What will winter do to your wood?

A comprehensive guide to calculating seasonal wood movement

By Hendrik Varju

Photo by Stacey Bradford

No comments

It's not the measuring, cutting or finishing that makes designing and building hardwood furniture so challenging; it's the phenomenon of seasonal wood movement, which would be irrelevant if wood expanded and contracted by the same rate in all three dimensions. But the rate of movement in width, thickness and length are all different; they also vary by the cut of lumber: flatsawn, riftsawn or quartersawn. Ignoring wood movement is a disaster waiting to happen. Warping and cracking will turn a beautiful design into a flea-market special in a hurry. The key to successful design, aside from aesthetics, is understanding and accounting for seasonal wood movement.

The basics

First, it's important to understand what the terms flatsawn, riftsawn and quartersawn mean. Flatsawn wood has U-shaped growth rings that run roughly parallel to the wide face. Riftsawn growth rings run on a 45° angle. Quartersawn lumber has vertical growth rings, running perpendicular to the wide face.

The most important point is that wood moves more along the growth ring than it does across it–roughly twice as much. So, a flatsawn board moves more in width than riftsawn or quartersawn lumber because the growth rings run roughly parallel to the wide face. Quartersawn lumber moves more than flatsawn in thickness, although it is still only a tiny amount if the board thickness is just 3/4". Riftsawn moves less than flatsawn in width, but more than quartersawn.

Movement in length, which is known as longitudinal movement, is considered zero for furniture design purposes. While it isn't actually zero, longitudinal movement is incredibly small for the board lengths commonly used in furniture.

Why does all this matter? It can mean the difference between heirloom quality-furniture and a project that will fail in the coming months. There are four simple mistakes beginners make:

Error #1:

If you edge-glue a 2"-thick quartersawn board to a 2"-thick flatsawn board when making a solid-wood panel, such as a tabletop, the quartersawn board will move more in thickness than the flatsawn one. While glue is flexible enough that the joint won't fail, a step will develop over time at the joint line, making it look like something is wrong.

 

Error #2

If you join an 8"-wide flatsawn rail to a vertical post, you have something called unacceptable cross-grain construction. The rail needs to expand and contract across its width but is joined to the post where it has zero longitudinal movement. Failure occurs when the rail wants to move but the post restrains it. The rail will crack during a period of low relative humidity when it tries to contract.

 

 

 

 

Error #3

If you attach a solid-wood tabletop to a table base so that the top can't expand and contract freely as relative humidity changes, the table will fail. The tabletop can expand enough in high humidity to blow apart the leg-to-apron joinery, or contract sufficiently in low humidity to crack down the middle. This principle also applies to frame-and-panel doors; a solid-wood panel must be allowed to float freely within the frame. Do not add glue to the joints.

 

Error #4

If you underestimate the amount an inset drawer front might expand vertically in high humidity, you may find that the drawer wedges itself into the case and can't be opened easily. I think we all have at least one piece of furniture like that at home.

 

 

 

 

Illustrations by Stephen Hutchings

Do…

- Allow wide panels, such as tabletops and door panels, to float freely as relative humidity and moisture content change.
- Pay attention to the boards you choose so that only boards with similar growth-ring orientation are edge-glued together to make panels.
- Choose quartersawn lumber, or at least riftsawn, for higher-risk cross-grain construction situations that come up in your designs.
- Use veneered plywood when necessary to eliminate wood movement.
- Make calculations to estimate wood movement so that you are able to design your projects around it intelligently.

Don’t…

- Design with wide areas of cross-grain construction in solid wood without engineering some method of free movement into the joinery.
- Glue a solid-panel tabletop to its base. Use cleats or some other method to allow the top to expand and contract freely.
- Glue panels into frame-and-panel door frames unless the panels are plywood.
- Glue plywood bookshelves into dados cut into solid-wood gables, or vice versa. Solid wood and plywood can only be combined in certain ways because solid wood moves and plywood restricts movement.
- Ignore wood movement or you’ll watch your project slowly self-destruct over time.

Calculating wood movement

Wood technologists have come up with a wood-movement formula that helps you calculate such things as how much a drawer front is likely to expand and contract annually or how much your 36"-wide tabletop will move. The following is the formula I use:

D = ΔD x R x ΔMC ÷ fsp

It means that the change in dimension (ΔD), which is what we're trying to calculate, equals the current dimension (D), multiplied by the rate of movement (R, which varies by species and the particular cut of lumber), multiplied by the change in moisture content the piece of wood is going to be subjected to (ΔMC, which depends entirely on relative humidity conditions), all divided by the fibre saturation point of that species (fsp).

Before you panic, let me simplify the formula by inserting one fixed number: 0.28 for fsp, a number that wood scientists consider a fair number to use for this variable. While it does vary slightly from one species to another, this is an acceptable number to use.

The ΔMC in the formula is dependent on the area in which you live, and you can calculate this number with the help of weather statistics or a book such as R. Bruce Hoadley's Understanding Wood: A craftsman's guide to wood technology. For the Toronto area, which isn't far from where my workshop is located, I use a figure of 0.05 for ΔMC, or 0.06 if I'm being extra-cautious. That means that I’m assuming that the wood in a finished project, stored indoors, will be subjected to a moisture-content change of no more than six per cent over the year. So, the wood might take on an moisture content of something like 10 per cent at the height of summer if you don't use air conditioning or a dehumidifier; four per cent would be quite possible in January or February in a forced-air furnace environment with little or no humidification.

The best way to illustrate the use of the formula is to give an example. Let's suppose you intend to build a frame-and-panel door with a panel that is 16" wide. It is made of flatsawn hard maple, which has a rate of movement of 9.9 per cent. (See the statistics in "Moving Around" on the next page.) The calculation would go as follows:

ΔD = 16" x 0.099 x 0.06 ÷ 0.28

The rate-of-movement value of 9.9 per cent was converted to a decimal to become 0.099.

 

 

When you do the math, the answer is given in inches: 0.339". That’s slightly less than 1 1/32", or between 5/16" and 3/8". Here is the tricky part, though. My own experiments have shown that the formula consistently overstates wood movement when the lumber is flatsawn. I have two explanations for this. First, wood scientists assume a completely flat growth ring when stating the rate of movement of flatsawn lumber. Because the growth ring is usually U-shaped, it means that each flatsawn board is actually riftsawn for more than half of the board. (And we know that riftsawn moves less in width than flatsawn.)

Second, wood scientists calculate wood movement with unfinished wood. All finishes let moisture through slowly, but many will slow moisture transfer enough that some of the extreme high or low humidity conditions that only last a day or two won't have full effect on the wood. In other words, the 0.06 number used for MC may very well be exaggerated.

There are many variables here and many assumptions. But, my own testing and experimentation leads me to cut in half the result from the formula when calculating wood movement for flatsawn lumber. When using quartersawn lumber, I trust the formula's result because growth rings any less than perfectly vertical result in increased movement in width, while the effect of the finish on the wood tends to cancel it out. So, for the example above, I would assume that the panel will move year-round by about 0.165" in width, or just under 3/16". We assume zero movement in length because it’s longitudinal movement.

The question now is what you do with the information. If your workshop is not humidity-controlled and relative humidity conditions are currently at an annual high in your area (such as August in the Toronto area), you can assume your lumber is at its largest, swollen, year-round size. So, you can assume that all 3/16" of annual wood movement will be contraction after the door is completed. In that case, make sure the dados holding the panel are well over 3/32" deep per side so that the panel can’t contract out of the frame. Pinning the panel at its centre point using a small dowel or nail on the back keeps the panel centred in the fame. Also, there is no need to have a lot of extra dado depth when inserting the panel because all of the yearly movement will be contraction. You still want to leave something small, perhaps 1/16" per side, in case your calculations are wrong.

Moving around

The chart bellow shows the rate of movement in width, based on various species and growth-ring orientation. The rates of movement are expressed as a percentage. To use the number in the wood-movement formula, move the decimal two places to the left. (So 7.1 per cent becomes .071 in the formula.)

Only rates of movement in width are given, since movement in width is always the biggest concern. To get the rate of movement in a thickness of flatsawn oak, use the number in the column showing you the rate of movement in width of quartersawn oak, as they are the same. Riftsawn lumber has the same rate of movement in both thickness and width. And remember: the rate of movement in length (longitudinal movement) is considered zero for all species for indoor furniture purposes.

Species Flatsawn Quartersawn Riftsawn

Ash, black 7.8% 5.0% 6.4%
Beech 11.9% 5.5% 8.7%
Cedar, western red 5.0% 2.4% 3.7%
Cherry, black 7.1% 3.7% 5.4%
Mahogany, South American 5.1% 3.7% 4.4%
Maple, hard 9.9% 4.8% 7.35%
Maple, soft 8.2% 4.0% 6.1%
Oak, red 8.6% 4.0% 6.3%
Oak, white 10.5% 5.6% 8.05%
Pine, Eastern white 6.1% 2.1% 4.1%
Teak 4.0% 2.2% 3.1%
Walnut, black 7.8% 5.5% 6.65%

 

On the other hand, if it is the lowest relative humidity season in your area (such as January or February in the Toronto area), you can assume your lumber is at its smallest year-round size. So, all 3/16" of annual wood movement will be expansion after the piece is completed. In that case, the panels only need to go a small distance into the dados to begin with and the dados need to be at least 3/32" deeper per side so that the panel can expand fully without bottoming out. Again, pinning the panel at its centre point is very useful or else the panel might slide one way, pulling out of the dado on the opposite side.

Whatever your calculations, it isn't a bad idea to overestimate anticipated wood movement if there is no structural or design reason that you can't. In other words, if you think the dados need to be 3/32" deep, why not make them 1/4" deep instead? It's far more than you'll ever need, but what is the downside to going deeper in this situation? What if this piece is put away in a storage facility for a year while you take a job in Australia and the non-humidity-controlled storage locker hits phenomenal humidity conditions in the summer? Or if you move from Regina to Vancouver?

In my workshop, I control the relative humidity all year to the approximate midpoint of annual relative humidity conditions in a typical home in my area of the province. So, whatever amount I calculate for annual wood movement, I assume half of the movement will be contraction and half expansion; then I design accordingly.

A shift in focus

You may find it difficult to get excited about wood movement. But, the first time you watch a hundred hours of your painstaking work get ruined by wood that warps, twists and cracks, you’ll develop an interest.

You don't need to be a scientist to work around wood movement, but you do have to understand the basics and think before you design. And once you start thinking in terms of seasonal movement, it will become second nature.


No comments

To leave a comment, please log in

Don't have an user account? Register for free

Poll

How do you heat your home?

Loading ... Loading ...

Recommendations