US2285598A - Multiplex telegraph distributor - Google Patents

Description

June `9, 1942. R. E. MA'rHl-:s

l MULTIPLEX TELEGRAIH DISTRIBUTOR Filed June 15. 1946 s sheets-sheet 1 ,lun,9, 1942.l R. E. MATHEs MULTIPLEX TELEGRAPH DISTRIBUTOR 3 Sheets-Sheet 2 Filed June 15, 1940 vINVENTOR www@ i. mr/ff:

June 9,1942. R E, MATHEs 23,285,598v

' MULTIPLEX TELEGRAPH' DISTRIBUTQR Filed June l5, 1940 3 Sheets-Sheet v/fa /22 /5/ fw /50 f5:

Fia-4i.

BYA

INVENTOR.

Patented June 9, 1942 llIULTIPLEX TELEGRAPH DISTRIBUTOR Richard E. Mathes, Westfield, N. J., assignor to 'Radio `Corporation of America, a corporation of Delaware Application June I1 5, 1940, Serial No. 340,657

Y(Cl. l178-51) 18 .'Claims.

This invention relates to multiplex telegraphy and particularly to distributors for connecting the various channels tothe ,sending or receiving apparatus in continuous succession.

An object of the invention is to eliminate all moving parts in the distributor mechanism by harmonic wave control of a switching medium.

Another object is to distribute the signals to or from the various channels by improved electron action, which may be termed electron switching.

Other objects of the invention YWill appear in the following description, reference being hadto the drawings in which:

Fig 1 is an electrical diagram of a distributor mechanism for a two channel system.

Fig. 2 is an electrical circuit for a three channel system.

Fig. 3 is a set of voltage curves `applying to Fig. 1.

Fig. 4 is a set of voltage curves applying to Fig. 2.

Figs. 5 to ll, are graphs of the signals with which the operationis explained.

Referring to Fig. 1, the invention has been illustrated in connection with the receiving end of the system, though it could have been shown equally as well with the transmitter. At the receiving station there is a standard frequency unit Fs maintained automatically in exact synchrcnism with the `standard frequency unit at the transmitter and also in the desired phase relation This is shown 1in block diagram and indicated by reference character I. Standard frequency units are old and well known in the art Yand there are various types described in the literature and in commercial use. It is, therefore, unnecessary to complicate this disclosure by eX- plaining the details of .such unit, since this per se forms no part of the invention.

Voltage Ifrom the standard frequency unit is fed through conductors 2 and 3 Ytoa frequency changer which in this case is a device for lowering the frequency a predetermined amount to a frequency F3. Frequency reducers or dividers are also well known in the art and need no description, `especially since any `frequency divider lwould suffice as far as my .invention is concerned. Reference, however, is made to a multivibrator frequency changer, which may be either a divider or multiplier, described in an article by Callahan, Whitaker and Shore in Radio Facsimile, volume I, vpages 91 and 92, 1938, published by R. C. A. Institutes Technical Press, New York.

second and frequency F1 60 cycles per second f and it also may be assumed that frequency F1 is the same as the channel keying frequency.

The output offrequency changer 4 is fed through conductors 9, I0 to primary II of transformer I2 and also to primary I3 of a similar transformer I4. The output of frequency changer 'I is fed through conductors I5, I6 to primaries I'I and I8 of transformers I2 and I4, respectively. The secondaries of the transformer I2 are connected to the grids of tubes I9, 23 of a push-pull stage, so that when one grid is positive the other is negative. A negative bias battery ZI or other suitable bias is connected to furnish thedesired negative potential to the grids. The primaries 22, 23 of output transformer 25 are 'joined together in push-pull and connected to the positive terminal of the plate supply, ythe negative terminal of this plate supply being connected vto the cathodes of the tubes. The secondaries of the transformer 25 areconnected to a full wave rectifier of any kind, but by way of example I have shown it as two Fleming tubes 26, 21 joined together in the hook-up ygenerally used to produce full wave rectification.

The positive terminal of the rectifier is connected preferably through a countervoltage source 28 to one terminal of a load resistance 28' and to the grids of vacuum tubes 29, 30, through condensers 3|, 32, which may have leak resistances as shown. Resistances 33, v31% are connected in series across the two grids. The junction point of these resistances is vconnected to the other terminal of the resistance 28 and to the negative terminal of the double rectifier. The cathodes of tubes 29 and 33 are connected to the junction point of the two resistances .33, 3A and to the negative terminal of the B supply. The negative b-ias 35 is sufcient to block .the tube 29 except when the voltage peak from the local frequency arrives and at that time thetube unblocks. yNegative bias 35 blocks tube 29 even when the local voltage peak arrives, but when kthe signal pulse arrives simultaneously with the local peak, Vthe tube conducts.

The plates of tubes 29 and 30 are connected through resistances 36, 3l to the positive terminal of the B supply, the negative terminal being connected to the cathodes. This may be grounded as shown. Two points on resistances 36A and 31 are connected by conductors 38 and 39 through resistances 40, 4| to the grids of tubes 42, 43, respectively, of a Finch locking circuit generally indicated at 44, which is well known in the art, being described, for example, in Finch Patent 1,844,950. The positive terminal of the B supply is connected to the plates of tubes 42 and 43 through resistances 41 and 48 and the negative terminal is grounded and connected to the cathodes. The grid of tube 42 is connected through resistance 49 to the plate of tube 43 and the grid of tube 43 is connected through resistance 50 to the plate of tube 42. Appropriate resistances 5I, 52 are connected between the grids and cathodes of tubes 42 and 43, respectively, through a negative bias source 53. Utilization device 54 in channel No. 1 may be connected in various Ways to the locking circuit 4|, but I have indicated it as adjustably connected across resistances 41, 48.

The sec-ondaries of transformer I4 are connected to a full wave rectiiier shown, by example, as consisting of Fleming tubes 55, 56. The positive terminal of the double rectier is connected through a counter voltage source 51 to one terminal of a load resistance 51 and to the grids of tubes 58 and 59 through grid condensers and leaks, as With tubes 29 and 80. These grids are connected together through resistances 60, 6I and the junction point of the resistances is connected to the other terminal of the load resistance 51' and to the negative terminal of the double rectifier, as well as to the cathodes of tubes 58 and 59 and to the negative terminal of the B source. Negative bias sources 62 and 62 have the same effect on tubes 58 and 59 as that already described in connection with tubes 29 and 30. The anodes of tubes 58, 59 are connected through resistances 63 and 64 to the positive B terminal.

The output terminals 65, 66 are connected between points in these resistances through resistances 61, 68 to the grids of the tubes 89, 19 of a locking circuit 1| for channel No. 2, like the one already described for channel No. I. Utilization device 12 may be connected to the plate resistances of tubes 69, 10, similar to the utilization device 54.

The incoming signal is applied to resistance 13 through line 14. It is assumed, merely by way of example, that this incoming signal consists of interrupted direct current for the marks and spaces of the telegraph signal, although it is common to have the signals sent from the distant radio receiver in the form of tone signals, which on reaching the local office are rectified and filtered, but this is immaterial as far as my invention is concerned. The wire 15 places the incoming signal on the grid of tube 30, and Wires 15 and 16 place it on the grid of tube 59.

The operation of the embodiment of Fig. 1 may be described as follows:

The standard frequency unit I is continually producing an alternating electromotive force of any desired form, but for explanatory purposes it Will be assumed to have the form of a sine Wave. It has already been assumed that this frequency has a value of '120 cycles per second. This frequency is reduced by the frequency divider 4 to 180 cycles at the terminals 5, 6 and the voltage divider 1 reduces it to 60 Cycles at terminals I5, I6. It will likewise be assumed that these voltages have a sine wave form.

The 180 cycle voltage is applied to transformers I2 and I4 and the 60 cycle voltage is likewise applied thereto. The 60 cycle voltage component induced in the secondary of transformer I2 will (by proper poling) be in phase with the 60 cycle voltage in the primary of transformer I4. Graph 11 of Fig. 3 represents both of these secondary voltage components, Also, by proper poling, the

` 180 cycle voltage component in the secondares has the phase relation indicated by graph 18 of Fig. 3. The relative values of the voltages of the two local sources will not affect the phase relation of the resultant voltage peaks, but for purposes of illustration it has been assumed that the 60 cycle voltage is three times that of the 180 cycle voltage. The combined output in the secondaries of transformers I2 and I4 is the sum of the two graphs and is represented by graph 19. This graph has one positive peak and one negative peak, one cycle only being considered. The output voltage of transformer I4, typified by graph 19, is fed into rectiers 55, 56 and the rectified voltage or current will appear as a pulsating output 19, 19 of the same polarity. While it is not a requirement, I prefer to use a counter E. M. F. 5'I to serve as a threshold for passing the rectified voltage only above the voltage of the counter E. M. F. The rectied Voltage then appears as in graph 80, having the same two peaks 8|, 82 as the graph 19, 19.

The rectifiers 26, 21 produce similar rected pulses, but they are out of phase with the pulses from rectiers 55, 56, due to the action of the vacuum tube push-pull stage with suitably designed transformers. Hence, the voltage pulses applied to the grids of tubes 29, 30 have the phase relation given by graph 83, 84, 85, unless no counter E. M. F. is used. In that case, the output would be like graphs 19, 19', but spaced 90 therefrom.

It is thus seen that the local standard frequency at the receiver is continually applying local pulses to the grids of each pair of keying tubes, but the pulses applied to one pair are 90 displaced from the pulses applied to the grids of the other pair.

Now referring to the message being sent, let it be assumed that at the transmitter the letter U is being transmitted in channel No. I and the letter D is being transmitted in channel No. 2 in the chance time-relation shown in Figs. 5 and 6. 'Ihe tWo signals Will be picked 01T at an intermediate part, for example at about the center of each half baud, at lines I, II, by the transmitter selector and will appear as short pulses, as shown by Fig. 7, the spaces usually being negative current and the marks positive, as shown. This pulsating current of Fig. '1 through an appropriate locking-keying circuit will key the radio transmitter and produce a radio wave, indicated by Fig. 8. The Wave continues from a positive pulse to the next negative pulse and is interrupted from that negative pulse to the next positive pulse, and so on. Reference is made to my Patent 2,214,642, Sept. 10, 1940, for a complete disclosure of a lockingkeying circuit in a transmitter, which, per se, is not claimed herein.

The combined signal Wave will be received and detected at the receiving station and may be assumed to appear in the hook-up of Fig. 1 as direct current pulses in the resistance 13, as shown in Fig, 9.

When local pulses84,85 arrive at thegridsof tubes 29 and 30 with no signal pulse, tube 29 unblocks and tube 30 .does not. The grid of locking circuit tube 43 therefore has high voltage due to the absence of drop in resistance 31. A reduced Voltage is applied to .the grid of locking circuit tube 42, as there fis a .voltage drop in resistance 30. Consequently, tube 43 conducts and this instantly blocks tube 42, due to the cross connection between the plate vand grids as shown. When the local voltage pulses pass, the locking circuit does not change. `The equilibrium can be changed only by changing the relative voltage of the grids.

The local voltage pulses 8|,.82 (Fig. 3) produce the same effect on locking tubes 69.and 10.

vWhen no signal arrives, utilization devices 54 and 12 will have the left-hand terminal positive in respect to the right-hand terminal and a space eifect is produced. This is because there is a voltage drop in resistance 48 and none in resistance 41.

When the time corresponding to the center of the rst half of the baud of mark signal 89 of channel No. I arrives at I (Fig. 9), it is, of course, applied to the grids of both tubes 30 and 59. Local pulse 8| has unblocked keying tube 29 and the signal pulse andlocal peak 8| unblocks tube 30. Neither tube 58 nor tube 59 is at this time unblocked, as their local voltage peak has not arrived. Due to the adjustment of sliders 38 and 39 or to the relative values of the plate current in tubes 29 and 30, slider`39 is at a lower potential than slider 38. This makes tube 42 conduct, which instantlyblocks tube 43 and reverses the voltage applied to utilization device 54. Utilization device 54 therefore starts to form the signal mark at the cene ter of the rst half baud, namely, at point '90 of Fig. l0. This mark will continue until the balance of the locking tubes is again disturbed.

When the center of the second half baud of signal mark 89 arrives at 1I, Fig-9, a local pulse is applied to the grid of tube 59, but no local pulse is applied to the grid of tube 30; hence, tube 59 is unblocked by the signal and local pulse. The balance of locking tubes 69 and 10 is thus disturbed and tube 69 conducts while tube 10 blocks. rlhis reverses the voltage applied to utilization device 12 and produces the beginning of signal mark 9| in channel No. 2 (Fig. 11).

Since the tubes 29 and 30 received no local pulse at the center of this second half baud, the equilibrium of the locking circuit in channel No. I was not disturbed; hence, the mark in channel No. continues until the center of the next baud is reached. At this time tube 30 will not be unblocked, for, as will be seen by referring to the combined received current in Fig. 9, there is no signal voltage on the line 15. Tube 29 is always unblocked by the `local pulses and this disturbs the balance and causes tube 43 to conduct and tube 42 to block. This reverses the voltage in utilization device -54 and terminates the mark signal 90.

Mark 9| of channel No. 2 will continue until the local voltage pulse is applied to the grid of tube 59, with a space condition of the signal. By comparing Figs. 9 and 11, it will 'be seen that this will occur for the rst time at the point 92, at which time tube 59 `will'not be unblocked. This will unbalance the locking circuit and cause tube 43 to conduct and block tube 42. 'Ihe potential in utilization device 12 in channel No. 2 now reverses and Yterminates the mark. .From this, it will be clear that the local pulses being timed with the channel keying cycles will separate the messages into the two channels :from 'the composite signal received.

In the modification of Fig. 2 a three channel installation is shown. In this figure, standard frequency unit I may be the same as in Fig. 1; that is, it may be assumed to have the frequency 720 cycles per second, though, of course, the frequency may be anything desired so long as it hasa frequency that is proportional to the frequency standard at the transmitter. 'Ihe first frequency reducer |0| would then be designed to lower the frequency to 120 cycles per second at output lines |02, |03 and the second divider |04 would lower the frequency to 60 cycles per second at the output lines |05, |06. In this modication the two frequencies, cycles and cycles, will be combined and rectified so as to produce six voltage peaks to switch the signals into the three channels. As before, the keying :frequency of each channel Vis 60 cycles per second.

The 60 cycle voltage isfed into a primary coil in each of the transformers |01, 08. The 120 cycle voltageis likewise applied to a primary coil in each of the transformers. The 60 cycle voltage Valone is applied to transformer |09.

The poling of the coils in the transformers |01, |08 is such that the two frequency components in the secondaries are in phase at the start'of the 60 cycle in one of the transformers, say |01, and out of phase in the other one, |08. This phase relation can be obtained in other ways, but for purposes of illustration I have indicated `it by reversing the connections from lines |05, |08 to the primary |I0. For explanatory purposes it hasrbeen assumed that the secondary coils III, |I2 of transformer |01 have their upper terminals simultaneously of the same sign and the same assumption has been made in respect to the secondary coils of transformer |08.

Fig. 4 gives the relation of the various components in transformers |01, |98 and |09. In the secondaries of transformer |01, the 60 cycle current is indicated by graph II5 and 120 cycle current by graph IIS. The amplitude values chosen for the two frequencies are such that the 60 cycle frequency is twice that of the 120 cycle frequency. When graph IIB is combined with graph'I I5, it produces a composite frequency indicated by graph I1, having a peak V| I8 at 60 f.' and another peak I I9 at 300.

In transformer |08, Where the 120 cycle voltage is reversed in phase, graph |20 would indicate this component. The combination of graph |20 with graph H5 produces inthe secondary coils of transformer |08 -a composite voltage |2| having a peak 122 at 120.and a peak |23 at 240.

,In the secondary coils of transformer |09, only the 60 cycle frequency is present, as indicated by graph v|I5. Due to the interposition of -the vacuum tube stage |24, as in Fig. l, the voltage peaks in the secondaries |25, |26 are shifted 90. Olne peak |21 will then appear at 180 Aand the other peak |28 will appear at 360. The amplification factor or the number of turns on :the transformer coils are preferably so chosen that the voltage peaks |21, |28 will have the ,same maximum value as the peaks IIB, |22, |23 ,and ||9. 'Thepeaks |21, |28 as kshown on the drawing indicate position only and not relative amplitude.

Rectifiers |29 and |30 are so connected to the secondaries of transformer |08 that |29 passes the positive half of the wave only, while the other passes the negative half of the wave only. The secondary coils of transformer |01 are likewise so connected to rectifiers |3| and |32 that one passes the positive half of the wave only, while the other passes the negative half of the wave only. In other Words, the pairs of rectifiers are so connected to their transformer as to produce full wave rectification. The rectifiers |33, |34 are also connected for full wave rectification.

The output of rectifiers |33 and |34 is connected to the load resistance |53 and to the grids and filaments of tubes |35, |36 in the same way as in Fig. 1 The circuit of these two tubes is otherwise similar to that of tubes 29 and 30 in Fig. 1 and need not be specifically described. Load resistance |54 of rectier |3| has its terminals connected to the input terminals of tubes |31, |38, the grid terminal being positive. Rectifier |30 channel has its output terminals connected to load resistance |54 and hence to the input terminals of tubes |31, |38, its positive terminal being connected to the grid terminal.

Load resistance |55 of rectifier |29 has its output terminals connected to the input terminals of tubes |39 and |40. Rectifier |32 also has its output terminals connected to resistance |55 and hence to the input terminals of tubes |39, |40. The grid terminal of resistance |55 is positive for both rectifiers.

Except for the differences required for the changes mentioned, the circuit arrangement of the keying tubes in Fig. 2 is similar to that of Fig. l. Also in Fig. 2 the signal input in resistance 13 is connected by wire |4I with the grid of tube |30, by wire |42 with the grid of tube |38 and by wire |43 with the grid of tube |40.

While this embodiment of my invention can be readily used without threshold rectification, it is preferable to use it and therefore I have indicated counter E. M. F. sources |56, |51 in circuit with these rectiers.

To simplify the wiring diagram, the negative end of the resistors |54 and |55 has been connected to the cathodes of their respective keying tubes through ground, but, of course, this not essential.

The keying tubes in Fig. 2 will be biased to operate in the same way as the keying tubes of Fig. 1 and the output terminals of the tubes will be connected to locking circuits in the same way as in Fig. 1, except, of course, there Will be three locking circuits and three utilization devices. The operation of the modification of Fig. 2 is as follows:

The local Voltage applied to the input of tubes |31 and |38 in each channel keying cycle will consist of peak ||8 at 60 and peak |23 at 240. Due to rectification this latter peak may be indicated by peak |49. The local voltage pulses applied to keying tubes |39 and |40 will consist of peak |22 at 120 and peak |50 at 300, the latter being the rectified peak I9.

The local voltage pulses applied to keying tubes |35 and |36 will be peaks |5|, |52, which are the peaks of the 60 cycle wave ||5 shifted 90 at |21 and |28 with the proper amplitude value and rectied.

It will be seen from the six peaks I8, |22, |40, |50 and |52 that the circuit connections will bring the keying tubes into action to switch the signals into the three channels as desired, but the action will be briefly referred to.

In a three channel system each baud or dot is divided into three parts and a middle portion of each sectional baud is picked off and transmitted, as indicated in the two channel system of Fig. 1. Except for the fact that there are three sections to each baud, the action is the same as already described for the two channel system. The cycle frequency, graph ||5, is, as before, the same as the channel keying frequency of the message. It will thus be seen that there are six equally-spaced peaks for each baud or dot cycle.

When the peak IIB is impressed on the grids of keying tubes |31 and |38, the signal lpulse in channel No. will arrive at the same time and the locking circuit of channel No. (not shown) will be keyed for a marking pulse. When peak |22 arrives in channel No. 2, the signal in channel No. 2 will operate its locking circuit and utilization device. When peak |5| arrives, the signal of locking circuit No. 2 will be operated and the utilization device of that channel will respond.V All three of these .peaks may, for example, be assumed to be positive peaks. When the rectified negative peak |49 arrives, it will operate the locking .circuit of channel No, Channel No. 2 will have its locking circuit operated when the next rectied negative peak 50 arrives simultaneously with the signal of channel No. 2 and finally the third rectified negative peak |52 will arrive simultaneously with the signal of the third channel land its utilization device will be operated by the third channel locking circuit.

To prevent possible confusion over the term channel keying frequency and related terms, it is explained that the smallest marking unit in a message is a dot. Since a space is required tordistinguish this dot from the other characters, the space immediately following the dot is .part of the dot cycle, the space being of equal length with the dot. The number of these dot cycles of a single channel, or their equivalent, fed through the multiplex system per second is called the channel keying frequency. In a multiplex system the channel keying frequency of all the channels is the same and the composite keying frequency is the sum of all the channel keying frequencies. For example, in a three channel system, if the channel keying frequency is 60 dot ,cycles per second, the composite keying frequency is 180 dot cycles per second. Both of these terms are used in multiplex telegraphy,

but I prefer to use the term channel keying frequency in explaining the principles of this invention.

The apparatus in the various channels may be interchanged; that is, Y in Fig. 1 the push-pull tubes could be used in channel No. 2 instead of channel No. and in the embodiment Fig. 2 the channels can be interchanged, except, of course, the rotation of the channels must be in the order of rotation at the transmitter.

I have illustrated my invention in connection with two channel and three channel operation, but it is not limited to this, as the principle may be applied to `produce operation of any additional number of channels.

The heating elements of the unipotential tubes in Figs. 1 `and 2 have been omitted from the drawings, as they are well known in the art.

from a fundamental wave and its third harmonic, displacing one derived 4wave 90 from the other and full-wave rectifying the displaced and undisplaced derived Waves.

3. The method of producing keying voltages for commutation of signals in multiplex telegraphy, which consists in combining a fundamental voltage wave Iand its third harmonic, deriving two control voltage'waves from the composite wave, each having .peaked waves spaced 180 apart and 90 from the peaks of the other, and full-wave rectifying the displaced and the undisplaced derived waves.

4. The method of producing Ykeying voltages for commutation of signals in multiplex telegraphy, which consists in combining a fundamental voltage wave and its third harmonic in phase opposition at the beginning of each fundamental cycle, deriving two voltage waves from the composite wave, each yhaving peaked waves spaced 180 apart, displacing one of the derived waves 90 and full-'wave rectifying the displaced and the undisplaced derived waves.

5. The method of producing control voltages for commutation of signals in multiplex telegraphy, which consists in combining a fundamental voltage wave yand its second harmonic in phase opposition at rthe beginning of each fundamental cycle to produce a control wave with peaks at 120 and 240, lcombining said fundamental and said harmonic in phase at said bef ginning to produce a control Wave with peaks at 60 and 300 and displacing the phase of said fundamental 90 to produce la control wave with peaks at 180 and 360.

6. The method of producing control voltages for commutation of signals in multiplex telegraphy, which consists in combining a fundamental voltage wave and its second harmonic in phase opposition at the beginning of each fundamental cycle to produce `a control wave with peaks at 120 and 240, ,combining said fundamental and said harmonic in phase at said beginning to produce a control wave with peaks at and 300, displacing the phase of said fundamental 90 to produce a control wave with peaks at 180 and 360, and full-wave rectifying all said control waves.

'7. The method of producing control voltages for commutation of signals in multiplex telegraphy, which consists in combining a fundamental and its second harmonic with relative amplitudes of two to one in phase opposition at the beginning of each fundamental cycle to produ-ce a control wave with peaks at 120 and 240, combining said fundamental and said harmonic in phase at said beginning to produce a control wave with peaks at 60 and 300, displacing the phase of said fundamental 90 to produce a control wave with peaks rat 180 and 360.

8. The method of producing control voltages for commutation of multiplex telegraph signals from aA fundamental voltage wave and its second harmonic, which consists in combining the fundamental land `the harmonic with relative amplitudes of two to one in phase opposition at the beginning of each fundamental cycle to produce a controlwavewith peaks at and 240, combining the fundamental and said harmonic in phase at said beginning to produce a control wave with peaks at 60 and 300, displacing the phase of the fundamental 90 to produce acontrol wave with peaks `at and 360, `and fullwave rectifying all said control waves.

9. A distributor for a multiplex telegraph system comprising a local source of alternating voltage having the channel keying frequency, asecond local source of alternating voltage having la frequency harmonic to said channel keying frequency, means to combine the two voltages to produce a -pair of positive and negative voltage peaks lfor each channel during each cycle of the channel keying frequency, means for rectifying said combined Vvoltages to produce peaks of the same sign, a pair of relay tubes for each channel connected to the signal source and means for applying said voltage peaks to the control electrode of said tubes in succession.

10. A distributor for a multiplex telegraph system comprising a local source of frequency equal to -the channel keying frequency, a second local source of frequency that is a third harmonic of said frequency, and means tocombine the frequencies from said local sources to producefour equally-spaced peaks for each of the keying cycles.

11. *A distributor for a two channel telegraph system comprising a local source of frequency equal to the channel keying frequency, a second local source of frequency that is a third harmonic of said frequency, means to combine the frequencies from said local sources toV produce foreach of said keying cycles four equallyspaced voltage peaks, a utilization circuit for each channel, means to apply the incoming signals to both of said circuits lin common and means to apply two of said voltage peaks spaced a half cycle apart to one circuit and the remaining two voltage peaks to the other circuit.

12,. A distributor for a multiplex telegraph system comprising a local source of frequency equal to the channel keying frequency, a second local source of frequency that is a second harmonic of said frequency, and means to combine the frequencies from said local sources to produce six equally-spaced peaks for each of the keying cycles.

13. A distributor for a three channel telegraph system comprising a local source of frequency equal to the channel keying frequency, a second local source of frequency that is a second harmonic of said frequency, means to combine the frequencies from said local sources to produce for each of said keying cycles six equally-spaced voltage peaks, a utilization cir-y cuit for each channel, means to apply the incoming signals to the three utilization circuits in common and means to apply exclusively to each of said circuits two of said voltage peaks spaced a half cycle apart.

14. In a multiplex telegraph system having a conductor adapted to contain the composite signal and two channel conductors, a local source of alternating voltage having the channel keying frequency, means for producing the third harmonic of said frequency, means for combining the fundamental and harmonic frequencies in phase opposition at the beginning of each fundamental cycle, means for displacing the resultant 90 out of phase, means for full-wave rectifying the resultant and displaced resultant and means controlled by the peaks of the rectified pulses. for connecting the firstmentioned conductor to said channel conductors in succession.

15. In a two channel telegraph system having a conductor adapted to contain the composite signals and a pair of channel conductors, means for producing alternating current waves having the fundamental and the third harmonic of the channel keying frequency, means for combining said fundamental and harmonic Waves in phase opposition at the beginning of each fundamental cycle, means for connecting the first conductor to one channel conductor at the extrema of the resultant wave, means for displacing the resultant 90 out of phase and means for connecting said first conductor to the other channel conductor at the extrema of the displaced resultant Wave.

16. In a three channel telegraph system having a conductor adapted to contain the composite signals and three channel conductors, means for producing alternating current waves having the channel keying frequency, means for producing alternating waves having a frequency that is a second harmonic of the first frequency, means for combining said Waves in phase at the beginning of each fundamental cycle, means for combining said waves in opposite phase at the beginning of each fundamental cycle, means for displacing the fundamental wave 90 out of phase, means for full-wave rectifying the two resultant waves and said displacedk wave and means controlled by the peaks of the rectified pulses for connecting the first-mentioned conductor to said channel conductors in succession.

17. In a three channel telegraph system having a conductor adapted to contain the composite signals and three channel conductors, means for producing alternating current Waves having the channel keying frequency, means for producing alternating current Waves having a frequency that is a second harmonic of the rst frequency, means for combining said frequencies in phase at the beginning of each fundamental cycle, means for combining said frequencies in opposite phase at the beginning of each fundamental cycle, means for displacing the fundamental wave out of phase, means for connecting thefirst conductorto one channel conductor at the extrema of the positive half cycles of the first resultant Wave and of the negative half cycles of the second, means for connecting the first conductor to a second channel conductor at the extrema of the negative half cycles of the first resultant wave and of the positive half cycles of the second, and means for connecting the first conductor to the third channel conductor at the extrema of the positive and negative half cycles of said wave of displaced phase.

18. In a three channel telegraph system having a conductor adapted to contain the composite signals and three channel conductors, means for producing alternating current Waves having the channel keying frequency, means for producing alternating current waves having a frequency that is a second harmonic of the first frequency, means for combining said waves in phase at the beginning of each fundamental cycle, means for combining said waves in oppo site phase at the beginning of each fundamental cycle, means for displacing the fundamental Wave 90 out of phase, means for full-wave rectifying the two resultant Waves and said displaced Wave, whereby six equally-spaced unidirectional peaks are produced in each channel keying cycle, means controlled by the first and fourth peaks for connecting the first conductor to one channel conductor, means controlled by the second and fifth peaks for connecting said first conductor to a second channel conductor, and means controlled by the third and sixth peaks for connecting the first conductor to the third channel conductor.

RICHARD E. MATHES.

Images