In metal 3D printing, “precision” is a single word doing several jobs at once. It bundles together accuracy, repeatability, resolution, and tolerance — related but genuinely different ideas. This guide defines each one, gives the tolerances the main metal processes can realistically hold, explains what drives and degrades precision (thermal distortion, layer thickness, sintering shrinkage, surface roughness), and shows how to design and finish parts so they land within spec.1,2
Accuracy, precision, resolution, tolerance — four different things
Most confusion about “precision” comes from collapsing four distinct concepts into one word. Keeping them separate is the first step to specifying a metal 3D-printed part correctly.1
- Accuracy — how close a finished dimension is to the nominal value in the CAD model. A part can be accurate on average yet vary from copy to copy.
- Precision / repeatability — how consistently the same feature comes out when the part is printed again, on the same machine or across machines. This is the statistical scatter, independent of whether the average is on target.
- Resolution — the finest detail the process can physically form, set by the laser (or binder-jet) spot size, the powder particle size, and the layer thickness. Resolution limits the smallest wall, hole, or feature you can print.
- Tolerance — the allowable band of deviation you specify around a nominal dimension (for example ± 0.1 mm). Tolerance is an engineering requirement; accuracy and repeatability determine whether the process can meet it.
In practice these interact: a process with fine resolution but poor repeatability will struggle to hold a tight tolerance, and a coarse layer thickness caps both surface fidelity and dimensional accuracy on sloped faces.2,3 When someone asks “how precise is metal 3D printing?”, the useful answer names which of these four they mean.
What tolerances can metal 3D printing actually hold?
For laser powder bed fusion (the process behind both SLM and DMLS), a widely reported general tolerance is about ± 0.1–0.2 mm (roughly 100–200 µm), with tighter values achievable on small, well-supported features and looser values on large spans, tall builds, and unsupported overhangs.4,5 Layer thicknesses are commonly 20–60 µm, while practical minimum wall, hole, and channel dimensions are usually considerably larger and depend on the alloy, orientation, powder, machine, and post-processing. Fine internal channels and lattices that cannot be machined become possible — at the cost of a rougher as-built surface than machining.6
These are process-typical figures, not guarantees. Dimensional accuracy in laser powder bed fusion is strongly governed by the specific alloy, part geometry and orientation, and the process parameters — there are more than fifty of them, and laser power and scan speed are among the most influential for both accuracy and surface finish.4,7 Machine builders reduce systematic deviations by calibrating beam offset and X–Y scaling to a benchmark artifact, which is why a qualified, calibrated machine holds tighter tolerances than a generic spec implies.4,5 For a given part, the achievable tolerance should always be confirmed with the provider rather than read off a table.
Interactive: tolerance & shrinkage visualizer
Pick a process and a nominal dimension to see the typical tolerance band, and — for binder jetting — how much the green part must be printed oversize to land on size after sintering shrinks it.
A schematic teaching aid: the tolerance band and shrinkage use representative values from the ranges discussed above, not a guarantee for any specific part, alloy, or machine.
What drives precision in metal 3D printing
Precision in metal additive manufacturing is the net result of many competing physical effects. The main ones:
Melt-pool behavior and process parameters
In laser powder bed fusion the laser creates a moving melt pool whose size and stability set how faithfully each track follows the intended contour. Too much energy drives a deep, unstable “keyhole” that traps pores and roughens the surface; too little leaves lack-of-fusion defects. Both degrade dimensional and surface quality, so the process window must be tuned per material.8,9,10,11,12,13 Vaporization, spatter, and melt flow also perturb the edge of the part.14
Layer thickness and the staircase effect
Because the part is built as a stack of finite layers, any sloped or curved surface is approximated by a series of steps — the staircase effect. It is unavoidable in principle; the only way to reduce it is to reduce the layer thickness, which sharpens contours and surface fidelity but increases build time.15 Layer thickness therefore couples resolution, accuracy on sloped faces, and speed all at once.
Thermal distortion and residual stress
The intense, localized heating and rapid cooling of laser melting build up steep temperature gradients and, with them, residual stress. When that stress exceeds what the growing part and its supports can constrain, the part warps — especially thin, tall, or unsupported regions — pulling dimensions off nominal and, in severe cases, curling enough to interfere with the recoater.16,15 This is why laser-melted parts are typically stress-relieved on the build plate before removal, and why distortion-prone geometries need supports and thoughtful orientation.17,18
Sintering shrinkage (especially binder jetting)
Binder jetting builds a bound “green” part with no melting, then sinters it in a furnace, where the metal densifies and the part shrinks — commonly by a few percent, and not always uniformly.19,20 That shrinkage must be predicted and compensated by printing the green part oversize; non-uniform shrinkage and slumping during sintering are a primary source of dimensional error in binder jetting, which is why its as-sintered tolerances are looser than laser powder bed fusion and depend heavily on the sinter cycle.20,21
Surface roughness
As-built metal AM surfaces are rougher than machined ones because partially melted powder adheres to the edges and the staircase effect textures sloped faces. Roughness is worst on down-facing overhang surfaces, where the melt pool sinks into loose powder and forms dross; it rises sharply once the overhang angle drops below roughly 45–50° from horizontal.16,22 Orientation, laser power, scan speed, and scan strategy all shift the achievable roughness, and re-melting or post-finishing can improve it.23,24,25
Supports, overhangs, and orientation
Supports anchor the part against warping and conduct heat away, but every support leaves a witness mark that must be removed — and removal itself can alter the local dimension and finish.15 Build orientation decides which faces are up-skin (smoother) versus down-skin (rougher), where the staircase effect lands, and how tall and stress-prone the build is, so it is one of the most powerful levers over precision.
Powder, calibration, and environment
Feedstock matters: powder particle size and packing set the achievable resolution and density, and poor spreading introduces defects.26,27 Machine calibration (beam positioning, focus, scaling) removes systematic offsets, and a controlled inert atmosphere limits oxidation and thermal variation that would otherwise shift dimensions and properties.4,28
Precision by process
The table contrasts the two dominant metal families for precision-relevant attributes. Read “SLM / DMLS” as one column (laser powder bed fusion). Values are typical published ranges and vary with alloy, machine, parameters, and post-processing.29
| Precision attribute | SLM / DMLS (Laser Powder Bed Fusion) | Binder Jetting |
|---|---|---|
| Typical tolerance4 | ~ ± 0.1–0.2 mm (tighter on small features) | Looser; dominated by sintering shrinkage compensation |
| Sintering shrinkage | None (part is fully melted in place) | ~ 1–3 %+, must be predicted & compensated |
| Layer thickness | ~ 20–60 µm | ~ 50–100+ µm (powder-dependent) |
| Minimum feature | Tens of microns; fine channels & lattices | Coarser; limited by powder & sintering |
| As-built surface | Moderate; down-skins rougher; often finish-machined | Can be uniform but sinter-dependent |
| Residual-stress distortion | Significant → supports + stress relief | Low (room-temperature printing); distortion is from sintering |
| Main precision limiter | Thermal distortion, staircase, down-skin roughness | Non-uniform shrinkage & slumping in the furnace |
| Tightening options | Orientation, supports, finish machining, HIP | Shrinkage modeling, sinter setters, machining, HIP/infiltration |
The pattern is consistent with the processes themselves: laser powder bed fusion fights thermal error and finishes rough surfaces by machining, while binder jetting fights shrinkage error and controls it through the sinter cycle. Post-processing such as hot isostatic pressing or infiltration mainly targets density but also interacts with final dimensions, so it is planned alongside the tolerance budget.30,31
The tradeoffs: precision vs speed, cost, and strength
Higher precision is never free. A few tradeoffs recur on almost every metal AM job:
- Precision vs speed. Thinner layers and finer scan strategies improve surface fidelity and reduce the staircase effect, but multiply build time.15
- Precision vs cost. Tighter tolerances usually mean more supports, careful orientation, slower builds, and finish machining — all of which add cost. Specify tight tolerances only where the function needs them.4
- Surface vs bulk. Parameters tuned purely for a smooth or precise contour are not always the parameters that maximize density and strength; the process window is a compromise between edge fidelity and sound bulk material.7,10
- Precision vs freedom. The processes that hold the tightest metal tolerances (laser powder bed fusion) constrain you to weldable, printable alloys and to supported geometries; looser processes may offer other advantages such as batch throughput.29
Designing and finishing for tolerance
Most tolerance problems are solved before the build starts, in design and process planning:
- Orient for what matters. Put critical, tight-tolerance faces on up-skin surfaces and away from shallow down-skin overhangs; orient tall features to limit distortion.16
- Avoid unsupported low-angle overhangs or add supports; overhangs below ~45–50° are where roughness and warping grow fastest.16,22
- Add machining stock to surfaces that must hit a tight tolerance or fine finish, and finish-machine them after stress relief. As-printed metal AM rarely meets a bearing-fit tolerance directly.24
- Model shrinkage for binder jetting (and plan sinter setters) so the green part is printed to the right oversize.20,21
- Apply GD&T and post-processing deliberately. Call out only the tolerances the function requires, and budget for HIP or infiltration where density and dimensional stability both matter.30,31
The bigger picture
Precision in metal 3D printing is not one number but a system: accuracy, repeatability, resolution, and tolerance, each shaped by melt-pool physics, layer thickness, thermal distortion, shrinkage, and surface roughness, and each traded against speed and cost. The practical path is to specify the tolerance the function actually needs, choose the process that can hold it, design and orient for it, and finish-machine the surfaces that must be exact.32
Because the two main processes limit precision in different ways, the choice between them is part of the tolerance decision — see our deeper comparison of SLM vs DMLS vs binder jetting. ACS Material’s metal 3D printing service is built on laser powder bed fusion and includes the stress relief, support strategy, and finish machining needed to hit demanding tolerances, at build volumes up to 2000 mm; we also supply the XDM metal 3D printers. Send us your drawing with its tolerance callouts and we will advise on what is achievable for your part.
FAQs
What is the average tolerance in metal 3D printing?
Laser powder bed fusion (SLM / DMLS) typically holds about ± 0.1–0.2 mm (100–200 µm), tighter on small, supported features and looser on large or unsupported ones, with layer thicknesses of 20–60 µm. These are process-typical ranges; the tolerance achievable for a specific part depends on the alloy, geometry, orientation, machine, and post-processing.
What is the most precise metal 3D printing method?
Among widely used industrial metal AM processes, laser powder bed fusion — whether branded SLM or DMLS — is usually the most practical choice for tight dimensional control and fine features, because it fully melts the powder in place and avoids the sintering shrinkage that binder jetting must compensate. The two are the same laser powder bed fusion process under different names.
Is precision the same as accuracy?
No. Accuracy is how close a dimension is to nominal; precision (repeatability) is how consistently the same result is produced; resolution is the finest detail the process can form; and tolerance is the allowable deviation you specify. A process needs both good accuracy and good repeatability to reliably hold a tight tolerance.
Why do binder-jetted parts shrink?
Binder jetting prints a bound green part with no melting, then sinters it in a furnace, where the metal densifies and the part shrinks — commonly by a few percent. The shrinkage is predicted and compensated by printing the green part oversize; non-uniform shrinkage is a main source of dimensional error in binder jetting.
How can I make metal 3D-printed parts more precise?
Specify only the tolerances the function needs; orient the part so critical faces are up-skin and away from shallow overhangs; add and manage supports; keep layers thin where surface fidelity matters; stress-relieve before removing from the plate; model sintering shrinkage for binder jetting; and finish-machine surfaces that must hit tight tolerances or fine finishes.
Do metal 3D-printed parts always need machining?
Not always, but critical fits and fine finishes usually do. As-built metal AM surfaces are rougher than machined ones, so mating faces, sealing surfaces, and bearing fits are typically machined after printing and stress relief; non-critical surfaces can often be left as-built.
References
This article is provided by ACS Material LLC for educational purposes and describes dimensional precision, accuracy, and tolerance in metal 3D printing (metal additive manufacturing). Tolerance ranges, layer thicknesses, shrinkage percentages, and surface-roughness behavior cited are typical figures from the referenced studies and general process characteristics; the precision achievable for any specific part depends on the alloy, machine, process parameters, orientation, and post-processing, and must be confirmed for your application. Consult product datasheets and safety data sheets for grade-specific specifications and handling guidance. The interactive tolerance and shrinkage visualizer is a schematic teaching tool based on representative values from the ranges described above, not predictive design or metrology software.