15 YEARS OF TERMINAL ANTENNA DESIGN AT LEMA : AN OVERVIEW A.K. Skrivervik, J.F. Zürcher Laboratoire d'Electromagnétisme et d'Acoustique, Ecole Polytechnique Fédérale de Lausanne, Station 11, CH-1015 Lausanne, Switzerland email@example.com Keywords: terminal antenna, multi-frequency antennas. Abstract The enlarged demand for electrically small antennas in the 900MHz – 5GHz range started in the early nineties, with the boom in mobile phones launched by the second generation of mobile phones. The laboratory of Electromagnetics and Acoustics of EPFL was involved in the design of dedicated antennas for mobile services since 1990, and has been involved in this topic since. More specifically, we have been interested in electrically very small antennas, multiband antennas, optimization, design rules and measurement. In this paper some of the key aspects encountered during these years of work on terminal antennas will be evoked, and illustrated by practical examples. 1 Introduction The boom of wireless services, which started in the early nineties with the democratization of the mobile phone has lead to the revival of a class of antenna which had been somewhat forgotten since world war II and the apparition of microwaves, namely the electrically small antenna. Indeed, even if the wavelength corresponding to these new wireless services (roughly 20 cm to 6 cm) remains very reasonable, the fact that the purpose of these services is to connect mobile users, a new paradigm has changed the viewpoint of antenna engineers : volume, weight and cost have become key issues for the antenna design. This trend is moreover increasing rapidly with time, as more and more services share the same mobile terminal. Thus, the nature of these new services, as well as the competition for the available frequency spectrum have broadened the range of requirements that are made for the antennas which are used on the mobile terminals. These requirements can be summarized as follows : • • • • • • • small dimensions low weight low induced SAR low cost high efficiency capability to handle multiple frequency bands broadband • • robust to changes in the environment optimize the use of the available channel capacity The first four requirements are typically "user defined requirements", the others being defined by the service provider or he network. Of course, depending on the considered service, the relative importance of all these requirements varies a lot. The size is for instance far less critical for WLAN system located in a laptop than for a DCS phone. The bandwidth and capacity however will be far more critical in the former example. Considering this, the design of terminal antennas is more than ever an art of defining the right compromise between all the requirement for a specific application. The Laboratoire d'Electromagétisme et d'Acoustique of EPFL has been active in the field of terminal antennas since the early nineties. Our first designs were related to the Swiss watch industry, which was very interested to check the feasibility of integrating various communication services like cordless phones, cell phones, pagers, GPS in a wrist watch (it works !), and some of these designs will be presented in section II. The following section, will be dedicated to multifrequency terminal antennas, and section IV to some critical aspects of antenna measurements. We will then conclude with new trends and perspectives. 2 Examples of very small antennas : integration in a wrist watch Our laboratory had a fruitful collaboration with the Swiss watch industry, as several mobile communication devices where planned to be incorporated in wristwatches. Some examples are shown below. 2.1 A miniature GPS antenna The main challenge in this design was to obtain a circular polarization is the very small space allotted to the antenna in this application. Indeed, the antenna was to be place at the top of the wrist watch, just beneath the hands, and the maximum available volume was of 30mm in diameter (meaning a radius of 0.075λo) and a height of 1.5 mm (0.0075 λo). The selected design was circular patch antenna with a slight notch on one axe to obtain circular polarization. In order to reach the specified dimensions, a dielectric substrate with a high permittivity was used (εr =10.5). This value of the dielectric constant was however not enough to obtain a resonance at GPS frequencies within the allotted space, so slots were etched in both axes of the patch in order to further reduce the latter. This option was preferred to selecting a higher permittivity, as the latter would have reduced the achievable bandwidth too much. The Antenna is depicted in figure 1, and its axial ratio in figure 2. The impedance bandwidth at -10dB of this antenna is rather small (0.6%), as was expected due too the very small size, but it is sufficient if a temperature stabilized dielectric material (like Rogers TMM for instance) is used. The axial ration bandwidth is lightly smaller (Figure 5), and the gain of the antenna was measured at – 5dBi . 29mm Figure 1 : GPS antenna for a wristwatch but the performances where nevertheless acceptable of the application in mind. 2.2 Miniature DECT and Bluetooth antennas A second design performed again, for an service to be integrated in a wrist watch was an antenna for a DECT phone [4, 5]. The overall dimension of the watch had to be comprised in a cylinder of 35 mm of diameter and 8mm height, this volume giving also space for the phone and the elements of the watch, including batteries. The antenna had to be in the watch itself, not in the bracelet, so there were only two possible locations : On the top (under the hands of the watch) or on the circumference of the cylinder. In order to achieve more easily the bandwidth required by DECT (1.881.9 GHz), we decided for the second option, thus to conform our antenna around the cylinder. The antenna type selected is a PIFA, as it has good bandwidth characteristics with respect to dimensions. The PIFA was first conformed around the cylinder, than integrated into it in order to avoid protuberances, as is sketched in figure 3. We named the final design SMILA (Small Monobloc Integrated L Antenna) Figure 3 : PIFA integrated in a wrist watch A realized prototype of the antenna is presented in Figure 4 2.5 MHz 2.6 MHz Figure 2 : axial ratio of the GPS antenna of fig. 1 Figure 4 : Prototype of the SMILA This performance can be considered as being quite satisfactory when compared to the fundamental limits for bandwidth and gain, as introduced by Chu , McLean  and Harrington , which are of @@% for the theoretical maximal achievable bandwidth and 1 dBi for the gain. We are of course quite far from these limits, due to the unfavourable aspect ratio of imposed by the specifications to the antenna, The measured performances of this antenna are presented in Figure 5. The – 10dB impedance bandwidth reached is of 4 %, and the measured gain of 0.5 dB, for a theoretical maximum of the gain after Harrington  of 2.5. This result is thus very satisfactory, taking into account the severe volume requirements, which shows us again all the potential of the PIFA for a compact antenna. Figure 5 : return loss of the SMILA antenna Figure 7 : PIFA integrated on the top of the watch The same antenna was designed for a Bluetooth application to be integrated in a wristwatch of same size. In this case, we measured also the effect of the users arm on the behaviour of the antenna . Unfortunately, the gain performances of the antenna deteriorate very drastically when the watch is worn on the wrist, as is illustrated by the measurements of figure 6. These measurements were done by placing the watch cylinder on a human arm phantom, with different spacing between the phantom and the watch. We see that when the antenna is too close to the arm, as is the case in practical situations, the performance of the antennas are degraded. Figure 8 : current distribution of the structure described in figure 7. 3. Multifrequency antennas Figure 6 : Gain versus frequency for the antenna at different distances of the arm We decided thus finally to modify our approach and to put the antenna on the top of the watch, keeping the integrated PIFA concept. The resulting design is illustrated in figure 7. The dimensions of the cylinder are the same as before. The measured gain a frequency of 2.42 GHz is 1.6 dB , which is an excellent result. Figure 8 shows the surface current distribution on the structure. The venue of new generations of voice services (DCS and UMTS to GSM for instance), and the offer of new services incorporated in phone terminals (like Bluetooth, GPS) require antennas which provides multiband possibilities. Indeed, a multiband antenna solution is often smaller and less costly than a solution with a distinct antenna for each frequency band. Several scenarii can occur, which are all illustrated on the well known PIFA antenna : • Multi bands but single feed [7-10]. The big advantage of having a common feed point for all the bands is that we do not need to care about mutual coupling problems. The drawback is that the radio front end has to discriminate the signal belonging to different services. This is the usual choice for dual band mobile phone handset antennas. An example of this kind of antenna is shown in figure 8. 4. On the Characterization of electrically small antennas Figure 8 : dual band PIFA antenna (GSM/DCS). • Multi bands with one feed per frequency band . The drawback is that the mutual coupling between the feed point can be high, which can degrade the overall performances in certain circumstances. However, each service is decoupled already at the antenna stage. This scenario is usually chosen when the two offered services are uncorrelated (voice and GSM for instance). This can also be a good choice for frequency bands which are far apart. An example of a double PIFA antenna is shown in figure 9. Figure 9 : dual PIFA antenna (GSM/DCS) • Multi bands with multi-feeds, however less feed than bands . This can be considered as a mixture of two former scenarios, and is often used to cope with the evolution of one type of service. For mobile voice in Europe for instance, one could use a multi band antenna with two feed points, one for the GSM/DCS bands and a second for UMTS. An example covering GSM/DCS on one port and UMTS on the other is shown in figure 3. During the middle nineties, the need for miniature antennas for mobile communication terminal became larger and larger, and in the euphoria of the fast miniaturization of chips and other electronically components due to the improvement of technologies, it was often forgotten that the fundamental limits linking antenna size and performances were not technological limits but physical limits. It was thus quite common to find, in specialized journals, press release on extremely small antennas boasting excellent (but unphysical) characteristics. Most of these adds were done in good faith by there authors, so what had happened ? The problem came from an incorrect measurement of the antennas. The difficulty in measuring electrically small antennas comes from the fact that they are often neither a symmetrical nor an asymmetrical structure. This is easily understood looking at the case of a microstrip patch antenna : on an infinite ground plane, the structure is clearly asymmetric and should be fed by an unbalanced transmission line. In the limiting case of an electrically small microstrip patch the ground plane had to be cut to the similar size than the patch, and the structure is symmetric, needing thus a balanced transmission line. In most practical cases an electrically small antenna will have a structure somewhere in between. As a result, there does not exist a transmission line which is able to feed these structures having no spurious current flowing on it. The consequence of this is that measuring an electrically small antenna using connecting cables on a conventional way, can lead to results as erroneous as when a dipole is measured connected to the equipment by a coaxial cable without balun. The errors introduced by the cable are particularly important when radiation characteristics of the antenna like the gain, pattern and efficiency are measured and can completely mask the true behaviour of the antenna . On the other hand, impedance measurements are less affected and often are still usable. In order to correctly measure the radiation characteristics of an electrically small antenna, following rules should be followed : • avoid the use of cable • perform system measurements, which means for the antenna to measure it in its final environment, taking into the casing, ground planes, batteries, etc. There has been much work on defining correct measurement procedures for electrically small antennas [5,13-17], and the topic is well understood today. Moreover, these procedures are evolving in order to take into account new terminal antenna features, like diversity gain, MIMO, etc. . Figure 10 : double PIFA for GSM/DCS (feed point to the left) and UMTS bands At EPFL-LEMA we also set up a measurement procedure dedicated to electrically small antennas . It is based on the principle described in figure 11, and is placed in an anechoic chamber. Figure 11 : Measurement principle The system, consists of a random positioner allowing to move the antenna and its casing in any direction (azimuth, elevation and polarization). The antenna is fed by a VCO inside the casing, and is thus transmitting. The receiving antenna is connected to a spectrum analyzer which records the maximum power received (for peak gain measurement) or the mean power received (for efficiency measurements). The antenna under test is then replaced by a known antenna (a dipole for instance) fed by known power, and the maximum gain or the efficiency are obtained by comparison. Note that the random positioner is made using only dielectric materials, the only metallic part being an electrical motor located at the bottom end of its foot, as depicted in figure 12.To obtain the maximum gain of the antenna under test, the random positioner turns the antenna until the maximum received power is detected by the spectrum analyzer. The Antenna under test is then replaced by a reference antenna which is powered so that the same level is reached on the spectrum analyzer. The maximum gain of the antenna under test is then given by : Pto _ dip + GainDip − Pow e rVCO = Gain AUT where GAUT is the maximum gain of the antenna under test. PVCO the known power delivered by the VCO, Pref the known power delivered to the reference antenna and Gref the gain of the reference antenna. The accuracy of the measurement is of ±0.5 dB, and depends mainly on the stability of the VCO and the precision of the reference antenna. The efficiency of the antenna can be obtained by measuring the average gain of the antenna instead of its peak gain. 5. Conclusions The demand for new mobile communication devices and systems will certainly continue for several years, increasing the market pressure for efficient electrically small antenna. These radiating devices should be optimized for each application, in order to get the best compromise between antenna volume, gain and bandwidth. In addition, more and more services will have to share the same terminal ; for this antenna, this means that a multifrequency or very broad band behaviour will be requested. Performances like some diversity or MIMO will also be introduced in order to improve the overall system gain. All this means that there remains a tremendous challenge for the antenna engineer to meet all these often antinomous specifications with one design. References  L.J. Chu, "Physical limitation on omni-directional antennas," Journal of Applied Physics, vol. 19, pp. 1163-1175, (Dec. 1948)  J.S. McLean, "A re-examination of the fundamental limits on the radiation Q of electrically small antennas," IEEE Trans. on AP, vol. 44, pp. 672-675, (May 1996).  R.F. Harrington, "Effect of antenna size on gain, bandwidth and efficiency," Journal of Research of the National Bureau of Standards-D. Radio Propagation, vol. 64D, pp. 1-12, (Jan-Feb 1960).  A.K. Skrivervik, J-F. Zürcher, O. Staub, J.R. Mosig, "PCS antenna design: the challenge of miniaturization",        IEEE Antennas and Propagation Magazine, vol. 43, Issue 4, pp. 12-27, (Aug. 2001). J-F. Zürcher, O. Staub, A.K. Skrivervik, "SMILA: a compact and efficient antenna for mobile communications", MOTL, vol. 27, no. 3, pp. 155-157, (Nov. 5 2000). J-F. Zürcher, O. Staub, A.K. Skrivervik, M. Hermanjat, "Accurate measurement of the maximum gain of electrically small antennas", MOTL, vol. 23, no. 6, 1999, pp. 328-331, (Dec. 20). G.Y. Lee and K.L. Wong, "Very low profile bent planar monopole antenna for GSM/DCS dual-band mobile phone", Microwave and Optical Technology Letters, Vol. 34, No. 6, pp. 406-409, (2001). P.L. Teng, H.T. Chen, "Multi-frequency planar monopole antenna for GSM/DCS/PCS/WLAN operation", Microwave and Optical Technology Letters, Vol. 36, No. 5, pp. 350-352, (2003). D. Kitchener, M.S. Smith, D.K. Power, J.G. Robson, L.P. Llewellyn, R.H. Johnston, "Multiband terminal and base station antennas for mobile communications", IEE 11th International conference on Antenna and Propagation (ICAP), 17-20. pp. 68-70, (2001). E. Lee, P.S. Hall and P. Gardner, "Dual band folded monopol/loop antenna for terrestrial communication system, Electronics letters, Vol. 36, pp. 1990-1991, (2000). J-F. Zürcher, A. skrivervik, O. Staub, S. Vaccaro, "A compact dual port, dual frequency printed antenna with high decoupling", MOTL, vol. 19, no.2, October 5, pp. 131-137, (1998).  F. Nunez, D. Llorens del Rio, J-F. Zürcher, A.K. Skrivervik, "Optimization of a tri-band mobile communication antenna using genetic algorithms", Journées Internationales de Nice sur les antennes (JINA'02), Nice, vol. II, pp. 65-68,( 2002).  Wheeler, H.A., "The Radiansphere Around a Small Antenna", Proc. IRE, pp. 1325-1331, (August 1959).  Pozar, D.M., Kaufman, B.: “Comparison of three methods for the measurement of printed antenna efficiency”, IEEE Transactions on Antennas and Propagation, vol. 36, no. 1, pp. 136 –139, (Jan. 1988).  Johnston, R.H., McRory, J.G.: “An improved small antenna radiation-efficiency measurement method”, IEEE Antennas and Propagation Magazine, vol. 40, no. 5, pp. 40 –48, Oct. 1998  K. Rosengren, P-S. Kildal, ”Characterization of antennas for mobile and wireless terminals in reverberation chambers : Improved accuracy by platform stirring”, Microwave and Optical Technology Letters, 20 (September 2001).  P-S. Kildal, K. Rosengren, J. Byun, J. Lee, “Definition of effective diversity gain and how to measure it in a reverberation chamber”, Microwave and Optical Technology Letters, Vol. 34, No 1, pp. 56-59, (July , 2002).  J. Carlsson, “Results and experiences from two years of benchmarking of measurements facilities for small antennas and terminals within ACE”., First European Conference on Antennas and Propagation (EuCAP), 2006, Nice, France, (6 - 10 November 2006).