Elegoo Centauri Carbon: Known Issues and Fixes
22 problems specific to the Elegoo Centauri Carbon: print head cable failures and error 103, a bed mesh forever probed at 60 °C, drifting Z-offset, an exhaust fan that never stops, ten firmware releases and a full error code reference straight from Elegoo's own procedures.
The Elegoo Centauri Carbon is a 2025 enclosed CoreXY printer with a 256×256×256 mm build volume, a 320 °C nozzle, a 110 °C bed and a claimed 500 mm/s top speed, sold assembled at open-bedslinger money. It is very much a living machine: ten firmware releases in eighteen months, and Elegoo has publicly owned up to at least two serious defects — the print head cable and the year-late multicolor system. Below are 22 problems that belong to this specific printer, with Elegoo's own repair procedures, real measured numbers and workarounds that actually hold up. Generic FDM troubles — stringing, warping, clogs as such, first-layer adhesion — live in separate guides linked at the end.
Current firmware at the time of writing is 1.4.49, released 29 July 2026. Nine versions came before it, and nearly every one fixed something while breaking something else: 1.1.25 is officially flagged for thermal runaway, 1.1.29 locked out parameter changes mid-print, Elegoo pulled 1.1.45 entirely, and 1.4.44 broke the connection to Elegoo's own slicer. So the first thing to do with a new Centauri Carbon is check what it is running: Touch Screen → Settings → Check for Updates.
The second thing is working out which hardware revision you own. Elegoo shipped three hardware revisions in three weeks during May 2025, and a fourth from November 2025 with a redesigned head cable. Their own wording: “Due to different manufacturing and shipping batches, the product you receive may not have the latest updates.” Translation: everyone's batch is different, and half the advice you find online describes somebody else's machine.
1. Print head cable failure and Error 103 / “Printhead anomaly detected”
The signature failure of this model. The screen throws “Printhead anomaly detected. Please troubleshoot the connections between the mainboard and printhead board” or error code 103. It most often bites in the front-left corner — exactly where the printer homes during calibration. A common variant is the nozzle temperature dropping 10–15 °C below target on the second layer, right when part cooling kicks in, and the print dying there. In the worst cases the machine boot-loops with no toolhead light at all.
The root cause is mechanical. The toolhead board connects over a USB-C cable carrying 24 V, and its factory bend radius is far too tight. Homing to the front left stretches that cable to the maximum travel distance, the conductors fatigue and the connector wears out. Hence the diagnostic that Elegoo itself publishes: the error appears and disappears when you wiggle the head by hand with the printer powered on. Failure timing is wildly inconsistent — some machines die after ten hours of printing, others are fine past seven hundred.
In November 2025 Elegoo published a detailed response to the community. The company ran the cable through two months of continuous printing plus five million cycles of the head shuttling between extreme positions twice per second, concluded that fire risk under normal use is “extremely low”, and listed six design changes:
- The cable-to-Type-C junction changed from perpendicular to parallel and the cable rerouted, cutting the bend angle at the four extreme positions
- Type-C fixing tab thickened from 3 mm to 6 mm to stop intermittent contact
- Insulation layer thickened by 0.1 mm and the core wires wrapped more tightly
- The cotton filler inside the cable upgraded to a V-0 flame-retardant material
- Cable length increased by 50 mm for more bending room
- Strain relief lengthened from 7 mm to 11 mm
- Wiggle the head with the printer powered on. If the error tracks the movement, it is the cable, not the heater
- Power down, loosen the two connector screws with a 1.5 mm hex key, reseat the cable and check whether the jacket has crept out of the right-angle plug
- Ask support for the revised cable — part number B0C0L1
- If you source a third-party USB-C cable, get a genuine 100 W or 240 W rated data cable, not a charge-only one. Some aftermarket plugs have an aluminum collar that stops the connector seating fully
- Print a strain relief — for example Centauri Carbon Print Head USB Strain Relief on Printables
- The structural fix: flash the OpenCentauri patched firmware, which homes to the front right instead. Their wording is blunt — the stock front-left home “causes wear and eventual failure over time”
- Stopgap while a cable is in transit: set part cooling to 0 % for the first two layers ramping to 100 % by layer six, or simply cap it at 20–40 %
2. Hotend heats up on its own at power-on — and why firmware 1.1.25 is dangerous
Three separate mechanisms produce this symptom, and all three are documented. Elegoo's wiki names one: “The temperature of the hotend of the printer rises automatically during startup. Reason: The hotend control board is damaged” — the fix is a new toolhead PCB, which they ship out. VoxelMatters documented a blown power stage: “a failed MOSFET on the toolhead PCB caused the hotend to heat uncontrollably on power-up, even with no print running”. And a German forum thread traced a machine that drove the bed to 100 °C and the nozzle to 300 °C — against setpoints of 60 and 220 — to a cold solder joint in the temperature control circuit, cured by reflowing it.
Firmware is a separate hazard here. Version 1.1.25, released 9 May 2025, carries an explicit note in the community archive: “Known Issue: Thermal Runaway. This firmware version has a known thermal runaway issue.” Extruder thermal protection only arrived in 1.1.29 on 18 June 2025. That makes the popular summer-2025 advice — “roll back to 1.1.25, everything worked there” — actively dangerous. Yes, 1.1.25 predates the mid-print lockout bug. It also predates thermal protection.
- Unplug the printer
- Check the firmware version. If it is 1.1.25, update. Below 1.1.42 the only safe target is 1.1.40, the official image Elegoo hands out for reinstalls; anything older than 1.1.29 has no thermal protection
- Email 3dp@elegoo.com with a video and your serial number — this defect gets a replacement toolhead PCB, part B0C091 or B0C0G1
- Preventive trick from owners: tape the ceramic heater with Kapton so the metal clip cannot short the heater against the thermistor — one such short cost an owner the entire toolhead board
3. The start G-code breaks leveling: mesh always at 60 °C, ooze from a 140 °C nozzle
Whatever bed temperature you set in the filament profile, the bed mesh is probed at 60 °C and the bed only reaches its working temperature afterwards. For ABS, ASA and polycarbonate at 100–110 °C that means the mesh was captured on an effectively cold plate, and the first layer is simply wrong. The temperature is hard-coded in the firmware source, in klippy/extras/auto_leveling.cpp. The tracker issue has been open since 29 September 2025 and is still open.
On the smooth plate side the effect reverses: the bed first heats to 60 °C for the mesh, then the M190 waits for it to cool back down to the 30–35 °C the profile asks for. That idle wait can run over half an hour before a print begins. Elegoo acknowledged it in the slicer tracker back on 19 May 2025: “Different heating and waiting strategies are needed for the B-side, which will be optimized soon.” A year later, nothing. The second flaw in the same start G-code: the nozzle is heated to 140 °C before probing, oozes, and a blob ends up under the sensor on contact. The probe here is a strain gauge — it needs no nozzle heat at all. Here is what a corrected start G-code should look like:
M104 S[nozzle_temperature_initial_layer]
M190 S[bed_temperature_initial_layer_single]
M109 S[nozzle_temperature_initial_layer]
G90
G28
M729 ; nozzle wipe - must come BEFORE probing
BED_MESH_CALIBRATE- Replace the machine start G-code with a plain BED_MESH_CALIBRATE at the real working temperature — the firmware accepts it, thanks to its Klipper lineage
- Alternative route: M8210 sets the mesh probing temperature, followed by BED_MESH_CLEAR and BED_MESH_CALIBRATE
- METHOD=fast reuses the stored 6×6 mesh from printer.cfg if you do not want a full re-probe every time
- Put M729 (nozzle wipe) before leveling so ooze never reaches the sensor
- Check that the wiper rocker moves freely and is not binding
4. Z-offset from the self-test is wrong and gets wiped on every re-level
Almost everyone walks this path. After the factory self-test the first layer looks fine, then a few prints later the nozzle starts scratching the plate and you nudge Z-offset by 0.05–0.075 mm. Then you re-run auto leveling and your saved value is gone. The underlying issue is probe repeatability: run-to-run variance of 0.03–0.05 mm and worse, caused by mechanical noise and vibration, no settling delay and too few samples. That is not owner folklore — it is the wording of a pull request against the official repository: “the Z consistency on the CC is absolutely awful with run to run variances >= 0.03 - 0.05”.
- Print a 100×100×0.2 mm test slab — Elegoo's own guide suggests 230×230×0.2 mm, but a smaller one is enough to dial it in
- Adjust Z-offset live, mid-print, in 0.01–0.05 mm steps. The officially suggested 0.1 mm step is too coarse and will overshoot the sweet spot
- Enter the value you land on under Settings → Z-axis offset
- Stop running the full auto-leveling routine afterwards, or you will be dialling it in again from scratch
5. Error 502: the nozzle probes the same spot forever and leveling aborts
During leveling the nozzle keeps probing one spot, or simply keeps travelling downward, and the screen shows 502 “leveling sensor data collection error” with the routine aborted. Elegoo lists three causes: the printer sitting on an unstable or uneven surface, loose bed locking screws, or a faulty strain gauge and its adapter board. Tools needed are 2.0 and 2.5 mm hex keys plus a ruler or vernier caliper.
- Set the printer down so all four feet contact the tabletop. A rolling cart will not do — if in doubt, put the machine on the floor and retest
- Remove the PEI plate and use two hex keys to check and tighten the lower locking screw clockwise in each of the four pairs
- Refit the plate, power on, tap Home and use a ruler or caliper to compare the distances from the three Z blocks to the platform mounts
- Reseat the five connectors on the strain gauge board. When swapping connectors to isolate a fault, only swap within the same row — never diagonally or across rows
6. Uneven bed: the plastic lip sits proud of the magnetic sheet
The symptom is squashed, skipped first layer along the front edge, and a mesh showing 0.2 to 0.6 mm of variation. The mechanism is simple: the lip of the plastic carrier stands above the magnetic sheet, so the PEI plate rests on plastic along one edge instead of sitting flat on the magnet. One owner measured a 0.6 mm lip. On top of that sits roughly 0.3 mm of genuine crown in the plate itself, which no amount of screw adjustment removes. A related defect is logged in Elegoo's own tracker: the plastic heater tray sits 0.55–0.78 mm low, and the mesh never probes closer than 20 mm to the edge anyway.
- Dump the mesh before touching a single screw: drop an M8803.gcode file on a USB stick and run it — the printer writes out a printer.cfg with the mesh, which the community browser visualizer will render for you
- Note that the mesh bounds are asymmetric: 20 minimum and 246 maximum on both axes, so it never describes the actual plate edges
- If you file the lip down, do it along its whole length. Trim only the front and the plate will foul the other edge when you flip it
- Shim with 0.5 mm washers between the plastic carrier and the aluminum
- Adjust the four spring-loaded screws underneath with a 2 mm hex key. They are sometimes simply left under-torqued at the factory
Two myths circle this topic. First, “never touch the bed screws, they hold the strain gauges”. The wiki does say to tighten all four, but they are spring-loaded, and an owner demonstrated experimentally that spring preload does not shift the strain gauge readings: meshes captured with the springs fully slack and fully compressed share the same relief. You can set the average plane with those screws; you cannot flatten 0.2–0.3 mm of crown with them. Second, “shim the low spots with Kapton”. Under a strain-gauge auto-level that only corrects the picture in the printer's head while physically bending the plate the other way. It works only if you never auto-level again.
7. The loader pushes filament out instead of pulling it in
You feed in a new spool and the extruder drives it straight back out, clicking against the filament tip. Elegoo logged this itself as a known issue of firmware 1.1.42, 18 September 2025: “Filament loading/unloading may turn filament the wrong way.” It survived into 1.1.46 and still surfaces on the community patched build. It hurts most during color changes, where the filament fails to load and the print dies. One tracker report puts it plainly: “You hear it clicking against the end of the filament as you push it in and the loader pushes it out.”
- Power cycle the printer — the symptom clears for the current session. That is the only working workaround
- Do not roll back to 1.1.25 over this: that build has thermal runaway
- Do not leave filament changes unattended on multi-color jobs
8. Layer shifting: 2 mm after a pause, plus six mechanical causes
Layer shifting on this machine comes from two independent places. The first is a firmware bug specific to the Centauri Carbon: after an M600 or an ordinary pause, the model resumes roughly 2 mm off on both axes. The tracker issue dates from 4 October 2025 and was confirmed again in February 2026 — “its ruined multiple long Hueforge prints”. One reporter needed hours of idle pause to trigger it, another only minutes. No cause has been established and there is no fix, which leaves exactly one practical rule: do not park the printer in a pause, swap filament immediately.
- A deformed hotend assembly. If the heat break is bent, replace the assembly
- A loose 3010 fan. Fasten it with M3×6 screws specifically — Elegoo states outright that the wrong screws leave the assembly insecure
- A sticking filament cutter that either fails to cut or fails to return
- A binding nozzle-wiper rocker
- The nozzle striking the top edge of the waste chute during wiping — start a print and watch it happen
- Over-tightened bottom cover screws distorting the chassis. These need loosening, not tightening
Check belt tension separately: loosen — do not remove — the four X and Y tensioner screws with a 2.0 mm hex key, walk the head by hand into all four corners so the belts self-tension, then retighten. The same procedure clears error 401. If shifting persists after all that, the printer is no longer the suspect — see our guide to layer shifting and ghosting.
9. Cancelling a print and manual jogging: two ways to bend your hotend
Cancel a print manually and the head drives home in X and Y without dropping the bed first. The nozzle ploughs straight through whatever you were printing. One tracker report ended with a bent heat break and the head fan cover torn off: “This caused a collision where the nozzle hit the print… and bent the hot end.” A normally completed print lowers the bed correctly — this is strictly a manual-cancel bug. As a bonus, pressing Cancel during leveling freezes the printer until leveling finishes, and if the bed is above 60 °C it waits for cool-down before it will even acknowledge the cancel.
- Never cancel a tall print. Pause it, wait for the lift, then stop the job
- Use Home All only — the X→Y→X→Y sequence is self-correcting
- Do not home X alone with the head parked at the back: you get a positioning error of roughly 60 mm, after which the left side of the bed becomes unreachable
10. Two fans you cannot switch off: exhaust and mainboard
The carbon-filter exhaust runs nonstop. Any non-zero speed in the slicer becomes 100 %, and switching it off from the screen resets on the next layer. The comedy is that the printer simultaneously asks you to open the door for PLA and PETG. The second irritant is the mainboard fan, which keeps buzzing indefinitely after a print ends, whines and squeals; one owner's started howling after 160 hours. The culprits are a cheap bearing and a mount fixed to a steel bracket that is itself soldered to the base plate, so vibration goes straight into the chassis.
- The dirty but effective exhaust trick: set the filament type in your profile to ABS and the fan never spins up
- The G-code hack: inject M106 P3 S0 every twenty lines
- Proper exhaust control only exists on the OpenCentauri patched firmware. Note that the stock OrcaSlicer profile still fires this fan with M106 P3 on PLA
- The mainboard fan whine is roughly halved by decoupling its body with thick double-sided tape
- A 24 V blower swap works, but buy a three-pin model: a two-pin fan without a tachometer triggers the “abnormal fan” error
11. Firmware roulette: what each version broke, and how to roll back
Elegoo publishes no firmware changelog. The Update Log section of the wiki tracks edits to wiki pages, not releases. The only complete archive — dates, changelog text and the raw responses from the update server — is maintained by the OpenCentauri community. Here it is, condensed.
| Version | Date | What matters |
|---|---|---|
| 1.1.18 | 31.03.2025 | Fixed abnormal strain gauge detection during printing and the printer getting stuck at target temperature without starting |
| 1.1.25 | 09.05.2025 | Officially flagged for thermal runaway. Do not run it |
| 1.1.29 | 18.06.2025 | Added extruder thermal protection during homing. Broke parameter changes during a print |
| 1.1.40 | 15.08.2025 | Base for the community patched firmware, and the build Elegoo hands out for USB reinstalls |
| 1.1.42 | 18.09.2025 | Fixed the web UI during prints and power-loss resumption. Broke part cooling below 10 % |
| 1.1.46 | 21.10.2025 | Reverse-feed bug still present, plus random model fan errors |
| 1.1.45 | — | pulled by Elegoo, release date never published |
| 1.4.44 | 28.04.2026 | Added CANVAS support. Its own changelog lists a known issue: “Issues with connecting to Elegoo Slicer” |
| 1.4.46 | 08.05.2026 | Fixed multiple system freezes. Broke manual filament change on two-color prints |
| 1.4.49 | 29.07.2026 | Current. Fixed fan control, rare freezes and known Wi-Fi instability |
One combination deserves its own warning: Elegoo Slicer 1.5.0.7 with firmware 1.4.44. The slicer reports Idle, the web UI and camera stream both work, and sending a print returns “Printer not connected”. Elegoo's official answer on 30 April 2026 was to roll the firmware back through support, or disable “Upload and Print” and start the job from the printer page instead. The clean fix is updating to 1.4.46 or newer.
- Download the build you want (the community keeps the full archive; support emails links to official files on request)
- Rename the file to update.bin and drop it in the root of a FAT32-formatted USB stick
- Insert the stick and power on — the printer offers the update itself, tap Update
- Remove the stick once it finishes, otherwise the update prompt reappears on every boot
- Rolling back uses exactly the same procedure; the firmware does not distinguish newer from older. Updates land on the inactive A/B slot, so the previous build physically stays in memory
12. Camera, file list and slicer control all disappear
The printer prints and reports Idle or Printing, but Elegoo Slicer and OrcaSlicer show nothing: no camera feed, no file list, dead buttons. Meanwhile the web UI on the printer's address and on port 8080 may answer normally. Elegoo support typically diagnoses this as “the web server on the board has failed” and replaces the mainboard under warranty. It is worth knowing that this is a software defect. One owner had it come back on the brand-new board and cured it with a firmware update instead, concluding on 3 June 2026 that the problem was never hardware in the first place.
- Update the firmware first — 1.4.46 and 1.4.49 fix it
- Reconnect Wi-Fi from the printer menu
- Move the printer to 2.4 GHz instead of 5 GHz — that alone fixed it for several owners
- Do a factory reset
- Only after all of that should you ask support for a mainboard
13. Manual color change silently does nothing without CANVAS
You insert a color change at a layer in Elegoo Slicer, the printer finishes the approach move, retracts 2 mm — and carries on printing in the same color. No pause, no error, no message. The answer is in the profile: change_filament_gcode for the Centauri Carbon contains M6211 T1, a command meant solely for the CANVAS multicolor unit. Firmware without CANVAS ignores it, along with T[next_extruder], and the profile provides no M600 fallback at all. Elegoo replied on 30 April 2026 with “recorded… currently under evaluation”; it remains unfixed.
- Write M600 into the machine filament-change G-code yourself
- Disable the prime tower — this part is not optional. After the jump from 1.1.46 to 1.4.46 the firmware ignores the whole WIPE_TOWER block, and the M600 nested inside it goes with it. Older advice that just says “insert M600” without this caveat does not work
- Expect the printer to sit in an undefined loading state afterwards and need a reboot once the job ends
14. The chamber is passive: 38–45 °C and a thermistor that reads 4–7 °C low
An enclosure is not a heated chamber. There is no active heating on the Centauri Carbon — the chamber warms from the bed alone, and it plateaus somewhere between 38 and 45 °C. Hence the headline complaint: large ABS, ASA and especially polycarbonate parts still lift their corners with the door shut and the bed at 100–110 °C. Tom's Hardware says it flatly about Prusament PC Blend: “despite letting the chamber heat up, using a 100℃ on the bed, the frame still lifted at the corners, but it's usable”.
| Conditions | Chamber temperature | Measured by |
|---|---|---|
| 20 min printing, fully closed, 20 °C room | 41 °C | Tom's Hardware |
| Top lid removed | 31 °C | Tom's Hardware |
| Door open | 27 °C | Tom's Hardware |
| 1 h printing, bed 100 °C, fully closed | 33 °C | Owner report |
| Fully closed, 29 °C room | 38 °C, reaching 42–45 °C after a couple of hours | Owner report |
| Ceiling without insulation | no higher than 45 °C | Owner on patched firmware |
One detail throws people off when they read the screen: the chamber thermistor is screwed to the aluminum frame with a steel screw, and thermal coupling to the frame makes it read 4–7 °C below actual chamber air temperature. So 38 °C on the display means roughly 42–45 °C inside. Nothing is broken — that is just where the sensor lives.
- Use a brim, run the bed at 100–110 °C, keep the door shut for the whole print and kill any drafts
- Insulation genuinely helps: 10 mm foil-backed foam or 4 mm acoustic mat on the side walls and lid. Thermal imaging points at the right wall — the one the fan blows onto — and the lid edge as the worst offenders
- Do not plan on PA-6: it wants a 130 °C bed and this printer tops out at 110 °C
- A realistic ceiling: ABS and ASA in modest sizes, PAHT-CF with caveats. PC, POM and PEEK are not happening in a passive chamber
- For warping in general, see our guide to warping and layer cracking
15. Heat creep on PLA and TPU: Elegoo itself tells you to open the lid
This is the enclosure's other face. Twenty minutes into a print the chamber sits at 41 °C, the hotend's heat break can no longer shed enough heat on PLA, and you get creeping under-extrusion that ends in a fully jammed hotend. This is not speculation: Elegoo prints the warning on a sticker on the lid and writes in its own support article “Keep the machine cover open during operation to ensure proper cooling and prevent nozzle clogging”. Making it worse, the printer has no clog detection in hardware — the spec sheet has a dash in that row.
- For PLA, PETG and TPU, take the top lid off or leave the door open. That is the intended operating mode, not a bodge
- Watch the hotend fan: error 702, “Abnormal heat break cooling fan”, means the heat break is not being cooled and heat creep is guaranteed
- For the full mechanics of jams, see our nozzle clogging guide
- If the filament has picked up moisture, start with our filament drying guide
16. Extruder clicking, coked nozzles, and why “Carbon” is not about the nozzle
The symptoms: rhythmic clicking from the head roughly every ten seconds, under-extrusion and delamination, and a machine that prints small models fine but falls apart on large ones. The dedicated error is “Extrusion Abnormality”. Inside sits a dual-gear direct drive with a NEMA14 motor, a TMC2209 driver, a 52:10 reduction and a 10 mm SDK11 tool-steel drive gear; the heater is a ceramic PTC element of roughly 9.6 ohm rated at 60 W.
- Heat to 250 °C and clear the nozzle with the cleaning needle — step one of Elegoo's official routine. One caveat: if the jam is clearly cooked PLA, stay below 205 °C or you will bake it solid, as covered in section 15
- Adjust extrusion force: remove the shroud, unplug the fan, loosen the four housing screws and turn the idler adjustment counterclockwise
- Open the gearbox and inspect the gears. They pack with filament dust and wear flat. Watch the idler spring orientation on reassembly
- A drive gear slipping on its shaft is cured with a drop of superglue
- For nozzle play, tighten the three screws. Hold the heat block with an 8 mm spanner after removing the silicone sock; the nozzle itself takes a 6 mm
- If the needle will not clear it, do a cold pull: heat above printing temperature, push filament through, drop to 100–150 °C and yank it out
About that nozzle. “Carbon” in the model name refers to the materials the printer can run, not to what ships in the box — which is a brass high-flow nozzle in a proprietary format, 0.4 mm. Tom's Hardware confirms: “high-flow, proprietary brass nozzle… can be swapped for a hardened steel available from Elegoo's website”. Owner reports converge on a blunt estimate: a brass nozzle is chewed out by a single spool of carbon-filled filament. A second observation worth repeating is that these nozzles are unusually prone to coking, particularly on PETG run too hot.
17. You do not get the full 256×256: the cutter stopper is in the way
The advertised build area is 256×256×256 mm, and out of the box you will not get it: the filament cutter stopper physically occupies part of the print area, so the slicer crops the plate. A genuine 256×256 mm footprint simply will not fit. Elegoo gave this its own wiki page — and was honest about the price of the fix.
- Undo the cutter stopper screw with a 2.0 mm hex key and remove the stopper
- Set “Excluded bed area” to 0×0 in the slicer
- Accept the trade-off: automatic filament unloading stops working. From then on you unload by hand — push the cutter lever back, then Prepare → Extruder → Unload
18. The printer burns tens of gigabytes of traffic while idle
Leave the machine doing nothing and it still generates abnormal outbound traffic. VoxelMatters measured around 28 GB across 48 hours of idling and called it “the clearest sign that the software side of the product still needs tightening” — pulling the printer off the internet stopped it dead. The mechanism is known: looping connectivity checks in the firmware. The community patched build lists it as a bullet point of its own, “Connectivity checks are blocked (web traffic issue)”. On a home network it is an annoyance; on a print farm it is a real problem.
- Block the printer's internet access at the router and keep it LAN-only — printing and the web UI are unaffected
- Or flash the OpenCentauri patched firmware, which blocks those requests out of the box
- An important caveat: there is no evidence that print content is being transmitted. This is a connectivity-check bug, not surveillance — do not repeat it as “the printer spies on you”
- If you want proper remote access without a cloud, see our write-up on remote access for 3D printers
19. Three revisions in three weeks, then a fourth: what to check on your own unit
The Centauri Carbon is not one machine but several. Elegoo rolled out three hardware revisions in three weeks during May 2025, then a fourth from November with the redesigned head cable. Their official phrasing: “Due to different manufacturing and shipping batches, the product you receive may not have the latest updates.” So before applying somebody else's advice, work out which batch you are holding.
- 12 May 2025 — a connector for the future multicolor unit appeared on the rear panel
- 14 May 2025 — LED strips added inside the chamber and the LED under the camera removed, because it was ruining the image
- 30 May 2025 — silicone pads between door and hinges for transport protection, a chamfer on the top frame to reduce PTFE friction, acoustic foam in the frame, nylon bed spacers replaced by springs, build plate thickness up from 0.4 to 0.5 mm and side panel from 0.6 to 0.8 mm
- from November 2025 — the new print head cable with the six changes from section one
The camera board exists in at least three revisions. The newest, V2.0.028, ships in machines with the 24-LED strip and has no LED on the camera itself. The camera is otherwise identical across batches and unremarkable: 720p, 16:9, 30 Hz maximum, fixed focus with a glued lens barrel, 90–110° field of view. For early machines there is a separate official procedure for swapping the old top frame for the new one, with different head-cable routing and new damping foam — Elegoo never says what it was fixing.
- Look for the multicolor connector on the rear panel — a rough date marker
- Count the chamber lights: one lamp or two
- Look under the bed: nylon spacers or springs
- Check how the cable enters the print head: perpendicular is the old revision, parallel is the new one
20. CANVAS on the Centauri Carbon: a year of waiting, a lost enclosure and a new error
The loudest story attached to this model. The launch landing page promised “available in Q3 2025” next to a picture of a four-color print. On 12 November 2025 Elegoo published a notice instead: “Due to some limitation of the hardware, so far the developing multicolor system has not yet delivered a performance that meets our expectation. And we cannot guarantee a release date for the multicolor system so far.” Plenty of people had bought the machine specifically for that module.
- 28 November 2025 — first image of the system
- 25 December 2025 — a video of it printing
- 6 January 2026 — the spool holder shown
- 26 January 2026 — “we have finished the hardware design… and are still optimizing its firmware”
- 27 April 2026 — CANVAS for the Centauri Carbon finally ships
Compensation came as a choice of three: a $50 filament coupon on orders over $100, a 365-day warranty extension, or an $80 coupon toward a Centauri Carbon 2 Combo. Claims were open from 10 November 2025 to 31 January 2026 — that window is closed, and there is no retroactive route to the bonus now.
The retrofit's biggest surprise: fitting CANVAS stops the printer being enclosed. The stock glass lid no longer fits the new arrangement, and the Centauri Carbon 2 lid does not fit either. Elegoo's official position is that a thermal hood is not included, and that ABS and engineering materials should be run inside an external fire-resistant enclosure of roughly 600×600×1000 mm. Installation means removing the toolhead front panel, the middle frame section and the rear cover, unscrewing a daughterboard and swapping the controller board; 3DWithUs clocked the job at about one hour ten minutes. Firmware 1.4.44 or newer is required.
Right after the upgrade some owners hit “Nozzle front cover detached” with the cover physically in place. The official cause: the new toolhead front cover contains a magnet the detection sensor looks for, and the old cover has none. The fix is swapping in the new cover.
- Four colors maximum, no expansion
- Every color change means a purge. Centauri Carbon 2 owners report waste outweighing the model itself
- The slicer underestimates job time: long multicolor prints run 3–5 hours over the estimate
- Carbon- and glass-filled composites are out of scope: the feed gears and sensors are not built for abrasives
- Spools sit in the open, with no dry storage
- Four extra motors and their fans make the machine louder
- If multicolor is your main use case, a Centauri Carbon 2 Combo is the more coherent purchase than retrofitting the first model
21. Noise, a rattling toolhead cover and idle whine
The spec sheet says “noise level ≤55 dB”, and that is true — with the fans off. Elegoo's own measurements at one meter put it at up to 60 dB with the fans running and up to 65 dB with the doors open. Independent numbers agree: reviewers logged 55–60 dB at maximum speed with the lid off, and 50–52 dB at moderate speeds. HowToGeek put it more vividly, comparing the sound to “a swarm of angry bees inside a beehive”, while Alex Reviews Tech described the enclosure as an effective echo chamber.
- The single most common complaint here is the toolhead front cover rattling at speed. Test it with a finger: press the cover and the clatter stops. Fix it with anti-squeak tape, foam or a thin rubber strip — owners describe it as abnormal but entirely ordinary on this machine
- Do not stand the printer on a metal cabinet — that is Elegoo's own advice, the chassis resonates
- Fit anti-vibration feet
- Clean and lubricate the shafts; the maintenance schedule calls for it every three months anyway
- Enable silent mode and drop speed on long overnight jobs
22. Arrived damaged: bent frame, jammed bed and a failed self-test
The self-test stalls at the Input Shaping step and goes no further — or the bed jams at the top and the machine emits ten seconds of loud grinding during the test. Elegoo names four causes: a frame deformed in transit, a tilted bed, transport screws still fitted under the bed, and a loose Z motor connector. Clearance sellers routinely list units with smashed glass, bent frames and cracked bodies, and a forum thread on a competitor's board calls bent frames and smashed glass the most common issue with this printer. Treat both as a biased sample — discounted listings are damaged units by definition — but the pattern is consistent enough to be worth checking for.
- Check whether the glass lid sits flat on the top frame. If it does not, the frame is deformed — that is a warranty case, go to support with photos immediately
- Remove the three factory transport screws under the bed. This is the most commonly missed step
- Lay the printer on its side, remove the ten base screws with a 2.0 mm hex key and pull the Z belt by hand: the bed should rise with no resistance
- Lubricate the rails with white lithium grease
- Reseat the Z motor connector
- If the self-test fails with code 401, tension the X and Y belts using the procedure from section eight
Practical advice for anyone ordering one: film the unboxing, measure the diagonals and inspect the base panel before first power-on. The base is thin sheet metal with the mainboard directly underneath it, so a dent there is not cosmetic. In fairness, packaging improved noticeably during 2026 — later batches ship with far more foam than the launch units did.
The maintenance schedule: every three months, and it is not a formality
Elegoo specifies servicing the Centauri Carbon every three months: clean and lubricate the X and Y plain shafts and the three Z shafts, the three Z lead screws, and clear the Z linear bearing holes with cleaning gel for one to two minutes. For that last step the company provides a printable tool, leadscrew_cleaning_tool.stl. The video tutorial on that page is still marked “Coming soon”.
- For general guidance on what to lubricate and when, see our 3D printer maintenance guide
- A caliper earns its keep here and during bed leveling — measuring from the Z blocks is exactly what Elegoo asks for
- The chamber filter is a consumable: there is a wiki page for replacing it, but Elegoo publishes no replacement interval
Elegoo support, spare parts, and the 110 V bed everyone forgets about
Here is a difference nobody mentions in reviews. The heated bed runs directly off mains and has a fixed resistance of about 48.4 ohm. That works out to 1000 W on 220 V but only 250 W on 110 V — Elegoo's own spec sheet lists rated power as 1100 W at 220 V and 350 W at 110 V. On a North American outlet the bed crawls toward 110 °C, taking well over seven minutes in testing, while a European machine gets there in a fraction of the time. Nothing is faulty; it is Ohm's law. If you print a lot of ABS on 110 V, budget the extra warm-up into every job.
- Support lives at 3dp@elegoo.com. They ask for a video reproducing the error, a video of the printer homing, the firmware version from Touch Screen → Settings → Check for Updates, and the serial number from the white label near the power switch
- Elegoo ships replacement parts free of charge, mainboards included — but you fit them yourself
- The free print head cable form at elegoo.com/pages/cc-support returns a 404 as of 20 August 2026, so email support directly
- The compensation window for the delayed multicolor system closed on 31 January 2026
- Expectation setting on the official GitHub tracker: 28 issues filed, 3 closed, and the 60 °C bed mesh bug open since September 2025. One participant's verdict — “I think Elegoo created this github to give people a place to complain, not to actually fix any of the issues”. Email gets results faster than the tracker
| Part | Elegoo part number |
|---|---|
| Nozzle | B0C0C1 |
| Hotend assembly | B0C121 / B0C131 / B0C141 / B0C151 |
| Thermistor and ceramic heater | B0C0H1 |
| Extruder gearbox | B0C011 |
| Print head PCB | B0C091 / B0C0G1 |
| Print head connection cable | B0C0L1 |
| Heat break cooling fan | B1C031 |
| Part cooling fan | B1C041 |
| Bed leveling sensor | B0C0A1 |
| Strain gauge adapter board | B0C031 |
| Mainboard | B0C001 |
| Chamber thermistor | B0C2N1 |
Quote those part numbers when you write in — they come straight from Elegoo's own replacement guides, and naming the part removes an entire round of back-and-forth about which component you actually need.
Elegoo Centauri Carbon error code reference
The first-generation Centauri Carbon has exactly nine documented numeric codes and two text messages — no more than that. Every cause and action below comes from Elegoo's official wiki.
| Code | Meaning | Cause | What to do |
|---|---|---|---|
| 101 | The heated bed did not heat up as expected | Low ambient temperature, faulty bed thermistor or heater, loose cable between boards | Confirm the room is 5–40 °C. Compare heating at 60 and 90 °C. Measure heater resistance across the N/L terminals of the OUTPUT connector — 40–60 ohm is normal. Reseat the HBED connectors |
| 102 | Anomaly in reading the heated bed NTC | Loose bed thermistor cable, damaged bed or mainboard | Remove the base cover (10 screws, 2.0 mm hex), reseat the BED port, then the BED-T cable into the BOX-T port. Wrong chamber temperature means replace the bed; correct means replace the mainboard |
| 103 | The print head did not heat up as expected | Damaged Type-C communication cable, adapter board pin, heater port cable or ceramic plate | Set 200 °C and move the head by hand: error only on movement means replace the cable. Otherwise measure heater resistance, around 10 ohm is normal |
| 104 | Anomaly in reading the printhead NTC | Poor adapter board contact, faulty nozzle thermistor | Reseat the thermistor port and measure resistance. Elegoo's table: 100 kohm at 25 °C, 80.39 kohm at 30 °C, 125.2 kohm at 20 °C |
| 401 | Accelerometer chip error | Loose X and Y belts, or a faulty accelerometer | Loosen the four tensioner screws with a 2.0 mm hex key, walk the head into all four corners, retighten and rerun the self-test |
| 502 | Leveling sensor data collection error | Unstable surface, loose bed locking screws, faulty strain gauge | Set the printer on a solid flat surface, tighten the four locking screw pairs, compare distances from the three Z blocks, reseat the five strain gauge board connectors |
| 701 | Abnormal mainboard fan | Fan cable or a failed board | Reseat the cable. While homing, watch the board LED: dark with a dead fan means replace the board; lit with a sluggish fan means replace the fan |
| 702 | Abnormal heat break cooling fan | Loose hotend fan cable | Reseat the cable on the toolhead PCB, remove the middle head housing, heat the nozzle to 220 °C and check rotation |
| 703 | Abnormal model fan | Loose part cooling connector or failed fan | Reseat the fan cable and the adapter on the toolhead PCB side. Elegoo also officially suggests reinstalling the firmware from USB. Verify via Function → Fans → Model Cooling Fan |
| Printhead anomaly detected | Print head communication fault | Intermittent head cable contact, failed toolhead PCB | Move the head by hand with the printer on. If the error tracks the movement, it is the cable. Diode-test with a multimeter: around 0.700 V is normal |
| Nozzle front cover detached | Front cover reported missing | Old toolhead cover without a magnet fitted after a CANVAS upgrade | Remove the cover and check for a magnet. No magnet means fit the new version |
Community firmware: what actually works and what does not
You cannot install Klipper on a Centauri Carbon, and that is Elegoo's official position rather than a guess: “No, the ELEGOO OS is not based on Klipper… the Centauri Carbon's hardware architecture uses only one chip. Currently, it is not compatible with open-source Klipper commands or some open-source plugins, such as Fluidd.” Moonraker ports are closed on the first-generation machine and there is no klippy.log, which trips up third-party automation. What did grow around this printer is a serious community that reverse-engineered the hardware down to board level and maintains the only complete firmware archive.
- OpenCentauri patched firmware, based on 1.1.40 — alive, in beta. Gives you SSH, a stoppable exhaust fan, front-right homing, Z-offset adjustment while idle, file uploads during prints, blocked outbound connectivity checks and G-code control of the chamber light. The Centauri Carbon 2 is not supported. On warranty the authors are careful: they have not seen a claim refused, but that is their observation, not Elegoo's policy
- COSMOS puts full Klipper on the stock board and is actively developed, but its own authors warn that “COSMOS is not currently stable so you should not install the early beta builds on any production machine”. The stock board is severely resource-limited, plugins are out, and CANVAS is not supported yet
- The official firmware source was released under GPL-3.0 in September 2025 after community pressure, and has not been updated since — there is no code for the 1.4.x branch
- The Home Assistant integration is alive and explicitly supports the Centauri Carbon. Known bug: it does not reconnect by itself after a power cut
- The second Home Assistant integration you will find in search results is abandoned — the author stated outright that he sold his Centauri Carbon and will not update it. Do not spend time on it
- There is no official Centauri Carbon profile in PrusaSlicer. Elegoo supports its own ElegooSlicer, OrcaSlicer and Cura — and since ElegooSlicer is an Orca fork, our OrcaSlicer guide transfers almost one to one
One more honest note about support. Elegoo's official GitHub tracker exists, but 3 of 28 issues are closed and the 60 °C bed mesh bug has been open since September 2025. A participant summed it up as “I think Elegoo created this github to give people a place to complain, not to actually fix any of the issues”. Set expectations accordingly: a specific problem gets resolved faster by emailing 3dp@elegoo.com than by filing a ticket.
General 3D printing problems on the Centauri Carbon
Beyond the Centauri Carbon's own quirks you will meet the ordinary problems of FDM printing, which have nothing to do with the model you own. We covered each one in a dedicated guide:
- First layer not sticking — the complete guide — PEI, Z-offset, temperatures, soap.
- Nozzle clogging in depth — cold pulls, heat creep, PTFE, nozzle swaps.
- Warping and layer cracking — enclosures, temperatures, materials.
- Stringing between parts — retraction, temperature, filament moisture.
- Layer shifting and ghosting — belt tension, Input Shaper, speeds.
- Under- and over-extrusion — flow ratio, drive gear, temperature.
- Printer maintenance schedule — what to lubricate, when, and how to clean it.
- How to dry filament — PLA, PETG, TPU, nylon, with temperatures and times.
- Fumes and ventilation — especially relevant when the lid has to stay off.
- Post-processing — support removal, sanding, painting.
Also worth a look: our OrcaSlicer settings guide, since ElegooSlicer is an Orca fork and almost everything carries over; our filament guide if you are still deciding what to run in a passive enclosure; and remote access for 3D printers if you would rather not leave this machine talking to the internet.
Elegoo Centauri Carbon FAQ
Sources
- ELEGOO Wiki — Centauri Carbon — error codes, hardware issues, step-by-step part replacement, part numbers.
- ELEGOO Wiki — Printhead anomaly detected — the official diagnostic tree for the print head cable.
- ELEGOO Official — response on the print head cable — testing, the six design changes, the replacement program.
- ELEGOO — Update Notice on Centauri Carbon — the official admission on the delayed multicolor system and owner compensation.
- OpenCentauri — Centauri Carbon firmware archive — every version with dates and changelogs, including the thermal runaway note on 1.1.25.
- OpenCentauri — Patched Firmware Features — the list of what the patch fixes in stock firmware.
- OpenCentauri — Centauri Carbon hardware — fan part numbers, chamber thermistor offset, bed specifications.
- Elegoo's official issue tracker — the 60 °C mesh, post-pause shifting, cancel without Z lift, jogging collisions.
- ElegooSlicer — manual color change on the Centauri Carbon — the M6211 T1 analysis and the M600 workaround.
- Tom's Hardware — Elegoo Centauri Carbon review — chamber temperature measurements, Benchy times, lighting, TPU, nozzle.
- VoxelMatters — Elegoo Centauri Carbon review — 28 GB of idle traffic in 48 hours, the failed MOSFET on the toolhead PCB.
- Tom's Hardware — Broken promises: Elegoo walks back planned AMS — the timeline of the delayed multicolor system.
- FauxHammer — CANVAS for Centauri Carbon review — losing the enclosure after fitting the module.
- drucktipps3d.de — Centauri Carbon leveling thread — drifting Z-offset and probing at 60 °C.
- HowToGeek — Elegoo Centauri Carbon review — firmware freezes across 100+ hours of testing, noise and lighting.
- Notebookcheck — Versions-Karussell — a breakdown of the three spring 2025 hardware revisions.
Printer Hub Team
We study official documentation and manufacturer guides, test mods on real printers, and analyze community experience from Reddit, Discord, Printables, and YouTube.
Your printer, in reach from anywhere
Remote access to Klipper and OctoPrint, AI print monitoring and your own model library — on a VPS from 172 ₽ for the first month.
- NVMe drives
- Anti-DDoS
- 24/7 support
- 🇷🇺 🇩🇪 🇳🇱 7 locations
First month with the promo code, then from 429 ₽/mo — Promo plan: 2 GB RAM, 30 GB NVMe, backups included.
Buying via this link supports Printer Hub 🤝