Wire array antennas

The antennas found in this Section are any of several array designs that use quarter-wavelength, half-wavelength, or other radiator elements arranged in a fixed array. There are several ways to categorize array antennas. A collinear array consists of two or more elements, ranging in length from 0.25 λ to 0.65 λ (some texts say 1l), fed such that the currents in the elements are in phase with each other.

Arrays are also classified according to the directivity: broadside arrays radiate and receive along a line perpendicular to the plane of the antenna, while end-fire arrays radiate and receive off the ends of the elements.

The down side of these types of antennas is that many of them require parallel open-line transmission line (although some can use 300 or 450 ohm twin-lead), and impedance matching stubs. But we can handle that problem.

Section 3 showed what parallel line and the two forms of twin-lead transmission line look like.

Some of the antennas can be fed directly with a 4:1 or 1:1 BALUN transformer and 75ohm coaxial cable, but in many cases an impedance matching stub is needed (see Section 11). The BALUN transformers can either be homemade, or store-bought. Although antenna and scanner sup plies distributors sell very high-frequency (VHF) BALUN transformers, you can also use television-style BALUN transformers. These may not be called BALUNs on the package, but are recognized by the fact that they have a coaxial 'F' connector at one end, and a piece of 300 ohm twin-lead at the other.

There are several antennas that fit either tightly or loosely into the general category of 'array antennas.' The selection of which antennas to put in this Section, and which to put in Section 5(dipoles, doublets), was not entirely clear in each case. This ambiguity was especially annoying when looking at antennas such as the double extended Zepp, which I put in Section 5. The problem is that antennas such as the double extended Zepp are quite properly called 'array' antennas even though they look suspiciously like doublets (which they are, of course). Other antennas were less difficult to place unambiguously in one Section or the other.

FIG. 1

ARRAY ANTENNAS

The idea in an array antenna is to provide directivity and gain by using two or more antenna elements in such a way that their fields combine and interact to focus the signal in one direction, or a limited number of directions (e.g. bidirectional, like a dipole). FIG. 1A shows an array not unlike some of those in this Section. It consists of two half-wavelength dipoles stacked one on top of another a distance (S) apart. The view here is broadside from the horizontal direction, so the dipole fed at A1-A2 is positioned above that which is fed at B1-B2.

One requirement of this form of antenna is that the two half-wavelength dipoles be driven in-phase with each other (some other antenna arrays want out-of-phase excitation). That means that the phase of the signal applied to A1-A2 is exactly the same as the phase of the signal applied to B1-B2. This is accomplished with either of the two methods shown in FIG. 1B and 8.1C. In FIG. 1B, a piece of parallel transmission line is connected between the feed-points such that A1 is connected to B1, and A2 is connected to B2. Another piece of parallel feedline from the receiver or transmitter is connected at the exact mid-point on the harness between the two dipoles.

When this is done correctly, the voltages appearing across A1-A2 are exactly the same amplitude and phase as the voltages across B1-B2.

In the method used in FIG. 1C, the parallel line from the rig or receiver is connected to the feedpoint of one dipole (in this case B1-B2).

In order to preserve the 1808 phase relationship between the elements, the feedline phasing harness between A1-A2 and B1-B2 must be twisted once such that A1 is connected to B2, and A2 is connected to B1.

The gain realized by stacking the two half-wavelength dipoles in this manner is a function of the spacing (S) between the elements. The approximate gain above isotropic that can be realized is shown in FIG. 1D (note: to convert to gain above a dipole, subtract 2.14 dB from each value). Note that the gain peaks at about 6.94 dBi (or 4.8 dBd) in the vicinity of 5j/8 (0.625j). This is the maximum gain realizable from this configuration. Although it looks good, there are sometimes other factors that deteriorate gain performance, so some experimentation with element spacing is usually in order for those who desire peak performance.

The antenna pattern produced by phased array antennas depends on the nature of the elements, the spacing between them and the ratio of the currents flowing in the elements. In the simple case of two elements spaced a half wavelength apart, we will see the general patterns of FIGs 1E and 8.1F. These patterns are azimuthal plots as seen from above; the antennas are omnidirectional vertical radiators. The pattern in FIG. 1E is for the case where the two antenna elements are fed in-phase with each other.

The maxima are perpendicular to a line between the two antennas, while the minima are off the ends. Feeding the antennas out-of-phase with each other ( FIG. 1F) flops the pattern over so that the maxima are along the line between the elements, and the minima are perpendicular to this line.

There are a large number of patterns for various spacings and phase angles. The originals were developed in 1929 by a chap named Brown, and published in the USA in The Proceedings of the Institute of Radio Engineers (now the Institute of Electrical and Electronics Engineers).

Brown's patterns are reproduced in The ARRL Antenna Hand guide and any of several antenna engineering handbooks. Now let us look at some antennas.

TWO-ELEMENT COLLINEAR ARRAY

One of the simplest forms of gain array antenna is the two-element collinear array (shown in FIG. 2 as a side view). It consists of two half-wavelength dipoles spaced a half wavelength apart. This antenna gives a gain of 3 dB, and is a broadside array (when viewed from above the pattern is perpendicular to the line between the antenna elements).

Each antenna element is a half wavelength, but the actual lengths and the equations used to calculate them are a little different for the HF and VHF/ UHF bands.

HF band lengths

On the HF bands (3-30MHz) the antenna is most likely mounted close to the Earth's surface, so there will be some end-effect reduction of the antenna length due to capacitive coupling. There is also a small reduction (2-3%) due to the velocity factor of the signal in the wire (this is set by the length/ diameter ratio of the wire). As a result, the length of the elements is typically reduced 4-5% compared with the free space length. The gap between the dipoles, however, is calculated at the free space value. Thus,

A = 143 F_MHz meters

and

B = 150 F_MHz meters

Typical lengths for popular bands are given in TBL.1.

FIG. 2

TBL.1 Band (MHz) A (m) B (m)

VHF/UHF band lengths

The physical length of the dipoles for VHF can be calculated from the free space version of the equation because these antennas are typically installed a sufficient number of wavelengths above the Earth's surface that the free space constant is reasonable. Thus, both A and B are the same physical length. These lengths are found as A_cm = B_cm = 2324=F_MHz, where F_MHz is the frequency in megahertz. Thus, at 150 MHz, the half wavelength is 15.49 cm; at 450 MHz it’s 5.16 cm; and at 800 MHz it’s 2.91 cm.

The antenna in FIG. 2 is a little inconvenient in one respect: it requires two transmission lines connected in parallel. This is easily accomplished by using 75ohm coaxial cable from each antenna element to a 'tee' connector, and then a piece of 52 ohm cable from the 'tee' connector to the receiver. At HF frequencies this antenna can take up a lot of 'footprint' space.

FIG. 3

FIG. 4

A different solution is shown in FIG. 3. The antenna elements are separated by only a few centimeters, rather than a half wavelength. The gain drops to 1.6 dB for two elements (although for three elements it’s 3 dB, and for four elements it’s 4.2 dB). The antenna is fed by 450-600 ohm parallel line or twin-lead. The feedpoint impedance will be as low as 1 kilo-ohm if tubing is used, and up to 4-6 kilo-ohms if wire is used. This is one of those antennas where a matching stub is needed.

The four-element version of the antenna is shown in FIG. 4 (it is called the collinear array or Franklin array). This antenna is similar to the two-element version of FIG. 3, but with an additional half-wavelength element tacked onto the ends. The additional elements are separated by quarter-wavelength phase reversal stubs. The stubs are needed to keep the phase of the currents flowing in the outer elements the same as the currents flowing in the center elements. These stubs are made from the same type of transmission line as the half-wavelength matching stubs, but are half as long.

Another collinear array is shown in FIG. 5. This antenna offers 3 dB of gain, and can be fed directly from 75ohm coaxial cable through a 4:1 BALUN transformer. The antenna consists of three half-wavelength elements, but the center element is fed at the center point (a quarter wavelength from each end). The quarter-wavelength phase reversal stubs are used between the center element and the two outer elements.

FOUR-ELEMENT BROADSIDE ARRAY

A four-element vertical broadside array is shown in FIG. 6. Although often implemented using aluminum tubing, the array can easily be erected using wire elements as well. This antenna has four half-wavelength elements spaced a half wavelength from each other. Like the antenna above, the radiation direction is perpendicular to the line between the antennas, or in and out of the page.

FIG. 5

The antenna elements are fed by a length of 600 ohm parallel line that connects all four together. But notice that the outside elements are fed by reversing the phase. This means that the transmission line must be twisted before being connected to the end elements; the unused conductor is left floating. The feedpoint impedance of this antenna is 200-300 ohms, so it can be fed with either 52 or 75 ohm coaxial cable if a 4:1 BALUN transformer is provided. Gains of the order of 6-8 dB are realized with this antenna.

FIG. 6

EIGHT-ELEMENT BROADSIDE ARRAY

The eight-element broadside array shown in FIG. 7 is built using similar methods to the previous antenna, but with two bays of elements rather than just one. The overall height of the array is one wavelength (although half wavelength versions can also be built at some cost to gain). As in the previous case, the elements are spaced a half wavelength apart and are fed at their bases through 600 ohm parallel line that is twisted at the ends. Rather than breaking one conductor to connect the feed, this antenna connects the BALUN transformer across the antenna phasing harness wires. Note also that a 1:1 BALUN is used here, rather than a 4:1 device as above. This antenna is capable of as much as 12 dB gain.

THE LAZY-H ARRAY

The lazy-H array is a special case of the doublet array in which the doublets making up the array are one wavelength long ( FIG. 8A), fed in-phase with each other. The directivity of this antenna is broadside, and so radiates in two directions into and out of the page. This view is as seen from the front or back, so note that the two one-wavelength doublets are stacked one on top of the other.

Gains for the lazy-H antenna vary with the spacing between the doublets (S), and will be as low as 4 dB at 3j/8 to almost 7 dB at 3j/4 spacing. The antenna should be installed at least a quarter wavelength above the Earth's surface.

The two doublets in the lazy-H must be fed in-phase with each other.

When the feed arrangement is as shown in FIG. 8A, i.e. with the feed point in the center, the transmission line phasing harness is connected straight across (no twisting). If only single-band operation is anticipated, then the feedpoint impedance will be a reasonable match to 75ohm coaxial cable; a 1:1 BALUN transformer is used to connect the coaxial cable to the phasing harness. For multiband operation, however, it’s necessary to use tuned feeders to an antenna tuning unit.

FIG. 7

FIG. 8

An alternative feed scheme is shown in FIG. 8B. In this arrangement, the feedpoint is at the feedpoint of one of the doublets, rather than in the center of the harness. In order to preserve the in-phase relationship, the harness must be twisted once as shown. The feedpoint impedance is around 3000 ohms, so a quarter-wavelength matching stub is needed in order to effect a reasonable match to the coaxial transmission line. Again, for multi band operation tuned feeders are advisable.

CURTAIN ARRAYS

The curtain arrays in this section are variations of the broadside arrays. The antenna shown in FIG. 9 is a small Sterba curtain array, larger versions of which are often used by high-power international short-wave broadcasters. It can be fed with either 300 ohm twin-lead, or with 75ohm coaxial cable if a 4:1 BALUN transformer is provided at the feedpoint.

The Sterba curtain antenna can be built of wire or aluminum tubing, although the wire option is probably the most popular at HF, and tubing at VHF. An advantage of these antennas is that they can be built for frequencies in the 6-7 MHz range (where wire construction is preferred), if you have enough room, and also well into the VHF region (in which case aluminum tubing construction is preferred).

The signal is bidirectional, and is broadside to the array (in and out of the page). Gain is of the order of 6 dB, although it rises with added numbers of sections.

FIG. 9

The antenna in FIG. 10 (our micro-Sterba) uses elements of two different lengths (j/2 and j/4), labeled A and B. These lengths can be calculated from

A = 149 F_MHz meters B = A 2

At 16 MHz, these lengths work out to be A = 9:31m, B = 4:66m; at 162 MHz, A = 0:92m and B = 0:46m.

The horizontal distance between horizontal elements should be 10.2- 15cm. In wire antennas, an ordinary end insulator placed between two elements will usually suffice.

FIG. 10

FIG. 12

The Sterba curtain concept can be extended by adding sections, which of course also increases gain. The version shown in FIG. 10 provides gain up to 8 dB, and is basically a pair of the previous antennas back to back.

The feedpoint is matched using a quarter wavelength stub and a 1:1 BALUN transformer. As with the other antennas, this is a broadside array.

The multi-section Sterba curtain array shown in FIG. 11 uses five loops, and is fed at the end with 300 ohm twin-lead (or a 4:1 BALUN connected to 75-ohm coaxial cable). Gain will be up to 9 dB or so, but at the cost of a lot of space (each A-section is a half wavelength long).

The Bruce array is shown in FIG. 12. This type of array is built using a long wire folded and fed in such a manner that the current nodes (i.e. points of high current) are in the centers of the vertical elements, while the voltage nodes (points of low current) are in the centers of the horizontal elements. About 9 dB can be achieved with this antenna.

A variant Bruce array is shown in FIG. 13. This antenna consists of two cross-connected sections, each of which are made of quarter-wavelength elements. The feedline is connected at the points marked X1-X2. The impedance at this point is around 700-800 ohms. If a 9:1 BALUN transformer is used, then this point can be fed with 75ohm coaxial cable with only a small voltage standing wave ratio (VSWR), assuming single-band operation. It’s also possible to feed this antenna on a single band with a quarter-wave length matching stub, along with a 1:1 BALUN transformer. If you want to operate on two harmonically related bands, however, tuned feeders and an antenna tuner are required.

FIG. 13

'SIX-SHOOTER' ARRAY

The array shown in FIG. 14A is sometimes called the 'six-shooter' because it consists of six half-wavelength elements. In structure it’s much like the Sterba curtain shown earlier, with the end elements open circuited rather than connected together. This antenna is capable of gains between 7 and 8 dB.

The lengths of the elements in FIG. 14A are calculated as shown in the figure. The horizontal elements are each calculated from:

L_meters = 145

F_MHz meters

FIG. 14

While the vertical separation between the two rows of horizontal elements is calculated from

L_meters = 296 F_MHz meters

These antennas are relatively easy to construct of either wire or tubing, and should be considered whenever you want gain on the cheap.

Like other broadside arrays, the 'six-shooter' is bidirectional, and sends or receives signals in and out of the page as you view the drawing. One variant that I saw on a 10 m amateur radio band 'six-shooter' is shown in FIG. 14B. In this version, a screen is placed a quarter wavelength behind the antenna. Signals traveling toward the screen from the antenna are reflected back toward the antenna element. Because the spacing is a quarter wavelength, the signal reflected arrives back at the antenna in phase, so reinforces the signal going in the other direction. A gain improvement of 3 dB over the gain of the unscreened version is realized.

The screen can be almost any form of reflective plane. Although one could use sheets of copper or aluminum, this approach is not the best for most purposes. The reason is that the reflector then has a tremendous 'sail area' and so will be susceptible to wind problems. A better solution might be to use metal window screening (if you can still get it), or 'chicken wire' (i.e. the kind of wire used to screen chicken coops). Alternatively, one could also create an effective screen by connecting wires in a square matrix, provided that the gaps are only a small fraction of a wavelength (i.e. _j/8).

THE BOBTAIL CURTAIN ARRAY

The bobtail curtain ( FIG. 15) is something of a favorite in the HF short wave bands (with both listeners and ham operators) when very low angles of radiation are required. A low angle is the desired condition for maximizing the distance received by any given antenna (a good 'DX' antenna has an angle of radiation that barely kisses the horizon). The bobtail curtain fits this requirement rather nicely. Although it’s easiest to construct for the upper HF bands, it’s not beyond reason for many sites for 75/80m band amateur radio operation or 60m band short-wave activity.

The bobtail curtain consists of three quarter-wavelength radiators (B) connected at their tops by half-wavelength sections (A). Because the antennas are a quarter wavelength, the proper lengths are of the order of 19.82- 21.34m for the 75/80m band and 14.63m for the 60m band. For the 13m band, on the other hand, the lengths of the vertical radiators are only 2.9- 3.05m. The connecting half-wavelength sections are twice these lengths.

Note that the current which is injected into the center section must split at the top, and only one-half of the line current of the center radiator flows in each side radiator. This results in a binomial current distribution between the elements.

The gain of the bobtail curtain is 7-10 dB, with a figure '8' pattern broadside to the antenna. In trials some years ago, I witnessed a well made bobtail curtain that only exhibited 5-6 dB (keeping in mind the difficulty of measuring antenna gain in situ), but even 5dB is relatively decent compared to a dipole (nearly an S-unit on conservatively specified receivers). The directivity of the bobtail curtain is better than the dipole, but not quite as good as a set of three vertical radiators fed in-phase with each other (which will produce nearly the same pattern).

The bobtail curtain is fed at the base of the center element through a parallel tuned LC tank circuit used for impedance matching. The coaxial cable is connected to either a tap on the inductor (as shown) or to a link or primary winding (a couple of turns of wire) to the main inductor. The values of the inductor and capacitor are set such that the reactances are equal to 1100 ohms. For operation up to 20MHz, a 50 pF variable capacitor will normally be satisfactory, although at higher frequencies the minimum capacitance of the typical 50 pF unit may be too high for the required capacitance (only a few picofarads). The inductor will vary from about 10 to 50 mH. The various inductances can be provided by either a variable inductor, or a fixed inductor with taps at the required inductance points.

FIG. 15

THE THORNE ARRAY BOBTAIL CURTAIN ANTENNA

The Thorne Array Bobtail Curtain (TABC), shown in FIG. 16, is basically a standard bobtail curtain array turned upside down, and fed in a different manner. Bidirectional gains up to 5dB have been measured on the 15m band, although I suspect that the gain is due largely to compression of the elevation lobe of the antenna. This antenna has a very low angle of radiation, so works well for long-DX in the upper end of the short-wave spectrum. On the 15m ham band, from a location in Texas, I have worked very loud Australian and New Zealand stations that were a lot less audible on the quarter-wavelength vertical and dipole antennas at the same location (and received much better signal reports of my own signal).

The TABC consists of three vertical quarter-wavelength (j/4) radiators (B) separated by half-wavelength (j/2) phasing elements (A). The two outer vertical elements are electrically connected to the corresponding phasing element, but insulated from the center vertical element. The two phasing elements are electrically connected to each other by a very short jumper wire. The center conductor of the 52 ohm coaxial cable to the receiver is connected to the base of the center vertical element, while the coaxial shield is connected to the jumper between phasing elements, and to ground.

The lengths of the two sections are A = 91=F_MHz meters and B = 46=F_MHz meters. For the 24 MHz band, the horizontal or phasing sections are 3.64 meters long while the vertical sections are 1.84 meters long.

Mechanically, the antenna is basically supported by dipole-like antenna end insulators and ropes to various masts or support structures (trees, buildings, etc.). If you prefer, the vertical sections can be replaced with 12_50mm aluminum tubing, but you will have to provide some form of mounting for the tubing.

FIG. 16

FOLDED DIPOLE 'X-ARRAY'

The array shown in FIG. 17 uses four folded dipole antennas arranged in an 'X-array' pattern to provide about 6 dB gain. The antenna elements are made from 300 ohm twin-lead. The ends of the twin-lead conductors are shorted together in this type of antenna. The arrangement is two elements in each of two bays. In each bay, the inner ends of the folded dipole elements are spaced 0.2 λ apart, while the two bays are spaced 1.25 λ apart. The transmission line harnesses (P) are 0.45 λ long. These transmission line harnesses are connected in phase with each other at the high impedance side of a 4:1 BALUN transformer; 75ohm coaxial cable carries the signal to the receiver. All four elements in FIG. 17 are in the same plane, and the direction of radiation/reception is perpendicular to that plane (i.e. in/out of the page).

PHASED VERTICALS

FIG. 17

FIG. 18 shows the basic two-element phased vertical array. This antenna uses two quarter-wavelength wire radiators erected in a vertical manner, spaced a half wavelength apart. Ropes or heavy twine can be used to sup port the antenna elements. The antenna element lengths are found from L_meters = 75=F_MHz, while the spacing between them is D_meters = 150=F_MHz.

There are two ways to feed this antenna, and both use a coaxial cable 'tee' connector to split the signal from the rig between the two antennas. If you want to feed the antennas in phase, then the two segments of cable from the 'tee' to the antenna elements are equal lengths. But if you want to feed the elements 1808 out of phase, then make one length j/4 and the other 3j/4. In the case of in-phase feed, the directivity and gain are in and out of the page, while when the antennas are in phase the directivity and gain are bidirectional on a line between the two antennas (see again FIGs 1E and 1F).

FIG. 18

W8JK ARRAY

The W8JK array ( FIG. 19) is sometimes called the 'flat-top' beam. It’s a bidirectional, end-fire array that consists of two half-wavelength in-phase collinear doublets, in parallel and fed 1808 out of phase with each other, spaced between j/8 and j/4 apart. Each element in each collinear doublet is j/4 long (dimension A). The view in FIG. 20 is from above, so the plane of the dipoles is horizontal to the Earth's surface. The gain will be from 5.7 dBi at j/4 spacing to 6.2 dBi at j/8 spacing. Note that the spacing between the elements of the same dipole is of the order of

C = 8:5 F_MHz meters

The length of each wire element in the dipoles is found from:

A = 71:5 F_MHz meters

The feedline phasing harness between the center-feed points on the two dipoles is made of 450-600 ohm parallel transmission line (or 450 ohm twin lead). The parallel feedline to the transmitter or receiver is attached to the phasing harness at points X1-X2, at a distance (B) of about S/2.

The impedance at the feedpoint (X1-X2) is several thousand ohms, so will necessarily create a rather high VSWR. An impedance matching stub can be provided at the feedpoint to improve the match to the transmission line. However, if tuned feeders are used then this antenna is capable of multiband (harmonically related) operation.

FIG. 19

FIG. 20

An end-fed version of the W8JK array is shown in FIG. 20. In this case, the C-dimension is the same as calculated above, but the A-dimension is not quite twice as long:

A = 110 F_MHz meters

As with the previous version, this antenna can be fed with a parallel tuned feeder, or with a matched line if an impedance matching stub is provided.

STACKED COLLINEAR ARRAYS

The collinear array antenna can be stacked as in FIG. 21. This antenna consists of two collinear arrays (similar to FIG. 5). The lengths of A, B and C are found from:

A = 144 F_MHz B = 73 F_MHz C = 150 F_MHz

…where A, B and C are in meters, and F_MHz is frequency in megahertz.

The two antennas are shown stacked vertically, although they can be stacked horizontally as well (although with a different pattern).

FIG. 21

LARGE REFLECTOR ARRAY ANTENNA

FIG. 22

FIG. 22 shows a large-scale reflector array. It consists of a group of sixteen half-wavelength elements stacked vertically and horizontally. The elements are cross-connected with phasing harnesses. The spacing between elements is half wavelength, although the formula used to calculate this half wavelength is different from the half wavelength of the wire elements:

Wire elements:

= 143 F_MHz

Spacing:

= 150 F_MHz

The reflector screen can be made of chicken wire or any other screening material in which the holes are not more than about _=12 in size. It’s sized at least _=8 larger than the outside dimensions of the array itself. The reflector screen is located a quarter wavelength behind the antenna array elements. The antenna array elements are made of either thin (#10 to #18) or thick wire (#000 to #10).

Prev. | Next

top of page   Home