Fifth generation of mobile cellular networks (5G) is going to create a bridge to the future into the world of wireless technology due to the promising features it provides including higher speed of data transmission, less network latency, and much more density of connected devices. From The radio frequency point of view, these advantages are all achieved thanks to the millimeter wave wireless technology. Extremely higher carrier frequency utilized in millimeter wave technology offers high performance capacity, throughput, and quality of service (QoS) compared with microwave frequencies which is applied in the current networks such as Long-Term Evaluation (LTE) wireless networks (at 2 GHz or 3.6 GHz) .
Although the features provided by mm-wave technology is promising, but the propagation characteristics are significantly different from that of microwave frequency bands in terms of path loss, diffraction and blockage, rain attenuation, atmospheric absorption, and foliage loss behaviors. In general, the overall loss of mm-wave systems is significantly larger than that of microwave systems for a point-to-point link . In order to overcome the problem of atmospheric wave absorption, Higher gains are required which could be achieved by introducing specific antenna array designs [3-6]. To have the array antenna radiating with desired gain and radiation pattern, each element needs to be fed through a feed network appropriately in both terms of phase and magnitude.
In this article we go through how to address one of the most important RF signal integrity issues while designing a printed circuit boards feed network for an 4×1 mm-wave array antenna. All transmission lines are Coplanar Waveguide with Lower Ground Plane (CPWG) and to split the power of the input with equal magnitude between four output ports, three Wilkinson power dividers are used in a cascade configuration to feed the antenna array elements in a one-to-four arrangement (Fig 1). The full-wave simulations are all caried out using CST Studio software.
Figure 1. Feed network of a 4×1 Array Antenna
In this study, the operating frequency band of our designed feed network is from 25 GHz to 30 GHz, so it is expected to have almost -6 dB of relative magnitude (considering the path loss) for the output power at ports number 2 to number 5 over the corresponding frequency band. Figures 2 and 3 demonstrate the feed network performance over the frequency band of 15 to 40 GHz.
Figure 2. Output S-parameters at Ports number 2 to 5
Figure 3. Return loss at the input of the feed network (Port 1)
As it is demonstrated in Figure 2, there is an average of 0.5 dB difference between the output power of the port 2 and port 3 (similarly between port 4 and port 5). As the array antenna elements are supposed to be fed with the same magnitude, this inequality would result in array antenna failure, principally in terms of radiation pattern and the antenna gain.
Considering the fact that all four outputs were quite identical for the similar cascade configuration of 3 Wilkinson power dividers that we already had designed for lower frequency band applications, it was considered to monitor the output signals over the wider frequency band from 10 to 50 GHz (Fig 4) since wider frequency range of response observation provides a more precise insight into the behavior of the feed network.
Figure 4. Output S-parameters at Ports number 2 to number 5 over 10 to 50 GHz
As the figure 4 illustrates, the higher frequency, the more difference between power level of output ports, so we came up with the idea that the main reason is related to the structure discontinuities of the feed network which are mainly negligible in the lower frequency bands but could have a huge impact on the output signals as the frequency goes higher and gets closer to the mm-wave spectrum. That is due to the fact that energy could be converted between different waveguide modes including the propagating and evanescent ones which might affect the S-parameters of the feed network as well. The evanescent modes are generated at the structure discontinuities, therefor it is needed to consider enough distance separating two successive discontinuities (parameter “L” in Fig 5) in order to let evanescent modes decay remarkably in the desired frequency band.
Figure 5. Parameter L, the distance between two successive discontinuities (indicated in red circles)
The CPWG transmission line used in our feed network design is simulated individually in order to evaluate the characteristics of the propagating mode (Fig 6) and the three first evanescent modes (Table 1). Evanescent modes are numbered in order of ascending cutoff frequency.
Figure 6. Wave characteristics and the E-field cross-sectional view of the propagating mode
Table 1. Wave characteristics of the three first evanescent mods
|Mode Type||Cutoff Frequency (GHz)||Attenuation Constant (α) (1/m)|
|Evanescent Mode NO 1||TE||60.01||1143.40|
|Evanescent Mode NO 2||TM||61.63||1180.72|
|Evanescent Mode NO 3||TM||76.90||1524.16|
Figure 7. Required distance for -40 dB decay of the three first evanescent modes at 25 GHz
According to the values indicated in Fig 7, we set the Parameter L in Fig 5 equal to 5.25 mm to meet all the required distances for -40 dB decay of the three first evanescent modes at 25 GHz. Fig 8 demonstrates a remarkable performance enhancement of the feed network as the difference of the output powers is less than 0.1 dB on average over the whole frequency band of interest (80 percent difference decrease in comparison with the results corresponding to the primitive value of L= 1.25 mm). The Return loss at the input of the feed network is illustrated as well in Fig 9 for both values 1.25 mm and 5.25 mm of parameter L. In case the energy conversion between different modes is still deteriorating the performance of the feed network, greater values of parameter L needs to be considered to let evanescent modes decay more than -40 dB
Figure 8. Output S-parameters at Ports number 2 to number 5
Figure 9. Return loss at the input of the feed network for L = 1.25 and 5.25 mm
Fifth generation of mobile cellular networks (5G) is getting more prominent in recent years thanks to the promising features it provides including higher speed of data transmission, less network latency, and much more density of connected devices. From The radio frequency point of view, these advantages are all achieved due to the capabilities offered by millimeter wave wireless technology. Although the features provided by mm-wave technology is promising, the overall loss of mm-wave systems is significantly larger than that of microwave systems.
One of the key solutions to compensate the signal loss of mm-wave systems is to take advantage of high gain antenna arrays. The array antenna elements are supposed to be fed precisely with the desired magnitudes and phases and any remarkable tolerance would result in deterioration of antenna performance. In this article we went through how to address signal integrity issues caused by the discontinuities in the physical structure of the array antenna feed network. Since The evanescent modes are generated at the structure discontinuities, it is needed to consider enough distance separating two successive discontinuities in order to let evanescent modes decay remarkably in the desired frequency band, otherwise the energy conversion between propagating and evanescent modes would affect the S-parameters of the feed network.
. Shaddad RQ, Al-Samman AM, Rassam MA. Utilization of Millimeter-Wave Spectrum in Wireless Networks (2018).
. Mumtaz S, Rodriguez J, Dai L. MmWave Massive MIMO: A Paradigm for 5G. Academic Press; 2016 Dec 2
. Agarwal S. High Gain Linear 1 x 4 X-slotted Microstrip Patch Antenna Array for 5G Mobile Technology. Journal of Telecommunications & Information Technology. 2020 Mar 1(1).
. Zhang J, Ge X, Li Q, Guizani M, Zhang Y. 5G millimeter-wave antenna array: Design and challenges. IEEE Wireless communications. 2016 Oct 19;24(2):106-12.
. Ashraf N, Haraz OM, Ali MM, Ashraf MA, Alshebili SA. Optimized broadband and dual-band printed slot antennas for future millimeter wave mobile communication. AEU-International Journal of Electronics and Communications. 2016 Mar 1;70(3):257-64.
. M. Agiwal, A. Roy and N. Saxena. Next Generation 5G Wireless Networks: A Comprehensive Survey. IEEE Communications Surveys & Tutorials, 2016 third quarter 18;3:1617-1655.
Disclaimer: The views and opinions expressed in this article are those of the author. It is intended only as a sharing of antenna design knowledge for educational purposes.