High-Gain and Narrow-Beam Antenna Array Design with Large Element Spacing at X Band

IMPACT STATEMENT

Antenna arrays with a more elements have been developed to obtain high directivity or high gain for use in several applications. However, the need for several elements and consequently more power splitters in feed networks of large arrays lead to high loss. Although some solutions such as waveguide feeds provide low loss, they have other drawbacks. As another issue, in all existing designs, it has been recommended to apply element spacing of less than one wavelength to avoid grating lobes. Thus, this causes considerable coupling effects between the elements, degrading the gain and the bandwidth. A new approach is used in the presented design to reduce the loss due to the use of several elements in large arrays. The design is implemented by applying a large element spacing, which also reduces couplings between the elements. No existing design in the literature uses element spacing more than λ, due to the rise of grating lobes. However, element spacing of 2λ is used in the presented design for the first time, having very low grating lobes benefiting from the use of directive antenna elements. Consequently, the same directive and narrow beam are achieved using fewer elements and, therefore, lower required power splitter and lower loss, compared to arrays with λ or λ/2 element spacing and having twice or four times a greater number of the elements. Finally, the measured HPBW of 5.7° is achieved by the presented 4-element array, which only has been obtained by existing 16-element arrays requiring fifteen power splitters, as demonstrated in the table in the manuscript. The design is implemented using only three power splitters, which assures lower loss in the feed network compared with existing arrays.

1. INTRODUCTION

Antenna arrays with more elements have been developed to obtain high directivity and high gain for use in several applications. This is majorly due to their ability to provide long-range coverage, better performance in terms of interference and better power efficiency [Citation1]. The recently hugely developed fifth generation (5G) of wireless communications is in urgent need of suitable high-gain antennas to compensate high propagation loss [Citation2]. Satellite technology is another widespread infrastructure where the use of high-gain antenna arrays has been demanded [Citation3].

Microstrip technology has been widely investigated for the design of high-gain antenna arrays [Citation4]. However, dielectric and ohmic losses and surface wave propagation loss in their feed network are unsatisfactorily high [Citation4,Citation5]. Additionally, the need for more power splitters in feed networks of large arrays leads to high insertion and isolation losses [Citation6]. Thus, shortening the length of transmission lines and lowering the number of power splitters in the feed networks of large arrays have been investigated extensively. Series feed solutions have rather narrow bandwidths [Citation7]. Waveguide feeds also are not cost-effective and have bulky dimensions [Citation4,Citation7]. In all existing designs, it has been recommended to arrange arrays’ elements with spacings of less than one wavelength to avoid grating lobes [Citation8,Citation9].

In this work, a new and simpler solution is presented to avoid these difficulties and limitations due to the use of several elements. The larger an array’s aperture, the higher its directivity [Citation2]. The enlargement of an array’s aperture is possible by either increasing the number of elements or the spacing between them [Citation2]. Element spacings in all existing designs have been ordered to be less than one wavelength. But, in the arrays based on directional antenna elements, the array’s size can be enlarged by lengthening the element spacing even greater than one wavelength, which is implemented for the first time in this design, instead of consecutively adding more elements. To accommodate the proposed design method with low SLL, a directional antenna element must be deployed with its radiation nulls toward the grating lobes of the array. In this design, the highly directional antenna element in [Citation10] is used to overcome the grating lobes.

2. ANTENNA ARRAY CONFIGURATION

To date, all available arrays have been designed based on the conventional limit of element spacing of less than one wavelength [Citation11,Citation12]. In that way, the only solution to widen an array’s size and consequently its directivity is to add more elements. However, it would be possible to break out this limit by the effective use of a directional antenna element.

2.1 Antenna Element Design

The antenna element is designed based on [Citation10]. The dimensions of the antenna were optimized considering the 10–12 GHz frequency range of the X band. The final prototype was fabricated on a Taconic substrate with a relative permittivity of 4.4, a dissipative factor of 0.003, and a height of 1.02 mm. Figure 1 demonstrates the antenna and the details of the parabolic slots in the ground plane by exposing their equations. The slots are formed by three curves C₁, C₂, and C₃, each one is described by the related expression in the inset of Figure 1(b). The expression for C₁ is described by the coordinate system with the Y and the Z axes and the center point O, whereas the expressions for C₂ and C₃ are described by the relative coordinate system with $Y^{\mathrm{\prime }}$ and the $Z^{\mathrm{\prime }}$ axes and the center point $O^{\mathrm{\prime }}$, which has L₅ and L₇ offsets from O. The optimal dimensions and the optimal values for the constant parameters are given in Table 1. It must be noted that those expressions for C₁, C₂, and C₃ are only described in and plotted for $Z\ge 0$ and $Z^{\mathrm{\prime }}\ge 0$ and the other sides of the curves in $Z<0$ and $Z^{\mathrm{\prime }}<0$ are their symmetrical complementary parts.

Figure 1: (a) Geometry of the antenna element based on [Citation10], which is optimized for the X band, with (b) the equations of its parabolic slots

Display full size

Table 1: The optimal dimensions and values for the constant parameters

Download CSVDisplay Table

S11 of the antenna element has two resonances around 8.2 and 11.5 GHz. 1.5 GHz is selected as the design frequency of the array. The radiation patterns are demonstrated in Figure 2. The normalized intensities for the E-plane at θ = 60° and θ = 120° in Figure 2(a) (marked with green lines) are −16 dB and −14 dB, respectively, which are low enough to mitigate the grating lobes in the beam of an array with 2λ element spacing.

Figure 2: Simulated radiation patterns of the antenna element in the (a) E-plane (y-z) and (b) the H-plane (x-y) and for both the co-polarization and the cross-polarization at 11.5 GHz

Display full size

2.2 Four-Element Array Design

The designed linear 4-element antenna array with 2λ element spacing (regarding the frequency of 11.5 GHz) is shown in Figure 3 with the detailed dimensions (in millimeters). The spacing between the elements, d, is 52 mm, which is equivalent to 2λ corresponding to the frequency of 11.5 GHz.

Figure 3: Geometric configuration of the linear 4-element antenna array at X band (all the dimensions are in millimeters)

Display full size

To better demonstrate the method, the 4-element array factor with 2λ element spacing, based on [Citation13], and the radiation pattern of the antenna element are demonstrated in Figure 4. Both patterns are normalized for better comparison. As can be seen in Figure 4(b), nulls of the radiation pattern of the antenna element are exactly in the directions of the grating lobes of the array factor so that they can be removed after multiplication. Consequently, this provides the possibility to design the array with 2λ element spacing without having grating lobes. The same demonstration is repeated in Figure 4(a) having different scales to see the intensity of sidelobes in more detail.

Figure 4: Demonstration of the multiplication of the array factor and the radiation pattern of the antenna element

Display full size

The achieved results and the comparison results with existing antenna arrays, provided in the next sections, confirm the value of $d=2\lambda$ as the optimum element spacing. However, the performance of the antenna array with different element spacings, d, is demonstrated in Figure 5.

Figure 5: Comparison of the simulated E-plane radiation pattern of the antenna array having 2λ element spacing with the arrays having 0.5λ element spacing (a), 0.7λ spacing (b), 1λ spacing (c), 1.5λ spacing (d), 2.5λ spacing (e), and the demonstration of gain and directivity versus the element spacing (f)

Display full size

According to [Citation13], the array factors with element spacings 0.5λ and 0.75λ have no major side lobes. Thus, the antenna arrays with 0.5λ and 0.75λ element spacings also have no major side lobes, as shown in Figure 5(a) and (b), respectively. However, the radiation patterns are less directive than the radiation pattern of the array with 2λ element spacing. The array factor with λ element spacing has major side lobes at θ = 0° and θ = 180°. Consequently, two major side lobs are observed in the same directions in the radiation pattern of the antenna array in Figure 5(c), which are not favorable. Also, the radiation pattern is again less directive than the pattern of the antenna array with 2λ element spacing. Those side lobes shift to around θ = 50° and θ = 130° in the array factor with 1.5λ element spacing. But, according to Figures 2 and 4, those side lobes are exactly in the direction of the nulls in the radiation pattern of the antenna element and consequently are eliminated in Figure 5(d). The radiation pattern in Figure 5(d) is then favorable in terms of side lobes. But, the radiation pattern of the antenna array with 2λ element spacing is still more directive, as shown in Figure 5(d). By increasing the element spacing to 2.5λ in Figure 5(e), the directivity is increased more. However, two larger side lobes are observed in the radiation pattern of the antenna array, due to having the side lobes in the same directions in the array factor, which are far from the nulls of the antenna element to be eliminated. Therefore, the optimum element spacing is seen as 2λ, as the best value to maintain the trade-off between the directivity and side lobs. Also, directivity and gain of the antenna versus element spacing are demonstrated in Figure 5(f). A high and favorable increase in the directivity and gain are seen by increasing the element spacing from 0.5λ to 1.5λ, which confirms the benefit of using larger element spacing. The increase continues slowly from 1.5λ to 2.5λ. However, the optimum value in this range must be chosen considering also the side lobe performance in the radiation pattern plots, which is chosen as 2λ.

3. FABRICATION AND MEASUREMENT RESULTS OF THE ARRAY

The prototype of the design is demonstrated in Figure 6. As seen in Figure 7, measured S11 of two fabricated prototypes confirms the simulated S11, mostly in the range 10–12 GHz.

Figure 6: Fabricated 4-element antenna array with its front and back sides

Display full size

Figure 7: Measured and simulated S11 parameters of the antenna array

Display full size

Figure 8: Simulated E-plane radiation patterns of the antenna array at different frequencies

Display full size

Figure 9: Simulated and measured radiation patterns of the 4-element antenna array in the E-plane at the frequency of 11.5 GHz

Display full size

Simulated E-plane ($\varphi ={90}^{\circ }$, y-z plane) radiation patterns of the antenna array at different frequencies are demonstrated in Figure 8. As it was expected, the larger side lobes at all the frequencies occurred near θ = 60° and θ = 120° directions, both of which are attributed to the grating lobes due to 2λ element spacing. The measured radiation pattern in the E-plane at the frequency of 11.5 GHz is demonstrated in Figure 9. The measured value of the HPBW in Figure 9 is 5.7°, compared with the simulated value of 6.2°. The gain of the antenna is 15.2 dBi. The intensities of the first side lobes are 12.4 dB less than that of the main lobe, as shown in Figure 9. However, the SLLs around the critical directions, $\theta$ = 60° and $\theta$ = 120° are even lower than −13.5 dB.

Finally, measured E-plane radiation patterns at different frequencies are demonstrated in Figure 10 within the range ${60}^{\circ }\ll \theta \ll {120}^{\circ }$. The SLLs are smaller than −10 dB at the frequencies of 11 GHz (Figure 10(b)) and 12 GHz (Figure 10(c)). Also, simulated and measured gains versus frequency are demonstrated in Figure 11.

Figure 10: Measured and simulated E-plane radiation patterns of the antenna array at the frequencies of (a) 10.5 GHz, (b) 11 GHz, (c) 12 GHz, and (d) 12.5 GHz

Display full size

Figure 11: Simulated and measured gains of the antenna

Display full size

As the presented antenna is a linear array and is designed by positioning the antenna elements in only the E-Plane direction (Not in two dimensions), the H-plane radiation pattern of the array is identical to the radiation pattern of the antenna element. Then, its analysis is out of the scope of this work. However, simulated and measured H-plane radiation patterns of the antenna array at the frequency of 11.5 GHz are demonstrated in Figure 12.

Figure 12: Measured and simulated H-plane radiation patterns of the antenna array at the frequency of 11.5 GHz

Display full size

To better demonstrate the contribution of the presented work, the achieved results are compared with those of some existing antenna arrays in adjacent frequency bands in Table 2.

In the planar m × n arrays in Table 2, the largest of m and n is considered for the comparison with the presented linear 4-element array. Where the amounts of HPBW or SLLs have not been described clearly, those values are demonstrated by “>” to provide an accurate and undoubted comparison. The SLL of the presented antenna array at the design frequency (11.5 GHz) is lower than −12.4 dB, which is better than the amounts achieved in the large arrays in [Citation5,Citation12]. The gain of the proposed design is demonstrated in Figure 10 at different frequencies with a maximum gain of 15.2 dBi. The length of the presented array is 8λ. This implies that at least sixteen elements are required in an array with an aperture length of 8λ and having $\lambda /2$ element spacings. In other words, the achieved HPBW in this work can be reached by using sixteen elements with $\lambda /2$ spacing. In Table 2, the amounts of the HPBWs in all the designs that include a linear array or sub-array of a maximum of eight elements in [Citation14–16], are larger than 12°, except in [Citation11] with the best amount of 8°. Similarly, the achieved HPBW of 5.7° in this work has been reached only in the designs that include a linear array of more than sixteen elements as in [Citation5,Citation12]. Then, a 16-element array will need fifteen power splitters, whereas this design is implemented using only three power splitters assuring lower loss and lower design complexity.

4. CONCLUSION

The 4-element linear array with the large element spacing of 2λ is presented at the X band. The very desirable HPBW of 5.7° was achieved with the maximum gain of 15.2 dBi and the SLL of −12.4 at the frequency of 11.5 GHz along with the acceptable characteristics across the bandwidth of 11–12 GHz using only four elements. The comparison results demonstrated that the achieved HPBW of 5.7° has been obtained by only linear arrays of more than sixteen elements requiring several power splitters and feed lines leading to high loss, design complexity or multilayer structures, whereas the presented design is implemented using only three power dividers on a single layer PCB.