US1600283A - Artificial line - Google Patents

Description

Sept. 21 1926.

H. w. HlTcHcocK ARTIFICIAL LINE Filed Oct. 20, 1921 2 Sheets-Sheet 1 INVENTOR I EWfiQ/afiwa ATTORN EY Patented Sept. 21, 1926.

UNITED STATES may w. nrrcncocx, or NEW YORK, N.'Y., ASSIGNOR To AMERICAN TELEPHONE PATENT. OFFICE.

TELEGRAPH COHPAJTY, A CORPORATION OF NEW YORK.

. AB'I'IrIcIAI. LINE.

The princi a1 object of my invention is to provide an e ectrical network which may made closely to simulate a given geographically extended signal transmitting line with respect to impedance, attenuation and phase type. Other objects of my invention relate to making the sections of properly graded length and electrical characteristics and providing for adjustment of the several sections to take care of local irregularities in the geographical line. All these objects of my invention and others will be appreciated upon consideration of a specific embodiment of its rinciple which. I now'proceed to describe in the following specification. It will be understood that the scope of my invention is defined in the appended claims.

Referring to the drawings, Figure 1 is a general diagram, Fig. 2 is a more detailed diagram of a. single section of the artificial line, Fig. 3 is a diagram for resolving the currents into their components in this section, Fig. 4 is a diagram showing the variation of impedance with frequency, and Fig.- 5 is a diagram showing how the section lengths and characteristics are graded along the artificial line.

The special example here presented to illustrate the principle of my invention is an artificial line designed and constructed to balance a'certain continuously loaded ocean cable of length about 100 nauts (nautical miles) and adapted for simultaneous superposed ordinary and carrier current telegraph and telephone operation, in which all current. frequencies up to 6,000 cycles per second Application 'flled October 20, 1921. Serial No.'508,945.

dicated at 14. This cable 11 is not of uniform physical character throughout its length, but its shore ends are specially armored to guard against shallow water hazards. Thus it is that the cable consists of a plurality of sections graded slightly in their characteristics from the middle toward the ends. Moreover, in the manufacture of such a cable, it is commonly made up in sections, each about two nauts long, which are spliced together, and it is impracticable in the manufacture of these sections to make them iden- .tical in their electrical characteristics. It is among the objects of my invention to provide facility for adjustment of the artificial line to match it to all. such irregularities. The shore end of the cable 11 with the terminal conductors 12 and 16 are in the cable but .17, where there is located a transformer 13 with a condenser 14 in the coil connected 1 to the conductors 12 and 16. The other coil of the transformer 13 is connected to the conductor pair 19. Branch conductors 18 for direct current telegraphy, with an interposed retardation coil 15 are connected to the i conductors 12 and 16 between. the cable end l rent telegraph channel are connected to the transformer 24, as shown. The frequency range for this telegraph channel lies above the normal essential voice frequency range. These input and output branches 22, 23, 25 and 26 have the usual suitable filters to separate the frequencies. Other frequency ranges may be utilized by other hybrid coil transformers in'parallel with 21 and 24, with suitable filters in their input and outputconnections. The parallel branches of the conductor pair 19, throughthe respective transformers 21 and 24, are united and lead through the network 27, which balances'tho part of the conductors 19 in the cable 20;

33 balances the part of the conductors 18 within the cable 20. A channel for ordinary duplex direct current telegraphy is connect- I ed to the terminals 31.

It will be seen that the apparatus at the cable hut 17 is balanced by the transformer 28, and the terminals 39 and 40 correspond respectively to 12 and 16. Connected to these terminals 39 and 40 is my, improved artificial line 43'which balances the cable 11.

This artificial line 43 consists of a large number of sections 35 connected in serial order. These sections are alike in general plan but differ in equivalent length and in ohmic resistance as will be pointed out presently. Each section 35'between the two terminals 39 and 40 on one end and 41 and 42 on the other end consists of a resistance 36 and inductance 37 in series on one side,

a like series combination of resistance 36' and inductance 37 on the other side, and a pair of equal cross-connected condensers 38.

The resistances 36' and inductances 37 are adjustable as indicated in Fig. 2.

When the resistances 36,. inductances 37 and the capacities 38 of the section of Fig.' 2 are given properv values, determined by.

the series resistance and inductance of the geographicahline and its distributed shunt capacity, then the artificial line section is practically equivalent to a length of the geographical line. These electrical characteristics are chosen to make the equivalent lengths of the sections comparatively short at the end near the ofiice and not so short at the remote end The exact dimensions in thisillustra-tive example of my invention are indicated in Figs. 5, which shows 20 halfnaut sections next to the office end, then 20 one-naut sections, followed by longer sections until the remote section with an equivalent length of twelve nauts is reached.

Assuming for the conductor pair of the geographical line that the ohmic resistance per nautis R ohms, inductance per naut is L henrys, and distributed shunt capacity per naut is C farads, if we make the network of Fig. 250 that each resistance 36 .is

-38 is 30, where is any number, then this network interposed in the geographical line is equivalent in impedance to a length of n v The definition of characteristic impedance assumes an infinite line, and accordingly the impedance across 39, 40 is the same as across 41, 42, as indicated by the'chal'acter Z in Fig. 2. Let the impedance of branch 39- 41 be and let the impedance of branch 3942 be i where Z and Z are respectively the series and shunt impedances per unit length of the smoothdine. Solving the network of Fig. 2 by the ordinary \Vheatstone bridge formula, the result is obtained that The last expression is the familiar formula for the characteristic impedance of the smooth geographical line whose constants are R,L and C. Thus it will be seen that the characteristic impedance is the same whether the sections of Fig. 2 are great or small in equivalent length. The advantages of making them small will be pointed out presently, and itwill be shown that they shouldbe shorter near the office end. 7 As to the propagation constant, we know that for the smooth line, it is given by where P is defined by the equation 6P:I1/I2 and where 1 /1 is the ratio of the currents at the input and output ends of a unit length of the smooth line.

Let the input and output currents for the section of Fig. 2 be .I, and 1,. By the symmetry of the structure. it. is seen that the currents through the branches 3942 and 41-40 are equal. Call these I. Resolving the currents in the network of Fig. 2 as shown in Fig. 3, we have by Kirchoifs law,

I1IZ/ IIIZ2I+(I!I+I2')Z1I and.

Eliminating IQ", substituting I 'i for Z" and simplifying, we get 1 E 1/ 2' 1/ T By analogy to the definition of P, let the foregoing expression equal e By definitIOII' of the hyperbolic tangent, tanh Substituting and reducing,we get By a simple transformation we get I I I When 5,- is small, tanh and are nearly equal. Hence we see that by making n small we make P and n .to be nearly equal, and thus we make the propagation constant of the section of Fig. 2 to be nearly the same as 12 units of length of the smooth line. Accordingly the size of the sections is determined largely by the consideration that they should be made'small enough to secure asufiiciently approximate equality of the propagation constant per; section with the corresponding length of the smooth line. The specific application of this principle in the present case' is exhibited in Fig. 5 where the gradation of section length is shown by appropriate legends on the drawing.

If the geographical line were perfectly smooth, the resistance and negative reactance components of its characteristic im pedance would vary with frequency some what according to the smooth curves of Fig.

4. Due to the inevitable slightirregulari-.

ties in the geographical line that have already been mentioned, a test of thecharacteristic impedance for varying frequencies gives a plot that may depart from the smooth curves, for example, somewhat as indicated by the dotted curves of Fig. 4. These departures are due to partial reflections of current waves from the junction points of sections of differing characteristics. They may be analyzed by graphical or analytical methods, and the actual distance of the irregularities from the .cable but may be determined in this way, and thereupon adjustments may be made in the proper artificial line sections to matclrthese irregularities. These adjustments will be made when the cable is started in operation and ,will be made once for all, because the irregularities in the geographical line are permanent, and therefore the corresponding adjustments of the artificial line are equally permanent.

For high frequencies, as already stated,

the sections should be small, in order to make the propagation constant nearly equal to that for the geographical line. But the high frequency currents are most rap dly attenuated, and for this reason the sections may be made longer at the distant end of the artificial line. Also because. the current is far less attenuated in theiiear sections than in the remote sections, much finer adjustment may be. necessary in the near sec tions.

current is so highly attenuated that, even if a considerable part of it were reflected back to the oflice, its effect there would scarcely be noticed. This is true for higher frequencies, say above 300 cycles per second, though lower frequencies would not be so much at-' 'tenuated and might be reflected back and At the remote end of the artificial line the 1 noticed atthe otlice- The artificial line in terminated at its remote end by a network 44,consisting of a series resistance 45 shunt edby a condenser. 46 and series resistance 47. The impedance of this combination approximates sufliciently to the impedance of the distant terminal apparatus when look ing into it from along the actual line.

In the discussion hitherto it has been tacitly assumed that the series-resistance R of the geographical line is constant, but in fact it is a'variable function of frequency.

For low frequencies the return current spreads out in the sea water, whereas for high frequencies it is largely confined to the cable sheath. thus increasing the circuit re- 'sistance.

Moreover the inductive loading involves more or less eddy current and hysteresis loss, which are functions of frequency and increase with frequency and therefore increase the apparent resistance to the current flow in the signaling circuit. It is impractical to construct'the artificial line to make itsresistance vary correspondingly. It is necessary to make the direct current resistances of the artificial line equal the cable resistance 'at about 1000 cycles per second. This is too high for direct current and therefore the resistance is reduced for the sections farther out, so that the total direct current resistance of the cable has the proper value. i a

In the foregoing specification, I have disclosed a network with sections of different equivalent length; by this term I mean the length of the extended lineto which the network section is approximately equivalent in its transmission properties. A. network of short equivalent length is one which is equivalentfto a short length of the extended line. y

I iclaim t 1. An artificial line consisting of sections in serial order, each intermediate section being connected with the adjacent section on one side at two points and with the adjacent section at the other side at two points, impedances connecting each point on one side with each point on the other side, one

-pair of non-adjacent impedances consisting each of aresistance and an inductance in series and the remaining pair of impedances consisting each of interposed condensers.

2. An artificial line of sections in serial order, each se' tion being a bridge type network, the sections near the input end being made small enough in equivalent length so that the propagation constant per sectlon shall be approximately the same as its limiting value when the size of the sections is made indefinitely small in equivalent length.

3. An artificial line of sections in serial order, each section being a bridge type network with the impedancevalues of its arms so related to-the characteristirs of a given smooth line that the sections simulate definite lengths of that line for characteristic impedance. the. sections being'made small enough in equivalent length so that the propagation constant per section shall be approximately the same as for the corresponding length of the smooth line.

4. An artificial ,line of sections in serial order, each section being a bridge type network. the sections being small in equivalent length and each section being matched closelv to a defin te finite length of a given smooth line. 5. An artificial line of sections in serial order, each section be nga bridge type network. the sections being small in equivalent length at the input end and in reasing in size away therefrom, and each section being closelv matched to a definite correspond .are alike, one pair of like lmpedances each consisting of a resistance and an lnductance ing finite length of a given smooth line.

6. -\n artificial line of sections in ser al order. each section being a bridge type networkbetween two pairs of terminals, each pair of terminals on one side being conneeted to a pair of the next succeedlng section on that side, and having a pair of its arms of adjustable impedance whereby they can be matched to local irregularities in an actual smooth line simulated by the artifici l line.

conductor, and an artificial line to balance it, said line consisting of bridge type net- 7. in combination, a submarine signaling 9; An artificial line of sections in serial order, each section being a. bridge type network liaving the series impedance of each arm of one pair of arms equal to half the series impedance of a definite length of a given smooth line and having the admittance of each arm of the other pair of arms equal to half the shunt admittance of the same length of the given smooth line, the equivalent lengths of said sections being graded in decreasingmagnitude toward the inputend.

10. In combination. a loaded signaling condu tor and an artificial line to balance it, said line-consisting of sections in serial order, the sections near the input end being made short enough in their equivalent length so that the propagation constant per section shall be approximately the same as its limit ing value when the size of the sections is made indefinitely short in equivalent length;

11. An artificial line of sections in serial ordexnrach intermediate section being conne ted at two points with the adjacent section on one side and at two other points with the adjacent section on the other side, the two points' on one side being connected with the two points on the other side by four impedances of which two non-adjacent impedances are alike and the remaining two in series and the reinaining pair each cons sting of capacities.

said lire consisting of sections-in serial order te made-short enough in their equivalent length so that the propagation constant per section shall be approximately the same as its limitingyahie when the size of the sections is made indefinitely short in equivalent length, and the sections farther away from the input end being made longer in equiva-- lent length. i

V In testimony whereof, I have signed my d name to this specification this 18th day of October, 1921.

HARRYWV. HITCHCOCK.

sections near the input end being

Images