De-Mystifying the Modern Dipole Curtain Array

L. B. Cebik, W4RNL (SK)

Long-wire antennas served primarily the needs of point-to-point HF communications in the first half of the 20th century. Although some rhombics remained in service within the short-wave broadcast (SWBC) industry, other antenna designs generally took over. SWBC tends to require a broader beamwidth than a rhombic provides. Although the rhombic had the frequency range necessary for frequency shifts in accord with changing HF skip conditions, other antennas could serve as well--or almost as well. Once aimed, the rhombic had a line of targets; SW broadcasters preferred a large region. Even if the target did not encompass the entire region, slewing the antenna's beam pattern could reduce costs by avoiding the need for second and third large high-gain arrays or complex turning mechanisms.

Antique and Modern Billboard Antennas

The solution to the needs of many SW broadcasters arrived with improvements to a very old antenna, once called the billboard. (See Kraus, Antennas, 2nd Ed., p. 547, for a representation of a billboard antenna.) The operational principle is simple. Any bi-directional antenna, such as a dipole, becomes a directional antenna when placed in front of a planar reflector. Planar reflectors find many contemporary uses in the VHF and UHF region today. Hence, we often overlook their continuing service for SW broadcasting. However, their current use depended upon a number of advances, standardizations, and combinatory techniques to give them the relative predominance that they now enjoy.

Fig. 1 sketches (with many missing details) an antenna about which I have received many inquiries. Vacationers encounter them in unexpected places from coast to coast (and well inland) in the U.S. The sketch is not to scale. The towers are much too fat for the array between them. The figure also lists other missing details that would obscure the main function of the antenna. For example, we would normally find many more guy wires for the towers and many more support and spacing wires and jigs for the key elements that form the antenna's radiation pattern.

In return for omitting some details, we can clearly see both the dipole elements in a 3-by-3 array and the reflective screen behind them. In many cases, the screen will consist only of horizontal wires, similar to the rod-based planar reflectors in my notes on that subject in past articles. Since the horizontal lines are very long, periodic vertical spacers are necessary to maintain the reflector shape.

The 3-by-3 array of dipoles also represents an evolution from some of the original billboard antennas that used center-fed full-wavelength elements. Note also that the dipoles are closely spaced. The overall reduction in billboard size per driver unit formed one of many reason why the modern dipole array is a primary short-wave broadcast antenna today. Even the original versions effected a major real estate saving over designs such as the rhombic. Of course, many of the original rhombics used timber supports. As steel tower structures became more common and less expensive, they not only replaced existing supports, but also made taller antennas more feasible. Hence, the billboard antenna traded vertical space for longitudinal space, cutting both purchase and maintenance costs. (Never volunteer to mow the lawn beneath a truly long rhombic.)

The dipole array rarely uses its original "billboard" name, although many folks call it a dipole curtain antenna. "Curtain" refers to the planar reflector behind the driven elements. They could move a bit in the wind. Early designs were not fully appreciated for several reasons. First, the high steel structures and copper wire were subject to corrosion. Breakage required more repair effort than splicing a rhombic leg. However, one of the electrical limitations of the billboard was its narrow operating bandwidth.

In the first half of the last (20th) century, almost all antenna designers strove to produce as much gain as might be feasible from a given design. This bad habit still infects much of the antenna design for amateur radio. We accept excessive problems in feedpoint matching by designing long-boom Yagis with the minimum number of elements necessary for a certain gain level. Even if we overcome that problem, we continue to accept relatively poor sidelobe suppression because we refuse to add a few more elements to the design. We continue to make excuses for antenna designs that are difficult to replicate due to their narrow operating bandwidth. (There are good reasons in certain circumstances for using a narrow beamwidth, but in general, it is usually a condition with which we are stuck for lack of design imagination.)

Early billboard antennas suffered from narrow operating bandwidth for several reasons. First, the driving elements used a spacing from the reflector screen that yielded maximum gain. Second, they looked for element-to-reflector spacings that left the feedpoint impedance unchanged relative to the same driver with no reflector. Third, they used driven element lengths and spacings that yielded maximum gain. For example, a collinear pair of 1/2-wavelength elements (or a center-fed full-wavelength element) yields a little more gain than a simple 1/2-wavelength dipole. (The high impedance of this type of element, of course, permitted the use of wide-spaced transmission line segment for feeding and phasing, a condition very suitable for high-power SWBC operations.) Although we knew that we might obtain even more gain with a vertical spacing of 5/8 wavelength, 1/2-wavelength became the standard for the ease of feeding a vertical collection of elements in phase. Rarely did we have the room to arrange the elements horizontally at optimal spacing. Initially, we used some very close spacing to reflector screens, sometimes as low as 1/8 wavelength. When we discovered that a wider spacing would yield more gain and weaker rear lobes, we opted to use that spacing despite the fact it still limited the operating bandwidth.

Modern dipole curtain arrays operate on other principles. Some common ones, found in Chapter 26 of Johnson's Radio Engineering Handbook, 3rd Ed., reappear in Fig. 2. We may note in passing that an engineer for TCI, a leading producer of dipole curtain arrays, wrote the 3rd edition version of the chapter on HF antennas. If the volume is not conveniently available, you may find some of the same data at the TCI website. Look at model 611 for a general description of their dipole curtain arrays.

The side view of the antenna shows the vertical heights generally used: 1/2 wavelength between dipoles of the array. Studies of planar reflectors strongly suggest that this antenna type achieves maximum gain for a given driver set when the reflector screen exceeds the driver assembly by 1/2 wavelength or so in every direction. Realities, including catenary effects on an all-wire assembly, usually dictate less reflector extension except perhaps at corners.

The face view shows the equally desirable horizontal reflector extension, although every extra foot of reflector screen adds to costs for perhaps marginal performance improvements. The most notable feature of the face view is the arrangement of the driver dipoles. Since a driven dipole is normally slightly less than a physical half wavelength, we may place the dipoles on 1/2-wavelength centers across the reflector. Because designers still wish to use wide-spaced transmission lines for feeding and phasing, the driven elements are usually some form (in some cases, an exotic form) of a folded dipole.

The final element to note from the sketch is the recommended spacing of the elements from the reflector screen: 0.3 wavelength. A simple dipole tended to show maximum gain and weakest rearward lobes with considerably closer spacing, but by accepting a lower gain per driver, the designer achieves a wider operating bandwidth. Before we close these notes, we shall look at the combination of ingredients that go into extending the operating bandwidth of a dipole array.

We (but not necessarily the designers) might express the overall goal in this manner: since wire elements are relatively light, we can obtain more performance by packing more elements within the available space rather than from seeking out the maximum performance from the minimum number of elements. Most of the array weight (but not necessarily stress in adverse weather) lies in the reflector lines or screen. The element spacings in the sketch--and any extension of the sketch--provide the most performance for a given space (side-to-side and vertical) occupied by the array. Performance here includes not only gain, but beamwidth, rear lobes, and feedline SWR for some specified reference impedance.

How the Dipole Array Achieves Its Performance

Let's back up a step and see how the modern dipole array achieves its performance. That step requires that we first examine dipoles on their own, that is, with no reflector screen. We shall survey in tabular form the maximum gain of various combinations of dipoles. Of course, the listed gain will be for a bi-directional array. We shall designate each combination by a code of the order mV-nH, indicating the number of dipoles stacked vertically (m) and horizontally (n). Each vertical dipole will be 1/2 wavelength from its neighbor, and horizontal dipole lines will be on 1/2-wavelength centers. The data include both the gain and the horizontal beamwidth. More correctly, the beamwidth is in the E-plane, since all values for this exercise are for free space. All dipoles consist of folded dipole made from AWG #10 copper wire. The test frequency is 10 MHz.

Free-Space Performance of Various Dipole Arrays

Array Size 1V-1H 1V-2H 1V-3H
Maximum Gain (dBi) 2.13 3.79 5.30
Beamwidth (degrees) 78.4 48.2 33.2

Array Size 2V-1H 2V-2H 2V-3H
Maximum Gain (dBi) 5.94 8.01 9.69
Beamwidth (degrees) 78.5 48.2 32.8

Array Size 3V-1H 3V-2H 3V-3H
Maximum Gain (dBi) 7.80 9.72 11.27
Beamwidth (degrees) 78.4 48.0 32.8

The 2V-2H configuration offers the greatest step-gain increase over versions with one less vertical or one less horizontal dipole. The gain steps are not smooth for two reasons. First, gain increases diminish as we add steps in a linear count. As well, the dipoles interact, so that gain is not strictly additive. Slightly different spacing values or even horizontal end-to-end distances may alter some of the numbers. Nevertheless, the overall progression of dipole maximum gain values is a fair representation of the potentials of dipole arrays on 1/2-wavelength centers.

One interesting fact about the progressions is that the E-plane beamwidth does not significantly change as we add dipoles vertically. Narrower beamwidths result from adding dipoles horizontally. Since the beamwidths of each level of horizontal stacking are constant, regardless of the size of the vertical stack, we can represent the array patterns with samples taken with a vertical stack of 2. Fig. 3 shows the pattern shapes for 1, 2, and 3 horizontal dipole stacks. The patterns for 1 and 2 horizontal dipoles are perfectly normal and familiar. The pattern for 3 dipoles resembles the pattern for a center-fed 1.25-wavelength extended double Zepp doublet.

Over ground, the E-plane patterns would not change shape significantly. The elevation pattern depends upon the height of the array above ground. Typically, an installation will adjust the dominant elevation angle for a design frequency by adjusting the bottom height for the array selected, which might be still larger than the samples used here. Although literature tends to use the average height of the array as a calculating point, the arrays equivalent height tends to be about 2/3 the distance between the height of the lower dipole and the height of the highest dipole. This figure does not vary much from the array's average height, but it does show up in vertically phased arrays (like the lazy-H) where the lowest height may be a large fraction of a wavelength above ground. Indeed, the lazy-H is a billboard antenna without the billboard, although some amateurs have added screen reflectors for increased directivity.

Adding a screen to the folded-dipole arrays that we have just surveyed creates a directive beam antenna. As shown in Fig. 2, the recommended spacing between the dipole arrays and the screen reflector is about 0.3 wavelength. The value is not optimal for maximum possible gain. In fact, designing a dipole array for maximum possible gain would require customizing each dipole element and the array spacing for every possible combination. The 0.3-wavelength spacing provides good gain and pattern shaping without regard to customizing the dipoles to account for their interaction. As a test, I created a screen for each folded dipole in the first sequence. Each screen consisted of a wire grid of standard modeling proportions (0.1-wavelength squares with a wire diameter that is the square side divided by PI). Each screen exceeds the dipole array dimensions both vertically and horizontally by 0.5 wavelength. The test frequency remains 10 MHz. To the data in the first table, I have added the 180-degree front-to-back ratio as a measure of rearward performance.

Free-Space Performance of Various Dipole Arrays with Screen Reflectors

Array Size 1V-1H 1V-2H 1V-3H
Maximum Gain (dBi) 7.45 8.68 10.11
Front-to-Back Ratio (dB) 19.24 21.18 21.78
Beamwidth (degrees) 69.6 48.8 33.0

Array Size 2V-1H 2V-2H 2V-3H
Maximum Gain (dBi) 9.77 11.21 12.85
Front-to-Back Ratio (dB) 21.44 28.73 28.90
Beamwidth (degrees) 72.4 48.8 32.8

Array Size 3V-1H 3V-2H 3V-3H
Maximum Gain (dBi) 11.72 13.22 14.87
Front-to-Back Ratio (dB) 21.58 29.33 29.34
Beamwidth (degrees) 72.0 48.6 32.8

Because the rear lobe structure changes, the gallery of E-plane patterns in Fig. 4 includes plots for all of the entries in the table.

A number of features of the patterns call for note. The beamwidth values do not change very much from the values without the reflector, except that they apply only to the single large forward lobe. The sidelobes of the versions with 3 horizontal dipoles are better than 20 dB lower than the main lobe regardless of the vertical stack size. A single bay consisting of 1 to 3 dipoles arranged either vertically or horizontally has a good front-to-back ratio. However, as soon as we add a second bay in one or the other direction, the ratio approaches 30 dB--even for the 2V-2H version of the array. For values over average ground with a base height of at least 1/2 wavelength, you may add about 5 dB to the gain for a ballpark total gain figure. The gain will slowly climb as we increase the base height of the array, of course, moving the screen upward with the dipoles.

One advantage that accrues to the dipole array is the ability to shift or slew the main direction of the beam by up to 30 degrees each way, depending on array size. Common installations employ "delay lines" that s hift the phase angle of the current for each vertical bay of dipoles. We may simulate this effect in models simply by using a current source and adjusting the source phase angle while holding the current magnitude constant. Fig. 5 shows the patterns for a 1V-2H array initially with both vertical dipoles in phase. The center pattern uses a phase angle of 30 degrees for the first dipole and 60 degrees for the second. The final pattern uses 60 degrees for the first dipole and 120 degrees for the second. The general rule is to change the phase angle of subsequent vertical dipole bays by a multiplier on the baseline phase angle according to the position of the dipole (or vertical bay of dipoles) relative to the first vertical dipole or bay.

The first move changed the heading of the main beam by 7 degrees, and the second changed it by 13 degrees. The angles would remain the same regardless of the size of the vertical bays in the array. By the correct selection of delay lines, we can achieve a relatively precise aim at a target of choice within the span of allowable slew angles. As we increase the angle of the main beam by these means, some distortion does appear in the form of forward and rearward sidelobes. At the angles in the sample, the distortions are not severe enough to void the use of slewing. However, they show that slewing has limits. Nevertheless, for a SWBC station that wishes to change its target from one session to the next, the process allows the change without physically altering the antenna or its position. Note that, when used within limits, the beam strength and beamwidth do not change to any noticeable degree.

The basic capabilities of a fixed position dipole curtain array are quite impressive, even using the ubiquitous amateur monoband Yagi as a standard of comparison. A 2V-2H assembly at a reasonable height above ground would easily match a 5-element Yagi, and delay-line slewing of the beam would permit coverage of all of Europe from the eastern U.S without need for a rotator. If we built equivalent dipole arrays on each side of the reflector, then we might cover Europe on one side and the Pacific on the other, at least from my location in the hills of Tennessee.

Broadbanding Techniques

The needs of SWBC stations are quite different from those of the average amateur station. SWBC stations tend to use very high power levels, up to 500 kW in some cases. Since we must provide energy to each dipole, the use of wide-spaced parallel transmission is fairly standard, indicating as well the use of high-impedance antenna feedpoints. A folded dipole of conventional construction--with equal diameter conductors throughout--goes part of the way toward the high-impedance goal. However, if we wish to raise the feedpoint impedance beyond about 280 Ohms, we must resort to more unconventional techniques. For example, if we use a smaller diameter wire for the line with the feedpoint and a much larger diameter wire for the other line, we increase the impedance transformation to almost any desired level within the limits of lines to match it. We may simulate very wide second wires using pairs or cages of wires so that the entire assembly remains lighter than it would be with a single fat wire or tube for the second conductor.

Obtaining a high impedance feedpoint does not resolve a second goal of dipole array designers: achieving a wide operating bandwidth. The gain of a dipole array changes slowly as we change the operating frequency as a function of the length of the elements relative to the operating frequency. However, being able to match the array over an extended bandwidth requires a combination of techniques. There is no magic to any of them, although amateurs rarely use them in complex combinations.

The first step is to begin with a wide-band folded dipole. The AWG #10 folded dipole used in our initial dipole array models has a 2:1 SWR bandwidth that runs from 9.6 to 10.5 MHz, a 0.9-MHz spread (given our test frequency of 10 MHz). We need to begin with a folded dipole array that has inherently a broader operating bandwidth. That is step 1 in the process. Most dipole array manufacturers have proprietary designs for their driven elements, designs to which I am not privy. (Even if I had access to one or more of them, I likely could not violate agreements that gave me such access.) So I shall begin with a moderately broadbanded driver of my own design. It will not have the full capability of some commercial driver elements, but it will be sufficient for our small demonstration.

A fan dipole with a 3:1 length-to-height ratio is capable of increasing the operating bandwidth over a conventional linear dipole. However, the feedpoint impedance is about 50 Ohms. If we create a pair of such fans, we only achieve the standard 4:1 impedance increase that is standard for a conventional folded dipole. However, if we use a single-wire as the fed portion of the folded dipole, the fan represents a much "fatter" second portion, for a significant increase in the feedpoint impedance. We connect the fed wire to the fan at the centers of the vertical sections, since that is the pair of points on the fan with minimum current. The free-space performance data for the arrangement is virtually identical to the standard folded dipole, but the feedpoint impedance at resonance is over 550 Ohms. A 600-Ohm SWR curves shows under 2:1 SWR from 9.4 to 10.8 MHz, a 1.4 MHz spread or about 1.56 times the spread of the standard folded dipole. Like the standard folded dipole, the folded-fan dipole is composed of AWG #10 copper wire in all of the models that we shall consider. Fig. 7 overlays the SWR plots for the standard folded dipole and the folded-fan dipole. Each curve uses its own reference impedance.

Step 2 in the process of broadening the operating bandwidth is to place the driver assembly ahead of the reflective screen and determine the best distance between the two. Like previous screens, the wire-grid structures used in this sample situation extend about 1/2 wavelength beyond the driver limits in all directions. Fig. 8 shows side and face views of the folded-fan dipole and its screen. In the EZNEC graphic, I have retained the segment and wire junctions dots to lend some color differentiation to the array pieces.

Fig. 8 shows no spacing value because I examined 2 cases. The first placed the driver 0.245 wavelength ahead of the screen. The free-space gain was 8.34 dBi with a front-to-back ratio of 18.77 dB. The feedpoint impedance with this spacing was close to 800 Ohms. Increasing the distance between a planar reflector generally has 3 easily noted effects. First, it raises the feedpoint impedance of the driver. Second, if the distance is greater than the maximum gain position, performance gradually declines relative to both gain and front-to-back ratio. Finally, increased spacing between the driver and screen tends to widen the operating bandwidth of the array. By increasing the spacing between the driver and screen to 0.3 wavelength, the feedpoint impedance rose toward 1000 Ohms. However, the maximum free-space gain dropped by 0.9 dB to 7.45 dBi, while the front-to-back ratio fell to 16.55 dB. Nevertheless, as shown in Fig. 9, the small increase in spacing widened the 2:1 SWR bandwidth, with each array design using its own reference impedance.

Although the curves appear similar, note the difference in the frequency limits of each graph. At a more optimal position for array gain, the passband runs from 9.2 to 10.9 MHz or 1.7 MHz. By increasing the spacing, the operating passband now extends from 9.2 to 11.9 MHz or 2.7 MHz. Notice that the SWR in neither case reaches a 1:1 value. That goal is often only an amateur fetish (but is not always a fetish by any stretch of the imagination). By selecting an acceptable reference impedance--generally one that reflects a transmission line that we can use with the system--we can often attain a wider passband within the upper limits of allowable SWR.

Stretching the operating passband in terms of SWR does not guarantee that the array pattern will be equally usable everywhere within the frequency limits. Fig. 10 presents sample E-plane patterns from the wider passband, using the upper, lower, and mid-band frequencies. Within the span of the antenna, the gain drops from 8.04 dBi down to 5.02 dBi as we raise frequency (and the spacing becomes wider as a function of a wavelength). The 180-degree front-to-back ratio tends to be stable in the 16-18-dB region. However, as we raise the operating frequency, the beamwidth broadens, especially toward the upper passband limit. At 11.9 MHz, the pattern shows twin peaks, although there is no noticeable null between them. However, for some applications, the beamwidth may have become too wide to meet operating criteria.

Our step-2 exercise has increased the frequency range for allowable operation. In the process, the exercise has also shown us that not every frequency that we can use is one that we can use well.

The third step on the road to broadening the passband of a dipole array involves what happens when we phase-feed more than 1 driver. For this rudimentary demonstration, I set 3 folded-fan dipoles (without a screen) at vertical intervals of 1/2 wavelength. The center dipole serves as the fed driver relative to the main transmission line. Each outer driver receives energy from a 1/2-wavelength transmission line connected to the center driver. The three drivers are now roughly in parallel. Hence, we can expect a reduction in both the resistance and the reactance at the feedpoint.

However, the drivers interact with each other. Outer drivers essentially interact with only one other driver, and mutual coupling shifts the feedpoint impedance of each of them by like amounts. However, the center driver mutually couples to both outer drivers and shows a different shift in impedance from the value it would have if used in isolation. By judicious sizing of the drivers we can overcome the impedance difference. However, let's size them in concert, that is, make them all the same size. The left portion of Fig. 11 shows the set-up in outline form.

Because the impedances on the outer drivers will be increasing for part of the frequency sweep while the center driver impedance decreases--and vice versa--we obtain an additional increment of passband broadening. Fig. 12, at the top shows the new passband, which even without a reflective screen extends from 9.55 to 11.8 MHz or 2.25 MHz. The reference impedance for the curve is 250 Ohms.

The bi-directional maximum gain of the drivers across the operating passband increases from 7.17 dBi to 8.03 dBi, with well-behaved lobes. Therefore, the driver set seems fit for combining steps 2 and 3, that is, using a phased set of 3 drivers ahead of a screen. The right portion of Fig. 11 shows the ultimate array (at least for our demonstration) in outline form. The lower portion of Fig. 12 reveals the 300-Ohm SWR passband, which now extends from 8.35 up to 11.5 MHz or 3.15 MHz. The following brief table samples the modeled free-space performance values at each ends of the passband and at the mid-band frequency.

Modeled Free-Space Performance of a Vertical Stack of 3 folded-Fan Dipoles 0.3 WL Ahead of a Screen Reflector
Frequency (MHz) 8.35 9.925 11.5
Maximum Forward Gain (dBi) 11.99 11.86 10.99
180-Degree Front-to-Back Ratio (dB) 19.81 20.15 20.00
E-Plane Beamwidth (degrees) 58.9 68.8 95.8

Except for not showing relative gain values, the patterns in Fig. 13 put a graphic face on the data in the table.

The array patterns show the same sort of development that we found in Fig. 10 for a single driver and screen. At the top of the passband, we find dual peak gain levels, but without a noticeable null between them. We also find a rapid rise in the beamwidth above the mid-band frequency--nearly 30 degrees. Although the same cautions apply to the expanded array as also apply to the single driver array, we should note that the operating passband is now about 50% greater.

Our goal has been only to demonstrate that obtaining a wide operating passband from a dipole curtain array requires a combination of ingredients or steps, as we have called them. The folded-fan dipole is far from being the ideal starting point for effecting the scheme, although its has served well in the demonstration. Some manufacturers of dipole screen arrays claim up to a 2:1 frequency range while also listing tighter SWR limits in their specification sheets. However, all specification sheets are incomplete performance records, so we have no idea of whether the patterns associated with frequencies across a given frequency range are all usable to the same degree as they are at mid-band.

Regardless of the performance obtained by any given maker, we have seen that through careful design (even at the crude level used in this exploratory account), a dipole array with a screen reflector lends itself to broadband service with gains that one might tailor by selecting the correct array size. A planar reflector, whether solid or composed of lines or screens, offers considerable flexibility in system design. If we give up the amateur habit of always seeking the highest gain from the fewest elements, we can achieve a number of other advantageous performance features in our arrays.

Back to the Billboard

Modern screen or curtain arrays employ dipoles. However, in the dim recesses of past times, the original standard driver billboard array was one or more sets of center-fed 1-wavelength elements, also called collinear half-wavelength elements. We neglected to test this arrangement to see if it offers any advantage over the use of dipoles. Let's pause before closing to see what happens if we replace dipole drivers with the 1-wavelength drivers.

To gain some perspective on the question, let's use a 2V-2H dipole driver set as one comparator. The corresponding billboard array would employ a 2V-1H driver set. Both driver arrays would have the same dimensions and require the same size reflector screen. In fact, we may also use the same spacing between the drivers and the screen for both antennas. The top portion of Fig. 14 shows the two outlines.

Between the two types of driver arrays we find under 0.1-dB difference in gain and only 1 degree difference in beamwidth. The front-to-back ratio of both arrays is near 28 dB. As the sample E-plane patterns show in Fig. 14, Nothing in either pattern gives one or the other array an advantage.

We have already explored some ins and outs of dipole impedance behavior in a planar-reflector array. The older billboard system shows individual feedpoint impedances that are very high--in excess of 3000 Ohms resistance. For a single frequency, the high impedances are no hindrance to the use of 1-wavelength elements. However, if we wish to change frequencies, we may encounter a need to retune the system. The center of a 1-wavelength element is a region of very large and rapid changes of impedance values as we shift the operating frequency. Unless all wires composing the system are very taut, storm winds may result in some impedance oscillations.

The much lower impedances of dipoles--even raised to several hundred Ohms--provide the array designer with a much more controllable situation with respect to impedance matching over a wider frequency range. The conversion to dipoles as drivers has made the dipole screen array a mainstay of short-wave broadcasting by providing the necessary gain and beamwidth (including slewing) with the ability to handle high power levels. Increased operating bandwidth supplies the final need of the SWBC industry.

In these notes, my aim has been to examine some of the features of an antenna type that we often encounter on our travels across the U.S., but seldom have occasion to use personally. The maze of wires--whether active antenna parts or support guys--gives these antennas an air of mystery--or at least considerable puzzlement. (Of course, not all mazes of wires strung between 2 tall support masts are dipole curtain arrays.) Hopefully, this small set of exercises has taken some of the mystery out of the arrays.

In the course of developing these notes, I have used published information on the antenna type, abetted by some first-order free-space modeling. Hence, my slant on certain features may differ markedly from the perspective brought to bear by an antenna engineer deeply involved in the design and implementation of such arrays. Indeed, I may well have overlooked numerous features that a manufacturer might consider critical and stressed others considered marginal or even trivial. Still, I hope that these notes contain enough analytical information to make the antenna type--the dipole curtain array--more familiar to and understandable by those who see one for the first time.

Return to Amateur Radio Page