Print speed vs quality: max volumetric flow is the real limit

You set 300 mm/s in the profile, then watch the printer run at about 130 mm/s. Nothing is broken. The slicer slowed it on purpose, because that is all the hotend can melt. This lesson explains what really sets print speed, how to work it out with a one-line formula, and what goes wrong first when you push past it.

Figures assume a 0.4 mm nozzle and 0.45 mm line width, the usual defaults.

Three speeds that are not the same thing

• Speed (mm/s): how fast the head moves while extruding, set per line type (outer wall, inner wall, infill, first layer).
• Acceleration (mm/s²): how quickly it ramps from standstill to that speed, which decides whether it ever reaches it.
• Volumetric flow (mm³/s): how much plastic leaves the nozzle per second, the ceiling the hotend imposes.

Most people only touch the first. The other two decide whether the print is actually fast and actually good.

The one formula to remember

One extruded line has a cross-section of roughly layer height × line width. Move it at some speed and plastic must flow out at cross-section times speed.

flow (mm³/s) = layer height (mm) × line width (mm) × speed (mm/s)

With 0.2 mm layers and 0.45 mm lines:

Set speed	Calculation	Flow
60 mm/s	0.2 × 0.45 × 60	5.4 mm³/s
100 mm/s	0.2 × 0.45 × 100	9.0 mm³/s
200 mm/s	0.2 × 0.45 × 200	18.0 mm³/s
300 mm/s	0.2 × 0.45 × 300	27.0 mm³/s

Now compare that with what hotends can do. Prusa’s knowledge base gives these ranges:

Hotend	Achievable flow (PLA)
Standard all-metal (E3D V6 style, older bed-slingers)	8-12 mm³/s
Newer hotends on fast machines (e.g. Prusa Nextruder)	15-20 mm³/s
High-flow nozzles (CHT type)	20-35 mm³/s
High-flow hotends (e.g. Revo HF, Rapido HF)	25-40 mm³/s

By material (Prusa’s ranges): PLA 12-20, PETG 8-15, ABS/ASA 10-16, Nylon 8-12, PC 6-10, TPU 2-5 mm³/s, lower on a standard hotend.

Worked example: the real speed of a standard-hotend printer

Say our machine has a standard hotend that tested at 12 mm³/s. Wall speed is set to 200 mm/s, 0.2 mm layers, 0.45 mm lines.

1. Flow requested = 0.2 × 0.45 × 200 = 18 mm³/s, above the 12 ceiling.
2. The slicer therefore caps the speed at 12 ÷ (0.2 × 0.45) = 12 ÷ 0.09 = 133 mm/s.
3. Switch to 0.28 mm layers to print faster, and the top speed drops to 12 ÷ (0.28 × 0.45) = 95 mm/s. Yet the work done per second is the same, because the same amount of plastic comes out.

Layer height	12 mm³/s ceiling	20 mm³/s ceiling
0.12 mm	222 mm/s	370 mm/s
0.20 mm	133 mm/s	222 mm/s
0.28 mm	95 mm/s	159 mm/s

The lesson: print time is set by mm³/s, not mm/s. Thin layers moving fast or thick layers moving slowly take about the same time if they push the same plastic. To print faster you raise the ceiling (hotend, bigger nozzle, higher temperature), not the speed number.

Prusa puts it bluntly: max volumetric speed never makes a print faster, it can only slow it down. Setting it to 0 removes the ceiling, and then the extruder simply cannot keep up.

Check it after slicing: in the preview choose Volumetric flow rate (PrusaSlicer) or Flow (OrcaSlicer, Bambu Studio). If the whole part is one colour at the ceiling value, the ceiling is limiting every line.

What happens when flow exceeds the ceiling

Plastic is pushed through the nozzle faster than the heater can melt it. Symptoms arrive in this order:

1. The surface goes matte instead of glossy. The OrcaSlicer docs name this as the first sign.
2. Under-extrusion: thin lines, gaps, holes in the top surface, broken infill.
3. Extruder clicking: the drive gear cannot push the filament and skips back.
4. Layers stop bonding: the plastic leaves cooler than set and the part snaps along layer lines (see the strong prints lesson).
5. Clogs: the gear grinds the filament to dust, or heat creeps up and it swells in the heatbreak.

The trap: on some fast printers the motors keep up while the hotend does not, and it looks like a filament problem when it is a flow problem.

Find your own ceiling

OrcaSlicer has a flow test under Calibration (the full routine for a new spool is in lesson 7).

1. Set start 5 mm³/s, end 20 mm³/s, step 0.5 (the defaults). It prints a tower whose flow rises with height.
2. Find the height where the surface starts to fail; measure it with calipers, say 19 mm.
3. Calculate: 5 + (19 × 0.5) = 14.5 mm³/s.
4. The docs recommend taking 10-20% off, because the test is a best case with no retractions or sharp corners: roughly 12-13 mm³/s. Save it in the filament settings as max volumetric speed.

The value depends on brand and even colour. To raise the ceiling, move the temperature toward the top of the maker’s range and test again.

Acceleration: small parts never reach the set speed

The head accelerates from a stop at every change of direction. The distance to reach speed v at acceleration a is v² ÷ 2a.

• At 200 mm/s with 2,000 mm/s²: 200² ÷ 4,000 = 10 mm to reach speed, and another 10 mm to brake. A 20 mm wall touches 200 mm/s for an instant and slows again.
• At 5,000 mm/s² the ramp shrinks to 4 mm; the same wall runs at 200 mm/s for 12 of its 20 mm.

This is why small, detailed parts do not print faster whatever speed you type, and why fast printers chase acceleration more than mm/s.

The price is vibration. Hard acceleration and braking leave the frame and head swinging for a fraction of a second after each corner, seen as ringing or ghosting: fine ripples on the wall just after a corner or raised lettering.

Input shaping and pressure advance

Input shaper is a firmware filter that analyses the planned motion and cancels the machine’s own resonance (Klipper, Prusa firmware 5.x and later, Bambu and newer Creality machines ship with it). The result is hard acceleration with most ripples gone. Prusa stresses two conditions:

• Before enabling it, the X and Y belts must be tight, the nozzle in good shape and the rods free of debris, or the measured values are wrong.
• You need the matching input-shaper slicer profile, and if the head’s weight changes (new hotend, a camera) it must be re-measured. Machines with a built-in accelerometer measure themselves; others need an external one.

Pressure advance (linear advance in Marlin) fixes a different problem. When the head brakes at a corner, pressure left in the nozzle keeps pushing plastic out, leaving a blob, and the line thins as it accelerates away. Pressure advance adjusts extrusion ahead of speed changes. The faster you print, the more it matters; OrcaSlicer has a test for it too.

What else suffers at speed, besides flow

Within the flow ceiling but still fast	Symptom	Fix
Overhangs and bridges	Fan cannot cool in time, edges droop	The slicer already slows these; do not turn that off (lesson 12)
First layer	Lines not pressed onto the bed, part comes loose	Keep the first layer at 20-30 mm/s, always
Outer wall	Ripples, visible seam	Outer wall at about half the inner wall speed, the default in every slicer
Small parts and thin pillars	Plastic has no time to cool, tips go soft and sag	Enable minimum layer time or slow down for small features
Tall thin parts	Wobble with the motion, shifted layers	Lower acceleration, or print several parts at once so layers get time to cool

Starting point: the printer maker’s profile, with only max volumetric speed changed to match your tested spool. Those profiles already split speed by line type sensibly.

Safety when printing fast

• A higher flow ceiling usually means a hotter nozzle and more fumes, especially with ABS/ASA. Ventilate.
• High acceleration shakes screws and belts loose faster. Check tension regularly; layer shifts in long prints often start here.
• A fast printer in an enclosure heats up quickly, motors and board included. Do not block its vents.
• Do not walk away during the first layers. At speed, a detached part can wrap around the nozzle in minutes.

Common mistakes

• Setting speed from the printer’s advertised spec. That is the maximum travel speed, not the extrusion speed the hotend supports.
• Setting max volumetric speed to 0 to “unlock” the printer. You get under-extrusion everywhere instead of the slicer braking for you.
• Using the generic PLA profile for every brand. Each spool has its own ceiling; test them one at a time.
• Moving to 0.28 mm layers and keeping the old speed. Flow jumps past the ceiling at once; listen for the extruder clicking.
• Raising acceleration without re-tuning input shaper. The ripples come back.
• Blaming wet filament when the matte, gappy surface appears only on the fast sections. That is flow, not moisture.

Sources

• Prusa Knowledge Base: Max volumetric speed
• OrcaSlicer wiki: Max volumetric speed calibration
• Prusa Knowledge Base: Input Shaper
• Prusa Knowledge Base: Layer separation and splitting (FDM)
• CNC Kitchen: The influence of extrusion temperature on layer adhesion