Soundboard question

[Nick Gravagne]

Hello List,

Beware, this reply is long but hopefully of interest and not long-winded.

RicB's original question as posed wondered about the "presumed benefits" of
compressing a soundboard panel prior to ribbing. Would the soundboard
assembly be less prone to RH reactive swings that a non-pre-compressed
assembly? And might not there be any advantages due to changed wood
properties?

RicB's question has prompted much give and take on the properties and
behaviors of wood, particularly as regards shrinking and swelling. So before
we could even take an educated guess at considering any "presumed benefits"
it seems that IMHO a review of some basics might help.

In the minds of some bellyman "compression set" has come to mean one thing
-- cellular crushing brought on by any force that has caused the wood to
exceed its elastic limit. The thought is that, be it a hammer blow to the
surface of wood, or the restrained expansion of wood due to swelling (which
is the case we are considering) compression set tends to be destructive,
probably crushing, and probably the cause of something called "compression
failure." 

If soundboard panel compression perpendicular to the grain is so great that
crushing does in fact result, the effect would not be evenly distributed
throughout, but would find its weakest links in light-colored earlywood or
softest bands. These (or certain of these) will crush first leaving the rest
of the wood healthy. Crushed wood is pulp and, as been pointed out in this
thread, has lost its structural integrity giving rise to possible cracking
due to subsequent drying. This is all true. But a compression or tension set
(they usually come in pairs) is not always destructive.

R. Bruce Hoadley's book, Understanding Wood, has become a highly valued
reference due to its comprehensive content and use of the scientific method.
But it is still only one source, and some of our experiences and other
anecdotal evidence do not always square (but that's another story).
Nonetheless, BH's tests tell us that beyond about 1% restrained swelling,
wood (in general) will exceed its elastic limit and begin to plastically
deform such that when the compressing stress is removed the strain is now
permanent and the piece has obtained a new dimension. This is called plastic
deformation and occurs in virtually all industrial materials (but for
different reasons). 

I haven't read the entire book in a while, but if we carefully read and
study the pertinent charts and data, it is clear that the author didn't
intend for us to think that straining wood beyond its elastic limit would
destroy or seriously compromise it (crush it) in every case at the cellular
level. He DOES say that once wood is strained beyond its proportional limit
"the piece is compacted more and resistance increases", and upon redrying
from a relatively swollen state a piece will "unload itself to a smaller
than original" dimension causing "permanent shrinkage". 

Hoadley refers to this as "compression set" and it can indicate, but doesn't
always indicate, crushing in a technical sense. And permanent shrinkage
doesn't necessarily mean permanent cellular destruction if it is simply
compaction. For our purposes it might be better to think more in terms of
plastic reformation rather than deformation.

The vast field of woodworking is rife with examples of wood bending and
reformation; every piece of wood so shaped has not been destroyed in part or
made pulpy in whole or in part. Indeed, all of industry shapes and reforms
materials of all kinds without destruction. Hoadley does address crushing
elsewhere, but does not always connect deformation, reformation, shrinking
or compression set to damaged wood (especially regarding compression
failures) saying, "There is considerable deformation, well beyond the
elastic limit, before total failure finally takes place."  Note the phrase,
"well beyond the elastic limit." For our purposes, failure means pulp in the
lengthwise grain due to compression perpendicular to the grain well beyond
the elastic limit.

One study by Gwo-shyong Hwang published (1997) in the Taiwan Journal of
Forest Science states, "The hardness, static bending strength, and shear
strength of compressed wood increased with the increasing compression set."
Note that the wood specimens used in this test were not particle board, but
solid China fir compressed in its radial direction. Also note here that the
term compression set indicates a non-destructive deformation / reformation
such that the wood retained useful and beneficial property changes.

Another paper, "Combined Densification and Thermo-Hydro-Mechanical
Processing of Wood" by Parviz Navi and Frédéric Heger (2004) says that a
"densification and a thermo-hydro-mechanical (THM) treatment can transform
wood into a new material with improved mechanical properties, decreased
sensitivity to moisture, increased durability, and no shape-memory effects."

We could write a book here including the intriguing prospect of measuring
shrinkage as a valuable tool for attacking the EMC issue, and I'm sure a
slew of questions if not objections are beginning to form.

But back to something more practical.

How does some of the math work out? Take for example a spruce specimen of a
4" wide by a 40" cross-grain length which has been living in a comfortable
shop setting of say, 70 degrees F by 40 to 50% RH. The EMC would be about 8
or 9%. Drying in the hotbox to 1% shrinkage (don't confuse this to mean
lowering EMC by 1%) would reduce the 40" length to 39.60" (40" - 0.400"). It
is not uncommon for bellymen to measure this kind of cross-grain dimensional
change. 

(Not to muddy the waters, but although we call this the 1% rule, BH's book
suggests that the strain in inches per inch is more like 0.008" per inch,
which means more precisely that the total shrinkage dimension in 40" would
be more like 0.320" across the panel).

Now begs the question: if we take the shrunken piece and reintroduce it to
the shop setting it will want to return to its 40" length. If we restrict it
completely, however, at its ends, and all along the surface of its length to
prevent buckling, the piece will certainly compact, but will not exceed its
elastic limit and theoretically there should be no destructive crushing.
That is if the 1% rule can be taken as gospel, and if other cellular
properties of the wood can be relied upon for uniformity, which they cannot.
Still, the compacted specimen, when removed from its constraints, should
measure out to the shorter 39.60", and this in the ambient shop setting and
at the relative EMC. But for how long? Will the piece spring back to its 40"
dimension almost immediately, or over time? Remember the shop conditions
must be maintained while the piece is relaxing.  

In any event, since no crushing or plastic wood deformation should have
taken place, the smaller dimension, if it does not recover, can only be
attributed to cellular and fibrous compaction and a non-destructive
"compression set". Non-destructive molecular and cellular compactions are
common processes in the industrial world in order to achieve density and
strength, which means more molecular and cellular material exists per unit
of volume. Piano hammers come to mind, as well as many of the processes
required to pack steel so as to increase its strength and hardness.  

Why not try this crude test. From an old soundboard cut out a cross-grain
length of wood between the ribs. Make it as long as possible but maybe only
2" wide. Whatever useful length you end up with, multiply it by 1% and
subtract that from the total length (e.g., 30" - 1% = 30 - 0.300" = 39.700)
Now shrink it in the hotbox to 39.7". Remove it and set it length-wise in an
immovable fixture such as a pipe or bar clamp so as to secure it from
expanding (do not set it in any kind of wood restraint). Don't forget to use
C-clamps and cauls along the length to prevent buckling. Let it sit for a
few days (hopefully in a reasonably stable shop environment), then take it
out and quickly measure it for immediate recovery. Continue to measure it
over the next few days. 

This experiment can be repeated using the same specimen subjected to more
and more shrinkage followed by relaxation. Compression set and compaction
will continue to occur until some point when crushing and failure obtain.

In partial answer to RicB's question then, pre-compression of the entire
panel, including its surfaces, suggests a significant challenge to practical
methodology. That said, studies suggest improved mechanical properties with
regard to hardness, static bending strength, and shear strength of
compressed wood due to increasing compression set. This begs the question;
in our panels do we require these "advantages"? Some may think so. As to RH
swings, I would be guessing that the densification of spruce would move its
properties closer to hardwood (or eastern and European white simply closer
to Sitka), and that contraction and expansion due to RH would be just that
-- relative. I would not guess any advantage here, but I would need to
research it further. 

Respectfully,   


Nick Gravagne, RPT
Piano Technicians Guild
Member Society Manufacturing Engineers