Welcome to the AMTA paper archive. Select a category, publication date or search by author.
(Note: Papers will always be listed by categories. To see ALL of the papers meeting your search criteria select the "AMTA Paper Archive" category after performing your search.)
The Compact Antenna Test Range (CATR) was initially conceived as an efficient way of testing electrically large antennas at very much reduced, fixed, range lengths than would otherwise be the case. However, when testing lower gain, physically smaller antennas, the measurements can become susceptible to inhomogeneities within the CATR QZ including phenomena associated with edge diffraction effects, feed spill-over, chamber multipath etc. Whilst it has been demonstrated experimentally that many of these measurement artefacts may be effectively mitigated using standard and modern more sophisticated post-processing techniques. This paper supports those findings through simulation of the direct and indirect far field ranges and by careful examination of the data processing chain. Results are presented, the relative success of the various techniques examined and the utility of this is set, and expounded, in the context of modern, i.e. 5G, communications systems.
Radiating performances of vehicle-installed antennas are typically performed in large spherical near-field systems able to accommodate the entire car. Due to the size and weight of the vehicle to be tested, such spherical systems are often nearly hemispherical, and the floor is conductive or covered with absorbers. The main advantage of the first is the ease of the accommodation of the vehicle under test. Conversely, the latter is more time consuming in the setup of the measurements because the absorbers need to be moved in order to be placed around the vehicle. On the other hand, the absorber-covered floors emulate a free-space environment which is a key enabling factor in performing accurate measurements at low frequencies (down to 70 MHz). Moreover, the availability of the free-space response allows easy emulation of the cars' behaviors over realistic automotive environments (e.g. roads, urban areas etc.) with commercially available tools. Such emulations are instead much more challenging when a conductive floor is considered. Furthermore, the raw measurements over conductive floors are a good approximation of realistic grounds (such as asphalts) only in a limited number of situations. For these reasons, when PEC-based automotive measurements are performed, it is often required to retrieve the free-space response, or equivalently, to remove the effect of the conductive ground. In this paper two spatial-filtering techniques (the spherical modal filtering and the equivalent currents) will be experimentally analyzed and compared to verify their effectiveness in removing the effect of the conductive floor. For this purpose, a scaled automotive PEC-based measurement setup has been implemented considering a small spherical multi-probe system and a 1:12 scaled car model. The two techniques will be analyzed considering two different heights of the scaled car model with respect to the conductive floor.
Electromagnetic Compatibility (EMC) radiated emissions measurements above 1 GHz are performed in a nominal free space environment as required by international standards, typically in an anechoic chamber. In an EMC chamber, the test zone consists of a circular region defined by a turn table, where an equipment under test is rotated and measured. The test zone is commonly referred to as quiet zone (QZ). Due to the non-ideal nature of absorbers, multiple reflections in the chamber affect the quality of the QZ. The constructive and destructive interferences from the reflections form standing waves in the QZ. The maximum value of the standing wave is used as the figure of merit for validation of testing facilities. Site Voltage Standing Wave Ratio (sVSWR) as specified in CISPR 16-1-4 is broadly used for the validation of test sites above 1 GHz. This method requires the measurement of six positions along a linear 40 cm transmission path at various locations in the QZ, with a frequency step of no greater than 50 MHz using an omnidirectional-like antenna (e.g. a dipole). Concerns have been raised that this method delivers an overly optimistic result due to both spatial and frequency domain undersampling. In this work, an alternative method to sVSWR for the validation of EMC chambers based on Spherical Mode Coefficients (SMC) is proposed. Two 90 •-rotated measurement cuts of an omnidirectional-like antenna are acquired around the periphery of the circular QZ. The measured situation and cut is replicated by applying translation and rotation of spherical waves to the known SMCs of the used omnidirectional-like antenna and transforming using the spherical wave expansion. The generated and measured cut are compared and the characteristics of the chamber are extracted. The major advantage of this method is the relatively high measurement speed and reliability.
The characterization of antenna radiation patterns in the millimeter wave band are particularly challenging. This is due to the fact that a misalignment of just a few millimeters between the probe and the antenna can generate substantial measurement errors. This paper describes a strategy to reduce measurement errors by introducing a highly precise measurement system using a 6-axis small robotic arm to characterize the performance of a phased array antenna operating at 60 GHz. The position accuracy of the robotic arm itself is approximately 20 m and a maximum far field distance of approximately 380 mm can be achieved. The robot is programmed to perform a spherical trajectory around the array with stops every 0.5⁰ along the path to gather the measured gain. It operates continuously by communicating with a computer, which triggers the network analyzer at preprogrammed locations. The system is tested initially using two horn antennas as the antenna under test (AUT), and the results are presented.
We present a novel, non-contact characterization technique for simultaneous characterization of conventional antenna parameters, including the antenna port input impedance, antenna gain and its radiation pattern, without requiring a network analyzer connection to the antenna port. The test antenna and the network analyzer are considered as a 2-port open-air fixture whose network representation corresponds to the desired antenna parameters. The unknown network parameters of the 2-port open-air fixture are determined via a novel calibration process using 4 offset-short termination standards. The error parameters determined by the calibration are then related to the test antenna port impedance and its gain as a function of frequency. Furthermore, the radiation pattern of the test antenna can also be characterized using measured reflection coefficient at the network analyzer port for two offset-short terminations of the test antenna port, while rotating the test antenna over the desired angular range. This novel technique is particularly attractive for installed-antenna applications where an active connection to the test antenna port is either difficult or undesirable, such as on-chip antennas and implanted antennas, to name a few. To demonstrate the efficacy our new method, we present the measured impedance, gain and radiation pattern of a diagonal-horn antenna operating over 360-450 GHz, and a lens-integrated planar butterfly antenna for the 220-325GHz band.
Antenna measurement errors occur due to reflections and diffractions within the measuring chamber. In order to extract and correct the undesired signals, a technique based on test-zone field compensation and spherical wave expansion is applied to Compact Antenna Test Range (CATR) and Spherical Near-Field (SNF) measurements of a base transceiver station antenna. The required spherical test-zone field is acquired by simulating the corresponding measurement environment with the multi-level fast multipole method. Due to the numerical complexity of the problem, only the parts of the chamber with a significant influence on the measurement results are modeled. Comparing the determined directivities after applying the correction method, an exact overlap is achieved between the SNF and CATR solution.
Wireless industry through 3GPP has standardized 5G in both FR1 (sub 6 GHz) and FR2 (24.25-52.6 GHz) frequency ranges. While FR1 will be using frequencies already in place for LTE-4G technology, FR2 is dealing with mmWave frequencies. Due to the high free space path loss (FSPL), 5G at mmWave would impose the use of directive antennas on both ends of the communication link, the User Equipment (UE) and the Base Station (BS). A black box approach (i.e. the location of the antenna within the device is unknown) has been agreed to be used for Over The Air (OTA) measurements. The physical center of the device must be aligned with the center of the measurement setup. Hence, the test antennas will likely be offset with respect to the center of the coordinate system. The measurement distance will be for most systems sufficient to minimize the amplitude error while will introduce a phase deviation between the actual spherical wave and the desired plane wave which may cause an effective phase shaping of the radiated beam of the small phased array under test. In this paper we will analyze the impact of the phase curvature on the beam antenna pattern and spherical coverage for the different testing environments. Specifically, simulation of a 5G terminal device with multiple beams will be considered and realistic spherical near field measurement at different finite distances will be emulated also taking into account different measurement antennas (probes).
The major drawback of Spherical Near-Field (SNF) measurements is the comparatively long measurement time, since the scanning of a whole sphere enclosing an Antenna Under Test (AUT) is required to calculate the Spherical Mode Coefficients (SMCs) required for the computation of the far field. Since the SMCs prove to be sparse under certain conditions, efforts have been made to apply compressed-sensing techniques to reduce the measurement time by acquiring a smaller number of sampling points. These approaches have been successfully tested in simulation using classically acquired measured data. This decouples the measurements from practical problems, such as basis mismatch due to the finite precision of the mechanical positioner and environment effects. In this paper, results from a sparse data acquisition performed with a physical system are reported. To decouple the error introduced by the approach itself from the error introduced by non-idealities in the measurement system, an AUT is measured using both traditional near-field sampling and compressed near-field sampling. The classically acquired data is used both as reference and as source to simulate a synthetic compressed measurement. The effects introduced by real considerations are calculated by comparison between the synthetic compressed measurement and the acquired one, while the error of both is evaluated by comparison to the reference measurement. The results further demonstrate the viability of this method to accelerate SNF measurements and pave the way for further research.
The Three-antenna polarization measurement technique is used to determine the axial ratio, tilt angle and sense of polarization of three antennas from measurements on each of three antenna pairs. The three antennas are generally nominally linearly polarized and the measurement data consists of the change in amplitude from the initial antenna orientation where they are co-polarized to the orientation where one of the antennas is rotated about its axis to the null amplitude position. The sign of the phase change is also noted and the phase change at the null position is known from theoretical calculations to be either plus or minus 90 degrees. The correct sign is determined from the sign of the phase change. For antennas with axial ratios in the range of 50 to 80 dB that will be used as near-field probes or as feeds for reflector antennas, it is imperative to measure the polarization parameters as accurately as possible. The primary source of uncertainty in the measurement is due to scattered signals in the measurement range that arise from multiple reflections between the two antennas and from the absorber on the chamber walls. For antennas with very large axial ratios, the scattered signals can be larger than the true measurement signal. These scattered signals can change the sign of the phase and produce large errors in the amplitude at the null. If the separation distance between the antennas is adjusted after rotating to the null to produce a maximum amplitude, the scattered signal is in phase with the true measurement signal. If the distance is adjusted for the minimum at the null, the scattered and true signals are out of phase. Measurements at these two positions will produce the best measurement of the phase sign and the true amplitude. But if measurements are being performed at a number of frequencies, the maximum and minimum amplitude positions will be different for each frequency, and this will complicate automated multifrequency measurements. New improvements have been developed in the details of the measurements that greatly improve the determination of the phase sign and the amplitude at the null for multiple frequency measurements and these will be described and illustrated in the following paper. With these improvements, the estimated uncertainty of a 60 dB axial ratio is on the order of 1.8 dB. A new technique has also been developed to improve the source correction of the pattern data for probes with large axial ratios that guarantees that the on-axis polarization of the pattern data will be identical to the results of the Three-antenna measurement. The probe correction processing will then produce the highest accuracy results for the polarization of the AUT.
This paper introduces the new Portable Laser Guided Robotic Metrology (PLGRM) system at the National Aeronautics and Space Administration's (NASA) Glenn Research Center. Previous work used industrial robots in fixed facilities to characterize antennas and required fixtures that do not lend themselves to portable applications. NASA's PLGRM system is designed for in-situ antenna measurements at a remote site. The system consists of a collaborative robot arm mounted on a vertical lift and a laser tracker, each on a mobile base. Together, they enable scanning a surface larger than the robot's reach. To accomplish this, the robot first collects all points within its reach, then the system is moved and the laser tracker is used to relocate the robot before additional points are captured. The PLGRM implementation will be discussed including how safety and planning are combined to effectively characterize antennas. Software defined triggering is a feature, for flexible integration of vector network analyzers and antenna controllers. Lastly, data will be shown to demonstrate system functionality and accuracy.
The aim of the paper is to address a relevant issue in the Near-Field (NF) measurements: the reduction of the measurement time. Generally speaking, for a given hardware, two main directions can be pursued. The first requires the adoption of an optimal field sampling strategy that reduces the number of sampling points, and the length of the scanning path, without impairing accuracy. The second strategy adopts an optimized control system able to exploit at the best the available hardware (scanning system and measurement instrument). Indeed, the latency of the instrument defines the maximum probe velocity during the field acquisition. Accordingly, unlike the conventional continuous scanning, an optimized controller can speed up the scanning by moving the probe along the measurement trajectory with a variable velocity, accelerating and decelerating between two consecutive sampling points, to increase the average speed. However, the use of an optimized controller is fruitful only when the optimized sampling scheme allows large distances between two consecutive sampling locations, to increase as much as possible the maximum probe speed. In this paper, by suitably using both the above strategies, it is proposed a fast NF system, implemented on a microcontroller Arduino Due, an extremely cheap and off the shelf hardware, that is able to handle the scanner and realize the synergy between the optimized sampling and the optimized control strategy. The simulation and experimental results show a dramatic reduction of the measurement time (up to one order of magnitude) with a high tracking precision (also in accordance with the proposed methodology), and of the costs with respect to standard solutions.
Full probe compensation techniques for Spherical Near Field (SNF) measurements have recently been proposed [1-5]. With such techniques, even antennas with more than decade bandwidth are suitable probes in most systems. The abolition of otherwise frequent probe changes during multi-service campaigns is a highly desirable feature for modern measurement applications such as automotive. In this paper, a standard dual-ridge horn with 15:1 bandwidth is investigated experimentally as probe in a SNF automotive range. The accuracy of the probe compensation technique is validated by comparison to standard single probe measurement.
The National Institute of Standards and Technology (NIST) has measured a WR-8, 3D printed horn at 112.25, 118.75, and 125.25 GHz using the near-field spherical scanning method. The data were processed with both the NIST standard software and the probe-position compensation software. We conclude that the positioning capability of the NIST Configurable Robotic Millimeter-wave Antenna System is so accurate that probe-position compensation is negligible at these frequencies.
An experimental case of spherical probe-corrected phaseless near-field measurements with the two-scans technique is presented, based on magnitude measurements at two surfaces of the VAST12 reflector antenna performed at the DTU-ESA Facility. Phase retrieval using strictly the directly measured near-field magnitude was unfeasible in this setup, due to the small sphere separation allowed by the probe positioner, which led to incorrect and excessively slow convergence. Phase retrieval with larger separation between spheres has shown remarkable results. For these tests a measured magnitude was used in combination with calculated near-field magnitudes at different (larger and smaller) spheres with larger separations than allowed by the experimental setup. It has been seen that larger separation between measurement spheres improves accuracy of phase retrieval. A measurement with a backprojected measurement with 3 m sphere separation is of particular interest because it can be potentially replicated in the DTU-ESA Facility assuming such range of movement was allowed, while being accurate down to an error of less than-35dB. Measurements with larger spheres show even better accuracy. These good results were obtained with the normal spatial sampling rate for complex measurements and with a very simple Hertzian dipole initial guess, and show the superior performance of spherical phaseless measurements with the two-scans technique, compared to a planar setup.
An algorithm is developed for the extraction of constitutive parameters from bi-layered uniaxial anisotropic materials backed by a conductive layer. A method of moments-based approach is used in conjunction with a previously-determined Green function. Possible challenges related to measurement diversity are highlighted and a possible mitigation path is proposed.
Specular reflectance data of indoor interior materials is a prerequisite to analysis of the channel characteristics for new millimeter and submillimeter indoor wireless communications. Antenna property such as gain and radiation pattern is one of the key measurement quantities in electromagnetic wave metrology. This paper describes a specular reflectance and antenna property measurement system and shows measurement results of the specular reflectance of an Acetal plate and the antenna property of a 24 dB horn antenna in 325-500 GHz frequency range.
In order to support near-field measurements of automobile antennas in as realistic as possible environments, an equivalent sources based near-field far-field transformation approach for near-field measurements above a possibly lossy dielectric half-space is presented and evaluated. Different possibilities for considering the half-space influence are discussed, where an approach with an appropriate half-space Green's function is found to be most accurate, as expected. The formulation of the equivalent sources transformation approach with the half-space Green's function and a formulation with free-space Green's function together with equivalent sources representation of the half-space influence are discussed and a variety of results are presented in order to corroborate the feasibility of the various approaches.
The Translated Spherical Wave Expansion (TSWE) has recently been proposed as a very effective Near-Field-to-Far-Field (NF/FF) transformation tool for down-sampled Spherical Near Field (SNF) measurements with offset Antenna Under Test (AUT). In case of electrically small probes and/or small AUT-probe view angles the TSWE can be accurately applied without compensating for the probe effect. Instead, when electrically larger probes and/or larger view angles are considered, the measured signal is affected by an averaging field effect that should be properly compensated to ensure a good accuracy. In this paper the TSWE technique is applied for the first time tacking into account the full effect of the measuring probe. To validate the proposed technique, a standard gain horn intentionally displaced in offset configuration have been measured in SNF geometry with a first order probe and two different wideband higher-order antennas as probe.
The unknown-thru calibration technique is being used to achieve a system level calibration at millimeter wave frequencies (>50 GHz) on the robotic ranges at NIST. This two-port calibration requires the use of a full bi-directional measurement, instead of a traditional single-direction antenna measurement. We explored the value of the additional data acquired. We find that we can use this information to verify antenna/scan alignment, image the scattering from the positioner/facility, and perform a first order correction to the transmission data for uncertainties due to LO cable flexure.
We present a preliminary study of a modal (partial wave) expansion of the field used to characterize a propagation channel. We assume that the measurements of the scalar, two-dimensional field from which the modal expansion coefficients are obtained, contain Gaussian phase noise with zero mean. Three spatial sampling patterns of the field are considered. We find that the accuracy of the reconstructed field is strongly influenced by the spatial sampling pattern.