USRE25064E - kruesi - Google Patents

Description

0d. 24, 1961 u s Re. 25,064

RADIO DIRECTION FINDER Original Filed June 17, 1957 7 Sheets-Sheet 1 F IG.

FIG 9A Z33 L3 INVIENTOR GEOFFREY GOTTUEB KRLESI.

BY mmmrlt, Cams- 3% ATTORN EYS.

7 Sheets-Sheet 2 ATTORNEYS.

G. G. KRUESI RADIO DIRECTION FINDER 0 WJ mm M m m y C mm I l G a v E w W G a w .m. lllllllllll I n K8 3 +21 I. B i, n R 4 I c Z.

Oct. 24, 1961 Original Filed June 17, 1957 I I l l l I l I I l I I I I I L l I I I I L Oct. 24, 1961 G. e. KRUESI RADIO DIRECTION FINDER Original Filed June 17, 1957 7 Sheets-Sheet 3 52 158 20.2582 H 3 R u 8 o m T B a H mm mm l u w o mz wx W. Y (REP m B a a 523% I 2.56 109 9 oh I B b fi F uz: zommimz E zzEz a #558 r 5 58:

mmmauz' ATTORIJEFZ Oct. 24, 1961 s. e. KRUESI RADIO DIRECTION FINDER Original Filed June 17. 1957 7 Sheets-Sheet 4 V. fi

3 3 t: w 11 m.

1g l l.

INVENTOR ATTORNEYS Oct. 24, 1961 G. a. KRUESI mm DIRECTION FINDER '7 Sheets-Sheet 5 Original Filed June 17, 1957 2 6 .l. 6 6 2 H. w x E W b. w H w c u Ix 4 e. P "v m B a INVENTOR GEa FREY 6077-4153 AAuEs/ ATTO R N EYS Oct. 24, 1961 G. s. KRUESI RADIO DIRECTION FINDER Original Filed June 17, 1957 INVENTOR ea,er ezv Gar? 1:1 Anuzsl BY m MJM ATTORNEYS Oct. 24, 1961 a. G. KRUESI Re. 25,064

RADIO DIRECTION FINDER Original Filed June 17. 1957 'T Sheets-Sheet 7 G are 1/0 eh m NETWOQK IN V E N TO R zorreiyg'arrusa Aeues/ United States Patent Ofifice Re. 25,064 Reissued Oct. 24, 1961 25,064 RADIO DIRECTION FINDER Geoffrey G. Kruesi, Los Altos, Calif., assignor of 10 percent to Albert C. Nolte, Sr., New York, NY.

Original No. 2,910,693, dated Oct. 27, 1959, Ser. No.

666,147, June 17, 1957. Application for reissue Apr.

6, 1961, Ser. No. 103,672

23 Claims. (Cl. 343113) Matter enclosed in heavy brackets appears in the original patent but forms no part of this reissue specification; matter printed in italics indicates the additions made by reissue.

This application is a continuation-in-part of my application Serial No. 269,218, filed January 31, 1952, now abandoned. I

The present invention relates to high frequency (HF) and ultra-high frequency (UHF) direction finders in which a HF mixing circuit or coupling structure between directional and nondirectional antenna to 2. HF receiverinput, or collector circuit to visually and aurally determine two directions 180 degrees apart, respectively, of an incident radio wave, both visibly and aurally indicated in the respective signal output channels of an evaluation network associated with the collector circuit.

The operation of direction finders of this kind is described in Patent No. 1,868,945. There is at present an inability to properly identify the received aural signal under otherwise standard signal-to-noise ratio conditions at field strengths of the incident radio Wave below the standardized, much too high values, varying from 25 to 50 microvolts/meter. Thus for a standard output signal in the visual amplifying channel 100 micro-amperes for a one degree deviation of the direction of the incident radio wave from the directional antennas zero signal pick-up direction easily readable bearing directions are obtained with standard airborne antenna while any further decrease in field strength from the above standard values causes the aural signal to disappear in the receiver noise and local interference noise. The somewhat unusual phenomenon, otherwise loosely termed frequency error constitutes in other words an appreciable drop in sensitivity of the bearing indication because of the limit in operating range set by the lack of aural identification of the signals. The lack in sensitivity is particularly demonstrated by making a comparison between the above performance and that of standard communication receivers which, as is well known, give very satisfactory results at field strength values of only 5 microvolts/ meter and less. Thus, briefly stated, the invention is concerned with improvements of the aural performance range by comparison with the visual radio compass range, respectively, at much lower HF carrier signal inputs to the two amplifying networks to render the same signal outputs to the aural and visual amplifying channels of the receiving equipment.

The coupling structure has been found to be the major source of trouble. It was found that the best possible optimum coupling conditions in the coupling structure to produce an optimum maximum output signal in the visual amplifying channel of the receiver do not correspond with those that would be necessary for the best possible aural reception, unless individual manually performed tuning adjustments in the two antenna amplifying networks not only differing from each other, but also at each frequency over a given tuning range of the receiver proper, are resorted to. Such a procedure, however, being inconsistent with electro-mechanical gauging of the HF resonance-tuned circuits in the three networks by means of a single tuning drive is completely out of the question.

The invention eliminates an inherent electro-magnetic condition of constraint among the total phase constants, a, 5, 1, respectively, of the non-directional and directional amplifying network and the collector circuit which requires the sum of these phase constants to remain constantly equal to electrical degrees for every frequency inside a given frequency range. This electrical condition of constraint is a direct consequence of a magnetic condition of constraint among the three possible coupling coefficients in the currently used HF mixing transformers according to which the ratio of the product of any two of the coefficients to the third can never be made smaller than the numerical value of V2. If, as in the case of a single tuning drive, the individual phase constants of the three networks are left to themselves over a given frequency range, obviously, they cannot undergo a change unless this change is eifected by the change in tuning constants of all three networks; the necessary degree of freedom required for any one of the three phase constants to vary independently from the other two being lacking.

This, by virtue of the elimination of the magnetic condiv tion of constraint, accomplished by the elimination of one of the three individual, generalized coupling pairs considered in opposite directions of energy flow, renders the required additional degree of freedom to bringabout highly desirable output terminal conditions in the two antenna amplifying networks by which an in-phase condition of the signal response functions in the collector circuit can be maintained over any given frequency range with the three networks electro-mechanically ganged. I am here referring to symmetrical terminal conditions as distinguished from the asymmetrical terminal conditions in the antenna amplifying networks of currently used radio compasses (right-O-left indicating instruments) and automatic radio direction finders 0 to 360 indicating instruments) which, in the presence of an unavoidable third shunt coupling in the coupling structure (HF three winding mixing transformer) are the reason for energy wasting circulating currents by which the HF carrier signal input to the two antennae has to be raised to give standard visual and aural channel outputs. Asymmetrical terminal conditions are also responsible for fre quency variations of the amplitude of the combined antenna signal in the collector circuit over and above the natural frequency variations of the effective antenna heights.

Briefly summarized, establishing coupling conditions that are free from any constraining relation makes available the choice of symmetrical output terminal conditions of the antenna amplifying networks by which the variation over a frequency range of the signal response function of one antenna amplifying network may bemade equal to that of the other considered in the direction of energy flow. In addition, feedback reactions of the combined signal in the collector circuit are prevented in said output terminations by uni-laterially conducting active, or, passive circuit elements with symmetrical electrical characteristics. Thus, any energy exchange between the HF input circuits over and, respectively, via the antenna amplifying networks and the collector circuit is effectively stopped. This then, constitutes the fundamental requirement for the obtainment of symmetrically varying signal response functions inthe collector circuit over quite extended frequency ranges. As a special selec tion of phase constants of the antenna networks are those whose difference remains constantly 90 degrees consistent with rendering optimum maximum antenna signal amplitudes in the collector circuit. This means that because of the compensated 90 degree phase difference between the antennae E.M.F.s the said signal response functions add up algebraically and consistently over a frequency range. If, as is usual practice, the signal amplitudes in the maximum signal pick-up direction of the directional antenna are made approximately equal, the combined antenna signal in the collector circuit'has a frequency independent cardioid receiving characteristic. This frequency independent performance has opened up possibilities to raise the operating frequencies into the 100 to 1000 megacycles and radar frequency regions so that novel applications of the present direction finder principle of coupling two antenna amplifying networks to a collector circuit, related to the Ilas, Vor, Tacan and Vortac instrument landing and airtrafiic control systems are made possible. An entire variety of coupling structures of, both, the inductive types (separate systems) and the direct coupled types (single physical systems) are made possible by the present invention. In fact, the electrical behavior of all the various coupling structures about to be enumerated is quite analogous to that of two co-axial cables, transmission lines or wave guides energized at the input ends with a source of energy each, respectively, 90 phase displaced, having physical lengths diifering electrically by 90 degrees and a common output termination with an impedance equal to the characteristic impedance of an infinite transmission line. Such an arrangement will also result in the total absence of reflected energy back into the two transmission lines and permit the in dependent flow of energy in each line toward the common output termination.

Of the above single physical systems which this invention makes possible there are three classifications of HF circuitry, as follows:

(A) balanced parallel input antenna networks with respect to a HF receiver-input or collector circuit, correcting for frequency errors, respectively, ahead of the point of coupling or collector circuit. These may include coupling structures of the inductive type of separate systems.

(B) balanced parallel input antenna networks not including coupling structures of the inductive type and, respectively, correcting for frequency errors after said point of coupling.

(C) balanced orthogonal-input antenna networks, such as bridge-adding networks.

All of the above coupling structures are typical of means to prevent any energy exchange between the input terminations of the antenna amplifying networks in opposite directions of energy flow, respectively, that could take place over said amplifiers and the point of coupling. If we denote the coupling coeflicients between the said output terminations of the nondirectional network I, the directional network II and the collector circuit III with (k g kg (k13=k 1) and (k23=k IS found that in eliminating the coupling coefiicient k =k by either a special three winding transformer construction, or, by devising coupling systems as noted under classifications, (B) and (C) above the constraining relation among the phase'coustants, a+fi+7=i90 reduces to a simpler one, called the principal requirement: (mm) ,Ba= i90=constant over a given frequency range, so that the phase constants, m and 6, either by reason of a source frequency variation with constant antenna network parameters, or, by reason of network parameter variations with a constant source frequency, can be made to vary independently from each other and the phase constant, '7, of the collector circuit.

Thus, the principal requirement (mm) is dependent upon the achievement of the following functional requirement, insofar as best possible performance is concerned:

In this connection the condition (mm) states in practical language, that the frequency variation of one phase constant may be made equal to that of the other; i.e., a positive increment, A 3, of a positive phase constant 8 during a frequency variation, Af, may be made equal to a positive increment Act of a negative phase constant, -oz.

The second stipulation in the functional requirement (I) states in practical language that, over a given frequency range, say, the voltage ratio in the maximum signal pick-up direction of the directional antenna between the sense antenna voltage response at the point of coupling, and the sense antenna input voltage of its network, will be equal to a corresponding voltage ratio between the directional antenna voltage response at said point of coupling and the directional antenna input voltage of its network.

Simultaneously therewith, the first stipulation in the functional requirement (I) states in practical language that, over a given frequency range, say, the voltage ratio, respectively, between the voltage response induced in one network (considered passive) and the voltage source at the input of the other network (considered the active network) must remain zero, while the passive receiver network is considered tie-coupled or open-circuited in both cases.

Since we denote with a and 3 the phase constants of the active antenna networks including the coupling structure, respectively, from the input nodes to the point of coupling and as such represent the phase functions of the respective system functions, it becomes clear that (Ba), the difference which according to (mm) is required to remain constant, relates also to the said common point of coupling, irrespective whether this common coupling point resides in the output termination of network I, or in that of the loop network II, or, in the input termination of the passive network III of the receiver. The stipulated difference between ,6 and a, in fact, is wholly due to an identical difference between the phase constants of the voltage or current sources (sense and loop antenna).

A few remarks will be made in regard to practical precautions that must be taken when designing single systems of the classifications (A), (B) and (C). In particular the remarks are related to a specific emb0di ment insofar as they are related to a particular embodiment of this invention belonging to the classification (A) of separate systems. For this reason this first, particular embodiment will be briefly described:

This embodiment is a high frequency H.F. mixing transformer, in which one of the mutual coupling co-- efficients k k k can be made to vary, independently from the other two, equalized, coupling coeflicients, respectively, from a positive value, over a zero value, to a negative value. The condition of constraint is removed at the precise moment when a certain value of the mutual inductance corresponding with the variable coupling coefiicient just suflices to counteract the value of the mutual capacitance in parallel with the mutual inductance. If the point of coupling resides in the passive collector circuit III, there-' fore, there are two symmetrically disposed equalized coupling coetficients, k k which remain between network III, respectively, and the output terminations of networks I and II. Exclusively for this case, then, the circuit arrangement may be compared with the method of operation of single systems, classified under (A) above. But, it cannot be strongly enough emphasized, that in this particular system the distributed stray capacitances between the active networks I and II are neutralized ahead of the common point of coupling in network III.

There are single systems of the classification (B) known in electronic computer engineering as adding circuit which strive to eliminate the stray capacitances between voltage sources that must be added. The voltage sources are each connected in series with balanced resistors, respectively, terminating into a common grid input of a negative feedback amplifier of the plate follower type, the idea being, to counteract the said mutual stray capacitances by the negative feedback action after the common point of coupling (grid input of the negative, feedback amplifier) and, respectively, prior to the re. ceiver input which amplifies further the added" voltages;

A second embodiment of this invention, therefore, consists of a high frequency input system for radio direction finders of the class described, in which said series resistors are replaced, respectively, by combinations of the said resistors connected to the outputs of the antennae networks in series, terminating, with the respective outer ends of said resistors connected together, into a common point of coupling, prior to applying said negative feedback' principle.

A third embodiment of this invention belonging to the classification (A) of single systems consists of a coupling structure which is quite similar to the operation of the mentioned first embodiment of this invention except that no mutual inductances are employed. The elimination of the undesirable system function is again taken care of ahead of the common point of coupling. The coupling structure is the dual of that of the first mentioned embodiment of this invention; i.e., the function of the remaining equalized coupling coeflicients k =k is taken over, instead of by mutual inductances, by equalized self inductances having an RF. resistance in common. The joined inner ends of these two equal coil halves form with the ground a node pair across which is connected the collector network, while the outer ends of the two coil halves connect to a differential variable tuning condenser and, respectively, to the high potential nodes of the output terminations of the active antenna networks I and II. At the resonance setting of the differential condenser the voltage drops across the node pair of the collector circuit III, and the one developed across the circuit branch comprising one of the coil halves connecting to the passively assumed high potential end of one of said output terminations (while the other high potential end is considered actively energized by the other antenna network) can be made equal and phase opposed, so that the system function in either direction between the antennae networks become zero This zero condition includes the presence of all capacitive stray couplings between the two antenna networks, because they are comprised (in parallel with the differential condenser) in the resonance tuning setting of the differential condenser, and this, at each and every frequency over the tuning range of the receiver.

A fourth specific embodiment of this invention, also belonging to the classification (A) of single systems, relies upon the rendering ineffective of the undesired system function between the antenna networks, respectively, before the common point of coupling, by a brute force method of overcoming the effects of mutual stray capacitances between the active networks in a differential amplifier modified to operate at radio and very high (V.H.F.) frequencies.

A fifth specific embodiment of this invention, again beloging to the classification (A) of single systems relies upon the rendering ineffective of the undesired system function, respectively, before the common point of coupling, by the use of a balanced cathode follower circuit prior to the common point of coupling to the node pair of the receiver network III. It is again, at best, called a brute force method. The grid-input capacitances (equalized) between the two grid nodes and ground, and, respectively, between said grid nodes and the common cathode terminal form the two couplings k and k between the collector network III and, respectively, the high potential ends of the output terminations of the active antenna networks I and II. By the cathode follower action said input capacitances are greatly reduced, hence, any mutual coupling between the active networks is confronted with a consequential high impedance in series therewith before the common point of coupling, which is the common cathode circuit in parallel with the receiver input network III.

A sixth specific embodiment of this invention, belonging to classification (B) of single systems eliminates the undesired system function indicated by the functional requirement (I) is initially removed by coupling the output terminations of the active antenna networks, respectively,

in orthogonal fashion, to the two node pairs of an equal arm bridge circuit. Mutual stray capacitances between the active antenna networks I and II which ordinarily will introduce a frequency error function may thus be counteracted by adequate unbalancing of the bridge network, thus creating a corrective frequency error function which opposes the former. Balancing a bridge network of this kind, and keeping said balance over a frequency range is in a way comparable with the establishment of a symmetrical circuit design of the active antenna networks I and II on the one hand, and the removal of constraint among the three coupling coeflicients, such as in the case of the first specific embodiment of this invention.

It is particularly emphasized that the present invention does not confine itself to the specific deficiencies of any one single element in the above enumerated circuit systems, but rather is concerned with the general relationships between all of the elements which lead to the undesirable frequency error between the response functions in the systems. Practical circuit evaluation, therefore, must touch upon:

(1) Symmetry of the equivalent network parameters, as represented by the individual system functions, or, as we have called them: generalized, individual couplings, respectively, referred to a common point of coupling between two of the three networks I, II and III, and the input node pairs of the voltage or current sources in the systems (sense and loop antenna).

(2) The possibilities of the elimination, or, otherwise rendering ineffective the individual system function between the active antenna networks I and II in the two opposite directions of antennae energy flow (generalized, individual coupling pair) either, before, or, after the common point of coupling in the system, and, furthermore, maintaining this condition over an extended frequency range.

(3) In conformity with (1) above, the attainment of an even energy distribution in the system (see second stipulation of the functional requirement (I) above).

(4) Making the total current losses in the high frequency input system of an optimum minimum which fact will result in optimum maximum response functions at the common point of coupling corresponding with an optimum maximum turn-indicator sensitivity over the range, as given by the principal requirement (mm).

The general design rules enumerated under paragraphs (1) to (4) above are found to be equally applicable to sinusoidal driving and response functions and transient, non-periodical driving functions (time modulated radio transmissions) Other objects and advantages of the novel invention and systems will become apparent, and the invention will be fully understood from the following description and drawing, in which:

FIG. 1 is a diagrammatic representation of a directional antenna network and a non-directional antenna network coupled to a common receiving device by means of a coupling structure (P) with two inputs and an output showing the three possible bilateral (assumed inductively coupled) coupling paths between each of the three possible combinations of two networks;

FIGS. 2A and 2C are vector diagrams showing voltage and current relationships of the loop and sense antenna input circuits, respectively;

FIG. 2B is an equivalent circuit diagram of the sense antenna input circuit;

FIG. 3 is an elevational view, partly broken away, showing a three winding coupling transformer suitable for use in practicing the invention;

FIG. 4 is a schematic circuit diagram of a radio direction finder using the coupling transformer shown in FIG. 3;

FIG. 5A is a generalized, mathematically equivalent, circuit diagram showing certain coupling paths among three mutually coupled networks such as found in the conventional direction finder circuits of the cardioid type;

FIG. B is a generalized, mathematically equivalent, circuit diagram showing a coupling system in accordance with the invention which uses the principle of pure parallel feed systems;

FIG. 5C is a generalized, mathematically equivalent, circuit diagram showing a coupling system in accordance with the invention which uses the principle of pure series feed systems";

FIG. 6 is a schematic circuit diagram of a radio direction finder using the coupling transformer of FIG. 3, showing a modified form of the invention;

FIG. 7 shows a typical high frequency mixing circuit arrangement commonly used in radio direction finders of the conventional cardioid type;

FIG. 8 shows a high frequency input circuit arrangement of the series feed type for a radio direction finder of the visual indicating type;

FIG. 9 shows a generalized equivalent double-dual circuit of the circuit arrangement shown in FIG. 1;

FIG. 10 shows a circuit arrangement including a balanced modulator connected to a modified form of the transformer shown in FIG. 3;

FIGS. 11 and 11a show circuits utilizing a negative feedback amplifier of the plate follower variety for coupling the antenna networks to the receiver network;

FIG. 12 shows an equivalent mathematical dual circuit of the coupling structure between the three networks shown in FIG. 1;

FIG. 13 shows the basic elements of a specific embodiment of the invention, derived from FIG. 12, in the form of a novel coupling structure for direction finders;

FIGS. 13a, 13b and 13c represent vector diagrams in connection with the operation of the coupling structure of FIG. 13;

FIG. 14 shows a schematic wiring diagram of a coupling structure using the principle of operation given by FIGS. 13a, 13b and 130;

FIG. 15 shows the wiring diagram of a radio direction finder which may be made to operate in the 100 megacycle frequency spectrum, using the principle of coupling shown in FIG. 13;

FIG. 16 shows the application of a differential amplifier as a coupling structure;

FIG. 17 shows a circuit using a cathode follower amplifier as a coupling structure between the antenna and the receiver networks;

FIG. 18 is a schematic diagram of a bridge type coupling circuit;

FIG. 19 is a circuit diagram of an embodiment of the bridge type coupling.

FIG. 1 shows a conventional high frequency input system of a radio direction finder of the class described in block form, consisting of two active antenna networks I and II, respectively, connected at their input terminations to the sense antenna 1, and a directional antenna 2, and a passive network III (designated with Z the input impedance of the collector circuit 3). All conventional high frequency input systems of this kind are mechanically and electrically ganged, i.e., high frequency resonance-tuned circuits in the three networks may be tuned to the operating frequency of the incident radio wave by means of a single mechanical control, or, by means of an electrical remote control device (not shown).

Between the output terminations (Z Z of the active networks I and II and the passive networks III is a coupling structure P" merely indicated by three mutual inductances as represented by the coupling coefiii2 21; ia ai; za 32- The already referred to magnetic condition of constraint among these coupling coeflicients is, symbolically expressed, as follows:

which relation, in turn, leads to a physical limitation among the ratios of shunt-to direct coupling efiects,

a2 21 l2 1 (1]) 31 en m) 2 2 i.e., no matter what is tried in the way of changing the number of turns, or, the lengths of reluctance paths of the magnetic circuit of a three winding mixing transformer, or, with two single, two winding transformers in which a third capacitive shunt coupling is present between at least two of the transformer windings, a residual third shunt coupling will always remain and thus bring all three coupled networks into mutual relationships. Hence, at no time is it possible to eliminate shunt coupling effects, as indicated by the products, k k or, k km of the above expression (ii).

If I designate the gains of the active networks, respectively, from the antennae input circuits to the point of coupling to collector network 3 and in the opposite direction of energy flow from network 3 to said input circuits with individual generalized coupling pairs considered in opposite directions of energy flow, it is readily seen that, since the above ordinary coupling coefficients must be contained therein, the foregoing conclusions derived as to the inability of separating the networks from each other must hold true likewise for said generalized coeflicients or individual system functions which, when multiplied with the respective driving functions result in the individual response functions. These coeflicients we shall distinguish in pairs, thus:

(K K respectively, between networks 1 and 2,

(K K between networks 1 and 3, and,

(K K between networks 2 and 3. The phase functions of these coefficients we have already referred to as the angles, at, B, and 'y, of networks 1, 2 and 3, while the amplitudef unctions are written thus:

( 12 21) etc.

The total gain of the two networks 1 and 2, as measured at the point of coupling, respectively, with only one driving function (antenna E.M.F.) active in the mutually coupled total system of networks 1, 2 and 3, we designate with: K considered in direction from the antenna 1 to the point of coupling 3, and K considered in direction from the antenna 2 to said point of coupling 3. They must, therefore, be some sort of an algebraic expression, as we shall see later in this specification, indicating a general relationship among all of the above individual system function pairs.

Each of the antennae 1 and 2 is characterized by a driving function (E.M.F.) which produces a response function, or, a response signal across the input nodes of the network III, and it is the purpose to obtain response functions which, in the maximum signal pick-up direction of the directional antenna 2, are of equal, phase aligned amplitudes, respectively, over a given operating, or, tuning range of the collector circuit 3. Expressed in other words: the amplitude and phase functions of the two response signals must have identical frequency responses the respective curves of which run parallel with each other. If the amplitudes remain equal and phase opposed (or phase aligned) in said loop direction over a frequency range, the true algebraic addition of said response signals results as we shall see, in a cardioid receiving characteristic.

The directional antenna 2 may consist of a loop, if the operating range is in the long wave and broadcast frequency spectrum of to 1750 kilocycles, or it may consist of an antenna array of the Adcock, or, Franklin types, if the operating range is in the VHF. and UHF. spectrum of 100 to 1000 megacycles, commonly employed in the Vor-Dem, Tacan and Vortac airtraffic control-and the Ilas instrument landing systems now in use, or proposed in this country. Either type of antenna 2 has a driving function, i.e., and

in which the angle of rotation of the directional antenna structure 2, respectively, 6=0 and +180 is defined to be lying in the maximum signal pick-up direction, being normal to the frontal area represented by the directional antenna conductors. This normal direction coincides with the minimum signal pick-up direction of the antennae for which the angle 6 is equal to, respectively, 6=+90 and +270. For this reason the normal direction is called the directional antennas zero, or, symmetry axis. The angle, (M, denotes the time reference phase" which, as we shall see, coincides with that of the driving function, or, E.M.F. E of the non-directional sense antenna 1.

The directional antenna 2, furthermore, is characterized as having an internal impedance, Z and feeds, over a high frequency feeder cable (not shown) an input coupling stage, Z an intermediate coupling stage, Z and the output stage, Z The chain of coupling stages is called the directional antenna amplifying networ and is designated with II. The stages are supposed to operate in true series fashion in which it is assumed that no individual shunt, or, feedback couplings between said stages will interfere with a true series operation of network II.

The above symbolic expression for the E.M.F. E shows the directional receiving characteristic to be that of a lemniscate, and it is further seen, that the phase of the induced E changes its sign from to when the direction, 5, passes through the value, 90, and that said phase changes its sign, respectively, from to when it passes through the value of +270", assuming in both cases the same sense of rotation. I speak, therefore, of a positive and a negative 180 phase reversal, respectively, in the "+90" forward, and, in the +270 rear direction of the incoming radio wave.

The non-directional antenna 1 has a driving function, or, induced E.M.F. E with its internal impedance Z (not shown). Its symbolic expression is given by:

in which the phase angle of (p is the already mentioned reference time phase. It is seen, that the receiving characteristic of this antenna does not change its sign, or, polarity, with respect to the direction, 6, of the incident radio wave. The said characteristic is usually assumed to be circular in shape, although any such statement should be regarded with precaution in the case of airborne antennae. As a rule a wire-or mast antenna is used in the case where the wind drag on medium speed, medium performance aircraft does not result in excessive mechanical vibrations, while, on high speed, high performance aircraft counter-sunk capacity plate antennas are used, properly insulated from its supporting frame which, in turn, is flush mounted with the aircrafts fuselage. The antenna 1 is fed, again over a high frequency feeder cable (not shown) a chain of series connected coupling stages, i.e., the input coupling stage, Z the intermediary coupling stage, Z",;, and the output stage, Z' Identical considerations hold with respect to the series operation of this network, denoted with nondirectional antenna amplifying network I.

If each of the output terminations of Z Z and the input termination of Z in FIG. 1 consist of the physically separate coils of a three winding H.F. mixing transformer usually employed in the conventional high frequency input system, we are not necessarily entitled to call such a system a separate system of coupled networks because of the individual character of mutual inductances between them. This is because of capacitive elements which are not noted in FIG. 1, and which may exist in the form of distributed capacity, inside the chassis of the direction finder, of which equivalent lumped capacitive elements may be thought of as lying in parallel to said individual mutual inductances, as expressed by the coupling coefiicient pairs, k -k k k kmh kaz. Especially in the case as in modern three winding H.F. mixing transformers, wherein said mutual inductances are relatively small and comparable in magnitude of the antennae capacitances of antenna 1 and 2, the influence of these capacitive shunt elements may seriously affect the operation of the coupling structure in a manner not ordinarily anticipated by the looks of the wiring diagram of the high frequency input system.

However, it is possible to form from the network configuration of FIG. 1 with the assumption that only mutual inductive couplings are present in the coupling structure, a dual coupling network such as shown in FIG. 12. The mutual inductances are inverted" thereby, and so are the self inductances of the three transformer coils. The inverted inductances receive the character of ordinary circuit elements which may now be added to said equivalent lumped capacitive elements, ordinarily not indicated on circuit diagrams. Mathematically it is the admittances which may now be added together, to result the admittances M12=M2b M13=M31 and M =M that are seen to be bi-lateral in character because the coupling structure is of a passive nature, there being neither voltage nor current sources comprised in these admittance branches.

They are frequency dependent and so are the corresponding coupling coefficients which, for this reason I now denote with, (k12 k21)+jw, (k =k and (k ='k since they can, generally, be expressed as the negative reactance of elastancm:

while the respective positive capacitive reactances may be expressed in the form: +j.w.C If we ascribe an R.F. resistance to the coil inductances of the transformer windings, say, R R, we may put the corresponding conductance, respectively, in parallel with the other two elements and obtain for the value M as the total admittance between nodes, i and j, in general, the symbolic expression:

in which and, W=2'1r'f the angular velocity of the radiated wave of frequency f.

FIG. 12 represents a mathematical equivalent of the high-frequency input system of FIG. 1 and may be analysed in accordance with standard methods found in text books (see Gardner & Barnes, Transients in Linear Systems" vol. 1, John Wiley & Sons, Inc., New York). The mathematical equivalent is seen to have current sources I of the sense antenna 1, and I of the directional antenna 2, with their internal admittances, respectively (Z11) and (Z while the output admittances of the active antenna networks denoted in block form with I and II, are indicated with Y, and Y The common point of coupling, node c represents together with the common datum junction (dash-dotted line) the input node pair to the collector circuit of network III.

The final analysis, however, is based on the loop method which necessitates forming a mathematical dual of the network configuration of FIG. 12 into that shown in FIG. 9A. The analog is seen to have three independent geometrical loops with the currents,

flowing therein, respectively, as the independent variables, two voltage sources E E already referred to as the sense and directional antenna E.M.F.s respectively, the individual transfer impedance Z Z of the active networks I and II, the input impedance Z of the passive network III (collector circuit) and the mutual impedances, Z12=Z21, Z13=Z31, and Z23=Z32 (the internal antennae impedances Z 2 are included [included] in said transfer impedances Z Z but not shown).

The mutual impedances no longer have the significance of the former mutual inductances above referred to since they are the duals of the said total admittances M as given by the definition (nn) above.

Referring now to FIG. 12, the initial assumption will be made that only the current source I is active, while the current source, I -cos 6, i inactive and replaced by its internal admittance, Y Network II is, therefore, passive and network I active. nal voltage, say, e i appearing between junction and ground. We define the ratio of the total (network) signal response due to the current source I with, 31,+ w= 31,+ w- P-L +I$31,+jw- Similarly, y suming I inactive and replaced by its internal admittance, Y while the current source I .cos 6, is active, another signal voltage, e .cos 6, will appear across junction c and ground, so that, for an arbitrarily chosen direction, 6, of the incident radio wave, I may define the total (network) signal response due to the current source I .cos 6 of network II, e .cos 6 (total system function, or, total generalized coupling) to the current source, I .cos 5, with, Kt =K .exp.; +j

K and K in other words, may be considered amplification gains (complex) of the total network configuration of network I, II and III with either I, or, respectively, I .cos5, active therein.

If it were not for the fact that the mutual coupling links, M M are the cause for an interaction" between networks I and II, obviously, we could then think of individual amplification gains, respectively, of network I and network II, counting the gains from the respective network input terminations to the common junction or coupling joint c of network III. Thus, if we had a means within the coupling structure to render ineffective the coupling effects due to the coupling links M M these individual amplification gains, which we define with the symbolic representation, K -jand K respectively, in opposite directions between the input terminations of network I and II, would become zero.

This fact in turn leaves only two of the total of three such individual, generalized coupling pairs, namely, (K13, K31)+ and (K23, K32), and IS seen from FIG. 12 that a possible energy transfer between the input terminations of networks I and II is now restricted to take place, respectively, from network I to network II over the series combination of coupling links, expressed by the product of M .M (indicated by arrows) and from network II to network I over the series combination of coupling links (indicated by arrows), expressed by the product of, M .M

Now, in order to render ineffective the interaction between the two networks I and II over said series coupling paths and, yet, preserve an energy transfer, respectively, in directions from the input terminations of network I and II to the common point of coupling (junction c) it is merely necessary to impede energy flow over the coupling links, M and M in directions from junction c to junction a and, respectively, b. One of the possible practical means to do this consists in providing the output terminations of networks I and II, to be sure, prior to the junctions a and b, with uni-laterally conducting circuit elements which permit the flow of energy toward the junction c, but not in the opposite directions, i.e., in directions c to a, and c to b.

Having accomplished this it is readily seen that the remaining individual, generalized couplings are, K and K They are the useful ones contributing to the combination of the two antenna signal responses, e and e cosa. The question now arises: (I) under which parameter conditions in networks I and II can we expect the amplitudes /e HW and /e .cos6/ There will then be a sigto remain in a constant ratio to each other while the operating frequency is varied over the usual range of approximately a 2.5 to 1 ratio, and (2) under which parameter conditions in networks I and II can we expect the phases of exp.4 p =u, and exp.L =fi, to assume a constant difference (13-00 as the said operating frequency is varied over said range? Then, the desirability for these amplitude and phase conditions should now be obvious: we strive to make the individual frequency variations of amplitudes and phases of the two signal responses, e and e .cos6, identical with respect to each other as the operating frequency is varied. This will insure the possibility of obtaining a in-phas'e condition of the signal responses in the maximum signal direction 6=0 direction, in which the amplitudes must be alike if the combined signal response, as a function of 6, is that of a cardioid characteristic. Or, by the same token, we may get the condition of true phase oppositions in the 6=+180 direction, giving a true eancelation and the desired cardioid minimum signal.

It is seen that we have arrived at the establishment of the principal requirement (mm) and that of (I) referred to at the outset of this specification through purely elementary consideration. Actually, they represent mathematical limiting conditions among the parameters of three integro-differential equations set up for the dual network shown in FIG. 9A.

Thus, the parameter conditions required of conventional high frequency input networks, such as shown in FIG. 1 are complicated by the necessity of unlike antenna energy transfers betwen network I and II that must be in a reciprocal complex amplitude and, respectively, phase relationship. This calls for asymmetrical network terminations at the junctions a and b and the actual presence of the mutual coupling links, M and M in FIG. 12. The undesirable feature of these arrangements, however, is that the network shown in FIG. 9a can only be solved for an in-phase condition, or, an out-of-phase condition between equal signal responses in the 6=0 and 6=+l80 directions for one single frequency inside a given frequency range. Theoretically, to maintain these conditions over the frequency range would require solving the said integro-differential equations for a new set of network parameters, every time the frequency is changed from one value to another. This, of course, is practically impossible. One has to be content with a in-phase solution at the mid-frequency point of the receiver range and permit deviations from the in-phase condition to either side of the frequency range for granted.

It has already been mentioned that one of the two networks I or II must contain a balanced modulator" circuit. While the term itself implies that modulator circuits of this kind must be properly balanced to function properly, i.e., to produce pure sideband frequency signals only, without the carrier frequency signal, it should be noted that the balanced carrier suppression must be automatically maintained over a given frequency range.

If the carrier leaks through, however small, interference noise and thermal noise entering the antenna will modulate the carrier to the extent that the receiver will amplify the noise-modulated carrier and cause instability of the cardioid-null signal. Indications of this null will become erratic as soon as the interference signal level becomes comparable in magnitude with the signal eve Aside from this highly undesirable phenomenon which creates unsteady, visual bearing direction indications, the mathematical analysis of FIG. 12 and the dual circuit arrangements of FIG. 9A shows, that instead of the principal requirement (mm) another more complicated phase relationship among the three phase constants, a, )9 and A, (A being the phase angle of the passive collcctor circuit of network III) must hold, namely,

(LL)a+fi+ t=i =constant over a frequency 13 range, so that the sum of the respective frequency variations of these phase angles must be zero, i.e.,

for all frequencies over the frequency range.

Now, to achieve the principal requirement (mm), the phase constants of the individual system functions K and K respectively, a and B of the active antenna networks I and II must have signs of opposite polarities, i.e., and -,8, or a and +5. Since we are primarily concerned with keeping said 90 phase difference, in the interest of a constant phase alignment, constant, we are to make the frequency variations d(+a)/df and d(fi)/df alike, for, only in this way can we be assured that at every frequency of the frequency range of the tuning dial of the radio direction finder, said 90 difference, and with it, said in-phase condition is sustained.

It is also known that said frequency variations of phase angles with opposite polarities are of the same sign, i.e., for equal, say, positive increments, at, of the operating frequency f, we have to account for finite variations d(+u)/df and d(B)/df that are both positive.

The second part of Equation LL on the other hand shows upon substituting positive variations +d(+a)/df and +d(fi)/df therein, that the frequency variation of the phase angle, 7, must be negative, i.e., equal to -d('y)/df, whereas a resonance-tuned collector circuit would indicate the value d(-y)/df to be equal to zero, since, 'y=0 over the whole frequency range. Or, if we maintain d('y)/df at the zero value, a positive variation +d(+oc)/df would have to be equal in magnitude to a negative variation d(fi)/df. This consideration is believed at the base of the many counter-measures that are proposed in the active antenna networks I and II in the form of phase corrective networks. But such is not possible except at one operating frequency only unless the frequency characteristics are shaped by additional mechanical gauging methods between the active networks. It is seen that the removal of the condition of constraint (ii) by nullifying the mutual coupling links M M (FIG. 12) is obviously a much more efficient and direct approach to the problem of meeting the principal requirement of phase alignment (mm). We are thereby giving the system an additional degree of freedom by which a variation of d(+a)/df and d(-B)/df may take place independently from the variation d('y)/df. The latter may therefore be kept constant.

Summarized, we may briefly state the conditions necessary to fill the functional requirement (I) and principal requirement (mm), respectively, among the three individual generalized coupling pairs as defined above: (see FIG. 12).

21 +Jw=K12 +wJ=0, consistent with M =M =0. K =K consistent with, M = M 2, and, 13,+1w=K23'+jw consistent with provision of uni-lateral circuit elements that are identical and form, in effect, the output terminations of network I and II,

In other Words, making M31=M32 and M12=M21=0, and, since these circuit branches of the coupling structure are included in the general concept of our individual system functions between, respectively, the points of coupling and the current sources, we have now proven the second stipulation in the functional requirement (I), namely, that K =K and, in the opposite directions, 13,+ w= 23,+jw'

Particular stress is laid upon the importance and far reaching significance of the extremely broad character of the above symbolized statement, because it embraces any kind of high frequency input system that we can possibly think of at the moment. The statement conveys the condition when, in direction of maximum signal pick-up of the directional antenna, the antenna energy distribution in 14 the system becomes even, in contrast with the claim made in the U8. Patent 2,142,133, whereby the antenna energy distribution in the system was defined un-even so long as asymmetrical output terminations (Y ,Y were resorted to in the active antenna networks I and II.

Under conditions of coupling, i.e., with the mutual circuit branches M M and M intact, the FIG. 12 shows clearly that the total current 1 flowing through the input admittance (Z =Y is equal to i plus i each of the latter of which is made up of two components, namely, i31=i11+i21 and, i3z=i12+i22- Simple deductions may be derived for asymmetrical terminations Y, and Y If, for a given potential e at output node a the admittance Y,, is small, a correspondingly larger current component i and current component i must result, while for a given potential e at the output node -b, respectively, equal to e a large admittance Y would entail the flow of smaller currents i and i There will be a difference between the currents, i.e., (i -4 flowing around the loop formed with nodes a, b and c, respectively, in the indicated direction of rotation. This is a so called gyrator action which manifests itself in the form ,of an energy sink (circulating current). The magnitude of this energy sink, obviously, can be controlled by the magnitude of the admittance M =M and if we could manage to make it equal to zero, no such circulating current could form at all.

If we for the moment assume, that the current components i and are, at a specified operating frequency f, in phase with each other, we may then compute the potential drops from the node c, respectively nodes a and b. Obviously these are largely governed by the output admittances Y and Y whose absolute values and phase angles, as in the case of conventional high frequency input systems, are, respectively, completely different from each other. For this reason the potentials developed across the nodes a and, respectively, b to the datum junction will differ with respect to both, phase and amplitude despite the fact that we have assumed currents i and i as having, respectively, equal phases and amplitudes. In other words we have dbtained a brute force phase alignment by the sole fact, that the difference between the phase constants, a and ,8, of the two networks I and II can no longer be if they were, such would not agree with the unalterable fact of a constant 90 phase difference between our current sources I and -I (sense antenna and directional antenna). At another operating frequency this in-phase condition between i and i will be immediately upset, of course. We can no longer speak of a sustained phase alignment, since with the foregoing required initial condition, (,Ba)%90 constant, we must conclude that the frequency variations d( +B)/df and d(a)/df are not identical. This means a divergence" from a desired parallelism between the current components i and i as the operating frequency is changed over the tuning range of the collector network III.

The fact that we have mechanically and electrically ganged resonance tuned high frequency circuits in the antenna networks I and II, respectively, with the resonance-tuned collector network III does not alter the above undesirable source for frequency errors to develop in the least. Then, it should be kept in mind that these circuits, in so far as their contribution to a finite phase angle component, respectively, of the whole phase constants, a and B, is concerned do not in any way enter into the consideration of the frequency variations. d(+ 3) /d'f and d(a)/df of the phase constants.

In asymmetrical network terminations the operation of the circuits is, in so far as the directional determinations in the 6=+90 and 6:+270 directions are concerned, de-

pendent only upon the geometrical configuration of the directional antenna. Hence the operation as such is not affected by the above undesirable features, directly so. But indirectly, it will be noted that while the current source I; of the loop antenna may be assumed rendering, say, a potential e at the node a, respectively, in the 6=0 direction, and the current source I, of the sense antenna we assume likewise the said potential e will become equal but negative in polarity in the 6=+180 direction. Instead of a current (i -i flowing around the loop formed with nodes a, b and c, respectively, in the indicated sense of rotation in FIG. 12, we have now a circulating current (i +i flowing in the opposite sense of rotation thereto. This means that the periodical reversal of a receiving characteristic which is not that of a constant cardioid to begin wit-h, becomes untrue. The turn-indicator sensitivity to either side of the symmetry axis of the directional antenna becomes, in other words, asymmetrical at greatly reduced absolute values thereof. This may affect the proper functioning of the loop follow-up mechanism, such as is used in the case of the automatic airborne radio direction finders with a to 360 bearing indication direction to above. (ADFs.)

Having now fully explained the physical significance of the individual system functions, or, generalized, individual coupling pairs, we may use simple diagrams in block form, to briefly summarize the characteristic features of the operation of conventional high frequency input systern. This is done in FIG. A. We imagine said networks placed at the corners of a triangle which may be thought of being a current triangle or a generalized triangular coupling system, both, depending on the view point which is taken. Thus, the sides of the triangle may be considered as generalized coupling path-s, symbolized by the individual system functions, (G ,K (K K L and (K ,K respectively, between the sense antenna network 1, the directional antenna network 2, and the receiver network 3. Considering at first only network 1 active with the others passive, it is easy to visualize a direct current flow i made possible in any manner whatever from network 1 into network 3 over the generalized, individual coupling K and, another, indirect current flow made possible from network 1, over network 2 into network 3, also in any way whatsover, respectively, by way of the two individual system functions, KazHw and KZLHW in series, as symbolized by the product, (K32,K21) This current we shall designate with i (because it flows in direction from network 1 to network 2) so that in the receiver network 3 we have now a resultant current i =i +i both, derived from the voltage source in the sense antenna network 1. It is hardly conceivable that the said two current components are either of equal magnitude or of equal phase. The resultant current i therefore, has a different ph ase from. either of its components i and i Assuming now the directional antenna network 2 active and networks 1 and 3 passive, we can again visualize a direct current flow i made possible in any manner whatsoever from network 2 into network 3 over the generalized, individual coupling, K and, another indirect current flow i made possible from network 2, over network 1 into network 3 in any arbitrary manner (without mentioning specific coupling paths from certain portions of network 2 to certain portions of network 3) so that in the receiver network 3 We have a second resultant current i =i +i both components derived from the voltage source in the directional antenna network 2. Also in this case it is hardly conceivable that the said two current components are either of the same magnitude or the same phase. The resultant current, therefore, has a different phase from that of either of its components i and i Under these arbitrary conditions, however, we can imagine a brute force in-phase condition between said resultant current components i and i Because of the two current triangles of which these form the closing sides, obviously, the direct current components i and i can no longer be assumed to be in an iii-phase condition, if such is stipulated for the resultant currents i and i in the receiver network 3.

Generally then, we have to account for four current components. In high frequency input systems, however, where the one current component, say, i from network 1 into network 2 is suppressed by virtue of, say, a unilateral element, such as a vacuum tube at the output of network 2, while no such suppression takes place with respect to the current flow i in the opposite direction, there are, principally only three current components pres ent in the collector network 3. This represents the much discussed asymmetrical coupling condition common to all high frequency input systems of the Dieckrnann-Hell system, including the most up-to-date systems of currently used automatic airborne radio direction finders. The resultant current i is, therefore, equal to iilzi igz depending on whether the direction of the incident radio wave is to the one or the other side of the directional antennas symmetry axis, respectively, relative to the current i =i whose polarity remains unaltered, of course, for any and all directions of the incident radio wave. Thus, to one side of said symmetry axis, the total current I is equal to: I =i +(-i +i and, on the other side of said symmetry axis it will be: I =i (i +i The phase of (-i +i being different from i it is seen that in no event is it possible to effect a cardoid minimum for which the difference, i (i +i would have to be zero.

To alleviate this very undesirable condition common to all such asymmetrically terminated high frequency input systems we may now show a generalized simple mathematical network, such as shown in FIG. 5B in accordance with this invention.

Herein the common point of coupling resides again in the collector network 3. However, both current components i and i are prevented from forming, thus leaving only the direct flowing current components, i and i in network 3. These flow, respectively, in direction from networks 1 and 2 into network 3, and the re-. spective energy transfers are symbolized by the individual system functions, KaLHw and Kgg d Referred to the common point of coupling residing in network 3 this is, therefore, a parallel-input network type belonging to classification '(A) of single systems, which classification may be extended to embrace inductively coupled circuits, provided, each circuit system 1, 2 and 3 has a common datum junction (common ground connection). There exist several practical circuit solutions bringing about either the elimination of (K ,K between networks 1 and 2, or rendering this generalized coupling pair ineffective with the collaboration of the other two generalized, individual couplings Kalntjw and K respectively, ahead of the actual point of coupling in network 3.

A line of distinction, however, must nevertheless be drawn between inductively and direct coupled systems in regard to keeping the residual coupling, i.e., after removal of the undesirable one (K ,K below a certain critical value. Generally, FIG. 5B shows the residual coupling between networks 1 and-2 to consist of the two series couplings, K ,K respectively, in direction from network 1, over network 3, into network 2, and, K ,K in direction from network 2, over network 3, into network 1.

We can visualize the blocks 1, 2 and 3, in a highly generalized sense, as the high potential nodes of the three networks 1, 2 and 3 to a common datum junction (ground, not shown). We can further assume (see FIG. 12) the individual system functions Kal'Hw and KazHw to contain the individual branches M and M as a result of inverted mutual inductances m and m with the addition, respectively, of capacitive elements, C

17 and C in parallel with the inverted mutual inductances D13 and [1133.

Thus, with further reference to FIG. B, any change of the current source, say, I in network 1 will develop a change in potential across the node 3 and ground because of the impedance Z interposed between the two nodes. This change in potential, in turn, will adversely affect the current source 1 in network 2. Similar considerations hold for changes of the current source, say, I in network 2 in so far as they affect the current source I Because of this residual coupling effect between networks 1 and 2 due to the impedance Z common to both networks at the point of coupling, a true addition of the efiects of the current sources is still limited in its accuracy. Fortunately, in the case of inductively coupled networks, however, the mutual inductance between the input winding and the output windings, respectively, of the network 3, as related to networks 1 and 2 may be made small enough to fall below the value of what is called the critical coupling coefiicients, say, k and k The critical coupling coefiicients are de-' fined by the expressions, k =K ,K /Q ,Q and aa= 2s, az Q22 Qa3 in which m al and zm aa are the absolute amplitude functions of our generalized incoupling pairs, (K13,K31)+ and (K 3,K and Q Q are the Q factors, respectively, of the equivalent two-terminal networks of the antennae networks 1 and 2. Thus, in meeting with the restriction that:

(oo)k ;+k 1, as a result of removing the effect of k by which fact the condition of constraint (ii) is no longer existing, and making, K ,K and, K ,K respectively, smaller than 1 we are able to keep the residual mutual coupling between networks 1 and 2 below the point where it might adversely affect the phase and amplitude of either current (or voltage) source.

In the case of direct coupled networks there are, as we shall see presently, other practical means to keep the residual coupling referred to at an absolute minimum. This may be done, both, before, and after the common point of coupling.

In conformity with the functional requirement (I) and in accordance with this invention, the following specific identity relations" determining the performance of these systems, symbolically represented by FIG. 5B, in accordance with the specified principal requirement (mm) may be written:

A sub-classification which we shall designate as Series Feed Systems in the case of inductively coupled circuits results from the above classification (A) of single systems, if, as in one of the specific embodiments of this invention, wherein a special variably adjustable three winding H.F. mixing transformer is employed for the purpose of rendering the circuit branch M ==M in FIG. 12 non-existent, the actual point of coupling is made to reside, instead of in network 3, in the output termination of either antenna network 1 or 2. The mathematical network is symbolized by FIG. 5C. The case in which the output termination of network 2 forms the common point of coupling is considered. FIG. 5C is best understood with reference to a comparison with FIG. 12 under special conditions in the coupling structure, i.e., with branch circuit M =M between the two junction pairs, bj and cj considered removed. Let us assume that the said removal was accomplished in a similar mannet as in the case of said specific embodiment of this invention where the mutual branch circuit M =M was removed by means of said variably adjustable transformer.

The actual point of coupling is node a in FIG. 12. The current source I, will cause a current i to flow through the output admittance Y respectively, in direction from node a to the datum junction j across which, in turn, a potential e will be developed. The current source I, will cause a current i to flow through the output termination Y from node b to the common node, or, datum junction j, across which, in turn, the potential e is developed. Between the junction pairs b and a-j there will be a voltage drop (e e causing a current i through the mutual branch M in direction from b to a. Resultant currents, i =i i will flow, say, in the 6=0 direction, and i =i +i respectively, in the 6=+180 direction of the incident radio wave, from node a (the common point of coupling) to node 0 of the collector network III. Obviously, for the obtainment of a cardioid the absolute values, fi and [i in said'maxi-, mum signal pick-up directions of the directional antenna must be equal, and, the phase difference between the node potentials 60:, 6 must be made equal to a constant, frequency independent value of provided the antenna networks I and II are designed symmetrically and, if the currents i and i are to remain in an in-phase condi tion over a frequency range. The current i between node b and node c is no longer possible because the branch M =M is'rendered ineffective. Thus, as in the case of FIG. 5B, only two current components are present at the two nodes a and e of the collector network III. Equal absolute values |i l and |i may be had by adjusting the amplification factors of networks I and H relative to each other, while the said required phase shift of 90' between the potentials c and E may be achieved by adequate terminations Y and Y Referring now to both, FIG. 12 and FIG. 5C, a comparison between the just now specified coupling conditions in FIG. 12 and those in FIG. 5C shows that they are, considered from a more general view point, identical. The directional antenna energy transfer from network II into network IH may be symbolized, as shown in FIG. 5C by the series path indicated thereon with respectively, by comparison with the non-directional antenna energy transfer from network I into network III symbolized by the individual system function, K FIG. SC, it should be emphasized, serves the purpose of illustrating a more general concept of the energy flow distribution than would be possible with the aid of FIG. 12.

It will be seen more clearly later in this specification how the elimination of the generalized, individual coupling pair (K ,K is rendered ineffective. Since the admittance termination Y, now represents the common coupling, or, residual coupling between networks I and II defined by the ratios, respectively, Z /Z,,+Z d T,. i w i z.= Y. d 1T... n: are the internal input impedances of networks I and II, respectively, as seen from the node a into said networks, a coupling effect of this kind may be rendered small, for instance, by making the branch admittance M (see FIG. 12) large by comparison with the stray capacitive admittance C between the networks I and II.

In conformity with the functional requirement (I), and, in accordance with thisinvention, the following specific identity relations" determining the performance of these systems in accordance with the principle requirement (mm) may be symbolically expressed:

( /K.= +,w=/K=./.,w. with (-b) net .1.

=/K =0 and the particular provision for inductively coupled networks of this kind, (c) k =k =0, and, (d) k -+k =1, Condition ((1) is the result of a substitution of (c) into the condition of constraint (ii).

In FIG. 3 is shown a variably adjustable H.F. three winding mixing transformer which enables the elimination of any of'the above referred to generalized, individual coupling pairs.

Wound upon a coil former 10, which is hollow in order to receive another coil former a, which is rotatable inside the coil former 10, is a crossed coil 12 and a one half of another coil 13, while a second crossed coil 11 and the other half 13, of the other coil are wound on the rotatable coil former 10a. In the relative position of the coil formers 10 and 10a shown in FIG. 3 the angular displacement between the two crossed coils is, making an angle which is greater than 90, but which can be made less than 90 by merely rotating the coil former 10a inside the coil former 10.

This will be more clearly understood when visualizing the variations of the included angle between the normal directions to the plane surfaces, or cross-sectional areas of the two crossed coils 11 and 12, so that, in their relative positions as shown in FIG. 3, the corresponding fluxes of the crossed coils interlink with each other, The flux of crossed coil 12, with an assumed direction and magnitude, will enter crossed coil 11 from one side, while, when turning coil former 10a relative to the other by a sufiicient amount of angular rotation, it will enter from the other side.

We may follow the process more closely by visualizing, instead of the angle, between areas of the cross section cut out in the cylindrical coil form 10 by the crossed coils 11 and 12 (elliptical configurations), the included angle (not shown) between the normal direction of said cross-sectional areas, which, unlike the angle, in the position of the two crossed coils 11 and 12 as shown in FIG. 3, is smaller than 9.0 degrees. Said normals are indicative of the directions of the electromagnetic fluxes produced by the currents flowing in said crossed coils, so that the included angle, (not shown) between the said normals is a direct measure of the amount of the interlinkage of the said fluxes.

Imagining for the moment the normal to the area of crossed coil 12 held stationary, and rotating coil former 10a in one direction or another, the normal to the area of crossed coil 11 will describe a conical surface with its apex located at the point of intersection of the central axis of coil former 10 with said normal direction as a rotating element. For a given angular displacement of coil former 10a, equal to, A6, a plane surface comprising the two normals in the new position appears, now not only twisted but tilted as well, while the included angle '7", between the normals has increased, say, from a value which was smaller than 90 degrees. In fact, depending upon the magnitude of the angle, '7", in the position of the cross-sectional areas as shown, quite a considerable angular rotation, As, is necessary before the included angle assumes a value of 90 degrees, and, quite an additional angular displacement, As", is necessary to make the included angle, '7', substantially greater than 90 degrees. The mechanical arrangements acts very much in the manner of a micrometer adjustmen of a variation of the electrical coupling effects between the crossed coils 11 and 12.

The outer ends of crossed coils 11 and 12 are connected to terminals A and B, respectively, while the other, inner ends are joined at the terminal G, which is grounded directly as shown in FIG. 3, or which may be grounded over a blocking condenser (not shown). The outer ends of the two coil halves 13 wound, respectively, on coil formers 10 and 10a are connected together as shown, or may lead to a common terminal (not shown).

It will also be noted that the said angular rotation does not in any way alter the direction or the magnitude of the electromagnetic flux interlinking crossed coil 11 with one of the winding halves 13, or crossed 12 with the other winding half 13 because of the inherent 20 symmetrical disposition of the crossed coils 11 and 12 relative to the winding halves 13.

The equivalent tota reactance, which, at some arbitrary operating frequency of the currents flowing through the coil system, might be found to exist, by, for instance, measuring the RF. potential between terminals A and B when feeding with a voltage source across terminals B and G, with a short circuit across terminals A and G and terminals C C left open-circuited, may be thought of consisting of four contributing factors, namely:

(1) the mutual inductance between crossed coils 11 and 12 (comparable, for instance, with the mutual inductance of two primary transformer coils, respectively, connected to the input terminal pairs of an assumed sixterminal network (coupling structure) P" in FIG. 1);

(2) the mutual capacitance between the two crossed coils, (distributed capacitances);

(3) the mutual capacitances between each of the two crossed coils 11 and 12 and, respectively, the two coil halves 13, (distributed capacitances); and,

(4) the equivalent lumped" capacitances, appearing individually and in parallel to the above capacitances (2) and (3), which are due to the amplified chassis efiect" as it appears between the two crossed coils.

For any arbitrary set of conditions (1) to (4) above, there will be found a natural period or frequency of oscillation at which the above inductive and capacitive reactances will neutralise each other, as, when, for instance the negative term in definition (nn) becomes equal to the positive term, while, at the same time, secondary resistances reflected into the parallel mutual circuit inductance and capacitance combination Will limit the electrostatic charging (circulating) current flowing therein. This appears somewhat of a nebulous statement because one is accustomed to visualize a closed circuit when we speak of a resonance condition. Actually, when the phenomenon, better termed anti-resonance" than neutralization is considered as such, it would be observed at the exciting end-terminals of the voltage source between B and G, that energy is being absorbed due to the anti-resonance phenomena to the extent of which reflected secondary resistances make their appearance (these resistances are given in definition (nn) represented by the conductance G The natural cmcillation may therefore be critically damped, resulting in a broadened anti-resonance eflect. It should be realized that at a given operating frequency at which the antiresonance phenomenon occurs, the said reflected secondary resistances may be small to the vanishing point, but, on the other hand, the situation changes rapidly as soon as we connect a load impedance" across the heretofore open-circuited terminals C and C such as a variable tuning condenser to tune the winding halves 13 to resonance. Clearly, under arbitrary conditions (1) to (4) above, the frequency range, as determined by the magnitude of the sum total of inductance of the two winding halves 13 and the difference between the minimum and maximum capacity settings of the said variable tuning condenser (being, for instance, that of the collector network III in FIG. 12, (not shown)), when considered isolated, cannot be the same as in the case when the above numerated reactances (l) to (4) enter into play. In all probability, in other words, the said frequency range will be shifted toward a region of lower frequencies.

But should we design a coil structure as shown in FIG. 3, such that said reactance components neutralize each other, say, by definition (nn) at the mid-frequency point of the frequency range in question, and we are able to maintain said state of neutralization by following up with the essential adjustment between the two crossed coils 11 and 12 in correspondance with the variable tuning condenser settings over an entire tuning range, then the frequency range is shifted bac to its former relative position along the frequency scale, as

21 though no coupling [coupling] with a marked detiining influence were present.

It has been found that the extent of the required relative angular rotation between the coil formers 10 and 10a during said follow-up procedure, as the frequency of the tuning of the collector input network III is changed, depends on the initial magnitudes of capacitive reactances involved, and that with careful design the said initial reactances can be reduced to values which render the required follow-up motion between the two coil formers so small that, in fact, very good results of an average consistent neutralization over the whole frequency range may be had, by merely locking the best adjustment position between the coil formers in place, preferably, at the mid-frequency point of the unaltered frequency range of the collector input circuit III. In this case the coupling structure is said to have a "zero phase shift characteristic.

If we denote the coupling pair between terminals A and B with k =kzn the one between the terminals A and, respectively, terminals C C; of the series connected winding 13, with k =k and the one between the terminal B and, respectively, terminals, C C of the series connected winding 13 with k ==k we may, under the above specified conditions of a sustained neutralization or anti-resonance condition over the whole of the frequency range, express this fact by saying that the generalized individual coupling pair (K, ,K has been rendered ineffective by virtue of the condition, k =k =0. Moreover, the ordinary coupling coefilcient pair, k =k, has been made to disappear independently from the coupling coefllcient pairs, k =k and k =k This is a positive criterion for having dispensed with the condition of constraint (ii). Careful attention should be given to the design, of course, of the electrical parameters and physical arrangement of the coupling transformer, in order to make the response curve of the voltage measured, say, between the terminals A and G, while the terminals B and G are fed with a high frequency voltage source, "run Hat with the frequency variation of said applied voltage source, having but negligible amplitudes over a given frequency range, respectively, at the most suitable setting between crossed coils 11 and 12 (locked in position during the measurement). Having thus obtained a satisfactory response curve with the load circuit of collector circuit III open, the latter will, upon being closed, track precisely with the remaining high frequency resonance-tuned circuits in the active networks I and II, provided the coupling values k and k are equal, respectively, with a magnitude which is equal, or, below to the critical coupling value. The symmetrical construction of the transformer will automatically bring about this required condition.

FIG. 4 shows the coupling transformer of FIG. 3 incorporated in a direction finder comprising a loop antenna 2 and a non-directional sense antenna 1. The coupling system is of the type represented by FIG. Sb. The sense antenna 1 is connected through an output stage Z" to the control grid of a pentode amplifier P, the output of which is fed to the coil 11 of the transformer. The loop antenna 2, which may consist of only a few turns, is connected throughseveral input stages 2"; and

to the control grid of a pentode amplifier P,., the output of which is coupled to the coil 12 of the transformer. One end of each of coils 11 and12 is connected to ground. The coupling circuit Z"',, for the loop antenna 2 preferably includes a transmission line as indicated in FIG. 8. Because the loop antenna inductance is small compared to the capacity of the transmission line in parallel thereto, the antenna input circuits can be readily adjusted so that the phases of the potentials applied to the control grids of pentodes P, and P, are the same. The terminals C and C, of coil 13 are connected over a large inductance U and a small trimmer condenser C'ss in series thereto, to the collector circuit (L C Thus the coupling circuit (L' C'ss) may be tuned to resonance at some intermediate point of the frequency band in use while the collector circuit (L C is kept continuously at resonance as the incoming signal frequency is varied. Due to the very high parallel impedance (L Q the resonance curve of the coupling circuit (L' C3 is exceedingly broad and the resulting small current amplitude flowing therein reacts only feebly with the currents flowing in the remaining coils of the coupling transformer. The voltage across the collector circuit C 1 is impressed on the control of pentode tube P the output of which is fed to the receiver. Coils 11 and 12 are adjusted relative to each other so that there is no coupling between the points A and B. The tuning of the various tuning condensers is, of course, ganged.

FIG. 6 shows still another form of the invention of the type represented in FIG. 5C. Herein one of the cross coils 12 is connected to the loop antenna circuit 2, the non-directional antenna 1 being connected to the two split windings 13-43. The connections from the antennas to the transformer are made through the pentode tubes P, and P The output of one of the amplifier stages, for example P is tuned to resonance by a condenser 63. The terminal B of the transformer is connected through a variable condenser C' and an inductor U to the control grid of a pentode tube 30. The output circuit of pentode tube 30 includes a resonant circuit consisting of a condenser C and inductance L A receiver 3 is connected across the resonant circuit. Anode potential is supplied to the loop antenna output stage P through the cross-coil 12, a suitable blocking capacitor being provided for the purpose of maintaining the common junction point of the two cross-coils 11 and 12 effectively at ground potential. The circuit C L is tuned to resonance. The cross-coils 11 and \12 are adjusted so as to neutralize the capacitive reactance 28 therebetween. Gang tuning may be provided as is conventional.

FIG. 7 shows an example of the H.F. mixing circuits presently in use in conventional radio direction finders. For reasons of simplicity the usual chain of coupling stages has been omitted. FIG. 7, in other words, represents schematically only one half of the usual balanced modulator circuit used in the output stage of the loop antenna network. Therefore, with reference to FIG. 7. P is a vacuum tube, usually consisting of a triode, but a pentode may be used to advantage. A three winding H.F. mixing transformer comprises three coupling coils, L L and L wound upon an iron core (not shown). In following the circuit parameters around the closed plate circuit of tube P there is encountered the plate resistance 1 between cathode K and plate electrode P,

' the coupling coil L with the reflected impedances of the receiver and sense antenna network terminations and the blocking condenser C A DC. source, marked furnishes the plate supply for P through the low potential end of coil L being properly filtered by the resistorcapacitance combination of R, and C to ground. Since no effort is made to eliminate the coupling coefiicient +K respectively, in direction from coil L to coil L a current component, i from the loop network flows directly into collector circuit 3, which fact makes the ratio, +K32.K21/+K31 a finite quantity. For this reason, the required phase condition (mm) does not remain constant over a frequency range.

FIG. 8 shows another circuit according to the invention wherein a sense antenna 1 is connected over a transmission line and a series inductance 123 to a grid resistor 122 and thus impresses received signals on the control grid of the pentode amplifier P Suitable potentials are applied to the screen grid through the resistor 117 and to the anode over the resistors 114 and 116.

'The'loop antenna 2 is connected by a transmission line 124 to primary coil 101, which is coupled to secondary 23 coil 102. Coil 102 and condenser 103 are tuned to resonance with the incoming signal. The potential across the resonant circuit is applied to the control grid of a pentode amplifier P,. The output of the pentode amplifier is connected to a point 100 forming a junction between condensers 74 and 75. Condensers 74 and 75 are connected to the control grids of a pair of pentode tubes 98 and 99. The control grids of these pentode tubes are also fed by a modulation oscillator B, through a pair of resistors 68 and 69 and a pair of blocking condensers 70 and 71. The output of the modulation oscillator E, is impressed across a pair of grid leads 72 and 73.. in the input circuit of pentode tubes 98, 99. The latter, in connection with the modulator-oscillator E and the coil assembly consisting of coils 12a, 12b, L and U constitute a balanced modulator circuit in which the generalized coupling pairs (K +K and (-K;;, +K respectively, between coil U and the coils 12a and 12b are neutralized, or, otherwise rendered inefiective while the generalized feedback couplings, K+ and +K respectively, between coil L and each of the coils 12a and 12b are vanishingly small by comparison with the respective forward" couplings, +K and +K The outputs of pentode tubes 98 and 99 are fed from the anodes of the tubes to two sections 12a and 12b of coil 12. The two sections of coil 12 have a common terminal which is connected to ground through condensers 83 and 84. The output of peutode P is connected across condenser 137 and coil L which are tuned to exact resonance. The other terminal of coil L is grounded. The output of the transformer is taken from coil U which has one terminal grounded and the other terminal connected through a condenser C' to a parallel resonant circuit C L forming the input of a suitable receiver. FIG. shows the reduction to practice of the said coil set in symbolic form connected to the balanced modulator 98, 99, of FIG. 8. A modified special transformer embodying the same principles as that of the transformer of FIG. 3 is used; the crossed coil 12 of the latter is merely split into two sections, 12a and 12b, the outer ends of which connect to the modulator tubes 98 and 99, and the inner ends of which form a junction which is grounded over the filter capacitors 85 and 86.

FIG. 10 shows circuit connections similar those of FIG. 8 between a blanced modulator 98, 99 and a receiver. The split coil 12 has sections 12a and 12b which are connected to the anodes of tubes 98 and 99. The coil 13 is connected to the ouput of pentode tube P,, as indicated in FIG. 8. Coil 11 of the transformer is connected to the control grid of a tube 146, preferably of the pentode type, through an inductance L A tunable condenser C is connected between the control grid and ground.

Referring now to FIG. 2A, there is shown a vector diagram for the equivalent circuit of the loop antenna, the primary current i flowing through coil 101 in FIG. 8, which is seen to be of a purely capacitive nature; because of the short loop feeder cable length, the circuit loss angle is negligible. The current i leads the induced loop E.M.F. E; by fully +90. The natural period of the equivalent circuit P is far above the upper end, r...- of the frequency range, since the loop consists of a few turns of extremely small diameter only. The parallel circuit combination becomes, therefore, equivalent to a capacitance.

Since the secondary circuit 102 and 103 (FIG. 8) is resonance tuned, its circulating current lags 90 behind the primary current i and the voltage across the variable tuning condenser 103 is equal to secondary circuit and w the period of oscillation in radians. Hence, E', lags 90 behind the loop E.M.F.

E and ought to, by this reason, fall into phase alignment with the induced sense antenna E .Referring now to FIG. 28, there is shown an equivalent circuit diagram of the sense antenna input circuit of FIG. 8. Due to considerable sense antenna feeder cable losses the sense antenna current i is not of the nature of a purely capacitive current but the phase P with respect to E is (90-Azp)= p (Scc FIG. 2C). The current i splits up into two components; a purely capacitive component i flowing across the cable capacity C and a component i flowing through the series connection of the R.F. coil 123 and the grid resistor 122.

In FIG. 2C, which represents a vector diagram of the events taking place in the equivalent circuit diagram of FIG. 2B, the current i is shown to have a leading phase with respect to the induced E.M.F. E being composed of the components i and i i being at a leading angle and i at a zero angle with E Without the R.F. inductance 123 the voltage drop across the terminals a and b in FIG. 2B would be 90 lagging with respect to the current i so that the voltage found across the grid resistor 122 would be of an identical phase. The same phase discrepancy, A(p appears again between the output voltage E' and the induced E It is possible to compensate for the said phase discrepancy, so that the output voltage E"' at the control grid of the pentode amplifier P falls exactly into phase alignment with the induced sense antenna E.M.F. E and,- of course, with the output voltage E', at the control grid of pentode P,. Then, in the final analysis this is precisely the result which 'we are after. Thus, the series combination of R.F. coil 123 and grid resistor 122 connected in parallel with the cable capacity C., respectively, between points a and b of FIG. 2B brings about the desired result as follows:

Since the amplitude and the phase of the very small current i are mainly governed by the generally high resistance 122, it will be practically at an angle of 90 with respect to the charging current i flowing across the cable. The voltage E across [points] points a and b will divide up into two components: i R and I jwL. Provided that we have selected the proper value of inductance for 123 (between 20 and 50 microhenries) the voltage drop across the grid resistor, which now has become the output voltage E can be made to fall in phase with ,the induced E.M.F., E and, 'with it, the output voltage 1'3 of the loop input circuit in FIG. 8.

With voltages E and E fed, respectively, in phase to the pentode amplifying tubes P and P,,, we may construct the respective output circuits of these tubes so that their no-load voltages at the plate electrodes are still in phase with each other. The input circuits, respectively, of the modulator tubes 98 and 99 represent, practically, resistive loads, consisting of grid resistors 72 and 73, since the series capacitors 75 and 72 have a negligible'reactance in comparison therewith. The plate currents flowing in the output circuit of said pentodes have for all practical purposes, the same phase a the respective control grid input voltages at terminals 76 and 77, because of the small inductance values assigned the two crossed coil halves 12a and 12b. From this it follows that said output currents flowing through the crossed coil halves have the identical phase as the phase of the voltage at the plate terminal of the loop preamplifying tube P,.

Since the plate load of the sense antenna pro-amplifying tube P, is equal to a resistance, i.e., the parallel impedance of the resonance tuned plate circuit comprising the coil halves 13 as load ind-uctances, the current flowing through the. parallel circuit is of the same phase as the current flowing through said crossed coil halves 12a and 12b. But between the circulating current flowing in the resonance tuned output circuit of tube P and said current flowing through the crossed coil halves 12a and 12b there exists a phase difference of precisely 90 25 which satisfies the phase relationship: fl=a+90 and corresponds, actually, very closely with the specific dual set of values: S=90, uz=), the constant 90 phase shift occurring between the respective output currents of the sense and loop antenna network in the special transformer of FIGS. 8 and 10.

Referring now again to FIG. 12 which, as we know, is derived by way of forming a mathematical dual of the inductive type of coupling structure such as indicated in FIG. 1, we would, if replacing the active networks I and II and the whole coupling structure with branch circuits M M and M with two resistors, say, R and R, with one end each joined together at the node c while the other two ends are connected individually to the current sources I and, respectively, I and, furthermore, if replacing the impedance Z between node c and the datum junction j with another resistor R obtain a network generally known in electronic computer engineering as a parallel-input adding network.

The assumption goes, in accordance with literature on this subject, that the current I, found at the common point of coupling c will then be equal to: I =a.I +b.I in which a and b are called the weighting factors."

In practice, however, the above linear equation between 1 and I is subjected to a frequency error because of a mutual coupling effect in the form of stray capacitances between the current sources 1;, I; that may exist, and because of the common resistance value R relative to the sums (R +R and, respectively, (R +R The purpose of these arrangements is to add a particular wave shape of the current source I to the current 1;, which may, or may not, be different from that of I in any event the idea consists of having a true addition" of the two wave shapes at the common point of coupling c.

Instead of interposing between the common point of coupling and a receiver input circuit a coupling structure such as exemplified by the circuit diagrams of FIGS. 4, 6, 8 and 10, computer engineering practices employ coupling structures of a different kind to overcome mutual coupling effects (interaction) between the voltage or current sources which are to be add in linear fashion.

One classification of computer circuits employs vacuum tube circuits with a, respectively, balanced input" using the effects of negative feedback either of the cathode or plate follower types and have the one feature in common with the circuits described so far, basically, that the correction for frequency errors (distortion of the individual wave shapes of the current or voltage sources to be added) is performed ahead of the point of coupling to the evalution networks. We have given these circuit arrangements the classification (A) of single systems because they are reducible, basically, to the network configuration shown in FIG. 12, and, in particular to the mathematical equivalent of FIG. B.

Another classification of computer circuits employs a single negative feedback amplifier of the plate follower type in which the correction for frequency errors during the addition of current and voltages sources of different or like wave shapes is performed after the point of coupling. This type of circuit arrangement is given the classification (B). Basically, they are, insofar asthe required parameter conditions of the input circuits, respectively, between the common point of coupling and said voltage or current sources are concerned, again reducible to the network configuration of FIG. 12, and to the mathematical equivalent of FIG. 5B.

A third classification of computer circuits employs orthogonal inputs of the voltage or current sources to be added, respectively, in bridge adding networks, in particular, of the equal arm bridge variety. The obvious intent here is to avoid, by the orthogonal method of coupling the voltage or current sources, a mutual undesirable coupling, respectively, between said source's, to

26 begin with, and this, without the use of negative feedback action above referred to. This classification of circuit arrangement is given the designation (C). Basically, the intent of rendering ineffective the generalized, individual coupling pair between the said sources is identical to that symbolized in the mathematical equivalent shown in FIG. 5B.

The junction network of FIG. 12 (in distinction to the loop network of FIG. 9A) can be used to advantage, however, to eliminate frequency errors without the use of negative feedback while, basically, the mathematical equivalent form is again that of FIG. 5B. A basic circuit diagram of this kind will now be described and reference is made to FIG. 13, which shows a variable resistor R (potentiometer) connected orthogonally to the midpoint of a resonance-tuned circuit consisting of a coil with inductance L and a differential condenser" C which is balanced to ground. Thus the capacity of said condenser C is represented by the branch M =M plus such stray capacitances C that may be found in the receiver chassis in parallel therewith (see FIG. 12). The branches M =M and M =M both, are represented by half the inductance, L/2, of the said inductance L, while R is thought of being represented by the circuit branch with impedance Z respectively, between node c and the datum junction j in FIG. 12. The nodes a and b connect, of course, to the internal output admittances Y and Y respectively, and it is clear that the sum of the voltage drops (e -e between nodes a and c, plus e between the node c and ground g, is equal to the output voltage c In distinction to the operation of the parallel feed system shown in FIG. 4 in which the coupling between nodes corresponding with those in FIG. 13, i.e., nodes a and b is eliminated by way of neutralization of such lumped chassis capacities that might appear between said nodes with an appropriately adjusted mutual inductance, the residual coupling consisting of the remaining two critically adjusted couplings, respectively, in series with the collector network III (synonymous with the product (K ,K in FIG. 5B) is also rendered ineffective by the circuit arrangement of FIG. 13.

The operation of the coupling structure is as follows: a current 'I,,, entering the resonance-tuned circuit L, C at the node a, must cause voltage drops, respectively, between nodes b and c, and between node c and ground g. These voltage drops are made exactly equal and phaseopposed, so that the voltage e between nodes b and g is equal to Zero. At the same time, if a current, -I,,, would be entering said circuit L, C at the node b, no voltage would be developed between nodes a and g. This operation is, however, true to fact, if in each instance we are entitled to assume a no loa condition, respectively, between nodes b and g with I active, and between nodes a and g with 1,, active; i.e., theoretically, the output admittances Y,, and Y should be equal to zero.

With these assumptions, therefore, the mutual antennae energy transfer from one active network, say, connected at the node a to another active network connected at node b is made effectively zero, or, we may say that the respective individual system functions (K ,,K between the antennae networks I and II fulfill the functional requirement (I) with respect to its first stipulation, and this, at each resonance setting of circuit L, C provided, as we shall see shortly, that the variable resistor R is adjusted at each operating frequency to a value of A of the parallel resonance impedance:

1 l nb Z' o- -Qo in which 1 W 2.1r.f and

Images