CHAPTER XIII

CURRENT SUPPLY TO TRANSMITTERS

The methods by which current is supplied to the transmitter of a telephone for energizing it, may be classified under two divisions: first, those where the battery or other source of current is located at the station with the transmitter which it supplies; and second, those where the battery or other source of current is located at a distant point from the transmitter, the battery in such cases serving as a common source of current for the supply of transmitters at a number of stations.

The advantages of putting the transmitter and the battery which supplies it with current in a local circuit with the primary of an induction coil, and placing the secondary of the induction coil in the line, have already been pointed out but may be briefly summarized as follows: When the transmitter is placed directly in the _line circuit_ and the line is of considerable length, the current which passes through the transmitter is necessarily rather small unless a battery of high potential is used; and, furthermore, the total change in resistance which the transmitter is capable of producing is but a small proportion of the total resistance of the line, and, therefore, the current changes produced by the transmitter are relatively small. On the other hand, when the transmitter is placed in a _local circuit_ with the battery, this circuit may be of small resistance and the current relatively large, even though supplied by a low-voltage battery; so that the transmitter is capable of producing relatively large changes in a relatively large current.

To draw a comparison between these two general classes of transmitter current supply, a number of cases will be considered in connection with the following figures, in each of which two stations connected by a telephone line are shown. Brief reference to the local battery method of supplying current will be made in order to make this chapter contain, as far as possible, all of the commonly used methods of current supply to transmitters.

Local Battery. In Fig. 125 two stations are shown connected by a grounded line wire. The transmitter of each station is included in a low-resistance primary circuit including a battery and the primary winding of an induction coil, the relation between the primary circuits and the line circuits being established by the inductive action between the primary and the secondary windings of induction coils, the secondary in each case being in the line circuits with the receivers.

Fig. 126 shows exactly the same arrangement but with a metallic circuit rather than a grounded circuit. The student should become accustomed to the replacing of one of the line wires of a metallic circuit by the earth, and to the method, employed in Figs. 125 and 126, of indicating a grounded circuit as distinguished from a metallic circuit.

In Fig. 127 is shown a slight modification of the circuit shown in Fig. 126, which consists of connecting one end of the primary winding to one end of the secondary winding of the induction coil, thus linking together the primary circuit and the line circuit, a portion of each of these circuits being common to a short piece of the local wiring. There is no difference whatever in the action of the circuits shown in Figs. 126 and 127, the latter being shown merely for the purpose of bringing out this fact. It is very common, particularly in local-battery circuits, to connect one end of the primary and the secondary windings, as by doing so it is often possible to save a contact point in the hook switch and also to simplify the wiring.

The advantages to be gained by employing a local battery at each subscriber's station associated with the transmitter in the primary circuit of an induction coil are attended by certain disadvantages from a commercial standpoint. The primary battery is not an economical way to generate electric energy. In all its commercial forms it involves the consumption of zinc and zinc is an expensive fuel. The actual amount of current in watts required by a telephone is small, however, and this disadvantage due to the inexpensive method of generating current would not in itself be of great importance. A more serious objection to the use of local batteries at subscribers' stations appears when the subject is considered from the standpoint of maintenance. Batteries, whether of the so-called "dry" or "wet" type, gradually deteriorate, even when not used, and in cases where the telephone is used many times a day the deterioration is comparatively rapid. This makes necessary the occasional renewals of the batteries with the attendant expense for new batteries or new material, and of labor and transportation in visiting the station. The labor item becomes more serious when the stations are scattered in a sparsely settled community, in which case the visiting of the stations, even for the performance of a task that would require but a few minutes' time, may consume some hours on the part of the employes in getting there and back.

Common Battery. _Advantages._ It would be more economical if all of the current for the subscribers' transmitters could be supplied from a single comparatively efficient generating source instead of from a multitude of inefficient small sources scattered throughout the community served by the exchange. The advantage of such centralization lies not only in more economic generating means, but also in having the common source of current located at one place, where it may be cared for with a minimum amount of expense. Such considerations have resulted in the so-called "common-battery system," wherein the current for all the subscribers' transmitters is furnished from a source located at the central office.

Where such a method of supplying current is practiced, the result has also been, in nearly all cases, the doing away with the subscriber's magneto generators, relying on the central-office source of current to furnish the energy for enabling the subscriber to signal the operator. Such systems, therefore, concentrate all of the sources of energy at the central office and for that reason they are frequently referred to as central-energy systems.

NOTE. In this chapter the central-energy or common-battery system will be considered only in so far as the supply of current for energizing the subscribers' transmitters is concerned, the discussion of the action of signaling being reserved for subsequent chapters.

_Series Battery._ If but a single pair of lines had to be considered, the arrangement shown in Fig. 128 might be employed. In this the battery is located at the central office and placed in series with the two grounded lines leading from the central office to the two subscribers' stations. The voltage of this battery is made sufficient to furnish the required current over the resistance of the entire line circuit with its included instruments. Obviously, changes in resistance in the transmitter at Station A will affect the flow of current in the entire line and the fluctuations resulting from the vibration of the transmitter diaphragm will, therefore, reproduce these sounds in the receiver at Station B, as well as in that at Station A.

An exactly similar arrangement applied to a metallic circuit is shown in Fig. 129. In thus placing the battery in series in the circuit between the two stations, as shown in Figs. 128 and 129, it is obvious that the transmitter at each station is compelled to vary the resistance of the entire circuit comprising the two lines in series, in order to affect the receiver at distant stations. This is in effect making the transmitter circuit twice as long as is necessary, as will be shown in the subsequent systems considered. Furthermore, the placing of the battery in series in the circuit of the two combined lines does not lend itself readily to the supply of current from a common source to more than a single pair of lines.

_Series Substation Circuit._ The arrangement at the substations--consisting in placing the transmitter and the receiver in series in the line circuit, as shown in Figs. 128 and 129--is the simplest possible one, and has been used to a considerable extent, but it has been subject to the serious objection, where receivers having permanent magnets were used, of making it necessary to so connect the receiver in the line circuit that the steady current from the battery would not set up a magnetization in the cores of the receiver in such a direction as to neutralize or oppose the magnetization of the permanent magnets. As long as the current flowed through the receiver coils in such a direction as to supplement the magnetization of the permanent magnets, no harm was usually done, but when the current flowed through the receiver coils in such a way as to neutralize or oppose the magnetizing force of the permanent magnets, the action of the receiver was greatly interfered with. As a result, it was necessary to always connect the receivers in the line circuit in a certain way, and this operation was called _poling_.

In order to obviate the necessity for poling and also to bring about other desirable features, it has been, until recently, almost universal practice to so arrange the receiver that it would be in the circuit of the voice currents passing over the line, but would not be traversed by direct currents, this condition being brought about by various arrangements of condensers, impedance coils, or induction coils, as will be shown later. During the year 1909, however, the adoption by several concerns of the so-called "direct-current" receiver has made it necessary for the direct current to flow through the receiver coils in order to give the proper magnetization to the receiver cores, and this has brought about a return to the very simple form of substation circuit, which includes the receiver and the transmitter directly in the circuit of the line. This illustrates well an occurrence that is frequently observed by those who have opportunity to watch closely the development of an art. At one time the conditions will be such as to call for complicated arrangements, and for years the aim of inventors will be to perfect these arrangements; then, after they are perfected, adopted, and standardized, a new idea, or a slight alteration in the practice in some other respect, will demand a return to the first principles and wipe out the necessity for the things that have been so arduously striven for.

_Bridging Battery with Repeating Coil._ As pointed out, the placing of the battery in series in the line circuit in the central office is not desirable, and, so far as we are aware, has never been extensively used. The universal practice, therefore, is to place it in a bridge path across the line circuit, and a number of arrangements employing this basic idea are in wide use. In Fig. 130 is shown the standard arrangement of the Western Electric Company, employed by practically all the Bell operating companies. In this the battery at the central office is connected in the middle of the two sides of a repeating coil so that the current from the battery is fed out to the two connected lines in multiple.

Referring to the middle portion of this figure showing the central-office apparatus, _1_ and _2_ may be considered as the two halves of one side of a repeating coil divided so that the battery may be cut into their circuit. Likewise, _3_ and _4_ may be considered as the two halves of the other side of the repeating coil similarly divided for the same purpose. The windings of this repeating coil are ordinarily alike; that is, _1_ and _2_ combined have the same resistance, number of turns, and impedance as _3_ and _4_ combined. The two sides of this coil are alternately used as primary and secondary, _1_ and _2_ forming the primary when Station A is talking, and _3_ and _4_, the secondary; and _vice versâ_ when Station B is talking.

As will be seen, the current flowing from the positive pole of the battery will divide and flow through the windings _2_ and _4_; thence over the upper limb of each line, through the transmitter at each station, and back over the lower limbs of the line, through the windings _1_ and _3_, where the two paths reunite and pass to the negative pole of the battery. It is evident that when neither transmitter is being used the current flowing through both lines will be a steady current and that, therefore, neither line will have an inductive effect on the other. When, however, the transmitter at Station A is used the variations in the resistance caused by it will cause undulations in the current. These undulations, passing through the windings _1_ and _2_ of the repeating coil, will cause, by electromagnetic induction, alternating currents to flow in the windings _3_ and _4_, and these alternating currents will be superimposed on the steady currents flowing in that line and will affect the receiver at Station B, as will be pointed out. The reverse conditions exist when Station B is talking.

_Bell Substation Arrangement._ The substation circuits at the stations in Fig. 130 are illustrative of one of the commonly employed methods of preventing the steady current from the battery from flowing through the receiver coil. This particular arrangement is that employed by the common-battery instruments of the various Bell companies. Considering the action at Station B, it is evident that the steady current will pass through the transmitter and through the secondary winding of the induction coil, and that as long as this current is steady no current will flow through the telephone receiver. The receiver, transmitter, and primary winding of the induction coil are, however, included in a local circuit with the condenser. The presence of the condenser precludes the possibility of direct current flowing in this path. Considering Station A as a receiving station, it is evident that the voice currents coming to the station over the line will pass through the secondary winding and will induce alternating currents in the primary winding which will circulate through the local circuit containing the receiver and the condenser, and thus actuate the receiver. The considerations are not so simple when the station is being treated as a transmitting station. Under this condition the steady current passes through the transmitter in an obvious manner. It is clear that if the local circuit containing the receiver did not exist, the circuit would be operative as a transmitting circuit because the transmitter would produce fluctuations in the steady current flowing in the line and thus be able to affect the distant station. The transmitter, therefore, has a direct action on the currents flowing in the line by the variation in resistance which it produces in the line circuit. There is, however, a subsidiary action in this circuit. Obviously, there is a drop of potential across the transmitter terminals due to the flow of steady current. This means that the upper terminal of the condenser will be charged to the same potential as the upper terminal of the transmitter, while the lower terminal of the condenser will be of the same potential as the lower terminal of the transmitter. When, now, the transmitter varies its resistance, a variation in the potential across its terminals will occur; and as a result, a variation in potential across the terminals of the condenser will occur, and this means that alternating currents will flow through the primary winding of the induction coil. The transmitter, therefore, by its action, causes alternating currents to flow through the primary of this induction coil and it causes, by direct action on the circuit of the line, fluctuations in the steady current flowing in the line. The alternating currents flowing in the primary of the coil induce currents in the secondary of the coil which supplement and augment the fluctuations produced by the direct action of the transmitter. This circuit may be looked at, therefore, in the light of combining the direct action which the transmitter produces in the current in the line with the action which the transmitter produces in the local circuit containing the primary of the induction coil, this action being repeated in the line circuit through the secondary of the induction coil.

The receiver in this circuit is placed in the local circuit, and is thus not traversed by the steady currents flowing in the line. There is thus no necessity for poling it. This circuit is very efficient, but is subject to the objection of producing a heavy side tone in the receiver of the transmitting station. By "side tone" is meant the noises which are produced in the receiver at a station by virtue of the action of the transmitter at that station. Side tone is objectionable for several reasons: first, it is sometimes annoying to the subscriber; second, and of more importance, the subscriber who is talking, hearing a very loud noise in his own receiver, unconsciously assumes that he is talking too loud and, therefore, lowers his voice, sometimes to such an extent that it will not properly reach the distant station.

_Bridging Battery with Impedance Coils._ The method of feeding current to the line from the common battery, shown in Fig. 130, is called the "split repeating-coil" method. As distinguished from this is the impedance-coil method which is shown in Fig. 131. In this the battery is bridged across the circuit of the combined lines in series with two impedance coils, _1_ and _2_, one on each side of the battery. The steady currents from the battery find ready path through these impedance coils which are of comparatively low ohmic resistance, and the current divides and passes in multiple over the circuits of the two lines. Voice currents, however, originating at either one of the stations, will not pass through the shunt across the line at the central office on account of the high impedance offered by these coils, and as a result they are compelled to pass on to the distant station and affect the receiver there, as desired.

This impedance-coil method seems to present the advantage of greater simplicity over the repeating-coil method shown in Fig. 130, and so far as talking efficiency is concerned, there is little to choose between the two. The repeating-coil method, however, has the advantage over this impedance-coil method, because by it the two lines are practically divided except by the inductive connection between the two windings, and as a result an unbalanced condition of one of the connected lines is not as likely to produce an unbalanced condition in the other as where the two lines are connected straight through, as with the impedance-coil method. The substation arrangement of Fig. 131 is the same as that of Fig. 130.

_Double Battery with Impedance Coils._ A modification of the impedance-coil method is used in all of the central-office work of the Kellogg Switchboard and Supply Company. This employs a combination of impedance coils and condensers, and in effect isolates the lines conductively from each other as completely as the repeating-coil method. It is characteristic of all the Kellogg common-battery systems that they employ two batteries instead of one, one of these being connected in all cases with the calling line of a pair of connected lines and the other in all cases with the called line. As shown in Fig. 132, the left-hand battery is connected with the line leading to Station A through the impedance coils _1_ and _2_. Likewise, the right-hand battery is connected to the line of Station B through the impedance coils _3_ and _4_. These four impedance coils are wound on separate cores and do not have any inductive relation whatsoever with each other. Condensers _5_ and _6_ are employed to completely isolate the lines conductively. Current from the left-hand battery, therefore, passes only to Station A, and current from the right-hand battery to Station B. Whenever the transmitter at Station A is actuated the undulations of current which it produces in the line cause a varying difference of potential across the outside terminals of the two impedance coils _1_ and _2_. This means that the two left-hand terminals of condensers _5_ and _6_ are subjected to a varying difference of potential and these, of course, by electrostatic induction, cause the right-hand terminals of these condensers to be subject to a correspondingly varying difference of potential. From this it follows that alternating currents will be impressed upon the right-hand line and these will affect the receiver at Station B.

A rough way of expressing the action of this circuit is to consider it in the same light as that of the impedance-coil circuit shown in Fig. 131, and to consider that the voice currents originating in one line are prevented from passing through the bridge paths at the central office on account of the impedance, and are, therefore, forced to continue on the line, being allowed to pass readily by the condensers in series between the two lines.

_Kellogg Substation Arrangement._ An interesting form of substation circuit which is employed by the Kellogg Company in all of its common-battery telephones is shown in Fig. 132. In passing, it may be well to state that almost any of the substation circuits shown in this chapter are capable of working with any of the central-office circuits. The different ones are shown for the purpose of giving a knowledge of the various substation circuits that are employed, and, as far as possible, to associate them with the particular central-office arrangements with which they are commonly used.

In this Kellogg substation arrangement the line circuit passes first through the transmitter and then divides, one branch passing through an impedance coil _7_ and the other through the receiver and the condenser _8_, in series. The steady current from the central-office battery finds ready path through the transmitter and the impedance coil, but is prevented from passing through the receiver by the barrier set up by the condenser _8_. Voice currents, however, coming over the line to the station, find ready path through the receiver and the condenser but are barred from passing through the impedance coil by virtue of its high impedance.

In considering the action of the station as a transmitting station, the variations set up by the transmitter pass through the condenser and the receiver at the same station, while the steady current which supplies the transmitter passes through the impedance coil. Impedance coils used for this purpose are made of low ohmic resistance but of a comparatively great number of turns, and, therefore, present a good path for steady currents and a difficult path for voice currents. This divided circuit arrangement employed by the Kellogg Company is one of the very simple ways of eliminating direct currents from the receiver path, at the same time allowing the free passage of voice currents.

_Dean Substation Arrangement._ In marked contrast to the scheme for keeping steady current out of the receiver circuit employed by the Kellogg Company, is that shown in Fig. 133, which has been largely used by the Dean Electric Company, of Elyria, Ohio. The central-office arrangement in this case is that using the split repeating coil, which needs no further description. The substation arrangement, however, is unique and is a beautiful example of what can be done in the way of preventing a flow of current through a path without in any way insulating that path or placing any barrier in the way of the current. It is an example of the prevention of the direct flow of current through the receiver by so arranging the circuits that there will always be an equal potential on each side of it, and, therefore, no tendency for current to flow through it.

In this substation arrangement four coils of wire--_1_, _2_, _3_, and _4_--are so arranged as to be connected in the circuit of the line, two in series and two in multiple. The current flowing from the battery at the central office, after passing through the transmitter, divides between the two paths containing, respectively, the coils _1_ and _3_ and the coils _2_ and _4_. The receiver is connected between the junction of the coils _2_ and _4_ and that of _1_ and _3_. The resistances of the coils are so chosen that the drop of potential through the coil _2_ will be equal to that through the coil _1_, and likewise that through the coil _4_ will be equal to that through the coil _3_. As a result, the receiver will be connected between two points of equal potential, and no direct current will flow through it. How, then, do voice currents find their way through the receiver, as they evidently must, if the circuit is to fulfill any useful function? The coils _2_ and _3_ are made to have high impedance, while _1_ and _4_ are so wound as to be non-inductive and, therefore, offer no impedance save that of their ohmic resistance. What is true, therefore, of direct currents does not hold for voice currents, and as a result, the voice currents, instead of taking the divided path which the direct currents pursued, are debarred from the coils _2_ and _3_ by their high impedance and thus pass through the non-inductive coil _1_, the receiver, and the non-inductive coil _4_.

This circuit employs a Wheatstone-bridge arrangement, adjusted to a state of balance with respect to direct currents, such currents being excluded from the receiver, not because the receiver circuit is in any sense opaque to such direct currents, but because there is no difference of potential between the terminals of the receiver circuit, and, therefore, no tendency for current to flow through the receiver. In order that fluctuating currents may not, for the same reason, be caused to pass by, rather than through, the receiver circuit, the diametrically-opposed arms of the Wheatstone bridge are made to possess, in large degree, self-induction, thereby giving these two arms a high impedance to fluctuating currents. The conditions which exist for direct currents do not, therefore, exist for fluctuating currents, and it is this distinction which allows alternating currents to pass through the receiver and at the same time excludes direct currents therefrom.

In practice, the coils _1_, _2_, _3_, and _4_ of the Dean substation circuit are wound on the same core, but coils _1_ and _4_--the non-inductive ones--are wound by doubling the wire back on itself so as to neutralize their self-induction.

_Stromberg-Carlson._ Another modification of the central-office arrangement and also of the subscribers' station circuits, is shown in Fig. 134, this being a simplified representation of the circuits commonly employed by the Stromberg-Carlson Telephone Manufacturing Company. The battery feed at the central office differs only from that shown in Fig. 132, in that a single battery rather than two batteries is used, the current being supplied to one of the lines through the impedance coils _1_ and _2_, and to the other line through the impedance coils _3_ and _4_; condensers _5_ and _6_ serve conductively to isolate the two lines. At the subscriber's station the line circuit passes through the secondary of an induction coil and the transmitter. The receiver is kept entirely in a local circuit so that there is no tendency for direct current to flow through it, but it is receptive to voice currents through the electromagnetic induction between the primary and the secondary of the induction coil.

_North._ Another arrangement of central-office battery feed is employed by the North Electric Company, and is shown in Fig. 135. In this two batteries are used which supply current respectively to the two connected lines, condensers being employed to conductively isolate the lines. This differs from the Kellogg arrangement shown in Fig. 132 in that the two coils _1_ and _2_ are wound on the same core, while the coils _3_ and _4_ are wound together upon another core. In this case, in order that the inductive action of one of the coils may not neutralize that of the other coil on the same core, the two coils are wound in such relative direction that their magnetizing influence will always be cumulative rather than differential.

The central-office arrangements discussed in Figs. 130 to 135, inclusive, are those which are in principal use in commercial practice in common-battery exchanges.

_Current Supply over Limbs of Line in Parallel._ As indicating further interesting possibilities in the method of supplying current from a common source to a number of substations, several other systems will be briefly referred to as being of interest, although these have not gone into wide commercial use. The system shown in Fig. 136 is one proposed by Dean in the early days of common-battery working, and this arrangement was put into actual service and gave satisfactory results, but was afterwards supplanted by the Bell equipment operating under the system shown in Fig. 130, which became standardized by that company. In this the current from the common battery at the central office is not fed over the two line wires in series, but in multiple, using a ground return from the subscriber's station to the central office. Across the metallic circuit formed by two connected lines there is bridged, at the central office, an impedance coil _1_, and between the center point of this impedance coil and the ground is connected the common battery. At the subscriber's station is placed an impedance coil _2_, also bridged across the two limbs of the line, and between the center point of this impedance coil and the ground is connected the transmitter, which is shunted by the primary winding of an induction coil. Connected between the two limbs of the line at the substation there is also the receiver and the secondary of an induction coil in series.

The action of this circuit at first seems a little complex, but if taken step by step may readily be understood. The transmitter supply circuit may be traced from the central-office battery through the two halves of the impedance coil _1_ in multiple; thence over the two limbs of the line in multiple to Station A, for instance; thence in multiple through the two halves of impedance coil _2_, to the center point of that coil; thence through the two paths offered respectively by the primary of the induction coil and by the transmitter; then to ground and back to the other pole of the central-office battery. By this circuit the transmitter at the substation is supplied with current.

Variations in the resistance of the transmitter when in action, cause complementary variations in the supply current flowing through the primary of the induction coil. These variations induce similar alternating currents in the secondary of this coil, which is in series in the line circuit. The currents, so induced in this secondary, flow in series through one side of the line to the distant station; thence through the secondary and the receiver at that station to the other side of the line and back through that side of the line to the receiver. These currents are not permitted to pass through the bridged paths across the metallic circuit that are offered by the impedance coils _1_ and _2_, because they are voice currents and are, therefore, debarred from these paths by virtue of the impedance.

An objection to this form of current supply and to other similar forms, wherein the transmitter current is fed over the two sides of the line in multiple with a ground return, is that the ground-return circuit formed by the two sides of the line in multiple is subject to inductive disturbances from other lines in the same way as an ordinary grounded line is subject to inductive disturbance. The current-supply circuit is thus subject to external disturbances and such disturbances find their way into the metallic circuit and, therefore, through the instruments by means of the electromagnetic induction between the primary and the secondary coils at the substations.

Another interesting method of current supply from a central-office battery is shown in Fig. 137. This, like the circuit just considered, feeds the energy to the subscriber's station over the two sides of the line in multiple with a ground return. In this case, however, a local circuit is provided at the substation, in which is placed a storage battery _1_ and the primary _2_ of an induction coil, together with the transmitter. The idea in this is that the current supply from the central office will pass through the storage battery and charge it. Upon the use of the transmitter, this storage battery acts to supply current to the local circuit containing the transmitter and the primary coil _2_ in exactly the same manner as in a local battery system. The fluctuating current so produced by the action of the transmitter in this local circuit acts on the secondary winding _3_ of the induction coil, and produces therein alternating currents which pass to the central office and are in turn repeated to the distant station.

_Supply Many Lines from Common Source._ We come now to the consideration of the arrangement by which a single battery may be made to supply current at the central office to a large number of pairs of connected lines simultaneously. Up to this point in this discussion it has been shown only how each battery served a single pair of connected lines and no others.

Repeating Coil:--In Fig. 138 is shown how a single battery supplies current simultaneously to four different pairs of lines, the lines of each pair being connected for conversation. It is seen that the pairs of lines shown in this figure are arranged in each case in accordance with the system shown in Fig. 130. Let us inquire why it is that, although all of these four pairs of lines are connected with a common source of energy and are, therefore, all conductively joined, the stations will be able to communicate in pairs without interference between the pairs. In other words, why is it that voice currents originating at Station A will pass only to the receiver at Station B and not to the receivers at Station C or Station H, for instance? The reason is that separate supply conductors lead from the points such as _1_ and _2_ at the junctions of the repeating-coil windings on each pair of circuits to the battery terminals, and the resistance and impedance of the battery itself and of the common leads to it are so small that although the feeble voice currents originating in the pair of lines connecting Station A and Station B pass through the battery, they are not able to alter the potential of the battery in any appreciable degree. As a result, therefore, the supply wires leading from the common-battery terminals to the points _7_ and _8_, for instance, cannot be subjected to any variations in potential by virtue of currents flowing through the battery from the points _1_ and _2_ of the lines joining Station A and Station B.

Retardation Coil--Single Battery:--In Fig. 139 is shown in similar manner the current supply from a single battery to four different pairs of lines, the battery being associated with the lines by the combined impedance coil and condenser method, which was specifically dealt with in connection with Fig. 133. The reasons why there will be no interference between the conversations carried on in the various pairs of connected lines in this case are the same as those just considered in connection with the system shown in Fig. 138. The impedance coils in this case serve to keep the telephone currents confined to their respective pairs of lines in which they originate, and this same consideration applies to the system of Fig. 138, for each of the separate repeating-coil windings of Fig. 138 is in itself an impedance coil with respect to such currents as might leak away from one pair of lines on to another.

Retardation Coil--Double Battery:--The arrangement of feeding a number of pairs of lines according to the Kellogg two-battery system is indicated in Fig. 140, which needs no further explanation in view of the description of the preceding figures. It is interesting to note in this case that the left-hand battery serves only the left-hand lines and the right-hand battery only the right-hand lines. As this is worked out in practice, the left-hand battery is always connected to those lines which originate a call and the right-hand battery always to those lines that are called for. The energy supplied to a calling line is always, therefore, from a different source than that which supplies a called line.

_Current Supply from Distant Point._ Sometimes it is convenient to supply current to a group of lines centering at a certain point from a source of current located at a distant point. This is often the case in the so-called private branch exchange, where a given business house or other institution is provided with its own switchboard for interconnecting the lines leading to the various telephones of that concern or institution among themselves, and also for connecting them with lines leading to the city exchange. It is not always easy or convenient to maintain at such private switchboards a separate battery for supplying the current needed by the local exchange.

In such cases the arrangement shown in Fig. 141 is sometimes employed. This shows two pairs of lines connected by the impedance-coil system with common terminals _1_ and _2_, between which ordinarily the common battery would be connected. Instead of putting a battery between these terminals, however, at the local exchange, a condenser of large capacity is connected between them and from these terminals circuit wires _3_ and _4_ are led to a battery of suitable voltage at a distant central office. The condenser in this case is used to afford a short-circuit path for the voice currents that leak from one side of one pair of lines to the other, through the impedance coils bridged across the line. In this way the effect of the necessarily high resistance in the common leads _3_ and _4_, leading to the storage battery, is overcome and the tendency to cross-talk between the various pairs of connected lines is eliminated. Frequently, instead of employing this arrangement, a storage battery of small capacity will be connected between the terminals _1_ and _2_, instead of the condenser, and these will be charged over the wires _3_ and _4_ from a source of current at a distant point.

A consideration of the various methods of supplying current from a common source to a number of lines will show that it is essential that the resistance of the battery itself be very low. It is also necessary that the resistance and the impedance of the common leads from the battery to the point of distribution to the various pairs of lines be very low, in order that the voice currents which flow through them, by virtue of the conversations going on in the different pairs of lines, shall not produce any appreciable alteration in the difference of potential between the battery terminals.