cc2-toolkit

Notes, calibration prints and scripts for the Elegoo Centauri Carbon 2 (CC2), from one owner's first weeks with the printer. It is a mix: some results come from real prints, some files are only checked in software (sliced, measured, not printed yet). Each calibration folder says which in its README and verification.json (physical_print_tested).

Your printer, bench and filament differ from mine. The guides show you how to find your numbers; my results live in examples/ as one worked example, not as recommended values.

What changed (2026-11)

• COSMOS: the author moved the CC2 to COSMOS (OpenCentauri's Klipper OS) on 2026-11-01. Motor currents ship as the CC1's, the Z baseline moves after a mesh, and TOOL= picks the CANVAS lane; the tools/ start and status scripts are stock-firmware only (guides/firmware-and-ssh.md).
• Z offset on stock firmware: a hidden -0.06 on every start and a screen that shows a remembered value; plus = farther (guides/gotchas.md item 29).
• Pressure advance: the G-code's value is the one that runs; the old "skip the PA test" advice is withdrawn, with line-test results per filament (guides/tuning.md).
• Wall streaks (open): what was ruled out and what helped (one wall, vase mode); outer-wall-first is the next test (guides/tuning.md).
• More gotchas: PETG brims at gap 0 and supports at 0.2 weld, minimum layer time changes wall speed per layer, TPU via the top-cover bypass, a loose plug behind the power trips (items 28-33).
• Filaments: SUNLU Rapid PETG record; Elegoo PETG HF rough tops traced to the flow limit (filaments/).
• Tools: send_cc2.py no longer greps a large G-code whole on the printer to read the colour table.

What changed (2026-10-17)

• Motor drivers: X/Y StealthChop can be switched live from G-code with no reboot (corrects the earlier note); a StealthChop setting left on shifted layers at 20000 accel (guides/firmware-and-ssh.md).
• Power cut: the screen's power-loss Resume does not measure Z again and printed in mid-air; how a layer recovery worked instead (guides/firmware-and-ssh.md).
• Z offset: the screen value went back to 0 after each print and adds to the G-code's z_offset; slice it in (guides/gotchas.md item 24).
• Noise: a second band with one belt at 42-52 mm/s; moving the printer moved the bands; after-move checks (guides/tuning.md).
• More gotchas: skirt off the bed on full plates, internal bridges at 40 mm/s, tall towers snapping in the base infill, top-surface flow 1.09.
• Filaments: Elegoo PETG HF record (rough tops still open); measured design clearances by hole direction (calibration/README.md Step 4).

What changed (2026-10)

Firmware 01.03.02.36 throughout.

• Accel: walls printed clean up to the printer's 20000 limit on rigid feet, well above the shaper's estimate (guides/find-your-numbers.md, examples/borja-floor-slab/).
• Feet tested (2026-10-18): Elegoo's anti-vibration feet vs printed rigid feet, dry head moves, phones on and around the printer: Elegoo's shook ~20% less under hard moves, a draw on everyday moves, no noise difference. Charts in guides/rigid-feet.md.
• Noise: belts ring in a speed band (50-62.5 mm/s on the author's printer); how to find yours, quiet profile and Z-hop notes (guides/tuning.md).
• New gotchas: a weeping hotend crust behind first-layer blobs, purge ooze on the first part, out-of-spec filament diameter, reboot only when cool (guides/gotchas.md).
• Firmware / SSH: log rotation, SFTP files missing from the app until reboot, queued commands during levelling, camera thread leak, TMC2209 StealthChop findings, COSMOS CC2 alpha (guides/firmware-and-ssh.md).
• Calibration: headless generators for the flow (YOLO) test, max volumetric speed tower, card v3 and temperature towers for any range; the v1 card's flow and temperature rows were unreliable; tolerance now uses a print-in-place hinge test (calibration/README.md).
• Filaments: Elegoo PLA Matte, DEEPLEE PLA+ 2.0 and SUNLU PLA+ records; stock 0.8 mm retraction held on every PLA-family filament tested (filaments/).

What's inside

Folder	What
guides/	Find your numbers (input shaper vs slicer accel), tuning and noise, gotchas, SSH / firmware access, rigid feet
calibration/	Printable tests with headless generators: temperature tower, YOLO flow, max volumetric speed, calibration card, ironing grid
filaments/	Per-filament results template + worked results (Elegoo PLA, PLA Matte, PLA+, PLA-CF, PETG HF; DEEPLEE and SUNLU PLA+)
tools/	Slice from the command line (quiet accel presets, adaptive layers), send and start over SSH, check G-code accel against your calibration
photo-measure/	Read mm and camera angle from a phone photo of a printed ArUco card (PM-2)
examples/	My measured numbers with full setup details
tests/	Offline regression tests for tools/

Share your results

Open an issue with the "CC2 results" template: firmware, bench, shaper results, the accel you chose and what the print looked like.

License

Code MIT (LICENSE), docs and models CC BY 4.0, with AGPL-3.0 exceptions for meshes and profiles derived from ElegooSlicer's calibration resources. Details: NOTICE.md.