Feature

Antennas: critical links in the wireless signal chain

A communications chain is no stronger than its weakest link. The right antennas can strengthen the chain by yielding better signal coverage, increased S/N ratio, reduced bit error rate, and lower power consumption-all at very low cost.

By Bill Schweber, Technical Editor -- EDN, 6/6/1996

Yet, good antennas, or "aerials," at the transmitter and receiver are vital. The right antenna can increase performance and overcome coverage weaknesses at little cost and with no increase in power consumption. You may even be able to reduce transmitter power by using a better antenna configuration.

A typical wireless application comprises a base station with a relatively fixed, planned-in-advance location. In contrast, the corresponding remote unit is often mobile and can be a wireless LAN, a handheld data-entry wand for taking inventory in a warehouse, a PC Card-based modem, or an in-plant communication link. A typical remote unit not only moves randomly within the overall area of coverage, but also varies in its orientation, must operate within local RF hot and shadow spots, and is typically less powerful and more design-constrained than is its base station.

Fortunately, today's 900-MHz and higher wireless applications offer designers a big technical advantage compared with the approximately 100-MHz applications of a decade ago. At 1 GHz, one wavelength (l) is just 30 cm, or about 1 ft. Therefore, a physically small and convenient antenna can be reasonably efficient and even electrically resonant at the frequency of operation, which makes providing a good antenna easier.

An effective antenna is more than just a conductor with the right size and shape. You also must consider the antenna interface to the transmitter power amplifier and to the receiver front end. Filtering and impedance matching let you realize an antenna's maximum potential and minimize sources of interference, which can severely degrade system performance.

Start with basic configurations

Although it may seem that antennas, like wallpaper patterns, come in a nearly infinite number of configurations, most antennas are based on one of four elements: the dipole, monopole, loop, and patch. More complex antennas combine these simple elements, use multiples of one type, or add reflectors or other structures to shape the propagation field pattern. Antennas have 360°: azimuth (or polar) radiation-gain and elevation patterns. These features are the primary parameters of performance. Other important criteria are beam width, bandwidth, front-to-back ratio, and polarization.

Antenna gain differs from amplifier gain, because antenna gain involves no active circuitry or increase in total signal strength. The gain of an antenna indicates how it shapes its radiation field pattern in decibels compared with an ideal isotropic antenna (dBi). An ideal isotropic antenna radiates or senses power uniformly over a spherical volume and is impossible to build. You can also compare an antenna's gain with a dipole, which you can build and use as a reference.

The oldest antenna structure is the dipole, or Hertz, which is usually fed by a transmission line at the antenna's center point. It is self-resonant at a length of one-half the operating wavelength, with an impedance of 72(ohm). In practice, the dipole's actual self-resonance is 3 to 5% shorter than the length in ideal theory, due to conductor size, the fringing at the ends of the antenna, and the effect of nearby objects. Figure 1 shows the radiation pattern for a vertical dipole.

Slightly younger than the dipole is the monopole, also called the "whip," or Marconi, antenna. The monopole is a vertical dipole; however, the phantom reflection of a conductive ground plane underneath the antenna replaces one leg of the dipole. This antenna is one-fourth-wavelength long, and its impedance is 36(ohm), one-half that of a dipole. Unfortunately, the ideal full conductive plane under the antenna usually is nonexistent or erratic. The actual azimuth pattern, thus, depends heavily on installation and use, in contrast to its theoretical circular pattern. The elevation radiation angle is also a function of the ground-plane situation and antenna's height above ground.

The third type of antenna, the loop, can be rectangular or circular and resonates at a perimeter length of one wavelength; it is fed by simply breaking anywhere into the loop (Figure 2). Although loops are often mechanically difficult to support at long wavelengths, they are practical for personal-communications-system (PCS) frequencies.

The patch antenna is a conducting surface separated from an underlying ground plane by a dielectric; a double-sided circuit board often works as a dielectric (Figure 3). Each edge is one-half wavelength at resonance, or you can use a circular patch with a radius of 0.3[lambda]. You feed the antenna through a small hole in the ground plane.

Antenna operation and coverage are the same whether the antenna is transmitting or receiving. From an analysis perspective, the reciprocity principle shows that the radiation pattern, gain, impedance, and other antenna parameters are the same in either mode. However, high-power transmitter applications need antenna structures and components that can carry higher currents. This requirement involves larger wires or tubing, higher power ratings on associated circuitry, and other structural enhancements.

Reality intrudes on free space

Antennas in actual mobile PCS applications are often smaller than the free-space or ideal-ground self-resonant dimensions indicate. In addition, the antenna is near other electronic circuitry, a user's body, an enclosure, power circuitry, and structures. You can consider these factors when designing the mobile antenna and its feed system, but you still must rely on a properly configured base station to overcome some coverage deficiencies.

Fortunately, antennas that are smaller than resonant size can still be effective radiators or energy receivers. Pagers, for example, use loop antennas that are about (1/10)[lambda]. However, the impedance-matching circuitry between the antenna and the power amplifier or front end causes losses and, thus, wasted power, reduced coverage, or weaker received-signal strength.

Because smaller antennas usually have a lower resistive component to their impedance, the dc resistance of the matching circuitry increases in significance. For example, an antenna with several ohms of resistance, fed by matching circuitry with a comparable resistance, wastes half the transmitted or received power in the matching circuitry, yielding a painful, 3-dB loss. The lower antenna resistance causes higher antenna currents and ohmic losses through matching components.

Short dipoles and monopoles have a capacitive impedance and, thus, have a lower resistive value than do resonant-length, free-space antennas, which have a purely resistive impedance. Therefore, the matching circuitry that transforms the antenna's complex impedance into an apparent resistance must introduce inductance to compensate. You implement this inductive loading in monopoles as a discrete wire coil at the antenna base, a coiling of the antenna whip at its base, or a continuously wound helix around a flexible core—the common, rugged, bendable "rubber-duck" antenna.

The whip antenna looks fairly simple. However, according to Centurion International, a leading manufacturer, combining the desired electrical, mechanical, connector, and impedance-matching characteristics often requires a custom design for best electrical performance, mechanical ruggedness, and lowest cost. A custom design originally for another OEM, though, can also meet your needs, especially if your requirements are flexible.

Most pagers and wireless wands use loop antennas. Unlike the dipole and monopole, the smaller-than-resonant loop antenna is inductive and needs capacitive compensation to yield the resistive result. Receivers usually have to compensate for the small loop's low efficiency. This compensation means putting more gain into the receiver's front end, which may be costly, or requiring lower noise, low-intermodulation-distortion components.

You can increase the loop efficiency with a few simple steps. Both theoretical analysis and field tests show that a loop antenna using flat-strip conductors, such as pc-board traces, has efficiency 1.5 to 2 dB greater than that of a loop antenna using round-wire conductors of the same cross-sectional area (Reference 1). This efficiency results from the fact that the antennas with flat-strip conductors have lower skin-effect loss in the rectangular conductor than that of antennas with round-wire conductors. This improvement is probably the cheapest and easiest-to-achieve you'll ever see in system design.

Similar analysis shows that using two coplanar loops instead of a single loop provides a 3-dB increase in efficiency, as long as you do not pack the loops closer than about 1/10 wavelength. However, in most pagers and similar designs, the loops are close, and the gain advantage decreases to about 1.5 to 2 dB or less—still impressive considering the low cost. Each loop turn, however, increases the inductance of the antenna, which makes impedance matching more difficult and sensitive to component tolerances.

You must also consider a user's effect on a mobile device's antenna. Capacitive coupling between the user's body or hand and the antenna increases antenna gain by 3 to 6 dB. Unfortunately, using that person as a coupled reflector can be bad news, because the user can arbitrarily vary the orientation of the mobile unit and its antenna. This variation can cause continual changes in signal levels and can affect the relative transmitter- and receiver-antenna signal polarization and the various multipath geometries.

Base station has more options

Your decisions for the base-station antenna are more critical than those for the mobile or remote antenna. You have much more control at the base station in placement, power output, antenna gain, radiation pattern, polarization, intended coverage, receiver sensitivity and performance, and, most important, consistency and repeatability. The base station is usually far less constrained in power output and antenna configurations. Its performance is more predictable than that of the mobile, randomly varying remote unit. Although you can fabricate the remote antenna within your circuitry, you usually purchase a base-station antenna from a vendor after analyzing your needs. Base-station vendors, such as Alpha Industries, Antennaco, Huber+Suhner, and M/A-COM, offer a wide variety of standard products; prices start at approximately $1000, depending on configuration, gain, power-handling capacity, and environmental rug-gedness.

The free-space path loss over the distance you need to cover at the intended frequency of operation shows the loss in signal strength. This loss is due to the inverse-square-law spreading of signal energy from a source that radiates uniformly over a spherical surface. You can determine the free-space path loss as follows:

where f is the frequency in megahertz, d is the distance in kilometers, and c is the speed of light. Expanding the log form of the right side and using 33108 m/sec for c yields:

An example shows how distance severely attenuates the transmitted signal, making the receiver's task difficult. At 1000 MHz, over a distance of 1.5 km (less than 1 mile), the free-space loss is 92.4 dB.

You can estimate and trade-off among the required gain of the antennas, the power output at the base station and remote units, and the minimum receiver sensitivities (see box, "How much gain do I need?"). If the free-space loss means that received-signal strength is too low for the receiver sensitivity, consider which of these factors to increase and by how much. You probably have more flexibility in some parameters than in others; for example, internal noise may make it impractical to increase receiver sensitivity, so increased transmitter power may be an easier, though more costly, choice.

You could also use an antenna with gain at the mobile unit, but, because the remote antenna's orientation is not usually fixed, the antenna cannot easily achieve consistent results. Increasing the remote antenna's gain may be an attractive option. However, any energy aimed in one direction comes at the expense of energy in another, because the total output and total received energy remains unchanged. Therefore, an antenna—with gain and consequent directionality—that points away from the target actually causes a loss, compared with an isotropic antenna.

Although a remote, mobile antenna should be simple, a base-station antenna can be more complex. By putting a simple plane or corner reflector behind the basic dipole, monopole, loop, or patch element, you create an aperture antenna with directional gain. A simple plane, for example, provides forward gain of about 9 dBi (relative to an isotropic antenna) and minimizes signals from the side and rear. The corner reflector provides gain of 12 to 15 dBi, with a sharper forward lobe.

These aperture antennas are relatively simple and effective, combining a single driven element and a passive reflector. (The connected element is called "driven," or "active," even if it's a receiver antenna.) For more complex shaping of the radiation pattern in both azimuth and elevation, you and the base-station-antenna vendor may consider an array. This antenna comprises multiple antenna elements driven with a calculated phase difference between adjacent elements. The constructive and destructive interference among the many radiated wave fronts shapes the overall antenna radiation pattern. By choosing the type, number, arrangement, and relative phasing of these driven elements, base-station vendors can tailor the radiation field to a site's requirements.

You must consider two circuit issues between your active circuitry and the antenna. First, you must make the antenna's impedance appear resistive and, thus, self-resonant. Physical dimensions, geometry, materials, and other mechanical factors determine the antenna's inherent impedance. You usually can't significantly change these factors without disturbing the desired field pattern and performance. The traditional technique for accommodating the antenna impedance is, first, add reactance to compensate for the reactance and, then, transform the remaining resistance to the needed value (Reference 2).

Impedance matching is necessary to keep the VSWR low enough for your application. Relatively low-power mobile units can often accept VSWR values as high as 1.5 or 2, although higher power base-station transmitters usually need VSWRs lower than 1.5 to prevent output-stage damage. Also, consider incorporating output-stage protection at the mobile unit in case the unit operates without its antenna, with a short circuit across the antenna connector, or with the antenna inadvertently touching a conducting surface, significantly changing its impedance. These things happen in the real world, especially with remote units.

You should also filter the RF signal to minimize interference and intermodulation. Fortunately, recent developments provide small, effective options for filtering the RF signal. Consider ceramic and SAW filters (Figure 4). Typical ceramic dielectric-resonator filters, such as the M-Block series from Murata Electronics, offer slightly less insertion loss, 2 dB, than do SAW devices at 3 dB. Ceramic filters typically operate at higher frequencies and are less expensive than SAW filters. In contrast, SAW devices, such as the MF1000 series from Mitsubishi Electronics, typically have sharper filter skirts and, thus, better rejection characteristics and are slightly smaller than ceramic filters. The available technology, though, is changing rapidly, and specifics vary among vendors. Both ceramic or SAW devices provide 25- to 50-dB rejection in surface-mount packages small enough for handheld wireless devices.

Also, consider whether multiple antennas can help your coverage. Using a "diversity arrangement," two antennas spaced at least ½[lambda] apart, increases the chances of at least one's receiving a signal with sufficient S/N ratio and uncorrupted by reflections, multipath, or skewed polarization. It's usually better to continuously sample and select the stronger signal, rather than simply sum the two antenna signals, which requires more complex front-end circuitry and also allows a distorted signal into your signal-processing chain. Fortunately, although impractical for mobile units, diversity antennas are efficient at the base-station receiver, which must capture the relatively fragile signal from the mobile unit.

Many vendors, such as Alpha Industries, Hittite Microwave Corp, M/A-COM, and Mini Circuits Laboratory, offer diversity switches. These GaAs ICs provide configurations ranging from a basic spdt switch, such as Alpha's AS103-12, to switches that support both diversity and transmitting/receiving, such as Mini-Circuits' 2-GHz MSWT-4-20 transfer switch (Figure 5). Look at key specifications, such as insertion loss and isolation, which are typically 0.5 to 2 dB and 25 to 45 dB, respectively, at 1 GHz. You should also factor these specs into your RF signal-level analysis. Some of these switches have built-in drivers; others require external driver circuitry.

Also, check the power levels that a switch handles. Received signals involve a negligible amount of power, but transmission requires a switch that can handle several watts of RF, depending on your design. Some of these switches also require a negative bias supply. Also, heed the effects of interference from nearby antennas. You can somewhat overcome these effects by adjusting the antenna directional characteristics and polarization. However, you may have little flexibility in this area due to conflicting goals.

How much gain do I need?

You can see the interaction and requirements among antenna gains, transmit power, required receive-signal level, distance, and frequency from this basic formula: where GT=transmitter-antenna gain, as a ratio; PR=required receiver power; PT=transmitter power; GR=receiver-antenna gain, as a ratio; R=range, and l=wavelength at operating frequency. Converting Equation 1 to logarithmic form produces the gain in dBi: GT=10 log PR+20 log (4pR)-10 log GR-20 log l. The free-space path analysis applies to line-of-sight propagation; you must correct for various other losses. These losses include absorption as signals pass through walls or earth; lossy reflection from surfaces, such as roadways and buildings; and diffraction, as the radiated signals scatter at the edges of RF-opaque objects, such as buildings. The total absorption varies, and base-station manufacturers provide charts, equations, and to help estimate these losses (Reference 3). These losses typically add as much as 40 dB to free-space value. Add correction terms in decibels from the above equation to account for diffraction, absorption, and other losses. For example, typical loss figures of 10 dB for diffraction, 7 dB for earth absorption, and 20 dB for structural absorption increase the required gain, GT, by 37 dB. In addition, consider additional losses due to misaligned polarization.

You can reach Technical Editor Bill Schweber at (617) 558-4484; fax (617)558-4470; e-mail:bill.schweber@cahners.com

References

1. Lau, Henry, "Performance concerns guide the design of pager antennas," Microwaves and RF, December 1995.
2. ARRL Handbook for the Radio Amateur, American Radio Relay League, 1995.
3. Rosol, George, Guide to Choosing a Wireless Basestation Antenna, M/A-COM Antenna and Cable Division.