US2336627A - System for overcoming thermal agitation noise and inductive disturbance noises in carrier systems - Google Patents

Description

Dem 1943. H. MQURADIAN 2,336,627

SYSTEM FOR OVERCOMING THERMAL AGITA'IION NOISE AND INDUCTIVE DISTURBANCE NOISES IN CARRIER SYSTEMS Filed March 31, 1942 4 Sheets-Sheet 1 TELEPHONE LINE 0R STATION LINE OR STATION T0 TELEPHONE LINE 052 STATION T0 TELEPHONE ,um: on STATION MINIMUM WORKING NET LOSS IN db N -b m m 4 ob (a 8 MILES wmaAm w VINVENTORI T0 TELEPHONE Dec. 14, 1943. MOURADlAN 2,336,627 SYSTEM FOR OVERCOMING THERMAL AGITATION NOISE AND INDUCTIVE DISTURBANCE NOISES IN CARRIER SYSTEMS Filed March 31, 1942 4 Sheets-Sheet 3 No.1 U TRANs. CHANNEL CHANNEL CARRIER SUPPLY DEM.

TELEPHONE m LINE 0R STATION P DEM I DEM a BB CHAN.

AMP FlLT.

CHAN NEL CARRIER SUPPLY MOD. CHAN Fl LT.

No.2, TRANS. CHANNEL I N VENT OR.

I Dec. 14, 1943.

H. MOURADIAN Filed March 31, 1942 7 SYSTEM FOR OVERCOMING THERMAL AGITAT ION NOISE AND INDUCTIVE DISTURBANCE NOISES IN CARRIER SYSTEMS I I l I j come GROUP GRXUP CA gLE FILT. FILT. AMP PAIRS k GROUP NM CARRIER REC. SUPPLY C CHANNEL (H 1 {I TO m TERMINALS (ab)FIG.4

u o g CHANNEL i%%fER n SUPPLY s T I \r COMB AUX. GROUP GROUP GROUP DIR, To FILT c DEM 0 CABLE AME FILT. AMR PAIRS FIG.8

Wau h INVENTOR.

Patented Dec. 14, 1943 anger? SYSTEM FOR @VERCQMKNG TEEREKAL AGE- TATION NOESE AND HNDUCTKVE DllSTUlEtB- ANCE NOISES IN CA3RIER SYSTEIWS Hughes Mouradian, Philadelphia, Pa.

Application March 31, 1942, Serial No. 4373. 12

(Cl. fill-78) 21 Claims.

3.. Noise from thermal agitation.

2. Noise from outside, sources of interference with commun cations. These include, among others, the following:

(a) Noise from heavy static or open wire taps to cable circuits.

(b) Noise from telephone and telegraph re- 9 peater oflices.

(c) Noise in amplifying systems directly connected to, and forming part of, the Wire communication systems.

3. Cross talk between circuits in the same cable or the same pole l ne. l. Babble between circuits in the same cable or the same pole line. i

In the case of the application of the unclerlyng principles of the present invention to radio transmission as fully outlined in U. S. A. 2.282,- 2919, granted to me, as previously cited, no problems of mutual interference between two channels of commun cation using the same wave length or frequency was involved. Items 3 and 4, mentioned hereinabove, relate therefore peculiarly to wire systems. As well known. same problem, in its radio aspects, is solved by the physical sepwhich is provided between two radio systems using the same frequency. in the case of wire transmission. systems, and particularly of carrier system in cable, it is not at all feasible to provide adequate separation between any two pairs using the same range of frequencies. It has, therefore. been necessary to superimpose upon the particular system disclosed in U. S. A. 2,282,299 granted to me an additional invention the dificul s mentioned under items .1 last me .oned dif'liculties can als .ircome, but not to the extent disclosed herethese specifications, by the use of methods v dy known to the art. These will be reviewed more in detail hereinbelow.

This invention will be more clearly understood by reference to the following description, taken in connection with the accompanying drawings in which Fig. 1 illustrates a new l-wire non-carrier transmission system for overcoming crosstalk and inductive interference from outside sources, Fi 2 illustrates the l-wire system of the present art. Fig. 3 illustrates the best performance characteristics of the two-wire facilities, of the type therein indicated, used in the present art. Fig. 4 indicates the new 4-wire system, as applied to either non-carrier or carrier systems, equipped with twoway repeaters; Fig. 5 indicates a new il-wire system, for overcoming crosstalk and interference from outside sources, a new equivalent in its line features of the present day wire system of Fig. 2 the drawings. Th terminating hybrid coils have not been shown. Fig. 6 represents the performance characteristics of the present day 12- channel cable carrier system; designated as the type K system; Fig. '7 and Fig. 8 indicate a new arrangement of the terminating equipment for carrier systems. Fig. '7 indicates the individual channel section, while Fig. 8 indicates the group channel arrangement.

The d'fficulties experienced in practice with the items enumerated at 1. 2, 3 and 4 increase rapidly with the amount of amplification required to satisfy transmission requirements from-end to end of the communication circuit. For this reason, the most valuable field of use or" the present invention will obtain in the field of carrier transmission over wires. This particular field of application is therefore outlined in greater detail, herein oelow. However, in the non-carrier field of wire transmission, crosstalk and babble already been to make their influence felt at the maxi.- mum frequencies used of 3,200 cycles per second for commercial telephone circuits. The features mentioned become still more prominent in the case of wire facilities used. common broadcasting purposes, between the various studios of a national broadcasting network, the maximum frequency transmitted reaching a value of 15,000 cycles per second in this case; In any event, some of the features of the present invention can best be illustrated and more easily understood by their application to ordinary wire circuits, since these last require simpler equipment arrangements. Fig. 1 of the drawings represents the simplest form of application of the present invention. It will be noted that the system, ther indicated, makes use of a-wires between the terminals a, b and c, d. It will also be'noted that the upper set of two wires has two transpositions'- i. e. transpositions 1, 2, 3, 4 at one end of the line and transpos t- ions 5, 6, 7, 8 at the other end. It is necessary, in order for the system illustrated on Fig. 1 of the drawings to function exactly and fully as intended; that the upper and lower pairs be as nearly identical as possible to each other in electrical characteristics and have, in addition, as nearly as possible ident cal coupling characteristics (both magnetic and static) with all other pairs in the same cable. There are a number of ways in which this requirement can be met, as outlined in some detail in these specifications hereinbelow. A new method covering the construction aspects of a cable system which meets the balance requirements as above outlined is described in my application S. N, 477,159 for Cable systems for high frequency transmission, filed February 25, 1943. Reference may be made to the disclosures in said application and particularly to Table XIX, last column which clearly shows the equality of unbalance obtained between pair 1, 3 of one cable unit and pairs 5, '7 and 6, 8 of an adjoining unit by means of construction procedures followed during the manufacture of the cable. Assuming the above said requirement to have been met, let us assume a physical pair in the same cable carrying speech currents. Such a physical pair will crosstalk both into the upper pair and the lower pair of Fig. 1 of the drawings. The amplitude and phase angles of the induced currents into the l-wire system of said Fig. 1 will be the same. The near-end crosstalk on the upper pair will be exactly equal, by construction, both in amplitude and phase, to the near-end crosstalk on the lower pair. In view, however, of the transposition 1, 2, 3, a on the upper pair, the two sets of induced currents will be in direct phase opposition and will cancel each other. The same situation holds true with reference to farend crosstalk. In this case, the induced currents in the upper and lower pairs of wires of the system illustrated on Fig. 1 of the drawings will arrive at terminals 0, d in direct phase opposition on account of transposition 5, 6, 7, 8 and will again cancel each other. Crosstalk, whether electromagnetic or electrostatic in nature, is thus eliminated both at terminals a, b and c, d. Babble, which is the resultant crosstalk from a multiplicity of inducing sources in the same cable, will also cancel out. If, now, instead of an ordinary pair in the same cable carrying speech currents, we were dealing with a second ii-wire system such as pictured on Fig. 1 of the drawings carrying a telephone conversation in the same cable, the final effect or crosstalk produced would still be zero. since each of the component pairs of this second a-wire system, taken individually, produces no effect upon the first l-wire system. The two pairs of conductors forming th proposed Ir-VH1? system are shown bridged together only at the two ends of the line on Fig. 1 of the drawings and only at the two ends of a repeater section on Fig, 4 of the drawings. There would be an advantage in bridging together the two pairs at more frequent and convenient intervals. Snce nductive eifects. which it is proposed to mitigate have longitudinal components far greater in magnitude than the residual metallic components, the bridging of the two pairs and the provision of the two transpositions at the bridging points automatically eliminates the longitudinal effects leaving only the residual metallic components.

The 4-wire system described above, and illustrated on Fig. 1 of the drawings, should not be confused with the l-wire system of the present art, illustrated on Fig. 2 of the drawings, which uses one pair of wires for transmission in one direction and a second pair of wires for transmission in the opposite direction. The restriction in the directions of transmission is obtained, as well known, through the use of one-way type telephone repeaters. In the proposed system of Fig. l of the drawings both pairs of wires are used simultaneously for transmission in the same direction. Two-way telephone repeaters would be inserted, whenever equired, in both the upper and lower pairs of wires of Fig. l of the drawings.

It will be noted that on Fig. l of the drawings both transpositions 1, 2, 3, 1 and 5, 6, '7, 8 are shown on the upper pair of wires. These could obviously be shifted to the lower pair. Another alternative arrangement would provide on transposition on the upper pair of wires at the transmitting end and a second transposition on the lower pair of wires at the receiving end, without departing from the spirit of the invention. The essential and controlling requirement is, that there should be one transposition at the transmitting end and a second one at the receiving end, it being immaterial on which pair of wires these transpositions are located. The only arrangement to avoid is to place both at the same end of the line, independently of their terminal location, on the upper and lower pair of wires.

Fig. 3 of the drawings shows the best performance of two-wire ci cuits that it has been possible to secure in the art. (See paper by A. 13. Clark and H. S. Osborne on Long d stance telephone circuits, Bell System Technical Journal, page 530, October 1932.) The ordinates indicate the lowest equivalent that it is possible to provide for the two-wire ci cuits. expressed in decibels, from end to end. The abscissae indicate the length in miles. The facilities used are 19 B. & S, gauge, B-88-50 and 1-1-88-50. It will be noted that crosstalk limits the equivalent of a circuit 100 miles in length to 7 decibels. This paticular limitation would be removed when use is made of the arrangement shown on Fig. 1 of the drawings, and it would be feasible to provide between two points 100 miles apart a 3 decibel circuit, with 13-88-50 facilities. Four different objectives were indicated hereinabove as representing the aims of the present invention. We have just shown that the 3rd and the 4th can be readily accomplished, i. e. crosstalk and babble can be eliminated for all practical purposes, with the use of the arrangement of Fig. l of the drawings. As well known, the level of transmission, with wire circuits not using carrier, is never allowed to dip more than 25 db. below the so-called 0 level of transmission of reference. For this reason the thermal agitation effect in the line conductors is not of importance for circuits of the type mentioned. On the other hand, noise from outside sources of interference, listed as the #2 objective, is as definitely eliminated as crosstalk and babble. Several sources of noise have been carefully measured for a wide range of frequencies and plotted in terms of the so-called reference noise, a noise which the human ear can barely hear. It has been taken equal to 1.0 watts, of 1-000 cycle power or roughly 90 db. below the level of 1 milliwatt of energy. These measurements and computations are shown on Fig, 14 of a paper by M. A. Weaver, S. Tucker and P. S. Darnell, bearing the title Crosstalk and noise features of cable carrier telephone system. (A. I. E. E. Transaction 1933, page 258.) For a 17 mile section of cable, the noise from thermal agitation in the line conductors, for the range of frequencies below 5600 cycles per second as herein considered, is inaudible, and hence of no importance as above stated for the case just considered. On Fig. 14 of the paper referred to above, this type of noise is shown on curve A. It is interesting to note that this noise is just at the threshold of hearing for 30,080 cycles. As clearly brought out by the above mentioned authors, the noise due to the repeaters which in reality should be classified as internal noise is also negligible. The noise effects referred to by the same authors as C, D, and E and illustrated on Fig. 14 of their paper, are:

C. Noise from voice frequ ncy telephone repeater cfiice.

D. Noi e from telephone and telegraph repeater office.

Noise from heavy static on open wire tap close to repeater input.

All of the above sources of noise are quite large, the source of noise E which is about 40 db. above reference point being the heaviest. This is the noise that is picked up by an open wire line connected to a pair in the same cable that is carrying, by assumption, the voice frequency i-wire circuit of 1 of the drawings. This type of noise completely elim nated by the arrangement of Fig. l of the drawings, since any current induced the upper pair of Fig, 1 of the drawings, by the cable pair carrying the heavy static, is completely balanced by corresponding induced current on the lower pair of Fig. l of the drawings. The same result, i. e. freedom from interference holds for the other sources of noise, designated as D and E. They are eliminated.

It is necessary to observe that the use of a second pair, in the new 4-wire systems illustrated on Fig. 1 and Fig. of the drawings, does not in way add to the efficiency of voice transmision between th terminals 0., b and c, d of the i-wire system, when said system is used in noncarrier operation. The provision of this second pair simply means additional expenditure to obtain some of the specific purposes mentioned hereinabove as the objectives of the present invention. It is well to point out, however, one interesting result. The characteristic impedance of the new l-wire system is one-hali of that of its two component parts, since these are in parallel. The potential impressed upon each of these two onent parts is of the potential which would have been impressed if these component parts had not been bridged a llel. The crosstalk, due to capacity unaction, by each of these component parts upon any other third circuit in the same cable therefore reduced in the ratio of or 3 decibels. The corresponding mutual reactance effect is also reduced in the same ratio. These are merely additional advantages to the more important effects of the -wire system in that the two component parts of said l-wire sysn act in opposition to each other upon any third .icuit. Conversely, any residual unbalance efect due to mutual reactance unbalance action be tween any third circuit and the new a-wire system is reduced in the ratio of on account of the design of inequality ratio coils T and T since these now match a characteristic impedance which is one-half of the characteristic impedance of each of the two component parts of the l-wire system, when considered separately. This is of importance in the carrier range, since mutual reactance in balance is controlling in that range.

If the arrangement of Fig. l of the drawings as indicated, is used, not only the investment in cable plant is doubled per communication circuit, but so are the telephone repeaters associated with the line. The additional cost of telephone repeaters can be eliminated if the system shown on Fig. 4 of the drawings is used. In this arrangement, the two-way telephone repeaters AA are used in common, for the two pairs of wires of the 4-wire system. Neutralization of the effects of outside interference crosstalk and babble, is, in this case, carried out for each repeater sec tion. Fig. 5 indicates the application of the principles of the present invention to the present day e-wire system of Fig. 2 of the drawings. It will be noted that the arrangement of Fig. 5 of the drawings actually uses 8 wires for ransmission purposes. As weil known, the repeaters in the present day 4-wire systems are of the one way type. The relative directions in which energy is amplified by these repeaters is shown by indicating arrows of the drawings. Fig. 5 shows only the line portions of the S-Wire system, the terminal equipment oi hyb id coils being the same as that shown on 2 of the drawings. Terminals i, 2, 5, 6, and 3, l, '5, 3 of Fig. 5 of the drawings correspond therefore to the same numbored terminals of Fig. 2 of the drawings.

As shown on Fig. i oi the previously cited paper by A. 13. Clark and S. Osborne, page 531, the present limitation as to minimum working net loss of the d-wire circuits, up to distances of 450 miles, is crosstalk. T is limitation makes it impossible to operate, as shown, present day type ii-wire circuits such as pictured on Fig. 2 of the attached drawings, at less than an equivalent of r 6 db. This limitation is removed by the S-Wire arrangement shown on Fig. 4 of the present drawings. Interference from outside sources is also eliminated. The other items or" interference, classified as l, 2, 3, i, hereinabove in these specifications, i. e. particularly noise from thermal agitation which is item 1 and item 20 are not of practical significance for the normal ranges of performance of 4-wire circuits of the non-carrier type.

The application of the underlying principles of the present invention is actually of greatest importance in conjunction with carrier circuits. As well known, such carrier circuits when in cable, make use of non-loaded pairs in order to secure high velocity of transmission. The advantages obtained by the use of the invention first outlined in these specifications, are greatest with carrier circuits, because of the fact that the difficulties experienced in practice with noise, crosstalk and babble increase rapidly with. frequency and with the amount of amplification required in with the communication circuits inno in non-carrier hereinbefore considered, the highest frequency used for commercial circuits is 3,230 cycles second, the highest frequency in the cable carrier range is 60,000 cycles per second, and in the open wire carrier range 140,000 cycles per second. In the low frequency non-carrier range, one type of facility quite generally used is 19-3-83. The

equivalent per mile of this type of facility is 0.28 db. On the contrary, where non-loaded #19 B. 81 S. gauge circuits are used, as they are so used in the cable carrier range today, the equivalent per mile varies from 2.5 db. to 3.8 db. The range of frequencies considered, corresponding to these equivalents lies between 12 and 60 kilocycles per second. In the low frequency range, -3200 cycles per second, the telephone repeaters are spaced 50 miles apart with a loss between repeaters for 2-wire l9-B-88 facilities equal to 14 db. Since telephone repeaters are adjusted to furnish speech energy at a level of +3 db., the lowest level of speech currents is +314=11 db. For levels of this type no difliculty is experienced with either thermal agitation in the line conductors or with the grid input circuits. In the cable carrier frequency range, on the other hand, the line losses are extremely large. If telephone repeaters were spaced 50 miles apart, as they are in the noncarrier range today, the line losses would vary from 125 db. to 190 db. per repeater section. The cable carrier system cannot simply be operated for line losses of this order, as the noise due to thermal agitation in the line conductors and the grid circuit of the first tube in the amplifying system would completely drown out or mask all speech currents. It is for this reason that telephone repeaters have to be located today, on

lines operated in the cable carrier frequency range or 12,000 to 60,000 cycles per second, every 17 to 18 miles. In order that this point may be made clear, it is pertinent to refer to the expression for the noise due to thermal agitation in the line conductors, as measured at the output of an amplifier with amplification factor G(f)- E r =mean square thermal noise voltage.

lc=Boltzmanns constant:1.5'7 X T=the temperature of the line conductors expressed in Kelvin degrees.

RU) =the resistive component in ohms of the line conductors as measured from the input terminals of the amplifying device.

G0) :the voltage amplification factor of the amplifier.

(Fa-4 1) :frequency band within which the amplification is considered.

Both the resistive component R0) and the amplification factor GU) are naturally functions f the frequency (f). If the above formula is plied to a line 17 miles in length in 19 B. & S. 'auge cable, operating within the range of 56,000 to 60,000 cycles, we would have for R0) the value of 132 ohms approximately. For the channel just mentioned, the value of Signal to noise ratio Repeater spacing Decibels Thus, with 51 miles between repeaters, noise would completely dominate over speech.

To the thermal agitation due to line conductors must be added the thermal agitation due to the grid circuit resistance of the first tube of the amplifier. According to data, as reported by Weaver, Tucker and Darnell, bearing the title Crosstalk and noise features of cable carrier telephone system, A. I. E. E. Transaction, May 3838, Fig. 14, page 258, the tube noise with the type of tubes used is about equal to the noise due to the line conductors, as the sum of the two noises is 3 db. higher than the noise due to t-..e line conductors alone. Reference must be made to the excellent paper of the above authors for a proper understanding of the various problems due to noise in the application of carrier in cable for voice communication purposes. Several other sources of noise, using the terminology of the above authors are mentioned, including C. Noise from voice telephone repeater ofiice.

1). Noise from telephone and telegraph repeater ofiice.

E. Noise from heavy static on open wire tap close to carrier repeater input.

It will be noted that while, by means of the special measures indicated at some length in the A. I. E. E. paper above referred to (pages 251-259 inclusive), the sources of noise C, D and E can be appreciably reduced, the sources of noise A and B, due to thermal agitation represent an irreducible minimum and are not affected by the arrangements therein described or as otherwise known to the art.

The present invention when applied to the socalled type K carrier system of the present art (see A. I. E. E. Transaction 1938, page 227, paper entitled A carrier telephone system for toll cables), will result in the virtual elimination of the irreducible minimum due to thermal agitation and of all the other sources of noise desig nated as C, D and E, hereinabove. Fig, 7 and Fig. 8 of the present drawings illustrate graphically how this may be accomplished. I propose, in order to achieve the purposes above men tioned, to use for transmission in any one direc tion, two single band carrier channels adjacent in frequency, instead of a single channel as at present in the specific manners described in detail hereunder. Fig. '7 and 8 illustrate the terminal equipment of a single two-way communication circuit, using for this purpose 4 carrier channels, two channels for each direction of transmission. This equipment is included between terminals a, I) connected to a telephone line of the non-carrier type on Fig. 7 and terminals c, d connected to a telephone toll cable or open wire line on Fig. 8. The terminal equipment at the distant end of the toll cable is exactly symmetrical with that shown on Fig. 7 and Fig. 8 of the drawings, and is not illustrated for that reason. On outgoing transmission, speech currents arriving over terminals a, b, divide themselves into two exactly equal parts. In the upper path or channel #1 as shown on the drawings are illustrated diagrammatically the 4- wire terminating hybrid coil, the channel modulating equipment, the carrier supply, the modulating channel filter, the transmitting'carrier bus on Fig. '7, and on Fig. 8, the compensating filter, the group modulating equipment, the group carrier supply, t .e group modulating filter, and the amplifier A, which connects this upper path with the toll cable at terminals c, d. In the lower path, or transmitting channel #2, as also shown on the drawings, are illustrated diagrammatically 1e l-wire terminating hybrid coil, the modulat .g equipment, the amplifier AA, the

.lter eq ment &c., all exactly similar in type to at snown in the upper path or channel #1.

The type of equipment thus bri fly outlined and as shown in either path is now in very extensive use and has been described in great detail in the A. I. Transactions, 1933, by R. W. Chestnut, L. M. Ilgenfritz and A. Kenner, on page 237. The description of the terminal equipment will not be therefore repeated here. The main points to note are, as essential new contributions l. The use of present, per communication circuit, in any one direction of transmission.

On Fig. 7 of the drawings, the two channels used for outward transmission are indicated as No. l and No. 2 and are both connected to outgoing bus bar U.

2. The use of a transposition in either one of the two channels.

On Fig. '7 of the drawings transpositionii, 22, 2'4 is shown in the No. 1 channel. An alternative arrangement would co-nsist in placing this in the No. 2 channel, without changing in any way the results obtainable.

3. The use of adjacent frequency for channels No. 1 and No. 2. Such use, while not absolutely essential, results in optimum effects and is the most advantageous.

7c=coefficient of proportionality.

C1=amplitude of channel carrier supply. It should be noted that this is the same for both channels.

C2=amplitude of group carrier supply.

gei=final frequency factor of carrier of channel two channels, instead of one, as at pz=final frequency factor of carrier of channel q=frequency factor of the speech currents.

01=phase constant.

S=amplitudeof speech waves.

Suppose a Wave train of the type given by Equation 2 is impressed upon the toll line or cable at the distant terminal and reaches terminals c, d of Fig. 8. We can make this supposition since, as stated, the terminal equipments at the two ends of the carrier transmission line are exactly similar. Such a wave train, would be amplified by group amplifier AA of Fig. 8 of the drawings, demodulated by a group demodulator, re-arnpliried by auxiliary group amplifier C reaching common receiving bus bar V. At this bus bar terminate the various individual receiving carrier channels. Since the wave train of (2) has components in two channels, -i. e. ii to f1-4000 cycles per second and ii to f1+4=000 cycles, this wave train divides itself into the receiving channels No. 1 and No. 2, as shown on the drawings, both of these receiving channels carrying the same communication message. After demodulation and amplification, the wave equations in the two channels will be, respectively- The coefficient Z and phase constant 02 have been assumed different from Z and 03 in View of the variable pad P and phase changing network Q in receiving channel No. 2. The speech wave reaching terminals a, b is given by The is now used in conjunction with the wave train in receiving channel No. l in View of the transposition 26, 22, -23, 2d, as shown on the drawings. We have, therefore, an addition of the speech waves in the. two receiving channels.

The main purpose of the arrangement of Fig. 7 and Fig. 8 of the drawings will now be described. First of all, the noise due to thermal agitation in the line conductors, as given by Formula 1 is equal to:

F2 =aTf Reman The noise power E /R(j) is equal to, if'Gdf) is replaced by its average value Ga(f) within the range of frequency of a given carrier channel by The average value Ga(f) of the amplification for the channel included between 12,000 and 16,-

000 cycles per second is 2.55 decibels per mile multiplied by the distance in miles between repeater stations. The average value of G8.(f) for the channel included between 16,000 and 20,000 cycles per second is 2.67 decibels per mile, multiplied by the distance in miles between repeater stations. If this distance is taken at 17 miles, then the dilference in noise power corresponding to equal frequency ranges in the two chanceiving channel No. l.

nels will be 2.04 db. For the other channels, the difference in amplification factor, per mile of distance between repeater stations will be approximately .10 of a decibel. Considering two components of the same frequency, due to thermal agitation, in receiving channels 1 and 2, and indicating the amplitude and phase of these components as (NM m) for channel No. 1 and (Naps) for channel No. 2 the expression for the noise effect at terminals a, b of these two equal frequency components will be, on received messages,

(6) 1\ a cos (pnH- oa)+Nb cos (Put-f-(pb) where pn 27Tfn. fn=the particular noise frequency component considered.

The negative sign of Na, in opposition to that of Nb, is due to the transposition 2 l, 22, 23, 24 in re- We have shown that practically a constant difference of 2 db. will exist between Na and Nb. How about the phase differences? While it appears amply evident that, at the point where the terminal agitation takes place the various frequency components of that agitation must all have the same phase, since these all start and end with the physical motion of electrons within the line conductors which gives rise to these various frequency components, it also appears quite evident that they will not have exactly the same phase at the input terminals 0, d of Fig. 8, since the speed of prepagation of waves along the line conductors varies with the absolute frequency Fn of such waves. To determine the phase differences, we apply the thermal agitation formula to an element of length (d!) of the line conductors and integrate e'r throughout the entire length of the lineafter simplification for the frequency range included in that channel and write zerG) F2 -F1) where 2 Tan a ihe phase change for ET is evidently The following table shows the computed values for the phase angle differences as given by Formula 8 TABLE I.PHASE ANGLE DIFFERENCES THERMAL AGITATION COMPONENTS Successive The differences in angle are smaller for the frequency channels higher than 30,000 cycles and have not been shown.

It is important to note that the phase angle differences shown in th above table are independent of the length of the line conductors, provided the line exceeds a certain minimum length. There will be no appreciable chang in the differences, as between two adjacent carrier channels, either in amplitude or phase differences whether the distance between successive repeaters is 17 miles or 50 miles as regards the noise power at the terminals of the input circuit, since these dilferences depend only upon the ratio which is independent of the length of the line. The same conclusions hold as to the amplitudes of the noise power due to thermal agitation again considered at the terminals of the input circuit before any amplification is applied. This will also be clear by reference to Formula 11 (see Limits to amplification by J. B. Johnson and F. B. Llewellyn, A. I. E. E. Transactions 19%, page 1449 and Formula 3, page 1450) wherein it is clear that before amplification the noise power depends only upon the absolute temperature T and frequency band width. The phase angle differences, as shown in the above table, can be readily taken care of by the phase change Q. The amplitude variations in th noise power are thus due only to the difference in the amplifications provided in the various channels of the carrier type amplifier, which differences are provided deliberately by design in order that the final demodulated speech product for each channel has approximately the same transmission equivalent. Where, as in the present instance, two carrier channels adjacent in frequency are used for the same communication or message, then the pad P can be readily adjusted to offset the higher amplifier gain required in the higher frequency channel of two adjacent channels, so that in Formula 6 the noise amplitude Na will be equal Nb. Furthermore, as already stated, phase change Q can be readily set to eliminate the small phase differences so that 0t= pa in the same Formula 6. Thus, the use of artificial pad P and phase changer Q in an arrangement wherein two adjacent channels are used for the same outward or inward message, will completely eliminate what has heretofore been considered an irreducible minimum" in the noise background of amplifiers and had been a bar to the securing of higher amplifications.

The conclusions reached hereinabove for the noise due to thermal agitation in th line conductors apply with equal force to the amplifier tube noises in receiving amplifiers D and 0, since both of these are also common to both channels. The corresponding type line and tube noises in the individual carrier channels can be greatly reduced but not completely eliminated, since these are no longer common to both channels. Reference is made here to amplifiers B and BB. The purpose of these is to offset the loss in the demodulating equipment with copper oxide discs which operate best at low levels of power. The amplifiers B and BB actually operate at low amplifications, of about 1'? db. from a level of 5 db. and the relative noise power due to these is cor espondingly quite small, and is further reduced by the effect of transposition 2i, 2, 23, This noise power could practically be eliminated by arranging so that all the amplification required is furnished by amplifiers C and D and regulating the copper disc type demodulators by demodulators of the vacuum tube in the receiving channels and operating these at zero amplification. The spacing between repeaters can therefore b increased to 50 miles, if needed, since neither the total amount of noise power nor the phase differences between two channels adjacent in frequency, increase when the spacing increases from 1'? to 50 miles or over. It is possible, of course, that there will be slight deviations in the actual cutoff points of the filters in each individual channel from the theoretically intended points and thereby corresponding variations introduced in amplitude and phase variations in th noise power. It is probable such variations will not exceed 1%.

The sources of noise indicated as C, D and E, in the A. I. E. E. paper by M. A. Weaver, R. S. Tucker and P. S. Darnell, already referred to, can also be reduced to a greater extent or virtually elimina without recourse to the special measures in cated by the authors. The sources of carrier frequency C and D, originating in telephone and telegraph repeater oflces are, shown on Fig. 14 of the authors just mentioned (reproduced as 6 of the present drawings) quite substantial in amount. To reduce the ef fect of these sources of noise, suppression coils are installed on both sides of the telephon or telegraph repea ers on all voice frequency circuits at such repeater offices. Fig. l of the same authors shows this point quite clearly. The effects of these sources of noise are automatically reduced, in fact can be virtually eliminated by means of pad P and phase changer Q, with the arrangement of 7 and Fig. 8 of the drawings.

t is necessary, in order to secure a complete and thorou h picture of the possibilities of balancing out the effect of the sources of noise C and D, by means of the methods first disclosed in these specifications, to develop fully not only the amplitudes pictured on Fig. 6 of the drawings, (Fig. 14 of the paper by Weaver, Tucker and Darnell) but the phase angles as well. As defined by the authors just mentioned, the typical magnitudes indicated as sources of noise 0 and D originate in existing telephone and telegraph repeaters at carrier frequencies and transmitted by the non-carrier pairs into the carrier pairs. Evident-1y the carrier frequency voltages generated at telephone and telegraph repeater stations are harmonic components of lower frequency D. C. telegraph, telephone speech, telephone signalling and power supply voltages. While clearly the various big 1' harmonic components of any given original lo frequency iundamental are originally in phas with each other, they will not remain in phase as they reach the final receiving channels of the carier system.

If wedesignate, as

before, by (Na, (pa) the amplitude and original phase constants of any single component of either source of noise C or D, then the far end induced noise on the carrier pairs is given by, for an unbalance at the distance I from the origin of disturbances.

(See Collected Papers by G. A. Campbell, page 526.) In the above formula i propagation constant of voice frequency circuits for carrier ire quency surges and v2 equals the propagation constant of the non-loaded pairs to carrier frequency currents. The loading coils in the voice frequency pairs in the cable act as filters since the carrier frequencies herein considered are above the cutoff frequencies of the 3-88 and 1-1-88 loading coils used for the two-wire circuits. The integration of Equation 9 gives the noise effect produced in the following form, after due simplification and elimination of negligible factors and application of the theory of probabilities.

Where Ya and Za are average values per mile of circuit.

I have carried out quantitatively in the light of the known data available for 19 B. & S. gauge cable, the integration procedure indicated by Formula 9, and expressed this by Equation 10. If the amplifiers associated with each cable section gains adjusted to equal the line losses (as they actually are) and are furthermore adjusted to wipe out the phase changes due to the line, then 6- 2 also compensated for its imagi nary components as well. Table II, reproduced below, gives the results of the computations which I have carried out as to the resultant changes imposed upon the original phase of the various noise components in function of their frequency. The second column in this table shows the part played in these changes by the crosstalk coupling factor i. e.

It is important to note the following facts, based upon the use of 13-88 facilities for the voice frequency pairs.

(at) The differences between successiv channels each 4000 cycles wide are quite small, varying between a minimum of 02520" and a maximum of IO52II2II (b) These changes in phase: are independent of the length of the line. Therefore, whether we are dealing with a repeater spacing of 17 miles or 50 miles, the phase change compensations required between adjacent channels, carrying the same message, will be independent of the repeater spacings. This is a very great advantage, since the balancing of the sources of noise C and D oifers no greater difficulty for a repeater spacing of 50 miles than for a spacing of 17 miles in so far as phase angles are concerned. In so far as amplitudes are concerned, these will be proportional to the square roots of the exposure lengths, so that a repeater spacing of 50 miles would result in a relative amplitude of as compared with a repeater spacing of 1'? miles. This means an increase of 9.3 decibels in absolute value, over the repeater spacing of 17 miles, the basis of Fig. 6 of the drawings (Fig. 14 of Weaver, Tucker and Darnell). The differences in the amplitudes of the noise components however, and between adjacent or successive channels will remain the same for a. 50 mile repeater spacing as for a 17 mile spacing as the increase in amplitude is the same for all frequencies. Therefore, the balancing out of the noise components due to sources C and D can be carried out as accurately for a 50 mile spacing between repeaters as for a 17 mile spacing. The values of the artificial pad P and phase changer Q will be substantially the same for both repeater spacings.

The final value of Ni in Formula 10, when not completely balanced out, depends upon the observed original values of Na. Weaver, Tucker and Darnell, in their paper already referred to, show the following average increases in amount of noise power, due to sources C and D for adjacent frequency channels.

Source of Source of In the case of thermal agitation, the corre sponding differences varied between 1.7 and 2.0 decibels. Thus, it is interesting to note, all of these amplitude changes, in function of frequency, while not exactly alike, are of about the same order of magnitude and will lend themselves to almost complete elimination by means of pad P. These differences between channels, furthermore, will not change with increased repeater spacings. Comparison of the phase angle data as shown on Table I for noise thermal agitation effects and Table II for noise effects due to sources C and D, indicate that these are not only quite small, but are nearly alike and do not change with the length of repeater spacings.

There remains the question of heavy static E, a source of occasional disturbances, on open wires directly connected to pairs in the cable through which are routed the carrier circuits. The arm plitudes of heavy static, induced into the carrier pairs through the mutual admittance and reactance of these pairs with the cable pairs connected to open wires is, according to the data presented by Weaver, Tucker and Darnell, included between 39 and 47 db. above reference noise for the range of frequencies between 15,000 and 55,000 cycles per second. The subject of static reduction was covered in great detail in U. S. A. 2,282,299 recently granted to me on Antistatic devices for a higher range of frequencies, i. e. the broadcasting range. It was shown in that application that reductions in static of the order of 3l-2 db. could be secured. It will be noted that the curve for source of noise E is quite fiat. This is obviously due to the fact that the intensity of static decreases as the frequency increases while the converse holds true for the amplifier gains provided at carrier repeater points. The phase differences, due to the sources of noise E will be practically identical with those given in Table II. These do not change, as already pointed out, with the distance between carrier repeater points. While the absolute magnitude of the noise E does change in accordance with the square root of the distance between repeaters, the differences in magnitude, as between adjacent channels do not change for the reasons already pointed out. The average change in magnitude, as between adjacent channels, due to source of noise E is about 0.8 db. and is the smallest of all five classified sources of noise A to E inclusive.

We still have to consider, the important question of crosstalk. There are several ways in which this problem can be met. One arrangement, which is the one in use today, would use the specific measures outlined very clearly by Weaver, Tucker and Darnell in their previously cited paper. The use of these measures would mean that on Fig. 8 of the drawings terminals e, f and g, 71. would not be connected together in parallel to terminals 0, d and there connected to a cable pair as shown in full line. Instead, terminals e, f would be connected to one cable pair in one cable and terminals g, h to another cable pair routed through a different cable, as shown in dotted line on said Fig. 8 of the present drawings. The directional filters would be removed. The use of conductors in two different cables eliminates, as well known, near-end crosstalk but leaves partly solved the still formidable problem of far-end crosstalk. The problem is formidable, as, at the high frequencies required for carrier operation, the far-end crosstalk due to capacity unbalance in relation to received transmission is 2.50 times and that due to reactance unbalance is 90 times greater than for the voice frequency circuits in commercial use between urban communities. The specific measures necessary to overcome far-end crosstalk have been outlined in great detail by the paper previously cited (pages 251 to 257, A. I. E. E. Transactions, 1938). Without going into the details of these measures, a brief statement of what is required and what is provided in the art today would be as follows:

50 49: 1225 networks This rather complicated procedure results in a total reduction of about 16 decibels, when used in conjunction with (a).

The second arrangement for overcoming crosstalk difiiculties and limitations in conjunction with carrier transmission makes use of both of the underlying conceptions first disclosed in these specifications.

(a) A line transmission system, as illustrated on Fig. 4 of the drawings, and as fully described hereinbefore.

(b) A terminal equipment system as illustrated on Fig. '7 and Fig. 8 of the drawings, and as fully outlined hereinabove.

In this second arrangement, therefore, terminals 0, d of Fig. 8 of the drawings connect to terminals a, b of Fig. 4 of the drawings which depicts the transmission system required for the line. The dotted lines at c, d on Fig. 8 and at a, b on Fig. 4 or the drawings show this interconnection of the terminal and line systems. At the other end of the line, as shown on Fig. 4, a twoway carrier repeater system is required, connected in the specific manner shown on said drawings to aforesaid line system. The connections at the distant terminal point of the entire carrier system will be completely symmetrical to those outlined for the nearby terminal point, equipped as shown on Fig. '7 and Fig. 8 of the drawings.

As previously explained and defined, use is still made under this arrangement of two adjacent side bands for transmitting purposes. One of the two side bands is reversed in phase by means of transpositions on either the odd numbered or the even numbered side bands. The main purpose of this arrangement, in the present instance, is to balance out all thermal agitation efiects and remove the need for the short spacing of carrier repeaters. This spacing is, at the present stage or the development of the art, restricted to a maximum of 17-18 miles.

On Fig. '7 of the drawings channel No. l is the one with phase reversal in view of transposition 2!, 22, 23 and 24. Channel No. 2, a channel adjacent in frequency and of higher frequency is not provided with transpositions.

In a combination system of the type just described, any current induced on the upper pair of the line system depicted on Fig. 4 of the drawings will find its exact counterpart on the lower pair of this same line system. The near end crosstalk on the upper pair will be exactly equal, v

by construction, both in amplitude and in phase, to the near end crosstalk on the lower pair. In view, however, of the transposition I, 2, 3, 4 on the upper pair, the two sets of induced currents will be in direct phase opposition and will cancel each other. The same situation holds true, and exactly for the same reasons, with reference to far end crosstalk. In this last case transposition 5, 6, I, 8 plays the same role that transposition l, 2, 3, 4 played in the case of near end crosstalk.

The above possibility of securing a radical solution to the extremely important problem of eliminating crosstalk in the carrier range of wire transmission makes it unnecessary to route the cable pairs, carrying the carrier message, in two distinct cables (see hereinabove (e) of the present art) and avoids the remaining measures (a) to (d) inclusive required by the art at the present time.

Not only does the use of the new line transmission system depicted on Fig. 4 of the drawings eliminate crosstalk, both near end and far end, but it also eliminates at the same time the important sources of noise designated as C, D and E. (See Fig. 6 of the drawings as to the magnitude of these efiects.) The reason'for this is quite evident. These efiects have their origin in the voice frequency repeater and also telegraph repeater ofiices and are transmitted through the voice frequency pairs, by mutual admittance and reactance unbalance, onto the carrier pairs working in the same cable through which are routed the voice frequency pairs. As just explained in connection with carrier frequency crosstalk from one system of carrier channels on one pair to a similar system on another pair, the induced effect due to any current carried by a voice frequency pair upon the upper pair of wires of the line transmission system of Fig. 4 of the drawings has an exact counterpart, but in phase opposition, on the lower pair of the same line system. Hence virtual elimination of the sources of noise C, D and E. If any residual effects are left, which in any event would be lower in amount than any such effects obtainablein the present art, they will receive a second setback and will be still further greatly reduced in view of the use of adjacent channels for the same telephone message and the phase opposition obtained through the effect of transposition 21, 2-2, 23, 24 of Fig. '7 of the drawings. In view of the fact that the residual effects, if any, of sources of noise C, D and E are already so far below'the noise efiects A and B due to thermal agitation, the use of balancing pad P and phase changer Q can be restricted to the balancing out of thermal agitation, thereby obtaining even greater ac-' curacy in the elimination of sources A and B. In illustration of the above point, if we were interested in determining what additional reduction can be obtained by the phase opposition between the two successive channels 12,000-16,000 cycles and 16,000-20,000 cycles, we would refer to Tables I and II. We note from Table I that to compensate for thermal agitation we must provide a phase change of (1 36' 30) in the higher of the two channels. Table II shows that to compensate for sources of noise C, D and E in the same two channels, the phase change required in the higher channelis .-(0 '25 20"). If, therefore, a change of +(1 36' 3,0) is provided for the benefit of sources A and B, then'the difference left uncompensated is increased to (-2 1' 50"). The pad P required to compensate for A and B being 2.00 db. and that for source C being 2.50 db., the two vectors representative of noise G in the two channels are, respectively and relatively to each other, 1.00 at angle zero, and 1.0593 at an angle of 2 1' 50". The resultant,after phase reversal, is equal to .0762 or a reduction of 22.3 decibels. Additional reductions of similar order will hold for source of noise D. These reductions are additional toreductions ,of the order of 45-50 db., due to the use of the arrangement of Fig. 4 of the drawings for theline conductors. The reduction that can be obtained for source of noise D will be greater and that for noise E, due to the opposition of phases in the adjacent channels, somewhat smaller. For source of noise E in the lower frequency channels the residual effect left is 19 db. lower. The total reduction that may be expected is, therefore, included between 64-69 db. for source of noise E.

It was stated, in the early part of this specification. that in order to obtain the full benefits of the line transmission system as depicted on Fig. 1 and Fig. 4 of the drawings, it would be desirable to route carrier circuits in a new type of toll cable. While this is true, it is possible, nevertheless, to realize the system shown on Fig. 1 and Fig. 4 of the drawings with present day type cables, though the reduction in crossstalk that can be obtained in this manner will not be as large as would be possible with a new type of cable specially designed for carrier operation. Emphasis has been laid, in present day practice, on the need of reducing to a minimum the unbalance between phantoms and their respective side circuits in quadded cables used for voice frequency operation. The phantoms are no longer desired for carrier operation. To obtain maximum benefits in crosstalk reduction, and also in the reduction of sources of noise C, D and E, it becomes desirable to modify the present attempt to obtain zero or a minimum unbalance between each pair and every other pair in the cable to the extent of requiring that the unbalance of side circuit No. 1 of quad No. 1 with respect to a third pair will be the same as that of side circuit No. 2 of quad No. 1 with respect to the same third pair. If a cable of the present type (i. e. duplex) is being installed new, one possibility for a line of action would consist in applying the capacity unbalance or reactance unbalance testing at intervals, perhaps not any more frequent than once per mile. At such points, the upper pair and the lower pair would be bridged together, after inserting transpositions at the ends of this section in accordance with Fig. 4 of the drawings. Evidently when equality of unbalance is obtained on the two pairs of the bridged system, the measured overall unbalance for the system as a whole is equal to zero. Therefore, from this point on, the testing can be carried on using present day procedures, i. e. adjacent sections throughout the length of the cable can be matched so as to reduce further the residual efiects. This means connecting the two wire (or the bridged) portions of the adjacent sections of the proposed 4- wire system, when applied to present day type cables, either straight through or through a transposition, depending upon the signs of the unbalances. This practice is too well known in the art to require elaboration here. In present practice, these preliminary unbalances are confined to a quadded group. Under the proposed arrangement, it appears desirable to consider for balancing purposes all six of the adjacent quads to the quad under test. It is unnecessary to consider more distant quads or pairs, as the proposed system is automatically balanced for such distant quads. Several alternatives present themselves for consideration. One alternative would be as follows. If the side circuits under test are those of quad No. 1 and the six adjacent quads are numerically designated as 2 to 7 inclusive, then the first unbalance test would be applied to obtain zero unbalance or reduced unbalance between the bridged pairs of quad No. 1 and quad No. 2, the second test would be applied to those of quad No. 1 with respect to quad No. 3 and so forth in sequence. This procedure would then be repeated. A second alternative line of action would consist in determining the average unbalance of the bridged pairs of quad No. 1 with respect to the bridged pairs of all six adjacent quads for two adjacent sections of line, and joining such two adjacent sections of squad No. 1 so as to re duce or oifset against each other the unbalances in these two adjoining sections. A third alternative would consist in carrying out the unbalance testing, as at present, for an entire repeater section, and also making use of the special balancing arrangement for each pair against every other pair. Before bridging together the two side circuits with their respective final residual unbalances, the coil setting on one side circuit would be deliberately changed from its minimum unbalance position so that the unbalance so created would equal the residual balance of the other side section. There are two ways of accomplishing this, since the coil settings can be made either positive or negative. In order to care for both near-end and far-end crosstalk, it may be necessary to carry out this balancing at both ends of a repeater section. The first balancing would take care of the far-end crosstalk. The second balancing would then be carried out to eliminate residual near-end crosstalk with the further possibility that it may be desirable to install a simple network consisting of a pad and phase changer on one or the other of the two side circuits to secure still greater refinement.

Other simple alternatives will occur to the mind of those versed in the art. It appears un-- necessary to cover them all.

The following statement gives a broad picture as to the effect of the application of the proposed system to present day type cables from the standpoint of crosstalk.

1) The side-to-side crosstalk problem, the most difiicult one, is automatically removed from consideration. This means a gain in the way of reduction in crosstalk of 10 db. approximately. The two pairs of a phantom being used in parallel, under the proposed system, the crosstalk between the two merely adds a small increment to What has been termed sidetone.

(2) The use of two channels, instead of one, allows the operation of each channel at a carrier frequency level of 3 db. less on the line, thus decreasing the crosstalk due to capacity unbalance by that amount and the reactance unbalance by the same amount since the change in line current amplitude is also 3 db. downward.

(3) The use of two pairs for the same message, in the manner shown on Fig. 4 of the drawings, results in a further decrease in crosstalk, as fully explained hereinabove, since the crosstalk on one pair is opposed to the crosstalk on the second pair.

All of the various measures described by Weaver, Tucker and Darnell result in a decrease of 16 db. in crosstalk. This is the highest figure and holds for 60,000 cycles, the highest frequency used. (See reduction from 67 db. to 83 db., Fig. 13, page 257, A. I. E. E. Transactions 1938.) The two steps alone, as above mentioned, bring about a reduction of 13 db. If nothing 'else was done and no account was taken of the improvement due to step 3 but the poling process was applied with an expected gain of 9 db. the same as in the present art, the total gain exclusive of step 3 would be 22 decibels. This confirms the View, that even with present day type cable there would be no diffi'culty in reducing far end as well as near end crosstalk to allowable limits.

With the use of the general methods described hereinabove for reducing noise and crosstalk, particularly when applied in conjunction with cables specially designed for carrier operation, it may even be possible to use the same carrier frequency for transmission in two directions and in the same cable. This is scrupulously avoided at present. The use of the same carrier frequency for transmission in both directions results, as well known, in the saving of costly terminal equipment.

It may be pointed out, furthermore, that on very long circuits where cumulative residual effects become important, the methods hereinabove indicated can be carried out one step further by terminating the carrier terminals used for such long haul trafi'ic at an intermediate point where carrier terminating equipment is available and carrying out additional balancing steps, in the manner indicated hereinabove, at such intermediate point.

The economic advantages of the system of carrier transmission, described for the first time in these specifications, will now be apparent. The proposed system will cost nearly twice as much in terminal equipment. On the other hand, the system first disclosed herein will use one-half and perhaps as few as one-third as many carrier repeater stations as required by the present art. As, in any carrier system installation there are just two points at which carrier terminal equipment is provided, while the number of carrier repeater stations is very large, the economic advantages will be correspondingly large. Furthermore, it will be superior to the present system in quality of service. In addition, there will be no need for installing a second toll cable in order to avoid undue crosstalk. This feature, in itself, represents a very important saving in cost of installation and in the extension of carrier systems. It should be remembered, however, that twice as many pairs will be required if use is made of Fig. of the drawings. This increase, in turn, can be avoided if it is found feasible to use the same frequency in both directions of transmission over the same cable pair.

, It is necessary to point out that with two adjacent channels used for each direction of transmission, the number of simultaneous messages would be reduced to six (6) with terminal equipment of the type K or 12 channel type. With the line conductors used on a two-way basis, which is the case, when these conductors are connected as shown on Fig. 4 of the drawings, 24 channels are required in the line. To secure these carrier channels and still continue to use the 12channel type carrier equipment, it is necessary to adopt the two-step group modulation and corresponding twostep group demodulation arrangement, instead of the single step modulation arrangement shown on Fig. '7 and Fig. 8 of the present drawings, which is satisfactory when the two directions of transmission are carried on separate sets of conductors as previously described. The-two-step modulation arrangement is fully described by B. W. Kendall and H. A. Aifel in a paper bearin the title A 12 channel carrier telephone system for open wire lines, published in A. I. E. E. Transaction, 1939, pages 351-353. Reference may be made to the paper just cited for an understanding of the details; It" has not been illustrated on the drawings, as the present invention can be fully understood Without eni-. cumbering these specifications with informationnot essential to such understanding. As applied to the system of two phase-opposed channel arrangement for each direction Of transmission, first disclosed hereinabove, the channels required in the line conductors would be included between 12 and 60 kilocycles in one direction of transmission and 68-116 kilocycles in the opp site direction. Considering, for instance, an individual channel of Fig. 7 of the drawings Working at 108- kilocycles foroutward transmission, it would be group-modulated at 340 and at 460 kilocycles. In the opposite direction of transmission, for inward transmission, the same channel would be demodulated at 340 and also at 380 kilocycles. The following table will make clear the-procedure required.

DIRECTION or TRANSMISSION BETWEEN TERMINALS A AND B A toB B to A (a) Outgoing: Kc. Kc.

Individual channel frequency 108 108 First step-group modulation... 340 340 Retained side band frequency 448 448 Second step-group modulation H. 460 380 Retained side band frequency transm ted over line 12- 68 (b) Incoming:

First step-group demodulation 460 380 Retained side band frequency. 448 448 Second step-group demodulation l. 340 340 Retained side band frequency 108 108 If, on the other hand, the line conductors are arranged on the basis of Fig. 5 of thedrawings, instead of Fig. 4, then only twelve channel will be required in each direction of transmission, making again a total of 24 channels, but in this case the twelve channels in the two directions of transmission can have the same frequency ranges, i. e. 12-60 kilocycles. The use of the higher frequencies with attendant higher line losses and correspondingly higher required amplification is avoided by thi means.

It is possible to secure the transmission of twelve (12) simultaneous messages with the system first disclosed in these specifications by increasing h r nge of frequencies in the individual channel terminal equipment from 12-60 kilocycles to 12-108 kilccycles and using either the arrangement of Fig. 4 or Fig. 5 of the drawings, preferably the latter. If use is made of Fig. 5 of the drawings for the line conductors, the top frequency in the lin conductor will not exceed 108 kilocycles. The increase from six to twelve simultaneous messages can be obtained with the arrangement last described, by: doubling the amount of the individual channel portion of the terminal equipment (diagrammatically shown on Fig. 7 of the drawings).

It will be noted that it is not necessary to provide two cables when use is made of the arrangement just described.

Another alternative method in the use of the present invention would consist in retaining. the use of short spacing between repeater to care for the sources of noise A and B in accordance with present practice in conjunction with the use of the line conductors in the manner indicated on Fig. 4 and Fig. 5 of the drawings. When this alternative arrangement is used, all sources of noise C to E inclusive, including also crosstalk are eliminated. The use of a new type of cable is assumed in making the above statement. The present-day type twelve-channel in dividual channel terminal equipment is sufi'icient to secure twelve simultaneous messages in this case.

It will be noted that with neither of the above two arrangements just described will it be necessary to provide two cables for the line conductors.

I claim:

1. In a system of carrier transmission over wires, a transmitting arrangement comprising means for modulating the same signal at two different carrier frequencies, means for selecting the upper side bands of the respective modulation products in each case, mean for reversing the phase of only one of said side bands, and means for impressing upon said wires both the reversed and unreversed side bands.

2. In a system of carrier transmission over wires, a transmitting arrangement comprising means for modulating the same signal at two different carrier frequencies, means for selecting the lower side bands of the respective modulation products in each case, means for reversing the phase of only one of said side bands, and

means for impressing upon said wires both the reversed and unreversed side bands.

3. In a system of carrier transmission over wires, a transmitting arrangement comprising means for modulating the same signal at two adjacent carrier frequencies, means for selecting the upper side bands of the respective modulation products in each case, means for reversing the phase of only one of said side bands, and means for impressing upon said wires both the reversed and unreversed side bands.

4. In a system of carrier transmission over wires, a transmitting arrangement comprising means for modulating the same signal at two adjacent carrier frequencies, means for selecting the lower side bands of the respective modulation products in each case, means for reversing the phase of only one of said side bands, and means for impressing upon said wires both the reversed and unreversed side bands.

5. In a system of carrier transmission over wires, a receiving circuit comprising means for separating incoming waves into pairs of channels, each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, means for segregating each such pair of incoming waves into two different frequency channels, means for separately demodulating said two waves with carrier supplied in said channels, means for retaining only the lower side bands of the respective demodulation products, means for reversing the phase of only one of said side bands, and means for recombining said two side bands thereby restoring the signal impulses in said side bands into a single reinforced signal impulse.

6. In a system of carrier transmission over wires, a receiving circuit comprising means for separating incoming waves into pairs of adjacent frequency channels, each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, means for segregating each such pair of incoming waves into two adjacent frequency channels, means for separately demodulating said two waves with carrier supplied in said channels, means for retaining only the lower side bands of the respective demodulation products, means for reversing the phase of only one of said side bands, and means for recombining said two side bands thereby restoring the signal impulses in said side bands into a single reinforced signal impulse.

7. In a system of carrier transmission over wires, a receiving circuit arranged for the reduction of thermal agitation disturbances in the line wires and disturbances in the amplifiers comprising means for separating incoming waves into pairs of channels, each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, means for segregating each such pair of incoming waves into two different frequency channels, means for separately demodulating said two wave with carrier supplied in said channels, means for retaining only the lower side bands of the respective demodulation products in each case, and means for reversing the phase of only one of the two resultant side bands, thereby producing two distinct disturbance waves in phase opposition, and means for recombining said two re-' sultant side bands into a greatly reduced disturbance wave and a reinforced signal wave.

8. In a system of carrier transmission over Wires, a receiving circuit arranged for the reduction of induced disturbances from adjoining voice frequency and telegraph circuits, comprising means for separating the received waves into pairs of channels, each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, means for segregating each such pair of incoming waves into two different frequency channels, means for separately demodulating said two waves with carrier supplied in said channels, means for retaining only the lower side bands of the respective demodulation products in each case, and means for reversing the phase of only one of the two resultant side bands, thereby producing two distinct disturbance waves in phase opposition, and means for recombining said two resultant side bands into a greatly reduced disturbance wave and a reinforced signal wave.

9. In a system of carrier transmission over wires, the method of reducing in the receiving circuits of said system the disturbances induced into and the disturbances originating in said wires, which consists in segregating said disturbances and the received carrier channel signal impulses into pairs of channels with each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, translating said disturbances into groups of two similar waves, each of these two waves comprising the same frequency interval and the same signal but with the first of said similar waves opposing the second and recombining said two waves into a single residual wave of reduced disturbances and a reinforced signal wave.

10. In a system of carrier transmission over wires, the method of reducing in the receiving circuits of said system the disturbances induced into and the disturbances originating in said wires, which consists in segregating said disturbances and the received carrier channel signal impulses into pairs of adjacent frequency channels with each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, translating said distrubances into groups of two similar Waves, each of these two waves comprising the same frequency interval and the same signal but with the first of said similar waves opposing the second and recombining said two waves into a single residual wave of reduced disturbances and a reinforced signal wave.

11. In a system of carrier transmission over wires, the method of reducing in the receiving circuits of said system the disturbances originating in the group demodulating circuits of said receiving circuits which consists in segregating said disturbances and the received carrier channel impulses into two sets of different frequency bands, translating said disturbances into groups of two similar waves, each of these two waves comprising the same frequency interval and the same signal but with the first of said similar waves opposing the second and recombining said two waves into a single residual wave of reduced disturbances and a reinforced signal wave.

12. In a system of carrier transmission over wires, the method of reducing in the receiving circuits of said system the disturbances originating in the group demodulating circuits of said receiving circuits, which consists in segregating said disturbances and the received carrier channel signal impulses into pairs of adjacent frequency channels with each such pair modulated by the same signal but in opposite phase relationship at the distant terminal, translating said disturbances into groups of two similar waves, each of these two waves comprising the same frequency interval and the same signal but with the first of said similar waves opposing the second and recombining said two waves into a single residual wave of reduced disturbances and a reinforced signal wave.

13. In a two-way transmission system between two points, a system of two sets of conductors, having equal mutual reactance and capacity admittance unbalances with all other similar sets of conductors, with means at each end of the transmission system for opposing the residual unbalances in the first set of conductors to the residual unbalances of the second set of conductors of the same system of two sets of conductors, thereby eliminating near-end and far-end crosstalk, near-end and far-end babble and all induced disturbances at both ends.

14. In a two-way wire transmission system between two points, consisting of two pairs of line conductors bridged in parallel at both ends, having equal mutual reactance and capacity admittance unbalances with all other line conductors, each pair arranged for transmission in both directions, only one of said two pairs being provided with phase reversal at one end of the transmission line and only the second pair being provided with phase reversal at the other end of the transmission line before being bridged with each other.

15. In a two-way wire transmission system between two points, consisting of two pairs of line conductors bridged in parallel at both ends, having equal mutual reactance and capacity admittance unbalances with all other line conductors, each pair arranged with transmission in both directions, only one of said two pairs being provided with phase reversals, one phase reversal at each end of the transmission line before being bridged with said second pair.

16. In a two-way system of carrier transmission over wires between two points, consisting of two pairs of line conductors bridged in parallel at both ends, having equal mutual reactance and capacity admittance unbalances with all other line conductors, each pair arranged for transmission in both directions, only one of said two pairs being provided with phase reversal at one end of the transmission line and only the second pair being provided with phase reversal at the other end of the transmission line before being bridged with each other.

17. In a two-way system of carrier transmission over wires between two points, consisting of two pairs of line conductors bridged in parallel at both ends, having equal mutual reactance and capacity admittance unbalances with all other line conductors, each pair arranged with transmission in both directions, only one of said two pairs being provided with phase reversals, one phase reversal at each end of the transmission line before being bridged with said second pair.

18. In a two-way wire transmission system between two points, consisting of two systems of line conductors with each systemarranged for one-way transmission but in opposite directions, each system consisting of two pairs of conductors between said points, having equal mutual reactance and capacity admittance unbalances with all other line conductors, both pairs of any one system arranged for transmission of speech in the same one-way direction, only one of said twopairs being provided with phase reversal at one end of the transmission line and only the second pair being provided with phase reversal at the other end of the transmission line before being bridged with each other.

19. In a two-way wire transmission system between two points, consisting of two systems of line conductors with each system arranged for one-way transmission but in opposite directions, each system consisting of two pairs of conductors between said point, having equal mutual reactance and capacity admittance unbalances with all other line conductors, both pairs of any one system arranged for transmission of speech in the same one-way direction, only one of said two pairs being provided with phase reversals, one phase reversal at each end of the transmission line before being bridged with said second pair.

20. In a two-way system of carrier transmission over wires between two points, consisting of two systems of line conductors with each system arranged for one-way transmission but in opposite directions, each system consisting of two pairs of conductors between said points, having equal mutual reactance and capacity admittance unbalances with all other line conductors, both pairs of any one system arranged for transmission of speech in the same one-way direction, only one of said two pairs being provided with phase reversal at one end of the transmission line and only the second pair being provided with phase reversal at the other end of the transmission line before being bridged with each other.

21. In a two-way system of carrier transmission over wires between two points, consisting of two systems of line conductors with each system arranged for one-way transmission but in opposite directions, each system consisting of two pairs of conductors between said points, having equal mutual reactance and capacity admittance unbalances with all other line conductors, both pairs of any one system arranged for transmission of speech in the same one-way direction, only one of said two pairs being provided with phase reversals, one phase reversal at each end of the transmission line before being bridged with said second pair.

HUGHES MOURADIAN.

Images