US 2401403 A
Description (OCR text may contain errors)
' June 4, 1946.
A. v. B EDFORD 2,401,403
SECRET SIGNALING Filed Dec. 15, 1943 3 Sheets-Sheet 2 Zsnnentor BEDFIJRD Gttorncg June 4, 1946. A. v. BEDFORD SECRET SIGNALING 3 Sheets-sheaf 5 Filed Dec.
- 3nnentor 'ALDH-V. BEDFIJRD (Ittorneg Patented June 4, 1946 SECRET SIGNALING Aida V. Bediord, Princeton, N. 3., assignor to Rsdio Corporation of America, a corporation of Delaware Application December 15, 1943, Serial No. 514,351
The present invention relates to secret telecommunications and more particularly to the method of concealing a message signal by a code or distorting signal and thereafter restoring the message signal at a point of reception.
One object of the present invention is to provide an improved system for secret communication.
Another object is to provide an improved method of transmitting information secretly.
Another object is to provide an improved method of transmitting and receiving a coded signal- Another object is to provide a novel generating and varying a code wave.
. A further object is to provide a novel means of synchronizing a receiving system with a transmitting system so that'a coded message can be made intelligible.
Further objects will hereinafter appear.
The present invention is directed to that class of communication wherein the communication signal S is modified by the code signal K to produce an unintelligible coded signal before transmission. At the receiving station a similar code wave is used to restore the unintelligible signal to its original form. It is difiicult or impossible for the transmitted signal to be made intelligible except by using the proper code wave; In accordance with the preferred form of the invention,
I the code signal K is produced by means of a multivibrator generating narrow pulses at a fixed frequency and operating in combination with a delay network which is excited, for example, by one hundred pulses per second to generate the code wave K.
In one system of this class, the coded wave is an audio-frequency wave having instantaneous ordinates which are proportional to the products of the signal wave S and the code wave K. This coded wave will be designated SK.
At the receiver, the signal S is obtained by multiplying the received signal SK by the reciprocal of the code signal, i. e., SK X 1/K=S. The source of the K wave at the receiver is of the same type as the source for the transmitter including the multivibrator and delay network, but
- the receiver utilizes in addition a circuit to produce the reciprocal of the sign l K n y, the decoding signal 1/K. When this decoding signal l/K is multiplied by the received signal SK, the
means of original signal S is obtained. An essential requirement for decoding is' that the two sources of the code waves K and l/K be synchronized with trigger accuracy, as otherwise, the message cannot be effectively reproduced.
In the accompanying drawings, Fig. 1 is a circuit diagram of one form of code Wave producer; Fig. 2 is a block diagram of a transmitting system embodying one form of the invention; Fig. 3 is a block diagram of a' receiving system operating in conjunction with the system of Fig. 2; Fig. 4 is a group of graphs that are referred to in explaining the production of the code wave; Fig. 5 represents a detail of the brush holder which forms one part of the switch mechanism; Fig. 6 represents a detail of one of the control discs forming the other part of the switch mechanism; Fig. 7 is a circuit diagram of a reciprocal circuit that may be employed in the system of Fig. 3; Fig. 8 is a block diagram of a form of multiplier that may be employed in the systems of Figs. 2'
and 3; Fig. 9 is a graph which is referred to in explaining the invention; and Fig. 10 is a circuit diagram of the multiplier of Fig. 8.
Referring to Fig. 1 of the drawings, one form of network is shown as a source of the code wave K and as this is common to both the transmitting circuit and the receiving circuit. it will be understood as forming a part of that portion of the respective block diagrams designated source of K in each instance. The network is fed by a multivibrator MV of any well-known type supplying narrow repeating pulses at a fixed fre-. quency to the input as shown by wave a in group of Fig. 4. contains eighty sections of series inductance II and shunt condensers IZ. For purposes of illustration, a fragmentary portion of the network is shown as all that is necessary for a clear understanding thereof. In view of the number of sections, it has been found advisable to insert a repeater and an equalizer at sixteen-section intervals to make up for attenuation along the network. As these devices are well'known, it is believed unnecessary to complicate the drawings and description with showing and explanation of them.
Adjustable taps in the form of single-pole double throw switches l3 are included in the respeotive circuit branches of'the inductances II and condensers it. As the switches l3 change from one position to another, either singly or in groups in a predetermined manner, the polarity at any tap it is either positive or negative and hence the voltage at the output P consists of the algebraic sums of the voltages along the network depending upon the switch positions. The common leads from the several taps are connected through In the preferred form, the network I 3 7 but the preferred details of construction and assembly will be disclosed hereinafter.
For selectively operating the switches i3 to obtain the required succession of pulses along the network, a preferred form of control is shown in the fragmentary views, Figs. 5 and 6, wherein a contactor disc [6 is mounted upon a tubular shaft l1 and keyed thereto, in this instance, by a diametrically disposed two-section telescoping pin it normallyexpanded by an intermediate comression spring 20 to project at opposite sides and enter a pair of notches 2i providedin th -inner face of the disc Hi. In effect, the pin it and the notches 2| comprise a detent.
concentrically arranged about the axis of the disc are three series of side face contactors 22, 23 and 24, referably in the form of silver segments insulated from each other and of three types located in irregular order of varying radial dimensions, but all having a portion in the same circular path. Thus, contactors 22 define neutral positions of the switches, contactors 23 define positive positions of the switches, and con tactors 24 represent negative positions of the switches. As shown, the inner edges of the contactors 23 aline circumferentially with the inner edges of the contactors 22, but the radial length of such contactors is approximately twice that of the contactor 22. On the other hand, the outer edges of the contactors 24 aline circumferentially with the outer edges of the contactors 22, but the radial length of such contactors 24 is also approximately twice that of the contactor 22.
' Hence, in the rotation of the disc the contactors 22 are arranged to wipe successively one series of brushes 25; the contactors 23 successively wipe the series of brushes 25, and a second series of brushes 26 in pairs, one brush from each series simultaneously; and the contactors 24 successively wipe the series of brushes 2! and a third series of brushes 21 in pairs, one brush from each series simultaneously.
By referring to Fig. 5 it will be seen that each series has five brushes arranged in sets of three on such an arc as to be wiped by the contactors on the contactor disc ii. The sets are mounted upon an insulating plate 28, which in assembled condition is in close proximity to the contactor side, of the disc I. in order that the respective contactors can contact with and bridge any pair of brushes to be selected to close a predetermined circuit for pulse transmission. Thus, it will be seen that each set of three brushes functions as a single-pole double-throw switch which is open, positive or negative, depending upon which contactor is against the brushes. Each of these switches controls two taps l4 on th delay network through two buffer resistances 30. In Fig. 1, which is a simplified diagram, only one bufl'er resistor is shown to avoid complication. PreferabLv, the order of theconnection between the consecutive taps and the switches is irregular.-
While but one disc l8 and its associated switch plate 28 are shown, it is to be understood that the shaft 11, in the unit here under consideration, mounts a total of eight discs so that this unit has forty switches it. Since each disc 18 can be turned relative to the shaft and held in any selected position by the detent, it is therefore possible to set manually each disc to its proper relative position in setting up a code. When so set, the shaft l1 can be turned step by step to new positions to produce new code waves. In the present instance, the disc is turned in increments of /200 of a revolution at intervals of than the original amplitudes shown by the dotted initially turning the shaft ll manually, the position of this rotary switch is brought into phase with the other rotary switch, it being understood as heretofore pointed out, that the receiving unit embodies one such-switch and the sending unit another such switch. The shaft I! may be'driven by any suitable step-by-step mechanism such, for example, as a solenoid ratchet unit wherein the solenoid is energized by a contact-making clock at intervals of one and one-half seconds.
From the foregoing, it will be seen that conjoint action of the multivibrator MV and the delay network iO serves as the source of the code wave K, whether in the sending unit or in the receiving unit. Fig. 4 illustrates the pulse frequency and how the code wave K is produced by combining with different changeable polarities the outputs from the taps along the delay network. Thus, the wave a is the repeating pulse supplied to the input of the delay network and has a frequency such that immediately after a pulse has reached the far end of the network, another pulse is fed into the near end. Thus, the voltages at tap points a, b, c, d, etc., are like a except delayed various amounts in time as shown at b, c, d, Fig. 4. Wave it illustrates various pulses which would combine to form a wave such as i having a variety of widths of lobes. Thus, a variety of complex waves are producedv having a repetition rate, in this instance, of one hundred times per second by the various switch settings described. Actually, the multi-vibrator MV tends to produce a very narrow repeating pulse, as shown at 1', instead of the smooth wide pulse shown at 0. However, the multivibrator shock excites the first section of the network to produce an oscillatory voltage wave as shown at 1. As the pulse progresses along the network, the phase distortion of the network (which is very pronounced shortly below cut-off frequency) causes the pulse to have an extended oscillatory tail" as shown by the wave m. The voltage at the output of the network, that is point P, is the sum of many components which are, themselves, quite complex in shape, and may have lobes with a large variety of amplitude, as well as widths as shown by wave n. While this wave n might serve as the code wave K, except it would have peak amplitudes far in excess of its average amplitude, the average utilization of the available transmitter power would be very low. Therefore, it is desirable, in order to improve the power efllciency of speech transmission, to limit the high peaks of wave n to amplitudes considerably lower peaks. This is accomplished by a suitable limiter so that the limiting action is gradual to retain some amplification in the code wave K. It is preferable to employ a small shunt condenser after the limiter to smooth over" the. discontinuities in the wave caused by limiting.
For coding and transmitting the me signal 8, reference is bad to Fig. 2 wherein the invention is applied to a radio apparatus wherein a microphone 3| and amplifier 82 are shown applyin the message signal to a multiplier unit 38,. which last may be of various forms, such as shown in applicant's co-pending application Serial No. 458,578, filed August 29, 1942, or as shown 15 in the co-pending application of Frank P. Wipfl,
aromas SerialNo. 484,303, filed April 23, 1943. This latter disclosure is the preferred form of multiplier and will be hereinafter described in detail. The code signal K produced as already described is applied to the multiplier unit 33, simultaneously with the message signal S and the output SK, and the 100 cycle signal is applied to a radio transmitter 36, or to a' wire line, if preferred. The transmitted signal SK may be made to hear so slight a resemblance to either the message signal S or the code signal K that neither one of these signals can be determined by unauthorized persons. The coded wave SK has neither the same shapes, frequencies or amplitudes as the original waves.
For receiving and decoding the signal SK, reference is had to Fig. 3, wherein the scrambled signal is picked up by a suitably tuned radio receiver 35 and applied to a multiplier unit 36, which is, preferably, a duplicate of the multiplier unit 33. The multiplier unit 36 is also supplied with the decoding signal l/K produced by the source of K 3? having its output K applied to a reciprocal circuit 38 having its output connected to the multiplier unit 36. The preferred reciprocal circuit is shown in the co-pending application of RCL 130 above mentioned.
In order that the decoding of the scrambled signal be made with the accuracy necessary for intelligent understanding of the transmitted message, it is important that the two sources of K, (respectively at the transmitter and the receiver) be synchronized with considerable precision. One form of such synchronization is shown in Figs. 2 and 3. wherein the source of K at the transmitter is energized by a 100-cycle sine-wave oscillator at, which also has an output to a mixer 5| located in the SK line. Between this mixer M and the multiplier unit 33, there is a IOO-cycle rejecting filter 38 serving to remove any 100- cycle components of the SK signal which would otherwise interfere with the synchronizing of the receiver. Thus, the signal SK leaving the mixer- 6i, includes the IOU-cycle synchronizing signal for synchronization of the source of K at the receiver. When the signal Skis picked up by the receiver 35, it is branched from the receiver output to the multiplier unit 38, and to a IOU-cycle selector filter 42, which latter synchronizes the multivibrator MV in the source of K to supply the synchronized code signal to the reciprocal circuit 38, and thence to the multiplier unit 36 where the received signal SK is decoded according to the equation SK (l/K) =8. The decoded vsignal S is then supplied to head phones 83 or a loud speaker, as the case may be.
Fig. 7, shows by way of example, one circuit that may be employed to obtain the reciprocal wave l/K. The circuit is described and claimed in the Meneley application heretofore mentioned. The code wave K, which might have a shape as by the curve K in the figure, is applied to the reciprocal circuit 38 at point K. The resistor 44 is of high enough resistance so that the driving source for the non-linear resistance unit 45 is of high impedance whereby there is only a slight variation in the current flow through unit 85. The unit 45 may consist of a pair of copper exide rectiflers 46 and 51 connected to conduct current in opposite directions.
The voltage appearing across the non-linear unit 45 is the voltage m having a slightly fiattened wave form. This voltage is amplified by a cathode-biased vacuum tube 48, which also reverses the polarity.
, 6 The rectangular wave n is produced at the grid of tube 53 by applying the output of the'tube d8 through a blocking capacitor 54 and a high impedance resistor 55 to a pair of diodes 5B and 51 which are connected to conduct in opposite directions. A biasing voltage drop for opposing current fiow through the diodes is produced across the resistor 60 by connecting a source of voltage (not shown) thereacross, aresistor 61 being in series with the voltage source. The condenser 54 becomes charged so that each diode is efiectively biased by half of the voltage drop across resistor 60. The diodes 5B and 5'! become conducting on alternate lobes when the voltage swing exceeds the D.-C. bias voltage. The clipping action is, therefore, equal for the positive and negative lobes of the wave. The rectangular wave n is amplified and reversed in polarity by the tube 53. The wave n and the flattened wave m (both with polarity reversed) are mixed or added together in network oi resistor 50, El and 52 to produce the desired reciprocal wave l/K.
If thewave m is flattened correctly and if the waves m and 'n are added with the correct relative amplitudes by proper adjustment of movable contact 62, the resulting signal will be substantially a true reciprocal of the wave K. The correct shaping of the flattened wave m may be obtained by selecting a non linear resistor unit #55 having a suitable voltage-resistance charac-. teristic and by adjusting the value of the resistor M. Diodes may be used instead of theoxide the reciprocal of an applied wave having one Wave form it will then always produce the reciprocal of an applied wave regardless of its wave form.
One of my preferred designs for the multiplier circuit 33 is illustrated in Figs. 8 and 10, and its mode of operation is illustrated in Fig. 9. The invention makes use of devices, designated Q1, Q2 and Q3, having a square law characteristic when used in a suitable circuit. Most such devices have a characteristic curve of the type shown in Fig. 9, i. e., voltages more negative than Eco (cutofi) produce no output. Hence, they must be operated with a bias voltage Eb of such value that negative signal swings will remain on the characteristic curve. In Fig. 9, El and E0 are the corresponding abscissa and ordinate values, respectively. at any point on the square-law curve. If a signal ea is applied to the input of a squaring circuit" Q3, the instantaneous input voltage will be equal to En+ea The instantaneous output voltage will be given by where A E E the bias measured from cutoli O=A constant If the signal voltage e. consists of the sum of two voltage waves S and K as indicated in Fig. 8, the output voltage will be 2SK is the desired term. The undesired terms are cancelled in the circuit of Fig. 8 by the use of two additional squaring devices Q1 and Q2 which square the waves (A+S) and (A+K), respectively. A- polarity reverse 59 is inserted in the output of Q3 so that cancellation, and not additiomwill result. The various outputs which are supplied through resistors 52, I3 and 54 to a common point are added with the proper polarity as indicated below:
Sum output A 2SK The term A is direct current and is dropped in the first resistor-capacitor coupled stage 101- lowing the multiplier.
Fig. 10 shows a circuit diagram corresponding to the block diagram of Fig. 8 in which vacuum tubes 63, it and 65 are utilized as the squaring circuits Q1, Q3 and Q2, respectively. These tubes must have plate current vs. grid voltage characteristic that follows substantially a square law. Vacuum tubes of the types 6J5, 6SN7, 605, and other triodes have suitable characteristics. The plate resistors for such tubes should be kept small, not more than about 1000 ohms.
The signals 8 and K are applied to the grids of tubes 63 and 65 through resistors 68 and 81, respectively. Both of the signals S and K are applied to the grid tube 64 through resistors 68 and 10, respectively. A suitable negative bias is applied to the grids of the tubes 63, 64 and 85 through grid resistors II and 12. The resistors 66, 61, 88, 10, II and I2 all have the same comparatively high resistance, such as between 0.1 megohm and '1 megohm, so as not to load the comparatively low impedance driving source too much. However, .the resistance is not made so high as to permit excessive phase shift due to grid-cathode capacity. Since the pairs of resistors 66--'H, 68-10, and 51-12 act as voltage dividers for the applied signals, the voltages applied to the grids of tubes 63, 64 and 85 will be ss+K 2 and 2 respectively.
The grid bias (-C) on tubes it, 04 and 65 should be approximately an amount required to cause the tube to operate about the middle of the straight portion oi the grid voltage vs. mutual' I conductance curve. For a GSN'I or 6.15 with 250 volts on the plate, this is about -14 volts.
A vacuum tube 13 serves to reverse the polarity of the output of tube 64 and should have approximateiy unity gain and very low distortion. The outputs of the tubes 03 and s5 and of the reversing tube I3 are added by supplying them to a commoniunction point 14 through the resistors ll,
of from 0.05 megohm to 0.1 megohm, for example. It everythingis properly adjusted, only the desired product term 2SK and the D.-C. component A appear at the junction. In practice it is necessary to adjust the system carefully tor minimum residual signal w e either the signal 8 or the signal K is removed. 7
It will now be apparent that a novel means for producing a code wave has been devised, by nlizing an electrical delay network to delay a pulse wave to form a number or components, reversing the polarity of some of these components-and combining the various components to form the predetermined code wave. cation this code wave is used to modify a message signal into an unintelligible coded signal and transmitted to a point 0! reception At the re- For secret communiception point a second code wave is produced, as a duplicate of the original code wave, which is then converted into its reciprocal wave and used to modify the received coded signal to restore the original message wave which is intelligible. In the production of the second code wave it should be noted that it is important to ensure that this wave be in exact synchronization with the original 'wave.
Having thus described by invention, I claim:
1. The method of producing a code wave which is a function of the voltages produced at successive time intervals in an electrical delay network, which comprises producing a series of pulses, applying said pulses to said network to derive delayed components, changing the polarity of at least some of said delayed components, and combining said delayed components and said delayed components of different polarities to form a code wave.
2. The method of producing a code wave which is a function of the voltages produced at successive time intervals in an electrical network, which comprises producing a series of pulses, applying said pulses to said network to derive delayed components, and combining said components with various polarities to form a code wave.
3. The method of producing a code wave which is the algebraic sum of the voltages produced at different points in an electrical network, which comprises creating a series of discrete electrical pulses, applying said pulses to said network to derive therefrom components at successive time intervals, and combining said components with 'various polarities thereby to form said code wave.
4. The method of producing a code wave which is a function of the voltages produced at successive time intervals in an electrical network, which comprises producing a series of pulses, applying It and 11. These resistors may have a resistance said pulses to said network to derive delayed components, combining said components with various polarities to form a code wave, and periodically changing said polarities in a prescribed irregular order to change said code wave. v
5. The method of secret communication which comprises producing a message signal, producing a series of pulses, applying said pulses to an electrical network to derive delayed components,
changing the polarity of at least some of the delayed components, combining the delayed components and the delayed components of different polarities to form a code wave, modifying the message signal with the code wave to obtain a coded signal, transmitting the coded signal to the point of reception, producing a second code wave at the point of reception which is substantially the reciprocal of the original code wave, and moditying the received coded signal 'by the second code wave.
6. The method of secret communication which comprises producing a message signal, producing a series of pulses, applying said pulses to an electrical network to derive delayed components, changing the polarity .of at least some of the delayed components, combining the delayed components and the delayed components of diflerent polarities to'iorm a code wave, multiplying the message signal with the code wave to obtain a coded signal, transmitting the coded signal to the point ofreception. producin a second code wave at the point of reception which is substantially ,the reciprocal of the original code wave. and multiplying the received coded signal by the second code wave.
7. The method 0! secret communication. which comprises producing a message signal, producing a series of pulses, applying said pulses to an electrical network to derive delayed components, changing the polarity of at least some of the delayed components, combining the delayed components and the delayed components of different polarities to form a code wave, modifying the message signal with the code wave to obtain a coded signal, transmitting the coded signal to the point of reception, producing a second code wave at the point of reception which is substantially the reciprocal of the original code wave, synchronizing the second code wave with the original code wave, and modifying the coded signal by the second code Wave to render the message signal intelligible.
8. The method of secret communication, which comprises producing a message signal, producing a series of pulses, applying said pulses to an electrical network to derive delayed components, changing the polarity of at least some of the delayed components, combining the delayed components and the delayed components of diiTerent polarities to'form a code wave, multiplying the message signal with the code wave to obtain a coded signal, transmitting the coded signal to the point of reception, producing a second code wave at the point of reception which is substantially the reciprocal of the original code Wave, synchronizing the second code wave with the original code wave, and multiplying the coded signal by the second code wave to render the message signal intelligible.
9. The method of secret communication which comprises producing a message signal, producing a. series of pulses, applying said pulses to an electric delay network whereby pulse waves having different delays are produced at the junction points in said network, mixing certain of the pulse waves at said points to form an irregular code wave, altering the message signal by the code wave to form an unintelligible coded wave, transmitting the coded Wave to a receiving station, producing a second code wave at the receiving station which is a function of said first mentioned code wave, altering said coded wave by said sec- 0nd code wave to produce an intelligible signal. 10. The method of producing an irregular code wave which comprises producing a series of regular pulses, applying said pulses to an electric delay network having junction points whereby pulse waves having different delays are produced at said junction points, and selecting and mixing certain of the delayed pulse waves to form said code wave.
11. The method of claim 10 characterized by the further step of synchronizing said regular pulses with the pulses of another similar system, whereby two similar code waves are produced.
12. The method of claim 10 characterized by the further step of reversing the polarity of certain of the pulses before mixing.
13. The method of claim 10 characterized by the further step of periodically changing those delayed pulse waves which are selected and mixed.
14. A system for secret communication comprising means for producing a message signal, a pulse generator for producing a series of pulses, a multisection electrical delay network, means for applying said pulses to said network to derive difierently delayed pulse components at each of said network sections, means for changing the polarity of at least some of said derived delayed components, means for combining said delayed pulse components and said delayed pulse components of different polarity to form acode wave, means for modifying said message signal by said code wave to obtain a coded signal, means for transmitting the coded signal to a point of reception, means for generating a second code wave at said point of reception which is substantially the reciprocal of the original code wave, and means for modifying the received coded signal by said second code wave.
- ALDA V. BEDFORD.