If the operator desires to call Station A, she will press key K^{1} which will ground the upper side of the line and connect the lower side of the line with the generator G^{1}, and this, obviously, will cause the bell at Station A to ring. The bell at Station B will not ring because it is not in the circuit. If, on the other hand, the operator desires to ring the bell at Station B, she will depress key K^{2}, which will allow the current from generator G^{2} to pass over the upper side of the line through the bell and condenser at Station B and return by the path through the ground. The object of grounding the opposite sides of the keys at the central office is to prevent cross-ringing, that is, ringing the wrong bell. Were the keys not grounded this might occur when a ringing current was being sent out while the receiver at one of the stations was off its hook; the ringing current from, say, generator G^{1} then passing not only through the bell at Station A as intended, but also through the bell at Station B by way of the bridge path through the receiver that happened to be connected across the line. With the ringing keys grounded as shown, it is obvious that this will not occur, since the path for the ringing current through the wrong bell will always be shunted by a direct path to ground on the same side of the line.

In such a two-party-line selective system the two generators G^{1} and G^{2} may be the same generator and may be of the ordinary alternating-current type. The bells likewise may be of the ordinary alternating-current type.

The two-party selective line just described virtually employs two separate circuits for ringing. Now each of these circuits alone may be employed to accomplish selective ringing between two stations by using two biased bells oppositely polarized, and employing pulsating ringing currents of one direction or the other according to which bell it is desired to ring. One side of a circuit so equipped is shown in Fig. 171. In this the two biased bells are at Station A and Station B, these being bridged to ground in each case and adapted to respond only to positive and negative impulses respectively. At the central office the two keys K^{1} and K^{2} are shown. A single alternating-current generator G is shown, having its brush 1 grounded and brush 2 connected to a commutator disk 3 mounted on the generator shaft so as to revolve therewith. One-half of the periphery of this disk is of insulating material so that the brushes 4 and 5, which bear against the disk, will be alternately connected with the disk and, therefore, with the brush 2 of the generator. Now the brush 2, being one terminal of an alternating-current machine, is alternately positive and negative, and the arrangement of the commutator is such that the disk, which is always at the potential of the brush 2, will be connected to the brush 5 only while it is positively charged and with the brush 4 only while it is negatively charged. As a result, brush 5 has a succession of positive impulses and brush 4 a succession of negative ones. Obviously, therefore, when key K^{1} is depressed only the bell at Station A will be rung, and likewise the depression of key K^{2} will result only in the ringing of the bell at Station B.



Four-Party Line. From the two foregoing two-party line systems it is evident that a four-party line system may be readily obtained, that is, by employing two oppositely polarized biased bells on each side of the metallic circuit. The selection of any of the four bells may be obtained, choosing between the pairs connected, respectively, with the two limbs of the line, by choosing the limb on which the current is to be sent, and choosing between the two bells of the pair on that side of the line by choosing which polarity of current to send.

Such a four-party line system is shown in Fig. 172. In this the generators are not shown, but the wires leading from the four keys are shown marked plus or minus, according to the terminal of the generator to which they are supposed to be connected. Likewise the two bells connected with the lower side of the line are marked positive and negative, as are the two bells connected with the upper side of the line. From the foregoing description of Figs. 170 and 171, it is clear that if key K^{1} is pressed the bell at Station A will be rung, and that bell only, since the bells at Station C and Station D are not in the circuit and the positive current sent over the lower side of the line is not of the proper polarity to ring the bell at Station B.

The system shown in Fig. 172 is subject to one rather grave defect. In subsequent chapters it will be pointed out that in common-battery systems the display of the line signal at the central office is affected by any one of the subscribers merely taking his receiver off its hook and thus establishing a connection between the two limbs of the metallic circuit. Such common-battery systems should have the two limbs of the line, normally, entirely insulated from each other. It is seen that this is not the case in the system just described, since there is a conducting path from one limb of the line through the two bells on that side to ground, and thence through the other pair of bells to the other limb of the line. This means that unless the resistance of the bell windings is made very high, the path of the signaling circuit will be of sufficiently low resistance to actuate the line signal at the central office.



It is not feasible to overcome this objection by the use of condensers in series with the bells, as was done in the system shown in Fig. 170, since the bells are necessarily biased and such bells, as may readily be seen, will not work properly through condensers, since the placing of a condenser in their circuit means that the current which passes through the bell is alternating rather than pulsating, although the original source may have been of pulsating nature only.



The remedy for this difficulty, therefore, has been to place in series with each bell a very high non-inductive resistance of about 15,000 or 20,000 ohms, and also to make the windings of the bells of comparatively high resistance, usually about 2,500 ohms. Even with this precaution there is a considerable leakage of the central-office battery current from one side of the line to the other through the two paths to ground in series. This method of selective signaling has, therefore, been more frequently used with magneto systems. An endeavor to apply this principle to common-battery systems without the objections noted above has led to the adoption of a modification, wherein a relay at each station normally holds the ground connection open. This is shown in Fig. 173 and is the standard four-party line ringing circuit employed by the American Telephone and Telegraph Company and their licensees.

In this system the biased bells are normally disconnected from the line, and, therefore, the leakage path through them from one side of the line to the other does not exist. At each station there is a relay winding adapted to be operated by the ringing current bridged across the line in series with a condenser. As a result, when ringing current is sent out on the line all of the relays, i.e., one at each station, are energized and attract their armatures. This establishes the connection of all the bells to line and really brings about temporarily a condition equivalent to that of Fig. 172. As a result, the sending of a positive current on the lower line with a ground return will cause the operation of the bell at Station A. It will not ring the bell at Station B because of the wrong polarity. It will not ring the bells of Station C and Station D because they are in the circuit between the other side of the line and ground. As soon as the ringing current ceases all of the relays release their armatures and disconnect all the bells from the line.

By this very simple device the trouble, due to marginal working of the line signal, is done away with, since normally there is no leakage from one side of the line to the other on account of the presence of the condensers in the bridge at each station.



In Fig. 174, the more complete connections of the central-office ringing keys are shown, by means of which the proper positive or negative ringing currents are sent to line in the proper way to cause the ringing of any one of the four bells on a party line of either of the types shown in Figs. 172 and 173.

In this the generator G and its commutator disk 3, with the various brushes, 1, 2, 4, and 5, are arranged in the same manner as is shown in Fig. 171. It is evident from what has been said that wire 6 leading from generator brush 2 and commutator disk 3 will carry alternating potential; that wire 7 will carry positive pulsations of potential; and that wire 8 will carry negative pulsations of potential. There are five keys in the set illustrated in Fig. 174, of which four, viz, K^{1}, K^{2}, K^{3}, and K^{4}, are connected in the same manner as diagrammatically indicated in Figs. 172 and 173, and will, obviously, serve to send the proper current over the proper limb of the line to ring one of the bells. Key K^{5}, the fifth one in the set, is added so as to enable the operator to ring an ordinary unbiased bell on a single party line when connection is made with such line. As the two outside contacts of this key are connected respectively to the two brushes of the alternating-current dynamo G, it is clear that it will impress an alternating current on the line when its contacts are closed.

Circuits of Two-Party Line Telephones. In Fig. 175 is shown in detail the wiring of the telephone set usually employed in connection with the party-line selective-ringing system illustrated in Fig. 170. In the wiring of this set and the two following, it must be borne in mind that the portion of the circuit used during conversation might be wired in a number of ways without affecting the principle of selective ringing employed; however, the circuits shown are those most commonly employed with the respective selective ringing systems which they are intended to illustrate. In connecting the circuits of this telephone instrument to the line, the two line conductors are connected to binding posts 1 and 2 and a ground connection is made to binding post 3. In practice, in order to avoid the necessity of changing the permanent wiring of the telephone set in connecting it as an A or B Station (Fig. 170), the line conductors are connected to the binding posts in reverse order at the two stations; that is, for Station A the upper conductor, Fig. 170, is connected to binding post 1 and the lower conductor to binding post 2, while at Station B the upper conductor is connected to binding post 2 and the lower conductor to binding post 1. The permanent wiring of this telephone set is the same as that frequently used for a set connected to a line having only one station, the proper ringing circuit being made by the method of connecting up the binding posts. For example, if this telephone set were to be used on a single station line, the binding posts 1 and 2 would be connected to the two conductors of the line as before, while binding post 3 would be connected to post 1 instead of being grounded.



Circuits of Four-Party-Line Telephones. The wiring of the telephone set used with the system illustrated in Fig. 172 is shown in detail in Fig. 176. The wiring of this set is arranged for local battery or magneto working, as this method of selective ringing is more frequently employed with magneto systems, on account of the objectionable features which arise when applied to common-battery systems. In this figure the line conductors are connected to binding posts 1 and 2, and a ground connection is made to binding post 3. In order that all sets may be wired alike and yet permit the instrument to be connected for any one of the various stations, the bell is not permanently wired to any portion of the circuit but has flexible connections which will allow of the set being properly connected for any desired station. The terminals of the bell are connected to binding posts 9 and 10, to which are connected flexible conductors terminating in terminals 7 and 8. These terminals may be connected to the binding posts 4, 5, and 6 in the proper manner to connect the set as an A, B, C, or D station, as required. For example, in connecting the set for Station A, Fig. 172, terminal 7 is connected to binding post 6 and 8 to 5. For connecting the set for Station B terminal 7 is connected to binding post 5 and 8 to 6. For connecting the set for Station C terminal 7 is connected to binding post 6 and 8 to 4. For connecting the set for Station D terminal 7 is connected to binding post 4 and 8 to 6.



The detailed wiring of the telephone set employed in connection with the system illustrated in Fig. 173 is shown in Fig. 177. The wiring of this set is arranged for a common-battery system, inasmuch as this arrangement of signaling circuit is more especially adapted for common-battery working. However, this arrangement is frequently adapted to magneto systems as even with magneto systems a permanent ground connection at a subscriber's station is objectionable inasmuch as it increases the difficulty of determining the existence or location of an accidental ground on one of the line conductors. The wiring of this set is also arranged so that one standard type of wiring may be employed and yet allow any telephone set to be connected as an A, B, C, or D station.

Harmonic Method. Principles. To best understand the principle of operation of the harmonic party-line signaling systems, it is to be remembered that a flexible reed, mounted rigidly at one end and having its other end free to vibrate, will, like a violin string, have a certain natural period of vibration; that is, if it be started in vibration, as by snapping it with the fingers, it will take up a certain rate of vibration which will continue at a uniform rate until the vibration ceases altogether. Such a reed will be most easily thrown into vibration by a series of impulses having a frequency corresponding exactly to the natural rate of vibration of the reed itself; it may be thrown into vibration by very slight impulses if they occur at exactly the proper times.

It is familiar to all that a person pushing another in a swing may cause a considerable amplitude of vibration with the exertion of but a small amount of force, if he will so time his pushes as to conform exactly to the natural rate of vibration of the swing. It is of course possible, however, to make the swing take up other rates of vibrations by the application of sufficient force. As another example, consider a clock pendulum beating seconds. By gentle blows furnished by the escapement at exactly the proper times, the heavy pendulum is kept in motion. However, if a person grasps the pendulum weight and shakes it, it may be made to vibrate at almost any desired rate, dependent on the strength and agility of the individual.

The conclusion is, therefore, that a reed or pendulum may be made to start and vibrate easily by the application of impulses at proper intervals, and only with great difficulty by the application of impulses at other than the proper intervals; and these facts form the basis on which harmonic-ringing systems rest.

The father of harmonic ringing in telephony was Jacob B. Currier, an undertaker of Lowell, Mass. His harmonic bells were placed in series in the telephone line, and were considerably used in New England in commercial practice in the early eighties. Somewhat later James A. Lighthipe of San Francisco independently invented a harmonic-ringing system, which was put in successful commercial use at Sacramento and a few other smaller California towns. Lighthipe polarized his bells and bridged them across the line in series with condensers, as in modern practice, and save for some crudities in design, his apparatus closely resembled, both in principle and construction, some of that in successful use today.

Lighthipe's system went out of use and was almost forgotten, when about 1903, Wm. W. Dean again independently redeveloped the harmonic system, and produced a bell astonishingly like that of Lighthipe, but of more refined design, thus starting the development which has resulted in the present wide use of this system.

The signal-receiving device in harmonic-ringing systems takes the form of a ringer, having its armature and striker mounted on a rather stiff spring rather than on trunnions. By this means the moving parts of the bell constitute in effect a reed tongue, which has a natural rate of vibration at which it may easily be made to vibrate with sufficient amplitude to strike the gongs. The harmonic ringer differs from the ordinary polarized bell or ringer, therefore, in that its armature will vibrate most easily at one particular rate, while the armature of the ordinary ringer is almost indifferent, between rather wide limits, as to the rate at which it vibrates.

As a rule harmonic party-line systems are limited to four stations on a line. The frequencies employed are usually 16-2/3, 33-1/3, 50, and 66-2/3 cycles per second, this corresponding to 1,000, 2,000, 3,000, and 4,000 cycles per minute. The reason why this particular set of frequencies was chosen is that they represent approximately the range of desirable frequencies, and that the first ringing-current machines in such systems were made by mounting the armatures of four different generators on a single shaft, these having, respectively, two poles, four poles, six poles, and eight poles each. The two-pole generator gave one cycle per revolution, the four-pole two, the six-pole three, and the eight-pole four, so that by running the shaft of the machine at exactly 1,000 revolutions per minute the frequencies before mentioned were attained. This range of frequencies having proved about right for general practice and the early ringers all having been attuned so as to operate on this basis, the practice of adhering to these numbers of vibrations has been kept up with one exception by all the manufacturers who make this type of ringer.

Tuning. The process of adjusting the armature of a ringer to a certain rate of vibration is called tuning, and it is customary to refer to a ringer as being tuned to a certain rate of vibration, just as it is customary to refer to a violin string as being tuned to a certain pitch or rate of vibration.

The physical difference between the ringers of the various frequencies consists mainly in the size of the weights at the end of the vibrating reed, that is, of the weights which form the tapper for the bell. The low-frequency ringers have the largest weights and the high-frequency the smallest, of course. The ringers are roughly tuned to the desired frequencies by merely placing on the tapper rod the desired weight and then a more refined tuning is given them by slightly altering the positions of the weights on the tapper rod. To make the reed have a slightly lower natural rate of vibration, the weight is moved further from the stationary end of the reed, while to give it a slightly higher natural rate of vibration the weight is moved toward the stationary. In this way very nice adjustments may be made, and the aim of the various factories manufacturing these bells is to make the adjustment permanent so that it will never have to be altered by the operating companies. Several years of experience with these bells has shown that when once properly assembled they maintain the same rate of vibration with great constancy.

There are two general methods of operating harmonic bells. One of these may be called the in-tune system and the other the under-tune system. The under-tune system was the first employed.



Under-Tune System. The early workers in the field of harmonic-selective signaling discovered that when the tapper of the reed struck against gongs the natural rate of vibration of the reed was changed, or more properly, the reed was made to have a different rate of vibration from its natural rate. This was caused by the fact that the elasticity of the gongs proved another factor in the set of conditions causing the reeds to take up a certain rate of vibration, and the effect of this added factor was always to accelerate the rate of vibration which the reed had when it was not striking the gongs. The rebound of the hammer from the gongs tended, in other words, to accelerate the rate of vibration, which, as might be expected, caused a serious difficulty in the practical operation of the bells. To illustrate: If a reed were to have a natural rate of vibration, when not striking the gongs, of 50 per second and a current of 50 cycles per second were impressed on the line, the reed would take up this rate of vibration easily, but when a sufficient amplitude of vibration was attained to cause the tapper to strike the gongs, the reed would be thrown out of tune, on account of the tendency of the gongs to make the reed vibrate at a higher rate. This caused irregular ringing and was frequently sufficient to make the bells cease ringing altogether or to ring in an entirely unsatisfactory manner.

In order to provide for this difficulty the early bells of Currier and Lighthipe were made on what has since been called the "under-tuned" principle. The first bells of the Kellogg Switchboard and Supply Company, developed by Dean, were based on this idea as their cardinal principle. The reeds were all given a natural rate of vibration, when not striking the gongs, somewhat below that of the current frequencies to be employed; and yet not sufficiently below the corresponding current frequency to make the bell so far out of tune that the current frequency would not be able to start it. This was done so that when the tapper began to strike the gongs the tapper would be accelerated and brought practically into tune with the current frequency, and the ringing would continue regularly as long as the current flowed. It will be seen that the under-tuned system was, therefore, one involving some difficulty in starting in order to provide for proper regularity while actually ringing.

Ringers of this kind were always made with but a single gong, it being found difficult to secure uniformity of ringing and uniformity of adjustment when two gongs were employed. Although no ringers of this type are being made at present, yet a large number of them are in use and they will consequently be described. Their action is interesting in throwing better light on the more improved types, if for no other reason.

Figs. 178 and 179 show, respectively, side and front views of the original Kellogg bell. The entire mechanism is self-contained, all parts being mounted on the base plate 1. The electromagnet is of the two-coil type, and is supported on the brackets 2 and 3. The bracket 2 is of iron so as to afford a magnetic yoke for the field of the electromagnet, while the bracket 3 is of brass so as not to short-circuit the magnetic lines across the air-gap. The reed tongue—consisting of the steel spring 5, the soft-iron armature pieces 6, the auxiliary spring 7, and the tapper ball 8, all of which are riveted together, as shown in Fig. 178—constitutes the only moving part of the bell. The steel spring 5 is rigidly mounted in the clamping piece 9 at the upper part of the bracket 3, and the reed tongue is permitted to vibrate only by the flexibility of this spring. The auxiliary spring 7 is much lighter than the spring 5 and has for its purpose the provision of a certain small amount of flexibility between the tapper ball and the more rigid portion of the armature formed by the iron strips 6-6. The front ends of the magnet pole pieces extend through the bracket 3 and are there provided with square soft-iron pole pieces 10 set at right angles to the magnet cores so as to form a rather narrow air-gap in which the armature may vibrate.



The cores of the magnet and also the reed tongue are polarized by means of the L-shaped bar magnet 4, mounted on the iron yoke 2 at one end in such manner that its other end will lie quite close to the end of the spring 5, which, being of steel, will afford a path for the lines of force to the armature proper. We see, therefore, that the two magnet cores are, by this permanent magnet, given one polarity, while the reed tongue itself is given the other polarity, this being exactly the condition that has already been described in connection with the regular polarized bell or ringer.

The electromagnetic action by which this reed tongue is made to vibrate is, therefore, exactly the same as that of an ordinary polarized ringer, but the difference between the two is that, in this harmonic ringer, the reed tongue will respond only to one particular rate of vibrations, while the regular polarized ringer will respond to almost any.

As shown in Fig. 178, the tapper ball strikes on the inside surface of the single gong. The function of the auxiliary spring 7 between the ball and the main portion of the armature is to allow some resilience between the ball and the balance of the armature so as to counteract in some measure the accelerating influence of the gong on the armature. In these bells, as already stated, the natural rate of vibration of the reed tongue was made somewhat lower than the rate at which the bell was to be operated, so that the reed tongue had to be started by a current slightly out of tune with it, and then, as the tapper struck the gong, the acceleration due to the gong would bring the vibration of the reed tongue, as modified by the gong, into tune with the current that was operating it. In ether words, in this system the ringing currents that were applied to the line had frequencies corresponding to what may be called the operative rates of vibration of the reed tongues, which operative rates of vibration were in each case the resultant of the natural pitch of the reed as modified by the action of the bell gong when struck.



In-Tune System. The more modern method of tuning is to make the natural rate of vibration of the reed tongue, that is, the rate at which it naturally vibrates when not striking the gongs, such as to accurately correspond to the rate of vibration at which the bells are to be operated—that is, the natural rate of vibration of the reed tongues is made the same as the operative rate. Thus the bells are attuned for easy starting, a great advantage over the under-tuned system. In the under-tuned system, the reeds being out of tune in starting require heavier starting current, and this is obviously conducive to cross-ringing, that is, to the response of bells to other than the intended frequency.

Again, easy starting is desirable because when the armature is at rest, or in very slight vibration, it is at a maximum distance from the poles of the electromagnet, and, therefore, subject to the weakest influence of the poles. A current, therefore, which is strong enough to start the vibration, will be strong enough to keep the bell ringing properly.



When with this "in-tune" mode of operation, the armature is thrown into sufficiently wide vibration to cause the tapper to strike the gong, the gong may tend to accelerate the vibration of the reed tongue, but the current impulses through the electromagnet coils continue at precisely the same rates as before. Under this condition of vibration, when the reed tongue has an amplitude of vibration wide enough to cause the tapper to strike the gongs, the ends of the armature come closest to the pole pieces, so that the pole pieces have their maximum magnetic effect on the armature, with the result that even if the accelerating tendency of the gongs were considerable, the comparatively large magnetic attractive impulses occurring at the same rate as the natural rate of vibration of the reed tongue, serve wholly to prevent any actual acceleration of the reed tongue. The magnetic attractions upon the ends of the armature, continuing at the initial rate, serve, therefore, as a check to offset any accelerating tendency which the striking of the gong may have upon the vibrating reed tongue.

It is obvious, therefore, that in the "in-tune" system the electromagnetic effect on the armature should, when the armature is closest to the pole pieces, be of such an overpowering nature as to prevent whatever accelerating tendency the gongs may have from throwing the armature out of its "stride" in step with the current. For this reason it is usual in this type to so adjust the armature that its ends will actually strike against the pole pieces of the electromagnet when thrown into vibration. Sufficient flexibility is given to the tapper rod to allow it to continue slightly beyond the point at which it would be brought to rest by the striking of the armature ends against the pole pieces and thus exert a whipping action so as to allow the ball to continue in its movement far enough to strike against the gongs. The rebound of the gong is then taken up by the elasticity of the tapper rod, which returns to an unflexed position, and at about this time the pole piece releases the armature so that it may swing over in the other direction to cause the tapper to strike the other gong.



The construction of the "in-tune" harmonic ringer employed by the Dean Electric Company, of Elyria, Ohio, is illustrated in Figs. 180, 181, and 182. It will be seen from Fig. 180 that the general arrangement of the magnet and armature is the same as that of the ordinary polarized ringer; the essential difference is that the armature is spring-mounted instead of pivoted. The armature and the tapper rod normally stand in the normal central position with reference to the pole pieces of the magnet and the gongs. Fig. 181 shows the complete vibrating parts of four ringers, adapted, respectively, to the four different frequencies of the system. The assembled armature, tapper rod, and tapper are all riveted together and are non-adjustable. All of the adjustment that is done upon them is done in the factory and is accomplished, first, by choosing the proper size of weight, and second, by forcing this weight into the proper position on the tapper rod to give exactly the rate of vibration that is desired.



An interesting feature of this Dean harmonic ringer is the gong adjustment. As will be seen, the gongs are mounted on posts which are carried on levers pivoted to the ringer frame. These levers have at their outer end a curved rack provided with gear teeth adapted to engage a worm or screw thread mounted on the ringer frame. Obviously, by turning this worm screw in one direction or the other, the gongs are moved slightly toward or from the armature or tapper. This affords a very delicate means of adjusting the gongs, and at the same time one which has no tendency to work loose or to get out of adjustment.



In Fig. 183 is shown a drawing of the "in-tune" harmonic ringer manufactured by the Kellogg Switchboard and Supply Company. This differs in no essential respect from that of the Dean Company, except in the gong adjustment, this latter being affected by a screw passing through a nut in the gong post, as clearly indicated.

In both the Kellogg and the Dean in-tune ringers, on account of the comparative stiffness of the armature springs and on account of the normal position of the armature with maximum air gaps and consequent minimum magnetic pull, the armature will practically not be affected unless the energizing current is accurately attuned to its own natural rate. When the proper current is thrown on to the line, the ball will be thrown into violent vibration, and the ends of the armature brought into actual contact with the pole pieces, which are of bare iron and shielded in no way. The armature in this position is very strongly attracted and comes to a sudden stop on the pole pieces. The gongs are so adjusted that the tapper ball will have to spring about one thirty-second of an inch in order to hit them. The armature is held against the pole piece while the tapper ball is engaged in striking the gong and in partially returning therefrom, and so strong is the pull of the pole piece on the armature in this position that the accelerating influence of the gong has no effect in accelerating the rate of vibration of the reed.



Circuits. In Fig. 184 are shown in simplified form the circuits of a four-station harmonic party line. It is seen that at the central office there are four ringing keys, adapted, respectively, to impress on the line ringing currents of four different frequencies. At the four stations on the line, lettered A, B, C, and D, there are four harmonic bells tuned accordingly. At Station A there is shown the talking apparatus employing the Wheatstone bridge arrangement. The talking apparatus at all of the other stations is exactly the same, but is omitted for the sake of simplicity. A condenser is placed in series with each of the bells in order that there may be no direct-current path from one side of the line to the other when all of the receivers are on their hooks at the several stations.

In Fig. 185 is shown exactly the same arrangement, with the exception that the talking apparatus illustrated in detail at Station A is that of the Kellogg Switchboard and Supply Company. Otherwise the circuits of the Dean and the Kellogg Company, and in fact of all the other companies manufacturing harmonic ringing systems, are the same.

Advantages. A great advantage of the harmonic party-line system is the simplicity of the apparatus at the subscriber's station. The harmonic bell is scarcely more complex than the ordinary polarized ringer, and the only difference between the harmonic-ringing telephone and the ordinary telephone is in the ringer itself. The absence of all relays and other mechanism and also the absence of the necessity for ground connections at the telephone are all points in favor of the harmonic system.



Limitations. As already stated, the harmonic systems of the various companies, with one exception, are limited to four frequencies. The exception is in the case of the North Electric Company, which sometimes employs four and sometimes five frequencies and thus gets a selection between five stations. In the four-party North system, the frequencies, unlike those in the Dean and Kellogg systems, wherein the higher frequencies are multiples of the lower, are arranged so as to be proportional to the whole numbers 5, 7, 9, and 11, which, of course, have no common denominator. The frequencies thus employed in the North system are, in cycles per second, 30.3, 42.4, 54.5, and 66.7. In the five-party system, the frequency of 16.7 is arbitrarily added.

While all of the commercial harmonic systems on the market are limited to four or five frequencies, it does not follow that a greater number than four or five stations may not be selectively rung. Double these numbers may be placed on a party line and selectively actuated, if the first set of four or five is bridged across the line and the second set of four or five is connected between one limb of the line and ground. The first set of these is selectively rung, as already described, by sending the ringing currents over the metallic circuit, while the second set may be likewise selectively rung by sending the ringing currents over one limb of the line with a ground return. This method is frequently employed with success on country lines, where it is desired to place a greater number of instruments on a line than four or five.

Step-by-Step Method. A very large number of step-by-step systems have been proposed and reduced to practice, but as yet they have not met with great success in commercial telephone work, and are nowhere near as commonly used as are the polarity and harmonic systems.

Principles. An idea of the general features of the step-by-step systems may be had by conceiving at each station on the line a ratchet wheel, having a pawl adapted to drive it one step at a time, this pawl being associated with the armature of an electromagnet which receives current impulses from the line circuit. There is thus one of these driving magnets at each station, each bridged across the line so that when a single impulse of current is sent out from the central office all of the ratchet wheels will be moved one step. Another impulse will move all of the ratchet wheels another step, and so on throughout any desired number of impulses. The ratchet wheels, therefore, are all stepped in unison.

Let us further conceive that all of these ratchet wheels are provided with a notch or a hole or a projection, alike in all respects at all stations save in the position which this notch or hole or projection occupies on the wheel. The thing to get clear in this part of the conception is that all of these notches, holes, or projections are alike on all of the wheels, but they occupy a different position on the wheel for each one of the stations.

Consider further that the bell circuit at each of the stations is normally open, but that in each case it is adapted to be closed when the notch, hole, or projection is brought to a certain point by the revolution of the wheel.

Let us conceive further that this distinguishing notch, hole, or projection is so arranged on the wheel of the first station as to close the bell circuit when one impulse has been sent, that that on the second station will close the bell circuit after the second impulse has been sent, and so on throughout the entire number of stations. It will, therefore, be apparent that the bell circuits at the various stations will, as the wheels are rotated in unison, be closed one after the other. In order to call a given station, therefore, it is only necessary to rotate all of the wheels in unison, by sending out the proper stepping impulses until they all occupy such a position that the one at the desired station is in such position as to close the bell circuit at that station. Since all of the notches, holes, or projections are arranged to close the bell circuits at their respective stations at different times, it follows that when the bell circuit at the desired station is closed those at all of the other stations will be open. If, therefore, after the proper number of stepping impulses has been sent to the line to close the bell circuit of the desired station, ringing current be applied to the line, it is obvious that the bell of that one station will be rung to the exclusion of all others. It is, of course, necessary that provision be made whereby the magnets which furnish the energy for stepping the wheels will not be energized by the ringing current. This is accomplished in one of several ways, the most common of which is to have the stepping magnets polarized or biased in one direction and the bells at the various stations oppositely biased, so that the ringing current will not affect the stepping magnet and the stepping current will not affect the ringer magnets.

After a conversation is finished, the line may be restored to its normal position in one of several ways. Usually so-called release magnets are employed, for operating on the releasing device at each station. These, when energized, will withdraw the holding pawls from the ratchets and allow them all to return to their normal positions. Sometimes these release magnets are operated by a long impulse of current, being made too sluggish in their action to respond to the quick-stepping impulses; sometimes the release magnets are tapped from one limb of the line to ground, so as not to be affected by the stepping or ringing currents sent over the metallic circuit; and sometimes other expedients are used for obtaining the release of the ratchets at the proper time, a large amount of ingenuity having been spent to this end.

As practically all step-by-step party-line systems in commercial use have also certain other features intended to assure privacy of conversation to the users, and, therefore, come under the general heading of lock-out party-line systems, the discussion of commercial examples of these systems will be left for the next chapter, which is devoted to such lock-out systems.

Broken-Line Method. The broken-line system, like the step-by-step system, is also essentially a lock-out system and for that reason only its general features, by which the selective ringing is accomplished, will be dealt with here.

Principles. In this system there are no tuned bells, no positively and negatively polarized bells bridged to ground on each side of the line, and no step-by-step devices in the ordinary sense, by which selective signaling has ordinarily been accomplished on party lines. Instead of this, each instrument on the line is exclusively brought into operative relation with the line, and then removed from such operative relation until the subscriber wanted is connected, at which time all of the other instruments are locked out and the line is not encumbered by any bridge circuits at any of the instruments that are not engaged in the conversation. Furthermore, in the selecting of a subscriber or the ringing of his bell there is no splitting up of current among the magnets at the various stations as in ordinary practice, but the operating current goes straight to the station desired and to that station alone where its entire strength is available for performing its proper work.

In order to make the system clear it may be stated at the outset that one side of the metallic circuit line is continued as in ordinary practice, passing through all of the stations as a continuous conductor. The other side of the line, however, is divided into sections, its continuity being broken at each of the subscriber's stations. Fig. 186 is intended to show in the simplest possible way how the circuit of the line may be extended from station to station in such manner that only the ringer of one station is in circuit at a time. The two sides of the line are shown in this figure, and it will be seen that limb _L_ extends from the central office on the left to the last station on the right without a break. The limb _R_, however, extends to the first station, at which point it is cut off from the extension _R_{x}_ by the open contacts of a switch. For the purpose of simplicity this switch is shown as an ordinary hand switch, but as a matter of fact it is a part of a relay, the operating coil of which is shown at _6_, just above it, in series with the ringer.



Obviously, if a proper ringing current is sent over the metallic circuit from the central office, only the bell at Station A will operate, since the bells at the other stations are not in the circuit. If by any means the switch lever _2_ at Station A were moved out of engagement with contact _1_ and into engagement with contact _3_, it is obvious that the bell of Station A would no longer be in circuit, but the limb _R_ of the line would be continued to the extension _R_{x}_ and the bell of Station B would be in circuit. Any current then sent over the circuit of the line from the central office would ring the bell of this station. In Fig. 187 the switches of both Station A and Station B have been thus operated, and Station C is thus placed in circuit. Inspection of this figure will show that the bells of Station A, Station B, and Station D are all cut out of circuit, and that, therefore, no current from the central office can affect them. This general scheme of selection is a new-comer in the field, and for certain classes of work it is of undoubted promise.



CHAPTER XVII

LOCK-OUT PARTY-LINE SYSTEMS

The party-line problem in rural districts is somewhat different from that within urban limits. In the latter cases, owing to the closer grouping of the subscribers, it is not now generally considered desirable, even from the standpoint of economy, to place more than four subscribers on a single line. For such a line selective ringing is simple, both from the standpoint of apparatus and operation; and moreover owing to the small number of stations on a line, and the small amount of traffic to and from such subscribers as usually take party-line service, the interference between parties on the same line is not a very serious matter.

For rural districts, particularly those tributary to small towns, these conditions do not exist. Owing to the remoteness of the stations from each other it is not feasible from the standpoint of line cost to limit the number of stations to four. A much greater number of stations is employed and the confusion resulting is distressing not only to the subscribers themselves but also to the management of the company. There exists then the need of a party-line system which will give the limited user in rural districts a service, at least approaching that which he would get if served by an individual line.

The principal investment necessary to provide facilities for telephone service is that required to produce the telephone line. In many cases the cost of instruments and apparatus is small in comparison with the cost of the line. By far the greater number of subscribers in rural districts are those who use their instruments a comparatively small number of times a day, and to maintain an expensive telephone line for the exclusive use of one such subscriber who will use it but a few minutes each day is on its face an economic waste. As a result, where individual line service is practiced exclusively one of two things must be true: either the average subscriber pays more for his service than he should, or else the operating company sells the service for less than it costs, or at best for an insufficient profit. Both of these conditions are unnatural and cannot be permanent.

The party-line method of giving service, by which a single line is made to serve a number of subscribers, offers a solution to this difficulty, but the ordinary non-selective or even selective party line has many undesirable features if the attempt is made to place on it such a large number of stations as is considered economically necessary in rural work. These undesirable features work to the detriment of both the user of the telephone and the operating company.

Many attempts have been made to overcome these disadvantages of the party line in sparsely settled communities, by producing what are commonly called lock-out systems. These, as their name implies, employ such an arrangement of parts that when the line is in use by any two parties, all other parties are locked out from the circuit and cannot gain access to it until the parties who are using it are through. System after system for accomplishing this purpose has been announced but for the most part these have involved such a degree of complexity and have introduced so many undesirable features as to seriously affect the smooth operation of the system and the reliability of the service.

We believe, however, in spite of numerous failures, that the lock-out selective-signaling party line has a real field of usefulness and that operating companies as well as manufacturing companies are beginning to appreciate this need, and as a result that the relief of the rural subscriber from the almost intolerable service he has often had to endure is at hand. A few of the most promising lock-out party-line systems now before the public will, therefore, be described in some detail.

Poole System. The Poole system is a lock-out system pure and simple, its devices being in the nature of a lock-out attachment for selective-signaling lines, either of the polarity or of the harmonic type wherein common-battery transmission is employed. It will be here described as employed in connection with an ordinary harmonic-ringing system.

In Fig. 188 there is shown a four-station party line equipped with Poole lock-out devices, it being assumed that the ringers at each station are harmonic and that the keys at the central office are the ordinary keys adapted to impress the proper frequency on the line for ringing any one of the stations. In addition to the ordinary talking and ringing apparatus at each subscriber's station, there is a relay of special form and also a push-button key.



Each of the relays has two windings, one of high resistance and the other of low resistance. Remembering that the system to which this device is applied is always a common-battery system, and that, therefore, the normal condition of the line will be one in which there is a difference of potential between the two limbs, it will be evident that whenever any subscriber on a line that is not in use raises his receiver from its hook, a circuit will be established from the upper contact of the hook through the lever of the hook to the high-resistance winding 1 of the relay and thence to the other side of the line by way of wire 6. This will result in current passing through the high-resistance winding of the relay and the relay will pull up its armature. As soon as it does so it establishes two other circuits by the closure of the relay armature against the contacts 4 and 5.

The closing of the contact 4 establishes a circuit from the upper side of the line through the upper contact of the switch hook, thence through the contacts of the push button 3, thence through the low-resistance winding 2 of the relay to the terminal 4, thence through the relay armature and the transmitter to the lower side of the line. This low-resistance path across the line serves to hold the relay armature attracted and also to furnish current to the transmitter for talking. The establishment of this low-resistance path across the line does another important thing, however; it practically short-circuits the line with respect to all the high-resistance relay windings, and thus prevents any of the other high-resistance relay windings from receiving enough current to actuate them, should the subscriber at any other station remove his receiver from the hook in an attempt to listen in or to make a call while the line is in use. As a subscriber can only establish the proper conditions for talking and listening by the attraction of this relay armature at his station, it is obvious that unless he can cause the pulling up of his relay armature he can not place himself in communication with the line.

The second thing that is accomplished by the pulling up of the relay armature is the closure of the contacts 5, and that completes the talking circuit through the condenser and receiver across the line in an obvious fashion. The result of this arrangement is that it is the first party who raises his receiver from its hook who is enabled to successfully establish a connection with the line, all subsequent efforts, by other subscribers, failing to do so because of the fact that the line is short-circuited by the path through the low-resistance winding and the transmitter of the station that is already connected with the line.

A little target is moved by the action of the relay so that a visual indication is given to the subscriber in making a call to show whether or not he is successful in getting the use of the line. If the relay operates and he secures control of the line, the target indicates the fact by its movement, while if someone else is using the line and the relay does not operate, the target, by its failure to move, indicates that fact.

When one party desires to converse with another on the same line, he depresses the button 3 at his station until after the called party has been rung and has responded. This holds the circuit of his low-resistance winding open, and thus prevents the lock-out from becoming effective until the called party is connected with the line. The relay armature of the calling party does not fall back with the establishment of the low-resistance path at the called station, because, even though shunted, it still receives sufficient current to hold its armature in its attracted position. After the called party has responded, the button at the calling station is released and both low-resistance holding coils act in multiple.



No induction coil is used in this system and the impedance of the holding coil is such that incoming voice currents flow through the condenser and the receiver, which, by reference to the figure, will be seen to be in shunt with the holding coil. The holding coil is in series with the local transmitter, thus making a circuit similar to that of the Kellogg common-battery talking circuit already discussed.

A possible defect in the use of this system is one that has been common to a great many other lock-out systems, depending for their operation on the same general plan of action. This appears when the instruments are used on a comparatively long line. Since the locking-out of all the instruments that are not in use by the one that is in use depends on the low-resistance shunt that is placed across the line by the instrument that is in use, it is obvious that, in the case of a long line, the resistance of the line wire will enter into the problem in such a way as to tend to defeat the locking-out function in some cases. Thus, where the first instrument to use the line is at the remote end of the line, the shunting effect that this instrument can exert with respect to another instrument near the central office is that due to the resistance of the line plus the resistance of the holding coil at the end instrument. The resistance of the line wire may be so high as to still allow a sufficient current to flow through the high-resistance coil at the nearer station to allow its operation, even though the more remote instrument is already in use.

Coming now to a consideration of the complete selective-signaling lock-out systems, wherein the selection of the party and the locking out of the others are both inherent features, a single example of the step-by-step, and of the broken-line selective lock-out systems will be discussed.

Step-by-Step System. The so-called K.B. system, manufactured by the Dayton Telephone Lock-out Manufacturing Company of Dayton, Ohio, operates on the step-by-step principle. The essential feature of the subscriber's telephone equipment in this system is the step-by-step actuating mechanism which performs also the functions of a relay. This device consists of an electromagnet having two cores, with a permanent polarizing magnet therebetween, the arrangement in this respect being the same as in an ordinary polarized bell. The armature of this magnet works a rocker arm, which, besides stepping the selector segment around, also, under certain conditions, closes the bell circuit and the talking circuit, as will be described.



Referring first to Fig. 189, which shows in simplified form a four-station K.B. lock-out line, the electromagnet is shown at 1 and the rocker arm at 2. The ratchet 3 in this case is not a complete wheel but rather a segment thereof, and it is provided with a series of notches of different depths. It is obvious that the depth of the notches will determine the degree of movement which the upper end of the rocker arm may have toward the left, this being dependent on the extent to which the pawl 6 is permitted to enter into the segment. The first or normal notch, i.e., the top notch, is always of such a depth that it will allow the rocker-arm lever 2 to engage the contact lever 4, but will not permit the rocker arm to swing far enough to the left to cause that contact to engage the bell contact 5. As will be shown later, the condition for the talking circuit to be closed is that the rocker arm 2 shall rest against the contact 4; and from this we see that the normal notch of each of the segments 3 is of such a depth as to allow the talking circuit at each station to be closed. The next notch, i.e., the second one in each disk, is always shallow, as are all of the other notches except one. A deep notch is placed on each disk anywhere from the third to the next to the last on the segment. This deep notch is called the selective notch, and it is the one that allows of contact being made with the ringer circuit of that station when the pawl 6 drops into it. The position of this notch differs on all of the segments on a line, and obviously, therefore, the ringer circuit at any station may be closed to the exclusion of all the others by stepping all of the segments in unison until the deep notch on the segment of the desired station lies opposite to the pawl 6, which will permit the rocker arm 2 to swing so far to the left as to close not only the circuit between 2 and 4, but also between 2, 4, and 5. In this position the talking and the ringing circuits are both closed.

The position of the deepest notch, i.e., the selective notch, on the circumference of the segment at any station depends upon the number of that station; thus, the segment of Station 4 will have a deep notch in the sixth position; the segment for Station 9 will have a deep notch in the eleventh position; the segment for any station will have a deep notch in the position corresponding to the number of that station plus two.

From what has been said, therefore, it is evident that the first, or normal, notch on each segment is of such a depth as to allow the moving pawl 6 to fall to such a depth in the segment as to permit the rocker arm 2 to close the talking circuit only. All of the other notches, except one, are comparatively shallow, and while they permit the moving pawl 6 under the influence of the rocker arm 2 to move the segment 3, yet they do not permit the rocker arm 2 to move so far to the left as to close even the talking circuit. The exception is the deep notch, or selective notch, which is of such depth as to permit the pawl 6 to fall so far into the segment as to allow the rocker arm 2 to close both the talking and the ringing circuits. Besides the moving pawl 6 there is a detent pawl 7. This always holds the segment 3 in the position to which it has been last moved by the moving pawl 6.

The actuating magnet 1, as has been stated, is polarized and when energized by currents in one direction, the rocker arm moves the pawl 6 so as to step the segment one notch. When this relay is energized by current in the opposite direction, the operation is such that both the moving pawl 6 and the detent pawl 7 will be pulled away from the segment, thus allowing the segment to return to its normal position by gravity. This is accomplished by the following mechanism: An armature stop is pivoted upon the face of the rocker arm so as to swing in a plane parallel to the pole faces of the relay, and is adapted, when the relay is actuated by selective impulses of one polarity, to be pulled towards one of the pole faces where it acts, through impact with a plate attached to the pole face of the relay, as a limiting means for the motion of the rocker arm when the rocker arm is actuated by the magnet. When, however, the relay is energized by current in the opposite direction, as on a releasing impulse, the armature stop swings upon its pivot towards the opposite pole face, in which position the lug on the end of the armature stop registers with a hole in the plate on the relay, thus allowing the full motion of the rocker arm when it is attracted by the magnet. This motion of the rocker arm withdraws the detent pawl from engagement with the segment as well as the moving pawl, and thereby permits the segment to return to its normal position. As will be seen from Fig. 189, each of the relay magnets 1 is permanently bridged across the two limbs of the line.

Each station is provided with a push button, not shown, by means of which the subscriber who makes a call may prevent the rocker arm of his instrument from being actuated while selective impulses are being sent over the line. The purpose of this is to enable one party to make a call for another on the same line, depressing his push button while the operator is selecting and ringing the called party. The segment at his own station, therefore, remains in its normal position, in which position, as we have already seen, his talking circuit is closed; all of the other segments are, however, stepped up until the ringing and talking circuits of the desired station are in proper position, at which time ringing current is sent over the line. The segments in Fig. 189, except at Station C, are shown as having been stepped up to the sixth position, which corresponds to the ringing position of the fourth station, or Station D. The condition shown in this figure corresponds to that in which the subscriber at Station C originated the call and pressed his button, thus retaining his own segment in its normal position so that the talking circuits would be established with Station D.

When the line is in normal position any subscriber may call central by his magneto generator, not shown in Fig. 189, which will operate the drop at central, but will not operate any of the subscribers' bells, because all bell circuits are normally open. When a subscriber desires connection with another line, the operator sends an impulse back on the line which steps up and locks out all instruments except that of the calling subscriber.



A complete K.B. lock-out telephone is shown in Fig. 190. This is the type of instrument that is usually furnished when new equipment is ordered. If, however, it is desired to use the K.B. system in connection with telephones of the ordinary bridging type that are already in service, the lock-out and selective mechanism, which is shown on the upper inner face of the door in Fig. 190, is furnished separately in a box that may be mounted close to the regular telephone and connected thereto by suitable wires, as shown in Fig. 191. It is seen that this instrument employs a local battery for talking and also a magneto generator for calling the central office.

The central-office equipment consists of a dial connected with an impulse wheel, together with suitable keys by which the various circuits may be manipulated. This dial and its associated mechanism may be mounted in the regular switchboard cabinet, or it may be furnished in a separate box and mounted alongside of the cabinet in either of the positions shown at 1 or 2 of Fig. 192.

In order to send the proper number of impulses to the line to call a given party, the operator places her finger in the hole in the dial that bears the number corresponding to the station wanted and rotates the dial until the finger is brought into engagement with the fixed stop shown at the bottom of the dial in Fig. 192. The dial is then allowed to return by the action of a spring to its normal position, and in doing so it operates a switch within the box to make and break the battery circuit the proper number of times.

Operation. A complete description of the operation may now be had in connection with Fig. 193, which is similar to Fig. 189, but contains the details of the calling arrangement at the central office and also of the talking circuits at the various subscribers' stations.



Referring to the central-office apparatus the usual ringing key is shown, the inside contacts of which lead to the listening key and to the operator's telephone set as in ordinary switchboard practice. Between the outside contact of this ringing key and the ringing generator there is interposed a pair of contact springs 8-8 and another pair 9-9. The contact springs 8 are adapted to be moved backward and forward by the impulse wheel which is directly controlled by the dial under the manipulation of the operator. When these springs 8 are in their normal position, the ringing circuit is continued through the release-key springs 9 to the ringing generator. These springs 8 occupy their normal position only when the dial is in its normal position, this being due to the notch 10 in the contact wheel. At all other times, i.e., while the impulse wheel is out of its normal position, the springs 8-8 are either depressed so as to engage the lower battery contacts, or else held in an intermediate position so as to engage neither the battery contacts nor the generator contacts.



When it is desired to call a given station, the operator pulls the subscriber's number on the dial and holds the ringing key closed, allowing the dial to return to normal. This connects the impulse battery to the subscriber's line as many times as is required to move the subscriber's sectors to the proper position, and in such direction as to cause the stepping movement of the various relays. As the impulse wheel comes to its normal position, the springs 8, associated with it, again engage their upper contacts, by virtue of the notch 10 in the impulse wheel, and this establishes the connection between the ringing generator and the subscriber's line, the ringing key being still held closed. The pulling of the transmitter dial and holding the ringing key closed, therefore, not only sends the stepping impulses to line, but also follows it by the ringing current. The sending of five impulses to line moves all of the sectors to the sixth notch, and this corresponds to the position necessary to make the fourth station operative. Such a condition is shown in Fig. 193, it being assumed that the subscriber at Station C originated the call and pressed his own button so as to prevent his sector from being moved out of its normal position. As a result of this, the talking circuit at Station C is left closed, and the talking and the ringing circuit of Station D, the called station, are closed, while both the talking and the ringing circuits of all the other stations are left open. Station D may, therefore, be rung and may communicate with Station C, while all of the other stations on the line are locked out, because of the fact that both their talking and ringing circuits are left open.



When conversation is ended, the operator is notified by the usual clearing-out signal, and she then depresses the release button, which brings the springs 9 out of engagement with the generator contact but into engagement with the battery contact in such relation as to send a battery current on the line in the reverse direction from that sent out by the impulse wheel. This sends current through all of the relays in such direction as to withdraw both the moving and the holding pawls from the segments and thus allow all of the segments to return to their normal positions. Of course, in thus establishing the release current, it is necessary for the operator to depress the ringing key as well as the release key.

A one-half microfarad condenser is placed in the receiver circuit at each station so that the line will not be tied up should some subscriber inadvertently leave his receiver off its hook. This permits the passage of voice currents, but not of the direct currents used in stepping the relays or in releasing them.

The circuit of Fig. 193 is somewhat simplified from that in actual practice, and it should be remembered that the hook switch, which is not shown in this figure, controls in the usual way the continuity of the receiver and the transmitter circuits as well as of the generator circuits, the generator being attached to the line as in an ordinary telephone.

Broken-Line System. The broken-line method of accomplishing selective signaling and locking-out on telephone party lines is due to Homer Roberts and his associates.



To understand just how the principles illustrated in Figs. 186 and 187 are put into effect, it will be necessary to understand the latching relay shown diagrammatically in its two possible positions in Fig. 194, and in perspective in Fig. 195. Referring to Fig. 194, the left-hand cut of which shows the line relay in its normal position, it is seen that the framework of the device resembles that of an ordinary polarized ringer. Under the influence of current in one direction flowing through the left-hand coil, the armature of this device depresses the hard rubber stud 4, and the springs 1, 2, and 3 are forced downwardly until the spring 2 has passed under the latch carried on the spring 5. When the operating current through the coil 6 ceases, the pressure of the armature on the spring 1 is relieved, allowing this spring to resume its normal position and spring 3 to engage with spring 2. The spring 2 cannot rise, since it is held by the latch 5, and the condition shown in the right-hand cut of Fig. 194 exists. It will be seen that the spring 2 has in this operation carried out just the same function as the switch lever performed as described in connection with Figs. 186 and 187. An analysis of this action will show that the normal contact between the springs 1 and 2, which contact controls the circuit through the relay coil and the bell, is not broken until the coil 6 is de-energized, which means that the magnet is effective until it has accomplished its work. It is impossible, therefore, for this relay to cut itself out of circuit before it has caused the spring 2 to engage under the latch 5. If current of the proper direction were sent through the coil 7 of the relay, the opposite end of the armature would be pulled down and the hard rubber stud at the left-hand end of the armature would bear against the bent portion of the spring 5 in such manner as to cause the latch of this spring to release the spring 2 and thus allow the relay to assume its normal, or unlatched, position.

A good idea of the mechanical construction of this relay may be obtained from Fig. 195. The entire selecting function of the Roberts system is performed by this simple piece of apparatus at each station.



The diagram of Fig. 196 shows, in simplified form, a four-station line, the circuits being given more in detail than in the diagrams of Chapter XVI.

It will be noticed that the ringer and the relay coil 6 at the first station are bridged across the sides of the line leading to the central office. In like manner the bell and the relay magnets are bridged across the two limbs of the line leading into each succeeding station, but this bridge at each of the stations beyond Station A is ineffective because the line extension R{x} is open at the next station nearest the central office.



In order to ring Station A it is only necessary to send out ringing current from the central office. This current is in such direction as not to cause the operation of the relay, although it passes through the coil 6. If, on the other hand, it is desired to ring Station B, a preliminary impulse would be sent over the metallic circuit from the central office, which impulse would be of such direction as to operate the relay at Station A, but not to operate the bell at that station. The operation of the relay at Station A causes the spring 2 of this relay to engage the spring 3, thus extending the line on to the second station. After the spring 2 at Station A has been forced into contact with the spring 3, it is caught by the latch of the spring 5 and held mechanically. When the impulse from the central office ceases, the spring 1 resumes its normal position, thus breaking the bridge circuit through the bell at that station. It is apparent now that the action of coil 6 at Station A has made the relay powerless to perform any further action, and at the same time the line has been extended on to the second station. A second similar impulse from the central office will cause the relay at Station B to extend the line on to Station C, and at the same time break the circuit through the operating coil and the bell at Station B. In this way any station may be picked out by sending the proper number of impulses to operate the line relays of all the stations between the station desired and the central office, and having picked out a station it is only necessary to send out ringing current, which current is in such direction as to ring the bell but not to operate the relay magnet at that station.

In Fig. 197, a four-station line, such as is shown in Fig. 196, is illustrated, but the condition shown in this is that existing when two preliminary impulses have been sent over the line, which caused the line relays at Station A and Station B to be operated. The bell at Station C is, therefore, the only one susceptible to ringing current from the central office.



Since only one bell and one relay are in circuit at any one time, it is obvious that all of the current that passes over the line is effective in operating a single bell or relay only. There is no splitting up of the current among a large number of bells as in the bridging system of operating step-by-step devices, which method sometimes so greatly reduces the effective current for each bell that it is with great difficulty made to respond. All the energy available is applied directly to the piece of apparatus at the time it is being operated. This has a tendency toward greater surety of action, and the adjustment of the various pieces of apparatus may be made with less delicacy than is required where many pieces of apparatus, each having considerable work to do, must necessarily be operated in multiple.

The method of unlatching the relays has been briefly referred to. After a connection has been established with a station in the manner already described, the operator may clear the line when it is proper to do so by sending impulses of such a nature as to cause the line relays of the stations beyond the one chosen to operate, thus continuing the circuit to the end of the line. The operation of the line relay at the last station brings into circuit the coil 8, Figs. 196 and 197, of a grounding device. This is similar to the line relay, but it holds its operating spring in a normally latched position so as to maintain the two limbs of the line disconnected from the ground. The next impulse following over the metallic circuit passes through the coil 8 and causes the operation of this grounding device which, by becoming unlatched, grounds the limb L of the line through the coil 8. This temporary ground at the end of the line makes it possible to send an unlocking or restoring current from the central office over the limb L, which current passes through all of the unlocking coils 7, shown in Figs. 194, 196, and 197, thus causing the simultaneous unlocking of all of the line relays and the restoration of the line to its normal condition, as shown in Fig. 196.



As has been stated, the windings 7 on the line relays are the unlatching windings. In Figs. 196 and 197, for the purpose of simplicity, these windings are not shown connected, but as a matter of fact each of them is included in series in the continuous limb L of the line. This would introduce a highly objectionable feature from the standpoint of talking over the line were it not for the balancing coils 7^{1}, each wound on the same core as the corresponding winding 7, and each included in series in the limb R of the line, and in such direction as to be differential thereto with respect to currents passing in series over the two limbs of the line.

The windings 7 are the true unlocking windings, while the windings 7^{1} have no other function than to neutralize the inductive effects of these unlocking windings necessarily placed in series in the talking circuit. All of these windings are of low ohmic resistance, a construction which, as has previously been noted, brings about the desired effect without introducing any self-induction in the line, and without producing any appreciable effect upon the transmission. A study of Fig. 198 will make clear the connections of these unlocking and balancing windings at each station.

The statement of operation so far given discloses the general method of building up the line in sections in order to choose any party and of again breaking it up into sections when the conversation is finished. It has been stated that the same operation which selects the party wanted also serves to give that party the use of the line and to lock the others off. That this is true will be understood when it is stated that the ringer is of such construction that when operated to ring the subscriber wanted, it also operates to unlatch a set of springs similar to those shown in Fig. 194, this unlatching causing the proper connection of the subscriber's talking circuit across the limbs of the line, and also closing the local circuit through his transmitter. The very first motion of the bell armature performs this unlatching operation after which the bell behaves exactly as an ordinary polarized biased ringer.



The construction of this ringer is interesting and is shown in its two possible positions in Fig. 199. The group of springs carried on its frame is entirely independent of the movement of the armature during the ringing operation. With reversed currents, however, the armature is moved in the opposite direction from that necessary to ring the bells, and this causes the latching of the springs into their normal position. In order that this device may perform the double function of ringer and relay the tapper rod of the bell is hinged on the armature so as to partake of the movements of the armature in one direction only. This has been called by the inventor and engineers of the Roberts system a broken-back ringer, a name suggestive of the movable relation between the armature and the tapper rod. The construction of the ringer is of the same nature as that of the standard polarized ringer universally employed, but a hinge action between the armature and the tapper rod, of such nature as to make the tapper partake positively of the movements of the armature in one direction, but to remain perfectly quiescent when the armature moves in the other direction, is provided.



How this broken-back ringer controls the talking and the locking-out conditions may best be understood in connection with Fig. 200. The ringer springs are normally latched at all stations. Under these conditions the receiver is short-circuited by the engagement of springs 10 and 11, the receiver circuit is open between springs 10 and 12, and the local-battery circuit is open between springs 9 and 12. The subscribers whose ringers are latched are, therefore, locked out in more ways than one.

When the bell is rung, the first stroke it makes unlatches the springs, which assume the position shown in the right-hand cut of Fig. 199, and this, it will be seen from Fig. 200, establishes proper conditions for enabling the subscriber to transmit and to receive speech.

The hook switch breaks both transmitter and receiver circuits when down and in raising it establishes a momentary circuit between the ground and the limb L of the line, both upper and lower hook contacts engaging the hook lever simultaneously during the rising of the hook.

The mechanism at the central office by which selection of the proper station is made in a rapid manner is shown in Fig. 201. It has already been stated that the selection of the proper subscriber is brought about by the sending of a predetermined number of impulses from the central office, these impulses passing in one direction only and over the metallic circuit. After the proper party has been reached, the ringing current is put on in the reverse direction.



The operator establishes the number of impulses to be sent by placing the pointer opposite the number on the dial corresponding to the station wanted. The ratchet wheel is stepped around automatically by each impulse of current from an ordinary pole changer such as is employed in ringing biased bells. When the required number of impulses has been sent, a projection, carried on a group of springs, drops into a notch on the drum of the selector shaft, which operation instantly stops the selecting current impulses and at the same time throws on the ringing current which consists of impulses in the reverse direction. So rapidly does this device operate that it will readily follow the impulses of an ordinary pole changer, even when this is adjusted to its maximum rate of vibration.



Operation. Space will not permit a full discussion of the details of the central-office selective apparatus, but a general resume of the operation of the system may now be given, with the aid of Fig. 202, which shows a four-station line with the circuits of three of the stations somewhat simplified. In this figure Station A, Station B, and Station D are shown in their locked-out positions, A and B having been passed by the selection and ringing of Station C, while Station D is inoperative because it was not reached in the selection and the line is still broken at Station C. Station C, therefore, has possession of the line.

When the subscriber at Station C raised his receiver in order to call central, a "flash" contact was made as the hook moved up, which momentarily grounded the limb L of the line. (See Fig. 200.) This "flash" contact is produced by the arrangement of the hook which assures that the lower contact shall, by virtue of its flexibility, follow up the hook lever until the hook lever engages the upper contact, after which the lower contact breaks. This results in the momentary connection of both the upper and the lower contacts of the hook with the lever, and, therefore, the momentary grounding of the limb L of the line. This limb always being continuous serves, when this "flash" contact is made, to actuate the line signal at the central office.



Since, however, all parties on the line are normally locked out of talking circuits, some means must be provided whereby the operator may place the signaling party in talking connection and leave all the other instruments on the line in their normally locked-out position. In fact, the operator must be able automatically to pick out the station that signaled in, and operate the ringer to unlatch the springs controlling the talking circuit of that station. Accordingly the operator sends impulses on the line, from a grounded battery, which are in the direction to operate the line relays and to continue the line circuit to the station calling. When, after a sufficient number of impulses, this current reaches that station it finds a path to ground from the limb L. This path is made possible by the fact that the subscriber's receiver is off its hook at that station. In order to understand just how this ground connection is made, it must be remembered that each of the ringer magnets is energized with each selecting impulse, but in such a direction as not to ring the bells, it being understood that all of the ringer mechanisms are normally latched. When the selecting impulse for Station C arrives, it passes through the ringer and the selecting relay coils at that station and starts to operate the remainder of the ringers sufficiently to cause the spring 12 to engage the spring 13. This establishes the ground connection from the limb L of the line, the circuit being traced through limb L through the upper contact of the switch, thence through springs 12 and 13 to ground, and this, before the line relay has time to latch, operates the quick-acting relay at the central office, which acts to cut off further impulses, and thus automatically stops at the calling station. Ringing current in the opposite direction is then sent to line; this unlatches the ringer springs and places the calling subscriber in talking circuit. When the operator has communicated with the calling subscriber, and found, for example, that another party on another similar line is desired, she turns the dial pointer on the selector to the number corresponding to the called-for party's number on that line, and presses the signal key. Pressing this key causes impulses to "run down the line," selecting the proper party and ringing his bell in the manner already described. The connection between the two parties is then established, and no one else can in any possible way, except by permission of the operator, obtain access to the line.

It is obvious that some means must be provided for restoring the selecting relays to normal after a conversation is finished. By referring to Fig. 194 it will be seen that the upper end of the latch spring 5 is bent over in such a manner that when the armature is attracted by current flowing through the coil 7, the knob on the left-hand end of the armature on rising engages with the bent cam surface and forces back the latch, permitting spring 2 to return to its normal position.

To restore the line the operator sends out sufficient additional selective impulses to extend the circuit to the end of the line, and thus brings the grounder into circuit. The winding of the grounder is connected in such a manner that the next passing impulse throws off its latch, permitting the long spring to contact with the ground spring. The operator now sends a grounded impulse over the continuous limb L of the line which passes through the restoring coils 7 at all the stations and through the right-hand coil of the grounding device to ground. The selecting relays are, therefore, simultaneously restored to normal. The grounder is also energized and restored to its normal position by the same current.

If a party in calling finds that his own line is busy and he cannot get central, he may leave his receiver off its hook. When the party who is using the line hangs up his receiver the fact that another party desires a connection is automatically indicated to the operator, who then locks out the instrument of the party who has just finished conversation and passes his station by. When the operator again throws the key, the waiting subscriber is automatically selected in the same manner as was the first party. If there are no subscribers waiting for service, the stop relay at central will not operate until the grounder end of the line is unlatched, the selecting relays being then restored automatically to normal.

The circuits are so organized that at all times whether the line is busy or not, the movement up and down of the switch hook, at any sub-station, operates a signal before the operator. Such a movement, when made slowly and repeatedly, indicates to the operator that the subscriber has an emergency call and she may use her judgment as to taking the line away from the parties who are using it, and finding out what the emergency call is for. If the operator finds that the subscriber has misused this privilege of making the emergency call, she may restore the connection to the parties previously engaged in conversation.

One of the salient points of this Roberts system is that the operator always has control of the line. A subscriber is not able even to use his own battery till permitted to do so. A subscriber who leaves his receiver off its hook in order that he may be signaled by the operator when the line is free, causes no deterioration of the local battery because the battery circuit is held open by the switch contacts carried on the ringer. It cannot be denied, however, that this system is complicated, and that it has other faults. For instance, as described herein, both sides of the line must be looped into each subscriber's station, thus requiring four drop, or service, wires instead of two. It is possible to overcome this objection by placing the line relays on the pole in a suitably protected casing, in which case it is sufficient to run but two drop wires from the nearer line to station. There are undoubtedly other objections to this system, and yet with all its faults it is of great interest, and although radical in many respects, it teaches lessons of undoubted value.



CHAPTER XVIII

ELECTRICAL HAZARDS

All telephone systems are exposed to certain electrical hazards. When these hazards become actively operative as causes, harmful results ensue. The harmful results are of two kinds: those causing damage to property and those causing damage to persons. The damage to persons may be so serious as to result in death. Damage to property may destroy the usefulness of a piece of apparatus or of some portion of the wire plant. Or the property damage may initiate itself as a harm to apparatus or wiring and may result in greater and extending damage by starting a fire.

Electrical currents which endanger life and property may be furnished by natural or artificial causes. Natural electricity which does such damage usually displays itself as lightning. In rare cases, currents tending to flow over grounded lines because of extraordinary differences of potential between sections of the earth's surface have damaged apparatus in such lines, or only have been prevented from causing such damage by the operation of protective devices.

Telegraph and telephone systems have been threatened by natural electrical hazards since the beginning of the arts and by artificial electrical hazards since the development of electric light and power systems. At the present time, contrary to the general supposition, it is in the artificial, and not in the natural electrical hazards that the greater variety and degree of danger lies.

Of the ways in which artificial electricity may injure a telephone system, the entrance of current from an external electrical power system is a greater menace than an abnormal flow of current from a source belonging to the telephone system itself. Yet modern practice provides opportunities for a telephone system to inflict damage upon itself in that way. Telephone engineering designs need to provide means for protecting all parts of a system against damage, from external ("foreign") as well as internal ("domestic") hazards, and to cause this protection to be inclusive enough to protect persons against injury and property from damage by any form of overheating or electrolytic action.

A part of a telephone system for which there is even a remote possibility of contact with an external source of electrical power, whether natural or artificial, is said to be exposed to electrical hazard. The degree or character of possible contact or other interference often is referred to in relative terms of exposure. The same terms are used concerning inductive relations between circuits. The whole tendency of design, particularly of wire plants, is to arrange the circuits in such a way as to limit the exposure as greatly as possible, the intent being to produce a condition in which all parts of the system will be unexposed to hazards.

Methods of design are not yet sufficiently advanced for any plant to be formed of circuits wholly unexposed, so that protective means are required to safeguard apparatus and circuits in case the hazard, however remote, becomes operative.

Lightning discharges between the clouds and earth frequently charge open wires to potentials sufficiently high to damage apparatus; and less frequently, to destroy the wires of the lines themselves. Lightning discharges between clouds frequently induce charges in lines sufficient to damage apparatus connected with the lines. Heavy rushes of current in lines, from lightning causes, occasionally induce damaging currents in adjacent lines not sufficiently exposed to the original cause to have been injured without this induction. The lightning hazard is least where the most lines are exposed. In a small city with all of the lines formed of exposed wires and all of them used as grounded circuits, a single lightning discharge may damage many switchboard signals and telephone ringers if there be but 100 or 200 lines, while the damage might have been nothing had there been 800 to 1,000 lines in the same area.