Your slicer assumes 100mm of filament comes out when it asks for 100mm. If your extruder is off by even 3%, every print is over- or under-extruded by that much — and no amount of slicer tweaking fixes a hardware calibration error. I’ve measured extruders that were off by 12% from the factory. Here’s how to fix it permanently.
E-Step Calibration: The Complete Procedure
E-steps (extruder steps per millimeter) tell the firmware how many stepper motor steps it takes to push exactly 1mm of filament. Get this wrong and your flow rate is fighting a losing battle.
Step 1: Mark and Measure
Heat the hotend to printing temperature for your filament (200°C for PLA, 230°C for PETG). You need the hotend hot so filament can actually extrude — cold calibration through an empty extruder produces different numbers because there’s no back-pressure.
Load filament. Use calipers or a ruler to measure exactly 120mm from a fixed reference point — the extruder body’s filament entry is ideal. Mark the filament at 100mm and 120mm from that reference. The 20mm margin prevents running filament completely out of the extruder during the test.
Command the extruder to feed exactly 100mm. In Marlin, use the LCD menu or send G1 E100 F100 via terminal. In Klipper, use the Mainsail/Fluidd interface or send the same G-code.
Step 2: Measure the Actual Output
After the extrusion completes, measure the remaining distance from the reference point to your 120mm mark. If the mark is 22mm away, you extruded 120 – 22 = 98mm instead of the commanded 100mm. Your extruder is under-extruding by 2%.
The formula:
New E-steps = Current E-steps × (100 / Actual Extruded Length)
Example: Current E-steps = 93.0, actual extruded = 96.5mm
New E-steps = 93.0 × (100 / 96.5) = 96.4
Step 3: Update Firmware and Verify
Marlin: Send M92 E96.4 then M500 to save to EEPROM. Or update in the LCD menu under Control → Motion → E-steps/mm.
Klipper: Update rotation_distance in the [extruder] section of printer.cfg. Klipper uses rotation_distance instead of steps/mm — the formula is rotation_distance = old_rotation_distance × (actual_extruded / 100). Then RESTART.
Verification: Run the 100mm test again. If the result is between 99.5mm and 100.5mm, you’re calibrated. If not, repeat — one calibration pass is rarely perfect because the first pass corrects the gross error, and a second pass dials it in.
Step 4: Separate E-Steps from Flow Rate
E-steps calibrate the extruder hardware. Flow rate (Extrusion Multiplier in some slicers) compensates for filament diameter variation and material-specific behavior. After E-steps are correct, print a single-wall calibration cube (0.4mm wall width), measure the actual wall thickness with calipers, and calculate flow rate:
New Flow Rate = Old Flow Rate × (Expected Wall Width / Measured Wall Width)
E-steps handle the mechanical accuracy. Flow rate handles the material behavior. Never use flow rate to fix a bad E-step calibration — that’s like adjusting your speedometer by driving faster.
E-Step Calibration Reference
| Extruder Type | Typical E-Steps (1.8° stepper) | Typical E-Steps (0.9° stepper) | Notes |
|---|---|---|---|
| Stock Creality (plastic) | 93 | 186 | Single-gear, no reduction |
| BMG Clone (dual drive) | 415 | 830 | 3:1 gear reduction |
| Bondtech LGX | 400 | 800 | Integrated gear reduction |
| Orbiter V2.0 | 690 | 1380 | 7.5:1 planetary gear |
| Titan (E3D) | 420 | 840 | 3:1 reduction |
| Sherpa Mini | 675 | 1350 | Nema 14, compact |
Common E-Step Calibration Mistakes
Mistake 1: Calibrating E-steps with the hotend cold.
The consequence: Without back-pressure from molten filament, the extruder pushes more filament per step. Your cold calibration reads 93, but under actual printing conditions the real value is 96 — a 3% under-extrusion baked into every print. The fix: Always calibrate with the hotend at printing temperature and filament loaded.
Mistake 2: Measuring from the spool instead of the extruder entry.
The consequence: Filament path friction between the spool and extruder varies with spool position. Measuring at the spool includes that variable friction in your measurement. The fix: Always measure at the extruder body entry — that’s the point where the extruder gear starts applying force, and it gives a measurement independent of spool tension.
Mistake 3: Using E-steps to fix flow issues caused by slicer settings.
The consequence: You set E-steps to 98 to “reduce over-extrusion,” then switch to a different filament that actually needs 100. Now you’re under-extruding and you’ve forgotten your E-steps aren’t at the hardware-correct value. The fix: E-steps are a hardware constant — set them once for the extruder and never touch them again. Use slicer flow rate for material-specific tuning.
Mistake 4: Not saving after M92.
The consequence: You send M92 E96.4, the extruder works perfectly for this print, and after power cycling it reverts to the old value because you forgot M500. The fix: Always follow M92 with M500 (Marlin) or update printer.cfg and RESTART (Klipper). Verify after reboot with M503 (Marlin) or check the printer.cfg value.
⚠️ Safety Notice: Always follow manufacturer safety guidelines when modifying your 3D printer’s firmware and calibration settings. Ensure your printer is on a non-flammable surface, has thermal runaway protection enabled, and is operated in a well-ventilated area. Never leave a running printer unattended.
E-step calibration is the foundation — once it’s right, other extrusion issues are diagnostic, not guesswork. Our over-extrusion diagnosis guide covers what to check when your prints look over-stuffed despite correct E-steps. And when prints are consistently weak or have gaps, our under-extrusion troubleshooting guide walks through nozzle clogs, extruder tension, and filament path issues.
For printers where the stock plastic extruder arm is the weak link, the Bondtech BMG dual-drive extruder is a drop-in upgrade that eliminates the filament slip that makes E-step calibration drift over time. The dual hobbed gears maintain grip even with flexible filaments, and the 3:1 gear ratio provides enough torque for consistent extrusion at any speed.