Your Nozzle Is Wearing Out — And Your Tolerances Are Paying For It

You track filament brand, nozzle temp, retraction distance, even ambient humidity. The one part actually touching every gram of plastic that leaves your hotend — the nozzle orifice itself — doesn't get tracked at all. It just gets used until something looks wrong, and by then the drift has been happening for days.

Nozzle wear isn't a failure mode. It's a wear curve, and it's silent. The orifice starts at 0.4mm and grows a little with every kilogram of abrasive material that passes through it. Holes print slightly small. Walls print slightly thick. Press-fit parts get looser. None of it looks like a nozzle problem, so it gets diagnosed as everything else first — a bad filament batch, a slicer flow ratio that needs recalibrating, humidity, a firmware update that "changed something."

Every Other Wear Part Announces Itself. This One Doesn't.

Belts stretch and start skipping teeth audibly. PTFE tubing degrades and lets filament ooze past the collet. Bearings get loud. The nozzle just quietly gets bigger, and a bigger orifice means more plastic comes out per unit of extruder rotation — which the slicer has no way of knowing, because it's still computing flow based on the diameter you told it when you set up the profile, not the diameter that's actually there now.

With unfilled PLA or PETG, this is a slow, mostly irrelevant process — plain brass can run for months without measurable change. The moment carbon fiber, glass fiber, or any filled filament enters the picture, the timeline compresses hard. Chopped carbon fiber strands are abrasive enough to put measurable wear on a stock 0.4mm brass orifice within 15 to 20 hours of print time. Left unaddressed, that same brass nozzle can go from 0.4mm to 0.6mm or wider well within a single kilogram spool, and at that point extrusion width is no longer something your slicer settings control — the hardware has already decided.

What Actually Wears, and How Fast

Nozzle material is the entire variable here, and the spread is bigger than most people assume when they buy whatever came in the box.

Checking Wear Without Guessing

You don't need a microscope to catch this before it costs you a batch of parts.

1. 1
   Pin gauge or drill bit check. A 0.40mm pin gauge (or the shank of a 0.40mm drill bit) should sit snug in a fresh nozzle and noticeably loose in a worn one. This is the single fastest confirm-or-rule-out test available.
2. 2
   Single-wall test cube. Print a simple single-wall box with a known nominal wall thickness and measure it with calipers. A wall reading meaningfully over spec on a profile that used to print to tolerance is your orifice telling you it's grown.
3. 3
   Log throughput by material, not just by nozzle. Track kilograms of CF/GF-filled filament run through each physical nozzle, separately from unfilled material. A brass nozzle that's seen 1.5 kg of PETG-CF is due for replacement even if it's only been installed a few weeks.
4. 4
   Don't chase it with flow ratio. If you're nudging flow percentage upward every few prints just to keep dimensions in spec, that's not calibration — that's compensating for a part that needs to be swapped.

• Pin-gauge or drill-bit test before assuming a slicer or filament problem
• Print and measure a single-wall test cube against known nominal thickness
• Log CF/GF kilograms per physical nozzle, not per calendar week
• Stop nudging flow ratio upward as a permanent fix
• Match nozzle material to filament type before starting a CF/GF job, not after

The PCTG Exception

The Real Fix Is a Parts Bin, Not a Recalibration

The fix here isn't a firmware tweak or a slicer profile — it's treating the nozzle like the consumable it is. Keep a stock of hardened steel nozzles on hand the moment any CF- or GF-filled material is in regular rotation, swap based on logged kilograms rather than waiting for a bad part to tell you, and reserve ruby or tungsten carbide for whichever printer actually lives on abrasive filament full time. A worn nozzle doesn't announce itself with a failed print — it announces itself as a part that's very slightly, very consistently, wrong, right up until someone measures it and finds out why.

If a part needs to hold tolerance across a full production run regardless of which material it's printed in, that's the kind of consistency DuffAM's process is built to guarantee before the part ever ships — see our realistic capabilities breakdown for where that line sits.