Flow rate and e-steps calibration, with the formulas

By Ned Walsch · Published November 9, 2026 · 5 min read

Why the order matters

Each step builds on the one before. The flow ratio corrects small differences between filaments. If the extruder itself pushes the wrong length, every filament profile carries that error. So fix the extruder once per printer, then tune flow once per filament. Shrinkage comes last, because it depends on the material.

E-steps calibration (Marlin)

1. Heat the hotend to printing temperature.
2. Send M503 and note the E value on the M92 line. That is the current e-steps.
3. Mark the filament 120 mm above the extruder entry.
4. Send M83, then G1 E100 F100. This extrudes 100 mm slowly, over about a minute.
5. Measure from the extruder to the mark again. If it reads 20 mm, exactly 100 mm went in.
6. Set the new value with M92 E…, then save it with M500.

For example, the old value is 93 and the mark ends 23 mm away. Then 97 mm went in, and the new e-steps are 93 × 100 ÷ 97 = 95.88. Run it once more to confirm.

Rotation distance calibration (Klipper)

Klipper sets the extruder with rotation_distance, in millimetres per motor turn. It works the other way round to e-steps: a larger value pushes less filament.

1. Heat the hotend. Mark the filament about 70 mm from the extruder intake, and measure the distance.
2. Send G91, then G1 E50 F60. The slow move takes about 50 seconds.
3. Measure the mark again. The actual length is the first measurement minus the second.
4. Work out the new value, put it in printer.cfg under [extruder], and restart.

For example, the old value is 7.700 and 48.6 mm went in: 7.700 × 48.6 ÷ 50 = 7.484. If the actual length is more than about 2 mm off, run it again after the change.

Converting e-steps to rotation distance

A 200-step motor at 16 microsteps with 93 e-steps gives 200 × 16 ÷ 93 = 34.409.

Flow rate calibration in OrcaSlicer

Select the filament you want to tune, then open Calibration ‣ Flow ratio.

• YOLO (recommended): one print of eleven blocks in an Archimedean chords pattern. Pick the block with the smoothest top and no gaps between the arcs. The new flow ratio is the old one plus that block's modifier: 0.98 + 0.01 = 0.99.
• 2-pass (legacy): Pass 1 prints nine blocks; Pass 2 refines from −9 to 0. After each pass, new = old × (100 + modifier) ÷ 100.

Save the filament profile after each change. On a Bambu Lab printer, leave the Flow calibration box unticked when you print the test.

Flow rate calibration in Bambu Studio

Bambu Lab says its defaults suit its printers and filaments. Run this when defects persist after Flow Dynamics calibration, or for third-party filament. Connect the printer, choose the nozzle, plate and filament preset, then start the flow rate calibration.

• Automatic: printers with Bambu's Micro Lidar read the result themselves.
• Manual, coarse: blocks from 80 % to 120 % of the preset, in 5 % steps.
• Manual, fine: blocks from 91 % to 100 %, in 1 % steps.

Pick the block with the smoothest top surface, then save it to a filament preset. Automatic results are less reliable with transparent, sparkle or very shiny filament.

The quick way: measure a wall

Without a slicer test, print a hollow cube with one or two walls and measure the wall with calipers.

For example, 100 % flow, an expected 0.90 mm wall and a measured 0.95 mm: 100 × 0.90 ÷ 0.95 = 94.7 %. Measure in several places and average. The slicer tests above are more reliable, because a wall's width also depends on how its lines overlap.

Volumetric flow: how fast can you print?

0.42 mm lines at 0.2 mm layers and 200 mm/s ask the hotend for 16.8 mm³/s. If the filament profile's maximum volumetric speed is 12 mm³/s, the fastest speed is 12 ÷ (0.42 × 0.2), about 143 mm/s. Use your slicer's filament profile value as the limit.

Shrinkage: why prints come out the wrong size

Plastic shrinks slightly as it cools: PLA barely, ABS and ASA more, nylon most. Print a part of a known size, measure it, and scale it in the slicer.

For example, designed 100 mm and measured 99.6 mm: 100 ÷ 99.6 × 100 = 100.4 %. Set each axis separately if they differ. For holes and fits, test in the same material.

What changed — November 9, 2026: first published.