Wood and 3D Printed Structures: How The Warp Tea Room Was Built

• 1
  What extrusion printing actually does with wood
• 2
  Why recycled wood changes the print
• 3
  Where the accuracy actually goes
• 4
  How print and CNC work together
• 5
  When printing beats machining, and when it does not
• 6
  Wood-composite printing vs CNC machining vs hybrid

Most people picture a desktop printer squeezing out plastic filament. Large-format extrusion printing works differently. A robotic arm or gantry carries a nozzle across a build volume that can reach several meters, and the nozzle deposits a paste-like compound in continuous beads. The bead is usually 8 mm to 30 mm wide. Each pass bonds to the one below it before the material sets.

When the feedstock is wood, the compound is a mix: recycled timber flour or sawdust bound in a binder, sometimes with a mineral filler to control shrinkage. The wood content typically runs 30% to 60% by weight. Higher wood content gives a warmer surface and better stiffness after curing, but it also raises viscosity and makes the bead harder to pump. Printers that handle this material run at lower speeds than plain concrete or clay extruders.

The Warp tea room sits at the small end of this scale. It is an architectural piece, not a load-bearing wall, so the print carries its own weight and little else. That distinction matters. A structure that only supports itself can accept a coarser bead and a looser tolerance than a machine bracket that has to hold a bearing in place.

• 1
  Bead widthTypical range 8–30 mm for architectural extrusion; thinner beads need slower feed and tighter rheology control.
• 2
  Wood fraction30–60% by weight. More wood means more warmth and stiffness but harder pumping.
• 3
  Cure behaviorBinder sets first, then the part dries. Shrinkage during drying is where most dimensional error appears.

Virgin timber has predictable grain, moisture and density. Recycled wood does not. Offcuts, pallet boards and demolition timber arrive with different species mixed together, different moisture content, and embedded metal. Before any of it becomes feedstock it gets screened, dried and ground to a consistent particle size. Moisture control is the step that decides whether the final part cracks.

Particle size distribution controls flow. Fine flour packs tightly and produces a dense, smooth bead. Coarse chips leave voids and a rougher surface, which can be attractive for a tea room wall but bad for a part that needs a sealing face. Most producers blend fine and coarse fractions to balance flow against shrinkage.

There is also a sustainability claim to examine carefully. Using recycled wood reduces demand for new timber, but the binder, the transport of feedstock and the energy used in drying all carry their own footprint. For a single architectural piece the material choice is defensible. For a high-volume part, run the numbers before assuming recycled wins.

• 1
  Species mixingDifferent woods shrink at different rates; blending without testing invites warping.
• 2
  Moisture targetFeedstock is usually dried to a narrow band before grinding, or the bead tears.
• 3
  Metal contaminationScrews and staples in reclaimed timber damage grinders, so screening comes first.

Extrusion printing is not a precision process. A printed wall that should be flat will show layer lines and slight bulging where the bead overlaps. On a large architectural print, dimensional variation of a few millimeters is normal and rarely matters. On a part that bolts to something else, it matters a lot.

The error comes from three places. First, the bead itself: it spreads under its own weight, so the top of a layer is never perfectly flat. Second, shrinkage during drying, which pulls the part inward and can bow long spans. Third, machine motion, which on a large gantry is less stiff than on a small mill.

The practical answer is to print oversize and machine the critical surfaces afterward. A printed blank with 2 mm to 5 mm of stock left on each functional face gives a CNC mill something to cut. The mill then holds ±0.005 mm on those faces and Ra 0.8–1.6 μm if the surface needs to seal or slide. Everything that stays printed keeps the soft, layered look.

• 1
  As-printedMillimeter-level variation; fine for sculpture, walls and non-mating surfaces.
• 2
  Machined after printing±0.005 mm on the cut faces, with a clean edge where the tool passed.
• 3
  Stock allowanceLeave 2–5 mm per functional face before machining, or the tool cuts into voids.

The most useful way to think about wood and 3D printed parts is as a hybrid: print the bulk shape, machine the interfaces. A tea room frame, a display plinth or an architectural panel uses the printer for its organic form and the mill for the flat seats, bolt holes and edges that have to line up.

Fixing the printed blank for machining is the hard part. Wood-composite prints are softer than aluminum and can crush in a vise. Vacuum fixtures, soft jaws and low clamping pressure work better. Cuts should be light and fast enough to clear chips, because the material abrades tooling faster than solid timber does.

Tooling choice follows the same logic. Carbide cutters with a sharp geometry and a modest helix angle clear the composite without pulling fibers. Coolant is often skipped; air blast keeps the dust moving and avoids swelling the binder. Feeds run slower than on aluminum, closer to what you would use on a plastic like POM.

After machining, the cut faces expose raw composite. A sealer or a light sanding restores the look, and the join between printed and machined surfaces becomes a deliberate design line rather than a defect.

• 1
  FixturingVacuum tables and soft jaws; clamp pressure low enough to avoid crushing the bead.
• 2
  ToolingSharp carbide, air blast instead of flood coolant, feeds similar to POM.

Printing wins when the shape is complex, the volume is low, and the load is modest. A one-off architectural element, a large curved panel, a form that would need a five-piece mold: these are good candidates. Printing also wins when the recycled-wood look is part of the brief and you want the material story to be visible.

Machining wins when the part has to fit something else. Bearing bores, sealing faces, threaded holes and anything measured in tens of microns belong on a mill or a lathe. If you need 10,000 identical brackets, printing is the wrong tool; die casting or injection molding will cost less per part.

A middle path covers a lot of real work. Print the form, machine the critical features, then finish. That is how a structural prototype with a wood-composite body and a machined metal insert gets made. The insert takes the load, the printed body takes the shape, and each process does what it is good at.

The decision should be made before the geometry is frozen. If you design a part assuming printed tolerance and then need a mating face, you may not have left enough stock to machine it. Bring the process choice into the first design review, not the last.

• 1
  Choose printingComplex geometry, low volume, light loads, visible material character.
• 2
  Choose machiningMating faces, tight bores, threaded joints, high volumes where tooling pays off.
• 3
  Choose bothOrganic body plus machined interfaces; leave stock on any face that must fit.