Screening 40-Meter Vertical Arrays

L. B. Cebik, W4RNL (SK)

The lower HF amateur bands tend to feature vertical antennas. We find a few horizontal beams, especially on 40 and 30 meters, but they tend to be very large, heavy, and expensive. In contrast, vertical arrays make use of wire and usually use much simpler construction. Hence, their maintenance requirements are also simpler.

At 40 meters (and by extension and scaling 30 meters as well) the vertical dipole becomes feasible, especially if we find a way to shorten it somewhat without losing significant performance. From the vertical dipole, we may create a large number of array types. In these notes, I want to examine in order of increasing performance capabilities a collection of vertical dipole beams and arrays. So that all comparisons will be fair, every vertical array will use AWG #12 copper wire, perhaps the most common material for amateur wire antennas. As well, all antennas will be over average soil (conductivity 0.005 S/m, relative permittivity 13). The performance of vertical antennas will change with the quality of the soil beneath them and in the far-field reflection zone, but the changes will tend to be consistent within any given soil type. Therefore, if we know the performance of an array over average soil and the relative performance for a simple vertical dipole over both average soil and the specific soil type for a given installation, we may extrapolate the performance values for a more complex antenna system.

Vertical Dipole Basics

The root antenna for what follows might seem to be a vertical dipole. Table 1 and Fig. 1 summarize the properties of a vertical dipole that extends from 1' above ground level to a top height of 67.6'. The total wire length is 66.6', but this exact value applies only to a resonant vertical dipole over average soil. Slight adjustments are likely for different soil types, especially since the lower end of the antenna is so close to the ground. As will be the case for all models used in these comparisons, the performance and dimension values do not account for the influence of objects in the immediate area of the antenna installation, that is, the so-called ground clutter. All vertical antennas require as much clearance from ground clutter as the installation site will permit.

For this and all following antennas, the test frequency will be 7.15 MHz. The clutterless radiation patterns show very normal characteristics--a circular azimuth pattern and a single elevation lobe with a low take-off (TO) angle. Hence, despite the low gain of the antenna, it finds general favor for its coverage and for its insensitivity to high-angle noise and signals. The simple outline sketch shows the current magnitude distribution along the antenna wire. One reason why the antenna patterns show a low TO angle stems from the relatively high position of the feedpoint or the region of highest current magnitude, just about 1/4-wavelength above ground. However, the close proximity to the ground at the lower end of the dipole does elevate its feedpoint impedance--from an expected 70-Ohm value up to the 98-Ohm value reported by NEC-4 for the sample antenna.

Many amateurs do not have 70' supports for a 40-meter vertical dipole in its simplest form. Individual circumstances vary, but let's suppose that the maximum support height is about 50'. Within this height restriction, we may still install a modified vertical dipole. Rather than accept the losses that center-loading or mid-element loading might create, we shall use end hats. An end hat or cap is a symmetrical structure at right angles to the main plane of the antenna. Radiation from the wires of the end cap largely self-cancels, leaving us with a vertical antenna in terms of the radiation patterns. Since the self-canceling portion of the radiation occurs in the low-current regions of the antenna, we preserve most of the performance that we might obtain from a full-length dipole. Fog. 2 shows the outlines and current distribution on one version of such a T-capped dipole/

The end hats on a vertical dipole can use any symmetrical arrangement. The more radial arms that we create, the shorter each one must be for a given length of vertical wire in the center. However, many-spoked hats set up very significant support requirements. The T-cap requires only two wires at each end of the dipole, and we may run these wires along the non-conductive ropes that we often use to support wire vertical dipoles between two posts or trees. Reducing the number of hat wires to only two per end does not affect the performance relative to using greater numbers of hat wires. For example, the azimuth pattern of the T-cap dipole in the sketch shows only 0.03-dB of gain variation as we check all 360 degrees of the horizon.

The T-cap dipole that Fig. 2 shows is only 35.5' long, stretched from 5' to 40.5' above ground. (One might easily raise the antenna by another 4 to 5 feet and still remain below the 50' ceiling that we set. The added height would also increase safety by raising the base wires with their high RF voltages above the level that family, friends, or even pets might touch.) The vertical section, then, is just over 1/4-wavelength. For all following 40-meter vertical antennas, we shall use the vertical section of the T-cap dipole. The two horizontal wires at the top and bottom of the antenna, various called arms or legs, are each exactly 10' long in the sample. Hence, the total width of the antenna is 20'. This width falls well within the clear area that we should have for any type of vertical dipole.

If we need to make adjustments for obtaining resonance in subsequent arrays that use the T-cap dipole, we shall adjust the length of the legs of the T, thereby leaving the center vertical section intact. The dimensions for the basic T-cap dipole appear in Table 2, along with some basic performance data. Within that data, only the first section is immediately relevant for comparison with the full-length dipole. First, note the 22-degree TO angle. This value is higher than the value for the full-length dipole, but the T-cap vertical's feedpoint is considerably lower (about 0.16-wavelength above ground). Partly as a function of the lower feedpoint height and partly as a function of the vertical-element shortening, the gain of our T-cap dipole is about 0.4-dB lower than the gain of the full-length dipole. Yet the T-cap dipole, as set up in the sample model, is nearly 30' lower in overall height, yielding what is for most amateur operators a much more manageable construction and maintenance situation.

In the table, we also find some performance values related to improvements that we may make in the local ground immediately beneath the vertical dipole. When we think of local ground improvements together with virtually any vertical antenna, most amateurs immediately think of radials. However, unlike the radials of a monopole, the radials beneath a vertical dipole (either full length or shortened) perform no antenna-completing function. Rather, they simply function to raise the conductivity of the soil immediately beneath the antenna. In cases 2 through 4 in the table, I created radial systems with the hub directly below the antenna. The 1/4-wavelength radials use AWG #12 wire buried 1' deep in the average ground. As the table shows, adding 4 radials amounts to wholly wasted effort, since the gain increase is only 0.04 dB. 16 radials provide a 0.3-dB gain increase, perhaps a marginal amount, considering the work involved. If radial installation is easy, we may increase the field to 64 radials and obtain nearly 0.7 dB gain increase. Fig. 3 provides a view of the 3 radial field along with the T-cap dipole above them. You may estimate the installation work from the sketches.

Note that none of the radial fields changes the TO angle of the elevation pattern. The TO angle is mostly a function of the far-field reflection zone, with is mostly well outside the radial limits. For the exercises, I did not change the antenna dimensions. Therefore, the major influence of the fields appears in the feedpoint impedance listings. As the size of the radial field increases, the resistive component of the impedance decreases and the reactance becomes more inductive. However, none of the changes in cases 2 through 4 present any operational concerns.

The last case presents an alternative method of local ground improvement. Instead of using a radial system, I modeled a wire-grid system below ground to simulate laying a screen of some sort below the antenna. The screen is square and 70' on a side. Fig. 4 overlays the 16-radial system on the screen to show the change in ground coverage in the screen corner region. (I used the 16-radials so as not to obscure the grid. However, the most relevant comparison would be between the screen and the 64-radial system.)

With the screen in place, we obtain almost 1.1 dB gain improvement over untreated soil, and about 0.4-dB improvement over the 64-radial system. The cost of the gain is a 1-degree increase in the TO angle and further drift in the feedpoint impedance. For antennas--like vertical dipoles--that do not require radials to function as the lower half of a dipole, screens may sometimes be the easiest and most effective means on improving soil quality immediately below the antenna. However, they do not provide that same benefits as living over very good soil, since the treatment does not also apply to the far-field reflection zone. For example, local area treatment may reduce ground losses below the antenna, but that results in a higher TO angle, because the improvement increases local reflection almost straight upward. The lower TO angles that we often associate with the same antenna over very good soil results from improved soil conductivity at considerable distances from the antenna.

In the end, soil improvement is a matter for the antenna user to decide after measuring the anticipated performance improvements against the amount and cost of work involved in local ground treatment. For the remainder of these notes, we shall use untreated soil beneath the antenna, so that the performance values in case 1 in Table 2 become the reference points for all that follows.

Before we leave the basic T-cap vertical dipole, we should introduce one more set of considerations. We are interested not only in the performance of any antenna at the test frequency, but also across the entire band. 40 meters is a wider (but not the widest) amateur allocation, with a 4.2% bandwidth. Some antennas will handle the entire band; others will not. A basic dipole--either full-length or T-capped--shows only a show change of gain across the band. The case-1 T-cap dipole, for example, changes gain by only 0.07 dB from 7.0 to 7.3 MHz. Note that our performance concerns include not only the SWR properties, shown in Fig. 5, but also such matters of forward gain and front-to-back ratio, both of which will become important as we add directionality to the basic dipole.

Although the resonant impedance of the basic T-cap dipole is in the vicinity of 75 Ohms, even the 50-Ohm SWR is below 2:1 from one band edge to the other. Hence, we may feed the dipole using either 50-Ohm or 70-Ohm coaxial cable. As usual, one should route the cable at right angles to the vertical dipole for as far as possible, providing supports so that the cable weight does not unduly stress the wire antenna. As well, I recommend the use of 2 common-mode current attenuators, one at the feedpoint to also serve as a balun, and the other at the entry to the operating building to attenuate any currents induced along the cable run.

A Bi-Directional Pair of T-Cap Dipoles

Because many of the arrays to follow will use a pair of T-cap dipoles, we should spend a moment two see what happens when we place a pair of these dipoles at a spacing of 1/2-wavelength (about 68.8') and feed them in phase. Because the in-phase fed pair of antennas will interact (or, otherwise put, exhibit mutual coupling), I extended the T arms to 10.31' each. As a result, the total array width, from outer T-tip to outer T-tip is about 89.4'. As shown in Table 3 and the outline portion of Fig. 6, the array height above ground has not changed relative to using a single T-cap dipole.

Perhaps the most surprising aspect of the array is the 4.1-dB gain improvement over a single T-cap vertical dipole. Of course, the added gain comes at a cost in the available beamwidth, now down to about 62 degrees. Nevertheless, the array is no taller and only a bit wider than a half-square, but provides more gain. The gain that is competitive with a bobtail curtain. In fact, the in-phase-fed pair of vertical dipoles lies at the theoretical core of all SCV (self-contained vertical) antennas, since the basic versions--ranging from deltas to rectangles to half-squares, all place two elements in phase, but at a spacing limited by the single-wire, single-feedpoint configuration. For further information on SCVs, see "Self-Contained Vertically Polarized Wire Antennas: A Family Album"

The independent feedpoint impedances of the two T-cap dipoles is about 56 Ohms. Therefore, we may equip the phased pair with a common feedpoint by using equal lengths of 70-Ohm feedline to a center point. Since the physical distance between each element's feedpoint and the center point is 1/4-wavelength, and since all 70-Ohm transmission lines have a velocity factor (VF) of well under 1.0, you may need to use 3/4-wavelength section to arrive at the required 100-Ohms at the junction, a value that becomes a good 50-Ohm match in a parallel connection. The lower portion of Fig. 6 provides the modeled 50-Ohm SWR curve. The array covers the entire 40-meter band with an SWR value that is less than 2:1.

We shall not linger over the in-phase-fed pair of T-cap dipoles and their broadside bi-directional pattern. However, we shall occasion to mention them once more before we close our screening survey. Nevertheless, we should call to attention one more time the gain value produced by the pair of elements. The maximum gain in each direction will exceed the forward gain value that we may achieve from some of the more basic beams that we consider when thinking about a pair of T-cap dipoles and their best use.

Parasitic Driver-Reflector Beams Using T-Cap Dipoles

For many operating needs, a front-to-back ratio of at least 10 dB may be more important than additions to the arrays forward gain. In such cases, we may create an endfire array using standard 2-element Yagi principles. In a Yagi with full-length elements, the reflector is normally longer and the driver normally shorter than a freestanding resonant dipole. In creating the driver for a useful Yagi-type beam, we may simply reduce the T-legs to an individual length of 9.6' (for a total element width of 19.2').

However, we shall not stop to develop a reflector for the T-cap array that uses a larger element. Instead, we shall simply load the reflector and use the same dimensions as for the driver. Reflector loading of a 2-element parasitic array normally does not reduce the forward gain, and it may actually improve the front-to-back performance over a full size reflector. Table 4 supplies the dimensions of the array, which spaces the elements 21' apart. Fig. 7 shows the outline and the radiation patterns that we might expect.

The azimuth plot overlays two patterns, one in each direction. One of the key advantages of loading the reflector is that--with very little effort--we can reverse the beam's direction. The reflector called for a 55-Ohm load to achieve the desired pattern. Instead of using an inductor as the loading element, we may use a shorted length of feedline, in this case, 50-Ohm cable to match the feedpoint impedance of the driver element. The electrical length would be about 18.24', but the physical length will be the electrical length times the cable's velocity factor. Even a solid dielectric cable will yield a physical length of at least 12', so that cable running from each element can meet in the center of the 21' element spacing. Then, with a suitable remote switch, we can short one cable to make the inductively reactive load for the reflector and connect the other line to the main feedline. With a flip of the switch, we have reversed the beam's direction.

When we first encounter vertically polarized parasitic arrays, we are sometimes surprised by the facts that their gain is lower and their beamwidth is much greater than the patterns that we find in horizontal arrays using the same number of elements. The beam's gain value is about 3.3 dB higher than for a single T-cap dipole, but less than the bi-directional gain of the phase-fed pair of verticals. The beamwidth is nearly 140 degrees, providing very good coverage in each of the beams two possible directions. Although ground losses have a role to play in setting the gain of all vertical antenna types that are close to the ground, the vertical array's gain would not catch up to the gain of a horizontal counterpart until we reached a height above 20 wavelengths. The key factor in the gain and beamwidth difference that we get from rotating the beam 90 degrees along its real or virtual boom is a function of geometry. The arrangement of element tips restricts the beamwidth in the E-plane, that is, in the plane of the elements. However, the H-plane has no such restrictive influence. Indeed, in free-space, H-plane patterns of a 2-element parasitic array look very much like the pattern of a single vertical dipole but displaced in the direction of the forward gain. A Yagi must have many elements (and be very long) before the H-plane beamwidth narrows significantly.

Nevertheless, the reversible 2-element vertical array with T-cap elements provides good service across the 40-meter band. As the 50-Ohm SWR curve at the bottom of Fig. 7 shows, the array will cover most of the band with less than 2:1 SWR. However, we early on noted that our bandwidth concerns covered more than just the SWR. We are also interested in how well the antenna performs in terms of gain and front-to-back ratio. Fig. 8 provides a partial answer to our questions.

The gain curve in the figure shows a 0.4-dB range of forward gain value across the band, a value that we may consider fairly stable for such a wide amateur band using elements composed of relatively thin wire. In contrast, the front-to-back ratio remains above 10 dB only for a small portion of the band. However, it never drops below 5 dB. In fact, with careful adjustment of the reflector load, we may be able to center better the front-to-back curve for relatively similar performance at the band edges. A slightly higher load reactance--meaning a slightly longer length for the shorted stub--would likely do the job. Alternatively, we may favor either the CW-digital end of the band or the phone end of the band, according to our operating needs.

One limitation of the 2-element driver-reflector array is that we leave some portions of the horizon only weakly covered. It might be useful if we could cover the entire horizon with less than a 2-dB drop in gain around the 360-degree span--and at the same time maintain at least the forward gain that we obtained from the reversible 2-element array. To achieve that goal, we may think in triangles.

Suppose that we set up a an equilateral triangle of T-cap vertical dipoles. We may designate the corner containing the driven element as the apex. The remaining corner will contain identical T-cap dipoles, but loaded to form reflectors. Table 5 provides the dimensions and test-frequency performance data, while Fig. 9 supplies a sketch and the radiation patterns.

The T-legs are each 9.5' long for the individual dipoles. Each triangle side is about 30.6' long, yielding a distance between the apex and the center of the triangle base of 26.5'. It does not matter how we orient the T-legs of the dipoles, so long as the legs form a straight line for each dipole. In the triangle shown--and other dimensions are also possible--each reflector dipole requires a j90-Ohm load. We shall again use shorted transmission-line stubs as the source of the required inductive reactance. Since the driver impedance is close to 70 Ohms, we shall use 70-Ohm lines, which require a 19.9' electrical length, with physical shortening that depends on the VF of the line used in the assembly. We shall bring the stubs to a center point within the triangle for switching. At any time, one of the stubs will actually be an extension of the main feedline to the driver, while the switch to form the required reflector loads shorts the other two lines.

The system gain is about 3.2 dBi, about 0.5-dB higher than for a standard 2-element parasitic array and about 3.7-dB higher than a single T-cap vertical dipole. The 15-dB front-to-back ratio is also several dB higher than we found for the 2-element parasitic beam. With its higher gain, the triangle shows a beamwidth that is about 5 degrees narrower than the beamwidth of the simpler beam. Perhaps the key advantage of the triangle is its ability to cover the entire horizon with only a 2-dB gain deficit at the overlap points between the forward lobes of the beam in each of its positions. Although the switching may be more complex for the triangle than for the reversible beam, the electronics are simple and cheaper than a rotator.

The lower portion of Fig. 9 shows the 50-Ohm and 70-Ohm SWR curves for the array. A 50-Ohm feedline from the switch junction to the equipment will be satisfactory if SWR values close to 2:1 are satisfactory at the band edges. However, a 70-Ohm line will reduce the maximum SWR value to about 1.5:1 at the band edges. With respect to the gain and front-to-back performance across the 40-meter band, Fig. 10 supplies the appropriate sweep data.

The performance bandwidth of the triangle is generally similar to the performance bandwidth of the reversible array, but with a smaller range of value change across the band. The gain varies by only about 0.2 dB. The front-to-back ratio remains at 10 dB or more for the entire band. The band-edge values may be equalized by slight adjustments to the load values, that is, the length of the shorted stubs. With the smaller range of performance change over the 300 kHz of amateur allocation (in the U.S.), one may design a triangle for whole-band use and expect only small deficits at each band edge.

A 2-Element Phased Beam Using T-Cap Dipoles

Obtaining the maximum possible front-to-back ratio from 2 elements is difficult if not impossible using parasitic techniques and 2 vertical elements that are close to the ground. However, it is easily possible to improve the front-to-back performance and to obtain a nearly cardioidal pattern by phasing 2 T-cap dipoles in an endfire arrangement. In fact, we may do so using commonly available feedline materials, although we may have to do some experimentation to find the optimal cable for the task. Experimenting with modeling software is more rapid and less costly than experimenting with lengths of actual cable.

Table 6 profiles the phased array that uses two standard T-cap dipoles with 9.6' T-legs and a spacing of 21' (the same spacing used for the reversible parasitic array). Both dipoles use identical construction. The phase line uses two separate lengths that reach a junction where we shall connect the main feedline. The section to elements 1 in Fig. 11 is 084' of RG62, 93-Ohm line with a VF of 0.84. Hence, the electrical length is 1'. The line does NOT undergo a half twist or reversal. However, the line from the junction to the rear element--from the same material--does undergo a reversal. It is 21' physically or 25' electrically.

The forward gain of the array is comparable to the gain of the reversible parasitic array. However, the front-to-back ratio climbs to over 26 dB at the design frequency. As well, the beamwidth increases to about 145 degrees, providing wide coverage in this essentially mono-directional array. The rearward quadrants of this array promise to be exceptionally quiet.

The cost of this type of performance lies in the odd feedpoint position before we add matching components. The low impedance has a high inductive reactance. However, the values are nearly optimal for a beta match, so long as we use an open or capacitively reactive stub across the feedpoint terminals. 14.3' of 50-Ohm VF 0.78 cable or its equivalent provides the necessary shunt component for the beta L-network that yields an impedance close to 50 Ohms. With the beta component in place, the SWR curve in Fig. 11 shows well under a 2:1 SWR value at the band edges.

Fig. 12 provides sweep data for the phased array relative to forward gain and front-to-back ratio. The gain rises steadily across the band over a 0.4-dB range. The front-to-back ratio peaks near the middle of the band and decreases to about 15 dB at the band edges. Compared to parasitic arrays, the whole-band front-to-back performance of the phased array rates very highly. Perhaps the major drawback to phasing is that the array does not succumb easily, if at all, to reversing direction.

Screen Reflectors and the T-Cap Driver

Screen or planar reflectors have been used since the 1920s in major arrays. Indeed, in their early years, some called them billboard reflectors. A solid surface properly sized will reflect radio waves in ways that are analogous to the reflection of light from a flat mirror. We find them in wide use in dipole arrays for short-wave broadcasting and in a wide variety of UHF antennas. More recently, I have recommended their use beneath a number of NVIS antennas to increase gain, essentially by elevating the quality of the reflective plane beneath the antenna. In fact, the wire-grid ground improvement technique noted earlier represents a different application of the same technology.

A wide range of experience shows that planar or screen reflectors are most effective in improving directional gain when they exceed the dimensions of the driver element by between 0.45 and 0.55 wavelength both horizontally and vertically. We can achieve this goal horizontally with our T-cap vertical dipole driver, but any screen reflector will be vertically challenged. The ground, of course, is one limit, preventing the reflector from extending below the driver by the desirable amount. Vertically, we shall be limited as well by some of the height restrictions that we set for this project. If we limit ourselves to a 50' top height, the forward gain will be only a little greater than the gain of the reversible beam. A top height of 100' yields perhaps a half-dB more gain than a 70' height. Therefore, for the comparisons that we shall show, the height of the screens will run from just above ground level to 70'. One might hang such a screen between two widely separated towers that support antennas for higher amateur bands.

The optimum width turns out to be just about 140', which is 1 wavelength at the test frequency (7.15 MHz) and a half-wavelength on each side of the driving dipole. Screen reflectors do not have to be solid surfaces to act like solid surfaces in the HF range. Chicken-wire fencing and open-weave materials will appear solid to 40-meter energy. The model for the screen-reflector array uses a wire-grid, as shown in Fig. 15. The dimensions and performance data appear in Table 7.

The selected distance from the driver to the reflector is 35' for an array using a single T-cap dipole driver and a 140' by 70' screen reflector. You may ground the screen bottom wires with no effect on performance, but with a considerable affect on safety. The driver uses 9.66' T-leg lengths for an 80-Ohm resonant impedance. With a planar reflector, you may vary two items to set a feedpoint impedance: the dipole dimensions and the spacing from the reflector. In general, closer spacing produces lower feedpoint impedance levels, but narrower operating bandwidths. Each space adjustment will change the coupling between the reflector surface and the driver, requiring adjustments to the T-leg lengths to return to resonance.

The chief merits of the planar reflector array are forward gain and operating bandwidth, when we compare the results to the reversible parasitic array. The top height of the reflector limits the front-to-back ratio to about 13 dB. A top height of 100' would have added another dB to the ratio, while a more ideal (and unrealistic) height of 140' would add a further dB or two. The array's forward gain is close to 4.7 dBi, about 1.8 dB higher than the 2-element parasitic array. The forward gain comes at the expense of the beamwidth, which is down to 83 degrees, about 50 degrees narrower than the beamwidth of the parasitic beam.

Despite the 80-Ohm resonant feedpoint impedance at the design frequency, the 50-Ohm SWR curve in Fig. 13 reveals a very low rate of impedance change across the 40-meter band. That curve shows less than 2:1 SWR at the band edges. The lower 80-Ohm curve shows less than 1.5:1 SWR at the band's upper and lower limits. The remaining prime operating parameters are equally slow to change, as shown in the sweep data in Fig. 14.

The gain across 40 meters changes by only about 0.1 dB, while the front-to-back ratio varies by just over 0.5 dB. It is possible to design driver elements with an inherently wider bandwidth and to use the array to cover both 40 and 30 meters with very little change in performance.

One common method of constructing a screen is to use a series of wires polarized as the driver element. To sample that option, I reconstructed the screen reflector using 29 AWG #12 wires at 5.0' intervals. The overall horizontal and vertical screen reflector dimensions remained the same. Doubling the wire size (to AWG #6) yielded no performance improvements. The only other change relative to the wire-grid screen was a 3' increase in the spacing of the driver from the reflector, as shown in the dimensions in Table 8.

As the data and the patterns in Fig. 15 show, the vertical-wire reflector is not quite as effective as the wire-grid version. Gain drops by about a half dB, while the front-to-back ratio decreases by well over 2 dB. Both decreases are indications that the wire version of the screen requires a taller top height or is perhaps "leakier" than the wire-grid. As well, the resonant impedance is 5 Ohms higher and does not produce 50-Ohm SWR values under 2:1 without further matching. However, the lower 75-Ohm SWR curve in the graph does track the wire-grid's 80-Ohm SWR curve very well.

The sweep information in Fig. 16 shows the same broad curves for both the forward gain and the front-to-back ratio. The gain range is about 0.15 dB across the band, while the front-to-back ratio changes by only 0.8 dB. Even with somewhat lesser performance than the wire-grid reflector, the vertical-wire screen still enjoys a considerable advantage over a parasitic reflector with respect to broadband characteristics.

To compensate, even if only partially, for ground losses, one may tilt the reflector back, as viewed from bottom to top. Using an 11' tilt (35' at the bottom and 46' at the top for the distance between the driver and the vertical-wire reflector), it is possible to add a few tenths of a dB to the forward gain and a similar amount to the front-to-back ratio. However, the exercise has its own consequences, as revealed in the overlaid patterns of Fig. 17.

The peak gain lines both forward and rearward are virtually indistinguishable in the patterns. However, note the increased high angle radiation, especially in the rearward lobes. As well, radiation directly overhead to the driver has also increased. In the end, a vertically oriented screen appears to yield the best combination of performance and patterns.

Earlier in these notes, I promised a second look at the in-phase-fed pair of T-cap dipoles spaced 1/2-wavelength apart. The dipole pair produced a bi-directional pattern with a maximum gain that was over 4 dB stronger than a single T-cap dipole. Since we can readily feed the pair of dipoles from a single feedpoint, using a pair of equal-length lines, we might wonder what would happen if we place the broadside array in front of a screen reflector.

Ideally, the screen should be over 200' wide. Understanding that we lose performance as we shrink a screen below its ideal proportions, let's continue to use the smaller 70' by 140' wire-grid screen. With the screen 40' behind the driver, the T-legs of each dipole are 10.16' long, longer than for a single dipole driver, but slightly shorter than the T-legs on the bi-directional array. As shown by the dimensions in Table 9 and the outline sketch in Fig. 18, we lose a good bit of the horizontal screen extension that optimizes performance. Nevertheless, the performance is noteworthy.

The 6.95-dBi forward gain value for the array may require some perspective. First, the value is almost 2.3-dB higher than the gain we obtain from a single driver and the same reflector. A sign that the screen is less than optimal in size comes from the fact that the screenless bi-directional array produces a gain value that was 4 dB greater than a single T-cap dipole. Nevertheless, the phased dipole array and screen produce almost 7.5-dB gain relative to a single T-cap omni-directional dipole. The forward beamwidth is down to 58 degrees, dictating careful aiming of the array.

The front-to-back ratio is not outstanding, at about 11.4 dB, another sign of using a small screen reflector. Because the unmatched feedpoint impedance of each dipole is in the vicinity of 90 Ohms, we may run equal lengths of RG-62 (VF 0.84) to a center point between the dipoles and obtain a net impedance of 48.9 Ohms. As the 50-Ohm SWR curve shows, the array easily covers 40 meters with a maximum SWR that is less than 1.5:1.

As the sweep data in Fig. 19 reveal, the dual-dipole-driver array is as stable across 40 meters as the other sample screen-reflector array. Both the gain and the front-to-back ratio show changes of just over 0.2 dB across the band.

The construction of a screen-reflector is the most difficult portion of the overall project. Therefore, one may well wish to place drivers on either side of the screen for reversible service using separate feedlines all of the way to the equipment location. Whichever driver is in use (whether single or double), the inactive one will remain invisible due to the screen's reflection characteristics. Screen reflector arrays may not be for everyone, but they may serve a few 40-meter operators.


The reason for our title is now clear. We have not only surveyed comparatively the performance of vertical antennas and arrays on 40 meters, but we have added screen reflector arrays to the list that we normally see. This last group of arrays presents serious construction challenges, but offers in return increased gain over the values that we may obtain from parasitic arrays.

All of the driving elements in our comparisons have used AWG #12 copper wire in a T-cap arrangement that extends from 5' to 40.5' above average ground. We changed the length of the T-legs to obtain a resonant feedpoint impedance. The T-cap dipoles provide a uniform element length to help validate the comparisons. Although scarcely longer than 1/4-wavelength, the T-cap dipoles lose very little performance relative to the reference full-length wire dipole with which we began these notes.

Just because we cannot obtain the gain level of a phase-fed dipole pair with a screen reflector does not relegate the intermediate designs to uselessness. Indeed, many operators use T-cap and similar 40-meter vertical antennas with surprisingly good results. The parasitic arrays offer reversibility and even--with the triangle--full horizonal coverage at the flip of a switch. For maximum front-to-back ratio, the phased array is difficult to surpass, even though it uses no complex networks to achieve a nearly cardioidal pattern.

By setting the entire range of vertical arrays on common ground at a common frequency with common construction, you may gather a more precise sense of the benefits and costs as you increase the complexity of a vertical array. Virtually all of the designs will scale directly to 30 meters, although bandwidth there is not a major concern beyond the realm of allowing rather casual construction without incurring performance losses. Scaling to 80 meters is also possible, although for most amateurs, the resulting element sizes may prove to be prohibitive. A T-cap dipole of the present design will extend from about 10' to about 81' at the top, depending on the precise frequency to which one scales the design. The T-legs can be trimmed for resonance if you stick with the AWG #12 wire rather than doubling its diameter. If the amount of trimming is not too great, adjusting only the lower T-legs will not disturb the centering of the feedpoint to any significant degree. (The presence of ground below the bottom of the antenna and essentially free space above the top already disturbs the balance that we presume when we physically center the feedpoint on a vertical dipole. Common-mode current attenuation devices are necessary adjuncts to any of these arrays.)

These notes have focused on performance comparisons and provided very few construction notes. Building any of these arrays, from the simplest to the most complex, will be an exercise in making use of available supports and in using what nature or prior antenna construction provides. In virtually all cases, the result will be far less expensive than a strong tower able to hold the weight of a horizontal beam that shows higher-angle elevation lobes and questionable operating bandwidth.

First printed in QRP Quarterly, January, 1997. Updated 01-22-97. 

Return to Amateur Radio Page