4.4.3 Measurement of the DIFA Against Conducting Boxes of Various Size. Gap

Transcription

1 4.4.3 Measurement of the DIFA Against Conducting Boxes of Various Size In Section 4.3.3, the IFA and DIFA were modeled numerically over wire mesh representations of conducting boxes. The IFA was modeled over a conducting box that was 6 cm wide, 10 cm long, and 1.5 cm high. The DIFA was modeled over the conducting boxes listed in Table 4.7. In this section, the validity of the wire grid model is tested by building and measuring the DIFA over box numbers 1 (6 X 8 cm), 3 (6 X 12 cm), and 4 (4 X 10 cm) in Table 4.7. The conducting boxes are 1.5 cm high, and are constructed using copper and brass. The DIFA is configured on the box as illustrated in Figure The antenna is fed using a 50 Ω coaxial cable that connects to a SMA flange-mount bulkhead connector. The connector is mounted on the bottom surface of the top face of the box. The antenna is built on the top surface of the top face of the box. The probe feed protrudes through the top face of the box and contacts the horizontal portion of the DIFA. A hole is cut in the bottom face of the box beneath the connector to allow attachment of the coaxial feedline. A cross-sectional representation of the feed is illustrated in Figure DIFA SMA Connector Gap Coaxial Feedline Figure The feed configuration for the DIFA operated against a conducting box. Using the feed configuration in Figure 4.40, the input impedance of the DIFA over box 3 in Table 4.7 (6 X 12 cm) was measured using an HP8720C network analyzer. The 139

2 reference plane is the SMA connector. The input impedance of the DIFA operated against box 3 is illustrated on the Smith Chart in Figure Figure The measured input impedance of the DIFA operated against box 3 (6 X 12 cm) in Table 4.7. Figure 4.41 shows that the DIFA operated against box 3 in Table 4.7 (6 X 12 cm) is resonant at 904 and 926 MHz. The input resistance at the first resonance is 34.5 Ω and at the second resonance is 76 Ω. The resonant frequencies are lower than those for the DIFA over a large ground plane. The VSWR of the DIFA over box 3 is compared to the VSWR modeled using NEC4 in Figure

3 Figure The measured VSWR of the DIFA operated against box 3 (6 X 12 cm) in Table 4.5 compared to the same VSWR modeled using NEC4. The measured impedance bandwidth of the DIFA on box 3 is 40 MHz, or 4.4 % of the center frequency. As shown in Figure 4.42, the measured impedance bandwidth is wider than modeled by NEC4. The VSWR characteristic is shifted down in frequency. The agreement was good between modeled and measured input impedance considering the circumstances of the measurement. The input impedance of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7 is shown on the Smith Chart in Figure

4 Figure The measured input impedance of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7. As illustrated in Figure 4.43, the resonant frequencies of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7 are 890 and 923 MHz. The input resistance at the first resonance is 21.8 Ω and at the second resonance is 51.1 Ω. Comparison of Figure 4.43 and Figure 4.41 shows that the effect of reducing the width of the box, W b, is to rotate the impedance locus clockwise around the center of the chart. The VSWR of the DIFA operated against box 4 is compared to the VSWR modeled by NEC4 in Figure

5 Figure The measured VSWR of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7 compared to the same VSWR modeled using NEC4. As shown in Figure 4.44, NEC failed to model the first resonance of the DIFA over box 4, but accurately modeled the second. A similar error occurred in the model of the RCDLA in Section Thus, the data in Figure 4.44 supports the fact that measurements should always be used to validate NEC impedance data. The measured impedance bandwidth below a VSWR of 2 is 27 MHz, or 3 % of the center frequency. The measured impedance bandwidth is the same as that modeled by NEC4. Overall, the agreement between the measured and modeled input impedance was good. The input impedance of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7 using the feed illustrated in Figure 4.40 is illustrated on the Smith Chart in Figure

6 Figure The measured input impedance of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7. Figure 4.45 shows that when the length of the box, L b, is decreased, the diameter of the impedance locus increases. The resonant frequencies of the DIFA become 919 and 942 MHz. The input resistance at the first resonance is 26 Ω and at the second resonance is 117 Ω. The VSWR of the DIFA operated against box 4 is compared to the VSWR modeled by NEC4 in Figure

7 Figure The measured VSWR of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7 compared to the same VSWR modeled using NEC4. The measured impedance bandwidth in Figure 4.46 for the DIFA operated against box 1 (6 X 8 cm) is 23 MHz, or 2.5% of the center frequency. This is just more than half the bandwidth of the DIFA operated on box 3 (6 X 12 cm). The measured impedance bandwidth was 43 % wider than that predicted by NEC4. The radiation patters of the DIFA operated over boxes 1, 3, and 4 in Table 4.7 were cut using the Virginia Tech Rooftop range and the configuration described in Section The notation on the patterns refers to the spherical coordinate system in Figure The source and test antennas were configured as illustrated in Figure In Figure 145

8 4.47, the radiation patterns of the DIFA operated against box 3 (6 X 12 cm) in Table 4.7 at 915 MHz are compared to the same radiation patterns modeled by NEC4. Figure The measured co- and cross-polarized far-field radiation patterns of the DIFA operated against box 3 (6 X 12 cm) in Table 4.7 at a frequency of f = 915 MHz. a) The pattern in the x-y plane. b) The pattern in the y-z plane. c) The pattern in the x-z plane. 146

9 Figure 4.47(a) shows that the DIFA operated over box 3 (6 X 12 cm) is somewhat directive in azimuth. The maximum vertically polarized gain in the equatorial plane of the DIFA is just over 0 db. The minimum vertically polarized gain in the equatorial plane is -8 db. The azimuthal pattern is directed toward the parasitic element. Figures 4.42(b&c) show that the vertically polarized gain has an absolute maximum of 3 db in the θ = 0 direction. The null that appeared in the model below the ground plane in the x-z plane and the E θ pattern was also formed in the measurement. The cross polarization components were higher than expected in the x-y and x-z planes. The measured and modeled patterns agreed at many points. The increase in spurious pattern effects is probably due to the introduction of feedline radiation. The size-reduced ground plane allows currents to couple to the exterior of the coaxial line. The resulting radiation causes constructive and destructive interference that shows up in the patterns. A tell-tale sign of feed radiation is offset nulls. This effect is seen in the cross polarized component of Figure 4.47(a). The θ = 0 null in the co-polarized component of Figure 4.47(b) is filled in completely. This is probably due to a combination of feedline and multi-path effects. The feedline radiation is eliminated if the DIFA is configured as the source antenna, and is fed using a small oscillator inside the box. Since the feed is entirely enclosed within the grounded shield, the feedline radiation cannot escape and is not included in the measurement. Despite feedline radiation, and range effects, the measured patterns of the DIFA over box 3 matched the model reasonably. In Figure 4.48, the radiation patterns of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7 are compared to the same radiation patterns modeled using NEC4. 147

10 Figure The measured co- and cross-polarized far-field radiation patterns of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7 at a frequency of f = 915 MHz. a) The pattern in the x-y plane. b) The pattern in the y-z plane. c) The pattern in the x-z plane. 148

11 The pattern in Figure 4.48(a) is less omni-directional than the pattern in Figure 4.47(a). The maximum vertically polarized gain in the equatorial plane of the DIFA operated against box 1 (6 X 8 cm) in Table 4.7 is 0 db. The minimum vertically polarized gain in the equatorial plane is -9 db. Figure 4.48(a) shows that the cross polarization component in the x-y plane of the DIFA increases as the length of the conducting box, L b, decreases. This was predicted by NEC4. The gain in the x-z plane was lower than predicted by NEC4, and a strong directive behavior appeared in the y-z plane that was not modeled. The nulls in the y-z plane appeared but were tilted by about 15. Both the directive effect and null offset are due to feedline radiation. In Figure 4.49, the radiation patterns of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7 are compared to the radiation patterns modeled using NEC4. Figure 4.49(a) shows that the measured maximum vertically polarized gain in the equatorial plane of the DIFA was 0 db. This value is higher than predicted by NEC4. In the x-z and y-z planes, the antenna was less directive than predicted by NEC4. The pattern in Figure 4.49(a) is more omni-directional than that in Figure 4.48(a), but contains a dip that makes the minimum vertically polarized gain in the equatorial plane -8 db. The cross polarization level was higher than expected in the x-y and x-z planes and lower than expected in the y-z plane. The null in the y-z plane was completely filled in. Once again, the patterns showed signs of spurious feedline radiation. Overall, the behavior of the DIFA operated against the conducting boxes matched the results of NEC4 modeling. There were a few deviations, including impedance bandwidths that were wider than expected and unpredicted pattern effects due to feedline and multi-path radiation. The characteristics of the three DIFA models are compared to the characteristics modeled using NEC4 in Table

12 Figure The measured co- and cross-polarized far-field radiation patterns of the DIFA operated against box 4 (4 X 10 cm) in Table 4.7 at a frequency of f = 915 MHz. a) The pattern in the x-y plane. b) The pattern in the y-z plane. c) The pattern in the x-z plane. 150

13 Table Comparison of the measured and modeled DIFA characteristics for operation over conducting boxes Quantity Symbol Measured Value Modeled (NEC4) Value Unit Operation against box 3 in Table 4.7. (6 X 12 cm) Resonant Frequencies f r1, f r2 903, 926 No Resonance MHz Respective Input Resistances R in1, R in2 34, Ω at Resonances Impedance Bandwidth % of Center Frequency BW MHz % Azimuthal Radiation -- Directed Omni -- Maximum / Minimum V-Pol Equatorial Gain G 0 / -8-3 db Operation against box 1 in Table 4.7. (6 X 8 cm) Resonant Frequencies f r1, f r2 919, 942 No Resonance MHz Respective Input Resistances R in1, R in2 26, Ω at Resonances Impedance Bandwidth % of Center Frequency BW MHz % Azimuthal Radiation -- Anomalous Omni -- Maximum / Minimum V-Pol Equatorial Gain G 0 / -9 0 db Operation against box 4 in Table 4.7. (4 X 10 cm) Resonant Frequencies f r1, f r2 889, 922 No Resonance MHz Respective Input Resistances R in1, R in2 52, Ω at Resonances Impedance Bandwidth % of Center Frequency BW MHz % Azimuthal Radiation -- Directed Omni -- Maximum / Minimum V-Pol Equatorial Gain G 0 / -8-2 db 151

14 In this section, the IFA and DIFA were measured over a large ground plane and the DIFA was measured over a number of conducting boxes. The results of the measurement were used to verify the model in Section 4.3. In the next section, concluding remarks are made involving the modeling and measurement of the IFA and DIFA Conclusions In this chapter, the Inverted-F Antenna (IFA) and Dual Inverted-F Antenna (DIFA) were characterized using numerical modeling and empirical measurement. The Method of Moments (MoM) was chosen as a modeling technique for wire antennas. In Section 4.2 an algorithm was developed that applied a point matching MoM technique to solve the currents on wire antennas of arbitrary geometry. In Section 4.3 the algorithm was applied to the IFA and DIFA using a commercial code called the Numerical Electromagnetics Code version 4 (NEC4). The results of the model showed the IFA and DIFA both exhibited azimuthally omni-directional radiation characteristics with similar maximum vertically polarized gain in the equatorial plane and high cross polarization components. The wide coverage and polarization insensitivity of the IFA and DIFA make them excellent candidates for application in the hand-held environment. However, the DIFA experiences an improvement in impedance bandwidth over the IFA. In Section 4.4 the DIFA was measured over a large ground plane and compared to similar measurements of the IFA. The results of the measurement were used to verify the model in Section 4.3. The design issues for the DIFA were studied over size-reduced conducting boxes. Section 4.4 showed that the numerical modeling technique used by NEC4 was more accurate for far-field radiation than input impedance, but gave a first cut solution that could be verified by measurement. The measured characteristics of the IFA and DIFA matched the NEC4 model well, and supported the fact that both antennas are well suited for applications in the hand-held environment. 152

15 REFERENCES FOR CHAPTER 4 1. Burke, G. J. Numerical Electromagnetics Code - NEC-4, Method of Moments, Part II: Program Description - Theory. Numerical Electromagnetics Code Documentation, January Stutzman, W. L., Thiele, G. A. Antenna Theory and Design. John Wiley and Sons, New York, Harrington, R. F. Field Computation by Moment Methods. Macmillan, New York, Burke, G. J. Numerical Electromagnetics Code - NEC-4, Method of Moments, Part I: User s Manual Numerical Electromagnetics Code Documentation, January Fuhl, J., Balducci, P., Nowak, P., Bonek, E. Internal Broadband Antenna for Hand-held Terminals with Reduced Gain in the Direction of the User s Head. Proceedings of the IEEE Vehicular Technology Conference, 1996, pp