Replies

1. 
   ErnieConover | Feb 09, 2007 04:40pm | #1

   As long as the grain runs in the same direction (you alternate the boards to get the equivalent of stripes) you can glue the different woods together. You cannot do the “plywood” routine! New (and some experienced, but unschooled) woodworkers treat wood like it is metal or plastic. As the understanding of the medium is paramount to the mastering of any craft, a good grounding in wood technology is useful. Understanding why wood does what wood does avoids common pitfalls. Let’s take a look at how wood behaves as it dries.
   When a tree is cut down it is 60 % or more water. The process of drying reduces this water content to ambient for the wood’s environment. Once cut into boards, or split into firewood, this water starts to evaporate. Regardless of the starting moisture content, nothing much happens until the wood reaches 28%, a figure called fiber saturation. As the water content drops below fiber saturation the actual cells that make up the wood start to lose water and shrink. Therefore, all shrinkage happens between 28% and the ambient relative humidity of its environment. This could be as low as 5% in a drying kiln. Since wood is hygroscopic it will gain moisture and expand if it is moved to a humid environment and this is usually the case when it comes out of the kiln. This means that the moisture content can vary between about 5% (a dry house in winter) and 15% (a very humid summer). Expansion and contraction of all woods occurs during these wet-dry cycles, in proportion to the original shrinkage from fiber saturation to ambient. During drying, the wood shrinks in three directions or planes. Subsequent expansion occurs in the same three planes.

   As we can see by the attached diagram that I drew for Turn a Bowl With Ernie Conover, longitudinal shrinkage (shrinkage along the grain) is very small. In fact, once dry, we can consider the figure to be zero. It is the reason houses can be timber framed from green wood. Since the size of the house is set with the long grain, the size of the house will remain constant. We can live with the shrinkage in the other two planes. Across the grain there is significant shrinkage, ranging from an average of 4.39% in the radial plane to 7.5% in the tangential plane. These are average figures, however, and American beech is going to shrink vastly more in the tangential plane than black walnut. Of more importance is the ratio of the two values. Beech with a tangential value of 11.9% and a radial value of 5.5% has a ratio of 2.2 to 1. A flatsawn beech board is ill disposed to remain flat with changes in humidity. Walnut with a tangential of 7.8% and a radial of 5.5% has a ratio of 1.4 to 1 and is much more prone to stability. All of the favored cabinet woods have a ratio of 1.75 to 1 or less. The best orientation of the grain for boards depends on what we plan to do with them
   A quartersawn board is one cut on a radius (or a near radius) of the tree such as in Drawing II A. Any board cut from within the cross hatched area would be considered quartersawn. Because the radial and tangential planes are lined up with the board it wants to stay flat. The greatest expansion and contraction tends to be in the tangential plane and this would be the thickness of our quartersawn bard. On average, thickness tends to be the smallest dimension of any board and so by having the largest value in this plane it will be the least problematic. A flatsawn board is one that is cut on a tangent to the growth rings such as in II B. If we are looking for figure in clear wood, the flatsawn board will generally provide better aesthetics. On the other hand it is much more prone to cupping because of the unequal tangential and radial forces acting with changes in humidity. The greater the tangential to radial ratio the more likely this is to happen. Also a thin board is much more prone to cupping than a thicker one because of the elastic limit of the material. (Every material has an elastic limit meaning how much it will bend before it breaks. And this leads to the fact that it is harder to bend a thick section over a thin one.) Alternating the direction of the growth rings in glued up panels can help to minimize the effects of cupping as well. Cherry, for some unknown reason, tends to remain flat no matter what plane it is cut in. I think this is the reason our forefathers chose cherry for drop-leaf tables.

   For figured wood such as curly or blistered, a quartersawn board tends to show figure better. This is because the grain of curly wood tends to undulate in the radial plane. In blistered wood it undulates in both radial and tangential. Cutting in the radial plane exposes this figure.

   Further complicating matters wood builds up internal stress during drying. The act of jointing and planning a board will inevitably change these stresses. A flat board can suddenly become cupped and twisted while a cupped board could spring flat. It is good practice to try to joint and plane equal amount off of each face of a board so as to equalize these stress changes as much as possible. Resawing one-third of a board away to get a thinner dimension often leads to cupping—especially in a wide flatsawn board.

   Bringing all this palaver to useful conclusion, we are better to choose quartersawn wood with favorable tangential to radial ratios for such things as thin panels, drawer sides and drawer supports. Flatsawn boards are better in thicker dimensions, but narrower widths, for table tops, trapped panels and such. Alternating the growth ring directions will also help in glue up panels. It is also better to sticker boards in your shop for a week or two before starting a project to allow them to adjust to ambient moisture as much as possible. During surfacing, try to take equal amounts from each face. Having said all of this I have seen quarter sawn boards warp and flatsawn boards that should have cupped stay flat. Further detailed information on wood technology can be found in Bruce Hoadley’s Understanding Wood, published by Taunton Press. There are even formulas for figuring the shrinkage in either plane for most species of wood.

   Your idea of building plywood will not work because you are using thick boards that are above the elastic limit of the material. Plywood only works with relatively thin thickness of material.

   With best regards,
   Ernie Conover

   
1. 
   GaryBowen | Feb 09, 2007 09:41pm | #2

   Ernie,

   Thank you for the thoughtful and complete response.  Let me make one more pass. 

   <!----><!----> <!---->

   The cocobolo will be glued on top of the walnut rather than side by side.  The only wood seen is cocobolo.  The cocobolo is just a very thick veneer.  In your example of beech and walnut, if I glued beech on top of walnut, the beech would be expanding at the 11.9% rate and the walnut at the 7.8% rate.  Would the beech rip off the walnut?

   <!----> <!---->

   <!----> <!---->

    

1. 
   ErnieConover | Feb 09, 2007 10:07pm | #3

   Much would depend on how the glue up was done. If you keep the boards in the glue up the same width. (i.e. 1/4 x 6 coco on top of 1 x 6 walnut, you will probably (and I say probably) o.k.) Just don't cross the grains of the two.Regards,
   Ernie conover