Spencer,  It is hard to see how an unfiltered power supply would help
contact wear without actually reversing polarity.  I can't see the
ripple doing anything to prevent or reduce contact build up.  However
unfiltered supplies are cheaper than filtered and in this case it
doesn't matter.

I would expect a DC supply with a standard bypass cap near the
contacts, much as in typical spark ignition circuits.  In a truly
modern organ the solenoids and relays would all have clamping diodes
but, alas, this will likely _not_ be the case in any real organ.

Thus you _will_ have large reverse emf from the coils, probably in
excess of 50v at the peak.  I would plan to clamp the input to your
board with something that will limit the peaks, either a fast Zener
diode or some sort of Schottky diode arrangement.

I would plan for _tons_ of contact bounce noise as these large
mechanical contacts are in no way related to the small, clean world of
PC-mount relays.  If you are debouncing in software then it is pretty
simple to choose a sample count to eliminate the bounce, e.g., three
samples in a row must be the same to be a valid input.  The same idea
can be accomplished with cascading flip-flops in hardware.  You could
even optically couple the inputs to avoid all the noise from the organ
environment but that might almost be overkill.

So I would plan for the worst, and I believe that your expectations
will be met sooner rather than later.

Regards,
Phil Duclos

 [ In a multiplexed organ the voltage at the keyboard contact may
 [ also respond to coupled keys, like the octave below.  I added
 [ a diode to isolate the key signal from the bus signals in my old
 [ vacuum tube Allen organ (+125 vdc).  Contact bounce and switch
 [ noise was filtered by 10 millisecond R-C filter and voltage
 [ divider followed by a Schmitt trigger inverter.  -- Robbie