Walls you can see daylight through. Solid infill that is not solid. Layers that sit next to each other without touching. The part is the right shape and the wrong density.
One question narrows this to half
Is the shortfall everywhere, or only in some places?
If slow outer perimeters are gappy along with everything else, the machine is not delivering the plastic it was told to. Look at calibration, grip and blockage.
If perimeters are fine but infill is thin, or the walls degrade only where the toolhead speeds up, or a part prints well small and badly large, nothing is broken. The profile is asking for more molten plastic per second than the hotend can produce.
That second case is under-diagnosed because everything on the machine is working correctly. It is a settings error that looks like a hardware fault.
The flow ceiling, with real numbers
There is an upper limit to how much solid filament a given hotend can turn into liquid in a second. That rate is a property of its geometry and heater — how much hot metal the filament passes and how long it spends there — and it is why identical profiles behave differently on two machines.
A 0.4 mm nozzle laying 0.45 mm-wide lines at 0.2 mm layer height, commanded to run at 300 mm/s, demands 27 mm³/s of molten plastic. A stock hotend of the kind fitted to most machines a few years old tops out somewhere near 12 mm³/s. Feed it that path and the real head speed collapses to about 133 mm/s — and if the firmware does not clamp the speed for you, the extruder simply falls short and the wall goes thin.
Those figures come from the print time estimator with exactly those inputs. Put your own nozzle, layer height, line width and speed into it before deciding your extruder is worn out.
When it really is the machine
Ranked by how often each turns out to be the cause:
- A partial clog. Flow that is short by a variable amount, usually with a curl to the extrusion as it leaves the nozzle. See clogged nozzle for the clearing procedure.
- Filament diameter wrong in the slicer. Every profile assumes a diameter and calculates volume from it. Measure the spool in three places with calipers; a batch running consistently under nominal produces a permanent shortfall of a few percent.
- Extruder tension or a worn drive gear. Teeth that have polished smooth cannot grip, and the filament slips without leaving the obvious chewed flat that grinding produces.
- Extrusion steps never calibrated. Ask the machine for 100 mm of filament and measure what it actually pulls in. This is a ten-minute job that has to be done before any flow multiplier is touched.
- Nozzle temperature too low for the speed. The same demand that a hotend meets at the top of a material's range it cannot meet at the bottom.
Fixes with a test attached
- Slow down by a quarter and reprint the same object. If the walls fill in, it was the flow ceiling and nothing is wrong with the printer.
- Raise nozzle temperature by 10 °C. Same test. More heat means more melting capacity, at some cost in stringing and overhang quality.
- Fit a larger nozzle if the parts are functional rather than fine. A 0.6 mm nozzle moves substantially more plastic for the same head speed, and on the example above it is the change that actually recovers the time.
- Calibrate extrusion steps, then flow, in that order. Reversed, the multiplier absorbs an error that belongs in the steps value and the machine is wrong everywhere else.
Do not turn the flow multiplier up
It is the fastest way to make one print look better and every future print worse. A multiplier that has to sit meaningfully above 100% is not a fix, it is a note that something upstream was never measured — and it stacks a dimensional error onto every part you print afterwards, because the extra plastic has to go somewhere.
Filaments that starve a hotend fastest
Anything filled. Carbon and glass fibre raise melt viscosity and abrade the orifice, wood-filled grades carry particles that lodge in it, and metal-filled grades are both abrasive and heavy. Flexible filaments under-extrude for a different reason entirely: a soft filament buckles between the drive gear and the melt instead of pushing, which is why TPU wants a short constrained path and a slow, patient profile.