Clogged Nozzle and Cold Pulls: Find Where the Plug Sits, Diagnosed in 4 Questions

The print is running, the head is moving — and nothing is coming out. Or almost nothing: a thin, curling dribble where a clean strand should be, an extruder clicking against a wall it cannot push through. That is a clog — but "clogged nozzle" is a location guess, not a diagnosis, and it is wrong often enough to matter: the plug can sit in the nozzle, above it in the heatbreak, or further up in the tube, and each address has a different cure. In the FixMyPrint community dataset, clogging and heat creep together account for 764 of more than 20,000 help posts — a count of community help posts in that analysis, not a measurement of how often nozzles clog across all printers. Instead of a fix list, this guide asks four triage questions: real clogs overlap (one jam can click and dribble and return tomorrow), so the questions set an order of suspicion — cheapest discriminating tests first, the plug's location established by evidence before anything is heated, pulled or unscrewed.

Compiled from published sources — Prusa's Knowledge Base, Bambu Lab's Wiki, Simplify3D's troubleshooting guide, the FixMyPrint dataset — plus clearly labelled community advice. Not personally tested on a printer in our workshop. Manufacturer figures are labelled; where no defensible universal number exists, the guide gives a test instead of a fake-precise setting.

Applies to: any FDM printer · direct-drive and Bowden machines (tests flag where they differ) · PrusaSlicer, Bambu Studio, OrcaSlicer or Cura. Scope: the filament path from the extruder's exit to the nozzle tip — nozzle, heatbreak, heatsink and tube. The extruder itself (gears, idler, feed calibration) is covered by our under-extrusion guide; this guide takes over where its feed test hands off. Materials scope: standard PLA, PETG and similar filaments; abrasive or filled filaments (carbon-fibre, wood-fill, glow-in-the-dark) wear brass nozzles and shed particles — they appear here as a clog cause and as Branch E's replacement case.

In this guide: Step 0: Read the failure · The 4-question triage · First: restore the baseline · Branch A: The plug is in the nozzle · Branch B: The plug is above the nozzle — heat creep · Branch C: The plug is in the path — tube and coupler · Branch D: It is not a clog at all · Branch E: The nozzle itself is finished — replace it properly · What NOT to do · FAQ

Key Takeaways

"Clogged nozzle" is a guess about location. Prusa's Knowledge Base says it plainly: a clogged hotend is not the same thing as a clogged nozzle, and changing the nozzle is not a universal solution.
Location is the diagnosis. A cold pull can only grab what the melt touches — it cannot cure a soft plug above the heatbreak or a deformed tube.
The hand-push test helps split the machine in half. Where your maker permits manual feeding, pushing filament by hand at printing temperature takes the extruder gear out of the push, and the resistance you feel helps distinguish a melt-path obstruction from a feed-side fault. It does not isolate them conclusively — and a free-flowing push does not exclude a partial clog, or a restriction that only appears at printing flow rates.
Cold-pull temperatures are model-specific. Prusa alone publishes ≈85 °C, ≈90–95 °C, ≈100 °C and an automated 80 °C cycle for different machines. Your maker's figure outranks every number here — and one Prusa nozzle line must never be cold-pulled at all.
Read what comes out. The cold pull's tip — a clean cone, a debris-laden one, or nothing at all — is one of the most informative clues in this guide, to be read alongside the loading, flow and cooling behaviour, not on its own.

Step 0: Read the Failure (60 Seconds)

Look at how the plastic stopped — and at what it is not doing:

What you see	What it most likely is	Where to start
Nothing comes out at all; the extruder clicks or ticks against the blockage	A full clog — the path is closed somewhere (Prusa's Knowledge Base describes the same picture, clicking included)	Question 1's full-clog path: Branch A, or Branch C if filament won't even load
Thin, gappy lines; the purge strand comes out thin, intermittent, or curls up and sticks to the nozzle	A partial clog — something narrows the nozzle's exit (Prusa and Bambu Lab both list the curling purge among the early signs)	Branch A
Prints well, then stops extruding at a similar point in print after print; the unloaded tip looks swollen or mushroomed	Heat creep — filament softening above the melt zone and swelling where it should stay solid	Branch B
Filament will not load at all — it stops well above the nozzle, hot or cold	An obstruction in the path, not the nozzle: tube, coupler, or a swollen filament tip	Branch C
Clicking and grinding — but a hand push at temperature flows freely	Probably not a clog: a free-flowing push makes a severe obstruction unlikely and points at the extruder — gear, idler or strand damage (a partial or flow-dependent restriction is not excluded by the push alone)	Branch D, then our under-extrusion guide
Nothing extrudes for the first layer or two, then the print runs normally	The nozzle started too close to the bed — its exit was capped by the plate, not blocked inside (Simplify3D's description)	Our first-layer guide
Popping or hissing at the nozzle, pitted surfaces — but flow is roughly normal	Wet filament is one suspect: moisture flashing to steam disrupts the stream without blocking it	Branch D, then our PLA drying guide
Gaps and weakness everywhere, but the nozzle purges cleanly	The wider under-extrusion family — feed, melt capacity or slicer flow, not necessarily a blockage	Our under-extrusion guide

Note when it began — right after a material change, after a long print or a hot idle, after abrasive filament, or out of nowhere? A change you made is a suspect; no change points at debris, wear and slow carbonisation.

The 4-Question Triage

A triage sequence, not a decision tree: causes overlap, and one fault wears another's face (a heat-creep plug clicks exactly like a nozzle plug; a slipping extruder starves a healthy nozzle). The answers are priority rules — cheapest discriminating tests first (the hand-push, where your maker allows it, and the strand test take minutes and help sort the path into three addresses) — and test results overrule the question order whenever they disagree. Throughout: same test print, same filament, same temperatures, one change at a time, each test before its fix.

Question 1: Does anything come out at all? Prusa's own branch point: if filament can still be loaded and unloaded, a partial clog can usually be melted out or cold-pulled; a complete clog that prevents even loading calls for the force-out path or a look inside (Branch A covers both, in that order). Nothing coming out and filament that will not enter the path at all points higher — Branch C.

Question 2: What does your hand feel at printing temperature? This test exists only where your manufacturer permits manual feeding. Heat the nozzle to the loaded filament's printing temperature and release the extruder's grip exactly as your maker's procedure describes, then push the filament by hand — Simplify3D's first move for a jammed nozzle, and the mirror of our under-extrusion guide's feed test: with the gear taken out of the push, the resistance you feel helps distinguish a melt-path obstruction from a feed-side fault — it does not isolate them conclusively. Keep hands on the filament only: never on or near the heater block, the nozzle, moving parts or exposed electrics. If manual feeding is inaccessible or unsupported on your machine, do not improvise access — use the printer's own load and extrusion controls, and read the same three outcomes from how the machine feeds. Flows freely: a severe obstruction is unlikely — suspicion moves to the push side, Branch D — but a partial clog, or a restriction that only appears at printing flow rates, is not excluded. Flows, but with hard, springy resistance: a partial obstruction — Branch A (or B, if Question 4's history is time-shaped). Will not move at all: a full plug — Question 3 locates it.

Question 3: What does the strand do — and what does a cold pull bring out? Purge into the air. A straight, even strand argues against a nozzle obstruction. A strand that curls sideways and climbs back onto the nozzle, runs ribbon-thin or pulses is a clue, not proof — persistent curl is the classic partial-clog signature in Prusa's and Bambu Lab's documentation, but residue on the nozzle's face can bend a stream too. The more informative evidence is the cold pull's tip (procedure and safety gate in Branch A): a cone carrying the nozzle's imprint, dark specks or a strand of the previous colour is contamination from inside the hotend, in your hand. A pull that comes out clean twice while the fault persists shifts suspicion upward — Branches B and C — without definitively clearing the nozzle: an intermittent or flow-dependent restriction can survive clean pulls. A pull whose filament stretches, necks down or breaks instead of releasing as one piece says the pull temperature or technique was wrong for the material — or that the plug is not where the pull can reach.

Question 4: What is the history? Began right after stepping down from a hotter material to a cooler one: leftover high-temperature residue that the cooler melt cannot clear — Branch A (Prusa's named mechanism; its prevention is loading the cooler filament at the previous filament's temperature first). Began deep into prints, at a similar elapsed time, or after the printer sat hot and idle: Branch B. Began after abrasive or filled filament: particles in the nozzle (Branch A) today, a worn bore (Branch E) over time. Began out of nowhere on an old nozzle: debris or carbonised build-up, Branch A — with Branch E waiting if cleaning stops working.

First: Restore the Baseline (Before Any Teardown)

Before anything is heated for a pull, needled or unscrewed, restore honest ground: a known-good manufacturer profile — stock printer profile plus matching stock filament profile — and verify nozzle diameter matches the nozzle fitted and the filament diameter is the standard value. Load a spool you trust (dry, untangled, previously well-behaved) — a clog diagnosis built on a suspect spool convicts the wrong part. Then run the two baseline tests: the hand-push test (Question 2 — where your maker permits manual feeding; otherwise the printer's own load and extrusion controls stand in for it) and the strand test (Question 3), and note what they showed. If your extruder's feed itself is in doubt, run our under-extrusion guide's 100 mm feed test before the hotend is opened: it measures how much filament the extruder actually moves under the conditions tested — a full-length result clears simple under-feeding, though it does not prove the extruder can sustain high-flow extrusion. Still blocked on baseline? The branches below start from evidence, not from a printer already mid-repair.

Two profile-level checks belong to the baseline too, because both imitate a clog perfectly. Flow demand: a nozzle can purge freely at hand-push rates yet starve at the volumetric flow a print demands — run a short, conservative test print (a small maker-supported calibration or first-layer print) and, if the failure returns only at printing speed, compare the profile's flow demand with the hotend's configured maximum volumetric flow; a profile asking for more melt than that limit delivers is a settings fault, not a blockage. Nozzle/profile match: on machines sold in standard and high-flow nozzle variants, verify the sliced profile matches the nozzle actually fitted — Prusa documents that an HF/non-HF profile mismatch can produce severe extrusion problems that present exactly like a clog. And one standing rule for everything below: unstable temperature readings, or a firmware heater or temperature-sensor warning, are stop conditions — never a reason to raise the temperature and push on. A hotend that cannot report its own temperature cannot be cleared safely.

Why It Happens (30 Seconds of Physics)

A hotend is a controlled boundary: filament must stay solid until just above the nozzle, then melt in a short zone and leave through one tiny hole. A clog is anything that breaks that arrangement, and the breaks group into two broad families — families, not sealed boxes: they interact, and heat creep, for one, can cook loose residue into a hard in-path plug. First, something is in the path — debris carried in on the filament, carbonised residue from plastic that sat too hot too long, a scrap of a previous, hotter material, or a filler particle bigger than the exit. Second, the path changes shape — heat climbs above the melt zone (heat creep), filament softens early and swells against the walls where it should slide as a solid rod; or a tube end deforms and narrows the passage. The two families explain why location is the whole diagnosis: a cold pull bonds to the plug only where plastic melts and re-solidifies around it — inside the melt zone. A plug above that zone, or a passage narrowed by a swollen strand or a crushed tube, is out of the pull's reach, and pulling harder does not extend it.

Branch A: The Plug Is in the Nozzle — Debris and Carbonised Material

Signs: partial flow with a curling or thinning strand, or a full stop with filament that still loads; any filament, any profile; often a history from Question 4 — a material change, filled filament, or a long hot soak that cooked residue onto the walls. Bambu Lab's Wiki lists the in-nozzle family plainly: particulates from the filament (carbon-fibre, glow-in-the-dark and sparkling filaments are its named examples), debris that rode in on the strand, and previous filament not fully purged when types changed.

The tests are the triage: the hand-push (hard, springy resistance), the strand test (curl or thin stream — a clue), and the cold pull below, whose tip is this branch's most informative single clue. If filament cannot be loaded at all, skip to the full-clog path after the pull procedure.

Printer-specific safety gate — read before any cold pull: check your model, hotend and nozzle type first. There is no universal pull temperature: Prusa's own procedures differ by model — the pull is made at roughly 85 °C (MK3 series), 90–95 °C (MINI), around 100 °C (CORE One, manual), and the CORE One's automated cycle pulls at 80 °C; Bambu Lab's X1/P1 procedure cools to 100 °C before retracting; Simplify3D reports users succeeding with a pull at 100 °C. And Prusa states plainly: do not cold-pull a Nozzle X — cleaning filament only, as the pull can damage that nozzle. The same exclusion covers needles: Prusa bars the cleaning needle on that nozzle too — cleaning filament only, for both tools. If your maker publishes a procedure for your machine (several do, some automated), follow theirs; it outranks everything below. What follows is an explanatory overview of how a manufacturer-approved manual pull works, so you can recognise and follow your own maker's version — it is not a licence to improvise one. Do not attempt a manual pull where no approved procedure exists for your hotend and nozzle combination: some machines clear by disassembly, an extruder release or their own automated cycle instead, and forcing the pattern below onto them invites exactly the damage this gate warns about.

Load light-coloured PLA or cleaning filament — debris shows on a pale pull. Heat to its normal printing temperature and hand-feed until plastic flows as steadily as it will.
Keep gentle forward pressure while the nozzle cools toward your maker's pull temperature (the labelled figures above span roughly 80–100 °C by model and method — use your maker's, not a midpoint). The plastic must fill the nozzle as it cools; a pull taken on a half-empty nozzle brings out nothing.
At the pull temperature, grip the filament with pliers and pull — firmly, steadily, straight up, in one motion. Prusa's cautions: a harsh jerk can snap the filament off inside the hotend or damage the printer, and an angled pull invites the same break. The plug releases suddenly; keep your free hand and face clear.
Read the tip. A good pull ends in a cone mirroring the nozzle's inside, imprint included (Prusa's standard for a complete pull). Dark specks, a gritty crust, or a strand of a previous colour in the cone is evidence of contamination inside the hotend, in your hand — not a guess, though on its own it does not prove the debris retrieved was the whole blockage. Repeat from step 1 until the tip comes out clean and full-shaped.

One family shows why the overview is not a recipe: Prusa's Nextruder machines (MK4, XL, CORE One) clear a clog by a different path — fresh filament pushed through from above, and an automatic cold-pull cycle on the models that support it — under their own manuals. Where your machine's clearing path is its own, that path is the procedure.

Safety: through most of this procedure the nozzle sits at or near printing temperature and will burn skin on contact — pliers and tools, never fingertips, and never judge the temperature by touch. Work over a heat-resistant surface. Never escalate force against a plug that will not release — and if the filament will not move at your maker's pull temperature, stop: do not pull harder, and do not improvise temperature increments either. Recheck the model-specific procedure — temperature, filament, release step; if it still will not release, the pull is over, and Branches B, C and E, or your maker's support, take it from there.

The needle — a narrow tool with strict rules: an acupuncture-style needle has one legitimate job here: dislodging a particle lodged at the orifice of a partially blocked nozzle, from below, while hot. The makers bound it tightly, and the bounds are model-specific: use only a cleaning needle approved for the installed nozzle type and diameter, at the direction and depth your maker specifies. Prusa supplies a 0.3–0.35 mm needle with its printers and directs, for its own machines: insert it into the nozzle from below only, 1–2 cm deep; never feed it in from the top, and never load filament while the needle is inserted. Those figures are Prusa's prescription for Prusa's needle — and a smaller nozzle does not license an improvised thinner needle: a nozzle under 0.4 mm takes only a needle its maker approves for it, or none. Bambu Lab's unclogging procedure adds the depth discipline from the other side — work the pin through the tip a few times, but do not push it too far in, as that can push molten filament higher and manufacture a harder clog — and skips the pin entirely for 0.2 mm nozzles. Both makers put heat-resistant gloves and a face kept clear on the step (Bambu Lab's wording): pressure built up behind a clog can make molten filament spray when it lets go. What the needle is not: a reamer. It must be smaller than the orifice it enters — that size gap is the whole safety margin — and a needle that has to be forced is the wrong tool, withdrawn, not pushed. On a Nozzle X the door is closed to needles as well as pulls — Prusa excludes both: cleaning filament only.

The full-clog path (filament will not load, nothing extrudes): a cold pull cannot start — there is no continuous strand for it to grab (Prusa: a cold pull can usually only be used if the filament is able to reach the nozzle). What follows is fenced, and the fence is part of the procedure: this heat soak and force-out is Prusa's directed route for the MK3-family machines covered by its clogged-nozzle article (linked in Sources) — follow it only on those models. Do not use these temperatures on any other printer unless its maker explicitly authorises them, and never exceed your maker's stated temperature limit for the installed hotend. If the printer reports a heater or temperature-sensor fault, stop — that is a stop-work condition, not a clearing problem. On those Prusa machines: unload what can be unloaded, set the nozzle to 260 °C for PLA or 280 °C for PETG/ABS (Prusa's figures, for its machines), wait two minutes for the plug to soften and begin to drip, work the bundled needle from below as above, then — only if needed — the advanced, last-resort step: a straight, rigid 1.5 mm wire down the path to feel how deep the plug starts and push softened material through, with Prusa's own caution to keep the wire off the drive gears and the PTFE tube. How far the wire travels before it stops is itself the location evidence: a plug that starts high is Branch C's, whatever the print looked like. Finish by loading PLA and purging until the stream runs clean. Check: the strand test runs straight and even, and the wire's story matches the branch you treated.

How you know it worked: re-run the strand test — straight and even — the hand-push flows with normal resistance, and the reprint's lines come back to full width with no setting changed. If two consecutive pulls come out clean and full-shaped yet the strand still misbehaves, take the same evidence to Branches B and C — without treating the nozzle as definitively cleared: an intermittent or flow-dependent restriction can survive clean pulls and still be sitting in it.

Branch B: The Plug Is Above the Nozzle — Heat Creep

Signs: the Question 4 time-signature — clean printing for a predictable stretch, then thinning, clicking, sometimes a full stall; a cool-down and restart prints well again for about the same stretch; the unloaded tip swollen or mushroomed. Prusa's Knowledge Base describes the mechanism: parts of the hotend above the heater block get too hot, filament starts to soften higher inside the hotend, and the softened, swelling strand becomes the clog. The same article adds the honest caveat this branch is built on: heat creep is only one of the ways a hotend clogs, and often not the most probable explanation — which is why it is tested, not assumed.

The confirming tests:

The restart test: cool the printer fully and restart the same print untouched. Fading again at roughly the same duration or layer height is consistent with a time- or heat-dependent fault — not proof of heat creep. Rivals wear the same timing: retraction-dense zones, spool drag, a heating extruder motor. The checks below separate them.
The fan check: with the hotend at printing temperature, confirm the hotend (heatsink) fan actually spins — Prusa calls a working heatsink fan the first check, and names insufficient airflow among heat creep's common causes. A fan that never starts, rattles, or is choked with dust is evidence with a name on it. (Prusa publishes an expected fan-speed range for its MK3's fan readout — 4,000–4,400 RPM in Extruder info; that figure is that model line's, not a standard.)
The tip read: unload mid-fault and inspect the tip before it cools. A swollen, mushroomed or rough-surfaced tip — wider than the strand that feeds it — supports heat creep when it arrives with this branch's failure history; on its own it is not decisive, because a tip can also deform while being unloaded against resistance — it is the combination, swollen tip plus the restart pattern, that makes the case. A clean taper with debris at its point sends you back to Branch A.

Fixes, in order:

Restore the heatsink's cooling. Clean the fan and blow the dust out of the heatsink fins (Prusa: accumulated dust measurably reduces the heatsink's effectiveness); replace a fan that does not run at temperature. Mind the room and the box: Prusa lists ambient temperatures above 35 °C (30 °C for some filaments) and closed, unventilated enclosures among the common causes, and warns that PLA — the lowest-melting common filament — in a fully closed enclosure is very likely to suffer. Print PLA with the enclosure open where your maker advises it. Check: the print passes its previous failure point at full thickness, in the enclosure state you actually print in.
Restore the thermal path. If the hotend has been apart, two assembly details decide this branch: the PTFE tube fully inserted and secured so its end sits flush at the heatbreak (Prusa's heat-creep article makes this a named check; Branch C covers the tube itself), and — where your maker specifies it for that assembly — thermal paste on the heatbreak thread where it meets the heatsink, if yours came apart: Prusa directs exactly this for its named reassemblies, with ordinary CPU-grade paste, but it is a maker-specified step for those hotends, not a universal one — if your maker's procedure does not call for paste, do not add it. Check: the fade point recedes or vanishes across two prints of the previously failing job.
Reduce the heat load from the print itself. Filament sitting nearly still in a hot hotend absorbs heat (Prusa's mechanism), so low-flow printing invites creep: the maker's own test levers are thicker layers (0.15–0.20 mm in Prusa's example), a modest speed increase (Prusa suggests testing +10 % via the printer's knob first), and a heatbed lowered by 5–10 °C where rising bed heat feeds the hotend. Heavy retraction over one small area pumps heat upward the same way (labelled community practice: reduce retraction frequency there, one test print at a time). Unload the filament while the hotend is still hot at a print's end — Prusa's stated good practice — instead of leaving it to soak. Check: each lever is tested singly; the one that moves the fade point is the one you keep.

If the fan runs, the tube seats, the thermal path is sound, and the fade still arrives on schedule, the fault sits beyond cleaning-and-settings territory: stop, and take the restart-pattern evidence to your maker's support.

Branch C: The Plug Is in the Path — Tube, Coupler, Swollen Strand

Signs: filament will not load, or loads to a hard stop well above the nozzle; a hand push meets a wall before the melt zone; trouble after a tube was refitted, a coupler disturbed, or on a Bowden machine with hours on its tube. Prusa's clogged-hotend article is explicit that the jam may live in the PTFE tube inside the heatsink — and that an unsecured tube that can move is itself a clog cause. This branch stays deliberately tight: the extruder-side half of any path complaint (grinding gears, crushed strand, feed calibration) belongs to our under-extrusion guide, and is not repeated here.

The confirming tests:

The stop-point test: with the hotend cold and the filament unloaded, feed a fresh, straight-cut strand down the path by hand and note where it stops. Stopping at a consistent point well above the nozzle — roughly where the tube ends or the coupler sits — locates the obstruction in the path rather than the melt zone. (On Prusa's force-out route in Branch A, the 1.5 mm wire tells the same story from the other side.)
The tube inspection, done cold: remove the tube by your maker's method and inspect both ends. A burnt, shrunk, melted-back or pinched hot end narrows the passage exactly where filament is warmest and softest; a tube that was never pushed fully home leaves a ledge where a plug forms. Filament found stuck inside the removed tube is the address, confirmed.
The strand read: measure the filament itself (our under-extrusion guide's caliper check): an oversized or oval section can wedge in the tube — Bambu Lab's Wiki lists inconsistent, oversize diameter among its clog causes. A strand that measures in-spec makes oversize filament less likely at the measured location — measure more than one spot along the strand before the filament is fully cleared; only then does the suspicion settle on the path.

Fixes, in order:

Repair the damaged tube end — only the way your maker approves. Where the manufacturer publishes a trimming procedure, a heat-damaged end is cut square exactly as it specifies; where it does not — some tubes are made to a factory chamfer and length that a hand cut cannot reproduce — fit the maker's pre-cut replacement part instead. A nibbled or angled cut recreates the ledge either way. Check: a cold strand slides the tube's full length by hand with even, light resistance.
Refit the tube fully home and locked. Prusa's securing detail for its machines: push the tube down while pulling the collet up about 1 mm, so the tube cannot move inside the hotend — a tube free to ride up and down gaps at the heatbreak and manufactures Branch B's symptoms as well as this branch's. A coupler whose teeth no longer grip is a replacement, not an adjustment. Check: the tube does not move under a firm hand pull, and the stop-point test now runs to the nozzle.
Clear a swollen-tip jam at its cause. If the strand's own swollen end is what wedges (a heat-creep souvenir), cut the damaged section away before reloading — and treat the swelling as Branch B evidence, or the new strand swells in its turn. Check: loading runs end-to-end, and the next unload's tip comes out a clean taper.

If the path is clear cold, the tube sound and locked, and filament still stops at temperature, the obstruction is in the melt zone after all — return to Branch A with the path's innocence established.

Branch D: It Is Not a Clog at All — the Look-Alikes

Signs: the clog toolkit has been honestly worked — pulls come out clean and full-shaped, the strand runs straight — yet the print still starves or stalls; or the hand-push flows freely while the machine cannot feed. Clean pulls and a straight strand reduce suspicion of a persistent nozzle obstruction — but neither proves unrestricted flow under print load, so believe them as far as they go, and look at the three impostors.

The impostors, with their tests:

The extruder is losing, not the nozzle. Clicking alongside a freely-flowing hand push moves the suspicion to the gear — it is failing to move filament the melt path accepts: packed gear teeth, an idler off its maker's setting, or a strand ground flat by earlier clicking. That is feed-side territory with its own tests — the 100 mm feed test and fixes live in our under-extrusion guide (its Branch B); work them there. One clog-specific chore belongs here, from Prusa's clogged-hotend article: after any real clog, clean the drive gears — grinding during the clog packs them with residue that reduces their grip afterwards, so yesterday's clog can impersonate today's.
The nozzle is capped from outside. Nothing extruded for the first layer or two, then normal printing as the head rises: the exit was blocked by the bed itself — Simplify3D's too-close start, remedied by nozzle height, not by cleaning. Our first-layer guide covers the height checks. An over-squished first layer can also choke flow and click the extruder for the opening minutes of every print — same address.
The filament is wet, not blocked. Popping and hissing, a pitted surface, a stream that spits rather than flows — moisture flashing to steam disrupts extrusion without obstructing it, and no pull fixes it. Weigh the full evidence (storage history, material, a comparison print on a known-dry spool) as our under-extrusion guide's Branch C sets out; if the spool is convicted, dry it at its maker's numbers — our PLA drying guide compares them. Check for all three: the named test moves its named sign — feed test moves full length, first layer lays connected beads, purge runs silent — with the nozzle untouched.

Branch E: The Nozzle Itself Is Finished — Replace It Properly

Signs: clogs that return immediately after thorough, verified cleaning; a strand that deflects even with clean pulls (a damaged or worn orifice bends the stream it should drop straight); a long history of abrasive filament through a brass nozzle. Prusa's clogged-hotend article sets the threshold honestly: if a nozzle has seen a lot of use, or is old, it may be better to simply replace it — and its nozzle guide adds the counterweight: unless abrasive materials were printed, a stock brass nozzle lasts many hundreds of printing hours, and changing a relatively new nozzle to resolve a clog is not the solution. Replacement is this branch's fix, not the guide's first reflex.

The confirming evidence: Branch A's pulls come out clean yet the fault is unchanged; the removed nozzle, inspected cold against the light, shows an orifice that is oval, enlarged, chipped or visibly scored; or the service history (abrasive hours on brass) makes wear the parsimonious explanation. One more inspection belongs here, done cold: filament leaked or baked around the heater block or the heatbreak joint — plastic escaping anywhere except the orifice is an assembly fault (a badly seated nozzle or heatbreak), and a leaking joint both causes clogs and mimics them; no cleaning round fixes it — it is corrected by reseating or replacing parts per the maker's procedure, which is this branch's territory. A nozzle that cleans up and prints clean was never this branch — re-check before spending.

The replacement — by your maker's procedure: there is no universal nozzle-change sequence. On Bambu Lab's X1/P1 the nozzle is part of the hotend assembly that is swapped as a unit; Prusa's Nextruder machines use their own nozzle system and procedure; every other assembly follows its own manufacturer's manual. The numbered sequence below is one worked example only — Prusa's procedure for its V6-based threaded-nozzle machines (MK3/S/+, MK2.5/S, MK3.5/S) — and its figures (the 280 °C heating step, the 2.5 N·m service torque, the ≈0.5 mm gap) belong to that example and those machines, not to threaded nozzles in general. Your maker's manual outranks this outline at every step:

Heat first. Unload the filament (a cold pull first is Prusa's optional step), then heat the nozzle — Prusa's Nozzle Change menu takes its machines to 280 °C. Heating is essential to the process: the nozzle must be tightened and locked against the heatbreak while heated, or the joint leaks.
Hold the block; mind the wires. Counter-hold the heater block with a spanner (17 mm on Prusa's machines) so it cannot twist, keeping clear of the fragile heater and thermistor wires — Prusa warns the spanner can break them off or short the heater.
Off with tools, onto a safe surface. Unscrew with pliers or the maker's socket — the nozzle is at full temperature; fingers never go on it, and it comes to rest on a non-flammable surface, not the bench, not the bed. Handling, threading by hand and inspection happen only once parts have cooled — a nozzle you need to examine is examined cold. The cold/hot split across this sequence is deliberate, not a contradiction: examination, handling and initial thread engagement belong to cooled parts, while the final tightening in step 4 is done hot on purpose, because the joint seals against the heatbreak at temperature — cold for hands and eyes, hot for the final lock, in the order the maker's sequence prescribes.
Fit and hot-tighten the new nozzle. Screw it in and tighten gently but firmly at temperature — Prusa's service teams use a torque wrench at 2.5 N·m on its V6 machines; use your maker's figure where one exists — and verify the assembly detail Prusa insists on: a gap of about 0.5 mm remains between the nozzle's hex and the heater block, because the seal is made against the heatbreak inside, not against the block's face. No gap means the joint is wrong, not tight.
Recalibrate. A new nozzle changes the first layer: Prusa directs redoing at least the First Layer Calibration after a change (full calibration if the extruder was dismantled). Check: the strand test runs straight, the first layer calibrates in range, and the previously failing print completes with no leak at the block.

Choosing the replacement: like-for-like diameter, and material matched to what you print. Brass is the standard for ordinary filaments; for abrasive and filled filaments, hardened steel is the makers' recommendation (Prusa's nozzle guide and Bambu Lab's hotend documentation agree) — the same particles that clog Branch A grind a brass bore open over time. If you fit a Nozzle X, its rule travels with it: cleaning filament only, no cold pulls, ever. And keep the old nozzle until the new one has printed: a "worn" nozzle that was actually innocent is your evidence that the fault is still in the machine.

About the Community Numbers in This Guide

The FixMyPrint dataset behind this guide is one analyst's scrape of 20,000+ help posts and roughly 60,000 fix suggestions from one of Reddit's largest 3D-printing help communities, collected over one year and classified by problem type, printer model, and whether the asker confirmed a fix. Clogging and heat creep are filed as one category of 764 posts. Read that number for exactly what it is: a count of community help posts in that analysis — not a measurement of how often nozzles clog across all printers, and the category's combined filing is itself this guide's point: the dataset's own author could not separate the nozzle plug from the heat-creep plug at the level of a help post, because askers usually cannot either. Our research notes for this guide record that post count but not a rank or a confirmed-fix rate for this category, so this guide quotes neither rather than printing numbers it cannot stand behind. That is why Branch A teaches the pull properly, with its makers' temperatures and exclusions, instead of waving at it: a pull done at the wrong temperature, or on the wrong nozzle, retrieves nothing — or does damage.

What NOT to Do

Do not cold-pull — or needle — a Nozzle X. Prusa's instruction is unqualified for both tools: cleaning filament only — the pull can damage that nozzle, and the needle is excluded with it. The same applies to any maker procedure that excludes your nozzle or hotend type; the gate in Branch A is read before the pull, not after.
Do not force a pull — and do not jerk one. The pull works inside its temperature window, in one firm, steady, straight motion. Escalating force against a stone-cold or unreleasing plug snaps the filament off inside the hotend (Prusa's warning) and turns a cleaning step into a teardown.
Do not use the needle from above, oversized, or deep. From below only (Prusa), smaller than the orifice, a few shallow passes (Bambu Lab); never with filament being loaded over it; never on a nozzle whose maker does not approve one — a Nozzle X included. A needle is a dislodger, not a drill.
Do not improvise temperatures above the spool's range to blast through a blockage. Overheated plastic degrades and carbonises into a harder clog, and the heat feeds heat creep (Branch B). The exception is a maker's own directed clearing procedure — Prusa's heat-soak force-out in Branch A is one — performed exactly as that maker specifies, and only on the models it specifies.
Do not swap the nozzle as the first response to every clog. A clogged hotend is not the same as a clogged nozzle (Prusa): if the plug sits in the tube, the new nozzle inherits the same blockage, and you have spent a part to learn nothing.
Do not disassemble the hotend hot, and do not handle a hot nozzle. Swaps follow the maker's hot-tightening procedure with tools and a counter-held block; handling, threading by hand and inspection wait for cold parts.
Do not change the filament, the temperature and the hardware in the same session and call the result a diagnosis. If it improves, you own three unverified changes; if it worsens, you cannot retreat. One change, one test, one verdict.

When to Stop and Ask for Help

If verified-clean pulls, a clear path (Branch C), sound cooling (Branch B) and a fresh nozzle (Branch E) still leave the same blockage returning, stop cleaning and gather the evidence you now own: which branch tests ran and what they showed, what the pull tips looked like, where the wire or the strand stopped, and the restart pattern if Branch B applied. Inspect the hotend cold, too, for filament leaked or baked around the heater block and the heatbreak joint: a leaking assembly is an assembly fault — it goes to the maker's seating procedure or to support, not to another cleaning round. A support ticket carrying locations and measurements gets a different conversation than "it keeps clogging". What lies beyond this guide's reach — a plug fused inside a heatbreak that your maker's procedure does not cover, damaged hotend threads, a heater or thermistor fault — belongs to the maker's manual and support.

Stop-work conditions — stop experimenting at once, whatever branch you are in: a burning smell, discoloured or heat-marked wires or connectors, a nozzle joint that leaks after a correct hot-tightening, or a heater that will not hold its set temperature. Power the printer off and contact support. In particular, never answer a suspected heater fault by raising temperatures further.

FAQ

Q: How often should I do a cold pull?
A: On evidence, not on a calendar: when the strand test curls or thins (safety gate first), and when stepping down from a hot material to a cooler one, so residue does not become tomorrow's Branch A plug. Prusa also lists the pull among its regular-maintenance techniques — but every pull is a procedure with a temperature window and a nozzle exclusion (Nozzle X), not a nervous habit.

Q: My cold pulls come out clean, but it still clogs. Why?
A: Because the pull only searches the melt zone. A plug that keeps forming above it is heat creep (Branch B); a passage narrowed above it is Branch C; a restriction that only bites at printing flow rates never shows at pull rates at all (the baseline's flow-demand check); and a fault that was never a plug is Branch D. The clean pull is not a failed test — it is evidence that relocates the search, without definitively clearing the nozzle itself.

Q: The clog appeared right after I switched from PETG to PLA. Coincidence?
A: Probably not — a material change makes retained higher-temperature residue a plausible cause, and it is Prusa's named mechanism for exactly this timing: residue of the hotter material remains in the nozzle, and the cooler material's temperatures cannot fully melt it out. Prevention, from the same source: load the cooler filament at the previous filament's temperature and purge, then reload at its own temperature. A cold pull after a downward material change is the standing cleanup.

Q: Is the needle safe to use?
A: Within its narrow brief, yes — and the brief is the safety. The figures are each maker's own: Prusa, for its bundled needle, specifies from below only, 0.3–0.35 mm for a 0.4 mm nozzle, 1–2 cm deep, never from the top, never while loading filament. Bambu Lab: a few passes through the tip, never pushed far enough to shove molten filament upward, skipped entirely on 0.2 mm nozzles, with gloves on and your face clear. Any other machine: only a needle your maker approves for your nozzle type and diameter, at the direction and depth it specifies — a thinner nozzle never licenses an improvised thinner needle. Outside those rules — or on a Nozzle X, where Prusa excludes needles as well as pulls and allows cleaning filament only — the answer is no.

Q: Can I drill or ream a clogged nozzle back to size?
A: No procedure in the manufacturer documentation behind this guide does that, and the reason is structural: the orifice's diameter and shape are what the printer's line width is calculated from, so a reamed nozzle is an uncalibrated one even if plastic flows again. Cleaning is pulls, purges and the bounded needle; a nozzle whose orifice is damaged or worn is Branch E — a replacement, fitted properly.

Summary: The Five Branches at a Glance

Diagnose before you heat: read the failure (Step 0), run the four triage questions, restore the baseline — then work the branch the plug's location points to:

Branch	Lead signal	Confirming test	First fix
A — Plug in the nozzle	Curling/thin strand; hard hand-push; after a material change	Strand test; read the cold-pull tip	Cold pull by your maker's method (safety gate first); bounded needle
B — Heat creep	Fades on a schedule; swollen unloaded tip; hot enclosure	Restart pattern; heatsink fan check; tip read	Restore cooling; thermal path; reduce heat load
C — Plug in the path	Filament stops well above the nozzle, cold or hot	Stop-point test; tube inspection (cold)	Repair tube end per maker's method; refit fully home and locked
D — Not a clog	Clean pulls, straight strand — or a free-flowing hand push	Feed test; first-layer pattern; moisture evidence	Work the impostor's own guide (feed, first layer, drying)
E — Nozzle finished	Clogs return after verified cleaning; abrasive history	Clean pulls + unchanged fault; cold orifice inspection	Replace by the maker's hot procedure; recalibrate

Every fix ends with its verification — if the named sign did not move, that was not your cause.

Honesty note: Compiled from manufacturer and community documentation — not personally tested on a printer in our workshop, and it does not pretend otherwise. Figures are labelled as such, and dataset figures carry their denominators where the research records them — where a figure is not recorded (this category's rank and confirmation rate), the guide says so instead of quoting one. Where the evidence ran out — a universal pull temperature, a universal needle rule — the guide says so instead of inventing a number.

Sources: FixMyPrint — dataset analysis (clogging and heat creep: 764 posts of 20,000+); Prusa Knowledge Base — Clogged nozzle/hotend (MK3.5/S, MK3S+, MK2.5S); Prusa Knowledge Base — Cold pull (MK3/S/+, MK2.5/S, MK3.5/S); Prusa Knowledge Base — Cold pull (MINI); Prusa Knowledge Base — Cold pull (CORE One); Prusa Knowledge Base — Extrusion stopped mid-print (heat creep); Prusa Knowledge Base — Changing or replacing the nozzle (MK2.5/S, MK3/S/+, MK3.5/S); Bambu Lab Wiki — Nozzle/Hotend Unclogging Procedure for X1/P1; Simplify3D — Print Quality Troubleshooting Guide (Clogged Extruder). All source pages opened and read on 2026-10-30.

Search This Blog