The Blanket Box That Opened at the Corner
A class is making blanket boxes about 900 × 450 × 450 mm from 20 mm solid pine. One student has built the carcase as butt joints held with 8 mm dowels and PVA only, and has made a solid one-piece lid 900 × 450 mm, glued along the front rail and fixed with ten 4 × 40 mm screws in tight clearance holes. She has fitted a strap hinge and a small brass stay, and the base is 6 mm ply nailed to the bottom rails. Six weeks later, in January, two splits have opened along the grain at the screw positions and a 3 mm gap has appeared at the front of the lid. The carcase is slightly out of square, and one corner joint opened a little when the box was dragged across the floor during a load test. The workshop is at 20 °C and 45% RH; the box is going to a bedroom heated to 21 °C, with a window that is often left open.
1. Diagnose the lid failure and estimate the amount of movement involved. Then redesign the lid fixing, explaining how the new detail works.
Read the worked answer
Diagnosis. The lid is a 450 mm-wide solid pine panel, so the grain runs along the 900 mm length and the 450 mm dimension is the cross-grain width. That is the direction in which the board will always try to move, and no finish, glue or screw will stop it. Pine moves roughly 0.25–0.3% of its width tangentially for each 1% change in moisture content, so across 450 mm each percentage point of moisture change is worth about 1.2 mm. A cold bedroom with an open window in January might sit at 60–70% RH, which is an EMC of around 12–13%, against roughly 8% in the workshop at 45% RH. That is four to five points, so 5–6 mm of swelling across the width - and the same amount of shrinkage if the room later dries out. The lid was glued to the front rail and screwed down in tight holes, so it was fully restrained. The restraint transfers the movement into tension across the grain, and the stress concentrates at the screw holes, which is where the splits started.
Redesign. Keep the solid lid but let it move. The most robust school solution is buttons or figure-of-eight fasteners in a groove or rebate in the rails: they hold the lid down but allow it to slide sideways. A simpler alternative is slotted holes with washer-head screws, with the slots running across the width and the screws only lightly nipped up. Glue nothing to the rails. Fix positively near the centre and let movement work outwards. If the design is changed to a frame-and-panel lid, the panel must sit in a groove with 1–2 mm clearance per 100 mm of panel width, and the frame carries the glue and the hinges.
Point to stress to the student. The failure is not that the wood moved; it is that the design assumed it would not. Removing movement is impossible - accommodating it is the whole craft.
2. Evaluate the carcase joint choice. Say what went wrong mechanically and specify a better joint with reasons.
Read the worked answer
What was used. Butt joints with 8 mm dowels and PVA form a joint that relies almost entirely on glue. Dowels into end grain of pine are a poor bet: end grain is porous and starved of glue, the glue line is brittle, and the dowel has little mechanical key beyond its own sides. Worse, the joint has almost no tensile strength across the corner, which is exactly what a box needs when it is pushed or dragged: the load racks the carcase and tries to pull the corners apart. The load test confirmed it. A 900 mm box also has a long lever arm, so even modest racking puts real force on those corners.
Better options, in order. Through dovetails are the traditional answer for a box: the interlocking geometry resists tension mechanically rather than by glue, and a well-cut set will survive a carcase being dropped. A lapped or stopped dovetail does the same job with less show. If dovetails are beyond the group's skill for this project, a loose tenon joint, or a rebate with screws and glue plus a glued corner block, is far stronger than dowels, and a rebate gives a gluing surface along the grain rather than the end. Because the carcase is being painted or stained, a rebated and screwed corner with a corner block is an honest, teachable alternative.
Also fix the rest of the structure. The 6 mm ply base nailed to rails does almost nothing for stiffness or squareness. A base panel in a groove cut into the sides and front/back holds the carcase square during glue-up and keeps it square forever - a much better use of a sheet material than nails. Check the whole design for a back panel or base that acts as a stiffener before blaming the joints alone.
3. The student wants to 'use more glue and longer screws'. Respond to that, and set out an inspection and test routine you would use before the next glue-up.
Read the worked answer
Why more glue and longer screws will not work. The problem is not the strength of the fixing; it is the direction of the fixing. A longer screw goes deeper into the same cross-grain, restrained condition and simply moves the stress concentration further into the board, where it can split a thicker section. More glue in a restrained cross-grain joint makes the failure more violent when it comes, because the joint can no longer creep to relieve itself. For the corners, more glue on dowels will still be in end grain and will still be starved. The student needs to change the detail, not increase the quantity.
Before the next glue-up - inspection and test routine. Measure the moisture content of the pine on a freshly planed face at several points and compare it with the EMC of the room the box is going to, using a hygrometer. Dry-assemble the carcase completely and check square by measuring both diagonals with a tape and comparing within 1–2 mm; then check for wind by sighting across the top corners or using winding sticks. Fit the lid dry, mark the fixings, and check that every fixing has room to slide across the width - measure the allowed movement against the expected movement calculated above, so the detail is designed rather than guessed. Load test the empty carcase by dragging it across the floor and by pressing on one top corner, watching for any movement at the joints. Photograph the dry assembly and record the checks. Then glue up once, re-check the diagonals while the glue is still open and pull them straight with a sash clamp, and inspect after the clamp comes off, not a term later.