Detailing for Wood Shrinkage

Like most builders, you probably back your work with a warranty under which you'll return to a customer's castle near the end of the first year to tweak doors, patch popped nails or silence squeaky stairs. Most of the routine repairs you'll make to trim, walls, floors, ceilings, stairs, and doors are related to the shrinkage and swelling of wood. By accommodating wood movement right from the start with smart construction details, you can reduce time and expense when making good on your promise.

Wood and water
In the living tree, wood is saturated with water. Some of it fills the cavities of wood's hollow, straw-like cells; some of it swells the cells' walls. To increase its stiffness, strength, dimensional stability, and usefulness as a construction material, the water must be removed. During air- and kiln-drying of green lumber, water evaporates first from cell cavities. But even when all the water in all the cavities is gone, the lumber still hasn't shrunk. Only once water starts to leave the swollen cell walls will wood's dimensions diminish. For almost all kinds of wood, the moisture content (MC) marking the onset of shrinkage and the lumber's greatest dimensions -the fiber saturation point- is about 30%. As moisture content falls below 30%, wood shrinks by about 1/30 of its total potential shrinkage for each one percentage point change in moisture content. The converse is true when dry wood picks up water and swells. Minimum dimensions are reached when wood is ovendry, or at 0% MC. Typically, the in-service moisture content of wood in heated buildings can range from about 4% to 16% annually.

Because wood's straw-like cells are laid down in concentric circles (the growth rings), with their length parallel to the trunk of the tree, green lumber shrinks by different percentages in length, width, and thickness during drying. With the exception of some kinds of abnormal wood, shortening along the grain, or longitudinal shrinkage, is so small (about 0.1% from green to ovendry, expressed as a percentage of the green dimension) that it usually can be ignored. But shrinkage across the grain, whether around the growth rings (tangential shrinkage) or across them (radial shrinkage), is substantial, and has to be accounted for in the design of just about anything made from wood. Though shrinkage values vary widely among woods, tangential shrinkage averages about 8%; radial shrinkage, about 4%.

Unequal shrinkage and swelling in the longitudinal, tangential, and radial directions gives rise to the bowing, crooking, twisting, cupping, and other forms of warpage commonly seen in lumber. It's also responsible for the wide checks and splits that open in large timbers used in post-and-beam construction. By cutting a saw kerf along the grain on a green timber's hidden face, you can encourage the widest check to open out-of-sight.

Shrinkage, not settlement
Contrary to popular phraseology, wood-frame buildings don't settle, they shrink. The year-round average equilibrium moisture content of studs, joists, and rafters in heated buildings is about 10%. But since framing lumber is exposed to outdoor relative humidity, and possibly precipitation too, during shipment, storage, and construction, it's usually sold at a moisture content of 15% to 19%, so some shrinkage and warpage is inevitable.

Beginning once the structure is weather-tight, most shrinkage takes place during the first heating season. A two-story, platform-framed home built with HEM-FIR lumber at 19% MC, for example, will shrink about 3/4 of an inch in height as it dries to 10% MC. Virtually all the shortening is due to across-the-grain shrinkage through the depth of the rim joists and the thickness of the wall plates. And that can lead to a multitude of headaches for builders.

For starters, joist and plate shrinkage can cause buckling of plywood siding panels outside or of drywall inside, especially in stairwells and spaces with cathedral ceilings. The problem arises when a panel crosses the rim joist between floors so that it's fastened to the studs above and below the joist. Vertical shrinkage of studs is virtually nil, but vertical shrinkage of joists and plates can be substantial. As the joist and plates shrink, studs on the two floors are drawn together, compressing the panel fastened to them. Being stiffer, plywood siding buckles, while drywall may buckle or crush. The solution is to break panels between floors. For drywall this may mean using an expansion joint at the joist and a control joint at the ceiling, or applying the drywall to resilient channels. For plywood siding, it means providing a flashed gap of about 1/4 in. at panel ends.

The initial shrinkage of framing can also lead to roof leaks when chimney flashing is rigidly -and thus incorrectly- connected to both the masonry and the wood frame. I've read one case history in which casement windows on the top floor of a three-story apartment building clad in brick wouldn't open after the first heating season because the platform-framed floors shrank below the openings in the masonry veneer.

Framing members that bulge out of the plane of a wall, floor, or ceiling as they dry often contain abnormal wood that shrinks excessively along the grain (ten or more times as much as normal wood), causing lumber to crook or kink. One kind, juvenile wood, forms around the center of trees for up to the first twenty years of growth, so just about all lumber sawn near the pith of a tree contains it. Another type, compression wood, forms on the bottom of branches and on the underside of leaning softwood trees. Lumber with lots of knots is apt to kink as it dries because of this. Cut excessively knotty or pith-containing lumber into cripples, blocking, and other short-length uses when you can.

Diagnosing diagonal cracks
Diagonal cracks occasionally appear in drywall at the corners over windows and interior doors. In some cases, overfastening is to blame; in others, the floor framing is at fault. If drywall is fastened to both header and studs around an opening, the header will pull down on the drywall as it shrinks. Fasteners in the studs resist the downward pull, placing the panel in tension, and presto! -the familiar diagonal crack. The remedy: around openings, fasten drywall to studs only.

Floor deck plans commonly call for 2x10 joists to bear on a central girder nail-laminated from 2x10s, and supported by metal columns. From a shrinkage perspective, the design is flawed. While the girder end of each joist bears on a beam 9 1/4 in. thick, the foundation end rests on a sill only 1 1/2 in. thick. The deck, though level when built, will slope towards the center after the framing shrinks because of the unequal depths of wood under opposite ends of the joists. Diagonal drywall cracks, racked interior door frames, and in severe cases, separation of interior partitions from floors, are possible results.

You can steer clear of these problems by mounting joists on the face of the girder with metal hangers, or by seating them on a ledger attached to a deeper girder. This way the entire deck area will be lowered uniformly as joists shrink. Using a steel carrying beam will also do the trick. But beware, there's one trap you can fall into when face-mounting joists to glulam, laminated veneer lumber, or other engineered structural wood beams. These products are typically somewhat drier (around 12 to 15% MC) than framing lumber when sold. If you install joists with their top edges flush with the top of an engineered beam, it's likely they'll shrink below the top of the beam, creating a ridge in the floor. By mounting joists slightly higher than the top of the beam, they're more likely to end up flush with its surface after shrinking. It's a good idea to do this even when built-up girders are used. One drawback is that face-mounted joists may make utility subs' jobs harder.

Stopping the popping
The familiar fastener pop is probably the most common drywall problem that crops up when studs and joists shrink. When first fastened, drywall is driven tightly against framing. But as the wood between the fastener tip, whose position is fixed, and the edge of the framing shrinks, it pulls away from the back of the panel, leaving a small gap between framing and panel. Pressure later applied to the panel face closes the gap, forcing the fastener head to lift the taping compound. Pops are fewer and less pronounced with screws versus nails. First, for the same holding power, screws are shorter than nails, so there is less wood between the screw tip and framing face to shrink. And secondly, it takes higher pressure to force drywall along a threaded shank than it does to slide it along a smooth one.

Pops frequently appear in ceilings near the perimeter because shrinking top plates force ceiling drywall down onto the upper edge of wall panels. Prevent these pops by not using fasteners in ceiling drywall within 16 inches of walls. Pops that appear when outlet and switch plate covers are screwed down, or when interior trim is applied, may be the result of overfastening or misplaced fasteners. You can reduce the potential for pops considerably by screwing and gluing drywall. The Gypsum Association, for example, extends its screw-only on-center spacing for walls from 16 in. to 24 in. when panels are screwed and glued.

Stop the pop that telegraphs through vinyl sheet flooring by using screws or ring shank nails long enough to fully penetrate the subfloor below underlayment, and by slightly recessing their heads as is done with drywall.

Annoying floor squeaks result when subflooring and stair treads rub against the shanks of fasteners popped from joists and stringers. Happily, prevention is easy. Just lay down a bead of gap-filling construction adhesive before installing subflooring and stair treads, and the culprit gap will never form. Using ring shank or coated nails seems to help too. Framing floors with drier, engineered wood joists, which shrink minimally after installation, is also a good solution.

Understanding withdrawal
Nail pops occur inside buildings because of the initial shrinkage of the framing. But cyclical shrinkage, swelling, and warping of exterior siding, trim, and deckboards can cause nails to be partially or completely withdrawn from framing. The holding power of nails driven into green wood that stays wet, or seasoned wood that stays dry, is essentially unchanged over time. But the withdrawal resistance of nails sunk into green wood that dries in place, or seasoned wood repeatedly wetted and dried, drops substantially over time. As exterior wood swells in thickness, it pulls on nails in direct withdrawal. The pulling action is amplified in flatsawn lumber whose edges lift as it cups.

Take the sight of nail heads protruding from water-borne preservative-treated deckboards. Saturated during treatment and sold essentially green, deckboards always shrink in thickness after installation, so nails flush with the surface when driven will later protrude. And regardless of whether they're laid "bark side up" or not, flatsawn deckboards almost always cup up as the sun dries their exposed tops faster and to a lower moisture content than their shaded bottoms. Cupping reverses itself when tops are wetted by rain. Repeated reversals can slowly pry nails from framing. I recently investigated a case in which cyclic cupping of flatsawn bevel siding caused nails to be withdrawn completely. Lack of backpriming, smooth nails that were too short, and butt joints that didn't fall over framing contributed to this callback.

Reduce the potential for nail withdrawal in exterior wood by using nails of proper size, ring shank nails, or where appropriate, screws. Apply a paint, stain, or water repellent as soon as possible to reduce cyclic dimensional changes. Choose quartersawn (vertical grain) rather than flatsawn, and narrow rather than wide, siding patterns when possible. Always backprime siding, and use fasteners that penetrate solid wood (sheathing and framing) at least 1 1/2 in. Minimize moisture content and dimensional changes after installation by buying siding, trim, and treated lumber ahead of time and letting it acclimate to site conditions elevated off the ground under a loosely draped tarp. Or consider using the newly available engineered wood fiber- and flake-based siding and trim products, which are touted as being more dimensionally stable and resistant to warping.

New techniques, new trouble
While solving old problems, new technology inevitably brings with it new ones. Such is the case with the two-decade-old truss-rising phenomenon that can cause cracks to open at wall/ceiling junctions during the heating season under roofs framed with metal plate connected wood trusses. An aesthetic problem only, truss-rising is usually associated with long span trusses (>26 ft.) of low slope (<6/12), and attic insulation more than 8 in. deep. Exposed to essentially the same air temperature and relative humidity, top and bottom truss chords have about the same moisture content for most of the year. But during the heating season the moisture content of the bottom chord, smothered in insulation and surrounded by warmer air at lower relative humidity, will drop. Meanwhile, the moisture content of the top chords, enveloped in much, much colder air at higher relative humidity, may increase. As a result, the bottom chord shortens slightly, while the top chords may grow a bit longer. Lengthening of the top chords forces the roof peak higher, while webs connecting top and bottom chords lift the bottom chord and ceiling drywall attached to it. Gaps close once the heating season ends as top and bottom chord moisture contents again equalize.

Though truss-rising can't be stopped, you can mask truss movement several ways. One option is to create a floating corner by holding back fasteners in ceiling drywall about 16 inches from partitions. Then use a drywall clip fastened only to the partition to make the ceiling/wall corner. Or, omit the clip, and hide the gap by fastening corner molding to the ceiling only. Another solution calls for 2x6 blocking to be fastened to the top of partitions, with no connection to the truss. Drywall edges are fastened to the blocking, but fasteners in the field are held back 16 inches from the edge, permitting drywall to flex between the edge and field fasteners as trusses rise. Yet another option is to make partition/truss connections using L-shaped brackets attached to the bottom chord with one fastener that slides in a slot as trusses arch upward. Here too, fasteners are held back 16 inches from partitions. Never rigidly attach trusses to partitions; this could induce bending forces trusses weren't designed to carry, or cause partitions to be lifted off the floor.

When panels bow
Buckling of plywood, oriented strandboard (OSB), and waferboard sheathing and subflooring panels almost always owes to edges being tightly -and thus improperly- butted during installation. Though considerably more dimensionally stable than solid lumber, wood-based panels are typically much drier (8 to 12% MC) when sold, and should be expected to increase in dimensions when exposed to outdoor relative humidity during construction. Because panels whose edges are tightly butted can't expand laterally, they accommodate expansion by buckling outward. That's why it's important to space panels according to the recommendation stamped on each sheet, usually 1/8 inch at ends and edges. The H-clips used between panels in roofs framed 24 in. o. c. space panels just about right. Builders have rightfully complained that when 48 x 96-inch panels are spaced as recommended, their ends don't fall on the framing after five or so sheets have been laid end to end. APA The Engineered Wood Association listened; its members now produce "Sized For Spacing" panels 47 7/8 x 95 7/8 inches that always line up with framing.

But even when properly spaced, panels soaked by rain during construction, or moistened by high relative or condensation in completed attics and crawl spaces can also buckle. Buckling occurs more readily with thinner panels and longer spans, and when fasteners miss framing. It even can be built in by applying panels to warped framing. And because it adsorbs water more readily, plywood made from southern yellow pine buckles much faster than that made from Douglas-fir.

What you should do about buckling depends on its cause. With tightly butted panels, create an expansion slot between panels by sawing a kerf along the unspaced edges. Swept free of ponded water, rain-buckled subflooring will usually flatten as it dries. Extra blocking below, and a few additional fasteners may be needed to coax it flat. Some panel makers notch the tongue in their T&G subflooring panels to encourage rain to drain. Buckling of roof sheathing before shingles are applied is usually due to rain. Buckling after roofing is in place often signals an interior moisture source and inadequate attic ventilation. The best defense against buckling is proper spacing of panels, but using glues, screws, and ring shank or coated nails also helps. Buy panels ahead of time, and let them acclimate to site conditions elevated off the ground under a loosely draped tarp.

Edge swelling can also occur, especially in OSB and waferboard, because end grain and voids exposed on these panels' edges adsorb water much faster than their faces. Edges that swell after installation may telegraph through roof shingles or vinyl sheet flooring, making permanently visible a faint outline of the panel. Water-based flooring adhesives applied to underlayment can produce the same effect, but in this case, the shadow usually disappears as the water disperses throughout the panel. Most OSB and waferboard makers now seal the edges of panels with brightly colored, low-permeability coatings to minimize moisture gain during storage, shipment, and construction.

Coping with seasonal change
The source of a home's beauty and a finish carpenter's pride, few things enrich an interior more than skillfully-executed trim, stairs, and floors. But the typical wide seasonal swings in indoor relative humidity can cause perfectly-mated joints to yawn, and cracks to gape in wood strip floors. During the dead of winter indoor relative humidity may drop below 30%, while with windows open in summer it may rise to the outdoor level of 70% to 80%. The result is that the moisture content of wood indoors, which averages about 8% year-round, may drop to as low as 4% during the heating season, and climb to as high as 16% during the cooling season.

Rough lumber for millwork and flooring is initially kiln-dried to 6 to 9% MC, but there's no guarantee that it's going to stay there during shipment and storage prior to sale. That's why it's critical that doors, trim, stair parts, and flooring be acclimated on-site for a few days with indoor temperature and relative humidity maintained near occupancy levels before becoming part of the building.

On-site conditioning can minimize wood's seasonal movement, but it can't stop it. Take the case of a mitered corner joint in profiled molding, which is tightly closed most of the year. As the width of the trim changes in response to seasonal relative humidity, the joint's outside corner opens in summer, while its inside corner opens in winter.

Aware that wood movement couldn't be stopped, our woodworking forefathers allowed it to happen harmlessly through judicious design. Framed panel construction, in which a wide, bevel-edged wood panel floats in an oversized groove inside a wood frame, is a classic technique still used today for doors, cabinets, and wall panels. Pinned to the frame only at mid-width, the panel is free to expand and contract without unduly pushing or pulling on the frame or itself. Solid wood countertops, like the 36-inch wide sugar maple slab capping the peninsula in my kitchen, should have the same freedom to move. L-clips screwed to the bottom of the counter, but free to slide in the grooved frame of the supporting cabinet, not only allow it to change in width, but hold it flat as well.

I know of no better advice about wood flooring than that given long ago by an anonymous author at the USDA Forest Products Laboratory: "The cure for cracks in a floor lies wholly in preventing them." Except when indoor relative humidity is mechanically controlled, narrow cracks (1/32 to 1/16 inch with 2 1/4 inch-wide flatsawn oak) should be expected to open between some courses in wood strip floors during the heating season. But wider-than-expected cracks can develop when flooring is allowed to adsorb excess moisture before or after being laid. Flooring acclimated on-site before concrete, masonry, drywall, or plaster is thoroughly dry, or before the heating plant is operating, will likely pick up moisture and swell. Edges butted at installation will shrink apart as flooring moisture content drops during the first few months of occupancy. By the middle of the first heating season, cracks become chasms.

Even if at the proper moisture content when laid, flooring that picks up excess moisture before or after finishing in a meagerly heated, unoccupied home, can later develop wide cracks due to a phenomenon known as compression set. As moisture is adsorbed, tightly butted edges prevent strips from widening, so no apparent swelling takes place. In reality, swelling is accommodated by partial crushing of the strips' edges. Though crushed, a compressed strip will still shrink by the same percentage as an uncompressed strip. But because its swollen width is narrower than that of an uncompressed strip, its shrunken width will be narrower too, making cracks between compressed strips wider. Subsequent swelling pressure during later periods of high relative humidity can increase the amount of compression set and the width of cracks. Compression set explains why old wood floors that were mopped with water often have gaping cracks. It's also why wooden tool handles continue to loosen after soaking them in water to tighten them. The tightening is temporary; subsequent drying produces even greater looseness. The solution is to keep flooring (and tool handles) dry so compression set can't develop.

Though it can't be eliminated, wood movement can be minimized, masked, and otherwise managed through attention to detail during design, installation, and finishing.

Stephen Smulski, Ph.D., is President of Wood Science Specialists Inc., a consulting firm in Shutesbury, Massachusetts that specializes in solving performance problems with wood products in all types of wood-frame construction.

Contact Information:
Dave Damery, Director
Building Materials and Wood Technology
120 Holdsworth Natural Resources Center
University of Massachusetts, Amherst, MA 01003
Tel: +1 (413) 545-1770

Publications Article