Wide-Band 40-Meter Yagis
Part 1: Standard and Non-Standard Designs

L. B. Cebik, W4RNL (SK)

In my Notes on Wide-Band Yagis, I discussed several 40-meter designs for full-band coverage (Volume 2, Chapter 7). Since the appearance of those notes, QEX has published an interesting new design by J. V. Evans, N3HBX. It promises (and delivers) the performance of a full-size wide-band 3-element Yagi in about 2/3 the boom length (31' vs. 45'). The key feature is the absence of a reflector element. Some years back, I noted in converting a 2-element pair of phased elements into a beam that a reflector was unnecessary, since the directors control both the gain and the front-to-back ratio. N3HBX provides further confirmation of this fact with his beam.

In addition, the beam also calls to our attention a facet of designing antennas by computer modeling techniques, one usually overlooked in the modeling process. One facet of converting a computer design into a physical reality is the sensitivity of the design to the inevitable small construction variations from the computer design. The more able a design is to absorb these variables without changing essential characteristics, the more buildable is the design. As we review some standard wide-band designs and the N3HBX design, we shall have an opportunity to examine this matter more closely and to show how we can use the design software to arrive at a probable answer to the question: "How buildable is the antenna?"

Some Essential Background on 40-Meter Yagis

The U.S. 40-meter allocation runs from 7.0 to 7.3 MHz. If we measure the band in kHz, it is narrower than 20, 15, and 10 meters. However, in terms of bandwidth, as the term applies to antennas (and other components and circuits), we cannot use a simple kHz-count. Instead, we should measure the bandwidth as a percentage of the center frequency. If we divide the 300 kHz bandwidth by 7.15 MHz (and multiply by 100, of course), we arrive at a bandwidth of 4.2%. Compare this value to the 2.5% bandwidth of 20, the 2.1% bandwidth of 15, and the 3.5% bandwidth of the first MHz of 10 meters. Obviously, covering the full width of the U.S. 40-meter band with a single effective Yagi is a challenge.

However, the challenge only begins with bandwidth. To an important degree, the element diameter influences the available bandwidth of a Yagi design. Suppose that we begin with a wide-band 10-meter design capable of covering the first MHz of that band. One way to obtain the coverage with relatively smooth performance figures is to use larger tubing. On 10, we might use an element taper schedule the begins with 3/4" tubing and goes down to 3/8" tubing. Such elements often have equivalent uniform diameters close to 0.5". If we scale the design to 40 meters, we must have an equivalent uniform diameter of about 2.0".

Unfortunately, we cannot simply scale the 10-meter elements. Rather, we must build tubular elements from many tubing sizes so that the element can support itself in a reasonable wind. The beefiest 40-meter element structure that I have so far encountered in U.S.-made beams begins with 2.5" tubing at the center and gradually reduces to 0.375" tubing. Fig. 1 shows the steps in this element. You may assume that where the steps are greater than 1/8", tubing has been doubled or tripled for greater strength in the sections that bear the highest loads. I shall not perform a full stress analysis on this element structure. However, I shall use it in all of the models that we shall review here. In that way, each model is subject to the same set of potentials and limitations related to potential physical structure.

The taper schedule of our standardized element yields an equivalent uniform diameter between 1.3" and 1.6", depending on the exact length of the element. This value falls far short of the requisite 2" to obtain equivalent operating bandwidth to a wide-band design that might appear at 10 meters. Therefore, we shall have to substitute design ingenuity for the convenience of electrically fat elements.

The extreme taper of the element presents a small challenge for modeling it. MININEC has no significant difficulties with stepped-diameter element construction, but NEC-2 is notorious for the errors that such element construction produces. NEC-4 revised the current calculation algorithms to overcome the problem, but it does so only if the steps are small and if the region of highest current has the fewest and smallest steps. If we look at the half-element diagram, we find that the inner half of the half-element decreases by a full inch in diameter, in steps that reach as high as 3/8". NEC-4 alone cannot handle this structure without error. Hence, whether using NEC-2 or NEC-4, one must invoke Leeson corrections and calculate on the basis of equivalent uniform-diameter elements.

We complete our list of design challenges with one more peculiar fact about wide-band Yagis on 40 meters. The operators who are most likely to invest high dollars in a large 40-meter Yagi, along with the massive support structure to hold it at an advantageous height and to rotate it around the horizon, tend to be avid contesters. Their equipment usually includes power amplifiers capable of the full legal amateur limit using any mode of operation. Many of these amplifiers contain protection circuits to ensure long life from the expensive components within them. One of those protection mechanisms is a circuit that cuts off the amplifier when the return voltage sample indicates an SWR value above a certain level. In amateur circles, we are most used to using a 2:1 SWR level as a convenient marker for operating bandwidth that will satisfy the modern generation of amateur transceivers. However, in many amplifiers, the cut-off limit is 1.5:1. Therefore, we must design with that limit in mind.

The Designs That We Shall Review

In the course of these notes, we shall examine the characteristics of 4 Yagi designs ranging from 2 to 4 elements. Fig. 2 shows the 4 designs in outline. The outlines are close to being in scale with each other. Hence, you can see the growth in the overall boomlength of each design as we move from 2 to 4 elements. The shortness of the N3HBX design becomes clearly evident.

Since every element uses essentially the same structure, we can supply the necessary additions to let you replicate the models. The following table lists the progressive element spacing, using the reflector as the baseline. It then lists the element half-length. Multiply by 2 for the total element length. The final column lists the tip length. A zero or a negative entry means that the element has no 3/8" tip section, and the value in the column represents an adjustment to the 0.5" element section shown in Fig. 1. You may also use that graphic to fill in the requisite wires for any models that you create from the following table.

Dimensions of the Yagis Discussed in These Notes

Note: All dimensions are in inches. Multiply by 2.54 for centimeters,
0.0254 for meters.

Antenna Element Spacing Half-Length Tip Length
2-Element Reflector 0 451 65
Driver 282 412 26
3-Element Reflector 0 463 77
Driver 305 428 42
Director 540 386 0
4-Element Reflector 0 449 63
Driver 245 436 50
Director 1 329 407 21
Director 2 622 379 -7
4-El N3HBX Slaved Driver 1 0 456.5 70.5
Fed Driver 62 440 54
Slaved Driver 2 107 420 34
Director 1 367 383.5 -2.5

For any given design, the required element lengths will change if you a. change the equivalent uniform diameter or b. change the taper schedule for the tubing. If the effective diameter does not change by much, you likely will not need to adjust the element spacing, since spacing is relatively less sensitive than is element length. Just how sensitive the element lengths are in a given design will be one of our later subjects.

2-, 3-, and 4-Element "Standard" Yagi Designs

In the 1980s, Bill Orr, W6SAI, presented a pair of designs for 10 meters. I have generally ascribed the designs to him, since he did not specify if he obtained them or generated them himself. The basic principle underlying both his 2-element and his 3-element design is straightforward. The reflector-driver spacing in a standard Yagi largely determines the feedpoint impedance level of the array. In the 1980s, most Yagis used fairly narrow spacing and ended up with feedpoint impedances ranging from about 18 to 25 Ohms.

In a 2-element Yagi that uses a reflector and a driver, widening the spacing actually has 3 effects. First, it lowers the gain and the front-to-back ratio by a small amount, but not enough to be operationally significant if 2-element performance is useful. Second, it raises the feedpoint impedance. A spacing of about 1/8 wavelength yields about the best performance compromise in a normal 2-element Yagi, but the feedpoint impedance is between 30 and 35 Ohms. Raising that spacing to a range of 0.14 to 0.17 wavelength increases the impedance to a good match for 50-Ohm coaxial cable. Third, the wider spacing also increases the SWR bandwidth. So the sample 2-element 40-meter Yagi uses a spacing of 0.17 wavelength. As a result, we obtain the following table of results, sampled at the band edges and at mid-band.

Frequency       Free-Space       Front-Back        Feedpoint Z               50-Ohm
MHz Gain dBi Ratio dB R +/- jX Ohms SWR
7.0 6.50 9.89 40.7 - j 19.4 1.60
7.15 5.96 10.70 53.1 + j 3.5 1.09
7.3 5.52 9.71 63.5 + j 24.2 1.63

To convert the raw numbers into a more graphical form, Fig. 3 presents curves that track the free-space gain, front-to-back ratio, resistance, reactance, and 50-Ohm SWR across the 40-meter band.

2-element driver-reflector Yagi gain naturally decreases with increasing frequency. The design creates a front-to-back peak ratio about mid-band. The feedpoint resistance rises only 23 Ohms across the band, while the reactance shows a 44-Ohm spread. As a result, we obtain acceptable 2-element Yagi performance with an SWR well below 2:1. However, we have missed the more desirable 1.5:1 ratio at the band edges. One standard commercial spacing used for 40-meter 2-element Yagis is 20'. Our wide-band Yagi uses a 23.5' boom.

The 3-element Yagi also uses an Orr-derived 10-meter design as its foundation. Once more, wider reflector-to-driver spacing increases the feedpoint impedance. By properly sizing and positioning the director, we can obtain good patterns and reasonable forward gain. However, the gain level is more appropriate to a shorter boomlength when designed for a lower impedance. For example, a boom length of about 0.22 wavelength in a conventional Yagi will yield about 7 dBi free-space gain, while a boomlength of about 0.33 wavelength will yield about 8 dBi gain, both with feedpoint impedances in the 25-Ohm range. Our 50-Ohm direct feed Yagi design yields about 7 dBi gain on the longer boom. However, we need that 45' boom to give us the wide-band performance values shown in the following table.

Frequency       Free-Space       Front-Back        Feedpoint Z               50-Ohm
MHz Gain dBi Ratio dB R +/- jX Ohms SWR
7.0 6.92 19.16 51.6 - j 17.6 1.41
7.15 7.03 21.54 48.9 - j 1.3 1.04
7.3 7.37 19.00 41.6 + j 18.8 1.57

We just miss having a 20-dB front-to-back ratio from one end of the band to the other. The 3-element Yagi almost reaches the goal of a 1.5:1 50-Ohm SWR all across the band. Fig. 4 provides the graphs and patterns to establish the utility of the design, even if the Yagi is twice as long as the 2-element version.

The gain curve is apropos to any Yagi having a single driver and at least one director: it rises as we increase frequency. The front-to-back ratio peaks about mid-band. The resistance and reactance excursions are between half and two-thirds of the amounts shown by the 2-element Yagi. Nevertheless, we have only come close to the most desired SWR limit.

The third design uses 4 elements on a 52' (0.38-wavelength) boom. The added element and increased boom length yields about a full dB of additional gain while preserving the front-to-back ratio levels. The design is an adaptation of a 10-meter design that I developed several years ago. The element called Director 1 in Fig. 2 also functions as a slaved driver to broaden the operating passband of the antenna. With the dual driver or master-slave system, the reflector tends to exert some control on the pattern properties at the low end of the band, while the remaining forward director has greater control of the high end of the band. The following table samples the modeled values.

Frequency       Free-Space       Front-Back        Feedpoint Z               50-Ohm
MHz Gain dBi Ratio dB R +/- jX Ohms SWR
7.0 7.90 18.08 46.8 - j 9.0 1.22
7.15 8.00 23.99 50.2 - j 0.2 1.01
7.3 8.21 18.64 44.9 + j 7.0 1.20

The most notable feature of the design is the SWR curve in association with the values of feedpoint resistance and reactance. Fig. 5 shows the data in graphical form. The 50-Ohm SWR never reaches 1.25:1 anywhere within the band. To obtain this level of performance, we must use a very long boom with very heavy elements on its ends. Hence, the support system for this beam would likely require heavy-duty trusses for both the individual elements and for the boom itself.

The 4-element wide-band Yagi has appeared in many guises, ranging from the lower HF range up to UHF. With suitably fat elements on 70 cm, the antenna is capable of covering the entire band with usable--if less than optimum--performance. The drawback of all of the designs is that they pay little heed to boomlength. On 40-meters, the shorter the boom, the more easily that we may convert a design into a physical antenna.

The 4-element N3HBX Yagi

N4HBX experimented with modeling 40-meter Yagi designs using a pair of slaved drivers that surround the fed driver. The triple combination, he reasoned, would offer even greater operating bandwidth. In the process of his design work, he discovered that he could eliminate the reflector. He lost some gain in the process, but the amount by which the boom became shorter more than offset the loss. His final boomlength is only about 0.22 wavelength or 31'.

Surrounding the fed driver with slaved drivers that use relatively tight spacing tends to give the slaved drivers control over the array's properties. Hence, the gain shows a mid-band dip (insignificantly small) as well as dips in the SWR curve. The front-to-back curve shows a small mid-band peak, but remains at 20 dB or better across 40 meters.

The cost is an impedance curve that varies around the 25-Ohm level. In the data table and impedance graphs, I have attached a quarter-wavelength section of 35-Ohm transmission line to match the impedance to a main 50-Ohm cable. You can construct the matching section from 35-Ohm coax or from paralleled lengths of 70-Ohm cable. The following table lists both the pre-match and post-match impedance and SWR values.

Frequency       Free-Space       Front-Back        Pre-Match Z        25-Ohm        Post-match Z       50-Ohm
MHz Gain dBi Ratio dB R +/- jX Ohms SWR R +/- jX Ohms SWR
7.0 7.21 20.91 16.3 - j 2.8 1.57 67.7 + j 20.5 1.59
7.15 7.18 25.20 31.7 + j 0.4 1.27 38.7 - j 0.1 1.29
7.3 7.46 20.69 15.7 + j 4.8 1.69 72.6 - j 20.4 1.65

The SWR values for the sampling points do not show the SWR minimum values that occur about 50 kHz from each band edge. These appear clearly in the graph in Fig. 6. The figure also includes the gain and front-to-back curves, as well as sampled patterns that correspond to those shown for the other beams in the collection. The gain reaches its minimum value at about the same frequency on which we find the maximum value of front-to-back ratio.

The N3HBX design achieves well under 2:1 matched SWR at the band edges. However, it does not achieve the ideal 1.5:1 level. The reason is that slaved drivers generally have a narrower operating bandwidth than a fed driver. Hence, the two slaved drivers lose control of the feedpoint impedance before we arrive at the band edges. Optimizing the driver lengths for a lower SWR value at the band edges results in a much higher mid-band SWR value, since the pair of slaved drivers effectively prevents the fed driver from controlling that portion of the band with respect to antenna properties. Therefore, we see only 2 SWR minimums and 2 gain maximums rather than 3. Nevertheless, the beam has generally very even performance from one end of 40 meters to the other--all on a relatively short boom.

Perhaps the key design element in the N3HBX design is the removal of the reflector. Many radio amateurs are wedded to the notions of "reflector" and "director" that suggest to the imagination that some real reflection and direction is occurring. In fact, all elements are directors in the sense of being positioned and sized so that the signal is highly directional. The reflector does not reflect in the sense that a flashlight reflector reflects. Rather, the length and position of the reflector sets up a current magnitude and phase angle that contributes to the overall directionality of the antenna's radiation field. Where we have only a driver and a reflector, the parasitic rear element plays a key role in the pattern shape. But as we saw in the 2-element Yagi design, the reflector only helps the gain and the front-to-back ratio a little bit. Directors are the key elements in controlling both the gain and the front-to-back ratio of a standard design Yagi. Once we add directors, the reflector's role is reduced largely to establishing the general level of the feedpoint impedance on the driver. Without a reflector, the N3HBX design could not achieve a direct 50-Ohm feedpoint connection, although the 25-Ohm level proves to be perfectly matchable to coaxial cable.

While exploring the properties of phased elements about 5 years ago (August, 1998), I experimented with adding a director to a pair of phased elements. I set the phased pair of elements to be nearly the same length and used a phasing line that set them up for very close to the maximum gain possible from 2 elements--a little over 7 dBi in free space. As Fig. 7 shows, that configuration yields a very poor front-to-back ratio compared to the very high ratio we can obtain from the same pair of elements at the same spacing just by altering the relative current magnitude and phase angles. (However, when set for maximum front-to-back ratio, the gain drops to just about 6 dBi.) The next step was to add a single director without altering the phased element pair. The 17.2' director (at 12 meters) is 9' from the phased element pair. The figure shows the resulting free-space E-plane pattern with 8.6 dBi gain and over 30-dB front-to-back ratio.

The experimental beam has quite narrow-band properties and requires a reverse beta match for a 50-Ohm cable (that is, a capacitor across the feedpoint to complete the L-network). The narrow-band properties of the phagi (or phased-Yagi combination) follow from the fact that the setting of a phased element pair for maximum gain is itself a narrow-band property. (For further information on the experimental beam, see Horizontal Phased Arrays with Parasitic Directors.) If this analysis is correct, then the driver set of the N3HBX array--without the director--should show some pattern properties that are as wide-band as the total Yagi itself. See Fig. 8 for confirmation of this condition. The patterns display values that would be quite usable in a 2-element Yagi. Unfortunately, in the absence of the director, the feedpoint impedances are not practical.

It does not matter whether we call the two un-fed elements slaved drivers or parasitic elements functioning as a very close-spaced set of reflector and director elements. In both cases, the elements have positions and lengths to effect current magnitudes and phase angles that result in a directional pattern. The director plays the key role in altering the patterns to their final desired shapes within the limits of what any parasitic element might do.

I am not advocating the general removal of reflectors. An operational Yagi is a blend and balance of many properties. Where one or more directors are present, the reflector remains a primary method of setting the general level of the feedpoint impedance to a convenient level for use with common feedlines. In most designs, that function is significant. In the case of the N3HBX array, shortening the overall boomlength proved to have a higher priority, and the resulting un-matched feedpoint proved to be quite usable without the reflector. For many other designs, doing away with the reflector would be highly impractical.

Moving from Model to Reality

One question normally overlooked by individuals who are new to designing antennas via computer software is the "buildability" of the resulting design. Even if we follow the computer-generated plan as closely as possible, the physical elements will rarely match the design dimensions exactly. Hardware will create "lumps and bumps" on elements. Even mounting hardware that we insulate from the elements may detune them by a very slight amount. Unless we have a very extensive or commercial-grade shop, our element-section lengths may vary a fractional amount from the model. Finally, even if we insulate the boom from the elements, we may encounter a small amount of coupling that the model cannot show due to its inability to handle transverse as well as axial current along elements.

Despite these potentials for a divergence between the model and reality, innumerable successful antennas have emerged from computer design processes. In most cases, the differences turn out to be negligible when measured in terms of predicted and actual antenna performance. However, that fact is not true of all designs. Therefore, the designer needs to develop a system for evaluating the likelihood that a given antenna will be sensitive to construction variables.

The need for such a system emerges from the modeling process itself. As we approach the dimensions of the "final" design, we tend to sneak up on the element lengths and positions. In other words, we vary each dimension in very small amounts until we optimize performance or at least achieve acceptable performance. Rarely do we keep track of the increments we use in varying element lengths. More important, we rarely take note of the significance of those design-variation increments. If we do not normally take this step during design, then we should systematically evaluate the design once we have frozen it.

No single system of pre-testing a design will satisfy every antenna on every band built in every situation. Commercial makers can usually cut tubing within a few thousandths of an inch, while the home-shop builder may be restricted to the accuracy of his eyes and the markings on the tape measure he uses. A half-inch variation limit in element length on 10 meters is equivalent to about 2" as a limit on 40 meters. Both figures represent about a quarter of 1% for the bands involved. Therefore, whatever tests we perform on the design models are likely to vary from one antenna design to another and from one part of the frequency spectrum to the next.

To obtain a first-order measure of how sensitive the 40-meter wide-band antennas are to construction variables, I set up a very simple test. Since the Yagis in this collection are set on 40 meters, I varied the individual element lengths by +2" and then by -2" from their design values. For each variation, I recorded the most telling data at the design frequency limits (7.0 and 7.3 MHz). We already have the values for the design dimensions in the tables associated with each Yagi. Of course, after looking at each element, I returned it to its design length before varying the length of the next element. After looking at the variations due to element length, I also varied the spacing between element by the same +2" and -2". Since element spacing is less sensitive to change than element length, I did not expect to find significant performance variations in that department. However, since some of the designs use parasitic or slaved drivers, the checks were necessary.

The "standard" designs all turn out to be quite buildable within the limits of the test. That is, changes of the order used in the test did not materially affect the radiation or the impedance performance by more than very small amounts. In no case did they jeopardize the utility of the antenna for its intended purpose. However, the N3HBX design raises a few question marks.

To abbreviate the test report, we may compare the N3HBX design with the longer 3-element Yagi that has similar performance. Fig. 9 shows the general parameters of the test runs.

The charts of values for the two antennas cover several sheets of paper. However, we can summarize the results by taking the maximum variation in each performance category. Very often a single element of spacing will account for the variation in a given category, and we may annotate the table accordingly. The following table provides the results for the standard 3-element design.

Results of Varying Dimension by +/-2" for the 3-Element Wide-Band 40-Meter Yagi

Category Variation Variation Notes
Range 7.0 MHz Range 7.3 MHz
Gain (dBi) 0.05 0.11 Equal range from reflector and director length variations
Front-Back Ratio (dB) 0.62 0.60 Sources variable
FP Resistance (Ohms) 1.11 2.68 Director length
FP Reactance (Ohms) 4.16 4.29 Driver length
50-Ohm SWR 0.115 0.117 Driver length

As expected, changes in the element spacing did not yield value changes that approached the high and low values that define the range of variation in each category. More significantly, the range of change in each category is very small. For that reason, I have characterized the 3-element wide-band design as quite buildable in the average home shop--assuming that the shop can handle the basic materials required. A commercial-grade shop or factory is not necessary to build this antenna successfully, if we again assume the same quality of components and basic construction techniques.

When we turn to the N3HBX design, we obtain a different set of values in each sampled category. The following table provides the data from the modeled variations. The model used contained the quarter-wavelength matching section. Hence, typical band-edge feedpoint resistance values are all above 50 Ohms.

Results of Varying Dimension by +/-2" for the 4-Element N3HBX Wide-Band 40-Meter Yagi

Category Variation Variation Notes
Range 7.0 MHz Range 7.3 MHz
Gain (dBi) 0.08 0.06 Rear-driver length at 7.0; director length at 7.3
Front-Back Ratio (dB) 3.78 0.98 Rear-driver length at 7.0; director length at 7.3
FP Resistance (Ohms) 9.11 18.99 Rear-driver and fed driver length at 7.0;
front-driver length at 7.3
FP Reactance (Ohms) 13.84 20.32 Fed-driver length at 7.0; front-driver length at 7.3
50-Ohm SWR 0.290 0.643 Rear-driver length at 7.0; front-driver length at 7.3

If we compare the 4-element Yagi table to the table for the 3-element design, we find greater variations among band-edge values in every category except forward gain. Gain, however, is rarely critical in a wide-band design. The N3HBX design shows modest changes in the front-to-back ratio at the upper band edge, since the director largely controls that value. At the low end of the band, the reflector normally performs the same function. However, the design lacks a true reflector, and the front-to-back value changes by almost 4 dB due to changes in the length of the rear slaved driver.

The feedpoint resistance changes by a factor of 9 at both ends of the band, relative to the changes we saw in the 3-element design. The rear slaved driver and the fed driver contribute equally to the resistance changes at 7.0 MHz, while the forward slaved driver accounts for the larger range of change we see at 7.3 MHz. The fed driver is the source of the widest reactance range at 7.0 MHz. At 7.3 MHz, the forward slaved driver again yields an even wider range of reactance values as we alter its length by the +/-2" amount defined for the test. Note that the forward slaved driver' length contributes most to the wider range of both resistance and reactance values. These values move in the same direction, resulting the very large change in SWR value. In fact, if we decrease the forward slaved driver length by 2", the band-edge 50-Ohm SWR drops to 1.39:1, while a 2" length increase yields an SWR increase to 2.03:1.

The overall picture for the N3HBX design is especially interesting, since a 2" length change to any one of the elements becomes the source of at least one tabulated value reaching a surveyed limit. The element with the most influence on the performance curves for the antenna is the forward slaved driver. The tests suggest that construction variables may well move the performance at 7.3 MHz outside the useful range of values.

The test establishes that the N3HBX 4-element design is more sensitive to slight variations in construction than we find for the 3-element design. We might have intuited this situation from the compressed spacing within the driver cell. Very closely spaced elements change their mutual coupling with only small changes in length and in spacing. Experience with master-slaved driver arrangements would also have taught us that the forward slaved driver tends to be the most sensitive element in any array using virtually any variation on that system. (Many wide-band VHF beams use first directors that in fact function as slaved drivers over half or more of the Yagi's passband, even though the spacing may not suggest this function. However, the function usually becomes clear if we track the relative current magnitudes on the fed driver and the first director across a given passband.)

These notes on sensitivity to change certainly do not condemn the N3HBX design. Indeed, it is potentially a highly competent beam that is competitive with 3-element designs that require much long booms. Instead, the notes simply offer a few cautions. A commercial shop could easily fabricate successful replicas of a field-adjusted prototype. In the home shop, where 1 and only 1 beam will emerge, the lesson is that we cannot simply build and mount the antenna with expectations that it will necessarily perform to specification. We have to be ready for field adjustments to bring the performance to specification. Since this procedure is wise for any Yagi, no matter how forgiving the computer design, we might alter the forewarning to this: for the N3HBX design, we should be prepared for extensive field adjustments.

The prospective size of the field adjustment task is one of the factors that does or should go into the building plans of anyone who constructs his own beams. How much weight we give this task varies from one builder to the next. The more sensitive a beam design is to small changes in its dimensions is a measure of the likely size of the adjustment task. These notes do not intend to establish the weight we give the task. Instead, they only provide one method of evaluating the size of the prospective task while the antenna design is still within the computer.


Adding the N3HBX 4-element reflector-less wide-band 40-meter Yagi to the collection of extant wide-band designs promises 3-element performance from a boom that is not much longer than we find on 2-element full-size 40-meter Yagis. The design is a good example of using a self-contained driver system--in this case, a master-slaved driver system--to achieve wide-band performance. Unexplored alternatives to the N3HBX drive system are direct driver coupling via very short lengths of transmission line and possibly a phased pair or triplet of drivers. Both systems--if successful--would add a few feet to the boomlength but still result in a beam that is quite a bit shorter than the comparably performing 3-element design.

The 4-element reflector-less design with a closely spaced driver cell has also provided me with a pretext for reminding home-designers of Yagi antennas to check the sensitivity of the antenna design to small changes in its dimensions. The results--especially when compared to results for a beam known to be forgiving of small dimension shifts--allows us to estimate the ease or difficulty of replicating the design using real aluminum tubes and stainless steel hardware. That information leads to better planning and decision-making within the overall task of constructing our own Yagis.

Go to Index