A rubber hammer'd probably do it too, but please don't take me
seriously; I was being a little tongue-in-cheek.

After all the Herculean efforts some folks have gone to (including
myself years ago) to re-engineer something that worked really well
to begin with.  Running the test roll through the 88-note note test
several times at really high vacuum usually works well enough if the
stack has been carefully rebuilt.  If the valve plates haven't been
re-set with new shellac, a few flakes may lodge where they won't come
out and the neither the test roll nor a sledge hammer is likely to
fix that.

That all said, I've also noticed about 10% of the round valves (and
about as many cross valves) per stack are warped, often enough to
produce substantial leaks.  I lap them, meaning I take a piece of flat
glass and some 600 sandpaper and carefully sand the seats flat, so
they're perfectly flat when I'm done.  Within the first few strokes
the warped seats quickly show an egg shape.  I then have them replated;
I don't paint them.  The cross valves are pretty ticklish to do as
it's easy to make them worse.  I'm happiest when I come across a set
of aluminum cross valves.

I always remove all of the valve seats & reset them with really thick
shellac. I keep them in order, along with all the screws.  Why not?

At a convention some time back I bought a bunch of cross valves to keep
on my shelves.  I also kept a set that a customer insisted I change
years ago.  Don't think I'd ever alter a cross valve stack again (been
25+ years since I did it).  I'm a born-again Luddite -- I still play
paper, too.

Pouches:  I'm doing a stack right now that had pretty good pouches, but
when I tested the wells (putting knotted tubes on the two pouches on
either side of the one to be tested) and then sucked and blew into a tube
to the one I was testing, I could cause both side pouches to inflate and
deflate fairly quickly.  The wells needed to be re-shellacked.  This
happens pretty often, so generally I replace all of the pouches with
.007" material I get from Organ Supply Co., Columbia, Penna.  It takes
only about an hour or two to do 88, and I think the leather's getting
too old to trust anyhow.

I wonder if this kind of leakage might have something to do with notes
dropping out on soft play in some pianos.  I can't bring myself to
believe it was sloppiness on the part of the recording department (e.g.,
Milne) as has been sometimes proposed.

Bruce Grimes