US 3452300 A
Abstract available in
Claims available in
Description (OCR text may contain errors)
June 24, 1969 J. D. CAPPUCCI ET AL 3,452,300
I FOUR PORT DIRECTIVE COUPLER HAVING ELECTRICAL SYMMETRY WITH RESPECT TO BOTH AXES Filed Aug. 11, 1965 Sheet of 2 F 1 FIG. 2A
4 --T" 0 EVEN MODE V; I/ZV. I i /2V 2 iz' l 3 1 J3 OPEN cm. ---|2 o.c.
I5 3 ooo MODE I 2| 3 I 3 W SHORT cm. 2 so. 23
June 24, 1969 J. 0. CAPPUCCI ET AL 3,452,300 FOUR PORT DIRECTIVE COUPLER HAVING ELECTRICAL SYMMETRY WITH RESPECT TO BOTH AXES Sheet Filed Aug. 11. 1965 FIG. 8A
EVEN MODE BISECTION ODD MODE BISECTION FIG. 9B
EVEN MODE EQUIVALENT ODD MODE EQUIVALENT INVENTORS JOSEPH D. CAPPUCCI HAROLD SEIDEL d 5 ATTO/RNEYS United States Patent 3,452,300 FOUR PORT DIRECTIVE COUPLER HAVING ELECTRICAL SYMMETRY WITH RESPECT TO BOTH AXES Joseph D. Cappucci, Wayne, and Harold Seidel, P.O. Martinsville, Warren Township, N.J., assignors to Merrimac Research and Development, Inc., Irviugton, N.J., a corporation of New York Filed Aug. 11, 1965, Ser. No. 478,930 Int. Cl. H01p /14; H03h 1/00 US. Cl. 333- 12 Claims ABSTRACT OF THE DISCLOSURE A symmetrical four port directive coupler for controlled operation over a wide band of frequencies in which two networks formed by lumped constant parameters are connected together by an electrical coupling means. The parameters of the two networks and the electrical coupling means are selected so that the normalized input impedance of the even mode bisection of the entire symmetrical directive coupler is equal to the nor malized input admittance of its odd mode bisection.
This invention relates to devices for coupling radio frequency energy and more particularly to couplers constructed in accordance with imposed conditions of duality.
An object is to provide coupling devices formed by symmetrical networks using imposed conditions of duality.
An additional object is to provide coupling devices formed by symmetrical networks and having desired coupling properties which are constructed using imposed conditions of duality in which the normalized input impedance of the even mode equivalent circuit bisection is equal to the normalized admittance of the odd mode equivalent circuit bisection.
'Another object is to provide devices for coupling radio frequency energy using lumped circuit components in which the values of the components at a particular frequency of operation are selected in accordance with imposed condition of duality.
Yet a further object is to provide radio frequency energy coupling devices having a plurality of obstacles using lumped constant components for both the elements of the coupler obstacles and for connecting the plurality of obstacles together.
An additional object is to provide radio frequency coupling devices having a plurality of obstacles and using lumped constant components for both the elements of the coupler obstacles and for connecting the obstacles together, such elements being selected in accordance with imposed conditions of duality, and means for correcting for dispersion of the devices.
In accordance with the present invention, radio frequency energy coupling networks are provided which are relatively simple to construct and have desired isolation, input match, coupled output, energy transmission and frequency responsive characteristics. These couplers are constructed as symmetric networks from lumped constant components whose values are selected in accordance with conditions of duality imposed upon the network. In the preferred embodiment of the invention the duality condition is that the input impedance of the even mode bisection of the network equals the input admittance of the odd mode bisection.
Other objects and advantages of the present invention will become more apparent upon reference to the following specification and annexed drawings in which: FIG. 1 shows a four terminal symmetric coupler network; FIGS. 2A and 2B, show the even and odd mode bisections of the network of FIG. 1; FIG. 3 is a schematic diagram of a four terminal symmetric coupler network formed of lumped constant components; FIG. 4 is a schematic diagram of the bisection of the network of FIG. 3; FIGS. 5A and 5B respectively are the odd mode and even mode equivalent circuits for the bisection of FIG. 4; FIG. 6 is a schematic diagram of a two obstacle coupler network; FIG. 7 is a schematic diagram of a two obstacle coupler using lumped constant components; FIGS. 8A and 8B respectively show the even and odd mode bisections of the coupler of FIG. 7; FIGS. 9A and 98 respectively show the equivalent transmission line type circuits for the bisections of FIGS. 8A and 8B; and FIG. 10 is a schematic diagram of a coupler of the type shown in FIG. 7 with additional elements provided for correcting dispersion.
FIGURE 1 shows in general block notation a symmetric four terminal network 10 of the type to be considered having components (not shown) which are frequency responsive to change their impedances and admittances. The network has four ports 1, 2, 3 and 4, and is of the type such that when an input signal is applied to port 1, a coupled output signal is produced at port 2, a transmitted output signal produced at port 3 in phase quadrature with the coupled output at port 2, and port 4 is isolated so that no output signal appears thereon. Many such networks are well known in the art.
Since network 10 is symmetrical, in accordance with theory of analysis it is said to have a plane of symmetry 12 so that a voltage V applied between input ports 1 and 2, with the other ports 3 and 4 terminated with the proper impedances, can be broken up into the two equivalent modes of excitation applied between ports 1 and 2. These two modes are, as shown in FIG. 1, an even mode excitation of /2V and /2V applied to the respective ports 1 and 2 and an odd mode excitation of /zV and /2V applied to the same two ports. The even and odd mode analysis is shown for example in an article by S. B. Cohn entitled Shielded Coupled-Strip Transmission Line in the October 1955 Transactions IRE, Volume PGMTI.
Bisecting the symmetrical network 10 along its plane of symmetry 12 produces the two bisected even mode and odd mode networks shown respectively in FIGS. 2A and 28. Reference numeral 15 indicates the characteristic impedance termination for port 3. Using the even and odd mode excitations of FIG. 1 for FIGS. 2A and 28 respectively, it can be seen that in FIG. 2A plane 12 is effectively an open circuit plane since excitations of V2V and /zV applied to terminals 1 and 2 produce no difference of potential therebetween and therefore no current flow between the two input ports 1 and 2. When the odd mode excitation of /fiV and -V2V is applied to terminals 1 and 2, plane 12 is considered to be short circuited since 81' difference of potential V exists between ports 1 and 2. This is illustrated by the odd mode network bisection network of FIG. 2B.
Letting Z be the normalized input impedance for the even mode bisection of FIG. 2A, where the normalized input impedance equals the input impedance of the network at any one frequency divided by the characteristic input impedance of the network, it can be shown by network analysis that:
where S is the reflected signal at port 1 and S is the signal transmitted between ports 1 and 3. I is the reflection coefficient of the even mode bisection at port 1 of the S term and [1I is the transfer coeflicient between ports 1 and 3 of the S term. The even mode bisection gives the same results for ports 2 and 4, so that:
where S and S correspond to Equations 1 and 2 for the even mode bisection of FIG. 2A. Since the odd mode bisection anticipates /2V at terminal 2 of the symmetrical network, then:
By imposing an condition for duality on the network 10 such that Z =Y then from Equations 1 and 5.
meaning that the reflection coeflicient of the even mode bisection is equal to the negative of the reflection coeflicient of the odd mode bisection. By using (9), the scattering coefiicients for the odd mode bisection in terms of I become:
To obtain the scattering coefficients for the entire four terminal network 10 of FIG. 1, Equations 1 through 4 and 5a through 8a are added, giving:
Equations 10 through 13 define the symmetric network 10 of FIG. 1 as a directional coupler having the following characteristics:
(a) Isolation-since S there is no signal transmission between ports 1 and 4.
(b) Input matchsince :0 there is no mismatch at the port 1 input.
(c) Coupled output of 1,,S =VI defines the coupling between ports 1 and 2,
(d) Transmission of [l-I S V[ 1 I 2] defines the transmission between ports 1 and 3.
Additionally, due to the symmetry of the network about a second bisecting plane 13, which is transverse to plane 12, the coupled output at port 2 can be shown to be in phase quadrature with the transmitted output at port 3.
All of the above desired characteristics are produced in the couplers of the present invention using imposed duality conditions with lumped circuit elements, such as .4 capacitors and inductances. By suitable connection and selection of the values of the elements, couplers of a relatively simple form are produced having desired coupling properties over a range of frequencies.
The principles of the present invention are discussed with respect to the coupler 20 using lumped constant elements shown in FIG. 3, which is called a single obstacle type coupler. Coupler 20 is a symmetrical network for-med by a lumped inductor 21 whose two coils respectively have their ends connected to ports 1 and 3 and ports 2 and 4. Ports 1 and 2 are shunted by a lumped capacitor 23 of capacitance C.
The wires of the two coils forming inductance 21 are preferably twisted together and both wires are wound on a coil form or a toroidal core to form a bifilar winding.
In any bifilar inductor, such as inductor 21, the odd, or anti-symmetric, mode inductance is less than the even, or symmetric, mode inductance. This is so because in the odd mode most of the electromagnetic field is contained between the wires while the even, or symmetric mode inductance remains large. A bifilar inductor can be wound to have a considerable difference between the even mode and odd mode inductances.
The odd mode inductance L of a bifilar inductor is measured as a series connection of the pair of wires, each having an inductance L forming the inductor. Since the wires each of inductance L are connected in series, the total odd mode inductance L equals /2L,,,. The even mode inductance L is measured as a shunt connection of two wires each having an inductance L so that L =2L Since network 20 of FIG. 3 is symmetrical, it can be bisected along its plane of symmetry 12 as in FIGS. 2A and 28 to give FIG. 4. The capacitance of capacitor 23a is 2C since half of the total capacitance C of capacitor 23 was taken and capacitors in series are added by adding their reciprocals. Resistor 25 designates the terminating impedance of the network, and is shown as having a value of one (1) unit. The value L is shown for the inductance of the one wire of bifilar inductor 21 assumed to be in the half of the network encompassed by the bisecting plane 12.
The odd mode equivalent circuit for the bisected network of FIG. 4 is shown in FIG. 5A. As explained with respect to FIG. 2B, bisecting plane 12 is considered to be a short circuit plane in the odd mode meaning that the capacitor 23a of value 20 is in parallel with the output impedance 25. The inductance of the inductor 21 is neglected since it is very small in the odd mode, as explained above. FIG. SB shows the even mode equivalent circuit for the bisected circuit of FIG. 4. Since bisecting plane 12 is an open circuit plane for the even mode, the capacitor 23a has no effect and the inductance of the coil 21 is the high value even mode inductance L For the odd mode equivalent circuit of FIG. 5A the normalized input admittance is:
Z designates the characteristic impedance of the circuit which is multiplied with the imaginary part of (14) rather than divided since admittance is the reciprocal of impedance. When terminated by its characteristic impedance Z the network 20 has an admittance looking into the circuit of 1/2 Using Equation 5 the reflection coefficient I of the circuit of FIG. 5A is:
and the scattering coefiicient S is:
Similarly, for the even mode equivalent circuit of FIG.
5B, the normalized input impedance is:
(17) jwL Z 1 To Here, the imaginary part of (17) is divided by Z to obtain the normalized impedance.
Since L =2L in the even mode, then:
w2L, Z;,, 1+] -1l- 7:c I Using Equation 1, the reflection coefllcient I of the circuit of FIG. B is:
2 jx and the scattering coefficient S e is: i
Imposing the duality condition on the circuit of FIG. 3 that its normalized input impedance for the even mode bisection equals its normalized input admittance for the even mode bisection gives from Equations 14 and 17:
and, from Equations '14 and 17a (21a) Using Equations 1 through 8 and 5a through 8a to obtain the even and odd mode scattering coefiicients for the network of FIGS. 3-5 in terms of the even mode coefiicient and adding the even and odd coefficients in accordance with Equations 10 through 13 to obtain the scattering coeflicients for the entire symmetrical four-port network gives:
The coupling k in db between ports 1 and 2 of the coupler of FIG. 3 is given as:
=10 log (1+?) This completely defines the coupler. Given the coupling value k necessary, Equation 26 is used to solve for x which, from Equation 21a, is equal to b. For any predetermined frequency of operation and input impedance level the value of the inductance 21 and capacitor 25 can be determined.
For example, consider that a 3 db coupler is to be constructed. Using 3 db as the value for k in (26) and solving for X gives X =b=2. Since, from (20) 6 and b=Z w2C=2 then and l Zo For an operating center frequency f of 30 mo. (w=21rf) and a characteristic impedance Z of 50 ohms, L =.265.3 p.11 and C=l06 ,u,uf. L gives the value of each coil of the bifilar inductor 21.
While the single obstacle coupler described above is quite useful and has many applications, it has a relatively narrow bandwith, like any other frequency responsive circuit with only a few components. To provide greater bandwidth, a two obstacle coupler is used. This coupler 30 is shown in FIG. 6. Here, two symmetrical single obstacle networks 20-1 and 20-2, of the same construction as network 20 of FIG. 3, are connected together by a coupling element 32 Network 20-1 has a capacitor 23-1 (C shunted across the ports 1 and 2 and a bifilar inductor 21-1 connected in the same manner as network 20 of FIG. 3. Network 20-2 is the same as 20-1 and has a bifilar inductor 21-2 and a capacitor 23-2, the latter of which is connected in shunt across output ports 3 and 4 and to the ends of the two wires of inductor 21-2.
The ideal elements 32 for coupling networks 20-1 and 20-2 together are, of course, transmission lines of the proper length and characteristics for matching the characteristic impedances of the two networks, so that the imposed duality condition -is not destroyed. A pair of such lines would be used, one connecting each of the corresponding coils of the two bifilar inductors 21-1 and 21-2. If the impedances of element 32 are non-dispersive, that is, have substantially linear phase shift and have no attenuation over the operating frequency range, then designing the two obstacle coupler 30 is a straight forward aplication of the design principles for the single obstacle coupler described above. At some frequencies, particularly higher frequencies, sort lengths of transmission lines possess substantially non-despersive properties, and are an ideal choice for elements 32. Such short line lengths would not adversely affect operation of the coupler 30 and the duality design criterion of Y =Z can be achieved.
At intermediate frequencies (I.F.), such as in the range from 1 to mc., the lengths of the two transmission line coupling elements 32 between networks 20-1 and 20-2 needed to preserve the duality condition would be fairly long and would make the coupler considerably bulky. To keep the coupler 30 compact, lumped constant components are preferably used for the coupling elements 32. At these lower I.F. frequencies the transmission lines have dispersive properties. Therefore, the coupler must also have the same dispersive properties at a particular frequency of operation as the equivalent lengths of transmission lines, in order to maintain the imposed duality conditions to achieve the desired characteristic for the coupler. Such a network is illustrated in FIG. 7.
FIG. 7 illustrates a symmetrical two obstacle coupler 40 with a coupling network formed by lumped constant components in the form of inductors and capacitors. The same reference numerals used previously are used here again for the same components. The lumped constant component equivalent for each connecting transmission line of FIG. 6 is shown as a T-circuit having two inductors 41 and 42 and a shunt capacitor 43. The upper T-circuit has series connected inductors 41-1, 42-1 and a shunt capacitor 43-1, whose lower end is connected to a point of reference potential such as ground 44, connecting the ends of the upper wires of bifilar inductors 21-1 and 21-2. Inductances 41-1 and 42-1 are single coil inductors. A similar T section formed by single coil series connected inductors 41-2, 42-2 and shunt capacitor 43-2 connects the ends of the lower wires of bifilar inductors 21-1 and 21-2. Capacitors 23-1 and 23-2 each have the same value C while capacitors 43-1 and 43-2 each have a value 2C. The single coil inductors 41-1, 41-2, 42-1 and 42-2 each have a value L FIGS. 8A and 8B respectively show the even and odd mode bisections of the coupler 40 of FIG. 7. As in FIGS. 3 and 4, the even mode inductance L of each bifilar inductor 21-1 and 21-2 is 2 L The odd mode inductance L,, for each of 21-1 and 21-2 is L /2L,,. The inductances of coupling coils 41 and 42 in each mode is L while the capacitance of capacitor 43 is 2C. Since the even mode bisection plane 12 is an open circuit plane, capacitors 23-1 and 23-2 do not appear in the even mode bisection of FIG. 8A. In the odd mode bisection of FIG. 8B, where plane 12 is a short circuit plane, capacitors 23-1 and 23-2 have a value of 2C. It should again be pointed out that the even and odd mode bisections are circuits with identical properties, although the circuits are shown in somewhat different form, since network 40 is symmetrical.
FIGS. 9A and 9B respectively show the even and odd mode equivalent circuits of FIGS. 8A and 8B. Referring first to FIG. 9A, for purposes of analysis, the even mode equivalent circuit has been redefined as an equivalent transmission line T-circuit with a pair of series connected inductors 44-1 and 45-1 of values L and L respectively, shunt capacitor 43-1 of value 2C connected to ground, and a second pair of series connected inductors 45-2 and 44-2 of values L and L connected to the junction of inductor 41-1 and capacitor 43-1. Here, the two inductors 45-1 and 45-2 of value L give the even mode inductance of the equivalent of T-section transmission line. The value L also includes the inductance of the bifilar transformers 21, which is not negligible in the case where two networks are connected together. The two inductors 44-1 and 44-2 of value of L are the coupling obstacles (inductances) of the equivalent transmission line. From FIGS 8A and 9A it is clear that:
The odd mode bisection of FIG. 8B is also shown in FIG. 9B for purposes of analysis as an equivalent transmission line T section. This section is formed by series connected coils 41-1 and 46-2 of value L connected at their junction by capacitor 43-1 of value 2C to ground. The unconnected ends of coils 46 are shunted to ground by the capcitors 23-1 and 23-2 of value 2C Here, capacitors 23-1 and 23-2 are the coupling obstacles of the equivalent line T sections and:
The even and odd mode bisections of FIGS. 8A and 8B as well as their corresponding equivalent circuits of FIGS. 9A and 9B are uncoupled to each other, due to the symmetry of original network 40 and the nature of the bisecting planes 12. These bisected circuits of FIGS. 8A and 8B are also identical, so that the even and odd mode T-section transmission line equivalents of FIGS. 9A and 9B are also identical so that:
Here, coils 41 and capacitors 23 (L and C are to be the coupling obstacles which are to be dualized with re spect to each other and with respect to the transmission line T-section equivalent circuits of FIGS. 9A and 9B to impose the condition of Z =Y Since the transmission lines are uncoupled from the coupling obstacles L and C and the entire network has a characteristic impedance Z which is the same as the terminating impedance 25 of FIG. 4, Z of the evenmode equivalent circuit of FIG. 9A is w(L )/Z and Y of the odd mode equivalent circuit is Z w(2C 8 In FIGS. 8 and 9 L is specified by subtracting (31a) and (31b) to give:
(3 LFZLFQ and L is determined from (31b) as:
As indicated above, the equivalent transmission line coupling elements between the two networks 20-1 and 20-2 are to have dispersive properties. To continue with the analysis presented in terms of an equivalent transmission line, such as the T-sections of elements 43 and 45 in FIG. 9A and 43 and 46 in FIG. 9B, the characteristic impedance of such lines can be defined as a quantity Z such that (35) Z =Z cos 0 where L 1 Z =\/g (n -m and sin 29 where Z is the characteristic impedance of the system at zero frequency and 0 is the length of the line in electrical degrees. Using Z as the characteristic impedance, then the coupling obstacles 44-1 and 44-2, of value L and 23-1 and 23-2, of value 2C must be dual to each other and dual to the T-sections to make Y =Z Since the even and odd mode equivalent circuit transmission line sections are T-networks of the constant k type, the inductor 44 (L can be made to have the same dispersion as these sections, by defining this inductors dispersion as:
X =K tan 0/2,
L F tan where K is a constant and 0 is the line length in electrical degrees.
For the even mode equivalent circuit of FIG. 9B it can be shown that its loss function L is:
The function A is derived from a matrix analysis of a series element X (here 44-1), a connecting transmission line section (here 45-1, 43-1 and 45-2) of dispersion Z =Z cos 0,
and another series connected element X (here 44-2).
For every coupling condition desired, where the cou pling is defined a and L (loss function) is given in (38), different values of K and 0 exist. For the specific case of d L d9 which gives the stationary point of the loss function of (38) with respect to line length 0, a plot of Here in Equations 41a, 41b and 410, L corresponds to an inductance 45-1 or 45-2 of FIG. 9A, C corresponds to one-half the value of capacitor 43-1, and w is the cutoif frequency of the T-section network.
Using Equations 41b and 41c in (37) gives:
w=w Sin /2 cos 0/2 from which from (34).
Also, from (42):
(44) L Z,, tan 0/2 and therefore 45 L 0:7} from (410).
(47) 21? which from (43) gives:
(48) 2C f=KC from (45) 0 Thus, for any given coupling value, all of the various parameters of the network of FIG. 7 are specified. As an example, using the 3 db coupling value given above in a 50 ohm system (Z operating at 30 mc., from the curves 0 :21.1"; K=8.17; X from (42) is 1.52158; w 188.496 cycles; L from (42) is .40361 10 henries; C from (43) is 180.72 10* farads; L from (43) is .0494 10" henries; and 2C from (45) is 39.52 10 farads. To get L if a bifilar inductor is wound to get equal to the value of L (.4036 10 henries), L can be calculated such that Any set of values of L and L can be used to fabricate L so long as L,,/2 is equal to or less than L,,. A perfect winding is when so that L =0 and four inductors are eliminated from the network.
The network 40 of FIG. 7 is dispersive, that is, its characteristic impedance is not constant and its phase shift is not linear with frequency, but the eifects of dispersion (Z Z cos 0) can be readily accounted for. Since 0 is known for any given coupling and Z is known, Z for the elements 32 coupling the two networks 20-1 and 20-2 together is also known. The same dispersion value Z can be realized by connecting equivalent dispersive elements to the network 40. Since the entire network 40 is a four terminal device with T-sections having constant k and frequency responsive properties, that is, it has a passband, it is effectively a constant K prototype filter. Therefore, other filters can be placed at each terminal 1-4 of network 40 to reduce the dispersion and make the network more linear. A preferred filter is the so-called m-derived end section (or half section) which can be connected to network 40 to correct the dispersion of the system.
FIG. 10 shows the network 40 of FIG. 6 with an mderived end section 60 connected to each port. Each end section '60 is shown in the conventional manner by the three generalized impedances 60a, 60b and 600 of the mderived 1r configuration half section. By suitably selecting the values of m the dispersion of the network 40 can be significantly reduced over a selected frequency range of operation. Design of m-derived filters is highly conventional in the art, so the details thereof are not given here.
The addition of m-derived sections makes the characteristic impedance of the network essentially constant at its terminals and makes its phase shift more nearly linear over the frequency band of concern. Thus, adding the mderived networks at the terminals of the coupler makes the response of the coupler more nearly that of a linear system.
While preferred embodiments of the invention have been described above, it will be understood that these are illustrative only, and the invention is limited solely by the appended claims.
What is claimed is:
1. A four port directive coupler for controlled response over a broad band of frequencies, said coupler possessing electrical symmetry with respect to both axes and comprising:
first network means formed substantially of lumped constant frequency responsive components, means connecting said components so that said first network means has a pair of input ports and a pair of output ports, one of said pair of input ports serving as the input port of the coupler and the other serving as the coupler coupled output port,
second network means formed substantially of lumped constant frequency responsive components having a pair of input ports and a pair of output ports, one of said output ports serving as the output port of the coupler and the other serving as the isolated port of the coupler,
said lumped constant components of said first and second network means formed by a pair of highly magnetically coupled conductors and a capacitor in shunt with two of the ends of the conductors,
and means for electrically coupling the pair of output ports of the first network means to the pair of input ports of the second network means, the parameters of the lumped constant components of the first and second network means together with the parameters of the coupling means forming a four port symmetrical directional coupler between the input ports of the first network means and the output ports of the second network means in which the two port networks formed by the even mode bisection of the entire coupler about the symmetry axis is substantially the dual of the two port networks formed by an odd mode bisection of the entire coupler about its symmetry axis over said band of frequencies.
2. A coupling network as set forth in claim 1 wherein said electrical coupling means includes at least one transmission line.
3. A symmetrical four port coupling network as in claim 1 further comprising means connected to at least one of said first and second network means for reducing the dispersion of the symmetrical coupler.
4. A symmetrical four port coupling network as in claim 1 wherein said electrical coupling means comprises a pair of transmission line means with a line being connected between the third and first and the fourth and second ports of the first and second network means respectively.
5. A coupling network as set forth in claim 4 wherein each of said transmission line means is substantially nondispersive over a substantial portion of the operating band of the directive coupler.
6. A symmetrical four port network as set forth in claim 1 wherein the components of each of said first and second network means comprises a bifilar wound inductor having one end of each winding electrically connected to a respective port of one of the pairs of input and output ports, a lumped capacitor means shunted across said one end of each of the two windings, and means connecting the other end of each winding of the bifilar inductor to a respective port of the other one of the pairs of input and output ports.
7. A symmetrical four port network as set forth in claim 6 wherein the electrical coupling means connecting the output ports of the first network means and the input ports of the second network means comprises a respective transmission line means connecting each of a pair of ends of the bifilar inductor of one network means to each of a pair of ends of the bifilar inductor of the other network means.
8. A symmetrical four port network as set forth in claim 6 wherein the other end of each of the windings of the bifilar inductor of the first network means is connected to a respective one of the pair of output ports and the other end of each of the windings of the bifilar inductor of the second network means is connected to a respective one of the pair of input ports of the second network means, said electrical coupling means between said first and second network means comprising a pair of series connected coils connected between. an output port of the first network means and an input port of the second network means, and a capacitor connected between the junction of each pair of series connected coils and a point of reference potential.
9. A symmetrical four port network as set forth in claim 6 wherein the electrical coupling means connecting the output ports of the first network means and the input ports of the second network means comprises a lumped constant transmission means formed of a cascade of series inductors and shunt capacitors connected between an output port of the first network means and an input port of the second network means.
.10. A coupling network as set forth in claim 7 in 'which each said transmission line means is substantially nondispersive over a substantial portion of the operating band of the directive coupler.
11. A symmetrical four port network as set forth in claim 8 further comprising means connected to a port of one of said network means for reducing the dispersion of the symmetrical coupler.
12. A coupling network as set forth in claim 11 wherein said means for reducing dispersion comprises an m-derived filter end section.
References Cited UNITED STATES PATENTS 2,263,461 11/1941 Hagen 333--6 X 2,975,381 3/1961 Reed et a1. 333-10 FOREIGN PATENTS 413,621 10/1932 Great Britain.
HERMAN KARL SAALBACK, Primary Examiner.
P. L. GENSLER, Assistant Examiner.
U.S. Cl. X.R. 333--24, 70
UNITED STATES PATENT OFFICE CERTIFICATE OF CORRECTION Patent No 3 ,452 300 June 24 1969 Joseph D. Cappucci et al It is certified that error appears in the above identified patent and that said Letters Patent are hereby corrected as shown below:
Column 2 equation (1] should appear as shown below:
in 5 =1/2v[ =1/2vr Column 3, line 36, at the beginning of the line insert (9) same line 36,
F 1 0 should read I -F 0 same column 3, equation (6a) should appear as shown below:
1 2 s =1/2v 1 r same column 3 line 66 section (d) should appear as shown below:
Column 8, line 27, "where Zp is the" should read where Z is the Signed and sealed this 29th day of June 1971.
EDWARD MFLETCHER,JR. WILLIAM E. SCHUYLER, J Attesting Officer Commissioner of Paten