A print starts perfectly. Somewhere between twenty minutes and an hour later extrusion falters and stops, at no particular height and on no particular feature. The extruder clicks, the filament will not push through by hand, and after everything has cooled down the machine works again — right up until the next print, which fails the same way at about the same time.
That clock is the diagnosis. A blockage that was already there fails immediately; a mechanical fault recurs at the same height; this one recurs at the same elapsed time.
The heatbreak is supposed to be a thermal choke
A hotend needs a short, sharp transition: filament rigid a few millimetres above the melt, fully molten a few millimetres below. That transition is created by fighting heat conduction. The heater block pushes heat upward through the metal; the heatsink and its fan pull that heat back out; between them sits a deliberately thin-walled section whose job is to resist the flow.
If removal stops keeping up with conduction, the boundary between soft and rigid migrates upward into the narrow bore. Filament there is warm enough to soften and expand but has nowhere to expand into, so it swells against the wall and locks. The extruder is now pushing a plug it created itself, and it will grind or click rather than move.
The delay is simply how long the assembly takes to reach that thermal state. That is also why the fault clears on its own after the machine cools, and why it comes back on the next print at the same point on the clock.
Triggers, in the order they turn out to be the cause
- The hotend cooling fan is not doing its job. Stopped, stalling under load, obstructed by a stray zip tie, or throttled by firmware that ramps it with temperature. This is the first thing to check and it accounts for most cases outright. It is not the part-cooling fan — the one that matters here blows across the heatsink fins, not at the print.
- Ambient temperature. An enclosure built for ABS makes a superb heat-creep generator when someone later prints PLA in it with the door shut. Anything that raises the air the heatsink is trying to reject heat into shortens the time to jam.
- An all-metal hotend with marginal airflow. All-metal designs conduct more heat upward by construction, which is the trade for being able to run high temperatures. On low-temperature materials that trade is unhelpful, and PLA in a poorly cooled all-metal hotend is the classic pairing.
- The heatbreak not seated properly. If it does not clamp hard against the top of the nozzle, there is a small gap where plastic collects and a poor thermal path where you wanted a good one.
- Very long retractions. Every retraction pulls softened plastic up into the cool zone. Whether this matters much is genuinely argued about: some experienced builders treat it as a leading cause, others measure little effect and point out that machines with no retraction at all still creep when the fan fails. The honest position is that it is a contributor rather than a root cause — worth reducing if your retraction is unusually long, not worth chasing if it is already modest.
What to do about it
- Watch the heatsink fan through the whole first twenty minutes of a print, not just at startup. A fan that spins on power-up and stalls when warm looks healthy exactly when you check it.
- Open the enclosure for low-temperature materials. If a machine only creeps with the door shut, that is your answer and it needs no parts.
- Duct air to the heatsink if the machine sits in a warm room or a cabinet.
- Check the heatbreak is tight against the nozzle by dismantling at temperature and reassembling so the nozzle is snugged down onto it.
- Shorten retraction to whatever still controls stringing, rather than to a number from a forum.
Success looks like a print that passes the hour mark. Anything shorter than that is not yet evidence, which makes this a slow fault to test — run a long, boring print for the trial rather than a short one.
Two instincts that make it worse
Raising the nozzle temperature is the natural response to something that will not extrude, and it is precisely backwards. A hotter block pushes the soft-to-rigid boundary further up the bore, which is the thing going wrong.
Wrapping insulation around the hotend is the other one. A silicone sock over the heater block is good practice and helps stability. Insulation above the block, on the heatsink or the heatbreak, blocks the escape route the design depends on.
The materials that creep
PLA leads, because it softens at a low temperature and any warmth in the wrong place reaches it first. Foaming LW-PLA is worse still — the whole point of the material is that it expands when heated, and it will happily do that inside the heatbreak. TPU is prone for a different reason: a soft filament deforms into the bore rather than sliding through it. Materials with high softening points are the least affected, which is why an enclosure that ruins PLA prints leaves polycarbonate untouched.