Predicting Wood Shrinkage

Imagine that you are making a pair of panel doors in flat-swan cherry. The panels are 16" wide with one and 1 1/4" inch rails and it's February in your heated shop where the wood has been stored for 6 months. The finished piece will be used in an unairconditioned home in Atlanta, Georgia. How much gap do you need to leave around the panel to avoid warpage or stress on the joints next August?

Or imagine you are placing a band on each end of a large 34" wide red oak trestle table for the same Atlanta customer. It's August and the kiln dried oak has been in storage for about a year. If you finish the bands even with the table edge, how much gap due to cross grain drying can your customer expect next winter?

These and other similar questions on wood movement due to normal humidity cycling are often asked by one woodworker or another without receiving very definitive answers. I have asked these questions and, while there were some "guesstimates", no one could tell me how to figure out a reasonable allowance for a particular project which is really what we all want to know.

As I started to research this problem, it got a lot more complicated before some simple solutions began to emerge. Terms like radial, tangential and longitudinal shrinkage, fiber saturation point and equilibrium moisture content had to be dealt with before I discovered why the books can't give a simple answer to these two seemingly simple questions.

The books referred to are "Wood Handbook: Wood as an Engineering Material" by the U.S. Forest Products Laboratory and "Understanding Wood" by Bruce Hoadley. The book by Hoadley was written for the interested layman and many portions have appeared in Fine Woodworking. This book is not only readable very readable on the physical characteristics and identification of wood but offers much useful information on the interaction between wood and common woodworking tools.

The reason these books can't give a direct answer to these two questions is that wood shrinkage and swelling is dictated by the local environment. Obviously, they can't furnish a table of sight cyclical dimensional change for every locale and condition to be found in the country. What they do provide is a table of approximate shrinkage, as a percentage, from green to oven-dry moisture content. Then Hoadley gives maps of the United States showing summer and winter averages for the moisture content for interior woodwork in a normally heated space. Now we can locate north Georgia on the map, pick off the variation in moisture content from summer to winter and then plug it (along with the appropriate data for our wood species and grain direction) into a formula applied by supplied by Hoadley.

Eureka! The answer comes out of the percent annual variation for my house in Atlanta, Georgia. From there, any woodworker can calculate 32nds or 8ths or whatever is needed. Now that's the simple answer we were looking for.

For the cherry panel, the total annual variation, as a calculated by the above process, is 0.144". Since these panels will be fastened in the center of the top and bottom stiles, the movement expected at each side will be 1/2 the total, or 0.072". To be a little conservative, I will allow 1/10 inch gap between the panel and each rail. While this gap can be left all around the panel, the panel and the rails will have nearly identical movement in the longitudinal (with the grain direction) and the gap in these stiles groove can be made quite small. For a single panel door with very narrow Stiles and rails, leaving the top and bottom gap very small improve the overall structural stability of the door.

For the oak table, the total winter summer-variation is 0.476" or almost 1/2 inch. Now I understand why some of my projects from more innocent days have persisted in self-destruction in spite of (or in some cases because of) screws, bolts, dowels and globs of glue. So, if I finished the main table plank flush with the end of the crossband, there will be an ugly gap of almost 1/4 inch next February. Since we cannot fool Mother Nature, either we accept the inevitability of such gaps or we changed the design of the piece. In the case of the panel door, this design was evolve precisely for the purpose of providing a large thin panel or door with stable exterior dimensions. In this case, it becomes clear that the floating panel designs were evolved by early woodworkers primarily to circumvent wood's natural instability -- rather than for esthetic reasons.

For the table, we can use a piece of the veneer plywood surrounded by a perimeter of solid red oak to achieve this desired visual effect...or leave it off the end bands...or use the end bands and accept the inevitable gaps...or design some kind of restraining system such as long threaded rods to prevent cross-grained expansion (though this has its own risks). In the last case, the table will have to be constructed of lumber at a minimum (winter) moisture content for the location.

A trip to the local furniture store will show the manufacturers usually opt for the plywood or chipboard solution. Since shrinkage and swelling for both plywood and long grain solid wood is very small, a joint can be constructed which will be stable in various humidity conditions around the country where the piece may be shipped.

Local woodworkers often take risks by creating a large cross-grained joint in solid wood and depending on the controlled airconditioned environment to keep them cup out of trouble. Most woodworkers should avoid this kind of joint unless the humidity cycle variation in dimension between long grain and cross grain is adequately handled. As a rule of thumb common half lap, bridal, mortise or similar joints having a cross grained dimension greater than about 3" should be treated with suspicion.

What I have done is to apply the data and formulas from the books to generate a table of annual dimensional variations for the north Georgia area. This is very easy to take the data from the table from tangential (flatsawn) or radial (quartersawn) lumber and determine how much shrinkage or swelling you can anticipate in a particular situation. As a quick rule of thumb, all this data can be reduced to the following:

1. In Atlanta, flatsawn hardwood will have an annual variation of 1 to 1.5 percent.
2. Quartersawn hardwood will have an annual variation of .5 to 1 percent.
3. Softwoods shrink about 2/3 as much as hardwoods.
4. Imported tropical hardwoods also shrink about 2/3 as much as domestic hardwoods.

In the table this is listed as 1.4% tangential shrinkage. All other data in the table was calculated in the same manner. So if you want to know the dimensional allowance required for shrinkage for any of the listed woods, just multiply the dimension of the piece by the percent variance (flatsawn or quatersawn) and you get the answer -- simple.

Predict Wood Shrinkage By Species

Domestic Hardwoods

Imported Hardwoods

Softwoods

There are a few omissions from this table for which total shrinkage data could not be found. Apple, pear, and dogwood are often used locally, but I could not find shrinkage information on these species.

Additional furniture making resources:

Standard Furniture Dimensions Table.
Some ABCs of Woodworking.