Wood movement as it Effects Joinery

Moisture’s effect on Wood

Most hobby woodworkers and professionals with small woodshop operations are aware of the fact that wood ‘moves’ both before and after being assembled into the final product of our efforts.  Most of us know that we can do things that will allow for this phenomenon and take the necessary steps to avoid problems.  Still, we often see examples of work that has not accommodated wood movement; for example: cleats, screwed and glued across tabletops in a deliberate attempt to restrict wood movement.   We base much of our knowledge of the techniques we use on our personal history.  “I’ve done this before, and it worked!”  Even so, it takes little effort to make the necessary adjustments to technique if we know just a few basics.  Unfortunately, these basics are usually shrouded in a cloak of scientific jargon and formulations that put off the average woodworker, and thus he/she just gives up on digging deeper.

The purpose of this article is to provide some basics without too much scientific jargon and to provide some usable tools for the woodworker.

Wood shrinks and swells because of its moisture content.  The moisture content of wood is closely related to the temperature and relative humidity of the atmosphere in which it is located.  Therefore, lumber can be expected to change in thickness, width, and (to a certain extent) length as the temperature and relative humidity of its surroundings change.  This is a state of nature, cannot be avoided, and must be taken into consideration in joining wood parts. We will skip the details and attempt to simplify the process of accounting for wood movement without going into the scientific and engineering formulations from which the calculations are derived.  This phenomenon is explained in detail in “Understanding Wood” by Bruce Hoadley and in “Wood handbook--Wood as an engineering material” from the Forest Products Laboratory in Madison, Wisconsin. The latter is on-line at:  http://www.fpl.fs.fed.us/documnts/FPLGTR/fplgtr113/fplgtr113.htm

First, consider a few facts and definitions.   Moisture in wood is called its Moisture Content (MC).  Wood will hold lots of water, but it will not start to change shape measurably until the wood has dried to its Fiber Saturation Point (FSP).  That is when the water within the cells (free water) has pretty much dissipated and the water held within the cell walls (bound water) starts to be removed.  Once the FSP has been reached in the drying process, the wood will shrink in an almost linear fashion down to the point where it will no longer shrink.  This is usually determined by placing it in an oven and heating it until it stops shrinking.  Thus, this point is called the oven-dry point. If a piece of wood is dried to the oven dry point and then exposed to a moist atmosphere, it will eventually return to a point known as the Equilibrium Moisture Content (EMC) for that environment.  At this point, it will not change until the relative humidity or temperature of the atmosphere changes.  For practical purposes, if the relative humidity of the surrounding air is 50 percent and the temperature is 70 degrees, the MC of the wood will be 9.2%. You can determine this from Chart 1. 

Chart 1.  Moisture content of wood in equilibrium with stated temperature and relative humidity.  (Chart from FPL Handbook page 3-7)

If the relative humidity increases from that point, the MC of the wood will increase along with it and will finally reach a new EMC.  From Chart 1 we can see that if the relative humidity increases to 70 percent and the temperature stays at 70 degrees, the MC of the wood would eventually reach 13.1 percent.  Now let’s reduce the relative humidity to 30 percent and keep the temperature at 70 degrees.  Note from Chart 1 that the MC will now go to 6.2 percent.  The plot of MC against relative humidity in Figure 1 shows the somewhat linear relationship described above. Figure 1 is a plot of the 70 degree line in Table 1. and shows the path that MC will follow in the drying process. 

Not shown in figure but interesting to note, the MC of the wood will decrease along a line as it dries but will not necessarily follow that same line back up as it increases.  This effect is called hysteresis.   While it may be important to consider hysteresis in scientific, or even some industrial, applications, it can be ignored in the applications of most woodworkers. 

The FSP for most woods is about 28 percent of moisture content but varies by species and within species.  In his book, “Understanding Wood”, Hoadley assumes a 28 percent FSP for his procedures.  The Forrest Products Lab procedures assume 30 percent.  The point is that this is not a very critical measurement for our purposes and, indeed, neither source offers a chart for FSP of the different species.

The formula for deriving the change in wood dimension with a change in moisture content is:      

Dc = Di x S x cMC) / FSP                   (1)

Where:

Dc        the change in the dimension of the piece of wood.

Di         the initial dimension

S          the total shrinkage (taken from wood shrinkage tables)

cMC    the change in moisture content of the wood

FSP     the Fiber Saturation Point

In other words, you can determine the change in wood dimension (Dc) by multiplying the current dimension of your workpiece (Di) by it’s shrinkage rate (S-- found in Appendix I) and then multiplying by the expected change in moisture content (cMC) and finally dividing the whole thing by the number representing the fiber saturation point (FSP).

Appendix I provides tables showing total shrinkage for common hardwoods and softwoods. The S values for the formula are shown both as a percentage and converted, by dividing the percentages by 100, for direct use in the formula. 

Appendix I also provides the shrinkage value for a 10-inch board in each species for a MC range of 6 to 13 percent.  The 10-inch board values are easily scaled to any whole inch width.  For instance, if a board were 5 inches in width, the tangential shrinkage over a 7 % range of MC would be 5/10ths of the value given for the 10 inch board; a one inch board, 1/10 the value; and so forth.

Since the FSP isn’t precisely known and usually estimated to be 28 percent, and the data in shrinkage tables is considered typical (not precise), the calculation from this formula must be considered an estimate of the wood movement.  To this end, a short cut is possible.  Use of the shortened formula below assumes a variation in MC of about 7 to 8 percent for the Tennessee area – from a winter indoor MC of about 5 to 6 percent and a summer indoor MC of about 12 to13 percent. 

The formula: 

Dc = Di x S / 4                         (2)

 is a close approximation for the wood movement.

 
In this short form, you can determine the approximate change in wood dimension (Dc) by multiplying the current dimension of your workpiece (Di) by its shrinkage rate (S-- found in Appendix I) and then divide that number by 4.

Wood changes in three dimensions relative to orientation within the tree.  Wood will shrink and swell about twice as much tangent to the annual rings as it will radially (i.e.; perpendicular to the rings).  See figure 2.   It shrinks very little in the length of the tree.   Which shrinkage value, radial or tangential, to use is another question.  See figure 3.  It is difficult and time consuming to establish just where the board in question was cut from the log.  In addition, not all boards have been cut precisely from the log so that they are oriented at their full length at the same distance from the pith.  It is usually the safest to assume the tangential dimension over the radial since it is usually about twice the radial dimension. 

Figure 2.  Tree Section showing Radial and Tangential Movement

Figure 3. Tree section showing lumber orientation.

For estimating the movement of most boards, the formula (2), and using the tangential value given in the tables in Appendix I, will get you close.  If a board were obviously quarter-sawn, you would use the radial value to compute the expansion across its width and the tangential value for its thickness.

There is another column in the table in Appendix I called T/R Ratio.  This is the ratio of the amount of shrinkage of the tangential dimension divided by that of the radial.  T/R is used to determine the propensity of a species to distort, and is important in understanding wood behavior before, during, and after assembly.  Wood with the larger ratios will distort more than those with smaller ratios will.  It is not practical, for our purposes, to go into all aspects of what this differential wood shrinkage means to the woodworker, but we offer an example to make the point.

Since wood will shrink about twice as much tangentially as radially, a board that is flat sawn from a log will cup with concavity away from the pith.  A board cut closer to the pith will cup more than one cut near the outside of the tree.  Take a board cut midway from the pith to the bark.  Dry the board to MC of 6 %, and note that there is some degree of cupping.  Now mill the board flat and square.  It will stay flat and square as long as you keep the MC at 6 %, but change the MC to 12 % and you will note that the board has cupped again – this time with concavity toward the pith. 

The above example is not as far fetched as it may seem.  If a board is milled to its final dimensions at a given MC and then put on the supplier’s shelf where it stays for several weeks at a higher relative humidity, the MC of the board will assume equilibrium with that environment.  When you buy that board, it will show the incumbent distortion.  However, when you allow it to equilibrate in your shop, it will reach another shape – that representing the temperature and relative humidity there.  This example suggests the better approach would be to buy oversize stock and mill it in your shop and to keep your shop at a temperature and relative humidity representative of the space in which the final object will be placed.  Alternatively, and more appropriately, design the piece so that it accommodates expected wood movement.

Coping with wood movement – or – “It happens deal with it!”

The four general rules of coping with wood movement from Hoadley’s book:

1. Season lumber to the average moisture content it will experience in service.
2. Design the work pieces to allow the movement caused by changes in moisture content expected in the location.
3. Coat the surfaces of the wood with the most impervious finish material that is suitable for the job. 
4. Use chemical or mechanical restraints only as a last resort.

Expansion on these rules:

1. Bring wood into your shop a few weeks or longer before milling it.  Let it reach the Equilibrium Moisture Content (EMC) in the shop.  Remember that if you let the wood reach EMC in your heated shop in the winter it will be drier than it will be in a typical house during summer; thus, tight joints (or drawers) will get tighter.
2. Pay close attention to wood orientation in the design and selection of wood to build the piece.  The orientation of the tangential, radial, and longitudinal dimensions of the wood must be such that the movement of one member doesn’t cause that member to invade the space of another.
3. Protect the wood while the project is in progress.  Once you mill your lumber to final shape take precaution against moisture exchange due to changing conditions within your shop.  If you’re not going to assemble the piece right away, put your pieces in a plastic bag or wrap them with paper to restrict moisture exchange, or put on the finish material.
4. If you coat the wood with a certain finish on the show side, put a comparable finish on the other side.  For example:  If you varnish the outside be sure to use a finish with a comparable resistance to moisture exchange on the inside.  If you put a heavy film on one side and a lighter film on the other (e.g., varnish outside and linseed oil inside) you encourage different rates of movement and may cause warping.

Joinery Techniques

The primary guideline is to, whenever possible, join wood where the grain direction will match; that is, tangential to tangential, radial to radial, or longitudinal to longitudinal.  This, of course, is neither always practical nor desired; however, adhering as close as possible to this guideline will help.  For instance, you know that tangential shrinkage will be about twice that of radial and that longitudinal shrinkage is almost always negligible, so when joining the long grain of a leg with the cross grain of an apron select a board which will match the radial grain with long grain.  See Figure 4.

Figure 4.  Joining Cross Grain Members

Another good example of a problem joint where the edge of a flat sawn board is joined to the edge of a quarter sawn board.  The tangential shrinkage of the quarter sawn board is twice that of the radial shrinkage of the flat sawn board and results in a ridge at the joint.  See Figure 5.

These are but a few of the joints singled out as examples.  They are by no means the most important nor egregious joints to consider.  They are intended to point out that thinking through how you orient your pieces will minimize problems.

Finishing

Various finishes and finishing techniques will result in different moisture exchange levels.  The following is a rough rule of thumb for effectiveness of different finishes.[1]

Highest level is 5

The level of resistance to moisture exchange depends to some degree on the thickness of the finish.  You must consider this in application technique and material.  One example: Since you cannot build oil finishes as thick as varnish, they will not form as solid a moisture barrier.   Another example:  Varnish, thinned and wiped on, will not form as thick a film as varnish brushed on full strength.

The most important thing to remember about how finishes affect moisture in wood is that even the most impervious finish will not stop moisture exchange.  Properly applied finishes will slow down moisture exchange so that the wood will not reach EMC in a new environment as quickly as it would without finish.  This means that wood movement will be less susceptible to short term changes in temperature and humidity, following instead the averages; for example:  Instead of shrinking and swelling on a daily basis, it will do so on a much longer cycle – perhaps seasonal.  Never the less, it will shrink and swell.

Properly applying finishes is as important as what kind you use.  As stated in rule 4 above, you should use similar finishes on the hidden side as used on the show side.  Also, do not forget that wood absorbs and desorbs moisture from the ends of boards several times faster than though the side and edge grains, so take appropriate steps to seal these with your finish of choice.

Antidotal success stories and why we get away with ignoring the rules:

“I’ve done this before, and it worked!” is a refrain we have all heard and used ourselves from time to time.  The fact is, we live in a very healthy area for wood furniture.  The middle southern part of the United States has, on average, a mild climate.  Winters are not generally too cold – too long, and summers are not too humid – too long.  The following was derived from the map in Hoadley’s chapter on Coping with Wood Movement.  It is provided to make this point.

 Middle South Relative Humidity and Moisture Content averages:

July:  Average minimum Relative Humidity is 50 %  == > Average Moisture Content is about 10 %.

January: Average temperature is 40 degrees F therefore average Relative Humidity, inside a heated space, might be expected to be 40 to 50 %  == > Average Moisture Content of wood in an interior space, heated to 70% F., would be about 8 to 9 %.

Remember that these are expected averages.  Some years with a particularly cold winter may result in much drier interior temperatures.  Also, remember that moisture sources within, and leakage from outside, will cause fluctuations in the relative humidity and thus changes in the moisture content beyond these levels.  While the finishes applied to the wood will decrease the effects of short-term fluctuations, the wood will eventually reach an EMC dictated by the average relative humidity.

Just remember that what we may have been able to get away with in the past due to our location may not work for us if we send a piece to Montana where the inside relative humidity in winter is probably close to 5 percent or to Florida where the summer humidity may be 85 percent.

Appendix I  -- Shrinkage Tables