Nozzle sizes explained: 0.2, 0.4, 0.6 and 0.8 mm
What changing nozzle size actually does to print time, strength and detail — and why 0.4 mm is the default for good reasons.
Nozzle diameter is the most consequential ten-pound decision in FDM printing, and the one most people never revisit. Almost every printer ships with 0.4 mm, and almost every slicer profile assumes it. That default is genuinely well chosen — but it is a compromise, and knowing which way to move it is useful.
What actually changes
A wider nozzle lays down a wider, thicker line. That means fewer lines to fill the same area, and it means each line is stronger. The trade is resolution: the nozzle cannot reproduce a feature narrower than the line it extrudes.
| Size | Good for | Cost |
|---|---|---|
| 0.2 mm | Miniatures, fine text, tiny mechanisms | Very slow; clogs easily; not worth it for most work |
| 0.4 mm | Everything, by default | The compromise, and a good one |
| 0.6 mm | Functional parts, faster prints, stronger walls | Loses fine detail; the best upgrade for most people |
| 0.8 mm | Large objects, vases, draft prints | Visible layer lines; poor at small features |
The one people underuse: 0.6 mm
If you mostly print brackets, enclosures, jigs and other functional things, a 0.6 mm nozzle is close to a free upgrade. Prints finish substantially faster because each perimeter covers more ground, and the parts come out stronger because thicker extrusions bond better. You give up detail you were probably not using.
A common setup is to keep 0.4 mm as the default and swap to 0.6 mm for anything large or structural. Nozzles cost very little; there is no reason to own only one.
Layer height is tied to nozzle size
The rough rule is that layer height should sit between roughly 25% and 75% of nozzle diameter. With a 0.4 mm nozzle that is about 0.1–0.3 mm. Going too thin risks poor bonding and inconsistent extrusion; too thick and layers do not adhere properly. A 0.8 mm nozzle at 0.1 mm layers is not “high detail” — it is a squashed, badly behaved print.
Material matters too
Some filaments effectively require a wider nozzle. Wood-fill, glitter and heavily filled composites contain particles that will bridge and block a 0.4 mm opening; 0.6 mm or wider is the usual advice. Carbon-fibre filaments will print through 0.4 mm but will chew a brass nozzle doing it — see hardened versus brass nozzles.
If you change it, tell the slicer
The most common nozzle-related failure is fitting a new size and forgetting to update the slicer. The printer will happily try to extrude 0.4 mm lines through a 0.8 mm hole, or force 0.8 mm of plastic through a 0.4 mm one. Change the nozzle diameter in the printer settings, re-slice, and check your first layer.
Nozzles are consumable regardless of size. Buy several at once; they are among the cheapest things in parts and spares.
Check current prices
More guides
-
Belts, tension and layer shifts
Ringing, wavy walls and prints that jump halfway through are usually mechanical. How to tension a belt by feel, and what else causes a shifted layer.
-
Build surfaces compared: PEI, glass and textured plates
Smooth PEI, textured PEI, glass and the rest — what each is actually good at, which filaments they fight with, and when a worn sheet is your real problem.
-
Calibration tools: calipers, gauges and what you actually need
One tool does most of the work and costs about a tenner. What digital calipers are for, and which other measuring kit is worth owning.
-
Finishing tools: deburring, sanding and cleanup
A handful of cheap tools turn a rough print into a finished part. What to buy first, the right sandpaper progression, and why wet sanding beats dry.
-
Hardened vs brass nozzles: when you actually need one
Abrasive filament destroys brass nozzles fast. Which filaments count, what hardened steel costs you in heat transfer, and when brass is still the right choice.
-
Print removal tools: what to use and what to never use
The wrong scraper destroys build surfaces and fingers. What actually gets a stubborn print off, starting with the trick that costs nothing.
Last reviewed 10 September 2026.