Directional beam antennas

THE DIRECTIONAL BEAM ANTENNA DOES SEVERAL JOBS. FIRST, IT PROVIDES AN APPARENT increase in radiated power, because it focuses available transmitter power into a single (or at worst limited) direction.

For this reason, a bidirectional dipole has a gain of approximately 2 dB over an isotropic radiator. Add one or more additional elements, and the focusing becomes nearly unidirectional, which increases the effective radiated power (ERP) even more.

Second, the beam increases the received signal available at the inputs of the receiver. Antennas are generally reciprocal, so they will work for receiving as they do for transmitting. Finally, the directivity of the beam antenna allows the operator to null interfering stations. In fact, it is the last attribute of the beam that is most useful on today’s crowded bands.

All in all, if your funds are too little to provide both increased RF power and a good antenna system, then spend what is available on the antenna—not on the power. In this chapter we will focus on directional antennas that can be built relatively easily.

It is assumed that most readers who want a triband multielement Yagi will prefer to buy a commercial product, rather than build a homebrew model. The material herein concentrates on homebrew projects that are within the reach and capabilities of most readers. The first of these is not a beam antenna at all, but rather a rotatable dipole.

Rotatable dipole 

The dipole is a bidirectional antenna with a figure-8 pattern (when viewed from above). The dipole is a half-wavelength and is usually installed horizontally, although vertical half-wavelength dipoles are known. Although the length of the dipole is too great for rotatability at the lower bands, it is within reason for the higher band. For example, the size of the halfwave dipole is approximately 16 ft on 10 m and 22 ft on 15 m. Even the 33-ft length on 20 m is not unreasonable for amateur constructors. The length of the dipole is found from

This length is approximate because of end effects and other phenomena, so some “cut and try” is required. Example 12-1 Find the length of a dipole antenna for a frequency of 24.930 MHz in the 12-m amateur radio band. Solution:

The half-wave dipole is fed in the center by coaxial cable. Each element of the dipole is one-half of the overall length (or, in the example given, about 9.4 ft). Figure 12-1 shows a rotatable dipole that can be designed for use on 15, 12, and 10 m. The radiator elements are made from 10-ft lengths of 3⁄4-in aluminum tubing. The tubing is mounted on “beehive” standoff insulators, which in turn are mounted on a 4-ft length of 2 2 lumber. 

The lumber should be varnished against weathering. In a real pinch, the elements can be mounted directly to the lumber without the insulators, but this is not the recommended practice. The mast is attached to the 2 2 lumber through any of several means. 

The preferred method is the use of a 1-in pipe flange. These devices are available at hardware stores under the names floor flange and right-angle flange. The 10-ft lengths of pipe are the standard lengths available in hardware stores, so it was selected as being closest to the required 22 ft for 15 m. A 0.14-µH loading coil is used at the center, between the elements, in order to make up for the short length. The dimensions of the coil are 4 to 5 turns, 0.5-in diameter, 4-in length. For low power levels, the coil can be made of no. 10 (or no. 12) solid wire—and, for higher levels, 1⁄8-in copper tubing. There are two basic ways to feed the antenna, and these are shown in details A and B in Fig. 12-1. The traditional method is to connect the coaxial cable (in parallel) across the inductor. This method is shown in Fig. 12-1, detail A. 

A second method is to link couple the coil to the line through a one- to three-turn loop (as needed for impedance matching). This is the method that would be used for a toroidal inductor. Lower frequencies can be accommodated by changing the dimensions of the coil. The coil cannot be scaled, simply because the relative length of the antenna changes as the frequency changes. But it is possible to cut and try by adding turns to the coil, one turn at a time, and remeasuring the resonant frequency. 

Adding inductance to the coil will make the antenna usable on 17 m and 20 m, as well as on 15 m. Another method for building a rotatable dipole for lower frequencies is to increase the element lengths. On 17 m, the overall length is approximately 27.4 ft, so each element length is 13.7 ft long. This length can be achieved by either of two methods. First, adjacent sizes of aluminum tubing are designed so that the smaller will be a slip-fit inside of the larger. What constitutes “adjacent sizes” depends on the wall thickness, but for one common brand, the 7⁄8-in is adjacent to the 3⁄4-in size. 

You can use two smaller lengths to make the larger lengths of pipe, and cut it to size. This method is only available to those readers who have a commercial or industrial metals distributor nearby, because the 16-ft lengths are not generally available from hardware stores. Bands higher than 15 m (i.e., 12 and 10 m) can be accommodated by using the 10-ft lengths of tubing, but without the inductor. The tubing is cut to the desired half-wavelength size and used directly.

360 Degree Directional Array Antenna

The phased vertical antenna concept can be used to provide round-the-compass control of the antenna pattern. Figure 11-5A shows how three quarter-wavelength verticals (arranged in a triangle that is a half-wavelength on each side) can be used to provide either end-fire or broadside patterns from any pair (A-B, A-C, or B-C).

Any given antenna (A, B, or C) will be grounded, fed at 0°, or fed with 180°. The table in Fig. 11-5B shows the relative phasing for each direction that was labelled in Fig. 11-5A. Either manual phase changing or switch-operated phase changing can be used, although the latter is preferred for convenience. Some international showcase broadcasters use antenna arrays formed into two or more concentric circles of vertical elements, with one element at the center. Selection of elements and phasing determines directivity and gain.

Feeding Phased Array Antenna

The second variation, shown in Fig. 11-2B, supposedly produces a 180° phase shift between antenna A and antenna B, when length L3 is an electrical half-wavelength. According to a much-publicized theory, the system of Fig. 11-2B

ought to produce the pattern of Fig. 11-1B—yet experience shows this claim is false. It seems that there are several problems with the system in Fig. 11-2B.

First, coax has a property called velocity factor(VF), which is the fraction of the speed of light at which signals in the cable propagate. The VF is a decimal fraction on the order of 0.66 to 0.90, depending upon the type of coax used. Unfortunately, the physical spacing between A and B is a real half-wavelength (L3 = 492/F), but the cable length is shorter by the velocity factor [L3' = (VF × 492) /F].

Consider an example. A 15-m phased vertical antenna system will have two 11-ft radiators, spaced 22 ft apart (approximately, depending upon exact frequency). If we use foam coax, with VF = 0.80, the cable length is 0.8 × 22 ft, or 17.6 ft. In other words, despite lots of publicity, the cable won’t fit between the towers! Second, the patterns shown in Fig. 11-1 are dependent upon one condition: the antenna currents are equal. If both of them are the same impedance, and are fed from the same transmitter, then it is reasonable to assume that the currents are equal—right? No, wrong! What about coax loss? Because of normal coax loss, which increases at higher frequencies, the power available to antenna Bin Fig. 11-1B is less than the power available to antenna A. Thus, the pattern will be somewhat distorted, because the current produced in Bis less than the current in A, when they should be equal.

The first problem is sometimes fixed by using unequal lengths for cables L1 and L2 (Fig. 11-2A), and using it for the out-of-phase case. For example, if we make L1 one-quarter wavelength and L2 three-quarter wavelength (Fig. 11-2C), antenna A is fed with a 90° phase lag (relative to the tee connector signal), while antenna B is fed with a 270° phase shift. The result is still a 180° phase difference.

Unfortunately, we have not solved the current level problem, and may have actually made it worse by adding still more lossy cable to the system. There is still another problem that is generic to the whole class of phased verticals. Once installed, the pattern is fixed. This problem doesn’t bother most point-topoint commercial stations, or broadcasters, because they tend to transmit in only one direction. But amateurs are likely to need a rotatable pattern. Neither the antennas in Fig. 11-1A nor that in Fig. 11-1B is rotatable without a lot of effort—like changing the coax feeds, or physically digging up the verticals and repositioning them. Fortunately, there is a single solution to all three problems. Figure 11-3 shows a two-port phasing transformer made from a toroidal balun kit. Use the kind of kit that makes a 1:1 balun transformer. Although we are not making a balun, we will need enough wire to make three windings, and that is the normal case for 1:1 baluns. Amidon Associates and others make toroidal balun kits.

Wind the three coils in trifilar style, according to the kit instructions. The dots in Fig. 11-3 show the “sense” of the coils, and they are important for correct phasing; call one end the “dot end” and the other end the “plain end” to keep them separate. If the dot end of the first coil is connected to J3 (and the transmitter), then connect the dot end of the second coil to the 0° output (J1, which goes to antenna A). The third coil is connected to a DPDT RF relay or switch. In the position shown, S1 causes the antennas to be 180° out of phase. In the other position, the “sense” of the third coil is reversed, so the antennas are in phase. Another phasing method is shown in Fig. 11-4. In this scheme, two convenient, but equal, lengths of coaxial cable (L1 and L2) are used to carry RF power to the antennas. One segment (L1) is fed directly from the transmitter’s coaxial cable (L3), while the other is fed from a phasing switch. The phasing switch is used to either by

pass or insert a phase-shifting length of coaxial cable (L4). For 180° phasing use the following equation to find the length (L4):

where L is the length of L4, in feet VF is the velocity factor (a decimal fraction) FMHz is the operating frequency, in megahertz

Some people use a series of switches to select varying amounts of phasing shift from 45° to 270°. Such a switch allows them to select any number of other patterns for special situations.

Directional phased vertical antennas

THE VERTICAL ANTENNA IS A PERENNIAL FAVORITE WITH RADIO COMMUNICATIONS users. The vertical is either praised, or cursed, depending upon the luck of the owner. “DXability” is usually the criterion for judging the antenna’s quality. Some amateurs can’t get out of their backyards with a vertical, and they let everyone within earshot know that such and such a brand is no good. Yet, another person routinely works New Zealand or Australia on 15 m using exactly the same brand of vertical. The proper installation of vertical antennas is dealt with in another chapter, so, for the present, let’s look at another problem attributed to vertical antennas. That problem is that vertical antennas are omnidirectionalin the azimuth aspect; that is, they send out and receive equally well from all directions. Some people moan that this pattern dissipates their power, and gives them a weaker signal “out where it counts” (true). However, the main disadvantage of the omnidirectional pattern is noise (QRN and QRM). “QRN” is natural noise from thunderstorms and other sources. “QRM” is man-made noise, and can consist of other stations or the many assorted forms of electrical filth that pollute the airwaves. All forms of noise, however, have one thing in common: they are directional with respect to the station. In other words, if you could null signals coming from the direction of the noise source (or undesired station), you would be able to hear desired stations much better. A directional antenna performs this task, so let’s look at some vertically polarized directional antennas. Although most amateurs seem to think that the effective radiated power (ERP) increase that the directional antenna gives them is the real reason to own one, the main benefit is actually on receive. Think about it for a moment. With anywhere from 100 to 1500 W available, the increase or decrease in signal strength (due to the directivity of the antenna) results in a minimal difference on the receive end, especially during good DX conditions. If we rotate the directional pattern, to null out interference, then we usually find that the change in our signal strength perceived by the other guy is small; the S meter reading of the desired station is minimally affected; but the amplitude of the interference source is greatly attenuated! The overall effect is an apparent increase in the other guy’s signal, even though the S meter tells a slightly different story. The improvement of signal-to-noise ratio (SNR) is tremendously improved.

Directivity and phasing 

So, how does a vertical antenna owner get the benefit of directivity without the kilobuck investment that a beam or quad costs? The usual solution is to use phased verticals. AM broadcast stations, with more than one tower, are using this type of system (although for different reasons than hams). The idea is to place two or more antennas in close proximity and feed them at specific phase angles to produce a desired radiation pattern. A lot of material is available in the literature on phased vertical antenna systems, and it is far too much to be reproduced here. There are “standard patterns” dating from before World War II that are created with different spacings and different phase angles of feed current. In this chapter, we will consider only one system. Figure 11-1 shows the patterns for a pair of quarter-wavelength vertical antennas spaced a half-wavelength (180°) apart. Without getting into complex phase shifting networks, there are basically two phasings that are easily obtained: 0° (antennas in phase) and 180° (antennas out of phase with each other). When the two antennas (A and B) are fed in phase with equal currents (Fig. 111A), the radiation pattern (shown somewhat idealized here) is a bidirectional figure 8 that is directionally perpendicular to the line of centers between the two antennas; this pattern is called a broadsidepattern. A sharp null exists along the line of centers (A-B). When the antennas are fed out of phase with each other by 180° (Fig. 11-1B), the pattern rotates 90° (a quarter way around the compass) and now exhibits directivity along the line of the centers (A-B); this is the “end fire” pattern. The interference cancelling null is now perpendicular to line A-B. It should be apparent that you can select your directivity by selecting the phase angle of the feed currents in the two antennas. Figure 11-2 shows the two feeding

systems usually cited for in-phase (Fig. 11-2A) and out-of-phase (Fig. 11-2B) systems. Figure 11-2A shows the coax from the transmitter coming to a coax tee connector. From the connector to the antenna feedpoints are two lengths of coax (L1 and L2) that are equal to each other, and identical. Given the variation between coaxial cables, I suspect that it would work better if the two cables were not merely the same length (L1 = L2), but also that they came from the same roll. 

Shortened coil-loaded dipoles

The half-wavelength dipole is too long for some applications where real estate is at a premium. The solution for many operators is to use a coil-loaded shortened dipole such as shown in Fig. 6-13. A shortened dipole (i.e., one which is less than a halfwavelength) is capacitive reactance. There is no reason why the loading coil cannot be any point along the radiator, but in Figs. 6-13A and 6-13B they are placed at 0 percent and 50 percent of the element length, respectively. The reason for this procedure is that it makes the calculation of coil inductances easier, and it also represents the most common practice.

Figure 6-13C shows a table of inductive reactances as a function of the percentage of a half-wavelength, represented by the shortened radiator. It is likely that the percentage figure will be imposed on you by the situation, but the general rule is to pick the largest figure consistent with the available space. For example, suppose you have about 40 ft available for a 40-m antenna that normally needs about 65 ft for a half-wavelength. Because 39 ft is 60 percent of 65 ft, you could use this value as the design point for this antenna. Looking on the chart, a 60 percent antenna with the loading coils at the midpoint of each radiator element wants to see an inductive reactance of 700 Ω. You can rearrange the standard inductive reactance equation (XL = 6.28 FL) to the form

where LµH is the required inductance, in microhenrys F is the frequency, in hertz (Hz) XL is the inductive reactance calculated from the table in Fig. 6-13C.
Example 6-2 Calculate the inductance required for a 60 percent antenna operating on 7.25 MHz. The table requires a reactance of 700 Ωfor a loaded dipole with the coils in the center of each element (Fig. 6-15B). Solution:


The inductance calculated above is approximate, and it might have to be altered by cut-and-try methods. The loaded dipole antenna is a very sharply tuned antenna. Because of this fact, you must either confine operation to one segment of the band, or provide an antenna tuner to compensate for the sharpness of the bandwidth characteristic. However, efficiency drops, markedly, far from resonance even with a transmission line tuner. The function of the tuner is to overcome the bad effects on the transmitter, but it does not alter the basic problem. Only a variable inductor in the antenna will do that trick (at least one commercial loaded dipole once used a motor-driven inductor at the center feedpoint). Figures 6-13D and E show two methods for making a coil-loaded dipole antenna. Figure 6-13D shows a pair of commercially available loading coils especially designed for this purpose. The ones shown here are for 40 m, but other models are also available. The inductor shown in Fig. 6-13E is a section of commercial coil stock connected to a standard end or center insulator. No structural stress is assumed by the coil—all forces are applied to the insulator, which is designed to take it.

Inductance values for other length antennas can be approximated from the graph in Fig. 6-14. This graph contains three curves for coil-loaded, shortened dipoles that are 10, 50, and 90 percent of the normal half-wavelength size. Find the proposed location of the coil, as a percentage of the wire element length, along the horizontal axis. Where the vertical line from that point intersects with one of the three curves, that intersection yields the inductive reactance required (see along vertical axis). Inductances for other overall lengths can be “rough-guessed” by interpolating between the three available curves, and then validated by cut and try.

Reference : Practical Antenna Handbook  - Joseph J. Carr