Elegoo Neptune 4 Pro: 28 Known Issues and Fixes

The Elegoo Neptune 4 Pro is a 2023 open-frame FDM printer with a 225×225×265 mm build volume, a 300 °C hotend, a 110 °C bed and Klipper running from the factory on a 64-bit board. What sets it apart from the plain Neptune 4 is the motion system — metal U-wheels riding steel guide rails on X and Y — and a segmented heatbed whose 120×120 mm centre zone heats separately from the perimeter. It is rated at 310 W and 500 mm/s at 20,000 mm/s², though Elegoo's own spec sheet recommends 250 mm/s.

Below are 28 problems specific to this machine — from two incompatible hardware revisions and a bed mesh profile with a bizarre name, to a hidden extruder tension screw that appears in no manual anywhere. Generic FDM troubles like stringing, warping and clogs live in their own guides, linked at the end. If you own the plain Neptune 4 without the Pro badge, some items overlap but the motion system is different: see our Elegoo Neptune 4 known issues.

1. Two Neptune 4 Pro revisions — ribbon cable and Type-C — and firmware does not cross over

The firmware you downloaded refuses to install, or the printer will not boot after the update. Sometimes it goes the other way: a machine delivered in 2026 already runs a build newer than the one listed in the main download column. The Neptune 4 series ships in two hardware versions, told apart by the printhead cable: the Ribbon-Cable Version and the Type-C Version, also known as the Neptune 4 Pro+.

They take different printhead parts and two incompatible firmware branches. Elegoo's own PDF says it plainly: “These two firmware versions are not compatible. Installing the wrong firmware will cause the printer to fail to start up.” The ribbon revision tops out at V1.1.3.3 with UI 1.2.18, dated 8 September 2025. Type-C has its own line: V1.3.1.2 on 28 September 2025, V1.3.1.3 on 25 November 2025 and the current V1.3.1.4 from 3 April 2026.

1. Look at the cable feeding the printhead: a flat ribbon with a horn-shaped connector is the old revision, a round cable ending in USB-C is the new one.
2. Second tell: the trace code in the lower left corner of the base. A plus after the model name means the newer machine.
3. Confirm the model from the serial on the white sticker at the bottom left of the bed — a Pro starts with NEP 4 PRO.
4. Download firmware strictly from your own column on the Elegoo download page. On Type-C machines you can only move to V1.3.1.4 from V1.3.1.3; earlier builds are incompatible.
5. Never apply the fix pack to a Type-C machine. Elegoo's own instructions carry the warning: “Do not update the Neptune 4+/Neptune 4 Pro+ with the Fix Pack, this may brick the printer.”

2. Update order: fix pack first, then the mainboard, then the screen

Elegoo still ships machines on old firmware, and nothing in the box explains the migration path. People drop the newest package straight onto the factory build and end up with a screen frozen on the logo, or an About Machine menu frozen solid — it registers touches and does nothing. The order is not optional: fix pack, then mainboard firmware, then the screen firmware separately.

The second trap sits inside the archive. For Neptune 4 and 4 Pro it holds two folders, 0.8A and 1.2A, for two different stepper motors: BJ42D15-26V77 draws 0.8 A in a 34 mm body, BJ42D22-53V04 draws 1.2 A in a 40 mm body. Only the 1.2A package is published for the Neptune 4 Pro, but checking your motor is still worth the minute — flashing the wrong folder is a classic way to brick a machine.

1. Use a drive under 16 GB, formatted FAT32, and make sure it is USB 2.0 — plenty of USB 3.0 sticks simply are not read. Screen firmware needs a microSD in FAT32 with a 4096-byte allocation unit and under 32 GB.
2. Check that more than 500 MB is free on the printer's own storage; Fluidd shows it. The fix pack wipes the job list, so copy anything you want to keep first.
3. Drop the ELEGOO_UPDATE_DIR folder into the root of the drive without renaming it. Power off, insert, power on, wait about two minutes.
4. Success is visible in files: the fix pack creates fix_klipper_succssed and fix_tag on the drive, and the mainboard update creates update.log.
5. The screen flashes on its own: .tft file in the card root, four 2.0 mm screws on the back of the screen, card in, power on, roughly 90 seconds to 100 %.
6. Afterwards, re-level the bed, reset the Z offset and re-run vibration tuning. The update rewrites the configuration and wipes every calibration — that is expected behaviour, not a fault.

3. The Update button in Fluidd bricks the printer

The web dashboard offers to update Klipper and Moonraker on its own. Pressing that button ends the same way every time: the printer will not start, Fluidd shows “Internal error during connect: CommandWrapper instance has no attribute '_get_command_tag'” and the configuration is scrambled. One owner talked support into shipping a replacement memory module; another was simply left with a dead machine.

The reason is that Elegoo ships its own Klipper fork with no published sources. A Klipper maintainer put it bluntly: “Elegoo seems to be shipping a modified Klipper version; I did not find any released sources for these modifications → Elegoo is in violation of the GPL... Support from the official Klipper project cannot be offered.” In practice, the Klipper forum will send you back to Elegoo.

1. If you already pressed it: connect the printer by Ethernet to a router with internet access — some recovery steps fail on a direct laptop link.
2. Apply the fix pack, then the mainboard firmware, then the screen firmware, in that order.
3. If it still will not boot, the remaining path is rewriting the eMMC image: an adapter, pulling the module off the board and flashing an image with a card writer.
4. Going forward: a clean Klipper on this machine only ever comes from replacing the whole image, never from a button in the browser.

4. The eMMC module dies and the printer never gets past the splash

Boot drags on for about three minutes, the screen then parks in About Machine with a blank firmware version, no IP is handed out and Fluidd is unreachable. The fix pack appears to run and changes nothing. Occasionally the menu even shows the wrong model. The printer's operating system lives on a removable eMMC module, and this is that module failing.

1. Rule out a short in the accessories first: power down, unplug the USB drive, the filament runout sensor and the gantry LED, then power up. If the menu appears, plug them back one at a time.
2. No luck — run the fix pack from a FAT32 drive under 16 GB with everything else deleted from it.
3. Next step is a fresh image: five screws on the bottom cover, unplug the internal board-fan connector, two screws holding the eMMC module, then flash the image through a USB eMMC adapter.
4. Many owners fit a 32 GB module at this point and keep the original untouched as a fallback. Sensible habit either way: image your working system once it is dialled in.
5. If the machine is under a month old, contact support — they ship replacement eMMC modules under warranty.

5. The Z offset “will not save”: Elegoo renamed the mesh profile to “6”

This one is sneaky. The Z offset does save, and after a reboot the field still shows the same number — say −1.530 — yet the real height is different. Yesterday −1.530 gave a perfect first layer; today it needs −1.100. Meanwhile the Klipper console flashes a message that profile “6” cannot be found.

In Elegoo's configuration the mesh profile is not called default, as every Klipper guide assumes, but “6” — and both autosave and autoload look for that exact name. A profile saved under your own name is simply never picked up. The name is baked into the update files from Elegoo's site too, so editing config files by hand does not help.

1. Run a fresh bed mesh.
2. When it asks for a profile name, type exactly 6 — the digit, no quotes, no spaces.
3. Reboot and confirm the console no longer complains about a missing profile.
4. Separately, calibrate the probe itself rather than the gcode Z offset: PROBE_CALIBRATE plus TESTZ, per the Klipper documentation. After that the on-screen Z offset is only a per-material nudge.

6. The mesh gets measured but never loaded — no slicer puts the command in

All 121 points probe, the map looks lovely on the screen and in Fluidd, and the first layer behaves as if nothing had been levelled: squashed in one corner, printing air in another. People blame the bed for months. The actual cause is that no start-gcode profile — including Elegoo's own bundled Cura — contains a command to load the saved mesh, and the Marlin habit of M420 does not exist in Klipper.

The other half of the same problem is coverage. The probe sits −24.25 mm on X and +20.45 mm on Y from the nozzle, so the factory mesh area in printer.cfg is mesh_min 10,21 to mesh_max 210,211 on a 225×225 mm bed. The edges are never measured, only extrapolated — owners notice probing starts roughly 30 mm in from the left edge.

1. Add two lines to your slicer's start gcode, after homing and before the prime line: BED_MESH_CLEAR and BED_MESH_PROFILE LOAD=6 (the profile name from the previous section).
2. Check which mode is probing: on the Level screen, the button in the top right toggles standard 36 points and professional 121 points, and you have to hit save afterwards.
3. Keep parts away from the very corners of the bed — those values are extrapolated, not measured. Widen mesh_min and mesh_max in printer.cfg through Fluidd if you need them.
4. Leave adaptive bed mesh off on this machine: owners report 0.2100 variance against 0.1340 from a plain 11×11 probe.

7. Z offset drifts: the probe maps the bed at 140 °C while you print at 220 °C

The saved offset does not match reality when the print starts: the nozzle either scrapes or floats. Re-probe right after saving and you get a different number. The inductive probe is precise — repeatability around 0.002 mm — but its reading moves with temperature, and the stock calibration routine only heats the nozzle to 140 °C and the bed to 60 °C.

One reviewer measured it: between a 140 °C nozzle and a 220 °C nozzle over a stable 60 °C bed, the probe reads 0.03 mm apart, which lands as 0.05–0.1 mm of first-layer error. The factory probe settings do not help either — samples 2, samples_tolerance 0.1, speed 10.0 — while Klipper's own documentation calls anything above 0.025 mm of spread insufficient.

1. Heat-soak the bed and nozzle for 15–20 minutes before calibrating — the thermistor hitting target does not mean the whole plate has.
2. Calibrate at the temperatures you actually print at, not at the stock 140 °C.
3. In printer.cfg through Fluidd raise samples to 3–4, drop samples_tolerance to 0.025 or lower and cut speed from 10 to 5 mm/s.
4. Check fade_end: some configurations ship with 30.0 where the Klipper docs suggest no more than 10.
5. Clear the play first: rock the head and bed by hand, verify the X gantry is parallel to the bed, and leave the screws holding the brass lead screw nuts slightly loose — tightened down, they bind the Z axis.

8. Do not level from the touchscreen and the web dashboard at once

Offsets wander, and the first layer alternates between floating and gouging the plate coating. The Elegoo screen and Fluidd write offsets to different places in the configuration and the two stack on top of each other. The advice from forum veterans is one line long: level from the screen only, or from the web only.

1. Pick one and stick to it. Not even a quick one-off touch-up on the other side.
2. Going the web route: probe the mesh, run PROBE_CALIBRATE, save the config. Klipper saves nothing on its own — every change needs SAVE_CONFIG.
3. Put the mesh-load and Z-offset lines in your slicer start gcode explicitly; they do not always load by themselves.
4. On newer firmware the Z offset cannot be written straight into printer.cfg anyway — Elegoo's service scripts zero it on reboot. One more reason to keep a single workflow.

9. The probe only sees metal: a melted hole in the sticker and a tilted sensor

This is an inductive proximity switch, not a strain gauge: it responds to the metal plate, not to contact. Elegoo states it directly — “Proximity switch can only recognize metal and cannot recognize other materials. Do not use the level and home functions without attaching the PEI sheet.” Home without the plate and the nozzle drives down to the mechanical limit, melting a crater into the magnetic sticker.

1. Print on the supplied steel plate only. Glass and mirror are invisible to the probe, and 4 mm glass under the plate also heats poorly.
2. If the head slams into the bed while homing, first confirm the nozzle is the stock one — a third-party nozzle breaks probe triggering, and that is Elegoo's own listed cause.
3. Inspect the ribbon pins on the printhead; a single bent pin produces the same symptom.
4. Check the probe itself: the small body behind and left of the extruder shifts in its mount, especially after the head cover has been off. Loosen the screw, set it perfectly vertical at stock height relative to the nozzle, tighten.
5. Reseat the probe connector on the mainboard under the bottom cover.
6. Verify the mainboard and screen firmware are a matched pair — a mismatch produces similar homing behaviour.

10. The hidden extruder tension screw nobody documents

The extruder clicks, the drive gear slips, feeding stops in patches — it looks and sounds exactly like a clogged nozzle, except the nozzle is clean. Walls and supports come out weak and under-extruded. The spring pressure on the drive gear is set arbitrarily at the factory, and it is a lottery: some machines arrive fine, others barely bite the filament at all.

There is an adjustment, it just appears in no manual and no wiki page. On the right side of the printhead housing there is an unmarked blind hole with a 2 mm hex screw inside. Counterclockwise increases the grip, clockwise reduces it. Intuition says otherwise, which is why so many people turn it the wrong way.

1. Find the hole on the right side of the printhead and insert a 2 mm hex key.
2. Turn counterclockwise. Two to four full turns is typical; owners in bad cases have gone as far as eight.
3. Aim for teeth marks biting roughly 10–15 % into the filament. Over-tightening backfires on flexibles: TPU buckles and jams, so for TPU you loosen instead.
4. No disassembly needed. But if you find yourself retightening every month, the screw is not the problem — the bearing on the end of the main drive shaft has failed and the gearbox needs replacing.
5. While you are there, check gear backlash: pull the filament, release the idler and turn the gear with a finger. Free play before the drive engages means backlash — loosen the three screws holding the feeder to the drive, shift it towards the drive shaft and retighten. The giveaway for this one is a double click on every retraction.

11. The stock rotation_distance is off, and that is your under-extrusion

Holes in walls, thin infill, translucent top layers — with sound mechanics, correct tension and a clean nozzle. The factory printer.cfg sets the extruder rotation_distance to 31.4, which is an average. The true figure varies machine to machine thanks to tolerances in the 5.2:1 gearbox.

Owners who bothered to calibrate report noticeably different numbers: 28.1947 and 28.73 on a Neptune 4 Pro, 27.7803 on a 4 Plus, 28.88 after correcting the factory 31.4. The gap between 31.4 and 28.7 is roughly 9 % of missing plastic, and no amount of flow tweaking in the slicer covers that honestly.

1. Heat the nozzle slightly above your filament's normal temperature and confirm the tension is holding — you should see teeth marks on the filament.
2. Measure and mark 100 mm of filament above the feeder entry.
3. Extrude exactly 100 mm slowly, at 1–5 mm/s: faster builds back pressure and skews the measurement.
4. Measure what actually went in, recalculate rotation_distance and enter the new value through Fluidd.
5. A result near 26.6 means you mismeasured — check the mark and the speed and try again.

12. The Pro's metal U-wheels grind to dust and howl without grease

After the first few prints a fine dark dust appears on the Y rails, sometimes rust-coloured, and the carriage starts to howl. One owner described it as “sounded like an animal was dying as it was going even at half speed”. This is the Pro's defining hardware difference: metal U-wheels on steel guide rails for X and Y, instead of POM wheels running on aluminium extrusion.

Elegoo's wiki states it outright: “The Neptune 4 Pro uses a dual-axis metal guide rail with metal U-shaped wheels.” That assembly needs grease that POM wheels never did — and it is a maintenance schedule, not a suggestion.

1. Power down. With a paper towel, wipe burrs and old grease off the metal U-wheels and both upper and lower X rails. Hold the towel against the wheel and roll the head back and forth so the whole wheel gets cleaned.
2. Apply grease to the wheels and rails, then run the head across a few times to spread it.
3. Repeat for the Y axis by moving the bed.
4. Wipe and grease both Z lead screws, power on and home the machine.
5. Elegoo's own intervals disagree: the general maintenance page says monthly, the U-wheel page says every two weeks. Go by noise and by the film on the rails.
6. Keep the printer somewhere dry — rust-coloured dust on the rails is exactly what it looks like.

13. The X carriage eccentric tightens nothing: the holes were drilled coaxial

The X carriage wheels refuse to preload against the uprights: the eccentric turns and nothing happens, the gantry has play and walls come out uneven. In the front and rear plates of the left carriage half, the wheel axle holes are coaxial, so the axle screw cannot shift and the eccentric does nothing. The flaw goes back to the Ender 3 lineage.

There is an opposite failure too. Over-tightened eccentrics show up in the bed mesh as regular ripples — every other row of points sitting lower than its neighbour, roughly 0.25 mm of variance while all four corners agree within 0.1 mm. Rotating the plate 180° changes nothing, because it is wheel runout, not the plate.

1. Try Elegoo's own procedure first: unplug the rear fan bar and remove it (three screws), then take off the extruder — four screws at the rear and two at the sides, all 2.0 mm hex, disconnecting the ribbon, then make small eccentric adjustments until the carriage stops wobbling but still glides.
2. If the eccentric spins with no effect, file the round axle hole into a slot, shielding neighbouring parts from filings.
3. Set every wheel to touch the extrusion with the same pressure. The rule of thumb: an eccentric nut should never spin freely, you should feel resistance.
4. If the mesh shows regular ripples, go the other way and back the eccentrics off slightly, both on the head and under the bed.
5. On the Neptune 4 Pro the bed wheels cannot be reached without stripping the bed: plate off, four knobs out, belt loosened and detached, bed carriage removed.
6. One officially acknowledged tolerance: of the six wheels on both sides of the X axis, one usually never tightens. Elegoo says the effect on operation and quality is minimal — keep the rest snug.

14. The gantry ships skewed: ovals instead of circles and a first layer that never lands

PLA prints fine while PETG refuses to lay a first layer at all; circles come out oval and mating parts do not fit. One owner fought it for a year and was about to order a replacement bed — the X beam was sitting a couple of millimetres higher on the left than on the right, and the bed itself was perfectly flat.

The same defect has a second flavour: the gantry upright screws tapped their own thread into the extrusion slot instead of landing in the proper holes, leaving one upright 3–4 mm further back than the other. From outside the machine looks correctly assembled. On top of that, some units arrive with the Z-axis mounts over-tightened, and the axis starts to bind.

1. Before touching bed levelling, measure from the X beam down to the bed on both sides. A difference means mechanics, not mesh.
2. Elegoo's routine: place two identical spacers under the beam, loosen the X mounting screws and the Z pulley brackets, lower the beam onto the spacers by turning both Z couplers, then retighten.
3. Check that the gantry upright screws actually went into the intended holes and not into the extrusion slot. If they tapped the slot, take it apart and reseat them.
4. For a wobbly base after shipping, loosen without removing the four screws on each side of the base with a 4.0 mm hex key, retighten them diagonally, then shake the base by hand and repeat if needed.
5. Only after the mechanics are square: run SCREWS_TILT_CALCULATE instead of the paper trick, then auto level.

15. The bed knobs unscrew themselves, and foil shims break the probe

A few hours into printing the bed geometry drifts, and on long jobs you keep nudging Z. Reviewers had the levelling wheels fall off the machine during an afternoon of speed tests. Elegoo brought back a spring-mounted bed with knobs even though the Neptune 3 Pro used a fixed platform, and bed-slinger vibration at 250 mm/s works them loose.

The springs are the other half. Resonance measurements show the Y base plate producing a clean graph while the bed itself smears — owners summarise it as “the bed is a shaking mess”.

1. Fit lock nuts to the levelling screws, or print knob locks: on the popular model one notch equals two minutes of rotation, which doubles as a precise adjustment step.
2. Next step is swapping springs for silicone spacers, then for aluminium spacers with a silicone and fibre washer. Do not overdo it — too rigid and the knobs lose their authority.
3. Re-check level every time you move the printer.
4. Replace the paper trick with SCREWS_TILT_CALCULATE. The stock config has no such section, so the Neptune 4 Pro screw coordinates go in by hand. Turn the knobs as it tells you until deviations sit within one minute, and re-check every five prints or so.

16. The Y belt tails bow the bed, and the plate itself can be 0.66 mm wavy

The bed is bowed, centre and edges disagree and the mesh cannot save it. Before blaming the plate, look underneath: the Y belt tails stick up and press into the bottom of the heated bed once the levelling screws are cranked down. Fix it by folding or trimming the tails — or simply by stopping the over-tightening.

If the tails are innocent, it is the plate. One Neptune 4 Pro owner started at 0.6653 mm of mesh variance with sharp drops at the edges and got it down to 0.1638 mm with shims placed by the height map: about an hour and a half of work, three auto-calibrations, 583 cm of tape and a 36-point probe. The forum benchmark is 0.10–0.20 mm, with 0.15 considered good. Another brand-new machine reported 0.2937 mm in the interface against 0.5055 mm peak to peak.

1. Take the plate off and inspect the Y belt crimps. If they stick up, fold or trim the excess.
2. Tighten the screws between the bed and its carriage — Elegoo names those as a direct cause of levelling that will not hold.
3. Probe the mesh on a heat-soaked bed and read the variance in Fluidd.
4. Add steel shims strictly according to the height map, re-probing after every change. Kapton tape beats copper tape here: copper tends to ooze adhesive and collect debris over time.
5. Do not chase zero. A realistic target is 0.10–0.20 mm and the mesh handles the rest.

17. Belt tensioners: X floats in its holder, Y cracks, and Z has none at all

A regular wave pattern shows up on walls and refuses to go away with input shaper tuning or belt re-tensioning in either direction. One owner traced it: the X tensioner assembly is not fixed inside its holder, so the belt climbs the pulley flange and creates variable resistance. The stopgap is shimming the pulley so it sits perfectly in the plane of the gantry.

The Y axis has a different failure: the plastic tensioner cracks and falls apart, usually after over-tensioning. Owners report the printed replacement is “unbreakable by hand” — the factory part is simply moulded too thin. And the Z synchronising belt has no tensioner at all: two lead screws driven by one motor through a closed loop, with the screw tops unsupported.

1. Tension the belts until the slack is gone and no further — an over-tight X belt causes layer shifts a few hours into a print.
2. Print a tensioner lock for the X axis. Note that the part for the Neptune 3 and the plain Neptune 4 does not fit the Pro, whose assembly geometry differs — pick a model marked specifically for the 4 Pro.
3. Replace a cracked Y tensioner with a reinforced printed one; the original is not cheap and cracks again.
4. For Z, print a belt tensioner and a stabiliser for the top of the lead screws — that kills screw wobble and the Z banding it produces.
5. The heavy-duty option is giving the second screw its own motor: a TMC2209 driver with the UART jumper, a dedicated cable, removing the top bearings and the belt, and replacing homing with Z_TILT in start gcode plus the matching printer.cfg section.

18. The nozzle is custom, and a cold-tightened one slowly walks out

A generic Ender nozzle threads in, but either the thread bottoms out in the heat block early or the finished nozzle sits low and crashes into the bed. Measurements explain it: 18 mm overall length, M6×1 thread 9 mm long, 6×4.5 mm hex flats. The teardown puts it as close to RepRap style except the threaded length is 2 mm longer, the overall length 5.5 mm longer and the profile slightly different. Elegoo confirmed by email: “The Neptune 4/4pro uses custom nozzles and is not universal.”

The second failure is gentler but pricier: a nozzle tightened cold slowly creeps out of the block as it heats. One owner read it as “drifting Z” through several successful prints — and then the hotend died. Same family of problems: the hotend mounting screws vibrate loose, the head starts to rattle and no Z offset tweak helps.

1. Buy nozzles made for the Neptune 4 and 4 Pro, and match your revision — the kits ship as two separate part numbers for the ribbon and Type-C versions.
2. Quick visual check: a nozzle bottomed out should still leave about half a thread turn showing.
3. To swap it: unload filament, power down, remove the head's front cover with a 2 mm hex key, unscrew the fan block (two screws each side), disconnect the probe connector and fold the block up.
4. Heat the hotend to 200 °C and back the nozzle out with a 6 mm wrench while holding the brass block with a second wrench so it cannot rotate.
5. Always torque the new nozzle HOT. Then redo the Z offset, PID, pressure advance and retraction.
6. Every couple of months, rock the head by hand: any knocking means the hotend mounting screws need retightening.

19. There is PTFE inside that “300 °C” hotend, and heat creep hits at 70–90 %

Great first layer, great adhesion, clean surfaces — and then at 70–90 % of the job plastic starts blobbing, the nozzle ploughs into the part and destroys it. That is heat creep: heat climbing the heat break until a plug forms above it, jammed up against the PTFE tube.

Elegoo markets a copper-titanium bimetal heat break and 300 °C, but the hotend kit also contains a PTFE tube. The teardown phrases it carefully: “This may be misleading, as the extruder feeds a PTFE tube, which feeds the throat-pipe.” So 300 °C is the heater's rated ceiling, not a setting for a long nylon print.

1. Confirm the heat break fan spins whenever the nozzle is hot. If it has stalled, a plug forms within five to ten minutes of printing.
2. Do not start a second print straight after the first: cooling is off during the first layer for adhesion, and the heat break is still soaked from the last job.
3. Torque the nozzle hot — plastic weeping past a loose nozzle reads exactly like heat creep.
4. Pull the PTFE tube, release the collet and reseat it fully home; chamfer the entry end with a knife so filament does not catch the lip. Treat the tube as a consumable — Elegoo documents the replacement procedure.
5. Check the filament separately: oversized filament jams BEFORE the hotend. The tell is a swollen slug inside the feeder above the gears and filament that snaps when you pull it. Switch to filament with sane tolerance and cut a clean, sharp end before loading.
6. For sustained printing above 250–260 °C, fit an all-metal hotend where nozzle and heat break are one brazed assembly: no leaks, cold nozzle swaps and no shift in Z height.

20. The blob of death buries the head and snaps hair-thin wires

A detached part sticks to the nozzle and within a couple of hours a mass of plastic has grown around the hotend. The nasty part starts when you try to pull it off: the head wiring runs into the ribbon straight through the extruder housing and ends up encased in the blob, and owners describe the thermistor leads as “as thin as human hair”. Snap them and the printer sits in standby and will not let you into the menu. Elegoo documents the removal procedure step by step.

1. Power on and read the hotend temperature on the screen: room temperature means the thermistor lives, a below-zero reading means it is dead.
2. Unplug the rear fan bar cable, remove the bar (three 2.5 mm screws), then the extruder (six 2.0 mm screws) and set it on the bed.
3. Set 200 °C on the screen, wait for it and then switch the printer off at the switch.
4. Pull the blob off with pliers while it is still warm, without tugging on any wires.
5. If the thermistor is dead, soften the blob with a heat gun, remove it and replace the thermistor — keeping the hot air off the extruder housing.
6. Reassemble, retorque the nozzle hot, re-level and reset the Z offset.
7. Prevention is dull but effective: correct Z offset, clean plate and supports for models with a small footprint.

21. The printhead ribbon: two different error codes from one cable

First flavour: X and Y home normally, then Z throws “Endstop z still triggered after reset” and the machine can neither level nor print. The probe reads as triggered across the whole bed regardless of height. Swapping heads between two machines proved the cable: the suspect head works on a healthy printer, and a healthy head fails on the suspect one.

Second flavour: about three hours into a print you get “Heater extruder not heating at expected rate”, and after that the heater will not gain a single degree even when set manually, while power shows 100 %. The thermistor, LED and fans keep working, which sends people chasing the thermistor. A power conductor has broken inside the same ribbon.

1. Meter it at the head: the thermistor should read around 89 kΩ at a room temperature of 27 °C and the heater around 12.8 Ω. Both normal.
2. Now measure voltage while heating: about 24 V at the mainboard output (23.8 V in practice) and the same at the hotend. If the board shows 23.8 V and the hotend shows 0, the ribbon is the culprit.
3. For the endstop error, swap motor or head connectors and see whether the fault travels with the part.
4. Some mainboard connectors are glued down at the factory — pry them gently.
5. Replace the printhead cable. On Type-C machines the same circuit reports “mcu THR: Unable to connect” instead: there you check the communication cable, the two connector screws and the retaining clip.

22. Hot glue from the board connectors drips onto the fan and kills a transistor

A month of flawless printing, then a fine first layer and a second layer that tears itself apart. Opening the base revealed the fan blades glued to a power wire and a blown transistor on the board. The mainboard connectors are dabbed with hot glue so they survive shipping; that glue often ends up resting on the stepper driver heatsinks, melts when they get hot and runs downward.

1. Any time the base is open, inspect the glue: if it is sitting on driver heatsinks, pick it off carefully.
2. Check that the mainboard fan spins freely. A stalled fan leads straight to driver overheating and warnings like Stepper_x OverHeat Warning.
3. Never run with the bottom cover off — it is part of the SoC cooling path.
4. Once a transistor has blown, the board gets replaced. If the machine is under a month old, talk to support first.

23. Four fans on the gantry: 60–70 dB and a part ripped off the bed

Cooling here comes in two stages: two 4015 blowers on the printhead and four 4020 fans on the X axis. With cooling on, the printer roars — 60 dB sustained with peaks to 70 by one reviewer's meter, 63 dB from half a metre by another's. With cooling off it drops to roughly 35 dB. The rated “no louder than 60 dB” refers to the quiet stepper drivers, not the fans.

The less obvious half: a strong stream across a tall part creates a temperature gradient and lifts it off the plate. One owner cycled the bed from 60 to 90 °C, the nozzle from 220 to 290 °C, tried manual and automatic levelling and washed the plate — then found the cause by accident, switching off the four gantry fans.

1. Switch the rear cooling mode on the screen: Silent runs at 60 %, Normal at 80 %, Sport at 100 %. Silent turns the rear bar off entirely.
2. For PETG, ABS, ASA and tall parts, turn it off deliberately.
3. If the fans refuse to spin instead, note that the bar has its own physical switch on the housing — check it and the cable.
4. Still nothing: power down, remove the bar (three 2.5 mm screws) and reseat the cables on the adapter board.
5. If neither the rear fans nor the head blowers run, the firmware has been modified incorrectly and needs reflashing.

24. The outer bed zone never turns on from OrcaSlicer or PrusaSlicer

A large part lifts at the edges while the screen shows the right temperature. The Pro's bed is segmented: a 120×120 mm centre zone at 100 W, 250 W for the whole plate, with the perimeter meant to switch on automatically when a model exceeds the centre. In printer.cfg the perimeter is declared as a separate generic heater with its own pin, thermistor and PID — and the usual bed-temperature commands from any slicer only drive the centre.

The side effect is twice as many failure points. The board carries two bed heater interfaces and two thermistors, TB0 and TB1. That is exactly why “MCU shutdown: ADC out of range” crops up more often on the Pro than on the plain Neptune 4 with its single zone. Plenty of owners consider the second zone pointless anyway: aluminium conducts heat from one zone to the other, but you now have twice the wiring and sensors to fail.

1. Either slice in Elegoo's own slicers, which know about the second zone, or add a macro that switches the outer zone on automatically for OrcaSlicer and PrusaSlicer.
2. Both zones can be toggled by hand on the screen — for small parts that saves power and warm-up time.
3. On an ADC out of range error, power down, reseat the extruder ribbon, wait a minute and power up; then inspect the hotend thermistor wiring and both bed thermistors.
4. Move the bed so the loom underneath flexes less and restart — if the error clears, the bed wiring is your problem.

25. The bed reads 5–7 °C high and refuses to pass 100 °C while printing

An infrared thermometer tells the story: the actual surface runs 5–7 °C below the setpoint, and the spread across the plate reaches 3 °C with no tidy centre-to-edge pattern. Set 105 or even 110 °C on the screen and the bed reaches 100, actually 95. Through the web you can ask for more than 100 and the bed does climb — but the moment printing starts it drops back to 100, even with 105 in the file.

There is a separate complaint too: on some machines PETG at 80 °C throws a heating error about ten minutes in, while 60 °C finishes fine. Support responds by replacing the heated bed — 110 °C is in spec, so this is not a feature.

1. Add 5–7 °C to whatever your filament's datasheet asks for.
2. For ABS and ASA, stop expecting more than 100 °C in practice — reach for an enclosure instead of a hotter bed.
3. Run a bed PID calibration: PID_CALIBRATE HEATER=heater_bed TARGET=65 takes about ten minutes, and the resulting coefficients go into the bed heater section. On the Pro the outer zone is calibrated separately.
4. Give the bed 15–20 minutes to soak before probing a mesh or starting a large part.
5. If the heating error repeats at 80 °C but never at 60, check the voltage selector on the rear panel (230 V for a 220 V supply), inspect the bed loom for kinks and burn marks and reseat the connectors. After that it is a warranty conversation.

26. The melt ceiling: 500 mm/s on the box, 25–27 mm³/s in practice

With 0.6 and 0.8 mm nozzles at a 0.4 mm layer, anything past 100 mm/s starts breaking up the extrusion line and tearing the surface. The advertised 500 mm/s and 20,000 mm/s² never translate into proportionally shorter prints, because the limit is volumetric flow through the hotend, not kinematics.

One Neptune 4 Pro owner ran a max flow test: skipping starts around 27 mm³/s, with 25 mm³/s reliable for PETG at 250 °C on the stock assembly with a 0.6 mm nozzle. Others put the stock ceiling at 15–20 mm³/s. Hitting 500 mm/s with a 0.4 mm nozzle at a 0.2 mm layer would need about 40 mm³/s, which simply is not there. The other half is mass: the Y axis carries roughly three times the load of X and accelerates about half as hard.

1. Stay in the 150–250 mm/s band. Elegoo's own spec sheet recommends 250, and perimeters realistically hold 150–200.
2. The compromise owners settle on: a 0.4 mm nozzle at a 0.3 mm layer, moving to 0.6 mm for large models.
3. Pushing 400 mm/s, raise the nozzle temperature by 10–15 °C or the plastic will not melt in time.
4. Do not try to lift the limits with SET_VELOCITY_LIMIT in start gcode: values inside the sliced file override anything set at print start, and within seconds you are back at 500 mm/s.
5. If you want more flow, change the hotend and cooling, not the numbers in a config file.

27. No Wi-Fi, not every dongle works, and there is a single USB port

Out of the box networking is Ethernet only: Wi-Fi ships on the Neptune 4 Plus and 4 Max, not on the 4 and 4 Pro. Exactly two adapters are officially verified — the Mercury MW150US and the TP-Link TL-WN725N, both in driver-free versions; anything else, dual-band included, is unsupported. The TL-WN725N specifically has to be revision V2 on the Realtek RTL8188EUS chip, because V3 uses a different one.

Even with the right dongle, owners report having to reseat it after every boot. The culprit is Elegoo's own /etc/wpa_supplicant/wpa_supplicant-wlan0.conf, which overrides user settings at every start. There is a deeper snag as well: the machine runs Armbian on Debian 10 Buster, long out of support, so installing network packages the normal way fails outright.

1. The official route: put your SSID and password into wpa_supplicant-wlan0.conf inside the ELEGOO_UPDATE_DIR folder, reflash the firmware, then insert the dongle and read the IP in the printer info screen.
2. 2.4 GHz only. If your 2.4 and 5 GHz bands share one SSID the printer will not find the network — split them and keep to channels 1–8.
3. If the dongle drops on every boot, SSH in and rename wpa_supplicant-wlan0.conf, then reboot.
4. On firmware V1.3.1.4 password SSH is closed and the network manager has been stripped out: access is enabled from a screen menu item, the login differs and the familiar network tools are gone.
5. Reserve the printer's IP in your router, otherwise the slicer loses the machine after every reboot.
6. The lazy fix that always works: plug the printer into a cheap access point by cable and never touch the system. Running a camera and a dongle together needs a USB hub.

28. The stock printer.cfg, and the borrowed config that cooks your motors

The factory printer.cfg runs the bed bang-bang: control = watermark, min_temp 0, max_temp 200 — a 200 °C safety ceiling on a bed that physically reaches 110. The author of one widely quoted community config puts it plainly: “Elegoo's default configuration is pretty unsafe out of the box with its heating profiles.” The input shaper section is identical on every machine — shaper_freq_x 66.66, shaper_freq_y 42.05, types mzv and ei — which are averages, not a measurement of your printer.

1. Identify your motor three ways: the label code (BJ42D15-26V77 is 0.8 A, BJ42D22-53V04 is 1.2 A), the body length (34 or 40 mm) and the X and Y run current in your own factory config.
2. Move the bed to PID control and calibrate separately for the nozzle, the bed and the outer bed zone. Drop the bed max_temp in the config to 110–120 °C.
3. Measure resonances on your own machine. The ADXL345 header is on the board and the factory config already contains the relevant sections — no firmware change needed, just the module. Only the ADXL345 is supported, Elegoo publishes no tutorial, so follow the Klipper documentation. Without an accelerometer, tune shapers with ringing towers.
4. Change one thing at a time and test-print after each. Back up the configuration before any firmware update.
5. If you want real Klipper, that means the community image — installed not over SSH but by replacing the memory module: an eMMC adapter, stripping the printer and knowing your motor and board revision exactly. You gain vanilla Klipper, adaptive meshing, screw tilt calculation and object exclusion. You lose the warranty, get a flaky stock screen and inherit open bugs like a Z offset that drifts between prints.

Neptune 4 Pro error message reference

The Neptune 4 Pro does not show numeric codes — it prints the Klipper error text on the screen and in Fluidd. Here are the messages that come up most often and what sits behind them.

Common 3D printing problems

Beyond the Neptune 4 Pro's own quirks you will meet the ordinary troubles of FDM printing. They behave the same on almost any machine, so each one has its own deep-dive guide:

• First layer not sticking — adhesion, cleaning PEI, first-layer height
• Clogged nozzle and how to clear it — cold pulls, hotend teardown, prevention
• Stringing and oozing — retraction, temperature, drying filament
• Warping and lifting corners — especially relevant to an open-frame Neptune 4 Pro
• Layer shifting and ghosting — belts, acceleration, resonance
• Under- and over-extrusion — flow, filament diameter, feed calibration
• How to dry filament — wet PETG and nylon mimic a clogged nozzle exactly
• Printer maintenance schedule — what to grease and tighten, and how often
• Fumes and ventilation — an open frame vents everything into the room

Also worth a look: setting up OrcaSlicer, the popular alternative to the bundled ELEGOO Cura, for which Elegoo also publishes Neptune 4 Pro profiles; choosing filament for a 300 °C hotend and a 110 °C bed; remote access to your printer if wired networking is not enough; and our Elegoo Neptune 4 known issues with its POM wheels, if you are still choosing between the two. Specs, prices and compatible parts live on the Elegoo Neptune 4 Pro page.