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.)
A desired feature of modern field probes is that the useable bandwidth should exceed that of the Antenna Under Test (AUT) [1]. Recent developments in probe and orthomode junctions (OMJ) technology has shown that bandwidths of up to 4:1 are achievable [2-5]. The probes are based on inverted ridge technology capable of maintaining the same high performance standards of traditional probes However, in typical Spherical Near Field (SNF) measurement scenarios, the applicable frequency range of the single probe can also be limited by the content of µ.1 spherical modes in the probe pattern [6-7]. This is because the traditional NFtoFF software applies probe correction under the assumption that the probe pattern is fully specified from knowledge of the E-and H-plane patterns only [8]. While this condition is guaranteed for virtually any type of probe for small illumination angles of the AUT and/or a long probe/AUT distance this assumption may lead to unacceptable errors in special cases. This paper describes the design and experimental verification of a Ka-band probe based on the inverted ridge technology. The probe is intended for high precision SNF measurements in special conditions that require less than -45dB higher order spherical mode content. This performance level has been accomplished through careful design of the probe and meticulous selection of the components used in the external balanced feeding scheme. The paper reports on the electrical and mechanical design considerations and the experimental verification of the modal content.
There are a number of near-field measurement situations where it is desirable to use a broad band probe to avoid the need to change the probe a number of times during a measurement. But most of the broad band probes do not have low cross polarization patterns over their full operating frequency range and this can cause large uncertainties in the AUT results. Calibration of the probe and the use of probe pattern data to perform probe correction can in principle reduce the uncertainties. This paper reports on a series of measurements that have been performed to demonstrate and quantify the cross polarization levels and associated uncertainties that can be measured with typical log periodic (LP) probes. Two different log periodic antennas were calibrated on a spherical near-field range using open ended waveguides (OEWG) as probes. Since the OEWG has an on-axis cross polarization that is typically at least 50 dB below the main component, and efforts were made to reduce measurement errors, the LP calibration should be very accurate. After the calibration, a series of standard gain horns (SGH) that covered the operating band of the LP probe were then installed on the spherical near-field range in the AUT position and measurements were made using both the LP probes and the OEWG in the probe position. The cross polarization results from measurements using the OEWG probes where then used as the standard to evaluate the results using the LP probes. Principal plane patterns, axial ratio and tilt angles across the full frequency range were compared to establish estimates of uncertainties. Examples of these results over frequency ranges from 300 MHz to 12 GHz will be presented.
The numerical analysis used for efficient processing of spherical near-field data requires that the far-field pattern of the probe can be expressed using only azimuthal modes with indices of µ = ±1. (1) If the probe satisfies this symmetry requirement, near-field data is only required for the two angles of probe rotation about its axis of . = 0 and 90 degrees and numerical integration in . is not required. This reduces both measurement and computation time and so it is desirable to use probes that will satisfy the µ = ±1 criteria. Circularly symmetric probes can be constructed that reduce the higher order modes to very low levels and for probes like open ended rectangular waveguides (OEWG) the effect of the higher order modes can be reduced by using a measurement radius that reduces the subtended angle of the AUT. Some analysis and simulation have been done to estimate the effect of using a probe with the higher order modes (2) – (6) and the following study is another effort to develop guidelines for the properties of the probe and the measurement radius that will reduce the effect of higher order modes to minimal levels. This study is based on the observation that since the higher order probe azimuthal modes are directly related to the probe properties for rotation about its axis, the near-field data that should be most sensitive to these modes is a near-field polarization measurement. This measurement is taken with the probe at a fixed (x,y,z) or (.,f,r) position and the probe is rotated about its axis by the angle .. The amplitude and phase received by the probe is measured as a function of the . rotation angle. A direct measurement using different probes would be desirable, but since the effect of the higher order modes is very small, other measurement errors would likely obscure the desired information. This study uses the plane-wave transmission equation (7) to calculate the received signal for an AUT/probe combination where the probe is at any specified position and orientation in the near-field. The plane wave spectrum for both the AUT and the probe are derived from measured planar or spherical near-field data. The plane wave spectrum for the AUT is the same for all calculations and the receiving spectrum for the probe at each . orientation is determined from the far-field pattern of the probe after it has been rotated by the angle .. The far-field pattern of the probe as derived from spherical near-field measurements can be filtered to include or exclude the higher order spherical modes, and the near-field polarization data can therefore be calculated to show the sensitivity to these higher order modes. This approach focuses on the effect of the higher order spherical modes and completely excludes the effect of measurement errors. The results of these calculations for different AUT/probe/measurement radius combinations will be shown.
Measurement of E and H plane far field patterns for an open-ended rectangular waveguide in the free air operating between the frequencies of 16 and 19 GHz are shown and compared with the simulated patterns derived by several authors. Although the theoretical expressions give a broader pattern for the E-plane than for the H-plane, which is observed, measurements exhibit a sharper decay in the E-plane than the one obtained by simulation. In this work, we calculate the errors associated with the use of the different models that fail to correctly approximate the E-plane. Finally, we introduce a parameter in the best model to adjust the effective aperture dimensions in order to obtain a more realistic representation of the measured far field.
A Plane Wave Synthesis Approach for mitigating errors in antenna measurements caused by stray signals and imperfections in the measurement range illuminating fields has been demonstrated previously for a 2D scalar source [1]. This paper presents algorithms developed for the Range Transfer Function (RTF) method for a 2D vector source. Vector basis functions for both the field representation and the AUT representation are implemented to provide a robust numerical solution. The new algorithms are more stable because the plane wave angles and the antenna measurement angles may be completely general, provided that Nyquist rules of sampling are observed during both the field probing (to obtain the plane wave coefficients) and the antenna measurement (to obtain the raw pattern data).
Far-field uncertainty due to probe errors in planar near-field measurements is analyzed for the fast irregular antenna field transformation algorithm. Results are compared with the classical technique employing two dimensional Fast Fourier Transform (2D FFT). Errors involving probe's relative pattern, alignment, transverse and longitudinal position, interaction with AUT etc. have been considered for planar measurements. The multiple reflections error originating from the interaction of the probe and the AUT tends to deteriorate the radiation pattern to a greater extent. Therefore, a novel technique which utilizes near-field measurements on two partial planes is presented to reduce the multiple reflections between the probe and the AUT.
This paper describes the behavior of the antenna radiation pattern for different planar near-field measurement errors superimposed on the near-field data. The disturbed radiating near field is processed using multilevel plane wave based near-field far-field transformation to determine the far-field. Errors like scan area truncation, transverse and longitudinal position inaccuracy of measurement points, and irregular sample spacing are analyzed for an electrically large parabolic reflector at 40 GHz. The error behavior is then compared with the standard transformation technique employing 2D Fast Fourier Transform (FFT) using the same near-field data. In order to exclude the effect of any other measurement or environmental error, electric dipoles with appropriate magnitude profile and geometrical arrangement are used to model the test antenna.
Large phased arrays have stringent beam-pointing accuracy requirements over their scan volume. Measuring the beam-pointing accuracy of a phased array with a planar near-field scanner is convenient but can lead to erroneous results if the near-field sampling grid is not well controlled. This paper describes numerical experiments that were carried out to assess the impact of various types of grid errors on the measurement of beam-pointing accuracy. The types of grid errors considered include skewing and curvature in the plane of the grid. The numerical experiments use infinitesimal dipoles as the radiating elements and assume an ideal probe. It is shown that beam-pointing errors induced by grid errors that can be described by an affine transformation can be estimated in closed form. For more complicated grid errors, the model is shown to be a useful tool in estimating their impact on measuring beam-pointing error. Finally, the amount of over-scan required for accurate beam-pointing measurements over a large scan volume is examined.
A novel probe design for measuring complex permittivity of soils in-situ from 10 to 1000 MHz without taking soil samples is presented. The dielectric constant and conductivity of soil is derived from step-frequency reflection taken inside a small freshly bored hole. As a result, permittivity at various depths with in-situ moisture content and soil texture can be obtained in the fields. A novel calibration method was developed to account for the frequency- and material-dependent geometrical factor which causes bias errors in conventional calibration methods. Experimental measurement results and simulation results are used to prove the efficiency and accuracy of this method.
Far field measurements of ground vehicle antennas in anechoic chambers often require the creation of a plane wave by near field hemispherical probing with associated mathematical transformations to the far field/plane wave result. Direct far field measurements can be done to save time when the frequency is low enough. This paper discusses a method of extending the frequency band where direct measurements can be done by synthesizing a plane wave using a small array of antennas. The use of an array to create a plane wave in an anechoic chamber usually results in errors due to the reflections from the walls of the chamber. The technique to be described in this paper is to model the wall reflections and the array antenna characteristics and to use optimization techniques to derive an antenna placement and power distribution scheme to optimize the plane wave. Several optimization techniques will be described and results from testing in a 1.2 meter long sub-scale chamber model will be shown. Improvements in the far field measurements will be discussed.
Highly accurate spherical near-field measurement systems require precise alignment of the probe antenna to the measurement surface. MI Technologies has designed and constructed a new spherical near field arch positioner with a 1.5 meter radius to support measurements requiring accurate knowledge of the probe phase center to within .0064 cm throughout its range of travel. To achieve this level of accuracy, several key design elements were considered. First, a highly robust mechanical design was considered and implemented. Second, a tracking laser interferometer system was included in the system for characterization of residual errors in the position of the probe. Third, a position control system was implemented that would automatically correct for the residual errors. The scanner includes a two position automated probe changer for automated measurements of multi-band antennas and a high accuracy azimuth axis. The azimuth axis includes an algorithm for correcting residual, repeatable positioning errors. This paper defines the spherical near-field system and relation of each axis to the global coordinate system, discusses their associated error sources and the effect on global positioning and presents achieved highly accurate results.
New developments in ortho-mode junctions (OMJ) and probe technologies has enabled near field probes with up to 1:4 bandwidth, while maintaining the high performance standards of traditional narrow band probes [1–3]. The new probe technologies are based on inverted ridge structures providing four symmetrical feeding points for external balanced feeding. The inverted ridge structure stabilizes the frequency dependence of the OMJ while the external balanced feeding is a crucial feature to achieve the desired high performances. This paper briefly review the theory of balanced feeding and derive performance guide lines on the external beam forming network for achieving high port-to-port isolation and matching on a wide bandwidth with the inverted ridge probe technology. The relationship between excitation errors in the balanced feeding scheme and the spherical mode index µ.1 content of the probe is also investigated and upper bounds on acceptable excitation errors are derived.
G/T is a standard figure of merit describing the sensitivity of receive, active, microwave antenna systems. It is therefore critical to have an accurate test technique for measuring the G/T of systems with varying levels of sensitivity. The Yfactor method is a commonly used and wellunderstood technique for characterizing receive microwave systems. While the concepts behind the Yfactor measurement techniques are generally understood, little quantitative work has been done establishing limits and rules regarding selection of measurement hardware and design of the measurement system. This paper seeks to provide examples of test systems and quantify the error induced into the measurement due to mismatched standard hardware. The variables and techniques involved in a G/T test are examined using simulated results of measurement systems performed in VSS (Visual System Simulator).
In this paper, a beam-steering computer design is explored for a large space-fed phased-array antenna. GTRI previously developed a beam-steering computer for a smaller phased-array antenna which accomplished spherical propagation focusing and multiple phase-only beam-broadening modes. In a subsequent effort, the beam-steering computer design was scaled for a large phased-array antenna to accomplish similar tasks. To verify the design, a series of far-field measurements was initiated to characterize the performance of the antenna by comparing with past measured near-field data and modeled results. One of the primary responsibilities of the beam-steering computer was the focusing of the spherical propagation wave front. A measurement technique is discussed to accomplish this focusing for the large space-fed phased-array antenna by correcting measurement errors in the spherical propagation routine of the beam-steering computer. Additional patterns were taken using the updated feed horn focal point for spherical propagation correction. By correcting the phase errors caused by spherical propagation defocusing in the original beam-steering computer, significantly better antenna performance was obtained, including higher peak gain, reduced nearby sidelobe levels, and removal of beam-pointing errors. Another important responsibility of the beam-steering computer was the ability to realize multiple antenna modes, including a focused pencil beam as well as defocused broadened-beam modes. A stochastic gradient descent algorithm was utilized to obtain several phase tapers to accomplish beam-broadening for the antenna modes. These modes were implemented in the beam-steering computer and tested on a far-field range. The antenna patterns were compared with modeled results and with previous measured data to ensure validity of the implementation.
Calibration of probes for planer near field range measurements is generally required to obtain accurate cross-polarization (xpol) data; however, probe calibration is costly and time consuming. Using analytical models in place of calibration is generally much more cost effective, but may result in larger measurement errors. In a previous paper [1], we showed that simple models of copol probe patterns with zero xpol can give accurate measured results, provided that the probe xpol is much better, generally 10-15 dB better, than the Antenna Under Test (AUT). The next question is “Can a lower performing (and cheaper) probe be used if both the copol and xpol probe patterns are modeled?” In this paper, we compute AUT xpol measurement errors that result from probe xpol errors, and we compare far field AUT patterns processed using probe models with patterns processed with calibrated probe files.
This paper elaborates on certain aspects of a new measurement process that permits an assessment of spherical near-field (SNF) measurement errors based on a set of practical tests that can be done as part of any SNF measurement. It provides error bars for a measured radiation pattern in an automated fashion.
– far-field transformation with cylindrical scanning are efficiently determined by using an optimal sampling interpolation algorithm. The comparison of the far-field patterns reconstructed from the acquired irregularly distributed measurements with those obtained from the data directly measured on the classical cylindrical grid assesses the effectiveness of the approach.
A stand-alone commercial program, performing advanced electromagnetic processing of measured data, is being developed by TICRA. The program reads the measured field and computes the extreme near field or the currents on the antenna surface. From the inspection of the extreme near field or currents, the program will solve typical antenna diagnostics problems, such as identification of array element failure and antenna surface errors, but also allow artificial removal of undesired contributions, such as currents on cables and fixtures, thereby saving valuable time and resources in the antenna design and validation process. The program will be based on two field reconstruction techniques, the SWE-PWE presented at AMTA in 2007, and a new and more accurate inverse higher-order Method of Moments (INV-MoM). The paper will illustrate the theory behind the two techniques and present numerical cases with simulated data.
The calibration and measurement of bistatic reflectivity at short range (3.3 m) presents challenges that are significantly different from the usual test range measurements (typically monostatic at 100 to 150 m). In order to overcome this an object-free calibration procedure has been applied, eliminating crosstalk, reducing other interferences and removing errors associated with the RCS and alignment of calibration objects. It is based on calibrating the transmitter and receiver antennas as a pair by directing the antennas toward each other. The method thus requires that the antennas can be separated. Furthermore the signal level needs to be handled e.g. by the separation distance or attenuators. The bistatic reflectivity measurements were performed by placing a clutter sample on a turntable which is located at the centre of a bistatic arc. This configuration enables us to do ISAR images. Background contributions were discriminated using a combination of synthetic resolution and zero-doppler filtration. The sensitivity variation across the antenna footprint was handled by calculating an equivalent area from measured off-axis antenna sensitivities. Reflectivities have been measured for a metallic test surface and for grass. The metallic test surface had been manufactured to correspond to typical theoretical bistatic clutter models.
A widely used time-gating technique can be effectively implemented in near-field (NF) antenna measurements to significantly improve the measurement accuracy. In particular, it can be implemented to reduce or remove the effects of the following measurement errors [1]: -multiple environmental reflections and leakage in outdoor or indoor NF ranges -edge diffraction effects on measurement accuracy of low gain antennas on a ground plane [3] In addition, reflectivity in the range can be precisely localized, separated and quantified by using the time – gating procedure with only one addition (a subtraction operation) added to the standard near-field to far-field (NF – FF) transformation algorithms. In this paper a step by step procedure is described which includes acquisition of near-field data, transformation of the raw near-field data from the frequency to the time domain, definition of the correct time gate, transformation of the gated time domain data back to the frequency domain, and the transformation of the time gated near-field data to the far-field. The time gated results, as already shown in [2], provides for more accurate far-field patterns. In this paper it is shown how the 3D reflectivity/multiple reflections in the measurement chamber or outdoor range can be determined by subtracting the time gated results from the un-gated data. This technique is illustrated through use of several measurement examples. It is demonstrated that the time gated method has a clear physical explanation, and, in contrast with other techniques [4,5] is less consuming (does not require mechanical AUT precise offset installation, additional measurement and processing time) and allows for a better localization and quantization of the sources of unwanted radiation. Therefore, this technique is a straightforward one and is much easier to implement. The main disadvantage cited by critics regarding use of the time gating technique is the narrow frequency bandwidth used in many NF measurements. However, it is shown, and illustrated by the examples, that the technique can be effectively implemented in NF systems with a standard probe bandwidth of 1.5:1 and an AUT having a bandwidth as low as 5% to 10%.