Antenna design

Different structures shown in Fig. 1 have been simulated and also analyzed by using HFSS which their comparison results of scattering parameter (|S ₁₁|) curves are illustrated in Fig. 2. First of all, a simple rectangular radiating patch with a rectangular feed line and a rectangular finite ground plane has been considered in structure 1 which is not appropriate for SWB or even UWB applications due to its low impedance BW. Consequently, a shovel-shaped radiating patch is designated in structure 2 to make the impedance BW a little wider which is useful for UWB applications. Subsequently, it can be seen in structure 3, the chosen symmetric stair-shaped ground plane instead of a rectangular one so that the entire SWB operating range of 2.8–40 GHz is achieved. As was investigated in previous works [Reference Dastranj23, Reference Dastranj24], the leaky-wave interaction between the ground plane and radiator can affect the impedance matching of the antenna. Lower leaky-wave from the ground plane and radiator leads to a better impedance matching and, hence, broadband characteristic for the antenna. The symmetric stair-shaped section decreases the leakage current distribution from the ground plane. As a result, the antenna performance becomes better [Reference Dastranj23, Reference Dastranj24]. Figure 3 illustrates the process of reaching DBN characteristics of the SWB antenna. As can be seen in Fig. 3(a), no slots have been etched on the radiating patch of the antenna. The band-notched characteristic of WiMAX band (3.7–4.7 GHz) which is the first band stop (BS) can be generated by etching a C-shaped slot on the shovel-shaped radiating patch in Fig. 3(b). It is shown that in Fig. 3(c) the second BS characteristic (WLAN (5.7–6.4 GHz)) can be generated by employing another C-shaped slot with smaller size radius. The final structure of the proposed antenna is shown in Fig. 3(d). In order to have a SWB antenna with DBN characteristics, two C-shaped slots and a circular slot have been etched on its radiating patch. The simulated scattering parameters (|S ₁₁|) of the structures 1–3 and the final structure are compared in Fig. 4. This figure clearly shows BW enhancement process and DBS characteristics of the antenna. As it can be observed from this figure, the final structure can reject unwanted frequencies at WiMAX and WLAN band stops. The geometry of the final optimized SWB DBN antenna with detailed and necessary dimensions that affect the impedance BW and DBN characteristics is shown in Fig. 5. Optimization of the dimensions has been accomplished by using HFSS software package. The antenna is compact which is printed on FR4 substrate with permittivity 4.4 and loss tangent of 0.02. The optimal dimensions (in mm) of the proposed antenna are as follows: W = 14, L = 20, h = 1.6, W _f = 1.5, L_g = 8, Lgn = 7.5, W ₁ = 1, W ₂ = 0.8, W ₃ = 0.47, W ₄ = 3.48, L ₁ = 4.8, L ₂ = 6.05, L ₃ = 6.7, L ₄ = 4.6, and g = 0.5. Important dimensions (in mm) which affect the DBN characteristics as shown in Fig. 5(b) respectfully are: r1 = 4.6, r2 = 4.1, r3 = 3.3, r4 = 2.8, and r5 = 1.12. The radii of the slots can be calculated approximately from the following formula. In these equations, f_{notch 1} and f_{notch 2} are the center frequencies of the rejected WiMAX (3.7–4.7 GHz) and WLAN (5.7–6.4 GHz) bands and c is the light speed at free space [Reference Liu, Ku and Yang25].

(1)$$r_{WiMAX} = \displaystyle{c \over {\,f_{notch\,\,1} \times 4 \times \sqrt {\varepsilon _{eff}} \times 2\pi \times 0.64}}\comma \;$$

(2)$$r_{WLAN} = \displaystyle{c \over {\,f_{notch\,\,2} \times 4 \times \sqrt {\varepsilon _{eff}} \times 2\pi \times 0.64}}\comma \;$$

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

Fig. 1. Impedance BW enhancement process.

Fig. 2. Comparison of the input impedance matching curves for three structures.

Fig. 3. (a) antenna structure without slots, (b) C-shaped slot used for WiMAX BS, (c) C-shaped slot used for WLAN BS, and (d) final structure.

Fig. 4. Scattering parameter (|S ₁₁|) simulation results of the structures mentioned in Fig. 3.

Fig. 5. (a) Geometry of the proposed antenna, (b) important parameters which affect the antenna's DBN characteristic.

As a parametric study of DBN characteristics of the proposed antenna, different values of r2, r3, r4, and r5 are considered to see their effects on band rejection in both WLAN and WiMAX. It can be concluded from Table 1 that r2 has an effect on first BS (WiMAX) while Table 2 indicates that r4 has an effect on the second BS, WLAN band. However, in Table 3, as it can be observed r3 effects both WiMAX and WLAN bands. Effects of r5 variation on the band rejection behavior of the antenna are tabulated in Table 4. This table shows that the circular slot in the center of the patch affects band rejection in both WLAN and WiMAX bands. The best values of the parameters achieved by optimization via HFSS simulator package are summarized as follows (in mm): r2 = 4.1, r3 = 3.3, r4 = 2.8, and r5 = 1.12.

Table 1. Effects of r2 on the first and second band stops

Table 2. Effects of r4 on the first and second band stops

Table 3. Effects of r3 on the first and second band stops

Table 4. Effects of r5 on the first and second band stops.

In this design, to achieve super-wideband operation from 2.8 to 40 GHz with high BW dimension ratio of 7427.4 and dual band-notched characteristics, other methods such as using split-ring resonators and metamaterials were examined. Based on the numerous simulations, the aforementioned characteristics were not achieved by these techniques. However, using C-shaped and circular slots lead to super-wide BW without interferences by WiMAX and WLAN bands.

The Comparison of |S ₁₁| parameters for the antennas printed on RT Duroid (with rescaled dimensions) and FR4 substrates is presented in Fig. 6. Notice that the relative permittivity values of FR4 and RT Duroid substrates are 4.4 and 2.2, respectively. Accordingly, a ratio of the square of 2 was taken in rescaling the antenna dimension for the RT Duroid material. It is seen that the antenna with RT Duroid substrate has |S ₁₁|>−10 dB over the large part of the desired band. Therefore, the antenna was printed on FR4 substrate. It should be noted that the worst input matching realized using RT Duroid substrate could be partially due to the lower losses of this substrate. In other words, the higher losses of FR4 are certainly a bad factor for the antenna gain, but they may play a good part to improve the input matching.

Fig. 6. Comparison of |S ₁₁| parameters for the antennas printed on RT Duroid and FR4 substrates.

Measured results and discussion

Figure 7 shows the photograph of the proposed antenna prototype. The antenna was simulated by using HFSS software package. Then, it was fabricated to validate the results which are carried out by simulation. The designed antenna is connected to a 50-ohm SMA connector for signal transmission. The part number of SMA female connector is SC8026 which normally operates from DC up to a frequency of 18 GHz and offers excellent VSWR of 1.23:1. However, this connector features VSWR of about 1.45:1 for the frequency range of 18–40 GHz. It is a low lost connector with reasonable lose.

Fig. 7. Photograph of the proposed antenna prototype.

Frequency-domain results

Measured scattering parameter (|S ₁₁|) comparing with the simulated one is shown in Fig. 8. There is a negligible difference between the simulated and measured results. As shown in this figure, the measured scattering parameter at some part of the frequency between 10 and 16 GHz goes greater than −10 dB. This is due to the measurement errors, fabrication tolerances, and SMA soldering effects. Hence, good agreement between measured and simulated results has been achieved which means the antenna is a very good candidate for SWB applications with DBN property. However, in order to further understand the utility of the proposed antenna over the entire operating BW, other radiation characteristics such as far-field patterns and gain must also be carefully investigated. The far-field radiation patterns of the SWB antenna in H (x-z)- and E (y-z)-plane at several frequencies were measured. For briefness, results at the lower band edge frequency (3 GHz), mid-frequency (18 GHz), and higher band edge frequency (40 GHz) are presented and compared with simulated ones (Fig. 9). As can be observed from Fig. 9(a), the antenna has an omnidirectional pattern in H-plane and bidirectional pattern in the E-plane at low frequencies with very low cross-polarization level. Figures 9(b) and (c) show that the radiation patters at higher frequencies are also reasonable. Notice that the cross-polarization level increases at higher frequencies due to excitation of higher-order modes. Notice that input matching and antenna gain are promising even beyond higher band edge frequency. However, at the frequencies higher than 40 GHz, the radiation patterns in both E-and H-plane start deteriorating. Also, cross-polarization significantly increases at these frequencies. Thus, the operation BW of the antenna is restricted by the increasing deterioration of the radiation patterns and the increasing cross-polarization at the higher frequencies.

Fig. 8. Comparison between the numerical and experimental |S ₁₁| curves of the antenna.

Fig. 9. Experimental and numerical far-field E (y-z)-and H (x-z)-plane patterns of the antenna (left: x-z plane, right: y-z plane) at (a) 3 GHz, (b) 18 GHz, and (c) 40 GHz.

The simulated and measured peak gain curves of the antenna versus frequency are illustrated in Fig. 10. Due to the omni-directional behavior of the antenna in x-z plane, the direction of the peak gain is not stable. As shown in this figure, the antenna gain sharply drops at the center frequencies of WiMAX and WLAN band stops (4 and 6 GHz). As plotted in Fig. 10, the measured gain at the notch frequencies of the band stops, i.e., 3.7, 4, 4.3, 4.7, 5.7, 6, and 6.4 GHz is −5.7, −9, −3, −1.4, −3.5, −5, and 0.42 dBi, respectively. Note that the measured gain is moderate at the operating frequency band respecting the small size of the antenna.

Fig. 10. Numerical and experimental gain of the antenna versus frequency.

In SWB antennas, BDR is an important parameter that the higher BDR signifies wider frequency band and compactness of the proposed antenna compared to the other structures. BDR indicates how much operating BW (in percentage) can be provided per electrical area unit. Equality is defined as follows [Reference Chen, Sim and Row15]:

(4)$$BDR = \displaystyle{{BW\lpar \percnt \rpar } \over {\lambda _{Length} \times \lambda _{Width}}}. $$

In this equation, λ is the wavelength at the lower cut-off frequency of the working BW. The results of the comparison between the designed antenna and other antenna structures studied in [Reference Singhal and Singh11–Reference Liu, Esselle, Hay and Zhong22], are presented in Table 5 on the basis of BDR. In spite of the small electrical dimension of the proposed antenna compared to the others, a large BDR of 7427.4 is exhibited. Accordingly, it can be concluded that the proposed SWB antenna can provide good BW ratio and very larger BDR characteristics with a much smaller size in comparison to the other antennas.

Table 5. Comparison of the proposed antenna with other SWB antennas (AG, average gain, GD, group delay). Note that the electrical dimension and the BDR are calculated at the wavelength at the lower cut-off frequency of the working BW

Time-domain results

Along with the frequency-domain analysis, time-domain performance should also be analyzed in order to be sure of the SWB operation [Reference Quintero, Zurcher and Skrivervik26, Reference Weisbeck, Adamiuk and Sturm27]. The time-domain analysis required two identical designed antennas, one as the transmitter and the other one as a receiver, in the adjustment of face-to-face and side-by-side. Time-domain analysis of both configurations was considered using CST Microwave Studio by a distance of 50 cm. Time taken by the antenna to receive the pulse is indicated by an important parameter named group delay. Group delay of face-to-face orientation is shown in Fig. 11 which its peak-to-peak variation is less than 2 ns over the entire frequency band. Although the result of the side-by-side configuration has not been discussed in this section, similar results were obtained which indicate an acceptable time-domain performance.

Fig. 11. Comparison between the experimental and numerical group delay results of the antenna versus frequency.

Another important parameter in time-domain analysis named the fidelity factor is used to calculate the correlation between transmitted and received pulses. By using the approach suggested in [Reference Wu, Jin, Geng and Ding28], the input signal is delivered to the antenna, and the far-field electric component is received by means of four virtual probes. To investigate the fidelity factor in both E- and H-plane, seven probes are placed at the angle equal to 0°, 15°, 30°, 45°, 60°, 75°, and 90°. Figure 12 presents this time-domain parameter for both planes. In a typical UWB system, the values of fidelity factor can vary between 0 and 100%. A fidelity factor value of 0% shows that the received and input pulses are completely different from each other, while a value of 100% indicates that the received and input signals are perfectly similar. As was mentioned in [Reference Quintero, Zurcher and Skrivervik26], a fidelity factor higher than 50% is the appropriate value for UWB systems. From Fig. 12, it is seen that the fidelity factor in both planes has acceptable values greater than 76%, making the proposed antenna very capable for use in SWB communication applications.

Fig. 12. Calculated fidelity factor of the antenna.