E-Steps & Extrusion Calibration Calculator
Correct your extruder steps in one pass — and see whether a 100 mm test can actually resolve the error you are chasing.
How long the test needs to be
A 100 mm test with a ruler is 2:1 on a 1% error
Resolution is simply the measurement error divided by the test length. Half a millimetre over 100 mm is 0.5%, and the error people are chasing is 1% — two-to-one signal against noise, which tells you the direction and leaves the magnitude a guess. Extruding 300 mm instead turns the same ruler into 0.17% and six-to-one. The fix costs a couple of metres of filament and nothing else, which makes the short test hard to justify — and if you'd rather not, callipers on a marked strand get the same improvement out of 100 mm.
And the diameter error swamps it anyway. Extruded volume goes as the square of the filament diameter, so a spool whose mean is 1.72 mm rather than 1.75 delivers 3.4% less plastic than the slicer assumes — over three times the E-steps error being hunted. Measuring your filament with callipers and entering the real figure is worth more than another decimal place of E-steps, and it takes a minute.
The correction is exact in one pass. New steps = old × asked ÷ actual, and the relationship is purely linear — starting from 93 with a true value of 96 gives exactly 96.0000 first time. Iterating is measuring your ruler, not your extruder. If a second pass disagrees with the first, the measurement is the problem.
- E-steps and flow multiplier are not the same correction. E-steps is measured in free air at slow speed with no back-pressure; gear slip under load, filament compression and die swell appear only when printing. E-steps makes the extruder feed the length it was asked for; flow multiplier makes the printed line the width it was meant to be.
- Reference the mark to something rigid. Measuring from the spool or the tube measures the flex in the Bowden, not the extruder. Cutting flush at the entry and measuring the offcut removes the parallax entirely.
- Extrude slowly and hot. At 100 mm/min or less you're measuring the feed mechanism rather than the hotend. Testing at print speed measures both at once and tells you about neither.
How to use
- Mark the filament against something rigid, not the spool or the tube.
- Extrude slowly and hot so you measure the feed rather than the hotend.
- Correct once — the relationship is linear and a second pass adds nothing.
- Measure your filament diameter, which is usually the larger error.
Frequently asked questions
How do I calibrate E-steps?
Mark the filament a known distance above the extruder entry, command a measured extrusion slowly, and measure what actually fed. New steps equal old times asked over actual — so 93 steps that delivered 98 mm when asked for 100 becomes 94.898. The relationship is purely linear, so one measurement gives the exact answer.
Is the 100 mm E-steps test long enough?
Barely. Resolution is the measurement error divided by the test length, so half a millimetre with a rule over 100 mm resolves 0.5 per cent — against the 1 per cent error people are usually chasing, that is two-to-one signal against noise. It tells you the direction and leaves the magnitude a guess.
How long should the E-steps test be?
Long enough for about five-to-one on the error you care about. With a steel rule at half a millimetre, that means around 250 to 300 mm rather than 100 — which turns 0.5 per cent resolution into 0.17. The extra filament costs almost nothing and it is the only variable you control for free.
Do I need to repeat the E-steps calibration?
No, and a second pass usually just measures your ruler again. The correction is exact in one iteration because the relationship is linear: starting from 93 steps with a true value of 96 lands on exactly 96.0000 first time. If a second pass disagrees with the first, the measurement is the problem rather than the extruder.
Why does my filament diameter matter more than E-steps?
Because extruded volume goes as the SQUARE of the diameter. A spool whose mean is 1.72 mm rather than the 1.75 the slicer assumes delivers 3.4 per cent less plastic — over three times the 1 per cent E-steps error most people hunt. Measuring with callipers and entering the real figure is a one-minute job with a larger effect.
What is the difference between E-steps and flow multiplier?
They correct different errors and neither excuses the other. E-steps makes the extruder feed the length it was asked for, measured in free air at slow speed with no back-pressure. Flow multiplier makes the printed line the width it was meant to be, which also has to absorb gear slip under load, filament compression and die swell — none of which an air test can see.
Should I calibrate E-steps or flow first?
E-steps, always, because flow multiplier is a correction on top of it. Tuning flow to compensate for wrong E-steps produces a number that only works at one speed and one temperature, and it hides the underlying error rather than fixing it. Get the feed right in free air first, then tune flow against a printed test.
Why measure with callipers instead of a ruler?
Because it multiplies the resolution by five without lengthening the test. A rule read by eye is honestly about half a millimetre; callipers on a marked and cut strand are about a tenth. On a 100 mm test that is the difference between 0.5 per cent resolution and 0.1 — the same improvement lengthening the test to 500 mm would give.
What speed should I extrude at for the test?
Slowly — around 100 mm a minute or less. At that rate the extruder is not fighting the melt, which is the point: you are measuring the feed mechanism rather than the hotend. Testing at printing speed measures both at once and tells you about neither, and a partly blocked nozzle turns the test into an expensive way to strip filament.
Where should I measure from on the extruder?
Something rigid on the extruder body. Measuring from the spool or from the far end of a Bowden tube measures the flex in the tube along with the extrusion, which is exactly the kind of error the test is meant to remove. Cutting the filament flush at the entry and measuring the offcut afterwards eliminates the parallax entirely.
My E-steps correction is over 5 per cent. Is that normal?
It suggests a wrong configuration value rather than a worn extruder — a stock figure for a different gear ratio, or a firmware default that never matched the hardware. Genuine mechanical drift is a per cent or two. Before accepting a large correction, check the configured value against the extruder actually fitted, since a mismatch there is far commoner than a machine that has changed.
Does E-steps calibration differ on Klipper?
The arithmetic is the same but the parameter is inverted. Klipper uses rotation_distance in millimetres per rotation rather than steps per millimetre, so where Marlin multiplies by asked over actual, Klipper multiplies by actual over asked. Everything about test length, measurement precision and the diameter error applies identically.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.