US2002192A - Electrical circuit arrangement - Google Patents

Description

' y 1935- N. M. RUST 2,002,192

ELECTRICAL C I RCUIT ARRANGEMENT Filed Dec. 18, 1931 5 Sheets-Sheet l INVENTOR NOEL MEYER RUST ATTORNEX May 21, 1935. N. M. RUST 4 ELECTRICAL CI IIRCUIT ARRANGEMENT Filed Dec. 18, 1931 5 Sheets+Sheet 2 INVENTOR NOEL MEYER RUST BY Z ATTORNEY May 21, 1935.

N. M. RUST ELECTRICAL CIRCUIT ARRANGEMENT Filed Dec. 18, 1931 5 SheetsSheet 3 INVENTOR NOEL MEYER RUST ATTORNEY May 21, 1935. N. M. RUST ELECTRICAL CIRCUIT ARRANGEMENT Filed Dec. 18, 1951 5 Sheets-Sheet 4 INVENTOR NOEL MEYER RUST BY 7 Q ATTORNEY May 21, v Q M, RUST 2,002,192

ELECTRICAL C IRCUIT ARRANGEMENT Filed Dec. 1 l95l I I 5 Sheets-Sheet 5 II """H MMAMMM-- g v v v v v W c2 vww2-vvw li y d5 R914 II II II ll INVENTOR A'ITORNEY 50 cuit characteristics of the key circuits shown in cated in Figure 1, and it will be seen that when Patented May 21, 1935 r 2,002,192, 7

UNITED STATES PATENT OFFICE ELECTRICAL CIRCUIT GEMENT Noel Meyer Rust, Chelms ford, England, assignor to Radio Corporation of. America, a corporationof Delaware 1 Application December- 18,1931, Serial No. 581,930

, lnGreat Britain December 24, 1930 t 9 Claims. 7 (01. 17844) 1 This invention relates to electrical impedance the resistances in the circuits of Figures 1 to 5. networks and has for its object to provide a cir- Figures 9 and 10 are vector diagrams showing cuit arrangement of substantially constant imthe principal vectors for the complex network of pedance, which is adapted to be employed for v Figure 8. Figure 11 illustrates a simplified comfrequency correction purposes. plex network diagram of a circuit arrangement 5 It is, of course, well known that the impedsuch as is shown in Figure 8. Figure 12 illustrates ance offered by a parallel circuit consisting of in- .a further circuit scheme embodyingthe princiductance, resistance and capacity or by a comples of the present invention. Figure 13 is a bination of such circuits is over-all non-reacvector diagram illustrating the effect of varying l0 tive (i. e. of zero total-reactance though having the circuit characteristics of Figure 12. Figure reactance in the component elements) and is 14 illustrates a phase correction circuit arequivalent to that of the ohmic resistance if the rangement embodying the principles of this ininductance, capacity and resistance are so provention. portioned that the last quantity. is equal to the Referring to Figure 1 which shows one way of 15 square root of the quotient of thecapacity into carrying out the inventioma correction circuit 15' the inductance. For example, a circuit consistcomprises a network consisting of two parallel ing of two parallel branches, the one comprising branches, the'one comprising an inductance! and inductance and resistance in series, andthe other, aresistance r in series, and the other a capacity-c I capacity and resistance in series, offers for all and aresistance r in. series. Means are provided frequencies a substantially constant impedance for tapping off the, voltageset up across the re- 20 equal to that of, a pure resistance of lu equal sistance portions in the two. parallel branches. to the actual ohmic resistancejpresent in each T se two Vo ta s are pp ied, each to one of branch, provided that the said ohmic resistance the primaries pr or 112 of a doub p ry trans- (in each branch respectively) is numerically equal former t across whose seconda y 8 W ll be ento the square root of the quotient of the capacity erated the voltages required for correction pur- 25 into the inductance. E poses. One primary 121 is thus connected between The present invention envisages the utilization e junction Point Of the res stance embers in of a substantially over-all non-reactive (i. e. of th p l l a e ui s and app s p in at zero total reactance though having rea tan i the other end'of the resistance r in one branch of the component elements) constant impedance that circuit, the Other p ypa g-S rly 30 circuit of this kind to provide voltages suitable cenhected across the resistance in t e other for use for freq'uency'correction purposes. branch of t e eilcuit- Preferably a reversing Acc rding to thi invention a frequency. switch S is included in the circuit of one primary re ti n ir it, comprises a, b t ti ll 'so that the'sense of combination of the voltage stant impedance over-all non-reactive network, Set D in the tWO primaries y be sed W en 35 means fo tapping. 1 lt s t up a o desired. The resistance-inductance-capacity i t nc i th ind tive d min capacity networkis so proportioned that each resistance branchesof said network, and means for combinis equal to the Square feet of the q ot e t ing Said voltages t give gmt m; Voltage of the capacity into the inductance. If the priwhich may be employed' for correction'purposes. meries he asSocieted With the Secondary in e 40 p eferablymeans are provided fo changing t sense the correction obtained will be (relatively) will the sense of the combination of the tappedan amplitude correction (see the vectorial analofi voltages. ysis given below) while with the primaries asso- The invention is illustrated and explained in ciated inthe other sense the correction obtained connection with. the accompanyingdrawings will be (relatively). a phase correction. In the 45 wherein Figure 1 illustrates a particular embodidrawings I indicates voltage input terminalsand ment of this invention and Figures 2 to 5 are fur- 11 output terminals. ther modifications thereof. Figure 6 is a vector Preferably. the tapping points upon the diagram chart of the circle type illustrating cirbranches of the network are movable as indi- Figures 6A and 6B. Figure 7 illustrates agroup the reversing switch S is in the position giving of curves which may be used toassist in the deconstantamplitude and no phase change, if. the sign of a correction circuit for any particular case. tapping point in the capacity branch of the cir- Figure 8 is a complex network diagram wherein cuit be moved down the resistance r in that branched sub-networks are shown substituted for branch (i.ge., towards the junction point of. the

resistances in the two branches) the effect will be to produce a response characteristic dropping in amplitude as the frequency is raised. If, on the other hand, this tapping point be kept stationary, and the other tapping point be varied the reverse effect i. e. a rising characteristic, will be produced, the amount of tilt produced being governed by the position of the tapping pointson the resistances. Obviously, the frequency at which correction begins to be obtained depends on the relative values of the inductive reactance and of the capacity reactance to the resistance. From this it follows that if a correction is required on low frequencies (e. g. for low notes) large values of inductance and capacity are required, whilst if a correction operative on higher frequencies (e. g. corresponding to high notes) is'required, smaller values of inductance and capacity must bechosen,

the value of the square root of the quotient ofthev capacity into the inductance being, of course, in all cases equal to the resistance. This arrangement is thus suitable for what may be termed amplitude correction. With the reversing switch in the other position what may be termed phase corrections may be obtained; or 'a combination of phase and amplitudecorrections may be obtained. I

It is not always necessary to employ a transformer for combining the voltage tapped off from the two resistances. For example, in the case of phase correction the transformer may be dispensed with and the circuit of Figure 2 adopted,

combined voltages being, of course, set up across theterminalsII. Figure 3:shows a. further form of circuit which-may be adopted in the case of amplitude correction. The "impedance of the circuit receiving the correcting voltages and connected at II should of course be of relativelyihigh impedance compared to the resistances r r In place of using the simple forms'of constant impedance overall-non-reactive network above described, many more complicated forms of substantially constant impedance overal l-non-reacti ve "networks may be employed. For example, such a network maybe as shown in Figure '4 and may comprise two parallel branches, one consisting of a' first inductance Z in series with a second inductance and capacity c the last mentioned two reactances being each shun-ted by a resistance 1 or r and the other consisting of a capacity c 'a second-capacity c and an inductance Z allfin series, the last mentioned two reactances "being each shunted by a resistance r or'r In this case the movable tapping points giving the combined voltages for the transformer primaries can betapped uponthe resistances shunting the second inductance and the second capacity respectively (or upon the other resistances respec- 'ti-velyaccording=to the effect it is desired to obtain) In yet another arrangement of overall-non-reactive substantially constant impedance network shown in Figure '5 the network comprises two parallel circuits in series, the first parallel consisting of a capacity 0 shunted by a resistance r and a capacity c in series, and also shunted by an inductance l and a resistance r in series. If desired, any of the resistances'included in any of the illustrated arrangements may be replaced by a branched sub-network and any of the resistances in the branched sub-networks may themselves be-replaced by further branched subnetworks and so on to any desired degree of complexity, the substituted branched sub-networks being, of course, themselves networks in accordance with this invention. It is not necessary that all the component branched networks making up a composition network should be all designed to halve the same quadrantal frequency and it may be desirable in many cases to design the various networks to have difierent quadrantal frequencies. The substitution of such branched sub-networks for simple resistances in (relatively) main networks permits of what may be termed multiplication of correction effects and further facilitates the obtaining of desired impedance values to accord with any given input-output impedance requirements. This substitution of sub-networks for resistances in (relatively) ma-in networks is illustrated in Figure 8 in which I r 0 21 represent the parts of a main network as shown in Figure 3. In Figure 8 the resistances r and T1 are shown in thin lines to indicate that these may be replaced by sub-networks similar to the main network (though not necessarily of the same quad-rantalfrequency) these substitution subnetworks being shown in Figure 8 in broken lines.

Similarly theresistances in the sub-networks may be replaced by further sub-networks as indicated in chain lines while, in their turn, the resistances inthe sub-networks shown in chain lines may be replaced by sub-networks as shown in dotted lines and so on. The brokenline network which is 'a'sub-network relatively to the network Z r c n is a main network relatively to the chain linenetwork, which is in turn a main network relatively to the dotted line'network. The various numbered points in Figure 8 will be referred to in the vectorial description later.

Arrangements in accordance with this invention arereadily applicable to line circuits where it will be found that the values of inductance and capacity generally required are quite convenient. The invention is also suitable for use in connection with valve circuits. 7 An incidental; but important, advantage in connection with the invention, is that owing to the fact that a correction device in accordance there with has a substantially constant impedance, it can conveniently be incorporated at the receiving end of a line, and where so incorporated may be so arranged as to give rise to substantially no difiiculties due to reflection for, in many cases, the constant input impedance may be chosen to be equal to the surge impedance of the line at whose end the device is connected.

The combining transformer or other device (if any) employed to combine the voltages set up in the branches of the substantially non-reactive network, should of course be so arranged that substantially no load is thrown back into the network,'for, if any appreciable load be thrown back,

the condition for constant input impedance is "proper.

Figure 6 is, in essence, a vector diagram chart of the well known circle type drawn for the cases shown in the key figures and in which the quadcircle 0, L, Z, is the vectorial representation of the resistance R. Then it may be shown that 0L represents, in Figure 6A, the impedance of the inductive section i. e. the impedance of L and. R. in parallel. Also it may be shown that 00 represents the impedance of the capacitative section CR. The overall impedance is therefore the resultant of 0L and OC=vectorial sum If, however, the tapping. upon the resistance shunting the capacity in Figure 6A be moved, say to include only 60% of the voltage across the capacity the voltage between the terminals II will no longer be given by the vector OZ but will obviously equalthe vectorial sum OL+Lz=Oz. This of course is for the case in which the voltages due to the inductance and the capacity are combined inlikesense. If, however, these voltages are combined in opposite sense i. e. that acrossthe condenser is reversed before combination as by employing a combining transformer as shown in Figure 1 with primaries in opposition the capacity vector LZ becomes LZ' La becomes L2 and OZ and Oz become OZ'and 02 respectively. Similar considerations apply to Figure 6B (the vectors for Figure GB at a frequency of 700 cycles are shown in broken lines) except of course that for Figure 6B the vectors within the circle OLZC apply to 'combinationof voltages in opposed sense and the vectors without the circle apply to voltages combined in the same sense. For the sake of varying the example, L121 and Liz/1 have been taken as 40% of L1Z and L1Z1 respectively. .In the chart the diametrical abscissa are percentage values of R and the peripheral ordinates values of frequency from 0 to The chart proper consists of the circles and the lines radiating from Z to the periphery of the largest circle. The actual vectors shown in Figure 6 relate, of course, to the particularcases just described. V

In Figure 7 are shown a number of curves illustrating the effect of various adjustments of the tapping point upon R in terms decibels of attenuation (ordinates) plotted against frequency, the

horizontal scale being logarithmic; The curves would have to equal to 10,000X21r. there is also the required impedance condition veniently employed to assist in the design of a correction circuit for any particular case as follows:,

The attenuation curve of the circuit to be corrected is plotted out on: the same logarithmic paperas the curves of Figure '7 and to the same scale in decibels attenuation but with attenuation values rising upwards from a zero line at the bottom of the paper instead of downwards from a zero at the top. The logarithmic paper employed is transparent, and the curve to be corrected for is moved over the curves of Figure 7 until it is superimposed upon that curve of Fig ure 7 which is found to be nearest to the curve to be corrected'for. The percentage correction is noted, and from the relative position of the frequency lines on thetwo sets of curves the required quadrantal frequency is immediately determinable. For'example, if it were found that the attenuation curve of Figure 7 nearest to the curve forwhich correction was required was the Z L C As however that 1 L and C are directly determinable.

Figures 9 and 10 are vector diagrams of the same basic type as that of Figure 6 but showing only the principal vectors for the complex network of Figure 8. The diagrams of Figures 9 and 10 are constructedon the same principle as that of Figure 6 and are obvious developments thereof. ;'I'he number references in Figures 9 and 10 indicate the vectors representing the voltages set up between the correspondingly numbered points in Figure 8. Figure 9 is a vector diagram drawn for the quadrantal frequency while Figure 10 is drawn for a frequency lower than the quadrantal frequency and such that the voltage across any particular inductance is one-half of that of the network (or sub-network as the case may be) of which it forms part. For the sake of simplicity Figures 9 and 10 have been drawn on the assumption that the quadrantal frequency for each network and sub-network is theisame though of course this is not a necessary condition. Also for the sake of simplicity in identifying the various vectors, parts of Figures 9 and 10 are drawn.

infull, broken, chain, or dotted lines according to whether they relate to networks shown in full, broken, chain or dotted lines in Figure 8.

In the case for the quadrantal frequency (Figure 9) the vector vector 02 vector 0-4 vector 08 vector 0 16 pointuponthe resistancerR in shuntacross the condenser C for the case of Figure 6A or to settings of the tapping point upon the resistance R. in series with the inductance L of Figure 6B. A set of curves as shown in Figure 7 may be con- 2 4 8 the phase shift for each step being The chart of Figure? may be utilized asfollows for the case of a complex network'giving a number-of stagesof correction.

The curve rising from the left handside of Figure '7 indicates the output voltage amplitude across the vector 0-8 as related to the input voltage vector 0-16. v

Now as the decibels attenuation scale is logarithmic the voltage-across vector 0-4 is the summation curve obtained by adding the 0% curve to itself, and as the same quadrantal frequency is used throughout this is equivalent to reading the 0% curve as though the attenuation scale had been doubled. Similarly the voltage across vector 0-2 is obtainable by adding to the summation curve the original 0% curve, or in other words by reading the 0% curve as though 7 work the curves have to be added, since the one curve cannot be" read to difierent scales.

[The same principle may be applied to obtain the phase angle and it is possible in the manner described to obtain the voltage across any two points from the curves.

For any specific requirements inspection will usually show what degree of complexity i. e. how many stages of correction are necessary, and it will then commonly be found possible to simplify the circuit and still meet the specific requirements.

For example it might be found that the correction curve best suiting conditions might be between 3 times and l times the 0% curve of Figure l 7. In such a case the circuit of Figure 8 could be simplified to that shown in'Figure 11. V

In Figure 11 the output is tapped off between 0 and A which is an adjustable tapping point on the resistance 1-2. The curve for any tapping position is estimated in this case by reading with a trebled attenuation scale the 0% curve of Figure '7 (above referred to) (of course for the correct quadrantal frequency) and adding to it the curve expressing the voltage relation between 0A and 02 from the group of curves.

Of course as above stated, where different quadrantal frequencies are used at each stage of correction i. e. in the different networks and subnetworks the curves for the corresponding quadrantal frequencies respectively must be added.

In any network or sub-network the inductance and condenser may be interchanged and it is also possible to combine networks and sub-networks in such a way as to multiply or add'in successive stages, one of the other percentage curves (20%, 4.0%, 60% and so forth).

Generally speaking it will be found that although the phase is rotated, in correcting for amplitude, the overall effect is such as to tend to correct the phase, and an overall delay effect is produced.

In designing correction circuits in accordance with this invention it is also possible to utilize the phenomena of resonance by designing one or more of the component networks of a circuit to be resonant within the range over which frequency correction is required. A' correction circuit so designed maybe advantageously employed in many cases where it is desired'to correct for a frequency characteristic showing a change occurring within a relatively narrow range of frequenciesz-for example, it might be required to correct a transmitter whose frequency characteristic showed a drop of 4 or 5 decibels between 6,000 and 10,000 cycles per second. An ordinary circuitcorrector of the simple resonant circuit type may, of course, be used for applying such a correction, but the present invention may also be adapted to give such a correction and offers the practical advantage that a circuit in accordance with the said invention is more readily calculable in its results and more flexible in its application than are simple resonant circuits.

Consider a circuit as shown in Figure 12 and consisting of a series branch L1C1R (inductance and capacity in series) and a parallel branch L2C2R2 (inductance and capacity in parallel) in series with one another. At frequencies below resonance the series branch behaves as a condenser ofiering at any frequency lower than the resonant frequency am an impedance given by the expression jww c while the parallel branch ductance of impedance Similarly at frequencies 601 above resonance the series branch behaves as an inductance of impedance while the parallel branch behaves as a capacity impedance behaves as an in- It follows therefore that the whole circuit will be overall non-reactive at frequencies below resonance if and-it will be overall non-reactive at frequencies above resonance if L -12. Therefore for the circuit to be overall non-reactive above and below resonance L1C1 should be made equal to L202 i. e. both branches should be made resonant to the same frequency; i. e. we should be of the same value for both branches.

Obviously there will be a different quadrantal frequency for the two cases of below resonance and above resonance the quadrantal frequency for the former case being found from the relation (w being an equivalent quadrantal velocity below resonance) while for the latter case the expression is employed. wl q is the corresponding equivalent quadrantal velocity above resonance. It will be clear that in the former case the equivaent angular velocity a (which is given by the expression cow 2 2 becomesinfinite when w0=w (at resonance) Whilst in the latter case when the actual velocity w1=wo range over which a desired correction occurs can (at resonance) the equivalent velocity, mkcorre sponding to ml and given by the expression becomes zero. ,The effect may. be expressedin the familiar vector diagram form as shownrin Figure 13 the numbered points upon whichcorrespond to the numbered points in Figure 12 inthe same .way as the numbers: in Figures 9,and IOcorrespond to the numbers in Figure 8. The circle in Figure 13'is the locus of the vector-potential 0-1. At w=0 the series branch offers in: finite impedance andpthe parallel branch zero impedance hence the vector potential 0- l.=vector potential 0-2. In Figure 13 the vectors aredrawn for the two quadrantal frequencies indicated at an; and mi; which, are the actual frequencies corresponding ;to the equivalent quadrant-a1 frequencies w and wi respectively,

,It will now be apparent thatthefrequency be fixed (l) by iixing the resonant frequency to determine the frequencies at'whichzero or maximum correction occurs and (2) by fixing the relative values of the products L201 and L102.

The chart of Figure 7; can be employed for the estimation of correction curves by computing the values of oh; and cql q and from this knowledge the attenuation for any correction tapping can be found in terms of w and 01 A curve can then be plotted showing the relation between these frequencies and the actual applied frequencies to (below resonance) and m (above resonance). For example, suppose it is required to impart a lift up of 6 decibels between 5,000 and 10,000

cycles, it being unimportant if an extra loss is effected above 10,000 cycles. Referring to the chart of Figure 7 it will be found that a tapping somewhere about 50% along the condenser branch is required i. e. in Figure 12 about half way along the resistance R in shunt across L1 and C1 the other tapping point being of course the point 2. Now the lift up will cease when w or w wo and therefore the resonant frequency we should be chosen at some frequency above 10,000, say 12,000. L201 would be chosen so that the correction started to come into operation at 5,000 cycles and to secure this result the first quadrantal frequency (17:; could be chosen at about 8,000 cycles the exact values being determined actually by fitting the curves. The a curve wocld not be used though it will be noted that the overall attenuation above 10,000 cycles will be actually increased. Under the conditions imposedhowever, this is allowable and in certain conditions might prove a definite advantage.

It is clear that the correction effects can be multiplied in exactly the same way as with the ordinary non-resonant branch circuits already described (see Figure 12) and that different resonant and quadrantal frequencies may be used at each step of correction if required.

it is also clear that a parallel branch type of circuit as shown in Figure 14 may be employed for phase correction and that by means of such I a circuit a 360 phase shift can be produced be- '1. In combination, a parallel circuit having two branches, one ofsaid branches comprising a resistance and inductance and the other of said branches comprising a capacity and resistance, the resistances in both said branches being equal and also equal to I where L is theindfuctance and C the capacity of the circuit, an input circuit across said parallel circuit, having a source of electrical energy containing a band of frequencies connected thereto, means for tapping off resultant voltages set up across both said resistances at points which give an output amplitude which is a function of the input frequency, and an output circuit connected to said means. p i

I 2. A circuit for correcting for frequency distortion comprising a substantially constant imped ance overall non-reactive network having capacitance, resistanceand inductance elements so related as togive a constant pure resistance impedance between a pair of terminals to which input voltages of different frequencies are applied, means for tapping off Voltages set up across the resistance elements in said network, and means forfco nbining said voltages to give an output voltage which-is distorted in inverse senseto th input voltage.

3. A circuit for correcting for frequency distortion comprising a substantially constant impedance overall non-reactive network having capacitance, resistance and inductance elements so related as togive a constant pure resistance impedance between a pair of terminals to which input voltages are applied, means for tapping off voltages set up across the resistances of said network, and means for combining said voltages to give an output voltage for amplitude correction purposes including a switching element for changing at will the sense of combination of the tapped off voltages.

4. A circuit for correcting for frequency distortion comprising a substantially constant impedance overall non-reactive network to which input voltages are applied, said network comprising two series sections, one consisting of an inductance shunted by a resistance and the other of a capacity shunted by a resistance, each resistance being equal to the square root of the quotient of the capacity into the inductance, and

means for tapping off the resultant voltages set up across the resistances substantially as described.

5. A circuit for correcting for frequency distortion comprising a substantially constant impedance overall non-reactive'network having capacitance, resistance and inductance elements so related as to give a constant pure resistance impedance between a pair of terminals to which input voltages are applied, means for tapping off voltages set up across the resistance elements in said network at points which give an output amplitude which isa function of the input frequency, and means for combining said voltages to give an output voltage for correction purposes in such fashion as to obtain. an output which varies with frequency, including. a transformer having two primary windings: one primary winding being in circuit with said inductance element and the other primary winding being in circuit with said capacitance element.

6. A circuit for correcting for frequency distortion comprising a substantially constant imment.

pedance overall non-reactive network to which input voltages are applied, said network comprising two parallel sections, one comprising an induc tance and a resistance in series and the other a capacity and a resistance in series, each resistance being equal to the square root of the quotient of the capacity into the inductance, and a double primary transformer connected to said sections for combining the voltages derived therefrom, one primary winding being across the resistance in the capacitance section and the other primary winding being across the resistance in the inductance section of said parallel arrange- 7. A circuit for correcting for frequency distortion comprising a substantially constant impedance overall non-reactive network to which input voltages are applied, said network comprising two parallel sections-one comprising an inductance and a resistance in series and the other a' capacity and resistance in series,each resistance being equal to the square root of the quotient of the capacity into the inductancdand a double primary transformer connected'to said sections for combining the voltages derived therefrom, one primary winding being in circuit with the capacitance section and the other primary tion of said parallel arrangement, including a reversing switch connected to one primary winding whereby the sense of combination may be reversed at willi 8. An equalizer circuit comprising a substantially constant impedance overall non-reactive network to which input voltages of different frequencies are applied, said network comprising a series branch including inductance and capacity in series, and a parallel branch including inductance and capacity in parallel, said branches being designed to be resonant at a frequency within a predetermined. range of applied frequencies, and means for combining the resultant voltages set up in said branches.

9. An equalizer circuit comprising a substantially constant impedance overall non-reactive network of two paths having capacitance, resistance, and inductance elements so related as to give a constant pure resistance impedance between a pair of terminals to which input voltages of different frequencies are applied, one of said paths including inductance and resistance, and the other of said paths including capacitance and resistance, the resistances in both of said paths being equal, and means for tapping oil" voltages set up across the resistances at points which are unsymmetrical with respect to the reactance elements.

NOEL MEYER RUST

Images