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.)
This paper presents the results of a recent study concerning the computational electromagnetic simulation of a spherical near-field (SNF) antenna test system. The new plane-wave scattering matrix approach [1, 2] allows many of the commonly encountered components within the range uncertainty budget, including range reflections, to be included within the model [3]. This paper presents the results of simulations that verify the utility of the spherical mathematical absorber reflection suppression (S-MARS) technique [3, 4] for the identification and subsequent extraction of artifacts resulting from range reflections. Although past verifications have been obtained using experimental techniques this paper, for the first time, corroborates these findings using purely computational methods. The use of MARS is particularly relevant in applications that inherently include scatterers within the test environment. Such cases include instances where a SNF test system is installed within an existing compact antenna test range (CATR) as is the configuration at the recently upgraded Queen Mary University of London (QMUL) Antenna Laboratory [5, 6]. Thus, this study focuses on this installation with results of CEM simulations and actual range measurements being presented. The method enables a quantitative measure of the levels of suppression offered by the MARS system. References A.C. Newell, S.F. Gregson, “Estimating the Effect of Higher Order Modes in Spherical Near-Field Probe Correction”, Antenna Measurement Techniques Association (AMTA) 34th Annual Meeting & Symposium, Bellevue, Washington October, 2012. A.C. Newell, S.F. Gregson, “Computational Electromagnetic Modelling Of Spherical Near-Field Antenna Test Systems Using Plane Wave Spectrum Scatting Matrix Approach”, Antenna Measurement Techniques Association (AMTA) 36th Annual Meeting & Symposium, Tucson, Arizona, October, 2014. C.G. Parini, S.F. Gregson, J. McCormick, D. Janse van Rensburg “Theory and Practice of Modern Antenna Range Measurements”, IET Press, 2014, ISBN 978-1-84919-560-7. G.E. Hindman, A.C. Newell, “Reflection Suppression in a large spherical near-field range”, Antenna Measurement Techniques Association (AMTA) 27th Annual Meeting & Symposium, Newport, RI, October. 2005. A.D. Olver, C.G. Parini, “Millimetre-wave Compact Antenna Test Range”, JINA Nice, November 1992. C.G. Parini, R. Dubrovka, S.F. Gregson, "CATR Quiet Zone Modelling and the Prediction of 'Measured' Radiation Pattern Errors: Comparison using a Variety of Electromagnetic Simulation Methods" Antenna Measurement Techniques Association (AMTA) 37th Annual Meeting & Symposium, Long Beach California, October 2015.
The accurate alignment of antennas and field probes is a critical aspect of modern antenna metrology systems, particularly in the millimeter-wave region of the spectrum.Commercial off-the-shelf robotic arms provide a sufficient level of positional accuracy for many industrial applications.The Antenna Metrology Project in the Communications Technology Laboratory at the National Institute of Standards and Technology has shown that path-corrected commercial robotic arms, both in hardware and software analysis, can be used to achieve sufficient positioning and alignment accuracies (positioning error ~ /50) for antenna characterization measurements such as gain extrapolation and near-field pattern out to 183 GHz [1]. Position correction is achieved using a laser tracker with a 6 degree of freedom sensor attached to the robot end effector.The end effector’s actual position, measured using the laser tracker, is compared to its commanded position and a path correction is iteratively applied to the robot until the desired level of accuracy is achieved in the frequency range of interest.At lower frequency ranges (< 40 GHz), sufficient positional accuracy can be achieved, without path correction, using a using a calibrated kinematic model of the robot alone [2].This kinematic model is based on knowledge of the link frame transformations between adjacent links and captures deviations due to gravitational loading on the joints and small mechanical offsets between the joints.Additionally, the calibration procedure locates the robot’s base frame in the coordinate system of the robot’s end effector.Each link frame is described by four physical quantities, known as Denavit-Hartenberg (DH) parameters [3]. We performed calibration measurements of our CROMMA system’s DH parameters over a working volume of ~1 m3.We then use the laser tracker to compare the robot’s positional accuracy over this working volume with and without the calibrated kinematic model applied.The path errors for the calibrated case set an upper frequency limit for uncorrected antenna characterization measurements. [1]D. R. Novotny, J.A. Gordon, J.R. Guerrieri, “Antenna Alignment and Positional Validation of a mm Wave Antenna System Using 6D Coordinate Metrology, ” Proceedings of the Antenna Measurements Techniques Association, pp 247-252, 2014 [2]R.Swanson, G. Balandran, S. Sandwith, “50-micron Hole Position Drilling Using Laser Tracker Controlled Robots, ” Journal of the CMSC, Vol 9, No 1, Spring 2014 [3].J.J. Craig, “Introduction to Robotics: Mechanics and Control, 3rd ed.,” New Jersey, Prentice Hall, 2004, pp. 62-69
At NIST, we have developed a precision, wide-band, mmWave modulated-signal source with traceability to primary standards. We are now extending the traceability path for this modulated-signal source into free space to be used for verifying over-the-air measurements in 5G, wireless receivers. However, to obtain a traceable modulated signal in free space, the full scattering matrix of the radiating antenna must be measured. We have extended the extrapolation methods used at NIST, based on the work of Newell, et al. [1]. The extrapolation measurement provides a very accurate, far-field, on-axis, scattering matrix between two antennas. When combined with scattering-matrix measurements made with permutations of pairs of three antennas, far-field scattering, and, thus, gain, is obtained for each antenna. This allows an accurate extrapolation of the antenna’s near-field pattern. We have incorporated the extrapolation fitting algorithms into a Monte Carlo uncertainty engine called the NIST Microwave Uncertainty Framework (MUF) [2]. The MUF provides a framework to cascade scattering matrices from various elements, while propagating uncertainties and maintaining any associated correlations. By incorporating the extrapolation measurements, and the three-antenna method into the MUF, we may provide traceability of all measurement associated with the gain, including the scattering parameters. In this process, we studied several aspects of the gain determination. In this work, we show simulations determining the efficacy of filtering to reduce the effect of multiple reflection on the extrapolation fits. We also show comparisons of using only amplitude (as is traditionally done) to using the full complex data to determine gain. Finally, we compare uncertainties associated with choices in the number of expansion terms, systematic alignment errors, uncertainties in vector network analyzer calibrations and measurements, and phase error introduced by cable movement. With these error mechanisms and their respective correlations, we illustrate the NIST MUF analysis of the antenna scattering-matrix with data at 118 GHz. [1] A. C. Newell, R. C. Baird, and P. Wacker “Accurate Measurement of Antenna Gain and Polarization at reduced distances by an extrapolation technique” IEEE Transactions on Antennas and Propagation. Vol. 21, No 4, July 1973 pp. 418-431. [2] D. F. Williams, NIST Microwave Uncertainty Framework, Beta Version. NIST, Boulder, CO, USA, Jun. 2014. [Online]. Available: http://www.nist.gov/pml/electromagnetics/related-software.cfm
Microstrip patch antennas are well known in the field of communications and other areas where antennas are used. They consist of a metallic conducting surface deposited onto a grounded dielectric substrate and are widely used in situations where a conformal antenna is desired. They are also popular antennas for array applications. But most patch antennas are typically resonant structures owing to the standing wave of current that forms on them. This resonant behavior limits the impedance bandwidth of the antenna to a few percent. In this paper we shall present an approach for improving the bandwidth of a resonant patch antenna which employs an engineered anisotropic superstrate. By proper design of this superstrate and its tensor, and proper alignment of it with the axis of the patch, an antenna with improved impedance bandwidth results. Some of the challenges associated with the measurement of the anisotropic superstrate will be discussed, ranging from 3D simulations to physical models tested in the laboratory. A final working model of the antenna will be discussed; this model consists of a stacked patch arrangement and was designed to operate at the GPS L1 and L2 frequencies. Data collected from 3D simulations using CST Microwave Studio along with laboratory and anechoic chamber measurements will be presented, showing how the bandwidth at both of these frequencies can be increased while maintaining circular polarization in both passbands. Tolerance to errors in alignment and fabrication will also be presented. Additionally, some lessons learned on anechoic chamber measurements of the antenna’s gain and axial ratio will be discussed.
Measurement of the radiation properties of low gain antenna operating at VHF frequencies is well known to be a challenging task. Such antennas are sometimes tested in outdoor Far Field (FF) ranges which are unfortunately subject to errors caused by the electromagnetic pollution and scattering from the environment. Near Field (NF) measurements performed in shielded anechoic chambers are thus preferable to outdoor ranges. However, also in such cases, the accuracy of the results may be compromised by the poor reflectivity of the absorbing material which might be not large enough wrt the VHF wavelength. Other source of errors may be caused by the truncation of the scanning area which generates ripple on the FF pattern after NF/FF transformation. Spherical multi-probe systems developed by MVG are optimal measurement solution for low directive Device Under Test (DUT). Such systems allow to perform a quasi-full spherical acquisition combining a rotation of the DUT along azimuth, with a fast electronically scanned multi-probe vertical arch. The DUT can be accommodated on masts made of polyester material which allows to minimize the interaction with the DUT. Measurements of low directive device above 400 MHz performed with such type of systems have been demonstrated to be accurate and extremely fast in previous publications. In this paper, measurements of a low directivity antenna, performed at VHF frequencies in a MVG spherical multi-probe system, will be presented. The antenna in this study is an array element, part of a larger array, which has been developed for space-born AIS applications. Gain and pattern accuracy of the measurement will be demonstrated by comparison with full wave simulation of the tested antenna.
This paper presents the results of a new measurement technique to determine antenna system noise temperature using data acquired from a planar near-field measurement. The ratio of antenna gain to system noise temperature (G/T) is usually determined in a single measurement when the antenna is alternately pointed towards the “cold sky” and a hot radio source such as the sun or a star with a known flux density. The antenna gain is routinely determined from near-field measurements and with the development of this new technique, the system noise temperature can also be determined. The ratio of G/T can therefore be determined from planar near-field data without moving the antenna to an outdoor range. The noise temperature is obtained by using the plane-wave spectrum of the planar near-field data and focusing on the portion of the spectrum in the evanescent or “imaginary space” portion of the spectrum. Near-field data is obtained using a data point spacing of l/4 or smaller and the plane-wave spectrum is calculated without applying any probe correction or Cos(q) factor. The spectrum is calculated over real space corresponding to propagating modes of the far-field pattern and also the evanescent or imaginary space region where . Actual evanescent modes are highly attenuated in the latter region and therefore the spectrum in this region must be produced by “errors” in the measured data. Some error sources such as multiple reflections will produce distinct localized lobes in the evanescent region and these are recognized and correctly identified by using a data point spacing of less than l/2 to avoid aliasing errors in the far-field pattern. It has been observed that the plane wave spectrum beyond these localized lobes becomes random with a uniform average power. This region of the spectrum must be produced by random noise in the near-field data that is produced by all sources of thermal noise in the electronics and radiated noise sources received by the antenna. By analysing and calibrating this portion of the spectrum in the evanescent region the near-field noise power can be deduced and the corresponding noise temperature determined. Simulated and measured data will be presented to illustrate and validate the measurement and analysis techniques. Keywords — Planar Near-Field, G/T, Figure-of-Merit Measurements, Simulation, Plane Wave Spectrum.
In high-gain array applications such as satellite communications and radars, low sidelobe level is a key performance requirement. Applying a magnitude tapering profile across the aperture is a common way to suppress the sidelobes. Other realizes sidelobe level control method via phase variation across the aperture. Common ways for implementing aperture magnitude taper include applying lumped resistors or introducing proper series feeding, etc. This paper discusses about 33-35 GHz fixed-beam, low-sidelobe array antenna design. In order to minimize the number of the feed network, the 11x32 Ka-band array is composed of 32 sequentially-fed 11-element subarrays. Also, to maximize the antenna efficiency, the Chebyshev tapering selected for achieving -35 dB sidelobe levels was accomplished using a combination of un-even power divider, mismatched feed lines, and different feedline lengths. This approach allows it to achieve magnitude taper control from -0.5 dB to -26 dB from 33 to 35 GHz without using resistors. The unequal power divider arrangement reduces mismatch loss my 1.3dB compared to conventional 2-way even dividers in the corporate feeding network. The paper also studies the impact of the amplitude and phase uncertainty among array elements on the degradation of the sidelobe performance. An 11x32 planar patch array prototype is designed, fabricated and measured. The array obtains a realized gain of 19 dBi with Azimuth 3dB beamwidth of 1.0 degree and -23 dB sidelobe.
Rotating the coordinate system of an antenna pattern can be problematic due to the need to interpolate complex data in spherical coordinates. Common approaches to 2D interpolation often introduce errors because of polarization discontinuities at the spherical coordinate system poles. To overcome these difficulties, it is possible to transform an antenna pattern from field-space into spherical mode-space, perform the desired coordinate system rotation in mode-space, and then transform the modes in the rotated coordinate system back into field-space. This method, while more computationally intensive, is exact and alleviates all of the interpolation-related issues associated with rotations in field-space. Although rotations in mode-space have been implemented in commercially available software (e.g., the ROSCOE algorithm provided by TICRA), these algorithms may not be well understood by the general antenna measurement community. Therefore, the first goal of this paper is to present an easy-to-understand algorithm for performing rotations in mode-space. Next, the paper will address the challenge of computing the rotation coefficients, which are required by the mode-space coordinate system rotation algorithm. Although J. E. Hansen presented a method for recursively computing the rotation coefficients, this method is numerically unstable for large values of N (where N is the upper limit of the polar index). Therefore, this paper will present a numerically stable method for the recursive computation of the rotation coefficients. Finally, this paper will show the relationships between Euler angles and both Az-over-El angles and El-over-Az angles. These relationships are quite useful because it is often desired to rotate an antenna pattern based on Elevation and Azimuth angles, whereas the inputs for the mode-space rotation algorithm are Euler angles. Knowing these relationships, the Euler angles may be computed from the Azimuth and Elevation angles, which can then be used as the inputs to the mode-space rotation algorithm.
GTRI has been developing a method for insertion phase calibration, as discussed in the paper “Insertion Phase Calibration of Space-Fed Arrays,” which was presented at AMTA in 2015 [1]. This method has been implemented to characterize the phase response of phase shifters in a system currently under fabrication at GTRI. One of the primary requirements for the phased-array antenna of this system is a maximum RMS phase error. The RMS phase error for this array is influenced by a variety of error sources, including phase shifter quantization, beam steering computer (BSC) algorithmic error, phase shifter unpredictability error, test fixture induced error, phase shifter thermal drift, and phase shifter frequency dependency. Each of these error sources has been categorized as either a non-deterministic error, whose behavior can be statistically characterized but not calibrated out, or as a deterministic error, whose behavior can be characterized and potentially calibrated out. The non-deterministic errors include element unpredictability, which is induced by the inability of an individual phase shifter to precisely repeat a given phase command, and errors induced by the calibration test fixture itself. The deterministic errors include phase shifter quantization error, which is a function of the phase state bit precision, BSC algorithmic error, which is driven by the numerical preciseness of calculation of the commanded phase states for each element, thermal driven phase drift, and phase shifter frequency dependency across the band of operation. To calibrate the insertion phase and phase-state response curves for all phase shifters used in the system, a custom-built calibration fixture was constructed into a septum wall that separates two semi-anechoic chambers. The realized phase-error budget of the system under fabrication was affected directly by the accuracy of both the calibration method and this fixture. We will present our analysis of all phase-error sources as they contribute to the overall phase-error design goal of the system. We have shown how the design and implementation of both the calibration fixture and methodology meet that goal.
The uncertainty of the far field, obtained from antenna planar near field measurements, against the dynamic range is investigated by means of statistical analysis. The dynamic range is usually limited by the noise floor for low level signals and by the receiver saturation for high level signals. The noise level could be important for high measurement rate, which requires the usage of a high signal level to ensure a sufficient signal to noise ratio. As a result the nonlinearities are increasing, thus a compromise must be accomplished. To evaluate the effects of the limited near field dynamic range on the far field, numerical simulations are performed for dipoles array. Initially, the synthetic near field data corresponding to a given antenna under test were generated and directly processed to yield the corresponding far field patterns. Many far field parameters such as gain, beam width, maximum sidelobe level, etc. are determined and recorded as the error-free values of these parameters. Afterwards, the synthetic near field data are deliberately corrupted by noise and receiver nonlinearities while varying the amplitude through small, medium and large values. The error-corrupted near field data are processed to yield the far field patterns, and the error-corrupted values of the far field parameters are calculated. Finally, a statistical analysis was conducted by means of comparison between the error-corrupted parameters and the error-free parameters to provide a quantitative evaluation of the effects of near field errors on the different far field parameters.
Measuring the RF, microwave or millimeter wave reflectivity of materials and components often requires a substantial length of transmission line or cables to connect the microwave source/receiver to the test apparatus. Such cables may be subject to environmental variations (e.g. temperature or pressure) that change the overall phase delay and amplitude of signals that travel through said cables. Furthermore, some testing requires physical motion of the cable, which is another source of phase and amplitude error. When possible, great care is often taken to design a test apparatus or methodology to minimize movement of the test cables so that these position-induced phase errors are small. However, in some measurements, such as those that require scanning sensors or antennas, position-induced phase and amplitude errors cannot be avoided. In some situations, temperature variations that change the cable phase response are also unavoidable. The problem of cable-induced errors has been a concern for many different applications and there have been previous attempts to address it. These previous methods have used specialized microwave circuitry or a separate phase-stable reference device to measure and compensate for phase errors. In this paper, a new correction method is described, which determines and corrects for phase and amplitude errors in transmission line cables. Unlike previous published methods, the present technique does not require any specialized circuitry at the device under test (DUT). Instead it utilizes in-situ reflections that already exist in the measurement apparatus to obtain a reference phase and amplitude signal. The described algorithm combines these reflections with frequency and time-domain signal processing to compensate for erroneous phase and amplitude shifts that occur during a measurement. This paper demonstrates the correction methodology with materials measurements examples. Additionally, this phase and amplitude correction may be applicable for scatter and antenna measurements. It can be applied to either reflection or transmission measurement data.
This paper presents a study on the low-frequency coaxial reflectometer measurement procedure. A time domain gating algorithm is developed by ETS-Lindgren and the results are validated after comparing to the Keysight 8753-time domain algorithm. The in-house time gating algorithm is then applied to the simulated reflectivity results of absorbers in reflectometer to the simulation results of the same absorbers with plane wave excitation using finite element method numerical computation. Based on the simulation results, the operable upper frequency limit and the minimum length of the straight coaxial section for the reflectometer are suggested. The errors introduced during measurement due to higher order modes are studied and the permissible limit for the errors is analyzed. The different higher order modes and their effects on field distribution are studied. The impact of the non-uniform field distribution on the absorber reflectivity measurement is also discussed.
Several recent articles [1-9] have focused on assessing spherical near field (SNF) errors induced by using a non-ideal probe, i.e. a probe that has modal content. This paper explores this issue from the perspective of the probe response constants, defined by [10], which are the mathematical representation of the effect of the antenna under test (AUT) subtending a finite angular portion of the probe pattern at measurement distance . The probe response constants are a function of the probe modal coefficients, the size of the AUT (i.e. the AUT minimum sphere radius ), and the measurement distance , and thus can be used to evaluate the relative contribution of probe content as both measurement distance and AUT size varies. After a brief introduction, the first section of this paper reviews the theory describing the probe response constants; the second section provides some examples of the probe response constants for a simulated broadband quad-ridge horn, and the final section examines measured AUT pattern errors induced by using the corresponding probe response constants in a conventional SNF-to-FF transform. References: [1] A. C. Newell and S. F. Gregson, “Effect of Higher Order Modes in Standard Spherical Near-Field Probe Correction,” in AMTA 2015 Proceedings, Long Beach, CA, 2015. [2] Y. Weitsch, T. F. Eibert, and L. G. T. van de Coevering, “Investigation of Higher Order Probe Corrected Near-Field Far-Field Transformation Algorithms for Preceise Measurement Results in Small Anechoic Chambers, in AMTA 2015 Proceedings, Long Beach, CA, 2015. [3] A. C. Newell and S. F. Gregson, “Estimating the Effect of Higher Order Azimuthal Modes in Spherical Near-Field Probe Correction,” in EuCAP 2014 Proceedings, The Hague, 2014. [4] A. C Newell and S. F. Gregson, “Higher Order Mode Probes in Spherical Near-Field Measurements, in EuCAP 2013 Proceedings, Gothenburg, 2013. [5] A. C. Newell and S. F. Gregson, “Estimating the Effect of Higher-Order Modes in Spherical Near-Field Probe Correction,” in AMTA 2012 Proceedings, Seattle, WA, 2012. [6] T. A. Laitinen and S. Pivnenko, “On the Truncation of the Azimuthal Mode Spectrum of High-Order Probes in Probe-Corrected Spherical Near-Field Antenna Measurements,” in AMTA 2011 Proceedings, Denver, CO, 2011. [7] T. A. Laitinen, S. Pivnenko, and O. Breinbjerg, “Theory and Practice of the FFT/Matrix Inversion Technique for Probe-Corrected Spherical Near-field Antenna Measurements with High-Order Probes,” IEEE Trans. Antennas and Prop., Vol. 58, No. 8, August 2010. [8] T. A. Laitinen, J. M. Nielsen, S. Pivnenko, and O. Breinbjerg, On the Application Range of General High-Order Probe Correction Technique in Spherical Near-Field Antenna Measurements,” in EuCAP 2007 Proceedings, Edinburgh, 2007. [9] T. A Laitinen, S. Pivnenko, and O. Breinbjerg, “Odd-Order Probe Correction Technique for Spherical Near-Field Antenna Measurements,” Radio Sci., Vol. 40, No. 5, 2005. [10] J. E. Hansen ed., Spherical Near-Field Antenna Measurements, London: Peregrinus, 1988.
Spherical near-field (SNF) antenna test systems offer unique advantages over other types of measurement configurations and have become increasingly popular over the years as a result. To yield high accuracy far-field radiation patterns, it is critical that the rotators of the SNF scanner are properly aligned. Many techniques using optical instruments, laser trackers, low cost devices or even electrical measurements [1 - 3] have been developed to align these systems. While these alignment procedures have been used in practice with great success, some residual alignment errors always remain. These errors can sometimes be quantified with high accuracy and low uncertainty (known error) or with large uncertainties (unknown error). In both cases, it is important to understand the impact that these SNF alignment errors will have on the far-field pattern calculated using near-field data acquired on an SNF scanner. The sensitivity to various alignment errors has been studied in the past [4 - 6]. These investigations concluded that altering the spherical acquisition sampling grid can drastically change the sensitivity to certain alignment errors. However, these investigations were limited in scope to a single type of measurement system. This paper will expand upon this work by analyzing the effects of spherical alignment errors for a variety of different measurement grids and for different SNF implementations (phi-over-theta, theta-over-phi) [7]. Results will be presented using a combination of physical alignment perturbations and errors induced via simulation in an attempt to better understand the sensitivity to SNF alignment errors for a variety of antenna types and orientations within the measurement sphere. Keywords: Spherical Near-Field, Alignment, Uncertainty, Errors. References [1] J. Demas, “Low cost and high accuracy alignment methods for cylindrical and spherical near-field measurement systems”, in the proceedings of the 27th annual Meeting and Symposium, Newport, RI, USA, 2005. [2] S. W. Zieg, “A precision optical range alignment tecnique”, in the proceedings of the 4th annual AMTA meeting and symposium, 1982. [3] A.C. Newell and G. Hindman, “The alignment of a spherical near-field rotator using electrical measurements”, in the proceedings of the 19th annual AMTA meeting and symposium, Boston, MA, USA, 1997. [4] A.C. Newell and G. Hindman, “Quantifying the effect of position errors in spherical near-field measurements”, in the proceedings of the 20th annual AMTA meeting and symposium, Montreal, Canada, 1998. [5] A.C. Newell, G. Hindman and C. Stubenrauch, “The effect of measurement geometry on alignment errors in spherical near-field measurements”, in the proceedings of the 21st annual AMTA meeting and symposium, Monterey, CA, USA, 1999. [6] G. Hindman, P. Pelland and G. Masters, “Spherical geometry selection used for error evaluation”, in the proceedings of the 37th annual AMTA meeting and symposium, Long Beach, CA, USA, 2015. [7] C. Parini, S. Gregson, J. McCormick and D. Janse van Rensburg, Theory and Practice of Modern Antenna Range Measurements. London, UK: The Institute of Engineering and Technology, 2015
Dual-calibration was first proposed by Chizever et al. in 1996 [AMTA'1996] and had get wide applications in evaluation of the uncertainty in radar cross section (RCS) measurement and calibration. In 2013, LaHaie proposed a new technique based on jointly minimizing the mean squared error (MMSE) [AMTA'2013] among the calibrated RCS of multiple calibration artifacts, which estimates both the calibration function and the calibration uncertainty for each artifact. MMSE greatly improves the estimation accuracy for the radar calibration function as well as results in lower residual and RCS calibration errors. This paper presents a modified version of LaHaie's MMSE by minimizing the weighted mean squared error (MWMSE) for RCS calibration processing from multiple calibrator measurements, which is related to the following functions and parameters: the calibration function; the theoretical and measured RCS; the number of calibration artifacts the number of frequency samples and the weight for ith calibration artifacts which may be defined in terms of the theoretical RCS of all the calibration artifacts. For example, if the weight is defined as the inverse of the total theoretical RCS of the ith calibration artifacts for all frequency samples, the error then represents the total relative calibration error instead of an absolute error as in MMSE. MWMSE then means that an optimal calibration function is found in terms of minimum total relative calibration error, which is expected for most applications. Numerical simulation results are presented to demonstrate the usefulness of the proposed technique.
ABSTRACT Spherical near-field error analysis is extremely useful in allowing engineers to attain high confidence in antenna measurement results. NSI has authored numerous papers on automated error analysis and spherical geometry choice related to near field measurement results. Prior work primarily relied on comparison of processed results from two different spherical geometries: Theta-Phi (0 =?= 180, -180 = f = 180) and Azimuth-Phi (-180 =?= 180, 0 = f = 180). Both datasets place the probe at appropriate points about the antenna to measure two different full spheres of data; however probe-to-antenna orientation differs in the two cases. In particular, geometry relative to chamber walls is different and can be used to provide insight into scattering and its reduction. When a single measurement is made which allows both axes to rotate by 360 degrees both spheres are acquired in the same measurement (redundant). They can then be extracted separately in post-processing. In actual fact, once a redundant measurement is made, there are not just two different full spheres that can be extracted, but a continuum of different (though overlapping) spherical datasets that can be derived from the single measurement. For example, if the spherical sample density in Phi is 5 degrees, one can select 72 different full sphere datasets by shifting the start of the dataset in increments of 5 degrees and extracting the corresponding single-sphere subset. These spherical subsets can then be processed and compared to help evaluate system errors by observing the variation in gain, sidelobe, cross pol, etc. with the different subset selections. This paper will show the usefulness of this technique along with a number of real world examples in spherical near field chambers. Inspection of the results can be instructive in some cases to allow selection of the appropriate spherical subset that gives the best antenna pattern accuracy while avoiding the corrupting influence of certain chamber artifacts like lights, doors, positioner supports, etc. Keywords: Spherical Near-Field, Reflection Suppression, Scattering, MARS. REFERENCES Newell, A.C., "The effect of measurement geometry on alignment errors in spherical near-field measurements", AMTA 21st Annual Meeting & Symposium, Monterey, California, Oct. 1999. G. Hindman, A. Newell, “Spherical Near-Field Self-Comparison Measurements”, Proc. Antenna Measurement Techniques Association (AMTA) Annual Symp., 2004. G. Hindman, A. Newell, “Simplified Spherical Near-Field Accuracy Assessment”, Proc. Antenna Measurement Techniques Association (AMTA) Annual Symp., 2006. G. Hindman & A. Newell, “Mathematical Absorber Reflection Suppression (MARS) for Anechoic Chamber Evaluation and Improvement”, Proc. Antenna Measurement Techniques Association (AMTA) Annual Symp., 2008. Pelland, Ethier, Janse van Rensburg, McNamara, Shafai, Mishra, “Towards Routine Automated Error Assessment in Antenna Spherical Near-Field Measurements”, The Fourth European Conference on Antennas and Propagation (EuCAP 2010) Pelland, Hindman, “Advances in Automated Error Assessment of Spherical Near-Field Antenna Measurements”, The 7th European Conference on Antennas and Propagation (EuCAP 2013)
Coupling a Chip Antenna to an Antenna Measurement System is typically achieved using a co-planar micro-probe. This micro-probe is attached to a probe positioner that is used to maneuver the micro-probe into position and land it on the chip. Through this process, the chip is held by a chuck. Intentional and unintentional radiation from the Chip Antenna will interact with the micro-probe and chuck. From design conception, the antenna designer must take steps to reduce currents on the chip surface to minimize unintended radiation that will interact with both the measurement setup and the surrounding components of the final design. Even with good design practices, residual currents will still remain and radiate from the chip. Combined with intentional radiation from the chip antenna in the upper hemisphere, these radiated fields will impinge on the micro-probe and the probe positioner. Reflections from both the micro-probe and its positioner will reflect and generate interference patterns with the desired signal in the spherical measurement probe. In this paper, we evaluate, to first order, these effects by experimentation on two types of micro-probes (ACP & Infinity). The residual errors are then evaluated using modal filtering tools that further reduce these effects and the results are presented. Finally the dielectric chuck is modeled in simulation to evaluate the effects of the chuck on antenna patterns at 60 GHz and the results are presented.
MIL STD 461 is the Department of Defense standard that states the requirements for the control of electromagnetic interference (EMI) in subsystems and equipment used by the armed forces. The standard requires users to measure the unintentional radiated emissions from equipment by placing a measuring antenna at one meter distance from the equipment under test (EUT). The performance of the antenna at 1m distance must be known for the antenna to measure objects located at this close proximity. MIL STD 461 requires the antennas to be calibrated at 1 m distance using the Society of Automotive Engineers (SAE) Aerospace Recommended Practice (ARP) 958. This SAE ARP 958 document describes a standard calibration method where two identical antennas are used at 1m distance to obtain the gain at 1m for each antenna. In this paper the authors show using simulations that the SAE ARP 958 approach introduces errors as high at 2 dB to the measured gain and AF. To eliminate this problem the authors introduce a new method for calibrating EMC antennas for MIL STD 461. The Method is based on the well-known extrapolation range technique. The process is to obtain the polynomial curve that is used to get the far field gain in the extrapolation gain procedure, and to perform an interpolation to get the gain at 1 m. The results show that some data in the far field must be collected during the extrapolation scan. When the polynomial is calculated the antenna performance values at shorter distances will be free of near field coupling. Measured results for a typical antenna required for emissions testing per the MIL STD 461 match well with the numerical results for the computed gain at 1 m distance. Future work is required to study the use of this technique for other short test distances used in other electromagnetic compatibility standards, such as the 3 m test distance used by the CISPR 16 standard. Keywords: Antenna Calibrations, EMC Measurements, Extrapolation Range Techniques
Achieving a very large quiet zone across a wide frequency band, in a compact range system, requires a physically large reflector with a suitable surface accuracy. The size of the required reflector dictates attention to several important processes, such as how to manufacture the desired surface across a large area and the practicality of transportation and installation. This inevitably leads to the segmentation of the reflector into multiple panels; which must be fabricated, installed, and aligned to each other to conform to the required geometry. Performance predictions must take into account not only the surface accuracy of the individual panels but also their alignment errors. This paper presents the design approach taken on a recent project for a compact range system utilizing a blended rolled-edge reflector that produces a 5 meter quiet zone across a frequency range of 350 MHz to 40 GHz. It discusses the physical segmentation strategy, the fabrication methodology, the intermediate qualification of panels, the panel alignment technique, and the laser-based metrology methodology employed. Performance analysis approach and results will be presented for the geometry as conceived and then for the realized panelized reflector as machined and aligned.
Antenna measurement accuracies or error budgets are related to the signal-to-noise ratio of the measurement receiver. Signal-to-noise can be modelled taking two vectors, the intended signal and an interfering signal (i.e. from unwanted multipath propagation or simple noise), which superpose to the actual measured quantity. Although this approach is widely used, it is rarely discussed to its full extent. Instead, the error of the measured quantity is often estimated for high signal-to-noise ratios by applying worst case assumptions to the unwanted part. This excludes not only any statistical nature of the interfering signal but also the probability of the error appearance represented by its standard deviation. Especially when considering several error contributions in a total budget the adequate combination of different error probabilities yields a much more realistic result than adding single worst cases. Within this paper probability functions are analytically derived for measurement errors depending on the signal-to-noise ratio according to the aforementioned model. This yields a more sophisticated analysis of amplitude and phase errors having standard deviations for measured quantities. The confidence intervals of measurement errors are given with respect to varying signal-to-noise ratios. Limitations of the white noise synthesis with uniform phase and amplitude distributions are explained. Further, the applicability of worst case assumptions to the analytical solutions is discussed in the presence of high and low dynamic ranges. The derived expressions are statistically tested using Matlab calculations and compared to measurements with a vector network analyzer. The results are interpreted with respect to practical applications.