Processing

Please wait...

Settings

Settings

Goto Application

1. WO2016153459 - PASSIVE SERIES-FED ELECTRONICALLY STEERED DIELECTRIC TRAVELLING WAVE ARRAY

Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters

[ EN ]

PASSIVE SERIES-FED ELECTRONICALLY STEERED DIELECTRIC TRAVELLING

WAVE ARRAY

CROSS-REFERENCE TO RELATED APPLICATIONS

The present application may be related to U.S. Patent Application Serial 5 Numbers US 13/372,122, US 13/372,117, US 13/357,448, and US 14/193,072 and published patent application documents including WO 2012/109652, US

2012/0274528, US2012/0206310 and US 2012/0744375.

BACKGROUND

Tedchnical Field

lo This patent relates to series-fed phased array antennas and in particular to

coupler(s) disposed between the radiating antenna elements of the array and waveguide(s) having an adjustable wave propagation constant.

Background

is Military, commercial, and consumer users are in need of ever increasing data bandwidth whether inside or outside of terrestrial data networks. Bandwidth demand is doubling every two (2) to three (3) years, and satellites intended to fill these needs are being launched to provide data service at economies of scale. While ample satellite bandwidth is available at this time, the existing Satcom on the Move (SOTM)

20 terminal antennas are either large/bulky (dish/positioners) or high cost (phased

arrays).

Phased array antennas have many applications in radio broadcast, military, space, radar, sonar, weather satellite, optical and other communication systems. A phased array is an array of radiating elements where the relative phases of respective 25 signals feeding the elements may be varied. As a result, the radiation pattern of the array can be reinforced in a desired direction and suppressed in undesired directions. The relative amplitudes of the signals radiated by the individual elements, through constructive and destructive interference effects, determines the effective radiation pattern. A phased array may be designed to point continuously in a fixed direction, or to scan rapidly in azimuth or elevation.

There are several different ways to feed the elements of a phased array. In a series-fed arrangement, the radiating elements are placed in series, progressively farther and farther away from a feed point. Series-fed arrays are thus simpler to construct than parallel arrays. On the other hand, parallel arrays typically require one feed for each element and a power dividing/combining arrangement.

However, series fed arrays are typically frequency sensitive therefore leading to bandwidth constraints. This is because when the operational frequency is changed, the phase between the radiating elements changes proportionally to the length of the feedline section. As a result the beam in a standard series-fed array tilts in a nonlinear manner.

SUMMARY

As will be understood from the discussion of particular embodiments that follows, we have realized that a series fed antenna array may utilize a number of coupling elements. The coupling elements extract a portion of the transmission power for each radiator from one, or preferably two, waveguides. Controlled phase shifters may optionally be placed at each coupler, to delay the amount of transmission power to each one of the respective phased array elements. The transmission line may also be terminated with a dummy load at the end opposite the feed to avoid reflections.

This arrangement is inherently frequency sensitive, since when the frequency is changed, so too is the phase at the respective radiating elements also changed. This change in phase is proportional to the length of its respective feedline section. While this effect can be used to advantage in frequency scanning, it is normally undesirable, since a phase controller must then also determine a change in the phase shift for each respective frequency change.

In one implementation, this shortcoming is avoided by using a waveguide having a variable wave propagation constant. In one example of a circularly polarized array implemented with such a waveguide, a single line of dual polarization couplers, or a pair of waveguides are used. Coupling between the variable dielectric waveguide and the antenna elements can be individually controlled providing accurate phasing of each element while keeping the Standing Wave Ratio (SWR) relatively low.

In still other aspects, multiple radiation modes may be used to extend a field of regard. Each of the radiation modes may be optimized for operation within a certain range of frequencies.

In still other arrangements, both to increase the instantaneous available bandwidth of the array and to allow maintaining direction of the main beam independent of frequency, progressive delay elements can be embedded in the waveguide couplers. In this arrangement coupler walls are placed along the variable dielectric waveguide. The coupler walls may be curved. These curved walls form focusing dielectric mirrors. These cause the energy entering the coupler to travel back and forth between the mirrors, accumulating delay, and thus effecting a further phase shift.

In one embodiment, the variable propogation constant of the waveguide is provided by adjusting an air gap between layers in the waveguide. There, the waveguide is generally configured as an elongated slab with a top surface, a bottom surface, a feed end, and a load end. The waveguide may be formed from dielectric material layers such as silicon nitride, silicon dioxide, magnesium fluoride, titanium dioxide or other materials suitable for propagation at the desired frequency of operation. Adjacent layers may be formed of materials with different dielectric constants. While gaps are formed between the layers, a control element is also provided to adjust a size of the gaps. The control element may be, for example, a piezoelectric, electroactive material or a mechanical position control. Such gaps may further be used to control the beamwidth and direction of the array.

In one refinement, delay elements for a number of feed points are positioned along the waveguide and fed with progressive delay elements. The delay elements may be embedded into or on the waveguide.

In another refinement, plated-through holes are formed along the waveguide orthogonal to the reconfigurable gap structure. Pins positioned in the plated-through holes allow the gap structure to mechanically slide up and down as the actuator gap changes size.

In yet another refinement, a 2-D circular or a rectangular travelling wave array is fed by waveguide(s) with multiple layers and actuator controlled gaps to provide high gain, hemispherical coverage.

The antenna solution described herein addresses the need for a steerable, wide bandwidth, low-profile antenna by using a variable effective dielectric traveling wave antenna array. By applying propagation velocity control to the traveling wave array technology, an efficient passive antenna array results which is linear, naturally bidirectional, and has no active, complex and expensive electronic devices. The dielectric traveling wave structure provides excitation of antenna elements to yield a cost-effective, high gain microwave antenna array that can handle the necessary high power levels. There are a multitude of possible applications for this phased array technology.

BRIEF DESCRIPTION OF THE DRAWINGS

The description below refers to the accompanying drawings, of which: Fig. 1 is a isometric view of of a unit cell waveguide coupler.

Fig. 2 is a side view of the unit cell.

Fig. 3 is a cross-section end view of the unit cell in an embodiment using a pair of variable dielectric waveguides feeding a patch radiator.

Fig. 4 is a embodiment using a single waveguide, with couplers for each array element; the couplers include matched reflection phase shifters as may be implemented with a quadrature hybrid.

Fig. 5 is a top view of an embodiment using a pair of waveguides with a constant phase shift provided by using dual quadrature couplers for each element.

Fig. 6 is a more detailed top view of one cell of the two-waveguide embodiment of Fig. 5.

Fig. 7 is a cross-sectional view of the unit cell for that same embodiment of Figs. 5 and 6.

Fig. 8 is a isometric, partial cutaway view showing detail of the same embodiment showing the coupled waveguide walls formed as plates.

Fig. 9 is another isometric view of the same embodiment with the walls implemented using pins.

Fig. 10 is a high level schematic diagram of a series fed phased array.

Fig. 11 illustrates e-field magnitude for different air gap sizes.

Fig. 12 illustrates scanning the array by charging the effective dielectric of the waveguide at a selected frequency.

Fig. 13 shows a microactuator in more detail.

Fig. 14 shows (A) epsilon versus scan angle; (B) effective epsilon versus air gap size and (C) dieletric constant versus frequency for a two layer dielectric waveguide.

Fig. 15 shows a gain versus scan angle plot.

Fig. 16 shows the zero directivity loss scanning through broadside.

Fig. 17 shows a feed arrangement for both right hand (RH) and left hand (LH) circularly polarizd (CP) elements.

Fig. 18 is a plot of gain and axial ratio for three frequencies.

Fig. 19 is a resulting elevation pattern for a 30 x 85 element array.

Fig. 20 is a parameter table for a Ka-band implementation.

Fig. 21 is an expected gain pattern.

Fig. 22 shows effective dielectric constant versus scan angle for three radiation modes.

Fig. 23 illustrates gain versus angle when multiple radiation modes are employed to extend a field of regard.

Figs. 24 and 25 are an isometric and cutaway side view of an implementation using curved walls disposed perpendicular to the propogation axis of the waveguide.

Fig. 26A illustrates a waveguide with variable effective propagation constant and crossed dupole radiator.

Fig. 26B illustrates an electrical connection diagram.

Fig. 27 is an exploded top view of a multilayer waveguide where the waveguide sidewalls are defined using sliding pins with plated through holes.

Fig. 28 is a side cross-sectional view of the Fig. 27 embodiment.

Fig. 29 is a bottom view of the same embodiment.

Fig. 30A is a top view of the same implementation.

Fig. 30B is a side view, again of the same implementation.

Figs. 31A, 31B, and 31C are cross-sectional, top and side views of the another implementation using circular array elements.

DETAILED DESCRIPTION OF AN EMBODIMENT

1. Introduction

In a microwave phased array antenna, it is desirable to simplify the design and manufacture of the power dividing phase network. In such components, individual phase controlling elements are placed between each radiating element in series. In this series fed configuration, a transmission line (which may be a waveguide or any other Transverse Electromagnetic Mode (TEM) line) contains all of the antenna element tap points which control power division and sidelobe levels, as well as the phase shifters which control the scan angle of the array. This arrangement provides a savings in the needed electronic circuitry as compared to a parallel feed structure which would typically require many more two-way power dividers to implement the same function.

By way of introduction, this simplification can be provided by performing the phase shift function by varying the wave propagation velocity of the transmission line, thereby inducing a change in electrical length between the elements. The resulting electrical length is given by


(Equations 1 , 2)

where L is the length of the transmission line between elements, and β is the wave propagation constant, inversely proportional to wave velocity, v. Wave velocity is conveniently controlled in certain types of waveguies by varying the dielectric constant of the material which in turn directly affects C, the capacitance per unit length of the transmission through the relationship

v = 1/ ^ (Equation 3)

with L' being the inductance per unit length. This arrangement however has the effect of changing the characteristic impedance of the line which equals

Z0 = y/L' (Equation 4)

The characteristic impedance of the transmission line is thus a fundamental parameter of the implementation, affecting power distribution, efficiency, input Voltage Standing Wave Ration (VSWR) and the like. The fact that line impedance and velocity are coupled in this way is typically considered a fundamental limitation of the series fed array. Thus, scan angle and power bandwidth are coupled together; two parameters that are normally independent in other antenna systems.

However if the variable waveguide / transmission line appears as a reflection type function, the desired phase shift may still be achieved using the same

fundamental type of C variation. In this case, reflections due to the characteristic impedance mismatch of the variable line are canceled at the input, as long as the two transmission line segments (of L) are equal. This arrangement occurs in many microwave circuits called "quadrature coupled" circuits. In this case, the approach is to provide a variable transmission line, with quadrature coupling to the radiating elements.

2. Waveguide Coupler / Coaxial Holes to L-probe-fed-in-quadrature patch

In one implementation of a Dielectric Travelling Wave Array (DTWA) herein, a quadrature coupler uses coaxial holes and an L-shaped probe to feed each radiating antenna element in a linear array. This arrangement solves the problem of how to control the coupling between the variable dielectric waveguide and the antenna elements to achieve accurate weighting of the antenna elements, while still keeping the Voltage Standing Wave Ratio (VSWR) low enough to eliminate the photonic band gap null for broad side angles.

One embodiment of such a waveguide coupler 101, shown in Fig. 1, is coupled to a variable dielectric waveguide 102 below it via several slots 103 formed in the broad walls of the main variable dielectric waveguide 102 and the coupler 101. The slots 103 may be provided in various orientations, numbers and sizes which control the coupling level into and/or out of the coupled waveguide.

Fig. 1 illustrates a unit waveguide coupler 101; each element of a multielement array requires one such unit coupler. In such an arrangement, as will be described below, the unit waveguide couplers 101 are periodically spaced along a main axis of the waveguide 102 according to the desired radiating element spacing on the top layer.

In one embodiment, the unit waveguide coupler 101 is formed in a Printed Circuit Board (PCB) with walls defined by vias or metal plates, but the unit coupler 101 can also be formed in a traditional waveguide structure. The waveguide coupler 101 need only be relatively short in length, as it is used to transfer a guided mode from the main waveguide structure 102, up to the radiating element.

The variable waveguide(s) 102 are formed from a dielectric material or mechanical configuration for which the propogration constant can be varied, either by using materials where dielectric constant is changed via a bias voltage, or through mechanical layer separation in multilayer waveguides. See the discussion below, as well as our related U.S. Patent Publication 2012/0206310 for more details of adjustable waveguide structures.

Fig. 2 shows a side view of the unit cell 101 geometry. On one end of the coupler (the end which feeds a patch antenna radiating element 104) there is a shorted pin 106 (via) that passes through a coaxial hole in the top of the waveguide, up through substrate layers and lands on an L-shaped probe 105 under the patch element 104. On the other side of the coupler 101 is another pin, serving as a matched load 107. Because the coupler 101 is directional, very little energy is dissipated in the matched load 107.

Above the L-probe 105 sits another substrate 108 and on top of that the patch radiator element 104. The L-probe 105 is capacitively coupled to the patch radiator 104. The shunt capacitance between the L-probe and ground plane is cancelled with the series inductance provided by the load pin 107.

Fig. 3 shows further details of the geometry of the feed for an embodiment with two waveguides 102-1, 102-2 arranged in parallel. When two respective L-probes 105-1, 105-2, waveguide couplers 101-1, 101-2, and main variable dielectric waveguides 102-1, 102-2 are situated with a single radiating patch 104 (as per Figs. 3 and 5), each radiating patch radiates a very wide, highly efficient antenna pattern as shown in Fig. 10. Any polarization can be achieved by controlling the phase shift and amplitude for the inputs to the two variable dielectric waveguides, as described below for certain example configurations.

3. Quadrature dielectric traveling wave antenna feeds

In one implementation, phase shift between two feeds changes along with change in a variable dielectric used to implant the main waveguide(s) 102.

Traditionally, to feed a dielectric traveling wave antenna, scatterers or couplers fed in series along the length of a waveguide. For a fixed propagation constant in that waveguide, this fixes the phase difference between the scatterers or couplers, which in turn radiate or couple energy onto another transmission line with that fixed phase difference. In a fixed beam circular polarization traveling wave antenna, this means two quadrature scatterers or couplers are spaced at λ/4 (where λ is the propogation frequency). This causes the phase shift between the two polarizations to be orthogonal, or 90 degrees apart.

However, when the propagation constant of a waveguide 102 can be varied, such as in the case of a dielectric traveling wave antenna described herein, this phase shift between the scatterers or couplers 101 varies with the imaginary component of gamma (and velocity of propagation). The impact of this variable phase shift causes the axial ratio of a Circularly Polarized (CP) antenna to degrade because the axial ratio has a term for phase difference in it. Typically, one would space the scatterers or couplers at such a spacing to cause the phase shift to be 90 degrees as the beam is crossing through broadside so 1) axial ratio would be optimum at broadside and 2) the photonic band gap reflection is cancelled within the waveguide.

An alternative to suffering this axial ratio degradation is to feed a quadrature radiating element (one example would be a dual input patch), as pictured in Fig. 5. Fig. 5 shows the two waveguides 102-1, 102-2 having a relative constant phase shift 110 placed before the feed. In the CP antenna example, this would be a constant phase shift of 90 degrees leading into one of the waveguides. In this way, the phase shift between pairs of scatterers or couplers 101 is fixed, and the change in propagation constant in the waveguide does not affect this phase shift (only the L-probes 105 are shown in Fig. 5 for the sake of clarity; it is understood that unit couplers 101 are associated with each radiating element 104 in this embodiment as were shown in Fig. 3).

The two waveguides 102-2, 102-2 can feed a single line of dual polarization, dual input radiators as per Fig. 4, or each waveguide can feed an individual line of single polarization radiators, as per Fig. 5.

This implementation solves an impedance mismatch when changing transmission line velocity.

As per Fig. 4, this implementation a) inserts an impedance transformer between each radiating element of the array and the following device; and 2) places two equivalent variable transmission lines on quadrature hybrid ports and using combined reflected waves at a fourth port as output.

The arrangement is motivated by the following factors: (a) High Voltage Staning Wave Ratio (VSWR) on travelling wave antennas scanned near boresight due

to admittances adding up when elements separated by half wavelength (λ/2); (b) characteristic impedance of series feeding transmission line changing as its velocity is changed to steer the array.

Prior approaches had several disadvantages including:

(a) VSWR buildup when antenna elements are separated by half wavelength. It is well known that impedance on a line repeats every half wavelength, effectively putting the elements in parallel. When N such impedances are placed in parallel, a high VSWR results.

(b) Characteristic impedance (Zo) of feed line changes as its velocity

(vp) is changed to steer the beam. Zo and vp are interrelated by Zo=sqrt(L'/C) and Vp=l/sqrt(L'*C')- It is impossible to change C without changing both Zo and vp.

The advantage of the Fig. 4 approach is that the addition of impedance transformer eliminates VSWR buildup; in addition, the reflectionless phase shifter decouples Zo and Vp.

As a result, the lowered VSWR will increase gain and improve system performance; and decoupled Vp and Zo will improve maximum scan angles for a given change in feedline parameter C.

More particularly, by inserting matched reflection type phase shifter(s) 120 into the line there is no variation in feedline Zo as the electrical lengths of the short circuited variable lines is changed.

Additionally, the impedance at the junction of each antenna element and the rest of the array can be made to equal 50 ohms by making the parallel combination of the element and feedline impedance 50 ohms. This is done by increasing the feedline impedance by using a quarter wave transformer, or other methods.

4. Two Waveguides and Directional Couplers Feeding Each Patch Radiator

Fig. 5 is a top cutaway view of one implementation of the two waveguide implementation. Fig. 6 shows the detail for one unit cell from a top view. A circular radiating element is implemented as a patch antenna 104. Two waveguide couplers 101-1, 101-2 feed the patch element 104 in quadrature as mentioned previously. The walls defining each of the unit waveguide couplers 101 may be implemented with a

"picket fence" of via pins 130 disposed, as shown, in a rectangular region about the unit cell. Also visible are the L-probes 105-1, 105-2, load pins 107-, 107-2, and coupling slots 103-1, 103-2.

Fig. 7 is a more detailed cross-sectional side view of the unit cell 101 for this embodiment showing the radiating patch, L- shaped probe 105, coaxial holes 112 that accommodate L-shaped probe 105, shorting pin 107, and section of the coupled waveguide 102. Example dimenstions and materials are also listed in Fig. 7 (in this view the vertical axes of the L-shaped probe 105 and shorting pin 107 are seen aligned with one another).

Figs. 8 and 9 are futher isometric views of a two waveguide embodiment showing the several radiating patches and unit couplers. Fig. 8 uses metal plates to define the unit cell walls; the Fig. 9 arrangement instead uses pins to accomplish the same end.

In the DTWA approach described herein there are cost and producibility advantages to building a passive power dividing and phase controlling network as shown in Fig. 10, as opposed to the more common corporate fed parallel structures, with a low noise amplifier (LNA) or power amplifier (PA), or transmit/receive (T/R) module at every element.

In the series fed configuration, a single transmission line (waveguide or TEM line) contains all the antenna element tap points which control the power division and sidelobe levels, and the phase shifters which control the scan angle of the array. This is a great savings in electronic circuitry over a corporate feed structure which would require many 2-way power dividers to perform the same function. In some instances there is a further simplification by performing the phase shifting function by varying the transmission line wave propagation velocity as mentioned above.

The DTWA array of Figs. 5, 6, 10 and elsewhere herein is a completely passive electronically scanned array, which provides numerous advantages over an active electronically scanned array. In one example a complete Ka-Band DTWA Tx/Rx array assembly including electronic components located behind the array can be fit into in the same housing.

To meet the stressing Ka-Band SOTM phased array requirements there are two (2) types of generic antenna implementations: 1) an Active Electronically Steered Array (AESA), or 2) a Passive Electronically Steered Array (PESA).

Although both approaches use arrays of antenna elements connected to variable delay

control circuitry to provide a steerable pattern, the differences in implementation are so great that they must be compared in detail.

The PESA implementation preferred herein is a much better approach due to its lower cost, ruggedness, and simplicity. The approach provides the same performance characteristics as a full AESA due to the unique architecture, but without the complexity and cost of active electronic modules. In a receive (Rx) AESA, the required gain over noise temperature (G/T) is obtained by placing many microwave Low Noise Amplifiers(LNA's) close to each receiving element so that the Signal to Noise Ratio (SNR) of the signal of interest is maintained high throughout the rest of the antenna, which can then have substantial radio frequency (RF) loss. The front end gain of the AESA needs to be sufficient to raise the signal above thermal noise throughout the downstream chain of components. Since the system can now have more insertion loss after the LNA, requirements on other components such as phase shifters and power combiners can be relaxed. At first this would seem like a good tradeoff especially for large arrays where dissipative loss in the power combiners has a major impact on cost and performance. However, consider the realistic impact of inserting an active device, usually several stages of GaAs Field Effect Transistor (FET) gain, at the very input to the system where the desired signal is weakest and the system literally interfaces to the RF environment. Without bandpass filtering ahead of the amplifiers which typically have very sensitive inputs, saturation and even Electrostatic Discharge (ESD)-destruction of the front end will occur in an

Electromagnetic Interference (EMI) environment of several systems operating close together. The filtering must be placed at the very input and therefore has a negative effect on Signal to Noise Ration (SNR), so size and weight are increased if the filters are to be low loss. The problem of phase and gain matching of each channel is greatly exacerbated by the filter/amplifier combination, especially at Ka-Band. AESA systems therefore must employ correctional phase shifters with enough resolution to accommodate these phase errors and are calibrated in software requiring large amounts of data storage. This calibration varies with temperature and time, which greatly complicates the system operation and makes control of the phase shifters a major networking challenge. Calibration times can even limit system concept of operation. There is also a small variation in phase due to the LNA's Amplitude Modulation/Phase Modulation (AM/PM) conversion at high signal levels which increases the Error Vector Magnitude (EVM) especially with Quadrature Amplitude

Modulation (QAM)-type modulation. Additionally, to achieve the required instantaneous bandwidth, true time delay elements are necessary in addition to the phase shifting. These component limitations within the AESA have been the main reason for their slow development over the years.

None of these problems exist with a PESA. As we shall see, the dynamic range of the approach preferred here is limited only by the breakdown voltage of waveguides and printed circuit components, which equate to hundreds of watts of power handling capability. In a transmitting AESA there are additional design issues associated with placing a power amplifier at each array element. First, there is the problem of phase and amplitude matching as described above, and means must be provided to calibrate the transmit path as well as receive. Small phase drifts during amplifier heating in the AESA may affect Error Vector Magnitude (EVM) and be impossible to correct. The phase shifters used for phase correction have to occur at the amplifier outputs, in which case their loss reduces efficiency and raises temperatures of active devices. AM/PM conversion is much more of an issue on transmit, especially if amplitude tapering is attempted in order to reduce transmit pattern sidelobes. The amplifiers must therefore be highly linear which means they are inefficient and output filtering to reduce broadband cosite noise further reduces the system efficiency. All of these components add size, weight and power. The fact that they are operating at a high power level and temperature reduces Mean Time Between Failures (MTBF). Even though AESA's are theoretically capable of "graceful degradation," it has been observed that PESA's are always more reliable. Passive electronically steered arrays (PESAs), especially of the type described herein, have absolutely none of the AESA disadvantages.

The preferred PESA implementation uses a micro-actuated control of the delay of a waveguide transmission medium feeding the elements to steer the beam of the array. There are no active devices in the path between any array element and the output. Array gain is achieved by coherently combining element outputs in very low loss, broadband waveguide directional couplers. There are only two (2) low loss microwave components in the beamformer; transmission lines and directional couplers. With this low complexity, reliability is maximized and design risk is reduced. Since the circuitry is passive and bi-directional, the same antenna array is used on receive and transmit with no difference in performance. In fact, full duplex operation is possible, which has been range tested and proven at Ku-Band.

Fig. 1 1 illustrates a modeled E-field magnitude for the DTWA for two different air gap thicknesses. Fig. 12 illustrates gain patterns versus elevation angle for different effective dielectrics (selected by chosing the size of the gap between dielectric layers).

The variable effective dielectric waveguides 102-1 , 102-2 provide series phase shifting for the radiating elements. In equation (5) if β„3 is the wavenumber of the propagating mode, Θ represents the scan angle along the array axis. wg is controlled by the variation of an air gap within the waveguide, changed by micro-actuators (d is element separation), m is described as an integer radiation mode number and can be any integer and λ is element spacing. As higher radiation modes are used, higher dielectric material are used to support such a slow wave.

cos Θ = —— (Equation 5)

o d

In a preferred arrangement as per Fig. 13 for reference, the variable effective dielectric waveguides 102 each may contain two (2) layers of dielectric. The space between the two (2) boards forms a single air gap in each waveguide, which is controlled by the micro-actuation. In this embodiment, the upper board remains fixed to a multi-layer PCB above it, while only the lower board is moved down to control the air gap height. As the air gap thickness is increased, wg increases, causing a change in Θ. This method has been described previously for phase shifter

applications, however the application of this technology in a traveling wave antenna believed to be unique to the preferred DTWA design.

Fig. 14 illustrates further results of an HFSS model of the single gap embodiment for Ka-band, showing (A) epsilon versus scan angle; (B) effective epsilon versus air gap size and (C) dielectric constant versus cutoff frequency.

A particular waveguide dispersion may cause very small beam squint; as well, element spacing may also cause beam squint. Those can be corrected by

implementation of progressive delay at each element (a frequency-dependent phase shift), improving bandwidth.

For most Ka-Band SOTM applications, Circular Polarization (CP) is desirable. To facilitate CP one can utilize two (2) variable dielectric waveguides, fed in quadrature feeding a single line of radiating patches as per Fig. 5. This approach provides better axial ratio over Field of Regard (FoR) than a single waveguide, as the Theta, Phi gain angle deltas are held constant at 90°, over the entire FoR.

Additionally, it allows for selectable polarization Left Hand (LH)/Right Hand (RH)/Circular Polarization (CP)/Linear (LHCP/RHCP/Linear). It is also possible to feed a single radiating CP patch from a single waveguide, if the feeds are spaced at λ/4, however, this requires the phase shift between the patch feeds to be fixed to the propagation velocity in the waveguide, so a low axial ratio is only possible over a narrow FoR.

The waveguide may be arranged to ensure the fundamental guided mode propagates and all other guided modes remain in cutoff for the entire frequency band of operation - and through the entire air gap range. As shown in Fig. 15, the open stop band, also known as photonic band gap, is eliminated in this traveling wave antenna because the return loss of each directional waveguide coupler is so low, the coherent sum of reflected power does not significantly increase the VSWR.

Fig. 16 illustrates the resulting zero directivity loss in a plot of RHCP gain versus theta for different gap sizes. There is an open stop band passing through broadside due to coupling of oppositely directed space hormones, also causing high VSWR at broadside. But the coupler maintains directivity, coupling value, and return loss throughout system bandwidth and waveguide characteristic impedence.

Fig. 17 illustrates a RH/LH CP feed circuit.

The waveguide-fed unit cell patch may also have a broad 3 dB elevation pattern to reduce scan loss over the required Field of Regard (FoR), and low axial ratio over a broad elevation pattern for maximum signal efficiency. Additionally, the patch fed from the waveguide directional coupler has extremely low return loss and high efficiency, allowing it to achieve a peak gain of 8.5 dBiC throughout the Ka-Band. The unit cell RHCP gain, LHCP gain and axial ratio patterns for three (3) frequencies are shown in Fig. 18, with performance represenative of the entire Rx Ka-Band. Depending on G/T and gain margin required over FoR, it is possible to customize the unit-cell pattern to fit the gain over FoR requirement more closely. For example, the array gain over FoR presented below has excess G/T at broadside for the Ka-Band SOTM application, and meets the G/T requirement at the FoR edges. If the array size/gain over FoR trade shows the G/T may not roll off below the requirement at the FoR edge, this unit-cell would be modified to have a lower peak gain, but

broader elevation pattern, which ultimately smooths out the G/T over scan angle, meeting the requirement over all scan angles, while reducing the required array size.

For each element, a directional coupler, adapted from work done by H. J. Riblet in "A New Type of Waveguide Directional Coupler," Proceedings of the IRE, 1948, pp. 61-63, was designed to provide: 1) a controllable coupling value for amplitude taper implementation, and 2) extremely low return loss in the waveguide, eliminating the photonic band stop passing through broadside. Additionally, the directional coupler directly feeds the patch feed, eliminating any lossy intermediary feedline. The photonic band gap phenomenon which results in large gain variation with frequency is alleviated by this method.

The implementation of waveguide directional couplers solves two (2) issues: 1) elimination of the photonic band gap effect, which causes a gain dropout at broadside on typical half-wave traveling wave arrays, and 2) allowing precise amplitude illumination for sidelobe level/beamwidth control. The directional couplers have extremely low return loss, essentially eliminating any reflection in the main waveguide, which is the source of the photonic band gap effect. The size and shape of the coupler elements control the level of energy coupled into a guided mode in a PCB integrated waveguide which in turn feeds the patch above it.

In Fig. 19 the sidelobe level control that is a result of a modified- Taylor amplitude series of waveguide directional couplers to form the desired current distribution along the array is shown. In the Ka-band satcom application, SLL and beamwidth control are critical because communications satellites are tightly placed in the geostationary orbital positions, as close as 1 degree. Strict sidelobe level and beamwidth compliance is required by 47 CFR FCC 25.209 and similar military requirements exist as well to prevent interference with adjacent orbital positions.

The guided mode within the dielectric waveguide may be excited or received directly from a PA, LNA or down/upconverter. Typically a short coaxial cable connects the antenna to these devices. For this application, a coaxial feed was designed to interface with the waveguide. A mode conversion from a coaxial cable to a waveguide guided mode, whose impedance and wavenumber is changing rapidly (over a factor 2: 1) as the beam is scanned through the FoR is achieved over the Rx Ka-Band bandwidth (19.2 - 21.2 GHz) using the waveguide feed developed for the Ka-Band DTWA. There is a quarter-wave cavity backing the opposite direction of

desired propagation. The feed achieves < 2:1 VSWR and < 0.6 dB insertion loss over all air gaps and throughout the band.

5, System Considerations

For the Ka-Band SOTM application, tracking the satellite location with respect to the platform location and aspect is critical. Although it is true the instantaneous bandwidth of a Wideband Global Satellite (WGS) transponder (a typical Ka-Band transponder), and potential communications channel bandwidth is 125 MHz, the nature of the waveform often used for this application requires tracking on a satellite beacon rather than the main communications waveform itself. A beacon signal typically accompanies the various transponder signals, either above, below, or within the band. A separate beacon receiver is used to monitor the beacon signal, and provide input to the core terminal on tracking movements. However, this beacon signal could be separated in frequency from the main channel bandwidth, requiring an effective instantaneous bandwidth greater than the communications channel itself. For this reason, it may be necessary to extend the instantaneous bandwidth of the antenna above what is actually required for the communications channel to accommodate this tracking method.

It is important to address this subtle system requirement because it has a large effect on system bandwidth requirements. Reuse of the DTWA for the beacon receiver channel provides little impact on the overall system performance. A single directional coupler per line can be placed either between elements or on the underside of the waveguide to couple energy from a subset of the line array into a beacon receiver. A lower risk implementation trades size, weight, power and cost to reduce the instantaneous bandwidth to only what is needed for the communications channel. With the addition of a lower gain beacon DTWA and beacon receiver, adjacent to the communications channel DTWA, tracking is performed side by side with the main array, at the cost of additional surface area.

To address an instantaneous bandwidth need there are several options. An end fed array provides angle scanning by varying the propagation velocity of the transmission line feeding the antenna elements, thereby controlling the differential phase between elements. This also has the result that angle scanning occurs as the operating frequency is changed, which is an undesirable effect for our application. In fact, antennas constructed from arrays of waveguide-fed slots exist more because of their ease of fabrication than by a desire to synthesize such a structure. In any case, the problem comes about due to the vectorial summation of element outputs; the direction in which an end fed array has maximum gain is a direct function of the phase through the medium of propagation between elements. If there is a fixed differential delay in a path between two (2) elements and the distant summing point, then the relative phase is a function of frequency. In corporate feed structures means are provided to equalize these delays so that array pointing is independent of frequency. In an end fed system, provisions have to be made to delay the arrival of signals from elements closer to the output to equal the delay from elements further away. For example, the delay from the sixth element to the output is six (6) times the delay from the first element to the output, where there is an excess phase shift of:

ΔΦ = 360Ndf/vp (Equation 6) degrees, where d is the (constant) separation between elements and N is the difference between the element number and the number of elements. The constant vp is the velocity of propagation in the medium and f is the frequency of interest. So, one simply needs to insert a differential delay of that amount between antenna 1 and coupler 1. The three (3) viable approaches to address this requirement, dependent on the total bandwidth required are:

1. Do not implement any delay or correction - Depending on bandwidth requirements and peak gain beamwidth, the far-field peak beam direction may only vary over a very small angle across the RF bandwidth. This undesired beam scanning with frequency causes a slight distortion in the gain over frequency curve, and the severity of that distortion depends on the beamwidth. This method is acceptable up to a 2.5% bandwidth, given the beamwidth is not extremely narrow.

2. Progressive delays embedded in the line arrays - A progressive delay approach allows equalization of delays and far-field pattern alignment over a 10% bandwidth. A delay element is inserted between the coupled waveguide and the radiating element. The delay element is designed N times for different delay values, and each is implemented separately along the line array. The limiting factor in the progressive delay element approach is loss per unit delay. As with the waveguide, loss in the delay element must be kept to a minimum. Various

circuit and material combination can realize the delay as a several section (~5) hairpin resonator coupled-line filter, implemented in microstrip on fused quartz is optimal. With this approach, loss per delay has been simulated as less than 1.0 dB/nS.

3. Dielectric wedge approach - A dielectric wedge may be placed on top of the array, and is integrated as part of the radome. The dielectric constant and shape of the wedge creates the progressive delay required along the line array. The advantage of the wedge is that it can be implemented in a low loss, high epsilon dielectric, providing a high delay to loss per unit length ratio. For this reason, it can achieve the highest bandwidth, >10 .

One specific implementation of the DTWA antenna was designed to meet the requirements shown in Fig. 20. These requirements were derived from a system level performance analysis for Ka-Band SOTM on move applications.

6. Multiple radiation modes extend field of regard in a traveling wave antenna.

The following equation shows the peak radiation scan angle for any traveling wave antenna:

β

cosQ = - m (Equation 7)

Po S

Θ is the scan angle

λ is the free space wavelength

s is the line array element spacing

βο is the free space propagation constant

β is the adjustable waveguide propagation constant ; and

m is the radiation mode

One can thus select multiple m (mode values) and find multiple solutions for theta for a certain range of β. For example, in the plot of Fig. 22, the x axis represents theta (scan angle), and the y-axis represents an "effective dielectric constant" which is related to beta. A solution to the equation is shown for three frequencies ( at the operating frequency band edges and at a middle frequency) for an element spacing of .525λ. As we change beta (the waveguide propagation constant), the solution to the equation scans along theta.

There are three radiation modes plotted (m=0, 1 ,2) in Fig. 22. It can easily be seen that to scan to a single theta value (such as theta indicated by the vertical arrow 1100), one could source the traveling wave antenna radiation from a waveguide with an effective dielectric constant of different values, and depending on that value, a certain mode would be selected. In the illustrated case, one could scan lower in theta along the thick line 1100 using up to an effective dielectric constant of 22.5, and if desired, continue scanning with a lower dielectric constant of 7.5. Using this method of mode switching, the FoR can be extended to 180 degrees.

This feature becomes useful when trying to achieve very high effective dielectric constants, where the gaps between waveguide laters must become very small. To alleviate this very small gap requirement, as the array is scanned in that direction, operation can switch to the next lowest mode to continue to the Field of Regard (FoR) edge with larger airgaps.

An HFSS (High Frequency Structured Simulator) model simulated this phenomenon and shows that multiple radiation modes can be used to extend the Field of Regard (FoR). See Fig. 23.

7. Progressive Delay Elements

To increase the instantaneous bandwidth of the array, i.e. to maintain the direction of the main beam independent of frequency, progressive delay elements may be embedded in or withthe waveguide couplers 101. One possible geometry is shown in Figs. 24 and 25. The input and output coupler faces 140 lying transverse to the axis of the variable dielectric waveguide 101 may be curved to form a pair of focusing dielectric mirrors 145. The energy entering the coupler 101 then travels back and forth (as shown by dashed lines 147) between the mirrors 145 much like the mirrors in a laser. The number of passes will depend upon the exact curvature of the mirrors 145. It is anticipated that a high dielectric material (e=36) may be used to accumulate the required delay. Delay will thus vary progressively along the array.

8. Other Design Considerations

In addition, there are further possibities with the phased array antenna(s) described herein

Do not implement any delay or correction. Depending on bandwidth requirements and peak gain beamwidth, the far- field beam direction may only scan over a very small angle across the bandwidth. This beam scanning with frequency causes a slight distortion in the gain over frequency curve, and the severity of that distortion depends on the beamwidth. This method is acceptable up to a 2.5% bandwidth, given the beamwidth is not extremely narrow.

Progressive delays embedded in the line arrays. The progressive delay approach allows equalization of delays and far-field pattern alignment over a 10% bandwidth. A delay element can be inserted between the coupled waveguide and the radiating element. The delay element is designed N times for different delay values, and each one is implemented separately along the line array. The limiting factor in the progressive delay element approach is loss per unit delay. As with the waveguide, loss in the delay element must be kept to a minimum.

Dielectric wedge approach. A dielectric wedge may be placed atop the array, and integrated as part of the radome. The dielectric constant and shape of the wedge performs time delay beamforming for each progressive element. The advantage of the wedge is that it can be implemented in a low loss, high epsilon dielectric, providing a high delay to loss per unit length ratio. For this reason, it can achieve the highest relative bandwidth, >10%.

9. Waveguide with Adjustable Propogation Constant and Progressive Delays

Conventional traveling wave fed phased arrays are inherently narrow band antennas. The equation governing the beam direction Θ is given by

cos(9) = beta (waveguide)/beta (free space) - m λ/d (Equation 8)

where beta (waveguide) is the propagation constant of the waveguide, beta (freespace) is the propagation constant in air, d is the array spacing, m is the mode number, and λ is the wavelength. The wavelength term limits the bandwidth.

Figs, 26A and 26B illustrate a refinement where the bandwidth limitations of travelling wave phased arrays are overcome by embedding progressive delays into array elements positioned on or in the waveguide. Here a variable propogation constant waveguide 1502 is formed of multiple layers, with gaps provided between the layers. Changing the size of the gaps has the effect of changing the effective propagation constant of the entire waveguide.

An array of antenna elements, here consisting of crossed bow ties 1504, are placed along the length of the top surface of the waveguide 1502. The antenna elements 1504 may each be fed with a quadrature hybrid combiner as for the other embodiments (not shown). The key to the wide band operation is a delay line 1525 embedded in or with each antenna element along the array. The delay line 1525 is a compact helical HE11 mode line using a high dielectric constant material such as titanium dioxide or barium tetratitinate.

As shown in Fig. 26B, the delays 1525 progressive decrease along the array. These delays cancel out the delays caused by the waveguide 1502 which allows the use of m = 0 in equation (1) and results in the equation:

cos (θ) = δ beta(waveguide)/beta(freespace) (Equation 9)

where δ beta( waveguide) is the additional delay (plus or minus) added to the waveguide to permit scanning. There are no frequency dependent terms, thus the scanning is wideband.

The additional delay is provided by changing the propagation constant in the waveguide with a gap structure.

10. 2-D Dielectric Travelling Wave Array Methodology for Implementaiton of Actuator-Controlled Beam Steering

In a second refinement, a waveguide has plated-through holes provided with a reconfigurable gap structure, with pins positioned in the plated-through holes. The pins allow the structure to slide up and down as the actuator gap changes size.

In order to facilitate beam steering in two dimensions with a 2-D configuration consisting of rows of 1-D traveling wave excited arrays of elements, a 2-D gap structure may utilize layers of dielectric slabs 1602 with rows of periodically spaced plated through holes 1610 and actuator strips 1620 of piezoelectric or electro active material. The rows of plated through holes define side walls of individual waveguide sections 1502. The slab waveguide 1600 arrangement is shown in Fig. 27.

Pins 1630 are placed along the actuator strips to:

1) ensure the alignment of the reconfigurable gaps 1603 as the gap spacing is increased to scan the beam;

2) add shielding between adjacent rows of 1-D arrays;

3) provide a DC path for control power to the actuator strips 1620; and

4) feedback to provide close loop control.

Strips of conducting material can be deposited on both sides of the piezoelectric layers 1620 to enable control voltages to be impressed upon the piezoelectric actuators through the pins 1630. The control voltages can be applied separately to each row or applied to the entire array by connecting the conducing strips together at one end of the structure.

Fig. 28 shows a side view of the same structure 1600 with an exciting horn antenna (feed) 1650 at one end. There will typically an array of horns, one for each row (e.g., for each waveguide). To facilitate beam steering in the direction orthogonal to the 1-D rows of elements, each horn is fed with a progressive phase shift. The radiationg patch(es) are placed in a layer 1650 above the slabs 1602.

Fig. 29 shows a bottom view of the same slab waveguide structure 1603 with the array of horn antennas 1650 now visible at one end. The reconfigurable gaps 1603 and the waveguide pins 1630 are also seen. The lower surface may have a printed circuit board 1680 that provides control and power circuits to the actuators which allows for control of the gap size(s). The control of the gaps changes the effective dielectric of the slab which allows for scanning of the beam without a change of frequency in the traveling wave array.

11. 2-D Dielectric Travelling Wave Antennas

In this refinement, 2-D circular and rectangular travelling wave arrays are fed by slab waveguides with multiple layers and actuator controlled gaps to provide high gain hemispherical coverage.

Traveling wave arrays would typically require a separate waveguide to provide exitation to each row of a 2-D traveling wave array. Here, a single waveguide provides an elevation steerable line array of elements with the line arrays configured side-by-side. A separate conventional feed system is used to excite each-line array with the proper phase or time delay to provide steerabiility in the azimuthal plane. The elevation steering of the traveling wave line arrays is accomplished by actuator controls gaps in the dielectric to control the propagation constant.

By using a two-dimensional slab waveguide with 2-D gaps controlled by actuators, it is possible to eliminate the need for separate waveguides and to provide high gain hemispherical coverage. The two geometries to be considered are (A) a Cartesian geometry using rectangular slabs and (B) a circularly symmetric geometry using circular slabs.

(A) Cartesian geometry case using rectangular slabs

As shown in Fig. 30A (a top view) and Fig. 30B (a side view), a square slab waveguide 1600 (again, formed of multiple dielectric layers as per Fig. 16 (27?) is used in which the exciting elements 1910 are mounted along the sides of the waveguide. The exciting elements (vertically polarized) 1940 of two adjacent sides are used to generate a plane wave excitation in the slab as shown by the dotted line 1960 in Fig. 30A. A plane wave 1620 in any direction can be generated by the use of the exciting elements 1910 on the appropriate two adjacent sides.

The exciting elements 1910 should have beam widths of 90° to guarantee uniform coverage over the azimuthal plane. Mounted on the top surface of the slab waveguide 1600 are so-called scattering elements 1940 which intercept a small amount of the plane wave excitation and reradiate the power. The system thus operates as a leaky wave structure.

The scattering elements 1940, which should exhibit hemispherical patterns, can be circularly polarized crossed dipoles are arranged in a Cartesian grid pattern, as shown.

As in the implementations described above, one can control the propagation constant in the slab using the actuators (not shown in Fig. 19 A), and thus determine the elevation angle of the beam, while here the direction of the plane wave in the azimuthal plane defines the azimuthal angle of the beam.

(B) Circular symmetry implementations

The implementations shown in Figs. 31 A, 3 IB and 31C provide circular symmetry as : 1) a "flat" circular slab version and 2) a "conical wedge" version.

The flat circular case in Figs. 31A and 3 IB uses a circular slab waveguide with a hole in the center for the exciting elements, a commutator, and a beam former. As in a generic circular array, the beam former feeds a sector of exciting vertically polarized elements 2010 to obtain a narrow beam in the direction of that sector, while the commutator 2020 selects the sector direction. The scattering elements are configured in concentric circles 2030 (only partially shown for clarity), keeping the number of elements in each concentric circle constant. The elevation angle of the beam is determined by the propagation constant of the slab waveguide 2002 with configurable gaps 2003 as determined by the gap width, which is controlled by the gap actuators. The azimuthal angle of the beam is determined by the position of the commutator 2020. As in the Cartesian case of Fig. 30A, the scattering elements 2050 should have a pattern providing hemispherical coverage.

The wedge version shown in Fig. 31C provides wideband coverage using a conical wedge 2080 as a progressive delay element. The wedge 2080 is situated on top of the circular slab waveguide 2090 with configurable gaps 2092. An exponential coupling layer 2095 is introduced between the wedge and the slab waveguide. The exponential layer 2095 is needed to generate a uniform plane wave across the wedge 2080. No scattering elements are needed since the layer and the high dielectric constant of the wedge provide a leaky structure. The elevation angle of the beam is, as in the flat slab version of Figs. 31A and 31B, determined by the propagation constant of the slab waveguide as determined by the gap width. Since no scattering elements are used, arbitrary polarization can be provided in the main beam by introducing circularly polarized exciting elements 2099, or combine vertical and horizontal elements such as crossed bowties.

What is claimed is: