[TowerTalk] Multi-station antenna selection patch panel

Jim Lux jimlux at earthlink.net
Sat Jan 18 14:16:19 EST 2014

On 1/18/14 4:07 AM, David Robbins wrote:
> Agree, any to any gives lots of relays no matter how you do it.  but you
> really don't need vacuum relays, most of the common switch boxes today just
> used regular frame relays.  If you want to keep cost down by using off the
> shelf parts try looking at the rcs-10 remote switches that do 8 to 1, I am
> using them for my rx antenna switching:
> http://wiki.k1ttt.net/2011%20Maintenance%20and%20Upgrade%20Blog.ashx#bevswit
> ch  and they work well... of course there is no lockout, you would have to
> come up with that in whatever provides the control power.  I think k1xm was
> working on something like that, but not sure how far it ever got.

I think it's the port to port isolation requirement that is going to 
drive the design..  Regular old power relays, if chosen wisely, will 
probably work just fine, but you might need to come up with switch 
architectures that cascade two or short unused ports, etc. to get enough 

I doubt a single relay will get you 60 dB of isolation for instance.

There's a fair number of posts to TT over the years with suggested part 
numbers for the relays, and for such a big project, I'd go out and get a 
few and rig up some test fitures before committing  (although, if you 
got a good deal on surplus and they cost $1/each in that quantity...)

Another aspect to consider is failure modes: if one of those dozens of 
relays sticks or fails to change positions you don't want to be blowing 
up radios with the output of the linear.  This is one of the advantages 
of the various schemes where you have two SPDT relays in the path: if 
one fails, you've still got the other one to prevent disaster.

I've found that it's the driving circuitry that is often the more 
complex aspect.  Although.. these days, there are people selling nifty 
computer controlled relay drivers AND decent software that can be 
configured appropriately, so you probably don't have to spend hours 
figuring out diode matrices.

More information about the TowerTalk mailing list