Why Does an MJF Nylon Part's Tolerance Call-Out Matter Less Past a Certain Feature Size Than Below It?

Because MJF tolerance is specified as a band that scales with size, ±0.3 mm or ±0.3% of the dimension, whichever is greater. Once a feature is large enough that the percentage term exceeds the fixed term, the practical tolerance stops being a flat number and starts growing with the part. Below that crossover, a small feature is governed by the fixed ±0.3 mm floor and the percentage term is irrelevant. Above it, the percentage term dominates, and a tight nominal dimension on the drawing doesn't get you anything tighter than what the process delivers.

Where the crossover sits

With a ±0.3 mm or ±0.3% rule, the two terms are equal at 100 mm. 0.3% of 100 mm is 0.3 mm. Below 100 mm, the fixed 0.3 mm term is larger and that's what governs. Above 100 mm, 0.3% of the dimension exceeds 0.3 mm and that governs instead.

A small bore or boss on a part sits firmly in the fixed-tolerance regime. Specifying a feature at 0.1 mm or 0.3 mm doesn't change the real tolerance band at all, since it's capped at ±0.3 mm regardless.

A large overall envelope dimension on the same build, say 250 mm across, is already in the percentage regime. The real tolerance there is closer to ±0.75 mm rather than the flat ±0.3 mm a designer might assume applies everywhere on the part.

Why this surprises people

Designers used to machined tolerances expect a single number to apply uniformly across a part, and powder-bed processes don't work that way. The physics driving dimensional error in MJF, thermal gradients across the powder bed, fusing energy spread over a large area, part orientation in the build, scales with distance across the part, not with the absolute size of any one feature.

A small feature near the middle of a large build isn't immune to that drift. The error contribution from a feature's own size is small compared to the error from where it sits relative to the rest of the geometry. That's also part of why a critical small feature's position relative to a distant datum can be less accurate than the feature's own local dimension.

The feature itself might measure fine. Its location relative to something 200 mm away drifted with the build.

What this means for a drawing

Tightening the tolerance call-out on a small feature below the 100 mm crossover point asks for something the process can't deliver tighter than ±0.3 mm. If a feature genuinely needs tighter than that, MJF isn't the right process for it. That's a case for a secondary machining operation on the printed part, or a different process entirely.

For large features where the percentage term governs, the practical move is designing mating geometry with enough clearance to absorb a tolerance that scales with size, rather than fighting the process for a flat number it won't hold.

The practical takeaway

Know which regime your feature sits in before writing a tolerance call-out for an MJF part. DigiForge's MJF and SLS tolerance is ±0.3 mm or ±0.3% of the dimension, whichever is greater. Below roughly 100 mm that's a flat ±0.3 mm. Above it the real number grows with the part, and no amount of precision on the drawing changes either.