GB1567477A - Digital switching for telephone communication systems - Google Patents

Description

(54) DIGITAl SWlTCf II NC; K R 'l'ELEPHONE COMMUNICATION SYSTEMS (71) We, DANRAY, INt. I corporation organized tinder the laws of the State of Texas, United States of America, of 1374() Neutron Road.Dallas, Tuxnss 7524(), United States of America, do hereby declare the invention. for which we pray that a patent may be granted to us, and the method by which it is to he performed. to be particularly described in and by the following statement: This invefltiou rclatcs to telephone communication systems. 'ind more particularly to stored program cont rolled exchanges; it is particularlv hut not exclusively applicable to private automatic branch exchanges (PABX's).

According to the present invention, a telephone exchange comprises, for each telephone line served thereby, means for converting analogue signals from the line to pulse width modulated digital signals, which signals are available on a first digital signal line associated with that telephone line, and means for converting pulse width modulated digital signals appearing on a second digital signal line associated with the telephone line to corresponding analogue signals, which are transmitted to the telephone line, the exchange also comprising a solid state switching matrix having a plurality of imputs connected one to each of the first digital signal lines, and a plurality of outputs connected one to each of the second digital signal lines, the switching matrix being arranged to transmit digital signals, without multiplexing, from any of its inputs to a selected one of its outputs, in accordance with switching command signals received from a stored program controlled processor, which processor in turn received information concerning the required connections from the telephones connected to the telephone lines, over signalling paths which do not pass through the switching matrix.

Because the digital signals being transmitted by the switching matrix are not multiplexed, a considerable bandwidth is available for these signals, and therefore the fact that pulse-width modulation is less economica! in its use of bandwidth than other forms of modulation is not important. On the other hand, however, the modulators and demodulators for use with pulse-width modulation can be relatively simple. Thus, the invention may in some cases permit an overall simplification of a telephone exchange, in comparison with an exchange which takes advantage of the fast switching capacity of a solid state switching matrix to handle signals in time-division multiplex.

The invention may be carried into practice in various ways. but one specific embodiment will now be described by way of example, with reference to the accompanying drawings, of which: Figure I is an overall block diagram of a PABX telephone communication system embodying the present invention; Figure 2 is a diagrammatic representation of the switch matrix of the system depicted in Figure 1; Figure 3 is a coordinate data representation of the switches or crosspoints of each switching module of the input, output, and intermediate switch stages of the switch matrix illustrated in Figure 2 along with the representation of the relationship between crosspoints of each state; Figures 4A, 4B and4C are graphic summaries of the computerized technique for the selection of specified crosspoints in the intermediate stage of the switching matrix;; Figure 5 is a flow chart denoting the programmed steps utilized in effecting the required communication path interconnection; Figure 6 is an illustrative example of a matrix interface for responding to output signals from the central processor and converting these into signals usable by the switch matrix for selectively closing crosspoints therein; Figure 7 is a diagrammatic representation of the parts of the switch matrix concerned with routing signals through the input stage of the matrix illustrated in Figure 2; Figure 8 is a diagrammatic representation of the parts of the switch matrix concerned with routing signals through the output stage of the matrix illustrated in Figure 2, and thereafter through digital-to-analogue converters; Figure 9 is a schematic logic diagram representation of a typical example of a digital switch utilized in the system shown in Figures 7 and 8;; Figure 10 is a schematic illustration of a preferred form of A to D converter; and Figure 11 is a schematic illustration of a preferred form of D to A converter.

Referring initially to Figure 1, private automatic branch exchange system (designated generally by the reference numeral 1) comprises a plurality of telephone sets or stations 10 having voice communication channels 11 respectively coupled by way of analog to digital (A to D) converters 200 and digital to analog (D to A) converters 500 to a switching matrix 12. For convenience in describing the invention herein. two such station (designated 10 and 10') have been illustrated in the drawing. it being understood that many such stations may be so coupled, the only limit on the number of stations being the available switching paths in the matrix 12.Each voice communication channels 11 actuallv functionallv comprises input and output channels 1 lea and lib, respectively coupled by way of analog to digital converter 200 and digital to analog converter 500 with the mouthpiece and earpiece of the telephone sets 10 respectively transmitting and receiving the voice signals between coupled stations. as subsequently described.

Each of the sets 10 has conventional pushbutton signal generators 15 for selectively producing signals which, in combination with a line scanner 20, will implement the command or control signals for the operation subsequently described. Alternatively. signals may be generated from a rotary dial mechanism of the type known in the art, the pushbutton digital signal generactors however. allowing a flexibility of signalling not conveniently available from rotary dail pulsing.

The telephone sets 10 may also include such additional features as signalling lamps and pushbuttons to allow such conventional operations as line-hold. busy signal indication, ringing indication. etc.. these additional operations not being essential to the present invention.

A switching matrix or network 12 is provided for selectively coupling, upon command, the input and output lines of the voice communication channels 11 of one or more of the stations 10 to establish a communication link therebetween thereby routing the voice or audio signals between the phones 10 desiring to "talk" to one another. The matrix 12, the internal details of which are subsequently described, comprise a plurality of alternative non-blocking switching paths, the establishment of switch connections through the matrix providing the necessary coupling or link between the voice communication channels of the calling and called stations.

The control of this switching matrix, and specifically the making and breaking of communication links therethrough, is computer controlled, the instructions for such communication linkage being initially implemented by appropriate control or instruction signals received from the stations 10. In this regard, each of the stations 10 has a channel 14 separate and distinct from the communication channels 11 for transmitting. on a substantially unlimited capability, signals representing the station being "cal led", each of the channels 14 being coupled intermediate the signal or tone generating means 15 and a continuous line scanner 20.

While the channels 14 are represented as separate control channels respectively coupled between the different subscriber stations 10 and the continuous line scanner 20, it is also possible to utilize conventional multiplexing and have a common control data bus to which each of the signal generators of the respective instruments 10 is coupled. Furthermore. the tone signals being generated may. if desired, travel along the same wires transmitting the audio signals of the phones 10 for a given distance, and then be separated for inputting to the scanner 20, the audio signals being inputted to the converter 200.

The principal function of the continuous line scanner 20 is to continually monitor each of the signal channels 14 coupled thereto to determine the presence of a "called party" instruction or control signal thereon identifying with which station the calling party wishes to communicate, to determine the "on-hook" or "off-hook" status of the "called-party", and thereafter to generate a coded signal on line 25 which designates (1) the identification of the "call ing-party-- (which may be obtained, for example. be sensing, identifying, and registering the particular input channel 14, and therefore which phone set 10, from which the "called-party" signal was received), and (2) the identification of the "called-party".

Various types of conventional digital apparatus may be employed for the line scanner 20 which continuously monitors the input channels 14 and generates digital words representing the "calling" and "called" party, as indicated.

Coupled to the output of the line scanner 20 is a stored program controlled computer basically including a central processing unit 21 having an associated memory unit 22, the output of central processor 21 being coupled through matrix interface means 71 to the switch matrix 12. The primary function of the central processor 21 is to respond to the instruction or control signals received from the continuous line scanner 20 over line 25 to establish the required communication link between the input and output paths of the communication channels 11 of the respective stations. Accordingly, the processor will have included within or associated therewith conventional encoders or decoders, standard registers and computing cle- ments, and suitable interface equipment.

An example of a stored program controlled computer, including a memory, which may be utilized in a preferred embodiment of this invention is the Nova 1200 computer manufactured by Data General Corporation, Southboro, Massachusetts. A description of this computer and the input/output requirements for interfacing therewith, may be found in the manual "How to Use the Nova Computers". published by Data General Corporation (DG NM-6, October, 1972).

Memory unit 22 has stored therein: (1) coded information corresponding to possible alternative switching paths within the matrix 12 defined by switching crosspoints therein which, if closed, will effect the required coupling between voice communication channels; (2) current status of availability of specific crosspoints; and (3) instruction programs for controlling and directing the operation of the central processor in response to the stored information and incoming control signals from line scanner 20.

The functions of central processor 21 and memory unit 22 can be briefly summarized as follows: (1) The generation of address signals to the matrix interface means 71 identifying the specific crosspoints of the matrix 12 to be "closed" to establish the desired linkage between the voice communication channels 11 of the specific "calling" and "called" stations 10 as identified by the instruction signal on lead 25.

(2) The continuing evaluation of data corresponding to the switch matrix 12 to determine the status of switch crosspoints therein. and the generation of information about defective crosspoints: (3) The control of special functions between internal stations 10: (4) The control and establishment of intercommunication between internal stations 10 and an external central office exchange; and (5) The control and generation of additional functions external to the system 1.

The matrix interface means 71 has its input coupled to the output from central processor 21 over leads 100 and a plurality of outputs respectively coupled over leads 7() to the control terminals of switches (identified by reference numeral 60 in Figure 2) disposed at each crosspoint of the matrix 12. The function of interface 71 is to accept the digitially encoded signals on leads 100 from the processor 21 (identifying crosspoints of the matrix 12 which are to be "closed") and convert these address signals to appropriate logic level signals on leads 70 (high or low) applied to the control terminals of switches 60 for closing those switches across the crosspoints to be "closed".Various types of conventional digital circuitry presently known in the art may be employed to construct the interface means 71 to effect the required conversion and steer the logic signals to the appropriate switch input, as will be described in detail hereinafter.

Establishment of comntu,lication between stations The essential function of the processing unit 21 is to control the switch matrix 12 to establish the desired linkage between the communication paths 11. In accordance with the instructions from a program, the details of which will be subsequently described, stored in the memory 22, the central processor 21 effects such control in the following sequential manner: (a) it receives the "connect" instructions on lead 25 from continuous line scanner 20, verifies its completeness. and determines from such instructions with stations 10 are to be coupled; (b) it determines the availability of the "called" party; (c) by analyzing data stored within the memory unit 22, it determines the optimum switching path within the matrix to effect the required coupling; (d) by reference to memory, it monitors information corresponding to the status of the switches 60 at the crosspoints within the matrix 12 defining the optimum switching path and determines the availability of these crosspoints; (e) if the crosspoints are not available, it repeats steps (c) and (d) until an available path (crosspoints) is found; and (f) it deploys the digital encoded address signal to the switch matrix interface means 71 to close the specific switches at these crosspoints to establish the necessary communication path therein. As is subsequently described, the processor 21 as a last resort, upon determination that one or more of the required crosspoints (or their subsitutes) are unavailable to effect the needed communi cation link, will reallocate crosspoints for existing communication paths in order to free up the desired path or link.

Diagnostic evaluation The central processor 21, in combination with the memory unit 22, also provides a diagnostic function to determine and identify the location of defective solid state switch crosspoints within the matrix 12.

Specifically, the existence of a defective switch is directed into memory 22, the location of the defective switch crosspoints thereby being stored therein. To effect the coupling of the communication channels just described, the central processor 21, under control of a separate stored program detailed hereinafter, then takes into account these defective switch points, establishes a path rerouting in view of these defective switches, and generates output information about the location of these defective switch points.

Special functions Under the control and direction of stored programs within the memory 22, the central processor 21 effects special functions, such as redirecting incoming calls to other stations in the event the called station line is busy, establishing priority breaks into communication links, etc. Each of these functions, as well as additional special functions, may be initiated in response to special code signals generated by the pulse means 15 of the instruments 10.

Intercommunication The private automatic branch exchange not only provides a wide range of communications between subscriber stations 10 within the system 1. but also provides intercommunication between these stations and phone stations coupled to an external central office exchange 30. This interconnection is effected by way of suitable interface equipment 31 coupled to the processor 21 and matrix 12.

External functions Additional auxiliary equipment may be coupled if desired, to the system 1, all under control of the processor 21. For example, an attendent monitor 32 may be utilized to monitor or override the automatic controls; a traffic monitor 33 may be employed to record the transactions carried out by the system: or other miscellaneous equipment 34 such as Teletypes (Registered Trade Mark). printers. etc, may be coupled thereto.

Referring now to Figure 2, the switch matrix 12 for establishing voice communica tion link or transfer of signals between selected telephones 10 is depicted as comprising an input switch stage (consisting of input switch modules Il, I2,...IN) and an output stage (consisting of output switch modules Oi, 02,...0N). The inputs to each of the input switch modules Il, I2, etc, are respectively coupled by way of A to D converters 200 to the mouthpieces of the different telephone sets 10 by way of the input channels 11a, which telephone sets have been designated for convenience as 1, 2, 3, etc.In similar manner, the outputs from output switch modules Oi, 02, etc, are respectively coupled by way of D to A converters 500 and output channels 11b to the earpieces of different telephone sets 10, which sets for convenience have also been identified as 1, 2, 3, etc.

In addition to the input and output stages, the matrix 12 includes an odd number, in this case one, of intermediate stages comprising a plurality of switch modules appropriately interconnected with the switch modules of the input and output stage. The intermediate or secondary stage is illustrated in Figure 2 as having switch modules S1, S2 ... SN=5- In the present example, each of the input switch modules I,, I., etc., and the output switch modules Oi, 02, etc., have the same number of inputs as outputs, it being understood that this is not essential to the operation of the system.Each input and output switch module has coupled thereto the input or output channels to five telephone sets, the switch matrix illustrated in Figure 2 thus being capable of establishing a communication link between the mouthpiece and earpiece of 25 telephone sets.

The outputs from any one input switch module are coupled to the inputs of different secondary switch modules, and the inputs to any one output switch module are coupled to the outputs of different secondary switch modules. Thus, as illustrated in Figure 2, the number of secondary switch modules S,, S., etc., is equal to the number of outputs from (or inputs to) the input or output modules. In the illustrated example, there are thus five secondary switch modules S,, S2, S3, S4. and S5, in the intermediate or secondary stage of the matrix 12. As a consequence of this interconnection, the number of available communication paths between a mouthpiece of any one given telephone set to the earpiece of any other given telephone set is equal to the number of secondary switches which, in the present example. is five.

Each switch module has an array of switches 60, selected one of which, when closed, selectively couple a respective input bus to an output bus. Such a switch array is illustrated with respect to input switch mod ule I. it being understood that a similar type switch array would be included within each of the input, secondary, and output stage switch modules. As subsequently described in greater detail, to establish a communication link between the input channel 11a coupled to one phone to the output channel tib of any of other phone requires the closing of a specified switch 60 in the input stage, a specified switch 60 in the secondary stage, and a specificed switch 60 in the output stage.In the specific illustrated example, there are five separate sets (a set composed of the three so specified switches) of switches which may be so closed (thus defining five separate paths) to establish the voice communication link.

Each of the switch modules may be represented as illustrated in Figure 3 by a matrix or array of circles 60 respectively representing a potential interconnection (crosspoint) between an input and output bus. The interconnection of the earpiece of any one telephone set to the mouthpiece of any other telephone set (including itself) is thus provided by a path (or a number of alternative paths) defined by a set of crosspoints respectively located in the input, secondary, and output stages, there being five such paths (five such sets) in the illustrated example of Figure 2.

Each crosspoint can be identified by coordinate data respectively representing the number of the switch module, its position in the horizontal direction, and its position in the vertical direction. Thus, when the input switch module is the first one (or I,) of the input stage, the crosspoint designated with an "x", for example, may be identified by the designation (1, 4, 2), the closing of this particular switch or crosspoint indicating the interconnection between the fourth input bus and second output bus.

Similarly, when the switch module is the first one (or O,) of the output stage the crosspoint designated by the "x" may be identified by the designation (1, 3, 3), indicating the interconnection between the third input and third output bus thereof.

The determination of the specific crosspoints (switches 60) necessary to establish one of the desired communication paths is effected by the central processor 21 of a general purpose computer (and an associated utilization storage device or memory 22) in response to command signals received from the line scanner 20 which indicate the particular phones (channels) to be interconnected. Such determination is now described.

Stored within the memory 22 is not only the set of instructions to the processor 21 necessary to carry out the operation as subsequently described, but also the status (availability) of a crosspoint (switch in use or defective). As previously mentioned each of the respective outputs from any one input switch module is respectively coupled to a corresponding input to a different one of the secondary switch modules; and respective inputs to any one output switch module are respectively coupled to corresponding outputs from the different secondary switch modules.

Specifically, and with reference to Figure 2, all of the output terminals A-E of the input switch modules I, - Is are connected to the corresponding lettered input terminals of the secondary switch modules; and the input terminals A-E to the output switch modules are also coupled to the corresponding lettered output terminals of the secondary switch modules. Additionally, it is noted that the first or A output terminal of the first input switch module I, is coupled to the first input terminal of the secondary switch module Si and that the first or A input terminal of the first output stage switch module O, is coupled to the first output terminal from switch module Sl.

As a consequence of this interconnection scheme of the matrix 12, any set of three crosspoints necessary to effect the required communication path respectively located in the input secondary, and output stages, are specifically related to one another in the manner illustrated in Figure 3. Accordingly, the coordinates of the secondary switch crosspoint are defined as (Ksec, Knout, Kin) where K,,,, and Kj" respectively represent the output switch module to which the desired output channel is coupled and the input switch module to which the desired input channel is coupled.Moreover, the horizontal position coordinate of the output switch module crosspoint is always equal to the vertical position coordinate of the input switch module crosspoint which in themselves will always be equal to the number of the particular second switch module Ksec being used.

The foregoing relationship can be utilized to define the sequence of steps by which the required communication paths are established. Since the sequence will be effected under the control of a set of stored program instructions, the method or program algorithm basically includes the following steps in response to the receipt of the command signals on lead 25 indicating which input and output channels are to be coupled: (a) initially determining which input stage switching module and which output stage switching module are required to effect the coupling of the selected input and output channels; (b) selecting an available crosspoint of the intermediate or secondary stage: (c) selecting or determining the necessary crosspoints of the input and output stage switch modules as a consequence of the selection of the particular secondary switch module crosspoint; and (d) generat ing the address signals defining, in digital code, the set of three chosen crosspoints to close the required switches 60 and establish the communication path.

Stored within the memory 22 are digital words, whose position within the memory and coding are indicative of the specific crosspoints of the switch modules and their status. For example, the rows of crosspoints of the primary switch modules may be defined by digital words, the location of a bit in the word representing the input terminal coupled thereto, and the status of the bit (0 or I) indicating the open or closed nature of the switch. The relative locations of the rows to one another (i.e. the location in memorv) can then define the particular output terminals. Similar coding may be effected for the crosspoints of the secondary and output modules, the digital words for each of these stages being disposed at different respective locations in memory.

The initial selection of which input and output switch modules are required may be effected bv calculntion initiated by the input command signals and constrained or determined bv the coded information. Alternatively. coded data indicating which input channels and which output channels are respectively coupled to which input and output modules mny be stored within the memory 22 and retrived in response to the input command signals.

The crosspoint of the intermediate or secondary stage may be selected from a total possible number of available secondary state crosspoints equal to the number of switch modules in the secondary stage. For the example illustrated in Figure 2. there are thus five possible acceptable crosspoints to choose from depending upon their respective availabilitv. The availability of any particular crosspoint switch requires that no other crosspoint be in use (either "closed" as a result of a previous command or a failure) in either the row or column in which the desired crosspoint is located. For the interconnection scheme shown in Figure 2, this availability will exist as long as neither the required input module nor the required output module is already connected to the secondary module of interest.This availabil ity can be detected by the computer interrogating the memory. preferably on a sequential basis. as to the status of each acceptable secondary stage crosspoint for each of the secondary modules.

This technique can be summarized by reference now to Figures 4A through 4C wherein there is depicted a chart of the combined status of secondary stage crosspoints for each secondary module. The abscissa of each of the graphs of Figures 4A-4C correspond to the required input modules, and the ordinate of the chart corresponds to the secondary stage modules. Accordingly, assume that it is desired to couple the input channel of telephone set 1 to the output channel of telephone set 5.

Since the telephone set 1 is coupled to, and thus requires the use of the input module II, the number 1 (which represents the required output module dictated by telephone set 5) may be placed (as one solution) within the northwesternmost block sine there is no other number in that block (thus indicating that the required input module is not already connected to the first secondary module) and since there is no other number "1" in the same row (thus indicating that the required output module is not already connected to the first secondary module). Thus, the chosen secondary crosspoint is defined as (1, 1, 1).

Since the input channel to one telephone set 5 should normally also be coupled to the output channel to telephone set 1, the determination of the required secondary crosspoint will now be effected for that condition. It is noted that both the first input module and first output module are still required to effect this coupling. However, the northwesternmost block of the graph Figure 4A is already taken (thus indicating the existing connection of input Ii module of the first secondary module) so the number "1" is inserted, for example, in the appropriate block in the second row, as illustrated in Figure 4A.

Assume now that it is desired to couple the input channel from telephone set 10 to the output channel of telephone set 3, thus requiring the use of the second input module 12 and the first output module 01. With reference to the chart of Figure 4A, it is noted that the block represented by the abscissa 12 and the ordinate Sl is available, thus indicating that the required input module 12 is not already connected to the first secondary module. However, the number "1" (indicating the required output module 0,) cannot be placed in this block since the same number already exists in the first (and second) secondary module, thus requiring the placement of the number "1" in another row, for example the third row (representing the third secondary module), as illustrated in Figure 4B. The secondary stage crosspoint is thus defined as (3, 2, 1).

In similar manner, the secondary stage crosspoints may be chosen to effect the required interconnection between the input and output channels of all telephone sets, one typical solution being summarized in Figure 4C. Such a solution allows the utilization of all necessary crosspoints in the secondary stage to effect the required coupling between telephone sets without blocking.

The computerized selection of the parti cular secondary stage crosspoints can be effected preferably in one of two manners.

First, the respective blocks of the chart represented in Figures 4A-4C may represent individual storage locations in memory 22 and the numbers 1 through 5 (corresponding to the required output modules) can be represented by multibit digital words directed into respective locations in storage.

Thus, under control of the stored instructions, the respective secondary stage crosspoints may be selected. Alternatively, by utilizing the digital word coding previously described, and storing the digital words representing the crosspoints already in use, a determination can be made each time a new coupling is to be effected as to the availability of a particular crosspoint in the desired secondary stage module.

After each secondary stage crosspoint is defined, the particular crosspoint of the required input and output modules can be calculated since the "M" coordinate point shown in Figure 3 is now defined by the number of the secondary module containing the selected secondary crosspoint. Thereafter, and at an appropriate time, a coded signal on leads 100 may be generated by the processor 21, which signal corresponds to the defined set of three crosspoints and matrix interface means 71 may convert this coded signal on leads 100 into a plurality of high (1) logic signals in leads 70 directed to the control input of the switches 60 defined by the crosspoints, the switches 60 thereby being closed to effect the required communication coupling.

In many instances, a solution to effect the coupling of telephone sets is not possible without the necessity of reallocation or rearrangement of any set of crosspoints.

Sometimes a defective crosspoint prevents the desired connection. Sometimes enough crosspoints are being utilized to block the setting up of the desired connection. In such instances, and under instructions of the stored program within the memory 22, the processor 21 will compute a new solution, and rearrange or reallocate another secondary crosspoint for an existing connection, thus freeing up the desired secondary crosspoint for the new connection being sought.

In effect, this means that as a consequence of the reallocation. a different, but acceptable, communication path is defined to maintain the existing connection, and the new communication path may be established.

Referring now to Figure 5. there is depicted a flow chart of the algorithm in accordance with a preferred embodiment of the present invention. Accordingly, the box 1000 indicates the generation of the command signals on lead 25 indicating the desired phones to be coupled. and the process box 101 indicates the determination of the required input and output modules pursuant to said command signals. The next step is denoted by box 102 whereby the secondary count is initially set at Ksec=l (first secondary module). The decision box 103 determines of the required input module is connected to the chosen secondary module. If the answer is "yes" with respect to Kse=1, then the Ksec is incremented to Used=2 =3, etc, as represented by process box 106, until the answer from decision box 103 is "no".Once a "no" answer is received, the next step is determined by the decision box 104 as to whether the required output module is connected to the chosen secondary module K Sec represented by the decision pursuant to box 103. If the answer is "yes", the output is directed back to the incrementing operation represented by the process box 106, and the cycle is repeated until a "no" is received at the output of box 104 (thus indicating the available secondary module). The process box 105 then represents the difinition of the particular crosspoints for the secondary, input, and output modules, a "no" answer from the failure inquiry process box 108 thereafter activating the switches 60 defined by the chosen crosspoints. as represented by the process box 110.The decision box 107 represents the determination of the availability of any secondary module Used, a "yes" answer therefrom indicating the necessity for reallocation, as previously described, as represented by the process box 109.

Referring now to Figures 6 through 11, Figure 6 shows the matrix interface 71 which accepts an address from central process 21 and converts this address into a form suitable for use by switch matrix 12 to close selected crosspoints therein. The output of the exemplary processor 21 is designed to interface with TTL circuitry requiring three volt logic levels. Switch matrix 12 is illustratively comprised of CMOS integrated circuit elements operating with 12 volt logic levels.

One of the functions of matrix interface 71 is therefore to provide the appropriate logic level conversions. Additionally. the exemplary processor 21 utilizes a single I/O bus which is time shared between the input and output signals. Therefore. a circuit such as that designated 601 is utilized to provide both the level conversion and the inputoutput routing. Since the exemplary processor 21 utilizes a 16 bit data word, 16 of the circuits referenced as 601 are required, one for each of the data bits. In order to simplify the drawing, the other 15 level shifters are shown below the one designated 601 and are referenced with the letter A in the lower right corners thereof. It should be noted that circuits 601 may be connected to other circuits, not shown, which send information back to processor 21.

At this time a brief discussion of the illustrative addressing scheme for addressing selected crosspoints is in order. Ten of the 16 bits are utilized to select one of the switch modules of matrix 12. Three of the bits are used to select a column of the selected module and the last three bits are utilized to select a row of the selected module, thereby fully defining a single crosspoint in switch matrix 12. In the illustrative embodiment where switch matrix 12 comprises three stages of switching, central processor 21 would send through matrix interface 71 three sets of address bits to define the three crosspoints in the input, intermediate, and output stages, respectively, which must be closed to establish a communication link through switch matrix 12.

After the level conversion by circuits 601, the 16 data bits are strobed into the temporary storage unit 605 which may be comprised of an array of any of the known types of storgae elements, such as flip-flops. Returning now to Figure 6, it is seen that the six bits defining a crosspoint within a switch module go directly from storage 605 to matrix 12 over leads designated X, Y. Z. P, Q. and R. The 10 bits designating a particular switch module of matrix 12 go from storage 605 to matrix 12 after being further decoded. Two sets of two bits each are decoded to one of four signals and two sets of three bits are decoded to one of eight signals, the last group of three bits being strobed and decoded after a delav from the computer output strobe signal so as to ensure that all bits in storage 605 have had an opportunity to change state.The four sets of decoded switch module address bits then go through switch module selection tree 610 which provides one lead for each switch module of matrix 12, up to a maximum of 1024.

Turning now to Figure 7, a plurality of analog to digital converters 200 are respectively disposed in each channel 1 lea and coupled between the phone sets (1. 2. 3, etc.) and the inputs to the switch matrix 12.

Each converter is effective to translate the respective audion analog signals inputted thereto to digital pulse signals whose widths are modulated in accordance with the analog information. Each of the outputs from the A to D converters 200 are then respectively coupled to the corresponding inputs of the input switch modules I. 1.. etc in the same manner as previously described with respect to Figure 2. In the illustrated example of Figure 7 there are eight such inputs to each of eight input switch modules, with specific reference, for convenience of de scription. being to one such input switch modules Ii. Similarly, for convenience, Figure 8 specifically depicts one of the eight output switch modules Os coupled to telephone sets 57, 58,... 64.

Coupled to the outputs from matrix interface means 71 (Figure 7) are a plurality of integrated circuit digital switching networks 301-308 for selectively coupling one of the input buses of module I, to an output bus therefrom. Accordingly, each input bus would be respectively coupled to an input terminal 1', 2', 3', etc. (corresponding to phone channels 1, 2, 3, etc.) of the networks 301-308, the output terminals of each switching network 301-308 being respectively coupled to the output terminals A-H. The leads P, Q, R pass through a 3 bit to 1 out of 8 row select decoder to enable only one of the switching networks 301-308, in accordance with which of the leads R1-R8 is logically high. The leads X, Y, Z go to all networks 301-308 to select one of the eight crosspoints within the enabled network, as will be described with reference to Figure 9.

While various types of circuitry may be employed for each of the switching networks 301, 302, 303, etc., one such type is depicted in Figure 9. As can be easily seen from a comparison of the network illustrated in Figure 9 and its use in the manner shown in Figure 7 (and 8), each switching network 301, 302, etc. functionally provides, in effect, an entire row of switches 60 (Figure 2) for the input switch module Ii.

Thus. in the example of an 8x8 switch module, network 301 provides the first row of switches 60, network 302 provides the second row, and digital switching network 308 would provide the last row thereof.

The "closing" of a crosspoint of the input switch module I, is now described. Accordingly. assume that the digitized audio signal from phone 3 is to be coupled to output B from module Ii. To effect such coupling, the address of the crosspoint to be "closed" in the input stage is that one at the intersection of the third input bus and the second output bus of switch module Ii. The signal on leads 100 would contain this address, and the matrix interface 71 will effect the required conversion of the digitally encoded address signal. The lead SELECT I, would be high: the leads P. Q, R. would be O, 0, 1 to indicate digital switching network 302; and the leads X, Y, Z would be O, 1, O to indicate the third input terminal. As a result, the input terminal 3' of switching network 302 would be connected with the output terminal therefrom, and the digitized audio signal would be transmitted from the third channel input bus to terminal B. In similar manner, the digitized audio signals can be switchably rounted through the intermediate and output stages of the switch matrix 12. Thereafter. the digital to analog converters 500 respectively reconvert the pulse width mod ulated digitized audio signals at the output of matrix 12 to analog signals which are coupled to the appropriate earpieces of the telephone stations 10.

In the present example there is no actual "opening" of a crosspoint. When a connection between telephone stations is to be terminated, processor 21 causes the appropriate stations to be disconnected from matrix 12, by means not shown. The crosspoints within matrix 12 defining that connection remain enabled until the establishment of a subsequent connection requires them to be opened; also, processor 21 alters memory 22 to remove those crosspoints from a "reversed" list in the memory so that those crosspoints may be reassigned for a subsequent connection.

Referring now to Figure 10, a preferred form of the analog to digital converter 200 is depicted. Accordingly, the converter comprises three high input impedance, large bandwidth inverting amplifier stages 201, 202, and 203, a capacitor 204 providing positive feedback from the output of amplifier stage 202 to the input of amplifier stage 201, a resistor 205 providing negative feedback from the output of amplifier 203 to the input of amplifier 201. A positive supply voltage preferably one-half the value of the voltage supply to the amplifiers is coupled across resistor 206 and capacitor 207.

The entire network 200 oscillates to generate a continuous square wave signal at the output terminal 210. When the analog audio signal is applied to the input terminal 211, the resulting variations in the magnitude thereof will alter the widths of each of the square wave pulses in accordance with the time varying analog signal.

Specifically, as the amplitude of the analog signal varies, the average voltage across capacitor 207 is correspondingly changed, which change is amplified by the stages 201, 202, and 203, to produce a corresponding change, in the opposite direction. at the output terminal 210. This varying voltage at the output terminal 210 is then coupled by way of resistor 205 to the high side of capacitor 207, readjusting the voltage across capacitor 207 toward its original value.

Thus, the variations in the magnitude of the analog signal will produce corresponding variations in the average value of the output voltage at terminal 210 which is reflected by the varying pulse widths of the digital signal therefrom. Thus, the output signal at 210 is pulse width modulated in accordance with the analog signal information.

Referring now to Figure 11, a preferred form of the D to A converter 500 comprises a suitable R-C filter network comprising resistor 501 and capacitor 502, the function of the converter 500 being to recover the average value of the voltage present in the digitized audio signal being inputted thereto and generate an analog signal representative thereof. This resulting analog signal thus contains the same audio information that was inputted to the switch matrix, and is coupled to the earphone of the associated telephone set.

WHAT WE CLAIM IS: 1. A telephone exchange comprising, for each telephone line served thereby, means for converting analogue signals from the line to pulse width modulated digital signals, which signals are available on a first digital signal line associated with that telephone line, and means for converting pulse width modulated digital signals appearing on a second digital signal line associated with the telephone line to corresponding analogue signals, which are transmitted to the telephone line, the exchange also comprising a solid state switching matrix having a plurality of inputs connected one to each of the first digital signal lines, and a plurality of outputs connected one to each of the second digital signal lines, the switching matrix being arranged to transmit digital signals, without multiplexing, from any of its inputs to be selected one of its outputs, in accordance with switching command signals received from a stored program controlled processor, which processor in turn receives information concerning the required connections from the telephones connected to the telephone lines, over signalling paths which do not pass through the switching matrix.

2. A telephone exchange as claimed in Claim 1, in which the processor is arranged to produce switching command signals which control the switching matrix to transmit digital signals both from the first digital signal line associated with a calling telephone line to the second digital signal line associated with a called telephone line, and from the first digital signal line associated with the called telephone line to the second digital signal line associated with the calling telephone line.

3. A telephone exchange as claimed in Claim 1 or Claim 2 in which each means for converting analogue signal to digital signals comprises a cascaded chain of two amply fiers. one amplifier being non-inverting, and being provided with a positive feedback path to provide a Schmitt action, while the other amplifier is an inverting amplifier, a capacitor being connected between the input of the amplifier chain and ground, while a negative feedback path is connected from the output to the input of the amplifier chain.

4. A telephone exchange as claimed in Claim 1 or Claim 2 or Claim 3 in which each means for convering digital signals to analogue signals comprises a low-pass R-C

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (7)

**WARNING** start of CLMS field may overlap end of DESC **. ulated digitized audio signals at the output of matrix 12 to analog signals which are coupled to the appropriate earpieces of the telephone stations 10. In the present example there is no actual "opening" of a crosspoint. When a connection between telephone stations is to be terminated, processor 21 causes the appropriate stations to be disconnected from matrix 12, by means not shown. The crosspoints within matrix 12 defining that connection remain enabled until the establishment of a subsequent connection requires them to be opened; also, processor 21 alters memory 22 to remove those crosspoints from a "reversed" list in the memory so that those crosspoints may be reassigned for a subsequent connection. Referring now to Figure 10, a preferred form of the analog to digital converter 200 is depicted. Accordingly, the converter comprises three high input impedance, large bandwidth inverting amplifier stages 201, 202, and 203, a capacitor 204 providing positive feedback from the output of amplifier stage 202 to the input of amplifier stage 201, a resistor 205 providing negative feedback from the output of amplifier 203 to the input of amplifier 201. A positive supply voltage preferably one-half the value of the voltage supply to the amplifiers is coupled across resistor 206 and capacitor 207. The entire network 200 oscillates to generate a continuous square wave signal at the output terminal 210. When the analog audio signal is applied to the input terminal 211, the resulting variations in the magnitude thereof will alter the widths of each of the square wave pulses in accordance with the time varying analog signal. Specifically, as the amplitude of the analog signal varies, the average voltage across capacitor 207 is correspondingly changed, which change is amplified by the stages 201, 202, and 203, to produce a corresponding change, in the opposite direction. at the output terminal 210. This varying voltage at the output terminal 210 is then coupled by way of resistor 205 to the high side of capacitor 207, readjusting the voltage across capacitor 207 toward its original value. Thus, the variations in the magnitude of the analog signal will produce corresponding variations in the average value of the output voltage at terminal 210 which is reflected by the varying pulse widths of the digital signal therefrom. Thus, the output signal at 210 is pulse width modulated in accordance with the analog signal information. Referring now to Figure 11, a preferred form of the D to A converter 500 comprises a suitable R-C filter network comprising resistor 501 and capacitor 502, the function of the converter 500 being to recover the average value of the voltage present in the digitized audio signal being inputted thereto and generate an analog signal representative thereof. This resulting analog signal thus contains the same audio information that was inputted to the switch matrix, and is coupled to the earphone of the associated telephone set. WHAT WE CLAIM IS:

1. A telephone exchange comprising, for each telephone line served thereby, means for converting analogue signals from the line to pulse width modulated digital signals, which signals are available on a first digital signal line associated with that telephone line, and means for converting pulse width modulated digital signals appearing on a second digital signal line associated with the telephone line to corresponding analogue signals, which are transmitted to the telephone line, the exchange also comprising a solid state switching matrix having a plurality of inputs connected one to each of the first digital signal lines, and a plurality of outputs connected one to each of the second digital signal lines, the switching matrix being arranged to transmit digital signals, without multiplexing, from any of its inputs to be selected one of its outputs, in accordance with switching command signals received from a stored program controlled processor, which processor in turn receives information concerning the required connections from the telephones connected to the telephone lines, over signalling paths which do not pass through the switching matrix.

2. A telephone exchange as claimed in Claim 1, in which the processor is arranged to produce switching command signals which control the switching matrix to transmit digital signals both from the first digital signal line associated with a calling telephone line to the second digital signal line associated with a called telephone line, and from the first digital signal line associated with the called telephone line to the second digital signal line associated with the calling telephone line.

3. A telephone exchange as claimed in Claim 1 or Claim 2 in which each means for converting analogue signal to digital signals comprises a cascaded chain of two amply fiers. one amplifier being non-inverting, and being provided with a positive feedback path to provide a Schmitt action, while the other amplifier is an inverting amplifier, a capacitor being connected between the input of the amplifier chain and ground, while a negative feedback path is connected from the output to the input of the amplifier chain.

4. A telephone exchange as claimed in Claim 1 or Claim 2 or Claim 3 in which each means for convering digital signals to analogue signals comprises a low-pass R-C

filter.

5. A telephone exchange as claimed in any of the preceding Claims, in which the processor is arranged, if a switch of the switching matrix which would otherwise be selected by the switching command signals is unavailable, to produce a different set of switching command signals avoiding use of the unavailable switch.

6. A telephone exchange as claimed in any of the preceding claims, which is a private automatic branch exchange.

7. A telephone exchange substantially as herein described, with reference to the accompanying drawings.