US3613034A

US3613034A - Waveguide structure with pseudocavity region for constraining pump and idler energies

Info

Publication number: US3613034A
Application number: US870903*A
Authority: US
Inventors: James N Linnd; William E Meyer
Original assignee: North American Rockwell Corp
Current assignee: Boeing North American Inc
Priority date: 1967-02-20
Filing date: 1969-08-14
Publication date: 1971-10-12
Anticipated expiration: 1988-10-12
Also published as: DE1591346A1; SE347094B; NL6715538A; GB1201946A; US3501706A

Abstract

A noel waveguide structure usable in a broadband microwave parametric amplifier is described which comprises a rectangular waveguide with a pair of substantially U-shaped channels extending oppositely from the top and bottom of the waveguide to a depth of 1/4 guide wavelength at the lower idler frequency. The channels define a pseudocavity region in the waveguide, within which region the idler and pump energy is constrained and parametric interaction occurs. The waveguide is of sufficient width to allow propagation of a signal in the lowest order transverse electric mode and contains a varactor diode as a nonlinear reactance to couple pump energy at a high frequency to a signal at a much lower frequency.

Description

United States Patent El Segundo, Calif.

Original application Feb. 20, 1967, Ser. No. 617,231, now Patent No. 3,501,706, dated Mar. 17, 1970. Divided and this application Aug. 14, 1969, Ser. No. 870,903

WAVEGUIDE STRUCTURE WITH PSEUDOCAVITY REGION FOR CONSTRAINING PUMP AND IDLER ENERGIES 8 Claims, 5 Drawing Figs.

US. Cl 333/83 R, 333/98 R Int. Cl I-l0lp 7/06 Field of Search. 333/73 W, 83, 95, 98

Primary ExaminerHerman Karl Saalbach Assistant Examiner-Paul L. Gensler Attorneys-William R. Lane and L. Lee l-lumphries ABSTRACT: A noel waveguide structure usable in a broad band microwave parametric amplifier is described which comprises a rectangular waveguide with a pair of substantially U- shaped channels extending oppositely from the top and bottom of the waveguide to a depth of V4 guide wavelength at the lower idler frequency. The channels define a pseudocavity region in the waveguide, within which region the idler and pump energy is constrained and parametric interaction occurs. The waveguide is of sufficient width to allow propagation of a signal in the lowest order transverse electric mode and contains a varactor diode as a nonlinear reactance to couple pump energy at a high frequency to a signal at a much lower frequency.

SHEET 10F 3 INVIiN'I'ORS JAMES N. LIND WILLIAM E. MEYER ATTORN Y PATENTEDum 12 19?! 3,6130 34 SHEET 3 BF 3 +ho O T FIG. 3

FIG. 4

INVIZNTOKS JAMES N, LIN WIQLIAM E, ME R I m I ATTORNEY WAVEGUIDE STRUCTURE WITH PSEUDOCAVITY REGION FOR CONSTRAINING PUMP AND IDLER ENERGIES CROSS REFERENCES TO RELATED APPLICATIONS This application is a division of a prior application by the same applicants, Ser. No. 617,231 (now US. Pat. No. 3,501,706), filed in Group 252 on Feb. 20, 1967.

BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates to a novel rectangular waveguide structure having substantially U-shaped channels, extending from its top and bottom, fonning a pseudocavity region in which pump and idler energies may be constrained. The structure facilitates parametric amplification with extremely large percentage bandwidths of microwave signals introduced into the waveguide.

' 2. Description of the Prior Art The principle factor limiting the sensitivity of microwave signal receivers is the noise inherent in the device used to amplify such signals prior to theirdetection. While low noise amplification may be achieved by the use of masers, these devices are restricted in their operation to natural frequencies of the maser material and usually must be operated in a cryogenic environment. A more acceptable approach is to use a parametric amplifier in which energy from an intense pump wave is coupled to the signal by a component exhibiting nonlinear reactance. Parametric amplification provides extremely low noise operation without the need for refrigeration.

Typical prior art microwave parametric amplifiers utilize structures having two adjacent waveguide cavities, one tuned to the signal frequency and a second tuned to the pump frequency. A back-biased diode having nonlinear capacitance, mounted in an aperture in a common wall between the two cavities, reactively couples energy between the pump and signal cavities.

Another typical parametric amplifier structure includes a rectangular waveguide to contain the pump energy. A varactor diode, mounted coaxially in a circular waveguide which intersects one wall of the rectangular guide, serves as the requisite nonlinear reactance. The signal may be introduced into the circular waveguide, and energy may be extracted either at the signal frequency via the circular waveguide or at the idler frequency via the rectangular waveguide.

While the parametric amplifiers described hereinabove may be tunable over a limited frequency range, they are essentially narrowband devices. Prior art broadband parametric amplification has required use of complex mechanical structures. Typical of such devices are the Coupled-Cavity Travelling- Wave Parametric Amplifiers described by K. P. Grabowski and R. D. Weglein in the Proceedings of the RE, Volume 48, No. 12, beginning at page 1973. These devices utilize a series of inductively coupled microwave cavities each containing individual signal, idler, and pump resonant chambers and each containing a varactor diode. The cavities are cascaded and pump energy is applied to the diodes in an appropriate phase relationship to allow travelling wave parametric operation. While broadband performance may be achieved, the system is cumbersome mechanically, requires a plurality of cavities, and demands considerable care to maintain the proper pump phase at each varactor diode.

This application utilizes a novel waveguide structure having a pseudocavity region which pennits pump and idler energy to be constrained to a small portion of a waveguide. This structure permits efiicient parametric interaction between a signal present in the waveguide at a relatively low frequency and a pump having a considerably higher frequency. The structure is mechanically simple, and when used in a parametric amplifier requires only one reactive element and permits amplification with large percentage bandwidths.

SUMMARY OF THE INVENTION The inventive waveguide structure has a pseudocavity region within which the pump and idler energy may be constrained. The structure comprises a rectangular waveguide having a width sufficiently large to support signal energy in the lowest order transverse electric mode. The rectangular waveguide includes substantially U-shaped channels extending from its top and bottom, and terminates in a second waveguide beyond cutofi'. The channels, which preferably exhibit a depth of one-quarter guide wavelength of the lowest idler frequency, define the pseudocavity region. The length and width of the region are selected to ensure that both pump and idler energies are constrained and that a maximum electric field occurs at the location of a varactor diode disposed within the region. This varactor diode couples energy from the pump to the signal. The amplified signal may be extracted from the rectangular waveguide utilizing a microwave circulator. Nondegenerate broadband parametric amplification is achieved using a pump frequency much higher than that of the signal.

It is thus an object of this invention to provide a waveguide structure having a pseudocavity region for constraining pump and idler energy to a portion only of a waveguide.

It is a further object of this invention to provide a microwave structure including a shorted waveguide containing substantially U-shaped channels extending from the top and bottom of the waveguide to form a pseudocavity region capable of containing therein signals substantially above the cutoff frequency of the waveguide.

Further objects and features of the invention will become apparent from the following description and drawings which are utilized for illustrative purposes only.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is an exploded elevation view of the inventive microwave waveguide structureincluding channels defining a pseudocavity region in the waveguide.

FIG. 2 is an elevation view showing the interior surface of the waveguide structure illustrated in FIG. 1.

FIG. 3 is a diagram of a possible electric field distribution within the channels of the pseudocavity as viewed in a plane generally along the line 3 -3 of FIG. 2.

FIG. 4 is a diagramof a possible electric field distribution within a portion of the pseudocavity as viewed in a plane generally along the line 44 of FIG. 1.

FIG. 5 is a graph showing the operational mode spectrum of the inventive broadband parametric amplifier.

FIG. 6 is a block diagram of a broadband parametric amplifier utilizing the inventive waveguide structure illustrated in FIGS. 1 and 2.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The novel waveguide structure which forms the subject matter of this application is illustrated in FIG. I. The waveguide structure comprises four

major sections

10a, 10b, 10c, and 10d, which may be assembled into a unitary structure. In a preferred embodiment each of the sections 10a, 10b, 10c, and 1M may be milled from a solid block of a metal such as copper having high conductivity. Dowel pins (not shown in the figures) may be used to ensure accurate alignment of the assembled sections. Alternately, the structure may be constructed as a single block for example, by casting. The interior surface of the assembled waveguide structure is illustrated in FIG. 2.

As illustrated by FIGS. 1 and 2, the waveguide structure comprises a waveguide section 20 which extends through section 10a, 10b, and 10c, and which terminates at shorting planewidth a and height b, the latter measured between top 21 and bottom 22 of waveguide section 20. As is well known to those skilled in the art, the lowest frequency signal which can be propagated in waveguide section 20 is one whose free space wavelength is equal to 2a. Waveguide section 40, whose width is considerably narrower than a, thus has a lowest propagation frequency considerably higher than that of waveguide section 20. Thus wall 24 and waveguide 40 appear as a short circuit termination to a signal propagating in waveguide section 20 in the lowest mode. I

Extending respectively from top 21 and bottom 22 of waveguide section 20 are substantially U-shaped channels 30a and 30b. These channels, which may terminate against wall 24, define pseudocavity region 35 of waveguide section 20. Channels 30a and 30b respectively comprise end sections 31a and 31b,

first side grooves

320 and 32b, and

second side grooves

33a and 33b. End sections 31a and 31b may be constructed by milling a rectangular opening 31 in section b as illustrated in FIG. 1. Grooves 32(a and b) and 33(a and b) may be constructed by milling appropriate slots through section 10c. In a preferred embodiment, shown most clearly in FIG. 1, the thickness of channels 30a and 30b is t; that is, end sections 31a and 31b, and side grooves 32 and 33, each have the same thickness t. The length and width of pseudocavity region 35 are given by I and w respectively, as shown in FIG. 1. Channels 30a and 30b each extend to a depth d above and below respective waveguide top 21 and bottom 22. Criteria for the selection of the various dimensions t, l, w and d are described in detail hereinbelow in conjunction with the operational description of pseudocavity region 35.

Varactor diode 12 is situated in pseudocavity region 35 and may be supplied with a DC bias via a wire introduced through hole 16b in block 100 (see FIG. 1). A second contact to varactor diode I2 is provided by point contact 14 (not shown in FIG. 2) which extends from bottom 22 of waveguide section 20. Point contact 14 may be fashioned at the end of a metal FIG. 1. In particular FIGS. 3 and 4'show typical electric field patterns which may be induced in

channels

300 and 30b when these channels are of the preferred dimensions.

In a preferred embodiment, the depth d of channels 30a and 30b essentially is equal to one-quarter guide wavelength of the energy desired to be constrained within pseudocavity region 35. For energy introduced in the appropriate mode, this ensures that the electric field (represented by arrows in FIGS. 3 and 4) is a maximum in the planes of top and

bottom surfaces

21 and 22 of waveguide section 20. Thus, a wave introduced into pseudocavity region 35 (e.g., via waveguide section 40) will see a very high impedance at the periphery of pseudocavity region 35 defined by channels 30.

Note that the energy will be constrained to pseudocavity re gion 35 and will not extend out into waveguide section beyond the region defined by substantially U-shaped channels a and 30b. In a preferred embodiment width w of pseudocavity region essentially is equal to an integral number of one-half guide wavelengths of the signal being constrained. Similarly the preferred length l of pseudocavity region 35 is equal to an integral number of one-half guide wavelengths plus one-quarter guide wavelength. These preferred values are illustrated in FIGS. 3 and 4, which show the electric field distribution for energy introduced into pseudocavity region 35 in the TE mode.

Using the configuration of FIGS. 3 and 4, varactor diode 12 may be placed three-quarters of a guide wavelength from wall 24 and midway between grooves 32a and 330. As illustrated, this will ensure an electric field maxima at the location of varactor diode 12 for energy introduced into pseudocavity region in the TB mode.

In a preferred embodiment, the groove thickness t of channels 30a and 30b should be less than one-half the depth d of grooves 32 and 33. An optimum value for thickness t is in the order of one-tenth guide wavelength. A greater thickness I may result in the excitation of spurious modes within channels 30a and 30b, resulting in a significant decrease in the ability to constrain energy to within pseudocavity region 35.

Since the thickness t is considerably less than the width 0 of waveguide section 20, a relatively low-frequency signal introduced into waveguide section 20 will see only a very small inductive perturbation in its field due to the existence of channels 30a and 30b. Thus a signal of a first relatively low frequency (introduced via port 25) may be present throughout waveguide section 20 at the same time that a wave at a second much higher frequency (introduced by way of port 45 and waveguide 40) may be present only in the pseudocavity region 35 of waveguide section 20.

Signals having a one-quarter guide wavelength (kg/4) slightly greater or less than d also will be constrained somewhat by channels 30a and 30b to pseudocavity region 35 However, the degree of containment will be slightly reduced from the optimum value obtained when the channels depth d is exactly )tg/4. Thus it is possible to use pseudocavity region 35 to constrain energy at several closely related frequencies.

In another embodiment of the inventive waveguide structure (not illustrated), a second pair of channels may extend from the top and bottom of waveguide 20 to form a second pseudocavity region surrounding the first pseudocavity region. This second region then may function to constrain any residual energy not completely contained by the first, smaller pseudocavity. In yet another embodiment, two adjacent pseudocavity regions may be formed in the same waveguide. If appropriate care is taken to prevent excitation of spurious modes, the side grooves of one pseudocavity simultaneously may be used as the side grooves of the adjacent channel, even though the two adjacent regions are of different size.

Further, it will be understood that various other modifications may be made to waveguide structure 20 within the spirit and scope of this invention. For example, although FIGS. 1 and 2 illustrate channels 30a and 30b as terminating against wall 24, this is not required. Rather channel sections 30a and 30b may be separated some distance from wall 24 and include yet another groove to form substantially rectangular shaped channels. These channels then will define a rectangular pseudocavity region within the main waveguide 20, into which region energy may be introduced, e.g., by way of a coaxial waveguide introduced through top 21 and bottom 22. Altemately, it may be possible to form pseudocavity region 35 using only one channel, extending either from top 21 or bottom 22. Such a configuration would provide somewhat less energy containment than the opposing channel embodiments illustrated, but may be advantageous in certain applications where it is impractical to prepare grooves in one waveguide wall.

The characteristics described hereinabove indicate that the waveguide structure illustrated by FIGS. 1 and 2 is well suited for application in a microwave parametric amplifier capable of broadband operation in the nondegenerate operational mode.

The theory of parametric amplification has been described widely in the literature, as for example, in the textbook entitled Coupled Mode and Parametric Electronics by William H. Louisell, published in 1960 by John Wiley and Sons, New York. Basically, parametric amplification involves the mixing of a signal at frequency w,=21rf, with an intense energy source called a pump having a frequency 01,. (In the following discussion, the term frequency" will be used to denote the circular frequency F21rflwhere f is the frequency of interest). When combined in the presence of a nonlinear reactance (e.g., a varactor diode) energy is coupled between the pump and the signal. The interaction also gives rise to energy at two additional frequencies w =w,,m, and al m i-0,; these are called idlers. Nondegenerate parametric operation occurs when m is not equal to 2:0,.

Fig. 5 is a graph illustrating a parametric operational mode spectrum; the Wideband parametric amplifier may be operated in the corresponding mode. As indicated in FIG. 5, solid vertical line 51 represents the power present at signal frequency (0,, the center frequency of the amplifier passband. Similarly, line 52 represents the power present in the pump wave at frequency (0,. Lines 53 and 54 respectively represent the power which will be present at the idler frequencies w and when parametric interaction occurs.

For optimum wideband parametric amplification with nearly constant gain across the band of interest, it is desirable to use an input band-pass filter to define the bandwidth of the signal to be amplified. Further, it is desirable to have extreme separation between frequencies 0:, and 0),. The theoretical considerations on which these factors are based are described in the article by George Matthaei, entitled A Study of the Optimum Design of Wideband Parametric Amplifiers and Upconverters," published in the IRE Transactions on Microwave Theory and Techniques, VolumeMIT-9, No. 1, Jan. 1961, pages 23-38.

As shown superimposed on the operational mode spectrum of FIG. 5, dashed curve 56 illustrates the preferred attenuation characteristics of an input band-pass filter which may be used as a component of the wide'band parametric amplifier. Note that the passband region is centered around frequency w, and that the width of the passband defines the bandwidth of the amplifier. Note also in FIG. that frequency (a, is widely separated from 0,. For example, if m, is selected to be 9 GI-Iz. and the passband of attenuation curve 56 is selected to extend from 8 GI-Iz. to Gl-Iz., then (u, may be selected to be a frequency of 94 GHz. In this example, the idler frequencies will be u =85 GHz. 1-] GHz. and m =103 GI-Iz. :1 GI-Iz. (Of course, it is to be understood that these frequencies are cited by way of example only, and that the parametric amplifier described herein may be operated at other frequencies as well.)

A block diagram of the broadband parametric amplifier 60 utilizing the inventive waveguide structure is shown in.FIG.'6; parametric interaction occurs in pseudocavity region 35,'in which region varactor diode I2 is located. The characteristics of the waveguide structure have been described hereinabove in conjunction with FIGS. 1 and 2. In a preferred embodiment, the depth d of channels 30a and 30b is selected to equal onequarter guide wavelength at the lower idler frequency a) Since the pump and upper idler frequencies (1, and a respectively) are not far removed from lower idler frequency m energy at these frequencies also will be constrained to within pseudocavity region 35, but to a slightly lesser degree. This characteristic is illustrated by curve 57 in FIG. 5 which represents the passband" of pseudocavity region 35. Note that at frequencies m, and w pseudocavity region 35 will present a slightly reactive load as viewed from waveguide 40.

The signal to be amplified is introduced into parametric amplifier 60 by way of signal input port 61 in microwave circulator 60. Microwave circulator 62 is of a type well known to those skilled in the art, and functions to allow a signal introduced through input port 61 to exit via common port 63 while insuring that a signal which enters via port 63 will leave circulator 62 via output port 64.

The signal to be amplified enters signal band-pass filter 65 from circulator common port 63. Signal band-pass filter 65, well known to those skilled in the art, may be of the type described, e.g., in chapter 9 of the book entitled Microwave Filters, Impedance Matching Networks, and Coupling Structures by George Matthaei, et al., McGraw-I-Iill Book Company, I964. In a preferred embodiment, signal band-pass filter 65 will exhibit the attenuation characteristics described generally by curve 56 in FIG. 5. That is, filter 65 will have minimum attenuation within the desired passband, and high attenuation at all other frequencies.

The signal from band-pass filter 65 next passes through signal impedance matching network 66, which to match the input signal to the small inductive perturbation which it will see in waveguide as a result of the presence of pseudocavity region 35. Impedance matching network 66 also serves to introduce the signal into waveguide 20 in a particular mode, e.g., in the TB mode, the lowest mode which can be supported by waveguide structure 20. The design of impedance matching network 66 is well known to those skilled in the art, and is described for example in Chapter 6 of the book Microwave Filters, Impedance-Matching Networks, and Coupling Structures," referenced above. In a preferred embodiment, waveguide structure 20 will have a width a (see FIG. 1) which is one-half free space wavelength at the lowest frequency of the signal passband. This will allow the signal to be introduced into waveguide structure 20 in the TB mode.

Pump energy may be supplied from pump source 70 which may comprisea reflex klystron oscillator and a microwave isolator, both operating at the pump frequency 10,. These components are well known to those skilled in the art. Energy from pump source 70 passes through pump impedance matching network 72, upper idler termination 74, and waveguide 40 into pseudocavity region 35. The function of impedance matching network 72 is to match the pump wave to the reactance exhibited by pseudocavity region 35 at the pump frequency. Network 72 performs the further function of ensuring that the pump energy is introduced into pseudocavity region 35 in an appropriate mode e.g., the TE, mode illustrated in FIGS. 3 and 4) to ensure that the pump electric field will be a maxima at the location of varactor diode 12.

In a preferred embodiment, the width 0' of waveguide 40 is selected to provide a cutoff frequency between lower idler frequency w and pump frequency 10,. Further, waveguide 40 should be sufficiently long so that pump impedance matching network 72 does not interact with energy at lower idler frequency an In a typical embodiment, a waveguide 40 length which will ensure 20 db. attenuation at idler frequency w is sufiicient for satisfactory operation. These considerations will allowpump energy from pump source 70 to be introduced into pseudocavity region '35 via waveguide 40, but will prevent energy at the lower idler frequency w from exiting pseudocavity region 35 .via waveguide 40. In effect, this will completely constrain energy at the lower idler frequency w to .within pseudocavity region 35. Upper idler termination 74 is a reactivetrap which functions to reflect energy at the upper idler frequency 0 (which energy will propagate through waveguide 40) back into pseudocavity region 35.

-With the parametric amplifier illustrated in FIG. 6, when a signal and a pump are introduced into waveguide structure 10, parametric amplification will occur and the amplified signal may be extracted via impedance matching networks 66, signal band-pass filter 65, circulator common port 63, circulator 62, and signal output port 64. The bandwidth of the amplifier will be defined by the passband of signal band-pass filter 65 (see curve 56 of FIG. 5). a

It will be apparent that the inventive waveguide structure described in conjunction with FIGS. 1 and 2 is not limited to .use in a broadband parametric amplifier, and may be employed in various other microwave devices. Moreover, while the invention has been described and illustrated in detail, it is to be clearly understood that the same is by way of illustration and example only, and is not to be taken by way of limitation, the spirit and scope of the invention being limited only by the terms of the appendedclaims.

We claim:

l. A rectangular waveguide having means for constraining energy at a given frequency to within a region of said waveguide,

said means comprising at least one channel extending orthogonally from a wider wall of said waveguide to a depth of essentially one-quarter guide wavelength of said frequency,

said channel being substantially U-shaped and comprising an end section and first and second side grooves, said channel forming a pseudocavity capable of supporting a mode of electrical signals at a frequency different from that of at least another mode supportable in said waveguide.

2. The structure as defined in claim 1 wherein a second end section is included at the open end of said grooves of said channel to form a substantially rectangular-shaped channel which defines a rectangular pseudocavity region.

3. In a waveguide having a lowest propagation frequency, means for constraining energy having a second frequency higher than said lowest propagation frequency to within a pseudocavity region of said waveguide, said means comprisin two channels extending oppositely one from the other from the top and bottom of said waveguide to a depth of essentially one-quarter guide wavelength of said second frequency,

said channels being substantially U-shaped and comprising an end section and first and second side grooves wherein the thickness of said end section and said side grooves measured in the plane of attachment to said waveguide is no greater than one-half the depth of said channels.

4. A structure as defined in claim 3 wherein the width of said pseudocavity region is an integral number of one-half guide wavelengths at said second frequency, and the length of said pseudocavity region is an integral number of one-half guide wavelengths plus one-quarter guide wavelengths at said second frequency, said width being the distance between said side grooves of said U-shaped channels and said length being the distance from the open end of a side groove to the remote side of the end section.

5. A structure as defined in claim 3 wherein one end of said waveguide is short circuit terminated, wherein said channels are substantially U-shaped, and wherein said terminated end of said waveguide provides one boundary of said pseudocavity re on.

i A structure as defined in claim 5 further comprising means for introducing energy into said pseudocavity region, said means comprising a second waveguide extending from said terminated end of said waveguide and having a cutoff frequency higher than said lowest propagation frequency.

7. A structure as defined in claim 5 wherein the thickness of said channels is approximately one-tenth guide wavelength at said second frequency.

8. A structure as defined in claim 7 wherein said pseudocavity region has a width of one guide wavelength at said second frequency, and a length of one and one-quarter guide wavelength at said second frequency.

Claims

1. A rectangular waveguide having means for constraining energy at a given frequency to within a region of said waveguide, said means comprising at least one channel extending orthogonally from a wider wall of said waveguide to a depth of essentially one-quarter guide wavelength of said frequency, said channel being substantially U-shaped and comprising an end section and first and second side grooves, said channel forming a pseudocavity capable of supporting a mode of electrical signals at a frequency different from that of at least another mode supportable in said waveguide.

3. In a waveguide having a lowest propagation frequency, means for constraining energy having a second frequency higher than said lowest propagation frequency to within a pseudocavity region of said waveguide, said means comprising, two channels extending oppositely one from the other from the top and bottom of said waveguide to a depth of essentially one-quarter guide wavelength of said second frequencY, said channels being substantially U-shaped and comprising an end section and first and second side grooves wherein the thickness of said end section and said side grooves measured in the plane of attachment to said waveguide is no greater than one-half the depth of said channels.

5. A structure as defined in claim 3 wherein one end of said waveguide is short circuit terminated, wherein said channels are substantially U-shaped, and wherein said terminated end of said waveguide provides one boundary of said pseudocavity region.

6. A structure as defined in claim 5 further comprising means for introducing energy into said pseudocavity region, said means comprising a second waveguide extending from said terminated end of said waveguide and having a cutoff frequency higher than said lowest propagation frequency.