Hole Clearances for 3D Printing

Do printed holes come out at the size in the CAD file?

No: Stratasys Direct's FDM Design Guide states that holes, including ones inside bosses, are generally fractionally undersized and are bored or reamed during post-processing to get an accurate diameter.

That is a direct statement from an OEM design guide, not a general rule of thumb someone repeated online. The guide describes the direction of the error, smaller than nominal, without putting a number on how much.

Why won't anyone publish a single undersize number for FDM holes?

Protolabs Network states plainly that the amount of undersize depends on the printer, the slicing software, the hole size, and the material, and recommends printing a hole undersized and drilling or reaming it to the finished diameter afterward.

No OEM, bureau or standard reviewed for this page publishes one offset figure that applies across machines and materials. A single number here would be a guess dressed up as a spec.

What printed-hole fit clearances has a manufacturer actually published?

Markforged's Composites Design Guide gives three named fit classes: press fit at 0.00 to 0.05 mm diametral clearance, close fit at 0.05 to 0.10 mm, and free fit at 0.10 to 0.20 mm. These figures are measured on Onyx material with 100 micron layers, not a material-agnostic FDM number.

How much clearance does a moving printed hole need?

Stratasys Direct's FDM Design Guide states 0.015 in, about 0.381 mm, of clearance in XY for a moving fit, with Z clearance set by slice thickness instead. Prusa's Knowledge Base gives a simpler rule of thumb of at least 0.3 mm for a movable part.

What is the smallest hole FDM can reliably print?

Markforged's Composites Design Guide sets a minimum hole size of 1.5 mm in XY and 1.0 mm in Z, and states that a hole on a horizontal surface, printed in Z, comes out more precise than one on a vertical surface printed in XY. Xometry's guide adds a simpler floor, that a hole under 1 mm is not reliable and may need to be drilled after printing.

What did one real ABS hole measurement show?

In one test setup, not a general rule, Zielinski and coauthors printed a 2 mm nominal ABS hole that measured 1.06 mm and a 5 mm nominal hole that measured 4.16 mm.

What did one real PLA gear bore measurement show?

In a separate single test setup, Wei and coauthors measured a printed PLA gear bore at 0.345 mm undersize of its nominal diameter. The nominal bore size itself is not stated in that source, only the deviation.

So what should I actually do about a hole that needs to fit something?

Design the hole for the operation that comes after printing: print it undersize and ream or drill it to the finished diameter when the fit actually matters, and cut a test coupon before committing a full run.