US2293507A - System for transmitting telegraph signals - Google Patents

Description

Aug. 18, 1942. E. HUDEC SYSTEM FOR TRANSMITTING TELEGRAPH SIGNALS Filed Aug'. 23, 1.938 5 Sheets-Sheet 1 l c)l|l|llll' 1942- E. HUDEC 2,293,507

SYSTEM FOR TRANSMITTING TELEGRAPHSIGNALS Filed Aug. 23, 1958 r 5 Sheets-Sheet 2.

afl m I'IIIIIJ m b II I III I I I III l CTIII-II 'I IIIIII'I of! II I I I'I I"I I'LI'LI I m bl III I II'II III C] I I I I I I I I I I I I I II an FL bl I I I I III .I flIIIIIII'IIIIIIIIIIIIIIIIIIIIIII CIIIIIIIIIIIIII'IIIIIIJ 'iIIIIIIIIIIIIIIlIIII IIIIIIIIIII lllllllll III III b a I e f 9 h Aug. 18, 1942. HUDEC 2,293,507

SYSTEM FOR TRANSMITTING TELEGRAPH SIGNALS Filed Aug. 23, 1938 5 Sheets-Sheet I5 Aug. 18, 1942. E. HUDEC SYSTEM FOR TRANSMITTING TELEGRAPH SIGNALS Filed Aug. 23, 1958 5 Sheets-Sheet 4 Aug. 13, 1942. E. HUDEC. zi2'9:', 507

TEM FOR TRANSMITTING TELEGRAPH SIGNALS Filed Aug. 23, 1938 5 Sheets-Sheet 5 Patented Aug. 18, 1942 2,293,507 SYSTEM FOR TRANSMITTING TELEGRAPH SIGNALS Erich Hudec, Berlin, Germany Application August 23,1938, Serial No. 226,259 Germany August 26, 1937 4 Claims. (Cl. 1787.1),

Thisinvention relates to telegraphy on elec-' tric waves and has for its principal object to decrease the distortions caused by fadings and echoes by transmitting the telegraph signals by pulses.

The invention more particularly-concerns a method and means for transmitting and receiving Morse signals, five-unit code signals, abbreviated Morse signals and picture signals.

A further object of my invention is to provide means at the transmitting side for converting telegraph signals and picture signals into pulses.

Another object of my invention is to provide means at the receiving side for producing pulses and for converting the pulses into telegraph signals or into picture signals.

It has been suggested to transmit only the beginning of black and white of a picture by two trains of waves of different frequencies and to change the recording from black to white or vice versa, whenever the frequency corresponding to black or to white arrives at the receiver. Although this method is based on the fact that fadings and echoes are brought about primarily by interference on the part of waves which have left the transmitter one after the other and reach the receiver by different paths, said method has not proved useful. The pulses received have,

been disturbed just as much or more than ordinary telegraph signals.

This difiiculty has beenovercome according to the present invention by producing short telegraph pulses of equal direction at the commencement and (or) at the end of each telegraph signal by transmitting them on one electric wave (one carrier frequency) and by providing means at the receiving end for converting the telegraph pulses into telegraph signals.

If the commencement as well as the termination of each telegraph signal are transmitted by short telegraph pulses on one wave, a relaxation device is provided for producing the telegraph signals at the receiving end, the current of said relaxation device being alternatively cut out and cut in by the pulses received on the same electric wave. A device for carrying out this method, which will be referred to in the follow ing as "double pulse method," is shown in Figs. 3 and 4 and will be explained later.

If only the commencement or only the termination of each telegraph signal is transmitted by a short pulse on one wave, peculiar pulses are produced locally at the receiving end,- these receiver pulses fixing the termination or the commencement of the telegraph signals which are to be reproduced. For carrying out this method, which will be referred to as single pulse method," the speed of the transmission has to be maintained constant with great accuracy.

For normal telegraphic purposes an accuracy of approximately 1 in 10 is sufllcient, as the chronological spacing of the locally generated initial or final pulses can be readjusted without dlfiiculty at the receiver, should the signals no longer have the correct length.

In image telegr phy an accuracy of approximately 1 in 10 is required. With a size of image of 13 x 18 cm. and a; line density of 4 lines per mm. the error in the length of the screen then amounts to approximately 10;. It is much less than the error caused by echoes heretofore in wireless operation.

In the accompanying drawings. embodiments of my invention are illustrated diagrammatically.

Fig. 1 shows a telegraph signal and the telegraph pulses at. the beginning and at the end of .the signal.

Fig. 2 shows a diagram of a device for producing atelegraph pulse at the beginning or at the end of the telegraph signal (single pulse method).

Fig. 3 shows a diagram of a device for producing telegraph pulses at the beginning and at the end of the telegraphsignal (double pulse method).

Fig. 4 shows a diagram of a device used at the receiving end for converting pulses into telegraph signals, pulses transmitted at the beginning as well as at the end of telegraph signals.

Fig. 5 shows a diagram of a device used at the receiving end for converting pulses into telegraph signals, pulses transmitted either at the beginning or at the end of a telegraph signal and pulses produced locally at the receiving end.

Fig. 6 shows the telegraph signals (Fig. 6a), the pulses produced at the transmitter (Fig. 6b) and the pulses produced locally at the receiver (Fig. 6c)

Fig. 7 shows a diagram of a device for reproducing telegraph signals for picture transmisslon.

Fig. 8 shows telegraph signals (Fig. 8a), transmission pulses (Fig. 8b), and receiver pulses (Fig. 8c) for transmitting Morse signals.

Fig. 9 shows telegraph signals (Fig. 9a) transmission pulses (Fig. 9b), and receiver pulses (Fig. for transmitting Morse signals according to another method.

Fig. 10 shows. telegraph signals (Fig. 10a), transmission pulses (Fig. 10b), and receiver method for producing the receiver pulses according to Fig. 110. 1

Fig. 13 shows substituted Morse signals for producing the transmission pulses according to Fig. 11b.

Fig. 14 shows a diagram of a device for producing the telegraph signals at the receiving end.

Fig. 15 shows a diagram of another device for producing the telegraph signals at the receiving end.

Fig, 16 shows a diagram of a device for producing telegraph pulses at the transmitting end.

Fig. 17 shows a diagram of another device for producing telegraph pulses at the transmitting end.

Fig. 18 shows a characteristic of the relaxation device in Figs. 15 and 17.

Fig. 19 shows a diagram for producing transmission pulses for transmitting pictures.

Fig. 20 shows-receiver pulses (Fig. 200.) transmission pulses (Fig. 201)), and telegraph signals (Fig. 206) reproduced at the receiving end.

Fig. 21 shows receiver pulses (Fig. 21a) transmission pulses (Fig. 21b), andtelegraph signals (Fig. 210) reproduced at the receiving end, if the transmission pulses are disturbed by' echoes;

Jig. 22 shows receiver pulses (Fig. 22a), transmission pulses (Fig. 22b), and telegraph signals (Fig. 220) reproduced at the receiving end. The echoes oi the transmission pulses are ineffective by a proper length and amplitude of the receiver m ses. i l

Fig. 23 diagrammatically shows a complete telegraph system operating with signals, as shown in Fig. 10.

Fig. 24 diagrammatically shows a telegraph transmitter for transmitting telegraph pulses, as shown in Figs. 6, 8 and 9.

Referring to Fig. 1a, there is shown a telegraph signal which increases from zero proportionally to time up to'a constant-maximum value and likewise decreases proportionally to time. This telegraph signal can be converted at the transmitter into telegraph pulses by electrical differentiation, as shown in Fig. 1b.

The electrical differentiation "can be carried out in the known manner with the aid of a resistance and a condenser or an inductance. In Fig. 2 the difl'erentiation is shown with the aid of an inductance. In the anode circuit of an electronic valve having a high internal resistance there is a resistance and also a transformer, which is not terminated on the secondary side by a resistance. If u is the control potential, D the reciprocal oi. the magnification factor, R1 the internal resistance and R- the external load resistance of the valve, andM the mutual inductance of the transformer, there results I nd, there is obtained approximately and accordingly the potential at the transformer du u -M a Ii the electronic valve in Fig. 2 is controlled by the telegraph signals there are obtained at the transformer the telegraph pulses according to Fig. 1b.

When using the single pulse method the transmitter is keyed telegraphically by the positive or negative pulse in Fig. 1b. The other pulse remains inefiective. This may be done by applying a suitable grid bias potential to the electronic valve IA of Fig. 2.

When using the double pulse method the negative pulses in Fig. 112 must be converted into positive pulses. For this purpose there is employed, for example, the arrangement accord ing to Fig. 3. The valve l corresponds to the electronic valve I in Fig. 2. The grids of the valves 2 and 3 are connected in push-pull, whilst the anode circuits are connected in parallel with the resistance R. The valve 2 is controlled by the first pulse in Fig. 11), whilst the valve 3 is controlled only by the second pulse in Fig. 1b. The resistance R is traversed both at the beginning as well as the end of the telegraph signal by telegraph pulses of the same direction (as shown by the broken line in Fig. 1b).

The most convenient duration of the telegraph pulses depends on the conditions of transmission. It is approximately of the order of ,400 sec. The desired duration of the pulses can be obtained by a suitable slope oi. the telegraph signal according to Fig.' 1a. A further possibility resides in the control of suitable relaxation oscillation by the pulses at the resistance R in Fig. 3. Other devices are described later on.

In Fig. 4 there is shown an arrangement by means of which the received telegraph pulses can be converted into normal telegraph signals. The arrangement comprises a relaxation device K in conjunction with an electronic valve connection. The telegraph pulses P act on the two resistances R1 and R2, which act as potentiometer. The direct potential acting on the relaxation device (the anode potential U including the grid potential) is situated just between the upper and lower relaxation potentials. To indicate the efiect of the arrangement in Fig. 4 there has been included a numerical example. The efl'ective direct potential amounts to +l0=100 v., the upper relaxation potential to v. and the lower to 90 v. The electronic valve is biased substantially to cutofl.

Owing to the telegraph pulses at R: the potential acting on K is raised to v., so that When the next telegraph pulse is received the condenser C is charged, so that the potential'at R: then has the value of 6 v. indicated in Fig. 4. The grid potential of the electronic valve then rises to the value of -1.0 v., so that a powerful electronic current flows, which produces at B4 a drop in potential amounting to 24 v. The potential acting on the relaxation device K is now 25+90+10--6+15=84 v., and it accordingly relaxes to the lower relaxation current. By means of the condenser C it is again accomplished that the grid potential does not sink too rapidly, so that the anode current continues to prevail until the cutting oil of the telegraph pulse.

In order that the time constant of the arrangement according to Fig. 4 does not become excessive it is desirable for the telegraph pulses to be as short as possible. If desired, they can be shortened in their duration, for example with the aid of a condenser discharge by way of a suitable resistance'notshown.

If merely the commencement or the end of the telegraph signal is transmitted by a pulse, the transmitted pulses in and the pulses us gen-' erated at the receiver are allowed to 'act in opposite directions on the relaxation device K in Fig. 5. As in Fig. 4, the potential U is situated between the upper and lower relaxation potentials of K. The amplitude of the pulses to and u: is so great that the upper potential is exceeded due to the pulses ur and the voltage reduced below the lower potential due to the pulses ua. Thus current flows except when the pulses a: and it: occur simultaneously. The desired telegraph signals result at the resistance R. To permit of adjustment of the proper length of the telegraph signals it must be possible to adjust the telegraph impulse at the receiver by the duration of a signal element.

In Fig. 6a. there are shown various telegraph signals and below the same in Fig. 6b the pulses to be transmitted by the transmitter and in Fig. 6c the pulses to be produced at the receiver. The last-mentioned telegraph pulses always have the same spacing, whilst the spacing of the transmitted pulses (Fig. 6b) varies corresponding to the length of the signals. The telegraph signals shown in Fig. 6a. are, for example, Morse signals.

According to Fig. 6b transmission pulses are transmitted at the beginning of each first and each. even numbered telegraph unit of each spacing interval between the telegraph signals. According to Fig. 6c receiver pulses are produced at the beginning of each odd numbered telegraph unit (signal and interval unit). According to Fig, 6a. the pulses transmitted at the commencement of each interval unit start the intervals of the signals reproduced at the receiver. Since the pulses transmitted at the beginning of each even numbered interval unit and the corresponding receiver pulses act simultaneously on the relaxation device in Fig. 5, they compensate each other in their effect. For reasons of safety the impulse generated at the receiver is made somewhat shorter and somewhat smaller in amplitude, so that the pulse transmitted by the transmitter preponderates.

The receiver pulses (according to Fig. 6b) are produced, for example, with the aid of an enforced relaxation oscillation, which is controlled by the synchronisation frequency. The discharge pulses by way 01' the relaxation device can be employed as telegraph pulses in direct fashion. The position of the pulses must 'be readily capable of being varied by an entire pulse spacing. This is carried out in its most simple form by reversing the phase of the synchronization frequency.

At the transmitter there can be employed the known perforated tape method. The speed of the perforated tape must be regulated to the requisite accuracy.

The method according to the invention is adapted for the transmission of pictures according' to the time modulation process. It is necessary, however, if it is desired to transmit only the commencement or the end of the image-telegraph signals, that the spacing between the commencement or the end of two adjacent image signals always remains the same.

A time modulation process of this kind is described ln detail .ln the Application Ser. No. 180,146-of Dec. 16, 1937.

In such a method, control pulses or screen pulses are developed periodically and are utilized to initiate or terminate reoccurring signals of equal duration or screen period, each comprising a marking interval and a spacing interval. The ratio of these two intervals is determined by the brightness of the picture element to be transmitted and is controlled in accordance with an image signal representative of that brightness.

More specifically, the marking intervals of such signals are initiated at intervals by the periodically reoccurring screen pulses and their termination is controlled by the magnitude of the image potential, or the marking intervals are initiated and controlled by the image potential and their termination is effected at regular intervals by means of the screen pulses.

The arrangement of Fig. 7 has been described in my above-mentioned ccpending application. In this device the condenser C is charged by the direct-current source U through the resistance W1, its potential increasing exponentially. In parallel with the condenser there are two relaxation devices Kr and K2. The upper relaxation potential U: of K: is greater than the upper relaxation potential in of K1; the lower relaxation potential I}: of K: is smaller than the lower relaxation potential 1 of K1: from. gz gn The relaxation device K: is caused to relax at equal intervals by short screen pulses p. In this way the condenser C is rapidly discharged downto the lower relaxation potential 2:.

It may be assumed for the moment that no image signal be applied and the terminals at up be conductively connected. Starting at the beginning of a screen period, the condenser is charged from the lower relaxation potential U: of K: until its potential reaches the upper r'e laxation potential Ch. K1 is then suddenly-traversed by a current, and the potential at the condenser 0 drops until the current i flowing by way of K1 is equal to the' charging current 1. The charging current, the dimensions of the relaxation device Kr and the resistance R are accordingly so chosen that the condenser 0 cannot be discharged by way of K1 down to the lower relaxation potential of K1.

The current i traversing the relaxation device K1 produces at the resistance R a drop of potential iR. It occurs at the moment of relaxation oi K1, and endures for such time until the condenser C at the end of the screen period has been discharged by way of K: to the lower relaxation potential of K2. The potential iR then suddenly returns to zero. This drop of potential supplies the rectangular telegraph signals which are necessary for image transmission on short waves.

The charging current i is so adjusted by the resistance W1 that the upper relaxation potential of K1 is reached just prior to the end of the screen period. The duration of the drop of potential iR at the resistance R then amounts, for example, to 5% of the screen period.

11', now, an image signal us, representative of the brightness of the picture element, is applied between the relaxation device K1 and the condenser C, as shown, the upper relaxation potential of K1 is attained sooner due to the sum of ue-i-ub, so that the duration of the potential drop iR across the resistor It becomes longer according to the brightness of the picture element tobetransmitted: i

The ends of the image'telegraph signals accordingly always have, in this method, the same spacing independently of the light intensity of the picture, whereas at the beginning the spacing varies according to the intensity.

1.When .using the single pulse method these compulses,- which causes-considerable diihculties inthe usual transmission methods.

The telegraph-pulses must be transmitted from the'receiving station to the picture apparatus with a correspondingly high carrier frequency (at least LOW-10,000 cycles), so that there are at least two to three oscillations of the carrier frequencyfor each pulse:

- In Fig. 6'there has already been set forth the manner in which Morse signals can be transmitted according to the invention. In the following figures the transmission of the diiferent telegraph code signals will. be described more thoroughly.

In Fig. 8 the transmission of Morse signals is shown once more. The receiver pulses (Fig. 8c) are generated at the commencement of each Baudot telegraph unit of odd number, i. e. after each first, third and fifth unit of a telegraph signal, and the transmission pulses (Fig. 81)) at the end of each Morse dot and each Morse dash. The Morse signals (Fig. 8a) are brought about at the receiver by the fact that a relaxation device is made, for example, current transmissive by the received transmission pulses and. nontransmissive by the receiver pulses. To produce the spacing of three units between the different letters and the spacing of five units between the words a pulse is transmitted by the transmitter in the intervals at the end of each telegraph unit of even number, i. e., after each second and fourth unit of a telegraph signal, which pulse serves to render the corresponding receiver pulse ineffective.

If the receiver pulses according to Fig. 9c commence with each telegraph unit of even number, transmission pulses (Fig. 9b) are transmitted at the commencement of each first and each second telegraph unit of the Morse dots and dashes.

The transmission of Morse signals according to Fig. 8 or Fig. 9-has the disadvantage that the rate of transmission is comparatively small owing to the large average number of current units in the Morse signals.

This drawback is eliminated according to the invention bythe fact that at the receiver, at the commencement of each signal element, there is produced a pulse which is employed for composing the telegraph signals.

The receiver pulses can have the same direction, as in Fig. 10c, or they can be directed alternately positively or negatively, as in Fig. 11c.

In Fig. 100 there are shown telegraph signals.

according to a five-unit code (the international telegraph alphabet No. 2). These signals are transmitted by transmission pulses at'the commencement of each signal element to the re-- ceiver, where they act in opposition to the receiver pulses on a relaxation device. The phase,

I the duration and the amplitude of the receiver pulses are so adjusted that each transmission -;.-'the relaxation device current-nontransmissive.

If a plurality of transmission pulses or a plurality of receiver pulses act successively on the currentreiaxation device, the second and each following pulse .do not vary their current-transmissive orcurrent-nontransmissive nature.

In order that the telegraph signals will be transmitted properly the received transmission pulses can be extended by echo efi'ect at the most by the duration of a signal element. The dura- -=tion of the signal element must accordingly be selected to be somewhat greater than the maximum duration of echo effect observed.

According to the above method it is possible to transmit Morse signals or also telegraph signals of any desired kind.

In Fig. 110 it is assumed that the receiver pulses are alternately positively and negatively directed. Pulses of this lzind are generated, for example, by means of two enforced relaxation --oscillations, of which the first is controlled by a synchronizing frequency and the second by the pulses of the first relaxation oscillation. The dimensions of the second relaxation device should be such that the condenser is discharged at the end of each second period of the first relaxation oscillation. If the pulses (Land b of the two relaxation oscillations according to Fig. 12 are connected in opposition to each other,

and if the amplitude of the pulses b is made to be twice the amplitude of the pulses a, the deof the Morse code .there is transmitted by the transmitter at, the commencement of the second signal element or at the commencement of the second interval element according to Fig. 11b a short pulse, which has the object of rendering the corresponding receiver pulse inefiective. Since the receiver pulse following thereon also has no effect on the relaxation device it remains in accordance with Fig. 11 current-transmissive -or non-transmissive for three current units or three interval units. For thetransmission of a longer interval, which serves in the Morse code for characterising the spacing between words, a further pulse is transmitted at the commencement of the fourth interval element.

The advantage of this arrangement, which corresponds to theGulstad relay in line telegraphy, resides in the fact that the receiver pulse remains ineffective at the commencement of each third current or interval unit, and that accordingly the received transmission pulses can be extended by echo efiect up to two telegraph units without interference being caused. The

smallest chronological spacing of the transmisslon pulses can accordingly be exactly as great as in the method according to Fig. 10. Consideration has been paid to this in, Fig. 11 by "su' i table'.selection of the time scale. Itshows comparison 'with Fig. that the Morse code according to the above method is somewhat more favourable than the five-letter code.

Inlthe case of the present method the receiver 'T o generate the telegraph pulses at the transm'itterthere. are required constant control pulses spaced according to a telegraph unit; Pulses of this kind are produced in the most simple form with the'aid of an enforced relaxation oscillation, which is controlled by the synchronizing frequency at .the transmitter, These control pulses are allowed to act together with the telegraph signals on the grid of a negatively biased electronic valve; The amplitude of the control pulses and of the telegraph signals and alsdthe grid bias are so selected that by means of the control pulses an anode current is initiated only when the telegraph signal controls the grid positively. In this way, in accordance with Fig. :10, a telegraph pulse is obtained uponeach current unit.

By reversing the directionof the telegraph signals acting on the grid of 'the relaxation de- {vice it c an be accomplished that,- vice versa, a telegraph pulse ls generated upon each interval In carrying out this method of generating the transmission pulses it is necessary that the phase of the-transmission pulses can be displaced in relation to the telegraph signals. The

phase is adjusted so that the pulses are produced approximately at the middle of the signal units. as in this way there can be compensated irregularities in the division and in the speed of the telegraph apparatus. The method according to Fig. 10 permits of the use of all synchronous telegraph apparatus,

particularly of the Siemens high-speed telegraph and also multiplexapparatus. su'ch as the Baudot apparatus or the Western or the Siemens Multiplex apparatus.

If there is reckoned a maximum echo duration of 3 0 sec. (under good conditions of transmission it amounts am, $80.), the signal duration can be set at 3 sec. There are then 250 signal units or 50 signals per second and 500 words per minute, whereas normally in wireless operation onlyf80 words per minute are transmitted.

Since by reason of the pulse method the quality and the speed of transmission are much inchange-over from pure postal eration.

In order that the transmission pulses can be generated in the mannerdescribed also in the method according to Fig. 11 there must be provided on the perforated tape in place of the Morse signals special signals which are produced from the normal Morse signals by the fact that only the second and third units of each Morse dash and only the second and third units of each interval following on a letter are recorded by a corresponding signal. In Fig. 13 the signals for a few letters are given by way of example. The spacing between letters is recorded by a dashjof the duration of two current units immediately appended tothe final signal unit of the word.

In Fig. 14 there isset-forth an arrangement for composing the telegraph signals at the receiver with the aid of a grid-controlled currentrelaxation device. The transmission pulses w and the receiver pulses its act on the controlgrid of the relaxation device in opposite directions. By means of the transmission pulses U1 then per and the lower relaxation potentials 1 and Q are, for example, reduced, and-by. the receiver pulses u: they are increased. If the direct potential U is selected just between the relaxation potentials I? and the relaxation device is made current-transmissive by the transmission pulses and non-transmissive by the receiver pulses; The only condition in this connection is that the amplitude of the pulses is sumciently great. so that the upper relaxation potential'is reduced by the transmission pulses to the amount'of the direct potential U, and that the lowerrelaxation potential is raised by the receiver pulses-to the potential I I-iR, i beingthe lower relaxation current. There then results at the resistance R the desired telegraph signals.

In Fig. l 15 there is shown the settingtogether of the telegraph signals by means of a pot ntial-relaxa'tlon' device K. The same in series .with two parallel-connected electronic valves is joined up with a direct potential U. Theivalve 2 is so biased that the current traversing the potential-relaxation device in the stationary condition is.between the upper and lower relaxation current I v I j In the stationary condition'the electronic. valve (is blocked. It is made current-transmissive by the transmission impulses ur and raises thecurrent traversing the potential-relaxation current. In this way the drop in potential atK'suddenly jumps to a considerably higher value, which also prevails after the end of the transmission pulses.

The receiver pulse's ua control 'thefelectrqnic valve downwards; they reduce the current traversing K to below the lower relaxation current.

The drop in potential at K in this way suddenly goes back almost to zero, and retains thislow value even when the valve 2 has again-become current-transmissive. The potential at K accordingly supplies the desired telegraph signals. It has been described to produce transmission pulses from telegraph signals by electrical .difierthe telegraphic signals and periodic entiation or by controlling a grid of avalve both pulses spaced according toa telegraph unit. a

The first method has thedisadvantage that the-length and amplitude of the telegraph signals can be little controlled, as both depend solely on thecommencement of the telegraph signals.

The second method is a very good one, but it is limited to telegraph signals which begin and end after a whole number of telegraph units.

The following method can be employed for producing transmission pulses from any kind of telegraph signals which start always after a constant time and vary their duration.continuously as in the. case of picture transmission. Those telegraph signals are illustrated by Fig. 20c. This method is an integration method, which consists in the fact that the telegraph sigi being the'telegraph current and Q a fixed amount of electricity. If all telegraph signals have'initially the same disposal independently of their length, the integration period t has the same value. As telegraph impulse there is employed in this method a potential existing beyond the integration period t.

According to the above method there are obtained only the initial pulses for the transmission of telegraph signals. In similar fashion it is possible, however, to produce also the end by reversing the telegraph signals prior tocarrying out the method, 1. e., by previously producing new telegraph signals which extend beyond the duration of the interval between the given signals and which are interrupted for the duration of the given signals. This reversal can be carried out in simple fashion with the aid of an electronic tube or also with the assistance of a relay with a resting contact.

In Fig. 16 there is shown an arrangement for realising the method according to the invention. In the anode circuit of an electronic valve there is a condenser C and in parallel therewith a relaxation' device K. The grid of the electronic valve, which is just well blocked by a negative bias, is controlled by the telegraph signals.

The condenser C is charged by each telegraph signal for such time until its potential reaches the upper relaxation potential I? of K.

The relaxation device then takes up the anode-current and prevents the condenser from being additionally charged. Since the anode-current is interrupted at the end of the telegraph signal, C in the following interval between the signals is discharged to the lower relaxation potential Q ofK.

The same occurrence is repeated at the commencement of the next telegraph signal. The

, condenser is. again initially charged from the lower relaxation potential U to the upper relaxation potential ii. If is is the anode current flowing to the condenser, the duration or the condenser charge t is determined by the equation L'aawcaiy) The integration period t can be adjusted to any desired value by suitable selection of t, C,

15 and E. It is independent of the duration of the telegraph signals, so far as thas the same curve at the commencement of all telegraph signals.

As telegraph pulse there is employed the charging current of the condenser or the drop in potential at the resistance R in Fig. 16. It initiates at the commencement of the telegraph signal and prevails for the time t. Its amplitude remains approximately constant during the charging it the diflerence between the upper and lower relaxation potentials is considerably smaller than the difference between the anode potential and the upper relaxation potential (l7 U-l7) and it the reciprocal of the ampliiication of the electronic valve is small.

The anode current I. is made to be so great in the arrangement according to Fig. 16 that the condenser C is unable to to be discharged to the lower relaxation potential already during the further course of the telegraph signal. This condition is fulfilled if the anode current is greater than the lower relaxation current which traverses the relaxation device K in the case of the lower relaxation potential. I! it is only little greater than the lower relaxation current, the condenser C, as the upper relaxation potential of K is exceeded, will be immediately discharged fairly strongly and its potential drops in the course of the telegraph signal practically to the lower relaxation potential, and after completion of the telegraph signal then only drops by the small amount down to the lower relaxation potential. In this case, shortly after completion of the telegraph pulse, an oppositely directed discharge current flows from the condenser C by way of the relaxation device. Later, at the end of the Y telegraph signal, there merely flows a small discharge current.

This condition varies all the more the greater the anode current i. may be in relation to the lower relaxation current. If la is still greater than the upper relaxation current, the condenser at the end of the telegraph pulse will be charged to a slight extent until the anode current a is equal to the relaxation current flowing by way of K. In this case C is not discharged until the end or the telegraph signal.

is is preferably selected between the lower and upper relaxation currents in order that after completion of the telegraph pulse there flow only discharge currents which are directed in opposition to the telegraph pulse and can readily be rendered ineffective.

{mother arrangement for carrying out the method according to the invention is illustrated in Fig. 17. The telegraph signals are connected to a series connection comprising an inductance and a potential-relaxation device.

The characteristic of the potential-relaxation device is illustrated in Fig. 18. It corresponds to the characteristic of a current-relaxation device, with the exception that current and potential are changed about. As the. current i increases the potential at the potential-relaxation device is at first very small. If, however, i exceeds the upper relaxation current I, it suddenly jumps 'to the very much higher value f]. It i is again reduced, it diminishes only very little to the value 2, and it is only then that it suddenly jumps back to the original small value after i has dropped below the lower relaxation current I.

In the arrangement according to Fig. 17 the amplitude U of the telegraph signals is preferably selected between the upper and lower relaxation potential I I U fi.

At the commencement of a telegraph signal practically the entire telegraph potential it acts on the inductance, as the drop in potential at the potential device is only small, There is accordingly approximated The current 1. traversing the inductance L increases for such time until it reaches the upper relaxation current I. Since at this moment the potential at the relaxation device jumps to the value U U, the potential at the inductance becomes uz.=U-Z'I O and has the effect that the current i slowly drops. It drops to the value Io in Fig. 18, for which the potential as at the relaxation device is equal to the telegraph potential U.

This condition prevails until the end of the telegraph signal. Since at this moment the telegraph potential drops to zero, the total relaxation potential acts in a negative direction on the inductance L and causes the current to decrease to the value I At this moment u: jumps from Q practically to zero.

At the commencement of the new telegraph sig nal i again increases from zero up to the value I. The time necessary for this is determined by the equation It can be adjusted to any desired value by selection of L and the characteristic of the relaxation device.

During this time the potential at the inductance L is approximately equal to the telegraph potential, during the further course of the telegraph signal it is approximately equal to zero and after completion of the telegraph signal for a certain time approximately equal to the negative value of the telegraph potential, and following thereon up to commencement of the new telegraph signal it is approximately equal to zero. Since the small deviations from zero have no bearing, and since the negative potential can be readily suppressed, the potential occurring at the inductance L is suitable as a telegraph pulse.

A small fluctuation in the pulse duration t is brought about by the fact that the characteristic of. the potential-relaxation device in the first ascending portion does not run as in the majority of current-relaxation devices along the ordinate axis, and that the losses in a currenttraversed inductance are considerably greater than those in a charged condenser. Upon the return relaxation the current I does accordingly not prevail until the commencement of the next telegraph signal; it continues slowly to decrease to an extent which is all the greater the longer the period of time may be until commencement of the next telegraph signal. The time t is accordingly satisfied only approximately by the last equation. It is determined more accurately by the equation I... being between E and O. The time t is accordingly dependent to a smaller extent on the length of the telegraph signals. Since, however, a small fluctuation in the duration of the telegraph pulses does not cause interference, this embodiment Of the integration method is also well suit able for the generation of telegraph pulses.

In Fig. 19 there is shown the complete execution of the method according to the invention for the example of image telegraphy according to the time modulation method. The connection system is divided into three parts, which are separated from each other by broken lines. The first part serves for generation of the timemodulated signals from the image current (which is supplied by the image-scanning device), in the second part there are generated the telegraph pulses, and in the third part they are converted into carrier-frequency pulses, so that they can be transmitted without distortion by way of a wire to' the receiver or to the wireless transmitter.

The first part of the connection system, which is similar to Fig. 7, comprises in substance a condenser C1, which is charged by way of a resistance W1, two relaxation devices connected in parallel with the condenser, e. g. the glow lamp G1 and a valve-relaxation device, which is formed by the pentode I, the resistances W: and W4 and the coupling condenser K. The image-telegraph At the beginning of a screen period, the con- I denser C1 starts to charge by way of resistors W1 and W: until the sum of the condenser voltage, and the image signal an have reached the upper relaxation potential of the glow lamp G1. The latter then becomes conductive and a voltage drop is developed across the resistor R1. The instant at which G1 becomes conductive is determined by the image signal us. a

An equilibrium condition is established. in which the current through the glow lamp and the condenser changing current are equal, so that the condenser C1 does not become. dis-. charged below the lower relaxation potential of G1. As long as these conditions prevail, a constant voltage drop'is developed across the resistor R1.

At the end of the screen period a positive screen pulse p is applied to the control grid of tube l, causing it to relax. The suppressor grid of tube l is normally held negative with respect to the cathode. The condenser K is normally charged and the tube cut oil. As soon .as a positive screen pulse is applied to the control grid, the tube draws current and screen current flows, which discharges the condenser K. The discharge current through the resistor W3 renders the suppressor grid more positive, thus aiding a sudden flow of maximum plate current through the tube, characteristic of a relaxation device.

As soon as tube l draws current, the condenser C1 becomes discharged below the lower relaxation potential of glow lamp G1 and the latter relaxes to the condition of non-transmission oi current, which terminates the voltage drop across the resistor R1.

Their end is initiated by periodically occurring scanning line pulses p and their commencement by the image signal an, which is generated by an image-scanning apparatus oi the usual kind. The initiation of the signals occurs all the earlier the higher the image potential may be. If the image to be transmitted is scanned in a dark section, there results at the .resistance R1 a short practically rectangular signal. The brighter the image the longer the image-telegraph signal at R1 will be.

The second part of the connection system according to Fig. 19 comprises the arrangement according to Fig. 16 and an amplifying valve following thereon. The grid of the second valve 2 is controlled by the image-telegraph signals which occur at the resistance R1. As currentrelaxation device there may be employed a glow lamp G2.

At the resistance R2 there occur the desired image pulses, which initiate simultaneously with the image-telegraph signals. The brighter the picture, the earlier they begin. They are negatively directed. -At the end of the pulse and at the end of the image-telegraph'signal there occur at R: short positive pulses, which are caused by the discharge of the condenser vC2. The grid of the thirdtube 3 is positively biased, so that the positive pulses are suppressed by the grid current.

The potential at theresistance R3 controls the grids 'of the two' electronic valves 4a and 4b, which formthe third" part of the arrangement; In

" addition-to the telegraph pulses there is also conwhilst the alternating, currentsare added together, so that at the output of the transformer ii; carrier-frequency image pulses are obtained, which can be transmitted by a line'to the receiver or by a line to the radio transmitter.

If picture signals are to be transmitted the duration and the amplitude of the receiver pulses are preferably so adjusted that each transmission pulse received makes a receiver pulse ineffective and also the relaxation device Current'transmissive, and that each receiver pulse alone makes the relaxation device current-non-transmissive.

Ilia transmission pulse is received at the moment when a receiver pulse begins, the length of the picturesignal producedat the receiver is zero.

If the transmission pulse arrives a little later at the receiver the length of the picture signal increases. In order that the picture signals will be produced properly at the receiver the received transmission pulses must arrive early enough that they do not extend by echo efiects to the commencement of the next receiver pulsess Thereforethe length of the telegraph signals is limited between zero and a maximum value, which often is much smaller than the length of the screen.

As it is very important to reproduce pictures with good details in the black parts, the short picture signals must correspond to the black parts of the picture. Therefore the pulses produced by a black part of the picture must begin earlier than those produced by a white part. a

The transmission pulses produced by a black part according to Fig. 19, on the contrary, begin later than those produced by a white part. This is illustrated in Fig. 20, in which a are the pulses generated at the receiver at the commencement of each telegraph signal,

b are the transmission pulses passed from the transmitter to the receiver (the pulses at the left and at the right are produced by a bright part of the picture, the pulses in themiddle are produced by a black part of the picture),

c are the telegraph signals which are'formed from the pulses a and b.

The transmission pulses in Fig. b are as long as the receiver pulses in Fig. 20a, but if they are expanded by multipath-echoes, wrong signals may be reproduced.

A case of this kind is illustrated in Fig. 21. The received pulses in b are considerably wider 'mitter to the receiver without-echoes.

than the receiver pulses in a. It will be seen from Fig. 21 that the first three short telegraph signals are not affected by the widening out of the transmission pulses. They have the same duration as if the pulses were transmitted from the. trans- Of the long telegraph signals, however, only the first is reproduced, and all of the remainder. are eliminated. as'the receiver pulses are rendered ineffective by the greatly extended transmission pulses. The same applies to the first short telegraph signal following the long signal in Fig. 2. Only the second and the following short telegraph signals are correctly reproduced.

These difiiculties can be partially eliminated if the received transmission pulses are shortened in their duration prior to the action on the relaxation device in Fig. 5,- 14 or 15, so that there are obtained approximately the pulses, shown in Fig. 20.

This method, however, is not completely reliable, since thereceived transmission pulses may also be broken in their duration. This is pos-- sible, for example, if the transmission pulses are shorter or longer than the difference in the time of transit of the waves 1, II,-lII, etc., of the different multipaths. In the event of their being shorter an interruption occurs continuously between wave trains. If the same are longer, the received pulses'are interrupted on eachoccasion when the waves I, II, III, etc., have approximately the pulses, and several pulses accordingly remain.

Theymay produce considerable interference in similar fashion as the extended'pulses in Fig. 21. To eliminate the interference according to Fig. 21 the amplitude of the receiver pulse is made so great that their eifect on'the relaxation devlcein Fig.'21 is not eliminated by the transmission pulses received, and the duration of the receiver pulses is made so long that the transmission pulses acting on the relaxation device never extend beyondthe next receiver pulse.

In Fig. 22 it is shown in what manner'it is-accomplished by selection of the receiver'pulses in accordance with the invention that the telegraph signals are reproduced in faultless manner exactly as in 20 with reception of the transmission pulses without echo.

The long telegraph signals in Fig. 22 exactly in the manner of the long signals in Fig. 20 could be longer to the extent of fully 5% without being falsified, The short telegraph signals in Fig. 22 'could be exactly 5% shorter. On the otherhand the short signals in Fig. 20 could have been more than 15% smaller without being falsified owing to the widening out of the telegraph pulses."

The highest permissible telegraphy speed is arrived at by the condition that the transmission pulses arriving at the receiver should not merge one into the other. Since the transmission pulses are situated most closely together in the cases illustrated at the end of Figs. 20, 21 and 22, wherein a short telegraph signal follows immediately on a long one, the chronological spacing of the transmission pulses must be in this case somewhat greater than the duration of the received transmission pulses. From the broken lines between Figs. 22b and c it is to be seen that-the total of the shortest duration of interruptionand the shortest duration of the telegraph signal must be at least equal to the duration of the received transmission pulse.

If, for example, the incoming transmission pulse has a duration of /400 sec. and if the smallest duration of interruption is to amount to and the shortest duration of the telegraph signal to 20%, the time between one receiver pulse and the next must amount to at least 4X /4o0= /1o0 sec.

In Fig. 23 there is shown a system for transmitting intelligence telegraphically comprising a transmitter and a receiver.

For the purpose of translating the intelligence to be transmitted into telegraph signals, there is provided a teletyper l0 ofi'conventional construction in transmitting position. A synchronizer II is provided for synchronization of the teletyper l0 and of a control pulse generator l2. The synchronizer II is of conventional design and may comprise a tuning fork, preferably synchronized with the power mains (not shown). The control pulse generator I2 is also of conventional design and preferably comprises a relaxation device such as a multivibrator.

A converter i3 is provided for converting the telegraph signals into telegraph pulses. This converter preferably comprises an electronic valve having a negatively biased grid to which the control pulses and the telegraph signals are applied in positive polarity,.as explained in connection with Fig. 10. For the purpose of applying telegraph signals and control pulses thereto, the converter I3 is connected to the control pulse generator I2 and the teletyper ill. The output of the converter I3 is connected to the modulation input circuit of a radio transmitter H for modulating or keying a carrier wave in accordance with the telegraph pulses to be transmitted.

At the receiving station, there is provided a radio receiver of conventional design, preferably of the superheterodyne type, for receiving, amplifying and demodulating the received modulated carrier wave. To the output side of the receiver 20 there is connected the input side of a converter 2i, of the type shown in Fig. 14 or 15, for converting the received telegraph pulses into telegraph signals. For the purpose of generating control pulses of the same frequency and of equal phase as those at the transmitter, there is provided a synchronizer 23, which may comprise a' tuning fork, preferably synchronized from the same power mains (not shown) as the synchronizer II at the transmitter. Connected to the output side of the synchronizer 22 there is connected the input side of the converter 2 I. For reproducing the intelligence to be received, there is provided a teletyper 24 connected to converter 2| for receiving telegraph signals.

In operation the teletyper ID generates telegraph signals as shown in curve a of Fig. 10. The control pulse generator l2 generates periodic control pulses, as shown in curve c of Fig. 10, with a periodicity of one telegraph unit. Both the telegraph signals and the control pulses are applied with positive polarity to the converter l3 and more specifically to the control grid of a negatively biased converter tube therein which generates telegraph pulses whenever marking units of telegraph signals and control pulses occur simultaneously.

In this manner a series of telegraph pulses, as shown in curve b of Fig. 10, are developed which are applied to the radio transmitter H to modulate or key the carrier wave developed therein. The modulated carrier wave is then radiated.

At the receiver 20, the modulated carrier wave is received and demodulated and .the telegraph pulses developed thereby are applied to the converter 2| in one polarity; The control pulse generator 22 generates control pulses, as shown in curve 0 of Fig. 10, which are applied to the converter in opposite polarity. The control pulses are of the same frequency as those generated at the transmitter, and are held in synchronism therewith by means of the synchronizer 23, which, in turn, is preferably synchronized by the same power mains at the synchronizer H, in order to maintain the proper phase relation between the received telegraph pulses and the control pulses generated at the receiver. This presents no difliculties since widespread power systems are operated in perfect synchronism.

The telegraph signals developed in the converter 2| are applied to the teletyper 24,- of conventional construction, which is synchronized by the incoming signals, 'and which reproduces the intelligence to be received.

In Fig. 24 there is shown a complete transmitter for transmitting telegraph pulses of the type illustrated in Figs. 6, 8 and 9, which comprises a teletyper 30 of conventional construction, synchronized by a synchronizer 3| which, in turn, is preferably synchronized by the power mains (not shown) as is well known in the art. It maybe assumed that the telegraph signals developed by the teletyper are to be converted into telegraph pulses according to curve b of Fig. 9, comprising actuation pulses occurring at the beginning of each marking unit of the telegraph signals and cancellation pulses occurring at the end of each odd unit of a telegraph signaland in the marking unit thereof. For the purpose of developing the actuation pulses at the beginning of each telegraph marking unit, there is provided a converter 33 comprising a device substantially as shown in Fig. 16. For the purpose of producing cancellation pulses occurring at the end of each odd unit of a telegraph signal and within the marking unit thereof, there is provided a control pulse generator 32 preferably comprising a 'multivibrator of well-known construction capable of delivering control pulses according to curve 0 of Fig. 9 at a periodicity of two telegraph units and coinciding with the ends of the odd telegraph units. For the purpose of synchronizing the control pulse generator 32, it is connected to the synchronizer 3|. For developing cancellation pulses according to curve c of Fig. 9, there is provided a second converter 3'4 preferably comprising an electronic valve having. a negatively biased grid to which the telegraph signals and the control pulses are applied in posi tive polarity. For transmitting complete telegraph pulses comprising actuation and cancellation pulses, the output sides of converters 33 and 34 are connected together and connected to a radio transmitter 35 for the purpose of modulating or keying the carrier wave generated therein. The modulated or keyed carrier wave is then radiated.

Likewise, telegraph pulses, according to curve b of Fig. 8, can be transmitted by means of an arrangement, as shown in Fig. 24. In this case the converter 33 is arranged to produce actuation pulses at the end of each telegraph marking unit, while the control pulses generated by the generator 32 are adjusted to be in phase with the begirming of the odd telegraph units of the telesgraph signals. The telegraph signals are applied in reversed polarity to the converter, that is,-

the spacing are of positive polarity. with these slight adjustments of the arrangement oi Fig. 24, telegraph pulses, in accordance with .curve I) of Fig.8, can be transmitted.

cancellation pulses. These terms are charac teristic of the action of these pulses at the converter of the receiver, As described above, both telegraph pulses and control pulses are applied thereto to convert the telegraph-pulses into telegraph signals. The actuation pulses are not coincident with control pulses and, therefore, actuate the converter. The cancellation pulses are coincident with certain control pulses and cancel their eifect upon the converter, since it'is undesired at those particular instants. This will be obvious to'those skilled in the art from an inspection of Figs. 6, 8 and 9.

It will be understood that the teletyper ID of Fig. 23, as well as the teletyper 30 of Fig. 24, can

be replaced by any other conventional means suitable for converting intelligence into telegraph signals, or it may be replaced by a circuit, as shown in Figs. 7 and 19, which is capable of producing a telegraph signal representative of conditions of light and shade of an elemental area of a picture to be transmitted. In the latter case, the screen pulses can be produced by a control pulse generator of the same type as the generator iii of Fig. 23 and the generator 32 01' Fig. 24.

I claim as my invention: 7

1. A transmitter for carrier wave transmission of intelligence comprising means for developing electrical signals representative of the intelli gence to be transmitted, 'said signals having a substantially rectangular wave shape having unequal and relatively long intervals of a first character and of a second character, means for deriving from said signals unidirectional pulses of relatively short and equal durations coincident with the changes of said wave shape from said first character to said second character, and means for modulating a carrier wave in accordance with said pulses.

2. A transmitter for carrier wave transmission of intelligence comprising means for developing electrical signals representative of the intelligence to be transmitted, said signals having a substantially rectangular wave shape having unequal and relatively long intervals of a first character and of a second character, means including a glow discharge device for deriving from said signals unidirectional pulses of relatively short and equal durationcoincident with the changes of said wave shape from said first character to said second character, and means for modulating a carrier wave in accordance with said pulses.

3. A transmitter for icarrier wave transmission of intelligence comprising means for developing electrical signals representative of the intelligence to be transmitted, said signals having a v substantially rectangular wave shape having unequal and relatively long intervals of a first character and of a second character, a condenser, an electron discharge device controlled by said signals for charging said condenser, a glow discharge tube connected in parallel relation to said condenser for discharging said condenser, a resistor connected in series relation with said condenser for developing a voltage drop in accordance with the current through said condenser, a second electron discharge device connected in circuit 50- as 'to exhibit a grid current limiting action, means for controlling said second device in accordance with said voltage drop to produce unidirectional pulses of relatively short and equal durations coincident with the changes of said wave shape fromsaid first character to said second character, and means for modulating a carrier wave in accordance with said pulses.

4. The method of carrier wave transmission of intelligence comprising the steps of developing electrical signals representative of the intelligence to be transmitted, said signals having a substantially rectangular wave shape having unequal and relativelylong intervals of a first c" acter .and of a second character, deriving 11. said signals unidirectional pulses of relativel short and equal durations "coincident with the changes of said wave shape from said first character to said second character, and modulating a carrier wave in accordance with said pulses.

= ERICH HUDEC.

v I CERTIFICATE OF CORRECTION; Patent No. 2,293,507. Au uet-ls, 19!;2.

- Enron HUDEC.

It is hereby certified thet error appears in the printed afiecificatibn o'tthe above numbered patent requiring correction emu follows: Page 2 sec- 0nd column, line 55, for the wordi'circle' read -ci'rcuit {page 8, sec

and column, line29, berere' "wave" insert --the different"; and that the said Letters Patent should be read with 12114.8 correction therein that the eams may confonn to the record of the case in the Patent Office.

Signed andeealed this 27th de of October, A. D. 19!;2.

7 Henry Van Arsdale,. (Seal) Acting Commieeioner of Ptents.=

Images