The motors are uncomfortable to touch after a long print, and prints have started developing missed steps or shifts late in long jobs while short jobs are fine.
Warm is normal, hot is a symptom
A stepper draws current continuously to hold position, whether or not it is moving. That current becomes heat, so a motor that is doing nothing is still dissipating power — which surprises people used to motors that are only warm under load.
A stepper at 50 °C is unremarkable and most are specified to tolerate considerably more. What matters is not whether it feels hot but whether it is still delivering torque: magnet strength falls as temperature rises, and a motor that is losing torque at hour six is a motor whose prints fail at hour six.
So treat "too hot to hold" as a signal to investigate rather than as a fault in itself, and treat missed steps late in long prints as the real evidence.
The four causes
Driver current set higher than the motor needs. Every surplus milliamp becomes heat and buys no useful torque. Machines shipped with a generic configuration are frequently set high for margin, and lowering it usually improves everything.
Mechanical binding. A stiff rail, a tight leadscrew nut, a belt over-tensioned to the point of loading the bearings — anything that raises the mechanical load raises the current the driver delivers to keep up. This is the cause that produces one hot motor while the others are fine.
Poor airflow around the electronics. The driver chips run hotter than the motors do, and many have thermal protection that reduces output when they overheat. That produces the confusing case of a machine that loses steps while the motors themselves feel reasonable.
An enclosure with the electronics inside it. A chamber at ABS temperatures is a poor environment for power electronics. This is a known design compromise on several enclosed machines and it is why some manufacturers moved the boards outside the heated volume.
Bringing the temperature down
- Lower the driver current in small steps until the machine starts to lose steps under a demanding print, then go back up one increment. The lowest reliable setting is the coolest one, and it is almost always below what the machine shipped with.
- Move every axis by hand with the motors disabled. It should feel smooth end to end, with no stiff spots and no gritty patches. Anything you can feel is load the motor is paying for continuously.
- Check belt tension is firm rather than tight. Over-tensioned belts load the motor bearings and heat both.
- Confirm the electronics fan is running. A stopped fan on the board is invisible from the front of the machine, and it produces exactly this pattern of late-print failures.
- Enable driver standstill current reduction if the firmware offers it, which cuts holding current when an axis is idle. Note that this can cause a Z axis to sag on some machines; test before trusting it on a long print.
The failure nobody expects
On machines with printed parts, the bracket holding a motor may be PLA — which softens around 55 to 60 °C. A motor running at that temperature will slowly deform its own mount, and the geometry of the machine changes with it: belt alignment shifts, a pulley starts rubbing, and prints go wrong in ways that appear to have nothing to do with heat.
If your machine has printed motor mounts and hot motors, inspect them. Reprinting the affected brackets in PETG, ASA or a fibre-filled grade is a permanent fix, and it is worth doing before the deformation has propagated into the rest of the assembly.
Which motor is which problem
The extruder motor hot — often a filament path with too much drag, or an extruder fighting a partial blockage. Check that before blaming the driver; see extruder clicking.
One axis motor hot, others cool — mechanical binding on that axis.
All motors hot — a global current setting, or ambient temperature.
Drivers hot, motors fine — cooling of the electronics bay.
What this shows up as in prints
Late shifts, most often — the pattern covered in detail on layer shifting, where "only on long prints" is one of the diagnostic branches. A Z motor losing torque can also produce a gradual loss of layer registration rather than a clean step, which is harder to spot and easier to blame on the slicer.
Materials make no direct difference here, but the profile does: high-acceleration profiles demand peak torque constantly, and a machine that runs a fast profile all day is asking considerably more of its motors than one printing at moderate speeds.