"3D Printing" Is Five Different Industries

Asking whether 3D printing suits your part is like asking whether machining suits it. The answer depends entirely on which process you mean. A part printed on a desktop filament machine and the same geometry printed on an industrial powder-bed system differ by an order of magnitude in accuracy, by a factor of thirty in cost, and by a complete reversal in which directions the part is strong.

Every additive process builds up in layers, and that is where the similarity ends. Five families cover essentially everything a product developer will encounter. Knowing what each one is physically doing tells you where it will fail you.

Material Extrusion: FDM and FFF

A heated nozzle lays down molten thermoplastic in beads. It is the cheapest, most available process and the one most people picture.

Where it wins: quick form checks, fixtures, large parts, and anything where a 0.008 to 0.020 in (0.2 to 0.5 mm) dimensional expectation is acceptable. Materials run from PLA and PETG through ABS, ASA, nylon, polycarbonate, and glass- or carbon-filled grades on industrial machines. A hand-sized part costs $10 to $60 and arrives in a day.

Where it fails: visible layer lines, poor small-feature resolution below about 0.03 in (0.8 mm), and anisotropy. Layers bond by partial remelting, so pulling a part apart along the Z axis typically yields 40 to 70 percent of the in-plane strength. That directional weakness is not a defect to tune out; it is physics, and it makes print orientation a load-bearing engineering decision rather than an operator preference.

Vat Photopolymerization: SLA, DLP, and LCD

A laser or projected light image cures liquid resin one layer at a time. Layers run 0.001 to 0.004 in (25 to 100 microns) and the surface comes off the machine nearly smooth.

Where it wins: fine detail, threaded features, small snap fits, clear parts, dental and jewelry work, and any model that a customer or investor will actually hold. Tolerances of ±0.005 in (0.13 mm) are realistic on small parts.

Where it fails: long-term mechanical performance. Standard resins are brittle, notch-sensitive, and lose stiffness in sunlight and heat. Tough and high-temperature resins narrow the gap but rarely close it against an injection-molded thermoplastic. Post-processing is also real work: parts need washing and UV cure, and support marks on cosmetic faces need sanding. The head-to-head tradeoffs are worked through in FDM versus SLA.

Powder Bed Fusion in Polymers: SLS and MJF

A laser or a fusing agent plus infrared lamps sinters nylon powder layer by layer, with unfused powder supporting the part. No support structures, which means geometry freedom nothing else matches.

Where it wins: functional parts that need to behave like real plastic. PA12 nylon prints tough, chemically resistant, and reasonably isotropic, typically 80 to 90 percent of in-plane strength through the Z axis. Living hinges survive. Enclosures with internal channels come out in one piece. Because the whole build volume nests parts in three dimensions, a batch of 200 small components is dramatically cheaper per unit than any other process short of tooling. Details are covered in SLS nylon printing.

Where it fails: surface finish is grainy and porous straight off the machine, natural color is grey or white and dyeing is the practical color route, and unit cost only makes sense at quantity. A single one-off part is usually cheaper in resin.

Metal Additive: DMLS, SLM, and Binder Jetting

Laser powder bed fusion melts metal powder into fully dense parts. Binder jetting glues powder and then sinters it in a furnace, trading some density and dimensional predictability for much lower cost at volume.

Where it wins: titanium and Inconel geometry that cannot be machined, conformal cooling channels inside mold inserts, topology-optimized brackets in aerospace, and patient-specific implants. Material properties after heat treatment approach wrought equivalents.

Where it fails: cost and process burden. A palm-sized titanium part commonly lands between $800 and $4,000 once you count support removal, stress relief, wire EDM off the build plate, and finish machining of critical surfaces. Metal printing almost never eliminates machining; it replaces the roughing. Before assuming it beats a mill, read when metal 3D printing actually makes sense.

Material Jetting and Multi-Material Printing

PolyJet and similar systems spray photopolymer droplets and cure them instantly, mixing materials within a single build. That allows a rigid frame and a soft rubber-like gasket in one part, full color, and transparent sections in the same model.

Where it wins: appearance models, overmold simulation before committing to two-shot tooling, and anatomical models with varying tissue stiffness.

Where it fails: the soft materials are simulations, not silicones. They tear, they creep, and they degrade under UV within months. Treat these parts as communication tools with a shelf life, not as test articles for durability claims.

How to Choose Without Overthinking It

Start from what the part has to prove, not from what a service bureau is promoting.

  • Does it need to look right? SLA or PolyJet, then paint and finish.
  • Does it need to work mechanically? SLS or MJF, or machined plastic if tolerances are tight.
  • Does it need to be cheap and fast today? FDM.
  • Does it need metal properties? Ask whether CNC gets there first; if the geometry blocks a cutter, then powder bed fusion.
  • Do you need 300 of them? Compare MJF against urethane casting against soft tooling. Additive stops winning somewhere between a few hundred and a few thousand units.

Material selection deserves its own pass alongside process selection, since the same nominal plastic behaves differently in every family; the materials guide covers what is actually available in each.

The Trap Everyone Falls Into

The dominant failure is treating a printed part as a preview of the molded one. It is not. Wall thickness rules differ, shrinkage differs, surface texture differs, and a printed part has no knit lines, no gate vestige, and no sink marks. A snap fit tuned on a nylon SLS part will be wrong in polycarbonate. Printed threads hold until they do not. Expect that translation gap and plan for it, and keep realistic tolerance expectations in mind when a part does not fit.

The second trap is sending geometry drawn for molding to a printer and vice versa. Additive has its own rules for overhangs, minimum features, holes, and orientation, and following design for 3D printing guidelines eliminates most reprints before the file leaves your desk.

Picking the Right Process for Your Part

Projects House specifies and sources printed parts across all five families for clients developing products, matching process to the question the part has to answer and to the budget available. Send your CAD or a description of the part and what it needs to prove through our contact form.