A range is the honest answer
Every reputable filament ships with a temperature range rather than a single figure, and that is not vagueness. The correct nozzle temperature depends on your hot end, how the thermistor is positioned, print speed, layer height, part geometry and cooling. Two printers running the same spool can legitimately land ten or fifteen degrees apart.
Use the ranges below to know roughly where to start. Use the filament's own data sheet to narrow it. Use a test print to settle it.
Typical published ranges
| Family | Nozzle | Bed | Enclosure | Part cooling |
|---|---|---|---|---|
| PLA | 190 to 220 °C | 20 to 60 °C | Not required | High |
| PETG | 220 to 250 °C | 70 to 90 °C | Helpful | Low to moderate |
| ASA | 240 to 260 °C | 90 to 110 °C | Recommended | Minimal |
| TPU | 210 to 230 °C | 30 to 60 °C | Not required | Low to moderate |
These are the ranges commonly published across the material families, not specifications for any particular product. A high-speed, matte, silk or composite grade can sit well outside them. Where the data sheet and this table disagree, the data sheet is right.
Bed temperature and first-layer adhesion
Most first-layer failures are adhesion problems rather than temperature problems. Before raising the bed temperature, check that the plate is clean, that the nozzle height is correct and that the surface suits the material.
Bed temperature interacts with the build surface. A textured plate, a smooth plate and an adhesive-coated plate can each want different figures for the same filament, and the plate manufacturer's guidance takes priority over the filament's. This matters most with PETG, where over-adhesion can make a part difficult to remove without damage.
Cooling is the setting people get backwards
PLA generally wants as much part cooling as the printer can provide. Overhangs and small features improve markedly with it.
The higher-temperature families want the opposite. ASA relies on staying warm to bond each layer to the last, so aggressive cooling is a common cause of layer splitting and warping. PETG sits in between: some cooling improves surface finish and bridging, too much weakens the part.
Speed, flow and the limits of a headline number
A printer advertising a high maximum speed does not mean every filament reaches it. The real ceiling is how fast the hot end can melt material consistently, which varies by filament and by hot end. Pushing past that produces under-extrusion, poor layer bonding and a part that looks finished but is not strong.
Flexible materials are the clearest case. TPU typically prints far slower than rigid filament regardless of what the printer can do, because the feed path rather than the hot end sets the limit.
Abrasive filament and nozzle wear
Carbon fibre, glass fibre and some wood-filled or metal-filled filaments wear a standard brass nozzle measurably, sometimes within a single large print. A hardened steel or ruby nozzle is the usual answer.
This is a compatibility question rather than a temperature one, but it belongs in the same pre-print check: confirm the nozzle material as well as the diameter before starting.
How to actually find your number
Print a temperature tower. It is a single test object that steps the nozzle temperature across its height, so one print shows you where layer bonding, stringing and surface finish are best on your machine with that specific spool.
Change one variable at a time and write down what worked. Keeping a short record of filament, colour, printer configuration and slicer profile turns a good result into a repeatable one, which matters more than any individual setting.