4.4. Experimental Results and Analysis

Transcription

1 4.4. Experimental Results and Analysis Measurement of the IFA Against a Large Ground Plane The Inverted-F Antenna (IFA) discussed in Section was modeled over an infinite ground plane using NEC4. In practice, a ground plane of infinite extent is impossible. However, a large ground plane is a known approximation to an infinite ground plane. In this section, the input impedance and radiation patterns of the IFA in Section are measured with the antenna positioned in the center of a flat, square aluminum ground plane that is 1.23 meters (3.75λ at 915 MHz) on a side. Figure 4.32 illustrates the constructed prototype of the IFA on the large ground plane. y x Aluminum L s 2 L s 2 L s = 3.75 λ Figure The IFA designed in Section configured on a large ground plane The feedline of the IFA (wire 3 in Figure 4.7) in Figure 4.32 is connected to a 50 Ω coaxial line using an SMA flange-mount bulkhead connector. The connector is the reference plane for input impedance measurements made between 800 and 1000 MHz using an HP8720C network analyzer. The measured input impedance of the IFA on the large ground plane is shown on the Smith Chart in Figure

2 Figure The measured impedance of the IFA over the ground plane in Figure Figure 4.33 shows that the IFA is resonant at 938 MHz with an input impedance of 17.7 Ω. The NEC4 model of the IFA illustrated in Figure 4.12 produced a resonant frequency that was 16 MHz lower than the measurement, and an input resistance at resonance of 12 Ω. This corresponds to a 1.7% frequency deviation and a 47.5% impedance deviation. This supports the claim in Section that NEC4 impedance data should always be verified through measurement. The radiation patterns measured for the IFA are shown in Figure 4.35 using the planes defined by Figure 4.14 and Table 4.7. The patterns were measured using the Scientific Atlanta 1780 system on the Virginia Tech Rooftop Antenna Range. The IFA is 130

3 configured as a receive antenna with a log periodic dipole array (LPDA) as a source antenna. The IFA was rotated in azimuth and the field strength was sampled every 2. The test is automated using Range Runner Version 1.2, an internally developed software package. Range Runner reduces and displays the data. Another internally developed routine, REDUCE, is used to port the data to other computer packages. The source and test antenna configurations are illustrated in Figure Source Test X - Y plane, E φ H - Plane, Cross-Pol X - Y plane, E θ H - Plane, Co-Pol Y - Z plane, E θ E - Plane, Co-Pol Y - Z plane, E φ E - Plane, Cross-Pol X - Z plane, E θ E - Plane, Co-Pol X - Z plane, E φ E - Plane, Cross-Pol Figure Source and test antenna configurations and the corresponding radiation plane. The radiation patterns of the IFA in Figure 4.34 were cut using the configurations in Figure 4.34 and compared to the same radiation patterns from the NEC4 model of the IFA over an infinite ground plane. The absolute levels of the patterns are in terms of gain. The NEC4 radiation patterns were calculated in terms of gain. The Virginia Tech Antenna Lab and Range Runner produce patterns that are normalized to a reference level. The reference level is a decibel ratio of transmitted to received power level. Unpredictable range effect components make it impossible to predict antenna gain based on reference level alone. To determine the gain of an antenna, the vertically and horizontally polarized 131

4 reference levels of a standard gain dipole are measured. Then, the reference levels of the test antenna, such as the IFA, are measured using an identical measurement configuration. Referencing the pattern of the dipole to the pattern of the test antenna gives gain in dbd, decibels above a dipole. The absolute level of the pattern in db is found by adding the gain of the dipole (2.15 db) to the dbd measurement. The result is the gain of the test antenna. This is the process used to determine the absolute level of the patterns illustrated in Figure As illustrated in Figure 4.35, the measured radiation patterns matched the results of the NEC4 model closely in the x-y plane. However, there was an unexpectedly high cross polarization component that NEC4 did not predict. In the x-z plane, the measured data fit the model well, but, again, there was a unpredicted cross polarization component. In the y-z plane, the null in the E φ pattern was offset about 30, and the null in the E θ pattern at θ = 0 was filled in. Since the IFA is an omni-directional antenna and the LPDA source is also wide beam, energy from oblique bounce paths leaks into the pattern nulls and causes them to fill in. In addition, a significant ground reflection is present. If the test antenna is not perfectly centered during azimuthal rotation, slight lateral variations in position change the phasing of the bounce paths and cause measurement error. These multi-path and range effects could account for the variation of the measurements from the model, and for the unexpectedly high cross polarization levels. The maximum vertically polarized (E θ ) gain of the IFA in the equatorial (x-y) plane was measured at 3 db, very close to the modeled value of 3.12 db. 132

5 Figure The measured co- and cross-polarized far-field radiation patterns of the IFA in Figure 4.32 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. 133

6 4.8. The measured and modeled data for the IFA is summarized and compared in Table Table 4.8. Comparison of the measured and modeled IFA characteristics Quantity Symbol Measured Modeled Unit Value (NEC4) Value Resonant Frequency f r MHz Input Resistance at R in Ω Resonance Azimuthal Radiation -- Omni Omni -- Maximum Gain G db Maximum Cross Polarization Level in the Principal Planes (Ratio: G co - G cross ) CPL x-y CPL y-z CPL x-z 8 Inf db -6 -Inf db 3 Inf db In the next section, the Dual Inverted-F Antenna (DIFA) is measured over the same ground plane and using the same measurement configuration that was applied to the IFA in this section Measurement of the DIFA Against a Large Ground Plane In Section the Dual Inverted-F Antenna (DIFA) was modeled over an infinite ground plane. In this section the input impedance and radiation patterns of a constructed prototype of the DIFA in Section are measured using the aluminum ground plane described in Section and illustrated in Figure

7 y x Aluminum L s 2 L s 2 L s = 3.75 λ Figure The DIFA designed in Section configured on a large ground plane. The DIFA in Figure 4.36 is connected to a 50 Ω coaxial feed and the VSWR is measured using the configuration in Section The result of the VSWR measurement is illustrated in Figure Figure The measured VSWR of the DIFA in Figure

8 The measured impedance bandwidth of the DIFA is 36 MHz or 4 % of the center frequency. The two resonant frequencies are 920 MHz and 938 MHz. The input resistance at 920 MHz is 52 Ω and at 938 MHz is 84 Ω. The measured impedances are higher than those modeled by NEC4. The measured VSWR is compared to the model calculations in Figure Figure Comparison of the measured VSWR of the DIFA to the VSWR modeled using NEC4. As shown in Figure 4.38, the measured impedance bandwidth of the DIFA is slightly wider than that predicted by NEC4. The measured center frequency of 920 MHz is 7 MHz lower than the modeled one. The overall agreement between the measured and modeled data is good. 136

9 The radiation patterns of the DIFA operated against the large ground plane, cut using the Virginia Tech Rooftop Antenna Range and the configuration described in Section for the IFA, are illustrated in Figure The patterns are cut using the source and test antenna configurations illustrated in Figure The absolute levels of the patterns are determined using the gain measurement process described in Section Figure 4.39(a) shows that the DIFA is omni-directional in azimuth with a gain of 3 db. The patterns of the DIFA are almost identical to those of the IFA. The E θ pattern in the x- y plane was close to the pattern modeled using NEC4, but had an unpredicted cross polarization component (E φ ). In the y-z plane, the null in the E θ pattern was offset by a few degrees. There was an unexpected null in the x-z plane pattern near θ = 0, and the cross polarization level was higher than expected. These unexpected features of the DIFA patterns are possibly attributable to the range and multi-path effects discussed in Section Overall, the measured behavior of the DIFA supported the results of the NEC4 model. The measured and modeled characteristics of the DIFA are summarized and compared in Table 4.9. In the next section the input impedance and radiation patterns of the DIFA are measured against conducting boxes of constrained size. Table 4.9. Comparison of the measured and modeled DIFA characteristics Quantity Symbol Measured Modeled Unit Value (NEC4) Value Resonant Frequencies f r1, f r2 920, , 930 MHz Respective Input Resistances R in1, R in2 52, 84 30, 60 Ω at Resonances Azimuthal Radiation -- Omni Omni -- Maximum Gain G db Maximum Cross Polarization Level in the Principal Planes (Ratio: G co - G cross ) CPL x-y CPL y-z CPL x-z 9 Inf db -10 -Inf db -10 Inf db 137

10 Figure The measured co- and cross-polarized far-field radiation patterns of the DIFA in Figure 4.36 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. 138