US1107137A - Telephone-exchange system. - Google Patents

Description

E. E. CLEMENT. TELEPHONE EXCHANGE SYSTEM. APPLICATION FILED AUG. 31, 1905.

1,107,137, Patented Aug. 11,1914.

8 SHEETS-SHEET 1.

q/Vihwoow E. E. CLEMENT.

TELEPHONE EXC Patented Aug. 11,1914.

6 SHEETS-SHEBT 2.

E. E. CLEMENT.

TELEPHONE EXCHANGE SYSTEM.

APPLICATION I ILED AUG. 31, 1905.

Patented Aug. 11, 1914.

we masses E. E. CLEMENT.

TELEPHONE EXCHANGE SYSTEM.

urmoumn FILED we. 31,. 1905.

alvanic Patented A113. 11

a snnnrmznnnr 4.

Q/Vdncooe E. B. CLEMENT.

TELEPHONE EXCHANGE SYSTEM,

APPLICATION FILED AUG. 31, 1905.

Patented Aug. 11, 1914.

luten kn 6 SHEETS-SHEHT 5 W; hwooee E. E. CLEMENT; TELEPHONE EXCHANGE SYSTEM. APPLICATION FILED AUG. 31,1905.

Patented Aug. 11,1914;

UNITED STATES PATENT OFFICE.-

ED\VARD E. CLEMENT, OF WASHINGTON, DISTRICT OF COLUMBIA, ASSIGNOB, BY

1 MESNE ASSIGNMENTS, TO FREDERICK G. STEVENS, OF ATTICA, NEV] YORK.

TELEPEONE-EXCHAIIGE SYSTEM.

Specificationof Letters fatent.

Patented Aug. 11, 1914.

Application filed August 31, 1905. Serial l lo. 276,634.

Ton?! ir/wni it mag concern:

Be it known that I, Eowann E. CLEMENT, a citizen of the United States, residing at ll'ashingtoin'in the District of Columbia, have invented certain new and useful Improvements in Telephone-Exchange 'Sys- 7 tenis, of which the following is a specification, reference being had therein to the accompanyin'g drawing.

' My inrention relates to telephone systems,

and particularly to exchange systems.

i 't' hasfor its object the improvement of such systems in general, in point of first cost, in point of cost of operation and maintenance, and in point of etliciency. I

X subsidiary object, though an important one, is the provision of means whereby existing plants may be reorganized without substantial change 1n present circuits or methods, so as to largely increase their ca-- p'acity and efiiciency.

I attain my objects by combining certain features of the so-called manual systems with certain features of the so-called auto matic, in a novel manner, the product being what I shall denominate for convenience a semi-automatic system. 7

K oping in mind the necessity for slightchangeespe'ciany at the subscribers stations, in changing a plant over with economy, I have designed my circuits so that an ordinary twouvire multiple line circuit may be used, no ground being required at the subhowever, I have added a break key, and in.

practice will provide a simple form of automatic device to produce the same results. At the central ofiice the subscribers lines terminate-in'jacks and each has a line sig na'l, asusualin manual installations.

Instead of the operator who answers a call'having a pair of plugs, cords, ringing and listening keys, etc., she has a series of single plugs each associated with a supervisory lamp signal and each connected to a trunk-"line which leads to a first selector switch of the standard automatic type.

. switches.

others. The ordinary automatic systems reqmre a certain number of impulses to be sent over one side of the line, then an impulse or impulses over the other side, then more impulses over the first side of line, and another over the second, and so on. In clearing out,.both sides of the line get a ground so that simultaneous impulses are produced. It is manifestly necessaryto have a somewhat complicated sending apparatus even in a full automatic system, to

produce the proper efi'ects at the proper,

-times,'and in practice this is so.

In automatic and semi-automatic systems, as in manual installations, party line service is not only desirable but necessary in order to meet. present day conditions and insure proper returns in the face of competition. My present invention not only includes such an arrangement of the best type, but makes it absolutely individual and selective. The party line features, moreover, may be added to any existing automatic installationwithout great expense and without any change in existing parts and connections.

I solve the problems ofcombination and selection by the use of time-element apparatus. In other words, I introduce as a substitute for varying circuits in the linecall a varying time factor, giving signal impulses different lengths as they are required to. operate the working switches or the side In its simplest form this arrangement consists in merely breaking the metallic. circuit and immediately closing itagain a determinate number of times, depending upon the number desired, then leaving it closed for a longer time than that of any break before making further breaks. The simplicity of the apparatus required fortransmitting a signal of this kind is apparent, being as I have stated either a sim- Heretoforeithas been impossible to combine ple break-key or the switch hook itself.

In order to translate the makes and breaks of the metalhc line clrcuit into current 1mpulses on one or the other side or the automatic circuit, I employ in conjunction with my cord circuit a pair of very sluggish relays, controlled by a simple double contact relay connected to that part of the cord which is brought into connection with the line in plugging in to answer a call. When the plug is'inserted one of these sluggish relays pulls up, when the first break is produced at the subscribers station, the other sluggish relay pulls up, and as the breaks rapidly. succeed each other both these relays remain up, that is, their armatures do not have .time to retract between the successive impulses which they receive 'from the rapidly vibrating line-controlled relay. At the end of the series of-breaks one sluggish re-- layremains energized, and the other one is given time to retract, owing to the relatively long interval I have mentioned. Thus, I have established different. conditions for my sluggish relays corresponding to the time differences in'the line transmission, and I cause the armatures of the two sluggish relays to control connections to the two sepa-.

rate sides of the automatic trunk sons to send over one or the other as desired. A ;third condition is possible when the line circuit remains broken after a conversation,

"when the calling subscriber has replaced his receiver upon its hook. The relative conditions of the sluggish relays are then reiversed, that is, the one which was first pulled up is now retracted, and the other is pulled up. In 'this" condition I cause their armatures to ground or send simultaneous impulses over both sides ofthe automatic trunk, .whereby the usual release magnets are operated.

I-solve'the problem ofpartly line selective service by making vuse of the same princi le of difi'erent time constants in myrelays. I' employ a ringing relay so sluggish in its action that the character of the current supplied to it, or of the circuit over which the current is supplied, may be changed during the preliminary operation of the ringing relay so that in finishing its operation it will apply the current so selected, or current over the circuit selected, to the. line which is to be called. I cause this ringing relay to pull up on the first impulse of a series whose number is determined by the calling subscriber and corresponds to what I' may callthe ringing number, or ring ing letter, of the called subscriber. The breaks in the line or trunk are caused to select the particular current required, and after once'pulling up I proportion the parts of the rin' 'ng relay so that-it will not again ,fall back til the last break has been sent and the line remains closed for a definite time. Then, in fallin back, and between the position of attraction and that of complete retraction, I cause the ringing relay to conncct the generator to line. In other words, the relay acts as a ringing key, and the accompanying selector acts like the selective keys in an ordinary manual set. lVhen the ringing relay armature is down all the way the selective device is restored, the stepper disk or arm being put back to zero or normal position.

In the system hereinafter described as embodying my invention I have incorporated what I now consider the best and most advantageous features from a commercial standpoint. The selection of the ringer desired is by means of a specific frequency of ringing current, according to the so-called harmonic idea. I do not limit myself to this, however, for I can also use, as well, what is known as the Leich system, or the Sabin system employing ground taps at the subscribers stations and pulsatory currents of opposite signs. In fact, I may denominate this part of the present invention 1, him constituting his ringing number or ringing letter. Thus, allnumbcr 1 parties-on polystation lines have the ringing number 1, all the next parties the number 2, and so on; or for letters the party line subscribers may be divided into classesA, K, F, Y, or the like. In sending a number a calling subscriber proceeds exactly as in a manual system, i. 0., sends the line number, then the ringing number. Thus, for example, the third party on line 2224 is 3 and his number appears in the directory as follows'i 2224-3. In sending this number, the calling subscriber, if using a key or single number sender such as commonly employed in Strowger systegns, sends 2, 2, 2, 4-, and then 3. The first four numbers select the line wanted. The first impulse of the fifth number lifts the ringing relay or solcnoid, which has then been brought into circuit at the connector, switch, and the next two impulses shift the ringing selector two steps, cutting ofl successively the normal or low current, and the next higher, leaving number 3 current 011 the key. Then the line remains undisturbed, as the calling subscriber simply waits, so the solenoid lets go its core, which in retracting connects the se-. lected current to line, and then after its dcscentwhich may be made anything desired in point of time, and which of course determines the duration of the ring-cuts off the current, recompletes the line and restores the ringing selector to normal, ready for a new operation. If the called subscriber does not immediately answer and a repetition of the ring is required, the caller simnlv sends the ringing number 3, 3

3 as often as re uired. This will do nothing after the complete operation of see lection but repeat the rin An ancillary advantage in my system is the absence of a-pnsh button or other calling means, so that for party line work I actually decrease the subscribers apparatus instead of increasing it.

' Other advantages'oit the invention will appearfrom the following description and the pulses. Fig. 2 is avertical section of the relay D in Fi 1, taken on the line Fig. 3 is a diagram showing a party line with four subscribers stat-ions thereon, an operators connecting cord circuit with a 7 pair of sluggish relays, and connections to a first selector switch. Fig. 4 isa diagram complementary to that of Fig. 3, showing the circuits of the first selector swltch and a; trunk with a second selector swich, a

trunktherefrom to a connector switch, selective ringing apparatus for the connector, and a connection from the connector bank contacts to the subscribers line of Fig. 3, to illustrate the scheme for ringing back on the calling line without getting a busy test.

Figs. 5 and 6 are diagrams similar to those in Figsfiand t, but showlng only one party on e.'ch line, and an ordinary ringlng relayfin a connector clrcult. F1g. 7 shows a modificationof the c rcuit of Fig. 6, and F1g.'8-1s adeta1l; 1

Before proceeding to detailed description 'of the circuits, I will first briefly describe the pieces of apparatus employed. These comprise the sluggish relays, D and D, the. automatic switches withtheir operating parts and magnet-s, the ringing relay D and its selector D D, and the subsidiary parts employed at'the s'ubscrihers station. I will then take up Figs. 5 and 6, from which the party line apparatus is omitted, finally passing to' Figs. 3 and 4, which show the fullsystem with party linearrangement.

The sluggish translating relays to which I have heretofore referred as distributing agents for the line impulses, are best shown in Figs land 2. Each consists of a solenoid with a long movable core, and each pair is mounted for convenience on'a panel (1*. The mechanical structure of each solenoid in a pair is the same, but their circuit springs are different in number and arrangement. The solenoid D closes its springs togetherv vmomentarily once in pulling up and. once in letting go. The "solenoid D. on the other hand, opens its'springs on pulling up and closes them together when its armature 1s fully retracted: Turning to the solenoid D, as shown in Figs. 1 and 2, it consists of a spool formed of a tube d with heads the upper end of the tube fitted with an iron plug (1 having a central threaded opening 'to receive the valve nut d, wlnch serves to secure the spool to a bracket d mounted on the base panel. The valve nut has a. central openingto receive the stem d of a check valve (5, which rests upon the upper face of the nut cl, and is faced with rubber,

leather, or other relatively soft material.

The opening (2 in the valve nut is large enough to afford a clearance space around the stem and the panel d is mounted vertically, so that the axes of the solenoid and the valves will also be vertical. Fitted snugly to the tube (Z 'is the iron core (Z, which normally rests in the position shown, partly withdrawn from the coil, leaving an air space above it. The lower end of the core is fitted with a projecting screw (1 carry- .ing. an insulating disk (l adjustable by Beneath means of upperand lower nuts. the end of the screw an adjustable stop' screw. (Z is mounted in a post (Z carried on the panel. This stop screw is fitted with a lock nut so that it can be set'accurately and then s locked, There is no retracting spring shown for the, cored, the force of gravity being depended upon, and it being-possible to load the core to any desired extent. I

The relay D is similar in every respect to therelay D, except, as. I have-stated. in the matter of its springs. Both the brackets (2 have depending arms (l9, upon which the respective contact springs are mounted;

[There are three of these springs, rl*, (2, and

( Z for the relay D, all normally separated.

but momentarily closed together when the core (Z goes up or down, the edge of the disk (Z in such case riding over the end of the spring (P. For the relay D there are four springs, (Z ,'(Z, and d, the first two of which tend to close together, andthe others to open when the core of the solenoid is up. lVhen the core is down,the insulating disk (Z rests against the spring (Z breaks its contact with (1. and forces together the three springs (Z d, (1 The screw 1] of the relay D is set and locked in its post so that the disk will remain in contact with the fitting the cores I may employ any desired 1 form of packing to prevent a-free escape of In a 1' I T an. The operation is as follows: hentic either solenoid receives current the core is at once sucked up into the tube, and this very fall far enough to reach the spring in either 25 rapidly, as the check valve (1 rises from case, however, for on the downward move-v ment the checkvalve (Z closes upon its seat and the air enters very slowly. During a succession of impulses therefore, it rapidly transmitted, the core d or d will simply vibrate slightly without descending, and" sisting of a solenoid with a long core d carrying a terminal insulating disk, which when down rests on springs (Z and cl, and when up pulls up springs d" and d". During its ascent and descent, and whileit is between its extreme positions, it leaves the springs closed in pairs as followsmi -d, ilia Z185 lll dlfi, d19 d20" d21 d22 hen all the way down, the last named pairs are opened, and cl d? and d d 'are closed. The springs (Z 22 are simultaneously opened, in Order to cut off battery current from wire 134 of the .detent magnet D, to deenergize the latter and permit the restoration to zero position byv its spring of the ringing selector switch.

Associated with the ringing solenoid is the switch to which I have referred as the selector or i ringing selector. This cons'ists of a pair of switch arms 14.4 and 145on a common spindle carrying a fixed ratchet wheel with which meshes a pawl on the end of the'lever of the armature of magnet D This lever is marked d, and the pawl is pivoted to its end so that in passing up it willyield and move over the teeth of the ratchet but in moving down it will turn the wheel. Then in disuse it remains out of engagement with the wheel. The ratchetwheel .has a detent or stop pawl carried on the armature lever of the magnet D When the magnets D and D are deenergized, both the working and the stop pawls are held out of engagement with the ratchet, by

gravity or suitable springs. ,When the mag-- net 'D isdeenergized both pawls are off and jso ' springs or studs, connected to the generator I 55. i

the ratchet wheel returns to zero er normal position under the influence of a suitable coiled spring. The circuit of magnet D* is controlled, as already stated, by the normally open springs J, 01 of the ringing solenoid, D Cooperating with the arms 14:4: and 1 15 are fixed contacts, which may he circuits. As a pair of'arms' starts to mdve around step by step, directed by theiatchet and'driven by the magnet D it connects first the generator G then G then G, then and "finally G, to. the wires l 3'5i'and 135* leading to the ringing solenoid.

The next apparatus to be touched upon is the typical automatic switch employed. It

is to be understood thatl do not lay claim herein to a particular form of switch, nor am I limited to any particular type, but the circuits and other apparatus have been de signed in the form illustrated for convenient operation with what I may designate as the Strowger type of switch. This switch was originally disclosed in Patent Number 447,918, granted MarchlO, 1891 to Almon B. Strowger; and has since been improved as shown in Patents Number 672,942, granted April 30, 1901, and Numbers 815,176 and 815,321,granted March 13, 1906, to Keith and Erickson.

In order not to depart from well-known forms of the circuits, and to gain in simplicity of description, I have chosen herein to adhere to the type mentioned, as the original two-motion switch. this connection is that since the Strowger improved switch is in wide commercial use my present circuit-s will be more readily perceived to be applicable to existing installa-' tions in the form thus chosen than if I set forth more unfamiliar forms.

The distinguishing feature of the switch.

is its two direction movement. The line contacts are. arranged in curved rows embedded or supported in insulating arcs or segments, and the switch wipers or movable contacts are supported and moved by a vertical spindle which may be stepped up by one magnet, called from this motion the vertical magnet, and which may be turned around its axis by another magnet called from the fact the rotary magnet. Originally these two magnets were controlled by two corresponding relays one connected to each side of line, and working so that when impulses were sent over one side of the line the vertical relay and its vertical magnet were operated, and when over the other side of the line the rotary relay and magnet were operated. From this method of operation it became customary to call the two corre sponding sides of the line the vertical side and the rotary side respectively. At the present time and according to the standard practice in the commercial automatic systems, the vertical relay alone does the stepping, operatingthe vertical and rotary magnets (which still retain their same -funct1ons as origmally) in succession, and

being shifted from one to the other for that purpose by a so-called. side-switch, which is controlled by what was, and is'still called,

Another point in the rotary relay. This arrangement was adopted because other things than the mere stepping of the spindle and contact wipers must be provided for and the side-switch is a convenient apparatus for the purpose.

The switch then consists of a frame having top and bottom bearings, as shown in the Keith and Erickson patent referred to, with a vertical spindle therein carrying contact wipers connected to a flexible cord attachedto terminalsgon the switch frame and thence to the trunk" leading to the switch.

Rows of multiple contacts connected to other trunks leading to swltches of a lower order or to subscribers. lines are arranged 1n curved lines around the spindle and concentric therewith. Thesideswit'ch eo'nsists o'f a set of switch'arms haylng a common pivot so they will all move together, these being shown herein at s, a, a, and a, of Fig' 3, and elsewhere. The arms are all put back into a normal or zero position, (in which they are illustrated herein), when the switch is at normal or out of use. After theswdt'ch has started they are held only by a step by step ""escapement controlled by a so-called private magnet, S, ore S Figs. 3 and 4,

and this magnetby its attraction and release of itsfarmature permits the switch to take step by step the successive positions indicated by the contacts -11, 42, 4:3, 38, 39, a0, 4A, 45, 46, and -19, a7, 4:8. The magnet thus designatedas a private magnet is really a test magnet, operating to select atrunk-or line contact that is free and not in use. If the ery first contact on which the wipers of the main spindle come happens to be free the side switch might not be moved, u'nless special provision is made, because the private magnet must,'as stated, make-two steps to work the escapement. For this purpose the rotary lua net. I shown at S? and S iii-Fi s.

3 and'), Whiclris a vibrator making and breaking its owncircu t as well as turning the p1ndle of the switch, has its armature so. adjusted that it will knock against and hold up the armature of the private magnet in its back stroke of closure; Thus when the rotary magnet starts it takes onestroke, the wipers move around one step from zero to the first contact pair, and then as the rdtary magnet armature retracts; 'ireparatory to. another n'uWement forward" the private magnet is left in control of the situation. If v the trunk connected to the contacts onwhich the wipers rest is'busy a ground will'be on them and theprivate magnet will therefore maintain its armature up, as the rotary mag net left it, and another stroke of the. latter magnet will move tl1e'\v1} )ers to the next con tact, 'Here the operation is repeated, and

so L on until-a pairot contacts are found which are not busy, when the private armaturo will fallback as soon as the rotary mag net retracts and the side switch will promptly shirta step. Tlns private magnet gized, so that only the retary ni joined together,the pinion teothed vertically so that a pawl of the rotary magnet may en? gage it at any elevation to whichthe vertical magnet may raise it. I This iswell shown in the patents-"I have mentioned 'and needs no detailed description.

There are usually provided several switches of each digit, that is duplicates'towhich lines or trunks of the next order-of switches are multipled. The reason for this-is'ob- I vious. In the selector switches the vertical movement selects arow containing the do sired thousands or hundreds as the case my be, and the rotary motion of the spindle then moves 1t around therow'until an idlepair' in that row, or for that digit, s reached,

The'multiples of a trunk orline are usually represented as shown in thedrawings herewith, as small parallel lines 10, 10 ,10", 'w Zetc. oined to the line ends; These are sup-' posed to be distributed "in the different switch banks of the same nuniber or digit.

' The spindle and its wipers are usually shown as in Figs. 3,4, 5, and (3, the former being lettered w and the latter respectively w, 1'0 and for the selectors. It'will be understood after the foregoing explanation that the spindle w or w is moved up step by step through-a .vertical 'ratchetcperated by the magnet S S", or S; and, it is turned around its axis by the rotary magnet S, S or S, as the case maybe; In the connector switches there is this difference over the selectors, that the rotary magnet is not a vibrator self actuated, butis'moved step by step just like the vertical magnet; Thisis because no idle trunk selecting is to vbedone m the connector, but aten's row and a defi nite unit 1n that rovv are to be picked out 1n-' stead.

Having thus set forth theprincipalme-fi chanicalfeatures of the switches, the circuit descriptions w ll be readily coi'nprehended;

it" remains to refer briefly to the subsidiary apparatus at the subscribers station. *In' Figs. 3 and 5 I have shownthe' usual transnutter, receiver, switclihoolnringer and condenser. These may be of any usual and prove-d type. They are siifiicientto produce all the desired effects in my present system without any other apparatus, especially in skilled j liands'; but as telephone s1.1bsc1'ibcrs and users are too often imsl'nlled, I hnd' t convenient, not to sayfn'eces'saryto provide some special pieces.

These .are' the signal Q and the key orsender K.' The signal is;

a polarized magnet with' iacard'or t'arget ijzo,

fastened to the extren'ieendof its clapper rod, exactly as the .clapp'er of for the target. In the faceot the telephone casing, as shown in Fig. 3 particularly, form a w ndow, and thej card of'target q.

lies behindlthis window. Tlie'jniagnt Q is'e tlier-bridged across th'e line and wound 7 nger 1s "1; secured, so as to geta long path ofitravel to aj.-'highresistance and impedance, jor' itisi (as; shown included in the talking circui branch with the transmitter and receiver,

. in which case it'is low wound so as'to offer currents.

as littleimpedance -as possible to the voice 7 It is polarized and-adjusted so the current normally on the line, that .'1S Wheti aplug is not in the answering jack,

,. wry simply, as willappear in'the circuit description. The target preferably bears upon its" face the word: Number? so that when it appears the word indicates that the sending apparatus is to'be manipulated toiselect the line wanted. This is literally the machine asking for the information 'ustally'given an operator and which is of course fed to a machine mechanicall Any other suitable word can of ,course lie use suchas Send or Select or the like, but

I ctinsider thefirst the best.

I lilo-specific description of the key is nec-, essary, but it will be observed that .as only momentary breaks of the circuit are required it is quite possible to omit the key and use the switchhook H, since the motion in either caseis precisely the same.

' I will now take up the circuits, and as a preliminary I may state briefly the opera tion. In calling, the subscriber takes down his receiver and waits until he gets a signal to go ahead. The operator gets the usual calling signal by means of a lamp or shutter, and inserts an, answering plug. sertion of this p-l ug 'energizes the cut-01f relay, which disconnects the line from the signal relay and its battery, and connects the line to main battery through the cord. The connections of the battery and the cord are the reverse of thoseon the line before a plug is inserted, hence the insertion ofthe plug, by sending reversed current to dine operatively energizes the subscribers signal,

displaying an indication for him to proceed. Thereupon he manipulates his key or switch hook,'or other make and break device as in an ordinary automatic system, sending the groups of impulses which make up the digits. of the number wanted' Since the answering plug has connected with it 'responsive relays corresponding to the supervisory relays of the ordinary manual system,

and since the cord circuit is connected (through the condensers for conversation) with a trunk leading to an automatic switch, I cause the impulses to act upon one or-more of the relays to repeat over the'trunk to the switch or switches and thus make connection'with fihe wanted line. If that line *is' a party line, the calling subscriber transmits a fifth or extra group of impulses after he has selected'the line, this group of impulses being effective to select and put upon. the called line suitable current to ring the bell The in- .sible to connect therewith. After completing the connection, the calling subscriber retains control thereof until he finally hangs up his receiver, although what are known as back release magnets may be placed in each connector switch so that the line of the called subscriber cannot be tied up. Upon the calling subscriber replacing his receiver,

his metallic circuit is broken and remains broken, which causes a reversal of the solenoid conditions in the cord circuit, to sendproper impulses over both sides of the metallic trunk circuit, to release the auto-v matic switches; and at the same time the cord relay, which works the solenoid, lights the supervisory lamp, informing the operator that disconnection is desired. It is to be noted that a subscriber may. hang up long enough to restore the switches, which re-'.

quires perhaps two seconds in the system I have operated, and may then take his receiver for a new call without further delay. If the operator has. pulled the plug his signal'to go ahead does not show until he is given another plug and set of switches, but if she has not pulled the plug'his signal will show. at once, and he can make his recall without delay.

Referring now to Figs.- 5 and 6, which show a system without party lines, and regarding the two figures as one continuous diagram for the purposes of this description, A is a subscribers station provided with the usual transmitter T, receiver R, ringer Q, condenser C, switch hook H, and connections- In addition to these I provide the polarized magnetcontrolling the signal target, lettered respectively Q" and g, and the signaling key K. Both the key and the signal are in the transmitter branch, and are normally out of use. They may be placed in a shunt of the transmitter branch, especially the signal, which should then be high Wound. In the present illustration, being in series with the talking apparatus, it is supposed to be -very low wound, or .of very lowimpedance. The line wires 1 and Qare normally closed at the sub-station through the ringer, condenser, switch hook, and key, by way of I the wires 3 and 4. When the hook is up the path'is shifted to 1' -3-H-R-T-5-Q-2. At the central station the line wires terminate respectively, 'at the contacts n and n of the cut-off relay N, which are normally in contact with terminals of the wires 7 and 8, leading, respectively, 'to the line N is excited the terminals 11, n are connccted to wires 9 and 10 leading to the connection. terminals which are in multiple and comprise the EIHSWQllIlgJflCk J and themultiple switch terminals w a which are sup-posed to be distributed in different connector switches-through which this line can :foe reached. The relay N is in a ground ta'pfrom the sleeve side 10 of theterminals. The relay M through its armature/ m controls-the circuit-"of the :line signal lamp "m.

(For answeringicalls the operator is provided with a plug P whose tip-and sleeve are connected respectively with the cord conductors 1'1 and 1 2,and thence through the'condensers G" and C to the trunk wires 13:" -a-n-d. 14 leading to the automatic first t selector switch 5W. Across the plug end of the cord'are bridged the relays S S which correspond very closely in their arrangement 'and funct on to the supervisory relays'cf an =ord1nary cord circuit. The mam v battery B is included in the bridge between them, l the path -;fro1n one side of thecord tothe -;other being-as follows: B, 21, S '18,

v andvfrointhe other side'll by'wire 17, S,

and ground to battery.

The relay S being'on the live"side of'battery,takes current as soon as the plug is inserted by the followingpath: B, 21,8 18, 12, P

g J 10, N,;g-round, to battery. This energizes both relays S 7 and N the armature s is attracted,,and the battery is connected to the armature s of the relay S. This relay S takes current returning from the subscrihe-rs station; and hence is energized only so flong-as' the subscriber keeps the circuit closed. This control by thesubscriber' enables me to use the relay S for my control ling relay, which it really is. Like the .su 'iervisory relay in a manual system, it

follows every motion out-of the subscribers hookor key It. has front and back 'contacts, former connected by wire 23 to the-solenoid D, through branch wire, 33,

and to ground ext-31. A branch 29 is takerr through the springs a? (150i thesolenoid Di, and thence to-the trunk wire 14 by, branch 34, 'iavmg also a branch 35 con-1 nectcd to the spring zi tof the solenoid D. From the backgcontact of the relay S the wllfcfiii goes to the supervisory or clearing out lamp so", also through a branch 26 to the solenoid D and ground at 31, with a second branch 30 'to the spring (1? of the solenoid. ll. The battery supply for; cer:

tain. portions of the solenoid. operation is tlmsrontrollerl by the relays, but for certain other portions it is independent of the relay, the wire '22 being connected direct through the branch 220 to the spring d".

of the solenoid D.? From the spring 0? of this solenoid a wire 36 goes to the trunk Commencing with 'the'wires 13 and 14,

the rest, of the diagram relates to purely automatic apparatus. I have shown afirst selector switch W, whose trunk line is 1314, and is permanently connected to, and associated with, the answering plug It is to be observed that the first selectorsare usuall the same in number as the subscribers lines, and that what I have; done therefore by putting these first selectors on plugs is to reduce the number of first. se-- lectors to that of the plugs, or from100 per cent. to not more than 15 percent, a saving of .85per cent. minimum. -Moreover, the manual equipment is that now employed inmany large installations, and this system enables me to increase their capacity enormously without other change than that of connecting multiple jacks to act as answering jacks," and adding a numher of automatic switches ranging from 15 to 35 percent. of the number of lines, according to the size of theplant, and the amount of business itcarries. The numberof operators is greatly reduced, since one, operator can handle at least five times the number of. lines in this way that she can. handle in a straight multiple system. For clearness I should add that in a full automatic system the number of switches ranges; from 120 to 135 per cent. of the number of lines. Taking the vmaximum therefore, I' save percent. as against an automatic system, the answering jacks beinga small off-set, and the number of operators employed costing much'less than the mechanics necessary to care for the, automatic switches.

I have already described the switch which j I have shown as a type, and I need only say therefore, that w is the spindle carrying thewipers 'w, w w which are supposed to sweep around over horizontal curvedrows of contacts to which they are vertically stepped up, these being typified by w, 20 w". The magnet which directly steps up the spindle is S, and the magnet which rotates it after it is stepped up to thedesired row of contacts is S. The so-called private magnctisS, whose armature 8 controls the circuit ESQ-of the vertical stepper magnet S andby u'eans of a mechanical connection not shown works the. escapement for the side switch whose arms a flsj s and 's are all rigidlyattached together and-move as a-unit- -onestep, i. 6. over one contact each, at a-time. T he side switch is shown in the, normal position it occupies when the selec tor is at rest, and to this positionit is v brought by a mechanical connection with the spindle when it falls to its normal position of disuse. The spindle is prevented from falling after being rotated on any row by a stop or detent pawl and also by the engagement of the wipers with the horizontal row of contacts, and it is held against turning back until released by a second arm on the same detent. This detentpawl is controlled by the release magnet S whose circuit is also controlled in part by the private relay 8*, and is finally controlled through the wiper w", as will appear in the statement of operation. The relay S is the .vertical relay, which first controls the vertical stepper magnet S then the rotary magnet S, and is-finally disconnected in the first and second selectors.

S is the rotary relay, which, in the first two positions of the side-switch, responds to single impulses over thetrunk wire 14 to close the circuit of the private magnet S in order to move the side switch.

The connection between the rotary magnet S and the private magnet S is as follows: when the rotary magnet starts to vibrate in the second stage in the operation of the selector, aprojection' carried on its armature kicks against the armature s so as to hold the latter up on the stroke and release it for testafterward. In case the first contact the wiper 10 comes on is idle this would be necessary because as previously stated it takes an up and back stroke of the armature s to release the escapement each step. Hence if no ground were encountered and no mechanical. means were provided, there would be no second forward movement of the armature 8". Since each full viin'ation of the armature of the rotary magnet S moves the spindle and wiper 20 the space of one contact, however, and since the kick of the rotary armature simultaneously moves up the armature a at the sane time, then if the wiper e0 finds no ground on a. contact the armature s is in a position to release the escapementat once andv so stop further rotation by letting go the side switch and cutting off the rotary-magnet at. contact 41-2, the switch arm a then passing oil? that contact to the last one :43, a. blank. If, on the other hand, the wiper finds ground, itwill hold up after the kick up, and keep the side switclrset on contact 42 until. another step has been taken, and so on.

The parts and connections in the second selector are all the same as in the first so lector, except. that the vertical and the rotary relays S and 8 by their joint action pulling up boththeir armatures at the same time close the two contacts 103 and 107 together and so close the release circuitas followsL'i'O, 10?), 107, i 75, 1000, battery, from the private bank contacts '11: and so back to the wiper or, wire 61 r ay 3 8 -18, 59., and ground. Thai.

first selector as well asv the second one. To do it requires the simultaneousoperation of both relays S and 8 by impulses over both wires (37 and 68. The vertical magnet of the second selector is marked 8", the rotary magnet S, the release magnet S, the private magnet S, the test relay S. The

side switch has arms 8", s, s, and a, all

working simultaneously, as before. The second selector as a whole is marked W".

The multiple bank contacts w, w and w of the second selector are connected to trunk wires 97, 98 and 99 of the connector switch W. This last is different in several respects from the selectors. In the first place,

the vertical relay S is not deprived of its control after the first set of impulses and the first step of the side switch arms, being in fact independent, as is also the rotary relay, of the side switch, their bridge wires 100 and 102 being permanently connected to the trunk wires 97 and 98. In the second place the rotary magnet S has a stepper pawl which turns the spindle w only step by step one step at a time, the magnet not being a vibrator like the others, S and S, It should be herestated that these magnets are shown'each with a wedge shaped insulating extension which as the armature is attracted passes between and breaks apart the terminals shown between. the windings. This is the usual diagrammatic way of representing these rotary magnets, but any other form of vibrating contact might be enuployed, such as the ordinary buzzer contacts closing on the back stroke. The magnet t? we no such contacts and responds only to pulses from the controlling relay 5. In the third place, instead of the side switch arm 14G coming to a blank on its third step, it connects the local wire'120 of the vertical relay S to the wire 12? of the ringing relay D, which thereafter remains under the control of the said vertical relay. Finally, there is no separate wiper for testing, on the spindle 10", the test being made directly on the talking wiper 'w, as will hereafter appear. The release magnet or relay is S. controlled as in the second selector by the joint action of the vertical and rotary relays S and S The wipers 'w and 'w of the connector switch cooperate with line terminals directly, to", w,

heretofore noticed in connection. with line l-EZ. I have shownairother line l-Q at the right of Fig. 6 which is supposed to be equipped at. central and the substation just as the line in Fig. already described.

Having thus presented. briciiy l .=i with an attempt atclcarncss, the arrangement of. the switches, which as stated in themselves form no part of the present invention, I will now proceed with the s'-:tatcmcut of operation, in which the dctail t' th entire "iieases the system will be brought out fully.

We will assume that the subscriber A desires connection with say the subscriber on the line to the right of Fig. 6, who may for convenience be called A. A takes down"- his receiver R and waits just as he would in a manual system. As soon as his hook H rises the following circuit is closed for the ground 8, closing the line wires 1 and 2 and so the lineto the cord, as followszB 1000, 21, 8 ,152, P, J, 10,n', 2, Q,, 5, T, R, H, a, K, 1, n, 9, J, P, '(tip), 11, 17, S, ground to B. This'energizes relays S and S, the relay N having already taken current however over the circuit traced. The first effect of this is to cause magnet Q at the substation to turn over'its armature, raising the target 9 so that it will show through the little window in the front of the telephone casing, the

.magnet being so polarized and connected that the current through line relay M was of the wrong direction and would not raise the'target, but now receiving reversed current from battery through the sleeve side of line. The second efiect to be noticed is that relay S put current on the wire 22 and hence on the armature of relay S, but the latter being now attracted thelamp wire 25 is cut oil? and the lamp remains dark. The branch wire 26 is also cut ofi,-but the wire 23 takes current through the front contact of the relay S and the solenoid D'- immediately pulls up its core, closing together-the springs 61 and d and opening the springs d 'd, and Z The former pair complete the wire connection 36, 30-, 26 to the back contact of relay S, so that if that relay were now deenergized without deie'nergizing the solenoid D, a pulse of current would be sent to wire 13. However, the parts remain unmoved until the subscriber does something at this point. It will be observed that the solenoid- D is deenergized and its springs d (i are open. The sub scriber getting the signal 9 set before him, preferably bearing the Word Number i, now sends a group of short breaks by means of the key, corresponding to the first digit of the number wanted; Suppose that numher to be 2223. He sends two short breaks by means of two taps on the key. The effect of this is to cause the relay S to let go lts armature twice in quick succession and pull it up again each time, the second time holding it up. On the first retraction the sole-1 noid D of course loses current, and so on 26,30, d", d, 36. The first of these the second, but neither break is long enough to permit the core a! to drop far enough to reach the springs, and each make pulls. it up again to the top. On the other hand the solenoid D takes current on the.first break, 7o

since its circuit is then closed as follows: B,

1000, 8 22, s, 26, D, 31. This solenoid pulls 'up its core d at once and fully, althoughthe'pulse of current it receives is short, because it is so constructed that the up stroke is very quick. ()n-the lineomake following the break the core (I of course starts to. descend, but there is not time before the other pulse comes due to the second break in the v line, and so the springsd, d remain open.

It should be noted that nothingis done by the momentary closure of the springs on the up stroke, as their source of current sup 1y is cut off, the armature 8 being back at t at instant. Each time the break comes in the line, however, a pulse is sent tothe wire 13,

(the vertical sideof the automatic circuit) by the following path: B, 1000,, 21, s, 22, s,

v H I asses ,alliiig the wire ,13 to the switch v and thriough its vertical relaygS by the following path; 13,8 381,- i9, S,"and-ground to B.

The relay S becomes energized and pulls up its armature, closing the following circuit; B, 1000, 55, 54., s 52, 65, c4,- 51, 50, ground to B. This energizes the magnet S, which immediately acts to step up the spindle 'w one step. The second pulse follows the same route, by the same transmutations 'from substation break to automatic pulsation, and the magnet steps the spindle up another step, so that the wipers w w are opposite the second row of contacts in the banks, these being connected to trunks leading-to a set of second selector switches numbered two thousand. Some of these may be busy, and it isnow the function of the switch W itself without any act on the part of the subscriberto get on an idle trunk and an idle second selector. For

this purpose the side switch must. be worked, a

and this is done by a pulse of current over the rotary trunk wire 14: as follows: \Vhen.

thesubscriber let go the key the second time the circuit remained closed at his station and the armature s consequently remained up. The solenoid D was therefore deprived. of

current, and as its core sinks down it passes the springs d, (i pressing them together and closing the following circuit, which is now possible because the armature s is up:

B, 1000, 21, 8 22, 8, 23-, 29, d", a 34, 14.,

s 20, S ground to B. This energizes the relay 8, which closes the circuit of the private ma-gnet as'follows: B, 1000,55, S, 500,

50, ground to B. The magnet S pulls up 'its armature s, cuts of! the vertical magnet S at 65 by opening the pair of springs 64 65, and as the impulse ceases to energize it, permits its armature toretract and move ,theescapement the secondhalf of itsstep .ter, owing to divergent resistances.

releasing-the arms 8*, 81, 8 s, so that- Y of the rotary magnet S is then closed as' follows: 13, 1000, 55, 54, S, 53, 42, 8, ground to B. The vibrator magnet S then commences tomove its armature back and forth and with it the projection which kicks thearmature s? of the private or escapement magnet, and the pawlwhich steps around the spindle in. "At the first step the spindle is-moved so. that the wipers come onto the first pair of contacts, the armature s is kicked. up to put the escapement in operative condition, and the wiper to if its contact is busy and grounded because another I connection already exists thereon, closes a circuit for the test magnet as follows:

ground, B, 1000, 55, S 470, 47, s, 60, S

v61, wiper w and so to the ground on said wiper, to battery B. The two magnets S and S itwill be observed are in series in this-circuit. The former will be effectively energized by current therein but not the lat- Hence -the armature s will hold up on the second half 'of the step after the armature of the rotary or vibrator relay '5 has retracted, ,and the second movement will be permitted. As the wipers come on the second contacts the operation is repeated. The armature of S kicks up the armature s, the wiper w tests, but suppose in this case there is no .other connection on that trunk thus tested,

and so there is no ground on the contacts to. The circuit last above traced does not exist, the magnet S gets no current as the armature of S retracts and since the kicker is Withdrawn there is nothing to hold up the armature s so it too retracts and lets go the escapement, which at once steps around the arms 8, s, s, and s the last step. The first named cuts ofi the rotary relay 8, and that instantly stops the rotation of the spin- ,dle to, which thereafter stands with its wipers on the contacts thus selected. The second named disconnects the wire 13 from wire 19 and connects it t. the wiper w through the wire 15. The third named. does a similar change, connecting wire 14 on 16 and Wiper 'w The last named connectsthe test magnet S tothe ground through contact 48, thus also connectin aground to the contact '10 (see Fig. 6, lefl: hand side), so that as longas the connection we are tracing exists that trunk will test busy and be free from use'by others. This leaves the line 1-2 connected through wires 13-'1567, and 14-1668 to the second selector \V.

The statement of the operation has taken some pages and considerable time, but of course in practice the subscriber has been waiting only a couple of seconds to send the second set of impulses. I have found in practice that the'time required, averaged in :a good many trials, for the rotary magnet to select an idle trunk, where most of the trials went beyond the fifth contact, was less than a second, and it is stated by the manufacturers of the type of switches I show that the whole ten contacts can he tested in onehalf a second. The movement of the spindle in rotating is continuous, there bein no separate steps as I have described tlem. Keeping this time in mind, it is apparent that the subscriber, if he leaves an interval at all, will leave a long enough one in the very natureof things to permit selection of an idle second selector. He then sends two more short breaks by two taps on the key K. The results in thelsecond selector circuit are absolutely the same as in that of the first selector, except that it is now the desired hundred that is picked out in stepping two atcd together, as will appear in the statcment of the release, then a circuit is closed to the release magnet S and also to contacts of relay S, the test relay. WVhen the latter operates it closes the circuit of release magnet S over conductor 96, conductor 60, and conductor 70 to ground. Inasmuch as terminals w of switch V are disconnected from conductor 66, switch W is not affected by this release. The release of the first selector would accompany or follow momentarily the release of the second selector, a special release circuit being closed over conductor 1,001 by solenoid 1) when the calling subscriber hangs up his receiver. Magnet S restores switch W, and relay S pulls up and closes circuit B, 1000, 8, ground to B, for the magnet S which restores switch XV. It is thus impossible for any subscriber to be tied up and lose control of his line, although if he should by mischance get into such a condition in my present system, he canget the operato1"s'attention by continually flashing his lamp 8. This is'one of the most striking features of the entire scheme, that the subscriber, while doing his own work, and free from listening in and interference on the part of the operator, can always get a human being to talk to and to get help from, when he so desires. It is not to be understood from thisthat the subscribers operator listens to his complaints etc. The operator merely connects the subscribers line with a cord circuit leading to a manual operator who has time and facilities for conversation. An ordinary plug and jack connection may be'used for this pur-f pose. I

The subscriber has now sent'the digits to select two thousand two hundred, and wishes to select twenty three. He gives a third series of two taps on the key, producing two breaks again. The same results as before follow as regards the solenoids D and D,

and two impulses come over the wires. 13-

07-97 through the branclr wire -100 ahrough relay S and by wire 101 to ground. Relay 8 "pulls up twice, and each time closes the) following circuit for the vertical stepper magnet S: B, 1000, 123,91 121, 146, 120,

springs of relay S, 119,110, and ground.

This steps up the spindle with its wipers, e0

and w. to the second row. The subscriber having paused with the key- K closed, the

solenoids act to send a single pulse over the rotary side of line as follows: B, 1000, 21, 8

. 22', 5323,28, 29, d, (1234,14, 08, 98, 102, S", 101 and groundto B; Rotary relay S pulls up on this impulse closing the side switch control or private magnet circuit as follows: B, 1000, S, 111, 110, ground to B.

'- Magnet S? pulls up and falls back letting the side switch arms 146, 147, 148, and 149 go ahead astep, the first 146 disconnecting.

the vertical nia 'net S and connecting the rotary'magnet t rough wire 122 and contact 151; the second and third passing onto blank 'contacts 154and 157; andthe fourth connecting the test Wire 175 from the test wiper w to the wire 231 leading through contacts of relay S which are normally open as shown, to wire 230- and through release magnet- S? to battery B. The subscriber now sends his last group of three breaks by tapping his key K. The circuit of the arm 146 having shifted from the verticalmagnet S to the rotary magnet S, the latter responds by stepping the spindle around until the wipers 'w and 10 rest .on the third pair of contacts in the row.- This is the pair desired.

Now, as the subscriber pauses for the last time, the. solenoids send their accustomed pulse; of current again over the rotary side of line,146898, and the private magnet S is again energized and denergized thereby to let go the side switch. As soon as it is energized, however, and before it .can be deenergiized, the line is tested. That is, the "closing of the contacts of the magnetS completes the following circuit: from wiper w by test wire 175, 149,, 160, 231, contacts, 231,230, S 1000-, and battery. B to round.

The wiperw is lookingfor a ground: Now "it will'be observed that if another line or v A switch is connected with the line wanted 9 there will be a ground on the tip side of the line. That is, to say, if another switch is i already on' the line' then the ground will be by the path from the wiper contact wto 5 2 the wiper on it, its wire to 155, 147, 130 and 129 to124,.relayIS and ground. If anan proceeds as usual.

swering plug has been inserted there will also be a groundon this tip side of the line from wire 0 (see Fig. 3), through the jack,

plug tip andwire 11 to branch 17 and through the relay S to ground. So in elther event t-he'relay or magnet S which is testing, gets current before the relay S can let go, and the switch is released without calling the line wanted or disturbing it. On the other hand if the line is not busy, that is if there is neither plugaror switch in con-...

nection with it, then'th'e test is clear, and the relay S can let go, passing the side switch to the third position with all its arms. The arm 140 connects the wire 120 from the vertical relay to the ringing relay wire 127;

the arm 147 connects the trunk wire'l29- 130 to the tip co11tact-wiper10 the arm 148 similarly connects the sleeve trunk wire 131.132' to the sleeve wiper w; and the arm 149 connects the test wire 175 'to a blank 161, so as not to unbalance the talking" circuit.

The calling'lin'e is now through to the line wanted, but ringing. must take place. This is accomplished by sending any number of impulses indeterminately with the key K. ta'ch impulse, or break, produces an attraction of the armature of the vertical relay S, which closes the ringing relay'cir cuit as follows: B, 1000, 123, D1127, 152,

140, 120, 119, 110, and gromrdfl'preferably make the relay D a solenoid, as will be pres-v ently described, and so the numbers of short breaks cause thecoreor armature-to be attracted to break the talking circuit at (P -(Z and d "cZ connecting the springs (Z and d to (Z and Z respectively, which are connected to the generator G. I11: this waythe generator is conncctedtothe wipers and ringing current sent to the wanted line. A.s the solenoid or relay lets go, the springs will again assume the positions shown, and complete the talking circuit. The called subscriber then answers and conversation It is true that thetalking circuit shown is through two pairs of condensers, but in another'portion of the description a means for cutting these out during connection willbe given. Y

\Vhen the subscribers have finished conversing, the receivers are returned to the hooks, andthat of subscriber A produces the .followin results: The metallic circuit is broken and remains broken, hence the relay S lets its armature stay back. This permits the other of the selector trunk.

matter to make the practising of the 11113311 through the branches and 102' and the ,magnets S and S" to ground. Both these magnets are energized at once and their joint action closes the circuit of the release as follows: I3, 1000, S, 230, 112, 108, 99, to, a, S 5, 9i and ground. .Iagnet S closes and releases the connector switch mechanism by raising the detent pawls which hold the spindle, allowing a retractile spring on the said spindle to turn it to zero position, when its own weight pulls it down to its normal level. At the same time relay S closes the circuit of its release magnet as follows: B, 1000, 75, S 96, 66, 70 and ground to B. This releases the switch iv and the release relay S of'switch IV is energized, switch being thereupon released as aforesaid. The retraction of the armature of relay S causes the lamps 3 to glow continuously, and the operator perceiving this signal pulls the plug P and thereupon the cut-oil? relay N and the control relay S in the cord both let go their armatures, and the battery is reversed on the line again, while it is removed entirely from all the cord apparatus. The solenoid. D is thus finally let down, and all is ready for another call.

Two points remain for me to mention in respect to this arrangement. One is the bridged battery for the called subscriber, which feeds current to his line from the main battery B through two impedance coils S and S The other is the proper arrangement of the trunk and feed wires with respect to the cord circuit wires 1112. lVith a break on theside toward the line and battery coming on the other side of the condenser it is not always quite easy to get clear transmission, but it the connections are all reversed, that is if all the battery is supplied for the vertical and rotary relays from the other ends, and the solenoid put on grounds where I have shown them putting on battery, this trouble is entirely avoided. Also, if the tip and sleeve contacts of the plug P are connected respectively to the wires 14: and 13, instead of as shown to the wires 13 and 14:, or in other words the tip goes through its condenser to the rotary side of the 1st select-or trunk instead of to the vertical side as 'snown, the result is quite satisfactory. The interchange of grounds for battery connections, and battery connections for grounds, is of course not a matter of invention, as long as the circuits are unchanged thereby, nor is the connection of one side or the other of the plug circuit to one or the I mention the tion perfectly easy and simple.- The changes specified and many others are within the scope and purview of the invention.

Referring again to Figs. 3 and .4: I will now describe the operation of that part of my system relating to selective party line ringing. The operation of the first and sec ond. selectors and connector switch is identical with. that described in Figs. and 6 and need not be described a second time. In Figs. 3 and i I have for illustration shown a subscriber calling a party on his own line. I have shown it in this manner so as to describe the method of calling back on the same line without its testing busy. Taking itfor granted that the calling subscriber hasscnt the last set of vertical impulses through the connector switch in the manner already described in connection with Figs. and 6, the wipers of the connector switch rest on the contacts leading to the desired line, in this case his own line. After the last set of vertical impulses have been sent in there follows a long make at the subscribers station as usual. The relay S, Fig. 3, being energized at this time opens the circuit of the solenoid D, allowing the core (1 to drop, momentarily closing the contact springs d, d and (1 An impulse is sent over the rotary side of the trunk as usual energizing the rotary relay S" in the connector switch, Fig. 4, which connects the test relay S to the test wiper 119 of the side switch and from there to the tip side of the desired line. Ordinarily the relay S would become energized on account of there being an answering plug in the jack. thereby placing a ground on the tip side of the line. In this case, however, where the calling subscriber is calling back on his own line the test relay S is shunted in the following manner; at the same time the momentary impulse is sent in over the rotary side of the trunk a similar impulse is sent in over the tip side of the line by the following path: Battery B, 21, 8 22, s, 23, 28, 29, 32, d", (1", 27, 11, 9, co Q05. This impulse being of the same polarity as that of the relay S neutralizes it, preventing it from becoming energized, thereby enabling a calling subscriber to call back on his own line. After this rotary impulse has been sent in and the line not testing busy the calling subscriber is connected to contacts in the connector switch leading to his own line. The operation in calling a party on some other line would be identical only the impulse sent in over the tip side of the calling line could not affect the test in the called line. Suppose he desiregto be connected with subscriber A" whose ring: ing number is 3, he manipulates the key K three times sending in that number'of impulses operating the relay S in the connector switch, Fig. 1, which in turn closes the circuit of the solenoid D? by the following path: Ground, 110, 109, 119, 115, 116, 120, 11-16, 152, 127, 126, D 125, 162, to battery. The solenoid D pulls up at the first impulse and simply vibrates as the other impulses come in. In pulling up it opens the contacts d d and d. disconnecting the calling end of the trunk from the called end to prevent generator current from flowing over the trunk to thecalling subscriber. The contacts (Z (P are closed completing the mrcuitotthe release magnet D* whose armature now engages the ratchet wheel controlling the arms 144, The operation of the relay S also-clo-ses'the circuit of the stepping magnet D by the following path:

Ground, 110,109, 119, 115, 116, 120, 146', 127,.

128, D 133, to batteryB. This stepping magnet responds to every impulse sent in o zerthe vertical side of the line. and astl frcc impulses were sent in it attracts its armature d carrying a pawl which engages the ratchet wheel, moving the armsl44 and.

145 three steps, therebyconnecting current terminals from the generator G to the contact springs (1 and (Z -0f the solenoid D 'Efhese contact springs normally engage the springs (Z and (Z but while the solenoid is energized these'contacts are broken preventing generator current from being connected to the called line until that of. the proper .frequency is actually selected. After the three impulses have come in and the right kind of ringing current has been selected the solenoid D becomes deenergized and the core tl slowly drops. At the'first movement the contacts (Z and d", (P are closed and generator current is supplied to the called line by the following path: Over No. '3 side station.

of the line from generator G ,139, 145, 135, (Z a-(Z (1 (Z 130,147, 155,20 ,10 9,11, 1, to station A; over No. 6 side of the line from generator G, 137, 144, 135, d, ,07 J, 148, 158, 'w", w ,,l0,'n', 2, 6, to station A, thusringing the bell Q, of the desired Generator current is supplied to the called line until the core (1 ofthe sole- ,noid D comes to a stop,.whei1 the contacts. ai cl and (1 (Z are opened and thecontacts c (5. and J, d are connected thus cutting oif the generator current and connecting the calling end of the trunk to the called end completing the talking circuit. At the same time the circuit of the release magnet D is opened by breaking the contacts (Z and d, whereupon this magnet becoming deenergized allows its armature to -fall back, and releases the arms 144, 145

which, under the tension of a coiled spring, return to their normal or starting posit-ion.

From the foregoing description of opera tion it will be observed that the calling subscribers talking set is left bridged across its line while ringingcurrent is supplied to-the called line. In party li'neservice where the calling subscriber rings back on his own line. this talking set is liable to shunt- Sllfl'lClQIliJ' current to prevent the bell of the desired party from ringing, since it is or low-resistance in comparison with the ringers. In

order to. provide or this contingency I add zur attacl11nent' to'the polarized signal Q Which upon vibrating 1. 1 response to the free side of the battery. Instead of doing this I can use the sleeve, side of the line for modification. oft-he connector switch illusgenerator current rapidly opens and closes its own circuit, winch passes through the ward violently by the armature and returns to'its normal position rather slowly and during this interval the coils of the signal Q39 are disconnected and the talking circuit is completed through its resistance coil X 9.

- This resistance coil is addedias. a means for keeping the relay S of the cord circuit energized while ringing, but it should not-edv that the ringing generator current itself flowing out-on the line will also flowback through the supervisoryrelay S to ground and so keep it energized. In place of the;

: hunt resistance I may use any variable re:

sistance such as a carbon button and point which can be acted upon directly by the 95, armature 9 It will also be noticed that for testing purposes I haveused the tipside of the line, the test relay being connected ,to the testing purposes by having the .test relay connected directly to the grounded side of the battery, as the cut-ofl line relayjinthis system is also connected to the grounded-side of the battery and would theretorenot interfere with testing.

At the end of the conversationthecalling subscriber upon hanging up his receiver releases the various switches in the same manner as already described in connection with Figs. 5 and 6 and the operator getting the disconnect signal withdraws the plug in the regular manner.

Referring now to Fig, 7d haveshow'n a H trated in Fig. 6; which, enables metoshort circuit the condensers G and Candalso disconnectthe bridged. release relays S? and S", thus giving me apical-talking cir- -cui't back to the condensers in the cord cir-f cuit Fig. 3. The changes I make do not change or interferewith the operatmnof' the switch. In place of the retardation co l '3 shown in Fig. 6, I substitute the relay S The terminals ogftheirelays S 9 and instead of beinggconnected di. any-t0 the trunk Wires 9? and 98 respectively are con; nected indirectly thronghthe contacts 201, v 202 and 205, 203 of the relay,v 9. 1 have alsoadded alocking relay S 2,onetcrniina'l 13o of which is connected to one of a pair of normally open contacts 210 and .209 which are controlled by the spindle 10*. The other terminal of -this relay is connected to a normally open contact 208 of the relay and the two contacts 230 and 231' respectirely are connected to the contacts 206. and 2070f the same relay. The operation of this switch up to the point of calling and ringin the desired party is the same as already described and need not be described again. The difference in operation arises after the called party answers. In removing his telephone he closes the line circuit and energizes the relay S. This relay thereupon attracting its armature opens the contacts 201, 202 and 205, 203, disconnecting the relays S and S" and closes the contacts 200. 202, and 203, 2%, short circuiting the condensers C and 0". 'It also closes the circuit'ot' the relay S by the following path: Battery B, 1000, 12s, 21s, 208, 21.1, 5 216, 210, to the contact 209 to ground. The contacts 209 and 210 were closed when'the spindle w made its first step. The relay S attracting its armature closes the contacts 2:20, 2 231,-

232 to battery, thereby placing battery on the contacts 206 and 207 of the relay S which are now open. It will now be noticed that the called-party has control of the releasing of the switches, as the relays S and S" which were controlled by the calling subscriber are now disconnected. The release is now'efi'ected in the following manner: The called subscriber in hanging up his telephone de'e nc-rgizes the relay S whose armatures upon falling back again connect the relaysS and S" to the trunk wires 97 and 98 respectively and remove the short circuit from the condensers. The contacts 208 and 218 which first close the circuit ot the relay S are now broken but it still remains energized and locked by the following path:

Ground, 209, 210, 210, 232, 220, 217, 123, to

battery. This relay remaining locked furnishes battery to the two relays S and S" respectively through the contacts 207, 202, 201 and 206, 203, 205. These relays becoming energized simultaneously, eil'ect the release of the connector and selector switches 1,000, 230, 112,108,99,w ,-10 ,8O, S 8

- 94:,and groundto battery. The release mag- I lease of the first selector net S immediately restores the connector switch w while the relay'S closes the circuit of the release magnet S of the second selector by way of conductor 96, contacts of relay S 66, 70, (39 and ground. The reis however as shown in Fig. 5 under the control of the calling subscriber. Not until the calling subscriber hangs up 'is' his switch released. At

that. time solenoid D is deenergizcd and by its lowermost contact d 'puts battery onto conductor 1,001 thereby energizing relay S by way of side switch contacts s, l8, and conductor 59 to ground. The spindle w" when it has returned to its normal position opens the contacts 209 and 2.10 and releases the relay S whose armature when falling back removes battery from the contacts .206 and 207 dct ncrgizing the. relays S and S". In case the called subscriber did not an swcr the relay S'- would not be operated and the relay S would not be afi'cctcd, and the calling subscriber in hanging up would operate the relays S and S restoring th switches in the usual manner.

It should be noted that the arrangement of relays S, S and S, shown in Fig. 7 may well be employed without the locking relay S or the release cireuitsconnected therewith. The function of relay S would then be merely to cut ott relays S and S" and shunt the condensers during conversation. It should also be observed that I have shown a special release wire 1001 in Fig. 5, connected at one end to the wire 61 of the release magnet S and at the other end to a back contact engaged by the battery wire contact (l of the relay D when down. By this means the calling subscriber has control of the switch V at all times and can restore it to make a recall. I consider this means of double release control to be original with me. In the ringing circuit herein it should be noted that a battery such as B will be connected through an impedance coil to the sleeve side of the ringing circuit during ringing in the ordinary way to hold up the cut off relay N. The ringing circuits shown in Fig. :i: can be replaced as to the selector switch arms le k-145 by any of the standard forms of circuits, with ground taps on opposite sides of line or the like.

I wish to make it plain that the apparatus and circuits herein described are fully applicable without any change requiring invention to full automatic. A subscriber's line circuit need only be connected permanently to the wires 11-12, and the control of test condition given the relay S. The ringing scheme is contemplated for any type of system, as well as the reverting busy test. I do not wish to limit myself therefore to the particular circuits described.

It will be noted that while I have shown means in Fig. 7 for short circuiting the Com denscrs C" and C" in the connector circuit, and for cutting oli th vertical and rotary relays during conversation, this showing is omitted from Fig. 4, in order to simplify the drawing. It is to be understood that the short circuiting and cut-oil relay may be substituted for the relay S in any circuit of this kind where it is desired to have a clean path for talking currents.

I do not herein claim broadly automatic switches controlled by relays having different time constants or through impulses of

Images