Collection of technical reports on the electromagnetic analysis and design of the Northern Cross oriented to the BEST (SKADS) test bed.

Transcription

1 Collection of technical reports on the electromagnetic analysis and design of the Northern Cross oriented to the BEST (SKADS) test bed. Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax This activity is supported by the European Community Framework Programme 6, Square Kilometre Array Design Studies (SKADS), contract no

2 Introduction This paper is a collection of the technical reports (reported in chronological order) realized in collaboration with the IRA-INAF (Istituto di RadioAstronomia Istituto Nazionale di AstroFisica), within the project SKADS/BEST. Index EM Analysis of Northern Cross E/W 48 MHz..3 (3 rd November 26) Upgrade of the Northern Cross E/W Arm for the MHz frequency band: general investigation and fat dipole solution...17 (1 th April 27) Upgrade of the Northern Cross E/W Arm for the MHz frequency band: dense array solution...39 (26 th April 27) Upgrade of the Northern Cross E/W Arm for the MHz frequency band: log periodic antenna solution..49 (19 th July 27) Measurements on the custom log periodic antenna for the MHz upgrade of the Northern Cross E/W Arm..71 (2 th December 27) Upgrade of the Northern Cross E/W Arm for the MHz and the MHz frequency bands: multi-feed solution..79 (1 st February 28) 2

3 Technical Report EM Analysis of Northern Cross E/W 48 MHz (Draft 2-EA/6113) Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 3 rd November 26 3

4 This document deals with the electromagnetic analysis of the Northern Cross E/W arm at the actual working frequency of 48 MHz. The antenna is composed of a main cylindrical offset parabolic reflector and a feeding structure. The latter, which is also called focal line, consists of a linear dipole array placed in a subreflector. With reference to the nominal geometrical dimensions, both radiation patterns of the feeding structure and the whole system are provided. The corresponding results are compared to the ones reported in [1]. Further simulations accounting for the real measured wire profile are also provided. Finally, design guidelines to enhance the efficiency of the existing system are discussed. The electromagnetic analyses have been carried out using a specific numerical code. This tool, which was entirely developed by IEIIT, is based on the solution of the Pocklington equations by means of the method of moments. Arbitrary 2-dimensional wire structures can be characterized with this method. Since a 2D analysis technique has been exploited, the structures and sources are considered infinite in the wire direction. Moreover, every radiation pattern corresponds to the H-plane of the antenna. The antenna parameters in the 2D case are: Directivity dp( Rˆ) D ( Rˆ) = dϑ P irr 2π where dp ( R ) dϑ is the radiated power density in the Rˆ direction and P irr is the total radiated power per unit length (W/m). Effective width D( Rˆ) λ W eff ( Rˆ) = 2π where λ is the wavelength. When the Rˆ -dependence of the directivity or efficiency is understood, maximum values are indicated. Width efficiency Weff η = W geo where W is the width the antenna aperture projection along the focal axis. geo Phase center (for the subreflector only) The phase center d is defined as the value that maximizes the functional F 2 ϑill / 2 irr jkod cos( ϑ ϑmax ) ( ϑill, d) E ( ϑ ϑmax ) e dϑ ϑ / 2 =, where ϑ ill is the full illumination angle (it ill depends on the main reflector geometry) and ϑ max is the angular position of the maximum irr of the subreflector radiated field E. The functional F (d) is proportional to the directivity D of the overall antenna system (subreflector-main reflector). 4

5 Analysis of the subreflector with nominal dimensions The geometry of the subreflector is depicted in Fig. 1. The nominal dimensions are the ones reported in [1]. The corresponding wire distribution and the source and phase center locations are shown in Fig. 2. The phase center d is.48 m far from the aperture. It has been computed according ϑ 75. to the definition presented in the previous section with = ill In [1], the computed phase center is.75 m far from the aperture. This discrepancy arises from the coarser definition adopted in [1]. If the definition in [1] had been used, the computed phase center would have been.65 m far from the aperture. The estimated directivity loss with respect to the correct definition (see previous section) is anyhow less than.5. Figure 1. Subreflector with nominal dimensions (m). Wire distribution, N (number of wires)= source phase center m.5 d m Figure 2. Wire distribution of the subreflector. The wire spacing is 2 mm, the wire diameter is.5 mm. 5

6 The computed radiation pattern of the structure is shown in Fig. 3. At ±3 the value of the normalized directivity is -13.6, in good agreement with [1]. Max Directivity= 1.52, Max.Pos.=., FHPBW = 3.21 Radiation Pattern () Figure 3. Primary radiation pattern (normalized directivity) in the H-plane (radiation pattern of the line source in presence of the subreflector). 6

7 Analysis of the whole antenna system with nominal dimensions and using a parabolic profile for the main reflector The nominal geometry of the main cylindrical offset parabolic reflector is depicted in Fig. 4. The dimensions were looked up from [2]. The reflector width W, perpendicular to the focal axis, is 29.4 m. The focal length is 16.5 m, the feed direction is 55 from the focal axis and the full illumination angle ϑ ill (angle seen from focus) is 75. The corresponding wire distribution is shown in Fig. 5. No approximation of the parabolic profile and a 2-mm wire spacing were used in the simulation. geo Figure 4. Nominal dimensions of the cylindrical offset parabolic reflector. The dimensions are in meters. Wire distribution, N (number of wires)= m m Figure 5. Wire distribution of the overall antenna system. The wire spacing is 2mm and the wire diameter is.5 mm. The center of the reference system is placed in the parabola focus. 7

8 The illumination distribution (normalized incident field) along the main reflector has been computed and reported in Fig. 6. The left side of the abscissa corresponds to the furthest wires from the vertex of the parabola. The edge taper is and -21 at the left and right side of the reflector depicted in Fig. 5, respectively. Illumination Distribution wire number Figure 6. Illumination distribution (normalized incident field) along the wires of the main reflector. Wire number 1 is located in the upper left corner of Fig. 5. The radiation pattern is shown in Figs. 7 and 8. The secondary lobe level is below for each direction owing to the high edge taper. The and 18 directions correspond to the zenith and the ground, respectively, when the antenna is pointed toward the zenith direction. According to Fig. 5, the 9 and -9 correspond to the horizon in the northern and southern directions, respectively. Max Directivity= 22.48, Max.Pos.=.1, FHPBW = 1.88 θ= Radiation Pattern () θ=9 θ θ=18 θ= Figure 7. Radiation pattern (normalized directivity) of the whole antenna system (secondary pattern) in the H-plane. 8

9 Max Directivity= 22.48, Max.Pos.=.1, FHPBW = 1.88 Radiation Pattern () Figure 8. Zoom of the radiation pattern (normalized directivity) of the whole antenna system in the H-plane. The Full Half Power Beam Width (FHPBW) is 1.88 and the maximum directivity is The latter leads to an antenna width efficiency η =.7. Finally, the comparison between primary and secondary radiation patterns in Figs. 9 and 1 show that the secondary lobes which do not belong to the back of the reflector are produced by the subreflector. Primary and Secondary Radiation patterns Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position θ=9 θ θ= θ= 9-4 θ=18-6 Back of the reflector Figure 9. Comparison between primary and secondary radiation patterns in the H-plane. 9

10 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular 3 Position θ= θ θ=9 θ= 9 θ= Figure 1. Comparison between primary and secondary radiation patterns in the H-plane. 1

11 Analysis of the whole antenna system with the real wire profile The real geometry of the whole antenna system has been looked up from [3], where measured data are stored. The subreflector real profile is compared to the nominal one (from [1]) in Fig. 11. It should be noted that a slight aperture reduction (about 5 mm) as well as a tilt angle (about.5 ) are present in the measured profile m 16.5 Measured wire distribution Source Nominal wire distribution m Figure 11. Wire distribution of the subreflector. The real wire spacing is approximately 1 mm (2 mm) in the part that is closer to (further from) the source. m Measured Profile Nominal Vertex (,) m Nominal Focus (,16.5) m Best Fit Vertex =(-11.3,44.8) mm Best Fit Focus =(-6.5,6.6) mm Best Fit Parabola θ=-.365, e RMS =6.5 mm, e MAX =24.3 mm m Figure 12. Wire distribution of the main reflector: measured profile and best fit parabola. 11

12 Fig. 12 shows the measured main reflector profile. The wires are arranged in a piecewise fashion with 35 segments. The spacing is 2 mm for the first 25 bars that are closer to the vertex and 3 mm for the others. The origin of the measurement coordinate system coincides with the nominal vertex of the parabola. Therefore the nominal focus position is (,16.5) m. A least square approximation has been carried out. The resulting best fit parabola is depicted in Fig. 12 as well. As reported in the figure legend, a slight tilt angle of about -.4 in the parabola axis is present. Moreover, the focus and vertex of the best fit parabola have different positions with respect to the nominal ones. The discrepancies between nominal and estimated positions for the focus and vertex might be due to either a shape error of the reflector or to the measurement uncertainties. In order to compute the EM characteristics of the real structure, a simulation has been carried out. The results of the latter and the nominal case ones (see previous section) are both reported in Fig. 13. It should be noted that the two diagrams are practically coincident at every angle but around the maximum, where the secondary lobes of the measured profile case are higher. This phenomenon is probably related to the piecewise approximation of the parabolic profile. Max Directivity= 22.58, Max.Pos.=.3, FHPBW = 1.8 Radiation Pattern () with measured profile with nominal profile θ=9 θ θ= θ=18 θ= Figure 13. Radiation pattern computed from the measured wire profile and from the nominal wire profile (H-plane). The close match between the two radiation patterns in Fig. 13 demonstrates that both manufacturing and measuring uncertainties are not significant at this operative frequency (48 MHz). Therefore, every subsequent design study concerning this antenna will be carried out on the basis of the nominal dimensions of the structure. 12

13 Optimization of the 48 MHz In order to enhance the overall antenna width efficiency η a design study has been carried out on the subreflector, only varying the aperture dimension. The optimized geometry is shown in Fig. 14. The angle between the two arms is the same as in nominal case, whereas the aperture is reduced of 74%. Hence, the subreflector length is reduced as well. If the phase center of the new structure (.15 m from the aperture) is placed in the focus of the main reflector, with the same 55 direction, the illumination distribution shown in Fig. 15 and the radiation pattern reported in Fig. 16 and 17 can be obtained. 1.5 Wire distribution, N (number of wires)= 196 source phase center 1 m m Figure 14. Geometry of the optimized subreflector. Illumination Distribution wire number Figure 15. Illumination distribution of the optimized subreflector. 13

14 Thanks to the lower edge taper, the antenna efficiency η is now approximately.85 and the FHPBW is In the nominal case, the same values were η =.7 and FHPBW=1.88, respectively. The secondary lobes that are far from the maximum do not exhibit any increase. The first ones instead raise up to, as shown in Fig. 17. If only the aperture reduction had been performed, without refocusing the feed, a width efficiency η =.8 would have been obtained. Max Directivity= 23.28, Max.Pos.=.1, FHPBW = 1.54 Radiation Pattern () with optimized feed with nominal feed θ=9 θ θ= θ=18 θ= Figure 16. Radiation pattern of the whole system (secondary radiation pattern) in the H-plane for the optimized and nominal case. Max Directivity= 23.28, Max.Pos.=.1, FHPBW = 1.54 with optimized feed with nominal feed Radiation Pattern () Figure 17. Zoom of the radiation pattern in the H-plane for the optimized and nominal case. 14

15 Primary and Secondary Radiation patterns Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position θ θ= -4 θ=9 θ= 9 θ= Figure 18. Primary and secondary radiation patterns in the H-plane for the optmized case Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. 3 Angular Position θ= θ=9 θ θ= 9 θ= Figure 19. Primary and secondary radiation patterns in the H-plane for the optmized case. Finally, the comparison between primary and secondary radiation patterns in Figs. 18 and 19 show that the secondary lobes which do not belong to either the back of the reflector or the maximum area are produced by the subreflector. 15

16 Conclusions The EM analysis reported in this document demonstrated the good characteristic of the Northern Cross E/W Arm at 48 MHz. Moreover, excellent match has been achieved between the simulations that were performed on the nominal and measured wire profiles. Therefore, every subsequent design study concerning this antenna will be carried out on the basis of the nominal dimensions of the structure. A possible efficiency enhancement from η =.7 to η =.85 has been demonstrated. It entails a reduction of the subreflector wires and a subsequent refocusing. Bibliography [1] Stephen N. La and L. James, Analysis of the Bologna Cross Radio Telescope, CSIRO Division of Radiophysics, 5 April 199 [2] Drawing K 52/9abcd, Centine per Radiotelescopio SAE S.p.A. Milano, Università di Bologna, Istituto di Fisica, 2 Febbraio 1962 [3] punti e-w.xls, 26 Ottobre 26 Microsoft Excel Worksheet 16

17 Technical Report Upgrade of the Northern Cross E/W Arm for the MHz frequency band: general investigation and fat dipole solution (Draft 1-EA/741) Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 1 th April 27 17

18 This document contains a general investigation for the MHz frequency band upgrade of the Medicina Northern Cross E/W arm. As described in [1], the E/W arm of the Northern Cross is a cylindrical offset parabolic reflector with a feeding structure made up with a linear dipole array placed in a wire subreflector. The current antenna working frequency is 48 MHz, corresponding to a wavelength of.735 m. At this frequency, the aperture of the present subreflector (1.69 m) is larger than two wavelengths providing a high edge tapering (more than 2 ) on the main cylindrical reflector. A new feed system working in the MHz frequency bandwidth has been studied with the same dipole and wire subreflector configuration. This structure is used to show the concepts and the critical aspects of a broadband reflector feed design as well as to demonstrate the incompatibility of the MHz working condition with the present subreflector dimensions. As a matter of fact, the wavelength that corresponds to the lower end of the MHz band is 2.5 m. In this case, a scaled version of the present subreflector would have a 5.75 m aperture. This value is obviously critical from the structural and manufacturing points of view. Moreover, it would also provide a 2 edge tapering which is not the optimum condition in terms of the overall antenna efficiency. In this work, the subreflector aperture is instead selected to obtain a minimum size structure which still provides a 5 8 edge tapering on the main reflector. However, the resulting value of 3 m is still far larger than the present 1.69 m. The design of antenna arrays with broadband operative conditions in both radiation pattern and impedance is not trivial. Both these features are required in order to feed a reflector with high efficiency in the overall frequency bandwidth. Owing to its complexity, the design problem has been attacked from different perspectives. A starting design was carried out with the 2-D analysis method described in [1]. The subreflector dimensions were determined to provide a proper edge tapering in the overall frequency bandwidth and according to reasonable dimensional constraints. The obtained geometry and the corresponding H-plane radiation patterns are reported in section 1. Subsequently, a 3-D method was used to design the linear array of radiators inside the subreflector. Single radiator dimensions as well as element spacing and optimum placement in the subreflector were obtained. The frequency behavior of the relevant quantities such as reflection coefficients, mutual couplings, feed tapering and the corresponding radiation patterns are reported in section 2. Finally, the radiation pattern of the whole antenna system, composed of the presented feed system and the main cylindrical reflector, are reported in section 3. 18

19 MHz Wire Subreflector Design The Pocklington s equation 2-D method [1] has been used in this first design stage to define the subreflector dimensions according to the H-plane primary radiation pattern requirements i.e. proper edge tapering of the main reflector in the MHz frequency bandwidth. The wire distribution of the designed subreflector is reported in Fig. 1. The red mark represents the position of the linear dipole array (modeled as a line source in this case). The corresponding H- plane radiation patterns at 12, 18 and 24 MHz are reported in Fig. 2. When the feed system is arranged as in Fig. 3 (see also [1]), the black dash-dot line of Fig. 2 at 37.5 corresponds to the angular position of the reflector edges in the feed polar coordinate system. As one can see from Fig. 2, the level of the feed system radiation pattern at 37.5 is 12 MHz. Since the spatial attenuation at the left and right reflector edges of Fig. 3 (normalized to the center) is approximately 2 and -1, respectively, the illumination distribution reported in Fig. 4 shows edge tapering levels of 8 and 5. It should be remembered that the illumination distribution is the incident field along the main reflector wires. Even if a slightly higher edge taper could be desirable from the efficiency point of view, the present aperture choice w=3 m (1.2 λ 1, where λ 1 =2.5 m) leads to reasonable dimensions of the feed system. A 2 d 1.5 m 1 b.5 w m Figure 2. 2D geometry of the wire subreflector. The linear array position (line source) is represented with the red mark. The aperture size w is 3 m, the total length b is 2.25 m, the size of the back side A is 1.25 m and the distance d of the dipole array from the back of the subreflector is.5 m. The wire spacing of 2 mm leads to a 34 wire distribution. The wire diameter is.5 mm. 19

20 H-plane Primary Radiation Pattern 12 MHz 18 MHz 24 MHz Figure 21.Primary radiation patterns in the H-plane computed with the Pocklington s equation method. The black dashdot line at 37.5 represents the main reflector edges when the feed system is pointed as in Fig.3. Figure 22. Feed system placement and relevant dimensions of the cylindrical offset parabolic reflector [1]. 2

21 Illumination distribution along the main reflector wires 12 MHz 18 MHz 24 MHz Wire Number Figure 23. Illumination distribution (normalized incident field) along the main reflector wires. Wire number 1 is the furthest one from the vertex of the main parabolic cylindrical reflector (left side of Fig. 3). The feed system is pointed as in Fig. 3. Another important feature is the frequency behavior of the radiation pattern. As can be noted from Fig. 2, it becomes narrower at higher frequencies (as expected). However, this narrowing effect is controlled so that the corresponding edge taper (Fig. 4) maintains proper values even at the upper end of the bandwidth. This phenomenon is highlighted in Fig. 5, where the H-plane radiation pattern value at 37.5 versus frequency is reported. This quantity shows a minimum level of -13 at 22 MHz, corresponding to a maximum edge taper of 15. The particular frequency behavior of Fig. 5 has been obtained by properly selecting the other three feed system dimensions (see Fig. 1). Roughly speaking, the feed system length b=2.25 m (.9 λ 1 ) was chosen to provide a proper quadratic phase error on the feed system aperture at higher frequencies. The choice of d= λ 1 /5 and A= λ 1 /2, besides providing a good matching behavior for the fat dipole array (see section 2), leads to a high radiation efficiency of the line source inside the subreflector at 12 MHz, but to a lower one at higher frequencies. This phenomenon produces aperture quadratic phase errors as well. Both these effects broaden the H-plane radiation pattern with respect to a non-quadratic-phased aperture. It should be noted that these phenomena have been exploited without significant distortions of the pattern such as split lobes, etc. (see Fig. 2). From Fig.2, secondary lobe levels and back radiation levels of the order of and, respectively, can be also estimated. Finally, the distance of the computed feed system phase center [1] from the aperture is reported in Fig. 6 as a function of frequency. The variation is not very large and, as can be noted in section 3, it does not produce any relevant defocusing loss on the whole antenna efficiency. 21

22 Feed Tapering at GHz Figure 24. Value of the Primary Radiation pattern at 37.5 in the H-plane vs frequency. 7 Phase center 6 5 mm GHz Figure 25. Distance of the computed phase center [1] from the aperture vs frequency. 22

23 2. Design of the Fat Dipole Linear array inside the Wire Subreflector A fat dipole linear array has been designed inside the subreflector of section 1, in order to provide a minimum complexity feed system solution. The radiator design as well as their optimum placement inside the wire subreflector has been carried out using a 3D EM simulator. Seven elements of the designed array configuration are shown in Fig. 7. Larger finite structures were also extensively simulated finding out that the relevant electromagnetic parameters can be already extrapolated from this 7-element array. For the sake of simulation simplicity, the subreflector wire distribution was substituted with a continuous PEC surface. Obviously, the agreement with the wire configuration has been verified. The fat dipoles were simulated as solid cylinders but, owing to their size, they can be built with metal pipes. The dipole length L =.42λ 1 = 1.5 m, the diameter is.1l=15 mm, the source gap (gap between the two dipole arms) is.3λ 1 =75 mm Figure 26. Seven fat dipole array inside the subreflector with discrete excitation ports. The period of the structure is λ 1 /2=1.25 m. This value is a compromise between the mutual coupling at the lower frequencies and the grating lobes at the higher frequencies. Mutual couplings and reflection coefficients at 12, 18 and 24 MHz i.e. the diagonal and offdiagonal terms of the array scattering matrix are reported in Figs. 8, 9, 1, respectively. The performances of both central and edge elements can be seen in these figures. The reflection coefficient 1 at 12 MHz (Fig. 8) is about -9 for the central elements 3, 4 and 5. This quantity increases for the edge elements owing to the truncation effect of both the array and subreflector. The mutual coupling between adjacent elements is for the central elements and slightly higher for the edge ones. The coupling between non-adjacent elements is of the order of -17. Thanks to this, the central elements 3, 4 and 5 are so slightly affected by the truncation effect that their behavior is independent of the array size. In other words, the data computed for these elements are still valid for larger arrays. 23

24 Analogous considerations apply for the higher frequencies (see Fig. 9 and 1). However, since the mutual coupling between adjacent elements quickly decreases with frequency (it is -19 and -26 at 18 MHz and 24 MHz, respectively), the truncation effect is less significant. Reflection coefficient of the central element as well as mutual coupling versus frequency is reported in Fig. 11. The reflection level is lower than -9 in the operative bandwidth. The mutual couplings are below for frequencies above 12 MHz. The phase of the reflection coefficient and the adjacent element mutual coupling is shown in Fig. 12. As can be observed, these terms are nearly counter-phased in the overall frequency bandwidth. Reflection Coefficients and Mutual 12 MHz S 1n S 2n S 3n S 4n S 5n S 6n S 7n Radiator Number n Figure 27. Reflection coefficients and mutual couplings for the 7 dipole array of Fig. 12 MHz. The reference impedance is 15 Ω. 24

25 Reflection Coefficients and Mutual 18 MHz S 1n S 2n S 3n S 4n S 5n S 6n S 7n Radiator Number n Figure 28. Reflection coefficients and mutual couplings for the 7 dipole array of Fig. 18 MHz. The reference impedance is 15 Ω. Reflection Coefficients and Mutual 24 MHz S 1n S 2n S 3n S 4n S 5n S 6n S 7n Radiator Number n Figure 29. Reflection coefficients and mutual couplings for the 7 dipole array of Fig. 24 MHz. The reference impedance is 15 Ω. 25

26 Reflection coefficient and mutual couplings (15 Ohm) -4 S44 (central element) S34 S24 S14 (end/central) GHz Figure 3. Magnitude of the reflection coefficient of the central element and mutual coupling to the other elements versus frequency. The reference impedance is 15 Ω Reflection coefficient and mutual couplings (15 Ohm) S44 (central element) S GHz Figure 31. Phase of the reflection coefficient of the central element and mutual coupling to the adjacent element versus frequency. The reference impedance is 15 Ω. 26

27 The previous characterization of the finite array is based on single element excitation. However, another way of studying antenna arrays is the infinite array approach. In this case, the unit cell of the periodic structure (i.e. infinite array) is analyzed for different incidence angles. The reflection coefficient of the unit cell for broadside incidence is reported in Fig. 13 in comparison with the reflection coefficient of the central element of the finite array. As expected, these quantities differ from each other because they refer to two different working conditions. From the receiver design point of view, the first can be interpreted as the reflection coefficient at the input port of a corporate beam forming network (for broadside operation) connected to the array. The second is instead the reflection coefficient at the input port of one of the array elements. Hence, the second parameter should be considered for the digital beam forming network receiver design. -2 Reflection coefficient (15 Ohm) Infinite array (broadside) 7 element array (central element) GHz Figure 32. Reflection coefficient of the infinite array unit cell and for single excitation in the 7-element array. The reference impedance is 15 Ω. The radiation pattern for each array element has also been computed. The H-plane cuts at 12, 18 and 24 MHz are shown in Figs. 14, 15 and 16, respectively. The H-plane patterns computed with the Pocklington method (same as Fig. 2) are reported in the corresponding figures for comparison. The discrepancies between the various patterns are due to the diffracted field arising from the edges of the array. Nevertheless, apart from the edge elements, the various patterns are still consistent with the Pocklington simulation in terms of feed system tapering at 37.5, profile of the secondary lobes and front-to-back ratio. Therefore, the designed structure can be used to efficiently feed the main reflector. 27

28 H-plane Primary radiation MHz Element 1 and 7(end) Element 2 and 6 Element 3 and 5 Element 4 (central) Pockligton Figure 33. H-plane radiation patterns for the various antennas of the 7-element MHz. H-plane Primary radiation MHz Element 1 and 7(end) Element 2 and 6 Element 3 and 5 Element 4 (central) Pockligton Figure 34. H-plane radiation patterns for the various antennas of the 7-element MHz. 28

29 H-plane Primary radiation MHz Element 1 and 7(end) Element 2 and 6 Element 3 and 5 Element 4 (central) Pockligton Figure 35. H-plane radiation patterns for the various antennas of the 7-element 24 MHz. The E-plane radiation patterns are reported in Figs For the sake of readability, only the first four elements are shown. The patterns of the last three elements can be obtained by mirroring the first three with respect to the vertical axis at. Even in this case, the discrepancies between the various curves are due to the truncation effect. In particular, the E-plane pattern of the first and last elements is very different from the central ones. Therefore, in a real installation, these edge radiators should be considered as dummy elements i.e. loaded with their characteristic impedance 15 Ω but without the necessity of a receiver. As far as the five central elements are concerned, the ripple of the patterns is due two both the truncation effect and the mutual coupling between dipoles. These patterns have been averaged (average directivity) to obtain the Field of View of the array at 12, 18 and 24 MHz. The results reported in Fig. 2 exhibit attenuation values of 2-3 at 3 in the operative frequency bandwidth. 29

30 E-plane Primary radiation MHz Element 1 (end) Element 2 Element 3 Element 4 (central) Figure 36. E-plane radiation patterns for the various antennas of the 7-element 12 MHz. E-plane Primary radiation MHz Element 1 (end) Element 2 Element 3 Element 4 (central) Figure 37. E-plane radiation patterns for the various antennas of the 7-element 18 MHz. 3

31 E-plane Primary radiation MHz Element 1 (end) Element 2 Element 3 Element 4 (central) Figure 38. E-plane radiation patterns for the various antennas of the 7-element 24 MHz. E-plane Mean Primary radiation pattern 12 MHz 18 MHz 24 MHz Figure 39. Field of view of the array element. The values at 3 (red dash-dotted line) are -1.9, -2.2 and -3.1 at 12, 18 and 24 MHz, respectively. 31

32 3. Analysis of the E/W arm with the new MHz feed system The 2D analysis method was used to compute the H-plane radiation patterns of the overall antenna system. Since the H-plane radiation patterns of the single array element (see Fig ) computed with the 3D method were consistent with the 2D ones, the latter were used for the final analysis. In this way, even the interaction between the sub and main reflectors has been taken into account. The wire distribution of the entire system is depicted in Fig. 21. The feed system direction is 55 from the focal axis as in Fig. 3. The phase center of the feed system at lower frequencies (.2 m far from the aperture) is placed in the focus of the parabolic reflector. Wire distribution, N (number of wires)= m m Figure 4. Wire distribution of the antenna system with the feed system (upper right corner) and main the cylindrical offset parabolic reflector. The computed H-plane patterns are shown in Figs The blue line refers to the whole antenna system (secondary pattern), the red line is the feed system pattern (primary pattern) and the green one is the main reflector scattered field. Since the feed system pattern is computed in presence of the main reflector, the interaction leads to a non symmetric primary pattern with respect to feed system pointing direction. Owing to the low edge tapering at 12 MHz (see Fig. 4), the secondary lobes and the spillover lobes in Figs. 22 and 23 are approximately -17. These quantities are instead below -26 at higher frequencies. 32

33 freq=.12 GHz, η=.846, Max Directivity= 17.96, Max.Pos.= -.12, FHPBW = 4.81 Spillover lobes -4 θ=9 θ θ= θ=18 θ= Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Figure 41. H-plane radiation patterns of the whole antenna MHz Primary pattern Main Refl. Scattered Field Secondary Pattern 3 Main Refl. Angular Position Figure 42. H-plane radiation patterns of the whole antenna MHz. 33

34 freq=.18 GHz, η=.844, Max Directivity= 19.71, Max.Pos.= -.1, FHPBW = 3.52 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position θ θ= -4 θ=9 θ= 9 θ= Figure 43. H-plane radiation patterns of the whole antenna MHz Primary pattern Main Refl. Scattered Field 6 Secondary Pattern 3 Main Refl. Angular Position Figure 44. H-plane radiation patterns of the whole antenna MHz. 34

35 freq=.24 GHz, η=.84, Max Directivity= 2.94, Max.Pos.= -.1, FHPBW = 2.66 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position -4 θ=9 θ θ= θ=18 θ= Figure 45. H-plane radiation patterns of the whole antenna MHz Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Figure 46. H-plane radiation patterns of the whole antenna MHz. The other antenna parameters such as directivity (see [1]), Full Half Power Beam Width (FHPBW) and efficiency are reported in Figs. 28 as a function of frequency. As one can see, no gap in the frequency behavior is present providing a real broadband operation condition. The ripple of the various curves is due to the interaction between sub and main reflectors. 35

36 Since the efficiency (reported in Fig. 3) is good in the whole bandwidth (η>.8), the directivity at the upper frequencies (21 ) is 3 higher than at the lower ones (18 ). The FHPBW (Fig. 29) is approximately 4.8 and 2.6 at 12 and 24 MHz, respectively. The curve of the antenna efficiency reported in Fig. 3 is related to the frequency behavior of the feed system tapering in Fig. 5. At the lower end of the bandwidth, the efficiency increases with frequency because the feed system directivity increases, reducing the spillover losses. The minimum at 22 MHz corresponds to the maximum feed system tapering of Fig Directivity Antenna Gain GHz Figure 47. Directivity [1] of the complete antenna system. 36

37 5 FHPBW GHz Figure 48. Full Half-Power Beam Width of the complete antenna system. 88 Aperture Efficiency % GHz Figure 49. Efficiency of the complete antenna system. 37

38 Conclusions A preliminary upgrade study for the illumination of the Northern Cross E/W arm in the MHz frequency band has been presented. It showed good performances with a minimum complexity structure. In the whole frequency bandwidth, the overall antenna efficiency in the H- plane is better than.8, the H-plane HPBW decrease from 4.8 to 2.6, the E-plane Field of View is about ± 3 for pattern attenuation lower than 3, the reflection coefficient is better than -9 when a 15 Ω impedance is used. Mutual coupling between the array elements has also been estimated providing a upper bound. Thanks to this value, only two dummy elements are needed in the final array configuration. Other radiator geometries will be presented in the following reports. References [1] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, EM Analysis of the Northern Cross E/W 48 MHz IEIIT Technical Report EA

39 Technical Report Upgrade of the Northern Cross E/W Arm for the MHz frequency band: dense array solution (Draft 1-EA/7426) Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 26 th April 27 39

40 A preliminary solution for the MHz upgrade of the Northern Cross E/W arm has been proposed in [1]. In such design study, fat dipoles were arranged in a sparse array configuration i.e. the inter-element spacing of the radiators was larger than half wavelength at the highest working frequency [2]. It was in fact half wavelength at the lowest frequency (λ 1 /2=1.25 m). Hence, grating lobes occur with that configuration at the higher frequencies, when the array beam is scanned away from broadside. A dense array configuration is instead investigated in the present document. It is based on a branched Vivaldi linear array arranged inside a subreflector. The inter-element spacing is λ 1 /4.5 =.556 m providing a grating-lobe free MHz frequency bandwidth for each scanning angle. Moreover, the designed array also operates in the present bandwidth of the Medicina Radiotelescope [3] centered at 48 MHz. The λ 1 /4.5 inter-element spacing guarantees a grating-lobe free condition at broadside up to 43 MHz and more. The 3D view of the proposed solution is reported in Fig. 1. As one can see, the branched Vivaldi radiators are connected to each other. Moreover, they are also directly connected to the subreflector through a metal plate. Circular stubs are drilled on this plate to create wideband open circuits. The excitation is provided at the junction between the circular stub and the tapered slotline section of each Vivaldi radiator. Thanks to the circular stub, even an unbalanced (coaxial cable) excitation can be directly connected to the radiators. The subreflector is simulated as a PEC surface. Its dimensions are reported in the side view of Fig. 2. Excitation Tapered slotline Metal plate with circular stubs Subreflector Branched Vivaldi radiators Figure 5. Linear array of branched Vivaldi radiators inside the subreflector. 4

41 W s L rad L sub L back Figure 51. Side view of the branched Vivaldi linear array inside the subreflector. The aperture W is equal to 1.2 λ 1 = 3 m (as in [1]), the length of the subreflector section L sub is equal to 1 λ 1 = 2.5 m, the radiator length L rad is equal to.82 λ 1 = 2.5 m. The transition section is A= λ 1 /4.5=.556 m wide and L back = λ 1 /6=.417 m long. The width of the radiator branches s is.2 λ 1 = 5 mm. A Such subreflector dimensions are comparable to the ones reported in [1]. It should be noted that the length of the radiators L rad =.82 λ 1 = 2.5 m is less than the subreflector one L sub = λ 1 = 2.5 m. The lower part of Fig. 2 shows the transition section wherein the circular stubs are placed. The slotline tapering starts at the interface between the transition and the subreflector sections. The other relevant dimensions are reported in Fig. 3 where only the radiators are shown. The interelement spacing D is equal to λ 1 /4.5 =.556 m. The branching period b is.8 λ 1 =.2 m and the branch thickness t is.1 λ 1 = 25 mm. All the various branches can be manufactured with metal bars having a 5x25 mm 2 section. It should be noted that these quantities could be modified in order to meet industrial standards for a possible final low-cost manufacturing-oriented design. 41

42 t b D h end Tapered slotline L rad Figure 52. 2D view of the branched Vivaldi linear array. The inter-element spacing D is λ 1 /4.5 =.556 m and the radiator length L rad is equal to.82 λ 1 = 2.5 m. The branching period b is.8 λ 1 =.2 m and the branch thickness t is.1 λ 1 = 25 mm. The radius of the circular stub is λ 1 /15 =.167 m. The tapering profile is elliptical and the maximum slotline size (aperture) is h end =.146 λ 1 =.365 m. The present configuration has been initially designed using an infinite array approach (where all the elements are simultaneously fed) [4]. In this framework, a physical interpretation of the antenna electromagnetic behavior based on modal analysis was possible, leading us to significant and useful design guidelines. The obtained results are reported in Fig. 4 (green curve) showing reflection levels better than -18 in the MHz frequency band and better than -13 from 1 to 44 MHz (more than 2 octaves) for broadside direction. Reflection Coefficient (9 Ohm) Infinite Array (in-phase all-element excitation) 11-Element Array (central radiator excited) GHz Figure 53. Reflection coefficient of the branched Vivaldi linear array configuration. The reference impedance is 9 Ω. The vertical black dash-dotted lines represent the two operative frequency bands MHz and MHz. However, this feeding configuration does not correspond to the relevant working condition of the present radiators. These antennas are in fact loaded with independent receivers. Therefore, a single element excitation analysis has to be carried out in order to find out the correct antenna impedance (reflection coefficient) as well as the mutual coupling coefficients. This analysis has been performed on the 11-elements array of Fig. 1. The corresponding results reported in Figs. 4 42

43 definitely show higher reflection values (up to -4 ) in the lower frequency band. This phenomenon is due to the small size of the radiating elements. The radiator size was in fact reduced to meet the grating-lobe free requirement of the array in the MHz frequency band (interelement spacing less than λ 1 /4). As expected, the single element starts to properly radiate when its size is approximately half wavelength i.e. above 24 MHz. Mutual couplings between the central element and the other array elements are reported in Figs The highest value ( ) occurs with the adjacent element, in the lower part of the MHz frequency band. In such a band the reflection coefficient and the adjacent element mutual couplings are almost counter-phased (see Fig. 6). These data explain the low reflection level obtained with the infinite array approach (Fig.4), since Γ inf.array ~ S S 56. As can be noted from Fig 7, the mutual coupling between non adjacent radiators, which distance from the source is larger than two elements, is below. Reflection Coefficient and Mutual Couplings (9 Ohm) Reflection S 66 Coupling S 56 Coupling S GHz Figure 54. Reflection coefficient and mutual couplings of the branched Vivaldi linear array configuration. The reference impedance is 9 Ω. The vertical black dash-dotted lines represent the two operative frequency bands MHz and MHz. 43

44 2 Phase of the mutual coupling ratio S 56 /S GHz Figure 55. Phase difference between reflection coefficient and mutual coupling to the adjacent element for the central branched Vivaldi radiator in the 11-element array. Reflection Coefficient and Mutual Couplings (9 Ohm) Reflection S 66 Coupling S 36 Coupling S 26 Coupling S GHz Figure 56. Reflection coefficient and mutual couplings of the branched Vivaldi linear array configuration. The reference impedance is 9 Ω. The vertical black dash-dotted lines represent the two operative frequency bands MHz and MHz. Radiation patterns in the H-plane were computed with both the infinite (all-elements in-phase excitation) and finite (single element excitation) approaches. The obtained results at 12, 24 and 48 MHz are reported in Figs. 8, 9 and 1, respectively. The discrepancies between the two curves, 44

45 which should nominally be the same, are due to the array truncation. As expected, this effect is less significant at higher frequencies. Hence, the green curve should be considered to estimate the H- plane radiation pattern of the elements that are far from the edges in larger finite arrays. Since the H-plane patterns computed with the infinite array approach show a similar behavior with respect to the ones reported in [1] i.e tapering from -6 to -12 and a high front-to-back ratio, the present structure can feed the Northern Cross E/W cylindrical offset parabolic reflector with great efficiency in the continuous bandwidth from 12 to 43 MHz. E-plane radiation patterns at 12, 24 and 48 MHz of the 11-element array central element are shown in Fig. 11. The ripple of the curves is due to the great interaction among the elements and the array truncation. The vertical dash-dotted line at 3 represents the 3- field of view of the fat dipole configuration reported in [1]. As can be noted, the present configuration exhibits a narrower field of view (approximately 2 ) with respect to the solution in [1]. As a matter of fact, the Vivaldi antenna in its relevant configuration presents a higher directivity because of the stronger aperture coupling with the adjacent elements. H-plane Radiation 12 MHz Infinite Array (in-phase all-element excitation) 11-Element Array (central radiator excited) GHz Figure 57. H-plane radiation pattern of the branched Vivaldi array inside the 12 MHz. The black dashdot line at 37.5 represents the main reflector edges when the feed system is pointed as in [1]. 45

46 H-plane Radiation 24 MHz Infinite Array (in-phase all-element excitation) 11-Element Array (central radiator excited) GHz Figure 58. H-plane radiation pattern of the branched Vivaldi array inside the 24 MHz. The black dashdot line at 37.5 represents the main reflector edges when the feed system is pointed as in [1]. H-plane Radiation 48 MHz Infinite Array (in-phase all-element excitation) 11-Element Array (central radiator excited) GHz Figure 59. H-plane radiation pattern of the branched Vivaldi array inside the 48 MHz. The black dashdot line at 37.5 represents the main reflector edges when the feed system is pointed as in [1]. 46

47 E-plane Radiation Pattern 12 MHz 24 MHz 48 MHz GHz Figure 6. E-plane radiation patterns of the branched Vivaldi array inside the subreflector (single element excitation). The black dash-dot line at 3 represents the Field of View of the fat dipole configuration reported in [1]. 47

48 Conclusions A dense array has been designed to feed the Northern Cross E/W arm main reflector. It is composed of branched Vivaldi radiators arranged inside a subreflector in a linear array configuration. This one is grating-lobe free for each scanning angle up to 24 MHz owing to the close spacing of the elements. It is also grating-lobe free at broadside up to 43 MHz and more. However, such close spacing implies reduced dimensions of the radiating elements, leading to a high reflection coefficient (up to -4 ) at 12 MHz. Moreover, a strong array truncation effect has been observed at 12 MHz for radiation pattern in both E- and H-planes. Therefore, more than one dummy element should be used at the array edges in order not to have great discrepancies among the various element patterns. Finally, it was observed that the E-plane field of view of this configuration is narrower with respect to the fat dipole one [1] owing to the own directivity of the Vivaldi-type radiators in their working condition. References [2] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, Upgrade of the Northern Cross E/W Arm for the MHz frequency band: preliminary investigation IEIIT Technical Report EA741 [3] W.A. van Cappellen, S.J. Wijnholds, J. D. Bregman, Sparse antenna array configurations in large aperture synthesis radio telescopes, Proceedings of the 3rd European Radar Conference, Manchester UK, September 26, pp [4] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, EM Analysis of the Northern Cross E/W 48 MHz IEIIT Technical Report EA6113 [5] J. Shin and D. H. Schaubert, A Parameter Study of Stripline-Fed Vivaldi Notch-Antenna Arrays, IEEE Transactions on Antenna and Propagation, Vol. 47, No. 5, May 1999, pp

49 Technical Report Upgrade of the Northern Cross E/W Arm for the MHz frequency band: log periodic antenna solution (Draft 1-EA/7719) Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 19 th July 27 49

50 This document contains a further investigation for the MHz upgrade of the Medicina Northern Cross E/W arm [1]. The fat dipole and the branched Vivaldi solutions were presented in [2] and [3], respectively. Both these configurations make use of a wire subreflector to achieve the directivity required to feed the main reflector. In this way, high illumination efficiency and large field of view were obtained with low spillover losses. However, the presence of the wire subreflector leads to an increased mechanical complexity, weight and cost of the feed system which is not suitable for a first lowfrequency test-bed. For this reason, a subreflectorless solution is presented in this work. It consists of an array of log periodic antennas (LPAs) placed in the focus of the main reflector. This solution was previously discarded by the IEIIT group since a radiation pattern which is broad in the E-plane (large field-ofview) and narrower in the H-plane (proper illumination of the main reflector) is needed in the present application. On the contrary, the LPA has instead a radiation pattern which is narrower in the E-plane than in the H-plane. Therefore, reduced performances should be expected with this configuration. In order to further reduce the costs, a commercial LPA has been initially selected. The computed reflection coefficient and radiation patterns for this antenna are reported in section 1. However, since the previous antenna is not the optimum LPA for the present application, another LPA has been designed in order to show the achievable performances of this subreflectorless solution. The corresponding simulated parameters are reported in section 2. Finally, the radiation pattern of the whole antenna system with the LPAs is reported in section 3. The comparison to the fat dipole solution (with subreflector) is also shown in order to quantify the pattern and efficiency degradations. 5

51 1. Commercial LPA 1 5 MHz, 1.2 m, 18 elements The LPA reported in Fig. 1 is a commercial model. Its operative frequency band is 1 5 MHz and the typical gain is 6. This antenna has been simulated in order to verify the nominal data and to obtain the radiation patterns at 12 and 24 MHz. Figure 61. Commercial LPA with 18 elements, total length 1.2 m, longer dipole length 1.55 m, shorter dipole length.215 m. The reflection coefficient in the 1 5 MHz band is reported in Fig. 2, confirming the 5-Ω impedance matching in the overall band. The antenna maximum gain is instead reported in Fig. 3. The level is even better than 6 but at some frequencies the gain drops to very low values. At those frequencies, the corresponding radiation patterns are very distorted. The radiation patterns in the H-plane and the E-plane are reported in Figs. 4 and 5, respectively. Thanks to the properties of the LPA and since this antenna is designed for a maximum working frequency of 5 MHz, the patterns at 12 MHz and 24 MHz are practically coincident. The H- plane pattern level at 37.5 (edge of the main reflector [2]) is approximately -1.1 leading to a poor tapering of the main reflector (see section 3). As expected, the E-plane pattern is narrower than the H-plane one and the front-to-back ratio is of the order of 2. 51

52 Reflection Coefficient () GHz Figure 62. Reflection coefficient of the commercial LPA. The reference impedance is 5 Ω Maximum Gain () GHz Figure 63. Maximum Gain of the commercial LPA. 52

53 H-plane 12 MHz 24 MHz Radiation Pattern () Figure 64. H-plane Radiation Pattern of the commercial LPA. The black dash-dot lines at 37.5 and represent the main reflector edges when the antenna is pointed at 55 with respect to the axis of the parabolic profile. The values of the pattern at 37.5 are -1.1 at both frequencies. E-plane 12 MHz 24 MHz Radiation Pattern () Figure 65. E-plane Radiation Pattern of the commercial LPA. The HPBW is approximately 7. 53

54 2. Custom LPA MHz, 2.2 m, 18 elements A new LPA has been designed in the MHz frequency band with a higher directivity with respect to the commercial one, in order to improve the main reflector illumination efficiency. The designed structure, which is shown in Fig. 6, has 18 elements and is 2.2 meters long. The estimated weight is 3.7 Kg. Figure 66. Custom LPA with 18 elements, length 2.2 m, longer dipole length m, shorter dipole length.49 m. The reflection coefficient is reported in Fig. 7, showing a good 5-Ω impedance matching in a frequency band that is slightly larger than MHz. The maximum gain reported in Fig. 8 is about 9.5 at the lower end of the band, and smoothly decreases to 7.5 at the upper end, without significant ripple. Due to the different gain values at 12 MHz and 24 MHz, the H-plane radiation patterns in Fig. 9 are different: the pattern levels at 37.5 are in fact -2.7 and -1.8, respectively. As it will be shown in section 3, this phenomenon is suitable for the present application because the narrower beamwidth at the lower frequencies provides a better tapering of the main reflector with a consequent reduction of the splillover lobes. Since the spillover lobes are nominally lower at the higher frequencies, this slightly broader MHz can be tolerated since it avoids an excessive LPA length and weight. Even in this case, the E-plane pattern is narrower than the H-plane one but it is quite constant with frequency. The front-to-back ratio is better than 2. 54

55 Reflection Coefficient () GHz Figure 67. Reflection coefficient of the custom LPA, the reference impedance is 5 Ω Maximum Gain () GHz Figure 68. Maximum gain of the custom LPA. 55

56 H-plane 12 MHz 24 MHz Radiation Pattern () Figure 69. H-plane Radiation Pattern of the custom LPA. The black dash-dot lines at 37.5 and represent the main reflector edges when the antenna is pointed at 55 with respect to the axis of the parabolic profile. The values of the pattern at 37.5 are -2.7 and -1.8 at 12 MHz and 24 MHz, respectively. E-plane 12 MHz 24 MHz Radiation Pattern () Figure 7. E-plane Radiation Pattern of the custom LPA. The HPBW is approximately 6. The custom LPA was also analyzed in a 7-element array configuration in order to compute the mutual couplings and estimate the field-of-view. The inter-element spacing was 1.25 m, as in [2] (sparse array). It has to be noted that this arrangement, where the inter-element spacing is slightly 56

57 smaller than the antenna maximum dimension (1.275 m) is possible due to the LPA geometry (the two arms of the dipoles are slightly displaced from the center of the LPA). The reflection coefficient of the central element of the array, reported in Fig. 11, does not show significant differences with respect to the single LPA case (Fig. 7). The mutual coupling between the central element and the neighbors is below. The maximum gain of the central element (shown in Fig. 12) is smaller than the one of the LPA alone (Fig. 8) in the lower part of the frequency band. This phenomenon is related to the coupling that broadens the E-plane pattern of the central element at lower frequencies. As reported in Fig. 13, the E-plane pattern at 12 MHz is in fact broader than the one of the single LPA (Fig. 1). On the contrary, the pattern at 24 MHz, where the coupling is less significant, remains almost the same as in the single LPA case. From Fig. 13, it has to be concluded that the field-of-view of the present feed system is comparable to the one of the fat dipole solution described in [2]. As far as the H-plane radiation pattern is concerned, only a slight narrowing effect is observed at 12 MHz. Reflection Coefficient () -4 S 44 (central) S 43 S GHz Figure 71. Reflection coefficient (S 44 ) of the central radiator of the 7-element custom LPA array, and mutual couplings (S 43 and S 42 ), the reference impedance is 5 Ω. 57

58 Maximum Gain () GHz Figure 72. Maximum gain of the central radiator of the 7-element custom LPA array. E-plane 12 MHz 24 MHz Radiation Pattern () Figure 73. E-plane radiation pattern of the central radiator of the 7-element custom LPA array. The minimum values in the 3 range are MHz and MHz. 58

59 H-plane 12 MHz 24 MHz Radiation Pattern () Figure 74. H-plane radiation pattern of the central radiator of the 7-element custom LPA array. The values of the pattern at 37.5 are -3.6 and -1.8 at 12 MHz and 24 MHz, respectively. 59

60 3. Analysis of the E/W arm with the new MHz feed systems A 2D analysis method was used to compute the H-plane radiation patterns of the overall antenna system. The wire distribution of the entire system is depicted in Fig Wire distribution, N (number of wires)= 1745 m m Figure 75. Wire distribution of the main the cylindrical offset parabolic reflector. The red marker represents the focus. Four different antennas were used to feed the main reflector of Fig. 15: - a line source with no subreflector - the commercial LPA - the custom LPA - the fat dipole with the subreflector (see [2]). The black and red vertical lines in the following radiation patterns represent the reflector edges [2] and the horizon, respectively. The radiation patterns with the line source are reported in Figs As one can see, since the primary pattern is omnidirectional, only a small part of the radiated energy is focused on the main lobe, leading to a width efficiency (see [1]) below 25%. Moreover, the energy which is not collected by the main reflector produces a constant ( ) spillover lobe level at 12 MHZ (24 MHz). 6

61 freq=.12 GHz, η=.182, Max Directivity= 11.3, Max.Pos.= -1.47, FHPBW = 4.39 Spillover lobes θ θ= θ=9 θ= 9 θ= Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Figure 76. H-plane radiation pattern of the main reflector with a line source feed (no 12 MHz Primary pattern Main Refl. Scattered Field 15 Secondary Pattern Main Refl. Angular Position Figure 77. H-plane radiation pattern of the main reflector with a line source feed (no 12 MHz. 61

62 freq=.24 GHz, η=.262, Max Directivity= 15.88, Max.Pos.= -.3, FHPBW = 1.83 Spillover lobes θ=9 θ θ= θ= Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position θ=18 Figure 78. H-plane radiation pattern of the main reflector with a line source feed (no 24 MHz Primary pattern Main Refl. Scattered 15 Field Secondary Pattern Main Refl. Angular Position Figure 79. H-plane radiation pattern of the main reflector with a line source feed (no 24 MHz. 62

63 The H-plane radiation patterns of the LPAs reported in Fig. 4 and 9 were introduced in the 2D simulator in order to obtain the corresponding radiation patterns and width efficiencies. Since the radiation patterns of the custom LPA arranged in a 7x1 array (Figs. 13 and 14) showed small differences with respect to the ones of the single LPA, the latter were used in the 2D simulations. In this way, since the single LPA exhibits less directivity than the LPA array configuration, more conservative results can be obtained. The pointing direction for the LPAs is 55 from the axis of the parabolic profile (same direction as the feed systems in [2] and [3]). The optimum placement along this direction was found with simulations. Optimum results were obtained by placing the tip of the commercial (custom) LPA at.2 m (.7 m) from the focus, toward the main reflector. The radiation patterns for the commercial LPA are reported in Figs Thanks to the front-toback ratio and to the directivity of this antenna, a better radiation pattern is obtained with respect to the line source case. The spillover lobes are below -12 and at 12 MHz and 24 MHz, respectively, and the width efficiency (see [1]) is about 6% (it has to be remembered that the with efficiency is not the overall radiation efficiency). The radiation patterns for the custom LPA are instead reported in Figs Thanks to the higher directivity of this antenna with respect to the previous cases, the spillover lobes become -14 at 12 MHz. Moreover, although the custom LPA gain decreases at 24 MHz (Fig. 8), it is still higher than the one of the commercial LPA (Fig. 3), providing a spillover lobe level better than.5. Finally, the radiation patterns for the fat dipole solution with the wire subreflector reported in [2] are shown in Figs This configuration produces an edge tapering of approximately 12 MHz on the main reflector that leads to a spillover lobe level with a 85% width efficiency. The edge tapering further increases at the upper frequencies, providing almost the same efficiency but with spillover lobe levels below All the significant data are summarized in Tab

64 freq=.12 GHz, η=.585, Max Directivity= 16.36, Max.Pos.= -.8, FHPBW = 4. Spillover lobes θ θ= θ=9 θ= Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position θ=18 Figure 8. H-plane radiation pattern of the main reflector with the commercial 12 MHz Primary pattern Main Refl. Scattered 15 Field Secondary Pattern Main Refl. Angular Position Figure 81. H-plane radiation pattern of the main reflector with the commercial 12 MHz. 64

65 freq=.24 GHz, η=.568, Max Directivity= 19.24, Max.Pos.= -.5, FHPBW = 2.13 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Spillover lobes θ=9 θ θ= θ= 9-4 θ= Figure 82. H-plane radiation pattern of the main reflector with the commercial 24 MHz Primary pattern Main Refl. Scattered Field 15 Secondary Pattern Main Refl. Angular Position Figure 83. H-plane radiation pattern of the main reflector with the commercial 24 MHz. 65

66 freq=.12 GHz, η=.746, Max Directivity= 17.42, Max.Pos.= -.15, FHPBW = 4.52 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Spillover lobes -4 θ=9 θ θ= θ=18 θ= Figure 84. H-plane radiation pattern of the main reflector with the custom 12 MHz Primary pattern Main Refl. Scattered 15 Field Secondary Pattern Main Refl. Angular Position Figure 85. H-plane radiation pattern of the main reflector with the custom 12 MHz. 66

67 freq=.24 GHz, η=.652, Max Directivity= 19.84, Max.Pos.=.2, FHPBW = 2.17 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Spillover lobes θ=9 θ θ= θ= 9-4 θ= Figure 86. H-plane radiation pattern of the main reflector with the custom 24 MHz Primary pattern Main Refl. Scattered Field Secondary Pattern 15 Main Refl. Angular Position Figure 87. H-plane radiation pattern of the main reflector with the custom 24 MHz. 67

68 freq=.12 GHz, η=.846, Max Directivity= 17.96, Max.Pos.= -.12, FHPBW = 4.81 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Spillover lobes θ θ= θ=9 θ= 9-4 θ= Figure 88. H-plane radiation pattern of the main reflector with the fat dipole and the wire 12 MHz Primary pattern Main Refl. Scattered 15 Field Secondary Pattern Main Refl. Angular Position Figure 89. H-plane radiation pattern of the main reflector with the fat dipole and the wire 12 MHz. 68

69 freq=.24 GHz, η=.84, Max Directivity= 2.94, Max.Pos.= -.1, FHPBW = 2.66 Primary pattern Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Spillover lobes θ=9 θ θ= θ= θ= Figure 9. H-plane radiation pattern of the main reflector with the fat dipole and the wire 24 MHz. Primary pattern 3 Main Refl. Scattered Field Secondary Pattern Main Refl. Angular Position Figure 91. H-plane radiation pattern of the main reflector with the fat dipole and the wire 24 MHz. 69

70 Freq. MHz Width η (%) HPBW () Secondary Lobes () Spillover Lobes () Line source (no subref) Commercial LPA Custom LPA Fat dipole subreflector Table 1. Comparison table between the different feed systems. Conclusions A low-cost feed system solution based on the LPA has been investigated. A commercial antenna was considered first. A width efficiency of approximately 6% and spillover lobes lower that -12 were found. On this basis, a custom LPA has been designed. It provides efficiency levels of 75% and 65% at the lower and upper ends of the MHz bandwidth, respectively. The spillover lobes are below -14. The comparison between these solutions and the ones that make use of a subreflector ([2] and [3]) highlights that the low-cost LPA feed system is feasible. However, it will not exploit the maximum capabilities of the Northern Cross E/W arm. Since the spacing between the custom LPAs is 1.25 m, 18 antennas can be arranged in an array configuration which is approximately 22.5-m long. The outermost elements can be used as dummy elements providing a better similarity of the remaining 16 channels. It has to be remembered that the present wire subreflector 48 MHz has to be removed in order to install the LPAs. Otherwise, a great distortion of the LPA radiation patterns would occur, reducing the overall efficiency and increasing the spillover lobes. References [6] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, EM Analysis of the Northern Cross E/W 48 MHz IEIIT Technical Report EA6113 [7] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, Upgrade of the Northern Cross E/W Arm for the MHz frequency band: general investigation and fat dipole solution IEIIT Technical Report EA741 [8] G. Virone, R. Tascone, O. A. Peverini, G. Addamo, A. Olivieri, Upgrade of the Northern Cross E/W Arm for the MHz frequency band: dense array solution IEIIT Technical Report EA7426 7

71 Technical Report Measurements on the custom log periodic antenna for the MHz upgrade of the Northern Cross E/W Arm (Draft 1-EA/7122) Giuseppe Virone, Giuseppe Addamo, Riccardo Tascone, Augusto Olivieri, Oscar Antonio Peverini IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 2 th December 27 71

72 This document reports the measurements carried out on two prototypes of the custom log periodic antenna described in the previous EA7719. One of the prototypes is shown in Fig. 1 along with one of the designers in order to let the reader directly quantify the antenna dimensions. Figure 92. Custom Log periodic antenna (without flange) The designed antenna will be mounted on the E/W arm feed supporting structure by means of the metallic flange reported in Fig. 2. Figure 93. Metallic flange to connect the antenna to the E/W arm supporting structure. 72

73 The reflection coefficient of both prototypes was measured with a vector network analyzer. The corresponding results are reported in Figs. 3 and 4, respectively. Both measurements are in good agreement with the simulated results confirming the reliability of the manufacturing procedure. The Aircell 5 coaxial cable which carries the signal from the tip of the antenna to the back is about 3.2 m long. Its measured transmission coefficient, which is reported in Fig. 5, can be used to compute the antenna losses. Simulation Measurement on prot GHz Figure 94. Reflection coefficient (5 Ω) for prototype 1. Simulation Measurement on prot GHz Figure 95. Reflection coefficient (5 Ω) for prototype 2. 73

74 -.3 S21 cable GHz Figure 96. Transmission coefficient of a 3.2 m-long Aircell 5 coaxial cable. The radiation pattern of one of the two prototypes has been characterized by means of an outdoor antenna test range. The prototype mounted on the fiberglass rotating pole is shown in Fig. 6. Figure 97. Custom Log-periodic antenna mounted on the measurement fiberglass rotating pole. 74

75 The computed and measured H-plane radiation patterns are reported in Figs Since these patterns are in agreement with respect to the simulated data (also reported in EA7719), these antennas are suitable for the illumination of the E/W arm in the MHz frequency band. The E-plane radiation patters are instead reported in Figs In this case, the good agreement between simulations and measurements confirms the possibility of scanning the beam in the E- plane when these antennas are mounted in a linear array configuration, with or without the E/W main reflector. The discrepancies between simulations and measurements are due to the residual errors of the outdoor antenna test range. 12 MHz Simulation Measurement Figure 98. H-plane radiation 12 MHz. 75

76 18 MHz Simulation Measurement Figure 99. H-plane radiation 18 MHz. 24 MHz Simulation Measurement Figure 1. H-plane radiation 24 MHz. 76

77 12 MHz Simulation Measurement Figure 11. E-plane radiation 12 MHz. 18 MHz Simulation Measurement Figure 12. E-plane radiation 18 MHz. 77

78 24 MHz Simulation Measurement Figure 13. E-plane radiation 24 MHz. Conclusion The measurements on the custom log periodic antenna prototypes confirmed the validity of both the design and the manufacturing technique. Therefore, the log periodic feeding array for the E/W arm of the Northern Cross in the MHz frequency band can be built using the developed antennas. 78

79 Technical Report Upgrade of the Northern Cross E/W Arm for the MHz and the MHz frequency bands: multi-feed solution (Draft 1-EA/821) Giuseppe Virone, Riccardo Tascone, Oscar Antonio Peverini, Giuseppe Addamo, Augusto Olivieri IEIIT-CNR Istituto di Elettronica ed Ingegneria dell Informatizione e delle Telecomunicazioni Politecnico di Torino Corso Duca degli Abruzzi 24, 1129-Torino (Italy) Tel , Fax riccardo.tascone@polito.it Torino 1 st February 28 79

80 This document is an extension to the previous EA/741 report entitled Upgrade of the Northern Cross E/W Arm for the MHz frequency band: general investigation and fat dipole solution. The main improvements with respect to the previous fat dipole solution consist of a smaller back side of the subreflector and a further dipole array working in the MHz frequency band. The former modification has been done in order to simplify the subreflector mounting on the existing feed supporting structure of the E/W arm of the Northern Cross, according to the meeting in Medicina on the 1 th of May 27. The second dipole array which is smaller and closer spaced than the fat dipole one is mounted as in Fig. 1 between the previous array and the back side of the subreflector. It has to be noted that the smaller array could be built using the dipoles of the present 48 MHz feed system. The present solution has been investigated in order to provide a low-cost upgrade of the Northern- Cross E/W arm with both the MHz and MHz frequency bands with a large field of view. The back side reduction and the presence of the smaller array required a new design of the overall system. The main geometrical data are reported in Fig. 2. The subreflector aperture size w and length b are the same as in the original design. The larger and smaller arrays are arranged and labeled as in Fig. 3. The spacing of the larger array is the same as in the original design (1.25 m). The spacing of the smaller array is instead.368 m i.e. half 48 MHz. Figure 14. Geometry of the subreflector with two dipole arrays. The larger ones work in the MHz frequency band. The smaller dipoles work in the MHz frequency band. 8

81 b d l d s a Smaller Dipoles Larger Dipoles w Figure 15. Side view of the multi-feed structure. The aperture size w is 3 m, the total length b is 2.25 m, the size of the back side of the reflector a is.44 m, the distance d l of the larger dipole array from the back of the subreflector is.625 m and d s is.184 m. L -1 L L +1 S -2 S -1 S S +1 S +2 Figure 16. Top view of the two dipole arrays. The spacing of the larger dipoles, labeled with L, L +1, L -1, is 1.25 m. The spacing of the smaller dipoles, labeled with S, S +1, S +2, S -1, S -2, is.368 m. 81