Balanced veneer

Why does a cabinetmaker glue an expensive sheet of veneer onto the hidden back of a panel nobody will ever see?

A cabinetmaker lays veneer on the face of a door panel, then lays a second sheet on the back — the side that will face a wall, in the dark, for the life of the piece. Asked why, the shop answer is "to balance it", which names the practice without explaining it.

It looks like a habit a trade keeps out of pride. So ask something narrower than "why balance": what force is the back sheet opposing, given that nothing is pushing on the panel at all?

Start with what veneer does after it is glued. It is thin wood, and wood changes dimension with the moisture it holds — markedly across the grain, very little along it, the difference being roughly an order of magnitude (a recalled ratio; species vary widely). The adhesive is usually water-based, so the sheet swells as it goes down, then contracts as it dries and the glue cures.

Now that sheet is stuck to a core that did not swell nearly as much. It wants to be shorter than the surface it is bonded to and cannot shrink, because the thing it is glued to is not shrinking with it. What can the assembly do with that mismatch?

Only two things, the same two available to any bonded pair of unlike layers: hold the difference as stress, or convert some of it into shape. A flat panel with a shrinking skin on one face resolves it by curving toward that face — the arithmetic of a bimetallic strip, with humidity in the role of temperature. The panel cups.

How much depends on geometry rather than on how hard the veneer pulls. For a thin skin on a thick substrate the curvature goes roughly as the inverse square of the core thickness (Stoney's classical result, which I am recalling rather than deriving). That predicts what a cabinetmaker recognises: a 6 mm panel cups visibly under a skin that leaves a 25 mm worktop flat, and the penalty is steep — not four times but about seventeen, for that ratio of thicknesses.

So what does a second veneer do? Notice what it does not do: relieve the stress. The face veneer still pulls, and now the back one pulls too. What has changed is symmetry — both pulls are equal and sit at equal distances either side of the panel's mid-plane, so their bending moments are equal, opposite, and cancel. A balanced panel is not a relaxed panel; it is a stressed panel with no reason to bend either way.

That distinction is not pedantry, because it fixes what "matching" must mean. The backer must match the face in whatever governs the size of its pull: species movement, thickness, grain direction, cut. Lay it with its grain crossed against the face and nothing has been balanced — it moves in a different direction by a different amount, and the panel twists as well as cups.

It also fixes what matching need not mean. Beauty is irrelevant to the mechanics, which is why the trade word is "balancer" and why the hidden sheet is usually a cheap veneer of matched movement. The question's premise holds only in one case: when the show face is an unusual species whose movement nothing cheap matches, the cabinetmaker really does bury money in the back.

The principle predicts something beyond veneer, too. A finish film is a layer — lacquer slows moisture exchange. Seal one face and leave the other bare and the two equilibrate at different rates with the seasons, so the panel cups though the veneers were matched. Hence the shop rule to finish both sides, which most people take as tidiness.

Is any of it inertia? Some. A panel screwed rigidly into a carcase is restrained by it and often gets away unbalanced; and suppliers require balanced construction as a warranty condition, so a backer sometimes goes on as insurance against a claim. The rule is tuned for the cases that matter and applied everywhere, because sorting them costs more thought than the veneer costs money.

The misconception to retire is that the backer resists the face veneer, as though bracing the panel. Nothing is being resisted. The two skins never touch and never oppose each other directly; they produce bending moments about the same neutral plane, and equal moments in opposite senses sum to nothing.

Two refinements. Symmetry is a property of the whole stack — core, glue lines, veneers and finish all count, which is why plywood is built in an odd number of plies with matched pairs either side of a centre. And the balanced stress is still there, still available: cut a balanced panel down the middle and each half may bow, because the cut releases restraint and the offcut is no longer geometrically identical about its own centre.

The test is whether cupping tracks veneer or tracks asymmetry. If symmetry about the mid-plane is the mechanism, a panel veneered both faces but finished on one should cup with the seasons; an unveneered board sealed on one side should cup too; and a thick core should cup far less than a thin one under the identical skin. Find single-veneered panels sitting flat through large humidity swings while matched double-veneered ones cup, or finishing asymmetry making no measurable difference, and this account is wrong — the cause would lie in the veneer or glue itself, and the back sheet would be doing nothing.

• Why solid-wood panels are left floating in their frames instead, and what that solves that balancing cannot.
• How the same symmetry argument sets the ply count and layup of plywood and laminated glass.

Every term the collection defines is gathered in the glossary.