where Z =i\\Q impedance = Zi J 'B?-\- 2nNL ^ N being the frequency of the alternating current. Since capacity and inductance produce opposite effects, they can be used to neutralize each other. If 2ttNL= o-nP ^^^^ equation becomes 7= -^ as for direct currents, E being the instantaneous value of the E. M. F. In circuits where the resistance and capacity are very small, and the self-induction comparatively large, as in primaiy sending circuits, 2r = approximately 27rNL, or the current depends almost entirely on the reactance of self-induction. The current in wireless telegraph sending circuits is sometimes governed by reactance regulators placed in the primary circuit. (See art. 30.) 137. The equation P = IE(2) is also derived from the relations existing between power, current and E. M. F. in direct current circuits. In alternating current circuits, the current and E. M. F. due to the effect of self-induction and capacity do not reach their highest and lowest points together except when as pointed out in the preceding article, 27rNL=- ,-,^ . The release of energy stored in an inductance creates 2TrNC -^ a current in the same direction as the inducing current. The release of energy stored in a capacity creates a current in the opposite direction to the charging current. Inductance in a circuit, therefore, delays the reversal of a current; i. e., causes the change to lag behind the electro- motive force which produces it. Capacity on the other hand, assists the reversal of a current; i. e., it produces a leading current, or one which is ahead of the electro-motive force which produces it. Since the power at any instant is equal to the current at that instant, multiplied by the E. ]\I. F. at the same instant, the product of the readings of the A. C. voltmeter and ammeter in such a circuit does not give the true power expended in the circuit, but only the apparent power. The true power can generally only be obtained by the use of a wattmeter, which auto- MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 87 matically multiplies the instantaneous values of voltage and current and indicates the product. The ratio of the true power to the apparent power is called the powei factor. SKIN EFFECT IN HIGH-FREQUENCY ALTERNATING CURRENTS. 138. Another effect of alternating currents on the apparent resistance of circuits is seen when the frequencies are above 100. It is called by Fleming the phenomenon of skin or surface resistance. The current seems to begin at the surface of a conductor and soak in, and to penetrate to the center it must have time. This is another instance of the time effect that must be kept in mind when dealing with alternating and oscil- lating currents. If the wire is of iron its comparative increase resistance for high-fre- quency currents is still greater than that of non-magnetic wire. The resistance of No. 16 wire for frequencies of a million is 6.5 times greater than its steady resistance. The larger the diameter of the wire the greater the proportional increase in resistance. Stranded wire, having proportionally greater surface than solid wire of the same area of cross section, offers less resistance to high-frequency currents. Flat ribbons, having larger surface, offer less resistance than circular wire of the same area of cross section. In the Stone receiving circuits, the inductance coils were wound with wire of such size that, for the frequency intended, the current penetrated to the center and there was no wasted material. Resistance is decreased by using a number of strands in parallel. Fleming advocates this for send- ing circuits, and Marconi uses very heavy stranded wire for inductances at high-powered stations. Currents in wireless telegraph circuits having a wave length of 300 meters penetrate about yV millimeter, or approximately ^^ inch, inside the surface of the conductor. If the wires are of iron the current pene- trates about -j-j^^^^inch. 139. We see, therefore, that in wireless telegraph circuits the resistance is not a constant, but depends on the frequency, i. e., on the wave length we are using. Generally speaking, the shorter the wave length the greater the resistance, though we have actually less length of wire. Dr. Austin has shown that for grounded circuits, such as aerials, there is a point where the resistance is lowest and an increase or decrease in wave length will increase the resistance. He finds that resistance varies nearly inversely as the square of the wave length up to a point which is slightly less than twice the natural wave length of the aerial. Beyond this point the resistance rises again nearly 88 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. in direct proportion to the wave length. This rise in resistance appears to be less on ships, where there is good ground, than on shore, where the ground is comparatively not as good. For instance, the aerial at the Bureau of Standards had a resistance of 32 ohms for wave lengths of 400 and 2000 meters, while its resistance for a wave length of 650 meters was about 12 ohms. In the case of a ship, the Maine's aerial had a resistance of 10 ohms for a 400-meter wave. A resistance of less than 5 ohms, for a 500-meter wave, and a resistance of but little more than 2 ohms for a 750-meter wave. For inductances such as are used in closed sending circuits and in receiving circuits he found that the resistance also varies inversely as the wave length. For very long wave lengths and small coils the resistance was not much greater than the direct current resistance, but generally the high frequency resistance increased faster than the number of turns of the coil, but not as fast as the self-induction, i. e., the self-induction of a circuit can be in- creased without increasing its high frequency resistance in the same proportion. MEASUREMENT OF INDUCTANCE AND CAPACITY IN OSCILLATING CIRCUITS. 140. By comparison with standard inductances and capacities, the capacity and self-induction of circuits can be measured and their periods calculated. Measured inductances and capacities connected together so as to form an oscillating circuit are made so that the capacity or inductance (usually the capacity) or both are variable. They can be calibrated so as to show directly either the period or wave length of the circuit for any position of the variable elements. If brought near another circuit in which electrical oscillations are taking place and adjusted so as to have a maximum of current induced the two circuits are said to be in tune or resonance. (They have the same electrical length.) When used as above, calibrated oscillating cir- cuits are called wave meters, ondameters or cymometers. Wave meters can be so arranged as to measure separately the induc- tance or capacity of oscillating circuits as well as their periods. If a spark gap forms part of the oscillating circuit, its period can also be directly measured by measuring the time between the successive surgings of the spark. This is done by photographing the sparks by reflection from the surface of a rapidly revolving mirror. The movement of the mirror between sparks separates their images on the photographic film, and knowing the number of revolutions of the mirror per second, the elapsed time between sparks can be calculated and hence the period of the circuit. Chapter V. POWER EXPENDITUEE AND EFFICIENCY OF SENDING AND EECEIVING APPARATUS. 141. With a given power the work that can be done per second is fixed (JVJ2 In charging a condenser TF= ___(5) . The number of times this is done per second gives the work per second, or the 'power expended. By increasing the frequency we can for a given power either reduce the voltage (length of gap) or the capacity of the condenser. For instance, at a frequency of 500 cycles, for the same power, the condenser need only be 1/10 the size of that for a frequency of 50 cycles. Or, keeping the capacity the same, the voltage can be reduced to 1/V 10 = approximately 1/3 of that necessary for the same power at 50 cycles. A table showing the capacities necessary for given powers at different frequencies and voltages is given in table 2, appendix A. MECHANICAL WORK DONE IN MAKING DOTS AND DASHES OF THE TELE GRAPH CODE. 142. We are now in position to speak in more specific terms of the work done in sending wireless telegrams. Let us suppose that we are delivering 2 kilowatts at 60 cycles and 110 volts to a transformer, which delivers it to a condenser at a maxi- mum potential of 30,000 volts. Two kilowatts =2000 watts = 2000 joules per second = 1474 foot-pounds per second. Since 60 cycles=120 alternations per second, the work equals approxi- mately 12.3 foot-pounds per alternation. If the work done on the condenser is in phase with the charging E. M. F., and if the spark gap is set to break down at a potential of 30,000 volts, the condenser will be discharged at the peak of the charging curve, or when one-half of the work that can be done in an alternation (12.3 foot-pounds) has been done on the condenser. The capacity of a condenser which takes 12.3 foot-pounds of work to charge it to 30,000 vclts = .0372 microfarad, or approximately eighteen 0.002 microfarad jars. Suppose we are sending at the rate of 20 words per minute, that the words average 5 letters each, and that each letter is made up of 3 char- acters equal in length to 9 dots, then a minute can be represented as equal to 20x5x9 = 900 dots=15 dots per second. In other words, 90 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. the length of a dot is one-fifteenth of a second. Now we have 120 alter- nations per second, so that we have about 8 alternations per dot when sending at the rate of 20 words per minute; therefore a dot is made up of 8 distinct sets of discharges of the condenser and a dash of three times that number. The condenser is doing work in producing ether waves at the rate of 12.3 foot-pounds per alternation, equaling, approximately, 100 foot-pounds per dot and 300 foot-pounds per dash. It will be noted from the text that at this sending rate the frequency necessary to give 1 alternation per dot and 2 alternations per dash is only 7^ cycles per second. It will be noted further that with one spark per alternation we cannot utilize 2 kilowatts continuously. We can only use it in charging the con- denser during the first half of each alternation. As soon as the discharge begins the condenser circuit oscillates in its own period as if entirely disconnected from the transformer. In this respect the charge and discharge of a condenser resembles the loading and firing of a gun. 143. Professor G. W. Pierce, of Harvard University, has measured tht period of some types of oscillating circuits used in wireless telegraphy, and it is from his published account of his experiments that the follow- ing description is derived. Suppose a spark gap set to break down at a potential of 10,000 volts,, to be used in a circuit where the maximum potential reached in the condenser is 30,000 volts. Let the curve of sines in fig. 18 represent the condenser potentials: of the oscillating circuit during 2 alternations, each lasting y^s of a second. The resistance of the spark gap is practically infinite before the poten- tial reaches 10,000 volts, and therefore no current passes. When the potential has risen to 10,000 volts the spark gap is ruptured. Its resist- ance decreases instantly to a fraction of an ohm, and during the first half of the oscillation the condenser is discharged to zero potential. Dur- ing the last half of the oscillation it is charged again in the opposite sense. The sparks pass first in one direction and then in the other, and the spark gap not regaining its resisting qualities, the oscillations or surgings continue until the potential (owing to losses due to the radiation of energy in the shape of electric waves, to heating the circuit, and the light and heat at the spark gap) does not rise high enough to disrupt the gap. The transformer Immediately recharges the condenser, which, as soon as it again reaches a potential of 10,000 volts, breaks down the spark gap again, and a second series of oscillations begins. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 91 In the circuit under consideration the maximum charging potential is 30,000 volts, so that a condenser with a spark gap breaking down at 10,000 volts may be charged and discharged several times during one- half cycle of the charging current. The spark acts like a trigger which suddenly releases the stored energy in the condenser, and as soon as this energy has been radiated, the trigger automatically resets itself and does not release again until the condenser is recharged. It is evident that if the spark gap in the circuit under consideration is adjusted to 30,000 volts, but one discharge of the condenser per alterna- tion will take place and but one train of waves will be sent out. Shorten- ing the gap will increase the number of discharges per alternation. The exact number for any spark-gap length will depend on the time of an alternation — i. e., the frequency, and on the length of time it takes the available power to charge the condenser to the voltage required to break down the gap. Less energy per wave train will be radiated on a short gap than on a long one, because the work done varies as the square of the voltage (see art. 130) ; but the total work done may be equal, on account of the greater number of discharges. If the spark gap is too short, an arc is formed and no oscillations take place except those due to the frequency of the charging current. Professor Pierce has shown that the interval between wave trains may vary on account of the residual charge left in the condenser. When the spark gap's original resistance is restored, the potential of the residual charge may be opposed to the potential of the transformer and delay the charging. He has shown also that the gap sometimes partly retains its conducting character and breaks down at a lower potential than its length would indicate. This makes the sparks and oscillations irregular in strength and number and produces ragged and poor signals. E 144. Keeping in mind our five equations — /= „ (1); P~IE (2): period (T) =2^VLC (3) ; 1F= -^ (4) ; and W= ^^'' (5). Let us consider a condenser having a capacity of 0.02 microfarad (10 standard jars in parallel) charged to a potential of 30,000 volts. Such a condenser would contain ynTYT^-fnyf-v =0.0006 coulomb, and 1, , ^^ f A ■ 1 i^2X 10«X 9X 10» o . , would be capable of domg work equal to ^ =9 joules = 6.64 foot-pounds. If this condenser is discharged through a circuit having a self-induc- tion of such value (0.00125 millihenry) as will give a wave length of 300 meters, the frequency of the circuit is 1,000,000, the alternations 2,000,000 per second, and 0.0006 coulomb will create in such a circuit 92 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. a momentary current having an effective value of approximately 2700 amperes.* If this energy is radiated in five complete oscillations, the rate of doing work, if the efficiency of conversion is unity, is 9 joules in Tu-olTr¥f second =1,800,000 per second = 1800 kilowatts. This shows that though the available energy is very small, the rate of doing work, that is, the power of a wireless telegraph sender, may be very great for an exceedingly short period of time. 145. Let us now consider the case of continuous oscillations. The Elaine's aerial, mentioned in art. 139, had a capacity of ,001 mf.. half of one standard jar. Considered as a condenser, charged to 1000 volts, it would contain .000001 coulomb, which would create in a circuit of a frequency of 1,000,000 (i. e., 2,000,000 alternations) a momentary, effective current of approximately four amperes. If this energy is radiated in five complete oscillations, the rate of doing work (that is, the power) is equal to .1 k. w. E A^ When oscillating, the potential of the charge in the Maine's aerial, at tho instant of reversal of the current, might be represented by the line E C, fig. 18e. The point C, being the ground at zero potential, the point A being the upper end of the aerial at the maximum potential. The total charge in the aerial will be its capacity multiplied by the mean potential from C to A, The total work it can perform will be its capacity multiplied by one-half the mean of the squares of the potentials from C to A. Suppose this mean square is 10000^, then this aerial in dissipating * Statement in preceding editions of this book, that the momentary current in such a case is 1200 amperes, was based on the average coulombs passing a given point in a half-cycle, and was incorrect. It was intended to show the necessity for large surface area in inductances of radio sets. At high-powered stations, Marconi installs condenser leads and inductances with very large surfaces. For low-powered stations, however, no such provision is made. But losses from small wires may be greater than we perceive, since recent experi- ments by Alexanderson have shown that corona losses may begin as low as 3000 volts, and he dissipated one-half kilowatt in two wires, each y*o -millimeter diameter, 1 meter long, in air, at a distance apart of two feet Difference of potential, 2700 volts; frequency 100,000 cycles. The wires did not heat, but were surrounded by the blue flame of the corona. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 93 such a charge would do work at the rate of 10 K. W. on a maximum voltage of about 14,000 volts. If the energy is supplied as fast as it is dissipated, the rate of doing work will be greater than 10 Iv. W., because the amount of work done during the first oscillation is greater than during any suc- ceeding oscillation. This is due to the fact stated in art. 100, that, with damped oscillations, the amplitude of each oscillation is a given fraction of the one preceding it and to the fact that the work done varies as the square of the amplitude. For instance, suppose the first oscillation has an amplitude of 10 and the next of 8; the fraction is j-^ and the third oscillation has an amplitude of 6.4. The work done is as the squares of the amplitudes, or as 100 to 64 to 40.96, etc. From which it will be seen that 36 per cent (100-64) of the total work is done during the first complete oscillation, about 23 per cent (64-40.96) during the second, and so on. 146. From the foregoing we see that if we know the capacity of an aerial, its damping, its maximum potential and its frequency, we can determine very closely the amount of power dissipated by it, if it is oscillating continuously. If oscillating intermittently, we must know also the frequency of tlie source of supply. Standard methods of measuring all of these quantities, except the maximum potential, have been developed. EFFICIENCY OF SENDING APPARATUS. 147. On account of unavoidable losses, we cannot supply to the aerial all of the energy supplied by the transformer, nor all of that stored in the condenser. With a certain wireless telegraph set, on which experimental measure- ments were made, Fleming found the actual power radiated to be about 10 per cent of that supplied to the transformer and 20 per cent of that sup- plied to the condenser.* The arc set is roughly 15 per cent efficient, over-all, on powers below 20 K. W., and 50-60 per cent efficient for sets between 50 and 100 Iv. W. Professor Fessenden and Dr. L. W. Austin, in the Brant Eock experi- ments, found that, with the set on which measurements were made, about 75 per cent of the power delivered to the spark gap was given to the aerial. Other experiments have shown that 80 per cent to 90 per cent of the energy delivered to the transformer is transferred to the closed circuit. We may conclude that 60 per cent (80x75) is a fair over-all efficiency for a wireless set, i. e., 60 per cent is delivered to the aerial, where it is dissipated partly in heat and partly in radiating electric waves. As yet we have no standard means of separating these two factors. *Journal Institution of Electrical Engineers, vol. 44, London, April, 1910. 94 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Very few complete investigations of the efficiency of wireless telegraph sets have been made.* LOSSES IN CONDENSERS. 148. When a piece of iron is magnetized and demagnetized — i. e., goes through a cycle of magnetization — a certain amount of energy is ex- pended, which appears in the shape of heat in the iron. It is supposed to be due to internal friction in the molecules of the iron and is called magnetic hysteresis. In the same way, to put a condenser through a cycle of charge and discharge requires the expenditure of a certain amount of energy, which appears as heat in the dielectric and is called dielectric hysteresis. The loss of energy due to this quality varies in different dielectrics and is a function of the frequency. In choosing condensers for the closed sending circuit, it is of great importance to find those which will absorb a minimum of energy and at the same time show no tendency to break down under the large differ- ences of potential impressed upon them. The losses of energy in condensers are of two kinds : internal losses produced by dielectric hysteresis, and external losses produced by the brush discharges at the edges of the conducting surfaces. The ideal dielectric in respect to the internal losses is air, as it is entirely free from internal energy absorption. When used at ordinary pressures, however, it is unable to bear any considerable difference of potential. It has been discovered that when the air pressure is increased to tlie neighborhood of 250 pounds, the dielectric strength becomes so great that it is suitable for use at any of the potentials ordinarily used in wireless telegraphy. Compressed air condensers are ordinarily made up in the form of a series of plates so connected that the alternate plates may be charged positively and negatively, and the whole set is enclosed in an air-tight steel tank which can be pumped up to the desired pressure. Such a condenser, while ideal in its electrical properties, is somewhat bulky, and difficulties are sometimes found in preventing leakage of the air. It is therefore common in stations, where the last degree of efficiency is not demanded, to make use of glass condensers, in the form of either flat plates or jars. The conducting surfaces of condensers are now generally formed of elec- trolytically deposited copper. It is generally stated that flint glass is the glass best suited to form the dielectric. Experiments which have been * At Tuckerton, N. J., the antenna resistance, at 7400-meter wave length, is 6 ohms. Tlie Goldschmidt high-frequency alternator produces 135 amperes in the antenna, with an input of 180 K. W., which on a CT'R basis gives an overall efficiency of 60 per cent. At the same station, an arc set with an input of approximately 90 K. W., produces approximately the same antenna current, with an apparent antenna resistance of only 3 ohms. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 95 made show that the internal losses of glass condensers in ordinary use amount to from 2 to 8 per cent of the total energy flowing through them. The losses due to the brush discharges from the edges of the conduct- ing surfaces, which sometimes amount to 30 per cent of the total energy, may be much reduced by immersing tlie condensers in oil or by placing several condensers in series, which reduces the individual potential differ- ence on each condenser, or by covering edges of foil (or copper) and plates with an insulating compound. 149. Dr. Austin has measured the losses in various types of sending condensers and expresses the total losses as an equivalent resistance. He summarizes his results as follows : 1. The losses, in the compressed air condenser used, amount, at a pres- sure of 15 atmospheres, to an equivalent resistance of between 0.1 and 0.2 ohms. 2. Condensers in which brushing is prevented by the nature of their con- struction show no change in resistance between the limits of observation, 4000 to 20,000 volts, indicating that the internal losses are independent of voltage. 3. Leyden jars of commercial types immersed wholly in oil show losses but slightly greater than those of the compressed-air condenser. 4. The paper and micanite condensers measured show very much larger losses. 5. The resistances of the different Leyden jars in air increase greatly over resistances when measured in oil, and vary between 1 and 1.8 ohms at 14,500 volts. Between 10,000 and 20,000 volts the equivalent resist- ance increases approximately in proportion to the square of the voltage. 6. Placing Leyden jars in parallel series, the capacity remaining the same, does not diminish their brushing losses below 20,000 volts. 7. Immersing only the edges of the conducting coatings of Leyden jars in oil gives an equivalent resistance midway between that observed when wholly in air or wholly in oil. 8. Brushing losses are much increased by any semiconducting material on the surface of the glass at the edges of the conducting coatings of Leyden jars. 9. The resistance of condensers increases nearly in proportion to the wave length. 10. The losses in mica condensers, using carefully selected mica, are con- siderably less than in glass-jar condensers. LOSSES IN CLOSED AND OPEN CIRCUITS. 150. The losses in the closed circuit inductance are those due to its high-frequency resistance (art. 138). To these we must add losses from the sound, light, and heat in the spark gap. Of these, the heat losses are considerable, as is shown by the necessity of using blowers for preventing 96 MANUAL OF RADIO TELEGRAPHY AND TELEPHONT. arcing in spark sets and in the necessity for water cooling the electrodes in arc sets. In arc sets, a non-inductive resistance, called a dead or ballast resistance (fig. 29d), must be used to steady the arc and this is an addi- tional source of loss. While, as previously stated, a closed circuit is a persistent oscillator and, therefore, a poor radiator, it does radiate some energy, and this is an additional loss, since the radiation from the closed circuit is not useful. 151. Losses in the open circuit are discussed in art. 147. We want the useful losses — those due to radiation — to be as large as possible. The faster the aerial radiates energy, in the form of electric waves, the less the loss in heat and the more efficient it is as a wave producer. But unless its oscillations are persistent enough to permit selective tuning of receiver circuits, it is a detriment to carrying on communication. We are, for this reason, restricted in spark sets to a certain limit of persistency, i. e., the damping of the oscillations must not exceed a certain limit, which is ex- pressed in the law (appendix D) as follows : " The logarithmic decrement per complete oscillation in the wave trains emitted by the transmitter shall not exceed two-tenths, except when sending distress signals." If successive values of the amplitude of damped oscillations are meas- ured and their relative values expressed in figures, the natural logarithms* of these figures will differ by a constant quantity called the decrement and it is this constant difference which it is specified by law shall not exceed .2. A decrement of .2 gives about 15 complete oscillations before the amplitude falls to yV of the maximum and permits good tuning. 152. Of course continuous oscillations permit much more selective tun- ing, because we can get along fairly well with only 15 oscillations in y^Vo of a second, while we would have 300 oscillations in the same length of time if they were continuous and we were sending a 1000-meter wave. A large decrement in an open circuit, used for continuous oscillations, is, therefore, not a defect; in fact, it serves to increase efficiency if its cause is rapid radiation and not high resistance. A whip-crack transmitter, which is not permissible with spark sets, is all right for arc or other apparatus producing continuous oscillations. RELATION BETWEEN HEIGHT OF AERIAL, OSCILLATING CURRENT, WAVE LENGTH AND DISTANCE OF TRANSMISSION. 153. From the results of experiments made at Brant Eock in 1909-1910, Dr. Austin has expressed the relation between the vertical height of the * The logarithm of a number is the power to which another number called the base must be raised to equal the number in question. For instance the logarithm of 100 to the base 10 is 2 because 10 must be squared to equal 100. The base of so-called natural logarithms is the number 2.7183 and is usually represented by the letter e. The log of base e to base 10 is 2.3026. Base e is the one to which the term logarithmic decrement (as used in the law) refers. The base of common logarithms is 10. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 97 aerial, the amount of oscillating current, the wave length and the distance of transmission in a formula, which will be found in appendix A, last part of table 7, and the working of the formula illustrated in tables 8, 9, 10, 11 and 12. Eef erring to table 9, it is found in practice that, while the relative distances given are approximately correct, usually the reliable working distance is about two-thirds of that given in the table. Subsequent experiments tend to prove that the formula is correct. Therefore, the tables (while not absolutely correct, except for good weather conditions and expert operators with the best receiving apparatus) are relatively correct and repay careful study. EFFICIENCY OF RECEIVING APPARATUS. 154. Fessenden in his published account of his experiments on the sensitiveness of wireless telegraph detectors states that in the most sen- sitive detectors the least amount of work which will render a signal readable is .007 erg per dot. If a dot lasts -zr- of a second this represents approximately .01 erg per second, or -^-r^ watts. Dr. L. W. Austin's tests of telephones in 1908, several years subsequent to Fessenden's experiment, indicated that in order to produce audible 3 sound in a«telephone not less than yy— watts were required. With tele- phones of the sensitiveness previously available it required not less than 3 . :j-^ watts. With test telephones of the sensitiveness available in 1914, Dr. Austin estimates the power required (at 1000 sparks per second) 4 to make the electrolytic detector function audibly, as ztt.— watts. Standard reed telephones in 1914, at 600 sparks per second give an Q audible sound with an expenditure of the order of -^rv. watts. lO^'* INCREASE OF EFFICIENCY DUE TO A HIGH SPARK FREQUENCY. 155. If two alternating currents of the same intensity but of different frequencies be sent through a telephone, it is found that the sound in the telephone produced by the current of higher frequency is much louder than that produced by the lower. This fact is due in part to the peculiarities of the human ear, which is more sensitive to high-pitched sounds than to low, also in part to the diaphragm of the telephone, which is usually of such a weight and size as to vibrate most readily to a sound of rather high pitch. This fact has an important bearing on wireless telegraphy, for the pitch of the sound produced in the telephone con- 7 98 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. nected to the detector at the receiving station depends simply on the number of wave trains per second at the sending station. In order to determine exactly what is the relation between the strength of current required to produce an audible sound in the telephone and the frequency, a series of experiments was carried out on a pair of head telephones of the type ordinarily used in wireless telegraphy, the results of which are shown in the table. 'requency per second. Volta to produce audible sound. Frequency per second. Volts to produce audible sound, 60 6200 X 10-T 540 80 X 10-7 120 2900 660 30 180 1700 780 11 300 600 900 6 420 170 In the first column are given the frequencies or the number of wave trains per second, and, in the second, the number of volts of alternating current which it would be necessary to apply to the terminals of the tele- phone to produce an audible sound. From this it is seen that it requires about a thousand times as much voltage at a frequency of 60 to produce a sound as is required at a frequency of 900. We may assume, therefore, that if the number of wave trains at the sending station be increased from 60 to 900 per second, and the spark length be kept the same, the effect at the receiving station would be increased one thousand times. If the number of sparks be increased in this way without reducing the spark length, it is evident that the energy made use of at the sending station must be greatly increased. It will be more interesting, therefore, to calculate what the increase in sending efficiency of the station will be with increasing spark frequency, if the total energy be kept constant. So if we assume that the energy is proportional to the number of wave trains, and divide the relative increase in loudness of sound in the telephone at the receiving station, for any frequency, by the relative increase of the number of wave trains per second, we will have a fair comparison of the efficiencies at the two frequencies. Streng-th Strength Frequency. of Frequencj'. of sigrnal. signal. 120 1 540 13 240 1.5 900 64 The results of such calculations are seen in the table, which shows that there would be very slight advantage in replacing a 60-cycle alternator giving 120 wave trains per second with a 120-cycle giving 240 wave trains, but that the advantage increases rapidly as the frequency is increased. The maximum sensitiveness of the telephone appears to lie in the neighborhood of 900. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 99 156. In addition to the increase of sensitiveness of the telephone at high frequencies, there are other quite independent advantages in the use of a high-pitched spark. First, it is found in practice that a high- pitched musical signal is much more readily distinguished at the receiving station in the midst of ordinary interference and atmospheric disturb- ances; and second, at the sending station a shorter spark gap, which would generally be used with a high frequency spark, puts less strain on the insulation of the condensers and other parts of the circuit, and re- duces the losses due to brush discharges, which in many stations amount to a considerable share of the total amount of power employed. A third advantage is that with a high spark frequency larger amounts of energy can be radiated from a moderate-sized aerial without sub- jecting it to excessively high potentials. Experiments have been recently carried out in which it has been shown that in moderate frequencies with stationary spark gaps there are nearly always secondary discharges, irregular, but giving very high tones, so that the real advantage of the high spark frequency, from the stand- point of telephone sensitiveness, is usually less than that indicated in the table. The advantages of ease of reading, the lessening of the strain on the condensers and insulators, and the increase in effective energy capacity of the antenna, especially when the latter is small, are very marked, so that it has been found possible to obtain the same results with small wireless sets of 2 K. W. capacity where formerly 5 to 10 K. W. were employed. The only difficulty involved in using very high spark frequencies lies in the cooling of the spark gap. For this purpose a rotary gap or some special refrigerating device must be used. For the reasons stated above, 500 cycles has been adopted as a standard frequency for the alternators of spark sets. When receiving from a sender like an arc, producing continuous oscil- lations, efficiency is secured in a manner which will be described under " Eeceiving Apparatus," chapter VII. COMPARISON OF EFFICIENCY USING DAMPED AND UNDAMPED WAVES. The undamped wave is superior in efficiency to the damped wave, for transmitting, due to some extent to the decrease in absorption using un- damped waves, but mainly to the efficiency of the receiving oscillating audion in comparison with the receiving methods used in reception of spark signals, and also to the fact that longer waves are used in trans- mitting with damped waves. LOSSES IN RECEIVING CIRCUITS. 157. The losses due to high frequency resistance in receiving circuits are the same as in sending circuits, except that there is less difference in 100 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. smaller wire between the direct current resistance and the high frequency resistance. Naval specifications require that the decrement of receiving circuits shall not exceed .3, while for sending sets the limit is .15. Specifying the high- est permissible decrement of the receiving circuits is a change from former practice, which was that the d. c. resistance should not exceed 4 ohms per millihenry. 158. Dr. Austin has shown that the most efficient coupling of the detector circuit and aerial is when the incoming energy is divided equally between the two and that then the loudest signals are produced. He has also shown that the loudness of the sound in a telephone varies directly with the power applied, that is, with the square of the potential or oscillating current (art. 117). This assumes, of course, as a primary essential of efficiency of receiving apparatus, accurate tuning of both the open and closed receiving circuits to the incoming wave length. Chapter VI. SENDING APPARATUS. GENERATORS. 159. Induction coils (fig. 14b) with hammer breaks operated by direct current have been used to a very limited extent for naval purposes. The vibrations of the hammer were difficult to regulate and the large size necessary to handle large currents made the frequency too low for successful work. Hammer breaks were soon discarded and make and break regulated by some form of rotary motion. The most successful form was the mercury turbine interrupter. This interrupter was installed in the circuit containing the sending key and the primary winding of the induction coil. The interrupter consisted of a direct-current motor driving a centrifugal pump which revolved in a chamber of mercury. The mercury was connected to one side of a break in one leg of the primary circuit. It was drawn up by the pump and delivered as a jet through a revolving nozzle. The mercury jet during a portion of each revolution struck a metallic segment connected to the other side of the break in the circuit, and, if the sending key was closed, thereby completed the circuit and built up a current in the induction coil which charged the sending condensers. When the jet passed the segments, the circuit was broken. (The jet passed through grain alcohol which absorbed the spark at break.) This make and break occurred once in each revolution. The motor made approximately 1800 revolutions per minute. Assuming that the condenser was discharged only on the break, this gave but 30 dis- charges per second, or a note two octaves * lower, as compared with 120 discharges from a 60-cycle alternator. The operation of these sets was much improved by increasing the number of segments and, therefore, the number of makes and breaks per second, as many as six being used, thus giving a spark note slightly higher than that of a 60-cycle alternator. The spark in the interrupter at break always carbonized some alcohol and the latter also became mixed with mercury and formed a more or less conducting carbon-mercury-alcohol emulsion, so that the interrupter and contents required frequent cleaning, washing, and filling. • The octave of a note is that differing from it by 8 notes of the scale — do-re-mi-fa-sol-la-si-do — the octave above having twice as many vibrations per second and the octave below having one-half as many vibrations as the note referred to. Standard tuning forks vibrate 256 times per second. The pitch of a note is the number of vibrations per second producing that note. 102 IfAl-TDAL OF RADIO TELEGRAPHY AND TELEPHONY. For small powers these sets, with care, gave good results, and being generally used with mechanical recording apparatus the spark note was not of marked importance. CONSIDERATIONS GOVERNING FREQUENCY OF GENERATORS. 160. Turbine interrupters were practically entirely replaced by 60- cycle alternating current generators operated by motors (on ships and at navy yards) or oil engines (isolated shore stations and light ships). These in turn have been replaced by 500-cycle generators operated by motors or engines as above. No special description of generators will be given. Sixty-cycle current was first selected because alternators of this fre- quency were commercial articles. When the use of telephones with receiv- ing sets became general it was realized that a sound of a higher note was desirable and that for the very best results the frequency (pitch) of this note should be that to which the telephone diaphragm or the operator's ear, or both, are most sensitive. A pure spark note is produced when the spark gap is so adjusted that the condenser discharges but once per alternation, thus sending out but one wave train per alternation. The pres- ent standard of 500 cycles affords general satisfaction, though the ears of some operators are more sensitive to a lower frequency (see art. 155) . 161. To ensure a perfectly regular condenser discharge and thus obtain but one wave train per alternation, some generators have a disk mounted on an extension of the main shaft and revolving with it. This disk carries projecting electrodes, one for each pole of the alternator, equally spaced like the spokes of a wheel and connected to one side of the closed circuit. (See fig. 52.) In revolving they pass very close to a fixed elec- trode, or spark point, connected to the other side of the circuit, sparking taking place as the points pass — one series of oscillations for each alter- nation. Generators carrying rotary spark points must of necessity be placed in or near the operating room and are to that extent objectionable, on account of the noise of the spark, the additional space required, and the noise of revolution which interferes with receiving. Many attempts have been made to muffle this type of spark gap, but without success. Motor generators, or generators driven by engines, except as stated above, are usually arranged for being started or stopped from a distance. The controlling apparatus is mounted on a switchboard which carries voltmeters and ammeters, one each for the supply current and one each for the generator current. A frequency meter is also part of the switch- board equipment. This with the field rheostat of the motor enables the operator to adjust the speed of revolution so as to give the required frequency. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 103 TRANSFORMERS. 162. A generator designed for a certain frequency works best in con- nection with a transformer designed for the same frequency. If the size of the condenser to be used in the closed circuit is fixed, and known to the designer of the transformer, the latter can be built so that the secondary winding and condenser form a circuit whose natural period is that of the generator frequency ; a few such transformers have been sup- plied and are preferred to those requiring reactance regulators. Neither generator nor transformer will work without overheating at a frequency much greater than that for which they are designed on account of the increase of heating in the iron cores and frames, with increase of cycles of magnetization per second. An examination of fig. 29 will show that the generator armature wind- ing and the primary winding of the transformer form one circuit, and the secondary winding and condenser another. The reactances of these circuits should be such as to maintain the charging E. M. F. and current in phase with each other. When 60-cycle current was the standard, transformer windings were designed to give a potential of from 25,000 to 30,000 volts in the secondary when the primary was supplied with 110-volt current. Standard transformers now have a maximum effective voltage of 25,000 when supplied with 220-volt current. For small sets both induction coils (open core) and closed core trans- formers are satisfactory; for large sets closed core transformers are pre- ferred. Transformers are fitted with safety spark gaps set at the maxi- mum safe sparking potential. REGULATION OF A. C. SENDING APPARATUS. 163. Sending sets work most efficiently when the interruptions or alternations of current are in resonance with the circuit formed by the secondary of the transformer and the sending condenser. When running on open circuit practically no work is being done by the motor or generator except that necessary to overcome friction. When the primary circuit is closed by the sending key, Avith the spark gap opened, so that no sparking takes place, the secondary of the trans- former charges the condenser during the first half of each alternation and receives current from the condenser during the second half of each alternation. The load thrown on the motor generator by pressing the key depends on the period in a cycle at which contact is made, but, generally speaking, it may be considered as instantaneous " full load." If the spark gap is set so that the condenser potential breaks it down, the oscillations of the closed sending circuit practically cut out the 104 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. secondary of the transformer, so that a condition of instantaneous " no load " exists as soon as the spark passes. As soon as these oscillations cease, the secondary again begins to charge the condenser, and a condi- tion of almost instantaneous full load is established. This interval is 60 short that the inertia of the moving parts of the motor generator pre- vents any change of speed or voltage, so that the instantaneous full load thrown on when the key is closed is the one affecting operation. When the key is closed the momentary current starting at that instant depends only on the voltage at that instant and on the reactance of the primary of the transformer and of the generator armature, since the re- sistance is very low. To control this sudden rush of current an adjustable choke coil, called a reactance regulator, may be placed in the primary circuit. This coil, on account of its inertia, acts as a buffer against sudden changes of cur- rent, and by means of its adjustability enables the phase relation of the E. M. F. and current in the circuit to be varied and thus the power expended to be controlled. Since the reactance regulator controls the power expended, it controls the secondary voltage and the maximum spark gap that can be used. By placing the sending key in shunt around it and having an inductive resistance in series with the key, the reactance regulator can be adjusted so that no sparking will take place, but by closing the key the current added through the shunt circuit is sufficient to cause sparking to take place. By means of this method the sudden changes from full to no load are avoided and the regulation improved, and since only a small portion of the total sending current is broken at the sending key, it is much easier to keep the contacts in good condition. A safety switch is placed in the primary lead when the method of con- trol described above is installed. This switch should only be closed when sending and should be opened at all other times when the motor genera- tor is running. A method of control tried with fair results is to have the sending key, by working auxiliary contacts, weaken the fields of the motor and strengthen the field of the alternator by varying the alternator resistance just before the primary circuit is closed. But no method has yet been developed that does not show some decrease both in frequency and voltage between no load and a long dash. In arc sets, a large ballast resistance (fig. 29d) is necessary in the D. C. current to steady the arc. The alcohol feed and cooling water supply must be regulated; in other respects the arc is automatic. Arc sets are operated by change of wave length only — not by make and break of arc current. The charge and discharge of the condenser when not sparking is indi- cated by a rustling sound, which signifies danger. This warning applies MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 105 equally to induction coils and transformers, both terminals of which are dangerous when using alternating current. On account of the small penetrating effect of high-frequency currents (art. 138), it is believed that high voltages when associated with fre- quencies of above 100,000 per second are not dangerous to human life, but low frequency, high-voltage currents are very dangerous, and it must be borne in mind that a condenser being charged and discharged at the alternator frequency is very much more dangerous than when it is dis- charging across the spark gap. SENDING KEYS. 164. The sending key, or the auxiliary key operated by it, except in arc sets, is placed in one leg of the primary circuit. In arc sets the sending key is used merely to short circuit part of the aerial inductance, i. e., it merely changes the wave-length ; the arc itself is not broken. This is neces- sary because it takes time to start an arc. When placed directly in the primary circuit, sending keys, in some cases, have condensers shunted around them to absorb the spark at break. Their contacts when used to break the primary current direct, are larger than in the ordinary telegraph key on account of the larger cur- rents handled. In other respects they resemble the telegraph key. When used to operate a relay, the ordinary telegraph key fills all requirements. The relay consists of a solenoid energized by the sending key, its arma- ture making and breaking the primary current in air or oil. Figs. 37 and 39 illustrate types of sending keys. The Slaby Arco keys shown in fig. 37 were of massive construction and very rugged. Fig. 38 shows a solenoid break, the connections for which are illustrated by fig. 38a. It will be noted that this is a positive break as well as make. Fig. 39 is practically the same as the ordinary telegraph key with large contacts. It operates direct or through a relay. For very large powers massive solenoid breaks in parallel are necessary, which require special attention in the design to permit rapid operation. Sending keys should be adjusted to have just sufficient movement to prevent arcing and permit well defined making and breaking. For direct breaking, though platinum contacts are largely used, com- paratively large brass or silver contacts are satisfactory. All contacts must be kept smooth and clean and their faces parallel. What is known as a " break key " is preferred. It was first used on the Stone sets, and is an ingenious and useful device for "listening in" while sending. An attachment to the sending key breaks the detector circuit just before the sending key makes contact. When the sending key is released the receiving circuit is automatically cut in, so that the receiver can " break " the sender by a call, which the latter can hear in the interval between his letters or words. 106 MAJSrUAL OF RADIO TELEGRAPHY AND TELEPHONY. r^ ^ T' £^ FiQ. 37.— Slaby Arco Key. Fig. 38. r»/A(^RAM or CONNECTIONS mj] (^0 -■ ^ ^ TRANSrORMER PI?|MARr ma TO no VOLTS D.C. ► Fig, 38a.— Solenoid Key. Fig. 39.— Wireless Specialty Apparatus Co. MANUAL OF KADIO TELEGRAPHY AND TELEPHONY. 107 For sending time signals, a Western Union relay closes a local battery having in circuit a solenoid, whose armature carries a lever which presses and releases the sending key in unison with the current impulses sent from the standard clock at the Naval Observatory in Washington. CLOSED CIRCUITS (INDUCTANCE, CONDENSER, SPARK GAP) . 165. Sending circuits are illustrated by elementary diagrams in figs. 29 to 29e and 40 to 48 inclusive. The symbol | indicates alternating current. The names under figures are those of the engineers proposing or designing sets with the connections shown. To render them capable of adjustment all wireless telegraph oscillat- ing circuits have either variable inductances or condensers or both. These condensers and inductances vary greatly in design. Those for sending circuits, especially on account of the high potentials to which they are subjected, are very different in construction and mounting from those used in receiving circuits. Fixed condensers and variable inductances are used in sending cir- cuits. The condensers may be single, two or more in series, or in parallel. Series-parallel installations may be made also, just as in primary bat- teries. (See figs. 28c and 28d.) The variable inductance usually consists of a helix or spiral of compara- tively large bare wire (round or flat) mounted on an insulating frame a foot or more in diameter, the turns of wire varying from about |" to 2" apart. (See figs. 72 and 73.) Previous to the introduction of quenched gaps, the greater number of sending circuits were direct connected, but inductively connected sets are equally efficient and have this advantage, that the coupling can be readily adjusted independently of the natural wave length of eitlier circuit. (Figs. 42, 43, and 45.) 166. In direct connected sets, three movable clips or sliders are usually provided, one for the closed and two for the open circuit (fig. 40). The closed circuit is permanently connected to one end of the helix and the circuit completed by means of the wire from the movable clip, which can be connected to any desired point. The open circuit has the ground and the aerial wire, respectively, attached to the other two clips and these are attached to such points of tlie helix as will give the open circuit the same natural period as the closed circuit and at the same time give the two circuits the number of turns in common necessary for the desired coupling. 167. In inductively connected sets, the closed circuit helix or spiral is the same as before, the open circuit helix is permanently attached to the ground lead and the aerial lead attached to whatever point is necessary. The mutual induction and coupling are varied by moving the open circuit helix as a whole. (Figs. 42 and 43.) Inductive connections are re- quired with naval spark sets. 108 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. TO AERIAL XO AERIAL k s X X DEFOREIST SHOEMAKEf^ FIG. 40 TO AERIAL TESSENDEN FIG. 41 -TO AERIAL stone: FIG. 42 TO AERfAL , MARCONI STONC FIG. 43 TO AEf?iAL ], sT massie: FIG. 44 . LOWENSTEIN FIG. 47 TO AERiAu s PIERCE SLABY ARCO FIG 48 Note. — Figures 40, 41, 42, 44, and 48 represent circuits that are not in general use to-day, and hence to the student have a historical rather than a practical value. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 109 Making the adjustments to particular wave lengths and couplings is called tuning and is discussed in Chapter VIII. 168. In some Telefunken sets (fig. 46) the sending inductance in both closed and open circuits consists of flat spirally wound coils, mounted parallel and close to each other in a frame. Alternate coils are con- nected to a lever by which their position relative to the others can be varied. The coils are connected so that currents in adjacent coils oppose each other and decrease the self induction of the whole, called by the manufacturers a variometer. By means of the lever the coils can be separated and the self-induction and consequently the period of the cir- cuit regulated. This is illustrated in fig. 46 by an arrow drawn diagonally across the inductances in which it is used. Fig. 74 shows the apparatus as manufactured. Open spiral inductances are preferable (Fig. 72). 169. For older direct connected sets connections shown in figs. 40 and 44 were preferred. In the figures referred to, the condenser is directly across the secondary terminals of the transformer, and the spark gap in one leg of the closed oscillating circuit, as contrasted with the spark gap being placed directly across the transformer terminals. (Fig. 48.) The former is considered to be a more symmetrical arrangement. Attention is invited to fig. 41, which shows one leg of the trans- former directly grounded and the other leg connected direct to the aerial. All other methods of connection afford direct path to ground and path through condenser and spark gap. This method of installation affords path to ground through condenser or spark gap only and affects tuning. If the aerial is touched when current is on the transformer, the latter, having one leg grounded, is short-circuited through the body and a severe shock may be experienced. Though this method of connection is no longer used, it is referred to here to show the necessity of giving careful consideration to the relative positions of ground, spark gap, and condenser. Errors in connections of direct connected sets are sometimes made so that the most direct path to ground is through tha spark gap. This induces potentials at the gap or condenser approximately equal to those at the upper end of the aerial and produces disagreeable inductive effects in the operating room. Some installers prefer to ground one leg of the secondary of the trans- former when the closed circuit is inductively connected, but this is con- sidered unnecessary. 170. Fig. 43 shows the preferred form of inductive connection or coup- ling of a helix, that is, one inductance above the other. This takes up less floor space and the coupling is varied by vertical instead of horizontal movement, as is necessary when the coils are side by side as illustrated in fig. 42. 110 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Spiral inductances (fig. 72) can be arranged in a small space above one another, side by side or at an angle, and are very convenient; they are preferred by some manufacturers. Inductances added to the aerial, such as those on the right of fig. 45, are added for the purpose of lengthening its period. They are called " loading coils." Fig. 47 indicates a method of connecting up sending sets so that the operator by moving a hand wheel or lever can change the wave length of the open and closed circuits the same amount without changing the coupling. This apparatus is just being introduced and should greatly facilitate the operator's control over his sending wave length. It is called a tune shifter. CONDENSERS. 171. For transmitters up to 5 K. W., standard coppered jars in air or oil are preferred. Tinfoil covered jars are no longer supplied. The standard jar has a capacity of .002 mf. Condenser racks or tanks are arranged to hold a number of jars somewhat greater than that necessary for the rated output (see Table 2, Appendix A). Inside connections to Leyden jars are best made by soldering one end of a strip of copper or brass gauze to the inner copper coating and clamp- ing the other end to the charging bus bar. Outside connections are made either by supporting all jars on a con- ducting plate connected to the other charging bus bar or connecting this bar to a strap of sheet brass or copper clamped around the jar. The important point about condenser connections is that they should make a good electrical contact, of comparatively large area, with the charging wire or bus bars and with the condenser jars or plates. A symmetrical arrangement of material giving as nearly as possible equal lengths of discharge paths should be made. Many kinds of springs and clips for condenser connections have been devised and are in use, but none are better than those just described. Less difficulty is experienced with connections on copper coated jars or plates than was the case when tinfoil was used exclusively for distributing the charge over the glass dielectric. 172. The condensers now in use are standard Leyden jars in air or oil. (Fig. 49.) CZa^s p/a/es in air or oil (glass dielectric) . Metal plates in compressed air (air dielectric) (fig. 50) and tinfoil (paper dielectric). For large powers, glass plates in oil or metal plates in compressed air are preferred. For small sets the most convenient for use, installation, and inspection, are the standard jars, in air or oil, or glass plates set ver- tically in oil. Fig. 49a is a special form of Leyden jar which is con- venient for some purposes. The low voltages associated with 500 cycle sending sets have made practicable the use of paper condensers as noted MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Ill Fig. 49b. — Mica Condenser — Dubilier Type. Fio. 49. — Leyden Jar Battery. UJ Fig. 49a. — Moscicki Tube. Fig. 50. — Compressed Air Condenser. 112 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. above for small powers, but they are relatively inefficient as compared with others mentioned. In some of the largest stations, galvanized iron plates have been hung up side by side under cover, forming huge condensers, which require very little attention. 173. Practically all other insulators have a greater specific inductive capacity than air at ordinary pressure, and nearly all of them have a greater dielectric strength than air (Table la, Appendix A). The Leyden jar, having long been used as a high-potential condenser, its method of manufacture being well known, and the best glass having not les than nine times the capacity of air, has been very generally used in wireless telegraph sending circuits. Air and oil, while requiring much larger volume to give the same capacity as glass, have the excellent property of mending them- selves after puncture by a spark, while all kinds of solid or semisolid die- lectrics require renewal after rupture. Mica has very great dielectric strength, as much as 5000 volts per mil, and has been used to some extent in condensers in the form of micanite. The semisolid dielectrics, such as beeswax and paraffin, have to be made up with considerable attention to the temperatures in which they are to be used, since they may melt in summer and crack in winter, but they are cheap and easily obtained. Dielectric strength of insulators per millimeter increases with decrease of thickness, except in oils, where it seems to decrease. Dielectric strength of air increases with increase of pressure. Dielectric strength of air decreases with decrease of pressure until the pressure is in the neighborhood of 1 millimeter of mercury, when it increases. Dielectric strength of a vacuum should be infinitely great. Fleming states that with the best flint glass it is possible to store about 45 foot-pounds of energy per cubic foot of glass. The limit is set by the dielectric strength of glass. He has shown that the lengths of discharge paths of all condenser elements should be equal. Capacity varies inversely and dielectric strength directly, as the thick- ness of the dielectric, but they do not vary in the same ratio. The dielectric strength of glass condensers decreases, that of oil con- densers increases, with the frequency. A table showing the specific inductive capacity of a number of dielec- trics and their dielectric strengths is given in appendix A. This table, la. is by no means complete. Data relative to the hysteresis losses of various dielectrics is almost lacking, and want of agreement is noted among different authorities. DIELECTRIC STRENGTH OF AIR. 174. The dielectric strength of air is considered to be about 4500 volts per millimeter for gaps of about 1 millimeter in length, and about 3000 volts per millimeter for gaps of the length of a centimeter or more. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 113 ' —1 ' -1 ■^ /| "^A r N 5 w ?l K IN G ; I 31 s L M 4C .E •- I i? TV ^^ Er 1 < >H ^Fi ,P N EE D _E F^ Ol MTl S fJT vi»L- M lTC n J? r ?'^ Ci] 1 ir 101 4 r d.A n tw rn m m R Jl ft Wf im ( { m. E / Hi T r / y / / / / IS 7 / l/ / / / y / / n / o /^ ^ ^ / / 2 •^ x" ^ ,^ / / a r •^ y r' / / S J ^ ^ / / ^ '2 e .-r* 1^ h^ / / ,/ / f US' ,/ r ,/ r fl f) > .^ / in 2 rt } o / / 4 y K I.C vc iiT e; / / 19 /' / / / 1 / / / / Q y r / y / / 7 u / / ^, /* C / / 6 r / 7 r / y / $■ / / / A /' 4 f / / ;^ 3 y / i / y Z / / /. y I • ^^ ^ t^ >^ K 1 r ,.0 IT s ^ ^ 2_ a_ --i a_ Jk ft- 5^ -£ e- _^ a_ -^ L. f J ft ?_ Jl ^ O- -a a- n. f— M £_ M B- FiQ. 51. 114 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 51 shows sparking distances for various voltages in air between needle points, as determined by experiment. These distances are usually greater than those obtained from equal voltage between the blunt spark points used in wireless telegraphy. The latter probably correspond more closely to table 1, appendix A. On the other hand, this table of spark dis- tances was determined by raising the voltage very gradually and exactly alike for each gap, while in oscillating circuits there is a convulsive rush which may produce very high potentials. This has been shown by intro- ducing a minute spark gap elsewhere in the circuit, the effect being to greatly increase the gap, which can be ruptured by a given transformer potential. The inertia of the charge carries it forward, and just as the inertia of water in a pipe produces a great pressure if its flow is suddenly checked, so the potentials in the sending circuits may, and usually do, rise much higher than is indicated by the transformer ratio. SPARK GAPS. 175. A great deal of thought and ingenuity has been expended on improving the action of spark gaps. For instance, the use of magnetic blowouts, induced and forced air drafts across the gap; dividing them into a series of short gaps; placing gaps in parallel; enclosing them in compressed air and in nitrogen gas; making the points hollow and cool- ing them with air or water. Until recently, no method of construction for small powers was mark- edly better than the ordinary gap in air between two zinc rods, i to i^ inch in diameter. There are two points in common for all good working gaps — (a) The sparking surfaces must be clean and fairly smooth; (b) They must be kept from heating. The increased radiation from cooled spark electrodes as compared with heated ones is very evident. Heated surfaces give off more metallic vapor and tend to the forma- tion of a low frequency arc. There is no doubt that much of the irregularity noted in sending is due to an improperly adjusted spark gap and the effect known as " soaring " or " swinging " is probably due to the inequalities in the action of the spark gap and condensers caused by heat. An open spark must be kept white and crackling and have considerable volume. If too long, it will be stringy ; if too short, an arc will be formed. All spark gaps are adjustable — either in length or in number. All should be well muffled for obvious reasons. The types of spark gaps now in use are shown in figs. 52-57. The only types now supplied are fig. 52, the synchronous rotating gap, and fig. 57, quenched gap. Fig. 57 illustrates only one form of the quenched gap. It is made in other equally efficient forms following the same principle; viz., a series of very short gaps (arts. 85 and 180). MANUAL OF EADIO TELEGRArHY AND TELEPHONY. 115 SPARK GAPS SHArx or ALTERNATOF^ 5 t SYNCHRONOUS ROTATING SPARK GJ\P NON-SYNCHRONOUS ROTATING; SPARK QAP FIG 52 FiG. 53 AIR '^f^ FIG 54 FIG. 55 u AIR BLAST QAP PARALLEL qAP <€ 4 M( A w f1\ "^ rh VU m w m CI 03 a QyeNCHCD SPARK aAP FIG. 57 =^^=l=D^ (0 (0 MARCONI DISC DISCHARGER FIG. 56 FIG. 58 T-ELEFUINIKEN (Now Obsolete) FIG. 59 STONEl (Now Obsolete) MASSIE QAP FOR COMPRESSED AlR FESSENDEN-J)E FOREST FIG. 60 (Now Obsolete) FIG. 61 (Now Obsolete) 116 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 176. The function of the spark gap in an oscillatory circuit is to allow the condenser to charge to the required potential, and then to break down and permit the charge to surge back and forth until its energy is dis- sipated. The ideal spark gap would be one which would insulate per- fectly while the condenser was charging and conduct perfectly while it was discharging, and the nearer these conditions can be fulfilled the more efficiently will the spark gap perform its duty. Either condition can be fulfilled alone, but the combination is somewhat difficult to obtain. The resistance of the spark gap when the discharge is passing depends upon two factors; it increases rapidly with the spark length, and de- creases rapidly with the oscillatory current, amounting with a half-inch gap to several hundred ohms when a fraction of an ampere passes, and a small fraction of an ohm when 50 or 60 amperes are flowing. With the spark length above half an inch, the resistance with the same oscillatory current flowing may be taken as roughly proportional to the spark length. But in a condenser circuit the amount of electricity stored up in the condenser, and hence the amount of oscillatory current, increases with the spark length. Thus we have two conditions working against each other as regards the influence of the spark length on the spark resistance; but we can increase the amount of current flowing without increasing the spark length by increasing the size of the condenser, and the most efficient form of circuit for a given power is that in which a moderate spark length and large condensers are used. When, after the condenser is charged, the spark gap breaks down, the gap becomes filled with metallic vapor and for the time being forms a high-frequency alternating current arc. It is the presence of the metallic vapor which produces the conductivity of the spark. After the discharge ceases, however, if this metallic vapor is not removed from the gap, the insulation will evidently be poor at the time that the condenser is next being charged, hence the first condition of spark efficiency would be want- ing. It is therefore necessary to remove this vapor completely as soon as possible after the surgings of the condenser charge cease. This is done partly by cooling the electrodes of the spark gap, thus stopping the vaporization, and in some cases by blowing the vapors out of the gap. 177. In the simple gap, such as is found in sets of small power, the vapor is usually sufficiently dissipated by the natural cooling of the electrodes and by ordinary air currents. Such a gap, however, not pro- vided with an air blast, should not be enclosed. For somewhat larger powers, an air blast is ordinarily considered necessary. This carries away the metallic vapors and at the same time cools the electrodes. Such an arrangement is shown in fig. 54. Another form of gap for small powers which gives good satisfaction is the parallel gap (see fig. 55), in which two cylinders of zinc or brass MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 117 are placed parallel to each other, and the spark runs from point to point, never jumping twice consecutively in the same place. This wandering of the spark is facilitated by a slight roughening of the electrodes with a file. The explanation of this phenomenon of the running spark is probably as follows: The spark jumps from a slight projection on the electrode which in the course of the oscillations is burned away, so that at the next discharge an easier path is found from some other projecting point. 178. For high powers, a good form of spark gap is the rotating 5301- chronous gap shown in fig. 52, This consists of one or more stationary members and a rotating member made up like a wheel with projecting spokes. This, in its best form, is attached directly to the shaft of the alternator, and is so adjusted that a spoke comes opposite a stationary member at the exact moment that the maximum of potential is obtained in the condenser. This insures one discharge for each alternation of the current, the complete absence of conducting vapors, and gives a satisfac- tory insulation for each spark. The regularity of discharge from this form of gap produces a pure musical note, which is of great importance in the telephonic reception of signals. (See art. 156.) 179. Another form of rotating gap, called the non-synchronous rotat- ing gap, is shown in fig. 53. In this the wheel is rotated rapidly by an independent motor without regard to synchronism with the alternator. The face of the stationary member of the gap forms an arc of a circle long enough to a little more than cover the distance between two spokes, thus alwa3's insuring the proper sparking distance. The rotating wheel itself forms an efficient fan. 180. What is called a " quenched gap " (fig. 57) may be made up of a number of copper discs accurately turned and separated by annular rings of mica about .01 inch thick. The spark is confined to the air tight space inside the mica-rings. It has almost entirely displaced open gaps (art. 175). This type of gap, if a proper number of discs are in series, also gives one discharge for each alternation of the current and produces the same pure musical note as the synchronous gap. It is almost noiseless and has the further advantage of (probably on account of its large cooling surface) quickly stopping the oscillations of the closed circuit, so that the open circuit is left free to vibrate in its own period, and it therefore radiates waves of but one length. This fact haa an important bearing on the tuning of wireless telegraph sets and also on the coupling, which can without change of wave length be made that which will transfer energy from the closed circuit to the open circuit with the least loss (art. 85) . 118 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. The quenched gap can not be depended upon to operate without artificial cooling of the discs when any but very small powers are used. Like all other gaps, its action is improved by an air blast. 181. In the case of the rotating gap the equivalent air velocity in a case of large power was about 20,000 feet per minute. Mr. J. Martin finds a very distinct gain in radiation from an air cooled gap with air pressures up to 15 pounds per square inch, which corresponds to a velocity of 82,000 feet per minute, or about 1400 feet per second. Take a single gap operating on a 1000-meter wave on the peak of the charging E. M. F. : If the coupling between the open and closed circuits is such that the closed circuit transfers all its energy to the open circuit in five complete vibrations the first group of sparks will last 1/30,000 second. To remove the conducting vapor from the gap in that time would require a minimum air velocity across the gap of 1000 feet per second if the electrodes were .4 inch (1 cm.) in diameter. From this point of view it would seem, therefore, that any gap will act as a quenched gap if the air velocity across the gap is sufficiently great, and that the required air velocity varies directly as the diameter of the spark electrodes — inversely as the wave length, directly as the damping — and (since it is known that close coupling increases the damping) directly as the percentage of coupling. Loose coupled circuits would require a lower air velocity than close coupled ones. Fig. 56 illustrates the Marconi disc discharger, which is practically the same in principle as fig. 53 — the non-synchronous rotating gap. A special motor is required to operate the discharger. It has also the dis- advantage of being as noisy as the synchronous gap. The disc discharger, like the synchronous rotating gap, is suitable for large powers and for use with direct as well as alternating current. It is fitted with an auxiliary stationary gap for use in case- of motor breakdown. USE OF THE ARC FOR PRODUCING UNDAMPED OSCILLATIONS. 182. The arc method of producing undamped oscillations with direct current was discovered by Professor Elihu Thompson in 1892 and has been developed by many other investigators. In order to prevent the oscillations from running back to the dynamo, choke coils or very high resistances must be placed in the D. C. leads. When the shunt containing inductance and capacity is closed around the arc in a circuit like that shown in fig. 29d, 62 or 91, a part of the cur- rent fiows into the condenser, thus robbing the arc of a part of its current ; but as the D. C. potential across an arc increases as the current decreases, this decrease in current increases the potential difference, and the con- denser continues to charge. At the next instant, however, the condenser MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 119 commences to discharge, increasing the direct arc current until it is entirely discharged ; then the process repeats itself. Oscillations can be produced in this way from almost any form of arc and over a wide range of voltages, but it is found that high frequency oscillations are best produced when the direct current voltage is high (500 volts or more), and when the positive arc electrode is capable of conducting away heat rapidly. Water is used as a cooling medium instead of air, as with a spark transmitter, and to facilitate its application, the positive (copper) electrode is made hollow. This rapid cooling of the arc plays a very important part in the production of the oscillations, as it causes the arc to die down rapidly and increases the suddenness with which the current flows into the condenser. It has also been found that when the PILOT LAMP Fig. 62. — Arc Arranged for Wireless Telephony. arc is formed in an atmosphere capable of assisting in this cooling, the energy of the oscillations is vastly increased. The best gaseous conductor of heat is hydrogen, and consequently the best results are obtained in an atmosphere of hydrogen or some mixed gas or vapor containing hydrogen. Common illuminating gas gives excellent results, and recently alcohol and ether introduced into the arc chamber drop by drop and vaporized by the heat of the arc has come into use. It has been suspected that these gases and vapors may have some effect on the electrical conductivity of the arc as well as on its cooling, but this point is still unsettled. The energy of the oscillations which can be obtained from the arc is mcreased by forming it in a magnetic field the lines of force of which are at right angles to the arc length. The action of the magnetic field is twofold ; first it deflects the arc to one side, increasing its length and consequently the difference of potential between the arc electrodes, and second, it blowa out of the field the conducting ions formed in the gas, thus decreasing the arc conductivity and still further increasing the difference of potential between the electrodes (fig. 29e, art. 94). 120 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. For the successful production of oscillations a correct relation must exist between the arc current, the arc length, and the strength of the magnetic field. This relation in general can be obtained only by ei- periment. If these adjustments are not correctly made several sets of useless superposed oscillations may be produced in the condenser circuit Therefore it is necessary in working with waves produced from the arc to examine its oscillations from time to time with the wave meter, in which, if the adjustment be correct, but one sharp and powerful maxi- mum will be found. THE FEDERAL-POULSEN SYSTEM OF RADIO TELEGRAPHY. This system uses, for radio transmission, an undamped or continuous electromagnetic wave, as distinguished from the damped oscillations pro- duced by the discharge of a condenser across a spark gap. One great advantage obtained by the use of undamped waves lies in the fact that when transmitting over long distances, where the daylight absorp- tion is great and a long wave length is desirable, the daylight absorption, if the wave length be over 3000 meters, is much less with an undamped wave than with a damped wave.* The frequency of these oscillations is controlled by the electrical char- acteristics (capacity and inductance) of the shunted circuit. When this latter consists of an antenna and ground, continuous or undamped waves are radiated. As previously explained undamped oscillations are generated in a circuit containing inductance and capacity shunted about an arc. To obtain the high frequency oscillations required in practice an arc between a copper anode (+ electrode) and a carbon cathode (—electrode) taking about 500 * Austin, Bui. Bur. Stds., Vol. 7, p. 341, the formula for daylight transmission of Damped Waves is where 7g = Sending antenna current in amperes. If = Received antenna current in amperes, Til = Effective height of antenna in kilometers, transmitting, hi = Effective height of antenna in kilometers, receiving, X = Wave length in kilometers, d = Distance in kilometers, a = Daylight absorption coefficient = 15 X 10'^ Fuller, " The Effect of Wave Length on the Absorption of Undamped Waves " gives Ir = 4.20 ^^ e \i.5, where all characteristics retain the same meaning, except that /3 = 45 X 10"*. This formula is derived from actual daylight tests made between San Francisco and Honolulu over a period of six months. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 121 volts is used. The arc is enclosed in an air-tight chamber kept filled with a gas containing hydrogen. This hydrogen vapor is obtained by allowing alcohol to drip into the arc chamber. A strong magnetic field is established across the arc by two coils in series with the arc circuit. The copper electrode is water cooled. The Federal Telegraph Company has developed this method of radio transmission and has in commercial operation, day and night, stations at Honolulu, San Francisco and other points. The distance from Honolulu to San Francisco is 2100 nautical miles and the 100-kilowatt equipment of these stations is similar to that being installed by this company at Darien in the Panama Canal Zone, for the Navy Department. Other stations use 30-kilowatt sets, such as the Navy Department is placing at Boston, Mass.; Point Isabel, Texas; New Orleans, La.; Guantanamo, Cuba, and the Great Lakes, while 12-kilowatt and 5-kilowatt sets are made for ships and smaller land stations. The wiring diagram and photograph of the 30 K. W. set, figs, 62a and 62b, are typical of arc sets. THE ANTENNA. The antenna at south San Francisco is supported by three guyed wooden towers, placed in a triangle, one of 608 feet and two of 440 feet in height, and is of the flat-top type. Its capacity is 0.010 microfarad, with a natural period of 2300 meters, and an efi:ective height of approximately 425 feet. The insulation of the antenna from the towers, and also that of the guys (which are insulated every 100 feet of length) is composed either of long wooden breaks or, in the case of the 608-foot tower, of stone blocks, 10 inches x 10 inches x 10 inches in size. THE GROUND. This is a radial network of wire extending beyond the projected area of the antenna on the earth. THE HELIX, WAVE-CHANGING SWITCHES, ETC. The antenna lead is brought with suitable insulation to a switch for transferring the antenna from the sending to the receiving circuits. The sending inductance is a helix of 1-inch copper tubing, 52 inches in diameter, and, in the case of the Darien installation, 16 feet 6 inches long. The cathode terminal of the arc being grounded, the anode is con- nected through a hot-wire ammeter to a number of wave-changing switches. Any one of these may be made to engage clips connected to points on the helix, giving, according to the number of turns utilized, a great range of wave lengths. No condensers are used. 122 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Hh HHi H -^i y^ CO - 1 -i-_JI I— j»- mm, MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 123 SIGNALS AND KEYS. In this system the oscillations in the antenna are continuously pro- duced by the arc, therefore the signals are not made by completely break- ing this oscillatory circuit, but by making a small change in wave length. This change is made by short circuiting part of the transmitting in- ductance by means of a multiple contact, solenoid operated relay key of rugged construction, capable of handling the heavy currents necessary to be broken. The contacts are cooled by an air blast. The current in the key solenoid is made and broken by a " key con- troller." This consists of two copper contacts in a sound-proof chamber, operated in a magnetic field in order to reduce arcing between them. One contact is moved by a second, exterior, solenoid, and this in turn, by a small current and an ordinary Morse telegraph key. The speed obtainable with this relay key is greater than that of the fastest hand sending. If it be desired to render signals audible on ordinary " spark " receiving apparatus using crystal detectors, and the like, a " chopper " is inserted in the key-antenna circuit, this being merely a rotary means of changing the continuous oscillations into wave trains of audible frequency. THE ARC. The arc itself is maintained in a water-cooled, air-tight chamber, within a strong magnetic field and in an atmosphere of hydrogen vapor. Means are provided, either by a spring lid or a poppet valve in the chamber, for releasing any undue pressure caused by striking the arc in an explosive mixture of air and hydrocarbons. The cathode electrode is of carbon, readily replaceable, and is rotated constantly while the arc is in operation, in order to keep its erosion even and the arc steady. The distance between it and the anode is regulated either by hand or motor control. The anode is copper, water cooled ; both it and the cathode being suitably insulated from the arc chamber. The necessary atmosphere of hydrogen is supplied by introducing into the chamber ordinary illuminating gas, alcohol, ether, water, steam or other compounds containing hydrogen. Projecting through the sides of the chamber are the poles of two power- ful electromagnets, in the field of which the arc is maintained. These are connected, either in series or parallel, to the jjower supply to the arc, usually 500 volts direct current. Choke coils are inserted in the power leads to protect the latter from high frequency current. A starting resistance is provided to take care of the momentary short circuiting of the power line caused by striking the arc. This resistance is rapidly cut out when the arc is established. The same result may be accom- plished by striking the arc at a low voltage and raising the same sub- sequently to full power. 134 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 62b.— 30 K. W. Arc Set. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 125 In practice, in the case of the lOO-kilowatt arc, the operations of turn- ing on or off of cooling water, gas or alcohol ; of starting and stopping the carbon drive, blower, tikker motors, etc., are all done automatically upon starting and stopping the arc, the latter being struck and its length ad- justed by remote control also. The operating room is usually entirely separate from that containing the arc and accessories. RECEIVING. In receiving, owing to the accurate tuning obtainable with undamped waves, a very loose coupled (and therefore highly selective) circuit is used. ^J^/7/7<:^ IS II l^r/aJp/(^ Cj2r7<:^f&'\izl '^ 1^r/}:P/77^/kr- 7/}^/^&r ^C. Fig. 62c. — Elementary Poulsen Receiving Circuit. An elementary diagram of the connections for the receiving circuit is shown in fig. 62c. The variometer being used to vary the wave length when long waves are being received, and the condenser is used to adapt the antenna for the reception of short waves. 186 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. The Federal-Poulsen tikker, the device rendering audible in the tele- phone undamped oscillations received by the antenna, consists of a revolv- ing brass disc upon which lightly impinges a fine steel wire and is very sensitive, the received watts in the antenna necessary for unit audibility (or where the dots and dashes of signals can just be differentiated) being 3.2 10"^° watts. Metals other than brass and steel may be used. The ele- mentary diagram fig. 62c shows the location of tikker in the receiving circuit. WIRELESS TELEPHONE TRANSMITTERS. 183. A method of arranging the arc circuit for wireless telephony is shown in fig. 62. Here the arc creates the continuous undamped oscilla- tions which are necessary for wireless telephony. The telephone trans- mitter in the aerial modifies (by its change of resistance when spoken into) the amplitude of successive oscillations in accordance with the vibrations of the sounds of the voice. These produce similar variations ia the receiver oscillations and thus reproduce speech. It has been im- practicable up to tlie present time to obtain carbon transmitters which can successfully carry large oscillating currents and vary them so as to reli- ably reproduce speech at long distances. On this account, the develop- ment of wireless telephony has been retarded. The undamped oscillations are of too high frequency to produce sound, but their slow variation produces sound. (See art. 201.) Assume that the undamped oscillations have a frequency of 700,000 and the notes of the human voice vary through two octaves (say from 300 to 1200 vibrations per second). The vibrations of the tele- phone diaphragm, by changing the resistance of the carbon, modifies the oscillating current in the aerial (and therefore the amplitude of the electric waves generated) in accordance with the vibrations of the voice of the person speaking. The ordinary receiving circuit having a crystal or electrolytic detector serves as well for undamped oscillations as for groups of wave trains, transforming the modified oscillations into human speech in the receiving telephone. The limit of mechanical or air vibrations recognized as sound is be- tween 30,000 and 40,000 per second. Although the undamped oscilla- tions are of a much higher frequency and therefore produce no sound in themselves, modifications of the amplitude of successive waves may be of such a nature as to produce sound by slower variations in the rise and fall of the received current. The transmitting telephone may be in the arc circuit instead of the aerial as shown, or it may be inductively connected to either the open or closed circuit. There is as yet no standard practice. The telephone transmitters are specially constructed to stand the voltage and current induced in the aerial or in the closed circuit. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 127 Fig. 62d.— Radio Telephone Set. 128 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. LIMITATIONS ON WAVE LENGTHS. 184. A certain amount of inductance is necessary in the closed circuit in order to transfer energy to the open circuit, whether the circuits are direct or inductively coupled. Since condensers of any desired capacity can readily be obtained, it is easy to make the closed circuit any electrical length we desire. There is, however, a lower limit to this, depending on the material and arrangement of the condenser and leads. Other things being equal, the larger the capacit}^ the longer the connecting leads; and the shortest wave length that can be obtained for a given capacity is that found when the leads from the condenser are connected in the most direct manner to those from the closed circuit and spark gap. The standard wave length for ships and shore stations, using damped waves, was first set at 320 meters. It is now 600-1000 meters for ships. Much longer wave lengths are used for undamped wave signalling. Experience shows that aerials with short wave lengths radiate more efficiently than those with long ones, and that up to several hundred miles short waves travel over salt water with no great absorption. When transmission over land is necessary and for long distances over water we gain more by the reduced absorption of long waves than we lose by decreased radiation efficiency. 185. The open circuit, while it has concentrated inductance like the closed circuit, has distributed capacity which is comparatively small, and though any electrical length we desire can be obtained by adding induc- tance, it is found that concentrated inductance beyond that necessary to receive energy from the closed circuit lessens the radiation, and on that account it is necessary to increase the period of the open circuit by adding capacity in the shape of additional wires to the aerial. We have seen that, unless they are quite a distance apart, two parallel wires do not have twice the capacity of one, so that it is practically difficult to get very long wave lengths in the open circuit, especially on shipboard. The wave lengths that we can efficiently use in the open circuit are, therefore, limited by practical considerations. Since the energy in any discharge varies as the square of the voltage, and since any desired voltage can readily be obtained, the work that can be stored in a condenser of given capacity depends only on the dielectric strength of the condenser material. But in the case of the open circuit, when the first transfer of energy is completed, unless it is radiated nearly as fast as received, the maximum voltage in the open circuit, on account of its capacity being very much MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 129 smaller, is much greater than that in the closed circuit. And we find that very high voltages, on account of difficulty of insulation, break out in sparks at all points of the circuit, that the aerial wire glows through- out its length, and the whole apparatus generally acts like a dry linen fire hose when subjected to a high water pressure — i. e., it spurts electricity at all points in all directions. So practical considerations limit the wave lengths that can be efficiently used on board ship, and also limit the power that can be used with them. 186. Eeferring to the closed circuit, it is probable that the best results with any given sender are obtained when the work necessary to charge the condenser to the transformer voltage is equal to that supplied by the available power of one-half alternation. This gives but one wave train per alternation, and, if true, fixes at once the capacity of the closed sending circuits for any given power. OPEN CIRCUIT — (aerial, INDUCTANCE, GROUND). 187. Aerials, with which the open circuit inductances of sending sets are connected, are shown diagrammatically in figs. 63 to 71 inclusive. The main principles to be remembered in connection with aerials (or antennae, as they are sometimes called) are that the higher the aerial the more efficiently the energy will be radiated in the form of electric waves and the larger the currents induced in the vertical part of the aerial the greater the amount of energy radiated. See tables 9, 10, 11, 12, appen- dix A. The total capacity of a ship aerial is usually less than one standard jar. To hold the same amount of energy as the condenser circuit, the aerial is, therefore, while oscillating, charged to a higher maximum potential than the closed circuit. 188. The form of aerial now generally used on ships and ashore is called the fat-top or inverted L (fig. 67). The leads to the operating room are taken from one end; the other (free) end is subject to high potentials and must be well insulated. Some T aerials are in use (fig. 70). They give greater relative capacity for the same amount of wire; but T aerials sag in the center, thus decreasing their effective height and they are subject to high potentials at both ends. The other types shown are, or have been, used on shore stations, except the special receiving aerial shown in fig. 63, which is the direction aerial used on ships as part of the Bellini-Tosi direction finder (art. 217). The umbrella aerial shown in fig. 65 has been used at some large shore stations. It is probably the best form that can be supported by a single mast. 9 130 iMANUAL OF RADIO TELEGUAPIIY AND TELEPHONY. K^SPAH FIG. 63 BELLINl-TOSI FIG. 64 MARCONI GOV3 FIG. 65 UMBRELLA TO SENDlNq Ht\.\]k FIG 67 auY9 ne: y Mt stone: t massiel jnsulated wire 4- bare wjrei FIG. 66 (Now Obsolete) TO SENpjMCi HELIX TO SEMPlNCi HELIX FIG. 68 FIG. 69 (Now Obsolete) (Now Obsolete) -TO SENPlNGv HELIX FIG. 70 I TO SENDlNt^ HELIX FIG. 71 Fig. 71a. — Latest Marconi High Power Antenna. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 131 Marconi found that a long flat-top aerial, like fig. 64 and fig. 71a, sends more strongly in the direction away from the free end of the aerial and receives more strongly from the direction in which it sends best. This has proved to be of practical use on shore, where the horizontal part can be made very long as compared with the vertical part and in his transatlantic stations his transmitting aerials are installed pointing away from one another and the receiving aerials (which are outside of, or at a distance from, the transmitting aerials) are also very long (1 mile) horizontally. No satisfactory explanation of the directive action has yet been given. 189. It will be noted that the diagrams of receiving sets (figs. 81 and 82) show an aerial in the form of a loop, beyond three spark points arranged in the form of a triangle. The lower one of these points was con- nected to the sending circuit inductance so that as far as sending is concerned this aerial was the same as any other, since the high potentials used in sending easily jump the short gaps between the two sides of the loop ; but for receiving it was different — the weak currents can not jump the gap, which is known as an anchor spark gap, so that the circuit was only looped for receiving and not sending. The anchor spark gap served to cut out the sending circuit when receiving. When sending, the volume and color of the sparks in the anchor gap indicated roughly whether the sending apparatus was work- ing properly. For receiving sets not requiring a looped circuit the two sides of the loop were joined below the gap and used as a single wire. A little consideration will show that the wave length of a loop is the same as that of half the loop on open circuit. A loop is, however, a persistent oscillator. Now that hot wire ammeters are installed on all sets, anchor gaps are no longer useful and are not supplied. 190. Except where they pass near conducting objects or through decks, all parts of the aerial wire are left bare on account of the lighter weight and smaller surfaces exposed to the wind as compared with insulated wire. The size of wire generally used is made up of seven strands of No. 20 B. & S. phosphor or silicon bronze wire or monnot metal having fairly high elastic strength. Stranded wire is more flexible, and the materials given above have fairly good conductivity and much greater elasticity than copper wire. The elasticity prevents permanent elongation and sagging after being hauled taut. For those parts of the aerial which pass through decks and for parts near decks a special heavily insulated flexible stranded wire, called rat-tail wire, is used. 191. The natural wave lengths of certain aerials of the flat top type (inverted L and T aerials, figs. 67 and 70) are given in tabular form below. To the aerial is added the necessary turns on the open circuit helix to bring the natural wave length to the standard. It usually requires 132 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. a number of turns of the helix or spiral to do this. When it is desired to greatly increase the sending wave length, special loading coils are added to the open circuit. (See figs. 45, 46 and 47.) Since the closed circuit has large capacity and small self-induction a turn or more of inductance added to the closed circuit makes a large percentage addition to its self-induction and, therefore, to its wave length. But the open circuit has small capacity and relatively large self-induction, so that each additional turn does not make such a large percentage addition to its self-induction and, therefore, its increase of wave length per additional turn is much less than that of the closed circuit. (See Adjustments, Chapter VIII.) Ship. Type. No. wires. Distance apart. Length of flat top. "Vertical length of lead to operating room. Total length. Natural wave (meters). Glacier T Mayflower T Dolphin 1 Louisiana 1 Chester 7 Birmingham q Connecticut 1 Maine T 2 feet 26 inchei 2 feet 3 " Baltimore Ouantanamo . 4 feet 4 '■ 170 feet 124 " 140 " 150 " 160 " 160 " 125 " 120 " 130 " Inverted pyramid. 82 feet 132 " 136 " 129 " 97 " 90 " 137 " 120 " 132 " 200 " 252 feet 256 " 276 " 279 " 257 " 250 " 262 " 240 " 262 " 330 360 330 426 395 385 360 330 370 900 In all aerials referred to above, except those of the Maine and Baltimore, the long wave contained the greater amount of energy. In the case of these two aerials the greater amount of energy was radiated on the short wave. The law now requires that the energy in the lesser wave shall not be more than 10 per cent of that in the greater. These sets (except the Shoemaker, whose closed circuit was designed to give loose direct coupling at 425 meters, and which did not require the use of the aerial loading coil for 425 meters) had no direct provision for changing the wave length of the aerial except in the coupling coil, and, therefore, when coupled gave a wider variation from the standard wave length than the Shoemaker sets. OPEN CIRCUIT INDUCTANCE. 192. With direct coupling the open circuit inductance forms part of the same helix as the closed circuit inductance, as has already been stated. (See fig. 40.) In inductively coupled sets the open circuit helix or spiral is movable, so that the coupling can be varied by moving the entire coil while keeping the same wave length. Provision is also made for a MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 133 Fia. 72.— Spiral Inductance. Fig. 73.— Helix and Spark Gap. 134 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. srariable connection to the helix so that the wave length can be varied. (Figs. 42 and 43.) It must not be forgotten that varying the wave length of either circuit by varying the inductance of the coupling coil or coils varies the mutual induction, as well as the self-induction, and also the coupling and damping, so that the most recent sets — Fessenden (fig. 45), Tele- funken (fig. 46), Lowenstein (fig. 47) — make provision for varying the wave length at some other part of the circuit than at the coupling coil, or, as in the Lowenstein sets, for automatically moving the coils so as to maintain the same coupling when the wave length is varied. These out- side coils are called loading coils, as distinguished from the coupling Fig. 74 Wave changer. coils, by means of which energy is transferred from the closed to the open circuit (and vice versa in sets not having quenched or properly air- cooled gaps). The method of building the Telefunken variometer coils, shown in fig. 46, is illustrated further in fig. 74. This method of varying the self-induction of a circuit has the advantage of not having any dead ends as in the old inductance helices, shown in fig. 73. However, the vari- ometer shown in fig. 74 is not suitable for inductive coupling. The spiral inductance, shown in fig. 72, is suitable for both direct and inductive coupling, and also for loading coils. It is convenient to manu- facture and to mount and is coming into use rapidly. The inductance shown in fig. 72 is an early DeForest type. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY, 135 The common method of shifting wave lengths quickly is hy means of the navy standard wave changer. A simple wave changer is shown in Fig. 74a. The coils of the primary and secondary circuits are mounted in such manner in the wave-changer that by a simple movement of one handle the inductance of both couple coils and loading coils is varied properly to give any one of several wave lengths for which the apparatus is adjusted. The wave changer is the invention of Mr. Guy Hill and G. H. Clark, radio expert aids of the navy. AERIAL ACCESSORIES. 193. A lightning switch (fig. 75) is installed outside the station, or where the aerial enters, by means of which it is grounded during thunder storms. I ^ Fig. 75.— Lightning Switch. Fig. 76. — Hot Wire Ammeter. The other aerial accessory — the hot wire ammeter (fig. 76) — is in- stalled in the ground lead; its uses are particularly referred to Id Chapter VIII. 136 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. GROUNDS AND GROUND CONNECTIONS. 194. As has been previously explained, wireless telegraphy makes use of earthed electric waves, as compared with the free waves discovered by Hertz and used by Marconi in his first experiments. It was soon found by Marconi that good connection to earth or to a large conducting body is essential to good working. On board ship the end of the aerial below the open circuit inductance (called the ground had) must be welJ soldered, bolted, or clamped to some portion of the hull. A grounded vertical wire, well earthed, has a wave length not less thau four times its natural length. At its free end there is a potential loop and a current node (maximum potential — no current). At its earthed end there is a current loop and potential node (maximum current — no potential). (See fig. 18d.) The same wire free at both ends has an electrical period equal to twice its length, and, if oscillating, has high potentials at both ends. If the ground connection is not good, there is a tendency to choke the current passing in and out of the earth and thus to cause a rise of potential and consequent sparking and reflection of energy at the earth connections, making the period irregular and impair- ing the sending qualities of the station. It should be possible to grasp the ground lead where it is soldered to the ship without injury. Inability to draw a spark there is proof of good connection. 195. At shore stations it is found that the resistance of the earth be- tween two earthed conductors, a given distance apart, varies widely in different localities and even in the same locality with moisture and tem- perature. Low ground resistance at a station is usually accompanied by good radiating qualities. Where and when the soil is very dry it is neces- sary to pay much greater attention to the area of the ground connections, and where the resistance of the earth in the vicinity of the station is high the station is a poor radiator unless an artificial ground called a " counterpoise " is installed. This can consist of any large conducting area laid on the ground or wires connected between the mast guys. The natural period of the counterpoise should be the same as that of the aerial. The U. S. Naval Eadio Station at Peking has its mast and counter- poise about 50 feet above ground, on top of the wall of the Tartar City. This set operates with fair eflficiency. Generally a good ground is made by connecting the ground lead to copper plates of large area in good contact with moist earth, or to radiating lines of galvanized iron telegraph wire ending in pipes driven to moist earth, or to wire netting spread on the ground and covered with earth. At stations on tops of buildings grounds are made to the steel frames of the building and to water and gas pipes. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 137 As previously stated, Dr. Austin has shown that the high frequency resistance of a well-installed ship aerial should not exceed 2 ohms at wave lengths slightly greater than twice the natural period of the aerial, and should not increase greatly for longer wave lengths. This phenomenon does not accord with theory and is somewhat compli- cated by the fact that we have no absolute means of measuring the true resistance and the radiation resistance separately. Their resultant is measured as a single quantity. We want the apparent resistance due to radiation of energy to be as high as possible and the true resistance to be as low as possible. The in- crease of resistance referred to above is not accompanied by any increase in radiation. It is, therefore, believed to be a reaction due to earth cur- rents, in the same way that other aerials or circuits having the same, or nearly the same, period react on the aerial and increase its apparent re sistance without improving the radiation. Chapter VII. EECEIVING APPARATUS. 196. Receiving and detector circuits are illustrated in figs. 77 to 92, inclusive. In all figures, the fixed condensers shown are for the purpose of pre- venting the direct current from the battery or detector from flowing through the inductance. Variable condensers, and variable inductances are used for changing the period (wave length) of the circuits. Referring to fig. 79, the cup-shaped construction under the word Pessenden indicates a detector and the construction shown above the fig- ures 79 indicates a telephone in all diagrams. The non-inductance resist- ance, with arrow-headed connection, is used to regulate the impressed voltage at the detector terminals. It is called a potentiometer. Other symbols used have been previously described. 197. Fig. 77 shows the detector (in this case a coherer — art. 211) in shunt in the open circuit, the open circuit having a variable tuning inductance. The remainder of the figure shows the coherer-tapper, call, and the relay for the Morse recorder. Fig. 77, like figs. 78 and 79, illustrates direct-connected receiving sets. They are not now generally used. Inductively connected sets, shown in figs. 86 and 88 to 92, are preferred. Fig. 92 shows a method of connecting both loading and coupling coils, which avoids dead ends and inductive effects on parts in circuit. It also shows connections for putting a condenser in series or parallel with the aerial, in accordance with present specifications. It will be noted that in fig. 83 provision is made for tuning the closed circuit with detector directly in circuit; while in the Fessenden interfer- ence preventer illustrated in fig. 85, no provision is made for tuning the detector circuit. In figs. 80 and 86 the detector is in shunt around a closed tuned circuit. In all inductively connected receiving sets, provision is made for vary- ing the mutual induction between the open and closed circuits. This whether the closed circuit is tuned or untuned. Professor Pierce's investigations of detector circuits, like those in fig. 83 (except that the closed circuit inductance was fixed and the condenser variable), indicate that, if the resistance of the detector is not too great, very much greater selectivity, with equal loudness of signals, is obtained by tuning the detector circuit, with the detector directly in the circuit as in fig. 83. No absolute figures are at hand as to the effect of shunting the detector around a closed tuned circuit as in figs. 80 and 86. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 139 ELEMENTARY DIAGRAMS, RECEIVING, AND DETECTOR CIRCUITS. Dc Forest JL (Now Obsolete) MASSIE ^^ir 1 IK rjOSO^ ^yOUQ^ ^^ FIG. 81 (Now Obsolete) Shoemaker \/ ryOSMYJf. v^^ U =^ FIG. 82 (Now Obsolete) pWB FIG- 83 140 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. but results obtained in distance of communication show this method equal if not superior to any other, and it would seem that, if the detector has a resistance such as would prevent sharp resonance from being obtained when placed directly in the closed circuit, shunting it will assist in producing sharp resonance and together with tuning the closed circuit make a more efficient arrangement. It is now the usual practice to tune the closed circuit and shunt the detector around it. 198. Eeceiving sets, such as shown in fig. 80, were first introduced by Stone and used later by Marconi. The intermediate tuned circuit in these sets is called the weeding out circuit. Provision is made for switch- ing the detector to the weeding out circuit when very sharp tuning is unnecessary, since there is loss of range, due to loss of energy in so many transfers. The Fessenden interference preventer, shown in fig. 86, attains select- ivity in a different manner from that just described. The currents induced in the aerial, from whatever cause, have two possible paths to earth; one of these paths is tuned to the wave length it is desired to receive, while it is supposed that waves of other lengths, or static dis- charges, out of tune with either leg, will divide themselves equally be- tween the two legs and produce no effect on the untuned detector circuit. Attention is invited to figs. 81 and 82, showing the DeForest and Shoe- maker looped receiving circuits. These differ from the other circuits illustrated in that the induced currents are in the same direction on the two sides of the loop and like a double-ended sending aerial induce a maximum potential at some point in the loop whose electrical distance from the point of origin of the disturbance is the same for each side. The wave length of a looped circuit ungrounded is therefore the same as that of one-half of it ungrounded and the wave length grounded is twice the electrical length of the loop. In other words, one-half the loop can be considered as a shunt of the same period as the other half. From this point of view the DeForest detector circuit is practically the same as fig. 83 and the Shoemaker circuit can be considered as one having a detector directly in the open circuit and shunted by a variable con- denser. 199. The variable inductance of fig. 83 is hinged so that the coupling can be varied by a combined movement of separation and rotation with reference to the fixed inductance (fig. 100). In fig. 86 the coupling is varied by sliding the closed circuit inductance on a graduated bar parallel to its axis (fig. 101a). Figs. 84 and 88 represent the valve and audion receiving circuits. In fig. 84, the valve detector is shown as shunted around a tuned, closed circuit. The audion shown in fig. 88 has tuned connections like the valve detector, but has a local battery in the telephone circuit. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 141 i, TO AERtAL y.-c oaflfiOyHMiS FIG. 84 l/ALVE RECEIVER -MARCDN I - TO AERIAL VMRELESS SPECIALTY APPARATUS CO. FIG. 85 FES5EUDZN INTERFERENCE PREVENTER TO AERIAL -h^ m €^ D=ie Fig. 87. — Magnetic Detector — Marconi. 142 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 92 represents circuits of a receiving set complying with present specifications. In all such sets closed circuits are calibrated and curves drawn or a scale furnished showing wave lengths for all settings from 200 A Fia. 88. — Audion. to 4000 meters, and additional calibrated loading coils for very long waves are supplied. With very loose coupling the wave length of received signals can thus be read directly. 200. Fig. 89 represents receiving circuits for continuous oscillations; it shows the receiving telephone connected across the terminals of a fixed condenser in series with the detector. Fig. 89.— Federal Co (Poulsen) In using undamped oscillations for wireless telegraphic purposes it must be remembered that the frequency of the oscillations themselves is too high to be heard in the telephone connected with the ordinary re- ceiving circuit, and when the circuit at the sending station is closed all that would be heard is a slight click, so that there is no way of telling a MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 143 dot from a dash. This makes it necessary to place a rapidly rotating circuit breaker either in the sending or the receiving circuit for the pur- pose of creating a buzz in the telephone at the receiving station when the sending circuit is closed. This circuit breaker may be placed in the send- ing aerial, while the sending key is placed either in the aerial or shunted around a few turns of the aerial inductance, in which case it serves merely to throw the aerial in and out of tune. (Art. 164.) If no interrupter is used in the sending apparatus, no signals can be read at a receiving station unless the wave trains are there broken up so as to produce a buzz in the telephone. For this purpose the Poulsen ticker is sometimes used, whicli at tlie same time does away with the need of any special receiver. It consists essentially of a circuit breaker actuated by a small magnetic vibrator, kept in action by a dry cell. In this receiver the closed circuit is coupled very loosely to the aerial, and this circuit is inter- mittently connected to a large condenser, of the order of a microfarad, by the ticker or a slipping contact detector (fig. 89). (Art. 210.) During the time of contact the condenser becomes charged, and when the contact is broken it discharges itself through the telephone, producing a note corresponding in tone to the frequency of the ticker. 201. Figs. 90 and 91 represent a Fessenden receiving set and hetero- dyne, respectively, the loop in the aerial being for the purpose of impress- ing the oscillations in the heterodyne circuit on the aerial and, conse- quently, on the closed receiving circuit. As will be seen from an inspection of fig. 91, the heterodyne is an arc sending set, without aerial and with variable inductances in the closed circuit for changing its wave length and the frequency of the continuous oscillations it produces. The heterodyne is valuable as a means of enabling us to read signals sent with undamped oscillations, without using an interrupter at the send- ing end or a ticker at the receiving end. It is also useful as an amplifier of signals from damped oscillations. The undamped oscillations in the receiving aerial produced by the heterodyne are combined with those from the transmitting aerial. If they are of exactly the same frequency, they simply tend to amplify or neutralize each other, depending on their relative phases. If not of the same frequency, beats are produced as in music. For instance, if the trans- mitter has a frequency of 300,000 per second and the heterodyne 301,000, when one has made 300 vibrations the other will have made exactly 301 ; so that they will exactly coincide once in each j-^ second, thus pro- ducing a third frequency of 1000 per second. Both transmitter and heterodyne have a frequency much greater than can be detected by the human ear, but their combination, producing a maximum of current in the aerial 1000 times per second, can be read easily. The resultant 144 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. frequency would be the same (1000) if the frequency of the heterodyne, were 299,000. If it were 302,000, the transmitter would make 150 vibra- tions while the heterodyne made 151. They would coincide 2000 times per second and the resultant note in the receiving telephone would have that frequency. Fig. 90. — Fessenden. SOO VOI-TS DC Fig. 91. — Fessenden Heterodyne. We see, therefore, that by varying the heterodyne by means of the variable condensers shown in fig. 91, we can produce any desired note at the receiving station. It is not necessary that the heterodjTie should be in the same room or even in the immediate vicinity of the receiving aerial. Two stations, miles apart, sending continuous oscillations, may heterodyne each other, if one is sending while the other is receiving. As an accessory of the receiving apparatus, the heterodyne promises to be very valuable. The positive electrode is copper and the negative electrode, carbon ; the same as in the ordinary arc transmitter. Ammeters are supplied with it for indicating the current in both the primary and oscillating circuits. Thus far, the best results in operation are obtained by keeping the readings of the two ammeters as nearly equal MANUAL OF RADIO TELEGKAPHT AND TELEPHONY. 145 as possible. With properly adjusted carbons, the instrument should operate an hour without attention and then requires only the careful cleaning and facing up of the carbons. 202. In spark sets we want the sending aerial to be a good radiator, but not so good that it will give a whip crack discharge. We want its oscilla- tions to be persistent enough to require for their best reception a receiving aerial tuned to the period of the sender, and as a present standard we have set for the sender a damping considerably lees than .2, so that it makes fifteen complete oscillations before the oscillating current falls to .1 of its original value. We want the receiving aerial to radiate as little UOADIMG COIl-S ^^<^ Fio. 92. — Standard Receiving Set. as possible; but to so combine the energy of the fifteen waves that the highest maximum is produced in the aerial, and transferred to the closed receiving circuit. If the sending aerial is coupled so as to send out waves of two lengths, there appears to be no question that the coupling of the receiving circuits should be such that, if they acted as senders, they would send out waves of these lengths, or so loosely coupled that their natural period is that of the arriving wave containing the most energy. If, in the case of very loosely coupled circuits or those supplied with quenched spark gaps, but one wave length is being radiated, receiving circuits should also be loosely coupled or should be coupled so that the transfer of energy from the open to the closed circuit and the damping of the latter (with the detector, however connected) is at such a rate that a maximum current in the closed circuit is reached at the instant the open circuit has come 10 146 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. to rest after being set into vibration by the passing wave train and after having radiated or transferred all its induced energy. This is analogous to the statement relative to the quenching of the closed sending circuit (throwing it out of tune) when the open circuit has reached its first maxi- mum. When the closed circuit has reached its first maximum the rectified current in the case of crystal detectors or the battery current in the case of electrolytic detectors has also reached its maximum. It must be remembered that change of coupling changes mutual induc- tion and, therefore, changes the tune so that retuning is necessary after every change of coupling. NAVY RECEIVING SET TYPE A. Coupling. — This receiver differs from all earlier sets in the naval service in that there is no electromagnetic coupling between the primary and sec- ondary circuits. In the ordinary type of set, the primary and secondary coils are electro- magnetically coupled, and the coupling is "increased" by moving the coils closer to each other, and " decreased " by separating the coils. In this type, the inductances in the primary and secondary systems are placed at right angles to each other, so that there is no mutual induction between them, and the coupling is obtained by means of small variable condensers connecting the two systems. Here, to increase the coupling, a greater value of the coupling condenser is used, and vice versa. The inductances are not moved at all, but are permanently located. The " coupling condenser " acts as a sort of " energy controller " or " valve " between the two systems (primary and secondary) . The greater the value of this condenser, the more energy is transferred from the antenna circuit to the detector circuit, and conversely. In this respect the analogy with the electromagnetic coupling method holds, but in other ways there seems to be a difference. It is interesting to note, from the theoretical standpoint, that the same effect can be produced if an inductance, or even a resistance, be used in place of the " coupling condenser," but the efficiency is very much less than with the condenser. The chief advantages of the " static " or " condenser " coupling over the electromagnetic method are : (a) Compactness of Set. — It is not necessary to separate the induc- tances to decrease coupling. For very loose coupling, with the old type of set, it was necessary to move the secondary coil quite a distance from the primary. Here, the coils are fixed in position. (b) Ease of Operation. — With electromagnetically coupled sets, the mechanical movement of the secondary coil always proved rather bother- some, especially when quick changes in coupling were necessary. This is obviated by the extremely easy operation of the coupling condenser. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 147 (c) Increased Efficiency of Inductances. — In the old type, it was neces- sary to wind the inductances in such a form that there would be sufficient coupling under any circumstances. This resulted in the cylindrical form being of necessity adopted, A much more efficient form of winding is adopted in the present sets, the " banked " winding for some of the coils, and the " square section " method for others. (d) Increased Over-All Efficiency of Receiver. — The electrostatic coup- ling gives equal efficiency to electromagnetic for short wave lengths, and very much greater for long wave lengths. (e) By a simple switching device, a " stand-by " or " pick-up " circuit greatly superior to any heretofore developed can be used. This pick-up circuit enables receiving to be done over a comparatively wide range of wave lengths. Hence the sharpness of tuning is sufficient for ordinary " pick-up " purposes without the usual defect of picking up every signal and disturbance in the vicinity. This receiver also contains a novel indicating mechanism, whereby the operator can set the receiver in advance to be in tune with any desired wave length. This is very important for battle radio work. The increased use of the transmitter wave changer renders some device necessary for the receiver, and this is the first apparatus to contain this feature. The method commonly in use for setting the receiver in advance to a given wave length consists in having the operator refer to a table for the proper value of primary coil (tens and units), primary loading coil, primary condenser, secondary coil, secondary condenser, and coupling. In this form, the primary can be adjusted by one movement, the secondary by another, and the coupling by a third. Later designs will probably reduce these movements to two, the primary and secondary being simultaneously adjusted as in the transmitter wave changer. In this connection it should be noted that in the transmitter wave changer all three variables, i. e., primary, secondary, and coupling, are varied simultaneously. In this case the damping is fixed, being the damp- ing of the transmitter set itself. In the receiver, however, the damping is determined in part by the distant transmitter and can be widely different when receiving from different stations, even on the same wave length. Hence it is not practicable to make all three adjustments with the receiver by one movement. The coupling must always be independent, to enable adjustment to suit the local conditions. Primary Circuit (see fig. 92a). — This consists of two inductances vari- able by small steps, a loading inductance and series condenser. The larger of the variable inductances (45-46) has a total value of 1.2 M. H., divided into 40 steps of equal inductance per step. (Note that the divisions are not of equal numbers of turns, but of equal values of inductance. This insures constant overlapping between sections.) 148 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 92a. — Navy Receiving Set. Type A. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 149 The smaller of the variable inductances (51-52) has a total value of 0.03 M. H., divided into 20 equal steps. Both coils are wound on the same core. The smaller is single layer, and the larger is wound banked. The loading coil (39-42) has a total value of 6 M. H., in 5 equal steps. All coils are wound with Litzendraht. The series condenser (7-8-9-10) may have three values, 0.0009, and 0.0015, also 0.0003, respectively. Secondary Circuit. — The secondary coil (16-24) is wound in the same manner as the larger of the variable inductances in the primary and has a total of 3 M. H. There are six steps, all but the first and second being equal and of value 0.6 M. H. The secondary variable air condenser (25-26) is of the balanced type and has a total capacity of 0.0025 Mfd., with a total value of .007 Mfd. 'Potentiometer. — The potentiometer (61-75) is wound with No. 38 German silver wire with a total resistance of 400 ohms. The various steps are 25-25-25-25-25-25-35-35-40-40-50-50 ohms, respectively. Buzzer. — The buzzer (67-68) for testing the detector is of the Ericsson type, and is mounted with thick felt on the top of the panel. The buzzer control key (71-72) enables the buzzer to be left in con- tinuoue operation, giving free use of both hands for other adjustments, or operated intermittently by hand as desired. A key (64-65) controls the potentiometer battery so that it can be cut off when the set is not in use. Separate batteries are used for the buzzer and potentiometer in order to obviate false signals in the telephones. The buzzer acts inductively on the antenna circuit, as shown in the schematic diagram, so that it not only can be used for tests for sensitiveness of detector, but also can be used to indicate approximately whether or not the antenna and detector circuits are in resonance with one another. No provision is made for detector or telephones other than to provide binding posts for these to be connected. The binding posts are so con- structed that two sets of telephones can be connected in parallel if desired. Circuits of Receiver. — Fig. 92b shows the schematic representation of the two-circuit electrostatically coupled receiver. The primary circuit, Cj, L^, is tuned to the incoming wave length. Similarly, the secondary coil L2, and condenser Co, are resonant to the same wave length. The energy is transferred from the one circuit to the other by means of the " coupling condensers " 00^ and CC,. These con- densers are in no sense of the word tuning condensers, and do not vary the adjustments of either primary or secondary. They are used for no other purpose than their name implies. It is possible to use only one condenser, CCj, the connection from ground to the secondary system being by a direct wire (CC, short circuited) . It is found, however, that this method is not as efficient as that shown in the figure. 150 MANUAL OF RADIO TELKGRAPHY AND TELEPHONY. Both foiulcnsors, CCj and CCo, are on the same shaft, and hence are simultaneously adjusted. As shown, the usual detector circuit is used, the telephones being placed around the stopping condenser. By a simple switching arrangement, the detector system can be directly attached to the primary circuit. This is used for a " stand-by " or " pick- up " circuit. Only one circuit has to be tuned in this case. Whereas, with all other pick-up circuits heretofore used, the broadness of tuning was too great to enable the circuit to be of any great use (as so much interference was received that it was impossible to pick out the signal desired), the circuit shown here is quite selective, and hence is not disturbed to so great an extent. CCi BC S^ CCo Fig. 92b. — Receiver Circuit. Fig. 92c. — Pick-up Circuit. An ideal pick-up circuit, it is true, is one that picks up any signal, regardless of wave length, but this would work only when the number of stations sending within range was very limited, and atmospheric distur- bances were at a minimum, on the other hand, a pick-up circuit that is strongly selective defeats the very purpose that its name implies. The circuit shown is a mean between those two extremes, and is rather broadly hined. It is very easy, with this circuit, to tell if any one is sending within the range of a wave length for which the receiver was designed, as a con- tinuous variation of the one-tuned circuit concerned can be made over the entire range in a very short time. Under many conditions, this " pick-up " circuit can be used for regular work, shifting to the two-circuit system when interference develops. In this form of circuit, the coupling condenser is again :ised in its func- tion of controlling the energy transferred to the detector. In general, this MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 151 condenser sliould be set at or near its maximum when " picking-up " is being done. The coupling would prevent the rectified currents from the detector from passing through the inductance L^ and thence through the telephones if the telephones were placed around the stopping condenser, as they are in the two-circuit system of sheet 13. This is obviated by placing the telephones around the detector in the pick-up circuit. This results in a slight decrease of efficiency. General Description. — Fig. 92e shows the front of box containing the set. Referring to the numbers on fig. 92e the function of the various parts is as follows : Knob A-7 controls the series condensers 8, 9 and 10 and load- ing coils 36. When on the first point the smallest series condenser is in antenna circuit, on second point the next condenser, etc. ; on the fourth point neither condensers nor loading coils are connected ; on the fifth point the smallest loading coil is in antenna circuit and so on to ninth point where loading coil of largest inductance is put in the circuit. This knob also controls through a system of gears and levers, pointer A-4 so that as A-7 goes from first to ninth point the pointer A-4 moves from outer circle on dial A-5 into inner circle. Dial A-5 is graduated in wave length similarly to dial A-26 using con- ventional lettering for standard wave length. Knob A-6 controls large inductance 45 to 46 and dial A-5 rotates with it. Knob A-8 controls small inductance 51-52 and is connected to pointer A-4 in such a way that it can swing the pointer through a small arc. In this way the dial A-5 after once being graduated for the particular antenna in use will indicate the wave length of antenna circuit directly. The same wave length will appear on several circles of dial A-5. In general for sharpest tuning the inner circle is used and for loudest signal the outer circle on which the desired wave length can be found. Knob A-9 controls the two coupling condensers CCi at 13-14 and CCj at 32-33. Knob A-10 controls stopping condenser BC at 54-55. Knob A-20 is switch at 60, 76. In position marked Sec, the switch shown on fig. 92a gives connection as shown in fig. 92b. In position marked Pri. the connections are as shown in fig. 92c for pick-up circuit. Knob A-27 controls secondary condenser C, at 25-26 and A-28 controls secondary inductance L, at 16. Dial A-26 indicates wave length of secondary. Knob A-21 controls potentiometer 61-75. A-11 and 13 are connections to antenna and ground with safety spark gap between them. A-16 and 17 are connections for detector battery. A-18 and 19 are switches in detector and buzzer circuits. A-22 is the buzzer. A-32 and 33 are telephone and AA-34 and 35 are detector connections. 152 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. A-1-3 and A-29-31 provide places in which extra inductances or capacities may be connected in primary or secondary circuits respectively. To use type A receiver for receiving undamped waves with the oscillating Audion, the telephone binding posts of the receiver are short circuited and the crystal detector removed. The detector binding posts are then con- nected to the Eo and Eg terminals of the Audion box (fig. 92d) . This box contains the grid condenser GC, the bridging condenser C3 and the battery of dry cells B, of 36 volts, and has terminals for attaching the telephones BAT. "B" iRc. ^ft£_ A-'' RECEIVING SET ^^im% 'LA u- I.I BAT. "A" Fig. 92d. — Diagram of Connections Type A Receiver, for Receiving Undamped Waves with the Oscillating Audion. and the wires from the plate (red), and grid (green), of the bulb. A storage battery of six volts for heating the filament in the bulb is then con- nected to the terminals A-\-A, taking care that the filament rheostat is placed in the position " in " before the current is turned on at the " on " and " off " switch. The rheostat is then adjusted so that the filament, if of the Hudson type, burns with about the brightness of the filament of a carbon incandescent lamp. The brightness should not exceed a dull red if the filament is coated with oxide. The grid condenser and the stopping condenser are set about one-fourth in, and the tuning condenser C2 set at zero, then if the Audion is to be used in ultra- Audion connection the Audion switch is set " arc " and the B battery potentiometer adjusted until oscillations are produced. This con- MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 153 y Receiving Set, Type A. 154 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. ditioii may be tested by touching with the moistened finger a bare portion of the wire connecting the receiver with Ha. If oscillations are present, touching the wire will produce a clicking or rustling noise in the tele- phones. The secondary condenser is next increased to give the required wave length and the Audion again tested for oscillations by touching the wire at Ea. If the set is provided with tickler or back-coupling coils the Audion switch is placed on " spark." The coil T2 of from 2 to 6 M. H, is con- nected to the Ijinding posts marked " tickler/' and T^, of 0.5 to 3 M. H. connected to the outside secondary loading coil terminals, and the coupling T^Tg adjusted. If oscillations are not produced the terminals of T^ should be reversed. The back-coupling connection usually gives more stable oscil- lations than the ultra-Audion connection. Considerable amplification of spark signals without the loss of musical note may be obtained by closing the back coupling to a point not quite close enough to produce oscillations. VARIABLE CONDENSERS. In fig. 92f is shown the type of variable condenser used in radio receiving circuits and wave meters. It consists essentially of two sets of semicircular plates one fixed and the other capable of being rotated so that its plates may Fig. 92f. — Variable Condenser. occupy positions between the plates of the fixed set. The capacity is varied by the relative position of the semicircular plates. The condenser shown is manufactured by the De Forest Eadio Telephone and Telegraph Co. for general radio work. MANUAL OF RADIO TELEGRAPHY AND TELEPTTONY. 155 CONDENSERS IN RECEIVING CIRCUITS. 203. A variable receiving condenser usually consists of semi-circular metal plates separated by air dielectric, alternate plates being fixed. The other plates are movable on an axis, by turning which, any desired amount of the movable plates can be included between the fixed plates. The axis carries a pointer which moves over a scale graduated in degrees or directly in microfarads. If used in connection with a fixed inductance, the scale, like a wave meter, which in this case it becomes, may be grad- uated directly in wave lengths. Some of the Stone receiving sets had sliding glass plate condensers, and the Pierce sets, step-by-step variable condensers in the receiving circuits, but the revolving plate type described above is practically a standard and is illustrated in fig. 109. Variable condensers now supplied have the limits of their capacity marked on the name plate in microfarads. Fixed condensers, in receiving circuits, are often called stopping con- densers (art. 196). They may be of any compact type, and (except in the case of a fixed condenser for use with the ticker detector, which should have a comparatively large capacity), the capacity may be quite small. INDUCTANCES IN RECEIVING CIRCUITS. 204. Variable inductances include the step-by-step, roller, and vari- ometer types. The first is made up of plug or dial steps, giving a limited number of changes, one section of a coil being varied at a time, or it may be a cylindrical coil of insulated wire wound on hard rubber, glass, or por- celain, one point in each turn being bare and co^nnections being made by a slider giving as many adjustments as there are turns of wire in the coil. In the DeForest pancaTce tuners the coil was a flat spiral of insulated wire on glass, one point in each turn being bared so as to form an arc of a circle, the end of an arm pivoted at the center of this circle making contact at any desired point. Shoemaker sets (fig. 82) have a single roller inductance, the bare wire being wound in a spiral groove on an ebonite cylinder. A sliding contact, on a rod parallel to the cylinder, works in the groove and is pressed against the wire by a spring. By revolving the cylinder an infinite num ber of adjustments can be obtained. Earlier Fessenden sets (figs. 79 and 82), have double roller inductances, by turning which, the wire can be reeled from one roller to another as desired. On one roller the turns are insulated from each other and on the other they are short circuited so that any desired length can be re- tained in the circuit. None of the above types of variable inductances can be readily mounted so as to vary the mutual induction between them by any definite amount. They are suitable for loading but not suitable for loose coupling. 156 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. For this reason the preferred types of receiving circuits are made up of fixed inductances (or those varied by plug or dial steps), mounted so that they can either be pulled apart or one coil revolved so as to change its plane and hence the mutual induction with reference to the other or others. The variometer type, mounted like variable condensers, are now being manufactured. Their self-induction can be varied quickly and con- veniently and close adjustment of period (tuning) made with them or with the variable condensers, but the entire coil is always in circuit. Each section of an inductance not in circuit should be opened at both ends, i. e., entirely disconnected, and if its natural period is large it should be mounted so that the inductive effect of active parts on it is a minimum. This applies especially to loading coils for long wave lengths. (Fig. 92.) All inductances are wound on hard rubber, porcelain or glass and so as to have a minimum high frequency resistance. The decrement of the entire circuit must not exceed .3. From the formula for damping d — it can readily be seen that a very pronounced natural period — a stiff circuit — can not be obtained unless the self-induction is large compared with the total resistance (in- cluding the radiation resistance) . DETECTORS. 205. There are two types of detectors now in general use, viz., the Audion, and the Crystal or rectifying detector. The electrolytic is used only as a standard for comparison. Coherers and microphones are prac- tically obsolete, and comparatively few of the magnetic detectors have been installed, but the use of audions and ultra-audions is increasing. Unlike coherer detectors, all types of crystal, magnetic, electrolytic and audion detectors are self-restoring. Generally speaking all should be put on open circuit while sending, to preserve them from injury due to induced potentials and currents. The ordinary detector serves as well for receiving the continuous modified oscillations of wireless telephony as for the groups of oscillations in ordinary wireless telegraphy. THE ELECTROLYTIC DETECTOR. ^ 206. It consists of a fine platinum wire just touching an electrolyte made either of a 20^ solution of nitric or sulphuric acid or an alkali. Of these the nitric acid solution is preferred. The other electrode is also of platinum. The containing cup (fig. 79) is made quite small so that the cohesive power of the electrolyte will prevent splashing in a sea way. The electrolytic detector must have the fine wire terminal connected to the positive pole of the local battery (fig. 79), otherwise the device is not operative. MANUAL OF RADIO TELEGRAPHY AND TKLEPHONY. 157 Dr. Austin states that the higher the frequency, the finer the wire should be, and that the depth of immersion does not matter if the detector is not directly in series in the closed circuit. When a current flows through the electrolyte, the latter is decomposed (the action being called electrolysis) liberating oxygen at the anode and hydrogen at the cathode. The accumulation of these non-conducting gases on the electrodes interferes with the passage of the current, which soon ceases to flow and the cell is then said to be polarized. The fine wire anode is then insulated by the oxygen, which forms the dielectric of a small condenser, of which one conducting surface is the electrolyte and the other the wire point. The critical potential of the detector is just below that necessary to break down this insulating layer of oxygen and is determined by increas- ing the potential at the detector terminals by means of the potenti- ometer until a bubbling or hissing sound is heard in the receiving telephone ; then resistance is cut in until this sound just ceases. When electric oscillations are impressed on this condenser, the polariza- tion layer breaks down and permits a pulse of direct current from the battery to pass through the cell and telephone. As soon as the oscilla- tions cease the polarization is restored. Except when they are very strong, the loudness of the sounds produced in the telephone is an exact measure of the energy of the oscillations passing through the cell. This constancy of action of the electrolytic cell is utilized as a means of comparing the sensitiveness of detectors, the standard being the sen- sitiveness of an unjacketed platinum wire electrode .0002 in diameter in a solution of 20% nitric acid.* Glass jacketed electrodes formed by sealing the wire in glass (the two having the same coefficient of heat expansion) have been used, but are less reliable and in general less sensitive and are no longer supplied. Some of these glass points were made hook shaped, the hook pointing upward to facilitate depolarizing but no increase in sensitiveness waa noted on this account. With the Shoemaker receiving sets was furnished what was called a primary cell detector. The electrolyte used was a 20^ solution of sul- phuric acid and the other electrode was a zinc rod amalgamated with mercury, which in the acid solution gave a difference of about .7 volt between zinc and platinum. No local battery was required (fig. 82). At times this detector compared favorably with the one just described, but was in general more irregular and less sensitive in its action. ♦ Dr. Austin has invented a " detector tester " which affords a means of direct comparison between detectors operating under the same conditions. 158 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. In all electrolytic detectors very strong signals or static discharges produce actual sparking or an explosive action in the electrolyte, which destroys the platinum point and an operator must be constantly on the lookout to protect his point from burning out. The best results in electrolytic detectors have been obtained with a distance between electrodes of approximately \ inch. RECTIFYING DETECTORS. 207. There are certain substances which when brought together in not too close contact, have the property of producing a direct current when an alternating current or electrical oscillations are sent through them. The cause of this action is not yet known. Among these substances are car- bon in contact with steel, tellurium with aluminum or galena, silicon with any of the ordinary metals, and certain crystals. The first of the crystal detectors to be supplied was General Dun- woody's carborundum crystal detector. Since rectifying detectors permit the passage of current in but one direction, they produce pulses of direct current. These pulses, if strong enough, can be heard in a telephone so that local batteries are not re- quired, although a slight increase of sensitiveness is noted in some de- tectors with an E. M. F. across the terminals of the detector of about 0.2 volt. Rectifying detectors are connected in receiving circuits in the same manner as the electrolytic. Their sensitiveness for general use is practically equal to the electro- lytic and their simplicity makes them the more suitable. They are in general less sensitive to injury from static discharges, strong signals, or induced currents from sending, than the electrolytic, but, like coherers, different crystals of the same material vary widely in sensitiveness and sensitive spots in any crystal have to be found by trial and when found are not constant. They are thus not as capable of quick readjustment as the electrolytic, but their other advantages are such as to be con- clusive as regards their use. The carborundum detector when first introduced was simply held between two points or wrapped with copper wire for one connection, with a needle, knife edge, or more blunt piece of metal for the other. It was later found that embedding a large part of the crystal in a conductor such as solder or a mercury paste, and thus limiting the rectification to one contact only, produced much better results and carborundum crystals have been found equal in sensitiveness to other crystals now generally utilized. Pickard's silicon detector followed the carborundum and is still in use but it has been largely superseded by the Perikon & Pyron, supplied MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 159 with the receiving set illustrated in fig. 86. The Perikon detector con- sists of two crystals, chalcopyrites and zincite. A number of zincite crystals are held in a conducting disc, a crystal of chalcopyrites is mounted so that it can be brought into contact with any part of any of the zincite crystals at will, and the pressure between the two regulated. In the adjustment for maximum intensity of signals, the exact degree of pressure and the most favorable points of contact are of importance. These can only be ascertained by trial and test with the testing buzzer. The sensitiveness of the Perikon may be approximately doubled by connecting a battery across its terminals so as to give approximately 0.2 volt. The positive pole must be connected to the single crystal. The Pyron consists of a crystal of iron pyrites in contact with a metal point like the silicon. This is very satisfactory for strong signals and constant in its action. The iron pyrites is more sensitive when the pressure of the metal point is adjustable. The area of contact is also a determining factor of sensitiveness; comparatively fine points will dis- cover sensitive places on irregular crystals, which blunt points will not. The Perikon is more sensitive and must be protected against strong signals. The zincite is the crystal injured by strong signals. It should not be subjected to heavy pressures or grinding from the chalcopyrite. When deadened the zincite crystals can be made operative by scrubbing them with a bristle brush wet with carbon bisulphide, then with soap and water and then rinsing with fresh water and drying. In damp weather or in tropical climates this detector is improved by spreading a drop of paraffin oil over the surface of the crystals. This comment applies to the silicon also. Galena, cerusite (a form of galena) and iron pyrites are all giving satis- factory use by mounting so as to have contact all over one surface and a very fine, flexible wire point, just touching the other surface. I VACUUM TUBE DETECTORS. 208. The two forms which have been used are kno\vn as the valve and the audion, illustrated in figs. 84 and 88. The valve was discovered by Fleming, and is sometimes called the Fleming valve. It is a rectifier, permitting the passage of current in one direction only. It consists of a special incandescent lamp (see fig. 8-i), operated by a 12-volt storage bat- tery and having a small sheet or cylinder oE metal held in the bulb near the filament. Lamp filaments when glowing emit negative electricity, which carries away part of the filament and causes the darkening of the bulb seen on old carbon lamps. The vacuum thus becomes a conductor in one direction only. It is not found to be a very sensitive one. 160 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. The audion (fig. 88) has a metal grid interposed between the metaJ plate and the lamp filament. In the valve, the metal plate is connected to the receiving circuit, but in the audion, the grid is connected to the receiving circuit, while the plate is connected to the telephone. In the valve, the variations of current in the receiving circuit produce differences of potential between the filament and the plate. In the audion, these differences of potential are between the filament and the grid ; or, as it is perhaps better to say, the grid is charged by the received oscillating cur- rents. In addition to this difference, the audion has a local battery with its positive pole connected to the metal plate, and its negative pole to the lamp filament. This battery, as well as the battery supplying the lamp, has a variable voltage. The battery voltage and lamp voltage must both be adjusted to secure the greatest sensitiveness of this detector ; but Fig. 88a. — Amplifying Audion. this adjustment is permanent for any given conditions. The charge on the grid, produced by the incoming signals, interferes with the flow of negative electricity between the filament and the plate. This flow of negative electricity, when the heat from the filament and the local battery voltage are properly adjusted, produces a current through t]ie audion of the order of a milliampere. This current flows through the receiving telephone and variations in it, produced by the varying charge on the grid, are what make the signals heard in the telephone. The audion has the further advantage over the valve, in that the tele- phone can be replaced by the primary of a transformer, the secondary of which is connected to another audion, with the result of amplifying the signals produced. (See fig. 88a.) Audion bulbs have tantalum lamp filaments, and plate and grid are usually double, one on each side of the filament. The plate and grid are of nickel. The audion can also be used as a heterodyne (art. 201). When connected for receiving continuous oscillations it is called " the ultra-audion." MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 161 We see, therefore, that the audion is suitable for use with arc, as well as spark sets, and also as an amplifier; thus combining within itself the qualities of all other detectors. THE OSCILLATING AUDION. If a low resistance circuit, consisting of a coil and condenser, be con- nected to an audion, a lead from one side of the condenser connecting through a stopping condenser to the grid, and a lead from the other end of the condenser to the plate, it is found that the audion produces oscilla- tions in the circuit, these oscillations being determined as to frequency by Fig. 92g. — DeForest Audion-Detector. the inductance and capacity of the associated circuit. This form of audion connection has been termed by Dr. DeForest the " TJltraudion." It was later found that oscillations could similarly be developed with the condenser connected to the grid (through stopping condenser) and the fila- ment respectively, by inserting an inductance in the B battery circuit lead- ing to the plate, and coupling this coil with the correct polarity to the in- ductance of the oscillating circuit. This coil has been variously termed the " back-coupling coil," the " feed-back coil," and " tickler coil." This form of oscillating audion is now generally used for receiving, using the method of beats. The secondary circuit of the radio receiver is connected as above to the ultraudion, and oscillations are set up. The frequency of these oscillations is made slightly different from the frequency of the incoming signals, and the two slightly differing frequencies combine 11 162 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. to form a third which is within the range of audibility. If the frequencies differ by 1000, for instance, the resultant frequency gives a note in the receiving telephone the same as from a 500-cycle spark set. This note is of course under the operator's control, as by varying the secondary condenser of the receiver the frequency of the oscillations generated by the ultraudion may be varied, and hence the difference in the two frequencies may be made greater or less than 1000. Larger size tubes are now in general use as transmitting devices, the general principle of generation of oscillations being as explained above. The transmitting units are sometimes known as " Oscillions." MAGNETIC DETECTORS. 209. The operation of magnetic detectors depends on the fact that when iron is being magnetized its magnetization is somewhat delayed in time behind the impressed magnetizing force, and when in this condition the iron is very sensitive to any change in the magnetizing force, a very small increase of which will produce a momentarily large increase in the strength of the magnetic field. Many patents have been issued for various forms of magnetic detectors, the best known and the most largely used of which is Marconi's, patented in England in 1902. It is not injured by strong sending, but is not as efficient as the crystal detector, the electrolytic nor the audion. In its present form it consists of a flexible band of silk-covered iron wires, moved by clockwork around two pulleys which support it. A glass tube, through which the band passes, has a primary winding of insulated wire in series with the aerial and a secondary winding forming a closed circuit through a telephone. Close to the secondary windings are placed similar poles of two horse-shoe magnets, which magnetize the iron band slowly moving under them. Electric' oscillations in the primary winding, produced by passing electric waves, produce momentary changes in the magnetization of the iron band under the magnets, and these changes induce oscillating currents in the secondary winding which produce sounds in the telephone. An elementary diagram of this magnetic detector is shown in fig. 87. It requires no local battery, and, not being subject to burn-outs except from very large currents, it is a very convenient instrument, but is not as sensitive as those previously described, especially for short wave lengths. There are other methods of connecting this detector than that shown in fig. 87, but since comparatively few magnetic detectors are in use they are not shown here. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 163 SLIPPING CONTACT DETECTOR. 210. In its present form this consists of a small bundle of fine wires, or a single wire resting lightly on the rim of a wheel of conducting material, which is revolved at high speed by a small motor. It is not, in general, as satisfactory as crystal detectors, but is suitable for receiving undamped oscillations (fig. 89). Its note is too low to permit good results in receiving, when there is any static. COHERERS AND LODGE-MUIRHEAD DETECTOR. 211. Coherers being practically obsolete are i»ot described. They are illustrated in fig. 93. Of the many other kinds of detectors that have been used, the Lodge- Muirhead, which would work either with a telephone or recorder, was the most sensitive and reliable. SLABY ARCO COHERER FiQ. 93. r~~i |_ . -!^-~V c c c 4 ) ) ) s LODGE-MUIRHEAD COHERER Fig. 93a. It is illustrated in fig, 93a, and consists of a polished steel disc rotated by clockwork, its edge just touching the edge of a globule of mercury covered by a film of oil. A pad which rubs against the disc keeps it clean and bright. This coherer may be direct or inductively connected in or to the aerial. Its conductivity changes sufficiently to relay a current for working a siphon recorder so that it is suitable for use in connection with determining longitudes by wireless telegraphy. It is also self-restoring and can therefore be used with a telephone. 164 MANUA]> OF IJAUIO TELKGRAPHY AND TELEPHONY. TESTING BUZZERS. 212. A testing buzzer with its battery of one cell, its condenser and circuit, is a miniature sending set and an important auxiliary of every receiving set. The oscillations set up in its circuit induce currents in the receiving circuits, which serve by their effect to determine the sensi- FiG. 94. — Wireless Telegraph Test Buzzer. For Ships. Hi y. Pl^TlNUM CONTACra, [ BUZZER Fig. 95. tiveness and readiness for operation of the detector. A testing buzzer outfit furnished with the Telefunken sets is shown in fig. 94, the con- nections of that supplied with the l-P-76 receiving sets in fig. 95. Tuned buzzer circuits are useful in measurements. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 165 RECEIVING TELEPHONES. 213. The low resistance telephones in ordinary use are not suitable for wireless work on account of the high resistance of the detectors, which may be several thousand ohms. Specially made telephones are required to produce the best effect. The magnet wire has very thin silk or enamel insulation. A length of wire whose resistance is approximately equal to that of the detector can be efficiently used. This is from 1000 to 2500 ohms in each of the double head telephones supplied. For low frequencies, telephones with adjustable diaphragms or magnets are found to be about ten times as sensitive as the ordinary type with a fixed distance between diaphragms and magnets. This advantage decreases as the frequency approaches the present standard of 500 cycles (1000 sets of sparks per second), but is still sufficient to warrant the retention of the adjustable diaphragm type. An improvement recently made, is the substitution of a single magnet, located under the center of the diaphragm, for the pair of magnets in the usual construction. Also, what are known as " reed " telephones are an improvement on the usual double magnet type. At stations where it is necessary to listen on two wave lengths at the same time, one half of the head set may be connected to one receiving set and the other half to the other receiving set. Attempts are being made to apply the principles of " tune shifters " to receiving sets also, so that an operator, by turning one wheel, can shift through his entire range, keeping both circuits in tune and maintaining proper coupling through- out; this wheel to be graduated in wave lengths. (Art. 170.) In some Marconi sets low resistance receiving telephones are used, connected through a step-down transformer. (See fig. 84.) Batteries and potentiometers are used with receiving telephones, theii connections being as shown in figs. 77 to 89. In order to produce sound, intermittent work must be done on a tele- phone diaphragm at a certain minimum rate. (See arts. 154 and 155.) In other words we must apply a certain power to it — power being rate oi doing work. The frequency must be within the limits of audibility. (Art. 183.) It appears that with crystal detectors we obtain this power directly from the aerial, while with the electrolytic, and audion the power from the aerial only works the detector as a relay — the power used in making sound in the telephone coming from the local battery. Difficulties surrounding accurate measurement of the very minute qu^tities involved make the above statement subject to modification. We do not yet know exactly how a detector acta. 166 MANUAL OF. RADIO TELEGRAPHY AND TELEPHONY. RELAYS OR AMPLIPHONES. 214. The amplifying qualities of the " audion " and " heterodyne " have already been referred to; the general term for microphonic and other detectors used for the purpose is " ampliphone/' If they prove to be constant and reliable they will be supplied for general use, (a) to enable ordinary messages to be read without the use of a head telephone, (b) as a call, (c) to increase the absolute difference between signals of different strengths thus enabling the message desired to be read through inter- ference or static, (d) to step-up signals so weak that they could not other- wise be read and thus increase the range of communication, (e) as a resonance device responding within limits to a single spark frequency, thus cutting out interference, (f) for separating signals of different wave train frequencies or different wave lengths so that several messages of different frequencies can be received at the same time on the same aerial, (g) to automatically record incoming signals. Coherer detectors change their resistance sufficiently to work a relay which actuates a call, tapper and recording apparatus. Generally, the induced currents rectified by crystal and valve detectors are too weak to produce visible material movement unless a " string " galvanometer is used with them. The same is true of the direct currents produced by the momentary depolarization of electrolytic detectors. Rectified currents will produce sufficient movement of the diaphragm of a receiving telephone to alter its pressure on a microphonic contact, this alteration being enough to change the conductivity, and thus increase or decrease the current in a circuit containing the contact, a battery and another telephone. This change in current moves the diaphragm of \he second telephone and its movements can either be read directly as sound or made to change the current in another circuit by change of pressure on another microphonic contact. One or more of these microphonic relays produces sufficient action in a loud speaking telephone to be heard in the operating room. When used as a resonance relay, the relay diaphragms are mounted so as to have a pronounced mechanical period of vibration and act as wave filters or weeding out circuits, responding most efficiently only to wave trains of a frequency the same as their own. The sound produced by the last one in circuit (the loud speaking telephone) may be intensified by attaching to it an air pipe whose note is the same as that of the diaphragm in vibration. The microphonic ampliphones, just described, have not proved as reliable as the audion ampliphone. RECORDING APPARATUS. 215. Recording apparatus went out of use with coherers. It is possi- ble that the ampliphones referred to in the preceding article will again permit the use of both recording and calling apparatus. Both tend MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 167 directly to economy in the operation of wireless stations, by reducing the number of operators to a minimum. It has already been found possible to use a galvanometer in connection with a photographic film for recording transatlantic messages. DIRECTION FINDERS. 216. The first experimental installations of direction finders have been withdrawn, it not being found practicable to operate them. The principle on which they operated was that two vertical wires parallel to the plane of movement of an electric wave, if half a wave length apart, would have electric currents of opposite phase induced in them, which could be made to double the receiving effect as compared with a single wire, while if at right angles to the plane of movement of the wave, the induced currents would be in the same direction and could be made to neutralize each other. If the plane of this direction finder pointed towards the sending station, the strength of the received signals would be a maximum. If at right angles to the sending station, it would be a minimum. By swinging the ship in azimuth, the compass heading, when the strength of signal was a maximum, would indicate the line of bearing of the sending station. The practical difficulty in the way of operating this system to the best advantage is the very short waves which are necessary on account of the comparatively short distances that can be obtained be- tween wires on board ship. Two methods of determining the bearing of a light-house from a ship are being tried out. Both depend on the application of Marconi's dis- covery (art. 188), that a horizontal aerial sends more strongly from the direction away from the free end of the aerial, and receives more strongly from the direction in which it sends best. A number of aerials radiating from the sending station (light-house) like the spokes of a wheel, are in- stalled on shore. If a ship sends, the spoke on which her signals are re- ceived the strongest is the one having the same bearing from the light- house, as the light-house has from the ship. If the light-house sends, it sends on each aerial in succession, on a time schedule known to all ships. For instance, the ship knows exactly at what times the light-house trans- mits from the aerial bearing north from the light-house. If received signals are loudest at those times, the light-house bears north from the ship. BELLINI-TOSI WIRELESS COMPASS. 217. Difference in the planes of direction of two similar aerials makes a difference in the strength of received signals whether the distance between wires is half a wave length or not. And this fact is utilized in the Bellini-Tosi apparatus where two such aerials are installed (see fig. 63) in planes at right angles to each other or nearly so. The open circuit receiving coils are mounted so that they make the same 168 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. angle with each other as their aerials and their angular position relative to the ship's head is measured and shown. The closed circuit coil can be placed in the plane of either aerial coil or in any intermediate position. Its plane, relative to that of the other coils when the strength of signals is a maximum, is approximately that of the passing waves and is indicated in degrees relative to the ship's head. PORTABLE AND AUXILIARY SETS. For the requirements of auxiliary sets see the law in appendix D. 218. Portable sets, as their name indicates, are special small sets which have their own source of power, such as a foot or hand operated generator or 96.— N. E Navy Portable Set. storage battery, and when used on shore have portable masts for supporting the aerial. On board ship this single wire aerial can be run up by signal halliards, and if insulated wire is used (since portable sets work usually at low voltages) no particular care need be taken to prevent the wire from touching the mast, deck or rigging. To secure good results with portable sets, careful tuning is required both for sending and receiving. A hot wire ammeter and wave meter are useful adjuncts of portable sets as well as of those of larger power. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 169 The suit-case type illustrated in fig. 96 weighs about 75 pounds com- plete. It has a motor generator for ship use, which lias an output of 50 watts and can be plugged in on any lighting circuit. Small gasolene driven generators are used for some portable shore sets, the entire send- ing and receiving apparatus being mounted on wheels. The power or hand operated generator set of the suit-case type is good for about 20 miles. A complete set is seen with condenser, inductance, and key in the left half; motor generator, quenched gap, transformer, and receiving apparatus in the right half of the case ; with the plug for connecting up with the lighting or power circuit at the upper left-hand corner. 219. To illustrate an actual wireless telegraph installation the station at Sitka, Alaska, has been selected. This station is situated on Japonski Island (see frontispiece). The masts, rigging and rigging insulators, aerial and buildings are shown in fig. 97; one unit of the generating sets in fig. 98; the receiving apparatus in fig. 99. These figures repay study as illustrating a neat and workmanlike installation. The sending and receiving apparatus is after the designs of Professor Pierce. Figs. 100 and 101 illustrate actual receiving sets of other types, the elementary diagrams of which are shown in figs. 83 and 86. The construction and arrangement of both sending and receiving apparatus will continue to vary, but a careful study of elementary dia- grams (figs. 29 to 29e, 40 to 48 and 77 to 92) in connection with installa- tion diagrams like figs. 102, 103, 104, 105, which accompany each set will enable an electrician to connect up and operate any set intelligently. There are too many types of apparatus in use to warrant a detailed description or illustration of each. Such description and instructions are furnished with each set. This manual has therefore been confined to the principles common to practically all wireless sets. AIRPLANE RADIO TRANSMITTER, The first airplane set used in the navy weighed about 60 pounds and had a sending range of about ten miles. The set was tested out at Annapolis, Md., in 1911. A counterpoise antenna, half on each plane, was used with this transmitter. The power was supplied from the propeller shaft, by belt- ing to the radio generator. In 1915 several types of airplane sets, ranging from one-quarter to one-half K. W. were tried out and used in service. The apparatus is mounted on the fusilage of the airplane, in front of the observer, in order to be easily operated. A small generator, propelled by a wind-driven fan, is used as the source of power. A counterpoise antenna is used, using a trailing wire. The weight of the complete set is approximately 100 pounds. Its transmitting range is roughly 75 miles. 170 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 97. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 171 ■ 1 Swif- _ M 1 iJ ■ 1 ^^^ ^B l^f ^ J r4=H J TW\ ^H HpiP fXn^ y7k:X\ 1 Ulft' ~T \-V ll "^"^"J^ 1 i^'iyW^ MyMj^ 1 : ■:. ^^minn M|tt| Fig. 97a. — Washington Station. 172 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 173 Fig. 97c. — Insulation at Base of One Leg of Tower. 174 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 175 FiQ. 99. 176 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Fig. 100. — ^Wireless Telegraph Receiver MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 177 TO AER/AL c-s^s^ vix£ ^PARK (SAP ^^0^ AMMETER PRIMARY REACTANCE KEyO= {=&J - — TEZLEFUNKEN. Fig 102. 5 ^ \XL ^ HOT WIRE. AMM- 5HIP MAINS CLOSED CIRCUIT OPEN CIRCUIT INPUCTANCC. INDUCTANCE FESSELNDEN FiQ. 103. 12 178 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 179 -* ■* — I—*' ♦■ Hi" iir Chapter VIII. INSTALLATION, ADJUSTMENTS AND MEASUREMENTS. INSTALLATION. 220. For installation ample room is available at all shore stations. On board ship, a room having about 100 square feet of floor space, with no dimension less than 6 feet, should be provided for the installa- tion and operation of a wireless telegraph set. The operating room should be well ventilated and lighted, as nearly sound-proof as practi- cable, and free from vibration. The exact location of the room is not of great importance, provided a good lead to it for the aerial can be ob- tained. The farther this lead is from large conducting bodies the better. The room should have a well-insulated entrance for the aerial and should be fitted with an operating table about 2^ feet wide, not less than 7 feet long, and of convenient height for working the sending key while sitting down. The table should be strongly built of dry, well-seasoned wood. The instruments should be mounted on the table so that they are at safe sparking distance from each other and from any part of the oper- ating room. The receiving instruments should be as far away from the sending instruments as practicable. The induction coil or transformer may be mounted on the bulkhead or under the table. In any case it should be where its terminals are not likely to be touched accidentally. The motor generator is preferably installed near the operating room, but outside of it. It may be installed in the operating room or in the dynamo room. The connections between all parts of the sending and receiving instru- ments should be as direct as possible, and in the case of sending instru- ments they should be of large surface and well insulated by air or other nonconductors. Sharp turns in connecting wires should be avoided on account of brush discharges, which always start at corners. The effect is the same as if the electricity were traveling too fast to turn corners. The necessity for bringing a number of leads to the combination switch for sending or receiving detracts considerably from the simplicity of the installation and to a slight extent from the efficiency of the set as a whole. High-potential leads should be kept well away from low-potential leads, and where they cross it should be nearly at right angles. V u Plate 1.— 600- Watt Radio Transmitter. Front View. PuiE 2.— 500-Watt Radio Transmitter. Side View. Plate 3.— 500-Watt Radio Transmitter. Rear View. Wiring Diagram. 500-Walt Transmitting Set. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 181 The ground connections should be electrically good and of large area. The receiver (and the transmitter when practicable) should be wired up before installation, requiring only to be secured in place and attached to aerial and ground. The sending appliances should be so arranged that the leads connecting the condenser, inductance, and spark gap of the transmitter will be of minimum length. At shore stations means should be provided outside the operating room for disconnecting the aerial from the operating circuit and connecting it direct to ground. On board ship a lightning switch should be installed which when in use will safely and completely disconnect the aerial from all of the re- ceiver and transmitter circuits and connect it direct to ground. The aerial should be well insulated where it enters the operating room and where it passes through decks or bulkheads. Porcelain or glass insulators are best for this purpose. When necessary to guy the aerial at any point an insulator should be used in the guy line. The suspending or hoisting halliards of the aerial should be insulated. Two types of suitable strain insulators for this purpose are shown m figs. 106 and 107. FiQ. 106.— Aerial Insulator— Buck Link — Strain 10. Fig. 107.— Aerial with Locke No. 105 Insulators. ^« "»^ rE 3.— 500-Watt 1 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 181 The ground connections should be electrically good and of large area. The receiver (and the transmitter when practicable) should be wired up before installation, requiring only to be secured in place and attached to aerial and ground. The sending appliances should be so arranged that the leads connecting the condenser, inductance, and spark gap of the transmitter will be of minimum length. At shore stations means should be provided outside the operating room for disconnecting the aerial from the operating circuit and connecting it direct to ground. On board ship a lightning switch should be installed which when in use will safely and completely disconnect the aerial from all of the re- ceiver and transmitter circuits and connect it direct to ground. The aerial should be well insulated where it enters the operating room and where it passes through decks or bulkheads. Porcelain or glass insulators are best for this purpose. When necessary to guy the aerial at any point an insulator should be used in the guy line. The suspending or hoisting halliards of the aerial should be insulated. Two types of suitable strain insulators for thia purpose are shown m figs. 106 and 107. FiQ. 106.— Aerial Insulator— Buck Link— Strain 10. P^Q. 107. — Aerial with Locke No. 105 Insulators. 182 MANUAL OP RADIO TELEGRAPHY AND TELEPHONY. 221. The large momentary currents in aerials produce large inductive effects in conductors near and parallel to them. This is more noticeably the case in wire stays or masts, shrouds, braces, etc. These should be grounded. It should also be noted that an aerial wire parallel and near to a long lighting or power lead may induce sufficiently high potentials in the lead to puncture the insulation and cause sparking between it and other con- ductors in the vicinity of combustible material, thereby causing fires. Or it may puncture the insulation and cause a bum-out of an armature, field, or transformer. All of these effects have been experienced. They are especially dangerous to the wireless sending apparatus. PROTECTIVE DEVICES. 222. Rigging of masts at shore stations is divided into short lengths by strain (usually locust) insulators. Wire braces were formerly served near the middle with chokes made of No. 26 B. & S. soft iron wire for a length of about 10 feet, the object being to ensure that no conductor of approxi- mately the same natural period as the aerial should be in its immediate viciniiy (art. 195). Now that longer wave lengths than that of any single piece of ship's rigging are in general use, it has been found beneficial to carefully ground to the hull all metal rigging, a slight increase in the effective capacity of the aerial resulting. Wire leads at shore stations are lead covered and the lead grounded. Wires in conduit on board ship are protected by the conduit being grounded. Wires in the open should have an armored cover well grounded. The protective devices shown in fig. 108 are installed to conduct to ground induced high potential at (a) Terminals of primary of transformers. (b) Terminals of armature of alternator. (c) Terminals of field of alternator. (d) Terminals of shunt field of motor. (e) Terminals of armature of blower motor. FUSE RESISTANCE VWNAA/\/WWW- I 25,000 OHMS Fig. 108. — Protective Appliance. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 183 One type of protective device consists of two l-microfarad condensers " B " connected in series, the middle connection grounded and the two outer terminals connected by strip copper leads (as short as possible) to the apparatus to be protected. In parallel with each condenser is a per- manently set spark gap " C " between .001 and .003 inch in width. Each leg of the device is fused with a 3-ampere cartridge fuse " A." A resist- ance rod "D" (25,000 ohms), with its middle point grounded, is also connected across the line. (See figs. 108a and 108c.) In addition to the above, safety spark gaps are fitted to receiving tele- phones. Secondary terminals of transformers are protected by chokes /7 /7 /9 C/tOOA/O ^T I Fig. 108a. made from the leads, and by safety spark gaps permanently set at the maximum safe sparking distance. Safety spark gaps will be fitted to aerial insulators at stations subject to lightning strokes. The usual standard protective device consists of two 0.05-microfarad mica condensers in series, connected across the circuit to be protected with the wire joining the two condensers grounded. No fuse or protective spark gap is used. The resistance rod is also no longer used, on account of the objection to grounding the direct current power circuit through this rod, thus giving a " ground " indication on the ground indicators at the main switchboard. Lugs are provided on bus bars, forming part of the pro- tective device, the lugs on one end being marked " line " and the other end " apparatus." Thus, the leads to the apparatus to be protected pass through the protective device; hence, if the protective device is removed the circuit is opened. This eliminates the possibility of disconnecting the 184 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. device and leaving the apparatus unprotected. In order to prevent short circuits, due to metal being placed across the bus bars, the entire pro- tective device is covered with an insulating cover. A photograph of this type of protective device is shown in fig. 108b. 223. All wireless telegraph sets are fitted with a multiple switch which in the sending position disconnects the receiving circuits from the aerial and ground and breaks detector and telephone connections as may be necessary to protect them from induced high potentials. When in the receiving position this switch opens the primary or sec- ondary circuit of the transformer and, if the motor generator is in the operating room, operates a relay for opening the field of the motor, or in some cases short circuits the armature to bring it to a stop quickly. This switch should also stop the blower motor. Fig. 108b. The necessity for the above detracts considerably from the simplicity of an installation. ADJUSTMENTS. 224. This includes calibration and tuning. A station is tuned when both sending and receiving circuits are correctly calibrated, coupled and adjusted to the standard damping and standard wave lengths. Since the periods of the open circuits of both sending and receiving sets depend on the aerial with which they are used and the constants of the latter can not usually be predetermined, the open circuit has to be calibrated after the set is installed. The closed circuit of receiving sets can readily be cali- brated before installation. Also the closed circuit of sending sets, if wired up before installation. The wave length of a circuit made up of a calibrated inductance and a calibrated capacity can be calculated from the formula : Wave length in meters = 1884.9 5 VCL. When C is in microfarads, L is in micro-henries, the formula being derived from the fundamental : T = 2TrVLC. For calibrating directly in wave lengths the aerial circuit and other circuits not already calibrated, wave meters are supplied, which are used as receivers to calibrate sending circuits and as senders to calibrate receiv- MANUAL OF RADIO TELEGRAPHY AND TELEPHONY, 185 ing circuits. After calibration the adjustment of these circuits to the same wave length and to the desired coupling is called tuning. 225. To be completely in tune, a spark sending set should have the cir- cuit made up of the A. C. armature winding, primary leads, and primary winding of transformer, in resonance (tune) with that formed by the sending condenser and secondary winding of the transformer. Both cir- cuits should be in resonance with the alternator frequency. The closed sending circuit should be in resonance with the open cir- cuit and the coupling and decrement of the open circuit such as to afford the necessary selectivity to the receiving circuits with the best efficiency of radiation. A rx L. I ^ ilX Ofioi/^o -"=" Fig. 108c. Receiving circuits to receive from such a sender should be in resonance with each other and with the sending circuits and should have the same coupling as the sending circuits. The telephone diaphragm should be in resonance with the wave triain (alternator) frequency and with the operator's ear. As was previously stated, instead of designing telephone diaphragms for resonance with alternators, we design the alternator for resonance with the telephone diaphragm or with the human ear. Resonance is thus seen to be a vital quality in wireless telegraph cir- cuits. (1) Resonance of alternator frequency with primary sending circuit. (2) Resonance of primary circuit with secondary sending cir- cuit. (3) Resonance of closed oscillating circuit with open radiating circuit. (4) Resonance of coupled receiving circuits with each other and with coupled sending circuits. (5) Resonance of telephone diaphragm with primary frequency. (6) Resonance of human ear with telephone diaphragm. Of these (1), (2) and (5) are elements of design and are not change- able at the will of the operator, (1) and (2) can be varied to a certain 186 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. extent by reactance regulators which in some sets are provided for both circuits; but it is preferable to cover this feature in the original design of the transformers. (3) and (4) are entirely under the operator's con- trol and on them the efficiency of the set depends. MEASUREMENTS, WAVE METERS AND THEIR USE. 226. Standard calibrated oscillating circuits called wave meters, which are adjustable at will to a great number of known wave lengths are used for calibrating and tuning. When adjusted to resonance with the circuit to be measured the fact is indicated according to the type of wave meter by a maximum of sound in a telephone, a maximum glow in a vacuum tube, a maximum reading of a hot wire ammeter or the brightness of a glow lamp. The wave meter is a necessary adjunct to every radio station. It is a closed oscillating circuit of which the capacity or inductance, or both, may be varied. It is very carefully calibrated so that the wave length corre- sponding to any capacity and inductance used may be read directly from an attached scale. When two circuits are in resonance any oscillations set up in one will have the maximum effect in the other. Thus, by placing a wave meter near any oscillatory circuit and varying the capacity and inductance of wave meter and thus gradually changing the period of the wave-meter circuit, the current-indicating device will give its maximum indication when the two circuits are in resonance. At this time the scale of wave meter will show the wave length of the resonant circuits. This is one of the principal uses of the wave meter. In this way the various circuits may be calibrated so as to permit the use of any wave length desired. The wave meter may be used as a receiver when any form of detector is included in its circuit. It may be used also as a miniature sending set with proper means of exciting it. When used in this way, it may set up, in a nearby oscillatory circuit, waves of any desired length within the limits of the meter. This method is always used in calibrating the receiving circuit. The Telefunken wave meter, large E. G. W. type, is extensively used in the naval service. The wave meter may be used for the following operations : 1. Tuning and measuring wave lengths of all the circuits of a radio telegraphic installation, i. e. : a. The closed sending circuit. b. The open sending circuit (aerial). c. The closed receiving circuit. d. The open receiving circuit. 2. Testing of transmitter tone. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 187 3, Testing of detectors for sensitiveness. 4. Determination of capacity, self-induction, coupling coefficients, long- distance wave lengths, etc. The Pierce wave meter uses a telephone exclusively. It is suitable for determining resonance only. The Donitz meter uses a hot wire ammeter or air thermometer, whose maximum reading indicates resonance and lower readings the relative amount of energy received when the wave length of the wave meter is varied. These readings can be plotted as a curve. Wave meters now furnished can be used either with a detector and telephone or with a hot wire ammeter or galvanometer for determining resonance and making other measurements. Wave meters are also fitted with small spark gaps and spark coils so that they can be used as senders for calibrating receiving circuits. Instructions for the use of wave meters are supplied with the instruments. The two wave meters illustrated A - binOnc posts. C - variablC capocitt. D- OrNiMOMLTCR. L- INOUTANa. P - POINTER. T - SWITCH. L - LONG WAVt LENGTHS s - ShORT WAVE LENGTHS PiQ. 109. (figs. 109 and 110) are early tj^es, but differ only in details from those now supplied. measur:ements of wave lengths. 227. Fig. 109 shows the Pierce wave meter referred to in art. 226 This meter is used for calibrating and determination of resonance by means of sound only. Fig. 110 illustrates the original Donitz wave meter with air thermom- eter. A hot wire ammeter or detector and galvanometer are now used with this instrument. Instructions for the use of the Pierce wave meter, which follow, are applicable in general to all wave meters. 188 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. IIIHIIIIIIUIIIIIIH ^ h FiQ. 110. — Slaby-Arco-Donitz-Wave Meter. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 189 iNSTRDCTiaNS FOR USINQ THE PlERCE WaVE MeTER OF THE MASSACHU- SETTS Wireless Equipment Co. A. — calibration of sending station. 1. To make the Instrument ready for use. — Take off the cover, fold back the hinged loop, and attach the leads of the telephone receiver (stowed in cover of box) to the two binding posts near together to the left of the metric scale. 2. Placing the Instruments. — Place the instrument near the circuit whose wave length is to be determined, and by turning the loop on its • projecting horizontal axis bring it in such a position (parallel) that it will be linked by the magnetic lines from the oscillating circuit. The proper distance from the loop to the oscillating circuit depends on the intensity of the oscillations. When the observations are to be made directly on the Leyden jar circuit the wave-meter loop may be one or two meters from the discharge circuit, while if observations are to be made on parts of the circuit in which the currents are feebler, this distance may be reduced to a few centimeters. In setting up a station the wave lengths of the various parts of the cir- cuit may be determined separately in the usual manner. When the station is already set up ready for use, the wave length or the two wave lengths it is radiating may be determined by placing the wave meter near the wire to ground or the wire to antenna with the loop of the instrument in the plane of the wire. 3. Regulation of Spark. — Make the spark of the station short and adjust the current in the discharge circuit so that the spark is clear and sharp. 4. Taking Observations. — Put the telephone receiver to the ear and with the hand holding the receiver, touch one of the metallic tips of the lead where it enters the receiver. This will shut out the general hum due to the alternating current in the transformer. If no such hum is present it is not necessary to touch the terminal in this manner. Now with the free hand turn the handle in the center of the instru- ment and set for a maximum in the telephone. In making these observations the switch to the right must be either on " L " or " S." This switch should be on " L " for long waves and on " S " for short waves. With the switch on " S " read the position of the pointer or red scale. The position of the pointer for a maximum sound in the telephone is the wave length in meters. If the switch is on " L'* the black scale should be read and gives the wave length in meters. In case the sounds in the telephonic receiver are too loud for accurate settings, their intensity may be reduced either by moving the instrument farther away, or more conveniently, by turning the receptor loop so that 190 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. the inductive action is diminished. In the final setting it is desirable to have the sound in the telephone just audible at resonance. 5. Use of Geissler Tube for Demonstrations. — If it is desired to use a Geissler tube with the instrument, leave the telephone connected in, con- nect one terminal of the tube to the nearer left-hand post along with the telephone lead and the other terminal of the tube to the idle binding post at the back of the instrument. The tube is then in parallel with the condenser of the wave meter and should glow at the proper setting. B. — CALIBRATION OF RECEIVING STATION. 6. Use of Wave Meter as Sending Station. — Take off the telephonic receiver of the wave meter and put in its place the spark-gap supplied with the apparatus. This attachment has a coil in its base of approximately the proper inductance to replace the telephone. Use the wave meter as a sending station and for any given wave length of the wave meter set the receiving station to resonance as in receiving messages from a distant source. 7. SparJc-Coil. — In using the wave meter as a sending station it should be actuated by a small spark-coil. Attach the leads from the secondary of the spark-coil to the two sides of the wave meter spark-gap. This gap should be opened not more than a few hundredths of an inch (.1 or .2 millimeters). When the gap is too wide, sparks occur inside the wave meter between the plates of the condenser. 8. Position. — The wave meter when used as a sending station should be placed about three meters from the receiving antenna, and should not be approached too closely by the observer who is listening at the telephone of the receiving station, since conduction or induction through the body of the observer and along his telephone leads will result in a general hum that can not be tuned out. C. — PRECAUTION AND CARE OF THE INSTRUMENT. 9. Do not attempt to open the telephone receiver, and do not change or break the leads of the telephone as injury to the telephone will disturb the calibration. Due to climatic conditions and other causes all wave meters are subject to changes of characters. They should be frequently checked with a standard meter and errors noted. 10. In stowing away the apparatus be careful to leave the pointer free from obstructions. To this end, whenever the instrument is to be trans- ported it is advisable to disconnect the telephone and place it in the clamp in the cover of the box with the leads secured under the wooden buttons. 11. The receptor loop should be folded in with hnob upma/rd so thai pointer can he rotated under the loop without interference. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 191 228. In determining wave lengths, three methods for fixing the con- denser reading for maximum current may be used. 1. For a rough determination the apparent position of maximum reading may be fixed by a single observation. 2. For a more accurate determination the maxi- mum reading of the hot-wire ammeter or galvanometer may be noted, and the condenser pointer be moved first to the right until the current reading falls by a certain amount, and then to the left of the maximum position until it falls to an equal amount. The position half way be- tween these two condenser readings may be taken as the true maximum. 3. The values of the current reading corresponding to a large number of condenser readings on each side of the maximum may be taken, and a curve plotted having condenser readings as abscissas and current readings as ordinates. From this curve the most accurate possible position of the maximum can be obtained. For measuring the wave length of any sending set as it is being used it is only necessary to bring the wave meter into position near a single loop in either the antenna or ground connection, taking care that there is no direct induction from the helix into the wave meter coil, close the key for a long dash and ascertain, by moving the pointer over the grad- uated scale, the position of resonance as indicated above, i. e., by tele- phone H. W. A. (hot-wire ammeter) or galvanometer. This will gen- erally be on the longer wave or " upper hump " since in stations sending out two waves the longer wave contains the most energy and is the most easily read.* To locate the short wave (" lower hump ") it may be neces- sary to couple the wave meter helix quite closely with the loop. In seta having loose coupling and those supplied with quenched spark gaps, hut one position of resonance should he found. To ascertain the wave length of the closed circuit, disconnect the aerial, couple the wave meter with the helix and proceed as before. But one wave length vdll be found. This, if the same as the first one measured, will show that there is but one length of wave being generated and radiated. The two operations above described can be performed in less than five minutes. To ascertain the wave length of the aerial is not so easy. To do so, disconnect the closed circuit, place a temporary spark gap in the ground lead of the aerial, connect leads from the transformer to each side of the gap and, in ordinary ship sets where the capacity of the aerial is small, also put a Leyden jar across the gap. Place the wave meter near the aerial inductance and adjust to resonance. This reading should be the same as that found for the closed circuit. The above meas- urements will show whether the open and closed circuits are in resonance and what wave length or lengths are being sent out. * See law relating to use of " pure wave," p. 248. 192 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. A rapid way to adjust a transmitter is to tune the closed circuit to the desired wave length with a wave meter and then tune the open circuit to resonance with the hot wire ammeter, but this has its limitations (see arts. 230 and 340). 229. If both open and closed circuits read 425 meters and the upper hump is found to be 450 and lower hump 390, the percentage of coup- ling * is 450-390 425 ^*"^- If but one hump is found and that at 425, with an ordinary spark gap, the circuits are very loosely coupled. This fact can also be determined approximately by an inspection of the sending helix. If direct con- nected and but one wave length is found, it will also be found that the number of turns common to the two circuits is very small or less than one turn. If inductively connected, that the active parts of the two helices are not close together, in other words, the mutual induction is very small. The single wave found on loose coupled sets using an ordi- nary gap is not as sharp as that found on the closed circuit read separately. Some mutual induction is necessary to transfer energy so thai the two waves can not quite merge into one. TUNING CURVES. 230. Tuning curves showing the wave length for any adjustment of each circuit should be made, plotting the wave meter readings as wave lengths (horizontally) on the bottom of a sheet of cross section paper (standard A sheet) and the number of turns of the helix for each read- ing on the side (vertically). Draw smooth curves (fig. Ill, curves (1) and (2) and fig. 115, curves marked "aerial" and "exciting") through the points thus found for both the open and closed circuits. An inspection of these curves will show how many turns of the helix must be included in each circuit for any given wave length. When set by these tuning curves to the same wave length the accuracy of the curves can be checked by the reading of the H. W. A. If the setting is correct, any change of inductance or capacity in either circuit will decrease the reading of the H. W. A. It must be remembered as stated elsewhere that it generally takes a change of several turns of inductance to change the wave length of the open circuit appreciably, while a change of less than one turn will change the * Percentage of coupling, as defined above, differs from coeflBcient of coup- ling, which is defined as the ratio of the mutual induction of two circuits to the square root of the product of their respective self inductions or M — 7= — coeflBcient of coupling. MANUAL OF RADIO TELEGRAPHY AND TELEPHONT. 193 »v<» f« /< tngfh . Mmtors. Oi ^. r- 1 1 i 1 1 1 I / I-" § s o s ^ 5 V V < >>- \ \ / / / • — 1 \ '5 \-*---5 1 1 I ; ^ 1=^ \" \ V \ \ -5^ 1 '•'C-> ?*'> 1 '^^ \ \ /-• =o^^ '1 - ""■~- .. ■ 5: P— Ce^/>/. •fl- vf^. fam ''-. '"-*. <■ ;^ — ^ 1 ,^ ■»\- (f <• '«• 31 \7 I r « 1 f 1 1 ^ \ O c S / N ^\- V f . \ 5 < • z > 5 1 1 1 \ Vi \ r 3 oa z > o A 1 1 1 \ \ \ 1 3- $ "i 1 / i M \ \^ \ \ \ I > 7 n r ( / / \ ■v ,/ n • 3 Q. 01 e 1 1 1 >* \ \ \ \ / ^ 3 O c ■0 r m Cb f 1 1 1 r X §4 9X 3 oa o > Is 1 1 >* V / / ^ \ v^ a • c < n \ 1 % / / \ s. '^ \ o z -^ / / \ \^ s / ( 1 i 1 K ? / / \ s s t-1 /^o^ AK/z-e Ammefar Sca/c Oer/. .... •# Oct/ a. • / /4/»jo 13 194 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. wave length of the closed circuit considerably, so that it is much easier to throw the two circuits out of resonance by changing the closed circuit turns than by changing the open circuit turns. It must also be remembered that the tuning curve for a closed circuit is only correct for the capacity in the circuit at the time the measure- ments were made. ^ The removal of a jar from the condenser; change of shape or length of leads to helix; bad connections to jars — each and all change the wave length of the closed circuit and throw it out of resonance with the open circuit with marked decrease in radiation. The H. W. A. can be used to adjust two circuits to the same wave length but it gives no indication of what that wave length 'is. When a wave meter is available, it is shown above, that to take a reading of the closed circuit wave length requires but a minute's work. The wave length of the open circuit with the same number of turns included varies little from any cause, and if the insulation and ground are good the causes of decreased radiation should be looked for in the spark gap or in bad connections, broken jars, etc., in other parts of the closed circuit. It has been proposed, where the coupling is such that two wave lengths are radiated, to throw the two circuits slightly out of resonance in order to increase the proportion of the total energy in the long wave ; but no dis- tinct gain in efficiency has been noted. It is better to loosen the coupling to the point where but one wave can be found, even if this is beyond, as it usually will be, the point where the highest hot wire ammeter reading is obtained. It must be remembered, however, that efficiency varies directly as the H. W. A. reading and the latter must be maintained as high as possible consistent with sending out waves of but one length. 231. In calibrating closed sending circuits, the shape, as well as the length of the leads, must be taken into consideration. This shape must be the permanent one. In sets now being supplied connections with the helix or spiral are made so as to avoid any change of shape with change of wave length. In calibrating the open circuit of receiving sets the same difficulty will be found in obtaining sharp resonance as when calibrating open sending circuits. Calibrating sending and receiving circuits enables us to select and send and set our instruments to receive definite known wave lengths and is the first requisite of tuning. In order that our receiving circuits may be selective, i. e., respond only to the wave lengths for which they are adjusted they must have comparatively large self-induction. In other words, they must be what are known as stiff or rigid circuits. In order MANUAL OF RADIO TKLEGRAPIIY AND TELEPHONY. 195 aeo 2.60 Z70 2B0 2.30 300 3IO 320 330 ^M-O 350 360 37o 3BO 3&0 WAVE LENGTHS IN METERS Fio. 112. 196 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. that a wave train may be long enough to build up current in a rigid re- ceiving circuit the sending circuit must be a persistent oscillator, i. e., it must be slowly damped. (Fig. 18h.) It must not be forgotten that the currents in a very persistent oscilla- tor like a closed sending circuit are mostly dissipated in heat, while we wish to have their energy radiated in electric waves. We must therefore strike a mean between the efficient very highly damped sending circuit which radiates nearly all the energy it receives in one or two waves, but which affects all receiving circuits alike and the inefficient persistently oscillating sending circuit which dissipates most of its energy in heat but which is favorable for selective receiving. Sharp tuning or selectivity depends, therefore, on self-induction in the radiating circuit as well as in the receiving circuits. 232. The air thermometer readings in the wave meter measure the received current in the same way that a hot wire ammeter in the aerial measures the sending current. The readings of both meters vary accord- ing to the heat generated by the currents and this heat varies as the square of the current. Dr. Austin finds that the loudness of signal in a receiving telephone is proportional to the square of the current and that, if a rectifying detector and galvanometer are used for measuring the received currents, the galva- nometer deflections are also proportional to the square of the oscillating currents ; so all these ways of measuring are directly comparable. RESONANCE AND AUDIBILITY CURVES. 233. Eeferring to the first paragraph of art. 228, the third method given for determining the position of maximum current in the receiving circuit (wave meter), and, therefore, the position of resonance between the wave meter circuit and the transmitting circuit results in a curve which may be called a resonance curve. Eesonance curves are illustrated in figs. Ill to 116 and particularly in Figs. 113 and 114. In fig. 113 the wave meter condenser readings are laid off horizontally to scale (these are the abscissas of points in the curve). The correspond- ing hot wire ammeter or galvanometer readings are laid off vertically (these are the ordinates of points in the curve) . Instead of the condenser readings, we might use directly the wave lengths which they represent, and instead of the galvanometer readings, we might use multiples of the least current, which would make a signal readable, or, rather, multiples of the least received power (see table 8, appendix A), which would make a signal readable. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 197 Power and audibility are proportional to the galvanometer readings, so that the shape of the curve would not be changed and it might be called an audibility curve as well as a resonance curve. 234. Audibility curves can ])e plotted with any calibrated receiving set as well as with a wave meter. The wave length at the lower or upper A = 750 M. Sl + 6 = 038. bU r r— 1 r\ / 50 / ] / 40 / / / '^f 30 \ 20 \ \ ]| y \j 10 \ \ / \ / / \ / / V / ^8 4 9 5 5 I 5 a CONDENSER SETTING IN DEGREES. FiQ. 113. — Resonance Curve Taken with Wave Meter. limit of audibility, divided by the resonant wave length, might, when receiving from a standard transmitter at a standard distance, be called the figure of merit of a receiving set, an indication of its selectivity. Or if the same receiving set is used, the audibility or resonance curves of different transmitters are a means of comparing their dampings and their suitability for selective receiving might be said to give the figure of 198 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. merit of the transmitters. For instance, in curve 7, fig. Ill, and curve III, fig. 112, let the portion above the heavy line X Y represent the range of audibility at any distance, say 100 miles from the transmitter. In curve 7 a change of 150 meters on either side of the position of maxi- mum loudness (resonance) would be required to render signals inaudible, while from the transmitter of curve III a change of only 12 meters either way from the position of resonance would cut out signals. Since the same or similar wave meters were used in plotting these curves, we conclude that the difference in their shape is due to differences in the transmitters, and that the transmitter of fig. 112 is a more persistent oscillator than that at Guantanamo (fig. Ill) . The maxima of these curves have no direct relation to each other, since they are produced by different amounts of radiated energy and different relative positions of the wave meter and transmitting circuits. It is their shapes alone which are the subject of comparison and discussion. The shape of each curve will remam the same, whatever the position of the receiving circuit. In neither of the curves under discussion would the lower hump audibly affect the receiving apparatus. 235. Where audibility changes rapidly with small change of wave length, the circuits are said to be sharply tuned. It will be noted from the resonance curve of the aerial (fig. 112) that sharp tuning with it alone is not possible, but that when coupled with the persistently oscillating closed circuit, the transmitter as a whole gives fairly sharp resonance. There is no more possibility of an escape from a whip-crack transmitter than from static. It is found that the shape of resonance curves depends on the damping of the receiver as well as that of the transmitter, and that sharp curves, like those in figs. 113 to 116, cannot be obtained without a stiff receiving circuit loosely coupled (par. 239). 236. The resonance curves of figs. Ill, 112, 115 and 116 are from direct connected transmitters. The transmitter of fig. 115 was so loosely coupled as to give but one hump in the curve. The transmitter of fig. 116 has a quenched gap. The lower hump (curve II), shown at about 860 meters, and the flat part of the curve, at 1075 meters, may indicate the energy radiated while building up (fig.lSh) before the transfer of energy to the open circuit is complete. The percen- tage of coupling indicated by these two humps is 22 per cent, but by far the largest part of the energy is radiated at the natural period of 975 meters. (The two curves in this figure have a different scale of ordinates, MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 199 CURVE! 1_ wave: L-CMCSTH DAtS/1P'|M<3 I SF'ARK 3S10 .021 2 AEC 2 COIL-S SS20 DIRCCT .OlS 1 1 . 2 "• J35.4-** 1 ' 1 /y v_ V cucve. -I cuizvEi - a 5A- 129 S6 131 S8 133 e,o 62 135 137 '10. 114. 66 68 14-3 200 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 60 that their coincidence is only apparent. The maximum readings are in reality smaller for the open than for the closed circuit.) 237. The resonance curves in fig. 114 are (curve 1) from a feebly damped, inductively coupled transmitter, with a synchronous spark gap, and (curve 2) from an undamped arc transmitter, with the arc directly in the open circuit. Theoretically, all the damping of curve 2 should be due to the receiver, but slight inequalities in wave lengths emitted by the arc can be found, which have the effect of an apparent damping; it tends to broaden the resonance curve. As has already been stated, on account of the time required to properly adjust the arc, arc sending sets are in operation all the time, sending con- sisting only of change of wave length, so that they can be tuned in on the wave length of the intervals instead of the dashes and dots. On account of the very steep resonance curves of such transmitters, when the range of audibility is near the upper part of the curve, it is very easy to miss when tuning, i. e., a sharply tuned transmitter is more difficult to find than a broadly tuned one, unless its wave length is known exactly. MEASUREMENT OF DAMPING. 238. In addition to being able to estimate damping from tuning or resonance curves we can measure it directly as follows ; 72 It was stated that the damping of any circuit 8= ^-y , where R is the resistance, n the frequency and L the self-induction of the circuit. The theory of coupled circuits shows that the sum of the damping of the C —C I P two circuits 8i + 82 = 7r — 7^ iJ -^ 2 _n » where Cm represents the position of the condenser in degrees for most perfect resonance, and Im the maximum current in the second circuit corresponding to the position of the condenser Cm, and where I represents the current in the circuit corresponding to any other position C of the variable con- denser. This formula becomes much simplified for practical purposes, and gives in general accurate enough results, if, instead of plotting a complete curve, we change the variable condenser so that for the reading C, I*=i Pm- The quantity under the radical then becomes C —G unity, and 8i-f-82 = 7r ^^^ . Two values of C should be observed, one on each side of Cm, and the mean of the two values of the damping taken. If the current is measured by means of a thermo-element or a perikon detector in connection with a galvanometer, the readings of the MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 201 ""■ ■~~ """ 1 —~ —"" — 1 — — "^ T i 1 i • s u n ! <5 V z S _i S) J rt I »!** 5 $ (0 sc 5 3 J^ IS c/i -1 ? vi 1 ? oc (0 J f $ 1 ' 5 S i 1 \ «c s >• \ < 1 ' ^ 1 \ i z hi > s. ( s J 1 \ \, I M u \ 1 1 \ 1 5 s ^ If s if ' 1 " 'V^ /^ 1 1 ^ i » 1 If u N s i z \ > 1 * (3 > V S 1 \ 1 ■ > > N s. 2 t \ X- 5= > 5 \ V U -<- \ 3 1' ridi ^1 i S < V I 4 g \ V 1 < •« > itua. V ^ ___ _ -^ s. ■n — 3» It » ^ ^ N s. V « k V z (A g ^ > s \ « i^ f— *i tt\ 1 ^ n i ^ ft-^ \ il V Ot'M i^if^ t $ ^ '^ s. i^\ y/ 1 tf 1 s s (t s ^s *** "- ? uni ^ >- rib i 1 9 ? < h "^ •«« - i - ^ 2 «/i i 3 i \ - w a k •> K «> •■ • ii ♦ h •< 202 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. galvanometer are proportional to P ; that is C is so chosen that the gal- vanometer deflection is reduced to one-half that observed with Cm- If the current is read with a hot-wire instrument reading directly in am- peres, then the reading of the meter corresponding to C should be -— — of that corresponding to Cm, since 1.41 = V2. This expression gives the true value of the dampings of the circuits only when the coupling between them is extremely loose. If the coupling is not very loose between the two circuits, the apparent value of the damping will be too large. The proper degree of coupling can be ascertained by observing the point beyond which loosening the coupling does not decrease the damping. If the damping of the wave p meter circuit be known or can be calculated from the formula 80= -^ — ^, by subtracting this from the sum of the two dampings we get at once the damping of the other circuit. If we wish to express damping in terms of wave length A instead of capacity or inductance, it may be shown mathematically that the sum of the damping 8^ + 8^ = Stt ^^ — ^ , where as before A is the wave length, which reduces the square of the received current to one-half of that found for resonance at A,n.* 239. From the results of damping measurements it has been found that very sharp tuning is impracticable when a wave train contains less than 15 oscillations. This corresponds to a decrement of .2 (fig. 18h). Hav- ing measured the damping of the open circuit as coupled and found it too large, it is necessary to add inductance in order to decrease it, or to weaken the coupling in order that the total resistance R may be decreased. If it is not practicable to change the wave length, the aerial must be shortened to decrease its capacity while retaining the same wave length by adding inductance. Loosening the coupling also decreases the damp- ing. Eeceiving circuits can be stiffened without changing the wave length by putting a condenser in series to decrease the capacity and then add- ing inductance to keep the same wave length. But the damping of sending circuits can not be conveniently changed in this way on account of the high potentials which would be induced in the series condenser. The method of measuring damping just described is applicable to receiv- ing as well as to sending circuits. Eeceiving circuits have In general greater high frequency resistance than sending circuits, but specifications * A form of wave meter, specially fitted for measuring damping, is called a " decremeter." Marconi and Kolster decremeters are in use. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 203 204 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. require that their decrement shall not exceed .3 for any wave length within the range of the set. The law requires that the decrement of transmitters shall not exceed .2. MEASUREMENT OF SENDING CURRENT. 240. For measuring the sending current a hot-wire ammeter is installed directly in the aerial just above the ground connection. Those now sup- plied are graduated to read directly in amperes. Curve 6, fig. Ill, shows hot wire ammeter readings in open circuit for various couplings and wave lengths at the Guantanamo station. The A- ' t ^' f 1440-1120 oo^ maximum reading is for a coupling of — t^^t^ = *3^« The highest hot-wire ammeter reading shows that the circuits are in resonance and is usually taken also to indicate the best coupling; but except for circuits with quenched gaps the highest H. W. A. reading is usually obtained with a coupling which causes the radiating circuit to be too highly damped. It is therefore best to loosen the coupling until the shape of the resonance curves, or actual measurements, show sufficiently small damping; and then, by careful adjustment to resonance, attention to connections, to spark gap, and to regulator, get the highest hot-wire ammeter reading that can be obtained with that coupling and wave length. In order that the performance of different sets can be compared it is necessary that all hot wire ammeters be calibrated for reading directly and correctly in amperes. A hot wire ammeter which reads correctly on direct current should be calibrated for high frequency as follows : First remove the shunt and send with reduced power so that the de- flections will approximately cover the scale. This can be done either by cutting down the actual power or by loosening the coupling between the closed circuit and aerial. Note the deflections. Then close the shunt and leaving everything else unchanged, send again and note the deflection. The relation between the two deflections gives the ratio, for this wave length, of the shunted to the unshunted readings. If any other wave length is used, the shunt must be recalibrated since its effective resistance depends on the frequency. Eeports of current in aerial should always read correctly in amperes and be accompanied by report of exact frequency and input to transformer in amperes and volts. It is found that the distance of transmission varies directly as the oscillating current in the aerial, so that it is important to ascertain correctly what this current is. manual of radio telegraphy and telephony. 205 the shunted telephone method of measuring the intensity (loudness) of signals. 241. It is often desirable to make quantitative determination of the intensity of incoming signals, especially when tests are being made of either sending or receiving apparatus. This can be done if the station is provided with an electrolytic receiver, preferably of the free-wire type, and a resistance box. The connections are shown in fig. 103. E> ■E u K Fig. 117. — Detector Circuit with Shunted Telephone. Here L and L^ are wires running to the receiving circuit, K a stopping condenser, D the electrolytic, T the telephone, R a resistance box in shunt across the telephones, P the potentiometer, and C a choke coil to prevent the oscillations running around through E and P instead of passing through D when the shunt R is closed. Two 60-ohm telephones form a suitable choke. Whatever choke coil is used, it should be tested by being placed across LL^. If the choke is perfect no oscillations will pass through it, and its presence across LL^ will not diminish the loudness of the signals in the telephones. The measurement of the intensity of signal is made as follows: After the receiving circuit and detector are adjusted to give maximum loudness in the telephone, the shunt resistance R is closed and the resistance regu- lated until the signal just remains audible. The value of the current pulses c in the telephone, which are proportional to the energy of the incoming waves in the detector, is expressed by the following formula, vchere r is the value of the shunt, and t is the resistance of the telephones, and c* the least current audible in the telephones : r-\-t , c* is the audibility current, and the signal is often expressed as being 80 many times audibility. With care a series of measurements of inten- sity may be made to agree among themselves to within 5 to 10 per cent. Resistance boxes specially made up and calibrated for this purpose are called "audibility boxes." measurement of INDUCTANCE AND CAPACITY AND TOTAL RESISTANCE. 242. Inductances and capacities can be directly measured by wave meters as follows: 206 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Inductance. — A circuit is formed containing the unknown inductance, a known capacity (one or more standard jars), and a small spark gap. This circuit is used to excite the wave meter, and the variable condenser is varied until a maximum current in the wave meter is obtained. The two circuits being then in resonance, the product of the inductance and capacity in each is the same; that is, LC = L^C^, or if L is the unknown quantity, L= —^. Capacity. — If the spark circuit is made up with a known inductance and unknown capacity, by the same process we determine that C= ~^- . Li Total Resistance. — This expresses all the losses in the oscillating circuit, and is determined from the formula previously given 8 = ^^ • Having measured 8 and L and found n the frequency from the wave length, we have B = 2nL8. As yet we have no standard method of separating R into the equivalent radiation resistance and true high frequency resistance, which are its principal parts. THE MEASUREMENT OF LOGARITHMIC DECREMENT. Considering the Kolster decremeter, fig. 117b, the operation for measur- ing the logarithmic decrement is as follows : The rotary condenser is first set at the position of complete resonance as indicated by the maximum deflection of the sensitive hot-wire instrument, the scale readings of which are proportional to the current squared. This maximum deflection is now reduced to one-half its value by decreasing or increasing the capacity of the rotary condenser. The decrement scale, which may be rotated independently, is now set at zero, then clamped so that when the condenser is again varied it will rotate with it. Starting at the zero setting of the decrement scale with the hot-wire instrument reading one-half the maximum deflection, the condenser is varied continuously in one direction until the needle of the hot-wire instrument makes a complete excursion from one-half deflection to maxi- mum deflection and back again to one-half deflection. The scale reading now opposite the index mark is the value of S^-f S,, 8i being the decre- ment of the circuit under test and So the known decrement of the instrument. It will be noted by referring to fig. 117a that it is desirable to make the zero setting of the decrement scale at the point of half deflection and also to take the final reading at the point of half deflection, because at these points the resonance curve is steep, and consequently the settings are sharply defined and easily made. In this connection it will be noted that the formula MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 207 does not involve the resonant value of capacity, Cr, but only those at the points of half deflection where the slope of the resonance curve is steep. This formula is therefore the most desirable one to use, and the decremeter is consequently operated in accordance with it. In Fig. 117d a schematic diagram of the circuit is shown. / is a single- turn coil which may be connected in the circuit under test, as, for example, the aerial circuit of a radio transmitter. The inductance of this single Fig. 117a. — Kolster Decremeter. turn is, in the majority of practical cases, small as compared with the total inductance of the circuit under test, and therefore will not affect the tuning adjustment. The coil L is the inductance of the decremeter circuit and is so arranged that the mutual inductance between it and coil I can be easily varied. It is very essential that the degree of coupling between the circuit under test and the decremeter circuit be small. Cv is the variable condenser to which the decrement scale is attached through gears. In parallel with Cv is a small condenser Cf which remains fixed in value after proper adjustment. 208 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. H represents the hot-wire instrument or indicating device, the scale of which is so marked that the readings are proportional to the square of the current passing through it. A crystal dector D is provided and the wave length of distant stations may be measured by using telephone receivers T. Fig. 117b. — Diagram showing relation between decrement scale and resonance curve. By means of a switch, the buzzer circuit B B E may be connected to the instrument for calibration purposes. When persistent oscillations of single frequency are emitted from a radio transmitting station much more selective receiving apparatus may be employed with advantage at receiving stations, permitting sharp tuning with consequent minimizing of interference caused by stations other than those with which communication is desired. MANUAL OP RADIO TELEGRAPHY AND TELEPHONY. 209 Since the logarithmic decrement is a measure of the decay of a train of waves, it is desirable that this decrement be made as small as possible in Fig. 117c. order that a series of decaying trains of waves may approach as near as possible to the condition of persistent oscillations. A wave train having a H Fig. 117d. — Diagram of connections. logarithmic decrement of two-tenths, the limit set by the Federal regula- tions, will have 24 complete oscillations before the amplitude of the last wave has decreased to 1 per cent of that of the first. 14 Chapter IX. CARE AND OPERATIOK. 243. At all stations, ship and shore, the best results are invariably obtained and the most satisfactory service given by alert and careful operators who take pride in the condition of their instruments. Eadio instruments like all others depend for their efficiency on their condition and amply repay good care. Furthermore a neat and clean outfit inspires higher efficiency of personnel. An excellent operator once said that no matter how good he thought his contacts and connections were he always found that by going over them he could make them better and increase his sending and receiving efficiency. A routine, which, if followed, will ensure the proper care of a wireless set, is given in Appendix E. All sliding contacts, especially in receiver tuning coils, should be clean and bright and free from foreign matter. Sending key contacts should be kept clean and smooth and with faces parallel to each other. Detectors must be kept in their most sensitive condition and frequently tested by means of the buzzer furnished for the purpose. When using audion detectors, care must be taken not to use too much battery current which would shorten the life of the bulb or burn out the filament. When any part of the condenser is injured it should be immediately replaced or repaired. Any change in closed or open circuit without a corresponding change in the other throws the two circuits out of reso- nance and greatly decreases the sending radius. If the capacity in the condenser must be decreased for any cause then in order to retain the same wave length the inductance in the closed circuit must be increased. 244. The following general instructions apply to all stations: The operator shall wear the double head receiver continuously while on watch, with the detector adjusted to maximum sensibility and tuner adjusted to proper wave length. He shall satisfy himself by frequent testing with the buzzer that his detector is sensitive, and while in the vicinity of other vessels or near shore stations and using a detector that may be injured by strong sending, he shall always be alert to protect it by weakening the coupling or by opening the receiving switch. In order to avoid interference, he shall make a practice of loosening the coupling of his receiving set after hearing a call to reduce signals to a point where they are just clearly readable. He shall familiarize himself with all sending and receiving connec- tions and adjustments and be able to tell when they are correct and to MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 211 renew them when necessary ; but he shall not make any changes in any of them without the knowledge and permission of the chief electrician or operator in charge. He shall be capable of adjusting the spark gap, motor and generator rheostats and reactance regulator, so as to obtain the necessary output for the communication to be made. He shall use the least power that will enable his messages to be clearly read. He shall be vigilant in noting and keeping in good condition all sending condenser connections and in keeping all articles or instruments which might be injured or cause a ground or sparking well clear of the sending apparatus at all times. He shall not, except in cases of emergency, call or send any message, when official messages are being sent or received by other vessels or stations in his vicinity. He shall be careful to file correct copies, on the official forms, of all messages sent and received by him, initialing each and filling in time and place and other information as called for on forms. He shall avoid a short and jerky style of sending. Dots and dashes and intervals must be of proper relative lengths as shown by the code in order that the sending may be clear and legible. Operators must en- deavor to attain fair speed, both in sending and receiving. Where heavy static is encountered, dots and dashes may be longer, but must preserve their relative length. The generator shall be run only during the time necessary to send messages. Where a number of tunes are ordered to be used the operators shall be careful to see that all circuits are correctly adjusted before attempting to send. Each operator shall turn over all orders to his relief and also turn over a clean and neat station. A sending set with all connections good, closed and open circuits in resonance, no sparking from edge of condenser jars or plates, no glow from aerial and no sparking to rigging, is utilizing its power more efficiently and will be heard farther than the same set pushed to the limit but out of resonance witli high resistance connections and sparking at all points. jMessages shall not be sent during the interval in which naval radio stations send the time signal for the use of navigators in comparing chronometers or when broadcasting. The officer or electrician in charge is responsible for the routine of the station and for the instruction of his assistants in the proper use of the sending and receiving apparatus and that they understand and carry out all orders. He should stand watch sufficiently to keep himself expert in sending and receiving and, in any case, not less than 2 hours daily. 212 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. CALLING. 245. If a station called does not answer the call, repeated three times at intervals of 2 minutes, the call should not be resumed until after an interval of 15 minutes, the station making the call having first made sure that no other communications will be interfered with. Repeated and continuous calling is one of the principal sources of inter- ference. In a fleet, when a ship does not answer after a first call, it saves time and interference to shift tune rather than continue to call on the first tune used. This does not apply to calling merchant ships or commercial coast stations, which should be called on their normal wave length. SENDING. 246. When a ship is within ten miles of another which is receiving faint signals, the first ship should not attempt to send until the receiving ship has finished, unless she sends on a widely different wave length and even then she should not use more than 1 kilowatt. Ships in the same vicinity (within 20 miles) should not use more than 1 ampere in the aerial when communicating. When a distress signal is heard, all ships and stations hearing it should at once cease all radio work and not attempt to communicate, even with the vessel in distress, unless specially requested by that vessel to do so. The vessel in distress should make ,a sufficient number of times to quiet all radio work and should follow this by a broadcast message stating, (1) the name of the ship, (2) the station or ship it desires to communicate with, (3) the nature of her distress, (4) her approximate position, (5) by a general call (inquiry) for any ship or station to answer. If the station or ship called by the vessel in distress does not answer, the General Call may be answered by anyone within hearing. As soon as communication with the vessel in distress has been established, every other operator should preserve absolute silence. When the vessel in distress has finished its communications her operator should send a broadcast message to that effect, so that other ships may go ahead with their work without interfering. 247. The sending operator should not attempt to attain high speed unless he knows that the receiving operator is fully capable of receiving at high speed. Like continuous calling, repetition is one of the principal sources of interference. Steady sending, at the rate of about 20 words per minute, will give the best results. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 213 When receiving code or cipher, an operator should habitually hand- print the letters rather than write them in the ordinary long hand. If this is done fewer mistakes will occur in decoding or in repeating a message, especially if the decoding or repeating of a message is done by an operator other than the one who did the receiving, due to the difficulty of the latter reading the former's handwriting. This is especially true of the letters n and u, u and v, m and w, z and g, and a and o, which are often misread for one another in ordinary handwriting. The fact that code words are once repeated allows the receiving operator time to take this extra precaution. Code and cipher should be sent at a speed about one-third slower than plain language and great care should be taken to leave an appreciable space between the groups of letters and the interval character separating these groups, otherwise the interval character is apt to be confused as part of a code group. When hand-printing, the letters and D should be made carefully, lest the operator make a D which will be read as an 0, or vice versa, and similarly for the letters G and Q, U and V, and V and Y. Operators should frequently practice rapid hand-printing in order that they may become expert. Special care should be taken with the address and signature. DUPLEX OPERATION. 247a. A high-powered sending set can be operated continuously for sending if the receiving for the same station is done on a different antenna at a distant point (several miles away), in order that the sending will not interfere with the receiving done simultaneously on a slightly different wave. In such a case the operator at the receiving station uses a small sending key connected electrically with a relay at the sending station, which operates a solenoid whose armature carries a lever which acts as the sending key for the main sending set. A sending station operating by distant control from a near-by operating station in this manner generally carries on what is known in radio as "duplex" operation with a siuiilar station, such that both stations are continuously sending to one another. HIGH-SPEED OPEEATIOX. 247b. High-powered stations, carrying on continuous conminnication with one another, are sometimes equipped with apparatus such that the sending and receiving is done automatically, at speeds up to about 150 words per minute. To accomplish this the sending-key solenoid is operated electrically by means of a current controlled by a make and break contact maker which, 214 MANUAL OF RADIO TELEGRAPHY AND TELEPPIONY. in turn, is controlled by perforations of dots and dashes on tape, which latter is fed through a Morse writer; the dots and dashes being previously punched in the tape in the same manner as is done for cable operation. To keep the contacts of high-speed sending keys cool, and to prevent arcing across these contacts, a jet of cold air under pressure is played on the contacts. The receiving of high-speed signals is done by means of a phonograph, on the record of which the signals " talk." After the signals are recorded on a record the latter is run at a slow speed such that it repeats the signals to an operator at a speed such that the signals can be read. 248. TO SEND A MESSAGE Example. — Ship Prairie (NQM) calls coast station Key West (NAR): NAR acknowledges call: NQM sends message: NAR acknowledges receipt: Qnish. Call. 1. m ^H % m Attention signal. 2. NAR NAR NAR Station called. 3. ^ • • • De. 4. NQM NQM NQM Calling station. Reply, 5. Bi ^ m % ^H Attention signal. 6. NQM NQM NQM Calling station. 7. Hi • • • De. 8. NAR Acknowledging station. 9. g_ BB Go ahead. Message. 10. ■■ % m # IB Attention signal. 11. Alert Office of origin. 12. 5 Number of radiogram. 13. XN Sending operator's sign. 14. 7 Check. 15. Twelfth Original date of message. 16 ■■ • • • WM Break. 17. Govt. Secnav Washington Address. 18. ■■ • • • ■■ Break 19. Apache Mazanilla Phalara Text. 20. ■■ • • • WM Break. 21. Alert Signature. 22. # ■■ # ■■ • End of message 23. NQM Call letters of sending station. 24. ^m • WM Go ahead. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 215 Acknowledgment. 25. NQM Station which sent message. 26. # ■■ # Received. 27. 5 No. of message. 28. NAR Receiving station. 29. SP Receiving operator's sign. 30. ■■ • ■■ Go ahead. Finished Signal. 31. • • • WU • ■■ Finished. 32. NQM Sending station. RECEIVING. 249. In ordinary circumstances, while listening in, the set may be kept closely-coupled, to broaden the tune ; but the aerial circuit should be stiff, i. e., having a considerable amount of inductance (art. 231). The aerial circuit should be tuned with a variable condenser in series for short waves and in parallel (around inductance) for long waves (fig. 92). When the calling station is well tuned in, loosen the coupling, if there is interference. This should be done gradually, adjusting both the open (aerial) and closed circuits with each change of coupling, until a point is reached where signals are readable through the disturbances. With fairly strong signals the coupling on a 1 P. 76 receiver should be not less than 16 on the coupling scale, and, at this point, for moderate wave lengths, the signal should not fall materially in intensity. For further improvements in tuning, the closed circuit condenser should be made as large as possible and the closed circuit inductance correspond- ingly reduced. The settings (for best coupling and of open and closed circuits) for all stations habitually communicated with should be recorded and posted for the use of the operator on watch. The practice of loosening coupling while receiving should be made obligatory on all operators. It not only cuts out existing interference, but prepares for any interference which may arise during reception. Owing to the change of effective self-induction, in both circuits, both require readjustment (retuning) with each change of coupling. Two aerials in the same immediate vicinity as on board the same ship have an influence on each other so that if both are used for receiving at the same time, the tuning of either will affect the other and this effect may be observed between aerials on different ships if they are very close together. 250. When two ships are close together, the one which is not doing the sending can generally assist in receiving faint signals. Eegardless of opening of circuits, high power sending lessens the sensitiveness of the 216 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. detector of the sendintr station when, as usual, the transmitter and detector are only a few feet apart. Quite frequently in the tropics, it occurs that signals " swing " in and out (par. 175) ; that is, they are easily readable for a few seconds, and a few seconds later become hardly audible, and alternately grow strong and weak in this manner. Sometimes this can be overcome by a slight shift of wave length, but more often by allow- ing the spark gap frequent cooling periods. After once establishing communication, under such conditions, do not call again unnecessarily, but commence at once with the message, otherwise the gap will heat up before the message is well underway. It is often advisable to pause after, say, every 10 words, under such conditions, to listen for the receiving operator's OK, otherwise many repetitions may be required ; these pauses also assist in keeping the gap cool. INTERFERENCE. 251. The foregoing articles indicate specifically how to avoid interfer- ing with others and how to work through interference, but additional con- certed methods must be employed to avoid interference other than static. The mere establishment of standard wave lengths promotes interference, in one sense, since it ensures that there will be stations, ship and shore, in the same vicinity, wishing to communicate at the same time on the same wave length, but it conduces to safety. As between ship and shore stations, the latter controls and decides upon the order in which she will take and send (clear) messages. As between two shore stations ia adjacent countries, or the same country, both having business with ships, or between men-of-war of dif- ferent nationalities, a division of time is arranged. For shore stations communicating only with other shore stations, specific tunes (wave lengths), widely different from the standard, are assigned. As between merchant ships in the same vicinity and merchant ships and men-of-war, mutual forbearance and patience are absolutely necessary. It is often of value, in working through interference, to reduce the fre- quency of the note by reducing the speed of the motor generator below normal. This is especially valuable when several sets are operating in the same general vicinity. Shifting the sending coupling is another thing which will sometimes give the receiving operator a better opportunity to receive through inter- ference. In the latter case the antenna inductance must be re-adjusted to give maximum radiation. In fleets, standard calling tunes are established and a wide range of wave lengths (differing by a percentage sufficient to avoid interference iu ordinary circumstances) are assigned for communicating ; these times are known by their letters, as A, B, C, etc. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 217 The vessel called, when she acknowledges the call, designates the com- municating tune as C or D to be used by the sending vessel, so as to avoid interference with other tunes audible in the receiver of the vessel called. If two operators " lose " one another in an attempt to shift to a different tune to avoid interference, both should return to the original tune, and thus avoid delay and confusion. It should be borne in mind in this connection that, although any one of a wide range of wave lengths may be selected, the waves close to the fundamental carry much greater distances than do the extreme wave lengths. When all sending and receiving sets on all ships are calibrated and con- structed so as to permit easy, rapid and definite changes of wave length, while remaining properly coupled, calling and communicating tunes might be established and designated by international agreement and assist in relieving the existing congestion. STATIC. 252. The use of undamped oscillations will materially assist in the sharp tuning necessary to prevent interference by the use of standard calling wave lengths and codified standard communicating wave lengths ; but neither undamped nor damped oscillations can be relied upon to com- pletely eliminate the effects of the vagrant waves and local electrification grouped under the name of " static." Every lightning discharge produces powerful electric waves which affect conductors at great distances, and since thunderstorms in warm climates, and especially in summer, are almost continuous in the sense of existing somewhere in the area in which they affect detectors, the interference caused by them is almost continuous. The waves created by lightning discharges vary greatly in length ; but are highly damped and affect all aerials more or less. Again, at every wireless station the air at the top and foot of the aerial is at different potentials. The atmospheric potential gradient at any station varies with the time of day, the season of the year, and the local weather con- ditions. It is usually steeper in summer. This difference of potential tends to equalize itself through the aerial. The upper air is usually positively electrified, the earth negatively. The amount and regularity of the discharge to ground at any time depend on the difference of potential between the upper air and the ground at the time and the amount of electrified air which comes in contact with the aerial. The discharges are usually intermittent and vary in strength. Some- times they produce a continuous roar in the telephone. In this respect the note of the spark affects reception and it is possible to read a 500-cycle note through static which would render a 60-cycle 218 • MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. note unintelligible. As previously stated, the use of the heterodyne per- mits the note of received signals to be controlled at will. Whatever tends to selectivity or inertia in receiving circuits, such as large inductances, also tends to decrease static interference. Inductively coupled receiving sets afford a direct path to ground, so that static charges do not accumulate on the aerial, and the inductive coupling weakens the energy transfer of all induced currents which are out of tune. A silicon detector, connecting the aerial to ground above the receiving inductance, has been used with fair success at Key West for cutting out static* We see that loose coupling, small damping and high frequency, which we desire for other reasons, are also desirable as tending to eliminate static interference.f CODES. 253. Prior to July 1, 1913, the date set for putting the London conven- tion into effect, commercial shore stations in the United States, and United States coasting vessels used American Morse. All foreign stations, ship and shore, public and private, used Conti- nental Morse. American Morse is a little faster. The Continental Morse is a dash and dot code throughout with a maximum of four elements in any letter. The American Morse uses five elements in the letter P, four elements and a space in Y, Z and &, and a long dash for the letter L. It has a relatively less number of dashes than the Continental code and is on that account faster. A, B, D, E, G, II, I, K, M, N, S, T, U, V and W (fifteen out of the twenty-six letters of the alphabet) are the same in both codes. It is to be hoped that the use of wireless telegraphy will eventually bring about an international agreement as to the elements for the re- maining eleven letters and thus provide a universal code ashore and afloat. When it is desired to communicate by the international signal book (as between two vessels whose operators do not use the same language) the " call " should be followed by the letters P R B in the Continental code (art. 255). It is important that operators aboard ship learn, and become expert in, the American Morse code, since they must use this code at shore stations in handling land wires. * This method of cutting out static is proposed by Dr. Austin. t The British Association for the Advancement of Science has undertaken a systematic investigation of the phenomena grouped under the head of " static." It states that as far as is yet known the natural electric waves reaching wireless telegraph stations in latitudes above 50° north appear to travel mostly from the south. MANUAL OF KADIO TELEGRAPHY AND TELEPHONY. ' 219 The international signal of distress is • • • ■■ ■■ ■i • • • > making the letters S S of the Continental code (appendix C). The two signals given above were adopted nt the International Wire- less Telegraph Conference at Berlin in IDOG. 254. INTERNATIONAL MORSE CODE SIGNALS. (To be used exclusively for all radio communications.) Spacing and Length of Signals. 1. A dash is equal to 3 dots. 3. The space between two letters is 2. The space between the signals equal to 3 dots. which form the same letter is 4. The space between two words is equal to one dot. equal to 5 dots. LETTERS. A # 1^ U • • ^ B ■i • • • V • • • ^ C WM • WM • W • ■■ Hi D IB • • X WM • • 1^ E • Y WM • WM Hi F • • ■■ • Z WM ^M • • G ■■ ■■ • a (German) H • • • • • WM • H I • • a or a (Spanish-Scandinavian) J • ■■ ■■ §■ • ^M ■■ • ■■ K ^M • ^m ch (German) L • ^ • • ■i ■■ ■■ Hi M ^m WM 6 (French) N tm • • •§■•• ■1 IH ^M n (Spanish) P • ^M WM • ■1 ^M • ■■ ^m Q ■1 ^M • ■■ 6 (German) R • ^ • ■■ ■■ ■■ • S • • • ii (German) T "" NUMERALS. 1 • ■■ IB Hi ■! 6 ^ • • • • 2 ^ ^ • • • 3 • • • ■■ Bl ■1 ^M ^M • • 4 • •••■■ ■■ 1^ ■■ ■■ 4 5 ■1 ■■ m ^ ■ PUNCTUATION AND OTHER SIGNS Full stop (.) •• •# ## Semicolon ( ; ) WM # ^H # WM # Comma (,) •■■•■§•■■ Colon (:) ^^■■••9 220 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. Interrogation or REPEAT (?) • • tm ■■ • • Exclamation (!) ^M ■■ # # ■■ ■■ Apostrophe (') # ■■ MM !■ j^l • Hyphen or dash (-) Hi • • • • HI Bar indicating fraction (/) Bi # # ■■ # Parenthesis (before and after words) ( ) ■■ • ^M ■■ # ■■ Quotation marks (before and after each word or each passage quoted) (" ") # ■■ # # ■■ # Underline (before and after words or part of phrase) ( — ) • • ■■ ^B • ^M ATTENTION (or call) WM • ■■ • IB Double dash or BREAK (signal separating preamble from address, address from text, and text from signature) (=) ■■ • • • ■■ UNDERSTOOD • • • WM • ERROR • • • • • • # • GO AHEAD IH • ^ END OF MESSAGE • ■■ • ^ • WAIT • ^ • • • RECEIVED (acknowledgment of receipt of message) # ■■ • FINISHED (end of work) •••■^•liB In official repetitions and in the preamble of radiograms, figures may be rendered by means of the following signals, which may also be used in the text of telegrams containing figures only. Radiograms must, in this case, bear the service instructions " in figures." 2««H1 1 mM • • 9 With the adherence of the United States to the Berlin Convention, which requires all commercial ships flying our flag to use the International code, and the adoption of that code for communication between the army and navy, the necessity for using the American Morse code for any purpose except land line telegraphy has been eliminated. The International Morse code, alphabet, numerals, punctuation, etc., shall be used exclusively for all radio communications. In addition to the letters of the English alpha- bet, the following foreign letters may be used by ships : a (German) # ■■ • ^B d or a, (Spanish-Scandinavian) # ■■ Hi # ■ ch (German) ■■ ■■ WM ■ 6 (Spanish) # # ■■ • # fi (Spanish) MM ■■ • ■■ (German) IB ■■ IH # u (German) # # Bi Hi MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 221 AMERICAN MORSE CODE USED BY LAND-LINE COMPANIES. A # WM J ■■ • ■■ • s • • • B ^ • • • K ■■ • WM T ^ C • • • L ■■i u • • ■■ D ■1 # • M ■i ■■ V • • • Bi E N ^m • W • ^ IB F # ■■ • • • X • ■■ • « G ■■ ■ ■ • P • • • •• Y • • • • H • • • • Q • • Hi • z • • • • I • • R • • • 1 2 • • ■ ■ • • • NUMERALS. 6 7 A t k A A A A w f P V V V V 3 • • • ■■ • 8 ■i • • • • 4 • • • • 1 ■i 9 ■1 • • Hi 5 fH ■ 1 ■ 1 ^H PUNCTUATION AND OTHER SIGNS. Full Stop (.) • • Bi ^M • • Semicolon (;) ••• •• Comma ( , ) • ■■ • ■■ Interrogation or repeat (?) HH • • IHI • Exclamation (!) ■■ ■■ ■■ # Beginning of bracket ([) • • • • • ■■ # Ending of bracket (]) ##### ## ## Hyphen or dash (-) •••• •^•^ Dollar mark ($) ••• •■■•• Quotation (" ") 9 # ^M # ■■ # HAVE YOU ANYTHING FOR ME? ■■ 1^ ^ I AM BUSY IH • • • • ABBREVIATIONS. 255. The following abbreviated signals will go into effect with the London Convention, July 1^ 1913, and will be used by ships of all nations which may ratify that convention. ■i • H # ■■ ■■ # ^M (CQ) Signal of inquiry, or General Call, made by a station desiring to communicate. ■■ # ■■ # (TR) Signal preceding position report; or " Send position report." ■■ ■■ • • ^M ^M (!) Signal indicating that a station is about to send at high power. 232 \ MANUAL OF RADIO TELEGRAPHY AND TELEPHONY, Abbre- via- tion. Question. Answer or notice. PRB QRA QRB QRC QRD QRF QRG QRH QRJ QRK QRL QRM QRN QRO QRP QRQ QRS QRT QRU QRV QRW QRX QRY QRZ QSA QSB QSC QSD QSF QSG QSH QSJ QSK QSL QSM QSN QSO QSP QSQ QSR QST QSU QSV QSW QSY QSX Do you wish to communicate by means of the International Signal Code? What ship or coast station is that ? What is your distance ? What is your true bearing ? Where are you bound for 1 Where are you bound from? What line do you belong to ? What is your wave length in meters? How many words have you to send ? How do you receive me? Are you receiving badly? Shall I send 20 • ••■■• for adjustment ? Are you being interfered with? Have you much static ? Shall I increase power ? Shall I decrease power ? Shall I send faster? Shall I send slower? Shall I stop sending? Have you anything for me ? Are you ready? Are you busy ? Shall I stand by? When will be my turn ? Are my signals weak ? Are my signals strong ? Is my tone bad ? Is my spark bad ? Is my spacing bad ? What is your time ? Is transmission to be in alternate order or in series ? What rate shall I collect for ? Is the last radiogram canceled ? Did you get my receipt ? What is your true course ? Are you in communication with land ? Are you in communication with any ship or station (or, with )? Shall I inform that you are calling him? Is calling me? Will you forward the radiogram ? Have you received the general call ? Please call me when you have finished (or) at o'clock. Is public correspondence' being handled ? Shall I increase my spark frequency? Shall I send on a wave length of meters ? Shall I decrease my spark frequency? I wish to communicate by means of the Inter- national Signal Code. This is My distance is My true bearing is degrees. I am bound for I am bound from I belong to the Line. My wave length is meters. I have words to send. I am receiving well. I am receiving badly. Please send 20 • • • Hi • for adjustment I am being interfered with. There is much static. Increase power. Decrease power. Send faster. Send slower. Stop sending. I have nothing"for you. I am ready. All right now. I am busy (or : I am busy with ). Please do not interfere. Stand by. I will call you when required. Your turn will be No Your signals are weak. Your signals are strong. Your tone is bad. Y'our spark is bad. Y'our spacing is bad. My time is Transmission will be in alternate order. Transmission will be in series of 5 messages. Transmission will be in series of 10 messages Collect for The last radiogram is canceled. Please acknowledge. My true course is degrees. I am not in communication with land. I am in communication with (through ) Inform that I am calling him. You are being called by I will forward the radiogram. General call to all stations. Will call when I have finished. Public correspondence' is being handled. Please do not interfere. Increase your spark frequency. Let us change to the wave length of meters. Decrease your spark frequency. 'Public correspondence is any radio work handled on the commercial tunes 300 or 600. Ad(iitional abbreviation proposed for international use, and authorized for naval stations: Abbre- via- tion. Question. Answer or notice. QSZ Send each word twice. I have difficulty in receiving you. When an abbreviation is followed by a mark of interrogation, it refers to the question indicated for that abbreviation. MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. 223 EXAMPLES. Station A. QRA? =Wliat is the name of your ship or station? Station B. QEA Celtic MLC =This is the Celtic. Her call is MLC. Station A. QRG? =To what line do you belong? Station B. QRG White Star =1 belong to the White Star line? QRZ =Your signals are weak. Station A then increases the power of its transmitter and sends : Station A. QRK? =How are you receiving? Station B. QRK =1 am receiving well. QRB 80 =My distance is 80 nautical miles. QRC 62 =My true bearing is 62 degrees, etc. COMMERCIAL OPERATION BY UNITED STATES NAVAL COMMUNICATION SERVICE. By an Act of Congress, approved August 13, 1912, entitled, " An Act to Regulate Radio Communication," the Secretary of the Navy was directed to open certain naval radio stations to general public service. This service involves the handling of commercial traffic to and from all ships at sea, and between certain fixed points on shore. The term " commercial," as applied to radiograms, is used to denote all messages other than of official nature. In addition to the coastal stations open to this service, all vessels of the navy are directed to handle com- mercial traffic for convenience of the officers, crew and the public. Arrangements have been made Avith the principal land line and cable com- panies, as well as the radio operating companies, for the interchanging of this traffic, so that whether it be land or sea, all telegraph facilities are linked with the Naval Communication Service. These arrangements include the handling of commercial traffic in and with Alaska. During interruptions to the Army Cable System, the work is taken up by the radio circuit. In the United States, during interruptions to the land line service, the Naval Communication Service fills in the breaks, carrying the traffic across these breaks and acting as an adjunct to the telegraph service when all other means of communication fail. The accounting in connection with the handling of commercial traffic is centralized into the office of the Director Naval Communications, and through this service the traffic is carefully checked and payments made by that office to the connecting lines. SOURCES OF INFORMATION. 256. The act regulating apparatus and operators on merchant steamers, which has been in effect since July 1, 1911, will be found in Appendix D. The act regulating licenses, wave lengths, decrement, etc., and author- izing the Secretary of the Navy to open certain naval wireless stations to 224 MANUAL OF RADIO TELEGRAPHY AND TELEPHONY. commercial business and to fix rates for this service, will also be found in Appendix D. This act went into effect December 13, 1912. The most important rules of the London Convention, which went into effect on July 1, 1913, will be found in Appendix C. The United States having ratified the London Convention, the operating rules provided by that convention are incorporated in the extracts from " Commercial Traffic Eegulations," U. S. Kaval Communication Service, printed in Appendix B. The conventional abbreviation signals, authorized by the London Con- vention, and which are used by naval shore stations when practicable, will be found in this chapter, after the code. (Art. 255.) Information — relative to naval shore stations open to commercial busi- ness, lists of such stations, message rates, names of stations and hours of sending out time signals for the use of navigators in comparing chronom- eters, weather reports and storm warnings — is issued in " Notices to Mariners " and shown in pilot charts published by the U. S. Naval Hydrographic Office. In accordance with the Berlin Convention, all stations have three call letters. Groups of call letters are assigned to nations for their exclusive use by the Bureau of the International Telegraph Union at Berne, Switzer- land. (All international cable accounts are settled through this Bureau.) Specific call letters from these groups are assigned to each ship and shore stations by the governments of the respective countries. To the United States has been assigned all combinations of letters beginning with N and W and the combinations KIA to KZZ. A list of call letters will be found in " Wireless Telegraph Stations of the World," published by the Department of Commerce. It can be obtained from the Superintendent of Public Documents at Washington. This will eventually be an international publication issued by the Berne office, Eules governing the licensing of commercial and private stations and operators are issued by the Department of Commerce and inspections to determine their compliance with the laws printed in Appendix D and with the London Convention, are made by that department. All matters pertaining to the operation of U. S. Naval Eadio stations, both high-powered and coastal stations, and their relations with com- mercial stations afloat and ashore, in the United States and foreign countries, are under the supervision of the Director Naval Communica- tions, Navy Department, Washington. APPENDICES. NOTE 1. The following list of metals is arranged in such order that any one will be the positive pole of the battery when used with the metal next below it on the list as a battery element and the negative pole when used with the element next above it, the difference of potential between any two being greater the farther apart they are in the series. Carbon, Silver. Lead. Zinc. Platinum. Copper. Cadmium. Magnesium, Gold. Iron. Tin. Sodium. The amount of potential difference also depends on the battery solution, and in some instances it may be reversed. Commercial primary batteries are of copper and zinc, with an E. M. F. of approximately 1 volt, and carbon and zinc, with an E. M. F. varying from 1.4 in LeclanchS cells and some dry cells to 2.1 in some types of wet cells, depending on the electrolyte. NOTE 2. The relations existing between electricity and matter have been most ex- haustively investigated by Prof. J. J. Thomson, who has proved that electric- ity has an atomic structure and that it can exist separately from an atom of matter. When a current is sent through a vacuum tube, the luminous beam pro- ceeding from the cathode has been shown to consist of particles projected from the cathode. These particles are capable of turning a small wheel. The cathode beam can be deflected by either a magnetic or an electric field, and it is found to consist of particles of negative electricity or of parts of the atom negatively charged, each having about one eighteen-hundredth of the mass of an atom of hydrogen. These particles are the same, no matter what gas is used in the vacuum tube. They are usually called electrons. "When an electron is broken off from an atom, the remaining part is positively charged. Currents of elec- tricity, however produced, are the result of the decomposition of atoms into positive and negative electric charges. There can be no electric current without movement of electrons. Conductors are bodies in which the break- ing up of atoms and movements of electrons take place more or less easily. Some free electrons exist in all bodies. It is by setting these into vibration and by means of this vibration making them break off similar particles from neighboring atoms, and thus propagate the disturbance throughout the mass of the conductor, that electric currents are generated. 15 326 APPENDICES. APPENDIX A. TABLE 1. [Extract from Fleming's Cantor lecture. Journal of Society of Arts, p, 1S6, January 5, 1906. Taken mostly from A. Heydweiller, " On Spark Poten- tials." Ann. der Physik, vol. 248, p. 235 (1898).] Spark Voltage Between Brass Balls 2 Centimeters in Diameter for Various Spark Lengths. Spark length (cms.). Spark voltag«3. 0.1 4,700 0.2 8,100 0.3 11,400 0.4 14,500 0.5 17,500 0.6 20,400 0.7 23,250 0.8 26,100 0.9 28,800 Spark length (cms.). Spark voltage. 1 31,300 1.5 40,300 2 47,400 2.5 53,000 3 57,500 3.5 61,100 4 64,200 4.5 67,200 5 69,800 TABLE lA. Material. Specific inductive capacity. Air 1 Hard rubber 2.29 India rubber 2.10 Mica 6.64 Micanite Typewriter linen paper Paraffin oil 2.71 Glass (crown) 6.96 Glass (plate) 8.45 Glass (ligbt flint) 6.72 Glass (extra dense flint) 9.86 Porcelain 4.38 Shellac 3.10 • Per millimeter for thicknesses up to 1 millimeter. 2 Per centimeter. Dielectric strength. Volts. ( M,500 ( « 3,000 • 40,000 » 30,000 » 60.000 • 40,000 • 45,000 • 7,000 20,000 r 9, tl6. 9,000 000 3 Per millimeter * Approximate. APPENDICES. 227 TABLE 2. Condenser Capacity Required to Give Full 1,000,000 TABLE 5. LoGABiTHiiic Decrement (5) of Wave Train and the Approximate Number OF Waves (N.) in the Train Before the Amplitude Falls to One-Tenth of the Maximum. i N 1.0 3.5 0.8 4.0 .6 6.0 .4 7.0 .3 8.5 .8 12.5 8 V .1 24.0 .08 30.0 .06 39.0 .04 58.0 .03 78.0 .02 116.0 Good tuning is not possible with less than fifteen waves in the train. 228 APPENDICES. TABLE 6. Some Common Units Expressed in Terms of Absolute Units, 1 microfarad = 1 . 10-ib c. g. s. 1 millihenry = 1 . 108 1 microhenry = 1 . lOs 1 volt = 1 . 108 1 ohm = 1 . 10» 1 ampere = 1 . lO-i 1 watt = 1 . 107 TABLE 7. Some Common Hiqh-Frequenct Equations. The time of oscillation of a condenser circuit is r= 27r »y LC seconds. (1/ in henries, in farads.) V = n'/. and T = — , where v is the velocity, n the frequency, and X the wave length. The wave length Is therefore A = 1.885 s/Za .109 meters. In a condenser charged ?^ times per second the energy passing through In one second is P = I -y^ N watts. (G in microfarads and F in volts.) The damping of a single circuit is (R In ohms and L in henries or both in absolute units.) The damping of two circuits by the resonance method. '^^'^--^ -^;;ry w-r^' The following equation and tables are the results of experiments conducted between Brant Rock station and the cruisers Salem and Birmingham In 1909-10. See " Some Quantitative Experiments in Long Distance Radio* Telegraphy," by L. W. Austin, Reprint No. 159, from Bulletin Bu. of Stand- ards, Vol. 7, No. 3, Feb. 1, 1911. APPENDICES. 229 Equation /„ = 4.25 X /o X hi Ag ad I = Antenna current, sending, in amperes. Zg = Antenna current, receiving, amperes through 25 ohms. h, = Height of flat-top antenna, sending station. In kilometers. h, — Height of flat-top antenna, receiving station, In kilometers. n = .0015. d = Distance in kilometers, / = Wave length In kilometers, e = 2,7183. 25 ohms = high-frequency resistance of ship aerial of 1000-meter wave length. The above equation covers the normal-day received current over salt water, through 25 ohms for two stations with flat-top aerials of any height, with any value of sending current and any wave length, provided the sending station is so coupled as to give but one wave length. The following tables (8, 9, 10 and 11, 12) illustrate the application of this equation : TABLE 8. For good communication received current should be equal to i«=40 X 10-« amperes through 25 ohms = 40 X lO-*" watts = * erg per second. For audible signals 7^=10 X lO-s amperes through 25 ohms = 2,5 X 10-» watts =jV6rg per second. TABLE 9. Calculated Relation between Antenna Current and Distance for Two Ships with Antenna Heights of 130 Feet, A = 1000 m. Antenna Current Working Distance 40.10-* amp. Extreme Distance of Audibility 10.10-«amp, Is. Day. Night, (Zero Absorption) Day. Night. (Zero Absorption) 1 amp. 76 miles 90 miles 200 miles 360 miles 3 135 180 300 720 3 180 270 375 1080 6 235 460 475 1800 7 280 630 550 2520 . 10 346 900 6.30 3600 16 420 1360 726 6400 20 475 1800 790 7200 26 535 2260 840 9000 30 666 2700 900 10800 40 630 3600 970 14400 60 686 4600 1026 18000 60 726 6400 1150 21600 230 APPENDICES. TABLE 10. Good Working Distance and Sending Current for Two Stations with Flat-Top Antennas 450 Feet High. Nautical Miles. A = 1000 m. A = 3500 m. A = 3760 m. A = 6000 m. 1000 15 amp. 13.5 amp. 15 amp. 17 amp. 1250 38 27 27 30 1500 91 49 44 46 1760 200 95 77 74 2000 490 155 122 105 2260 246 200 160 2600 470 314 236 2750 500 336 3000 ... 775 600 APPENDICES. 231 CO en S 00 t- o ^ g ?,' 00 lO IM OS «- '"' " " " «c 03 CO ^_^ S ?§ s «o CO CO e» a « ^ ,_, f^-* tn CO CO tl 00 l-H 05 pa 10 CO ffi '"' ^ »< (M II II •* ^-t 00 CO CO ■# -< ^i^ s js CO CO 5 00 s ?i -** U3 5 §§ s C-l t- ■* (M to (Y) la 00 -H g B CO -H 00 >C OS fl Si Oh <-> m c? t- CO ,—4 II ^ CO 00 Wta i kn ,— , •« g e< 00 «n ;* ■* -< bj 2? ?i f« ;?; ■* ©1 t- so CO '^ OS ■* C5 CO ffi a « s 3 (— , j^ <-) V 03 e^ ->j< to CO ,^ ^ 2 00 CO ou S II ©I II w s 10 ,« 5 ?s < W|-o s - rrt 10 t- -* (M <-) S8 ?g ^ -* w cs CO II -< II Wl-a g § g i 8 g 11 10 5f? s g IM g ?2 g OS CO g a ci. a: g 8 S g «? to 10 III t- C'l II tr 10 «o •^ 11 W'ts g i S 8 8 s 2 y 05 f? 8 g t- 10 « •5 a ?i s 8 g g 8 8 8 8 8 g g g B s? '"' '"' '^ M ei CO Z "X « g 00 '« o> ee •