US2457986A - Synchronization of time division multiplex communication system - Google Patents

Description

Jan. 4, 1949. r.1. o. EDsoN 2,457,986

SYNCHRONIZATION. OF TIME DIVISION MULTIPLEX COMMUNICATION SYSTEM 1 Filed Dec. 11, 1945 2 sheets-sheet 1 /NVE/v Tof? J 0 E DSON ATTORNEY Jan. 4, 1949. J. o. L-:DsoN 2,457,986

SYNCHRONIZATION OF ,TIME DIVISION MULTIPLEX COMMUNICATION` SYSTEM Filed Dec. l1, 1945 -2 Sheets-Sheet 2 N R0 mw NLE M0. J V B um wm Tlplv R R -Ears mm N @www WE s wm m. m l m mm u J| .v..o\k.

j df( ATTORNEY Patented Jan. 4, 1949 0F TIME DIVISION MULTIPLEX COMIMUNICATION SYSTEM James 0. Edson, Great Kills, N. Y., assignor to Bell Telephone Laboratories, Incorporated, New York, N. Y., a corporation of New York Application December 11, 1945, Serial No. 634,359

SYNCHRONIZATION 15 Claims.

This invention relates to time division multiplex, two-way radio communication systems and particularly to such systems for `communicating between relatively moving stations.

Time division multiplex, like frequency multiplex, can be used not only for multichannel com munication between two terminals but also for communication between a main station and a grou-p of substations in which case individual multiplex channels are utilized for communication with the respective substations. For twoway communication, dilerent time division channels may be assigned to the two directions of transmission or a combination of time division and frequency multiplex may be utilized in which communication to the. two directions is carried out on separate frequencies and time division multiplex is utilized to afford separate channels for each substation.

Particularly where a synchronizing pulse is utilized, no dimculty is encountered in communication from the main station to a group of subtations located at different distances from the main station, even where the location of the stations is continuously changing. However, communication from `the substations to `the main station creates a problem of avoiding interference between the signals arriving from the different substaticns. Transmission from each substation must be so timed that the signals in the several channels arrive at the main station in the proper time order to permit their separation. In the case of moving stations this timing must be continuously regulated.

An object of the invention is to provide time mission of signals therefrom. In particular, the I signals from the main station are modulated by a pilot current having a characteristic proportional to the time of arrival of the signal from the corresponding substation relative to its assigned arrival time. At the substation. the received pilot current is utilized to regulate the phasing of a timing wave that controls the transmission of signals therefrom.

These and other objects, features and aspects of the invention can be better understood by reierence to the following detailed description in connection with the drawing in which:

Figs. 1 and 2 are block schematic diagrams oi a main station and a substation respectively of a radio telephone system embodying the invention; A

Figs. 3 to 6 are explanatory diag-rams; and

Fig. 7 is a schematic diagram of a modification of a portion of the circuit of Fig. l.

The` invention is illustrated herein as applied to a time division multiplex system in which the signals are transmitted by pulse position modulation. However, it should be emphasized that it is not limited to this type of modulation but is equally applicable to any time division multiplex system irrespective of the type of modulation used.

Fig. 1 shows a fixed radio station for communicating with a group of mobile stations on the individual channels of a. time division multiplex. In such a system one frequency is used for each direction of transmission and transmission to the individual subsidiary stations is on a shared time or time division multiplex basis. Fig. 2 shows one of the mobile subsidiary radio stations. `The others would be identical except of course that the gate generator of each station would be designed to select the particular time channel assigned to that station.

The system herein shown utilizes the principle of pulse position modulation for transmitting the signal currents. In pulse position modulation the time of occurrence of a pulse with respect to its average time of occurrence is varied in accordance with the amplitude of the signal current. Such a system is readily adaptable to time division multiplex operation by interlacing several channel pulses so modulated. My oopending application Serial No. 559,354, filed October 19, 1944, discloses a complete pulse position modulation multiplex system. Reference thereto is made for the details of the terminal modulating and demodulating circuits for the present system. A similar system is described in an article in Electronic Industries for December 1945, volume 4, No. 12, beginning at page 82.

While the method of and systems for pulse position modulation are not new in the art, they have not at the present time found wide application. Accordingly, for the purpose of providing a more comprehensive and self-sufcient description for those not completely familiar with sucn prior art disclosures a brief descriptio-n of the over-all system and certain more detailed features of the particular type of system of my copending application will be given. However, reference is made to that application for circuit and operational details. It being understood that the present invention is not limited to that detailed type of system which is utilized herein only for the purpose of illustrating the operation of the present invention.

In the system of my copending application there is employed the method of pulse position or pulse phase modulation in accordance with Which uniform pulses of very short duration recurring at a rate at least twice the highest signal frequency are modulated in time position in accordance with the signal to be transmitted. For telephone communication Where the speech band may be restricted to the frequency range below about 3500 cycles per second, e. pulse recurrence rate of about 8000 cycles per second and a pulse length of about one microsecond is typical.

For the purpose of multiplex operation the period corresponding to the recurrence rate is divided into a number of equal intervals corresponding to the desired number of channels with an additional shorter interval for use in synchronizing. Thus for the eight-channel system of my copending application the recurrenceperiod of 125 microseconds is divided in eight channel periods of 15 microseconds each and a synchronizing period of 5 microseconds. The recul'- rence frequency is determined by a stable control oscillator from Whose output there are derived a train of short pulses, eight for each cycle, timed to occur normally at the mid-points of the channel periods and a longer pulse occurring in the synchronizing period and marking the beginning of each cycle or frame. This synchronizing or marker pulse may be considered as establishing a time datum against Which the time occurrence or position of each channel pulse is compared for the purpose of regenerating the modulating signal at the receiver. (In other systems of the prior art sometimes referred to as double pulse systems a marker or datum pulse is transmitted for each channel but the two types of system are essentially equivalent in their broader aspects.) At the transmitter the modulation of the channel pulses by the respective signals varies their time positions with respect to ytheir normal mid-period times. The degree of modulation is limited so that the pulse for any particular channel never moves outside the channel period, thus avoiding crosstall: between channels.

At the receiver the synchronizing pulse is readily separated from the channel pulses by virtue of its greater length and is employed to generate gating pulses coincident and coeXtensive with the respective channel periods. By means of such gating pulses the various channel pulses are separated. The selected position modulated channel pulses are converted into length modulated pulses, in particular lasting from the time of occurrence of the channel pulse to the end of the corresponding gate pulse. Such length modulated pulses contain the signal and any other modulations Which may be obtained by filtering out all components of the length modulated pulses outside of the modulating frequency range.

In the system of the present invention as will more readily appear later each substation receiver produces only one gating pulse corresponding to the channel period assigned to that substation but the generation of that gating pulse and the handling of the channel pulse selected thereby is the same as for the corresponding channel of the complete multiplex receiver of my copending application as used in the main station of the present system.

Other methods of detecting the modulations of lthe channel pulses are found in the prior art and are essentially applicable to the system of the present invention. However, for the purposes of illustration of its operation that method just described and found in my copending application has been chosen.

At the main or fixed station as shown in Fig. l there is provided a channel pulse generator and position modulator l0 in which the eight pulses corresponding to the eight channels are generated and modulated in their time position by separate signals. l'n addition to the eight channel pulses there is also transmitted a marker pulse for synchronizing. This marker pulse is produced by the marker pulse generator Il. The marker pulse is longer than the channel pulses in order that it may be separated from them at the receivers. In the particular system herein illustrated the time set aside for the marker pulse is somewhat shorter than the channel time allotted to each signal pulse. The pulse array is shown in Fig. 3 from Which it Will be observed that each frame comprises ay marker pulse 50 Vand eight channel pulses 5l to 53. This frame repeats itself regularly, providing a sample of the signal current of each channel at the rate of the frame recurrence. In Fig. 3 the channel pulses are shown in their mean position at which they occur when no signal current is present in the respective channels. When a signal current is present the corresponding pulse will vary in its position in either direction and in an amount depending upon the polarity and amplitude of the signal current. It is a recognized principle in such systems that the frame frequency which is the average rate of recurrence of each channel pulse should be at least twice the maximum frequency of the signal current transmitted.

An oscillator I2 supplies reference oscillations of the frame frequency that are used for controlling the channel pulsegeneratcr lll and the marker pulse generator H. The output of the oscillator l2 is also used for controlling the receiving equipment as will be described later.

Reference is made to my copending application for circuit and operating details of the channel pulse generator and position modulator lil the general functions of which have been described above. However, the general make-up and operation of this unit Will be described with particular reference to a typical channel, namely, channel l, the numbering corresponding to the time order of the channel in a frame beginning with the marker pulse lThe signal currents from the line l5 enter through a bandpass lter l5 which.

serves not lonly to separate the signal currents from any low frequency currents such as are used for ringing and to limit the modulating signal currents to a frequency band extending from approximately 200 to 3500 cycles per second but also to prevent the low frequency control currents in the circuit 35 (to be referred to later) from being impressed on the line I5. Both the signal current output of the filter I6 and the control currents from the circuit 35 after ampliiication are impressed on a pulse modulator where oscillator outputs.

they operate to cause the time modulation of the channel pulse. i

The pulse modulator receives short pulses at fixed instants in each multiplex frame and from this generates; a length modulated pulse `timed to end normally at the mid-point ofv the respective channel period but in the presence of modulating signal or control currents vary in length within the limits of the channel period under the influence of such modulating currents.` The abrupt ending of the length modulated pulse is caused to energize a pulse generating and shaping circuit for the production of the ultimate position modulated pulse. Such position :modulated pulses for all channels are of uniform shape and energy and since the pulses occur serially the shaping system may be shared by one `or more channels.

The short pulses supplied to each position modulator at xed intervals are derived from the master oscillator i2. This oscillator is stable and operates at the frame recurrence frequency of 80D() cycles per second delivering an output voltage of substantially rectangular wave form. Sharp pulses are obtaining by differentiating the These are used directly for controlling the length modulators of certain channels. For others an intermediate exciter stage is employed. By this method in combination with a control of the time constants of the individual modulators, each is caused to produce a pulse normally ending at the center position of its respective channel time.`

The outputs of the pulse generators and modulator Il) and the marker pulse generator II are supplied to the amplifier I3 and hence to the radio transmitter I4 which `they control or modulate to transmit pulses of radio frequency oscillations `of frequency f1.

The receiving equipment at the main station of Fig. 1 comprises a radio receiver 2| `for receiving radio waves of frequency fz transmitted from the mobile stations including that shown in Fig. 2. The output of the receiver `2| is amplified in an amplifier 22 and supplied to the various channel selectors and converters. Only the selector and converter 23 forchannel No. I from the station of Fig. 2 is shown. The others together with their associated control equipment are identical and dilfer in operation only due to the difference in the gating pulses supplied thereto which vary from each other in their time of occurrence to permit the selection of the respective channel pulses. The gating pulse is produced in the gate pulse generator 24 which is controlled by square waves from the generator 34 which is itself' controlled by the reference oscillator I2.

Again reference is made to my copending application for details of the channel selector and converter 23 and similar unitsfor the other channels. In the system of the present invention the receiving multiplex diners from that of my copending application in one main detail that however does not affect the general method of op eration. In the present 4system the incoming radio signals do not contain any marker pulses. Accordingly, the gating puise generators usually controlled by the incoming marker pulse are instead controlled by the marker pulse from generator Il I Thus both the transmitting and receiving system at the main station of Fig. 1 aretimed by the master oscillator I2. The gating pulses are utilized for separating out the various channel pulses.

For the purpose of producing the various gating` pulses a square `wave generator 34 is controlled by the marker` pulse from the generator Il. The output of the square wave generator 34 is nearly symmetrical dividing the frame period into two parts and is used to start the generation ;ofithefrespective gating pulses at the proper times so that theseveral gates will correspond in time to the respective channel periods. Since the first half of the square wave output of gener ator 34 starts at the end of ;the marker pulse which is also the beginning of the period for channel No; I, the gate generator 24 for that channel is controlled directly by the output of generator 34; Channel No. 5 is similar since the begining of its period coincides with the start of the second half of the output Wave lof generator 34. The generation of the gating pulses for the other channels cannot be directly started by the square wave but must be delayed by various amounts so that theirespective gates correspond. in time with the respective -channel periods. This is accomplished by including in each of those gate generators an exciter that produces a sweep voltage and arranging to start the generation of the gating pulse at a time when the sweep voltage crosses a certain reference voltage. For the purpose of providing the proper timing the various sweep exciters are provided with different time constants in order to lcontrol the rate of rise of the various sweepvoltages and accordingly to adjust the periods of the respective gates.

The output of each gating pulse generator is supplied to the respective pulse converter in par allel with the output of an ampliiier to which all the received channel pulses are supplied. The respective channel pulse will therefore be super* imposed on the gating pulse. The pulse converters are arranged to respond to a voltage equal to the Sum of the gate and channel pulse voltages but to be unaieected by either the gate or channed pulse voltage alone. i

The pulse converters are relaxation circuits similar to those used `as position modulators in the transmitting system. Their purpose is to convert the position modulation of the respective channel pulses to the original modulating signal (speech and control current). This is accomplished by converting the pulse position modulations to pulse length modulations from which the modulating frequency may be filtered out. The converters are arranged to be started only by a voltage equal to the sum of the gating and channel pulse voltages and toistop at the end of the gating pulse. Consequently they produce pulses that start with the occurrence of the appropriate channel pulse and stop at the end of the gating pulse and are therefore modulated in length inthe Same manner as the channel pulse is modulated in its time position.

The output of the selector and converter 23 is divided into two branches 25 and 25. The branch 25 containsa band-pass filter 2l which eliminates the low frequency control currents and high frequency disturbances from the signal current supplied to the outputsignal line 28.

The loranch`26 includes a time control network 30 and a switch 29. The Contact of switch 2S is connected to a second time control network 3|. For normal communication the switch `2li is in the open position as shown, and only the slow time control network 30 in the path 2B. During the` `period of establishing communication with the subsidiary station of Fig. 2 the switch 29 is operated to its closedposition connecting the fast time control networki'BI in circuit in 7 shunt to the network 30. The use of this fast time control network reduces the interference' with communication being vcarried out on other channels `'during thev establishment of commumcation as will be described in detail later. The

outputs of both networks 3l! and 3| is connected to the channel I input of the channel pulse gen-v erator and position modulator I at the output of filter I6.

At the subsidiary station as shown in Fig. 2 there is provided a radio receiver 40 for receiving Waves of the frequency transmitted by the transmitter I4. The-signal or video output of the receiver 40 is supplied to an amplifier 4I from which it is divided into two branches One of these branches includes the marker pulse selector 42 which operates to select the wide marker pulse 50 (see Fig. 3) to the exclusion of the narrow channel pulses I to 58. The marker pulse in the output of the selector 42 controls the gate generator 43. The other branch of the output of amplifier 4I leads to the channel selector and converter 44 which is lcontrolled by the gate pulse rom the generator 43 and'is identical in makeup and operation to the converter 23 of Fig. 1.

The signal output ci the converter 44 like that of converter 23 is supplied to the line 46 through a band-pass lter 45.

The output of the converter 44 is also branched off to a control circuit. This includes a control voltage frequency shifting network 41 the output of which is suppliedl to the reactance tube 49 which controls the frequency of the reference oscillator 5S).y

For transmission from the station of Fig. 2 the signal currents from the line 6I are suppliedY to the pulse generator and position modulator 62 which is controlled by the reference oscillations from the generator 59. The generator and modulator 62 operate in the same way as the corresponding unit I0 of Fig. 1 except of course that it is required to generate and modulate only a single channel pulse.

The position modulated pulse from the generator and modulator 62 is amplied in the ampliiier 63 and supplied to the radio transmitter 64 to cause the transmission of corresponding p-ulses of radio waves of a frequency f2.

operation The operation of the system can best be understood by rst considering the explanatory diagrams of Figs. 3 to 6 which show the pulse characteristics at various points in the system.

As previously stated Fig. 3 shows the pulses as generated and transmitted from the station of Fig. 1. Assuming that the distance separating the stations of Fig. 1 and Fig. 2 is such that the time required for radio waves to be transmitted from one to the other is t, then the pulses arriving at` the station of Fig. 2 will be as shown in Fig. 4. The shape oi the array of pulsesremains unchanged by the transmission delay so that the spacing between the synchronizing marker pulse and the various channel pulses remains xed, the complete array being shifted in time.

Now if the generator v62 were operated in synchronism with the spectrum of pulses received from the generator I0 at the station of Fig. 1 the rst channel pulse would arrive at the station of Fig. 1 at a time indicated by the position of the dotted line pulse 'II of Fig. 6, the transmission from the station of Fig. 2 to that of Fig. 1 requiring the same time t as the transmission in the opposite direction.- By comparing this with 8 the' pulse` array of Fig. 3"it will be seenrthat it does not fall in the period allotted t channel I but elsewhere in the'frame period so as to interfere with pulses from another station. Furthermore,as the station of Fig. 2 moved with respect tothat of Fig. 1' this position would vary. The same type of effect would be produced on the pulse from each station and there would be a complete jumble of received pulses at the station' Now if the pulse transmitted from the station of Fig. 2 is delayed by al time d the pulse transmitted therefrom will be indicated by the pulses 'I2 of Fig. 5 and the pulse arriving at the station f of Fig. l will be pulse 'I3 of Fig. 6. Ascan be example by an operator at the station of Fig. 1

observing the received pulse with respectto the transmitted array on an oscilloscope and then transmitting to an operator at the station of Fig. 2 the information required to adjust the delay. Such a method would be cumbersome particularly where there are frequent or rapid variations in the distance between stations. In accordance with a feature of this invention there is provided a method of `maintaining an automatic regulation of the delay. y However the manual method is one which could be used for establishing contact between stations after which the automatic control could take over.` However, in the present embodimentthe process of establishing contact as well Vas that of maintaining the proper adjustment after contact hasbeen established is carried on automatically.

In order to achieve this automatic control there is transmitted from the main station of Fig. 1 as a modulation ofthe channel pulse a pilot that is a measure of the error in the mean position of the received channel pulse. In particular the receivers at both stations are caused to give out a direct current voltage depending on the mean position of the channelpulse but .having the voice (or other signal) frequencies ltered out. At the main station the outgoing channel pulse is modulated by having its position varied by this direct current voltage. The transmitter at the subsidiary station (Fig. 2) is caused to emit its pulse at a time or phase with respect to the received synchronizing pulse that is controlled by the corresponding direct current output of its receiver. Thus if the mean position of the channel pulse received at the master station is late, the channel pulse transmitted therefrom will be caused to be late.` In turn if such a late pulse is received at the subsidiary station the effect will be to cause the pulse transmitted therefrom to be earlier.`

The Vregulation of the transmitter at the subsidiary station isin fact a matter of time control and consequently could be achieved by putting the received synchronizing pulse through ai controllable delay network vand then utilizing it to control the generation of the outgoing pulse.

.` However if'it is necessary to take care of any wi-de variation in separation between the stations a' large variation in delay or phasefshift would attese have to be provided. Thus if the total range over which the system is to operate is one hundred miles and the frame or recurrence frequency is 8 `kilocycles, the phase shifter would have to cover a range of 3100 degrees or nearly nine revolutions of a 36o-degree phase shifter. The maximum range variations that could be covered by a single revolution of such a 36D- degree phase shifter would be about twelve miles. For larger variations the operating range would have to be divided into zones of this width and each time the boundary f zone traversed there would be a loss of contact and a new searching operation would have to be undertaken.

In accordance with a feature of this invention this limit on the range of control avalable'by the use of a phase shifter is avoided by using a local reference oscillator 59 of controllable frequency. The recurrence rate of the outgoing pulse is controlled by this oscillator instead of by the received synchronizing pulse. By controlling the frequency of the oscillator the phase of the outgoing pulse may be regulated to any extent desired. Proper regulation is achieved by utilizing the pilot `or control voltage output of the receiver to control the frequency of the oscillator. rlihus if conditions require that the phase of the outgoing pulse be retarded, the frequency of the oscillator would be momentarily reduced by a small amount and then brought back to the correct value when the phase was correctly established. The action is quite analogous to that of synchronizing two alternators to be connected in parallel.

The control circuit is in fact a closed loop and as has been recognized in connection with other similar control circuits the problem of yits stable operation is analogous to the problem of stability in feedback ampliers. l i

In considering the problem of establishing contact it can in general be assumed that when it is desired to put a subsidiary station in communication with the main station there is no information available by which the timing of the transmitted pulse may be correctly adjusted. Thus it will be necessary to turn on the subsidiary` transmitter and search by sweeping the channel pulse until contact is established. The probabilities are greatly in favor of the pulse not being in its assigned period at the main station.

As the pulse is thus swept it will of course pass through the time periods assigned to other channels which may already be in communication with the master station. When the searching pulse sweeps intosuch a time interval it will affect the converter for that channel causing it to report a pulse of incorrect time. r r The effect would be to produce a pilot current that corresponds to this incorrect time and which will aifect the subsidiary transmitter for that channel to throw it out of correct timing. Accordingly if the dynamic properties of the control circuits of the two channels are identical the established channel would be driven out of synchronism (correct timing) before the searching pulse had traversed that channel time interval. In order to avoid this diinculty the control circuit of a channel establishing contact is given a different dynamic characteristic from that of channels already in communication. This is accomplished by providing the control` circuit of each channel withiilter networks of Widely different time constants. During search, a wide band control circuit is used andthe searching pulse is made to sweep rapidly across the channels. yOnce contact is established the Wide band network is disconnected from the 10 circuit and the control is effected by a narrow band network thus limiting its band Width to `a fe'w cycles per second. Such a control is sumcient t'o maintain `synchronisii'i once it has been established. This sluggish control Will not be' thrown out by the rapidly moving sweeping pulse. All that can happen is that a brief burst of noise of abouti Inillisecld duration will appear in e'ch channel as it is swept across by the searching' pulse. i

, vafter Contact is established the tand Width of the control circuit required to maintain control depends upon the rate of change of the relatii'fe position of the two communicating statioi'ls and `hperltheinherent Stability 0f the circuits. If the 'relative notion is assumed to be 550 miles" per hour the path length will vary at the rate of 6000 microseconds per hour or 100 microseconds perminute. The effect of such operations on ban-d width can probably best be demonstrated by assuming certain conditions of control modulatitn'i.` Thus, assuming that the pulse were modulated at an amplitude of 5 microsecondswith a sine wave of frequency of 0.05

cycle per second, the` maximum rate of change wouldbe L57 microseconds per second. This is somewhat less than 100 microseconds per minute. Accordingly a control circuit having a band width of l or 2 cycles per second is entirely capable of maintaining control once it is established The conditions encountered in establishing control require that the pulse be swept through the frame time and that the control conditions of the sweeping cir-cuit be such as not to interfere with channels already in communication. In the specific system herein disclosed this is accomplished by having the searching pulse sweep through one frame (complete recurrence period) in approximately 1/100 of asecond. This of course requires that in order tol be able to seize contro-l ofy a searching pulse varying at 4this rate, the control cir-cuit must pass a band of about cycles per second. r i r For this purpose the switch 29 at the main Station of Fig'. 1 is .Clos-ed. This connects the; net,- work" 3i into'- the control circuit. This network 3l has a band Widthof approximately 100 cycles per second. For normal operation after contact has been established the network 30 is used. This network has a band width of about 2 cycles per second which is suicient to take care of variations due to relative motion of the stations and .the like as was pointed out above.

Since stations already incommunication are operating with e. control circuit having a `band width of 2 cycles per second the pulse sweeping"` at` the rate of `1G() cycles per second will not affect the control cir-cuit to throw the stations out of synchronism. The only effect of the sweeping pulse is `to cause a burst oi noise lasting about 1 millisecond as it passes through the time period of such a station.

Automatic seizure Fig. 7 shows a Inodication of the circuit of Fig. l in which .the change from a condition for establishing contact to one for communication is established automatically, rather than bythe use of the manually operated switch 29 of Fig. l. I-n the circuit of Fig. 'l certain of units of Fig. l are shown and are given the -s'amerei'erence numerals as inlig l. r

In' ythe circuit of Fig. 7 the capacitor au and resistor 8| are the equivalent of the fast net- Work' 3|` of Fig. 1 in the "condition in which 1` 1l sistor 82 is shorted by the armature and contact of relay 83. When the relay 83 is released, as shown, the network comprises the capacitor 88 and the two series resistors 8l and 82 and has a relatively long time constant thus being equivalent to the slow network 30 oiFig. 1.

The choice of the two networks is controlled by the relay 83 which is operated by the vacuum tube 84. For the purpose of controlling the tube 8d, there is obtained from the vchannel selector and converter 23 a voltage that is negative in the presence of received pulses (either modulated or unmodulated) and that disappears or becomes posi-tive in the absence of a received pulse in the respective channel. In the system of my copending appli-cation previously identied such a voltage is available at the output of the rectifier of the ringing circuit. This voltage is fed through the connection 85 and the resistance-capacity lter 86 to the grid of the tube 84.

In the absence of any received pulse in the channel, there will be no voltage in the output lead 85 an-d the tube 84 will be conducting, holding the relay 83 in its o-perated position. The grid battery 8'! provides the proper voltage to cause the flow `of plate -current required for the operation of relay 83. Under these conditions the resistor 82 will be short-circuited and the short Ktime constant network Bil-8l will provide the fast operating time required for searching.

When a signal pulse is received in the channel, the voltage on Ithe lead 85 becomes negative and the plate current of the tube 84 is blocked causing the release of the relay 82. This results in the removal of the short circuit across the resistor 82 and the introduction into the circuit of the slow network comprising the capacitor 80 and the two resistors 8| and 82.

In order to permit the establishment of a steady voltage across the capacitor 80 on `the appearance of a pulse in the channel and before the change in the time constant of the circuit, the network 86 is used in the grid circuit of the tube 84; This network 86 has such a time constant that the release of the relay 83 is relatively slow permitting the voltage on the capacitor 88 to cordingly, the continuous attention of an operator to manipulate the switch 29 of Fig. 1 when any channel comes into operation or goes off is not required.

What is claimed is:

v l. In a system for two-way communication between a main station and a, plurality of substations by time division multiplex, the method of regulating the time of transmission from each substation to prevent interference between the channel pulses from the respective substations received at the main station which comprises, transmitting from the main station for each channel a pulse having a characteristic proportional to the time position of the received pulse of that channel, and regulating the transmission 1,2 from each respective substation in accordance with said characteristic of the respective pulse received from the main station.

2. In a system for two-way communication between a main station and a plurality of substa tions by time division multiplex, a time standard at the main station, means controlled by said time standard for determining the periods allotted to the various outgoing channels, means controlled by said time standard for separating the received channels at said main station, means for transmitting in each channel period a signal having a characteristic proportioned to the time of the received signal of the corresponding incoming channel, and means for regulating the time of transmission from each substation in accordance with said characteristic of said signal having a characteristic proportional to the time of the received signal.

3. In a system for communication between a main station and a plurality of substations by time division multiplex, the method of regulating the time of transmission from each substation to prevent'interferenoe between the respective chane nel sig-nais received at the main station which comprises,` transmitting from the main station a pulse for each channel in time position proportional to the time position of the received pulse of that channel, and regulating the time of transmission from each substation in accordance with the time position of the corresponding said pulse received from the main station.

4. In a time division multiplex radio system for two-way communication between a main station and a pluralityy of substations relatively movable with respect thereto, means for transmitting from said main station pilot signals indicative of the time of arrival of signals from the substation, and means at each substation responsive to said pilot signals for regulating the time of transmission fromrthe substation.

5. In a time division multiplex radio system for two-way radio communication between a main station and a plurality of substations relatively movable with respect thereto, means for transmitting from said main station pilot signals individual to each ymultiplexchannel and indicative of the timing of signals received from the substation operating on the respective channel, and means at each of said substations responsive to the respective pilot signals for regulating the time of transmission therefrom. y

6. In a time division multiplex radio system for two-way communication between a main station and aplurality of `substations relatively movable with respect thereto, a recurrence frequency oscillator for timing the transmission from eachv substation, means for transmitting from said main station pilot signals indicative ofthe time of arrival of the signals from the substations, and means at each substation responsive to said pilot signals for regulating the phase of said recurrence frequency oscillator. 4` y 7. In a time division multiplex radio system for two-way communication between a main station and a plurality of substations relatively movable with respect thereto, a control loop for regulating the time of transmission from each substation comprising the radio transmitters` andre'ceivers at said main station and the respective substation,a recurrence frequency oscillator for determining the time of transmission from each substation, means for momentarily varying the frequency of said oscillator Vand means for producing a voltage for controlling ksaidnieans and having a characteristic proportional to the time of transmission'from said substation relative to the .required time therefor to prevent interference at the vreceiver of the main station between signals from the `respective substation and those from other substations.

8. In a system for two-way communication between a main station and a plurality of substations by `time division multiplex, timing means at the main station for regulating the transmission of channels, means at the main station for separating the received channel signals under control of said timing means, means at the main station for producing a pilot signal having a characteristic proportional to the time of arrival of each channel signal relative to its assigned time, means for transmitting each said pilot signal in its corresponding time and means at each substation ior regulating the. time of transmission in accordance with the received pilot signal.

9. In a system for two-way communication between a main station and a plurality of substations by time division multiplex employing pulses modulated in time position, a main station comprising means for separating the received pulses on the basis of the time distribution of the pulses transmitted from the main station, means for producing a pilot current proportional to the average deviation of each pulse from its normal time position, means for modulating each corresponding transmitted pulse in accordance with said pilot current, and a plurality of substations each comprisingI means for regulating the normal time of transmission in accordance with the pilot current modulation of the pulse received at the respective substation.

10. The method of two-way communication between a main station and a plurality of substations which comprises transmitting from the main stations to the substations successive frames of pulses one pulse for each substation and each modulated by the signal for said substation, said pulses each occurring within a predetermined channel period in said frames, receiving at the main station successive frames of pulses in synchronism with the transmitted frames, transmitting 'from said main station pulses the time of occurrence of which within their respective channel times is proportional to the time of occurrence of the respective received pulses, transmitting from each substation a pulse modulated by the signal for the main station, and regulating the time of transmission from each substation of said pulses in accordance with the time of said pulse received at the respective substation.

11. The method of two-way communication between a main station and a plurality of substations by time division multiplex comprising establishing at the main station a time pattern of recurrent frames each divided into channel periods one for each substation, transmitting signals to each substation during corresponding channel periods of said frames, separating signals received at said main station on the basis of said time pattern, transmitting from said main station a pilot signal indicative of the time of arrival of the signal from each substation, and regulating the time of transmission from each substation in accordance with the respective pilot signal.

l2. The method of multiplex communication between a main station and a plurality of substations which comprises, producing at the main station successive frames of pulses each frame comprising, an initial pulse of relatively long i. pulses, producing under control of the separated initial pulses a gating pulse occurring at a time corresponding to the channel period of the channel pulse for that substation, combining the received channel pulses with said gating pulse to separate and detect the signal for that substation, generating a pulse to be transmitted to the main station, regulating the normal time of occurrence of said pulse in accordance with the average time of. occurrence of the separated chani nel pulses, modulating said pulse by the signal for said main station, separating the pulses received at the main station, and establishing the average time position of each pulse transmitted from said main station in accordance with the average time of arrival of the pulse from the corresponding substation.

13. The method of two-way communication between a main station and a plurality of substations which comprises, producing at the` main station successive frames of pulses each frame comprising a synchronizing pulse of long duration and a succession of short uniform channel pulses one for each substation and occurring normally at equal time intervals, modulating the time of oc- I currence of the channel pulses in accordance with the respective signals to be transmitted to the substations, receiving at the main station similar frames of channel pulses one from each substation, producing under control of the transmitted synchronizing pulse a plurality of gating pulses at successive times corresponding to the several channel periods, combining the channel pulses received at the main station with the respective gating pulses to separate and detect the signals in the different channels, further modulating the time of occurrence of each transmitted channel pulse in accordance with the average time of occurrence of the corresponding channel pulse received at the main station, receiving the pulses at each substation after transmission through a common medium, separating the synchronizing pulse, generating under control of said synchronizing pulse a gating pulse at a time corresponding to the channel period for that substation, combining the received pulses with said gating pulse to separate the respective channel pulse, generating an outgoing channel pulse, regulating the normal time of occurrence of said outgoing channel pulse in accordance with the average time of said separated channel pulses, and modulating the time of occurrence of said outgoing channel pulse in accordance with the signal to be transmitted.

14. A time division multiplex system 4for communication between a main station and a plurality of substations in which the main station comprises means for producing recurrent frames of pulses each frame comprising a marker pulse and a plurality of channel pulses one for each communication channel and each having an assigned normal time position in each frame, means for modulating the time position of each of said channel pulses in accordance with the signal of the assigned channel, a receiver for similar recurrent channel pulses from the substations,

.15 means synchronized with the transmitted marker pulse for separating the channel pulses from each substation from those of other substations, and means responsive to any deviation of the average time position of each received pulse from its assigned normal time position for producing a corresponding variation in the average time posi,- tion of the transmitted channel pulse of the corresponding channel, and each substation comprises means controlled by the received marker pulses for separating the channel pulses assigned to that substation from the other channel pulses, means for generating a channel pulse for trans` mission to said main station, and means responsive to the average time position of the received 16 mal .rate of recurrence of the channel pulses, means for producing a voltage proportional to the average time position of the received channel pulses and means for regulating the frequency of said oscillator in accordance with said voltage.

J'AMES O. EDSON.

REFERENCES CITED The following references are of record in the le of this patent:

UNITED STATES PATENTS

Images