Key takeaways
- Calibration belongs to the spool, not the printer. The OrcaSlicer documentation states that even within one material type, different brands and colours change maximum flow.
- The order is part of the method. Temperature changes viscosity, so running the temperature test after the flow test means rerunning the flow test.
- Every test gives one number and one place to save it: the material profile, not the printer profile. Name the profile by filament batch and you won't have to guess next time.
- No calibration fixes wet filament, slack belts or a worn nozzle. If the retraction tower is still stringy at its top step, the OrcaSlicer documentation says to dry the filament and check the nozzle, not to keep increasing retraction.
Most printers today print well straight out of the box, and with mainstream PLA the maker's default profile is usually good enough. Calibration starts to matter elsewhere: when you switch to an unfamiliar spool, when you want to print faster and keep the surface, or when a part has to hit fit dimensions. In all three cases, what's missing isn't a secret setting but a few measurements on that specific spool.
Why the numbers belong to the spool
A printer profile describes mechanics: toolhead mass, travel, extruder type. A material profile describes what flows through the nozzle, and that's what changes between spools. OrcaSlicer's calibration documentation states that maximum volumetric speed depends on the material, printer, nozzle diameter and even the extruder — and that within the same material type, brand and colour also shift it, read 07/09/2026. See sources.
That sounds minor but has real consequences: a heavily pigmented black PLA and a white PLA from the same maker may not share the same speed ceiling. How to read a data sheet to see what the maker actually published is in how to read a filament data sheet.
The order the documentation recommends
| Order | Test | What it measures | Where the result is saved |
|---|---|---|---|
| 1 | Nozzle temperature | The temperature for the best surface, layer bonding and stringing | Material profile |
| 2 | Maximum volumetric speed | The mm³/s ceiling the nozzle can still melt | Material profile |
| 3 | Pressure advance | Compensation for pressure lag as the toolhead speeds up and slows down | Material profile |
| 4 | Flow ratio | Actual plastic delivered versus what the software calculates | Material profile |
| 5 | Retraction | The shortest retraction that still prints clean | Material profile |
| 6 | Motion: jerk, input shaping, VFA | Mechanical vibration and ripple marks | Printer profile |
| 7 | Tolerance | The real clearance between two parts that fit together | Drawing and print profile |
What each test measures, and how to read it
Temperature tower
A tall block of several tiers, each printed at a different temperature. Once printed, you look at each tier and weigh the four signs the OrcaSlicer documentation lists: lower temperatures usually string less; higher temperatures usually bond layers better; overhangs and bridges need enough heat to flow but sag when too hot; surface finish is a balance between temperature and flow. If a whole band of temperatures looks good, take the middle of the band — and if you plan to print fast, lean towards the top of it.
Maximum volumetric speed
The most important test for anyone who wants to print fast, because it's the real ceiling of the nozzle-and-spool pair. The test piece increases flow with height; you measure the height where quality starts to break down and calculate the value. With the dialog defaults — start 5 mm³/s, end 20 mm³/s, step 0.5 — a piece that fails at 19 mm height gives 5 + 19 × 0.5 = 14.5 mm³/s, exactly the documentation's example.
Pressure advance
As the toolhead slows into a corner, pressure in the melt zone doesn't drop at once, so plastic keeps coming out and the corner bulges. Pressure advance is the value that compensates for that lag. OrcaSlicer has three methods — tower, pattern and line — and the documentation notes each has separate versions for direct-drive and Bowden extruders; you must pick the right one. With the tower method, the value rises evenly with height and the result is the start value plus the step times the measured height: the documentation's example is 0 + 0.002 × 8 = 0.016.
- Printers running Marlin need Linear Advance (M900) enabled in firmware — the documentation notes not every printer has it on.
- On Bambu Lab printers, don't enable the Flow calibration option when running this test. The OrcaSlicer documentation warns about this in both the pressure advance and flow ratio tests, because the printer adjusts itself mid-print and spoils the measurement.
- Save the result in the material profile, not the printer profile — pressure lag depends on the plastic that's flowing.
Flow ratio
Flow ratio decides how much plastic actually comes out versus what the software calculates. Too little and the top surface has gaps, layers bond poorly and the part is weak; too much and the surface ripples and dimensions are off. OrcaSlicer's current recommended method is a single print with an adjustment range of −0.05 to +0.05 in 0.01 steps; the version for perfectionists uses −0.04 to +0.035 in 0.005 steps. Pick the block with the smoothest top and add or subtract that block's adjustment to the old value.
Retraction
This test prints a tower of steps, each with a different retraction length, and you pick the shortest step that's still clean — shortest, not cleanest, because longer retraction raises the risk of clogs and filament grinding. The dialog default is 0 to 2 mm in 0.1 mm steps, suited to direct drive; the documentation recommends changing to 1 to 6 mm in 0.2 mm steps for Bowden extruders.
Four numbers to have before printing a batch
- Nozzle temperature
- The middle of the good band on the temperature tower
- Maximum flow
- The measured value, minus 10 – 20 %
- Flow ratio
- The old value plus the adjustment of the smoothest block
- Retraction length
- The shortest step that's still clean
Faults no calibration can fix
This is the most skipped part, and it's why many people run the tests over and over and still get contradictory results. A measurement only means something when everything else is held still.
- Wet filament. The OrcaSlicer documentation is blunt: if there's still a lot of stringing at the top of the retraction tower, dry the filament and check the nozzle is seated tight. Dry first, measure after — the routine is in how to dry filament.
- A worn or partly clogged nozzle. A bore worn wide gives wider lines at every flow ratio. With fibre-filled filament it's a question of when, not if — see the nozzle article.
- Loose mechanics. Slack belts and wobbly wheels create ripples no pressure advance can compensate. Check before running the motion tests.
- An unstable first layer. Several tests are read on the printed surface; if the first layer is faulty, the results aren't trustworthy — see first-layer adhesion.
Calibration turns an unfamiliar spool into a named profile. If the result isn't saved, you just printed a tower of plastic to throw away.
Saving results so they're usable next time
- 01
Create your own material profile; don't edit the default
The default profile is overwritten when the software updates, and you won't know which values were yours. Duplicate it and give it a new name.
- 02
Name it by maker, line, colour and batch
For example: maker — line — colour — batch code. Colour is in the name because it changes maximum flow; the batch code because formulations change between batches.
- 03
Record the date and nozzle diameter too
The same spool on a 0.4 mm and a 0.6 mm nozzle is two different sets of numbers. If the software has no notes field, put it in the profile name.
- 04
Rerun after any of three changes
A spool from a different batch, a new nozzle, or a significant change in target print speed. Outside those three, there's no need for periodic re-measurement.
- 05
Leave calibration mode before printing real work
The OrcaSlicer documentation reminds you to start a new project after running the tests to exit calibration mode. Skip that and your next print still carries the test piece's settings.
The shortest way to see it: calibration isn't a ritual for every spool, but a tool for answering a specific question — why this part is off-size, why this surface ripples, why this speed won't go higher. Without such a question, the material maker's default profile is usually a reasonable starting point; starting windows for each plastic are in the handbook.
Frequently asked
My Bambu Lab printer self-calibrates — do I need to do anything else?
The printer runs some steps before each print, but those don't replace measurements that belong to the spool, like maximum flow or optimal temperature. One important note: when running OrcaSlicer's flow ratio or pressure advance tests on a Bambu Lab printer, the documentation advises not enabling the Flow calibration option, since the printer intervenes mid-print and spoils the measurement.
Do I need to recalibrate when changing colour within the same line?
At least recheck maximum flow. The OrcaSlicer documentation states that within one material type, brand and colour noticeably affect maximum flow, because pigments and additives change how the plastic flows. If you print well below the ceiling, you usually won't see a difference.
After calibrating, will parts be dimensionally correct?
Not necessarily, which is why the tolerance test is last. The right flow ratio gives walls the right width, but clearance between mating parts also depends on shape, orientation and material shrinkage. Starting clearances are in the handbook, and annealed parts must be re-measured per annealing 3D prints.
How do I run these tests without OrcaSlicer?
The principles are the same, and equivalent test pieces exist as models or G-code for most slicers. This article follows the OrcaSlicer documentation because that's where the whole set is packaged with guidance on reading the results; the default numbers here are that software's and may differ elsewhere.
How long does a full calibration take for one spool?
The first four — temperature, flow, pressure advance, flow ratio — usually fit in one working session, since each test piece is a small block. Motion tests belong to the printer, so run them once and keep them. For a production workshop the time pays back quickly; for someone printing the occasional decorative model, the default profile is usually enough.
Should I calibrate first or dry the filament first?
Dry first, always. Wet filament throws off almost every measurement on the list: it strings, makes the extrusion bubbly and pitted, and makes actual flow unstable. A set of numbers measured on a wet spool is wrong as soon as you dry it.