Why is my CNC router losing steps?
The machine starts a job on-zero and finishes off it. A second pass does not line up with the first, a pocket walks sideways, or the whole job shifts after a fast move. Nothing is physically stuck - the controller simply commanded moves the motors did not fully make, and because open-loop steppers have no feedback, it never knew.
Quick answer. The most common cause by a wide margin is acceleration or jerk set too high - the motor is asked to change speed faster than it can and slips. Reduce acceleration first; it is free and fixes the majority of cases. The next most common is simply not enough stepper torque (or current) for the cut being attempted.
Work down the list, most likely first
These are ordered by how often they are the answer, not by severity. Check each in turn and stop at the first one that confirms.
- 1
Acceleration or jerk set too high
Confirm it's this
Steps are lost at the moment of a direction change or the start of a rapid, not in the middle of a steady cut. The faster and jerkier the toolpath, the worse it gets.The fix
Lower the acceleration setting in your controller (GRBL $120/$121/$122, or the equivalent) in stages until the loss stops, then keep a margin. This is the first thing to try because it costs nothing and resolves most lost-step complaints. - 2
Insufficient stepper torque or drive current for the cut
Confirm it's this
Steps are lost only under cutting load - a deep or fast pass in hard material - and never during no-load moves. The motors may feel hot or the loss tracks how aggressive the cut is.The fix
Reduce the cut load (shallower depth of cut, slower feed - use the calculator to find a sane chip load) and, if your driver allows, raise the motor current toward its rated limit without overheating. A machine simply at the edge of its torque envelope will lose steps until the cut is eased. - 3
Loss specifically on direction changes
Confirm it's this
The offset appears right where the toolpath reverses or corners hard. It is the combination of the load's inertia and a motor low on torque at that instant.The fix
Lower acceleration and, on heavy gantries, reduce cornering speed. If it persists, the machine is torque-limited for that mass - see the closed-loop note below. - 4
Mechanical binding
Confirm it's this
Push the axis by hand with the machine off - it should glide. If it catches, drags, or is tight at one end of travel, the motor is fighting friction and stalling.The fix
Find and clear the bind: misaligned rails, an over-tight V-wheel or gib, debris in a leadscrew, or a rail that is not parallel. The axis must move freely by hand before any electrical tuning is meaningful. - 5
Loose motor couplings or pulley set screws
Confirm it's this
The motor shaft turns slightly without the leadscrew or pulley following, or a set screw is not seated on the flat of the shaft. Often intermittent and position-dependent.The fix
Check every coupling and pulley set screw; seat set screws on the shaft flat and add thread-locker. A slipping coupling mimics lost steps exactly but is purely mechanical. - 6
Worn bearings or worn drive components
Confirm it's this
Play or notchiness you can feel by hand, uneven resistance through travel, or noise localised to a bearing block.The fix
Replace the worn bearing or component. Wear raises friction and slop until the motor can no longer hold position reliably. - 7
Electrical noise (EMI) into the controller
Confirm it's this
Losses are random and unrelated to load or toolpath, sometimes only when the spindle/VFD or a dust collector is running. Unshielded step/direction wiring run alongside spindle power is the classic setup.The fix
Separate signal wiring from power wiring, use shielded cable grounded at one end, add ferrites, and ensure a clean common ground. Noise makes the controller see or miss pulses that were never commanded. - 8
Open-loop steppers at their limit
Confirm it's this
You have addressed acceleration, binding, couplings and noise, and the machine still loses steps on demanding cuts it is otherwise sized for.The fix
This is the ceiling of open-loop motion: there is no feedback to catch a slip. The upgrade is closed-loop steppers or servos, which detect and correct a missed step instead of silently dropping it.
Stopping it coming back
- Tune acceleration conservatively and leave headroom rather than running at the edge.
- Keep feeds and depth of cut inside a sane chip-load window for the material - run the numbers in the calculator rather than guessing.
- Re-check couplings, set screws and belt tension periodically; they loosen with use.
- Route signal wiring away from spindle and VFD power, shielded and grounded.
When it's the machine, not your setup
Some of this is setup, but some is the machine. A light gantry on open-loop steppers has a hard torque ceiling and no feedback to recover from a slip, so it will lose steps on cuts a heavier, closed-loop machine takes without noticing. If you have genuinely worked through acceleration, binding and noise and it still slips under normal cuts, you are at a hardware limit, not a tuning one. Rigidity and drive type are exactly what our scoring weighs, and the machines built to cut metal are the ones with the mass and the closed-loop drives to hold position under load.
FoxAlien XE-Ultra 8080
$1,799.00If you've worked through tuning and it still slips under normal cuts, the honest fix is a machine that can't. This one runs closed-loop steppers - they detect and correct a missed step instead of dropping it silently - on a stiffer aluminium frame, which is exactly the combination that ends chronic lost steps.
Frequently asked
- How do I know if I'm losing steps or my file is wrong?
- Run the same job twice from the same zero. If the second pass lines up with the first, the toolpath is fine and any single-run drift was lost steps. If both runs are wrong in the same way, it is the file or the zero, not the motion system.
- Does losing steps during rapids mean something different from losing them while cutting?
- Yes, and it is the single most useful tell. Steps lost during rapids or no-load moves point to acceleration set too high, mechanical binding, or electrical noise. Steps lost only under cutting load point to insufficient torque or too aggressive feeds and depth of cut.
- Will closed-loop steppers fix lost steps?
- They fix the class of problem where an open-loop motor silently slips and never recovers, because closed-loop drives detect the missed step and correct it. They do not excuse a machine binding mechanically or drowning in electrical noise - fix those first, since they are free.
Keep going
Diagnostic patterns are distilled from manufacturer documentation and established tooling references; where the community converges on a real-world tell, we state it. Exact feeds and speeds always come from the calculator, never invented here.