LIBRARY 
 
 9 
 
 UNIVERSITY OF CALIFORNIA. 
 
 Received. j^____4e 4&^2 > 18 . 
 
 A c cessions No*&- tZlafjJi- Shelf No 
 
QUADRUPL 
 
 X 
 
 BY 
 
 WM. MAVER, JB., AND MINOR M. DAVIS. 
 i 
 
 WITH CHAPTERS ON 
 
 The Dynamo-Electric Machine in Relation to the Quadru- 
 plex. The Practical Working of the Quadruplex. 
 Telegraph Repeaters, and the Wheat- 
 stone Automatic Telegraph. 
 
 BY WM. MAVER, JR. 
 
 NEW YORK: 
 THE W. J. JOHNSTON CO., LIMITED, 
 
 TIMES BUILDING, 
 
 1890. 
 

 
 Copyright, 1884, by "W. J. JOHNSTON. 
 
PREFACE. 
 
 WHEN the authors of the following description of the Edison 
 quadruplex system of telegraphy undertook to write a clear and 
 complete explanation of that system, and of the instruments con- 
 nected therewith, they felt convinced that in doing so they would 
 meet with encouragement from many of their fellow operators, 
 especially those who had not had a favorable opportunity to study 
 the theory or to investigate the practical working of the system. 
 
 They were aware that among the employes of the various tele- 
 graph companies there were many who were and are deterred from 
 gaining a closer acquaintance; with; the more complicated systems of 
 telegraphy, for the reason that the text books are, as a rule, beyond 
 their comprehension, and, therefore, the information they are in 
 search of can only be obtained by persistently questioning those 
 who are better informed than themselves, a method requiring much 
 patience, and not always an agreeable one. 
 
 To meet the wants of this class was their principal object, and 
 they have been pleased to find that, judging by the reception which 
 greeted the appearance of the articles in the pages of The Operator, 
 their efforts have been successful. 
 
 In the following description of the quadruplex, as well as in the 
 other articles contained in this book, no attempt has been made 
 to display erudition. On the contrary, it has been the aim of the 
 writers, if they erred at all, to err on the side of simplicity. 
 
 It is a book written by operators for operators, and numerous 
 
4 PREFA CE. 
 
 assurances have been received, that, as such, it has been useful. 
 But while such is the case, they also hope, and have reason to be- 
 lieve, that it will be found an easy source of information to those 
 outside of the telegraphic fraternity, who may desire to form an 
 acquaintance with these various systems and instruments. 
 
 The authors take this opportunity to return acknowledgments 
 to Mr. C. O. Mailloux for valuable hints as to the construction of 
 the diagrams. The majority of the diagrams were prepared for 
 the engraver by the same gentleman. 
 
 In conclusion, the writers hope that their book will prove to be 
 free from serious blunders, and that all who find profit in reading 
 what is herein written may be induced thereby to seek to increase 
 their knowledge of electrical matters by a study of more elaborate 
 works on such subjects. 
 
CONTENTS. 
 
 PAGE 
 The Quadruplex. 
 
 Chapter 1 Development of the Quadruplex, - 7 
 
 " 2 Introduction and Explanatory, 14 
 
 3 The Transmitter, Rheostat, and Condenser, - - - 18 
 
 " 4 Stearns Duplex, 25 
 
 " 5 Instruments of the Polar Duplex, 29 
 
 6 The Polar Duplex, 35 
 
 7_The Quadruplex, ... 40 
 g The Dynamo-Electric Machine in relation to The 
 
 Quadruplex, 55 
 
 9 The Practical Working of The Quadruplex, - - - 73 
 
 Telegraph Repeaters, .-- 86 
 
 The Wheatstone Automatic Telegraph, --- 110 
 
hfl 
 
THE QUADRUPLEX. 
 
 CHAPTER I. 
 
 DEVELOPMENT OF THE QUADRUPLEX. 
 
 In 1874, Mr. Thomas A. Edison succeeded in getting his quadru- 
 plex into such shape that it was considered fit to be used on wires 
 for the transaction of commercial business. 
 
 The first quadruplex circuit on which business was actually 
 transmitted, was between New York and Boston. At this time the 
 quadruplex was worked on what is known as the bridge system, as 
 seen in Fig. 1, that is, the receiving instruments were placed in the 
 arm usually allotted to the galvanometer in the Wheatstone bridge, 
 and of course were operated by the excess of current passing 
 through that arm precisely as a galvanometer would be. 
 
 At this time the instruments were not wound differentially. 
 
 The No. 2 relay was then, as now, a neutral relay, with a soft 
 iron armature, but with an ordinary long core. 
 
 At that early stage of the history of the quadruplex system, it 
 cannot be said that it was a complete practical success. 
 
 The principal difficulty appears to have been caused by the un- 
 steadiness of No. 2 relay, and this was attributed to the falling back 
 of the armature of the No. 2 relay at the moment when there was 
 no battery at the distant end, namely, while the change of polarity 
 was taking place ; but, in the light of subsequent experience, it may 
 perhaps be questioned whether this was really the case. 
 
 To assist in obviating this trouble, the armature lever of that re- 
 lay was made very long, and at its other end it was furnished with 
 an armature placed opposite another electro-magnet, which electro- 
 magnet was in a shunt circuit in which was also placed a condenser. 
 
8 
 
 THE QUADRITPLEX. 
 
 This condenser, at the moment of change of polarity at distant end, 
 discharged into the shunt circuit, and, by thus magnetizing the extra 
 electro-magnet, held the lever firmly against the contact point for 
 the desired time. 
 
 In 1875 a change was made in the foregoing arrangement, the 
 same principles being retained, but the differential method was 
 adopted and both Nos. 1 and 2 relays were polarized. 
 
 The armature of the No. 2 relay of course responded to all changes 
 of polarity, but its contact points were so arranged that when the 
 full strength of current was to the line, the retractile springs were 
 overcome and the local circuit broken as seen at R 2 , Fig. 2. 
 
 Fig- 2. 
 
 The next alteration, shown in Fig. 3, consisted of a novel arrange- 
 ment of the circuit, which did not require the polarity to be reversed 
 during the transmission of a signal. The diagram shows the con- 
 
THE QUADRUPLEX. 9 
 
 nections necessary for working this device on the diplex system, 
 that is, for sending two messages in the same direction simulta- 
 neously. By the adoption of any of the duplex systems, this method 
 can be converted into the quadruplex system. Mr. Gerritt Smith 
 was the inventor of this ingenious arrangement. 
 
 Fig 3. 
 
 Following this came a supposed improvement of the original 
 arrangements of the quadruples (Fig. 4), which, as may be seen, is a 
 combination of the bridge and differential methods. The No. 1 relay 
 was placed in the bridge, and the No. 2 relay (now transformed 
 into a compound polarized relay, also an invention of Mr. Gerritt 
 Smith,) was placed in the line and the rheostat wires, as at present. 
 
10 
 
 As in the former case the No. 2 relay, Sr in the figure, responded 
 to either polarity, but only broke the local circuit when full battery 
 was to the line at distant end. Subsequently both relays were 
 again placed differentially, a polarized relay, for increase and de- 
 crease of strength, still being retained. (Fig. 5.) 
 
 LIN t 
 
 GROUND 
 
 Fig. 4. 
 
 About this time a condenser was placed between the two relays, 
 one series of its plates being connected to the rheostat wire, the 
 other series to the line wire, which was found to be of service. 
 This arrangement was devised by Mr. F. W. Jones, at that time in 
 Chicago. 
 
 About this time also a combination relay, shown in Fig. 6, capable 
 of performing the service of the present No. 1 and No. 2 relays was 
 
THE qiTADRUPLEX. 
 
 11 
 
 introduced. It may be seen by following out the local connections 
 that the armature will give the correct signal for increase and de- 
 crease of strength, regardless of the side on which it may be giving 
 a polarity signal. This device worked successfully, but it was found 
 
 Fig. 5. 
 
 more desirable to have two separate relays so that each receiving 
 operator could have control of the adjustment of his own instru- 
 ment. 
 
 Fig. 7 shows the stage at which the system had arrived in 1878, 
 four years after its first appearance. 
 
 At that time the compound polarized relay was still used as the 
 
12 
 
 THE QUADRUPLEX. 
 
 No. 2 relay. The local contact points, as heretofore, were so 
 arranged as to close the local circuit only when full battery at dis- 
 tant end was to the line. 
 
 Then followed the introduction of the short core No. 2 relay, upon 
 the advent of which a marked improvement in the working of the 
 system was discernible, and not only in the working of the system, 
 but in the facility with which the No. 2 relay could be adjusted. 
 
 c 
 
 Since that time no further improvements have been added to the 
 system. Indeed, it is difficult to imagine in what respects it could 
 now be improved. In fair weather and under good conditions as 
 regards management, etc., the system approaches very near perfec- 
 tion. 
 
 The continuity preserving transmitter has retained its original 
 shape throughout, but the pole changer has seen about as many 
 changes in the form of its construction as there have been methods 
 
THE QUADRUPLEX. 
 
 13 
 
 employed in the endeavor to bring the quadruplex system to its 
 present shape. 
 
 We are mainly indebted to Prescott's description of the quadru- 
 plex for the foregoing information. 
 
II. 
 
 INTRODUCTORY AND EXPLANATORY. 
 
 It is purposed in these articles to present a description of the 
 Edison or Western Union standard quadruplex in such a manner 
 as will be easily understood by every one. It is not expected that 
 the explanation can be made so plain that it will be understood 
 without study on the part of the reader ; but it is hoped that the 
 amount of application necessary will be reduced to a minimum by 
 the use of simple terms and lucid illustrations. 
 
 This quadruplex is, with some slight changes, virtually a combi- 
 nation of the Stearns and polar duplexes. If, therefore, we ac- 
 
 Fig. 8. 
 
 quire a distinct knowledge of the theories of these duplexes, it will 
 be comparatively an easy task to comprehend the manner of their 
 combination in the quadruplex. 
 
 The instruments used in the working of these duplexes are 
 similar to those required in the quadruplex, namely, the single-cur- 
 rent transmitter and neutral relay of the Stearns system, and the 
 pole changer and polarized relay of the polar. 
 
 The function of the single-current transmitter is to increase and 
 
THE QUADRUPLEX. 
 
 15 
 
 decrease the strength of current on the line : that of the pole 
 changer is to change the direction of the current. 
 
 The office of the neutral relay is to respond to the increase or 
 decrease of current on the line, while that of the polarized relay is 
 to respond to the changes of polarity of the current. Both relays 
 are wound differentially, as will be explained. 
 
 Here are two distinct principles which, when properly combined, 
 give us the quadruplex. 
 
 We will proceed to inspect the manner of working the Stearns du- 
 plex, but, before doing so, it will be necessary to examine some of 
 the laws of electricity and magnetism on which its operation is based. 
 
 Fig. 9. 
 
 If electricity be passed through a coil of wire encircling a bar of 
 soft iron, the bar will become a magnet, the polarity of which will 
 be determined by the direction in which the current passes around 
 the bar. The iron will, however, lose its magnetism the moment 
 the current ceases. If the wire be wound around the bar, as shown 
 in Fig. 8, and if the current be passing in the direction indicated 
 by the arrow S, the poles will be as indicated in the figure. 
 
 Winding the wire in the opposite direction around the bar, 
 the current retaining the same direction, reverses the polarity of 
 the magnet, as shown in Fig. 9. 
 
 It follows that if two equal currents are caused to pass around 
 the bar in opposite directions, tending to establish exactly opposite 
 
16 
 
 THE QITADRUPLEX. 
 
 effects, they will neutralize each other, and the bar will remain non- 
 magnetic. 
 
 If we bend the bar to the form of a letter U, both poles of the 
 magnet will exert whatever attractive force is developed, upon an 
 unmagnetized cross piece, or armature, as may* be seen in ordinary 
 relay and sounder magnet. 
 
 If we connect one pole of a battery with the earth and from the 
 other pole run to the earth two wires of equal resistance (Fig. 10), 
 equal currents will traverse the wires. 
 
 Fig. 10. 
 
 These principles are employed in the Stearns differential system 
 in the following manner : Suppose a wire A (Fig. 11) is led from 
 a battery around a bar of soft iron from left to right, over the bar, 
 and another wire of equal resistance B around from right to left, 
 two equal opposing actions will be set up in the bar, one exactly 
 offsetting the other, and no magnetic effect will be produced. A 
 relay thus wound is called a differential relay. 
 
THE QUADRUPLEX. 
 
 17 
 
 Now, suppose a battery is introduced at the distant end of B 
 (Fig. 12), and note the result. 
 
 Battery C supplies each wire with an equal current, having a 
 
 strength of, say, 2. Battery D adds to the strength of the current 
 
 traversing wire B an additional strength of 2, making a strength of 
 
 4 upon wire J5, a surplus of 2 above that on A, which is sufficient 
 
 B 
 
 Fig. 12. 
 
 to neutralize the magnetism which the current in wire A would 
 tend to set up, and to establish, by means of its excess of strength, 
 a magnet, the polarity of which is determined, as in the previous 
 cases, by the direction of the current around the bar. 
 
 Let us take an illustration frequently used. Suppose a wheel 
 (Fig. 13) upon the opposite sides of which two streams of water 
 
 Fig. 13. 
 
 are falling with equal force. So long as these forces continue equal 
 the wheel does not move, but upon increasing the pressure on one 
 side, the force acting upon the other will be overcome, and the 
 wheel will revolve. In a similar manner the excess of electrical 
 force in one of the coils overcomes the opposing force in the other. 
 
CHARTKR III. 
 THE TRANSMITTER, RHEOSTAT AND CONDENSER. 
 
 In order to produce rapidly and conveniently the magnetic effect 
 described, instruments specially adapted to the work are used. 
 Let us first examine the transmitter. 
 
 THE TRANSMITTER. 
 
 In the Stearns duplex, the office of the transmitter, or continuity 
 preserving key, is to alternately place the battery and the earth in 
 contact with the line, substituting one the instant the other is re- 
 moved ; thus preserving, with the exception of a slight interval 
 noted further on, a continuous circuit at all times. 
 
 A (Fig. 14) is an electro-magnet controlling, by means of local 
 battery L B and key K, the action of armature lever B. is a thin 
 tongue of elastic metal attached to the lever 5, but insulated from it. 
 The lever B is bent at one end, as shown, and the tongue extends 
 just below the arm of the bend in such a position as to come in con- 
 tact with it at the point X. The post D is insulated from the base 
 upon which it stands, and supports the screw 6r. This screw is so 
 adjusted as to intercept the upward movement of the tongue when 
 the transmitter is closed, establishing contact at E and breaking it 
 at X. Now, if the battery be connected to D, the ground to j5, and 
 the line to the tongue C", the instant we close the transmitter, screw 
 Cr places the battery in contact with the line at E, while it pushes 
 the tongue downward at JT, breaking contact at that point. The 
 screw S regulates the play of the armature lever. The relay used 
 in the Stearns duplex is differentially wound, but is otherwise con- 
 structed like those in ordinary use. 
 
THE QITADRUPLEX. 
 THE RHEOSTAT. 
 
 19 
 
 In the systems of duplexes and quadruplex which we are 
 studying, we have two wires wound around the cores of electro- 
 magnets in opposite directions, so that the currents from the home 
 battery have no effect upon the cores, as one current neutralizes the 
 other ; in other words, one wire balances the other. 
 
 By placing another battery at the distant end of one of these 
 wires, however, we have seen that we can diminish or increase the 
 
 Line 
 
 |crouiui[ [flrouii(t| 
 
 Fig. 14. THE TRANSMITTER. 
 
 strength of the current passing through that wire so that the equi- 
 librium which the home battery tends to maintain in the cores is 
 overcome, and we can thus operate tlie home electro-magnets from 
 the distant end, at will. Now, as the object of the multiplex sys- 
 tems is to double or quadruple the capacity of the existing wires 
 between any two places, it is evident that one of the wires which 
 we use to preserve the equilibrium in the electro-magnets must be 
 comparatively of no expense, or nothing will be gained. For in- 
 stance, if we desired to set up a duplex between any two points, 
 
20 THE QUADRITPLEX. 
 
 and were to use two of the wires already running to those points, 
 one wire to balance the other, it is manifest that, although we 
 should literally make one wire do the work of two, we should act- 
 ually still be using two wires to accomplish the feat. 
 
 It is a law of electricity that with a given electro-motive force the 
 strength of current will depend on the resistance of the conductors 
 through which the electricity is required to pass. 
 
 It is not necessary that the two wires should be of the same 
 length, size, or material. Provided the resistance of the conductors 
 is the same, the strength of current will be equal in each. Certain 
 metals offer a much less resistance to electricity than others. Thus 
 the capacity of iron wire as a conductor is more than twice that of 
 German silver. This is supposing the metals to be of equal thick- 
 ness. The thicker we make a metal, the more we increase its 
 capacity as a conductor. Conversely, the thinner we make a metal 
 the poorer a conductor it becomes ; in other words, its resistance is 
 greater. So it is possible, by using iron of a certain thickness for 
 one of the wires under consideration, and for the other a very much 
 thinner German silver wire, to make, say, 500 miles of iron wire 
 equal in point of resistance to perhaps but half a mile of German 
 silver wire, and a given battery will produce as great a strength of 
 current in the iron wire as in the much shorter German silver wire. 
 This fact, then, is taken advantage of in these multiplex systems, 
 and an artificial wire consisting of coils of fine German silver con- 
 tained in a box termed a rheostat is used as a balancing wire. 
 
 Fig. 15 shows the manner in which the coils are connected to the 
 brass discs on the cover of the box and to each other. A brass 
 plug inserted between any two of the discs virtually cuts out one 
 of the coils, as it provides a route of practically no resistance for 
 the current. The coils are wound double upon themselves, as 
 shown in the figure, which prevents any inductive effects, the law 
 of electricity bearing on this point being " that a circuit doubled 
 back upon itself so that the current flows back along a path close 
 to itself exerts no force upon external points.'* 
 
 Thus we have a simple means of obtaining a wire which we can 
 
THE QUADEUPLEX. 
 
 21 
 
 adjust to correspond with a line wire of almost any resistance. If, 
 for example, we have a wire, as in the figure, 3,000 ohms resist- 
 ance, we simply unplug as many coils in the rheostat as may be 
 necessary to bring the artificial wire up to the same resistance, and 
 if by reason of wet weather the resistance of the line wire de- 
 creases, we reduce the resistance of the rheostat by inserting more 
 plugs between the discs ; for, if we do not preserve the balance be- 
 
 400 300 200 
 
 Fig. 15. 
 
 tween the two wires, the home battery will operate the home relays 
 and thus interfere with the signals from the distant end. The fig- 
 ures on the brass discs represent the amount of the resistance in 
 ohms of the coils attached to them. 
 
 THE CONDENSER. 
 
 As without the use of condensers or some other instrument capa- 
 ble of performing a similar service, it would be impossible, for rea- 
 sons which will be explained hereafter, to successfully work the 
 multiplex systems on lines of any great length, it will be proper to 
 devote some space to a description of their action and manner of 
 construction : 
 
 It is a well known law of both electricity and magnetism that like 
 poles repel and unlike poles attract each other. 
 
22 
 
 THE QITADRUPLEX. 
 
 If by any means we electrify any insulated conductor with elec- 
 tricity of one polarity, that electricity will tend to attract toward 
 itself in surrounding objects sufficient electricity of opposite polarity 
 to establish equilibrium, or what is termed a non-electric state. 
 
 Fig. 16. 
 
 Electricity thus attracted in surrounding conducting objects is 
 called induced electricity, and this property of inducing electricity 
 in other objects is termed induction. All substances are not 
 equally susceptible to inductive influence. Good conductors of 
 electricity generally are, while insulators are not. In Fig. 16, A 
 and B are supposed to be insulated metal spheres remote from other 
 conducting bodies. If A be charged with positive electricity from 
 an electrical machine, it will at once receive all the electricity which 
 it will apparently hold. If B be now brought close to A, it will be 
 found that A will receive more electricity, because the positive 
 electricity on A has attracted a corresponding quantity of negative 
 electricity on the side of B nearest A, and has repelled an equal 
 amount of positive electricity to the opposite side of J9, and in the 
 
THE QUADRUPLED. 
 
 23 
 
 effort of the two polarities to unite they have become gathered or 
 condensed upon the sides of the spheres nearest each other. If we 
 now connect B with the earth, thus allowing the positive electricity 
 on B to escape, it will be found that A will receive a still further 
 charge of positive electricity, and B will take up a corresponding 
 amount of negative electricity. So long as A is kept electrically 
 charged, the electricity which has been induced in B will be held 
 
 Lino 
 
 Condenser 
 
 Fig. 17. 
 
 in .B, but the instant we provide a conductor to the earth for A, the 
 induced electricity in B will simultaneously flow to the earth. 
 
 We cannot interpose any known substance between the spheres 
 which will prevent the above described inductive action from tak- 
 ing place. The least inductive effect is produced when the plates 
 are separated merely by air ; the greatest when they are separated 
 or insulated by mica. The closer the plates can be brought with- 
 out touching, the greater the inductive effects will be. 
 
24 
 
 THE QUADRUPLEX. 
 
 Condensers are constructed in accordance with the foregoing 
 laws. They are, therefore, generally composed of sheets of tin-foil 
 connected alternately to a battery wire and to the earth, as seen in 
 Fig. 17. Each sheet is separated from the other by some insulating 
 substance, preferably mica, on account of its thinness and high in- 
 ductive capacity, although paraffine paper is frequently used. Sup- 
 posing the lines in Fig. 17 to represent the foil, and the spaces the 
 insulating material, the plan of a condenser will be understood. 
 The current passing to the line charges one set of plates as shown, 
 which induces electricity of an opposite polarity in the ground 
 plates, and this in turn condenses electricity in the line plates, ac- 
 cumulating therein a charge in proportion to the number of plates 
 in action. 
 
 ? 
 
 4 
 
 C T 
 
 3 2 
 8 
 
 D "q 
 
 f6 
 
 3 l 
 32 
 
 3 
 
 to 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Fig. 18. 
 
 In Fig. 18 the general plan of the condenser is shown. All the 
 lower plates are connected with the thumb-screw B, which can be 
 readily connected with the earth, and the upper plates to J., to 
 which the line can be attached. By inserting plugs at a a a a a, 
 more plates may be connected, and thus the amount of charge in- 
 creased as desired. 
 
CHAPTKR IV. 
 
 THE STEARNS DUPLEX. 
 
 Fig. 19 is a diagram showing the theoretical arrangements neces- 
 sary to the working of this duplex system. 
 
 In it T and T' are transmitters worked by means of a local battery 
 and key. R and R' are relays differentially wound. E and E' are 
 resistance coils, or rheostats, which furnish a method of making an 
 artificial resistance equal to that of the line. ^Tand H' are condens- 
 ers, the manner of use of which will be explained later on. 
 
 Having made the rheostat resistance equal to the line resistance, 
 we have at each end the conditions described in Fig. 12; the line 
 wire representing wire B of that figure and the rheostat wire A, the 
 core of each relay being equivalent to our soft iron bar. 
 
 Now let us see what takes place. Suppose T to be closed, and 
 T' open. (In the diagram both transmitters are represented as 
 closed.) In this position of the transmitters it is seen that battery 
 F is in contact with the line, while battery I" will be cut off at 6r'. 
 
 Thus equal currents from battery F will pass in opposite direc- 
 tions around the cores of relay 72, one portion going to the ground 
 through coil A of relay R, and the rheostat ; the other, via the line 
 and through coil B' of relay R' to the ground at Q'. Thus battery 
 F produces no effect on relay R, but in traversing coil B' of R' it 
 magnetizes the core of the latter, and its armature is attracted, re- 
 cording a signal. Let us say that the current from battery F in 
 passing through coil B' of R' produces a magnetic strength of 2 
 in the core of the latter. Let us also grant, for the sake of illustra- 
 tion, that it is passing through coils A and B of relay R with a 
 strength of 2 each but owing to the fact that this strength is 
 exerted in opposite directions no effect is perceptible and let this 
 
26 
 
 THE Q UADR UPLEX. 
 
THE QUADRUPLEX. 27 
 
 also be true of the effect of battery F' on the cores of relay R\ 
 when F' is placed in the circuit, viz., that it has a strength of 2. 
 
 Should transmitter T' be now closed, as in the figure, breaking 
 circuit at x', and placing battery F' in the circuit, wire a' of relay 
 R f will now carry a current of a strength of 2, but an additional 
 strength of 2 has been given to the current traversing wire B 1 . 
 Wire B' of relay R' therefore still carries a greater current than 
 wire A' of the same relay, viz., as 4 to 2, consequently that relay re- 
 mains closed, while the current in wire B of relay R is also in- 
 creased by the current of 2 from battery .F', and is now of greater 
 strength than that upon wire A of that relay, hence magnetism is 
 produced in its core, and the armature is attracted, both relays be- 
 ing now closed. If transmitter T be now opened, it will be seen 
 that the current from battery F is removed from the line, and the 
 currents in wires A' and B' of relay R' become equal, consequently 
 the core of that relay loses its magnetism, and the armature is 
 pulled back by the retractile spring, opening the local circuit ; but, 
 as there is still a current in coil B of relay R, and none in coil A, 
 that relay remains closed. Thus it is seen that under all circum- 
 stances battery F produces an active effect only on relay R', and 
 that battery F' only operates relay R, and hence two messages may 
 simultaneously be sent in opposite directions. 
 
 A line wire acts to a certain extent like a condenser. When it 
 is connected with a battery it becomes charged, and whenever the 
 proper conditions exist along its route it will induce electricity in 
 adjoining conducting objects, and thus is enabled to take up a 
 greater charge. In this duplex system, when we open the trans- 
 mitter, and thus suddenly cut off the battery, and substitute a path 
 to the earth, a portion of the charge will come back, and pass to 
 the earth by this route rather than overcome the greater resistance 
 to the distant end. 
 
 This discharge varies in quantity with the strength of the charg- 
 ing battery and the length and size of the wire. Thus there are 
 many short duplex and quadruplex circuits on which this line dis- 
 charge, or static discharge, as it is often called, is so small or dis- 
 
28 THE QITADRUPLEX. 
 
 charges itself so quickly that no effect is produced on the instru- 
 ment. But on long circuits it becomes so great that in passing 
 through the line coil of the relay a momentary magnetization of the 
 core is caused, which attracts the armature and makes a false sig- 
 nal. In order to counteract this effect, a condenser, which can be 
 adjusted to equal the accumulative capacity of the line wire, is con- 
 nected to the rheostat coil of the relay (as seen in the diagram) so 
 that its discharge may traverse that coil in an opposite direction 
 around the core, and at the same instant that the line discharge trav- 
 erses the line coil, thus neutralizing the effect of that discharge. 
 Further allusion will be made to the action of the condenser when 
 we consider the quadruplex. 
 
 The elements of batteries, that is, the zinc, copper and liquids, 
 offer resistance to the passage of the current through them ; this is 
 called "internal resistance." In order to keep the resistance the 
 same whether the circuit is completed through the battery, or 
 through the armature lever of the transmitter, to the earth, resist- 
 ance coils Q and Q' (Fig. 19), having a resistance equal to the in- 
 ternal resistance of the home battery, are inserted between the 
 levers of the transmitters and the earth. 
 
V. 
 
 INSTRUMENTS OF THE POLAR DUPLEX. 
 
 THE POLE CHANGER. 
 
 Figure 20 presents an end view of a pole changer, with its the- 
 oretic battery connections. 
 
 The square B represents one end of the armature lever, which 
 protrudes through a hole in the center of the disc K. The squares 
 on either side are screwed to the metallic disc JT, but the tongues 
 T and T', each having two contact points, while supported by the 
 
 linn 
 
 Fig. 20. 
 
 disc JT, are insulated from it. Each tongue is connected to a pole 
 of the battery. The bar B, or armature lever, is connected to earth. 
 The line wire is connected to the disc K. In the figure, the key 
 of local circuit controlling the pole changer is presumed to be 
 
30 
 
 THE QVADRVPLEX. 
 
 closed, hence the lever B has made contact with the upper tongue 
 T, raising it from the square S on the left hand side, and has at the 
 same time released the lower tongue which has risen by its own 
 tension and made contact with the square on right hand side. The 
 circuit is thus complete, from the earth at E, through lever, to the 
 upper tongue T, to the negative pole of the battery, through bat- 
 tery to the lower tongue T', thence to the line. Thus in this case 
 the positive pole of battery is to the line, and the direction of the 
 current is shown by the arrows. 
 
 Fig. 21. 
 
 Fig. 21 shows the pole changer with key open. The lever B has 
 now made contact with the lower tongue T', detaching it from the 
 right hand square S. It has at the same time released the upper 
 tongue T, which now descends and connects with the left hand 
 square S. The circuit is again complete, from the line to square S, 
 to tongue T, to negative pole of battery, through battery to lower 
 tongue T' and earth, and the direction of current is now changed 
 as shown ^y arrows. Thus at each motion of the lever, up or 
 down, the poles are changed and, consequently, the direction of 
 the current. 
 
THE QUADRUPLEX. 
 
 31 
 
 In Figs. 8 and 9 we have seen that according to the direction 
 of the current around the bar is determined the magnetic polarity 
 of the bar. In those cases the direction of the current around the 
 bar has been changed by winding the coils in opposite directions, 
 so that the current has been made to go around the bars in differ- 
 ent ways. We might have obtained the same results by changing 
 the direction of the current on the wire and leaving the coil wound 
 in one way. If we know the direction in which a current is trav- 
 ersing a bar, we can tell the position of the magnetic poles of the 
 bar by the following rule : Suppose that we are looking at one 
 
 Fig. 22. 
 
 end of the bar ; if the current is circulating around the bar (with 
 relation to the position in which we are looking at it) in the direc- 
 tion in which the hands of a clock move, then the end at which we 
 are looking is a south pole, and vice versa. 
 
 When a positive pole of a battery is placed to a line wire, the 
 current is supposed to flow out on the line ; when a negative pole 
 of a battery is to the line, the current is supposed to flow in from 
 the line. If the reader who is beginring the study of the polar du- 
 plex will bear this in mind, it will simplify the study considerably, 
 
32 
 
 THE QUADRUPLEX. 
 
 and it will be well to remember that when we have, say, 50 positive 
 cells of battery at one end of the wire and 50 cells of positive bat- 
 tery to the line at the other end, there will be no flow of current on 
 the line, as these batteries neutralize each other. 
 
 In Fig. 22 let us suppose that B is a bar of soft iron, bent into 
 the shape shown, encircled by a coil of wire, also in manner shown, 
 and that M is a strip of magnetized iron, freely suspended, with its 
 
 Fig. 23. 
 
 north pole N between the ends of the iron bar. If a current of 
 electricity be sent through the wire in the .direction indicated by the 
 arrows, the bar will become a magnet with its poles as marked. 
 Hence, as like poles repel and unlike poles of magnets attract each 
 other, the north pole of the magnet M will be attracted to S and 
 repelled from N. 
 
 Now if the direction of the current be reversed, as shown by the 
 arrow in Fig. 23, it will of course traverse the coil and go around 
 the bar in a direction opposite to that shown in Fig. 22. Conse- 
 quently, as the polarity of the bar is determined by the direction of 
 the current around it, its polarity will be changed, and the poles 
 
THE QUADRUPLEX. 33 
 
 will now be as marked in Fig. 23. Therefore the magnet M will 
 now be repelled from its last position and attracted to the other 
 side, as shown. Thus we can see that as often as we change the 
 direction of the current around the bar, we change the position of 
 the poles of the bar, and at the same time, by the alternate attrac- 
 tion and repulsion, change the position of the magnet M suspended 
 between the poles of the bar. It follows that if we cause the mag- 
 net M to control a local circuit, and provide an instrument to 
 reverse the direction of the current, we have the necessary condi- 
 tions for a system of telegraphy, and this principle is taken advan- 
 tage of in the polar duplex, by means of the polar relay and pole 
 changer. 
 
 THE POLAR RELAY. 
 
 The polar relay is constructed with a view of complying with the 
 above conditions. 
 
 When an iron bar is magnetized by electricity circulating around 
 it, it is termed an electro-magnet. 
 
 Fig. 24 gives a front view of a polar relay. 
 
 P is a curved piece of steel, which is permanently magnetized to 
 the polarity marked. E M\& an electro-magnet, the core or bar of 
 which is somewhat similar in shape to the bars shown in Figs. 22 
 and 23, but wound differentially. This electro-magnet is screwed 
 firmly to the south pole of the permanent magnet. The armature 
 is a soft iron strip or lever hinged at one end to the north pole of 
 the permanent magnet P, as shown in dotted lines, its other end 
 projecting between and beyond the poles of the electro-magnet 
 E M. 
 
 Where there is no current traversing the coils of the electro-mag- 
 net E M, the poles of its core become, by induction, the south poles 
 of the permanent magnet P. Likewise the outward end of the 
 lever M. has become and remains the north pole of the permanent 
 magnet P. Thus the lever M, when adjusted in the center, will be 
 attracted equally by either end of the electro-magnet, as both of its 
 ends are, at present, south poles. 
 
34 
 
 THE Q ITADR UPLEX. 
 
 When a current is sent thrc . jh one coil of the electro-magnet E 
 M, the electro-magnetism thus generated in its core, being of 
 greater strength, will overcome the magnetism induced by the per- 
 manent magnet P, and the poles will be established according to 
 the direction of the current around the core. The polarity of the 
 lever M will, however, remain constant, and it will be attracted 
 
 Fig. 24. 
 
 and repelled by the alternate changes of the polarity in the electro- 
 magnet, and, as the lever is provided with a contact point to break 
 and close the local circuit shown in the diagram, a signal will be 
 recorded on the local sounder at each change of the direction of the 
 current around the core of the electro-magnet. 
 
CHAPTKR VI. 
 
 THE POLAR DUPLEX. 
 
 Fig. 25 is a diagram of the polar duplex set up at both ends. 
 
 For the purpose of illustration, we shall speak of the currents as 
 circulating around the cores of the polar relays in a direction from 
 right to left, or left to right, as the case may be ; and, in the dia- 
 gram, the preponderance of the current in either of those direc- 
 tions may be supposed to have magnetized the cores to the polari- 
 ties marked. 
 
 A and A' are pole changers, R and R' are polar relays. B is a 
 battery having, in the present position of pole changer A, its posi- 
 tive pole to line. B 1 is another battery having, in present position 
 of pole changer A', its negative pole to line. C and C'are condens- 
 ers adjusted to give a static discharge equal to that from the line 
 at the moment of the change of the battery polarities, and thus 
 prevent a " kick " similar to that mentioned in th.e Stearns duplex. 
 H and H' are rheostats adjusted to equal the resistance of the line 
 wire. The electro-magnets of the polar relays R and R' are wound 
 differentially. Hence a current from battery B (supposing for the 
 moment that there is no battery at the distant end) will divide at 
 the split S in two equal portions, each portion we shall say with a 
 strength of 2, as explained in the Stearns duplex, one portion 
 going to the ground through the line, and the other to the ground 
 through the rheostat H, in opposite directions around the core of 
 the relay R, producing no magnetism in it. 
 
 If now we suppose that battery B' be placed in the circuit, with 
 its negative pole to the line, it will give the current traversing the 
 line coil a of R, from left to right, a strength of 4, and this addi- 
 tional strength overcoming the strength of 2 in coil b will mag- 
 netize the core of R to the polarity marked, consequently the 
 
36 
 
 THE QUADRUPLEX. 
 
THE qUADRUPLEX. 37 
 
 outward end of lever M being a north pole will be attracted to S and 
 repelled from N. A somewhat similar effect takes place in relay 
 R'. The positive current from B is augmented by the negative 
 current of B' and thus there is a current of a strength of four trav- 
 ersing coil a' and going around the core of R' from right to left, 
 also .magnetizing it to the polarity marked and attracting the 
 lever M. 1 
 
 Now, if while the negative pole of battery B' is still to the line, 
 we change the polarity of the home battery B by opening the pole 
 changer A, what will be the effect on the home relay R ? It 
 should be nothing. Let us see, and here the reader is recom- 
 mended to draw for himself diagrams corresponding to the conse- 
 quent changes. This change puts the negative pole of battery B 
 to the line. Thus batteries B and B' are now opposed to each 
 other, and, being of equal strength, there is no flow of current 
 through the line coil a of relay R, but as the direction of the cur- 
 rent of B is changed, and while it is true that there is now no cur- 
 rent on the line coil of jR, there is a current with a strength of 2 
 flowing in from the rheostat to the ground through coil 5, and, as it 
 thus flows, we may see that the electricity is still going around the 
 core of R, in a direction similar to that in which it was circulating 
 before the change of polarity at -5, namely from left to right, and 
 hence the polarity of the core is unchanged, and the lever M re- 
 mains as before. This shows that the magnetic polarity of the 
 home differential polar relay is not affected by the changes of polar- 
 ity in the home battery. 
 
 Let us now examine the effect on the relay R of a change of 
 polarity in battery B'. It should reverse the magnetic polarity of 
 the core of relay R. The negative pole of battery B' has been to 
 the line. It will now be the positive pole. Let B have positive 
 pole to line as in Fig. 25. In this position, the positive pole of B 
 now neutralizes the positive pole of 5, and therefore there is no 
 flow through the line coil a, but there is a current of a strength of 
 2 now flowing through the rheostat ff to the ground through the 
 coil 6, and as this current is circulating around the core in an oppo- 
 
38 THE QUADRUPLEX. 
 
 site direction to that in which the previous magnetizing current had 
 been flowing, namely from right to left, the polarity of R is now 
 changed, whereby the lever M is repelled from its former position 
 and attracted to the other end of the core. Thus we observe that 
 the relay R responds to every change of the polarity at the distant 
 end of the line, regardless of the polarity of the home battery B. 
 It is needless to add that the relay R' will in a like manner respond 
 to every change of polarity in battery B. Such being the case, we 
 may simultaneously change the polarity at each end as frequently 
 as desired, and the various signals will be recorded appropriately. 
 
 There is probably nothing in the operation of the polar duplex 
 or quadruplex more difficult of comprehension to the beginner than 
 the effect produced by the placing of similar poles of batteries 
 against each other. For instance, if we take a battery of 100 cells 
 at one end of a line and one of 50 cells at the other end, and place 
 the positive poles of each battery to the line, there will be, roughly 
 speaking, a strength on the line equal to that from a single battery 
 of 50 cells, and the beginner is generally at a loss to understand 
 what becomes of the strength of the remaining 100 cells. To tell 
 him that 50 cells at one end are neutralized by 50 at the other end 
 is not always satisfactory. 
 
 Perhaps the following homely illustration will make it somewhat 
 plainer : 
 
 Let us imagine a railway train with a locomotive at each end of 
 the train, each locomotive to have a strength of 50. If we oppose 
 the strength of these engines against each other, we know that the 
 train will not move. If now we bring another locomotive having 
 also a strength of 50 and place it at one of the ends, the train will 
 now move, but urged only by a strength in all of 50, although there 
 is a locomotive force of 150 being exerted on the train. In other 
 words, the power of two of the locomotives is consumed, one in 
 opposing another of equal power, as in the case of the battery 
 referred to. 
 
 We may indeed use the same homely illustration in a variety of 
 ways to show the action of electricity on a wire. 
 
THE QUADRUPLEX. 39 
 
 We have heretofore spoken of the current or flow of electricity 
 when \ve have meant to signify the direction in which the electricity 
 acts. Thus, we have said that when the positive pole of a Battery 
 is placed to the line its action is, invariably, toward the line, as 
 shown by a galvanometer, and the action of a negative pole when 
 placed to the line is as invariably inward from the line. 
 
 Now, if we say that when a locomotive is pushing the train it is 
 exerting a positive strength, and when it is pulling the train that it 
 is exerting a negative strength, it is easy to understand that if we 
 cause the engine at one end of the train to exert a positive and the 
 one at the other a negative strength upon the train, there will be 
 the full effect of two engines acting upon the train in the same di- 
 rection, and the train will move accordingly. 
 
 On the contrary, if we cause both to exert a positive strength, 
 or both to exert a negative strength upon the train, that is to say, 
 if we cause them both to push or both to pull the train, it will not 
 move. 
 
 Thus it is, practically, with electric batteries at each end of a 
 wire, with regard to their effect on electro-magnets. If we place 
 their positive poles to the line, the action of each is toward the line 
 and no effect is produced. If we place their negative poles to the 
 line, their action is inward from the line and no result follows ; but 
 if we put the negative pole of the battery to the line at one end 
 and the positive pole at the other end, the strength of both is now 
 virtually being exerted in the same direction on the wire and the 
 strength of the two batteries is obtained. 
 
 In addition to the weight of the train which the engines have to 
 move, they also have to use part of their strength in moving them- 
 selves. . 
 
 The resistance which a battery has to overcome in its own ele- 
 ments and which is called its internal resistance, as before stated, 
 may be likened to the weight of the engine. Thus the lower the 
 internal resistance of a battery can be made without reducing its 
 electro-motive force, the gfeater will be the resultant strength up^n 
 a given wire. 
 
CHA.PTBR VII. 
 
 THE QUADRUPLEX. 
 
 Having studied the principles upon which the Stearns and polar 
 duplexes are operated, it will perhaps render the study of the 
 quadruplex less complex if we first examine separately the modifi- 
 cations necessary to the combination of these duplex principles in 
 the quadruplex. 
 
 We have seen that the Stearns duplex is operated primarily upon 
 two conditions, namely, battery to the line and no battery to the 
 line. Consequently, there are intervals in the working of this du- 
 plex when there is no current on the line. It will at once suggest 
 itself to the student that such a condition as this on the polar du- 
 plex (which is operated, not by a withdrawal of the current at any 
 time, but by a change in its direction,) would render it inopera- 
 tive. Therefore, if we are to combine these two principles, we 
 must in some way preserve at all times a current on the wire to 
 insure the working of the polar system. 
 
 The manner in which this is accomplished will be shown in the 
 description of Figs. 27 and 28. 
 
 We have seen also that the polar relay in the polar duplex sys- 
 tem is actuated by the reversals of polarity, and that the neutral 
 relay in the Stearns duplex is attracted by a current of either 
 polarity. 
 
 The following illustration will explain how these different results 
 are obtained on the same wire at the same time, as they are in the 
 quadruplex system, without interfering with each other. 
 
 In Fig. 26 we have a single wire encircling the cores of a neutral 
 and a polarized relay in such a manner that a positive current 
 with, we will again assume, a strength of 2 from battery C will 
 pass around their cores in a direction tending to magnetize them 
 
THE qUADRUPLEX. 
 
 41 
 
42 
 
 THE QUADRVPLEX. 
 
 to the polarity marked that is to say, if we are looking at the ends 
 marked $ (the south pole), the current will be circulating around 
 the cores in a direction similar to that of the hands of a watch. 
 Hence the north pole of lever M is attracted to S and repelled from 
 N, and the armature M 1 is attracted to both poles of its magnet, as 
 in the ordinary Morse relay or sounder. Let us now increase the 
 tension of the retractile spring of M' just enough to overcome the 
 attraction of the magnetism generated in R' by the existing cur- 
 rent having a strength of 2 ; so that M' is withdrawn from R'. 
 If now, by means of a pole changer, we change the direction of the 
 
 J) 
 
 Illlllllll 
 
 Fig. 27. 
 
 current from battery (7, we shall find that the poles of both relays 
 have changed, and that the armature or lever M has changed its 
 position because of this ; but we find that the armature M' of the 
 neutral relay is unmoved, being still retained by the retractile 
 spring, for the reason that, although the direction of current has 
 
THE QUADRUPLEX. ^ 
 
 changed, its strength has not, and consequently the strength of the 
 magnetism of R 1 has not been augmented. 
 
 Again referring to Fig. 26. If now we increase the strength of 
 the current from 2 to say 4, without changing the direction of 
 the current, we find that the increased magnetism in relay Rf has 
 attracted armature M', but the only effect produced on the lever M, 
 by the additional strength of current, is that it is more strongly 
 attracted to S, and repelled from N. 
 
 line 
 
 Fig. 28. 
 
 To carry the illustration one point further. If, while the strength 
 of 4 is on the line, we change the direction of the current, we find 
 that the lever M has again been affected, and that it has changed 
 its position, while the armature M' is unmoved for the reason be- 
 fore given. 
 
 The foregoing embodies the manner in which the two main prin- 
 ciples of the duplexes are combined in the quadruplex. 
 
 In our study of the Stearns duplex we learned that the function 
 
44 THE QUADRUPLEX. 
 
 of the transmitter was to alternately place the line to the battery 
 arid to the ground. In the quadruplex system its function is to 
 alternately place the whole battery and a small portion of the bat- 
 tery to the line, in the manner hereafter described. 
 
 The armature of the neutral relay in the quadruplex is adjusted 
 so that the smaller portion of the distant battery will have no prac- 
 tical effect upon it, and as this smaller portion of the current is re- 
 duced to the lowest point compatible with the effective working of 
 the polar relay, it follows that, as the polar relay is designed to 
 work uninterruptedly and smoothly through all changes in the 
 strength of current, it should be a more sensitive instrument than 
 the neutral relay, and such is the case 
 
 As sensitiveness is not so much desired in the neutral relay, its 
 core is made very short, and it is wound with coarser wire than 
 the polar relay, hence it has fewer convolutions of wire around its 
 cores than the latter instrument. Therefore a given strength of 
 current will have much more magnetic effect upon the polar relay 
 than upon the neutral relay. But apart from the above noted dif- 
 ferences in the construction of the two instruments the fact that the 
 armature of the polar relay aids in its own movements is an impor- 
 tant factor in rendering it more sensitive than the neutral relay. 
 
 The usual resistance of each coil of the polar relay is about 400 
 ohms, that of the neutral relay 200 ohms. 
 
 Figs. 27 and 28 show the manner in which the transmitter 
 changes the strength of current on the wire. In the figures the in- 
 struments unnecessary to show this action are omitted. 
 
 Fig. 27 shows a wire running from the earth to the nega- 
 tive pole of battery B. There is also another wire from one of the 
 positive cells of the battery which runs from P to the bar of the 
 transmitter, which is represented as open Another wire is led from 
 the positive end of the battery to the post D of the transmitter, 
 and the line wire is connected as heretofore to the tongue of that 
 instrument. The same connections occur in Fig. 28. In Fig. 27, 
 if we follow the circuit through the full battery to post D, we can 
 see that it leads to an open point near X, so that no current can 
 
THE QUADRUPLEX. 45 
 
 pass. On the contrary, if we follow the wire, which taps the bat- 
 tery at P, we may see that there is here a route for the current 
 through the bar to the tongue of the transmitter, and thence to the 
 line ; but in this case it is plain that that portion of the battery be- 
 tween P and F is cat off, and thus only a small portion of the bat- 
 tery is placed to the line, hence the neutral or No. 2 relay at dis- 
 tant end will not have'attracted its armature. Referring now to 
 Fig. 28, the transmitter is shown as closed. The figure shows 
 that the bar of the transmitter has broken contact with the tongue 
 at X, but instead the tongue has made contact with the post D'. 
 Thus the full' battery is now placed to the line, and the armature 
 of neutral relay at distant end will be attracted to the core. Thus, 
 as often as the transmitter is opened or closed, the full battery, or 
 only a portion thereof, will be placed to the line, but in either case 
 there must always be sufficient current left on the line to operate 
 the polar relays. 
 
 Fig. 29 shows the theoretical manner in which the various prin- 
 ciples that we have been considering in the duplexes, etc., are 
 grouped to produce the Edison quadruplex. 
 
 P and P' are pole changers, 2* and T' are transmitters, -ZV^and N' 
 are neutral relays, R and R f are polarized relays. The manner in 
 which the relays are wound in the figure is not the precise method 
 adopted in actual practice, but the result is the same. We have 
 chosen the manner of winding shown in the diagram with a view to 
 making the action of the current on the cores of the electro-magnets 
 easier of comprehension. 
 
 The figure shows the quadruplex theoretically set up at both 
 ends, namely A and B. 
 
 At A the keys of both sides are supposed to be closed, at B 
 open. 
 
 It might be supposed that having discovered the foregoing princi- 
 ples of electricity and magnetism, it only remained for the inventor 
 to manufacture and arrange the instruments necessary to 
 produce the increase and decrease of current and the changes of 
 polarity, with instruments to respond to those changes, when he 
 
Fig. 29. THEORETICAL DIAGRAM OF QUADRUPLED 
 
THE QUADRUPLEX. 47 
 
 would have a successful working quadruplex system, and probably 
 this was the conclusion arrived at by many inventors and experi- 
 menters, but actual experience taught them that more than this was 
 required. For instance, it was found, as in the duplexes, that 
 between each change of polarity of the home battery the line on 
 long circuits gave back a return or static discharge, which, by mo- 
 mentarily charging the cores of the relay with opposite magnetism, 
 produced a click in the home relays, which rendered the signals 
 from the distant end unreadable. Hence it was necessary to insert 
 in the balancing wire between the rheostats and the relays a con- 
 denser ((7 and 0' as in figure), which would accumulate a quantity 
 of electricity equal to that accumulated on the line, and which 
 would also discharge itself during the interval between the reversals 
 of the poles of the batteries, and thus neutralize the effect of the 
 line discharge. This was found effectual on comparatively short 
 circuits, but on longer circuits it became apparent that the condenser 
 discharged itself before the Mne had completely discharged, and 
 thus the result desired was not quite attained. This necessitated 
 the insertion of the resistance coils H" and H'" between the artifi- 
 cial wire of the rheostat and the condenser, to retard the discharge 
 of the latter, which proved successful. 
 
 It was also seen that at every motion of the transmitter which 
 cuts off the larger portion of the home battery, the internal resist- 
 ance of said portion of the battery, which is equal to from two and 
 one half, to three ohms per cell, was taken from the wire, thus de- 
 ducting from the total resistance of the line, say, from 100 to 500 
 ohms, according to the number of cells cut off. This would have 
 produced a variation in the resistance, which, especially in bad 
 weather, when there was a small working margin of current, would 
 more than likely, by making the balance at the distant end un- 
 steady, prevent the successful working of the quadruplex. To com- 
 pensate for this, resistance coils (A and A'), equal to the internal 
 resistance of the cells thus cut off, were inserted between the taps 
 Z and Z' and the bars of the transmitters. 
 
 Another obstacle to the working of the quadruplex was also en- 
 
48 THE QUADHUPLEX. 
 
 countered. It was found that while the full battery was to the line 
 (as at A in the figure, and the armature of neutral relay N' at dis- 
 tant end B was thus attracted to its magnet), if the polarity at sta- 
 tion A was changed, thereby, as we have seen, for an instant taking 
 off the battery at JL, from the line, at that instant the retractile 
 spring of the neutral relay N' would tend to withdraw the armature 
 from the magnet, thus making a false "clip," and breaking up the 
 regular signals. To obviate this trouble the local contact of the 
 neutral relay is placed on the back stop, and a repeating sounder 
 is in many cases also added to the local circuit, as is shown in cut 
 (Fig. 30). The instant when the battery is cut off from the line 
 is, however, of so short duration that the neutral relay has appar- 
 ently not time to become entirely demagnetized, so that it regains 
 its maximum magnetism and re-attracts its armature before the lat- 
 ter has made a firm contact on the back stroke, and hence the re- 
 peating sounder is not fully charged, consequently its armature is 
 not sufficiently affected by the false signals to interfere with the 
 regular sounder, which is operated by this armature, and thus the 
 signals intended for the neutral relay, although undoubtedly muti- 
 lated in the course of transmission while the polar side is in opera- 
 tion, are received, by means of the above arrangement, unbroken, 
 and on the front stop of the regular local sounder. When the re- 
 peating sounder is dispensed with, as it is now on many circuits, 
 the connections are so arranged at the distant end that in reality 
 the transmitter sends out the signals on the " back stroke," but as 
 the local contact point of the neutral relay is placed, as before 
 stated, on the back stop, the receiving operator receives the signals 
 as usual. 
 
 It may be asked, why is not the polarized relay affected in the 
 same manner at the instant of change of polarity of the distant bat- 
 tery ? In fact it is, that is to say, it also partly loses the magnet- 
 ism due to the distant battery, but at that instant the permanent 
 magnetism of the polarized relay comes into operation, and, as has 
 been shown, as the ends of the core of the electro-magnet of the 
 polarized relay are in that case both south poles, the north pole of 
 
THE QUADRUPLED. 49 
 
 its armature will be retained in its present position until the actual 
 change of polarity ensues. But aside from this, there would be no 
 tendency on the part of the armature of the polarized relay to leave 
 its existing position, owing to the fact that its inertia would be 
 sufficient to keep it where last placed ; it being, as we have said, 
 the retractile spring of the neutral re 7 ay which tends to draw it 
 away at the time referred to. 
 
 Even after all these defects were discovered and remedied the 
 quadruplex was not yet a success, and it was not until after many 
 months of actual working and experiment that the modifications of 
 the instruments first used, which have brought the system to its 
 present state of usefulness, were invented and applied. 
 
 The polar side of the quadruplex is frequently called the No. 1 
 side ; and the increase and decrease of strength side, the No. 2 side. 
 
 The polar relay and neutral relay are also spoken of as the No. 1 
 and No. 2 relay, respectively. 
 
 When the compound polarized relay was used on the No. 2 side 
 of the quadruplex it was found that a condenser inserted between 
 that relay and the polar relay, by absorbing or counteracting the 
 extra currents generated by the demagnetization of the cores of the 
 relay, steadied that relay somewhat during the changes of polarity, 
 but as the utility of that condenser seems to have ceased with the 
 introduction of the short core relay, arid as it is now dispensed 
 with entirely on many circuits, we have not included it in the dia- 
 grams. 
 
 There now remains but little to be said as to the working of the 
 quadruplex if the reader has followed us by the different stages 
 that have led up to this point. 
 
 We shall, however, for the benefit of the beginner, trace the 
 course of the circuit in Fig. 29. Beginning at E, station J., the cir- 
 cuit is led to the bar of the pole changer, thence via the upper 
 tongue of the same to the negative pole of the battery at -4, through 
 the entire battery to the post and tongue of the transmitter, to the 
 lower tongue of the pole changer and thence to the disc thereof. 
 Here the circuit is divided, one portion going through the neutral 
 
50 THE QUADRUPLEX. 
 
 relay, the polar relay, and to the line and to the ground at distant 
 end B. The other portion of the circuit goes through the neutral 
 and polar relays and through the rheostat ^Tto the ground. 
 
 Beginning at E, station .5, the circuit is led to the bar of the 
 pole changer, thence to the lower tongue of that instrument, to the 
 positive pole of battery at -B, through the small end of this battery 
 to Z, thence to the bar of T', and via the tongue of T to the upper 
 tongue of P' and to the disc thereof, where the current divides as 
 at A. 
 
 In the present positions of the transmitters and pole changers at 
 both ends, the polar relays and neutral relays are attracted or with- 
 drawn, as shown. That such results are produced by these conditions 
 is well known to all who have had practical acquaintance with the 
 working of the quadruplex, but for the information of those who wish 
 to trace the question we shall endeavor to analyze one or two of the 
 many combinations that occur on the quadruplex, which could not 
 occur on either the Stearns or the polar duplex alone, promising 
 the beginner, however, that it may require considerable study on his 
 part to thoroughly understand them. We will suppose that the 
 small ends of both batteries consist of 50 cells and the large ends 
 100 cells each, making the total of each battery 150 cells. Thus, as 
 the full battery at A is to the line with its positive pole, and as it 
 is augmented by 50 cells from the negative pole of that at B, there 
 is virtually a strength from 200 cells flowing through the coil a of 
 R, from left to right to the line and a strength from 150 cells flow- 
 ing through coil b from right to left to the rheostat. (It is of 
 course understood that we speak of a strength of 200 cells on the 
 line merely for purpose of illustration.) This attracts, as we see, 
 the armature of the neutral relay N 1 at B, and the negative pole of 
 battery at B being to the line, the armature of R is as placed. Let 
 us now close the pole changer P 1 at B. This places 50 cells of posi- 
 tive polarity to the line at this point B, and these 50 cells oppose 
 and neutralize 50 positive cells of battery at A, making the total 
 strength of current now on the line equal to 100 cells. This now 
 leaves, therefore, a strength of 100 cells flowing through coil a of 
 
THE QUADRUPLEX. 51 
 
 the polar relay R to line from left to right, while there is now a 
 strength of 150 cells flowing from the battery at A through the 
 coil b of the polar relay to rheostat from right to left. This leaves 
 a preponderance of current flowing in the latter direction, and, con- 
 sequently, the polarity of the polar relay R is reversed, and its ar- 
 mature changes its position. 
 
 This position of the pole changer at j5, as we saw, reduces the 
 strength of current coming from battery at A to 100 cells. Why, 
 therefore, does not the retractile spring withdraw the armature of 
 N, which we saw was held to the armature by the full battery of 
 J., viz., 150 cells being to the line? It is because there is now a 
 strength of 50 cells from the positive pole at B going through the 
 neutral relay N' to the rheostat and ground, and this strength of 
 50 cells, owing to the manner in which the coils are wound, is cir- 
 culating around the core of the relay in the same direction as that 
 of the current of 100 cells strength from J., thereby still preserving 
 a strength of 150 cells on neutral relay M 1 , and of course this ac- 
 counts for the fact that its armature remains attracted. There are 
 many such combinations continually occurring on the quadruplex, 
 which at first sight would seem to be enough to entirely prevent 
 the working of the system, but an analysis of each combination will 
 prove, what is known to be the case, that they do not so prevent. 
 
 In studying the combinations that occur on these multiplex sys- 
 tems, it should be remembered by the beginner that the currents 
 which flow to or from the rheostat, or artificial wire, are always de- 
 pendent on the home battery. If, for instance, 50 cells of positive 
 polarity be placed to the line from battery at B in Fig. 29, a posi- 
 tive current of 50-cells strength will flow to the ground through 
 the rheostat, even though a positive battery of 150 cells should be 
 opposed to the same battery at the distant end, A. In other 
 words, the wire through the rheostat may be considered as a sepa- 
 rate wire, excepting that it is led through one coil of the various 
 differential relays. 
 
 Fig. 30 is a diagram showing the actual connections at one end 
 of a quadruplex system. The polar relay and the neutral relay are 
 
52 THE QUADRUPLEX. 
 
 shown by the bare coils. The repeating sounder previously men- 
 tioned adjoins the neutral relay. Q is the resistance to compen- 
 sate for that portion of the battery which is cut out when the small 
 end of the battery is to the line. The coil G- in the rheostat R is 
 the resistance equal to the internal resistance of the entire bat- 
 tery which is placed in the circuit when the wire is grounded for 
 a balance. The wire is grounded for this purpose by turning the 
 three-button switch B to the right. It will be noticed that this ac- 
 tion cuts off the battery, the pole changer and transmitter. 
 
 Q' is the resistance placed between the condenser and rheostat 
 wire on long circuits, for reasons already explained. 
 
 The route of the circuit in the present position of the pole changer 
 and transmitter is traced by the figures 1, 2, 3, 4, etc., up to 13, where 
 the circuit divides, and letters then show the route to ground 
 through the rheostat. 
 
 It may be stated in passing that there has been no special attempt 
 made in these diagrams to show the instruments in their actual re- 
 lations to each other as to size, the main idea being to present the 
 instruments in the clearest and simplest position. Thus the rheo- 
 stat box is shown as apparently smaller than the transmitter, 
 whereas, in reality, as we all know, it is several times larger. 
 
 BALANCING THE QUADRUPLEX. 
 
 To balance a quadruplex, ask the distant station to ground for 
 you. In doing so he cuts out his pole changer and entire battery, 
 as may be seen in the diagram of the complete quadruplex with 
 connections. It will also be seen that in grounding he interposes 
 a resistance equal to the internal resistance of his entire battery, 
 otherwise when he again cuts in his battery your balance would be 
 several hundred ohms too low, according to the size of his battery. 
 You now also ground. Now the wire is without battery at either 
 end. Then adjust the armature of the polarized relay so that it 
 will stay on whichever side it is placed. Now cut in your full bat- 
 tery and insert or withdraw plugs from your rheostat until the arma- 
 ture of the polarized relay again remains on whichever side placed, 
 
THE QUADRUPLEX. 53 
 
 this showing that an equal strength of electricity is flowing in oppo- 
 site directions around the core of the relay, as no magnetic effect is 
 produced, and proving that the resistances of the line and rheostat 
 are equal. Now open and close the pole changer and note the 
 action of the polar relay. If it responds clearly to each motion of 
 your pole changer, you either have too much or not enough dis- 
 charge from your condenser. Therefore adjust your condenser 
 until its discharge is sufficient to neutralize the static discharge 
 from the line, which will be shown by the failure of the polarized 
 relay now to respond to the movement of the pole changer. 
 
 The polar duplex is balanced in the same manner as the quadru- 
 plex. The Stearns duplex is balanced by asking the distant station 
 to open his key, which is virtually asking him to " ground." Then 
 dot on your own key and adjust your rheostat and condenser until 
 your relay remains passive under the movements of your trans- 
 mitter. 
 
 
 UITIVETISITT 
 
THE QUADRUPLEX. 
 
CHAPTER VIII. 
 
 THE DYNAMO-ELECTRIC MACHINE IN RELATION TO THE 
 
 QUADRUPLEX. 
 
 BY WM. MAVER, JR. 
 
 Upon the introduction of the dynamo-electric machines into the 
 Western Union building as a means of supplying electro-motive 
 force, it was found necessary, if this source of supply was to be 
 availed of in the working of the polar duplex and the quadruplex 
 systems, to alter considerably the arrangement of the instruments 
 used in the working of those systems by the gravity battery. 
 
 Before proceeding to describe these alterations it will perhaps 
 be well to interpose the following description of the manner in 
 which the dynamo machines supply the electro-motive force to the 
 various wires. The description will, I think, tend to make plainer 
 some of the statements hereafter made. 
 
 I may state that this account of the dynamo machines is taken 
 almost verbatim from an article on the Western Union building, 
 written by me some time ago, for The Electrical World. 
 
 The electro-motive force for the wires leaving the Western Union 
 main office is furnished by dynamo-electric machines, of which there 
 are fifteen on the battery floor, in series of five machines. Thus there 
 are three series. Two of these series are kept in operation day and 
 night, the other series being reserved for emergencies or to allow 
 of necessary repairs. The motive power for these machines is sup- 
 plied by three safety steam-power engines of six horse-power each, 
 which derive their steam from boilers in the sub-cellar. Of these 
 series or gangs of five machines, one machine is used to generate a 
 magnetic field for itself and the four other machines. It also deter- 
 mines the polarity of the series according to the direction of its cur- 
 rent around the field-magnets of the other machines. One of the act- 
 ive series supplies negative, the other positive, polarity to the wires. 
 
56 THE QUADRUPLED 
 
 The polarity of the spare series may easily be changed to positive 
 or negative by means of a switch arranged for the purpose in the 
 battery room. 
 
 Four grades of potential are furnished from each series of these 
 machines, in the manner hereafter described. 
 
 Artificial resistances, equal to three ohms per volt, composed of 
 coils of German silver wire wound on cylinders of prepared plaster of 
 paris, are interposed between each machine and every wire leaving 
 the switch, as shown in the engraving. Thus the resistance of each 
 coil interposed between the first potential and the wires is about 
 200 ohms ; between the second potential and wires, 400 ohms ; 
 between the third potential and wire's, 600 ohms, and between 
 the fourth potential and wires, 800 ohms. These coils serve to 
 diminish the intensity of the dynamo current, thereby obviating 
 sparks, and to prevent, in a measure, the burning of instruments in 
 case of the short circuiting of a wire. The first, second, and third 
 potentials are conducted to the main switch, the first and second 
 potentials only to the city line and Long Island switch. 
 
 The fourth potential is used principally to supply very long quad- 
 ruplex circuits, such as those to Bangor, Me. ; Lynchburg, Va., etc. 
 
 Wires having an average resistance of about 3,000 ohms are sup- 
 plied from the first potential. Those having an average resistance 
 of about 5,500, from the second potential. The third potential is 
 mainly used for duplex and quadruplex circuits. The New York City, 
 Brooklyn, and Jersey City wires, being generally of low resistance, 
 are also supplied from the first potential ; but the resistances of 
 those wires are increased by means of resistance plugs, inserted at 
 the switch in their circuits, to an average resistance of from 3,000 
 to 4,000 ohms. 
 
 The illustration, Fig. 31, gives a top view of one series of the 
 dynamo machines, showing the manner in which the several poten- 
 tials are derived and the method of interposing the artificial resist- 
 ance coils previously mentioned. 
 
 A, .B, C, _D, and E are the five dynamo machines of one series. 
 A is the feeder which furnishes the magnetic fields for itself and the 
 
THE QUADRUPLEX. 
 
 57 
 
58 THE QUADRUPLEX. 
 
 four other machines. A', B', C', D', and E' are the armature coils of 
 each machine. The rear of a portion of one section of the main 
 switch in the operating room is shown above the machines. To 
 simplify the illustration, neither the belting gear of the machinery 
 nor the polarity changing switch in the battery room are shown. 
 
 The plaster of paris cylinders, around which the artificial resist- 
 ance coils are wound, rest on horizontal strips of brass. The lower 
 terminal of each coil is attached to the brass support and the upper 
 terminal is conducted to a small disc on the switch. The perpen- 
 dicular strips on the switch are each connected to one line wire, so 
 that by means of a brass plug inserted between a disc arid a strip 
 the first, second, or third potential of current may be placed to the 
 line as desired. 
 
 The second machine B supplies the first potential of about 70 
 volts, which is conducted to the brass supports of the resistance 
 coils R, and thence to the switch, as shown. 
 
 The third machine C also supplies an electro-motive force of 70 
 volts, which being reinforced by the 70 volts generated by j5, raises 
 the net potential to 140 volts, which is conducted to the switch as 
 the second potential. In like manner the fourth machine D sup- 
 plies another 70 volts. This being added to the electro-motive force 
 of B and (7, increases the ^otential to 210 volts, which is also con- 
 ducted to the switch as before, excepting where it is used to supply 
 a quadruplex or duplex circuit, when it is conducted directly to the 
 instruments, after passing through special resistance coils. 
 
 The last machine E also generates an electro-motive force of 70 
 volts, thereby increasing the total electro-motive force of the series 
 approximately to 280 volts potential. As previously stated, this 
 fourth grade is used only for the longest quadruplex and duplex 
 circuits. Some of the machines have fewer convolutions on the 
 armatures than others, but these are run faster to compensate. 
 The internal resistance of each of these machines is about I of an 
 ohm. 
 
 The quantity of current generated by these machines would 
 appear to be almost limitless. Already over 300 wires of an aver- 
 
THE QUADRUPLEX. 59 
 
 age resistance of about 3,000 ohms are supplied from the first po- 
 tential without any perceptible diminution of the current or appre- 
 ciable effect on the working of the machine, hence the great economy 
 of the system, of which Mr. Stephen D. Field is the inventor. 
 
 I have said that upon the introduction of these dynamo machines 
 into the Western Union building it was necessary to alter some- 
 what the connections and arrangements of the quadruplex, etc. 
 This was owing to the fact, that as the dynamo machines are run in 
 two series, one supplying positive and the other negative polarity, 
 it is impracticable to tap a series at each end for change of polarity, 
 as in the gravity batteries ; and as each machine of a series furnishes 
 a certain electro-motive force, about 70 volts, it was likewise 
 impossible to regulate the small and long ends of the battery at 
 will, as is done with the gravity cells. These difficulties were over- 
 come : firstly, by changing the connections of the pole changer so 
 that at every alternate motion of its lever it drew from first one 
 series of dynamos, and then another, thus obtaining the necessary 
 changes of polarity ; and, secondly, by connecting the transmitter 
 so that in one position of its lever it allows the full current to go 
 to the line ; while in the other the current is forced to pass through 
 an added resistance before reaching the line, and is also given an- 
 other and generally a shorter route to the earth, in the manner 
 hereafter explained, so that the current passing to the line is greatly 
 weakened. 
 
 The same practical result is obtained by this device as by the 
 cutting off of a portion of the battery, namely, the current on the 
 line is increased or decreased at each movement of the transmitter. 
 
 In one case the strength of current is decreased by a reduction of 
 the electro-motive force, while in the other case, as we shall dis- 
 cover, the electro-motive force remains constant, but the circuit 
 is varied. 
 
 The retractile spring of the neutral or No. 2 relay is therefore 
 so adjusted that it withdraws the armature from the magnet when 
 the combination of the added resistance and the extra route to the 
 ground is in operation, but when this combination is inoperative 
 
60 THE Q UADE UPLEX. 
 
 the consequent increased strength of current going to the line 
 attracts the armature precisely as when in the gravity battery 
 system the entire battery is placed to the line. 
 
 In a former part of this book it has been stated that the strength 
 of current on any two wires supplied from the same battery will 
 be the same on each wire, provided the resistances of the wires are 
 equal. This is equally true if twenty or more wires of equal resist- 
 ance are attached to the same battery, and is in accordance with 
 the following law of electricity : 
 
 " The relative strengths of current in the branches of a divided cir- 
 cuit will be inversely proportional to their resistances" For example, 
 if we have a circuit divided into, say, two branches, the resistance of 
 one being 12 ohms and the other 6 ohms, twice as much of the cur- 
 rent will flow through the branch of 6 ohms as through the one of 
 12 ohms, because the latter route offers double the resistance to the 
 current. Hence the route offering 6 ohms resistance will carry, in 
 this case, two-thirds of the current, and the other route one-third. 
 But there will not be as much total resistance offered to the current 
 by the two routes together that is to say, by their joint resistance 
 as by either route alone. 
 
 By joint resistance is meant the total resistance offered to a cur- 
 rent supplied from one battery. For example, if two wires of 6 
 ohms each are led from one battery, the joint resistance of the two 
 wires will be 3 ohms. Three wires of 6 ohms each will have a joint 
 resistance of 2 ohms, etc. 
 
 This may be explained as follows : If we have a barrel filled 
 with water and attach one faucet at the bottom a certain amount 
 of resistance will be offered to the flow of water, another faucet 
 will decrease this resistance, and still another faucet by providing 
 an additional channel will yet further reduce the resistance. 
 
 The following is a simple rule for finding the joint resistance of 
 wires in a circuit. Divide the product of the resistances of the two 
 wires by their sum. 
 
 If there are more than two wires in the circuit, first find the joint 
 resistance of any two of the wires. Then use this joint resistance 
 
THE QUADRUPLEX. 
 
 61 
 
 of the two wires as though it were the resistance of a single wire, 
 and divide the product of said joint resistance and the resistance of 
 the third wire by the sum of the same, which will give the joint re- 
 sistance of the three wires. The joint resistance of the three wires 
 may now be used to find the joint resistance of a fourth wire, and 
 so on indefinitely. 
 
 The strength of a current of electricity, which is the amount of 
 
 1 electricity flowing through a circuit in a given time, is equal to the 
 
 number of volts of electro-motive force divided by the number of 
 
 ohms of resistance in the entire circuit. This is in effect Ohm's 
 
 law, viz. : 
 
 Electro-motive force, ~ E 
 Current = or (J =^ fr 
 
 Resistance H 
 
 All of these laws are availed of in the operation of the quadru- 
 plex when the currents from the dynamo machines are used. 
 Hence the necessity of a full understanding of them before pro- 
 ceeding to explain the latter system. 
 
 Fig. 32. 
 
 A preliminary example will now be given of the law of joint 
 resistance, the law relating to the strength of currents in divided 
 circuits, and Ohm's law. 
 
 In Fig. 32 let us suppose that up to the point Jf, there is a resist- 
 ance of 6 ohms. At X the circuit divides into two branches, A 
 
62 
 
 THE QITADRUPLEX. 
 
 and J5, each having 36 ohms resistance. By the rule given above, 
 we find the joint resistance of these two branches, A and J5, to be 
 18 ohms, which added to the 6 ohms resistance up to the point X 
 makes the total resistance of the circuit 24 ohms. Supposing the 
 electro-motive force to be 24 volts (a vplt being the unit of electro- 
 motive force), we find that as the strength of current on a circuit 
 is equal to the electro-motive force divided by the total resistance, 
 in this case it will be f = 1. Up to the point Jf, therefore, the 
 strength of current on the wire is 1, but at this point the current 
 divides equally between A and B, consequently on each of these 
 branches the strength of current will be . Now let us suppose 
 that this strength is quite sufficient to magnetize strongly a relay 
 at the distant end of A or B, and that in consequence its armature 
 has been drawn to the cores of the magnet, as is the case, for in- 
 stance, when the full battery is to the line in the Stearns duplex. 
 
 Now referring to Fig. 33. 
 
 We have interposed an additional resistance of 12 ohms before 
 the point X, and have inserted into the circuit at that point another 
 
 Fig. 33. 
 
 branch, (7, having a resistance of only 9 ohms. In this case there 
 are 18 ohms resistance before X. Now we require to find the joint 
 resistance of A, B, and C. We have seen that the joint resistance 
 
THE QUADRUPLEX. 
 
 63 
 
 of A and B is 18 ohms, therefore, by the rule already quoted, when 
 we find the joint resistance of 18 ohms and 9 ohms, which is 6 
 ohms, we have the joint resistance of the branches A, B, and 0. 
 These 6 ohms added to the 18 ohms inserted before X make a total 
 resistance for this circuit of 24 ohms, the same as in Fig. 32 The 
 strength of current, therefore, on this circuit up to X is also 1 ; but 
 now this current has to be divided proportionately between A, B, 
 and C. Hence, as the branch C offers only 9 ohms resistance while 
 the joint resistance of A and B is 18 ohms, C will receive twice as 
 much as both A and B. C will thus take two-thirds of the current, 
 while one-third remains to be divided between A and B. This gives 
 each of these branches a current of one-sixth, or only one-third 
 of what it was in Fig. 31, which we will assume so weakens the 
 magnetism of the relay at the distant end of A that the retractile 
 spring withdraws the armature. 
 
 Fig. 34. 
 
 Now if we can arrange our connections so that at one position of 
 the transmitter the circuit is virtually as shown in Fig. 32 while in 
 the other position the circuit is as seen in Fig. 33, it is evident that 
 we have the necessary arrangements for the increase and decrease 
 side of the quadruplex. In other words, we may accomplish in 
 
64 
 
 THE qUADRUPLEX. 
 
 this way what we performed previously by cutting off a portion of 
 the battery, as already explained ; and this is, in fact, the principle 
 on which the dynamo currents are used in the quadruplex system. 
 
 Fig. 34 will show the manner in which these connections are 
 made. In the present position of the transmitter in that figure the 
 current flows to the post of the transmitter, thence to the tongue 
 of that instrument, and after passing through the relays (not shown 
 in this figure) to the line and rheostat. It will be seen that in this 
 position of the transmitter there is also another route of 1,200 ohms 
 resistance, but, as the present route via the transmitter offers prac- 
 tically no resistance, all the current will pass to the line by the 
 latter route, where it divides equally between the two branches as 
 described in Fig. 32. 
 
 When the transmitter is open, as in Fig. 35, the current neces- 
 sarily flows through the 600 ohms at battery (which are inserted to 
 
 Fig. 35. 
 
 diminish the intensity of the dynamo currents, as mentioned in the 
 description of the dynamo machines), and the 1,200 ohms added 
 resistance, to the point JT, where it divides as explained in Fig. 33, 
 two parts of the entire current going to the ground via the trans- 
 mitter and the 900 ohms resistance, (which is technically termed 
 
THE QUADRUPLEX. 65 
 
 the " leak,") and one part dividing between the line and the rheostat 
 branches. 
 
 Since we know that by adding more resistance to a circuit having 
 a stated electro-motive force we can reduce the strength of current 
 on the circuit, the question may arise in the mind of the reader, 
 why not simply arrange the connections so that when the trans- 
 mitter is closed the added resistance should be out of the circuit, 
 and when the transmitter is open the added resistance should be 
 included in the circuit, thus dispensing with the third branch C or 
 " leak " ? The reason is, briefly, that the introduction of resistance 
 for the reduction of the strength of current has been found to 
 operate unsatisfactorily on quadruplex circuits ; which is mainly 
 owing to the fact, that so much resistance must be inserted in the 
 circuit to cause the instruments affected by the increase and de- 
 crease of strength, to act quickly, that it has a detrimental effect 
 upon the instruments worked by the changes of polarity. 
 
 The problem, therefore, to be solved was, how to arrange the con- 
 nections for the dynamo machine currents, so that whether the 
 transmitters were open or closed the resistance of the line should 
 be the same, and the problem was finally solved by the ingenious 
 addition of the leak in combination with the added resistance. 
 
 That this is so we may see by referring to Fig. 34 and following 
 out the routes. 
 
 A current coming from A, in the position of the transmitter in 
 this figure, would encounter, after passing the point X, 600 ohms, 
 not counting the 1,200 ohms added resistance. In Fig. 35, after 
 passing point X there is now a choice of two routes, namely, one 
 via the 1,200 ohms added resistance and the 600 ohms at the ma- 
 chine (making 1,800 ohms), and the other via the transmitter and 
 through the leak of 900 ohms. We know that the joint resistance 
 of 1,800 and 900 ohms is 600 ; thus the total resistance from the 
 point Xto the various grounds at the home end remains the same 
 in either position of the transmitter, while, as we have seen, the 
 strength of current reaching the distant end A or B is materially 
 altered at each change in the position of that instrument. Let the 
 
63 THE QUADRUPLEX. 
 
 reader bear in mind, therefore, that, notwithstanding the insertion 
 and practical removal of these various resistances from the line, the 
 actual total resistance of the quadruplex circuit is not changed. It 
 is the insertion of the branch circuit, or leak to ground, that is the 
 prime factor in obtaining the desired increase and decrease of 
 strength. 
 
 In applying machine currents to the quadruplex, almost the same 
 instruments as are used with a gravity battery will answer on short 
 lines, where comparatively weak currents are used, but, where con- 
 siderable strength of current is required, the spark produced at the 
 contact points of the ordinary continuity preserving pole changer, 
 at the moment of reversal of polarity, is sufficient to interfere with 
 its satisfactory working. 
 
 For this reason a pole changer, such as is shown in Fig. 36, is 
 used on long quadruplex circuits. 
 
 Its construction is very simple. The straight lever L is pivoted 
 on the standards S, so that when either of the ends of the lever 
 is depressed, it will come in contact with the upright post Z or C 
 beneath it. The ends of the lever and the tops of these posts are 
 furnished with platina contact points. The posts are insulated from 
 the plate upon which they stand. The retractile spring is also 
 insulated from the hook at the point P to prevent short circuiting. 
 The, line is connected to the lever via the standards S, and the two 
 posts Z and (7 are connected, one to a positive series of dynamo 
 machines and the other to a negative series. It may thus be seen 
 that if the right hand end be depressed, making contact at C, con- 
 tact will be broken at Z and a positive current will flow to the 
 line. 
 
 With the left hand end depressed, contact will be made at Z, 
 and the polarity will be reversed. 
 
 It will be noticed that, with this style of pole changer, contact is 
 not made with one post until after the other has broken contact, 
 and thus short circuiting of the battery is entirely obviated. 
 
 The figures which I have heretofore quoted, showing the 
 amount of resistance placed in the " added resistance " as 1,200 
 
THE QUADRUPLEX. 
 
 67 
 
 ohms, and that in the " leak " as 900 ohms, give, as we have seen, 
 a difference equal to 3 to 1, in the strength of current which reaches 
 the distant station, in the different positions of the transmitter. 
 
 This margin is found sufficient on circuits of medium length, but 
 on very long circuits better results are obtained when the margin 
 
 Fig. 36. 
 
 is made greater, for instance, 4 to 1. With a gravity battery the 
 division of the electro-motive force necessary to obtain this margin 
 is easily made by apportioning the " long " and " short " ends of. 
 the battery, respectively, the requisite number of cells to produce 
 the desired result. 
 
 Thus in a battery of 300 cells, the " short end " would be appor- 
 
68 THE qUADRUPLEX. 
 
 tioned 75 cells to obtain this margin. But, as I have previously 
 stated, it is not practicable to alter thus the electro-motive force 
 furnished by the dynamo machines ; therefore, when it is desired to 
 increase the margin as above, namely, to make 'the ratio as 4 to 1, 
 the resistance in the leak is made still lower (thus enabling it to 
 carry a greater portion of the current away from the line), and at 
 the same time the "added resistance " is increased. Of course these 
 resistances must be so chosen that whether the transmitter is open 
 or closed the total resistance of the line will be the same. 
 
 In actual practice the figures adopted for obtaining the ratio of 
 4 to 1 are, for the "leak," 800 ohms, and for the " added resistance," 
 1,800 ohms. It will be found that the joint resistance of 2,400 
 ohms (viz. : the 600 ohms at the machines and the 1,800 ohms 
 " added resistance ") and the resistance of the " leak," 800 ohms, is 
 600 ohms. 
 
 Since it is desirable to have the variable resistances under con- 
 venient control in order to work lines of any length or resistance 
 upon the same set of instruments, the added resistance and leak 
 resistance are usually placed in the same box, as shown in Fig 37. 
 
 Fig. 37. 
 
 The leak consists of two resistance coils of 100 and 800 ohms, 
 respectively, arranged in such a manner that by inserting a metallic 
 plug at L the 100 ohm coil may be cut out. The added resistance 
 coils measure 600 and 1,200 ohms, respectively, arranged so that 
 the 600 ohm coil may be cut out by a plug inserted at R. Thus 
 
THE QUADRUPLEX. 
 
 69 
 
 the leak may be readily changed from 800 to 900 ohms, and the 
 added resistance from 1,200 to 1,800 ohms. 
 
 When machine currents are used it is convenient to have some 
 means of quickly opening the circuits between the machines and 
 the home instruments, not only to facilitate any repairs or altera^ 
 tions that may require to be made, but also to cut out the instru- 
 ments in case of short circuiting of the wires at any point. For 
 this purpose each quadruplex set is furnished with a switch similar 
 to the one shown in Fig. 38. The brass strips B and E" are con- 
 nected at their top ends directly with the dynamo machines, as may 
 
 be seen in Fig. 40, while the strip B' is connected to the line. The 
 lower ends of B and B" are led to the posts of the pole changer. 
 By moving B and B" to the right they are placed on the dead 
 points, Z>, Z>', and of course the machine currents are thus taken off 
 the instruments. By placing the strip B' on the point 6r, the line 
 is placed to the ground through a resistance equal to the resistance 
 of 600 ohms at the machines. The dotted lines represent the con- 
 nections between the buttons under the wood of the switch. 
 
 Fig. 39 is a diagram of the quadruplex arranged to be operated 
 at station A by dynamo machine currents, and at B by the usual 
 gravity battery currents. It will be seen that the latter's connec- 
 
70 
 
 THE QUADRUPLEX. 
 
THE qiTADRUPLEX. 
 
 71 
 
72 THE QUADRUPLEX. 
 
 tions remain exactly the same as when both ends of the circuit are 
 supplied by gravity batteries. 
 
 There is no necessity for a further explanation of the operation 
 of this arrangement, as the effects produced are precisely analogous 
 to those already stated in the description of the regular quadruplex. 
 
 In Fig. 40 may be seen the actual connections required by the 
 arrangement under consideration, the only parts omitted being 
 the local sounder of the polarized relay and the resistances that are 
 interposed between the dynamo machines and the switch S. 
 
 Each instrument is plainly indicated by its name. 
 
 It only remains, therefore, to point out the route of the current 
 in the different positions of the transmitter, and these are located 
 by consecutive figures and letters in the diagram. For instance, 
 when the transmitter is closed, as in the diagram, the route is 
 shown from the dynamos by figures 1 up to 8, where the current 
 splits, as heretofore explained. 
 
 When the transmitter is open, the current divides at A, and that 
 portion which proceeds to the line passes through the added resist- 
 ance to 5, thence to D, where it again splits, as in the former case. 
 That portion of the current which seeks a ground through the leak 
 passes from A to the tongue of the lever, thence to (?' and to the 
 ground. 
 
IX. 
 
 THE PRACTICAL WORKING OF THE QUADRUPLEX. 
 
 BY WM. HAVER, JR. 
 
 After first-class instruments, perfect connections, and well kept 
 batteries, nothing conduces so much to the successful working of 
 the quadruplex as intelligent handling. 
 
 By intelligent handling, 1 mean that the persons in charge of a 
 quadruplex set or sets should have a thorough theoretical and prac- 
 tical knowledge of the entire system. The theory may be mastered 
 by study, but the practical knowledge can, as a rule, only be gained 
 by experience. 
 
 It is not, however, always practicable to have an experienced 
 person in charge of a system, for it frequently happens that a quad- 
 ruplex circuit is installed in an office and placed under the care of a 
 manager or chief operator who may never have seen a quadruplex 
 set before. 
 
 An enterprising manager, in such a case, will, of course, at once 
 endeavor to post himself in all the details of the system, but these 
 are more numerous than some people suppose, and while the man- 
 ager is acquiring his knowledge the best results cannot be expected 
 from that particular quadruplex system. 
 
 In the hope that my experience may be of service to this class in 
 the management of their quadruplex, as well as to others who may 
 wish information on this subject, I propose to jot down as many ot 
 the causes of trouble, and the symptoms occasioned by certain kinds 
 of trouble, as I can recall, together with any hints as to the manner 
 of obtaining the best results from the system, that may occur to me. 
 
 The following are some of the technical terms and expressions 
 -used in speaking of the quadruplex and its connections : 
 
 The " tap " or " tap wire " is that wire which taps the battery and 
 
74 THE QUADRUPLEX. 
 
 is led to the bar of the transmitter. It may be seen at point Z in 
 Fig. 28. The "short end" of the battery is that portion between 
 the ground and Z in the same figure. The short end is generally 
 to the line when the transmitter is open. The " long end " of the 
 battery refers to the portion between the points P and F, also in 
 the same figure. This is referred to chiefly when there is apparent 
 trouble in that portion of the battery, as explained hereafter. To 
 " reverse " means to open and close the pole changer, which instru- 
 ment, as is well known, reverses the battery at each motion. 
 
 The " ground coil " is that resistance which is placed between 
 the three point switch B in Fig. 30, and the ground, to compensate 
 for the internal resistance of the entire battery. It is sometimes 
 called the "spark coil." 
 
 The " polar relay " is the No. 1 relay. The " neutral relay " is 
 the No. 2 relay. 
 
 To " dot " or " write on both sides " is to write on the pole 
 changer and transmitter simultaneously. 
 
 Care should be taken that the compensating battery resistances, 
 such as that between the tap of the battery and the bar of the 
 transmitter, are correct. 
 
 It may be assumed that two and one-half ohms per cell is the 
 average internal resistance of a battery cell. 
 
 On circuits of, say 100 miles, it is common to divide the battery 
 into two equal portions, namely, (supposing that there are 100 cells 
 in the battery,) of fifty cells each. 
 
 Thus when the transmitter is open fifty cells are placed to the 
 line, and when it is closed, 100 cells. This is called giving a margin 
 of two to one. 
 
 On longer circuits the margin is made wider, until on some cir- 
 cuits, such for example as that between Bangor, Me., and New York, 
 it is four to one, that is to say, (supposing that there are in all 400 
 cells of battery,) there would be 100 cells to the line when the 
 transmitter is open, and 400 when closed. 
 
 This increase of the margin is rendered necessary by the fact 
 that on long lines there is considerable retardation and prolongation 
 
THE qUADRITPLEX. 75 
 
 of a current, and it is thus essential that the difference of potential 
 between the entire battery and the " short end " of it should be 
 greater, so that the variations in the amount of current reaching 
 the distant neutral relay may be more marked. 
 
 The method of balancing a quadruplex has already been ex- 
 plained in the foregoing chapters, but it is a subject which I am sure 
 will bear amplification. 
 
 As already stated, the distant station, which for the sake of brevity 
 I shall frequently hereafter call -Z?, is first requested to "ground." 
 Having done so, he should avail himself of the opportunity to see 
 that the number of ohms resistance in his " ground coil " is correct. 
 The ground coil is that resistance which is interposed between the 
 line and the ground, when one " grounds." As previously men- 
 tioned, the ground coil is put in to compensate for the " internal " 
 resistance of your battery. It very often happens that an inexperi- 
 enced "balancer" will unwittingly alter this resistance, to the seri- 
 ous detriment of your balance. For instance, if the internal resist- 
 ance of a battery is 600 ohms, and by inadvertence 1,200 has been 
 inserted, it will divert an unequal portion of your battery current 
 to the line, and thus render your balance unsteady. Of course it is 
 equally necessary that you should examine your own ground coil. 
 
 When B has grounded, you also ground. 
 
 The next thing to be done is to place the armature of your polar 
 or No. 1 relay in the "center," that is, so to adjust it that it will 
 remain on whichever side it is placed. You then " cut in " your 
 battery and change the resistance in your rheostat until the lever 
 again remains where it is placed. 
 
 Sometimes it happens that, owing to heavy inductive effects on 
 the wire, it is impossible to find this center, the armature keeping 
 up a constant clatter. In this case you must use your judgment 
 and adjust the armature in that position where the attraction to 
 either pole of the relay is about even. You then, as before, place 
 your battery to the line. If the wire is much out of balance, the 
 armature of the polar relay will now be strongly attracted to one 
 side, in that case, alter the rheostat resistance again. As you ap- 
 
76 THE qUADRUPLEX. 
 
 proach the balance, the inductive effects of the line will again begin 
 to be felt, and again your judgment must be exercised as to when 
 you have a nearly perfect balance. When this is obtained, request 
 the distant station to " cut in " his battery ; when he does so you 
 will very likely find that the inductive effects have disappeared, 
 having now been overcome by the distant station's battery. 
 
 On wires not exceeding from 70 to 100 miles the static discharge 
 from the line is scarcely sufficient to interfere with the working of 
 Morse circuits, therefore, on such wires condensers are unnecessary. 
 
 A few words more, here, explanatory of the static discharge or 
 return current, may be of use to the beginner. 
 
 To illustrate, let us suppose that we have a pipe between any two 
 points, somewhat remote, and let us further suppose that we have 
 at one end of this pipe an engine pumping water through it. Also 
 let us suppose that we have at each end of the pipe a wheel, which 
 will be revolved in a certain'direction accordingly as the water flows. 
 Now if we suddenly stop the pumping engine, leaving at the same 
 moment both ends of the pipe open, it is quite evident that the 
 water in the pipe will empty itself at both ends, and the water at 
 the end where we had supposed the engine to be will run out in an 
 opposite direction to that in which it had been flowing into the 
 pipe, and if there is enough water so left in the pipe, as there 
 probably will be if the pipe is sufficiently long, this returning water 
 will reverse the direction of the wheel. 
 
 If we could imagine that the sides of this pipe were of a spongy 
 nature, it would be easy to believe that the capacity of the pipe for 
 holding water would be increased, and, consequently, the greater 
 would be the outflow of the return current when we should sud- 
 denly stop the engine, and the greater would be the effect of this 
 outflow upon the wheel. 
 
 Now the foregoing action of water in a pipe is partly analogous 
 to the action that takes place in a telegraph wire. The battery 
 supplies the electric current on the wire. The action of induction, 
 previously explained under the heading of condensers, gives to the 
 wire a greater capacity for retaining electricity. The pole changer 
 
THE QUADRUPLEX. 77 
 
 momentarily cuts off this supply of electricity. The electricity left 
 in the wire at that moment seeks to return to the earth, and in 
 doing so temporarily charges the relay, giving a short false signal. 
 
 The condenser obviates this trouble in the manner already ex- 
 plained. It should be stated that the coils in the rheostat which 
 form the artificial wire have a very small capacity for retaining 
 electricity, and hence, although they offer, when properly adjusted, 
 as much resistance to the electric current as the main line, do not, 
 at the moment when the battery is cut off, send back anything like 
 so large a return 'current. 
 
 The condenser being placed in the artificial line, as shown in the 
 diagrams, when rightly adjusted gives that wire an artificial capacity 
 equal to that of the main line, and as the condenser discharges at 
 the same moment as the main line, but goes around the core in an op- 
 posite direction to the line discharge, the home relay is not affected. 
 
 After having balanced the quadruplex, it now remains to adjust the 
 condenser so that it shall give a discharge equal to that from the 
 line. This requires a greater degree of skill than simply to balance 
 a quadruplex and almost every expert has his own way of proceeding. 
 I have found that I could always eliminate the static effects by simply 
 closing the No. 2 key, (so that the full battery will be to the line,) 
 and dotting on the pole changer, in the meantime adjusting the 
 condenser, by taking out or putting in plugs, until the motion of 
 the polar relay in response to my pole changer was reduced to a 
 sort of a light double click. 
 
 So long as the polar relay responds clearly to each motion of your 
 pole changer, the static balance may be considered imperfect. 
 
 Another method for adjusting the condenser is, after balancing, to 
 give the adjusting screw of the polar relay a bias to one side or the 
 other until a well defined click is heard, then to insert or remove 
 plugs in the condenser (at the same time bringing back the arma- 
 ture of the polar relay to its former position) until this click is 
 eliminated. 
 
 I consider that this method has a defect, especially in the hands 
 of beginners, as it alters the " center " of the polar relay armature 
 
78 THE QVADRVPLEX. 
 
 and may thus interfere with the satisfactory working of the polar 
 side. 
 
 Another method, and a very good one I think, is, after balancing, 
 to request B to put his battery on, and to so place his key that your 
 polar relay armature rests on the local contact point. See that 
 your own full battery is also to the line. Now dot on your pole 
 changer and adjust your condenser until it has no effect whatever 
 on your polar relay's armature. If the static discharge has no effect 
 on your polar relay under the above circumstances, it will not be 
 likely to effect it under any other circumstances. 
 
 The reason I advise noticing that your full battery is to the line, 
 when you are adjusting for " static," is, that the static discharge 
 varies with the amount of battery to the line. The greater the 
 amount of battery to the line the greater will be the return dis- 
 charge. 
 
 After this has been done and the distant station has also balanced, 
 the following test should be made to ascertain whether the adjust- 
 ments are correct : 
 
 Ask B to dot on his No. 1 side and to write on the No. 2 side. 
 Now listen to see that B's writing on No. 2 side is coming all right, 
 at the same time open and close your own pole changer and trans- 
 mitter. If you get his writing clearly under these circumstances, 
 which of course is equivalent to having all sides working at once, 
 your balance is correct. 
 
 The distant station should then go through the same performance. 
 
 If it should transpire that while B is writing on both sides you do 
 not get him clearly on No. 2 side, after you have carefully adjusted, 
 request him to write on No. 2 side only. If, while he is so writing, 
 his pole changer should be open, and you get him O. K., next ask 
 him to close his pole changer and to continue writing on No. 2 side. 
 If you also get him O. K. with pole changer closed, it would appeal 
 that the trouble is in his pole changer's points, and in that case an 
 examination of them will probably reveal that the trouble was 
 caused by dirty points, or uneven contacts. 
 
 When your No. 2 side is broken up by the movements of B's pole 
 
THE qiTADRUPLEX. 79 
 
 changer, it is technically said that his " reversals " break up or con- 
 fuse No. 2 side. 
 
 A small file, adapted to go under the contact points of the pole 
 changer and transmitter, is a tool every manager of a quadruplex set 
 should possess. 
 
 If, after B has cleaned his points, the trouble be not removed, try 
 your balance again carefully. This is especially necessary in bad 
 weather. 
 
 If you get B's writing " light " on either side, ask him to clean 
 his points on both sides. Should he complain that you come 
 " light," likewise clean your points " all around." When the con- 
 tacts become oxidized it is well known that this increases the re- 
 sistance at the points of contact, and consequently the strength of 
 current is reduced. 
 
 If, while B is dotting on the pole changer, you are unable to read 
 him on No. 2 side, again get him to write on No. 2 side only, and 
 request him to give you one pole at a time. This he does by first 
 closing and then opening his pole changer, as you direct. If you 
 find that when he closes his pole changer, for instance, you do not 
 get him at all on No. 2 side, while you do get him while his pole 
 changer is open, it indicates that one of the tongues of his pole 
 changer is not making connection ; call his attention to that fact. 
 
 If while B is dotting on No. 2 side you cannot get him readably on 
 No. 1 side, likewise ask him to give you his full battery and his short 
 end alternately. This he does by first closing his transmitter (giv- 
 ing you all of his battery). If now you do not get him on polar 
 side, next ask him to give you his short end, which he does by open- 
 ing his transmitter ; if now you get him, it indicates that there is 
 trouble between the " tap " of his battery and the post of his trans- 
 mitter, probably a bad or broken connection in battery ; or it may 
 be that the tongue of his transmitter is not making proper connec- 
 tion with the post of transmitter. 
 
 Should the trouble be that you do not get him when his trans- 
 mitter is open, while you do when it is closed, the trouble would 
 appear to be that there is either a bad connection between where 
 
80 THE QUADRUPLEX. 
 
 his battery is tapped and the bar of his transmitter, or that per- 
 haps the tongue of the transmitter is not making proper contact 
 with the bar of the transmitter. 
 
 By reference to the diagrams of the quadruplex', Fig. 28, it will 
 be seen to what I allude. 
 
 If it happens when all sides are working that you do not get B 
 well on No. 2 side, notice carefully whether you do get him all right 
 on that side when your No. 1 side is quiet. If so, the trouble, most 
 likely, is caused by an imperfect static balance at your end. If you 
 find you do not get B well on No. 1 side when your transmitter is 
 closed, but that you do when it is open, this trouble, too, is proba- 
 bly occasioned by bad static balance. 
 
 If you find that your condenser seems to have no effect on 
 the static discharge, in the way of removing it, examine the con- 
 nections of condenser ; it may be that they have become detached. 
 It is as bad to have too many of the condenser plates in action, as 
 too few. Therefore such an occurrence should be avoided. 
 
 It occasionally happens that when one wire becomes crossed with 
 another, one of the wires is cleared by opening the other at both 
 ends. 
 
 If the wire so cleared should be a quadruplex or duplex wire, the 
 effect on the static discharge will be at once noticeable, namely, it 
 will be much increased. It is in some cases necessary to add another 
 condenser to offset this abnormal discharge ; an additional amount of 
 resistance placed between the condensers and the artificial wire may 
 also be thus rendered necessary, as the static discharge in this case 
 is somewhat more gradual than ordinarily. 
 
 When a wire has not been balanced since the previous day, and if 
 in the meantime the weather has considerably changed, or if there 
 has been a decided change in the weather at any time, it may occur 
 that you will not hear B calling or answering your calls, the wire 
 being so much out of balance. In this case it is a good plan to call 
 B a few times, and then suddenly ground your end. If B is there, 
 and answers you, you will be sure to hear him now. If so, cut in 
 again, ask him to ground, and proceed with your balance as usual. 
 
THE QUADRUPLEX. 81 
 
 When there has been a change in the weather from dry to wet, it 
 will probably be necessary to lower, very considerably, the tension 
 of the spring of No 2 relay, before you will be able to get B. This 
 is because the margin of his current reaching you has fallen in pro- 
 portion to the state of the weather, as well as the condition of the 
 wire insulation. 
 
 On the contrary, if the weather changes from wet to dry, it will 
 be found necessary to increase the tension of the No. 2 relay, the 
 margin from B having increased. 
 
 After a balance has been taken on the quadruplex, it is, generally 
 speaking, best not to alter the adjustment of the polar relay under any 
 circumstances. I am aware that some quadruplex experts favor giv- 
 ing the armature of this instrument a bias to the local contact point, 
 but I do not think the theory of the quadruplex will justify this 
 action. It should be considered that the object of placing the arma- 
 ture in the " center," before balancing, is to make certain that the 
 amount of current flowing through each coil from the home bat- 
 tery shall be equal ; and that the quadruplex is actually worked by 
 the production of inequalities of the currents flowing through these 
 coils, and thus, that any bias to one side, which may be given to the 
 lever, is equivalent to sending more current through that side from 
 the home battery. Furthermore, the fact of the lever being exactly 
 in the center is relied upon to keep it perfectly still when the dis- 
 tant battery is momentarily cut off during the changes of polarity 
 at the distant station, for at that instant, as we know, the induced 
 magnetism of the polar relay comes into play and retains the lever 
 at whichever side it may happen to be at that moment. Conse- 
 quently, if the lever is biased, it is possible that should the lever 
 happen to be on the unbiased side it might tend to move to the 
 other side and thus make a " click." 
 
 It is true that if but one side of the quadruplex is working, a 
 change in the balance may be tided over by simply changing the ad- 
 justment of the levers, but this is only a makeshift at best. Should 
 writing from B come light on the polar side, the true remedy does 
 not consist in biasing the armature. First ascertain that the 
 
82 THE QUADRUPLEX. 
 
 trouble is not in a weak local, dirty contact points, bad adjustment 
 of sounders, or in your balance ; after that, request the distant sta- 
 tion to diminish the tension of the spring on his pole changer. 
 If none of these resorts obviates the trouble, and the batteries and 
 wire are all right, it might be well to try what result would ensue 
 from a change of the sending operator. It has often been noticed 
 that some operators do not seem to have the correct method of 
 sending signals through on duplex or quadruplex circuits. 
 
 Where a repeating sounder is used on the No. 2 side, trouble 
 is sometimes caused by its contact points sticking, or becoming 
 loose ; they should be looked after. 
 
 In balancing you may sometimes find it impossible to get " a 
 balance," and that no amount of changing of the rheostat resistance 
 produces any effect on the relay. This, however, does not occur 
 very frequently. It may be owing to several causes. For instance, 
 the polar relay may have become defective ; burned out, as it is 
 called, by an unusually strong current, such as lightning, etc., or the 
 rheostat may have failed, some of the coils thereof having broken, 
 perhaps; or the trouble may be in the condenser. This would 
 be the case if a ground plate should touch a line plate. This may 
 be tested for by temporarily detaching the wire from the condenser. 
 
 The relay may be tested by taking out all the wires from its 
 thumb screws but the battery wire, and by feeling with wet fingers, 
 or a piece of wire, to ascertain whether the current comes through to 
 any of the other thumb screws than the one it naturally should 
 come to. The permanent magnet should also be tested to ensure 
 that it is not crossed, by way of the core of the electro-magnet 
 of the relay, with either of the coils. It is understood, of course, 
 that if the fingers are used for this test, the other hand must be 
 touching a ground wire, or if a wire be used, one end of it must be 
 to ground. The wire is the best, as it will show, by a spark, a lower 
 current than the fingers would detect. 
 
 The neutral relay may also be affected and tested in the same 
 way. 
 
 If the current should be found in other than the thumb screw 
 
THE QUADRUPLEX. 83 
 
 which is the terminus of the coil attached to the battery as above, 
 it shows the relay to be defective, probably the coils are crossed 
 with each other. 
 
 Another source of trouble is the crossing of a portion of one coil 
 with another portion of itself, which would make the resistance in 
 the crossed coil of the relay much lower than the uncrossed one, as 
 the wire between the crossed portions of the crossed coil would be 
 virtually cut out. 
 
 The" effect of such an occurrence is this : Suppose that it is the 
 line coil which is crossed with itself. In consequence of this cross 
 there is less aggregate current flowing in that coil around the 
 relay core than througli the rheostat coil. It will thus follow that 
 the strength of current flowing through the line coil must be in- 
 creased in order to enable us to balance the relay lever. To produce 
 this result we shall find that we will have to largely increase the re- 
 sistance in the rheostat to divert a greater portion of the current to 
 the line ; but, of course, a quadruplex will not work well on a bal- 
 ance obtained in this way. 
 
 When an abnormal balance, such as the above, is obtained, it is 
 a good plan to ask the distant station how many ohms resistance he 
 obtains a balance with. If his balance is normal you may suspect 
 either of your relays, and examine them, although it does not follow 
 always that the trouble is there. 
 
 It is presumed that every person in charge of a circuit knows the 
 normal resistance of the wire under his charge. Therefore, if the 
 balance should give about twice the usual resistance, it is likely that 
 the ground coil at the distant end has become detached, and in con- 
 sequence you are not only balancing against the line but also his 
 rheostat, through which you are getting your "ground." This 
 trouble is comparatively easy of detection. 
 
 When it happens that B does not get any battery from you on a 
 certain . quadruplex, and has informed you to that effect, via 
 another wire perhaps, it is best, first to ascertain whether your 
 battery reaches the switch, testing for it in both positions of your 
 pole changer and transmitter. If it does not, then the fault is, of 
 
84 THE QUADRUPLEX. 
 
 course, located between the switch and the quadruplex battery. 
 Next test for the battery at the screw-post of the polar relay from 
 which the wire runs to the switch ; next at the screw-post of the 
 same instrument where the line wire enters the relay ; next at the 
 screw-post of the No. 2 relay whence the line wire makes its exit, 
 and afterwards at the screw-post at which the line wire enters 
 that instrument. Suppose that at this latter post you feel battery, 
 but do not at the exit screw-post, in such an event it locates the 
 fault between these two points. It is probably a broken wire under 
 the base board of the instrument, one of the fine wires. But, 
 wherever the trouble may be, it is most quickly and certainly found 
 by this point-to-point test. A correct diagram of the connections is, 
 therefore, very essential to enable those who cannot carry the con- 
 nections in mind to follow out the route of the wires. 
 
 It should be borne in mind that if the trouble is occasioned by 
 the current having found a short route to the ground, as sometimes 
 happens, it will be almost impossible to locate the trouble with the 
 fingers, as the metallic route to the ground offers so little resist- 
 ance to the current, in comparison with that offered by the body, 
 that not enough current will pass by the body to make itself felt. 
 
 When " deep seated " trouble has overtaken a quadruplex cir- 
 cuit, and it is uncertain at which end of the wire the trouble is (the 
 wire having, of course, been changed for one known to be reliable, 
 so that the fault cannot be laid to it), and each station claiming 
 that the fault or " bug " (as a fault on multiplex systems is now 
 generally technically termed), is at the other end, the best plan, by 
 all odds, is to ask the distant, station to test your quadruplex set 
 against another set in his office, if he has more than one set. If 
 not, test your set against a set known to be free from trouble in 
 some other station. Should the same trouble still appear in your 
 set, that, of course, proves conclusively that the trouble is in your 
 set. If your set now works O. K., however, it as certainly shows 
 that the seat of the fault is in the distant quadruplex, and the 
 manager thereof should be informed of the result of said test ; 
 whereupon, it is to be supposed that he will set about with renewed 
 
THE QUADRUPLEX. 85 
 
 vigor to locate the fault, and generally with better success. You 
 need not be surprised, however, after the trouble has disappeared, 
 should the distant station calmly ask you, notwithstanding the cir- 
 cumstances, "What did you find? " The importance of keeping all 
 the locals connected with a quadruplex set in first-class condition, 
 and all the local contact points well cleaned, cannot be mentioned 
 too often. The locals should not be allowed to run down until 
 they actually fail, before renewal, nor should the local contact 
 points be permitted to remain unfiled until they actually stick. 
 
 It has frequently happened that quadruplex sets have been ren- 
 dered virtually useless for days at a time, the trouble being attrib- 
 uted to main line or main battery faults, instead of which it has 
 been an unheeded weak local or faulty local contact point. 
 
 It would be impossible to enumerate all the different sources of 
 trouble which are likely to arise on a quadruplex system. Those 
 that I have alluded to are among the most frequent. 
 
 It would undoubtedly add much to the successful handling of 
 the quadruplex if every one connected with it would frankly state 
 the cause of trouble, when he has found it, to the operator or per- 
 son in charge at the other end. The latter has seen the symptoms 
 occasioned by the trouble. If he were at once notified as to the 
 exact nature of the fault, he would be able at once to direct the dis- 
 tant station to the seat of trouble, when next the same symptoms 
 should manifest themselves, and thus time would be saved. Instead 
 of so doing, however, it has become almost a rule for every man- 
 ager to hide any trouble happening in his own set, and merely to 
 say that probably a loose wire " shook out," or something equally 
 vague. This pernicious custom is probably due to the fact that some 
 superintendents require so many explanations as to the reason why 
 such and such trouble was not obviated, or not discovered sooner, 
 that a manager rarely allows a fault to be definitely located in his 
 office more than once. It is thus, perhaps, an open question whether 
 it is really to the benefit of the service that managers, who, as a 
 rule, are not apt to willfully allow trouble to occur, or to sit idly 
 by when it has occurred, should be held so strictly to task. 
 
TELEGRAPH REPEATERS. 
 
 BY WM. MAVEK, JR. 
 
 After Columbus had shown the wise men of Spain how an egg 
 could be made to stand on end, they doubtless conceded that it 
 was a very simple feat, although they had previously spent some 
 time and thought in the attempt to solve it ; and it is likely that 
 had any one ventured on a second explanation of the mystery, the 
 wise men would have considered it quite unnecessary. 
 
 In like manner it is possible that to those who, by study or 
 explanation, have familiarized themselves with the principle and 
 operation of the standard repeaters, an article on that subject at 
 this late date may appear superfluous ; but to such as have not 
 been shown how the egg may be made to stand on end, so to speak, 
 the subject is as obtuse as it ever was. 
 
 It is very questionable whether, on an average, one out of ten 
 operators throughout the country is familiar with the principle or 
 mode of operation of any of the repeaters now in use. Indeed, the 
 writer has been told by several of his friends that they have found 
 the action of repeaters as difficult to comprehend as that of the 
 quadruplex. 
 
 This is, no doubt, an exaggerated statement, but it is nevertheless 
 true that to many the manner in which the repeater performs its 
 functions is a mystery. 
 
 This condition of affairs is probably due in many cases to a total 
 lack of interest in the matter. Again, it may be owing to a habit 
 which some inventors or draughtsmen have of displaying their 
 diagrams to the very worst advantage, with regard to clearness, so 
 that to a novice it becomes an almost insurmountable task to follow 
 out the labyrinth of lines, crossed in some places so carelessly that 
 it is difficult to determine whether a joint is not meant ; and in 
 
REPEATERS. 87 
 
 other places so poorly jointed that it becomes a question whether 
 the lines are not meant to cross each other there. In other diagrams 
 the student traces a wire until it is brought up to the screw post of 
 a certain instrument, and there it stops short. He sees that several 
 other wires lead out of that instrument by other screw posts, but 
 which of those is the one he wishes to trace is a discouraging 
 conundrum. 
 
 These, at least, are some of the difficulties which met the writer 
 in his search after knowledge on this and like subjects, and, con- 
 sequently, in the following descriptions he will strive to avoid fall- 
 ing into similar errors. 
 
 In the days of stage coaching, as we are all aware, a pair, or may- 
 be two pairs, of horses would be detailed to draw the coach to a 
 certain station, having reached which point they were relieved by a 
 fresh set of horses termed a relay. The stage was then drawn by 
 these horses to the next relay station, when another change of 
 horses would be made, and so on. 
 
 The obvious reason for these changes or relays was that the 
 horses had become fatigued, and although they might have gone 
 farther, the work would not have been done so quickly or so well 
 as by the fresh set of horses. 
 
 Now, what the relay of horses did for the stage coach, repeaters 
 practically do for the telegraph message. They take it off the 
 hands of a tired wire, if such an expression may be used, and pass 
 it on with fresh vigor to another wire. 
 
 It is a question whether the name " repeater," as applied to the 
 repeater of to-day, is not a misnomer, or at least somewhat inappro- 
 priate. The instrument which we know as a " relay " is, perhaps, 
 more correctly a repeater. It repeats to the sounder what is pass- 
 ing on the main line. The instruments composing an automatic 
 repeater do relay the message from one wire to another, which is 
 exactly what the relay did when its present name was given to it 
 by Morse, as will be shown later. 
 
 Repeaters are rendered necessary on telegraph wires, because it 
 is impossible to insulate a wire so perfectly that a portion of the 
 
88 REPEATERS. 
 
 electricity on the wire will not escape at each point of insulation. 
 Hence a certain amount of electricity passes to the ground at every 
 pole along a line, and consequently the longer a line may be, the 
 more electricity will escape to the earth in this manner, so that on 
 long lines (or even on short lines in wet weather, for then each 
 insulator becomes comparatively a better conductor, and so permits 
 more current to escape,) the current that reaches the distant end 
 becomes quite feeble. But, even if the question of imperfect insu- 
 lation were set aside, the operation of Ohm's well-known law of 
 electricity, viz., that the strength of current on a line is equal to 
 the electro-motive force divided by the resistance of the wire, would 
 confine the successful working of a circuit within certain limits. 
 Therefore it is found advantageous on circuits over a certain length 
 to insert midway, or at stated intervals," sets of repeaters, with fresh 
 batteries, etc. The intervals at which repeaters are placed of course 
 depend to a great extent on the conducting capacity of the wire used. 
 
 Professor Morse at first had an idea that the best effects could be 
 obtained from an electro-magnet (designed for telegraphic pur- 
 poses) by winding it with wire of the same size and material as that 
 used on the line. The consequences were that the instruments 
 were not only very unwieldy, but very poor magnets. It is easy 
 to imagine that such was the case when it is stated that the first 
 pair of magnets made for Professor Morse weighed, including attach- 
 ments, three hundred pounds each. Mr. Charles T. Smith, now 
 employed by the Western Union Telegraph Company in New York 
 city, who was connected with the Morse telegraph system in its 
 earliest days, and to whom I am partly indebted for this interest- 
 ing information, subsequently made two electro-magnets coiled with 
 a somewhat smaller gauge of wire than those just mentioned. The 
 coils of Mr. Smith's magnets were first covered with gum shellac, 
 and afterwards with a coating of thick glue. They only weighed 
 seventy-five pounds each. 
 
 Professor Morse very early in his experiments discovered that he 
 could not generate sufficient magnetism to operate his electro-mag- 
 nets at a greater distance than from fifteen to twenty miles, the 
 
REPEATERS. 89 
 
 cause of which, as we know now, and as Morse afterward learned, 
 was that his electro-magnets were wound with such coarse wire 
 that there were comparatively few convolutions around the core of 
 the magnet, and as the strength of an electro-magnet increases with 
 the number of convolutions of wire around the core as well as with 
 the strength of current flowing through the coils, the resultant mag- 
 netism in the professor's first instruments was very feeble. The 
 idea struck him, however, that while he could not satisfactorily 
 operate his magnet on a line of any great length, yet he might cause 
 a magnet operated at a limited distance to transmit into another 
 wire of the same length, and then cause that second wire to repeat 
 into a third wire, and so, as he thought and said, in this way en- 
 circle the globe with a wire. 
 
 Fig. 41 shows the manner in which the professor proposed to pro- 
 long his circuit. It is seen that line A relays into line B by means 
 of the armature of R', and B into C by the aid of R", and so on. 
 But by this arrangement it will be observed that there is no way 
 of sending messages back from D to A. To overcome this, Mr. 
 Morse used another wire, as seen in Fig. 42, giving R f in that fig- 
 ure control of line B, etc. The electro-magnets which performed 
 that service Mr. Morse properly termed relays, and that name all 
 line electro-magnets have since retained, so that when automatic 
 relaying instruments were invented, the word repeater was probably 
 applied to them as a distinguishing name. A working model of this 
 method was publicly shown in the University of New York, but 
 it never went into practical operation. Mr. Morse, however, 
 adapted this invention to the work of operating the secondary or 
 local circuit, and the manner of operating the relays was maintained 
 a profound secret for a long time, the magnet and its connections 
 being enclosed in a locked box. 
 
 THE BUTTON REPEATEE. 
 
 In 1846, the Woods button repeater was introduced; but in the 
 meantime, by the use of improved electro-magnets, the distance at 
 which relays could be practically worked had much increased. 
 
90 
 
 REPEATERS. 
 
REPEATERS. 91 
 
 Fig. 43 is a diagram of an improved form of button repeater now 
 in use. 
 
 Before describing its action let us first investigate the necessity 
 for such an instrument, and for this purpose the reader will please 
 consider that for the present the dotted lines D D D' D', and the 
 button switch B R are out of the circuit. 
 
 By examining the diagram it will be seen that line No. 1 runs 
 through relay R' and is led through the armature of the repeating 
 sounder R S' to the battery B' and ground ; and that line No. 2 
 goes through R" and through the, armature of R S to its battery B 
 and ground. By this arrangement R', having control of R S by 
 means of the local circuit, should, it would seem, be able to open and 
 close No. 2 circuit at the point x at pleasure, and in like manner 
 R", having control of R S', should be able to close and open 
 No. 1 circuit. But it is not so, for let us suppose that an operator 
 on No. 1 circuit opens his key ; this allows the armature of R' to 
 fall back, which in turn opens R S, and its armature, being drawn 
 up by the retractile spring, opens the No. 2 circuit at x. This 
 action opens relay R", which in turn opens R S', and the latter 
 breaks the circuit of No. 2 line. This is all right so far, appar- 
 ently, for we have both circuits open, but the operator on No. 1 line 
 now wishes to close his circuit, and consequently closes his key, 
 but that does not close his circuit, as it is still open at R S', and 
 thus No. 1 circuit remains open, precisely as if an operator at a 
 receiving station had his key open, and in that case No. 2 circuit is 
 also kept open ; hence a deadlock must ensue. This was probably 
 the reason why Morse suggested the use of two wires for his system 
 of relaying. At all events it is a sufficient reason. 
 
 Now, to describe in what manner the button repeater overcomes 
 this trouble, let us consider that the dotted lines D D D' D' and 
 the button switch B R are now in the circuit. "We will assume 
 that No. 2 line wishes to send into No. 1. 
 
 For this purpose the operator attending the button repeater at 
 the repeating station A turns the button to the left as shown in the 
 figure, so that a strip of metal indicated by the short black lines on 
 
92 
 
 REPEATERS. 
 
REPEATERS. 93 
 
 the handle of the button switch joins the metal points 1 and 2 
 together. This gives line No. 2 two routes to battery B and 
 ground, one via the dotted lines D D, and the other via the arma- 
 ture of R S. Again assume that No. 2 opens his key ; the 
 armature of R" falls back and opens the local circuit of R S', 
 which breaks line No. 1 at x', and this again opens R', which of 
 course opens R S and breaks line No. 2 at a;, 'but the route via D 
 D remains intact, so that No. 2 can close his own circuit when 
 next he closes his key, which in turn closes R", which closes the 
 local circuit of R S', and this again closes N6. 1 circuit. Thus, as 
 long as the button switch is held in that position, No. 2 can repeat 
 into No. 1. Should No. 1 wish to break No. 2, the attending opera- 
 tor at A turns the button so that it will cover contact points 3 and 
 4, which gives No. 1 the same control over No. 2 as the latter pre- 
 viously had over the former. 
 
 One advantage that this form of button repeater has over some 
 others is, that when the handle is directly in the center, each line 
 may be worked from station A as independent circuits. 
 
 Button repeaters are also useful when it is necessary to send 
 copies of the same press reports over two different circuits from a 
 central circuit, J., for instance, in the diagram. When used for 
 this purpose a key is inserted in either No. 1 or No. 2 circuit at A, 
 and thus the operator at A, by moving the handle to the proper 
 side, can transmit into both circuits. 
 
 AUTOMATIC REPEATERS. 
 
 In the description of the button repeater, Fig. 43, it was shown that 
 it was necessary to keep the circuit of the sending office intact at 
 the repeating station, at the same time that the said sending office 
 was enabled to repeat into another circuit by having control of a 
 repeating sounder, through the armature of which the wire of the 
 circuit to be repeated into was led. We saw in the explanation of the 
 button repeater that this object was accomplished by means of the 
 metal strip on the handle of the button switch, and the dotted-line 
 circuit, which insured a continuous circuit to the sending office. 
 
94 REPEATERS. 
 
 This plan for repeating, however, while quite successful, had a 
 bad feature, inasmuch as it required the constant presence of an 
 operator at the button repeater to watch for breaks and to turn the 
 handle at the right moment. It will therefore be readily surmised 
 that an instrument, or a combination of instruments, which would 
 dispense with the services of this operator, by automatically pre- 
 serving intact the circuit of the sending station, at the same time 
 allowing each station to break the other when desired, was much 
 sought after. 
 
 Mr. Charles S. Bulkleyhas the credit of having invented the first 
 automatic repeater. It was called an open circuit repeater, but as 
 it has been obsolete for years I shall not describe its action here. 
 
 Mr. Bulkley imitated Professor Morse in the manner of keeping 
 his invention a profound secret from curious investigators, by fitting 
 up a labyrinth of wire coils around his instruments. But it was 
 not long before other styles of automatic repeaters were invented 
 by Farmer, Woodman, Hicks, Haskins, Milliken, Catlin, Toye, and 
 a host of others, the majority of which repeaters, although quite 
 efficient, never came into general use. Of these the Milliken, which 
 was adopted as the standard automatic repeater of the Western 
 Union Company, and the Toye repeater, which is quite extensively 
 employed in Canada, are about the only repeaters now in general 
 use. For this reason I shall limit the descriptions of the various 
 repeaters to these two, more especially as they are widely different 
 in their methods of performing a similar function. 
 
 Fig. 44 is a diagram of the connections of the Milliken repeater. 
 
 In the figure, the western, or No. 1 circuit is led to the tongue 
 of transmitter T l and thence, when that transmitter is closed, to the 
 battery B l . The eastern circuit is led to transmitter T l and 
 through battery -B, also to the ground. It is seen that the arma- 
 ture of relay R l controls the transmitter T l by a local circuit shown 
 by dotted lines, and that the armature of relay It? in like manner 
 controls transmitter T l . Relay R l thus can repeat into the eastern 
 circuit at will, for at each motion of transmitter T [ the eastern 
 circuit is made or broken at the point X 1 . In the same way relay 
 
REPEATERS. 
 
 95 
 
96 REPEATERS. 
 
 R 2 can repeat into the western circuit by operating the transmitter 
 T\ as it opens or closes the western circuit at the point X. 
 
 We see that there are also two other local circuits in the diagram, 
 one from the extra relay RX Z to the bar of transmitter T\ and 
 completed through the insulated screw /S" ; and the other, from 
 relay R X 1 to the bar of T-, completed through the insulated screw 
 S. These extra relays are supported in the position shown in the 
 figure by brass standards. It is the function of each of these extra 
 relays RX 1 and RX* to keep the armature of the adjoining line 
 relay closed while its circuit is being repeated into. They do so 
 in the following manner : Let us assume that the east opens his 
 key. This allows the armature of R l to fall back, which action 
 opens the local circuit of T'*, thus permitting that instrument to 
 open, in doing which it removes battery B l from the western circuit, 
 and opens that circuit. This of course demagnetizes the relay M\ 
 and if unrestrained its armature would also fall back and open the 
 local circuit of T l , which in turn would break the eastern circuit 
 at the point JT 1 , and give us the exact conditions which were de- 
 scribed in the early part of the description of Fig. 43. But the 
 armature of R l is not allowed to fall back, for at the moment that 
 the transmitter T 2 was allowed to open, that instant the local cir- 
 cuit of RX 1 was broken at S, and this allowed the armature of 
 RX { to fall back against the armature of R\ and as the retractile- 
 spring of RX 1 is purposely made stronger than that of the relay R 1 , 
 it keeps the armature of R 1 firmly against its local contact point, 
 and thus transmitter T l is kept closed, and by this means the route 
 of the eastern circuit through T l is preserved intact, as the diagram 
 shows, at the repeating station, so that when the east again wishes 
 to close the western circuit he closes his own key, which by clos- 
 ing the eastern circuit attracts the armature of R 2 , which by clos- 
 ing T 2 closes the western circuit, and also the local circuit of RX 1 , 
 thereby also withdrawing its armature from the armature of R 1 . 
 But as now the western circuit is closed, there is no tendency on 
 the part of the latter armature to fall back. When the west wishes 
 to repeat into the east, the above described action is simply reversed. 
 
REPEATERS. 97 
 
 Supposing that the east now has his key open with the result 
 shown in the diagram, when the west wishes to break he will open 
 his key. This does not yet change the position of the instruments 
 at the repeating station, as the western circuit is still open at Jf, 
 transmitter T' 2 ; when, however, the east next closes his key, in so 
 doing he, of course, allows the relay 72 2 to magnetize. This 
 attracts its armature, which, by closing the local circuit, closes the 
 transmitter T 72 , and the western circuit is now closed at JT, and the 
 local circuit of RX 1 is closed at S. This action withdraws the 
 armature of RX 1 from the armature of relay R \ and^the key of 
 the western operator being open for the purpose of breaking, the 
 western circuit still remains open. Therefore the armature of R l 
 falls back, opening the local circuit of T l , and breaking the eastern 
 circuit at X 1 . The eastern operator, now finding his circuit open, 
 keeps his key closed, and thus the west is permitted to make any 
 necessary remarks by causing relay R l to operate the transmitter 
 5 71 , which, as we have seen, breaks or closes the eastern circuit at 
 X 1 at each motion. 
 
 In practice the transmitters are furnished with buttons, not 
 shown in the diagram, by means of which the automatic repeating 
 portions are cut off, thus allowing each of the circuits to be worked 
 as a single wire from the repeating station. 
 
 I have said that the local circuits of RX 1 and RX 2 are broken 
 at the instant that the respective main circuits are broken at X and 
 X 1 . In reality the local circuits at R X 1 and RX- are broken an 
 instant before the main circuits, for the moment that the levers of 
 the transmitters begin to move, that moment the local circuits of 
 RX l and RX- are broken, whereas the tongues of T l and T* fol- 
 low the contact screws until prevented by the bent arms AA\ and 
 by this arrangement the holding fast at the proper time of the arma- 
 tures of R l and R' 1 by the armature of RX* and R X 1 is insured. 
 
 In the transmitters used in the Milliken repeater, the bent arms 
 are insulated from the tongues, and the bar from the screws H and 
 H\ so that any possil lity of interference between the local and 
 line batteries of RX 1 a.vd RX 2 is removed. From this description 
 
98 REPEATERS. 
 
 it will be observed that the action is in a measure mechanical, the 
 armatures of the extra relays RX 1 , and RX 2 holding the armatures 
 of the line relays as firmly and promptly against their contact points 
 at the proper moment, as would be possible for the finger of an 
 operator to do ; indeed, incomparably more promptly and firmly. 
 
 The Milliken repeater was used for a long time between Jackson- 
 ville, Fla., and Havana, on the Cuba cable, the repeating stations 
 being at Punta Rassa, Fla., and Key West, a total distance of 560 
 miles, 320 of which is through a cable. The Milliken has, however, 
 recently been supplanted on this route by the Johnston repeater. 
 
 THE TOYE REPEATER. 
 
 As I have said, the operation of the Toye repeater is entirely 
 different from the Milliken. It is considered much simpler. It 
 certainly possesses an advantage over the Milliken, inasmuch as 
 the instruments used in its operation are ordinary line relays and 
 continuity-preserving transmitters, and need but two local bat- 
 teries. The Milliken repeater requires four local batteries in its 
 operation. The Milliken, however, has the reputation of being the 
 most generally efficient of repeaters. 
 
 Fig. 45 shows the connections of the Toye repeater. The main 
 circuits are first led from the line through the posts of the trans- 
 mitters, thence to the tongues of those instruments, to the line 
 relays, the batteries and ground. The bar of each transmitter is 
 connected with a rheostat, the object of which will shortly be 
 explained. 
 
 I will assume that the west wishes to repeat into the eastern 
 line. He has opened his key, which opens R l and permits its 
 armature to fall back, opening the local circuit of T, and breaking 
 the eastern circuit, as seen in the figure at the point -X" 1 . 
 
 Just prior to the moment, however, that the eastern circuit was 
 broken at X 1 , the battery B was shunted, via the bar or lever of 
 Tinto the rheostat H, as shown in the diagram. Thus the relay 
 R has not lost its magnetism, and its armature is still retained 
 against its contact point, thereby keeping the transmitter T l sta- 
 
REPEA TERS. 
 
100 REPEATERS. 
 
 tionary, and consequently the circuit of the sending office is kept 
 unbroken at the repeating station, which, as we have already seen, 
 is the desideratum in all automatic repeaters. This action is also 
 reversed when the east wishes to send to the western circuit. 
 
 The resistance of the rheostats is adjusted to correspond with the 
 resistance of the line circuits, so that the batteries meet with a 
 similar resistance in either position of the transmitters. 
 
 This description makes it plain that it would be difficult to con- 
 struct a more simple repeater than the Toye. It, however, has two 
 slight drawbacks, one of which is, that as it depends on the pres- 
 ence of the main batteries to keep the relays magnetized at the 
 proper time, it can only be employed at the terminals of lines, where 
 such batteries are. The other drawback consists of the fact that 
 the batteries are kept constantly closed. 
 
 COMBINATION DUPLEX AND SINGLE CIRCUIT REPEATERS. 
 
 The introduction of the Stearns Duplex, and subsequently the 
 quadruplex systems of telegraphy, dispensed with the services of 
 many sets of automatic repeaters. 
 
 These multiplex sj^stems did not render the wires capable of 
 being worked at any greater distance than heretofore, but it 'was 
 found a very easy matter to cause one duplex or quadruplex set to 
 repeat into another quadruplex or duplex set, without the use of 
 any of the regular automatic repeaters. 
 
 The manner in which this repetition is effected is very simple ; 
 but, as it may not be generally known, I shall herewith describe it. 
 
 Let us take, for example,the cable quadruplex circuit between New 
 York and Sydney, C. B.,with the quadruplex repeaters at Bangor, Me. 
 
 At the repeating station, Bangor, there are two quadruplex sets. 
 One of these works direct with New York, the other with Sydney. 
 
 I can, I think, simplify the description if the reader will take a 
 piece of blank paper and roughly draw on the left-hand side of the 
 paper a neutral relay and a transmitter, and immediately benei th 
 these instruments, a polar relay and i\ pole changtr, ani call this t, e 
 New York set. 
 
REPEATERS. 101 
 
 On the opposite side of the paper draw, in the same way, another 
 neutral relay and transmitter, a polar relay and pole changer, and 
 call this the Sydney set. 
 
 Now draw lines representing a local circuit from the contact 
 points of the neutral relay of the New York set to the magnet of the 
 pole changer of the Sydney set, and insert a local battery. This 
 gives the neutral relay control of the pole changer. 
 
 In the same manner give the neutral relay of the Sydney set con- 
 trol of the pole changer of the New York set. 
 
 Draw lines again from the polar relay of the New York set to the 
 transmitter of the Sydney set, and in like manner connect the arma- 
 ture of the polar relay of the Sydney set with the transmitter of the 
 New York set. 
 
 It is now seen that when the neutral relay of the New York set is 
 operated, its armature operates the pole changer of the Sydney set, 
 and when the neutral relay of the latter set is worked, it operates the 
 pole changer of the New York set ; that the polar relay of the New 
 York set operates the transmitter of the Sydney set, and the polar re- 
 lay of the latter in turn operates the transmitter of the New York set. 
 
 In other words, the armatures of the different relays act simply 
 as keys to operate the various pole changers^and transmitters of the 
 opposite sets. 
 
 I think it will now be quite plain to the reader that when, for 
 instance, the operator in the New York office, who may be sending 
 on the second side, manipulates his transmitter, his signals will 
 operate the neutral relay of the New York set in Bangor, which in 
 turn, by operating the pole changer of the Sydney set in Bangor, 
 repeats the signals on to the Sydney office, and vice versa. 
 
 Of course a duplex set can be caused to repeat into another 
 duplex set in the same way. 
 
 It will be noticed by the foregoing that the No. 1 side of each 
 set is made to repeat into the No. 2 side of the other set. 
 
 This arrangement is not absolutely necessary, as the first side of 
 one set could be just as readily made to repeat into the first side 
 of the other set. 
 
102 REPEATERS. 
 
 The first-mentioned arrangement is, however, the one generally 
 adopted, probably to neutralize any atmospheric condition which 
 would perhaps tend, on a long circuit, to make one side work bet- 
 ter than the other. 
 
 It is quite frequently found that all of a quadruplex set is not 
 required for the handling of business between certain points, and 
 that therefore half of a set, that is one side of it, can be spared for 
 other purposes. 
 
 Thus there is a quadruplex circuit between New York and 
 Elmira, one side of which is set apart for the local business, while 
 the other side is connected with a single wire to Bradford, Pa. ; and 
 there are many other similar circuits. 
 
 The utility of the quadruplex system is here seen, inasmuch as 
 it saves one wire for the distance covered by the quadruplex sys- 
 tem for instance, in the case referred to, a wire between New 
 York and Elmira. 
 
 A combination circuit of a quadruplex and single circuit requires 
 an automatic repeater at the point where the junction is made, and 
 several arrangements have been devised for this purpose, one of the 
 first of which, known as the Waterbury repeater, I shall here 
 describe. 
 
 In Fig. 46, T and N R are, respectively, the transmitter and 
 neutral relay of the second side of a quadruplex set, from which 
 all but the local connections have been omitted. R T is a repeat- 
 ing transmitter ; R R is a repeating relay. 
 
 The single wire is led, as may be seen in the diagram, through 
 the repeating relay to the tongue of R T, and thence to the battery 
 and ground. 
 
 There is also an extra circuit shown by dotted lines from R R to 
 the bar of R T, in which is inserted a local battery, also as shown. 
 
 When the transmitter 72 Tis open (as it will be when the neutral 
 relay N R is open), this extra circuit is completed through R R, 
 and via the bar and tongue of the transmitter at x', so that when- 
 ever R T is open, the repeating relay will be kept closed. 
 
 It is, of course, understood that the operator at the distant end 
 
REPEATERS. 
 
 103 
 
 TO 
 
104 REPEATERS. 
 
 of the quadruplex, whom I shall hereafter call the quadruplex 
 operator, can only send or receive one way at a time. 
 
 The manner in which this arrangement acts is as follows : 
 
 Suppose that the operator at the distant end of the single wire, 
 whom I shall refer to as the Morse operator, wishes to send ; in 
 that case the quadruplex operator must keep his key closed, which 
 we are all aware will keep the neutral relay at the repeating station 
 closed, as seen in the diagram. Now when the Morse operator 
 opens his key that action opens the repeating relay R R, for it may 
 be seen that the extra circuit is at present open at x'. This allows 
 its armature to fall back, opening the local circuit of the quadruplex 
 transmitter jP, and thereby opening the neutral relay at the dis- 
 tant end. 
 
 When the Morse operator again closes his key, thus closing the 
 single wire circuitj the armature of R R is attracted, and thereby 
 again closes the transmitter T. Thus as often as the Morse opera- 
 tor opens or closes his circuit, the relay R R opens or closes the 
 quadruplex transmitter. 
 
 When the quadruplex operator wishes to send, he simply operates 
 his transmitter, which in turn opens or closes the neutral relay N 
 R at the repeating station. 
 
 This action, by allowing the armature of N R to fall back, opens 
 the transmitter R T, which thus cuts off the single-wire battery 
 from the single wire at x', thus opening that wire. The relay R R 
 does not, however, open at this time, for, as I have already shown, 
 the extra circuit, with its local battery, comes into play the moment 
 that R T opens, and, consequently, the quadruplex operator does 
 not get his own writing back, as he would if R R were not kept 
 closed. When the quadruplex operator again closes his transmitter 
 the neutral relay N R is again closed. This in turn closes R T, 
 again connecting the single battery with the single wire and clos- 
 ing that circuit. 
 
 In this way the quadruplex operator, by controlling the neutral 
 relay N R at the repeating station, which, as we have seen, con- 
 trols the repeating transmitter R T, can operate the single wire at 
 
REPEATERS. 105 
 
 will, for at each motion of R T the single wire's circuit is opened 
 or closed at the point x'. 
 
 This arrangement has been found very serviceable in good 
 weather, but during stormy weather its action is nqt so satisfactory. 
 This is owing to the fact that the increased magnetism in the sin- 
 gle-wire relays, in wet weather, necessitates the pulling up of the 
 relay spring to such an extent that the local battery of the circuit 
 becomes unable to hold the relay closed against the increased ten- 
 sion of the relay spring, and hence the quadruplex operator gets his 
 own writing back. 
 
 It may have occurred to the reader that this function could be 
 performed by the use of one-half of either of the automatic repeat- 
 ers which I have previously described, and such is the case. 
 
 There are many circuits on which one-half of the Milliken or 
 Toye repeaters is used for this purpose with entire success. 
 
 It is well known that there are a great many quadruplex sets 
 leased to broker firms, two firms generally being assigned to one 
 set, each firm using one side of a set. 
 
 In such cases the quadruplex sets are stationed, in the main 
 offices of the company, while a single wire known as the " short or 
 branch wire" is run direct from the quadruplex sets to the 
 brokers' offices. 
 
 A repeating arrangement is also essential to the working of this 
 plan, and a number of different methods are employed, including a 
 modification of the Toye and other repeaters. 
 
 The method employed in New York city is known as the D. R. 
 Downer repeater .shown in Fig. 47 a very simple and effective 
 arrangement, which I shall proceed to describe. 
 
 In the figure, ^and N R are the transmitter and neutral relay of 
 a quadruplex set. They may, however, be supposed to represent 
 equally well a pole changer and polarized relay. R T is a repeat- 
 ing transmitter. 
 
 It is seen that the neutral relay has control of R T\)j means of 
 the local circuit shown, and that the broker's single wire has con- 
 trol of the quadruplex transmitter T by the route as represented in 
 
106 
 
 REPEATERS. 
 
REPEATERS. 107 
 
 the diagram through the post of R T, to the tongue of that instru- 
 ment, thence to the button switch S, through the magnet of T to 
 the battery and ground. When the broker at this end (referring 
 to diagram) wishes to send, it is necessary that the broker at dis- 
 tant end should keep his key closed. This keeps N R closed, and 
 consequently also R T ; thus -tli2 broker has unobstructed control 
 of the quadruplex transmitter T, for it is evident that when he 
 opens his key it must open the transmitter T by breaking the circuit 
 of the branch wire battery. On the contrary, it is quite as essen- 
 tial that, when the distant broker is sending to the broker at this 
 end, the quadruplex transmitter T at this end should be kept closed, 
 and that object is effected as follows : 
 
 When the distant broker desires to send, he opens his key ; this- 
 opens N R in the diagram, this in turn opens R T, which breaks 
 the branch wire circuit at the point Jf, and opens the sounder in the 
 broker's office at this end. At the moment, however, that the 
 broker's wire is opened at X, another extra circuit, T W, shown by 
 the dotted lines, is brought into play, namely, via the bar of R Tto 
 the point X' in the battery, through this small portion of the battery, 
 which is known as the " tap," to the magnet of T and back to the 
 bar of R T by way of the tongue of that instrument. By this 
 means it will be observed that the quadruplex transmitter is kept 
 closed at the right moment, without the use of any additional local 
 battery. When the distant broker again closes his key or trans- 
 mitter, the neutral relay N R is again closed, the repeating trans- 
 mitter R T is thereby also closed, and the battery is placed once 
 more to the branch wire at the point X, while the tap wire is 
 disconnected at that point at the same time. 
 
 It is sometimes necessary to cut off the branch wire to facilitate 
 the testing or balancing of the quadruplex. For this purpose the 
 switch S is inserted. By turning the switch to the lower button it 
 may be noticed that there is a short metallic circuit formed, enclos- 
 ing the key JSTand controlling T. 
 
 It is, I suppose, well known that the above described arrange- 
 ment dispenses with the use of one wire to the branch office. 
 
108 REPEATERS. 
 
 In some cases, however, the repeating device is not used, but in- 
 stead two short wires are furnished, one of which is in direct connec- 
 tion with the transmitter, and the other with the neutral rela} r , both 
 wires being grounded at the branch office. 
 
 In some main offices one-half of the Toye repeater is also used to 
 perform the necessary repeating work from the quadruplex to a 
 branch wire. 
 
 By reference to Fig. 47 it will be seen that this can be done very 
 easily, as follows : Remove the tap wire T IF from the battery at X', 
 and conduct it through a rheostat to the ground. No other change 
 is necessary. Of course the resistance of the rheostat will require 
 to be made equal to the limited resistance of the branch wire. 
 
 The only advantage that I know of in the Downer arrangement 
 over the Toye, for this purpose, is that mentioned in the description 
 of Che Toye repeater, namely, that the entire battery is not always 
 kept closed. 
 
 By the Downer plan the greater portion of the battery is open 
 half the working time. 
 
 I shall now bring this chapter to a close by giving one or two 
 hints on the practical management of repeaters of the latter class. 
 
 In the first place, it should be a rule to see that the contact points 
 are cleaned regularly, and that the local batteries are kept up to a 
 standard strength. It is a good plan to test the strength of the local 
 batteries occasionally, by withdrawing the armature of the instru- 
 ments against the attractive force of the magnets. When there is 
 a decided sign of weakening, the locals should be at once renewed. 
 
 It is a somewhat common occurrence to find that the writing on 
 the neutral relay is being repeated back on the transmitter. 
 
 This fact is sometimes brought to your attention by the state- 
 ment of the distant end that he is getting his own writing back. 
 
 This effect may be produced by one of several causes. For in- 
 stance, the branch office may have his key open, or the transmitter 
 points at X in Fig. 47 may not be making proper contact, or they 
 may need cleaning. Your own key may be open, or the battery 
 mav have failed. To locate the trouble, first see that the repeating 
 
REPEA TERS. 109 
 
 transmitter is acting properly, and that the points are clean. File 
 them. Turn the button switch to the left. If this closes the trans- 
 mitter, it shows that the local portion of the battery is all right. 
 Next take a piece of wire and attach one end of it to the post of 
 the repeating transmitter, the other end to a ground. There are 
 always plenty of good grounds around a quadruples set. If this 
 closes the quadruples transmitter, it proves that your battery is in 
 order, and that the circuit is open between you and the branch 
 office. Care should be taken, however, to notice that the trans- 
 mitter is closed with somewhat more than the usual strength, be- 
 cause it is possible that the battery may have partially run down, 
 and while too weak to close the transmitter when the full resist- 
 ance of the branch wire was in the circuit, is still strong enough to 
 close the transmitter with the branch circuit cut off. These tests 
 can be made within one minute when one is familiar with the mode 
 of proceeding. 
 
 Besides those repeaters which I have described, and also those 
 that I have referred to, there have been other forms of repeaters 
 invented. For instance, repeaters that allow one wire to repeat 
 into two distinct circuits, and give each of those two circuits the 
 power of automatically breaking the sending circuit. Such repeat- 
 ers have been termed " three-corner repeaters." 
 
 Mr. R. C. Edwards, of New York, has recently invented an 
 octuplex repeater, which enables the sending circuit to repeat into 
 eight distinct circuits, and also allows each of those circuits to 
 automatically break the sending circuit. 
 
 There may also be other systems which have not come under my 
 observation. 
 
 I have myself invented a very simple multiple repeater, by means 
 of which one circuit can repeat into one, fifty, or a hundred or 
 more other circuits, each of which other circuits can automatically 
 break the sending circuit, or it can be arranged so that one of a 
 system of, say fifty or one hundred circuits, can repeat into fifty or 
 one hundred circuits of another system, and any one of the latter 
 system can automatically repeat into all of the former system. 
 
WHEATSTOBTE AUTOMATIC TELEGRAPH. 
 
 BY WM. MAVER, JR. 
 
 A general interest has been created among operators concerning 
 the Wheatstone automatic telegraph since its adoption by the 
 Western Union Telegraph Company as its rapid telegraph system. 
 A desire having been evinced to learn more about its manner of 
 working, I shall endeavor in this chapter to give a detailed descrip- 
 tion of the method of its operation. 
 
 It is, I suppose, generally known that the Wheatstone automatic 
 system differs from the American Rapid Company's automatic, or 
 any other system of fast automatic telegraphy in this country in 
 which the Morse alphabet is used, inasmuch as it is what may be 
 termed a mechanical automatic telegraph, while the others are 
 chemical automatic systems. 
 
 The chemical systems that most resemble the Wheatstone are 
 those that use prepared perforated paper for the purpose of trans- 
 mitting signals; for instance, the American Rapid system, which 
 was fully described in the columns of The Operator some time ago. 
 
 In the latter system the perforated paper is passed over a metal 
 cylinder, which is a continuation of the main circuit, and two 
 needles, or styles, connected with the negative and positive ends of 
 batteries respectively, are placed on the top of the paper in such a 
 position that when the punctured portions of the paper pass, the 
 needles are allowed to connect with the cylinder beneath the paper, 
 and thus permit a positive or a negative current to flow to the line, 
 according to the situation of the punctures. This is called direct 
 contact. 
 
 At the receiving station of this system there is another metal cyl- 
 inder (which is also a part of the main circuit), over which passes a 
 
THE WHEATS TONE AUTOMATIC. Ill 
 
 ribbon of chemically prepared paper. Resting on this paper is a 
 needle, likewise a part of the circuit. As this paper is moistened 
 with a chemical preparation, it constitutes a good conductor, and 
 when a current of electricity is sent to the wire from the sending end 
 it passes through the needle and the paper to the cylinder and 
 ground. While the current is thus passing through the paper, a 
 chemical decomposition of the ingredients composing the solution in 
 which the paper has been previously soaked takes place, and in con- 
 sequence a mark, varying in color according to the constituents of 
 the solution, is recorded on the paper as long as electricity is allowed 
 to flow through the circuit. 
 
 Thus, by varying the duration of the current on the line, which 
 is the function of the perforated paper, dots and dashes can be sig- 
 naled as required. 
 
 In the Wheatstone system the perforated paper serves to send 
 currents of either polarity to the line, and also governs the length 
 of the mark recorded at the distant end but in an entirely differ- 
 ent manner from that just reviewed. 
 
 Beneath and pressing up against the perforated paper in the 
 Wheatstone system are two needles, or small levers, the lower ends 
 of which are connected by springs and "other mechanism to a 
 delicately constructed pole changer. 
 
 When there are no punctures in the paper passing above the small 
 levers, the latter are prevented from moving to their full extent, 
 and thus fail to actuate the pole changer. When either of these 
 needles comes to a hole in the paper it passes through it and makes 
 its full motion, and so moves its connections that a certain pole is 
 placed to the line. 
 
 At the receiving end a polarized relay is placed, whose armature 
 responds to the changes of polarity at the sending end, and operates 
 an inking apparatus, which, when in a certain position, records a 
 mark on the receiving paper. 
 
 This will probably be sufficient to show the distinction between 
 a mechanical and chemical automatic system. Each of these re- 
 spective systems has its advantages and disadvantages, but as it is 
 
112 
 
 THE WHEA TSTONE A UTOMA TIG. 
 
 not intended in this chapter to discuss these, I shall merely mention 
 one or two of the characteristic features of each. 
 
 In some of the chemical methods there seems to be almost no 
 limit to the rapidity with which signals may be sent and recorded, 
 as, to a certain extent, it simply depends on the frequency with 
 which the line can be charged and discharged. This is one of its 
 advantages. One of its disadvantages lies in the fact that occa- 
 sionally the perforations in the paper are not clearly punctured, and 
 thus small pieces of paper intervene between the direct contact 
 needle and the cylinder, and thereby cause confusion. 
 
 Fig 48. 
 
 The speed at which a mechanical system can be operated is lim- 
 ited by the rapidity with which the electro-magnet can be charged 
 and discharged, and overcome the inertia of its armature, so that 
 distinct signals shall be recorded. One of the advantages of the 
 latter system, and it is a valuable one, is that the receiving signals 
 are received on blue stiff paper, in clear Morse characters. Another 
 is that as the speed at which signals arrive by this method is not 
 beyond the rate at which the receiving instrument clerk can follow 
 them, any errors which may occur are at once detected and cor- 
 rected. 
 
 It may, perhaps, aid the reader to an understanding of the work- 
 ing of the Wheatstone system to remember that the operation of 
 
THE WHEATSTONE AUTOMATIC. 113 
 
 the various parts of its mechanism all tend to the transmission and 
 reception of the Morse characters. What operators do by the hand 
 this machinery does automatically. We are all aware that when we 
 are sending on a polar duplex, or the polar side of a quadruplex 
 (the number one side), to make a dot we have to cause the key to 
 make two motions, one down and one up. In doing so we have 
 changed the polarity of the current twice. By closing the key, 
 which is, of course, equivalent to closing the pole changer, we send 
 a current to line, which attracts the armature of the distant polar- 
 ized relay to the local contact point. By opening the key again, 
 we place a different pole to the line, and it repels the armature from 
 the local contact point. To make a dash we simply allow our fin- 
 
 Fig. 49. 
 
 gers to linger an instant longer on the key when it is closed. To 
 make the space we keep the key open a moment. What we thus 
 do by the guidance of our faculties, the automatic is required to do 
 by the arrangement of its component parts. By recollecting what 
 is required of the mechanism, we may be the better able to compre- 
 hend the manner in which it accomplishes this purpose, when it 
 is explained. 
 
 Let us suppose that when we close the key on the polar duplex 
 or quadruplex, we send a positive current to the line, and that the 
 polarized relay is so arranged that it is the positive current that 
 attracts the armature of the relay to the local contact point, and 
 
114 
 
 THE WHEATSTONE AUTOMATIC. 
 
 that when we open the key we place a negative pole to the line, 
 which repels the armature from the local points. 
 
 Now, if we attach a pen to the armature and cause a strip of 
 paper to pass in close proximity to it, if the key is closed we shall 
 have a mark on the paper until we again open the key. This is 
 similar to what occurs in the Wheatstone system by the combined 
 operation of the perforated paper, which acts the part of the opera- 
 tor's fingers, inasmuch as it controls the levers under the paper, and 
 the transmitting mechanism. These levers correspond to the Morse 
 key, as they actuate the pole changer, which latter in turn oper- 
 ates the distant polarized relay with its inking attachment, as will 
 be explained in the detailed description of the Wheatstone trans- 
 mitter and receiver. 
 
 Fig. 50. 
 
 The Wheatstone automatic apparatus consists of three parts, 
 namely : the perforator, the transmitter, arid the receiver. 
 
 The perforator is the instrument by means of which the message 
 is prepared for transmission, and is shown in full in Fig. 48. It 
 consists of three distinct boxes, the largest of which supports the 
 stand D, pn which messages intended for preparation are placed 
 for the convenience of the punching operator. It also contains the 
 paper to be punched. The smallest box F contains the punching 
 apparatus proper. Box Cr serves to support the latter box, and to 
 receive the residue of the perforations which rapidly accumulate. 
 The paper which is seen issuing from the largest box is of a stiff 
 white pattern, and has been previously prepared for punching by 
 having been dipped in olive oil, so as to dispense as much as possi- 
 ble with the paper dust, which, unless the paper is so prepared, is 
 found to cause trouble by clogging up the punching apparatus. 
 The oil also serves to lubricate the machinery of the perforator. 
 
THE WHEATSTONE AUTOMATIC. 115 
 
 Figure 49 is a top view of box F, Fig. 48. The buttons or discs 
 A B and C are mechanically connected to three levers or keys be- 
 neath the cover of the box, and these levers are again respectively 
 connected in such a manner to three, one, and four sharp, hollow 
 cylinders, that when either of the discs A y B or C is depressed, the 
 corresponding sharp cylinder is urged forward through the paper 
 ribbon, cutting clear round holes therein. 
 
 Fig. 50 shows these hollow cylinders as arranged. 
 
 The depression of disc A causes three of these cylinders to punch 
 the paper in one vertical line (see Fig. 51), which corresponds to a 
 dot ; B causes one only to punch (Fig. 52), which is the spacing 
 mark ; C causes four of them to perforate the paper (see Fig. 53), 
 which correspond to the dash. There is another and important piece 
 of mechanism in this box, not shown in the diagram. It consists of 
 a small star wheel, the teeth of which fit closely into the central or 
 
 Fig. 51. Fig. 52. Fig. 53. 
 
 space holes. The same action that impels any of the punching 
 cylinders forward jerks the star wheel around, and it is thus made 
 the medium for carrying the paper along as it is perforated. The 
 punching operator is provided with two small iron mallets tipped 
 with rubber, by means of which he depresses the discs. ' In practice 
 the dot and dash discs have a dot and dash engraved on them, but, 
 of course, this is not necessary to distinguish them after a few 
 hours practice. The dot disc is on the left side, the dash disc on 
 the right side, and the space disc is in the middle. To make the 
 letter (7, for instance, the dot disc is struck twice in succession, then 
 the space disc, and again the dot. The spacing disc is struck twice 
 between words, and once between each letter. This method of 
 punching entails a considerable amount of manual labor, and soon 
 becomes very tedious. The most expert punchers do not maintain 
 a speed of more than thirty-five words per minute for any length of 
 
116 
 
 THE WHEATSTONE AUTOMATIC. 
 
 time. As it would be impracticable to make 
 a dash long enough for the American Morse 
 signal for the letter L, it has been given the 
 characters of two dots and two dashes. 
 
 j^ Fig. 54 shows the words The Operator pre- 
 
 pared for transmission in the manner above de- 
 scribed. 
 
 i ^ 
 
 , The Wheatstone transmitter, Fig. 55, is con- 
 
 tained in a brass box about eight inches long, 
 
 I ** five inches high, and three wide. 
 
 . Fig. 56 is an exterior view of the box, show- 
 
 ing the transmitting apparatus proper. 
 
 The paper is represented by the straight 
 
 i ^ line P, and moves in the direction indicated 
 
 t by the arrow. 
 
 T# 
 
 , ^ * The levers S and M are kept in their up- 
 right positions by the retractile -springs shown. 
 1 These levers are so adjusted by the screw posts 
 
 | ^, F and F', that the lever M is allowed to go 
 slightly farther to the right than the lever S. 
 1 This apparently trifling arrangement should 
 
 be remembered, as it will be seen farther on 
 that it serves a useful purpose. 
 
 R is a beam of ebonite or other insulating 
 1 material, supported by a pivot in its center, 
 
 ^} running backward into the box. This pivot 
 is connected with the clock-work contained in 
 j &s the box, in such a manner that when the clock- 
 
 work is in operation it imparts to the beam 
 a rocking motion, similar to the Avalking beam 
 of a steamboat. An upward tendency is given 
 to the levers S and M by the springs S and S' 
 attachpd to the crank levers A and B, but the extent of the up- 
 ward motion of the levers $and Mis regulated by the metal pins 1 
 and 2 inserted in the beam, and also by the prepared paper P. 
 
THE WHEATSTONE A ITTOMA TIC. 
 
 117 
 
118 
 
 THE WHEATSTONE AUTOMATIC 
 
 Attached to the crank levers A and B are the straight levers H 
 and H'. The ends of these pass through the projecting pieces P 
 and P', which are fixed to the movable disc D. H and H 1 are 
 provided with small collars or collets K and K, which prevent their 
 ends from going beyond a certain distance through the projections 
 P and P 1 . 
 
 The disc D is pivoted to the frame-work of the box, and is 
 divided by an insulating material into two parts, one of which is 
 connected to the ground, and the other to the line. Metal pins are 
 also inserted in this disc, and on one or other of these pins, accord- 
 ing to the position of the disc D, the crank levers C and Z are made 
 
 Fig. 56. 
 
 to rest by the action of the retractile springs E and E'. In the present 
 instance the beam R has moved upward, thus permitting the lever 
 M to ascend under the influence of spring S'. The lever H' has 
 thus pushed the disc D over to its present position. The same 
 motion of the beam has, by aid of the pin 1, pushed the lever S 
 down, and this motion of lever S has withdrawn the lever H from 
 P', so that the lever H' was unopposed in pushing over disc D. 
 The small roller E, being pivoted on a tension spring, aids the 
 motion of the disc in this wise : when the piece P, going either 
 way, passes the center of the roller E, the latter gives the disc D a 
 jerk over. In Fig. 56 the lever M has passed through a hole in 
 the paper, and thus has made its full motion with the above effect. 
 It may also be seen that this position of disc D places a positive 
 
THE WHEATSTONE AUTOMATIC. 119 
 
 pole to the line. If, now, the beam make another motion whereby 
 the lever S is permitted to ascend, and the lever M is made to de- 
 scend, it may be seen that the collet of lever H will be driven 
 against the piece P', and as the collet of the lever H' is now with- 
 drawn from P, the position of the disc D is changed, and the nega- 
 tive pole sent to the line, as shown in Fig. 57. Thus we can see, 
 presuming there is no paper passing above to limit the motion of 
 the levers M and S, that as often as the beam is moved up and 
 down, the polarity of the battery is changed. 
 
 Fig. 57. 
 
 Directly under the paper P, in Fig. 56, is a small star wheel, not 
 shown in the diagram. This wheel is geared with the clock-work 
 also, and in such a way that when the beam R makes one motion, 
 the star wheel moves around the space of one tooth. The teeth of 
 this star wheel, as in the punching apparatus, fit into the central 
 holes of the prepared paper, and as it revolves carry the paper 
 forward. 
 
 Let us now follow the operation performed by this apparatus in 
 making the letter R, as seen in Fig. 58, namely, two vertical holes, 
 a blank, and a series of two sets of two vertical holes. As the right 
 hand end of beam R moves upward, the lever M passes up through 
 the topmost hole in the paper, as it appears in Fig. 58, or the 
 hrtle nearest the frame of the box. This, as I have already said, 
 i-ends a marking current to the line, to which the armature at 
 
120 THE WHEATSTONE AUTOMATIC. 
 
 distant end responds. The right end of the beam moves down- 
 ward ; the left end moves upward, and with it the lever S. At the 
 same time the star wheel moves the paper forward one space, and 
 this movement of the paper brings the lowest of the two vertical 
 holes now opposite the lever S (for it will be remembered that the 
 lever S is adjusted a little to the left of J/), which passes through 
 that hole. Thus the polarity is again changed, and the inker is with- 
 drawn from the paper at the distant end. The beam makes another 
 motion, and the star wheel carries the paper one tooth forward ; but 
 it is seen that the lever M meets with no hole, and its upward 
 course is, therefore, blocked by the paper, and the pin 2 moves up 
 unfollowed by M ; as the lever M consequently has not made its 
 full motion, the lever H' has not pushed the disc D over, and thus 
 the same pole is continued to the line, and a blank space is left on 
 the receiving paper. Another motion of the beam brings the lever 
 S up 7 to the paper, but it, too, in the same way, is prevented from 
 making its full phase, and the disc D remains as it was, which fur- 
 ther increases the space on receiving paper. A further motion of 
 the beam now brings the lever M again up to the paper, and also 
 opposite the topmost hole of the second series of holes, which the 
 star wheel has advanced. The lever thus makes a full movement, 
 and this sends a marking current once more to the line. Again 
 lever S is brought up to the paper, and, also finding a hole, passes 
 through it, again changing the polarity, and the marker is withdrawn ; 
 
 R. A. 
 
 Fig. 58. Fig. 59 
 
 thus another dot is recorded. The polarity is again changed twice 
 as before, and another dot is marked, which now gives us on the 
 receiving paper a dot, a space, and two dots. 
 
 The diagonal position of the dash perforations simply reverses the 
 action necessary for a spacing current. Take, for instance, the let- 
 
THE WHEATSTONE A UTOMATIC. 
 
 121 
 
 ter A, Fig. 59. The upper hole allows a marking current to be 
 sent. Immediately the lower hole reverses it, and a dot is recorded. 
 The second upper hole sends another marking current to line, and, 
 as owing to the position of the next lower hole the beam will fe- 
 quire to make two motions before the lever S reaches it, the mark- 
 ing current continues to the line, and a dash is allowed to be re- 
 corded on the receiving paper. 
 
 Of course these actions take place very rapidly, so much so that 
 when the apparatus is at full speed, namely, about 250 words per 
 minute, it is impossible to follow the motion of the levers. This 
 statement may be better understood when it is reflected that at 
 this rate of speed the levers move about 4,000 times per minute. 
 
 Fig. 60. 
 
 THE WHEATSTONE RECEIVER. 
 
 The receiving instrument is contained in n. brnss box of about 
 the same dimensions as the transmitter. See Fig. 60. Within the 
 box are the polarized relay, previously mentioned, and the clock- 
 work, hereafter referred to, which is operated by a spring. 
 
 Th f 3 receiving paper passes over a roller which is in close proxim- 
 
122 
 
 THE WHEATSTONE AUTOMATIC. 
 
 ity to the inking apparatus. The roller is actuated by the clock-work, 
 and is the means by which the receiving paper is carried along. 
 
 The instrument is supplied with a governor, which regulates the 
 speed of the clock-work. 
 
 Fig. 61 is a diagram of a part of the interior of the receiver. 
 
 P is a permanent magnet, in the ends or poles of which semi- 
 circular notches are made, into which the armatures of the polar- 
 ized relay M are loosely fitted. A is an axis to which these 
 
 Fig. 61. 
 
 armatures are fixed. It extends above and below the electro-mag- 
 nets. At its lower end connections are arranged for the operation 
 of a local sounder S, which is used when the circuit is switched 
 from the automatic to manual working. The upper extension of 
 the axis A is bent in the manner shown. The poles of the electro- 
 magnets M are placed opposite each other. One of the upper poles 
 of M is omitted in the diagram to show the position of the arma- 
 tures between the poles. The armatures oscillate between these 
 poles in response to the changes of polarity at the distant end. As 
 the extension E is also attached firmly to the axis A, it is evident 
 that its motion-must correspond to that of the armatures. When 
 
THE WHEA TS TONE A UTOMA TIC. 123 
 
 there is no current on the line, and when the armatures are prop- 
 erly adjusted, they remain on whichever side last placed, in the 
 same manner as the armature of the quadruplex polarized relay. 
 
 His the axle of a cog wheel which is geared with the clock- 
 work. To the right end of this axle, which protrudes through the 
 frame work of the box, is attached a small wheel IF, which rotates 
 opposite the center of the roller, over which the paper is seen pass- 
 ing in Fig. 56. This small wheel also rotates within, but does not 
 touch the groove of another wheel F, which revolves in an ink well/. 
 In this groove sufficient ink is brought up by capillary attraction to 
 the wheel W to keep its edge supplied without friction. 
 
 The axle X is given a slight tension toward the up-curled end of 
 the extension E. There is a small notch in this up-turned end 
 through which the axle X passes. This slight support is all that 
 the axle receives in addition to that from its cog wheel. 
 
 When a negative or spacing current is placed to the line the ex- 
 tension E, moving, as I have said, in unison with the armatures, 
 holds back the axle X, and, consequently, the wheel IF, from the 
 receiving paper. When a positive or marking current is sent to 
 the line, the extension E is withdrawn from the axle X, and thus 
 the wheel IF is allowed to touch and mark the paper. 
 
 The polarized relay being practically inaccessible within the box, 
 the armature is adjusted by means of apparatus placed outside of 
 the box. 
 
 The play of the armatures, and likewise the marking wheel, is 
 very limited, being about ^ of an inch. This is necessarily so, on 
 account of the rapid motions made when the apparatus is working 
 at full speed. For the same reason all of the above mechanism is 
 of very light construction, more so than the diagram would, per- 
 haps, indicate. It may be further stated, as showing the need of 
 delicate apparatus for fast working, that the sounder S in the local 
 circuit, which is adjusted as ordinarily, ceases to record intelligibly 
 after the speed exceeds say 125 words per minute, which is owing 
 chiefly to the fact that its armature is unable to travel the distance 
 to which it is adjusted, in the now diminished time. 
 
124 
 
 THE WHEATSTONE AUTOMATIC. 
 
THE WHEATSTONE AUTOMATIC. 125 
 
 The Wheatstoue automatic system may be worked on a single or 
 duplex wire. The foregoing description of the transmitter and 
 receiver applies to the actions that take place when it is worked as a 
 duplex. On the Western Union lines it is operated almost en- 
 tirely on the duplex system. By this method the capacity of a wire 
 is increased to about 500 words per minute. 
 
 Fig. 62 shows the connections necessary for working the Wheat- 
 stone duplex, and also the connections required to change the sys- 
 tem from the automatic pole changer to the Morse pole changer. 
 This change becomes necessary when it is desired to hold con- 
 versation or to balance the wire, and is made by means of the triple 
 switch (F) contained within the transmitter. In the diagram the 
 switch is arranged for the automatic system. One motion of the 
 lever controlling the switch transfers it to the Morse. The connec- 
 tions may be readily traced by any one familiar with the duplex 
 system. The circuit is as follows : From the earth at E E to the 
 disc D, to lever Z, to right-hand button of switch to negative 
 pole of battery, to left-hand button of switch and lever (7, to disc D 
 and central button of switch, thence to the receiver, where the cir- 
 cuit divides, one portion going through the galvanometor Cr to line, 
 and the other through the galvanometer to rheostat and ground. 
 
 As the polarized relay is inaccessible, owing to its position in the 
 receiver, the manner of obtaining a balance of the wire is different 
 from that on the quadruples. The distant station is requested to 
 open or close his key. This presents a certain pole to line and gives 
 the home galvanometer needle a certain deflection. The polarity of 
 the home battery is now reversed, and if a different deflection is 
 thereby caused, it is an evidence that the lume station is out of 
 balance, and the rheostat is adjusted until the galvanometer needle 
 remains stationary under the changes of home polarity, when the 
 balance is correct. 
 
 Fig. 63 shows the Wheatstone connections, as arranged for work- 
 ing the system on single wires. In this arrangement it will be seen 
 that the circuit is from the earth to the galvanometer, thence to 
 the middle button of the triple switch. Here the current has two 
 
126 
 
 THE WHEA TSTONE A UTOMA TIC. 
 
 routes ; one is to the pin 1 on the rocking beam R, thence to the 
 crank lever A, thence to pin 2 on rocking beam, thence to disc D, 
 to the battery, to the right-hand button on switch, thence to 
 lever Z and disc I), and to the line. The other route is from the 
 middle button of switch to an interposed resistance R, thence to 
 the disc Z>, etc., as before. This interposed resistance is generally 
 very high, so that when the route via the rocking beam is intact, 
 only a very small portion of the current will go through the resist- 
 ance, hence the full strength of current virtually goes to line. But 
 when it happens, for instance, that the lever J/is prevented, by rea- 
 son of there being no hole in the prepared paper, from following pin 
 
 It); ' 4. I Ht= 
 
 j^J-xfe).- ;- .1...,-., ' I I i i T] 7 
 
 Fig. 63. 
 
 2 in the beam R in its upward motion, the circuit is broken at that 
 point, and the current is compelled to go through the resistance R, 
 and the strength of the current is thereby greatly diminished ; but, 
 as we have seen that when either of the levers S or M"does not 
 make its full motion, the polarity of the current is not changed, and 
 as we also know that the tendency of the armature of the receiving 
 relay is to stay where last placed, this diminished current will be 
 
THE WHEATSTONE AUTOMATIC. 127 
 
 quite sufficient to keep the armature in its present position until 
 the polarity of the current is again changed. 
 
 When the Wheatstone was first invented this resistance was not 
 used, the wire being allowed to remain open while either of the 
 levers M or S was detached from its respective pin 1 or 2, but it 
 was found that greater speed could be attained by introducing 
 the above resistance, as the wire is thus kept at the same potential 
 until the next actual reversal of polarity. 
 
 The key shown in Fig. 63 is a double current or pole changing 
 
 key. It is still used on the Wheatstone system in England, but 
 
 has been displaced here by the ordinary pole changer, as shown in 
 
 . Fig. 62, it having been found that firmer contacts could be made by 
 
 the latter method. 
 
 In Figs. 62 and 63 the transmitter switch is placed for automatic 
 sending. 
 
 The punching instruments in the Wheatstone department at 
 195 Broadway, by means of which messages are prepared for auto- 
 matic transmission, have recently been supplied with pneumatic 
 pressure. Heretofore the operation of punching the paper has been 
 quite arduous, as considerable force is required to perforate the 
 paper with the precision which is essential. The old punching in- 
 strument, with its three discs, is still retained, but it is now placed 
 within a tight-fitting iron frame. A small portion of the upper part 
 of this frame is boxed in, and the air pressure is allowed to enter 
 the box at one of the sides. Three cylindrical pistons protrude 
 from the bottom of the box, and these pistons are placed directly 
 over the discs of the ordinary punching instrument. The iron 
 frame-work is also furnished with three keys, representing the three 
 discs of the punching apparatus, and these keys control, by means 
 of a stiff wire, three valves within the inclosed box. A slight de- 
 pression of one of these keys thus opens its corresponding valve and 
 allows the air to rush in, which forces the desired piston down with 
 a quick, strong motion upon the disc beneath it, causing it to perfo- 
 rate the paper. The air pressure at present employed is six and 
 one half pounds to the square inch. 
 
128 THE WHEATSTONE AUTOMATIC. 
 
 Messages intended for transmission by the Wheatstone system 
 are first handed to the perforators or punchers for preparation. 
 The prepared paper is then placed on the transmitter by the clerk 
 in charge of that instrument. Another clerk is in charge of the 
 receiving instrument. As the messages are received on the receiv- 
 ing paper they are given to transcribing clerks, who translate the 
 Morse characters into writing. These clerks are termed " writers." 
 
 Thus, by the automatic system, each message in the course of 
 transmission and reception passes through the hands of four clerks, 
 as against two by the Morse system. 
 
 The Wheatstone automatic telegraph is the standard automatic 
 system of the British Postal Telegraph, and it is in extensive opera- 
 tion in that country. It is found very useful there, especially for 
 press reports, on account of the fact that many similar special dis- 
 patches are sent to several different cities, and, as when the paper 
 is once prepared, it may be used any number of times, much work 
 is thus dispensed with. The automatic system is also found very 
 serviceable in case of prostration of the wires, owing to storms or 
 from any other cause, as in the meantime the business may be pre- 
 pared for automatic transmission, and as soon as one or two wires, 
 are obtained, may be quickly disposed of. 
 
UNIVEESITY OF CALIFOENIA LIBEAKY 
 BEEKELEY 
 
 THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 TJooks not returned on time are subject to a fine of 
 
 s?^SS ! Sfc i HSs*&gs 
 
 demand may be renewed if application is made before 
 expiration of loan period. - 
 
 
 
 * 
 
 50m-7.'16 
 
YD 04194