US1776310A - Two-way negative-impedance repeater - Google Patents

Description

Sept. 23, 1930. I c ss o 1,776,310

TWO-WAY NEGATIVE IMPEEDANCE REPEATER Filed May 19, 1928 2 Shoots-Sheet l INVENTOR. 6'. Cm/ssol/z/ Sept. 23, 1930.

c. cmssou TWO-WAY NEGATIVE IMPEDANCE REPEATER 2 sheets-sheet 2 Filed May 19, 1928 ar 9 a /J INVEN TOR.

6. Cicalssan/ A TTORNE Y Fatented ept. 23, 1930 rran STATES PATENT OFFICE GEORGE CRISSON, OF EAST ORANGE, NEW JERSEY, ASSIGNOR TO AMERICAN TELEPHONE AN D TELEGRAPH COMPANY, A CORPORATION OF NEW YORK Application filed ma is, 1928. Serial No. 279,060.

This invention relates to repeater systems, and particularly to one characterized by the use of two negative impedances.

A positive impedance, as the term is commonly used, includes a resistance component which is positive and a reactance component which ma be positive or negative. A negative impe ance includes a'negative resistance and a reactancecomponent which may be negative or positive. The negative impedance corresponding to a given positive impedance is one in which each component is of equal magnitude but opposite sign to the corresponding component of the positive impedance.

When an electric current flows in a part of a circuit whose impedance includes a positive resistance, electrical ener is absorbed. It may be changedinto some 0t erform of energy .as heat, light or mechanical energy, or it may be transferred as electrical energy into some other circuit. When the resistance is negative, electrical energy is developed. A part of a circuit having a negative resistance must include some means, for example, a vacuum tube, by which the energy is introduced.

A form of negative resistance which is well known in theart consists of a one-way amplifier whose input and output circuits are coupled together. I have discovered that. there are two kinds of negative resistance whose properties diifer with respect to their tendency to sink when connected to an external circuit. The type of negative resistance resulting from the coupling of the input and output circuits depends upon the manner in which such connection is made. The ..-types and the manner in which they are created is more fully described hereinafter. I havefurther discovered that by employing dissimilar types of negative resistances with a hybrid coil network, it is possible to produce a' repeater having very desirable characteristics not attainable by either of the negative resistances alone.

My invention resides in the combination of negative resistances having dissimilar characteristics and so proportioned that the amplified energy will not be transmitted back 50 toward the transmitting station.

Other objects of my invention will be apparent from the following description, when read in connection with the attached drawing, of which Figure 1 shows schematically an impedance network which illustrates the description of the principle underlying my invention; Fig. 2 is a modification of Fig. 1 Figs. 3 and 3 illustrate the shunt type, and Figs. 4 and 4 the series type'of negative resistances by means of which important results are attainable; and Figs. 5 and 6, respectively show in detail the application of negative impedances of different types. as' repeaters in physical and phantom circuits.

For a complete understanding of my invention, it seems desirable to describe somewhat in detail the transmission principles upon which it is based, which are illustrated schematically in Figs. 1 and 2. Fig. 1 shows a hybrid set connected between sections of a line designated E and W, which sections have the same characteristic impedance Z The hybrid coil comprises the line windings in series between the sections E and V and also the winding connected with the impedance 0. One of the windings of the-transformer b is connected between the midpoints of the line windings of the hybrid coil and the other windings of that transformer is connected with the impedance (Z. The hybrid coil a and the transformer b are assumed to be designed so that complete balance and no reflection loss exist when equal impedances are connected with the four 'pairs of terminals. although the variable impendances c and d may be given other desired normal values by altering the ratios of the ,tr ansformers to suit the"'impedances employed. When each of theimpedances c and d has the normal VaIueZ there is an infinite transmission loss from either section of the line to the other, but at the same time the impedance presented by the hybrid set toward each line, assuming perfect transformers, is equal to Z values of the impedances c and (Z are varied reciprocally from their normal values by any ratio, that is, if one is multiplied by the factor N (areal number), and the other is di-' vided by. N, the transmission loss in either If the direction becomes finite, being given in TU by theformula I lines remains equal to Z It therefore follows that by giving the impedances 0 and (Z suitable negative values, the hybrid set becomes a repeater adapted to amplify trans-' -mission in' both directions between the sections E and W;

It is desired to make clear thata hybrid coil has been described as the means for connecting the negative impedances with the line because such connecting means constitutes the preferred type owing to the conjugate relationship established between apparatus connected with the terminals of the hybrid coil. It is to be understood, however, that the invention is not limited to an arrangement employing a hybrid coil inasmuch as other means for connecting the negative impedances with the line may be employed without departing from the claims defining the invention.

While certain practical advantages of mynegative impedance repeater will be. more fully described hereinafter, it is desirable fied power will be the full output of both tubes, which is four times or 6 TU more than the useful outputot the present 21- or 22-type repeater.

The manner in which the amplified energy is prevented from passing to the transmitter will be made clear by describing the arrangement shown in Fig. 2. With the switches in the positions shown, the gen.- erator g sends a current through the resisttimes 2 and 1",, representing the sections of the line in the direction shown by the solid arrows. By changing the switches 8 and a to the lower contacts, the negative resist ance r, is put in series between 7' and 73. This increases the current in the circuit by reducing the total resistance, the increase by the solid arrows.

being in the direction shown, for example,

The negative resistance 7",. is arranged to be connected by the switch 8 with the tap at the midpoint of the retardation coil f, which offers a high inductive impedance to currents .flowing from one part of the line to the other but only a low non-inductive resistance to currents flowing in' or out at the tap. The direction of the potential difference across 7", is indicated by the plus and minum signs. Due to its negative character, this element sends out current in the direction indicated by the dotted arrows. It will accordingly be seen that the current from 92, increases the current in the section. 1' and decreases the current in the section 91. By properly proportioning W and 1'4, as described above, their currents cancel each'othe'r in the sec- 'tion 7' and aid each other in the section 1 so that there is no transmission of the amplified current. back to thev section n. It

. will furthermore be seen that, due to the way in which 1, is connected to the circuit, neither of the two negative resistances receives current from the other as long as the iinped ances 1' and 7- are equal.

The. character of the negative impedancesr and 4",, which permits their operation without singing, will be apparent from considering Figs. 3, 3", 4: and 4. What has been designated the shunt type is illustrated by Figs. 3 and 3, and the series type by Figs. 4: and 41.

Referring to Fig. 3 representing the shunt type, if an E. E is impressed on the terminals of the amplifier A, a current flows in the input circuit, which causes a voltage ML to act in the output circuit. M is the unilateralymutual impedance, representing the ratio of the voltage generated in the output circuit to the input current, and I is taken as positive when the F. generated in the outpnt circuit aids the current n th s circuit. The current in the output circult is then The total current in the amplifier is:

-=2 Eiw I d 0 R; T R0 RgR from which I .Z R.+R +M U As M increases from zero, the resistance R desreases but it remains positive, never reaching zero.

By making appropriatechanges in the amplifier or the connections, such as by reverstor G.

making R, large or opening it, feeding back pecomes possible and the amplifier can oscilate. i The shunt negative resistance will therefore sing when it is smaller than the external positive resistance R, but will not sing when it is larger. In this respect its properties are the reverse of the series resistance which will now be fully described. Fig. 3 shows a' particular arrangement of a vacuum tube and associated apparatus which will exhibit in its Winding 1 a negative impedance having the shunt characteristics described above in connection with Fig. 3.

Referring to Fig. 4, representingthe series type, a one-way amplifier A is connected with the circuit that includes the generator G and the external resistance R This amplifier has its input and output terminals connected in series with each other and with the external circuit in the manner shown.

ing the two wires connected to the input or to the output terminals, M can be made negative. If now M is increased, the current I in the output circuit decreases and the resistance of the whole amplifier R increases. When 1l1'= li,-+R the current I becomes zero and the resistance R becomes infinite. At this point the amplifier just supplies'its own losses and the current in the resistance R, falls to zero. By increasing M a little more, R becomes negative but very large and the amplifier sends a small reversed current through the resistance R supplying energy to it. The potential impressed upon the input circuit of the amplifier by. the output circuit will be too small to sustain itself, however, without the assistance of the genera- Putting E for the E. M. F. generated within-theoutput circuit by the current I, in the input circuit,

E, M1, MR;

In order that the voltage E may be sustained without assistance from the genera- (4) of the device A is represented by R and the resistance between the output terminals by B When current flows in the input circuit,

an electromotive force is generated in the tor G output circuit. The unilateral mutual im- R R pedance. which is the ratio of thls electro- 8 L motive force to the current in the input cir- E= E0 1- (5) cuit, is designated M. M is, in this case, con- R sidered positive when thedirection of the R,+ R, electromotive force is such as to oppose the current, and negative when it aids the cur- Substltutmg for 0 1133 Value from q rent in the input circuit. If the generator G tion (4) sends a current I through the circuit, the po- R,R, tential difference across the whole amplifier R.+B E E (R,+R.,+M I 8 R1 e i V o Rel-R1 a and the resultant resistance offered by the Solving for B amplifier to the passage of the current is o E RC- (R;+RO M) R R I R: +Ro

that is, when the negative resistance is equal to the positive external resistance, as will be seen in the case of the series negative resistance, a disturbance becomes seltsustaining.

By increasing M still further, which reduces -R still more, the amplifier supplies more than enough energy to sustain a disturbance once started and the disturbance grows until overloadingreduces -M or losses in the system increase to such an extent that R,,=R. after. which the disturbance porsists with constant magnitude.

Physically it is readily seen that by makin g R, small, the connection between the output and inputterminals of the amplifier is short circuited so that feeding back cannot occur and the generator G must supply a voltage in order tomake the system active even though the amplifier may actually be supplying energ to the generator. By

lhe amplifier now supplies enough energy to the system to overcome its own internal losses but supplies no energy to the external resistance R The larger M is made, as, lor example, by

M: Ii, +1i,,, the voltage it across the whole amplifier and the resistance R become zero. I

By making i /l still larger, i l and it The resistance between the input terminals verse in sign and the amplifier supplies energy to the external'resistance R When the amplifier supplies all energy required for the circuit and if the generator G is removed, leaving B in the circuit the oscillation will continue indefinitely.

If the circuit is opened, and the amplifier is adjusted so that -M R ,+R,-+R and the circuit again closed, any disturbance, however small, which causes a current in the input of the amplifier, will result in a voltage in the output large enough and properly directed to increase the current. The disturbance will then grow until it becomes so large that M is reduced by overloading (or the resistance of some part of the circuit increases)- to such an extent that after which the disturbance will persist at a constant value.

Looking at the matter from a physical standpoint, if R, is made very large, the E.

'M. F. developed in the output circuit of the a vacuum tube and associated apparatus which will exhibit in its windings 1-71 a negative impedance having the series characteristics described above in connection with ra e.

When the series type negative resistance is substituted for m of Fig. 2 and the shunt type for 73, the two-element repeater thus constituted will function without singing. The gain of such a repeater may be adjusted by varying the ratio N. This may be done either by changing the value of the negative resistance or by changing the ratios of the transformers a and I) by means of taps. The latter is probably the preferable form and its use is shown in greater detail in Fig. G.

The application of the invention to physical and phantom circuits is shown schematically in Figs. 5 and 6, respectively. In these figures, L and L represent parts of one side circuit of a phantom group which are connected through the negative impedance repeater, and L and L. represent parts of the second side circuit. Connected with the side circuits are hybrid coils a and a having the transformers b and b bridged across the line windings thereof. Negative resistances c and 0' of the series type are connected to the windings of a and a, and negative resistances d and d of the shunt type are connected across windings of the transformers I) and b. The midpoints of bridged windings of the transformers b and b are connected with the winding of the transformer 13, with which is connected a negative resistance D which is of the shunt type and functions in connection with the phantom circuit. Furthermore, it will be seen that each conductor of each line circuit contains two other windings, those connected with L and L being designated e and those with L and L being designated 6 These windings are inductively related with the windings f and f with which is connected the negative resistance C, which is of the series type and functions in connection with the phantom circuit. When transmission takes place from L to L or from L to L, the negative impedances 0 and d co-act to send an amplified wave into the receiving line without returning a wave toward the speaker. The other negative impedances are not affected. Similarly, L to L or from one part of phantom to the other.

As will be seen in Fig. 6, the apparatus windings of the hybrid coils and of the phantom transformers are arranged with a plurality of taps in order to control the gain of the repeaters. One repeater, comprising the negative resistances A and A is connected with the line sections L and L another repeater, comprising the negative resistances A and A is similarly connected with the sections L and L a third repeater, comprising the negative resistances A and A is connected with the phantom circuit. The

gain adjustment of each of the repeaters is independent of that of the others, but the taps of both transformers of any repeater or the values of the corresponding negative resistances must be changed at the same time. The filters and the networks shown in connectionrwith the repeaters in Fig. 6 are intended to adapt the repeaters for loaded cable circuits.

The strpctures disclosed hereinbefore are capable of producingnegative resistances having the desired series or shunt characteristics. Since in practice the repeaters are used in line sections, the impedances of which vary with change of frequency, it is necessary that the negative impedance of the repeater should likewise vary. If the negative impedances of the repeater were made constant in value, the gain of the repeater would vary with frequency. Moreover, the effects of series and shunt impedances would not balance each other at all frequencies, and consequently, waves would return toward the talker at some frequencies. To avoid such ditficulty it is Thenetwork that is associated with each negative res stance for accomplishing the purpose yust stated comprises a condenser 4 in parallel with the inductance 5, which com- 5 bination is in series with the branch circuit effectively connecting the negative resistance A acrossthe line L,.- Bridged across the branch circuit is a condenser 6 and connected in series with the said branch circuit is an- 1: other combination of elements, namely, the resistance 7 and the'inductance 8, the purpose of all of which elements and combinations will be presently described. As mentioned before, it is necessary that each negative impedance shall vary with frequency in a manner depending upon the variations of the line impedance with frequency in order to avoid the detrimental effects mentioned. If the line with which the repeater is connected were distortionless so that its impedance would be a pure constant resistance, it would be unnecessary to use additional apparatus in order to make the simple negative resistances function properly throughout any range of frequencies. Such networks are necessary when the-negative resistances are used in connection with lines whose impedances are not pure constant resistances.

It has been found and shown by Hoyt in his Patent 1,243,066, dated October 16, 1917,

that the resistance component of the im pedance of a coil-loaded line, measured from.

a point .8 of a section away from the first loading coil, is approximately constant over a wide frequency range. The negative of this "resistance component may therefore be closely approximated by the negative resistances'described above. He also found that the reactance component of the line impedance can be neutralized by a network consisting of an inductance and condenser in parallel, which combination is connected in series with the line. Thereactance of this combination of inductance and capacity is therefore the negative of the reactance of the line and'by connecting the combination in series with the negative resistance the negative of the impedance of the line at .8 section v is obtained.

five resistance A to any desired end section of a loaded line, preferably a half section. This result is accomplished by connecting the condenser 6 across thebra'nch circuit that extends from the transformer b to the negative. resistance A This is equivalent to removing negative bridged capacity ftom the negative impedance and gives the negative of the line impedance for end sections less than .8 section. I

The combination made up of the eleinents 7 and 8 that forms part of the adjusting network is an additional refinement. The effect of conductor resistance of a line is to cause a rise in the resistance component of the line It is of course desirable toadapt the negaimpedance at low frequencies and to introduce a negative reactauce also at low frequencies. By introducing the elements 7 and 8 and increasing the value of the negative resistance to offset the positive resistance? at high frequenciesthe negative impedance of the combination is caused to increase at low frequencies and a positive reactance increasing as the frequency is lowered is introduced.

These effects make it possible to simulate the By omitting the condensers Owing to the fact, however, that the si mulation afforded by the network ust described becomes imperfect at frequencies near zero and near the cut-off frequency of the line, and that the return loss of the line tends to become low as the frequency rises, it is necessary to provide means such as filter 9 to modify the negative resistances such as A at the extreme frequencies to prevent singing and to permit satisfactory gains to he obtained in the useful range of voice frequencies.

While I have described only the network and the filter associated with the branch circuit containing the negative resistance A it is to be understood that the networks and filters shown in connection with the other branch circuits of Fig. 6 are similar to those just described. I

The negative impedance repeater disclosed in this application not only possesses the advantage found in the 21-type repeater, which consists in having the line conductors continuous through the repeater circuit but it also possesses an advantage which has hitherto been found only in the 22-type repeater, namely, that itdoes not send speech waves back toward 1 the transmitting station.

The result is that the conductors of the circuits, as shown, may be employed for telegraph cr other unidirectional or' low frequency alternating current-s without the use of composite sets to carry" such currents around the repeater. It is merely necessary to usethe condensers to prevent the shortcircuiting of these currents through the bridged circuits. Furthermore, the undesired reaction toward the transmitter which is found in the 21-type of repeater is avoided. This arrangement also avoids the necessity for using phantom repeating coils to separate the side and phantom circuits in order to insert the repeaters.

While the invention has been disclosed as ISO embodied in particular forms, it is capable of embodiment in other and differentforms 1. A repeater comprising two negative i111 pcdunws: one of the series type and the other of the shunt type, one of which impedances is effectively connected in series with the line and the other impedance effectively in shunt with the line.

The combination with a line of a repeater connected between sections thereof, the said repeater comprising two negative impedances, one of the series type and the other of the shunt type, one of which impedances is effectively connected in series with the line and the other impedance effectively in shunt with the line.

3. The combination with a line of a repeater connected between sections of the said line, the said repeater comprising two negative im pedances, one of the series type and the other of the shunt type, the said impedances-being so connected with the said sections that the current in one of said sections arising from one negative impedance opposes the current in that section arising from the other negative impedance.

4. The combination with a line of a repeater connected between sections of the said line,

' the said repeater comprising two negative impedances, one of the series type and the other of the shunt type, the said impedances being so connected with the said sections that the current in one of said sections arising from one negative impedance neutralizes the current in that section arising from the other negative impedance.

5. The combination with a hybrid coil between two sections of a line of a series type negative impedance serially coupled to the line through the hybrid coil and a shunt type negative impedance connected in parallel relatiou to the line at the bridge points of the hybrid coil, the vector magnitudes of said negative impedance being inreciprocal relation to the line impedance.

(3. The combination with a hybrid coil between two sections of a line, of a series type negative impedance serially coupled to the line through the hybrid coil, a shunt type negativeimpedance connected in parallel relation to the line at the bridge points of the hybrid coil, the vector magnitudes of said negative impedance being in reciprocal relation to the line impedance, and means for varying the magnitude of said impedances without destroying the reciprocal relation.

7. The combination with a line, of a series type negative impedance serially associated with the line and a shunt type negative impedan ce associated in parallel relation with the line.

8. The combination with a line, of a series type negative impedance serially associated therewith, and a shunt type negative impedance associated in parallel relation to the line, the said negative impedances being in conjugate relation to each other.

9. The combination with a phantom group comprising two side circuits and a phantom circuit, of series type negative impedances serially associated with each circuit, and shunt type negative impedances associated in parallel with each circuit, the series and shunt type negative resistances of each circuit being inconjugate relation to each other.

' 10. A repeater comprising a hybrid coil connected between two line sections, a negative impedance serially related to the line through the windings of the hybrid coil, the said negative impedance comprising reactance elements and a negative resistance obtained by connecting the input and outputterminals of a one-way amplifier efiectively in series with each other; a second negative impedance bridged across the line .at the midpoints of the hybrid. coil, saidsec- 0nd negative impedance comprising reactance elements and a negative resistance obtained by connecting the input and output terminals of a second one-way amplifier etfectively in parallel with each other, the magnitudes and phase angles of said negative impedances being in reciprocal relation to the line impedance at all important frequencies, and means for etfeetively warying the magnitudes of said negative impedances without destroying the said reciprocal relation.

11. The combination with a phantom group circuits and a phantom circuit, of three repeaters in accordance with claim 10 connected between the parts of said phantom group extending in opposite directions, each of said repeaters being adapted to amplify theoitaves traversing one of said circuits.

12. A repeater comprising negative impedances, one of the series type and the other of the shunt type, the said series type impedance being eflectively connected in series the shunt type, the said series type impedance being effectively connected in series with, the line and the shunt type impedance lpeing effectively connected in shunt with the me. i

In testimony whereof, I have signed m name to this specification this 17th day of May, 1928.

GEORGE CRT'SSON.

- of transmission circuits comprising two sides llfi

Images