Walls that bulge outward. Top surfaces that come out raised, rough and ridged. A nozzle that visibly ploughs through the previous layer and leaves a burnished trail behind it. Parts that measure larger than the model in every direction.
Over-extrusion is the tidiest fault on this site, because it has almost no ambiguity: more material is entering the part than the geometry has room for, and the excess is pushed up and sideways.
Where the excess comes from
The slicer works out a volume for each path from the line width and layer height it was told to use. The firmware converts that volume into a length of filament, using the diameter it was told the filament is and the steps-per-millimetre it was told the extruder has. Any error in that chain scales every extrusion on the print by the same proportion.
That gives exactly three numbers to be wrong, plus one physical part that changes on its own:
- Extruder steps per millimetre, which is a property of the drive gear and the gearbox and is often left at a default that does not match the parts fitted.
- Filament diameter, which the profile assumes and reality varies.
- The flow multiplier, which is a correction factor people reach for first and should correct last.
- The nozzle, which wears. A brass 0.4 mm nozzle that has printed a spool of carbon-fibre filament can measure meaningfully larger, and it lays a wider bead than anything in the profile expects.
Measure it before adjusting it
Guessing at a flow multiplier is how machines end up correct on one test cube and wrong on everything else. The measurement is straightforward.
- Print a hollow cube with a single perimeter, no top or bottom layers and no infill — most slicers have a calibration object for this, or set walls to one and infill to zero on any cube.
- Let it cool completely.
- Measure the wall thickness with calipers at six places, spread across all four faces and at different heights. Do not measure at the seam.
- Average the six readings.
- Compare that average against the extrusion width the slicer used, not against the nozzle diameter — they are usually different numbers.
New flow multiplier = old multiplier × (intended width ÷ measured width). If the walls came out 0.5 mm where the slicer intended 0.45 mm, the extruder is delivering about eleven percent too much.
Correct the three numbers in this order
Steps per millimetre first. Mark the filament, command a known length — 100 mm is the convention — and measure what actually went in. Correcting this fixes the machine rather than one profile, and every material benefits.
Diameter second. Measure the spool with calipers in three places at least twenty centimetres apart, in two directions at each place, and enter the average. Nominal 1.75 mm filament from a good manufacturer measures very close to that; cheap filament does not, and no other setting compensates for stock that varies along its own length.
The multiplier last, and small. Once the first two are right, whatever residual remains is a genuine per-material correction — some filled and glossy grades genuinely need a slightly different figure.
Reversed, the multiplier silently absorbs an error that belongs to the steps value, and the machine is then wrong the moment anything else changes.
Suspecting a worn orifice
Suspect wear if the machine was accurate and gradually stopped being, if it has printed anything filled, and especially if the change accompanied a drop in surface quality on fine features. A worn orifice is not round, so it also lays an inconsistent bead. Nozzles are consumables; measuring one accurately is difficult, and replacing one is cheap.
What it breaks elsewhere
Correcting flow is worth doing early because so many other complaints are downstream of it: a rough top surface, parts measuring oversized, infill printing through the walls, supports that fuse to the part because the excess closes the gap, and a nozzle that starts striking the print because the surface is higher than the machine believes.
Chasing any of those individually while flow is wrong produces settings that each cancel one symptom and collectively make no sense.
Do not fix it with size compensation
X/Y size compensation exists to remove a small residual error after calibration. Used to cancel a flow error, it shrinks the outline while the bead stays too wide, so the walls thicken inward, holes close further, and the internal geometry of the part quietly goes wrong while the external dimensions read correctly on the calipers. Measure the wall, fix the flow, and leave compensation for what is left over.