Reversible Short Heliax Beverage Antenna

I have a lot of scrap coax. What to do with it? Sure, I could throw it out but most of it still works pretty well at low frequencies even though a full spectrum sweep can look especially ugly. I persist in trying to come up with a use for it all. There are hundreds of meters of the stuff.

At left is an SWR plot of a 40-year old 40 meter length of LDF4 terminated with 50 Ω. It has seen some rough handling over the years. I have 3 of these rolls in my scrap pile. They take space that I'd like to reclaim, but I haven't been able to talk myself into hauling them to the local landfill for disposal.

The plan

I would like a second set of receive antennas for multi-op contests. The existing 6-direction Beverage system can only be used by one station at a time since only one of the 6 directions can be selected. It can be fed to both stations if the operators are happy to use the same direction, but that's unreasonable. All I could do for the recent CQ WW contests was to switch the system to one station or the other. The one without had to make do. 

That isn't so terrible since the 80 meter yagi provides pretty good RDF (receive directivity factor) on receive. Unfortunately the yagi only functions between 3500 and 3650 kHz. That isn't enough for a phone contest. Extending the yagi's range to 3800 kHz may be done this year, but is more likely to be delayed until 2025.

I have a need and I have a lot of junk coax. I decided to use the coax to make a reversible short Beverage pointing east and west. It's short so that the azimuth pattern is broad, close to 90° in each direction, which covers half the compass, and with a low but useful RDF -- I can add a south Beverage or other small directional receive antenna later. 

I am basing the antenna on my experience with a short, 90 meter long uni-directional Beverage that I had for a short time a few years ago. It worked very well. In fact, I put the new antenna in almost the very same place, along the fence and tree line that separates two hay fields. To make it reversible, I used the same design as for the reversible north-south Beverage made from 150 meters of RG6. The only difference is in the transformers to accommodate the change from 75 Ω to 50 Ω coax.

I picked the two best of the 40 meter lengths of LDF4-50A. The low frequency SWR sweep of both are nearly identical. The third roll (wide sweep shown further above) is decidedly worse. I measured both lengths and then dug into my piles of short Heliax scraps to make up the difference to reach about 90 meters of total length. 

The two lengths were placed on top of the fence line with the help of a friend and joined them with an N barrel connector. I needed his help to thread them through the trees without kinking the cable or damaging the jacket. For half the distance the bushes are thick and thorny so it wasn't much fun. 

To keep the cable from drooping between supports (the log fence zigzags), rope was tied on at both ends, pulled taut and the ropes tied to conveniently placed trees.

That was in early fall. There it languished until mid-December when I had the time to resume work on it. Unfortunately it could not be prepared in time for the CQ WW contests. We made do with switching the existing Beverage system (6 directions) between stations. I would like to have the new Beverage ready for the ARRL DX contests.

Modelling

Before proceeding to wind the transformers and build the switching system I turned to modelling. Although 90 meters is a good length for a short Beverage, it is unusually low and "fat". I expected these factors to affect the optimum length, and I was right.

I tweaked the length until the F/B was best. That was 86 meters. I selected one of the scrap lengths of Heliax pictured above and added it at the east end of the Beverage to make it that length. Of course the NEC2 model likely has some inaccuracy due to the ground proximity and largely unknown composition, but it doesn't hurt to at least try to get it right. 

If it turns out not to be optimum the impact is small since the RDF is unaffected by small length differences, just the F/B noticably changes. The RDF in the EZNEC model can be calculated from the difference between the gain (-15.1 dbi) and the average gain (not shown) of -24 db. Therefore the RDF is approximately 9 db. That isn't great but it is better than almost all compact receive loops like the pennant, EWE and others.

Although the 90° azimuth beam width seems poor, in my use case that's an advantage. The objective is to keep it simple for the operator for the second station in a contest. This is not my main Beverage system where better performance is needed for both contests and DXing on top band. The RDF is fine for cases where the reception is poor on the omni-directional 160 meter transmit antenna. It is less helpful on 80 metes due to the narrower beam width though, as we'll see, it is still useful.

In time I would like to supplement the east and west coverage with a short south Beverage. It would have to be a winter only antenna since it almost certainly have to be run across a hay field. There would be no point in making it reversible to the north because there's little for us to work in that direction. There is the long north-south reversible Beverage in the main receive antenna system for those brief periods when we look north towards Asia.

Construction

The transformers are the key to good performance. I wound the transformers before building the rest of the electronics. I used the same binocular Fair-Rite transformers as in the past, and are typically used by pretty much everyone else building Beverages. This time, rather than use Teflon liners, I used insulated wire salvaged from scrap Cat5 cable for the winding with the fewest turns. The insulation served as a protective bed for the second and longer winding made from coated magnet wire.

The annotated diagram from ON4UN's Low Band DXing book shows the transformer turns for this implementation. The same book estimates that the surge impedance for this Beverage should be approximately 350 Ω. The reflected signal (west direction) travels through the 50 Ω Heliax and the feed line 75 Ω, made from a combination of old RG59 and new RG6. I used the handy table in the ON4UN book for winding turns for the various transformation ratios I needed.

All of the transformers were tested on a VNA using a suitable termination resistor. I didn't bother testing the insertion loss since it was certainly better than -1 db in all cases. It is not critical for a receive antenna as long as there's enough gain that atmospheric noise can be heard in the receiver. Beverages rarely require auxiliary amplification. Use the rig's pre-amp if necessary, and a BPF if there's a strong broadcast station nearby. In my remote QTH there are no nearby broadcasters so the only strong signal is from my second station during contests. I haven't yet needed a BPF to deal with it.

I selected enclosures for the head end electronics and for the reflection transformer. I reused a box from a retired Beverage head end for the new one. I enlarged the hole for the F connector to install an N connector and added a stud for the ground wire. For the reflection transformer I chose a very small plastic box. It is almost dwarfed by the N connector that barely fits on its surface. But it was cheap and the transformer is small.

The completed head end demonstrates that sloppy construction is usually acceptable at low RF frequencies. I like that since I'm a disaster waiting to happen when working on circuitry. A small proto-board loosely secured supports all but the GDT assemblies for lightning protection. There is generous wiring to the connectors to ease removal for service. There is no need for metal enclosures that will complicate construction since the coax connectors don't share the same ground.

I made mistakes: one cold solder joint, one no solder joint and one transformer installed backwards. All were easy to fix. I'm prone to carelessness when working on circuitry, whether it's commercial equipment or my own projects. Bedlam usually ensues when I open up a transceiver.

For testing, I built a bias-T interface on a proto-board to inject 12 VDC to energize the reversing relay in the head end. It consists of a 0.1 μF capacitor and an RFC. Although, as mentioned above, sloppy wiring is acceptable at low frequencies, you can see the rising SWR at the right of the analyzer screen (7 MHz) due to the inductance of the long leads used for testing. Resistors (330 Ω and 50 Ω) represent the Beverage loads. The nominal transmission line impedance is 70 Ω or 75 Ω, so an SWR of 1.5 is what to expect on a 50 Ω analyzer.

The head end was installed at the west end of the Beverage. I used RG59 for the feed line into the house since it was the right length and I had no other use for it. I tested it beforehand to ensure it was still okay, at least at low frequencies. For the rest of the run into the shack I used RG6. I used F twist-on connectors and F barrel connectors between coax segments. The RG59 was laid on the ground at a right angle to the Beverage to minimize interactions. I'll lift it onto an aerial messenger cable later this winter or in the spring to get it off the ground.

The tiny enclosure for the reflection transformer is dwarfed by the Heliax it plugs into at the east end of the Beverage. Both ends are grounded via 4' copper-clad ground rods for the terminations. These have proved adequate for all of my Beverages even though their ground impedance isn't great. Beverages are pretty tolerant, which is one reason I like them. In other installations it may be necessary to improve the ground rods or add short radials.

Despite all the below grade rocks and tree roots I only had to make 6 trials to drive in the two ground rods. I got lucky. In one case a few years ago it took twice as many tries to drive in one ground rod. You pull it out and try again a foot away. Trial and error is necessary.

The squirrels began exploring and abusing the Beverage as soon as it was pulled along the fence line. The Heliax became a convenient rodent highway that perhaps saved them the effort of jumping from log to log. The ⅛" nylon rope at the head end was soon chewed through so I replaced it with scrap insulated copper wire. 

I suspect that it was rodents rather than the usual deer that recently chewed through one of the radials on the 160 meter transmit antenna. That was easy to fix. I'm thankful that they have never (yet) chewed on the many hundreds of meters of RG6 running through the bush to the various Beverages.

When I first connected the antenna it didn't work. After some frantic troubleshooting I discovered that the ground wire to the head end had fallen off. Once I screwed it back on the Beverage worked as expected in the east direction. The SWR curves are hardly perfect and the scraps of feed line didn't help. I've found that Beverage antennas are very tolerant of impedance matching challenges like these.

The impedance centres near 75 Ω (SWR of 1.5), which is per the design. The cycling is most likely due to a surge impedance that differs from the design objective of 350 Ω. There are many factors to consider, though the most likely one in this case is the ground connection. The ESR (equivalent series resistance) of these ground rods in my soil is typically about 100 Ω. I could have wound the transformers for the higher net impedance of about 450 Ω, but that isn't necessary. 

The impedance in the reverse direction displays more anomalies. Those are almost certainly due to the poor quality of the Heliax. It is only in the reverse direction that the Heliax is used as a transmission line to transport signals coming from the west via the reflection transformer at the east end of the antenna.

I made a switch-operated bias-T to inject 12 VDC onto the coax for the reversing function. The F connectors are for the Beverage systems and the BNC connectors are for the receive antenna ports of two radios. At the moment the new Beverage is switched manually. The switch on the left, not yet operational, will be used to manually swap Beverage systems between the two radios. The rightmost F connector is for a possible south Beverage to be used in conjunction with the new east-west short Beverage.

Assuming that the system works well, a PCB with relays will be installed for software control of system swapping and direction control of the new Beverage. It will be point and click as for the main Beverage system. Implementation is straight forward but it is not a priority. I'll have more to say about this project after it is built.

Testing

I compared the new Beverage to the higher performance Beverage system on 160 meters. I used two receivers for the test so that I could easily listen to both antennas concurrently. Comparing the reception of signals was interesting.

Signals from Europe (northeast) were about the same on the new Beverage as on the longer east Beverage. The long east Beverage has a sharper beam width so its gain falls off to the northeast. The long northeast Beverage is, unsurprisingly, the clear winner for European signals. 

Signals are stronger on the long Beverage since Beverage gain increases with length. However, that does not indicate quality of copy since both signal and noise increase. Direction and RDF are what matter. Use the receiver pre-amp if necessary to raise the antenna noise level above the receiver noise floor.

Testing the F/B and the west direction had to wait because at first the antenna didn't work in the reverse direction. Troubleshooting found that the N panel jack on the enclosure with the reflection transformer had a broken centre pin, the ground lug was loose and the transformer was wired backwards. So many mistakes in such a small box! Well, that's not unusual for me.

When I did get it working the F/B was tested on 160 meters with stations in Europe and W7. It works very well in both directions. Performance to southerly directions was unpredictable, depending on where the signal fell into the off-the-side nodes and nulls of the Beverages being compared.

I next turned to 80 meters. It is interesting to compare the patterns of the 3-element vertical yagi and the short Beverage. There is of course a large gain difference but that is not entirely relevant. It may matter when switching from one to the other since it may also require turning the receiver pre-amp on or off. 

I live with it even though I could install an external amplifier for the Beverage to better equalize signal levels. I don't because it is common for a pre-amp, even one with a BPF in front of it, to be overloaded by a nearby kilowatt transmitter. That's only a concern for SO2R and multi-op contests, which is different than for most low band operators.

In my brief testing on 80 meters the short Beverage works well. The RDF of about 11 db, which is 1.5 db better than the yagi, may prove useful for digging out weak stations during contests when the main Beverage system is being used on 160 meters.

Done

With this antenna completed, that's one more check mark on my long list of winter projects. It is unlikely that I'll pursue manual or software swapping of Beverage systems during this contest season. I will instead turn to other projects.

If you have the space and one or more old rolls of retired coax, consider a project like this rather than tossing them into the garbage bin. It's a simple and inexpensive way to improve low band receive performance. The improvement can be especially impressive if your main antenna is not directive, such as a vertical or inverted vee. A 90 meter (300') long antenna will fit within many urban fringe or rural properties where a full size Beverage twice that length will not.