<Click Pictures to Enlarge>

W9FI - Jim


A Remote Antenna Switching Project, by Jim Wysocki, W9FI

Now that I had the time to think about it, I knew that I needed a remote antenna switch for my post-retirement antenna system.  I didn't want to bring a lot of coax cables into the house for lightning protection purposes.  I also didn't want to spend a fortune on coax cable, since the distance from the antenna field to the house was going to be over 200 feet.  The new cables also had to be buried to protect them from the desert critters who like to gnaw on exposed wires.  So after the jackhammer dust had settled, there were two coax lines and a pair of telecommunications control cables laid down inside of some 3-inch flexible plastic drain pipe that was buried six inches deep.  The basics were in place.  Then it came time for the remote switch decision.  Exactly what switch should be used and what should be its operating characteristics?  

I had some unpleasant past experiences with remote switches that injected control voltages through the coax, in that they blew out in a "not-so-high SWR but high-power" situation.  So they weren't worthy of further consideration.  Separate control lines would have to be used in order to avoid this problem in the future.  Research showed that there are several commercially available switches that would meet my requirements.  But some brands had a poor reputation for quality control, and the higher-quality switches that remained under review came with a healthy dose of sticker shock.  There were also several homebrew designs available.  Most of them were designed around one-off, exotic, expensive, or hard-to-find junkbox parts.  

However there was one design by Ron, KK1L, that turned out to exactly fit my requirements.  It features a pair of double-sided circuit boards and uses commonly available parts.  There is a parts list at one of the electronic parts distributors (Mouser) that took all of the guesswork out of what was needed.   A previous builder had even measured the isolation on the various ports and found that at 55 to 61 dBm isolation between ports, so it was as good as or better than most of the commercially-made units.  After several weeks of investigating possibilities I had found what I needed.  And I could build it myself and save a few bucks.  

First I checked with Ron, verified their availability, and ordered the PC boards.  Then I ordered the components.  With a few key strokes I deleted the parts that I had on hand and orded the rest from the list.  All of them were off-the-shelf components: no custom-programmed memory chips, limited-edition integrated circuits, rewired relays, or modified electromechanical devices were required.


It's a double-sided, plated through the holes, epoxy board.  All traces are properly masked with a transparent green insulated coating.  Component placing is clearly labeled.  The holes are properly sized and spaced to match the exact part numbers as specified in the Mouser project checkoff list.

The same is true of the relay board.  Extra holes are drilled to accept the lead spacing of three different relays.  If the recommended part is out of stock, other relays can be ordered which will fit the board.  That's a nice design option that should be encountered more often than it is.


I have no specific metalworking skills and this project took shape with nothing other than hand tools, files, and drills.  I planned out every step beforehand, drew up the plans before I started, repeatedly redesigned the switch modules on paper, and took my time in building them.  The final results turned out better than I expected.  


There is a lot of hidden design and workmanship that's necessary to make something that's simple to use and easy on the eyes.  In this case KK1L's web site gave several examples of completed switch projects and one of them really stood out.  Charlie, WA3UTC, did an exceptional job of ergonomic design and had such quality metalwork that I contacted him for advice.  He was kind enough to give me some construction tips, and also to share with me some computer files of his front panel designs.  I modified one of them to use as my standard.  The back panel I designed myself, using LibreOffice's freeware "Draw" software.  These designs were printed on clear decal paper using a laser printer.  Then they were cut and transferred to the front and back panels, by soaking them off of their backing papers and sliding them into place on the finished panels.  Finally, the panels were heat-treated with a hot air gun.   As will be seen, this process was not without its problems.  Here's a view of the work in progress, just after a decal was melted into the back panel and before the first of three thin coats of clear spray was applied.



I built two controller units.  The first one was for the antenna switching control but another was needed.  When I bought an older but mint condition TenTec automatic antenna tuner I soon realized that its settings could be controlled automatically by connecting a second controller box to the first.  So I ordered a second board, another blank enclosure, the necessary parts, and built it.  The second controller's output and back panel had to be redesigned slightly and that enabled the two controllers to work together seamlessly.


Here's an interior view of the first remote switch controller.  It's a three-port version.  Toward the bottom of the picture are the power switch, the Radio "A" switch and the Radio "B" switch.  Between the two switches is an indicator board of LED lights.  The LED board is a custom made one from WA3UTC. The antennas can be manually switched, but there is a switch setting that will allow the transceiver band data output to automatically switch to the correct antenna.  

The control board itself is above the switches.  At the top can be seen the two power connectors.  The control box can be powered by either an AC wall wart or a DC power supply, in an either/or arrangement.  There are three DB-25 control connectors, two mounted on the PC board for input control signals (band data) from the radios and one mounted externally for output power to the remote relays.   Ribbon cable was used for the interconnections in order to present a more orderly appearance.  This particular control box has the LED driver wires on the top of the circuit board.


The switching detail can better be seen in a view of the front panel.  The control logic is set up so that the first radio that gets an open antenna port holds on to it until it is released.  Both radios cannot share the same antenna at the same time, which eliminates the possibility of one radio transmitting into the others' RF input section.  The top six switch positions are for manual overrides.  The bottom position is for automatic switch selection, which is controlled by the transceiver's band data commands.  In this case Radio "A" has control of port 1 while Radio "B" has control of port 2.  If the Radio "A" switch were to be flipped to port 2, Radio "A" would be disconnected from all of the antennas.  This would be shown by no LEDs being illuminated for this radio.  Were Radio "B" to be manually switched to another antenna position, Radio "A" would then be allowed to grab antenna 2 and Radio "B" would be locked out of antenna 2.  If either or both radios were switched to automatic control, one or both yellow "Auto" lights would be turned on.  This situation is illustrated in one of the pictures to follow.


Here's what the back of the first control box looks like.  The top DB-25 connector goes to the remote relay control cable.  It carries the control signals to the remote switch box.  Below this connector are the input connectors from the two transceivers.  They contain the band coding signals, in this case using Yaesu-formatted control commands.  The top right connectors are for the AC or DC input.  If I were to do this over again I'd probably eliminate the AC input jack, since I always use a 13.8 VDC power supply to feed the controllers.  At the bottom is an RCA connector for external Push To Talk (PTT) switching, also unused at present.  The station ground connection is at the lower right.


Here's a closeup look at the inside of the remote switch.  The board is mounted upside down in order to test it.  After testing the board was flipped and mounted inside of the relay box.

The switch has been measured and been found to have a slight capacitive reactance of several picofarads on the 15 and 10 meter bands.  The semicircular red wires add a very small inductance to counteract a little bit of it.  There are six 50-ohm noninductive resistors that ground out each input connection when it's not selected, but they are hidden by the back row of orange relays.


This is a view of the finished remote switch box just after its final testing.  This one is installed inside of a couple of boxes, one of clear plastic and one of metal, to give it some weatherproofing.  It's gone through its first monsoon season without any problems.  It's even survived a burial attempt by a couple of field mice because of its "box within a box within a box" configuration.  

They're not particularly visible in this picture, but all of the connectors have IDs stamped next to them on the metal plate.  Numbers 1 through 6 show the antenna port connections, and letters A and B show show the transceiver control connections.  This switch eventually will become part of an SO2R (single operator, two radio) contest station.  Both cables A and B run to the radio shack but for now only cable A is connected to a radio.  The bandpass filters have to be finished before cable B goes live.


The next picture shows the partially assembled four-port controller box.  It has an extra DB-25  connector at the top of the rear panel, to be connected in parallel with the “Radio A” band data port.  This parallel configuration allows the transceiver's band data commands to also control the automatic antenna tuner's memory settings, thereby keeping the antenna tuner in synch with the antenna being used.  

There was trouble with the front panel decal pulling and stretching a bit when it was slid onto the front panel.  All of the holes had to be drilled beforehand because the decal would have been ripped if it had been applied first.  The power switch and pilot light holes were drilled spot-on, but during its application the decal stretched away from them a bit.  The "Auto" label got stretched as well.  It started out straight but ended up crooked.  Fixing this would require sanding everything off, repriming and repainting the panel, and then reprinting and replacing the label decal.  Consequently I left the panel alone because I didn't want to spend even more time on this part of the job.  The goal was to build an easy to use and reliable switch and not necessarily a cosmetically perfect product that rivaled a high-end commercial switch.  If I ever build a third switch, the lessons learned on the first two will get it closer to commercial quality.  Even as it is, this controller still looks cosmetically better than some of the remote switches that have been marketed in the past.


Let's move on to an interior view of the second controller.  Shown here is the completed four-port version, with the internal patch board all programmed, and the whole unit having just passed its quality control tests.  The LED indicator wires were installed on the bottom of this circuit board in order to make the testing process easier.  It also looks a little less cluttered as well.  The LED board is a commonly available prototype board with 0.1 inch spacing.  The board was squared up to the front panel and held in place with three dabs of hot-melt glue a few days after this picture was taken.  The friction fit of the LEDs into their sockets was good and tight as it was, but as is shown below they still could be accidentally pushed back into the chassis.  The hot-melt glue treatment resolved that problem.


The back of controller two gives a better view of the band data inputs and output, as well as the relay control connector.  The rest of the connections are the same as those of the first controller box.  This picture shows the back view of the switches, LED board, and power indicator LED.  They're fully wired and in their final configuration.  All that remained was to put on the top cover.


This project was undertaken to provide a fully automated way to control both the antenna being used and the antenna tuner settings.  The controllers are connected in a daisy-chain configuration.  This final picture shows the two controllers connected to a single transceiver, an antenna tuner, and a remote antenna switch.  Both the tuner and the switch are being controlled by the radio's band data settings.  Controller two at the top center is automatically selecting the antenna that is connected to input six.  It passes the band data information along to controller one, which is to it's right.  In this configuration the other controller is telling the antenna tuner to use its' memory position one.  But since the antenna is well-matched at this frequency (it's going to the triband yagi on 20 meters), the antenna tuner is powered down and bypassed at the moment.  Thus the tuner is ignoring the controller's commands right now, but whenever it is switched on and in-circuit it follows the commands of the controller.  There is an internal patch board within the controller card that can combine several band commands into one output port.  This is useful when one has all three bands of a triband yagi going to the same coax cable.  In this case switching the transceiver from 20 meters to 15 or 10 meters means that the radio's band data commands (20, 15, or 10) need to be forced into the same antenna port: port 6 in this case.  The indicator lights show a different pattern because their internal patch boards have been configured differently.  One is set up as a remote antena switch.  The other is set up as an autotuner memory controller.