A) Antenna configuration

The proposed antenna system shown in Fig. 1 consists of two identical elements placed in mirror configuration about Y-axis on Roger 5870 substrate with volume 35 × 30 × 0.8 mm³.

Fig. 1.

Antenna configuration. (a) CSRR, (b) front side, and (c) back side.

Antenna system consists of two identical cup-shaped monopole radiators to achieve UWB characteristics. They share common ground with rectangular slots and a ground stub to achieve better isolation. Both antennas are fed with 50 Ω microstrip lines spaced by distance, ds (=0.3 λg/2 at 3 GHz) from each other. The antenna system has been optimized using CST Microwave Studio, the optimized dimensions of the fabricated prototype is as follows (millimeters): W = 35, L = 30, Lg = 13, Wa = 10, Wb = 5.5, Wm = 0.4, W1 = 8.4, W2 = 3.9, g = 0.5, Lj = 0.4, We = 1.5, Wf = 2.4, Le = 9.8, Lf = 4, Wn = 1.5, Ln = 4, Lk = 3, Wk = 0.5, L1 = 7, Wp = 1.2, Lh = 5, Lu = 4, Wt = 6, r = 6, Wr = 12, Lr = 5.5, th = 1, Wh = 4.4, S = 1.5. The effect of these dimensions and effectiveness of slots and stubs are presented below. The fabricated prototype antenna is shown in Fig. 2.

Fig. 2.

Photograph of prototype antenna.

B) Ground stub and rectangular slots

The rectangular slots in both the radiators next to feed lines are etched out in order to achieve impedance matching. For further matching of impedance two rectangular slots are made in the ground plane under the feed lines as shown in Fig. 2. A novel H-shaped slot is also cut in the ground plane between two radiators to improve return loss and isolation at around 7.3 GHz. By introducing H-slot, the return loss improves from 9.9 to 13 dB and isolation is increased to 23 dB at 7.3 GHz as observed in Fig. 3.

Fig. 3.

Effect of H-slot and length Lh on (a) |S11| and (b) |S12|.

A novel T-shaped ground stub separating two radiators is added to achieve better isolation. This stub acts as monopole and resonates at frequency defined by its length. In order to increase the length of the T-stub so that it may reflect the lower frequencies between ports 1 and 2, this is done by etching the slots next to the stub in the ground plane. In this way we can achieve a good isolation at lower frequencies without any increase in the overall antenna size. This is an effective way to maintain the good isolation at this compact size. As shown in Fig. 2, the length of the T-stub is controlled by the length L1. It can be seen from Fig. 4(a) that with L1 = 0, lower cut-off frequency of the antenna is 4 GHz. With the increase in length L1 the return loss of antenna reduces at about 3 GHz. Another effect of increasing length L1 is, as the length of stub is increased isolation at lower frequencies (around 3 GHz) also increases. This could be due to the fact that the stub acts as a monopole and the distribution of surface current taking place around the stub gives rise to increased isolation. By etching slot1, two edges are formed in the ground plane. One edge controls the length of ground stub and the other acts as monopole increasing the isolation at higher frequency around 7.5 GHz as seen in Fig. 4.

Fig. 4.

Effect of variations in length L1 on (a) |S11| and (b) |S21|.

C) CSRR slot

Figure 5(a) shows the S-parameters of the antenna without complementary split ring resonator (CSRR) slot. The notched characteristics at the desired frequencies 3.3 and 5.8 GHz are introduced by two complementary SRR slots etched in the radiators. As seen from the input impedance, in Fig. 6(a), antenna acts as short and open at notch frequencies 3.3 and 5.8 GHz, respectively. The outer split ring slot in CSRR act like two parallel open guided quarter wavelength transmission line to short the antenna at the desired frequency [Reference Kerkhoff and Ling18] and the inner ring working is analogous to two parallel shorted quarter wavelength transmission line, because outer and inner split rings are complementary to each other. The total length of the slot is chosen to be λ/2 at the designated frequency and can be calculated by

(1) $$Lr = \displaystyle{c \over {2f.\sqrt {{\varepsilon _{eff}}}}} $$

(2) $${\varepsilon _{eff}} = \displaystyle{{{\varepsilon _r} + 1} \over 2},$$

Fig. 5.

(a) Antenna S-parameters without CSRR. (b) Effect of variation in CSRR length Lr on |S11|.

Fig. 6.

(a) Input impedance plot. (b) Effect of variation in CSRR length Lr2 on return loss.

where ‘c’ is the speed of light, ε _eff is the effective dielectric constant and Lr (Lr = 2(w1 + w2 + Lj) − g − 5 × Wm) is the average length of the SRR slot. According to (1), the length corresponding to 5.8 GHz comes out to be 20.04 mm.

Figure 5(b) shows the return loss at three different values of Lr = 20, 22.8, and 23 mm. It may be seen from the figure that, by increasing the length of SRR, the notched band moves to lower frequencies. In order to reject the center frequency 5.8 GHz the practical length of slot is found 22.8 mm.

Similarly, the average length of the outer SRR slot, Lr2 (Lr2 = 2(wa + wb) − g − 4 × Wm) controls the notched frequency band around 3.3 GHz. Figure 6(b) shows the effect of varying the length Lr2 on the return loss. The practical length of the slot for band notching at 3.3 GHz is found to be 28.9 mm.

A) Impedance performance

The antenna is designed using CST MWS and a prototype is fabricated with the optimized dimensions as above. The return loss of the antenna system is plotted in Fig. 7(a). It could be seen that the system covers the entire band with notched characteristics at 3.3 and 5.8 GHz. The isolation between the two radiators is shown in Fig. 7(b), the measured isolation is greater than 20 dB in the entire band. In the measured results as in Fig. 7(a), there is a shift in the upper notch frequency to 6 GHz. The increased width (Wm) in the fabricated prototype, which affects the average length (Lr) is the reason why the notch has shifted to higher frequency. Other discrepancies between the simulated and measured results might be due to the fact one the antenna system has been fabricated using chemical etching process and second the antenna ground plane is smaller in comparison with connectors [Reference Chen, Yang, Guo and Chia10].

Fig. 7.

Measured and simulated S-parameters of dual-element dual-notched antenna system. (a) Return loss. (b) Isolation.

B) Radiation performance

The radiation patterns of the prototype antenna are shown in Fig. 8 when it is excited at port1 and port2 at 5 and 8 GHz, respectively. As it can be seen the radiations patterns in the XY-plane are different when excited at port1 and port2 owing to pattern diversity, while as the patterns in the YZ-plane are mirror transformation of each other. The radiation patterns are stable in the entire UWB.

Fig. 8.

Measured radiation pattern of the dual-notch proposed antenna at 5 and 8 GHz. (a) XZ-plane, (b) YZ-plane, and (c) XY-plane.

The simulated efficiency and gain of the antenna are plotted in Fig. 9(a). Evidently the antenna radiates efficiently with a stable gain in the entire UWB.

Fig. 9.

(a) Simulated efficiency and gain plot. (b) Measured and simulated ECC of the antenna.

C) Diversity performance

Diversity is use of multiple numbers of antennas at transmitter and receiver end to increase the number of possibilities for reception of low fade signals. This performance is measured in terms of ECC and is given below [Reference Hallbjörner9]:

(3) $$ECC = \displaystyle{{{{\left \vert {S_{11}^{\ast} S12 + S_{21}^{\ast} S22} \right \vert} ^2}} \over {(1 - ( \vert S11{ \vert ^2} + \vert S21{ \vert ^2}))(1 - ( \vert S22{ \vert ^2} + \vert S12{ \vert ^2}))}}.$$

Measured and simulated results show that the ECC is below 0.015 in the entire UWB as shown in Fig. 9(b), thus the proposed antenna is a potential candidate for diversity applications.