The last five years the reverberation chamber has been developed to an accurate instrument for measuring the performance of small antennas and active mobile terminals in Rayleigh fading. The small antennas are characterized in terms of radiation efficiency, two-port antennas in terms of their diversity gain, and multi-port antennas for MIMO systems in terms of the maximum available capacity [4], all of which can be measured. The active mobile terminals are characterized in terms of the measured total radiated power, receiver sensitivity, and realized diversity gain. The reverberation chamber is known to emulate Rayleigh fading with a uniform distribution of directions of propagation of the waves, both in azimuth and elevation.

The present paper will for the first time give a gathered presentation of the research done by the authors in relation to achieving an accuracy of 0.5 dB RMS or better, when measuring efficiency, power and receiver sensitivity. The accuracy has been verified by comparison with measurements in anechoic chambers [3], and between reverberation chambers of different size [3], and from participation in benchmarking of measurement ranges done within ACE (Antenna Center of Excellence, a European Network of Excellence). The paper will describe the development of new and efficient stirrer techniques (platform [1] and polarization stirring [2]) to obtain the required accuracy with as small chamber size as possible, the systematic use of frequency stirring (averaging) to improve accuracy at the same time as resolution is kept under control, and the use of loading the chamber to allow measurements of active terminals for wideband systems such as UMTS. GSM terminals do not require more loading than a head phantom, whereas UMTS requires additional loading. Computer simulations of a numerical chamber of the same size as the actual chamber [5] are used to determine the location of the wall probes, and the amount of chamber loading to preserve accuracy.

The recent work is done in relation to the development of a new strongly shielded chamber for measurement of receiver sensitivity, i.e. TIS (Total Isotropic Sensitivity) and AFS (Average Fading Sensitivity). The performance of this chamber will be presented and compared with other chambers.

References

[1] K. Rosengren et al, "....Improved accuracy by platform stirring", MOTL, Sep. 2001. [2] P-S. Kildal, C. Carlsson, "Detection of a polarization imbalance..", MOTL, July 2002. [3] P-S. Kildal, C. Carlsson, TCP of 20 phones measured in reverberation chamber, Technical report, Bluetest AB, February 2002. [4] P.-S. Kildal, K. Rosengren, "Correlation and capacity of MIMO systems ..", IEEE Comm. Magazine, Dec. 2004. [5] U. Carlberg, P.-S. Kildal, and J. Carlsson, ".. using method of moments with cavity Green`s function calculated by Ewald summation", IEEE Trans. EMC, Nov. 2005.

Hologram CATR has been developed and demonstrated as most feasible new means of measuring mm and sub-mm wave antennas when truly approaching THz operation frequencies. Such a CATR facility and related techniques have been specifically designed for measuring antennas with a large aperture but can be applied also for other antennas. A hologram CATR overcomes many of the sub-mm wave operation problems associated with other or more traditional CATR methods. The Hologram CATR principle has been demonstrated also in practice at 322 GHz when measuring a 1.5 m aperture antenna. Currently a project for demonstrating the capability at 650 GHz is being carried out as reported separately in another paper submitted to this conference.

Even after the arrival of the hologram CATR, antenna measurements at high mm and sub-mm wave frequencies still present some challenges especially when considering the measurement of antenna pattern sidelobes. In some applications (e.g., antenna of the Planck Surveyor) there exist needs for measurement of very low-level sidelobes with an unprecedented accuracy. For such measurement needs, evolved Antenna Pattern Comparison techniques have been developed to provide additional accuracy to combat against the residual interfering factors available in any measurement application at THz frequencies. These techniques include feed-scanning and frequency-shift based methods. These techniques have currently been verified in small-scale demonstrations but they are also planned to be used as measured pattern correction techniques in the on-going 650 GHz large scale antenna measurement project carried out for ESA/ESTEC.

The paper will cover briefly the operation and features of a hologram CATR and related measurement applications but will mainly focus on other newest developments including the evolved techniques aimed especially at enhancing low-level sidelobe measurement accuracy.

1 - Introduction

Since several years IETR have been conducting active research activities on the modelling and design of millimetric radiating structures and devices. The experimental validation of innovative designs and home-made electromagnetic CAD tools is of uppermost importance. To this end, prototyping capabilities and complementary sets of antenna test facilities at millimetre waves are available for routine experiments (e.g. [18-110] GHz far field chamber, [2-110]GHz coaxial vector network analyser). In such a context, antenna measurements are considered as a key step of the research process. Even though our measurements facilities are not used for certification purposes, they need to provide reproducible, reliable and accurate data for a wide diversity of antenna topologies and dimensions. Then, definition of the same coordinates system for measurements and simulations remains a sensitive point. We present in this paper an easy-to-use positioning procedure based on phase measurements from Ka-band to W-band. This technique is briefly described in Section 2. Various experimental results are described in Section 3.

Defining the local coordinate system is a usual problem in antenna measurement systems. Conventional procedures using lasers and theodolite have been employed successfully but require expensive tools and/or very precise mechanical holders. In addition optical references might be inadequate due the small size of antennas under test. As a consequence, the implementation of a positioning procedure based on the determination of the phase center location is of particular interest. This is easily achieved by minimizing phase derivatives simultaneously in main beam H- and E-planes for antenna having the same phase center in both planes. For antenna configurations that do not fulfil this condition (e.g. horn antennas), an additional antenna should be used to determine initial adjustments and then according to these settings, the position of the device under test is determined.

This measurement technique has been applied successfully for various radiating structures. In the final paper, several experimental results will be described in detail, namely two integrated lens antennas (ILAs) operating at 28 GHz ( Fig. 1 and 2 ) and 58.5 GHz for shaped beam applications, an inhomogeneous lens based on a half Maxwell fish-eye lens, a lens horn designed for automotive radars at 77 GHz and a multilayer planar antenna array ( Fig. 3 ) designed for one-feed-per-beam reflector antennas during the MIPA European project ( MEMS based Integrated Phased Array ).

We present in this paper a simple positioning procedure. Its application on the characterisation of different kinds of antenna configurations has been evidenced by comparing measured and computed amplitude and phase patterns from 28 GHz to 77 GHz.

Antenna far-field reconstruction from the knowledge of complex field distribution in the radiating near-field region is a well-established technique for testing electrically large antennas. However, difficulties are encountered as long as the frequency increases, due to the fact that accuracy of phase measurements is strongly compromised unless complex and expensive test facilities are used. In order to have reliable results even in the presence of large radiating systems working at millimeter frequencies, methods evaluating far-field pattern from amplitude-only near-field measurements have been recently investigated in literature [1]. They are substantially based on a formulation in terms of a non-linear estimation problem for the retrieval of the near-field phase as result of the minimization of a suitable functional, starting from the knowledge of near-field amplitude over two or more distinct surfaces. A new hybrid approach is proposed in this paper to perform amplitude-only near-field measurements on a single scanning surface. Two co-planar probes are used to obtain phaseless data which are subsequently processed by an interferometric procedure retrieving the near-field phase information. In order to collect the necessary amplitude information, a microstrip circuit is designed which receives the complex signals from the two probes and transmits them to a couple of hybrids giving the amplitude-only signals as revealed by detector diodes. A prototype of the microstrip circuit working on the X-band is illustrated in Fig.1, where the interconnected hybrids are also shown. Once amplitude-only data are known, an interferometric formula is applied which gives the near-field phase distribution apart from a certain number of unknown phase shifts, depending on the distance between the probes, which are subsequently retrieved by a suitable minimization procedure. As a further investigation, the application of the integrated probe to new strategic geometries, such as the helicoidal one (Fig.2), is also considered to reduce both the acquisition and the computation time, by developing an efficient transformation algorithm, based on the use of the fast Fourier transform and the related shift property, which gives the far-field pattern directly from the knowledge of amplitude-only near-field samples on the helicoidal curve. This avoids any intermediate interpolation on cylindrical surfaces, which is usually performed to apply the well-known cylindrical near-field to far-field transformation. As a validation example, numerical simulation on a linear array of 37 z-polarized Huyghens sources is considered. Amplitude-only data are collected on a cylindrical helix of radius ro=21.5 wavelenghts and height h=120 wavelenghts, at an angular step d=2.38°. An excellent agreement can be observed in Fig.3 between the retrieved and the exact near-field phase.

References

1. O. M. Bucci, G. D’Elia, G. Leone, R. Pierri, "Far-field pattern determination by amplitude-only near-field measurements", Proc. 11th ESTEC Workshop on Antenna Measurements, Gothenburg, Sweden, 1998. Fig. 1 Microstrip circuit with interconnected hybrids Fig. 2 Helicoidal configuration Fig. 3 Comparison between exact and retrieved near-field phase on the helicoidal curve for a simulated arrayof 37 Huyghens sources

Miniature antennas are comparable in electrical size to common test equipment, which perturb the normal behaviour of the antenna. Current methods can be considered non-invasive, however there are still some limitations that hinder an accurate characterization of small antennas. The wheeler cup method requires the use of a metallic plane, being suitable for characterizing antennas with a ground plane. The transmission power budget method entails the use of a VCO connected to the antenna port, what perturbs the measurement when dealing with extremely small antennas.

The way to overcome the current limitations entails measuring the antenna under conditions that faithfully reproduce its actual working environment. This paper compares two novel gain-efficiency measurement methods for electrically small antennas which fulfil the required characteristic of non invasiveness, reproducing at the same time the actual antenna operation scenario.

The measurement method based on the MST principle consists on measuring the scattered field created by the antenna when it is illuminated by an incident field which is radiated from a calibrated antenna with transmitted power PT and gain GT. The signal scattered by the small antenna is modulated (with a modulation index Ąm) by a photodiode, and a coherent detection in the receiver extracts the desired signal (modulated reflected scattering), PR, from the noisy background (unmodulated reflected scattering). The antenna gain, GAUT, can then be extracted from the measured parameters as: (1)

Figure 1: Measurement set-up for efficiency measurement using MST

The calorimetric method is based on the measurement of the power dissipation in the antenna under test ([1]) when it is located in a thermally insulated box whose walls are transparent to microwave radiation. The accuracy of the method has been proved to be better than 1 %. Measurement results will be discussed and compared to those of wheeler cup and transmission power budget methods.

References: [1] Schroeder, W.L. and Gapski, D. "Direct Calorimetric Measurement of Small Antenna Radiation Efficiency". IEEE APS/URSI International Symposium 2005.

We are in the era of wireless communications and devices. The antennas that enable these technologies are electrically small antennas that can be challenging to test and analyze. Yet, the industry is becoming more standardized, and so too are the tests and certifications being adopted to validate these antennas. These antennas must undergo "antenna measurements" to characterize such information as far-field patterns and gain. Additionally, hand-held devices must satisfy requirements of the Over-the-Air (OTA) performance measurements as specified by the Cellular Telecommunication and Internet Association (CTIA). To perform the tests required by the wireless industry, a system that can accurately collect data on a spherical surface enclosing the AUT is needed. This system also has to provide the appropriate data analysis capabilities and has to be constructed from dielectric materials to minimize reflections.

This paper describes a 3D measurement system to address these challenges. It is an ideal system for measuring medium and low gain antennas and is well suited for wireless antenna testing. It has successfully demonstrated testing and characterization of far-field angular spectra, CTIA OTA performance data, EIRP, and RFID testing.

As well known, the directivity of an antenna denotes its capacity to concentrate the electromagnetic (EM) radiation in a desired direction. The computation of the total radiated power, which is obtained by integrating the radiation intensity on the far-field (FF) spherical surface, represents the crucial point in the directivity evaluation. By properly exploiting the nonredundant representations of EM fields [1], an efficient Sampling Interpolation (SI) formula has been proposed in [2] for determining the total radiated power from the knowledge of a nonredundant number of the radiation intensity samples. Since near-field (NF) measurements allow the antenna testing in a controlled environment, as an anechoic chamber, it can be indispensable to recover the FF data via NF–FF transformation techniques. In particular, the NF-FF transformations employing planar scannings are well-suited for directive antennas which radiate pencil beam patterns. In such a context, an effective NF–FF transformation with planar spiral scanning has been proposed in [3] for reducing the time needed for data acquisition by means of continuous movements of the positioning systems of the probe and the antenna under test.

S
q
d
z
x y
a
P( ) j
O

Fig.1 - Planar spiral scanning.

The aim of this work is to give a full procedure for determining the antenna directivity from a nonredundant number of NF measurements acquired via a planar spiral scannning facility. In fact, for pencil beam antennas, the power radiated in the half-space z < 0 is negligible and the total radiated power coincides practically with that radiated in the forward half-space. The proposed procedure consists of three steps: a) acquisition of the NF samples; b) evaluation of the radiation intensity samples on the FF spherical surface via the aforementioned NF-FF transformation; c) computation of the total power radiated in the halfspace z > 0 by means of the here developed SI formula. It is worthy to note that the step b) is split in two substeps. In the former, as a result of the fast Fourier transform (FFT) algorithm employed in the NF-FF transformation, we get FF samples uniformly spaced in kx =b q f sin cos and ky =b q f sin sin , b being the wavenumber. In the latter, a two-dimensional optimal sampling expansion of central type [4] is properly employed to reconstruct the radiation intensity data at the sampling points (uniformly spaced in q and f) required by the here developed SI formula. This last is different from that proposed in [2] since the radiation intensity samples are known only on the forward FF half-sphere and, accordingly, the related q-integral must be restricted to [0, p/2]. Numerical examples assessing the effectiveness of the proposed technique will be reported.

[1] O.M.Bucci, C.Gennarelli, C.Savarese, "Representation of electromagnetic fields over arbitrary surfaces by a finite and non redundant number of samples," IEEE Trans. Antennas Propagat., vol. 46, pp. 351-359, 1998.
[2] C.Gennarelli, G.Riccio, C.Savarese, "Closed form evaluation of the antenna directivity via sampling expansion," JEMWA, vol. 16, pp. 861-870, 2002.
[3] O.M.Bucci, F.D'Agostino, C.Gennarelli, G.Riccio, C.Savarese, "Probe compensated FF reconstruction by NF planar spiral scanning," IEE Proc. - Microw., Antennas Propagat., vol. 149, pp. 119- 123, 2002.
[4] C.Gennarelli, G.Riccio, F.D’Agostino, F.Ferrara, Near-Field - Far-Field Transformation Techniques, Salerno, CUES, 2004.

Hansen's spherical near-field (SNF) to far-field transformation theory utilizes a modal normalization that gives the power in a transmitted mode as ½ times the square of the modulus of the complex modal coefficient corresponding to that mode.[1] This normalization factor was chosen to make the expression for total transmitted power as simple as possible. P = ½ Ósmn |Qsmn|2 , where Qsmn is the complex modal amplitude for mode s,m,n in the expansion of the antenna's electric field. Hansen also employs the concept of a scattering matrix to represent the general characteristic of an arbitrary linear antenna. The corresponding scattering matrix characteristics are written as Tsmn and Rsmn, for the transmitting and receiving coefficients respectively.[2] The subscripts s,m,n are the same as the spherical modal indices that serve as labels for the various vector spherical modes. Using, Hansen's normalization convention, employed in Spherical Near-Field Antenna Measurements, one must make the following conversion between Tsmn and Rsmn for a reciprocal antenna. [3]: Rsmn = (-)m Ts(-m)n . In this presentation I show how a small modification in the choice of mode normalization changes the transmitting-receiving conversion to an equality for a reciprocal antenna.[4] This change affords us the simpler expression Rsmn = Tsmn ; and, as a result the opportunity to avoid confusion when manipulating the scattering coefficients. For this relation to hold, associated with this alternative normalization, is a useful convention to define the antenna's fiducial coordinate system, which I also discuss. This alternative choice preserves the power normalization described above and preserves the form of the SNF transmission equation. It also simplifies the expression of the translation theorem for the probe antenna coefficients.

References

[1] Spherical Near-Field Antenna Meaasurements, J.E.Hansen, Editor, Peter Peregrinus Press, London, UK, 1988, pp.13-14 & Eqn. 2.24.
[2] Ibid. pp.13-14 , Section 2.3, pp.27- 48.
[3] Ibid. pp.13-14 , pp.13-14 & Eqn. 2.107.
[4] Spherical Near-Field Range Configuration, Scientific-Atlanta, Inc., (now MI Technologies, Suwanee, Ga, USA.) 1982. Sections 3.3 - 3.4.

Accurate antenna pattern characterization by probe-corrected spherical near-field antenna measurements at the DTU-ESA Spherical Near-Field Antenna Test Facility [1] is currently based on using first-order (|\mu| = 1) probes. The applied probe correction technique for these probes is the traditional first-order probe correction technique [2]. The probes are conical horns fed by a circular waveguide. Since these probes are operated only in a relatively narrow frequency band, an undesirably large set of probes is required to cover the whole frequency range of the DTU-ESA Facility. The probes are also heavy and large at low frequencies (e.g. f < 3 GHz).

Spherical near-field antenna pattern characterization procedure, that employs a high-order probe, is currently being developed at the DTU-ESA Facility [3-5]. Among other things, this includes developing a general high-order probe correction technique [5]. This new procedure will enable, in theory, (almost) an arbitrary probe to be used in the measurements, and thus, compared to the existing procedure, it creates a greater flexibility to choose a light-weight and a wideband probe for antenna measurements.

Although the possibility of using a high-order probe provides certain clear advantages, some new requirements and restrictions will arise, that require special attention. The purpose of this paper is to discuss the challenges and practical issues related to the accurate antenna pattern characterization using a high-order probe. In particular, aspects related to the requirements for the scanning grid, for a dual-port probe and for the channel balance calibration procedure are raised.

[1] http://www.emi.dtu.dk/research/afg/snf/SNF.html
[2] J. E. Hansen (ed.), Spherical Near-Field Antenna Measurements, Peter Peregrinus, Ltd., London, 1988.
[3] T. Laitinen, S. Pivnenko, O. Breinbjerg, "Iterative probe correction technique for spherical near-field antenna measurements," IEEE Antennas and Wireless Propagation Letters, Vol. 4, 2005, pp. 221-223.
[4] T. Laitinen, S. Pivnenko, O. Breinbjerg, "Odd-order probe correction technique for spherical near-field antenna measurements," Radio Science, Vol. 40, No. 5, 2005, pp. 3009-3019.
[5] T. A. Laitinen, S. Pivnenko, O. Breinbjerg, "High-order probe correction for a square waveguide probe in spherical near-field antenna measurements," abstract accepted and final paper submitted to AMTA-Europe 2006, Munich, Germany, May 1-4, 2006.

Spherical near field measurement techniques combined with probe array technology offer a fast and accurate way to measure antenna performances. The use of increasingly higher frequencies and reduced testing time in all modern antenna application has increased the need for probe array based measurement systems at higher frequencies.

The spherical near-field measurement system (Starlab) has recently evolved to cover the frequency band from 800Mhz to 18GHz. The original probe array containing 15 probes covering the 800MHz to 6GHz band has been augmented with a second set of 16 dual polarized probes covering the 6GHz to 18Ghz range. The two probe array are interleaved and fully integrated in the original structure of the StarLab. The new system offers wider bandwidth, the speed advantages of a probe array while the mechanical rotation allows for unlimited angular resolution over the full sphere.

Several important design changes have been implemented. New high frequency dual polarized probes have been developed, the mechanical positioning system has been upgraded for higher precision and new efficient software for the spherical near-field to far-field transformation with non-uniform near field sampling has been developed.

This paper discuss the important system design aspects of the evolved measurement system and present results from the validation campaign.