Introduction
Millimeter-wave (mm-wave) frequencies are actively used in various wireless systems, including radar systems, satellite communication, and 5G mobile communication [1, 2]. Due to the saturation of the low-frequency spectrum, the frequency spectrum being used is becoming higher. For instance, in recent times, W-band applications in the 75 GHz to 110 GHz frequency band have been widely developed, including automotive radar, weather and monopulse radar, and satellite communications [3–6].
The front end of most mm-wave radio systems comprises antennas. These antennas require high gain and beam steering capabilities to compensate for significant free space path loss. Usually, antenna array structures or large aperture antennas are employed to achieve high gain, while mechanical steering devices or phased array antennas are utilized for beam steering. As a result, antennas in mm-wave applications are becoming increasingly complex in terms of structure and functionality compared to conventional antennas [7, 8].
In this context, W-band monopulse antennas mounted on radar systems—one of the most actively developed W-band applications—are characterized by high gain and a large aperture. These antennas typically employ a Cassegrain reflector antenna structure and a monopulse feed structure [9, 10]. The monopulse radar determines the direction of a target by analyzing the magnitude and phase differences between the sum and difference channels.
The performance of the monopulse antenna dramatically affects the performance of the entire monopulse radar system. Therefore, developing an efficient measurement system to evaluate the performance of the W-band monopulse antenna is crucial. Additionally, the presence of multiple input/output ports in the monopulse antenna necessitates comprehensive measurements, which far exceed the requirements of simpler antenna structures, to ensure accuracy and effectiveness [11, 12].
Among the primary performance parameters of an antenna, the gain and radiation pattern are measured using far-field and near-field measurement systems within anechoic chambers [13–16]. With regard to measurement speed, far-field measurements can generally be acquired faster than near-field measurements with regard to measuring planar radiation patterns. Moreover, when measuring large aperture antennas, such as W-band monopulse antennas, the required far-field distance increases. Therefore, constructing a far-field range antenna measurement system requires a large amount of space, such as an open area test site or a large anechoic chamber [17, 18]. Such expansive spaces also necessitate the use of longer cables in the measurement system, leading to increased signal loss.
Drawing on this background, this paper proposes a far-field range antenna measurement system capable of efficiently and accurately measuring a W-band monopulse antenna gain with a large aperture and multiple input/output ports. The proposed measurement system uses a multi-channel rotary joint, a pair of diplexers, and multiple vector network analyzer extenders (VNAXs). Furthermore, an RF sub-system that we previously developed [16] is employed for phase measurement. This subsystem utilizes a shared local oscillator (LO) and external power amplifiers (PA) to compensate for cable losses. Overall, the proposed structure enables the simultaneous measurement of the magnitude and phase of the signals received by the multiple ports of the W-band monopulse antenna.
The proposed W-band multi-channel antenna measurement system was built inside a large-sized anechoic chamber. A W-band monopulse antenna, developed in-house, was utilized as the antenna under test (AUT) to verify the measurement performance of the system. Notably, by simultaneously measuring the sum/difference channels of the monopulse antenna according to the azimuth angle, the relative magnitude and phase of the sum/difference signals could be measured both accurately and rapidly, thus minimizing the impact of phase drift. Furthermore, the accuracy of the simultaneous dual-channel measurement was validated through a comparison with the results obtained from the near-field measurements.
Antenna Measurement System Requirements
1. Angular Resolution
Fig. 1 shows a block diagram of a typical monopulse antenna used in radar systems to identify the direction of a target [19]. It comprises arrays and a feeding sub-system. The feeding subsystem is composed of a comparator and four hybrid junctions. The monopulse comparator generates three channels: a sum channel, a difference channel for the azimuth, and a difference channel for elevation. Notably, a monopulse detects the direction of the target by analyzing the magnitudes and phase differences of the sum and difference channels. This indicates that it is crucial to determine whether the magnitude and phase of the sum and difference patterns satisfy the design values. In particular, the angular resolution near the main beam and the null of the difference pattern should be precisely measured. Notably, the method proposed in this paper requires an angular resolution of about 0.01° for the measured AUT.
2. RF Sub-system
We used WR10 frequency extension modules such as VNAX to extend the measurement frequency range of the VNA. This VNAX generates W-band and intermediate frequency (IF) signals by multiplying the RF (12.5–18.4 GHz) by 6 (×6) and LO signals (9.3–13.8 GHz) by 8 (×8).
The AUT—the W-band monopulse antenna developed in-house—featured a large aperture of approximately 37.5λ0 × 37.5λ0, where λ0 is the wavelength at the center frequency. Considering its aperture size and operating frequency band, the AUT required a measurement distance of at least 15 m to satisfy the far-field conditions (2D2/λ). Therefore, 10 m or longer RF and LO cables were required for each transmit (TX) and receive (RX) path, which would result in considerable loss. To reduce path loss in LO cables, a previous study [15] proposed an RF sub-system that minimizes the length of LO cables by placing two distributed LO sources synchronized by 10 MHz as close to the TX and RX antennas as possible. However, measuring the phase in the W-band was difficult due to the phase drift occurring between the two LO sources [20].
Therefore, in this study, we applied an RF sub-system developed in our previous research [16] for phase measurement. This sub-system uses a shared LO and compensates for path loss using external power amplifiers to improve the phase stability and synchronization of the LO signals of the TX and RX paths, as shown in Fig. 2 [16]. As seen in the schematic in Fig. 2, a double external PA is symmetrically applied to both the TX and RX paths. This approach places the LO source at the center of the overall measurement system. As a result, the TX and RX paths have the same length, and the path loss of each path is compensated by adding an external PA to each path.
The transmission phase drift of the RF sub-system was verified by measuring insertion loss S21 [16]. After warming up and calibrating the RF and fixing the LO cables, S21 was measured at 23.5±1°C by directly connecting ports 1 and 2 of the VNAXs at 1-hour intervals for 24 hours. For comparison, the measurement was repeated for a basic RF sub-system configuration using short cables without an external PA. The measurement results showed that the RF sub-system in Fig. 2 is sufficiently suitable for phase synchronization and stabilization compared to the basic RF sub-system configuration [16].
3. Design of the Multi-Channel Antenna Measurement System
Another issue that should be considered in the context of this study is measurement time. As shown in Fig. 1, the monopulse antenna comprises several input ports. Measuring the pattern for each port is bound to take considerable time. In particular, measurements conducted in high-frequency bands, such as the W-band, often cause drifts, even in environments with a stable temperature. Therefore, wherever possible, it is important to reduce the measurement time by measuring several channels simultaneously.
To address this issue, we propose a W-band multi-channel antenna measurement system structure, as shown in Fig. 3, that enables simultaneous measurement of the multiple ports of an AUT using a multi-channel rotary joint, a pair of diplexers, and several VNAXs.
This section explains the multi-channel measurement system in detail. The TX antenna used in the proposed system is the same as that used in a conventional AUT measurement system. The TX VNAX is mounted on top of the TX mast, while the TX antenna is mounted on the TX VNAX. The TX antenna uses a W-band standard gain horn antenna. Furthermore, in the RX path, two or more VNAXs are mounted on top of the RX mast, with the ports of each RX VNAX connected to two or more ports of the AUT. The RX mast is then mounted on the positioner for rotation in azimuth directions, with the range being 360°. Furthermore, since the cable connected to the receiving antenna is expected to vibrate or twist during rotation, a two-channel rotary joint is installed in the rotation axis of the positioner.
Fig. 4 shows the signal flow graph of the proposed measurement system, considering a case where the system has to measure two channels simultaneously. The VNA is located at the center of the entire system, with the external PAs compensating for the loss caused by the long RF and LO cables. The identical external PAs symmetrically installed on each TX and RX path serve to amplify the divided LO signals with a reduction of the phase drift. The TX VNAX multiplies the LO and RF signals from the VNA and downconverts them to the IF signal. Subsequently, the reference IF captures the magnitude and phase information of the input signal to the AUT. The AUT then receives the W-band signal radiated from the TX antenna and outputs a plurality of RF channel signals (RF CH1, RF CH2, RF CH3) through the monopulse comparator, as shown in Fig. 2. Since the proposed system can measure two channels simultaneously, RF CH1 and RF CH2, for example, are connected to RX VNAX1 and RX VNAX2, respectively.
Similar to the TX VNAX, the RX VNAXs generate test IFs, which are sent to the VNA through the rotary joint. In this case, a three-channel rotary joint is required to transmit the LO, TEST IF1, and TEST IF2 signals. Moreover, as described above, the frequency of LO is in the 9.3–13.8 GHz range, while the frequency of IF is tens to hundreds of MHz.
We applied a diplexer to combine the LO and TEST IF signals into one cable and then separated the signal through the rotary joint. As a result, we could use a commercial two-channel rotary joint that is readily available.
This paper describes a dual-channel antenna measurement system as an example, demonstrating that the system can be expanded to measure more than two channels simultaneously. However, this is limited by the number of channels that a rotary joint can support. Notably, diplexers can reduce the number of channels required for a rotary joint.
4. Stability Analysis of the Dual-Channel Measurement System
An experiment was conducted to verify whether the two receiving channels of the proposed system could conduct stable measurements without deviation. Fig. 5(a) presents a photograph of the experimental environment for verifying the long-term stability of the dual-channel antenna measurement system. The signal from the VNA was multiplied by the W-band through the TX VNAX. In addition, the insertion loss of the W-band signal was measured by directly connecting the port of the TX VNAX to the two RX VNAX ports. In particular, the signal of the TX VNAX was distributed using a 10 dB W-band directional coupler to the two RX VNAXs. The TX VNAX, RX VNAX1, and RX VNAX2 are denoted as port numbers 1, 2, and 3, respectively. Finally, S21 and S31 were measured for 10 hours in a laboratory environment of 23.5±1°C.
Fig. 5(b) and 5(c) show the change in the relative magnitude difference and relative phase difference between S21 and S31 at the center frequency over time, respectively. In both figures, all magnitude and phase values are normalized based on the start-time value. Fig. 5(b) indicates that the deviation in the magnitude of the signals measured by the two RX VNAXs remained stable within 0.05 dB for 10 hours. Similarly, Fig. 5(c) shows that the phase deviation of the signals measured by the two RX VNAXs remained stable (within 1.2°) for 10 hours. In particular, it remained extremely stable (within 0.4°) after the initial warm-up time of 2 hours. These results emphasize that the same values can be consistently obtained by calibration, even when the ports of an arbitrary AUT are alternately measured using two RX VNAXs. In other words, it is possible to measure the two channels of the AUT simultaneously by constructing a two-channel antenna measurement system.
Measurement of Monopulse Antenna
In general, the gain and radiation pattern measurements of W-band antennas use either far-field or near-field test facilities. In the case of far-field measurement, a standard gain horn antenna is usually employed as the reference antenna, while an open-ended waveguide is used for near-field measurement. In this context, the accurate calibration of the reference antenna is a significant factor concerning measurement uncertainty.
In this context, the easiest way to ensure accurate measurement is to calibrate the gain of a reference antenna using a standard antenna gain calibration system with very low uncertainty, and then apply it to the far-field or near-field measurements.
As reported in previous studies [13, 15], we have previously developed a standard antenna gain calibration system. A three-antenna method based on extrapolation was employed to calibrate the far-field gain of a standard gain horn antenna with low uncertainty. Applying this calibrated standard gain horn antenna to the proposed antenna measurement system would help measure AUT performance accurately. Notably, the uncertainty of the proposed W-band standard antenna gain calibration system was less than 0.3 dB (95% confidence interval). In the present study, the gain of the reference antenna was evaluated using a standard antenna gain calibration system [15], and then applied to the W-band monopulse antenna measurement system to enable accurate measurement.
As shown in Fig. 6(a), we constructed an mm-wave far-field range antenna measurement system for a W-band monopulse antenna. The far-field gain and radiation pattern of the W-band monopulse antenna were measured using the two-antenna method. The proposed W-band multi-channel antenna measurement system was then installed inside a large-sized anechoic chamber of dimensions 24.0 m × 12.0 m × 12.0 m. To measure the radiation pattern of the W-band monopulse antenna, the positioner was rotated in the azimuth direction from the reference angle. At the same time, the measurement system measured the magnitude and phase of the received signal. Since the two channels could be measured simultaneously, the radiation pattern of the sum/difference channels of the monopulse antenna could be quickly measured according to the azimuth angle.
Since the beam width of the W-band monopulse antenna was very narrow when measuring the null depth of the difference pattern, a narrow measurement range had to be specified, and the angular resolution had to be precisely measured. Therefore, the antenna measurement range was measured by dividing it into a ±45° range with 0.1° resolution to observe the overall radiation pattern and a ±3° range with 0.01° resolution to closely observe the surroundings of the main beam. The average measurement times for the ±45° range with 0.1° resolution and the ±3° range with 0.01° resolution were about 6 and 5 minutes, respectively.
As shown in Fig. 6(b), the near-field measurement system used in this study is a planar near-field system. The far-field radiation pattern of the AUT was obtained from its near-field scan results through near-field to far-field transformation. This transformation incorporated techniques such as fast Fourier transform and probe correction [21]. Subsequently, the results of the near-field measurement system were compared with those of the dual-channel far-field measurement system to verify the measurement performance.
The near-field measurement system was installed inside an anechoic chamber measuring 7.0 m × 10.0 m × 6.0 m. The distance between the probe and the AUT was set to 12 mm, corresponding to the near-field region. The planar scanning area was set to a square area of about 63.0λ0 × 63.0λ0 in the xy-plane. The x- and y-axes sampling intervals were set to about 0.45λ0, where λ0 is the wavelength at the center frequency. With 141 sampling points in each axis direction, the total number of points measured across the entire range was 141 × 141 = 19,881. The average measurement time for near-field measurements in this range was approximately 5 hours.
Table 1 summarizes the specifications of the two measurement systems used in this study. However, it is important to note that the systems do not necessarily represent the performance of all far-field and near-field measurement systems.
Figs. 7 and 8 present the measurement results of the sum channel and difference channels of the W-band monopulse antenna, respectively. Far-field measurement (FF Meas.), indicated by the red solid line, was obtained by simultaneously measuring the sum and difference channels using the proposed dual-channel W-band antenna measurement system. Meanwhile, near-field measurement (NF Meas.), indicated by the blue dashed line, was obtained by separately measuring the sum and difference channels using the near-field measurement system.
Figs. 7(a) and (8a) show the normalized radiation pattern measurement results in the ±45° range with 0.1° resolution for the sum and difference channels, respectively. It is observed that the far-field results obtained using the proposed W-band multi-channel far-field antenna measurement system largely align well with the near-field measurement results. In particular, since the magnitude of the signal near the main beam is large, the two results coincide with almost no distortion.
It is evident that as the azimuth angle increases, the magnitude of the signal decreases, leading to the appearance of some ripples caused by the effect of the sensitivity and dynamic range of the measurement system, although the overall patterns remain similar.
Figs. 7(b) and 8(b) show the normalized radiation pattern measurement results in the ±3° range with 0.01° resolution for the sum and difference channels, respectively. Notably, this range is situated near the main lobe, which contains the most critical information pertaining to the performance of the monopulse antenna. It was confirmed that the near-field and far-field measurements agreed well with both the sum and difference channel measurement results in Figs. 7(b) and 8(b). In particular, the resolution of 0.01° made the measurements possible, due to which the null depth of the difference pattern could be expressed very precisely. Furthermore, it was confirmed that the near-field and far-field patterns aligned precisely within 0.1°.
Figs. 7(c) and 8(c) show the phase pattern measurement results in the ±3° range with 0.01° resolution for the sum and difference channels, respectively, confirming that the phase change tendencies of the near-field and far-field measurement results are similar. Except for the section where the phase value changes rapidly, the values show good agreement. The difference observed in the part featuring rapid changes in the phase value can likely be attributed to the relatively large conversion error that occurs during the near-field-to-far-field transformation near the phase inflection point during the near-field measurement. Notably, in our existing W-band standard antenna gain calibration system [15], phase measurement of the far-field pattern is impossible. In contrast, the proposed system applied the RF sub-system topology shown in Fig. 2 to enable phase measurement [16]. As a result, the proposed antenna measurement system was able to measure the phase of the far-field pattern, as demonstrated by the experimental results. A similar trend was confirmed when comparing the results to those of the near-field pattern.
In terms of measurement time, when measuring the W-band monopulse antenna AUT, the near-field measurement took several hours, whereas the far-field measurement took several minutes to tens of minutes. In an antenna structure with multiple input/output ports, such as a monopulse antenna, the measurement time increases proportionally with the increase in the number of measurement ports. Notably, reducing measurement time is essential because, for instance, when mass-produced expensive high-tech equipment require comprehensive inspection, the proposed multi-channel far-field range antenna measurement system offers the advantage of dramatically reducing measurement time by measuring several channels simultaneously.
Conclusion
In this paper, we propose an mm-wave far-field range antenna measurement system for W-band monopulse antennas. The proposed dual-channel antenna measurement system is based on a W-band RF sub-system [16] capable of measuring the phase of a radiation pattern. The proposed system accounts for conditions requiring multiple measurements due to the presence of multiple input/output ports, such as a W-band monopulse antenna, enabling precise measurements while efficiently reducing the measurement time. When used with more RX VNAXs and rotary joints with more channels, the proposed system is sufficiently scalable beyond a dual-channel system to become a multi-channel antenna measurement system.
The sum/difference patterns of the W-band monopulse antenna were simultaneously measured based on the azimuth angle using the proposed system. The measurement results were verified through a comparison with the near-field measurement results. In particular, the proposed system exhibited an advantage over conventional near-field and far-field measurement systems in terms of measurement time.
Apart from the W-band monopulse radar system, many other applications in the mm-wave band utilize multiple antennas or feature multiple input/output ports. Since most of these applications involve expensive high-tech equipment, comprehensive inspection is required during mass production, the measurement process for which is costly and time-consuming. Utilizing the proposed multi-channel antenna measurement system is expected to effectively reduce the time and cost involved in these measurements.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
