US2437707A - Communication system employing pulse code modulation - Google Patents

Description

March 16, 1948. J. R. PIERCE 2,437,707`

COMMUNICATION SYSTEM EMPLOYING PULSE CODE MODULATION By zdf/naw' ATTORNEY Marh 16, 1948. J. R. PIERCE COMMUNICATION SYSTEM EMPLOYING PULSE CODE MODULATION 5 snets-shet 2 Filed Dec. 27, 1945 WW Hl w E I+ www@ ww?. mm Q s). t W mw Y QM Wonu TJ. IT .35.5% F. .r J W e@ l r lr L| M P M DVD l Y T www www www www ArTOR/VEV March 16, 1948. J. R. Flr-:RCE

COMMUNICATION SYSTEM EMPLOYING PULSE CODE MODULATION Filed Dec. 27, 1945 3 Sheets-Sheet 3 /A/ VEA/TOR J R. P/ERCE ATTORNEY Patented Mir.'` 16, 194s OFFICE-t COMMUNICATION' SYTEM EMPLOYING PULSE CODE MODULATION v John R. Pierce, Millburn, N. J., asslgnor to Bell v Telephone Laboratories,

Incorporated, New

York, N. Y., a corporation of New York Application December 27, 1945, Serial No. 637,386

13 Claims. (Cl. 179-15) 'I'his invention relates to a communication system for the transmission of complex wave forms ofthe type encountered in speech, music, sound, mechanical vibrations and picture transmission by means of code groups of a uniform number of signal pulses of a plurality of different types of signaling conditions transmitted at high speed. Systems of this type have sometimes been identiiied as PCM or pulse code modulation systems.

The object of the present invention is to provide a communication system capable of transmitting and reproducing a complex wave form over an electrical transmission path such as a microwave radio path, a wave guide, cable circuit, etc., in such manner that the signal-tonoise ratio of the received signal is materially improved over the signal-to-noise ratio of the transmission path.

Another object of this invention is to provide improved and simplified apparatus capable of transmitting and receiving signal pulses over a path having a low signal-to-noise ratio.

More specifically, it is the object of the present invention to provide circuits and apparatus for transmitting a group of pulses in sequence over a given channel representative of the amplitude of a complex wave at successive instants of time.

Still another object of the present invention is toI transform a series of pulses representing the amplitude of a complex wave at a giveninstant of time into a single pulse having an amplitude which is a function of the amplitude of the original complex wave at the given instant and to recombine a succession of such single pulses of varying amplitude in a manner to reconstruct a wave form of substantially the same shape as the wave form to be transmitted.

A feature of the invention is that of accomplishing the above objects with a small number of pulses requiring a minimum of apparatus and equipment, taking advantage of the considerable distortion permissible in speech without loss of intelligibility The invention is similar in some'- respects to my copending application, Serial N0. 592,961, iled May 10, 1945, and Serial No. 603,934, led July 9, 1945, as will be pointed out hereinafter.

Other features of the invention relate to synchronizing and coordinating the `various circuits and equipment at the transmitting terminal with each other and with the circuits and equipment of the receiving end so 'as to secure proper-operation of the system as a whole.

Briei'ly, in accordance with the present invention, equipment is provided for generating a group of control pulses of predetermined time relation one with another. Apparatus-is also provided for rapidly sampling a wave function, the sampling v means beirg under control of the controlpulse generator. Furthermore, for each of the control pulses, a code element timing circuit generates a-series or cycle of code element timing pulses and these in combination with the sample amplitude first test the polarity of the sample and then compare itsamplitude under different conditions with known voltages..

In the meantime there is set up and transmitted a code of pulses corresponding to the polarity information and to the. amplitude information, which code thus characterizes the polarity and the amplitude of the complex wave sample. At the receiving station a control pulse generator, code element timing generator and associated apparatus are provided whereby each cycle or group of received pulses is employed to produce a pulse having a magnitude proportional to the magnitude of the complex wave sample at the transmission end of the system. 'I'he complex wave is then reconstructed from a succession of such reproduced pulses of varying magnitude.

A special feature of the system consists in measuring the sample amplitude on a decibel basis and in this respect the system is similar to that described in my application Ser. No. 592,961 as contrasted with the system described in my application Ser. No. 603,934 which operates on` a. linear or proportional basis. A

'I'he invention both as to its organization and method of operation, together with other features thereof, will be better understood from thefol.- lowing description when read with reference to the accompanying drawings in which:

Figs. 1 and 2 show in functional block form the various elements of the transmitter and of the receiver and the manner inwhich they cooperate to form an exemplary communicating system embodying the present invention;

Fig. 3 illustrates the timing and nature of the pulses charcteristic of the system at the transmitting end;

Figs. 4A and 4B are explanatory of the operation of the system; and

signal function may be obtained from any suitable source either directly 4or over any suitable transmission path. The signal source and transmission path may include the usual types of transmission, switching, and repeating equipment and operate in their usual and well understood manner.

A pulse generator CTi sets up a plurality of cycles of' pulses, the cycles coming in rapid succession, perhaps about 8,000 persecond. The rst pulse of each cycle, which is a relatively broad positive pulse, activates an electronic switch which causes the signalling wave or function to be sampled, whereby a storage condenser C is charged to a potential equal to the amplitude of the wave function at the instance of the pulse. From the electronic switch there is then sent to the transmitter information as to the polarity of the sample signal. This rst pulse of each group or series is also applied to a coding circuit to initiate action therein or operation thereof. The next two pulses, which are narrow positive pulses, from the generator are applied to the coder and give rise to information, to be transmitted, bearing on the amplitude of the sample. Finally va negative fourth pulse from the generator resets the circuit in preparation for the next cycle of operation. As a result of these operations there is passed on to the transmitter a binary representation of the polarity and amplitude of the sample digit by digit, the first digit defining the polarity and the next two the amplitude on a non-linear or decibel basis.

At the receiver these pulses are received and trigger oil a pulse generator to generate a series of pulses similar to and in synchronism with those at the transmitter. These pulses, cooperating with the corresponding received signal pulses, adjust the gain in decibels of a train o'f amplifiers to a total gain corresponding to the decibel amplitude of the original sam'ple with respect to its reference voltage. Following this a pulse from the generator is impressed on the input of the train for amplification and at the same time the polarity pulse, appropriately delayed, operates to control the direction of the output. A suitable resetting pulse is also provided. The output resulting from the successive cycles are then passed through a low pass lter to substantially reproduce the original wave.

In speaking of transmitted or received pulses it is understood that reference is had to two signaling conditions, such as +1 and -1 or 1 and 0, but even in the latter case an ol or signal is a signal pulse.

A more detailed understanding of the transmitter end of my system will be had by reference to Fig. 5. This circuit shows the sampling apparatus and the coding unit along with the pulse generator, but since the operation of thefirsty two are dependent on the pulse generator and since ,the generator operation is independent of these two, it will be described first.

Pulse generator The controlling element in the pulse generator is a relaxation oscillator comprising a gas tube I0. This oscillator is of a `form Well known in the art and includes a resistance 5I I for charging a condenser 5I2. Assuming that the condenser 5I2 is discharged, then on closure of the circuit it is charged at a rate determined in part by the resistance 5I I. When the potential of the condenser and the plate of tube 5I0 rises to a firing value the vcondenser suddenly discharges through the tube and resistor 5I4. The duration of the discharge is short and gives rise to a sharp positive pulse across resistor 5I4. The duration of this pulse and the rate at which it is followed 'by identical pulses can be completely controlled by the parameters of the circuit; in particular by the values of the elements 5II to 5I4 taken with the potential of the grid of the tube 5I0 as determined by the potentiometer 5I5. Any of several forms of relaxation oscillators will function equally well for the controlling oscillator or element and may be so used, the one shown being simple and satisfactory.

The positive pulse formed across impedance element 5I4 is now used to control the emission of pulses to various parts of the circuit. This is accomplished by connecting across the element 5I4 a delay line, network, circuit or time stick 5I6. A suitable form of delay device or network is shown in the drawing and is made up of sections of inductance and capacitance connected seriatum. The positive pulse travels down this delay network and `upon the arrival at each section gives rise to corresponding positive pulses at points I to 4 of the network spaced in time one after another by an interval determined by the number and contents of the elements in the preceding sections of the time stick. These pulses will control the operation of additional pulse circuits and, through them, the sampling circuit and the coder, as will be pointed out hereinafter. The delay network is terminated by a load /5I8 of proper value to suppress any relected wave.

The parameters of the relaxation oscillator' may be adjusted so that pulses are delivered across resistor 5I4 at any frequency desired. For the purpose of my invention it is preferred to have a sampling frequency higher than that of the highest frequency component in the complex wave to be` transmitted. If, for example, this Wave is to be a speech wave and it is desired to transmit all components up to 3,000 cycles then there should be at least two samples per cycle for this highest frequency component. A suitable value, therefore, for the relaxation oscillator frequency which becomes the sampling frequency, would be 6,000 cycles although a higher value may be used if desired.

At the time tm when the pulse is formed across 5I4 it is transmitted immediately to triode 520 and appears at the plate thereof as a reversed or a negative pulse. Furthermore, by virtue of the condenser 522 in parallel to the load resistor 523 the length of the pulse is increased to correspond to the initial wide pulse Pi shown in Fig. 3 for purposes to be hereinafter described. This pulse is now applied to the grid of triode 525 and appears as a positive pulse, amplified if desired, across the load resistor 526.

When the pulse reaches the point 2 on the time stick, it is transferred to the grid of triode 530 and appears as a positive pulse across the cathode resistor 53|, which pulse is then available for use. Also, when it reaches the point 3 on the time stick it is applied to grid of tube 540 and appears as a positive pulse across cathode resistor 54|. Finally when it reaches the point 4 on the time stick it is applied to the grid of tube 550 and appears as a negative pulse across the plate resistor 526.

Sampling circuit Associated with the sampling circuit of Fig. 5

now to be described there is shown a'soui'ce of complex wave -to be transmitted, such as 'a speech wave. This source may comprise a microphone or other suitable source 566 and suitable terminal equipment 56| including a. transmission line to transfer the signal function M to the trans"` forr er 562. In sampling the signaling wave the Y Vend the mid-point of the secondary of transformer 562 is connected to ,ground through resistor 510 and the condenser is supplied (from both ends of the secondary through the pair of diodes 51| and 512 so poled that charge ows to the condenser 566 in the same direction without regard to the polarity of the M function. Furthermore, the circuit is provided with diode 514 to discharge condenser -566 at the end of al cycle and triode 515 which controls the charging of the condenser The operation of the sampling circuit will now be described. When, at the beginning of a cycle, a positive pulse appears across resistor 526 current flows through the primary of transformer 516. One secondary 511 of that transformer is connected to the grid of tube 515 being so poled as to render the grid negative, 'whereupon the plate of that tube rises in potential by `such an amount that current ilorws through diodes 51| and 512 to charge condenser 566 to a potential substantially equal to the instantaneous potential of the M function. The condenser 566 is so connected that it retains its charge :with no appreciable loss during one cycle.

One end of the secondary of transformer 562 is connected through diode 516 in series with a protecting resistor 519 and a transformer to the transmitter. Accordingly, if the polarity of the M function relative to a zero or reference potential such as illustrated by line 30| of Fig. 3 is such as to give current in the secondary in the direction indicated by the arrow then a negative pulse will be sent to the transmitter indicating what may be designated as a positive polarity. If the polarity of the M function is negative relative to the reference potential, then no pulse, i. e., a zero pulse, will be transmitted. Thus, information is provided to indicate the polarity of the signal, a condition which is necessary if measurement of sample amplitude is to be on a decibel rather than a linear or proportional basis. 'I'his off or on pulse is the iirst pulse in tlie train of a cycle to be transmitted to the remote receiving point. The duration of this pulse,v as

already mentioned and as indicated in Fig. 3', is.

longer than that of the other pulses in the cycle.

Coder l The coder comprises 4three tubes, T1, T2 and T3, shown as triodes. Battery B1 supplies voltage to the plates of these tubes through a bleeder circuit shown as comprising the resistors r1, rz

. i 6 -g10und. The two triodes T1 'and T2 link the storage condenser 666 to a pair of diodes D2 and D3. The plate of T1 is connected through protecting resistor 680 to the positive end ci battery B1 and its cathode is connected through cathode resistors R1 and R2 in series to ground. 'I'he plate 0f T2 is connected to the -point b of the bleeder and its cathode includes as its cathode vresistor the portion R2 of the cathode resistor oi T1. The cathode'of Tais connected to ground through its cathode resistor R3. The grids of tubes T1 and T2 are connected to the upper side of storage condenser 566, which is at a positive potential when a sample is stored on it. Current is therefore normally flowing through T1. Current will or will not ow through T2 depending on whether the point b is at a positive or a negative potential with respect to its cathode. Current will or will not ow through Ta depending on the potential of its grid.

It will be evident that there will be some point along the bleeder circuit which will be at the same potential as the grounded point of battery B1. The resistors r1, r2 and r3 are so proportioned that when tube T3 is turned -oif that point is at some .place along r3 such as c. Consequently the plate of T2 is positive and current tends to flow through tube T2 under control of its grid potential. When Tais turned on the increased current through r1 is such that the ground point is moved to some such place as d and consequently the plate of T2 is Vnegative with respect' to its cathode and no current flows to the plate substantially independently of the potential difference between its grid and cathode. A small current may ow to thegrid .but the resistance Ro is made largey enough so that the reduction of the charge of the storage condenser is not appreciable during the length of one cycle.

These triodes can now function in two ways. If the grid of T3 is somewhat positive current iiows therethrough, the plate of T2 is negative and the voltage across R2 is V2 aVs where Vs is the voltage on the storage condenser and a is a constant having a value which a function of the circuit and tube parameters. If,

however, the grid of T3 is quite negative the plate of T2 is positive and V2=bVs where b is a different constant which is greater than a. It may b'e shown that if the tubes are similar, if the cathode resistors are large compared to the tube resistance, and if the gain of the tulbes is substantially the same and large compared to unity; that Tl; approaches and thus the ratio of b to a can be changed by altering the ratio R1/R2. 'I'he ratio of b/a may also be changed by changing other circuit and tube parameters as is well understood by persons skilled in the art.

The operation of the triodes in the. coder is controlled by the diodes Di, D2, and D3, which may be rendered conducting by positive pulses initiated at the time stick or delay line and operating through transformers N1, N2 and N3 all in a manner to be now described.

Tube 515 is biased so that, except when 9, neg. ative potential or pulse is applied toits grid, appreciable current flows between its anode and cathode and thus through resistor 51|).v The potential drop across resistance 510 makes the plat of tube 515 negative with respectto ground bu still positive with respect to its cathode. The plate of tube 515 is connected to the center of the secondary winding of transformer 552. Consequently, the negative potential of the anode of tube 515 with respect to ground maintains the diodes 51| and 512 non-conducting.

At the beginning of a cycle of operation a positive pulse initiated at the point I on the time stick and acting through transformer 515 to give a, negative potential to the grid of tube 515 removes the drop across resistance 510 so the potential of the mid-point of the secondary or transformer 552 returns to vsubstantially ground potential and allows the complex signaling wave operating through the diodes 51|, 512 to charge the storage condenser 566 to a plus value proportional of the wave amplitude.

At the same time if the plate of diode 512 is positive with respect to ground a positive pulse is transmitted through diode 518 to the transmitter; if not, no pulse is sent out. Obviously, this positive pulse may be changed to a negativepulse by means of a transformer 590 or vacuum tube if desired.

At the same time a positive pulse acting through transformer N1 on the diode D1 overcomes the bias battery 582 and causes current to flow through diode D1 and charge condenser 584 which in turn renders the grid of T3 positive by an amount V2. Current thereupon flows through Ta and lowers the potential of point b so that the potential of the plate T2 is negative and the voltage across R2 is aVs. Shortly thereafter, a positive pulse initiated from the point 2 of the time stick and operating through transformer N2 is applied to the diode D2 through condenser 584. Let the pulse voltage from transformer Na be Vp. If Vpl-Vg aVs current will flow through diode D2 and on the recession of the pulse the grid of Ta will go strongly negative', T3 will be turned olf, T2 will be turned on and aVs will change to bVs. If the current flows it will iloW through resistor Re and send a negative pulse to the transmitter, meaning that the sample on the condenser is less than a certain level, that is Using Vo as a reference level, let

through this diode and a pulse will be transmiti ted if cVp (a, b) Vs. Here the factor multiplying Vs may be either a or b depending on the result of the iirst operation, as will be pointed out later.

The ratio of b/a is so adjusted by adjustment i ofthe various circuit and tube parameters that 2o 10g b/a =2X and the ratio of transformer and other circuit pameters is chosen that c is given auch a value t t represent the following numbers ot decibels above V1:

l 1 0 decibels 1 0 X decibels 0 1 2X decibels 0 0 3X decibels This will be better understood by reference to Figs. 4A and 4B. In these gures a number of signal voltages are plotted on a decibel scale against time; in particular one sample Vn is taken as of an amplitude Vbetween 0 and X decibels, with respect to the reference level Vo. Examples are also taken of an amplitude sample Vez lying between X and 2X decibels, V23 between 2X and 3X decibels and V54 above 3X decibels.

Case 1, Fig. 4A (V2 between 0 and X 'decibels above reference level Vo). When V21 as shown by line M0 of Fig. 4A is between 0 and X decibels `above the reference level Vo, V21 will be less than Consequently when pulse Vp, which is illustrated by the shaded area 420 of Fig. 4A, from point 2 of the delay line is applied to the circuit of the diode D2 the voltage drop aVn over R2 is small so that current ilows through D2 and a pulse of short .duration is transmitted by the radio transmitter. During the pulse, condenser 584 will be discharged to a voltage diierence equal to the difference between Vp and aVsi. On the termination of this pulse the upper plate of condenser 584 and therefore the grid of Ts, falls to a more negative value so tube Ts is turned olf for the remainder of the cycle. Current immediately flows through T2 increasing the drop across Rz from aVsi to bVsi and consequently raising the potential of the cathodes of tubes 'Ii and T2 at point K by 2X decibels. In other words the potential of point K changes from the value aVsi to a value bVsi. The amplitudes are plotted in Figs. 4A and 4B on a logarithmic or db scale above a reference voltage Vo.'- Hence as shown in the drawings the lines represent the ratio of the various quantities to Vo. Shortly thereafter the pulse cVp, illustrated by the shaded area 422 of Fig. 4A, arriving at diode D3 nds that enlarged voltage bVn is still so low that current flows through Ds transmitting a pulse. The code for this amplitude then, is made up of the digits 1 1 and transmits the information that the sample voltage is to be taken as 0 decibels.

Case 2, Fig. 4A (Vs between X and 2X decibels above reference level Vo). Here aVsz as illustrated by line 4H of Fig. 4A is less than Vp-i-Vg, consequently, current will flow through Dz. yielding a, digit l, rendering the grid of T3 negative, turning Ta offand T2 on, whereupon the potential at point K is raised 2X decibelsto bVsg. It

will be observed that this potential is higher than cVp and consequently the pulse cVp arriving at Da transmits no pulse. The code thus transmitted is l 0 meaning that the signal voltage is less than 2X but more than 1X decibels above reference voltage Va.

Case 3, Fig. 4B (V. between 2X and 3X decibels above reference level Vn) In this case aVss is greater than Vp+Vg, consequently no pulse flows through D2 and no pulse is transmitted. The tube T3 is not turned off and T2 is not Vturned on; therefore, the potential of point K remains avia, where 'V53 is a fairly large value. 'However,-aVsa is less than cVp and consequently on the arrival of that pulse current flows throughsDa and a pulse is transmitted. Thus, the code transmitted is 1 meaning that the signal is more than 2X but less than 3X decibels above reference level.

Case 4, Fig. 4B (Vs above `3X decibels above reference level Vo). The potential drop aVsi is now quite large and no pulse is passed through D2, consequently tube T2 remains olf and the potential of the point K does notalter. Also, since aVsi is in excess of cVp, no pulse is transmitted during the application of cVp to the circuit of diode-Ds and the code message is 00, indicating that V54 is' in excess of 3X decibels. From the above it is seen that the following combinations represent the following number of decibels above 1 l 0 decibels l 0 X decibels 0 1 2X decibels 0 0 3X decibels Acates that the amplitude is more than 3X vabove reference level.

After the transmission of this code of three 6I4. It possesses the same control elements as at the transmitter end and the timestick similarly gives rise to pulses P1, P2, P3, and P4.

y Signal pulses from amplifier and pulse shaping equipment 60| are transmitted over conductor 603 to the grid of tube 610 but the relaxation circuit is so adjusted that it will be triggered of! by the broad cr longer pulses representing the polarity signal only. To accomplish this there is included in the path to the grid of 6 I 0 a pentode 605, the control grid of which receives the signal pulses and a limiting tube 615. The output circuit of tube 605 includes inductance 606 and condenser 601. The pentode, serving essentially as a constant current source, will build upa voltage e over the condenser 601 approximately proportional to theduration of the pulse on the control grid. Consequently, the long polarity pulses will causea greater voltage change across condenser 601 than the short pulses. These pulses areyap'- plied to tube 6|5 which is biased by batteryy 6l6 so that it will not repeat the short pulses but will repeat the longer pulses and apply a voltage to the grid of tube 6I0 suilicient to trigger it'. In this way definite synchronization is maintained between receiver and transmitter.

Incoming polarity signal pulses will also be conducted over lead 603 through the delay circuit 63| of any suitable type to operate on amplifier tube Tis in a manner and for a purpose herein after described.

Briefly, the operation oi the receiver circuit consists in adjusting the gain ci' two amplifiers T14 and Tis by a cooperative operation of incoming signal pulses with locallygenerated pulses.

. The gain of these two tubes having'been adpulsesI the pulse from thepoint 4 on the time stick and operating through triode 550 is applied as a negative pulse to transformer 516. This` renders tube 514 conducting and the condenser iusted in correspondence with the received code, a pulse from the generator is applied to them resulting in an output pulse of amplitude determined in accordance with the received code pulses. 'Ihis output pulse is applied to a gating device controlled by the received polarity pulse and the final output is accordingly a pulse of amplitude and polarity corresponding to the original sam- 566 is then discharged or reset for the next cycle of operations. The timing of the various pulses is indicated in Fig. 3 on the lines Pi, P2 and P3 as they come from the pulse generator and on line T there is shown the timing of the transmitted pulses. The polarity pulse and the reset or restoring pulse are also represented in Figs. 4A and 4B; 420 and 425 representing the polarity pulse and 423 and 428 representing the restoring pulse. d

Receiver indicated by CTs in Fign2; this pulse generator being similar Ito the one at the transmitter end and comprising a relaxation circuit with a gas tube 6l0v and a timestick or delay line. specically the relaxation circuit should have a high degree of frequency stability and includes resistor'6ll capacitance 6I2 and cathode resistor More,

ple at the transmitter.

'I'his will be made more clear by the following description, in which reference is first made to the output portion. 'I'he output is derived from the triodes Tie and T11 an essential function of which is to determine the polarity. These are high n triodes with a vslight negative bias so that no anode current ilows through them except when the grids are driven positive by the output from tube T15. If tube T15 alone drives the grids of tubes 'I'ie and T11 positive, there still will be no output to the low pass filter 640 because the plates of Tm and Tri are connected in push-pull and the grids are connected in parallel.v In addition to the pulse applied tc the grids of tubesTis and Trr, one or two other pulses are substantially simultaneously applied to the'- cathodes of tubes Tie and T17.

Foreach code group or cycle of pulses a pulse P4 from the delay line through tubeP4 is applied to the cathodes of tubes Tis and T11 through transformer 64I in such a direction and of such magnitude that the cathode of tube Tie is made isufficiently positive to prevent the flow of anode current through this tube even when the most positive pulse from the tube T15 is applied to its grid. This pulse is applied to the cathode of tube T11 in the opposite direction, an( its cathode is made sufficiently negative to provide a proper bias nitude which is proportional to or a function of the magnitude of the pulse or potential applied to its grid at thattime. 4

The received pulses appearing across resistor 602 are passed through the delay line 63! and selection tubes 625 and 635. Tubes 625 and 635 function to repeat the long polarity pulses but suppress and do not repeat the shorter pulses. The o eration of tubes 62.5 and 635 in suppressing the/s orter pulses and repeating the longer pulses is similar to the operation of tubes 605 and 615 respectively as described above. The circuits may also include pulse shaping networks and apparatus.

The delayed polarity pulse is applied to the 4cathodesoi tubes T15 and T11 over lead 630. This -which is controlled by the potential applied to the grid of tube T15.

Inasmuch as the anodes or output circuits of tubes T1s and-T12 are connected to supply pulses .of opposite polarity to the low pass filter. the polarity of the impulse which the low pass iilter receives depends on the presence or absence of a polarity pulse in the signal. 'I'he amplitude of the pulse that the low pass lter receives depends on how positive the triode T drives the grids of the triodes T15 and T11.

It should be noted that .there are resistances R15 and R11 in series with the cathode leads of T14 and T11. Thus when the grids are driven positive the cathodes rise in potential and the grids do not become positive with respect to the cathodes and therefore do not draw current. They merely become positive with respect to ground.

The positive pulse supplied by T15 is controlled on a decibel basis by the received signal pulses. I For this purpose the gain of both tubes T14 and T15 is controlled by received -pulses acting through tubes T12 and T12. Let'J us consider that the grids of triodes T12 and T13 have been rendered negative by application of a positivev reset pulse which causes electrons to flow to the grids leaving them with a negative charge when the reset pulse has subsided. This is accomplished by a pulse P1 from the timestick applied across the resistor R11 and through condensers C12 and C13. This reset pulse yoccurs immediately with the triggering of the relaxation oscillator 610 and therefore is substantially coincident with the arrival of the polarity pulse across 602. Initially the gain of triode T14 will be A decibels where A may have any desired arbitrary value and is determined largely by tube constants, the cathode resistor R14 and load resistor R24. Connected in parallel with the resistor R14 is a circuit comprising the plate of tube T12 and resistor R12. Normally when the grid of T12 is negative the tube will be substantially opencircuited and the gain produced by tube T14 will be the normal A decibels. If however, triode T12 is rendered conducting by bringing its grid to substantially zero potential, then the gain of T14 is modified by the circuit connected in shunt to re'- sistor R14 and becomes A+2X decibels. This may be achieved by adjusting the relative magnitudes 12 of resistor-5R14 and R12 and selecting suitable tubes and equipment.

Following the polarity pulsea positive signal -pulse may be received from amplifier and pulse shaping apparatus 60| and together with pulse P2 is applied to diode D12. Because of a bias battery 604 neither the signal pulse nor the pulse P2 alone is suilicient to render diode D12 conducting and to bring the gri-d of T12 to zero potential. Hence operation by spurious pulses or static or by the polarity pulse alone .is prevented. However, when the signal pulse representing a change of 2X-decibe1s occurs in combination with the pulse P2, the bias of battery 604 is overcome, current flows in diode D12 bringing the grid of T12 to zero potential and the gain of T14 is increased by 2X decibels. Similarly triode T13 which was rendered non-conducting by the reset pulse can be renderedv conducting only by a combination of the signal pulse representing X decibels and the pulse Pa' to change of triode T15 from the non-conducting to the conducting condition. Tubes T13 and T15 and their associated circuits are so designed that when tube T1a`is rendered conductive the gain of tube T15 increases by X decibels from its gain when tube T15 is non-conducting.

It is now evident that the gain of triodes T14 and T15 will have been adjusted so that the 'in- 'crease in their gain is in accordance with the two amplitude pulses arriving from the transmitter. Shortly-thereafter, the pulse from P4 on the timestick arrives on the grid of T14 and apv pears as a positive pulse at the plate of T15v with a gain corresponding to the received signaling.

pulses. This positive pulse operating on the grids of T15 and T11 in parallel will give rise to a flow of current corresponding thereto in one of those tubes. Simultaneously with this pulse from T15 the pulse4 P4 will have been transmitted to resistor 643 and the polarity pulse, if present, will have been transmitted to resistor 642. If both these pulses are present the resulting eiect is to cause tube T15 to repeat the pulse from T15 in the manner described above and to apply an output pulse of one polarity to the low pass filter 640, whereas if no polarity signal arrives and pulse P4 alone is present tube T12 repeats the pulse from tube T15 and applies a pulse of the opposite polarity to the output circuit.

From the above it is seen that there is nally delivered to the low pass lter 640 and the receiver 64I, a current pulse the polarity and amplitude of which are in accordance with the original sample. A series of such pulses following each other will yield a reproduction of the original complex wave within the limits of granularity contemplated in the system.

It will be observed also that the action 'in the tubes T15 and T17 calls for the applicationof the polarity pulse thereto simultaneously with the application of the pulse P4 thereto and to the grid of T14. Accordingly, the delay circuit 63| should be adjusted to this end. The pulse P4 from the timestick would ordinarily be adjusted to come very shortly after the termination of the pulse Pa which gives the gain setting for triode T15. Shortly after this last action the next code cycle will start coming in and its rst pulse being a polarity pulse will trigger the relaxation oscillation 6| 0 and the resulting pulse P1'will reset the system in preparation for the next train of pulses.

In order to avoid the production of a positive pulse on the plate of T15 when the grid of tube Tu is changed from negative to zero potential in response to a signaling pulse, the grid of Tu must be biased beyond cut-oir and the tube must be made conducting only when pulse P4 is applied to its grid. Changing the grid o f T13 from nega` tive to zero potential will cause a negative pulse on the plate of T15 and hence will cause no output. The reset pulse will cause a positive pulse at the plate of T15 but since no pulse is applied to transformer 64I at the time of the reset pulse the tubes Tis and T11 are balanced and hence the reset pulse causes no output.

It is apparent that many variations may ber made in the circuits of my system without departing from the spirit of the invention. For example, instead of introducing the delay for the polarity pulse in the receiver, it could have been introduced at the transmitter end in the lead from the diode 518 so that for a cycle of operations the polarity pulse would have come at-.the end instead of at the beginning. In any case, the delay line would have a delay slightly less than or equal to the reciprocal of the sampling rate thus delaying the polarity pulse just one cycle of operations.

Also the delay device may be of any suitable type such as any delay network of one or more sections, a. transmission line, an artificial transmission line. a transmission path through any suitable medium, etc.

What is claimed is:

1. In `a communication system, means for measuring the voltage on a condenser comprising a first ampller tube and a second amplifier tube with grids connected in parallel with each other and to said condenser, the two tubes having a common cathode resistor, a source of pulses of predetermined value and time spacing, means by which a rst pulse will turn oi the second tube, a second pulse which will turn on the second tube if the drop over the common cathode resistor is less than a given amount and will leave it oil if more than the given amount, a third pulse of larger amplitude than the second which will determine whether the drop over the cathode resistor is in excess of or below the voltage of said third pulse.

2. In a signaling system for transmitting information relating to a complex wave, a transmitting station comprising a circuit for periodically sampling the wave to charge a storage condenser to the instantaneous voltage Vs, which is a function of the instantaneous amplitude of said wave, an ampliiier tube subject to said condenser, said tube including a cathode resistor giving a sample voltage drop aV over one portion of the cathode resistor, a gain means for changing the sample drop over the said cathode resistor to bVs. where b is in excess of a, a source of pulses of predetermined value, a rst pulse rendering the said means inactive, a second pulse testing whether aV; is greater or less than the said second pulse amplitude plus a constant and if larger then leaving the said means inactive but if smaller then rendering the said means active to raise the sample voltage over the cathode resistor to bvs, a third pulse to `test whether the aV or the bVs, whichever is present, is greater or less than the amplitude of the third pulse.

3. In a signaling system for transmitting information relating to a complex wave, a transmitting station comprising a circuit for periodically sampling the wave to charge a storage condenser to the instantaneous wave voltage Vs, an

amplifier tube subject to said condenser, said tube circuit including a cathode resistor giving a sample voltage drop aVs over a portion of the cathode resistorwhere a is less than unity, a gain '5 device for changing the sample drop over the said cathode resistor to bVs where b/a is a gain of 2X decibels, a pulse generator giving rise to cycles of pulses, ,a iirst pulse cooperating to charge the condenser and to render the said gain device inactive, a.' second pulse testing whether aVs is greater or less than the second pulse amplitude plus a constant and if larger then leaving the said device inactive but if smaller then rendering said device active to raise the sample voltage over the cathode resistor to bVs, a third larger pulse to test whether the aVs or the bVs, whichever is present, is greater or less than the' amplitude of a third pulse.

4. A combination of claim 3 with the added feature of means whereby a signal pulse is transmittcd by the second pulse if the vgain device is not rendered active and by the third pulse if it is in excess of the voltage drop over the cathode resistor, the'two pulses thus characterizing the amplitude of the condenser voltage.

5. A combination of claim 3 characterized by the fact that the gain device is a second amplier tube with a cathode resistor which is the portion of the cathode resistor referred to in connection with the rst named amplifier tube, the grid of this second tube being also subject to the condenser.

6. A combination of claim 3 characterized by the fact that the gain device is a second amplifier tube with a cathode resistor which is the portion of the cathode resistor referred to in connection with the ilrst named amplifier tube, the grid of this second tube being also subject to the condenser, and characterized further by a third amplifier tube on which the iirst pulse operates such thatwhen the third tube is on the second tube is rendered non-conducting and when the third tube is turned olf, the second tube is rendered conducting to raise the sample drop across the common cathode resistor.

7. In a signaling system for transmitting information as to the shape of a complex wave, a receiver station adapted to receive cycles of three pulse codes of two diiierent signaling conditions, one pulse designating the polarity of the voltage and the other two pulses characterizing the amplitude on a, decibel basis, a train of amplifiers in tandem, the two amplitude pulses cooperating to adjust the gain of the train of amplifiers to conform with the amplitude which the said two pulses characterize, the polarity pulse cooperating v to control the direction of the amplier output in conformity with the polarity designated.

8. A combination according to claim 'I including a pulse generator for producing control pulses synchronous with the received said two pulses characterizing the amplitude, and means for preventing said adjustment of the gain of the train of amplifiers except upon the coincidence of a received amplitude pulse and a corresponding pulse from said pulse generator.

9. A combination according to claim 8 including means responsive to the rst amplitude pulse -for adjusting the gain of said train of ampliers 70 by 2X decibels, and means responsive to the second amplitude pulse for adjusting the gain of said train of amplifiers by X decibels, thus providing four amplitude steps of X decibels each' from zero to 3X decibels with respect to a reference voltage.

10. In a signaling system for transmitting in- "f formation on the shape of a complex wave, means at the transmitter for sampling the wave amplitude at frequent intervals and for transmitting cycles of three pulse codes of two different signaling conditions timed in accordance with a pulse generator, one of the three pulses designating the polarity of the sample and two characterizing the amplitude on a decibel basis, a receiver station adapted to receive said three pulse codes and comprising a pulse generator, a train of three stages of amplifiers the last being a normally balanced push-pull pair with no output, the polarity pulse serving to synchronize the pulse generator and operating over a delay circuit to the push-pull circuit, the received first amplitude pulse cooperating to adjust the gain of the first stage of the amplier depending on whether the said received pulse is an on or an oil" pulse, the received second amplitude pulse similarly -operating on the second stage whereupon the delayed polarity pulse operates to unbalance the pushpull circuit in one direction or the other depending on whether the polarity pulse is an on or oi y vlding four amplitude steps of X decibels each 16 in response to the combination' of said second amplitude pulse and the corresnding pulse from said pulse generator, means to unbalance said push-pull circuit in response to said delayed polarity pulse and the corresponding pulse from said pulse generator, and means for transmitting through said train of three stages of ampliers the last-mentioned pulse from said pulse generator.

12. A combnation'according to claim 11 including means responsive to the rst amplitude pulse for adjusting the gain of said first stage of the amplifier by 2X decibels, and means responsive to the second amplitude'pulse for adjusting said second stage by X decibels, thus profrom zero to 3X decibels with respect to a reference voltage.

413. A combination of claim 10 characterized by the fact that the rst stage in the train of amplifiers is normally below cut-ofi'and is activated by a pulse from the generator after the gain adjustments have been made,.the said last pulse cooperating with the delay polarity pulse to at the same. time unbalance thepush-pull stage.

JOHN R. PIERCE.

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

UNITED STATES PATENTS Name Date KOch May 7, 1940 Number

Images