Rebuild of the Beverage Antenna Head Ends

Lightning destroyed 2 out of 3 of the Beverage antenna head ends. These home brew devices select the normal and reverse direction. They also terminate the unused direction in a dummy load (resistor) to prevent reflections that would cause a bidirectional pattern. They contain transformers, relays, resistor and a bias-T to separate RF and the DC reversing voltage that share the feed line. 

There are articles for each of these devices and their matching reflection transformers. Rather than sprinkle the text with links to the articles I'll list them here. Refer to them for technical details since I will not repeat that information here. In this article the focus on the rebuild and design differences.

• North-south RG6 reversible Beverage
• Northeast-southwest open wire reversible Beverage
• East-west open wire reversible Beverage
  

The north-south Beverage head end was taken indoors before the lightning strike so it didn't suffer the same fate. That was because it, too, was malfunctioning. After completing a variety of other higher priority jobs the time had come to deal with the broken units. The top band season is well underway and I need the Beverages to hear stations that my big transmit antenna attracts.

Many of the components in the affected devices were destroyed, even the PCB copper traces. Since there was little to salvage I opted for a complete rebuild of the electronics. The plastic (ABS) enclosures and hardware were okay so I cleaned and reused them. Connectors and their wires were removed from the old boards and reused.

Hoping for good news, I removed the transformers and tested them. Several were okay but the rest had melted or shorted windings. Winding miniature binocular ferrite cores is unpleasant work and I wanted to avoid as much of it if as possible. The Fair-Rite cores are ½" on a side and the holes are ⅛" wide. I did not reuse the cores for the discarded transformers although they are probably okay. Time is of the essence and they're inexpensive.

I started with the north-south RG6 Beverage since it had no lightning damage. I removed and tested the transformers and they were both good. That was fortunate. The fault appears to have been one of the SPDT reed relays. It had a lot of material between it and the PCB from an insect infestation. A thorough cleaning did not resolve the problem. By holding my ear close to it while applying voltage (they switch very quietly) I could tell it wasn't healthy. They're sealed so I'll never know. The coil resistance tests normal.

I selected new parts and ordered them. The proto boards come from my stock. On a cold and miserable day I got down to work.

The new proto board happens to be designed to fit the Hammond 1590 enclosure. It's quite a common enclosure and it is available as plastic or cast aluminum. Plastic is easier to work with and unlike many other RF devices there is negligible benefit from an RF tight enclosure. Please note that the relay has a wiring error in this picture that I discovered during testing. Don't use it as a template! I was lazy and didn't bother to take another picture after correcting the error. You might find it educational to refer to the schematic (earlier linked article) and try to find my mistake.

Unlike many proto boards, the copper traces only connect two adjacent holes. To simplify the wiring there are several soldered joints above the board. Where I could stuff two wires into one hole I did. I am using the same type of RF choke and low voltage bypass capacitor in the bias-T. The new DC blocking ceramic capacitors are rated for 630 volts. They replace the 40 volt devices that lightning destroyed. Of course lightning induces a far higher voltage but these will have a fighting chance if the pulses are short enough. The difference in price is small and worth the experiment.

Reed relays are small and very good for small signal switching. But they're relatively expensive and two are needed since DPDT reed relays are rare or nonexistent. I decided to go with the larger and cheaper Omron G5V DPDT relay. It's small and sealed, and it is found in many commercial receive antenna products. That's good enough for me.

There is no flyback diode on the relay coil. I was in a rush and it can be easily added later. Alternatively, the diode can be placed in the shack controller, either on each control line or the positive line from the power supply.

Wiring and layout isn't critical for receive application below 2 MHz. The proto board is larger than what I used for the first Beverage head ends, which leaves lots of room. The wider separation of components eases construction.

When fully wired the head end is tested. I constructed a test jig on a small "push in" proto board, The analyzer or VNA is connected on one side and the head end on the other. A blocking capacitor protects the analyzer from the applied +12 VDC. The alligator clips connect to the power supply. 

To reduce stray inductance use shorter wires to the power supply or, better, insert RF chokes. For rough testing this setup worked well enough, but don't expect accurate reactance measurements. Resistors are placed across the antenna terminals to simulate the common (normal) and differential (reverse) modes of the Beverage. 

This isn't a perfect test since there is no true antenna, with a long wire (transmission line) and reversing transformer. The 75 Ω load resistor should see a matching resistance at its port via its transformer to minimize inaccuracy on measured mode. Again, for a rough bench test it isn't strictly necessary, but it may be helpful when interpretting measurements that differ from perfection. We'll see shortly that it did lead me astray in one instance.

The rebuilt RG6 reversible Beverage head end was reinstalled. The wiring error was discovered that evening. It was retrieved and repaired the next day. The next evening it worked as it should. That was one head end down and two to go.

If you refer to the article where I twinned the northeast Beverage to make it reversible you'll notice in the schematic that there is only one of the transformers connects to the Beverage wires. The other takes the common mode (normal direction) from the first transformer's centre tap. 

It was no surprise that when the transformers were removed from the lightning damaged head end that the first transformer was dead. The windings were burnt and shorted. The other transformer tested good and was reused.

Unfortunately the 4:12 transformer is more difficult to wind that the others. This is due to the large number of turns (16 in total) and care in construction to achieve best balance in the centre-tapped primary winding. I used a new binocular core and teflon sheeting. Teflon liners prevent wire abrasion by the hard ferrite (it's a ceramic). 

I purchased a small roll of 28 AWG magnet wire rather than the scavenged enamel wire I used previously. The enamel coating may have been abraded when it was removed from an old transformer and a few transformer failed due to shorted turns.

In retrospect it would have been better to use 24 AWG magnetic wire because the thinner wire is difficult to handle. It is also easy to damage the wire when stripping the enamel off the wire ends. For the number of winding turns needed for Beverage transformers there is no need for very thin wire. I often use insulated wire from discarded Cat5 cable for windings with only a few turns. When I do that I can often make do without the teflon liners.

I must be getting good at winding these tiny transformers because I got it right the first time. The impedance transformation of 9:1 (the turns ratio of 3:1 is squared) took the 330 Ω resistor to 37 Ω, which shows as a flat SWR of about 1.4 on the analyzer.

The new transformer was mounted on the PCB. The old transformer is also in the picture. After a bench test the head end for the northeast-southwest Beverage was reassembled and installed. It was tested that evening and although it worked there was a problem with the antenna itself that made it bidirectional. Nevertheless it was nice to have it back.

Suspecting lightning damage to the reflection transformer I trudged through the newly fallen snow to retrieve the transformer. It tested okay so I soldered it back in place and trudge back to reinstall it. I took the opportunity to replace the hastily repaired ground wire that an animal had torn off sometime over the summer months.

The next step was to retrieve and test the head end. After removing the cover it took less than 15 seconds to spot the problem. Instead of a wiring error it was a wiring omission. The picture above was taken before the repair so, again, you have an opportunity to troubleshoot it yourself. If you decide to take the challenge, stop reading now because I'm about to tell you what it was.

One of the two Beverage wire wasn't connected to the transformer. I had cavalierly dismissed a puzzling R component of the impedance measured while bench testing the reverse direction (differential mode). I was sloppy and I paid for it. 

The wire was installed and, without further testing, installed. An SWR sweep looked good and that evening it performed perfectly. That was a relief since CQ WW CW is this weekend.

I have yet to rebuild the east-west Beverage head end because I depleted my stock of proto boards. More will be ordered. With my temporary head end I now have 5 out of 6 directions working. It is east that is missing, and I can live with that for now since there are not many stations in that direction.

I plugged the weep holes I had drilled in the old plastic enclosures. The enclosures seal well on their own and the weep holes only encourage insects to make their homes inside. The moisture and detritus the critters introduce cause corrosion. We'll see how well this works. In any case, next summer I will likely remove the electronics. I want to be prepared for a third(!) lightning strike.