Global 3D Printing Technology: How the Processes Work and Where They Fit

• 1
  What global 3D printing technology actually does layer by layer
• 2
  The main process families and what each one is good at
• 3
  Where the cost curve actually sits
• 4
  Where printing stops being the right answer
• 5
  How printing and machining combine in production
• 6
  Process selection by geometry, quantity, and tolerance

Every additive machine does the same three things: it slices a solid model into layers, it deposits or solidifies material inside each layer, and it bonds that layer to the one below. The differences between processes come down to how the material is held and how the bond is formed. That single distinction explains almost every cost and tolerance difference you will see in a quote.

Powder bed fusion spreads a thin layer of metal or polymer powder and melts it with a laser or electron beam. Material extrusion, the FDM family, pushes a softened filament through a nozzle. Vat photopolymerization cures liquid resin with a light source. Binder jetting glues powder together and sinters it later. Each route has a different thermal history, and thermal history is what sets residual stress.

Layer thickness controls the trade-off between surface finish and build time. A 0.1 mm layer halves the step height of a 0.2 mm layer but roughly doubles the number of passes. For metal powder bed parts, layer thickness also changes how much energy reaches the previous layer, which affects porosity. You cannot simply dial layers thinner without rechecking the melt parameters.

The bond between layers is rarely as strong as the bulk material. In extrusion, the interface is a partial weld. In powder bed fusion, it is a re-melted region with its own grain structure. This is why printed parts often show anisotropic strength: strong along the layer plane, weaker across it. Orientation on the build plate is therefore a design decision, not a machine setting.

• 1
  Layer thickness sets finish and time0.05–0.2 mm is the common band for metal; thinner layers cost build hours.
• 2
  Thermal history drives stressFast cooling leaves residual stress that can warp thin walls.
• 3
  Anisotropy is realStrength differs between in-plane and cross-layer directions.

Material extrusion is the cheapest entry point and the most forgiving of poor geometry. It handles large parts, internal channels, and prototypes that only need to fit. Its limits are obvious: visible layer lines, limited resolution around 0.2 mm, and weak interlayer bonding under load. For jigs and fixtures that never see high stress, it is often the right answer.

Vat photopolymerization gives the best surface finish of the resin processes. A 0.05 mm layer produces a smooth skin that can be sanded or plated. The catch is material behavior: most resins creep under sustained load and degrade under UV. They are excellent for form-and-fit models, molds for vacuum casting, and visual prototypes, but poor for functional parts that carry load for months.

Powder bed fusion is where metal parts become real. Laser powder bed fusion reaches densities above 99% in alloys such as 316L, Ti-6Al-4V, and Inconel, with mechanical properties close to wrought material after heat treatment. It is the process that made conformal cooling, lattice structures, and consolidated assemblies practical. It is also the most expensive per kilogram of feedstock and the most sensitive to support strategy.

Binder jetting sits between the two. It builds fast because the whole layer is deposited at once and no melt pool is involved, then sinters in a furnace. Distortion during sintering is the main risk, and it scales with part size. For small, complex parts produced in volume, the economics can beat powder bed fusion. For one-off large parts, it usually does not.

• 1
  ExtrusionCheap, large, low resolution. Good for fit checks and fixtures.
• 2
  Vat photopolymerizationFine finish, weak long-term creep resistance.
• 3
  Powder bed fusionDense metal, high cost, support-sensitive.
• 4
  Binder jettingFast build, furnace sintering distortion risk.

Additive cost has almost no tooling component. That is the whole point. A machined part needs fixturing, program setup, and first-article inspection before the first good part appears. A printed part needs a build file and a support strategy. At quantity one, printing wins on setup alone, which is why it dominates prototype work.

The curve crosses over somewhere in the tens to low hundreds of parts, and the crossing point depends on geometry. A simple bracket with no internal features is cheaper to machine once you pass a few dozen pieces, because the cycle time is short and the setup is amortized. A manifold with internal channels that cannot be reached by a cutter stays cheaper to print far longer, because the alternative is joining several pieces.

Material cost per cubic centimeter favors machining for common alloys. Aluminium 6061 bar stock is inexpensive and widely available. Metal powder for powder bed fusion costs far more per kilogram and much of it is lost to the build plate and supports. This is why printed metal parts are usually small and feature-dense rather than large and simple.

Post-processing is the hidden line item. Support removal, stress relief, heat treatment, surface finishing, and sometimes HIP all add cost and lead time. A printed metal part that needs support removal by hand in a tight internal channel can cost more to finish than to build. Budget for finishing before you commit to a printed design.

• 1
  No toolingPrinting wins at low quantity because setup cost is near zero.
• 2
  Crossover by geometrySimple parts cross to machining sooner than feature-dense ones.
• 3
  Post-processing is realSupport removal and heat treatment can exceed build cost.

Tolerance is the first boundary. As-built powder bed fusion holds roughly ±0.1 mm on small features and worse on long dimensions, and the surface sits around Ra 8–15 μm. If your drawing calls for ±0.005 mm or Ra 0.8–1.6 μm, the printed part is a blank, not a finished part. It has to go onto a CNC for the critical faces, bores, and sealing surfaces.

Surface integrity is the second boundary. Printed surfaces in load-bearing applications carry as-built roughness that acts as a stress riser. Fatigue life in printed aluminium and titanium improves substantially after machining the critical surfaces and applying stress relief. For rotating or cyclically loaded parts, printing the near-net shape and machining the functional surfaces is the standard route.

Size is the third boundary. Build envelopes cap part size, and larger envelopes usually mean coarser resolution and more residual stress. When a part exceeds the envelope, you either split it and join it, or you machine it. Splitting introduces a joint whose strength and leak-tightness you now have to qualify. That cost is easy to underestimate.

Material certification is the fourth. Regulated industries need traceable feedstock, qualified parameters, and documented heat treatment. Not every print shop can supply that. If the part goes into a medical device or an automotive safety system, the process qualification matters as much as the geometry.

• 1
  ToleranceAs-built ±0.1 mm; critical faces still need machining.
• 2
  FatigueAs-built roughness reduces fatigue life in loaded parts.
• 3
  SizeEnvelope limits force splitting or a switch to machining.
• 4
  CertificationTraceable powder and qualified parameters are not universal.

The most practical use of additive in production is as a near-net shaping step. Printing gets you a blank with internal channels, lattices, or consolidated geometry that would be expensive or impossible to cut. Machining then establishes the datums, bores, threads, and sealing faces that carry tolerance. Neither process does the other's job.

Conformal cooling in injection molds is the clearest example. A printed mold insert with cooling channels that follow the cavity can cut cycle time and reduce warpage. But the parting line, the locating features, and the cavity surface still need to be machined and polished to the required finish. The insert is printed, then finished on a 3-axis or 5-axis mill.

At GreatLight, printed and machined work runs on the same floor. We operate 127 high-precision CNC machines, including 16 simultaneous 5-axis machining centers and 16 mill-turn centers, with a maximum processing size of 4,000 mm. Machining tolerance is held to ±0.005 mm and finishes down to Ra 0.2–0.8 μm when the drawing calls for it. We also offer custom 3D printing as a standalone service.

The decision rule is straightforward. If the geometry is simple and the quantity is more than a handful, machine it. If the geometry is complex, the quantity is low, or the internal features cannot be reached, print it and machine the critical surfaces. If the part must carry cyclic load or seal, assume the functional surfaces will be cut.

• 1
  Print for shapeInternal channels and lattices come free with additive.
• 2
  Machine for toleranceDatums, bores, threads, and seals still need cutting.
• 3
  One floorKeeping both processes together shortens the loop.