Beverage Lightning Strike 3: Am I Out?

In late summer there was a DXpedition (I forget which) that prompted me to look for them on 160 meters. I had just modified my antenna selection software to include the 160 meter mode of the 80 meter array. It's not a great antenna on 160 (base loaded 80 meter vertical) but at least I have it; the primary 160 meter antenna is not available during farming season because the radials are rolled away.

The northeast-southwest and east-west Beverages were intermittent and the north-south Beverage didn't seem to work at all. I suspected a fault in the remote Beverage switch. I did bother attempting a repair until after the hay harvest, and then I was distracted by more urgent antenna projects. 160 meters was not a priority until early fall and wasn't free to look into the problem until late October. By then I had to move quickly because major contests were on the horizon and there were several ongoing DXpeditions that I needed on 160.

[Note: Rather than repeat technical detail, please follow the links to the articles where that detail is provided. This includes each Beverage, the switching system, lightning damage and protection measures. Without that background, parts of this article may be mystifying.]

My first step was to open the head end for the north-south Beverage. There was some copper-coloured water pooling in the bottom. I noticed that the orientation of the box had the weep holes on top, which let water in but not out. I poured out the water and brought it indoors for a closer inspection.

That was when I noticed that the problem went well beyond a bit of water. The components and wires near the corner where the water pooled were easily pulled off the PCB. I removed the head end from its enclosure and turned over the PCB.

Yikes! It was lightning, not water damage. This was the third time in just a few short years. Here are links to the first and second instances.

Tracing suggested that the lightning travelled from (or to) both inner and outer conductors of the RG6 run to the Beverage switch. There is a GDT (gas discharge tube) from the centre conductor to the ground terminal but lightning from the outer conductor jumped across the PCB pads to reach the GDT. The solder was blown off and so were a couple of component leads. The picture was taken after I removed damaged components and cleaned most of the char from the pads.

I walked to the basement patch panel and did a resistance check of the control lines to the Beverage switch. There was an open circuit to the relay for the north-south Beverage, while the others had the expected resistance for the relay coils. I disconnected the switch and brought it indoors. There was no visible lightning damage. I repeated the resistance check and got the same result. At least the long run of control cable appeared to have survived.

I don't know when the lightning damage occurred. Perhaps I wasn't home at the time or I didn't notice a particularly close strike. After protecting all the head ends with GDTs, I chose to leave them in place over the summer in case I needed them to chase the occasional DX with the base-loaded vertical. That may have been a mistake, or perhaps the mistake was that I had not yet placed GDTs on the control lines or the Beverage ports at the switch. But I really should have done both.

Direct testing of the other two Beverages at the switch showed that they were behaving normally. I patched the northeast-southwest Beverage to the transmission line back to the shack and it tested good -- that meant that I could temporarily get by with that and the short east-west Beverage. None of the other coax runs or head ends were damaged, or at least not to the point of malfunction. Lightning may have reached them and they are protected by their GDTs. My time was limited so I decided to ignore them until later. An inspection is warranted despite their proper functioning.

Fearing more damage, I retrieved the reflection transformer from the end of the north-south Beverage. Other than some insect debris that I cleaned out, the transformer looked and tested good.

Here is where the peculiarities of lightning surprised me. Other than the components and solder being blown off the PCB and leads burned off, there was no damage! One of the two 0.1 μF coupling capacitors had one lead burned off right to the capacitor body. Yet it tested good, as did the other. The RFC was similarly fine despite one blackened lead. The DPDT relay also worked. There was an unrelated cold solder joint that was dislodged by my handling, not by lightning, which I repaired.

Since the PCB has plenty of space, I cleaned and then soldered the damaged wires and components to clean pads. With that done the head end worked as designed. You will notice that the GDT and resistor assemblies were removed. Although GDT will handle numerous lightning surges, the number of them it can endure depends on their peak current and duration. The parallel 33 kΩ resistor increases GDT life by providing a continuous discharge path for charges that can build up on the long Beverage wires from precipitation static and distant lightning.

The datasheet for the Bourns 2057-07 provides estimates for the type and number of surges the GDT will survive. Unfortunately there is no easy way to learn its condition when it's deployed in the field. Since a GDT fails open, it will pass a resistance test.

The GDT displayed evidence of a strong surge. The picture is of the one that I believe took the brunt of it. The GDT-resistor assembly measures open despite the resistor. When isolated, the resistor tests normal. Although my phone camera won't focus at a very close distance, there is just enough resolution to explain the measurements. It is clearer with inspected with magnifying glasses.

The solder on the front (bottom) lead is mostly gone! It is better on the other side but there was a powdery white residue that was easily scraped off with a pointed tool. Since there was enough energy through the GDT to melt or ablate the solder on the leads, I believe the GDT ought to be replaced. 

I will do the other two as well just to be certain. They are inexpensive to replace -- I bought 100 of them -- but a few minutes must be spent doing the replacement, including the resistors. I am also thinking of adding a fourth GDT to the coax centre conductor for complete coverage of the wires entering the head end.

Since we're past lightning season and I was pressed for time leading up to CQ WW CW, I deferred the job. I will also have to inspect the head ends of the other Beverages for signs of lightning damage since they share a common ground at the remote switch. There is time before the 2025 storm season.

We're not done yet. The remote switch was damaged and also had to be repaired. There was no visible damage despite knowing that it had failed. It was easy to confirm with an ohmmeter that all of the relays needed to be replaced.

I cut out and discarded the damaged SPDT reed relays. The copper windings of the output transformer were so corroded that I cut them off the ferrite binocular core and wound a new one from insulated tinned wire. The repeated insect infestations brought a lot of moisture into the enclosure. It's an impossible choice between weep holes and blocking entry to the tiny insects that can fit through the weep holes.

Reed relays work well for small signal switching, and that's what I went with the previous time the switch was damaged. However they are typically twice as expensive as conventional sealed relays. I have a stock of Omron G5V DPDT relays, and with only 3 days until CQ WW CW I decided to use them. There are better choices available even if the G5V is good enough.

I selected a small proto board to mount the components. It fits into the small space between the barrier strip and F connectors. To ease maintenance I used 16-pin DIP sockets to mount the relays. I had exactly three left from my decades old stock!

Since the relays are DPDT and the switch requires SPDT, I wired the two sides in parallel. If one set of contacts develops excess resistance, my hope is that the other set will provide redundancy. That is, if they don't age at the same rate! In case of degradation or lightning damage, the sockets allow for easy substitution.

I tested the RF and DC paths once it was assembled. There are 4 DC control lines for the coils of the 3 Beverage selector relays and one that feeds DC via a bias-T circuit to reverse the Beverage. There is a fifth control line for ground. The 5 lines pass through RF chokes.

All of the conductors into the enclosure ought to be protected with GDT. I will do that later. It's so tight inside the enclosure that I may need to move the switch to a larger one.

Here's a picture of the completed switch. The circuit is unchanged but it has a different format. The 3 control lines on the left are for selecting each Beverage, and the next 3 are currently unused. the rightmost control lines are for reversing the selected Beverage (via a bias-T) and ground. The PCB sits freely since the small proto board has no mounting holes. Cardboard isolates the PCB from the enclosure surface (ground). Plastic is a better choice since cardboard will absorb moisture. I'll deal with that when I pull the switch after the winter season for further work.

The Cat5 cable is about one foot long and is connected to the long Cat5 cable to the shack with crimp connectors. The reason for this change is to make it easier to remove the switch for service. The day before CQ WW CW was close to freezing and it was raining. I had no choice but to work outside in the miserable weather, but I tried to simplify the task with those crimps. I was relieved that the Beverage system was once again operational. It came in very handy on 160 meters during the contest.

Strike 3: am I out?

Beverages are simple and effective directional receive antennas for the low bands. That's hard to beat. However they have their quirks. One is the risk to the long wires from weather, wildlife and trees (growth and dead fall). Another is lightning. Strikes from nearby lightning induces a large voltage and current flow along those long conductors, and that energy has to go somewhere. Direct strikes are of course worse though less likely.

I can keep repairing the antennas and continue to improve lightning mitigation measures. They can be disconnected during the spring to early fall storm months but that would require temporary reconnection or other receive antenna solutions for summer contests and DXpeditions.

The obvious alternative is vertical arrays. These range from compact pennants to large receive 4-squares and 8-circle arrays. They are more immune to lightning and the latter two can outperform Beverages. They also have disadvantages. They are so complex that in most cases it is better to buy an expensive commercial product than to build one from scratch. 

A second problem is that they require pre-amps since they are very inefficient. For the top band DXer that is not an impediment, but it is for multi-op and SO2R contests. Filters and other measures are typically required to prevent the other stations' transmissions from overloading the pre-amps of the vertical receive array. I know contesters that had removed these systems from their stations for this very reason despite their superior performance.

I have enough land that I could install a vertical receive array far from the transmit antennas. Perhaps I will. If it works out I could decide whether to abandon the Beverage or to use the new antenna to complement the Beverages. 

Future lightning repairs to the Beverages might eventually push me to make the switch.