Michael Wang

Founder & Mechanical Engineer

As the founder of the company and a mechanical engineer, he has extensive experience in advanced manufacturing technologies, including CNC machining, 3D printing, urethane casting, rapid tooling, injection molding, metal casting, sheet metal, and extrusion.

Table Of Contents

Explore 3D Printing Design, Materials & Finishing




Printing Tolerances Are Process-Specific

3D printing tolerances are process-specific: SLA holds finer detail than FDM, SLS and MJF sit in between, and metal processes have their own behavior. Specifying a printed part’s tolerance means matching the requirement to the process’s real capability and reserving critical fits for post-printing machining. This guide provides the tolerance ranges by process, explains the silent variables of shrinkage and warpage, and shows how to call out tolerances that inspection can verify.

There is no single “3D printing tolerance.” Each process builds parts differently, laser-cured, sintered, melted, or extruded, and each has its own dimensional behavior. SLA parts hold fine detail; SLS and MJF parts hold functional tolerance in nylon; FDM parts are the least precise; metal parts combine tight potential with significant thermal behavior. The practical consequence is that the tolerance on the drawing must be matched to the process. A tolerance that is natural for SLA may be impossible for FDM, and a fit that needs machining should be called out as a machined surface rather than hoped for from the printer. The same part can also print differently on two builds of the same process, so the range, not the ideal number, is the honest specification.

Design Tolerance Ranges by Process

As a reference, the stated design tolerances by process are approximately:

  • FDM. ±0.2 to ±0.5 mm.
  • SLA. ±0.05 to ±0.1 mm.
  • SLS. ±0.2 to ±0.3 mm.
  • SLM (metal). ±0.2 mm.
  • MJF. ±0.3 mm.

These are stated design ranges, not guarantees for every geometry; the actual value depends on the part’s size, orientation, and features. The table is read with the part’s features in mind. A large flat surface behaves differently from a small boss or a thin wall, and the tolerance on each is set by the process and the geometry together. The buyer should not apply the table’s range to every dimension: the critical features are called out specifically, and the rest runs at the process’s practical value. The tolerance plan is per feature, and the table is the starting point.

The feature-specific tolerance is confirmed with the supplier. A bore that must fit a bearing, a slot that must guide a slide, and a face that must mate with a seal each carry a requirement, and the supplier confirms the achievable value on the actual geometry. The confirmation is part of the quote, and it is the basis for the machining or the post-processing that follows. The tolerance that is confirmed is the tolerance that is held.

SLA 3D printing process curing a resin part

Shrinkage and Warpage: The Silent Variables

Beyond the process tolerance, printed parts shrink and warp as they cool. Shrinkage is accounted for in the process, but its consistency depends on the geometry and the build; warpage appears where sections are long, thin, or uneven. The result is that a feature that should be at tolerance can drift, especially in large or asymmetric parts.

The design response is to control the variables: orient the part to minimize warpage, keep sections uniform, and avoid features that pull the part out of shape. Shrinkage is compensated in the process, but the compensation is not perfect for every geometry, so the critical features should be designed with the shrinkage in mind: symmetric sections, uniform walls, and features that do not fight the material’s contraction. The design that helps the process is the design that holds the tolerance.

Warpage is managed with the orientation and the supports. A long thin feature that spans the build can curl; the orientation that shortens the span and the supports that hold it reduce the warpage. Flag the long thin features on the drawing, so the orientation and the support plan address them. The feature that stays straight is the one whose warp was planned. Cooling strategy is part of the same conversation: a part pulled from the build early, or one cooled too fast, carries residual stress that shows up after the first day, not at the first measurement. Where dimensional stability matters, ask the supplier when and how the part is cooled and conditioned before the critical dimensions are verified.

FDM 3D printing process extruding filament layers

Fits That Need Machining After Printing

The engineering practice for critical fits is to print the part slightly oversized and machine the fitting surface after printing. A bearing seat, a precision bore, or a mating face that must hold a tight tolerance is machined to the final value, while the rest of the part runs at the printing tolerance. This hybrid approach, print for geometry, machine for fits, is the standard answer to printing tolerance limits. The drawing should mark which surfaces are machined-after-printing, so the process and the inspection follow the same plan. The machined face is then held to the same standard as any CNC feature: the tolerance, the datum, and the tool that produces it are confirmed against the part’s function, not against the printer’s range. Oversize allowances follow the same logic: the printed stock is sized so the machining pass takes material evenly, which keeps the tool loads low and the final surface predictable.

For a printed part with machined features, the tolerance callouts are the control. A boss that must fit a bearing is called out with the final diameter and the machined-after-printing note; a face that must seal is called out with the flatness; a slot that guides a slide is called out with its width and position. Each callout ties the requirement to the process step that delivers it. The printed dimensions that are not called out run at the process’s practical range, and the buyer should not expect better. Review the drawing for the unmarked critical dimensions, because each one is a surprise waiting at inspection.

Inspection for Printed Parts

Printed parts are inspected like machined parts for the features that matter: critical dimensions are measured, machined surfaces are verified, and the datum scheme guides the measurement. The inspection depth should match the application. A functional fit needs dimensional verification; a display model needs less. An equal-limit approach, in which every surface carries the same tolerance, is usually the wrong tool for printed parts, because it spends the inspection effort on geometry that the process controls and ignores the interfaces that decide the fit.

Measurement method matters as much as measurement frequency. A caliper is right for a bore diameter, but a flatness or an angular truth on a warped face needs a surface plate or a CMM, and the method should be named with the tolerance. The measurement conventions behind such reports follow the definitions maintained by standards bodies such as NIST measurement and standards guidance.

The buyer’s question is scope: which dimensions need verification, and what report comes with the part. Specifying the inspection in the RFQ keeps the process and the price aligned, because every verified dimension adds inspection time. The critical fits are measured; the cosmetic and structural surfaces run on the process control. The inspection report should show the measured values against the tolerances, tied to the drawing, so the buyer can use it to accept the part or to drive the next revision. Ask for the report format with the order, so the inspection arrives in a usable form.

Specifying Tolerances on the Drawing

The drawing is the process plan. Separate the surfaces: printed surfaces run at the process range, machined-after-printing surfaces run at their final tolerance, and the datum scheme ties both to the part. State the nominal dimension, the acceptable range, and, where a feature is machined, the order in which the machining happens, because a machined boss that must be concentric with a printed bore depends on the datum, not on luck.

When a part needs a tight fit on a feature the printer cannot hold, the options are limited but clear: machine the feature after printing, add a transition step such as a pressed bushing that is straight from machining, or redesign the interface around the process’s strength. The choice belongs on the drawing before the quote. The tolerance conversation with the supplier is a two-way review: the buyer states the functional dimensions and the process, and the supplier confirms the achievable values and flags the features that need machining or adjustment. For parts that mix processes, the CNC tolerance guide defines the machining side of the hybrid equation.

Frequently Asked Questions

What tolerance can 3D printing hold?

It depends on the process. As a reference, design tolerances are roughly ±0.05–0.1 mm for SLA, ±0.2–0.3 mm for SLS, ±0.2–0.5 mm for FDM, and ±0.2–0.3 mm for metal and MJF.

Why do printed parts drift from nominal?

Shrinkage and warpage affect dimensions as the part cools. The consistency depends on geometry, orientation, and section uniformity, so critical features should be designed for stability.

Can printed parts hold tight fits?

Yes, by machining the fitting surfaces after printing. Print the geometry, machine the bearing seats and precision bores, and mark those surfaces on the drawing.

How should I specify tolerances on a printed part?

Separate the surfaces: printed surfaces at the process range, machined-after-printing surfaces at their final tolerance, with a datum scheme and the inspection scope defined.