The Open-Day Display Stand That Split Its Joints
Liam's team has three workshop days to build a display stand for the science department's open day. The brief: 900 mm tall, 400 × 300 mm footprint, three shelves, the middle one carrying the department's 4 kg rock collection, and demountable so it fits a 600 mm deep cupboard. They have one 2440 × 1220 × 6 mm MDF sheet, PVA, 8 mm birch dowel, 4 × 50 mm woodscrews and €15 for anything else. The workshop has a pillar drill, router table, scroll saw and an FDM 3D printer. Their first build used four 8 mm dowels per shelf, drilled 30 mm into the panel edges, glued and clamped for 40 minutes. After two days under load the middle shelf has sagged 6 mm, the rear dowels have split the MDF around the holes, and the stand rocks on a floor that falls 3 mm across the footprint. The teacher will not accept a fixed, non-demountable frame, and the open day cannot move.
1. Explain, using the material and the joint geometry, why the dowel joints split the MDF. What does the failure tell you about the load path through the stand?
Read the worked answer
Material. MDF is a homogeneous fibre board with no grain. Its strength across the face (tension perpendicular to the surface) is low, roughly 0.4–0.6 N/mm², and its edge is the weakest zone because there is no continuous fibre running along it. A glued dowel relies on the glue line and the surrounding fibre holding in tension.
Geometry. An 8 mm dowel cannot be housed in a 6 mm panel edge: the hole is wider than the panel is thick. Drilling it breaks out through both faces, leaving a crescent of material less than 2 mm thick on each side and a glue line on only part of the circumference. The effective glued area is perhaps 200–250 mm² instead of the 750 mm² a full cylinder would give.
Load path. The 4 kg load deflects the 6 mm shelf, which tries to rotate at its ends. Rigid dowel joints cannot rotate, so the rotation is resisted by the rear dowels being pulled straight out of the panel face - the direction MDF resists least. The dowels act as levers prying chunks of fibre out, which is exactly the split seen.
Conclusion. The shelf load is travelling into the joints as tension rather than being carried in shear and bearing. A 6 mm MDF shelf spanning 400 mm will also creep permanently under a sustained 4 kg, so the 6 mm sag is partly the shelf, not only the joints.
2. Redesign the shelf-to-side connection so the stand can still be built in three days, stays within budget and can be taken apart. Give at least three specific dimensions.
Read the worked answer
Principle. Take the load out of the MDF edge in tension and put it into bearing on timber rails and cleats, then make the frame demountable with bolts rather than glue.
Proposed build.
- Glue and screw a 44 × 44 mm softwood rail along the inside face of each side panel, 25 mm below each shelf line (6 × 40 mm screws at 100 mm centres).
- Fix 20 × 20 mm softwood cleats to the rails; the shelf drops onto the cleats, so the load is carried in shear and bearing, not tension.
- Stiffen each shelf by gluing and pinning a 20 × 20 mm softwood lipping to the front and rear edges. This turns the 6 mm panel into a shallow beam and cuts deflection dramatically.
- Add one 20 × 20 mm rib under the middle shelf at mid-span.
- Join sides to rails with M6 × 60 mm bolts and cross dowels, two per rail end, so the frame comes apart with a hex key.
Why it works. The cleat bears on the rail over its full length, so the joint no longer pries the panel face. Bolted joints are removable and can be re-tightened as the timber seasons.
Cost and time. Roughly 3.5 m of 44 × 44 and 4 m of 20 × 20 softwood, well inside €15; cutting, drilling and assembly is about four hours, so it fits the three days. No 8 mm dowels are used at all.
3. A teammate suggests printing the corner joints on the FDM printer instead, in PLA, to save assembly time. Give a reasoned judgement.
Read the worked answer
Verdict: not for the load-bearing joints.
Reasoning.
- Creep. PLA creeps under sustained load at room temperature. The 4 kg sits on the stand for days, not seconds, so a printed bracket that holds when first assembled will relax and loosen. The load is exactly the kind (long-term, static) that printed parts handle worst.
- Heat. PLA's heat-deflection temperature is around 55 °C. A stand near a sunny window or radiator can soften.
- Anisotropy. FDM parts are weakest between layers. A corner bracket loaded in peel will split along layer lines unless printed flat so the layers run across the load - and even then it is brittle.
- Time. Four brackets at 40–60 minutes each is three to four hours of printing, so the time saving is largely imaginary.
Where printing is justified. A small cable clip, a shelf-edge label holder, or a locating spigot that takes no load. If a printed part must be used structurally, print at 100% infill with five perimeters, orient the layers across the load, and put a steel bolt or pin through it as the real load path.
Better use of the printer. Print a drilling jig for the M6 bolt holes so the demountable frame lines up accurately on assembly.