Hacking the Alpha Delta DX-DD

Matt Roberts - matt-at-kk5jy-dot-net

Published: 2018-04-06

Updated: 2022-08-19



What's a DX-DD?

Alpha Delta DX-DD There are several antenna designs that can be used for NVIS communication on 40m and 80m, but the one I keep coming back to is the DX-DD, from Alpha Delta.  The antenna is smaller than a full-sized dipole for 80m — and even shorter than the popular G5RV — but it still has good high-angle performance on both bands.  The loading devices are simple wire coils, which means no capacitors to burn out (or drift) like you have with a typical trap antenna.

The antenna design is actually quite simple, with coil forms made from PVC pipe and magnet wire wrapped in a close-wound inductor.  Using coils like this for multiband dipole antennas is an idea that has been around for a while, but Alpha Delta has commercialized their production, offering this 40m/80m antenna, as well as a 40m/80m/160m version, the DX-LB.


Why the DX-DD?

This antenna is a good one for "typical" NVIS usage.  During a solar peak, that means 40m during the day, and 80m in the evening and during the overnight hours.  Now that we are entering a solar minimum, NVIS mostly means 80m through the day and early evening, and 160m at night — sometimes.  I chose the two-band DX-DD over the three-band DX-LB, mostly for reasons explained in the sidebar at right.

What About 160m?

Using 160m for NVIS contacts during a solar minimum is a more complicated subject than with higher bands, since that band experiences a number of different propagation mechanisms throughout the day.

Many daytime 160m contacts within typical NVIS ranges are actually accomplished using E-layer skip at much lower angles.  As a result, daytime 160m work during the sunspot low is often better accomplished with a vertical antenna.  Indeed, the E-layer sometimes even presents in strength overnight, allowing full darkness contacts within the skip zone on 160m using a veritcal antenna.

A good 160m veritcal antenna can also cover several dozen miles radius via ground wave propagation.  This mechanism isn't dependent at all on the ionosphere, and also benefits most from a vertical antenna.

After sundown, the F-layer critical frequency drops from above 80m to 160m (or less) very quickly, sometimes rendering the evening usability of high-angle F-layer 160m propagation only marginally better than that available on 80m.  On those nights when the F-layer critical frequency hovers between 160m and 80m, a low dipole for 160m could be very useful for overnight contacts.

So the best antenna strategy for including 160m for skip-zone contacts depends on one's needs.  For daytime 160m operation, a vertical is often preferred, and can provide both ground-wave and E-layer skip into the skip zone.  For night-time NVIS support on 160m, a vertical can often continue to provide regional contacts through a combination of E-layer skip, lower-height F-layer skip, and ground wave propagation.

In any case, while 160m can be quite useful during a sunspot minimum, the utility of including 160m support in a multiband dipole is marginal, at best.  As a result, many installations would do better to use a combination of a 160m vertical antenna, and a 40m/80m dipole such as the DX-DD or the G5RV.  Such a combination should provide the best 24-hour coverage of the skip zone throughout the solar cycle.
Although 80m is the consistent workhorse band for NVIS throughout the sunspot cycle, having an effective antenna for 80m can be a bit of a challenge.  For one thing, an 80m dipole is long — around 133' long at the bottom of the band.  Another challenge with 80m is elevating the antenna to a height where ground losses are kept in check.  Most people advocate an installation height on 80m somewhere between 15' and 35' above the ground.  There are advantages to installing NVIS antennas on the low end of that scale, but doing so requires lifting the entire antenna (not just the center) to keep ground losses under control.  So the larger the antenna is, the more work it takes to get its entire length elevated to the proper height.

That's why I have come to settle on the DX-DD design for NVIS work: For NVIS work, the DX-DD is a great balance between mechanical and electrical needs.  However, the DX-DD does have a significant drawback — cost.

Retail pricing for the DX-DD is $150.  The antenna does not include a choke for the feedpoint, so by the time you add a commercial coaxial choke that is capable of 80m operation, the cost can easily exceed $200.  While a prefabricated antenna can be very convenient, that kind of cost is simply not necessary for an antenna that is essentially wire and plastic.

I set out to make my own copy of the DX-DD, using materials that I had on-hand, to see if this very capable antenna design could be done on the cheap, but still maintain the characteristics that make it such a good performer for NVIS work.

Spoiler alert: it's not difficult at all.

Be aware that this news story is still developing.  I have a working antenna, and I'm satisfied with the results, but I'm still playing with it, so this article will likely change as time goes by.


Requirements

There are two main components to the antenna — one is the feedpoint, and one is the loading coils.  Being able to clone the antenna successfully is dependent upon proper construction of these two components, because they determine the electrical performance of the antenna.  Everything else is just mechanical.  So the first step is to look at these two components in more detail.


The Loading Coils

The original DX-DD manual describes the loading coils as ISO-RES, whatever that means.  The devices are obviously simple wire inductors, made from enamel-coated magnet wire.  The operation of the coils in the circuit is not described by the manual, but it does claim that they are not traps.  They are certainly not classical traps, with independent inductor and capacitor sections, nor are they the coaxial trap designs used by many hams, where the coax windings provide both the L and C values intrinsically needed to form the filter that is a typical trap.

I'll write later about what these devices might actually be doing, but for now, it is sufficient to simply describe how to faithfully reproduce them.  Using a set of precision calipers, I determined that the dimensions of a finished ISO-RES coil for the DX-DD has the following characteristics: I found some thin-walled PVC at Lowe's that just happened to have the exact OD required, and the fact that it was thin-walled is a bonus for weight savings.  I used the same wire (#18 enameled) to wind the inductor, and the dimensions in the end were very close to those measured on the commercial product.

I also tooks some small liberties with the new design, each of which turned out to work just fine.  First, I used 70 turns to form the inductor.  The extra L allowed the finished antenna to be slightly shorter than the commercial version, but resulted in resonances that were nearly identical.  AD used two sets of fasteners on each coil — one to secure the coil windings, and a second to attach the wire spans.  Since the fasteners are not load bearing, I used one set of #4 machine screws for both functions.  I also drilled some small holes at the ends of the coil itself, and threaded the magnet wire through these holes to help hold it in place while I installed the hardware.

The manual for the DX-DD says that the coils are coated with a UV-resistant coating, but doesn't identify it.  The coating that I used was a clear enamel, available in a spray can in the paint section of Lowe's or any other home improvement store.  The product name is Rust-oleum Crystal Clear Enamel, and the manufacturer's website says that their enamels have significant UV resistance.  Once the new coils were completely constructed, I masked the stainless hardware with tape (since that's where the wires attach), and then applied one generous coat of spray enamel to the entire coil assembly.  The resulting coil was very nice and shiny looking, and should be reasonably resistant to exposure to the elements.

The new coil is shown below, next to the commercial version for comparison.

Loading Coil
Figure 1: Completed Home-made Inductor

Loading Coil
Figure 2: Original Alpha-Delta Inductor

The loops formed to secure the longer inner spans to the center insulator (described below) were constructeed similarly to those on the loads.  I left just enough slack in the pigtails to allow sufficient wire movement due to wind, installation, etc.


The Feedpoint

This is where we have to discuss the difference between the ideal world and reality.  The DX-DD manual claims that most installations shouldn't require a choke on the antenna feedline.

Rubbish. :-)

All amateur HF antennas (especially the so-called balanced antennas like dipoles and loops) require a feedline choke when fed directly with coaxial cable.  People will argue with me about this — I don't care.  The vast majority of amateur radio HF antennas are compromised in one way or another.  That's why they call it amateur radio.  As a result, I3 current on the feedline is always possible, and not always easy to chase down.  Many antenna installations just shunt the I3 current to the ground rods at the house entry point.  That can easily hide the problem, but it just wastes power.  Even the DX-DD manual concedes that "if" you have RF common-mode issues without a choke, you should add a choke, and they describe how to build a simple one out of loops of coaxial cable.  In practical terms the implication is simple — use a coaxial choke with this antenna design.

There are two common ways to make a coaxial choke, sometimes called a 1:1 balun or a 1:1 current balun, for a dipole antenna feedpoint.  Which one you choose is mostly a mechanical consideration, but I'll describe both.

The first is some form of what some call the Ugly Balun.  It is an air-core, coax-wound, inductive choke that uses several turns of coaxial cable to choke the outer surface of the feedline.  They are very simple, use common feedline and tie-wraps, and can be assembled in minutes.  Some people wind these around a large piece of PVC pipe or other such form to make the inductor turn spacing uniform.  The downside to these devices is that an air core inductor has limited inductance, and at 80m, you need a lot of choke impedance to prevent significant "RF on the coax."  That means that you need a lot of turns to get an effective 80m choke, and all that coax starts to add up quickly — not just in terms of extra losses, but particularly in the amount of mechanical weight that must be supported to place such a device near the center of the dipole.

The second type of choke commonly used with dipoles are the toroid chokes.  These are the same kind of device as the air-core coax chokes, but they are wound around a large ferrite toroid (ring), which greatly increases the amount of choke impedance that is added with each turn.  As a result, such devices tend to be much smaller, much lighter, and much more electrically effective (more choke Z, but less coax loss — win win).  The downside to these devices is that they tend to require special coaxial cable for the tight windings, and a proper selection of the core material.  Examples of good toroidal choke designs can be found on the Balun Designs 1:1 Choke Product Selector.  Their devices include plots of impedance and choke impedance, which is very helpful when selecting a choke for a specific antenna.  Such devices can also be built by hand if you have access to the right materials.

For both my commercial and hand-made DX-DD variants, I chose the BD 1115 coaxial choke.  The 'T' and 'S' variants of this device have studs for attaching wires.  I also opted for an N connector, to make the coaxial connection waterproof.  This device has around 4,000 ohms of choke Z on both 80m and 40m — perfect for this antenna.  The current cost is just under $70, so it may not be for everyone, but the design of this device is one of the best for this kind of antenna.  I also used their ABS mounting plate, which I also used as a strain relief for the dipole wires.  Together, these two parts replace the plastic center insulator that came with the DX-DD, and I used a second one for my DX-DD clone.

If you use a different kind of choke for the feedpoint, do some reading to select a design that is effective for both 80m and 40m.  This will save you from a lot of unexpected issues later on.  The DX-DD manual describes construction of a minimal air-core choke made from normal coaxial cable, but for good 80m performance, I would suggest you use at least double the number of turns they recommend.

So now that we have covered the needs for the two main electrical components, let's consider the other materials involved.


Little Stuff - Wire, Clamps, Insulators, ...

The rest of the antenna is quite straightforward.  The tuning of the 40m section is largely accomplished by adjusting the longer wires, while the tuning of the 80m resonance is accomplished by adjusting the shorter wires at the end.  Adjusting either the inner or the outer wires moves the 80m resonance quickly, so small adjustments are recommended.  The outer wires also have a small effect on the 40m resonance, but not much.  These considerations will influence the material selection for the rest of the antenna.

AD has been shipping porcelain insulators with their antennas for quite some time.  This is important if you want to run SSB or CW at the kW level.  The voltages that build at the ends of the antenna on 80m can be substantial at high power levels.  Porcelain insulators are available at just about any hamfest, or online, and they are relatively cheap, so I would avoid polymer or rope-based insulators unless you are confident of the dielectric abilities of the material.  PVC insulators or rope alone might be sufficient for low power levels.

Since both wire lengths need to be adjusted, it's a good idea to make it easy to do so when you attach wires.  I adjusted the 80m sections at the insulated ends of the antenna, where it is easy to move the wire through the insulator, and then fold back the unused portion.

The 40m section is a bit more of a challenge, but since the 40m section is also fairly broad, the tuning isn't as touchy  You can adjust the wire length at the load, or at the feedpoint, whichever is easier.

The selection of wire is also largely a preference.  The commercial version uses solid (and hard) copper wire that is roughly #12 or #14.  This wire has the advantage that you can easily wind a few turns of it around itself to form a loop at the load or the feedpoint, simplifying the assembly and adjustment of mechanical joints.  It is more difficult to wrangle during assembly and disassembly, and it doesn't have the resistance to repeated flexing that stranded wire has.

I chose stranded wire, for ease of handling and improved flexibility, but that brought its own challenges.

You can't wrap tight loops of stranded wire back on itself like you can with solid wire, so to form the load-bearing loops on the ends of the wire spans, I used tiny saddle-clamps, that are normally used for making loops in steel ropes.  These are common at hardware stores, and relatively inexpensive.  You could accomplish a similar feat with tie wraps if they were large and very snug.  The little saddle clamps did a great job of holding wire loops together, though.  I bought enough to have two on each loop, but I found that one was enough to hold the antenna together, even under significant tension.

The ARRL Antenna Book has a chapter on materials, and a whole section devoted to wire selection with respect to dipole antennas.  It is worth a read.


So How Does It Play?

40m SWR Sweep
Figure 3: 40m SWR Sweep
(500kHz)

80m SWR Sweep
Figure 4: 80m SWR Sweep
(500kHz)

The antenna I constructed ended up with an inner span length of 34', measured clamp to clamp, and not including the support loops at the ends.  The outer spans were 6' measured similarly, resulting in an antenna that was about 84' overall by the time you include the loads, insulators, and feedpoint panel.  This is slightly less than the commercial version installed at the same height, but to be expected, since each of my coils had one extra turn.

With those measurements, the antenna resonated in locations nearly identical to the commercial version, when installed as an inverted V, with the center at 20' above ground, and the ends at 6'.

The SWR plots are shown at right, taken with a RigExpert AA-54.

As with the commercial version, the 40m bandwidth is enough that the entire band can be made to operate under 2:1 SWR.  I resonated the antenna at the bottom of the band since I don't usually work SSB, but moving the antenna up to the center of the band is a simple adjustment.

As with all shortened dipoles, the 80m bandwidth is considerably sharper.  I find that my radio's 3:1 tuner can cover probably 100kHz on 80m with this design.  That's enough to cover the data/CW bands, or any 100kHz of the voice segments.  For wider coverage, you'll need a wide-range tuner, or similar matching device.

The EZNEC+ model I made for this antenna shows that the feedpoint match can be improved by adding a small inductor, sometimes called a "hairpin," across the feedpoint.  The value needed is extremely site- and installation-dependent, but if you want to bring the match as close to 1.0 as you can without a tuner, experimenting with a shunt inductor is probably a good idea.  The model suggests a value between 1μH and 2μH as a starting place.

Even with the reduced bandwidth, the antenna is a good match for most NVIS operations, which tend to be concentrated on a few closely-spaced "watering hole" channels.  The antenna's shorter wingspan also makes it easier to install, since smaller spaces can be considered.  Since the antenna covers both 40m and 80m in a single wire span, the installation location can be long and thin (like alongside a road), and doesn't require the area required by independent crossed dipole installations used by the military and many hams.

As with many horizontal wire antennas, proximity to the ground affects the resonant frequencies of this design.  As the antenna is raised to increasing heights, the resonant frequency will also rise.  As the antenna is lowered, the resonances will fall.  As a result, it is best to raise the antenna to its full intended height and configuration before making irreversible tuning changes (i.e., before cutting wires to their final lengths).  Raise the antenna, measure it, then lower it to adjust the lengths as needed.

This project is ongoing, but the first draft is a winner.  This design has proven itself to me repeatedly for high-angle contacts.  Even when installed at relatively low heights, the antenna has enough intrinsic efficiency to provide the kinds of high signal levels expected by NVIS stations.

More to come.

Copyright (C) 2018-2022 by Matt Roberts, KK5JY.
All Rights Reserved.