6x2 Coaxial Remote Antenna Switch

Coax Switch
6x2 Antenna Coax Switch
Completed unit showing weatherproof relay box and station control unit.

- All Photos and Graphs Clickable for a Larger View -


This coaxial antenna selector is different than most other remote coax switches because it has two 'common' ports instead of just one. This allows the operator to select one of six antennas to either of two transceivers. I've only seen two commercial source of these switches, one from Array Solutions, the other from Microham. I wanted two of them but at $400 or more each, that's a pretty expensive antenna selection solution. Multiplexing several antennas to 2 transceivers can be done with manual A/B and 1 of N switches, but the downside of that is the quantity and cost of switches required (and inherent space they occupy), and the mess and expense of all the coax cables that would have to be run to the shack. In my installation it cut the quantity of cables run through the wall from 10 down to 2. It also saved the amount of coax that was required, which is the primary reason one would want to use a remote coax switch. And, it puts the common ground point away from the shack while reducing the number of coaxial lightning arrestors required. I combined the two switch boxes together, giving me the ability to select any one of 10 antennas to either of my 2 HF rigs. (See a photo of the final configuration at my Hamshack page.)

The benefits of coax matrix switchboxes are:

These switchboxes are popular with the DX'ers to be able to share big arrays with multiple radios in contest operation. I'm just a casual operator but the ability to select antennas and swap rigs with a twist of a switch is very appealing.

Construction Notes

I searched the net in hopes of finding remote coax switch construction details. There wasn't much, I suppose there aren't many hams that have built them. There was a nice page that inspired me to give it a try. Specifically it gave details on how to make DC relays more suitable for switching RF by soldering a strap across two wipers, which shortens the path the signal has to go through. It just so happens I have a box of several dozen identical 4PDT Omron relays which would save me a lot of money over having to buy all new relays for this project (I need a total of 24). To test it out, I modified one relay and soldered coax to it's terminals, then connected it to an HP8753E network analyzer to see how it perform over frequency. Below is a plot of the results:

Relay Modification
Relay Plot
Relay Modification Detail
Here is a closeup photo of the shorting strap soldered across the relay wiper arms.
Relay Performance Plot
This plot shows the insertion loss versus frequency of a single relay.

Suprisingly, this 'DC' relay exhibited useful switching capability up through VHF. For a coax switch I only expect to use through 30 MHz, these relays will work fine. The only anticipated limits are high power and high VSWR. The Array Solutions unit uses beefy relays and can handle a kilowatt, but I only have 100W radios so it shouldn't be an issue. High VSWR could develop high voltage so the plan is to only connect well matched antennas, with a separate loop to connect tuner-matched random wires and loops. These were minor concessions that fit within my station setup plans.

Relay Box Details

Now that I had my relays figured out, the rest of the design revolved around the mechanical attributes of fitting those relays in a box. Having my own PCB fabrication capabilities made the relay connection issues easy. Having a couple extra switch poles in each relay allowed me to implement an electrical interlock between each antenna port relay pair. This makes it less likely that the two transceivers can be connected to the same antenna at the same time. It also simplified the control box by not having to implement any interlock cicuitry there. Drill patterns for the SO-239's were generated from the PCB artwork, and holes were drilled in the metal enclosures. Each SO-239's mounting holes were tapped with an 8-32 thread for mounting in the box.

Relay PCB
Relay Box
Finished Relay PCB's
Finished Relay Enclosures

Finished PCB
8-32 Tapped SO-239
Populated PCB
With relays and SO-239's, ready to install inside the enclosure.

Control Unit Details

I investigated using a variety of schemes for controlling the relays from logic cicuits to diode muxing. Ultimately I decided to stick with brute force- 1 wire for each relay with a common return. This aided in keeping things simple, and minimizes the risk that something could go wrong later down the line. Other factors that drove this decision is this is how Array Solutions does it (don't argue with success) and it also leaves the option for easy integration of a future automatic antenna selection unit. You have to run a control cable whether you use 2 or 20 conductors, no matter what. So having fewer wires does't really save that much work. I decided to use DB15 connectors due to their price, availability, and 15 connections provided ll the necessary connections plus a few spares. By using double-pole rotary selector switches, I could drive the relays with one pole and the indicators with the other, which made it easier to use the red/green LED's I had intended. By assigning one transciever to red and the other to green, you can tell at a glance which radio is connected to what antenna.

Control Unit
Control Unit
Coax Switch Control Unit

I built the control unit into a salvaged computer A/B switchbox. I fabricated new 'ends' out of aluminum plate. This photo shows the left control has selected antenna 1, and the right control is connected to antenna 6.

You can't do that...
Should the same antenna be selected by both transceivers at the same time, both red and green elements of the LED will illuminate kind of orange. Because of the interlock wiring in the relay box, whichever transceiver had selected that port first 'wins'. The other port will have no relay selected and is unterminated.

Performance Testing

After completing the unit I tested it for insertion loss and isolation. While I tested it up to 50MHz, the primary range of interest is 1.6-30MHz. The results shown are nominal, as each port is (marginally) different from the other.

Agilent 8753C
Analyzer Test Setup
Here is a photo of the Agilent 8753C network analyzer setup used to characterize the 6x2 coax switch.

Switch Loss
Return Loss
Coax Switch Insertion Loss
The plot above shows the insertion loss from one of the transceiver ports to a selected antenna.
Coax Switch Return Loss
This plot represents return loss from one transceiver port to a selected antenna port. Return loss is another term for match efficiency, the higher the number the better. As a point of note, the 15.7dB of return loss measured at 30MHz translates to about 1.4:1 SWR.

Port Isolation
Unterm Isolation
Terminated Coax Switch Port Isolation
How much energy 'leaks' through the switch to the alternate radio? This measurement shows the coupling from one radio port to the other with adjacent relays activated, and unconnected ports terminated.
Unterminated Coax Switch Port Isolation
Here's a plot showing the port isolation with only the driving port terminated. The alternate transceiver port and the rest of the antenna ports are unselected and unterminated.


It might seem like a lot of trouble to go through all of this work just for a remote coax switch but if you have a need for such a thing, options to purchase one are limited. Now that I've done the project, I can see that the Array Solution's SixPak at $400 is not an excessive price. The lesson learned here is that the Omron MY4 relay doesn't work that great above 10MHz or so. If it weren't for the fact I had many of the components already in my junkbox, it wouldn't have been cost effective to build myself based just on parts cost alone. But it provided an interesting project that resulted in a useful station accessory.

Design Files

Click on the icon below to dowload a zipped collection of PDF files for all the parts, and PCB information to build a 6x2 coax switch like the one documented above.

Zipped Design Files

Back Arrow
Back to the Main
Project Page
Back to the K5LXP
Home Page
OS/2 Warp
© 2003 Pane Relief Computer Services