Reversible RG6 Beverage Antenna

I am on course to install more Beverages as my preferred low band receive antennas. As I mentioned in a recent article on my revised receive antenna site plan I had wavered on whether to use Beverages or vertical arrays. I have the room to do either, or both for that matter.

My plan relies on reversible unidirectional Beverage antennas plus a remote switching system conveniently controlled by the operator. Although less of a challenge than what I faced with my 80 meter vertical yagi I wanted to be sure I got it right before going further. I built one of these antennas and a rudimentary direction selector to determine whether it works, how well it works and to understand the challenges involved.

At this point I am convinced so I will proceed with one more of these in the coming weeks. The remote switching system has been designed and will be added this spring if time allows or in early autumn. All the required components are in hand.

Reversible Beverages are well covered in the amateur radio literature and there are commercial products for those who value convenience. Beyond a brief description of how these antennas work I'll focus on my choices, technical approach and installation challenges that may be less well covered elsewhere. I hope that that will be most useful to others and encourage hams to build one of these antennas. There is great satisfaction in building rather than buying.

Overview

Beverage antennas can be bidirectional, unidirectional or reversible unidirectional. Reversible Beverages cut down on cost and and land use at the price of electronics for switching and tuning. The antenna "wire" is a transmission line. Traditionally this has been open wire line or commercial ladder line. Coax is perfectly usable and can be cost effective and convenient. The electronics (transformers) are easier however the physical challenge can be greater.

A nice graphic of a reversible coaxial Beverage antenna can be found in ON4UN's Low-Band DXing book. Any coaxial cable can be used provided it can withstand the rigours of horizontal suspension over a long distance. The transformer turn counts depend on the coax impedance and diameter, as I'll discuss below. If both directions are not needed simultaneously a single feed line and a switch are used. That's how I built my antenna.

From ON4UN's Low-Band DXing

Now a few notes on the basic functioning of a Beverage antenna. It is little more than a long wire suspended a short distance above the ground. Imperfect ground is required for it to work. Without delving into details, the Beverage has gain in both directions off the ends of the wire.

For the typical unidirectional Beverage the signal energy of the unwanted direction is dissipated in the termination resistor (including resistance of the ground connection). Left unterminated the signal reflects from that end due to the high SWR and that feature makes the antenna bidirectional. Some hams see value in this though most choose unidirectional for its higher RDF (receiving directivity factor).

The wanted direction is terminated with the impedance of the receiving system consisting of the transmission line and the receiver pre-amplifier. Terminations must closely match the surge impedance of the antenna -- typically 400 Ω to 700 Ω -- or some signal will be reflected and affect performance. Transformers match the Beverage impedance to the load. Since the Beverage is a non-resonant antenna with a flat impedance over a large frequency range transformers are perfect for this job.

A bidirectional Beverage works in the same fashion as a unidirectional Beverage. Rather than a termination resistance the signal at the far end is carried by a transmission line to the receiver. It can be a separate transmission line or, in the antenna discussed here, the coax being used as the antenna. The Beverage antenna "wire" is the large diameter outer conductor (shield) of the cable, and that is true for both directions.

The reflection transformer (T3) transforms the high surge impedance of the antenna (as measured between the outer conductor and ground) to the impedance of the coax and delivers it to the other end of the antenna. There it is picked up by T2 and delivered to a load. T1 picks up the signal in the usual Beverage direction, transforms the impedance and also delivers it to a load.

A reversible Beverage is actually two unidirectional Beverages. Follow the signal paths in the schematic from the ON4UN book until you understand how it works. It can be perplexing at first glance.

Note that the signal (either direction) does not see the electronics at the end of the antenna it first encounters; the signal energy at that point is zero and builds gradually along the wire until picked up at the far end. For the same reason increasing Beverage length increases gain.

Other than a very short Beverage there is no need for an amplifier, and therefore no added risk of amplifier inter-modulation products due to nearby transmit antennas. For a multi-op or SO2R contest that can be a considerable benefit.

Physical construction

Careful consideration to the coaxial cable mechanical properties is necessary when it is put under tension and suspended horizontally between supports. It is not easy to find these specifications for a lot of the cheaper RG6 on the market. The mechanical specification of a maximum of 69 lb (31 kg) tensile strength of Belden RG6 is at the upper end of the range.

I recommend no more than 10 to 20 lb tension for most RG6. This provides an allowance for poor quality and wind and ice loads. With closer spacing of supports the required tension is reduced. A messenger wire, cable or rope, or RG6 with an integral messenger cable are alternatives. All of these can double the cost of the antenna.

Gripping small coaxial cable for applying tension and anchoring is not difficult with small diameter rope. Some pros call it a "thousand mile knot" and they use it to hang large diameter hard line from towers is the equivalent (and expensive!) commercial grips are not handy. One product is the Kellem grip. The used grips for Heliax are far too large for RG6. Guy wire grips are not recommended since getting the right size that holds well and doesn't over-compress the coax is challenging.

If you want to try this start by bending the rope into a long narrow U shape with the open end facing the anchor. Lace the cable by wrapping both arms of the U at the same time around the coax, and repeating until you have at least 6 or 8 wraps. The U must be long enough that there is enough rope left to tie a couple of knots. Done properly there are no stress points on the coax and the grip is good enough for slippery PVC jackets. The rope easily slides along the coax when there is no tension but almost impossible with moderate tension.

Notice the F barrel connector on the RG6 termination. It protected the exposed centre conductor until the electronics were attached.

I put in about a day's work spread over a week to clear bush along the 150 meter run of the Beverage. The tree line that looks tidy in the satellite image (see previous article) is thick with trees, bushes, thorns and deadwood. Despite being only 4 meters wide it is difficult work. Winter is the best season for doing this due to the snow cover, no foliage and no bugs.

I selected trees along the route for hanging the RG6 then removed limbs, tree trunks and bushes that either impeded me, the coax or at risk of damaging the coax. After unrolling the coax along the path I installed hangers on the trees and cleared vegetation along the aerial path of the cable.

The hangers are made from PVC pipe and seem to work. At first the coax sits on the nails since the hangers can grab and kink the coax when tension is applied. When the job was complete I slipped the coax into the hanger slots. The wire retainers are optional and only used where there was risk of the coax popping out.

The span between hangers ranges from 8 to 15 meters. I am uncertain whether the coax will weather the elements without eventually deforming or kinking. I may have to resort to installing a messenger cable. I'll inspect it in the fall and decide what to do based on how well it handles the abuse. Foam dielectric is more easily crushed than solid dielectric such as in RG213.

Later the cleared vegetation was cut into small lengths and thrown back into the bush where it can rot in peace. None of it is suitable for firewood.As you can see from the pictures this work was done in the cold weather with snow on the ground.

During installation of the electronics 4' ground rods were pounded into the ground at each end of the Beverage. This was difficult since there are tree roots to avoid and at the north end the bedrock was close to the surface. In the latter case it took a dozen attempts to find a gap in rock that would permit the full length to penetrate. I am not using radials and based on performance (see below) they may be unnecessary.

Transformers

Several years ago I built a unidirectional Beverage matching transformer for practice. I put it to use in my first Beverage, the 170 meter long northeast unidirectional Beverage. That experience came in handy for this project. A few enhancements are needed to wind T2 and T3 to deal with the increased number of turns and a smaller wire gauge.

Following the guidance from ON4UN's book I purchased PTFE sheet to line the binocular cores -- Fair-Rite 2873000202. The plastic prevents abrasion and shorting of enamel coated wire on the edges of the core. A lifetime supply of PTFE sheets can be found online for a few dollars. The 0.1 mm thick sheets I selected works well in this application.

I am sure that even good quality vinyl electrical tape and non-adhesive plastic sheets will suffice for low frequency use. Use what you will. A better view of the winding method can be seen in the close up of the switching unit further below.

I use AWG 24 insulated wire from scrap Cat5 cable for the low impedance winding. This wire is easy to work and provides a "bed" for the high impedance winding made from AWG 30 enamel wire. Enamel wire is easy to scrounge from old RF chokes and from power and audio transformers. I use a jeweller's screwdriver to press the wire against the inside walls of the core to make room for additional turns. Otherwise the wire quickly consumes the narrow space.

Be sure to wind the primary and secondary in the same direction (clockwise or anticlockwise). Winding sense isn't important but for the reflection transformer you must connect together the same ends of the primary and secondary windings. Refer to the schematic above.

After winding each transformer I test it with an analyzer. For a secondary load I use a 470 Ω resistor to represent the Beverage surge impedance. The ideal input impedance is 75 + j0 Ω (SWR 1.5) across the frequencies of interest. This transformer measures almost exactly 81 Ω from 1 to 7 MHz with a load resistor that measure 510 Ω -- carbon composition resistors increase with age.

Direction switching

The ON4UN book includes no explicit description of a switching system. It is not difficult although there important details are easy to overlook. The final product fits in a small box and has ports for the feed line, Beverage coax and ground wire. The hand drawn schematic should be easy to follow.

A few ideas for my design come from the ON4UN schematic above and the RemoteQTH schematic for an open-wire reversible Beverage by OK2ZAW.

Layout is not critical. At 1.8 MHz a little bit of sloppy wiring does not affect performance. The transformers wound on these binocular ferrite cores have essentially no coupling to each other or anything else. The enclosure must be non-conducting since earth ground and the outer conductors of the both coax ports are at different potentials or DC isolated. With the enclosure closed you can tell which is the antenna port since it has DC connectivity to earth ground via transformer T1 (lower right). Choose UV resistant plastic for the enclosures.

On the left DC and RF are separated by a capacitor and choke. The coupling capacitor should be at least 0.1 μF so that the impedance is small enough not to cause signal attenuation at 1.8 MHz. I am using 0.2 μF and 0.5 μF will do well down through the AM broadcast band. The choke should be at least 100 μH to prevent leakage of RF into the DC circuit. I am using a 1000 μH choke with a self-resonant frequency of 1.2 MHz, well outside all amateur bands, and a current capacity of 300 mA which is far more than the relay coils draw. Make sure the self-resonant frequency is outside the amateur bands before you buy.

The SPDT reed relays are suited for small signal applications and rated to -40° C. A better alternative is a DPDT small signal relay such as the OMRON G5V-2. It is specified for the aforementioned RemoteQTH design. I have a bunch of reed relays on hand so I used them. The 12 VDC relay coils operate in parallel to switch direction.

Some relays have a built-in reverse EMF protection diode. These don't so I used a 1N4001. My SPST reed relays do have them so it is important to apply the correct DC polarity. I did not use a diode to protect against reverse polarity since it will be located in the remote switch that feeds switching power to the Beverage..

The south signal flows through the top transformer and relay and the north signals flow through the bottom transformer and relay. The NC and NO contacts of the north and south relays, respectively, connect to the feed line via the coupling capacitor. The respective NO and NC contacts connect the unused direction to the 75 Ω load to prevent reflection.

The load resistor ought to be 2 watts carbon composition. From my junk box I wired two ½ watt carbon composition resistors in parallel to get 75 Ω and 1 watt rating. The 2 watt rating can be had with 4 inexpensive ½ watt 75 Ω carbon composition resistors. The solder terminals for the resistors are raised above the board so they are easily accessed for replacement if a nearby lightning strike destroys them.

Hardware is stainless and spacers under the bread board are plastic. The wing nut on the ground wire lug uses a wing nut so a wrench is not needed. The bread board has two horizontal rails at the centre of the long dimension. I used one for feed line ground (coax shield) and the other for the +12 VDC relay voltage.

All the DC switching paths were tested without the transformers on the board and winding continuity was tested when the transformers were mounted. The reed relays are so quiet an ohmmeter was the only reliable way to determine whether the relays were operating. The RF signal path could not be easily tested and as it turns out there was a wiring error.

Getting the bugs out

There is nothing electrically complex about this antenna. Other than ground rods all of the important components are visible in the adjacent photo. Nevertheless it is easy to make mistakes, and those mistakes can be difficult to troubleshoot.

I made one that resulted in nothing heard. It turned out to be a simple wiring mistake. I diagnosed it by looking at the photo of the switching unit. As an exercise try to discover the error yourself before reading further (see earlier closeup).

With everything connected I went back to the shack, turned on the rig and listened. Since it was daylight I didn't expect to hear much but I heard nothing at all. My antenna analyzer showed an SWR plot of a long and unterminated coax. Applying voltage to switch direction had no effect.

With an ohmmeter I confirmed that the coax was connected to switching unit by the presence of relay coil resistance. Before running out to the field to retrieve the units I studied the photograph I of the switching unit. To my surprise that was enough to fully diagnose the problem. With sunset approaching I brought the unit indoors and confirmed that my suspicion was true. A soldering gun and a bit of wire later the unit was ready to go again.

The mistake was that I had taken the wire from the relay contacts for the receive path to a bread board pad and then soldered in the DC blocking capacitors. But these went to different columns of pads. The new wire connects the capacitors to the signal path wire. This was faster than pulling the board and redoing the job properly.

With that done I tried again and had success! It was sunset so testing could began immediately. More on this below. Despite many decades of being a ham and endless antenna projects I still get a thrill when a new home brew antenna is connected and it works. There was a smile on my face as I repeatedly switched Beverage direction, compared signals strengths and contemplated how a hunk of wire barely higher than my head could perform radio reception magic.

Temporary direction control

The default direction is south, that is, the direction with no relay power applied. North will be used relatively rarely to work on sunrise openings to Asia and during other brief propagation opportunities. This is a good strategy for all directional switching systems and not only for receive antennas.

I built a temporary device to inject 12 VDC into the coax from the operating position. It doesn't get any simpler than this. I removed the resistor network from a QRP attenuator and soldered in a 0.2 μH capacitor and a 100 μH RF choke. A wires snakes around the sheet metal where it can be accessed.

To switch the antenna to north 12 VDC is applied between the chassis and wire. Polarity matters since the diode across the relay coil will short with the wrong polarity. I didn't bother with polarity protection for this temporary device.

The temporary switching system is ugly. The red clip lead is attached to the wire to point the antenna north. Two 9 volt batteries provide operating voltage of about 15 volts (one of the batteries is weak) which is well below the maximum voltage specification for the relay coils.

Some care is needed to avoid shorts. It is critical to get the cabling right otherwise DC will be injected into the receiver front end and that can be a very expensive mistake.

Performance

The first test was the SWR. When properly built and adjusted the impedance should be 75 + j0 Ω across a broad spectrum. Of course that ideal is never reached so it is fortunate that perfection isn't necessary for excellent RDF. Beverages are more forgiving antennas than driven vertical arrays.

This is very good! Some ripple of impedance versus frequency is typical. The flattening of the SWR at higher frequencies could be surge impedance variation with frequency or, more likely, due to increasing transmission line attenuation with frequency. The transmission line is approximately 150 meters of RG6 plus another 150 meters of RG6 in the south direction since that signal comes back to the feed point via the Beverage's RG6 span. Coax loss reduces signal level but does not affect the antenna pattern.

Signal level seems comparable to the 170 meter long northeast Beverage. A direct comparison isn't possible yet since there is no Beverage switching system. Theoretically the gain should be perhaps 1 db worse for north and 2 db worse for south. There is no need for amplification.

The biggest obstacle to listening tests was finding enough stations! Despite the pandemic confining hams to their homes the springtime decrease of low band activity is striking. North is always a problem since it is really only Asia in that direction and that is a difficult path at the best of times. To the south there is little activity in the Caribbean and South America and US station location cannot be determined from their call signs.

Despite that difficulty I cound enough stations to determine that the antenna does work, and it works well. To the north on 160 meters the main lobe is wide enough to show improved SNR on European signals and high rejection when pointed south. I found more Caribbean and South American activity on 80 and 40 meters, where the main lobe narrows, to give the antenna a good workout. However the very narrowness of the main lobe and the many minor lobes on 40 meter made comparisons a challenge.

All the South American activity I've monitored on 160 meters is FT8 so that's what I tested with. The evaluation was qualitative since signal strength readings in WSJT-X can be deceptive. Good signals from PY, LU and VP8 in the south direction disappeared or became too weak to decode in the north direction.

Comparisons on 80 and 40 to my yagis -- both of which have only modest F/B -- was very promising. The Beverage greatly improved SNR for stations to the south. Due to the narrow main lobe I wasn't able to find suitable stations to the north on those bands for make a definitive conclusion. I didn't get up early enough to look for east Asia activity (north direction).

Summing up

The antenna works to my expectations so I will continue with the rest of the Beverage plan. I will try to build the east-west reversible Beverage and the remote switch this month before the bush if full of ticks. It is possible to take protective measures but it is not worth the effort and risk. If I still have enough time the northeast Beverage will be "twinned" to make it reversible. What doesn't get done now I can continue in the fall.

The north-south Beverage has been disconnected and the feed line moved back to the northeast Beverage. Since I can only have one of the Beverages connected for the time being this is the one that is most useful. Soon the primary 160 meter transmit antenna will come down for the summer and my low band activity will slow considerably.

Next winter I will have a higher performance low band antenna system, both receive and transmit. I expect it to be well worth the time and effort invested.