OrcaSlicer Speed Settings: What Each Limit Actually Controls
There is no single "OrcaSlicer speed." The number you type into a speed field is a request, and several other limits can quietly overrule it. Your print runs at the lowest ceiling that applies to that feature at that moment.
The main ceilings are: the maximum volumetric speed your hotend and filament can actually melt, the acceleration and jerk or junction-deviation limits the firmware will allow, the minimum layer time and cooling rules that slow a small layer down so it can set, and the per-feature speed split — outer wall, inner wall, infill, bridge, support, travel and first layer are separate values, not one shared number.
So the practical rule is: change one ceiling at a time, re-slice, and compare the estimate your slicer reports and the quality of a small test print. Raising a speed field that was never the binding limit changes nothing except the risk.
Note: Reading this page will not give you one magic number, because no honest one exists — flow capability, hotend design, firmware limits, part geometry and material all move the ceiling. What it does give you is the order to check them in, and the point at which to stop and check your printer's manual instead.
How to slow down print speed in OrcaSlicer
Slowing a print down is a profile edit, not a printer edit. Work on a copy so the profile you already trust stays intact, change the field that owns the feature you are unhappy with, and confirm the result in the preview before you commit filament to it.
- Duplicate the profile you are printing with. Duplicate the current user preset and rename the copy so you can tell it apart — for example "PLA - slower walls". Edit the copy, never your only copy. Where user presets live and how they are managed is covered in the official User Profiles documentation.
- Select that profile, then open the Process speed settings for the feature you mean. Per-feature speeds live in the Process (print) settings, in the Speed group, where the fields are organised by what they control: initial layer, other layers, overhang, travel and acceleration.
- Lower only the speeds you actually want lower. The usual targets are outer wall speed for surface finish, inner wall or sparse infill speed for strength and time, and initial layer speed for adhesion. Leave the rest alone — a broad cut across every field makes the result impossible to read. (If the first layer is what you are fighting, speed and Z offset are tuned together; see the Z offset calibration guide.)
- Re-slice. Slice the model again so the toolpaths are regenerated from the new values. A profile change that has not been re-sliced does not exist in the g-code yet.
- Inspect the preview before you print. Use the preview's speed legend — and the volumetric flow view where your build offers it — to check what the slicer actually planned along the path. This is where you see whether your new value was applied at all, and whether the moves are long enough to reach it.
- Compare the estimated time. Note the new estimate against the previous one for the same model and the same settings. If it did not move, stop turning that dial and go to why the estimate stays put below.
- Test small. Print a short verification piece that contains the feature you changed — a small cube with walls, or just the top of the model — before committing the whole part.
Field labels move between builds. OrcaSlicer is actively developed, so a control can be renamed, regrouped, or moved between the Process and Filament tabs from one release to the next. Every step above is described by what the control does, not by a version number. Read the label in your own build before you change it.
Extrusion speed, travel speed and acceleration are three different things
Only the first of those families changes how fast plastic comes out of the nozzle; the other two still change how the print turns out.
| Control family | What it moves | What changing it does |
|---|---|---|
| Extrusion speed — print speed and the per-feature speeds | The head while filament is being extruded: outer wall, inner wall, sparse infill, initial layer, bridge, support, gap fill | Changes how much plastic is laid down per second. This is the family that trades time against flow, cooling and surface finish. |
| Travel speed | The head while it is not extruding | Travel speed controls non-extruding moves. It is not the first control for changing extrusion speed, but it can affect travel-related artifacts; keep it within machine limits. |
| Acceleration, jerk and junction deviation | How quickly the head changes speed (acceleration), and how much speed is carried through a corner (jerk or junction deviation) | Decides how much of a short move is actually spent at the requested speed. Lower acceleration at an unchanged speed means the machine ramps for more of every move, so the part gets slower while the number in the speed field stays the same. Cornering is not an acceleration question: jerk or junction deviation governs how the head behaves through a turn, and what that means in practice depends on the firmware. |
One approximate flow calculation, so the numbers mean something
Extrusion speed and volumetric flow are not the same quantity. A useful first-order relation is:
volumetric flow ≈ line width × layer height × extrusion speed
Take a 0.45 mm line width and 0.20 mm layers: the cross-section is about 0.09 mm². A filament limited to 12 mm³/s therefore corresponds to roughly 133 mm/s of extrusion speed (12 ÷ 0.09 ≈ 133), before acceleration, cooling or any other limit is applied.
That is arithmetic, not a recommendation. The real extrusion cross-section is not a perfect rectangle — the slicer's own flow model is more careful than the relation above — and the figure ignores acceleration, cooling, pressure advance and the machine's own limits. 133 mm/s is not a good speed, a safe speed or a calibration result for any printer. It is only what the relation produces, and it is here so you can see why a flow cap and a linear speed cap are not interchangeable.
Which limit is actually capping you
Find your symptom in the left column, then take the first move on the right. Change one thing, re-slice, and look again.
| What you are seeing | What is usually binding | First move |
|---|---|---|
| Walls look thin, gappy, or matte at the higher speed, and the extruder clicks or skips | The hotend cannot melt that much filament per second | Lower the speed back and find the real ceiling with the official volumetric speed calibration |
| Corners are rounded and the surface ripples or rings after a direction change | Jerk or junction deviation (cornering), or acceleration, is the limit being hit — not top speed | Reduce the speed near corners, or bring acceleration and junction behaviour in line with the requested speed, then re-test |
| Small, short layers come out soft, glossy, or deformed | Minimum layer time and cooling rules are slowing that layer down — sometimes they are the whole story | Read the cooling and layer-time settings before touching speed |
| One region is slow or messy while the rest is fine | That feature has its own speed value (outer wall, bridge, support, first layer, travel) | Change that feature's own field, not the global speed |
| The print is fast but the tall thin mast wobbles | The machine is moving faster than the frame and part can stay rigid at | Slow down for that geometry or add support for the part itself; keep within the machine's documented limits |
The limits, control by control
Every row is a ceiling, and the second column names the settings group that owns it, because that is the part people get wrong. Your print speed is the lowest ceiling that applies.
| Limit | Settings group that owns it | What it caps | Conservative first move |
|---|---|---|---|
| Maximum volumetric speed | Filament settings — a separate limit from every speed field | The volume of plastic the hotend can melt per second, whatever the linear speed says. Documented in the official maximum volumetric speed page | Find your real ceiling with the built-in volumetric speed calibration and set the value from that test rather than from a guess |
| Print speed and the per-feature speeds | Process settings — the Speed group | The requested linear speed for that feature only: outer wall, inner wall, sparse infill, gap fill, bridge and support each have their own value | Change the feature that is actually wrong, not the global value. If supports are the feature you are fighting, that is a different problem — see supports too hard to remove |
| Initial (first) layer speed | Process settings — the Speed group, initial layer rows | The speed of the layer that decides whether the part sticks at all | Treat it as a pair with the Z offset rather than as a speed question; the two are set together in the Z offset calibration guide |
| Travel speed | Process settings — the Speed group, travel rows (also see the official travel speed page) | How fast the head moves without extruding | Travel speed controls non-extruding moves. It is not the first control for changing extrusion speed, but it can affect travel-related artifacts; keep it within machine limits, and treat the official page as the authority on what it moves |
| Acceleration | Process settings — the Speed group, acceleration rows; the official acceleration page is the reference | How fast the head changes speed. Acceleration is about the rate of change of speed, not about cornering — corners are governed by jerk or junction deviation | Set acceleration to a value your machine handles without ringing, and keep it within the manufacturer's documented limits; adjust the requested speed near corners alongside it |
| Jerk / junction deviation | Printer settings or firmware — see the official jerk page | The instantaneous direction-change allowance at corners | Treat it as a firmware-level limit and change it against your firmware's documentation, not by feel |
| Cooling, minimum layer time and slow-down rules | Cooling settings — a separate limit from every speed field, documented on the official material cooling page | How long a layer must take before the next one starts, so the plastic can set. The slicer lowers the speed itself to reach that time | Adjust cooling and layer-time rules first; speed is not the culprit here. If the part also includes a colour change at a set height, that pause is part of the same layer-time picture — see colour change at a layer |
| Overhang and bridge slow-down | Process settings — the Speed group, overhang rows (official overhang speed page) | The speed on unsupported and bridging moves | Leave the built-in slow-down in place and fix cooling or orientation instead |
| Firmware and machine limits | Printer settings and the printer's own firmware — official motion ability page | The maximum the machine's own firmware will execute, regardless of the slicer value | Read the printer's configured limits in its firmware or its manufacturer's documentation before assuming the slicer is wrong |
Order matters more than values. Flow ceiling first, then acceleration and cornering, then cooling and layer time, then the per-feature speeds. Note also that if Flow Ratio or Pressure Advance is wrong, faster printing mostly makes existing defects more visible — fix those before you draw conclusions from a speed change. The official calibration guide lists the tests and their intended order.
Why changing a speed does not change the estimate
A speed field only changes the print time when that field's moves were setting the pace, and when the machine could already reach the value you set. These are the ordinary reasons the estimate stays put — none of them is a mystery, and none of them is fixed by turning the same dial harder.
- The field you changed does not apply to this print. Support, bridge, gap-fill and brim speeds are only used if the model actually contains those features. A speed that is never read cannot change the time.
- The moves are too short to reach the speed. On small parts, most moves end before the head gets anywhere near the requested value because acceleration is what is really setting the pace.
- The filament's maximum volumetric speed caps the flow first. Lowering a requested speed changes little only while the new request stays above the effective cap — or while that feature contributes little to the total time. If the new request is lower than the cap, the planned speed can fall and the time can increase, because the flow is no longer what limits that move.
- Cooling and minimum layer time are holding the layer back. On short layers the slicer slows the print down on purpose so the plastic can set. That cap is independent of the speed field you edited.
- The machine's own firmware or motion limits cap it. The printer's configured maximum speed or acceleration can be lower than the slicer's request, in which case the slicer number is never executed.
| Symptom | Setting group to inspect | Safe check |
|---|---|---|
| Estimate unchanged after you lowered a wall speed | Process — Speed group, per-feature rows (outer wall, inner wall, sparse infill, initial layer) | Re-slice, then confirm in the preview that the toolpath for that feature exists in this model at all |
| Estimate unchanged even though the part clearly has that feature | Filament — maximum volumetric speed | Compare the flow the preview plans for that move against the filament's flow ceiling. While your request stays above that ceiling, lowering it changes little; once the request is below the ceiling, the planned speed can fall and the estimate can rise |
| A small or short part barely changes at all | Cooling — minimum layer time and slow-down rules; also acceleration | Read the layer time the preview reports for the slowest layer, and check how far each move actually accelerates. Leave the slow-down protections enabled while you look |
| The preview shows a lower speed than the one you set | Preview legend first (speed and flow views), then the printer profile's motion limits | Establish which layer of the stack is doing the capping — slicer-side limit or machine-side limit — before changing either one |
| Nothing changes anywhere, ever | The profile that is actually selected for the plate | Confirm the duplicate you edited is the profile the plate is sliced with, and that the change was saved into it |
Do not buy time by removing protections. Disabling the cooling and minimum-layer-time slow-down can reduce the estimated time, because the slicer stops holding those layers back — but it does that by printing each layer before it has set, so it can compromise cooling and print quality. Raising the machine's own maximum speed or acceleration limits in firmware trades the same way. Do not disable cooling protections or raise firmware limits just to chase a time estimate. If the estimate will not move within the limits your hardware documents, the honest conclusion is that this geometry, material or hotend has a slower floor, not that a setting is hiding somewhere.
Common mistakes
Treating one speed field as "the" speed
Outer wall, inner wall, infill, support, bridge, gap fill, first layer and travel are separate values. Editing one and expecting the whole print to change is the most common dead end.
Setting the filament's volumetric speed above what the hotend can melt
It hides the real limit inside the slicer and shows up later as under-extrusion, weak walls, or a clicking extruder.
Chasing corners with the speed field
Rounded corners and ringing come from how the machine changes direction and speed — jerk or junction deviation for the corner, acceleration for the ramps — not from the top speed field. Raising top speed usually makes them worse, not better.
Changing several fields before re-slicing once
Two simultaneous changes cannot tell you which one worked, and the preview will not show you which value was applied where.
Edge cases and when to stop
Small parts and short layers
Minimum layer time and cooling rules can be the binding limit for the entire print, so raising or lowering speed may change nothing. Work on cooling, orientation and layer height instead.
Tall, thin geometry
A tall thin section can flex even within machine limits. Slow that geometry down or add support for the part — the speed field is not the real constraint.
Flexible and brittle materials
TPU and filled materials have their own feed behaviour and their own flow ceiling. Start from the manufacturer's recommendations and change one value at a time.
Bowden versus direct drive
A long Bowden path reacts differently to speed, pressure and retraction. Do not carry values between the two without re-testing.
Worn hardware
Loose belts, a worn nozzle or a partial clog can look exactly like a speed problem. Fix the mechanical cause first; no profile change compensates for it.
When to stop
Stop when the binding limit is the machine's own documented maximum, or when a change produces skipped steps, grinding, or an extruder that is audibly past its flow capability. Both are answers about hardware, not invitations to raise a limit.
FAQ
Duplicate the profile you are printing with, open the Process speed settings for the feature you mean, and lower only those per-feature values — typically outer wall, sparse infill and initial layer. Re-slice, then check in the preview that the new values are actually applied and that the estimated time changed. If the estimate did not change, that field was not what was pacing the print.
No. It is a request. The machine runs at the lowest ceiling that applies to that move, so volumetric flow, acceleration, cornering behaviour, cooling and layer-time rules, the printer's own firmware limits, and the feature's own speed field can all pull the real speed below the number you typed.
There is no fixed answer, because it depends on the hotend, the filament, the geometry and the machine — and inventing a "usual suspect" is how people end up tuning the wrong field. What you can do is find out which one is capping your print: slice, read the speed and flow the preview actually planned, and compare that against the field you changed. The section above on unchanged estimates walks through the possibilities in order.
It is the volume of filament the hotend can melt per second, and OrcaSlicer stores it as a filament setting rather than a speed field — see the official maximum volumetric speed documentation. The calibration suite includes a test for it; run that with your actual nozzle and material to find your real ceiling.
Because the layer has to cool before the next one goes down. Minimum layer time and the cooling slow-down rules deliberately hold those layers back — that is a quality feature, not a bug, and it is a separate limit from anything in the speed group.
In time, yes — the moves where no plastic is laid down take longer. Travel speed controls non-extruding moves. It is not the first control for changing extrusion speed, but it can affect travel-related artifacts, so keep it within machine limits.
Where to go next
Speed is one ceiling among several. If the print is coming out wrong rather than slow, the problem is usually somewhere else in the profile.