Material Behaviour in Fabrication

When a tool meets a material, a handful of properties govern what happens - and knowing them lets you predict the result before you press start. Grain and anisotropy: wood, plywood, laminated composites and even 3D-printed parts are directional - stronger along the grain or layer, weaker and prone to tear-out across it. The same cut is clean one way and splintered the other. Hardness sets how hard the tool has to work: MDF and acrylic cut easily, aluminium needs sharper tooling and slower feeds, steel needs rigid machines and coolant. Brittleness versus ductility: acrylic and cast resin crack if you stress a thin section; aluminium and mild steel yield and bend. A living hinge that survives in plywood shatters in cast acrylic.

Thermal behaviour is the one beginners miss. A laser is heat - so it chars wood, melts and flame-polishes acrylic, and simply cannot cut most metals on a hobby machine. A mill generates heat through friction, which work-hardens some alloys and gums up thermoplastics. Moisture matters for anything organic: timber and MDF swell and shrink with humidity, so a joint cut tight in a dry workshop binds in a monsoon. Read these six - grain, hardness, brittleness, thermal, moisture and, next, spring-back - and most surprises disappear.

Bend a strip of aluminium or steel to a right angle and let go, and it will not stay at ninety degrees - it relaxes to something shallower. This is spring-back, and it is elastic recovery: part of the deformation you forced into the metal was elastic and springs out the moment the tool releases. The harder and springier the alloy, and the larger the bend radius relative to thickness, the more it recovers. A soft, dead-annealed aluminium barely moves; a spring-tempered stainless can lose several degrees.

The fix is not to fight the material but to account for it: you over-bend. If you need a finished 90 degrees and the material springs back 5 degrees, you bend to 95. Press-brake operators keep spring-back tables and test-bend on offcuts before committing the real sheet. The same logic reaches beyond metal - laser-cut living hinges relax, thermoformed acrylic pulls back toward flat as it cools, and even 3D prints warp as layers shrink. The maker mindset is to predict the recovery and pre-compensate in the file or the setup, then confirm on a test piece. Assume the material will move; design for where it lands.

Spring-back has a deeper lesson hiding in it: a material stores energy when you deform it, and it gives some of that energy back the instant you let go. That is true of a bent metal strip, a compressed foam, a stressed acrylic tab and a cooling 3D print alike. Once you start looking for it, you see the same effect everywhere - which is why experienced makers instinctively add a margin, run the first piece as a test, and trust the offcut over the datasheet. The number in a table is a starting guess; the piece in your hand is the truth.

Here is the lesson in one experiment. Draw a single tab-and-slot detail with a nominal 6.0 mm slot and cut it in three materials. In 3 mm plywood on a laser, the beam burns a kerf of roughly 0.1-0.2 mm and leaves a charred, slightly tapered edge; the ply also has internal voids and grain, so the real slot width varies along its length and the fit is a little loose and sooty. In 3 mm cast acrylic, the same laser melts rather than burns - the edge comes out glassy and flame-polished, the kerf is clean and consistent, but the material is brittle, so a tight press-fit that is fine in ply can crack the acrylic. In 3 mm aluminium, the laser is out; you switch to a CNC mill, which leaves a crisp but burred edge, needs coolant or air blast to clear chips and stop the tool welding to the metal, and holds a far tighter tolerance - but takes far longer and costs more.

Same file, three genuinely different objects. This is why matching material to process is a design decision, not an afterthought. You either pick the material to suit the geometry, or redraw the geometry - slot width, fillet radii, tab length - to suit the material. Ignore this and the fit, finish and even the structural integrity drift away from what you drew.

Every subtractive process removes a finite width of material - the kerf on a laser or waterjet, the tool diameter on a mill - and that width lives between your lines whether you planned for it or not. If your CAD says two 6.0 mm parts meet in a 6.0 mm slot, the kerf makes the slot slightly wider and the tab slightly narrower, so the joint ends up loose. Different materials, and even different sheets of the same material, shift the kerf: denser ply, thicker acrylic, a dirty lens or a worn mill bit all change it by tenths of a millimetre. That is why fablabs keep kerf and tolerance test strips - a comb of slots in 0.1 mm steps - and why a good maker cuts one before committing a whole sheet.

The number to internalise is that fabrication tolerance is a range, not a point. A hobby laser might hold plus or minus 0.1 mm; a good CNC router plus or minus 0.05 mm; a precision mill much tighter. Design your fits with that reality: a press-fit wants the tab a hair oversized, a clearance fit a hair undersized, and both depend on the material springing, swelling or charring. Decide up front which surfaces must be accurate and which can be loose - you cannot make everything tight, and chasing tolerance the material cannot hold just wastes time. A joint that must click home wants care; a panel hidden behind a wall does not, and spending precision where it never shows is its own quiet kind of waste.

Put it together into a habit. Before any job, run the material through the six questions - grain direction, hardness, brittleness, thermal response, moisture sensitivity and spring-back - and let the answers set your process, your settings and your geometry. A grained sheet? Orient the strong axis along the load and keep delicate cuts along the grain. A brittle sheet? Fillet inside corners so stress does not concentrate and crack it. A hygroscopic sheet? Cut, then acclimatise before you assemble, and leave a movement gap in the joint. A springy metal? Over-bend and test. A thermoplastic on a laser? Expect a melted edge and design the fit around it.

This is the difference between a file that might work and a part that does. The best fabricators are not the ones who know the fanciest machine - they are the ones who can pick up an unfamiliar blank, guess how it will behave under a given tool, cut a small test to confirm, and then commit with confidence. The rest of this module gives you the material families to reason with; this lesson gives you the questions to ask of any of them. Ask them every single time, on scrap first, and the machine stops surprising you.