Two nozzles that look identical from across the room behave completely differently once abrasive filament runs through them. Choosing between hardened steel and brass 3D printer nozzles comes down to one trade-off: brass moves heat faster, hardened steel survives what brass cannot.
Quick Answer
Brass is the better everyday nozzle for PLA and PETG because it transfers heat into the plastic more efficiently, which helps at high flow rates. Hardened steel gives that up in exchange for wear resistance, which matters the moment carbon fibre, glass fibre or glow-in-the-dark filament enters the picture. Those abrasive blends grind a brass tip out of round within a kilogram or two of filament, which is exactly why the R13,999 Creality K1 Max ships with a 300C hardened steel nozzle instead of brass.
🔩 Why brass wears out on abrasive filament
Brass is a soft metal by design, which is what makes it easy to machine into a precise 0.4mm orifice and also what makes it defenceless against anything gritty passing through that orifice at speed. Chopped carbon fibre and glass fibre are essentially fine abrasive particles suspended in plastic, and every pass through a brass nozzle sands the opening a little wider.
The failure is gradual rather than sudden, which makes it worse. Extrusion width creeps up slowly, dimensional accuracy drifts, and by the time the change is obvious the nozzle has already ruined several prints without anyone noticing why.
🛡️ What hardened steel trades away, and what it gets back
Hardened steel conducts heat noticeably slower than brass, which used to mean choosing between a nozzle that survives abrasives and one that flows plastic quickly. That trade-off is why plain hardened steel nozzles were historically reserved for slow, careful prints only.
What actually survives abrasive filament for 30 to 50kg rather than the 1 to 2kg a plain brass tip manages is a hardened wear surface paired with a core that still moves heat efficiently, which is the reasoning behind multi-metal nozzle designs generally. Whatever the construction, the number that matters is how many kilograms of carbon-fibre filament it survives before the orifice measurably widens, not the metal's name on the box.
🧪 Matching the nozzle to the filament, not the other way round
PLA, PETG and TPU are gentle on any nozzle and print fine on brass. The moment a spool of carbon-fibre nylon, glass-fibre PETG or a metallic-look filament shows up, check the nozzle first before printing a single layer, because running abrasives through brass is the fastest way to need a new one within days.
Keep a spare on hand regardless of which material you print most, since a nozzle is a wear part on any printer and running out mid-project stalls a job for longer than the part itself costs. Browse filament options before committing to an abrasive blend, and check whichever printer you are considering for what its stock nozzle is actually rated for. Resin machines such as Elegoo's range sidestep the nozzle question entirely, since they cure liquid rather than extrude filament.
🔧 Swapping a nozzle without ruining the next print
Changing a nozzle is a 5 minute job done carelessly and a 5 minute job done properly, and the difference shows up on the very next print rather than immediately. Always swap a nozzle hot, with the hotend at or near 220 to 250C, since a cold nozzle threads unevenly and can leave a gap that leaks plastic around the heat break for weeks afterwards.
Tighten the new nozzle against the hotend while it is hot, then let the whole assembly cool slightly before printing, which lets the metal contract onto a snug seal rather than one that was tightened once and never checked again. A nozzle that leaks slightly at the seat produces a fine, hard-to-diagnose stringing and inconsistent extrusion that looks identical to a filament problem, wasting far more troubleshooting time than the swap itself took.
🧪 Reading nozzle wear before it shows up on a print
Extrusion width is the earliest, most reliable signal of nozzle wear, and it is worth checking before a project rather than after it fails. Print a short single-wall test at a known 0.4mm line width, measure it with callipers, and compare the result against the nozzle's rated size; a wall reading 0.45mm or wider is the first sign the orifice has opened up from abrasive wear.
Keep a simple log of which nozzle, in which printer, has run which materials and for roughly how long. That record turns a vague sense that "prints have been coming out a bit off lately" into an actual answer about which specific nozzle is due for replacement, rather than guessing across an entire workshop's worth of machines and spools, and it also tells you, months later, which supplier's abrasive filament chewed through a nozzle fastest.
Frequently Asked Questions
How long does a brass nozzle last on carbon-fibre filament?
Often only a kilogram or two before the orifice widens enough to affect dimensional accuracy, compared to tens of kilograms on a hardened surface.
Is brass ever the wrong choice for standard filaments?
No, for PLA, PETG and TPU it remains a solid, inexpensive choice with good heat transfer.
Can a worn nozzle be fixed rather than replaced?
No, a widened orifice is permanent wear; nozzles are a consumable and get swapped, not repaired.
Does nozzle material affect print speed?
Indirectly, through heat transfer: a nozzle that moves heat into plastic faster supports higher flow rates before the plastic starts under-melting.
How do I know a nozzle swap actually sealed properly?
Watch the first layer of the very next print closely; a fine haze of plastic around the base of the nozzle, or an unexpected smell of burning filament near the heat break, both point to a seal that needs redoing while the hotend is still hot rather than after it cools.
Is it worth buying nozzles in bulk rather than one at a time?
For a workshop running abrasive filament regularly, yes, since ordering several at once avoids the delivery wait becoming the bottleneck on a project the moment a single nozzle finally wears through, and it is cheaper per unit than repeat single orders in most cases anyway.
Printing something that will actually test a nozzle?
Match the nozzle material to the filament before the first layer, not after the print fails.