US 1792273 A
Abstract available in
Claims available in
Description (OCR text may contain errors)
Feb. 10, 1931. A. BYK ETAL ELECTRICAL CONDUCTOR Fil-ed Nov.
Paemed Feb. 1o, 1931 ALFRED BYK, 0F BERLIN-CHARLOTTENBURG, AND
HANS JORDAN, 0F BERLIN- KARLSHORST, GERMANY, ASSIGNORS T0 GENERAL ELECTRIC COMPANY, A CORPO- RATION OF NEW YORK ELECTRICAL coNnucToR Application filedA November 14, 1927, Serial No. 232,988, and in Germany March 30, 1927.
This invention relates to electrical conductors, and particularly to multiplex cable circuits such as may be employed in telephony.
,-ln a cable quad having two pairs of wires which may be used in phantom circuit arrangement, the relative positions ofthe four wires is of importance. Their position largely determines the capacities of the side circuits of the phantom circuit as well as the disturbing capacitative coupling between the side circuits and the phantom. As regards the capacitative relationships, it is'very important to guard against 'too great a capacity in .the phantom circuit. A low phantom capacity may be realized by twisting the two separatelytwisted pairs of conductors, one about the other, as is done in multiple twin cabling. Such `a twisted quad has a veryv irregular shape and it is difficult to constructv such a quad to meet the req-uirements of precision in manufacture whlch are necessary to insure its freedom from high capacitative 2 coupling. Far lgreater regularity in construction may be obtained in a quad' of square cross-sectional area in which the conductors located at adjacent corners of the square, or the conductors located at diagonally opposite corners, are employed as one `side circuit, the other remaining two conductors in either case being employed as the other side circuit. By pairing diagonally opposite `conductors of a quad, it is possible to operateiboth the side circuits and the phantom circuit without substantial interference. Yet, any undesirably high phantom circuit capacity of the latterl arrangement may be greatly reduced and possibly completely avoided if adjacent conductors of the cablel quad are paired to represent each of the side circuits. In the latter case, it will be clear that both side circuits run .constantly parallel with eachother and are coupled with one another.
To neutralize the coupling in side circuits which run constantly parallel to each other, transpositions may be employed as in open wire construction in which one or both side circuits may be transposed in accordance with definite rules in order that the couplings of individual sections may be mutual. If
such transpositions be employed in the conductors of a cable quad in which the wires of each pair run parallel to the wires of the other pair, then the problem of interference elimination becomes more complex due particularly to the closer proximity of the side circuits resulting in greater interaction therebetween. Thus, it becomes apparent that special transposition schemes will be required for cable circuits which will differ considerably from those of open wire lines.
t Any practical transposition scheme must provide as much freedom from coupling in its ideal construction as is provided by the well-knownA form of star quadded coupling, i. e., a-quad in which the conductors are spirally wound about a common axis, the diagonally opposite conductors forming the respective side circuits. There is a greater necessity for care in the case of cable circuits than in open wire lines to reduce the detrimental eifects of unavoidable manufacturing irregularities to reduce any couplings which may arise through irregular or even systematic deviations of the wiring from their theortical positions. Any manufacturing irregularities may, in practice, introduce couplings only to a permissible and `predetermined extent. A transposition has a finite extensionand has a definite influence on the coupling relationships as well` as on the mutual capacities. Any useful transposition scheme must take into account any unavoidable imperfections in the actual construction of a cable quad.
The primary advantage of this invention is to provide a transposition scheme which guaranteesA substantial freedom from coupling in the practical construction of the cable quad as well as in its ideal construction.
This invention maybe better understood from the detailed description `hereinafter following, when read 'in connection with the accompanying drawing, in which Figure 1 represents two side circuits of a quad having an equal number of transpositions; Figure 2 shows a similar arrangement of side circuits in which one has four transposition sections as compared to the one transpos1t1on section in the other; Figure 3 shows a similar arbe appreciable, l and, after n introduces-an additive effect of considerable A transposition design Y has eleven transposition sections and the other seven; and Figure 6 shows a modification of the arrangement in Figure 5 in which the transposition sections in one side circuit are displaced with respect to those in the other.
for the conductors of the side circuits of a cable quad which are parallel to each other and, under ideal conditions, are free Jfrom coupling, involves an arrangement in which the transpositions take place in even pitch. A transposition arrangement in even pitch is shown in Figure l of the drawing. The transposition sections on both side circuits are equal inr length, a transposition on one side circuittating place midway between two transposition's on the other side circuit. In practicing this transposition scheme, it may occur through some necessary constructional measure or through a possible inaccuracy of slight order, that one of the movements of the transposing device employed in actual practice may be eiected so as to introduce a regularly recurring error. The transposing device may produce a transposition regularly at an angle slightly more or less than 180 degrees. Such an error may be repeated continually under the same condiytions, introducing a systematic coupling disturbance, which in a single section may not but which occurs constantly passing through several sections,
The two. side circuits of a 'cable may be provided with transposition sections of different lengths and still be in even pitch, i.-e.,
the number of transposition schemes on one side circuit for a` given .length may be some even multiple of the -number of transposition positions take place at dierent lengths in the respective side circuits, a transposition on the side circuithaving sparsely locatedtrans- -positions may coincide with a transposition on the other side circuit having a greater number of transpositions at regular intervals. Such a transposition'scheme is shown in Figure 2 of the drawing. In this embodiment, each transposition section of the lower side circuit is four times as long as y a transposition section in the upper side circuit.
However, in such an arrangement, manufacturing irregularities, 'such as are described hereinabove, may also be presented. In addition to these manufacturing'v irregularities,
anv irregular distribution of the transposicoupling disturbances resulting -sunrof the lengths of transpositions.
tions on the respective side circuits may be a further source of error introducing a disturbance in the balance between the side circuits. Thus, there is some disadvantage in having a pair of side circuits in which one of the side circuits has considerably fewer transposition ysections than the other of the side circuits by virtue of the fact that small and ordinarily inappreciable coupling disturbances may occur at each transposition.
In the transposition scheme of this invention, it is intended to greatly minimize any from mechanical imperfections. This is accomplished by having the relative positions of the transpositions on the `respective side circuits as diverse as possible in order that any possible schematic coupling errors, which may be inappreciable in a single transposition section, may not additively exceed some definite amount. At the same time it is intended to suitably transpose each side circuit at regular intervals, so that the number of transposition sections on one `side circuit will not greatly preponderate over the number of transposition sections on the other side `circuit. This may be accomplished by providing certain transposition schemes in uneven pitch in place of the schemes in even pitch. In order to more clearly understand the special transposition schemes in uneven pitch, which may bring about the above-mentioned advantages, even in view of practical manufacturc ing difiiculties, it becomes necessary to eX- amine somewhat closely diiferent possible forms of transposition schemes.
In quadded cabling in which the wires at two adjacent cornersof a square cross-section represent one side circuit, the wires at the remaining two ladjacent corners representing the other side circuit, the substantial elimination of crosstalk disturbances 'requires the, use of a transposition scheme, in which the all transposition sections which are separated from the first sec.- tion by an even number of transpositions, in-
cluding the first section itself, be equal to the sum of the lengths of all sections separated from the first section by an uneven number of Thus, 'the lengths of the transposition sections of both side circuits are tol be added together, and the diderence between the sum of all sections of the first (positive) kind and the sum of all sections of the second (negative) kind, will be called the effective disturbing length for that particular section of cable.
In the case of transpositions in even pitch, although transposition sections may be different from each 4other on the respective side circuits, these sections are individually equal to each other on the sameside circuit. The lengths of thetransposition sectionson both side circuits may, however be incommensurable. An incommensurable relationship between the lengths of the transposition sections may, in practice, be closely simulated by some commensurable relationship. `Accordingly, the length of a transposition section may be so chosen in a quad in whichthe side circuits run parallel 4to each other, that the transpositions at the ends of a definite length of cable coincide, while there is no coincidence of transpositions anywhere within that length. The length of the cable quad having equal sections on the respective side circuits providing coincident transpositions at its beginning and at its end, will be referred to hereinafter as the basic interval.
The basic interval may be considered as composed of m and n transposition sections on the respective side circuits, m and n being whole numbers. lt will be assumed hereinafter that m is greater than n. Moreover, a basic inter- -tion section in a side circuit having m sections is proportional to n, and conversely, the length of al transposition section in the other side circuit is proportional to m.
In the case of transposition fields of diiferent lengths arranged in even pitch, the following relationship will exist:
sleIsiH l in which c is a positive whole number and CZ is an odd number. n must necessarily be unity, for the right side of the Equation (l) Vwould not have an integral value unless m were divisible by n and unless m and n were relatively prime numbers. This, obviously, can only be true when n is unity. Transpositions in uneven pitch are those in which the relationship between m and n, 1. e. E is not represented by an even integer. An arrangement in which the number of transposition sections in the basic interval is even in one side circuit and uneven in the other, the uneven number ditfering from unity, or vice versa, represents a case in uneven pitch. Also, an arrangement in which the transposition sections in the basic interval are uneven integers in the respective side circuits, even though the smaller number be unity, represents another case of uneven pitch. Of course, an arrangement in which the transpositions in both side circuits are even integers is not here considered'because two even numbers `are not relatively prime.
Suppose m to be even and n uneven and suppose that the initial transpositions'on the two side circuits are coincident, then the transposition l -l-cm) of the side circuit with shorter spacings will coincide with the transposition (I+/7cm) ofthe side circuit with longer spacings, 7c being a positive whole number. Such an arrangement is illustrated iii'FigHre 3 of the drawing in which m equals 8 and n equals There a transposition on the side circuit of shorter spacing will lie between two transpositions on the side circuit of longer spacing, two short transpositions following each other consecutively. To calculate the effective disturbance, i. e., the effective disturbing length within a basic interval, it may be desirable to divide the basic interval into short and long transposition sections following each other alternately. Figure 3 shows portions of the arrangement enclosed by rectangles deined by dotted lines. The entire basic interval is symmetrical about a plane intersecting the quad at its center. Since only one transposition is located atl the center o f the basic interval, that transposition being in the sidecircuit of shorter spacing in which m is even, the two symmetrical transposition sections on both sides of the center line will be separated from each other by an uneven number of transpositions. Consequently, their disturbances will be mutually compensated, andthe total effective disturbing length of the basic interval will be equal to zero.V The same conclusions may be derived for the case in which lm is uneven and n is even. rlhus, in general, in the basic interval in which either m or n is even in one of the side circuits and uneven in the other of the side circuits, there is produced a balanced basic interval, and, consequently, a balanced cable quad, the relative positions of the transpositions on the respective side circuits being considerably more diverse than they are in the case of even pitch.
Figure 4 shows another arrangement in which m equals 3 and n'equals l. Since n equals 1, the conclusions derived hereinabove do not apply, and this particular case of uneven transpositions becomes of little interest.
Figure 5 shows an arrangement in which m and n are both uneven. Here m equals 11 and n equals 7. Moreover, the last transposif tions on the respective side circuits are coincident. Where m and n are uneven, no transposition is located at the center of the basic interval. Two transposition sections which are symmetrical about the center are separated from each other by an even number of transpositions and, consequently, their disturbances are additive. In such an arrangement, each basic interval may be considered to have a transposition section exhibiting a positive charge at its beginning. The disturbances of the individual bas-ic intervals are Aalso additive. To eliminate interferences, it
\ tribute to the eiective disturbance by the It is to be noted in the arrangement shown in Figure 5 that the transposition sections in the lower side circuit number 7 and yalternate in respectto the character of the electricalcharge. Since'the number` of transposition sections within the basic intervalis uneven, the effective disturbance may be considered proportional to 7. It may be shown that in any arrangement in which m and n are both odd integers the effective disturbance is +71? It is also to be noted that at theleft of the center, the lengths of the transposition sections in the space intervals of the irst side circuit all appear as whole numbers, each greater than zero and less than 7. In other words, the numbers 1, 2, 3, 4, 5 and Y6 occur, each occurring but once. Here, also,
the even integers exhibit negative charges and,`
and the negative transposition sections Vcon\y amount Thus, theentire disturbance, including both the positive and negative contributions, is
2 Consequently, the etl'ective disturbance with- 1n a basic lnterval in any arrangement in whichy both mand n are odd integers can never vanish.. Thus, it becomes impossible to eliminate the effect of crosstalk disturb! ances by means of transpositions where m and n are both odd-integers, when'terminal transpositions within the basic interval in both said circuits are coincident, as assumed.
It is possible, however, to displace the basic interval on one side circuit with respect to the same interval on the other side circuit. It .may be generally shown that in an arrangement in Which m and n are both odd,
the displacement of the terminal transposition of one side circuit with respect, to the terminal transposition of the other side ,circuit to the yextent of one-half the common measure of the transposition sections causes The l positive thev effective disturbance to become zero. A displacement of the basic interval in one side circuit to the same interval in the other side circuit by 1/271l may be made.
Figure 6 of the drawing shows the arrangement of Figure 5 where m equals^11 and n equals 7 ,A the terminal transpositions being displaced by one-half with respect yto each other. The arrangementin Figurev 6j provides the followingl spacing within the basicv interval upon the side circuit having `fewer transpositions: 0; 0.5; 7; 11.5; 14; 21;
.The spacing provided by the arrangement of Figure-'6 producesthe Ifollowing disturbances z-l- 55.5) (66.5 63) (77-70) -f +6.5+2.5+ 1.5+5.5+7+4.5+o.5+3.5+7=38.5. y
Thus, the net positive disturbance is equalto one-half of the length of the entire basic interval: However, the negative transposition sections introduce a negative effective disturbance of 38.5, the negative disturbance exactly compensating the positive disturbance. y l The particular advantages to be derived from an arrangement having an odd number Aof transposition sections become even more evident as the number of transposition sections with the `basic. interval increases. Thus, it will be shown that there is considerable advantage in having mand n comparfatively 'large numbers, these numbers representing one number` of transposition sections within the basic interval in the respecof the relative positions taking place within that interval.
It is'to be noted that a great increase in the magnitude of Y'm and n does not necessarily imply a corresponding enlargement of the transposition sections. The length of the transposition section on one of the side circuits may be selected freely, the length tive side circuits, no periodicity or repetition of the sections on the other side circuits being ldetermined from the relationship existing between fm and n. If m or n beof substantially the same order of magnitude as their difference, then it may occur as a result of the practically insignificant yariatio'ns in the length of the transposition sections, that fm, and n have a common divisor, whereupon the ratio of m to n becomes considerably smaller than anticipated. Thus, when m equals 1000 and n'equals 499, the complete repetition of the relative' positions of the transpositions on the two side circuits will, of course, occur only after 499 long transposition sections or after 1000 short transposition sections, respectively. Yet, if the4 longer transposition section be decreased in length by, for example, 1/500th of its design length, then the ratio of m to 'n may become 1000/500=2. Accordingly, the relative positions of the transposition sections will be repeated after two short sections or one of the longer sections, respectively, and the advantages mentioned hereinabove in 'connection with the employment of a great number of transposition sections Within a basic interval disappears.
If, however, m and n be comparatively large numbers and their difference be small as compared to these numbers, thenthe dithculty outlined hereinabove may be avoided.
' i If, for example, m equals 1000and n equals 989, then, should n be increased by 1, vthe ra.- tio of m to n would become 1000/990= 100/ 99. Here an increase in ln, by l causes areduction in the number of transposition sections Withlin the vbasic interval by almost 90%.
Then, if the transposition sections be properly accomplished in practice, resulting in a reduction in the values of m and n, the
reduced magnitude of m suchas m By reducing the value ofe, the order of magnitude of lml and also that of n, initially of the same order of magnitude, may be safely kept at a suitably high value. Consequently, the number ofv transpositions on the side circuits should not diii'er from each other too much, so vthat any unequal distribution of the transpositions on the respective side circuits Will not result in disturbing capacitative/unbalances. In general, a difference in the number of transpositions on the respective side circuits in excess of 20% may become objectionable. rllhis permissible percentage limitation depends on the accuracy to be acconiplished in practice in the formation of the transpositions as Well as on'the degree of immunity from interference required.
If m and n be comparatively large numbers Whose difference is small as compared to these numbers, as mentioned hereinabove, then the various cases of transpositions in uneven pitch, i. e., those in Which m is even and n uneven, or vice versa; those in Which m and n are both uneven, havin terminal transpositions Which are coincid nt Within the basic interval and which have imperfect interference elimination in practice; and those in which m and n are both uneven, hav- Thepossibility of a decrease in the ratio of m to" ing terminal transpositions which are relatively displaced, are practically not distinguishable. Where m, for example,equals 1000 anda is 989, the first transpositions Within the basic interval being coincident, a general limiting case of'an irregular transposition scheme is provided in practice, corresponding to an arrangement inwhich m equals 100 and n equals 99. Interference may be eliminated Within a basic interval including 100 short vtransposition sections or 99 long transposition sections, respectively, as mentioned hereinabove.
Taking into consideration the unavoidable irregularities of the sections in practice, if a case be assumed in rWhich fm, equals 999 and n equals 989, these numbers being relatively prime with respect to each other, and
`the terminal transpositions Within the basic interval being coincident, there Will be a disturbance Within the basic interval which Will not vanish. As may be assumed in all cases in which both m and n are uneven, lthe disturbance may amount toA -i- 1. In other Words, the disturbing length Within the basic interval Will amount to substantially l 1,000,000th of the length of the basic interval. In the case assumed, the denominator becomes the product of 989 andy 999v or 980,011. This arrangement, having uneven values for m and n and coincident terminal transpositions, is
fpractica'lly free from interference. However, by displacing the terminal transpositions by 1/2 With respect to each other, the disturbance may be completely eliminated. Accordingly, in the three types of transpositions in uneven pitch specifically enumerated hereinabove, in all of which there are the same relative positions of the transpositions on the respective side circuits Within the basic interval, the basic interval containing essentially 99 long sections and 100 short sections and coincidentV terminal transpositions, the
disturbance or. effective disturbing length may be completely eliminated.
The following table shows the characteristic values for the three types of transposition schemesmentioned hereinabove:
It may frequently be desirable to have the cable as free from interference as possible,
consist of an integral multiple of individually balanced basic intervals. The disturbing length in the latter case may, however,
vbe substantially'suppressed, if desired, by allotting to the greatest common measure of the transposition sections on the two side circuits whose absolute magnitude is arbitrary, a very small value. rIhus, it appears desirable to maintain the transposition sections on the side circuits comparatively small.
To increase the possible number of mutual positions of the conductors of the quad, the method ot accomplishing the transposition scheme may be changed to thereby increase its eiiectiveness. Accordingly, a single side cir` cuit may be transposed, iirst in one direction and then in the other. Thus, the tirst transposition may be accomplished by turningone y conductor about the other to the right through an angle ot 180 degrees, followed by a turn in the reverse direction through 180 degrees; then, by a transposition to the leftthrough 180 degrees; then by a turn in the reverse direction, and so on. Accordingly, one of the conductors of one side circuit andy then the other of its conductors are alternately located adjacent to the second circuit. Such an arrangement m'ay be brought to an even higher efficiency by preventing the coincidence of transposition sections exhibiting the same relative positions in both side circuits.
Whilethis invention has been shown and described in one particular embodiment merely for the purpose of illustration, it will be understood that the general principles of this invention may be embodied in other and widely varied organizations without departing from the spirit of the invention and the scope of the appended claims.
What we claim as new and desire to secure by Letters Patent of the United States, is,-
1. The method of constructing a cable quad which consists in twisting the four wires of the quad about a common axis, transposing one pair of wires so as to form m sections per unit of length, and transposing the other pair of wires so as to form fn, sections in the same unit o length, mand 'nI representing relatively different, prime odd integers.
2. The method of constructing a cable quad which consists in transposing one pair of wires of a quad representing one side circuit so as to form anodd number of sections within a predetermined interval, and transposing the other pair ofwires of the quad representing the other side circuit so as to form a different odd number of sections within the predetermined length, the number of sections in the respective side circuits being relatively prime numbers.
3. The method ot constructing a cable quad which consists in transposing one pair of wires representing one side circuit so as to form a definite number of sections within a predetermined interval, transposing the other pair of wires of the quad representing the other side circuit so as to form a diierent number ot sections within the same predetermined interval, the number of sections in one side circuit bearing a ratio to the number of sections in the other side circuit as do two odd prime integers, and displacing one pair of Wires with respect to the other pair of wires by a predetermined distance.
4. A cable quad to be employed for the transmission and reception of telephone messages, comprising four wires Wound about a common axis having point-like transpositions equally spaced in each pair of wires,thelength of the transposition sections in one pair of wires bearing the relation to the lengths of the transposition sections in the other pair of wires in accordance with two prime unequal integers, one of which is an odd integer diterent from unity, one pair of wires being displaced with respect to the other pair ot wires by a predetermined distance.
In witness whereof, we have hereunto set our hands this 25th day of October, 1927.
` ALFRED BYK.