DE589521C - Process for the production of active multi-pole circuits with a prescribed frequency dependence of their complex constants - Google Patents

Description

GERMAN EMPIRE

ISSUED ON JANUARY 15, 1934

REICH PATENT OFFICE

PATENT LETTERING

CLASS-21a ² GROUP 41 oi

General Electricity Society in Berlin *)

Patented in the German Empire on September 20, 1928

For the production of double acting amplifier circuits in transmission lines So far, two different paths have been taken. One has a common one for both directions of transmission or an amplifier each is provided with the help of hybrid circuits and, if necessary Line replicas have been introduced into the line, or attempts have been made to do so to switch on the energy supplying tubes in the form of negative resistances in the lines. Fig. Ι shows an amplifier of the former that transmits to both sides Art, Fig. 2 one of the latter

In practice, the first method has so far been used almost exclusively because it It was very difficult to get the negative resistances into the lines free of reflections turn on, and such reflections in the case of negative resistances with self-excitation are connected.

The invention now relates to a method which allows active multiple poles with the help build up of negative resistances, which are reflection-free in transmission lines with a given wave resistance Let it switch on and take over the amplification of the transmitted currents.

In the most common practical cases, due to the attenuation distortion of the transmission line and the frequency dependence of the characteristic impedance, a very specific frequency dependency of the corresponding values b and Q of the active quadrupole is required. The production method of such active four-terminal network according to the invention now uses methods already described earlier for extension lines. The difference between the known extension lines and the active quadrupole lines considered here consists only in the reversal of the sign of the attenuation, so that the active quadrupole lines according to the invention can accordingly be referred to as shortening lines. In the following it will be shown that and Avie this transfer of the method used for the extension lines to active four-pole connections is possible if, in addition to the switching elements considered for the extension lines (ohmic

*) The patent seeker indicated the following as the inventors:

Dr. Hans Busch in Darmstadt and Dr. Alfred Byk in Berlin,

Resistances, inductances, mutual inductances and capacitances) real resistances negative sign used, for example can be produced with the help of electron tubes with a suitable characteristic.

The method according to the invention has the advantage over the known method that only the real component has to have a negative sign, while the ίο imaginary component can remain positive. The proposals that have become known so far that is to say, they went to switching four-pole lines into lines, to which one such The degree of transmission has been given that it equals the degree of reproduction of the negative Lines is. This' leads to the need for negative impedances including Inductivities and capacitances, that is, to extremely complicated circuit structures. According to the invention, however, an extension line with the desired Characteristic impedance values and positive damping wanted. The conversion into a negative resistance then takes place by replacing only the positive real part of the resistance with a negative one real part. In the invention there is therefore no conversion of the positive impedances including the imaginary parts in negative impedances required, rather it is sufficient to merely add the complex conjugate impedances with negative To pass sign.

In the following two sections I and II, under I, a number of examples of four-pole circuits with negative resistances are given, which have a frequency-independent negative attenuation (gain) and whose characteristic impedance can be adapted to the impedance of the lines without this the gain is impaired, while in Section II the solutions to the general problem are given to calculate the required switching elements of the active quadrupole for prescribed values of § and b. These rules apply to both Γ, LT and cross connections, of which the T and / Z connections are shown in principle in Figs. 3 and 4.

As you know, the linear equations are

between voltage and current at the beginning and at the end of a symmetrical four-terminal network »1 = Söffe) -s _s + S · ©« (?) - 3 », W

where g - b + α i is the generally complex rate of propagation and § is the wave resistance of the four-pole, which is also generally complex.

The voltage equations for the T circuit now depend on the values R ₁ and R _{% of} the here negative series and transverse resistance as follows:

S ₁ =

2Ä,

R,

(4)

By comparing the corresponding coefficients of (3) and (4) with those of Equations (1) and (2) give the relationships

(5)

(6)

If one sets ag = b in accordance with the disappearance of the angular measure, then with real 3 for arbitrarily prescribed § and b one has purely real resistances R ₁ and R ₂ , which are negative with negative b, as required by the amplification.

For the il circuit of Fig. 4, the are Voltage equations

(7)

which gives by comparison with (1) and (2)

(9) ^(IO)

Here, too, with vanishing angular dimensions and purely real g and negative b, purely real negative values of Rl and Rt, result.

According to known formulas that allow the conversion of T and 17 circuits into cross circuits, one also obtains certain required negative, real resistances in the series and cross branches of a cross circuit for any real Q, negative b and vanishing α.

To save on negative resistances, it is sometimes advisable to use circuits in which a single negative resistor is inductively coupled to different parts of the circuit. Let it be B. in a T-circuit according to Fig. 5, the negative real resistance R is coupled via the coils L [ and L ' ₂ to the coils L ₁ and L _{2 of} the longitudinal and lateral branches of the T-circuit, with each half of the coil

L _{1 has} the inductance - ^l -,

If one applies Kirchhoff's rules to the circuits I, II and III in Fig. 5, so is obtained when the individual chip

L ₁
co ~~ r + im solutions and currents which can be seen from the figure, the equations

ico

; j - - SS ₁

ioo

- —3s

in the

4- % \ IrL] / L [U ₁ -im '

, L ₁

+ imLo _> - ^ ₃ .

(11)

(12)

(13)

If the system of equations is solved by means of determinants, one obtains

3,

L [ L ₂ L ₂ " + L [L "+ - L ₂ i? - in the R (L ₁ + L ₂ ) Li ( 4 ⁴ U Vl ₁ L ₁ Uo 4th L \ i
Ί

(14)

R + im {Ll + Li)

:( (- + L ₂ -imLoR

If one uses very large inductances, compared to whose inductive resistance R disappears, then according to (14) © of ig) becomes purely real and independent of frequency. Even g itself is not only frequency-independent, but also purely real, if one ensures that Eof (§ ·)> ι by suitable measurement of the individual self-inductions.

Even if R does not vanish against the inductive resistances, a purely real measure of propagation can be obtained if the self-inductions are chosen so that in Eof (g) im. Numerator and denominator the real and imaginary parts are proportional to each other. This gives the condition

the introduction of

L '-L

= χ ^ζ and

2 ■ & changes to

O.

Since negative values of χ are inadmissible due to its physical meaning, the only possible solution to this equation is

If the transmission ratios of the two transformers are dimensioned according to (15), becomes easy

so Since, as is well known, for the circuit in Fig. 3 <Sof (g) = i + (% £ -) (17)

I ^ ₃ /

is then used for the calculation of the attenuation or gain in the T-circuit Fig. S instead of the ratio of the real resistances of series and cross connections simply the ratio of the self-inductions of the corresponding branches of the current. · The fulfillment of the condition

R.,

589 5M

which brings the negative attenuation of the two circuits to match, but in no way leads to a simultaneous match of the wave resistances. Because while the characteristic impedance of Fig. 3 is known to depend on R ₁ and i? T, and is therefore frequency-independent if these two quantities are independent of frequency, this is not the case for the circuit in Fig. 5 according to equation (14). Rather, if one takes into account the vanishing of α,

= -βίηδ

H. + U-icaLlR

Since the ratio of the imaginary parts in the numerator and denominator is frequency-independent, the ratio but the real part is frequency-dependent, the phase angle of § frequency-dependent, and § can generally no longer be purely real. The circuit of the Fig. 5 thus gives the possibility of frequency-dependent wave resistances with frequency-independent Unite reinforcement.

Fig. 5 only offers the simplest example of one coupled to several branches at the same time negative resistance. Depending on the requirements for the frequency dependence of amplification and characteristic impedance the most varied changes are possible here, e.g. B. the activation of inductances in circuit II that are not coupled to circuits I and III.

II

For the production of any active four-pole circuits or shortening cables with prescribed frequency dependency of the characteristic impedance and the attenuation For the sake of simplicity, consider first of all symmetrical quadrupoles, such as those in the Connection of two similar lines, such as Pupin lines, to one another an amplifier is required; but the considerations can be made without further ado generalize to asymmetrical quadrupoles, as they are dissimilar when connecting Lines, such as a pupin cable with an overhead line, become necessary.

For example, a four-pole circuit can be selected as shown in Fig. 4, in which the adaptation of the characteristic impedance to the Pupin line, which usually begins with half the coil spacing, is easier than with a T circuit, for example. If W ₁ = q [· e, ^l * ^{v is} the impedance of the series circuit, 25Ro ■ = ρό · e '· ¹ ''' that of each of the two cross circuits according to Fig. 4, whereby the now generally complex series and cross resistances with In this case, the equations analogous to equations (7) and (8) apply

(7a)

By comparing the corresponding coefficients of (7a) and (8a) with the general ones Equations (1) and (2) for the symmetrical quadrupole are obtained here as follows Relationships:

JR ₁ '= 3. © in (g),

(9a)

(ioa)

From the equations (9a) and (ioa) the rules for the dimensioning of ΜΙ and 5R «for given values of § and b can be taken. It is essential that, in accordance with the negative values of b in these amplifier circuits, in contrast to the known extension lines, negative real parts of 5Ri and 5Ri are also used in accordance with the invention for realizing these values.

In order to assess the extent to which the conditions (9a) and (ioa) meet the conditions (9a) and (ioa) for certain circuits, it is expedient to represent § and b as a function of üi [ and 5Ro in accordance with these equations, whereby the value of α is usually not important, the artificial representation of a can, if necessary, also be taken into account. If one introduces their absolute amounts and phases _{for 5R 1} and 5R _{2, one obtains}

cos

ι sm

ψι + φ ·> ι ^ i

—- ■ —-— - r, arc tg -

² ν

(18)

Since the amplifier is terminated with a passive resistor, the sign is in (18) to be chosen so that the real part of § becomes positive. So is the cos with If the argument in square brackets is positive, then the upper sign applies, it is negative, the lower one. For the gain results

© tea = - | / _δο_ &.

The negative sign for the formation of the square root is necessary because in the determination of ? R [ and Sii for an amplifier according to (9a) and (ioa) b was assumed to be negative.

Since it often proves to be easier, matched distortion-free amplifiers with relatively To produce small reinforcements, it is advantageous to use several similar or dissimilar ones Connect ZT circuits in series to achieve the desired overall gain. With similar individual links the middle cross-circuits can be combined with one another in such a way that the Self-inductances as well as the mutual inductances and the positive and inductive real ones Resistances are doubled, but the capacities are halved.

An amplifier constructed according to the invention is shown in FIG. 6, the negative resistances being indicated by the signature. It is a two-part circuit that has a negative resistance -J i? ₀ in series with a parallel connection of self-induction i L ₀ and capacitance 2 C ₀ . The self-induction 4 · L ₀ can itself contain more or less positive real resistance. It can also be preceded by a further positive resistance within the oscillation circuit, possibly also a further negative resistance. In some cases the capacitance 2 C ₀ in the series connection can also be omitted. The cross circuits each contain a parallel connection of capacitance C ₁ and negative resistance W ₁ , which parallel connection is preceded by a further negative resistor W. The middle cross-circuits are combined with one another, the negative resistances W and W ₁ being halved in relation to the outer cross-circuits, but the capacitances C _{1 being} doubled. The positive resistances within the oscillation circuit can be kept so small in relation to the negative resistances that the active character of the quadrupole is not "touched". For reasons of symmetry, the series branches are evenly distributed on the outward and return path, with the coils on a common Core can be wound.

In order to adapt the amplifier to a Pupin trunk line of the German Reichspost with a diameter of 0.9 mm (S 0.9) and to achieve a distortion-free amplification & ₅₀₀₀ = - 1.5 ■, the following switching elements can be used for the circuit in Fig. 11, the a single link of the amplifier according to Fig. 6 without distribution of the switching elements on the back and forth of the series branches, select:

R ₀ = - 1348 ß,

W ₁ = - 5000 Ω,

W = - 180 Ω,

C ₀ = 0.0098 μΥ,

C ₁ - 0.014 ^ wF,

L ₀ = o, i63 H..

The solid curve I in Fig. 7 contains the setpoint values of the gain as a function of the frequency ze /, which are required _{to compensate for the attenuation of a line section from the go attenuation ^ 5000} - +1.5. The dashed curve II gives the gain factor of the amplifier, based on the termination with its own characteristic impedance. Fig. 8 shows in curves I and II the real and imaginary components of the wave impedance of the Pupin line, broken lines in curves III and IV show the wave impedance of the amplifier. As can be seen, the agreement achieved is satisfactory.

Another, generally two-part circuit, which is suitable as an amplifier, is shown in FIG. Here there are no switching elements other than a negative resistor W in each of the shunt branches. The series branches consist of the parallel connection of a capacitance C ₀ and a negative resistor R ₀ , which is in series with a parallel connection of self-induction L ₁ and capacitance C ₁ . The latter capacitance C ₁ can be omitted under certain circumstances. _{In front of the coil L 1 of} the resonant circuit of the series circuit containing a more or less large positive resistance, a further positive or negative real resistance can be placed as required. In this circuit, too, the above rule applies to the combination of the colliding cross members. The series connection can also be carried out in the same way as above

specified to be distributed on a round trip, with the coils on a single Core could be wound. The above remark also applies here that the positive Resistance of the oscillatory circuit only a small part of that of the negative Resistance generated energy may consume.

A third, generally also two-part amplifier circuit is shown in Fig. Io, which differs from the two preceding ones in that it contains negative resistances W only in the cross circuits, but positive R 'in the series circuit. The cross-circuit, consisting of negative resistance W, capacitance C ₂ , inductance L ₃ and capacitance C ₃ , corresponds in its structure, although generally not in its dimensioning, to the series circuit in Fig. 9. The series circuit, consisting of positive resistance R ', Capacitance C ₀ , inductance L ₁ and capacitance C ₁ , show the same type as the cross-connection, only that in general the values of the switching elements are dimensioned differently and the negative resistance in front of the resonant circuit is replaced by a positive resistance R' . The same remarks apply as in the two previous cases about the combination of the colliding cross-connections, the division of the series connections into forward and return lines and the application of the winding of the coils to a common core as well as about the possibility of omitting the capacitors C ₁ and C _{3 of} the oscillation circuits in the series and in the outer circuits. Here, too, positive real resistances in the oscillating _{circuit of the series circuit can be placed in front of the coils L 1} , positive or negative real resistances in the oscillating circuit of the direct circuit can be placed in front of the coils L ₃ . In contrast to the first two circuits, the positive resistances here not only represent corrections, but are just as important for the circuit as the negative ones. The equations (9a) and (ioa) show that with a suitable disposal of the angular dimension α with a passive 3V ₁ and an active 31 'one actually comes to an active quadrupole. If one sets α === π, one obtains

3V ₁ ~ 3 6in (5 + πί) - - S Sin h, (20)

If, as can be assumed with good approximation for the Pupin introductions beginning with half the coil spacing, § is almost real and positive, then for an amplifier, ie negative b, fR [ and 3ΐό real, namely 3Ϊ £ positive, 31% ^a ^ ^er negative. This results in a positive real series resistance in addition to a negative real transverse resistance. An adaptation of the frequency response of 3V ₁ and $ 1% to the frequency response of g and b prescribed by the right-hand sides of (20) and (21) ensures that the circuit does indeed have negative b , i.e. becomes active. The circuit of Fig. 10, however, has the required, essentially purely real frequency curve of the impedance of its series and cross circuit.

The usual amplifiers use sieve chains outside the transmission area or choke chains to reduce the unnecessary frequencies. These are with the The amplifiers described here can also be used. You can also have one or more Connect extension cables in front of the amplifier or connect the amplifier in several Dissolve links if, in special cases, a gradation of the reinforcing effect is required will. In general, all the usual measures in amplifier operation, such as use shared batteries for the individual electron tubes, as far as they are not ' become superfluous due to the nature of the new amplifier, to apply analogously to it.

The application of the method is not limited to II-circuits, but other four-pole circuits, such as T- and cross-connections, can also be used Easy to change the degree of coupling to set different amplification levels, inductive couplings can also be used for the same purpose with other circuits They can also include upstream circuits, such as ring transformers, etc. Also, one need not restrict oneself to four-pole connections, but rather 'can construct active circuits with more than four open ends, multipoles, with prescribed values of their complex constants, such as them for passive Vi elpole are already known. An active six-pole is z. B. represented by an amplifier with three outgoing double lines, wherein the current coming from each double line is to be passed on in each of the other two lines in an amplified manner. One can also

with multipoles, not only with quadrupoles, which amplifies the incoming electrical energy pass on, generally speak of negative damping, if of course also im It remains to be clarified by the sign of which characteristic size of the Multipole the energy gain can be seen.

The interior of such a multipole, in particular a quadrupole, can consist of any Combination of active and passive elementary bipoles, if necessary with the addition of inductive couplings, the two or a larger number of coils couple with each other exist. We want the borderline case to be more pure lossless inductances, capacitances and mutual inductances under the passive quadrupole calculate. The negative resistance of the two-pole connected electron tubes does not need to be frequency-independent , but can have an arbitrary frequency dependency, which can be differentiated by suitable Choice of characteristics and passive connections within the active Bipolar and can be achieved by other means. Yes it can also be used within the active Any other generator of electrical energy, such as an alternating current

30 dynamo machine. The frequency dependence of the active and passive Switching elements can be designed so that the four-pole or multi-pole only acts as an amplifier in certain frequency ranges, while it acts as an amplifier in others has positive damping.

Claims (10)

Patent claims: 1. A method for the production of active multi-pole circuits in T, Π or cross connection with prescribed frequency dependence of the characteristic impedance and the damping, characterized in that - starting from a .passive multi-pole with the same characteristic impedance and positive damping - only the positive real ( Ohmic) resistances can be replaced by negative real resistances of the same size. 2. The method according to claim 1, characterized by application to the longitudinal and / or transverse resistances of symmetrical T-, Π- or cross-connections, which can optionally consist of several links of the same type and fused to one another and which optionally have sieve chains in a manner known per se, Throttle chains, extension lines and / or the like can be connected upstream. 3. Switching arrangement according to the method of claim 1 and 2, characterized in that; marked, that the negative resistances are inductively coupled, possibly simultaneously to the series and shunt branches. 4. Switching arrangement according to claim 3, characterized in that the degree of coupling is changeable. 5. ZT circuit according to the method according to claim 1 and 2, characterized by a series branch consisting of a current resonance circuit or a pure inductance is in series with a negative resistance, and two are equal to each other Cross branches consisting of a negative resistor in series with a parallel connection of another negative Resistance with a capacitance exist (Fig. 11). 6. i7 circuit according to the method of claim 1 and 2, characterized by a series branch consisting of a current resonance circuit or a pure inductance in series with a negative resistance and a capacitance in parallel with this series circuit, and two equal cross-branches that each contain only one negative resistance (Fig. 9). go 7. i7 circuit according to the method of claim 1 and 2, characterized by two equal cross branches in the manner of the series branch of the circuit according to claim 6 and one series branch, the differs from the cross-branches only in that it replaces one negative real resistance a positive real resistance in series with the current resonance circuit or the inductance contains (Fig. 10). 8. Arrangement according to claim 5 to 7, characterized in that the inductors positive real resistances are connected upstream. · 9. Arrangement according to claim. 5 to 8, characterized in that the longitudinal branches are divided in a manner known per se on both wires of a double line. no 10. Arrangement according to claim 1 to 9, characterized in that the negative Resistors wholly or partly have common energy sources. 1 sheet of drawings berlin Printed in the reichsdruckeuei