Segmented Bowl Glue Line Creep
Russ, Good description and good solution it appears. I'm glad you put the
phrase "glue creep" in parenthesis, since this may not be the culprit.
According to mechanical engineers (I'm not one) creep is the permanent
distortion of a material under a compression force that is less than its
compression strength. I'm not sure that creep is at work since the large
forces necessary to make a plastic creep would also crush wood.
Plastics engineers know that a plastic part will creep away from the
pressure of an assembly screw given time. That's the reason many assembly
screws have a built-in washer, to lower the compression force and eliminate
creep.
I've tested some materials on an ASTM device that measures creep. This
included polyester resin loaded to 2000psi for two days. Except for an
initial creep when the resin lost the last remaining amounts of sytrene
solvent, it did not creep any more than aluminum, which is to say almost
none at all. Unfortunately, I didn't test epoxy resin or resorcinol, which
I think wouldn't creep much. I suspect the water-borne adhesives like PVA
and aliphatic resin would exhibit creep. But again, it needs to be a large
force. The question is, how large and is that force present?
Also, creep ocurrs under compression. Does compression exist in a glued
object? I'd hazard a guess that once the clamps are removed the forces on
the glue joints move from being under compression to either shear or
tension. All in all, I doubt if creep is the culprit. More likely something
related to moisture.
What we call "glue creep" is really a couple different things that are
happening.
It doesn't matter what type of glue it is - white, yellow, hide, epoxy,
plastic
resin, or whatever, ALL useable wood glues have some degree of elasticity
and
plastic properties. If it didn't the glue joint would be brittle and
break
apart with the seasonal wood movement. An example of a brittle glue would
be
CA, and for that reason it is not a good adhesive for a segmented turning.
At
the other end of the spectrun we will find the Elmers "White Glue" which
retains a considerable "rubbery" consistency after it has dried. The
Titebonds
are somewhere in the middle, making them a more suitable adhesive for our
work.
There are two different things happening. One is that the plasticity of
the
glue in the joint allows the two pieces of wood to move relative to each
other.
This is what happens when we use woods with different expansion and
hardness
characteristics on opposite sides of a joint.
The other is a property called "hysteresis", a common property of all
elastomers. This word describes the fact that when an elastomer is
stretched,
it will not retract totally to its original form. When a segmented bowl
expands
to a larger size, it pulls the glue in the joint along with it. When the
wood
retracts back to its original size, the glue does not return all the way.
This
leaves the joint with a little ridge of glue sticking out from it.
If there is a considerable difference in the characteristics of the wood,
a
step or offset can develop between the pieces across the joint.
These things are a fact of life when we glue two pieces of wood together,
and
nowhere are they more obvious than in a segmented turning. How much is
acceptable is a matter of definition. Many woodturners and the public that
looks at our work have seen so many bad joints that they are becoming
acceptable as the "norm" for segmented turnings.
However, there are some things that we can do to all but totally eliminate
these results of the wood movements. We can make sure that the joint is
tight,
and that the glue is totally cured. Then we can a few things to keep wood
movement to a minimum. A tight joint will have a minimum of glue in the
joint.
The glue cure and wood movement can be handled with care and time.
I use what I call the "30-30-30 Day Rule" as the minimum time between the
major
steps in the process.
All wood is allowed to sit in the shop for 30 days before it is sawn. This
allows the moisture to stabilize to equilibrium with the atmosphere in my
shop.
The bowl is allowed to sit for another 30 days after it is assembled. We
have
added a lot of moisture into the wood around the glue joints. This wait
allows
the assembly to achieve a uniform moisture content and for the adhesive to
TOTALLY cure. Yes, glues develop enough strength that it can be handled in
a
short period and we hear folks talking about such things as waiting 5, 10,
20,
or 60 minutes. But, a 100% cure of the adhesive will take up to 30 days or
more.
The last 30-day wait comes after the bowl is finished and rough sanded.
Turning
and sanding heat the wood and lower its moisture content. This wait allows
the
wood to stabilize with the atmosphere around it. It also allows the wood
to get
used to its new shape and relieve any internal stresses that were released
by
the turning, and some additional time for the glue in the joints to cure.
Now, the piece can be finish sanded and finished. Most of the joint
problems
will be sanded away.
We can also do a lot by choosing woods that behave the same way with
temperature and humidity changes. It isn't by accident that most of my
segmented turnings are made from Honduras Mahogany, Walnut, Birch, and
Holly.
As for all of this "waiting". Is it overkill? Maybe. But, my most stable
bowls
are those that have had a year between their assembly and turning. I have
been
able to move several of these several times between the humidity of FL
and
Western WA, and the dry CA desert, without seeing any changes in the
smoothness
of the wood across their joints.
Others may disagree with my methods, but this is what works for me.
Russ Fairfield
Post Falls, Idaho
http:/www.woodturnerruss.com/
|