Tap drill chart for 3D printing

Every tap drill chart written for metal says the same thing: for a coarse metric thread, drill the nominal diameter minus the pitch. M3 × 0.5 gets a 2.5 mm hole. That arithmetic is correct and it is also useless in a slicer, because the hole you model is not the hole you get. An FDM bore comes out undersize — the deposited bead is pushed inward by the curvature of the perimeter, and the effect does not shrink as the hole does.

So a printed thread needs two numbers: the classical tap drill, which says how much material the screw must find; and the process allowance, which says how much larger to draw it. This page tabulates both, for self-tapping screws and for clearance holes, M2 through M12. Pitches are ISO 261 coarse series and clearance holes are ISO 273 medium — both external standards. Every modelled figure comes from the same solver as the calculator.

Self-tapping screws into printed bosses

A self-tapping screw forms its own thread by displacing plastic. Leave too little material and the screw spins without biting; leave too much and hoop stress splits the boss on the first turn. The classical d − P target is a good one for plastic, and the job here is simply to hit it after the process has had its say.

Tap drill is the classical d − P figure: the hole you want to exist in the finished part. Model at is what to draw in CAD. Allowance is the difference between the two, and its behaviour is the whole story of this page: it barely moves. From M2 to M12 the target bore grows 6.4× while the correction stays inside 0.33–0.40 mm, peaking around M4 and then easing off.

Why small threads are a different problem

Compare the process undersize to the hole it is applied to and the difficulty inverts. The column below is the raw FDM undersize term — bead displacement and curvature only, before shrinkage is inverted — which is what isolates the geometric penalty.

An M2 hole is corrected by a quarter of its own diameter. A 50 mm bore is corrected by half a percent. Skip the correction on the large bore and you have a slightly loose fit; skip it on the M2 and the hole is very nearly not there. This is why small fasteners in printed parts fail so much more often than large ones, and it is not a calibration problem — it is geometry.

Clearance holes for machine screws

A clearance hole is the easy case: the screw passes through, nothing threads, and the only requirement is that the hole not end up smaller than the fastener. The same allowance applies, so a nominal ISO 273 medium hole still has to be drawn larger.

Clearance holes are forgiving in one direction only. Going oversize by a tenth costs you nothing but a slightly sloppier joint; going undersize means reaching for a drill. When in doubt on a through-hole, round up.

Nozzle and layer height

The allowance is driven by extrusion width, which is driven by nozzle diameter. Layer height contributes a smaller vertical-stepping term. Everything above assumes 0.4 mm and 0.2 mm; here is the same M3 self-tapping hole across a realistic range.

Halving the nozzle from 0.4 to 0.2 mm halves the allowance. Tripling the layer height at fixed nozzle moves it by 0.03 mm. If small threaded features matter, change the nozzle; layer height is close to noise here.

Material

Shrinkage is proportional to diameter, so on a 2.5 mm hole it is nearly irrelevant: across the full filament range the modelled figure moves 0.04 mm, well inside the ±0.08 mm repeatability band of a typical machine.

The lesson generalises: for fastener holes, pick the number by nozzle, not by filament. Material choice starts to dominate around 20 mm and is decisive by 40 — see the bearing fit chart, where the same swap moves the answer by 0.58 mm.

Frequently asked

Do I need to model the allowance if I drill the hole afterwards?

No. If the hole will be drilled or reamed, model it undersize on purpose and let the drill define the final diameter — that is the most accurate hole an FDM part can have. Model roughly 0.3 mm below target so there is material for the drill to cut without it wandering into a void.

Why are the M2 and M4 corrections nearly identical when the holes differ by 1.7 mm?

The curvature factor saturates. Below about 3 mm the inward bead displacement stops tracking diameter in any useful way, so the model holds the multiplier at its cap rather than pretending to resolve differences it cannot. Both holes end up corrected by roughly 0.40 mm; what differs is what that correction means relative to the hole, which is the point of the second table.

Can I use fine-pitch threads instead?

The pitches here are the ISO 261 coarse series, which is what almost all printed-part hardware uses. Fine pitch leaves more material (d − P is larger) but the shallower thread strips more easily in plastic. For self-tapping into FDM, coarse is the right default; fine pitch belongs in metal.

These holes come out oval, not undersize. What is wrong?

Ovality is a motion or flow problem, not a compensation one, and no allowance will fix it. Usual suspects are belt tension, excessive print speed on small perimeters, and pressure advance being untuned. Fix the machine first — a hole that is oval by 0.2 mm makes a 0.02 mm argument about allowance meaningless.

What about threading a hole with a real metal tap?

It works, and the numbers on this page are the right starting point — use the self-tapping column, since a cut thread wants the same d − P material. Go slow, back the tap out often to clear chips, and expect the plastic to be more forgiving than metal about a slightly generous hole and much less forgiving about a tight one.