The printer stops partway through a job and displays a thermal runaway, heating failed, or temperature deviation message. Usually everything shuts down and the machine requires a power cycle to restart.
Read this before searching for a workaround
Firmware watches whether the measured temperature responds to heater output the way it should. If the heater is on full and the reading is not climbing, or the reading falls while heating, the firmware concludes that it can no longer trust the relationship between the heater and the sensor — and shuts everything down.
The reason that check exists is that a heater cartridge with no working feedback will keep heating. A hotend has no mechanical limit to how hot it can get; a heater that stays on because its sensor fell out of the block will melt the block's surroundings, ignite plastic and set fire to whatever the machine is standing on. It has happened, more than once, and it is why every mainstream firmware ships this protection enabled.
Do not disable thermal protection. Do not widen the tolerance window until the error stops appearing. Do not follow instructions to comment it out of a firmware build. The error is the only thing standing between a hardware fault and an unattended fire, and a machine that trips it is telling you something real about itself.
What actually triggers it
A thermistor working loose. The commonest cause by a distance. The sensor sits in a small hole in the heater block, held by a grub screw, a clip or a smear of thermal cement. If it backs out even slightly it reads the air near the block rather than the block, so the reading lags, the firmware commands more power, and the discrepancy grows until the check fires.
Air blowing on the heater block. A part-cooling fan aimed slightly wrong, or a missing silicone sock, chills the block whenever the fan ramps up. The classic signature is a halt that occurs a few layers in — exactly when cooling comes on — rather than at startup.
A failing heater cartridge. Intermittent internal connections produce a heater that works, then does not, then does. If the machine takes noticeably longer to reach temperature than it used to, suspect this before anything else.
Wire fatigue in the toolhead loom. Cables that flex thousands of times per print break strand by strand. A thermistor lead with a nearly broken conductor gives wild readings whenever the head moves in a particular direction.
A bed that cannot hold its setpoint in a cold room, on a machine whose firmware treats slow heating as a fault.
Finding it
- With the machine cold, tug gently on the thermistor and heater leads where they leave the block. Any movement at the block itself is the fault.
- Heat the hotend and watch the reported temperature while you flex the cable loom by hand. A spike or a dropout as you move it is a broken conductor, and it is the test that finds intermittent faults that pass every static inspection.
- Watch the temperature graph rather than the number. Most interfaces plot it. A healthy hotend climbs smoothly and holds a flat line; a failing one shows noise, steps or a slow drift.
- Check the silicone sock is present and seated, and that the part-cooling duct points at the print rather than at the block.
- Run a PID tune afterwards. Any change to the sock, the cartridge, the thermistor or the fan changes the thermal behaviour the tuning was calculated for, and an untuned hotend oscillates enough to trip a deviation check on its own.
The bed version
The same protection applies to the heated bed, and there it is more often a genuine limit than a fault. A large plate in a cold garage may simply be unable to reach 110 °C, and the firmware sees a heater that is on with a temperature that has stopped rising.
The answer is to raise the room temperature, enclose the machine, or accept a lower bed setpoint — not to extend the timeout. A machine that cannot reach the temperature a material needs is a machine that should not be printing that material.
What is not this error
A reading of a wildly wrong number at startup — often shown as a minimum or maximum temperature error — usually means an open or shorted sensor circuit rather than a runaway condition. It is still a wiring fault, and it still wants finding rather than bypassing.
Intermittent extrusion problems with no halt at all are a different matter entirely; see extruder clicking, where a thermistor reading high is one of the candidates.
While you are in there
Two habits worth adopting on any machine that has tripped this once: check the toolhead loom's strain relief so the cable bends in a loop rather than at a point, and keep a spare thermistor and heater cartridge in a drawer. Both are inexpensive, both fail eventually, and both fail in the middle of a long print rather than at a convenient moment.