Topband: Engineering ?? @ relays

[Ford Peterson]

Jos wrote:

>We encountered this problem in audio freq switching years ago, even with
> sealed relays. The cure, let flow a DC current of some  mA
> similtaneously with the micro Amps of the functional AC signal.
>
> Jos on4kj 

Jos, and others...

My quest for suitable relays continues.  Many have suggested that I simply use vacuum devices.  This bullet-proof approach would be the natural 'fix' to the problem of unreliable relay contacts.  At $38 each, the solution would require 8 devices, costing around $300US.  All of my projects include an 'engineering review' by the XYL, who is particularly thrifty with respect to the family 'purse.'  After she watched me bury $500US of wire in the back yard 2 weeks ago, this issue requires some sensitivity on my part.  I suspect others can relate to my concerns.

Several have voiced similar opinions as Jos.  A redesign of my project concluded that I could fabricate my box using an idle current.  Last night, I configured a test fixture using one of my seemingly worthless relays (albeit still $24 worth) purchased at the surplus emporium.  The purpose was to run an idling current through the relay while simultaneously checking the contacts using an Autek VA1.

The fixture included a DPDT design, a balun bifilar wired as a monoband isolation transformer rather than the wideband Guanella style, and approximatly 100ma of DC.  The relays performed flawlessly the first time and every time, even with the low level (+10dBm) signal being used.  Removing the DC idle current rendered the relay worthless once again.  How little current can be used and still get reliable connections?  The DC load in my fixture is approximately 1.2 watts, (i.e.12v @ 0.1A), which seems wasteful and likely excessive.  Can the same trick be performed using uA?  How much idle current is generally required to achieve reliable connections?

Thanks to all for the input.  I (and my XYL) appreciate the help.

Ford-N0FP
ford at cmgate.com