Achieving Antenna Isolation Within Wireless Systems

Antennas embedded within compact devices, such as handheld computers and cellular telephones, must be optimized to limit interactions with surroundings. Such antenna isolation permits good efficiency within different enclosures and reduces the engineering time needed to incorporate antennas within new enclosures. Evaluation of antenna performance can be performed on an antenna within an anechoic chamber or measurements can be made on the complete wireless system of which the antenna is a part. The isolation can then be characterized by measuring the resonance frequency shift when parasitic elements are near the antenna. Such measurements will show that antenna isolation is a critical parameter when evaluating wireless designs as part of multiple systems within the same enclosure.

For practical reasons, multiple or multiband antennas are often implemented within the same enclosure. For efficiency and ease in integrating multi-band antennas, it is essential for their radiating elements to be highly isolated.

Antennas have been studied for over 50 years.¹ Antennas radiate electromagnetic (EM) waves that interact with resonant and absorbing materials nearby, including the enclosure or package in which they are mounted.² This interaction can reduce the overall efficiency of an antenna, with real-world antenna patterns representing some reflections due to the interference between directly radiated antenna waves and waves reradiated by different parts of an enclosure (Fig. 1). The radiation pattern of an isolated antenna can be smoothed and its efficiency improved by shaping its near-field pattern away from perturbators and absorbers. Moreover, the coupling between different antennas mounted inside the same enclosure is reduced.

An example will help to illustrate the importance of antenna isolation, based on experimental results obtained with a wireless communications system.³ Intrinsically, antennas interact with the surroundings as they radiate the EM energy carrying the information. Figure 2 presents the radiation patterns from two different antennas mounted on the side of a laptop computer (as in Fig. 1). The radiation pattern of antenna 1 (a printed stub) has many ripples whereas the pattern of antenna 2 is very smooth. The ripples are due to resonances created by the near-field energy of the antenna interacting with the enclosure.⁴ Part of this energy is reradiated and interferes with the direct emission in the far field. By measuring the full three-dimensional radiation patterns of both antennas and integrating the gain, it is possible to determine the efficiency difference between them.⁵ Such measurements reveal 4-dB better efficiency for antenna 2 compared to antenna 1, due to a dramatic amount of EM energy from antenna 1 absorbed by the enclosure. These results indicate that antenna 2 is better isolated than antenna 1.

Anechoic-chamber measurements are sometimes critical, as the antenna mounted onto the enclosure may not be exactly as it would be mounted onto a system.⁶ In the case of antennas 1 and 2, both antennas were tested with a wireless network system based on the Home-RF protocol, connected to laptop computers by means of PCMCIA cards. The measurements consisted of timing the transfer of a calibrated file (for example, a 1-MB file) for different antennas. This measurement can be expressed as an "active-bit-rate" function of the transmitted distance. Antenna 1, which is sold with the Home-RF system, and antenna 2 were measured inside a parking lot in order to reduce multipath effects. Figure 3 shows the evolution of the active-bit-rate measurement as a function of the distance. Whereas the bit rate obtained with antenna 1 drops very fast, either when the laptop computers are facing each other or when they are away from each other, the bit rate obtained with antenna 2 remains constant to a distance of 60 feet. At that point, it drops slowly, finally vanishing for a transmit range twice that of antenna 1. This measurement also shows that even in a reduced multipath configuration, the directionality of the antenna is not critical, at least for a Home-RF system. Unfortunately, this measurement does not give any figure of merit concerning the efficiency difference between antennas.

Page Title

To evaluate the efficiencies of the two antennas, another measurement setup was created. Using the same timing procedure, an attenuator was added between the RF output and the antenna. The other antenna, on the receiving laptop computer, was directly plugged into the output. Transfer times were then measured for different attenuation values from 0 to 11 dB. Figure 4 presents the results obtained for a dipole held away from the laptop computer, over the top of the lid, corresponding to an approximate 100-percent efficiency. Figure 4 also shows the bit rate obtained with antennas 1 and 2. A linear approximation gives the efficiency difference between both antennas as 4.5 dB in favor of the Ethertronics design. This result is consistent with the 4-dB difference obtained in the anechoic chamber measurements.

This example shows that when an antenna is poorly isolated, part of the radiated energy interacts with the case. This interaction can be observed as ripples in the radiation pattern, and can result in a dramatic drop in efficiency (4 to 4.5 dB), evidenced by reduced transmission range and system bit rate. Unfortunately, these measurements are time- and space-consuming and provide only limited information. Therefore, it is necessary to define some tools to evaluate antenna isolation in a more practical way.

A key requirement for antennas used within handheld devices is minimum use of real estate. Because of the increasing number of functions and shrinking sizes of handheld devices, antennas must be mounted close to other components, especially shields. A challenge for the antenna designer is then to tune the antenna inside the enclosure, taking into account all the interactions. The tuning and the matching of a patch or a planar inverted F antenna (PIFA) depend a great deal on the near-field environment of the antenna.⁷ To evaluate this influence, a solution consists in looking at the resonant frequency shift when a shield is moved around an antenna mounted on a ground plane. Mapping can be obtained to show the sensitive area around an antenna. This mapping also shows isolation differences between antennas. Figure 5 shows an example of mapping obtained with antenna 2.

The enclosure is not the only source of disturbances for embedded antennas inside handheld devices. The user's body is probably a bigger concern since the absorption/reflection configuration can change from one moment to another.⁸ Therefore, the influence of the body and particularly the interaction with the head and the hand must be studied in an analysis of antennas in handheld devices. The influence of the user's head is related to the antenna's shielding, a front-to-back ratio of the radiated power in the antenna's near-field pattern. This notion is quite different from the front-to-back ratio measured on the radiation pattern, corresponding to the far-field pattern. The hand's influence is linked to the isolation of the antenna since a user's hand is placed either beside or over the antenna. Therefore, it is necessary to study the influence of the hand on the resonant frequency of the antenna. Experiments have been carried out holding a cellular telephone in different hand positions while studying the resonant frequency. Figure 6 shows a cellular telephone in which a Global Positioning System (GPS) antenna is mounted. It also shows the return loss of the same antenna away from the hand and when the hand covers part of it. The resonant frequency does not really shift, it is mainly the matching which is disturbed, changing from −25 to −8 dB. But the antenna remains in an optimum configuration with the target GPS band even with the hand covering part of it.

Different tools can help provide a better understanding of the antenna's isolation either when the antenna is mounted inside an enclosure or simply on a printed-circuit board (PCB). This understanding can then help to improve antenna isolation and, as shown, also improve overall communications system performance. These methods are based on the current distribution analysis in the close environment of the antenna or even on the wireless device enclosure. This can be performed either numerically, using different in-house or commercial software,⁹ or experimentally using the near-field measurement setup as described below.

EM-simulation software is useful in antenna analysis since it can be used to estimate current and near-field distributions. By simulating an antenna on a finite-size ground plane, antenna isolation can be demonstrated. Figure 7 shows the current distributions of two different antennas, a PIFA type and a highly isolated Ethertronics IMD design. Results obtained with the PIFA show that currents are running on the ground plane, creating resonances on the edges. With the IMD antenna, currents are confined to the antenna area, with low currents on the edges.

Page Title

Software-simulation tools allow multiple results to be obtained at one time. Radiation patterns show the diffraction due to the edges even if, in this case, the limited size of the ground plane does not create a large phase difference and no dip null appears. Simulations can also be carried out when a shield is placed near the antenna.¹⁰ According to the simulations, a shield will cause a shift in the resonant frequency for both antennas. The resonant frequency of the PIFA shifts by 9 percent, but the resonant frequency of the isolated antenna shifts only by 3 percent.

Moreover, by looking at the current distributions, the PIFA antenna shows a strong coupling to the shield. But even if some currents appear by the shield with the IMD antenna, those currents are very low. EM-simulation tools play a big role in antenna development, in order to make sure that a new design can meet system requirements, including the use of small antennas inside handheld devices supporting multiple applications with simultaneous transmission and reception.

In order to obtain experimental results in support of the simulations, it was necessary to move a magnetic probe around each antenna.¹¹ For these measurements, the probe could be moved along the x-, y- or z-axis. A simple program using a GPIB interface can be written in order to move the probe near an antenna under test. The probe must be small to reduce any kind of interactions between it and the antenna under test. Using such a probe, experiments were carried out to study the current repartition over an entire cellular-telephone case. For a poorly isolated antenna, currents are running all over the case. Currents are higher on the edges of the case, as they are diffracted and create resonances in those areas. This is because part of the energy is reradiated and interferes with the antenna's direct radiation, and part of the energy will just be lost in heat inside the case. With an isolated antenna, currents are localized within the antenna area, reducing losses due to the enclosure and reducing antenna tuning/implementation time within the enclosure.

The first benefit of enhanced isolation is improved antenna efficiency, when an antenna is implemented within a handheld device. This greater efficiency translates into improved communication quality or better transmission range, longer battery operating lifetime, and the capability of using smaller batteries. It is also easier to tune an isolated antenna from one housing to another, potentially reducing product time to market. Finally, a properly isolated antenna can be used as a standard design for numerous handheld platforms, with only one or two parameters changing between designs.

REFERENCES

1. N. Marcuvitz, Waveguide Handbook, New York: McGraw-Hill, New York, 1951.
2. Ethertronics internal report 10019A, Antenna loss inside enclosures, 2000.
3. White paper on consumer use of Home-RF Technologies, www.homerf.org/data/tech/consumerwhitepaper.pdf.
4. G. Poilasne, Antennas and Photonic Band-Gap Materials, Ph.D. thesis, University of Rennes, France, 1999.
5. W.L. Stutzman and G. Thiele, Antenna Theory and Design, Artech House, 2nd Ed. NY 1998.
6. G. Hindman, Anechoic Diagnostic Imaging, AMTA Symposium, Antenna Measurement Techniques Association, Columbus, OH, 1992.
7. Z.N. Chen, Experimental Investigation of Impedance Characteristics of Patch Antennas with Finite-Size Ground Substrates, Microwave and Optical Technology Letters, Vol.25, No.2, April 20, 2000, pp-107-111
8. K. Ogawa, T. Uwano, M. Takahashi, Shoulder-Mounted Planar Antenna for Mobile Radio Applications, IEEE Transactions on Vehicular Technology, Vol. 49, No.3, May 2000, pp.1041-1044.
9. Version 8.0 of Ensemble, Ansoft Corp., Pittsburgh, PA.
10. G. Poilasne S. Rowson, L. Declos, Antenna Isolation, Ansoft World Wide Workshop series, Ride The Wave, San Diego,CA, 10/12/01.
11. D. Slater, Nearfield Antenna Measurements, Artech House, MA, 1991.