Crack fillers - old soundboards

[Ron Nossaman]

>What is the cause of the original cracking? 

The original board was likely compression crowned, so it was originally
dried down plenty - and them some. With the panel expansion supplying
crown, and the ribs and string downbearing resisting it, the panel is
pretty close to it's compression limit from day one. Throw in fifty or so
humidity cycles, with the expanding panel crushing because it has nowhere
to go, and a tuner pulling it back up to pitch in the dry season (insuring
further crushing next damp cycle), and pretty soon the panel wood is too
crushed, and has been held near the crush point for too long (compression
set) to spring back and maintain compression crown in the dry season. The
crown goes flat and the panel cracks because it's been crushed to a
dimension narrower than the rib length after shrinkage. Shims tend to cause
cracks alongside because the wood they're put in is already destroyed by
compression, even if it looks fine. Epoxy seeps into the panel wood a bit
and strengthens it adjacent to the repair, so epoxy shims tend to not
develop cracks beside them. Re crowning soundboards with new shims, or
epoxy filling the cracks, just plain doesn't work. Shims, whatever they are
made of and however and in what phase of the moon they are installed, are
cosmetic. A crown may seem to be restored in the short term by shimming
(until it gets out of the shop), but is very temporary. Unless you can find
a way to add enough compression resistance to the entire panel to restore
the ability of the panel to support crown and string bearing, you haven't
restored anything but the looks with shims. This may be possible, but shims
ain't it. 

This has been gone over lots of times.



>I do agree that shimming is largely cosmetic. But I also refinish
>the plate when I take it out.

Do you shim the cracks in the plate as casually as those in the soundboard?
They both perform functions at levels far beyond their surface appearance,
and making a cosmetic repair of a crack in either case doesn't necessarily
restore the performance potential.


Ron N