A Novel Methodology to Feed Phased Array Antennas

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 55, NO. 9, SEPTEMBER 2007 2489 A Novel Methodology to Feed Phased Array Antennas Diego Betancourt and Carlos del Río Bocio, Member, IEEE Abstract—A new methodology to design beam-forming networks (BFN) to feed antenna arrays is introduced. Using this methodology is feasible to reduce the complexity of the associate control by antenna array could be conof a phased array, since, an trolled to steer the beam using four phase shifters instead the 2 conventionally used. A prototype was designed, built and measured as proof of concept. The prototype consists on 3 by 3 Quasi-Yagi antennas fed by four input-ports. The measurements show that the main beam of an antenna array fed by this BFN can be steered to any desired direction. N N N Index Terms—Beam forming network, electronic scan, phased array. I. INTRODUCTION P HASED array antennas are called to play a fundamental role in high-performance communications systems, thanks to its characteristics of integration and versatility. Usually, beam steering is made by means of phase shifters and attenuators associated to each one of the radiating elements that conforms the antenna array [1]. This is an effective but costly and redundant way to define a Phased Antenna array, since complexity of systems grows with the number of elements usually required to ensure the appropriate angular resolution of the system. Several alternatives are used to dismiss phased antenna array complexity, such are: a) introduce a phase-shift to a group of radiators instead one per each element [2], reducing scanning possibilities to certain direction in space; b) use special beam-forming networks (BFNs), with special behaviors, i.e., Rotman lenses, Blass matrices or Butler matrix [3], generally applied to linear arrays; and c) use a combination of mechanical with electronic components to redirect a beam to the desired position [4], moving physically the antenna array. Others possibilities, are related to the use of optic components [5] and mathematical simplifications applied to BFN [6]. In this paper, we introduce a new methodology based on coherently radiating periodic structures (CORPS) concepts [7] applied to BFNs, that we call C-BFNs. Specifically, our interest is centred on the scan capability of a Phased Array antenna and how to reduce the complexity of common systems. In general terms, the beam steering control is usually independent of the Manuscript received January 6, 2007; revised May 10, 2007. This work has been supported by the Spanish Government by the project TIC2003-09317C03-01. The authors are with the Public University of Navarra, 31006 Pamplona, Spain (e-mail: [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TAP.2007.904133 Fig. 1. C-BFN of two layers, with N input-ports and N + 2 output-ports. power distribution system or feeding network. In the methodology proposed in this paper, we combine the beam steering control and the power distribution network, reducing the complexity of the conventional phased antenna array systems. Our C-BFN, not only distributes the power to each element of antenna array but also contributes to define the appropriate phase distribution overall the radiating aperture to steer the beam in the desired direction. With the methodology proposed is feasible to steer a beam of an by antenna array using only four input-ports. This paper is organized as follows. First, we state the working principle that governs C-BFNs ideal behavior, from linear to 2-D array systems. Second, we present a way to implement a C-BFN in microstrip technology using a power combiner/divider of three ports. Third, the build prototype is shown in order to corroborate the behavior characteristics of C-BFN applied to antenna arrays, and finally, the measurements and results obtained by the prototype are introduced. II. WORKING PRINCIPLE A C-BFN is defined by the alternative iteration of split (S) and recombination (R) nodes. The S-node receive power and redistribute it equally among its outputs. S-nodes are also responsible for the interconnection between layers of the network. The R-nodes make possible to recombine the power entered through its input-ports. The recombination process is, in fact, a vectorial sum of the field throughout a R-node. A general structure configuration for a planar C-BFN of two layers, input-ports and output-ports is shown in Fig. 1. Considering the conservation of energy and taking into count that C-BFN consists only on lossless passive components, we can calculate the behavior of this structure as a result of the characteristics of general split and recombination nodes. General-split nodes are such that have one input port and output-ports. At each output port is delivered a th part of the power introduced at input port, as can be corroborated using the following expression [9]: 0018-926X/$25.00 © 2007 IEEE (1) 2490 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 55, NO. 9, SEPTEMBER 2007 is the power delivered by the output ports of split Here, and with are the magnitude and node, phase of each of the output-ports respectively. For our split node case, is 2 and (1) will be (2) With the real part of the admittance seen at output ports. General-recombination nodes are such that have one output-port and several sets of two input-ports. The power at output-port can be calculated, using Fig. 2. C-BFN structure of three layers. The structure has one input and four outputs. Inset boxes show power values obtained after each layer. (3) and , with , are the field magnitudes Here, and phases of each set of input-ports respectively, and is the admittance seen at output. We can do some simplifications on (3) if we note that our R-node has only 1 set of input-ports , thus (4) the admittance seen at output. The last term in exBeing pression (4) represents the correlation between input fields at R-node that depends on its arrival phase. When fields have different phase, the calculus realized with (4) take into account how fields interact among them. Considering the energy conservation, that an S-node can act like R-node and input fields with equal magnitude and phase, we can establish GS as on half of GR. In other cases the phase of fields should also be taken into account. Phases are calculated doing the vectorial summation of field vectors at input-ports of R-node. In a S-node case, each output will have the same phase shift as related to the phase of the input. Additionally, for a C-BFN structure, as shown in Fig. 1, is important to note that the power-in at solid-grey port never reaches the most right-located R-node neither in the layer 1 or 2. This property of C-BFN is useful to define phase planes for beamsteering arrays, if we note that the phase of the most left-located output-port is the same as the most left-located input-port, and vice versa. By simple inspection, we can conclude that after layers, at output-ports is created a sampled (interpolated) version of the phase shift defined by the input-ports. This property will be explored more in detail in this paper. A C-BFN configuration of three layers, 1 input-port and 4 output-ports is shown in Fig. 2. Power values are shown in inset boxes in this figure, for a C-BFN with one input-port active and three layers. In general the results obtained for power values after each layer correspond with the Pascal’s Triangle normalized by the sum of its in-row coefficients. To know, Pascal’s triangle coefficients can be obtained using a binomial expansion, as follows [8]: (5) Being the number of layers used. Therefore, in a C-BFN, of layers, the amplitudes at the output-ports will have Binomial shape. On the same way, when equal magnitude and Fig. 3. N by 1 antenna array feed by N 0 2 layer C-BFN. in-phase field arrives at each input port of a R-node, at outputs of a R-node can be calculated using either (4) or (5). A. 2-D Feed Network Next step in the design is to consider an by array antenna and let it be feed by our C-BFN. We can use for this purpose C-BFNs that feed, each one, a particular row of a by antenna array, thus the extracted array , , will array. define an 1 From C-BFN point of view, there are outputs (to antennas) and 2 inputs, having layers, as it is shown in Fig. 3. Then, layers to feed the after using identical C-BFNs with inputs, two per each th C-BFN. antenna array, there will be Let define , , 2 the input-ports that feed the th and arrays of input ports. We C-BFN and consider could apply the same process to feed these input-ports as was used for feed each one of array . and are 1 arrays, so, we can feed Note that layers to obtain four them each one by an C-BFN of input-ports, two for each C-BFN respectively. Summarizing, after applying C-BFNs of layers to an by array, is possible to feed it using only four ports. The complete configuration for feed an by array is shown in Fig. 4. Geometrically, is a matter of fact that only three points are necessary to define a plane in Cartesian space. In our case, using only four input ports to define a plane phase front, we will be quite close to the geometrically optimal solution. This is possible thanks to special properties of C-BFN to make a linear representation of phase shift entered to input-ports and driven BETANCOURT AND DEL RÍO BOCIO: A NOVEL METHODOLOGY TO FEED PHASED ARRAY ANTENNAS 2491 Fig. 5. (a) Gysel cell. (b) Complete C-BFN, S-node and R-node are implemented with Gysel cells. N N N02 Fig. 4. by antenna array feed by only four input-ports. The input-ports C-BFN necessaries to feed this antenna array, are reduced to four by using of layers. N+2 through the output-ports. [10] shows this issue for a lineal array. In an by array antenna case, a plane defined with the four input-ports will propagate throughout the feed network, defining a sampled version of phased plane (with by points ) at the aperture plane, this plane can be used to steering the beam. III. C-BFN IMPLEMENTATION One of the most important challenges at time to implement a feed network based on C-BFN is to found a circuit that represents the desired characteristics of S-node and R-node. In our case we are going to work with micro-strip planar technology, therefore, both S-node and R-node would be implemented by mean of a Gysel Power Divider [11]. A Gysel Power Divider is ring impedance transformer, which consist of five ports, a as it can be seen in Fig. 5(a). The scattering matrix of a Gysel cell is shown as (6) As can be inferred from (6), a Gysel cell can acts like an S-node because it delivers equal in-phase power ratio in ports 2 and 3, while isolates ports 4 and 5. This is, there is no interaction between input signals since S-nodes and R-nodes have and . In order to effectively isolate port 4 and 5 of Gysel cell, defining a three port device, such ports are connected to ground via chip resistors of impedance Zo. Additionally, using the reciprocity properties of S-nodes and R-nodes, Gysel cell can also be used as R-node only changing its relative position in the circuit. So, a complete C-BFN can be implemented using Gysel cells. A C-BFN configuration of 1 layer, based on this circuit is shown in Fig. 5(b). phase shift to the fields Note that a Gysel cell adds a put into it, independently if it is working as R-node or S-node, as can be extracted from (6). In order to adjust the behavior of a C-BFN composed by Gysel cells to an Ideal C-BFN, is necessary to introduce, in each layer, an additional phase shift to the outer branches of the implemented C-BNF. An additional is introduced to these branches by adding extra phase of length. Thus, all outputs of each layer will be in-phase and the behavior of a Gysel cell based C-BFN can be predicted as is made in an ideal C-BFN. Our main objective, at time of implementation is to experimentally verify the ideal behavior of a C-BFN and its possible applications in antenna field. Specifically, we want to explore the possibility of made beam-steering of an by array only using four input-ports. Taking in mind this goal and once defined the prototype to use, we propose a reduced system of 3 by 3 antenna array fed by four input ports. In order to feed a 3 by 3 antenna array using only four inputports is necessary to use C-BFN of three outputs and two inputs, that is, C-BFN of one layer. In total, there will be used five C-BFN of one layer interconnected among each other, to finally have four input-ports to steering the beam. This configuration and . can be extracted from Fig. 4. for A prototype was build, using microstrip technology, designed to work at 2.9 GHz. The design is modular based, so, each individual C-BFN was printed on glass fiber slab ( 4.5) and assembled one with another one using SMA Jack-to-jack connectors. The whole prototype is shown in Fig. 6. As radiator for the antenna array were used nine Quasi-Yagi in vertical plane antennas [12]. The radiators are separated in horizontal plane. To assure best radiation charand acteristics, antenna characteristics were optimized. Besides, for implementation procedures, antennas are embedded to C-BFN slabs, so, three Quasi-Yagi antennas are printed join with fed circuit in each slab, see Fig. 6. IV. MEASUREMENTS AND RESULTS There are three experimental setups to work with, in order to verify beam steering properties of C-BFN and antenna array applications. Thus, the main beam is configured to point to boresight, to be steer in or plane and to an arbitrary direction (45 2492 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 55, NO. 9, SEPTEMBER 2007 Fig. 8. Return loss measured for the frequency range of observation. The color strip shows good matching of C-BFN system in each one of setups proposed. Fig. 6. Antenna array prototype. The array is conformed by nine Quasi-Yagi antennas fed by five C-BFN interconnected to each others by SMA connectors. The prototype is printed on fiber glass slabs of " 4.5. Fig. 9. Radiation pattern measured and calculated for C-BFN system antenna array in bore-sight direction (setup 1, ' 0 ). The prototype input-ports are feeding using (7) and simulated 3 by 3 antenna array are feeding using complex amplitudes defined in (8). Calculated and measured radiation pattern demonstrates good approximation between theory and experimental results for C-BFN. A. Setup 1: Bore-Sight For bore-sight feed, the port must be divided in four equal phase signals. So, phase-shifter tool is set to deliver the same phase at each branch, then, feed matrix is defined by (7) Fig. 7. Measurement setup. deg, plane). Measurement results obtained are then compared to ideal ones to verify the predicted behavior of system. Measurements are realized in an anechoic chamber, using a network analyzer (NA). The measurement setup is as shown in is used the Antenna array fed by the C-BFN Fig. 7. As a system and as a horn antenna that works in the same frequency band as antenna array. Note that in measurement setup we have, from NA, only two and purposes respectively. So, if we available ports for want to feed four input-ports of our C-BFN System, is necessary to additionally build a phase-shifter tool. The phase-shifter tool is a system with one input that is designed to deliver four outputs with equal power and the necessary phases to carry out each one of experiments proposed here. In Fig. 8, the matching of the complete system for each experimental setup proposed, that is, the C-BFN System plus phase-shifter tools. In the range of measurements from 2.8 to 3 GHz, see color strip, the Return validating our measurement Loss measured is less than setup. Were amplitude is normalized to 1 at each input-port and, at each branch, is necessary zero-phase to define the phase-plane to achieve a bore-sight beam direction. To calculate the complex amplitude that fed the antenna array is easy using (1) to (4). The antenna array feed matrix is (8) So the plane defined by phases at input-ports, as in matrix (7), is translated to antenna feeds, as is shown in matrix (8), and bore-sight radiation is assured at this point. The power present at antenna input is concentrated toward its central element, given the Binomial shape expected for amplitude distribution at this layer. Measured and calculated radiation pattern are shown in Fig. 9, for a plane defined by 0 . The agreements corroborate the bore-sight working of C-BFN System. B. Setup 2: Beam Steering to or Plane In order to steer the C-BFN System antenna beam in plane is necessary to setup the input matrix to define a phase-plane BETANCOURT AND DEL RÍO BOCIO: A NOVEL METHODOLOGY TO FEED PHASED ARRAY ANTENNAS Fig. 10. Radiation pattern measured and calculated for C-BFN system antenna array steering to E plane (setup 2, ' 0 ). The prototype input-ports are feeding using (9) and simulated 3 by 3 antenna array are feeding using complex amplitudes defined in (10). Calculated and measured radiation pattern shows scan angle obtained for this configuration of about 14 . 2493 Fig. 12. Radiation pattern measured and calculated for C-BFN system antenna array steering to a 45 phase-plane (setup 3, ' 45 ). The prototype input-ports and antenna array are feeding using (12) and (13) respectively. The scan angle obtained for calculated and measured radiation pattern is about 13 for a 45 phase-plane. plane, so, we can define the scan angle of the beam as a function of this phase and the physical distance between elements, as follows. (11) Fig. 11. Radiation pattern measured and calculated for C-BFN system antenna array steering to H plane (setup 2, ' 90 ). The prototype input-ports are feeding using the transpose of (9). Calculated scan angle is about 21 and measured scan angle is approximately 12 . that points to desire plane. The better way to define this phase plane is introducing a phase-shift between columns. Therefore, the phase-shifter tool is configured to deliver 0 phase at row 1 and 90 phase at column 2. Input matrix is defined by (9) On the other hand, to achieve a phase-plane to beam steering in Plane, the input matrix must be the transpose of (9), this is, the phase –shifter tool must be configured to deliver 0 phase at row 1 and 90 phase at row 2. After propagate the signal throughout our C-BFN System, using (1) to (4) we can estimate the resulting amplitude and phase necessaries to feed the 3 by 3 antenna array, as follows: (10) Note that resulting phase-shift in (10) is a sampled (interpolated) version of (9). Figs. 10 and 11 show the radiation patterns of a calculated and measured configuration of a beam steering in and plane, respectively. From radiation point of view, we are introducing a progressive phase-shift of 45 between neighborhood elements in or With the scan angle obtained, the wave-number at free space and the physical separation between two radiating elements or antennas. So, for a progressive phase-shift of 45 and a separain -plane and in -plane, tion between antennas of the scan angle is about 14.5 and 21 , respectively. Looking in detail the radiation pattern measured for C-BFN System steering to -plane, shown in Fig. 11, is evident a mismatch between measured and calculated values due to fabrication inaccuracy. This mistakes causes the main beam be deviate about 10 . More refined building process can dismiss both pointing and shape errors presented in this prototype. C. Setup 3: Beam Steering to a 45 Phase-Plane Using the same procedures described before in setup 1 and 2 we can define a 45 phase-plane at input-ports and at antenna feed, as follows: (12) (13) Here, as in -plane and -Plane cases, amplitudes are concentrated in central node but there are a small lose of power as a result of interaction among fields on R-nodes of C-BFN System. In Fig. 12 is shown the calculated and measured radiation patterns for a 45 phase- plane. Using (11), with equal to 22.5 , from (13). The scan angle obtained in and plane will be approximately 7.1 and 10.5 respectively. Composing these angles to a 45 cut plane, the scan angle obtained is about 13 . This scan angle is obtained by our C-BFN System antenna, corroborating its good operation. 2494 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 55, NO. 9, SEPTEMBER 2007 REFERENCES Fig. 13. Copolar polarized radiation power pattern measured for the C-BFN system antenna array. Main beam steering to (a) bore-sight; (b) E-plane; (c) H-plane and (d) 45 plane. Finally, the fully automatic measurement setup gives us the complete power radiated by each one of cases proposed for our experiment. Thus, in Fig. 13 is shown the copolar polarized radiation power measured of the C-BFN system. In this figure is easy to see the main beam steering to -plane, -plane and to an arbitrary direction (45 plane). For these figures goes from to 90 and goes from 0 to 360 . V. CONCLUSIONS In this paper was introduced a novel methodology to scan a beam radiated from phased antenna array. With this methodology, the beam obtained can be steered to any desired direction using only four input-ports. This methodology can be used as alternative for feed phased array antennas in order to reduce complexity of the BFN and associated electronic devices to control the beam steering of a Phased Array. A prototype was designed, fabricated and measured as proof of concept of C-BFN applied to Antenna Arrays. In general terms, behavior of C-BFN Systems agrees with proposed theory concepts. ACKNOWLEDGMENT The authors thank J. M. G. Arbesu and his team at the Polytechnic University of Cataluña, for measurement facilities. [1] R. J. Mailloux, Phased Array Antenna Hand Book. Norwood, MA: Artech House, 1994. [2] R. Garg, P. Bhartia, I. Bahl, and A. Ittipiboon, Microstrip Antenna Design HandBook. Norwood, MA: Artech House, 2001. [3] R. C. Hansen, Phased Array Antennas. New York: Willey-Interscience, 1998. [4] G. Zoran, D. Sasa, and C. Zoran, “A K- and Ka-band vehicular phasedarray antenna,” Microw. J., vol. 47, no. 1, pp. 58–74, Jan. 2004. [5] A. K. Bhattacharyya, Phased Array Antennas: Floquet Analysis, Synthesis, BFNs and Active Array Systems. New York: Willey-Interscience, 2006. [6] L. Thomas and P. Haskell, “Phased array antenna system with controllable electrical tilt,” Supplement Australian Official J. Patents, vol. 20, no. 18, p. 1789, May 2006. [7] R. Garcia, D. Betancourt, A. Ibañez, and C. del Rio, “Coherently periodic radiation structures (CORPS): A step towards high resolution radiations systems,” presented at the IEEE Symp. on Antenna and Propagation, Washington, Jul. 3-8, 2005. [8] R. L. Scheaffer and J. T. McClave, Probability and Statistics for Engineers. Pacific Grove, CA: Duxbury Press, 1995. [9] R. C. Hansen, Microwave Scanning Antennas. New York: Academic Press, 1966, vol. III, Array Systems,, pp. 241–242. [10] D. Betancourt and C. del Rio, “Designing feeding networks with CORPS: Coherently radiating periodic structures,” Microw. Opt. Tech. Lett., vol. 48, no. 8, pp. 1599–602, Aug. 2006. [11] B. Ooi, W. Palei, and M. S. Leong, “Broad-banding technique for in-phase hybrid ring equal power divider,” IEEE Trans. Microw. Theory Tech., vol. 50, no. 7, Jul. 2002. [12] W. Deal, N. Kaneda, J. Sor, Y. Qian, and T. Itoh, “A new Quasi-Yagi antenna for planar active antenna arrays,” IEEE Trans. Microw. Theory Tech., vol. 48, no. 6, Jun. 2000. Diego Betancourt was born in Bogotá, Colombia, in 1975. He received the B.Sc. and M.Sc. degrees in electrical and electronic engineering from the University of los Andes, Colombia, in 1998 and 2002 respectively. He is currently working toward the Ph.D. degree at the Public University of Navarra, Pamplona, Spain. He is currently a Research Student with the Antenna Group at the school of Electric and Electronic Engineering at the Public University of Navarra. Carlos del Río Bocio (M’94) was born in Reus, Spain, in 1970. He received the degree of Ingeniero Técnico de Telecomunicación and the degree of Electronic Engineer from the Ramon Lull University, Barcelona, Spain, in 1991 and 1993, respectively, and the Ph.D. in telecomunicación (with honors) from the Public University of Navarra, Pamplona, Spain in 1996. Since 1993, he develops his research and docent activities in the Electric and Electronic Engineering Department at the Public University of Navarra, where he is currently an Associate Professor. His research interest include horn antenna design in general, satellite and terrestrial communications, study of periodic metallo-dielectric structures, also known as “electromagnetic band gap” structures, development of numerical computation software based on mode matching and scattering matrix, antenna measurements, far-and near field measurement chambers, and compact ranges, etc.