Flow Rate Calibration: Slicer Tests and Dimensional Accuracy

Flow rate calibration adjusts the amount of filament requested by a slicer. The useful result is a saved material profile that produces joined lines and a reasonably even surface at your chosen temperature. It can improve dimensional accuracy, but an oversized hole or a short Z dimension does not, by itself, identify a flow error.

Use this procedure after the hardware extrusion baseline and nozzle temperature are settled. It covers OrcaSlicer’s generated flow tests, manual checks in PrusaSlicer and Cura, and the firmware boundary on Klipper. The calculations below are worked examples, not measurements from this site.

Flow Ratio, E-Steps and Volumetric Flow

These three settings answer different questions. E-steps, or Klipper’s rotation distance, describe filament movement through the drive. The slicer flow ratio scales the requested extrusion. Maximum volumetric flow limits how quickly the hotend is asked to process material.

A ratio of 0.96 means 96% of the slicer’s baseline request. It does not mean that the hotend can melt only 96% of its rated output. A volumetric limit, by contrast, uses cubic millimetres per second. Raising a flow ratio to compensate for starvation at excessive speed can overfill the same part when the printer slows.

The extruder calibration guide covers the hardware baseline. Keep it stable during the material test. A factory-configured printer that does not expose extruder calibration should use its supported filament calibration workflow; do not invent a hidden e-steps adjustment.

Step 1: Establish a Repeatable Baseline

Start with a mechanically sound printer, a clear nozzle and filament in suitable condition. Choose the correct printer and nozzle profile. Use the same spool throughout the comparison, and record its material, brand and colour so the saved result has a useful identity.

Find a suitable temperature with a temperature tower. Then leave temperature, cooling, layer height and speed unchanged while comparing flow. Otherwise, a better-looking patch could reflect a different thermal condition instead of a better multiplier.

The plate must support a consistent first layer. If a test detaches, curls or drags under the nozzle, fix that before reading its top surface. The bed-leveling guide explains the separate roles of leveling and nozzle offset.

Save a copy of your current profile. Write down its starting ratio and any printer-side flow override. For a controlled comparison, avoid stacking a temporary firmware adjustment on top of a slicer change. Keep ironing and special surface-flow overrides out of the baseline test because they change the surface being judged.

Step 2: Choose the Flow Calibration Method

Klipper is firmware, so it does not replace the slicer in this table. You can run an OrcaSlicer or manual coupon procedure on a Klipper printer using a compatible machine profile.

Prefer a visual coupon when the main complaint is rough solid surfaces. A wall measurement is an additional dimensional check when the tool and geometry support repeatable measurements. Do not average results from different methods into one unexplained number; save the settings and judge a verification part.

OrcaSlicer Flow Ratio Test

Select the printer, filament and process, then open the flow-ratio calibration menu. The OrcaSlicer project documentation describes both YOLO and the legacy two-pass method. They use different modifier arithmetic.

For YOLO, print the generated blocks and choose the surface with joined arcs and minimal excess at their boundary. Apply the signed modifier directly: a starting ratio of 0.97 and a chosen modifier of +0.01 give 0.98. Save the filament profile.

For the two-pass method, generate Pass 1 and choose the smoothest block without gaps. Calculate:

new ratio = old ratio × (100 + selected modifier) / 100

Save that result before generating Pass 2, then apply the second modifier to the updated ratio. Do not reuse the original baseline for both passes. On Bambu printers, follow the project’s instruction to leave the separate print-time flow-calibration option unchecked for this test.

These are visual block comparisons. They do not ask you to measure a single wall and automatically convert that measurement into a profile.

PrusaSlicer Manual Method

In PrusaSlicer, the relevant filament setting is Extrusion multiplier. Start with the visual calibration model and profile described in Prusa’s calibration instructions. Inspect the solid top after it cools. Reduce the multiplier if material builds up; increase it if adjacent lines remain separated.

Use small changes, around 1–2%, and repeat with the same geometry. Keep a note of the old and new values beside each coupon. A multiplier of 0.98 represents 98%; entering 98 in a ratio field would be a completely different request.

Prusa also documents a precision wall method. For its stated 0.4 mm nozzle profile, the example extrusion width is 0.45 mm; other nozzle and layer-height pairings change that width. Use the actual width from your own controlled test, take several measurements away from irregularities, and average them. Inexpensive calipers may not resolve a single wall reliably. Validate the resulting multiplier with a surface print before accepting it.

Cura Manual Method

Cura exposes global Flow as a percentage, plus values for particular features. Search the settings panel for Flow, and inspect which wall, skin and initial-layer values inherit it. Cura’s setting definitions show these relationships; a previously overridden skin value can stay different from the global value.

Use a simple coupon with a flat solid top. Slice a baseline copy, then another with a small global Flow change, keeping the rest of the profile identical. Sequential prints are easy to label and avoid accidentally comparing different per-object settings. Compare the broad surface, not only its edge or seam.

For example, compare 100% with 98% when the baseline shows excess material. If the lower setting creates gaps, the direction or size of the change was unsuitable. Return toward the baseline and check the prerequisites instead of repeatedly lowering Flow.

Once the global setting is established, restore any intentional feature-specific settings separately. An initial-layer defect belongs to the first-layer check; it should not dictate how every later wall is extruded.

Klipper: Keep Rotation Distance Separate

Klipper’s rotation-distance procedure establishes how far the extruder moves filament. Correct that when the hardware baseline is wrong, then leave it alone during per-filament flow calibration.

Klipper supports M221 S100 to set the extrusion factor to 100%. If a previous session left a different factor active, it can scale the output of the slicer’s test. Check your macros and start G-code so the factor used for calibration is intentional and repeatable.

For routine material tuning, save the result in the slicer filament profile. A temporary M221 adjustment is useful for comparison, but record it explicitly and avoid leaving two compensations active. For example, 0.96 in the slicer combined with 95% firmware flow produces an effective factor of 0.912.

Do not change rotation_distance to make one brand’s top surface look better. That changes the hardware baseline beneath every material profile. Klipper pressure advance is another separate setting and belongs to the later timing test.

Step 3: Calculate and Save the Correction

Here is a complete two-pass arithmetic example with hypothetical coupon choices:

The final result is about 0.9673, or 96.73%. The second calculation uses the result of the first. The examples are not recommended values for PLA, PETG or a particular printer; your coupon determines the modifier.

For a controlled wall check, the proportional estimate is:

new flow = current flow × expected wall width / measured wall width

If a test specifies 0.45 mm and repeated readings average 0.47 mm, a current 100% becomes approximately 95.74%. This estimate assumes the discrepancy represents extrusion quantity. A seam blob, flexible wall or inappropriate width assumption breaks that connection.

Keep more decimal places during the calculation than you enter in the interface. Record the rounded value actually saved, then regenerate the verification G-code. Reprinting an old export only repeats the old setting, even if the profile shown on screen has changed.

Step 4: Read the Defect-to-Flow Lookup Table

Use the table to choose a direction for the next comparison. It cannot diagnose a universal flow percentage from a photograph. The small changes are trial increments, not material specifications.

Compare the same region of each coupon. An edge beside a perimeter and a broad central patch represent different interactions. Pick the best overall surface, then check that the choice has not introduced gaps elsewhere.

If every sample is poor, widen the investigation before widening the flow range. The test needs repeatable extrusion to distinguish settings. A slipping drive can make one coupon appear underfilled regardless of the multiplier attached to it.

Step 5: Verify Dimensional Accuracy on Another Part

Print a second geometry with the chosen profile. Include a flat top and walls; include a hole or mating feature if that is what the real job needs. Let the part cool before taking dimensional readings, and measure at consistent locations.

A cube is useful for recording dimensions, but a cube alone cannot separate extrusion quantity from geometry, shrinkage or first-layer effects. Do not lower global flow until every external dimension matches while allowing visible gaps to appear in the surface.

Keep dimensional measurements and surface judgments in separate columns. For example, record the intended size, measured size, location measured and whether the top lines are joined. That makes it possible to see whether an apparent dimensional improvement came at the expense of a sound surface.

If a dimensional error is concentrated at corners, use the pressure advance calibration procedure. If it is concentrated at the base, return to leveling and first-layer settings. These patterns are reasons for another check, not proof that one setting must be wrong.

No universal tolerance follows from a flow calibration. Choose acceptance criteria for the actual mating parts, measurement tools and use case. A decorative model and a sliding fit require different evidence.

Step 6: Record the Profile and Continue in Order

Save the filament profile with a useful name and keep a short record of the printer, nozzle diameter, layer height, temperature, starting ratio, chosen result and test method. Note the firmware extrusion factor too. Without it, the saved ratio may not reproduce the coupon.

After a material, nozzle or hotend change, verify that the existing profile still behaves as intended. Hardware changes can require earlier calibration stages; a material change can require a new temperature and flow result. The calibration order maps those dependencies.

Keep the test export with the profile when practical. A record such as “PLA, 0.4 mm nozzle, two-pass flow test, baseline 0.98, final 0.9673, firmware factor 100%” explains how the example above was obtained. Add the actual temperature and date from your own session. A label that says only “best flow” loses those dependencies and makes a later comparison harder to interpret. When two coupons look equivalent, record that uncertainty instead of claiming the saved ratio is an exact physical constant.

Once flow is stable, tune pressure advance, then retraction under the same conditions. Revisit the surface check if those changes alter the result. Use the reference tables when choosing nozzle and layer-height combinations or changing the chamber setup.

FAQ

Is flow calibration the same as a maximum flow-rate test?

No. A flow-ratio test adjusts extrusion quantity. A maximum volumetric-speed test explores the rate at which the hotend can process material. A successful low-speed coupon does not establish that speed ceiling.

Should I use the same ratio for every PLA spool?

Keep the existing profile as a starting point and verify the new spool. The calibration belongs to the filament and printing conditions. A material label alone does not identify a universal multiplier.

Can a flow test replace e-steps calibration?

No. The slicer multiplier and the extruder movement baseline serve different roles. Correct a demonstrated hardware movement error through the firmware’s documented procedure, then evaluate the material profile.

What if the selected patch is between two modifiers?

Use a finer comparison around those values while keeping the conditions unchanged. Record the selection as a practical visual judgment. Extra decimal places do not make the surface comparison more precise.