When rebuilding a Mills Violano, I decided that safety was far more
important than absolute authenticity where the circuit was concerned, so
I used relays for a number of jobs, and discovered that today's relays
don't do a good job in high voltage dc.

For example, all I could find was silver contacts, and discovered that
the 110 v. dc was creating a "plasma" arc across the contacts when the
points open.  Then it would continue to arc until it had (in one place)
melted a relay down to an unrecognizable plastic lump.  Naturally, I could
not find oil-filled relays or anything fancy.  The biggest offender was
the staccato bow wheel circuit.  R-C quench circuits weren't much help,
either.  While they worked after a fashion, they would probably extend the
life of the relay to a couple of years, max.  So that wasn't the answer,
either.

What I had to do was modify a DPDT relay.  The primary problem was the
resting distance between open contacts, which was 'way too close.  So I
wired the contacts in series, doubling the distance.  That fixed the
problem, but I also spread the contacts a bit, since I had plenty of
throw, and from that point on, I never even began to see that bright
green arc as the magnet's back emf tried to start the on-board barbecue.

Once I got a half-a-dozen good relays in the machine -- including some
piano expression relays, etc. and finally a power switch relay that just
uses the carbon contacts to activate its coil -- I felt reasonably safe
at last, and I'm sure the rolls will last many more years as a result.  I
could not see even the tiniest spark at the roll fingers of these contact
positions.

Now all the owner needs is an automaton on a pedestal that swings around
and pulls the plug out of the wall when he's through playing it.

Craig Brougher

 [ That's a good description of a very practical remedy, Craig.  I hope
 [ we will also hear from the folks who have used solid-state relays,
 [ especially to learn if they are as reliable as the mechanical relay.
 [ -- Robbie