Beverage Remote Switch Repair (Again!)

Since earlier this spring I experienced intermittent functioning of the Beverages. With the 160 meter winter season winding down the problem wasn't high on my priority list. In September the problem worsened to the point that the Beverages were unusable. The top band season has now begun so it was time to deal with it.

This is not the first time I've had trouble with the Beverage remote switch which sits on the edge of a hay field not far from the "head ends" of the 3 reversible Beverages (6 directions). When the hay is high and the weather is warm the ticks are on the hunt. In any case the radials on the big shunt-fed tower are absent during the growing season to clear the ground for my neighbour's farm equipment. Radials and farming don't mix!

In early October I went out to the switch to troubleshoot. I had already determined that the long run of RG6 out to the field was fine. It is important to sectionalize a system and test each component if you hope to isolate and identify the cause of a fault. I next inspected the F connectors on the RG6 to each Beverage and tested them. The analyzer confirmed that the Beverages were all fine -- exclusive of damage due to summer growth and other damage. That's a different problem that I am dealing with in parallel to this one.

The last remedy I tried is one that should be approached with caution. Relay contacts can develop high resistance over time if they are not periodically wetted. Wetting means switching and conducting a modest current to break the thin layer of oxidation or contamination that builds over time. The small signal from a receive antenna is insufficient to wet the contacts.

The Omron G5V relays are good in this application but are not impervious to contact corrosion. Due to the design of the switch the DC reversing current can't wet the contact because the DC is injected at the relay outputs (the selected Beverage via a bias-T). Wetting of the relays requires RF. I moved the transmission line from the rig's receive antenna port to the main antenna port and transmitted a signal. 

Use only low power for the wetting attempt since the receive antenna electronics can't handle 100 watts. I gradually increased power to 10 watts while manually switching the Beverage selection. The problem persisted so that was unlikely to be the problem.

The picture above right (taken after the switch was moved indoors) shows how I most recently tried to keep insects out by taping shut the weep holes. The tape over the one on the right had been eaten through; ants always find a way. I tore it off before I took the picture. I prepared to find insect damage, again. 

Having localized the fault to the switch itself I opened its enclosure. It didn't look good at first glance. There is clear evidence of corrosion and insect activity. As we'll see in a moment it was worse than what could be seen with just the lid removed. 

I disconnected all the cables and brought the switch indoors for a closer inspection. When I disassembled it what I found was worse than I expected. We all make mistakes and hopefully you will learn from mine. I am not shy about exposing my failures in the blog since that is how we (me and my readers) learn.

This is what I saw when I turned the PCB over.

What a mess! Compare these pictures to what the switch looked like when I rebuilt it after a lightning strike two years ago. Can you spot the construction choice I made that led to the failure? Despite how ugly it looked, I took a few minutes to seat and reseat the relays in their sockets. That also didn't help. It was time to deal with the messy interior. 

Since the small protoboard has no mounting holes I needed something to insulate it from the aluminum enclosure. I looked around the table where I was working and saw a thin piece of cardboard that was exactly the right size. I knew that it was not a great decision but used it anyway since I was in a hurry to move on to other projects on my priority list. Insects and moisture decomposed the cardboard to it original components: plant fibre (tree pulp) plus a chemical soup.

I scrubbed the PCB clean, including using a wire to push out debris that clogged many of the pad holes. All copper pads and wires were corroded, and some pads were entirely missing; soldered pads survived. Tinned wires and component leads were fine. The stranded copper wires (salvaged from an old Cat5 cable) to the F connectors snapped off the PCB since there was little metal left for them to be structurally sound. As it turned out that was the proximate cause of the failure: intermittent or entirely disrupted continuity through those wires to the chassis F connectors for the 3 Beverages. 

The solid Cat5 wires on the leftmost connector were fine. Stranded wire corrodes faster since they have more surface area. For comparison, the tinned stranded wires to the barrier strip, though coated in insect debris were in good condition.

Many commercial products intended for outdoor use have a sprayed on dielectric coating on the underside of PCBs to prevent corrosion. As mentioned, tinned wires are preferred. It is not surprising that many electronic components use tinned leads.

After cleaning the board and sanding away stains on the enclosure I soldered tinned wires from the PCB to the Beverage connectors. With that done I proceeded to test the switch. It worked perfectly.

The Beverages switched, reversing current flowed and the impedance was per the original design. I was ready to put it back together and move on to the next item on my long to-do list.

Rather than repeat my earlier mistakes and oversights, I made he following changes during the rebuild:

• Tinned wire for the leads to the F connectors
• Silicone sealant on all the weep holes and around the (replaced) fibre rimmed washers on the transmission line connector since they degraded in the weather and needed moisture protection
• Thick teflon (PTFE) insulation under the PCB, replacing the cardboard I foolishly used last time
• GDT added for lightning protection on the control leads and each of the incoming Beverage connectors; these are complementary to the GDT in each Beverage head end

Aside from one cold solder joint it worked as it should. It has now been reinstalled and all is well. I can listen to activity on 160 meters even though my transmit capability is poor until I redeploy the radials on the shunt-fed tower. That has to wait a little longer to avoid tripping hazards until work on the HF yagis is done. The northeast-southwest Beverage still needs repair when I can safely enter the bush. The other two Beverages have already been repaired, other than needing to clear the bush after the ground has frozen.

With more time I would have made other changes to the remote Beverage switch. One is to put it into a larger enclosure. It's difficult to work on since the fit is tight. Also the layout of components is not ideal due to the limited space. 

Another change would be to use a plastic enclosure. There really isn't that much risk of coupling among the Beverages without a surrounding metal cage and plastic is both easier to work with and removes the need for isolating the transmission line connector from the metal enclosure with those flimsy fibre insulators. When the switch was first built I chose the enclosure material based on what seemed sensible at the time.

In closing I'll note that my Beverage articles are among the least popular on the blog. I write about the work I do on my own station, including my equipment and operating preferences, not what I think might be popular. Few hams have the land for Beverages and large top band antennas, which likely explains the low interest. Nevertheless the lessons learned can be applied to any antenna project. That, at least, makes writing these articles worthwhile.