Supports, Orientation & Failure

Because each layer is laid on the one below, a surface that leans outward as it rises - an overhang - gives every new layer less and less to sit on. Up to a point the printer copes: each layer overhangs the last by a fraction of a bead, and the part climbs at an angle happily. Past a critical steepness it does not, and the fresh material droops into empty space.

The rule of thumb is the 45-degree rule: an overhang shallower than about 45 degrees from vertical will usually self-support, while anything steeper needs help. A 45-degree wall means each layer is offset by roughly half a bead - still mostly on solid material. At 60 or 70 degrees, most of each new bead hangs over air and sags. The exact angle depends on process, cooling and settings, but 45 is the number to carry in your head.

A related trick is bridging: a printer can span a short horizontal gap between two supported points by stretching a bead across, with no support beneath, as long as the span is modest and cooling is good. So a flat roof over a gap can sometimes print unsupported where a slow-rising overhang cannot. Learn to spot overhangs in a model before you print - they are where trouble and supports both come from.

Why 45 degrees, and not some other number? Because at that angle each new layer sits half on the layer below and half over air, so there is just enough solid material beneath the fresh bead to catch and hold it. Push past that and the unsupported fraction grows until the plastic sags before it sets. The exact limit shifts with the material, the cooling and how thin you slice - a well-cooled PLA print may hold 50 or 55 degrees cleanly, a hot ABS one less - but 45 is the safe figure to design and reason with, and it applies, in spirit, to every layered process from FDM to resin.

When geometry does exceed what self-supports, the slicer builds temporary scaffolding. Supports are sacrificial pillars or lattices printed under overhangs and bridges to give each layer something to land on; you snap or dissolve them off afterwards. They are a necessary evil: they cost material and time, they scar the surface where they touch (the contact points always leave marks), and they can be fiddly to remove from delicate features. Good slicers let you tune support density and place them only where needed - and dual-material machines can print dissolvable supports (like PVA) that wash away cleanly, invaluable for enclosed geometry.

Two more helpers address the base rather than overhangs. A brim is a flat single-layer skirt printed around the part, fused to it, to widen its footprint and resist lifting - cheap insurance for tall or small-footprint prints. A raft is a thicker printed platform the whole part sits on, improving adhesion and giving an even base on an imperfect bed, at the cost of material and a rougher bottom face.

The craft is to need as few supports as possible - because the best support is one you designed out by orienting or shaping the part well, which is exactly the next idea.

How you place a part on the bed is one decision that quietly controls three things at once, and getting it right is the single highest-leverage choice in additive.

First, strength. Layered parts are anisotropic - the bond between layers is the weak direction. So orient the part so that the main load runs across the layers, not pulling them apart. A hook printed lying down (load across layers) is far stronger than the same hook printed standing up (load along the weak layer bonds). This one choice can double or halve a part effective strength.

Second, surface and detail. The top and bottom faces, and any near-vertical walls, come out cleanest; steep overhangs and the support contact points come out worst. So orient the most important face - the one people see, or a mating surface that must be accurate - where the process treats it best.

Third, supports. A clever orientation can put overhangs where they self-support or where supports are easy to remove, sometimes eliminating supports entirely. These three often conflict - the strongest orientation may need the most supports or spoil the show face - and resolving that trade-off deliberately, rather than accepting the default, is what separates a considered print from a lucky one.

When prints fail, the causes are a short list you can learn to diagnose by sight. Warping is corners or edges lifting off the bed as the part cools and shrinks unevenly, most common in ABS and large flat parts. Fixes: a heated bed, a clean level surface, a brim, and an enclosure to hold heat so the part cools slowly and evenly.

Stringing (or oozing) is thin wisps of plastic strung between separate features, left when the nozzle drags molten filament across gaps. Fixes: enable or increase retraction (pull the filament back during travel moves), lower the nozzle temperature a little, and dry the filament - damp filament stringing is very common.

Delamination (layer separation) is layers splitting apart, either mid-print or under load - a sign the layers never bonded properly. Fixes: raise the nozzle temperature, slow down, and stop draughts and over-aggressive part cooling, especially on ABS. Other regulars include elephant foot (a bulging squashed first layer - lower first-layer temperature or bed height), under-extrusion (gaps and thin lines - clogged nozzle, low flow or damp filament), and layer shifting (a stepped offset - a mechanical or belt problem).

The universal discipline is the same one from the FDM lesson: read the print, change one variable at a time, reprint a small test, and log what happened. Most trouble traces to temperature, adhesion or cooling - and a methodical operator fixes it far faster than a panicking one.

The deepest fix for supports, orientation headaches and failures is upstream: shape the model so the printer barely has to fight. This is design for additive manufacturing, and a handful of habits remove most trouble before it starts. Chamfer, do not overhang: replace a flat horizontal underside with a 45-degree chamfer and it self-supports - a tiny change that deletes a raft of supports. Add fillets at the base of tall features to spread stress and cut warping at sharp inside corners. Respect the minimum feature size: do not model a wall or pin thinner than the process can print (roughly a nozzle width in FDM), or it simply will not appear. Keep large flat areas off the bed where you can, since they are the most warp-prone, or break them up.

Two more moves punch above their weight. Split and rejoin: a model that is a nightmare in one piece - deep overhangs, a show face that fights the strong orientation - often becomes trivial as two well-oriented halves glued or pinned together, a standard professional trick rather than a cheat. And design the clearance in: parts that must mate - a lid on a box, a peg in a hole - need a deliberate gap (often a few tenths of a millimetre) because prints run slightly oversize, so test the fit on a small coupon before committing to the whole part. Think of every model as a set of instructions to a machine that can only build upward, layer on layer - design with that grain, not against it, and reliability stops being luck.