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 optimized approach to Near-Field sampling based on Singular Value Optimization allowing the definition of the optimal distribution of the number and measurement locations is here applied to the case of aperture antenna and the hemispherical scanning geometry. The numerical validation shows the performance of the approach and its robustness against the noise on the data. A numerical validation is presented.
Anechoic ranges require constant temperature and humidity, proper lighting to be able to work inside the range and closed-circuit television (CCTV) cameras to monitor the system while the measurement is being done. In addition, anechoic chambers require fire detection and suppression. Traditionally these penetrations are minimized and placed in non-critical areas. But the true effect of them has not been fully investigated. In this paper, antenna measurements as simulated in an indoor far field range. The approach to model the measurement is like the one the author presented in [1] and [2]. Thus, a range antenna (or near- field probe) and an antenna under test (AUT) are placed in free space and the AUT is rotated at discrete angles as it was done in [1]. Then a second model includes CCTV cameras, HVAC vents, light fixtures and both air sampling tubes and fire suppression nozzles and placed around. The simulation with these disruptions is repeated at the given discrete angles. The model does not include the absorber on the range. The model assumes a perfect absorber and the results of the simulated antenna measurement are compared to an ideal case with no disruptions. The results, while being approximations, provide a worst-case error for those disruptions of the RF-absorber layout. The results can be used to estimate the potential uncertainty on the measurement caused by the different systems that must be part of the anechoic enclosure. The technique is applied here to indoor far field measurements, and for near-field systems. Results show that for your typical roll over azimuth positioner, the effects of the penetrations on the ceiling are very small with differences in the -35 to -40 dB levels.
The first mention of a Robot for near-field measurements of antennas appears is by Jeff Snow in [1]. This was a simple robotic arm to do planar measurements. About 7 years later, the use of off-the-shelf industrial robotic arms for doing antenna measurements is introduced [2]. Since then, industrial- robot-arm based antenna measurement systems have become increasingly popular due to their flexibility to measure over different surfaces allowing the system to do planar, spherical and cylindrical. The use of other methods to perform the transform, by numerically compute the currents on an arbitrary surface from the measured fields has helped in the growing popularity of robotic systems. This is related that the measurement surface does no longer have to be a canonical surface but can be any shape. However, the flexibility of the robots may be limited by the RF absorber coverage used in treating them. In this paper, the authors explore the potential scattering from the robotic arm in different positions and its effect on the probe illuminations. This is an area of research on the use of absorber that has not been explored until recently [3]. Numerical experiments are conducted to explore the effects of RF absorbers in the 300 MHz to 3 GHz range. Open ended waveguides (OEWG) as well as dual ridged horns (see Figure 1) are used as the probes. The results suggest that some areas of the arm need to be treated while others can be left bare. The analyses performed suggest that optimized treatment of robotic arms to maintain the flexibility of the technique while also reducing effects on the probe illuminations are possible.
This paper proposes a novel approach based on a Multi-Probe Near-Field Scanner system utilizing bipolar and phased array antenna technology to measure antenna arrays with highly reliable precision. The innovative system incorporates a phased array antenna with full amplitude and phase control, enabling complete manipulation of the polarization state. This advanced capability allows the system to accurately measure and characterize antenna arrays while addressing and correcting polarization distortions caused by polar motion. By using the multi-probe configuration, the system significantly enhances measurement efficiency by capturing near-field data simultaneously across multiple probes. The integration of bipolar technology [1] ensures robust signal processing, while the phased array design enables the electronic rotation of the polarization of the probe array antenna. This feature is critical for mitigating errors introduced by polarization mismatches or distortions, ensuring high-fidelity measurements even in complex scenarios. The proposed system demonstrates superiority over traditional single-probe scanners by reducing measurement time, providing comprehensive polarization control and size reduction of the controlled environment. The paper discusses the system’s design, implementation, and performance.
The approach of non-redundant near-field sampling has been available for many years. A general and automated approach that yields the expected time reduction for an arbitrary antenna volume, however, has been elusive. One of the more practical approaches is the “PNF wide-mesh” sampling, where the probe grid is separable in x and y, and this approach is the one explored in this paper. A fundamental step in non-redundant sampling is to identify a volume that fully contains the AUT. Constraints imposed by theory have typically led this volume to be rotationally symmetric about a z-oriented line, and often also require that the volume be more spherical (less oblate) than a volume circumscribing the AUT. That larger volume generally results in more acquisition time than would a conformal volume, but allows those samples to be readily interpolated to the conventional half-wavelength PNF grid. This paper examines the impacts of relaxing those constraints in order to further reduce the required sampling time for a box-shaped AUT. It then looks for ways to reduce or remove those impacts. The implementation of this algorithm involved a minor reformulation, specific to the PNF (or linear-axis) geometry, of the underlying non-redundant sampling theory. That reformulation is briefly described herein. A new family of tunable AUT-volume edge treatments similar to the existing “double-bowl” is also described. The paper will show minor reductions in predicted acquisition time compared to non-redundant sampling with a circular double-bowl volume. Each non-redundant approach typically offers a 40-60% reduction with a rectangular AUT volume compared to a full conventional scan. A more notable advantage of the new approach is a significant reduction in pre-acquisition activity defining the several parameters that govern the non-redundant acquisition and processing.
A new sparse sampling and compressive sensing based reconstruction and near-field imaging technique is introduced for the measurement of electrically large production test and diagnostics of nose-mounted commercial radomes. Simulated results are presented, where it is demonstrated that far- field results with an equivalent multi-path level of better than - 60dB can be obtained from circa 10% of the points required by a classical Nyquist equiangular spherical near-field acquisition scheme for the case of an electrically large, i.e. full size, commercial airliner nose-mounted radome enclosing an x-band weather radar. Furthermore, a new method for the rapid noninvasive nondestructive imaging and identification of defects within these radomes is presented that provides significantly clearer fault detection at a far earlier stage within the radome measurement campaign than has previously been possible.
In this work, a novel two-steps technique to correct 3D probe positioning errors affecting the non-redundant (NR) spherical near-field/far-field transformation (NF/FFT) technique for quasi-planar antennas, modeled with a double bowl, is presented. The developed approach benefits from the synergy of two complementary correction steps. In the first one, the so-called NR correction compensates for the phase errors due to the deviations of the sampling points from the nominal scanning sphere. This is achieved through an ad hoc phase correction scheme, whereby an optimal phase factor is extracted from the acquired samples, accounting for their actual (i.e., erroneous) position. In the second step, an iterative approach is exploited to manage the residual 2-D errors affecting the data attained at the end of the previous step, allowing for an accurate and efficient restoration of the correctly located samples. Ultimately, the large number of input data for the classical spherical NF/FFT, uniformly spaced on the scanning spherical grid, are obtained via an optimal sampling interpolation formula. Numerical results showing the effectiveness of the proposed method in compensating even severe 3-D positioning errors are reported.
The practical implementation of an improved, 3- axis, magnetic field probe is presented. The 3-axis probe facilitates the simultaneous measurement of magnetic field on three orthogonal axes in the time and frequency domains. Each shielded loop employs two diametric feeds and a triaxial infinite balun to suppress the electric dipole mode. The triaxial infinite balun is shown to be more effective at suppressing the electric dipole mode than the magnetic choke balun used in a previous design. Also, it is shown that the magnetic balun is subject to saturation by common-mode (CM) voltage in turn caused by incident electric fields. Finally, it is shown that the triaxial infinite balun can be made to fit entirely within the loop conductor, whereas the external magnetic choke balun employed a conductive housing which caused a significant perturbation to the electromagnetic field. Discussion of the relation of the electric dipole moments to the loop isolation is presented in light of a recently-published design employing single-feed shielded loops. It is shown that the single-feed loops are not completely decoupled even when they are orthogonal and co-located unless the feeds are arranged so that the electric dipole moments are forced to be orthogonal. While this arrangement provides good isolation, the single-feed loops exhibit sensitivity to electric fields.
This paper explores a Phase Retrieval (PR) method for automotive Over-The-Air (OTA) spherical Near-Field (NF) measurements. Two signals are acquired phase-coherently: one from a fixed reference antenna attached to the vehicle, and the other from a conventional NF probe. The phase information is retrieved from the relationship between the two signals. An off-the-shelf Software-Defined Radio (SDR) two-channel receiver module is used. This setup enables one-step NF OTA measurements, potentially improving time- and cost-efficiency. Moreover, it solves the inaccessibility of antenna connectors. This approach has been tested on a vehicle in an actual automotive OTA measurement setup with a modulated signal centered at 810 MHz, which presents a challenging and realistic scenario. This work demonstrates that phase retrieval is feasible with the proposed setup. Moreover, good agreement with a reference measurement taken using a Vector Network Analyzer (VNA) is achieved at high Signal-to-Noise Ratio (SNR) levels. E.g., phase errors with respect to the reference measurement in the NF of less than 10° at 90% of the angular sampling points can be obtained with an SNR of at least 27 dB. The corresponding transformed Far-Field (FF) EIRP shows also good agreement between the PR and VNA measurements, with an Equivalent Noise Level (ENL) of -32 dB. In the evaluated example, the direct OTA measurement in the NF yields a much less similar pattern to the reference FF than with the proposed method.
This work presents an automated antenna measurement approach using an adaptive grid strategy. Designed for planar acquisitions with a robotic-arm system, the method significantly reduces the number of required samples and enables fully automated measurement of the Antenna Under Test (AUT), without user intervention or prior knowledge of the AUT. The system operates iteratively, selecting optimal sampling points at each acquisition. It uses a matrix inversion method for near-field to far-field transformation, allowing far-field computation from irregular measured data. Singular Value Decomposition (SVD) is used to decompose the transformation matrix, representing the far-field as a weighted summation of aperture modes. After each iteration, the acquired field is analyzed to identify the next set of samples that will induce the largest variation in the modal coefficients, thereby maximizing the update in the reconstructed aperture field. The measurement process automatically finishes when the variation of the values is below a fixed threshold. The proposed technique has been evaluated through numerical simulations at 30 GHz using an ideal aperture model, as well as through experimental measurements of a Standard Gain Horn (SGH) at 22 GHz.
We present the Derivative Singular Value Optimization (DSVO), namely, the combined use of derivative probes and SVO to reduce the number of sampling locations in Near-Field (NF) antenna characterization. A derivative probe provides, at the same measurement location, the simultaneous acquisition of quantities related to the values of the field and of the field first derivatives. The additional information conveyed by the derivative probe allows a reduction in the measurement points required by DSVO compared to SVO. Numerical results for planar scanning and aperture antennas show that DSVO indeed enables a significant reduction in the number of sampling locations, without loss of accuracy.
Near-Field (NF) antenna measurement ranges have evolved as an alternative to Far-Field (FF) ranges to be the prominent method for pattern estimation from measurement. This is because NF ranges are more convenient to use and require much less space in laboratories [1]. The Microwave Vision Group (MVG) StarGate series, such as the legacy Satimo StarGate 64 (SG64), employs an array of multiplexed probes to perform measurements, sampling across a synthetic aperture created by moving the test antenna with respect to the probe array. This example of a commercial system does not use a Vector Network Analyzer (VNA), and was not designed to be updated. This means that it cannot benefit from improvements made in RF measurement equipment, requiring instead that a new system be purchased, which is unrealistic for many users. This work describes the initial process of updating a legacy SG64 to use a VNA. It includes characterization of the control signals, the transmit and receive paths, as well as the potential improvements to performance and lifetime that upgrading the legacy system to a VNA-based configuration offers.
This study presents the design, fabrication, and testing of a 2P3T switch module for a 5G high-speed measurement system. The module is developed to switch the polarization signals received from multiple probes to a measurement program for 5G certification testing or to multiple receivers for modulated signal detection. Three signal paths are used to control the module. The module includes a low-noise amplifier (LNA) and uses a grounded coplanar waveguide (GCPW) structure to integrate the integrated circuit (IC) chip. The switch paths are controlled using a LabVIEW program. The performance of the switch module is evaluated as satisfactory.
An automated scan plane alignment technique for robot-based planar near-field antenna measurements enables precise and efficient calibration of unknown and arbitrarily oriented antennas under test (AUTs). By integrating a high- resolution laser line profile sensor with a robotic arm, the system dynamically determines the AUT’s position, orientation, and outline without requiring detailed prior knowledge. A real- time feedback loop guides the robot to adaptively align the scan plane based on measured surface profiles, taking into account tilts or non-ideal AUT placements. Edge detection and reference mark identification further enhance accuracy, allowing to precisely align the scan center with the AUT’s geometric center. The method is validated using a reference metal plate and is particularly suited for spatially flat antennas or radomes. Beyond alignment, the same setup enables high-resolution optical inspection, capable of detecting fine surface details such as cracks, dents, or even the thickness of ink from printing. The approach significantly reduces setup time by eliminating manual alignment steps, and broadens the functionality of robot-based measurement systems by combining self-alignment and optical inspection into a single automated process.
This paper presents an extended uncertainty analysis of a multiprobe antenna measurement system developed for large platform testing across the 64 MHz to 6 GHz frequency range. Installed at the Pulsaart by AGC facility in Belgium, the system enables fast and accurate characterization of complex structures integrating multiple antennas. Building on previous studies, the analysis expands the uncertainty budget by including a broader set of antennas, such as monocones operating down to 50 MHz, and evaluating key figures of merit including radiation pattern, gain, efficiency, and cross-polarization. Particular emphasis is placed on reflectivity-related uncertainty, which is a dominant factor at lower frequencies due to chamber electrical size and absorber limitations. The methodology incorporates modal filtering and spatial displacement of antennas to isolate the environmental effects. The results offer detailed insights into antenna-dependent uncertainties and, for the first time, provide complete uncertainty estimations for the aforementioned metrics across the full operating frequency range.
A commonly employed technique in the field of phaseless antenna measurements is the two-sphere method, wherein complex spherical mode coefficients are reconstructed from amplitude-only near-field measurements taken on two spheres that are separated by a specific distance. Recent numerical studies have indicated that the second sphere can also be substituted with a polyhedral surface. However, traditional positioning systems are incapable of measuring such configurations. This work demonstrates that a robotic antenna measurement system can efficiently measure the near-field on the surface of an octahedron. Two approaches for scanning a spiral grid with a robot are presented. It is shown that the additional degrees of freedom offered by the robot can reduce measurement time by up to 15 % compared to conventional spiral measurements. Furthermore, the far-field pattern derived from the complex near-field data on an octahedron produces highly accurate results, achieving an average equivalent error signal of −56.6 dB when compared to standard spherical measurements. A comparison between phaseless reconstruction using two spheres and a combination of a sphere and an octahedron reveals that both methods yield comparable accuracy.
This paper examines the uncertainty contributions associated with the spherical Near-Field to Far-Field (NF/FF) transformation process when applied to electrically large antennas. The transformation is based on the Spherical Wave Expansion (SWE) and implemented through the Transmission Formula (TXF), which provides a mathematically rigorous and computationally efficient framework. The TXF supports multiple levels of Probe Correction (PC), each with varying complexity and accuracy. However, applying the TXF to electrically large antennas (e.g. larger than 500 wavelengths) present significant computational challenges. The large number of spherical harmonics required increases the processing burden, and the accurate evaluation of the rotation and translation operators becomes critical. These operators must be computed using suitable recurrence relations to avoid instabilities. Additionally, the use of probes with arbitrary patterns can further complicate the probe correction process, potentially introducing numerical instabilities that must be carefully controlled. This work investigates the accuracy of the NF/FF transformation for electrically large antennas by considering both idealized cases without PC, and more realistic scenarios with full PC. The ability to compensate for large tapering effect introduced by the probe will be addressed for the first time.
This paper explores the use of the commercially available software Altair Feko® to model the on-axis directivity and gain of a commercial off-the-shelf (COTS) X-band (8 GHz – 12 GHz) standard gain horn (SGH) with and without waveguide adapter, specifically the SGH820 by the Microwave Vision Group (MVG). SGH820 antenna was simulated using full wave (Method of Moments (MoM) and Multilevel Fast Multipole Method (MLFMM)) solvers that are available in Feko. Reflection coefficient, boresight gain, and directivity parameters are computed with and without waveguide adapter. Both MoM and MLFMM solvers in Feko show good agreement and are shown to be within ±0.008 dB of each other for boresight gain and directivity calculations. Computational resource comparisons for simulations of the SGH820 antenna with an adapter using the MoM and MLFMM solvers are presented. The results show a significant reduction in computational time when using the MLFMM solver compared to the MoM solver without any degradation in accuracy.
NASA has commissioned a 10 m × 10 m Horizontal Planar Near-Field (HPNF) measurement facility to ad- vance the precision characterization of spaceborne antennas for aerospace and defense applications. Developed by MVG-OATI and NASA Glenn Research Center, the system accommodates diverse antenna types, including active electronically scanned arrays (AESAs) and parabolic reflectors, across a 1–110 GHz frequency range. It employs planar near-field techniques to measure gain, directivity, 3D radiation patterns, sidelobe levels, and advanced metrics like effective isotropic radiated power (EIRP), gain-to-noise-temperature (G/T), intermodulation distortion (IMD), and gain compression. The horizontal orientation simplifies installation and mitigates gravity-induced distortions, critical for zero-gravity space environments. A laser tracker ensures precise alignment and surface measurements, enhancing millimeter-wave accuracy. This paper provides an overview of the facility’s design, measurement methodology, and preliminary performance, highlighting its role in next-generation antenna testing.
The expanded availability of low-cost vector networks analyzers has changed the economics of microwave sensing techniques, allowing for use in non-traditional areas such as food science. For example, poultry producers struggle to identify myopathies in chicken breast that compromise the meat’s texture. The current industry standard test is to manually feel each chicken breast by hand to test texture. This method has the disadvantages of being both labor intensive and subjective, and an automated system to scan for defects in chicken breast is desired. This paper details our advances in adapting a commercial dielectric sensor, the epsilon measurement probe, into a tool for in- line detection of defects in chicken breast. The sensor operates in a similar manner as an open-ended coaxial line but uses a computational electromagnetic inversion method for correlating its signal to dielectric properties of the test medium. Using this method, we demonstrate that the measured properties are sufficient to classify the grade of the chicken breast as measured by expert graders. The ability to detect defects requires a combination of accurate reflection phase/amplitude, low enough frequency to penetrate into the chicken meat, and a wide enough bandwidth to resolve relevant dispersion features in the chicken permittivity. In addition, we demonstrate how the probe can overcome challenges related to in-line processing, such as alignment of specimens or imperfect contact with the sensor.