0 Fork 0

mirror of https://github.com/Klipper4FlashForge/firmware.git synced 2026-10-23 05:42:17 -05:00

Files

T

Oleksandr ShyshatskyiandClaude Opus 5 d05565284d Sections E, F, G: wrong values, stale claims, overclaims

E -- values and counts. The START_PRINT prologue has no M400 and an idle
timeout of 864000, not 1800000. The nozzle-clean temperature comes from the
material table FIRST, so NOZZLE=240 with no per-tool temp cleans at 220 --
documented backwards before. TOOLS is the paired 0:220,2:240 form. The wipe Z
is the raw eddy frame, not the homed one. The runout M117 strings are
"T<n> clog" / "T<n> out of filament", not E-codes. Grab-verify raises
E0051+tool, not E0143. Docks are at X~297; 250/280 are staging waypoints. The
grab sequence is documented in the order _grab actually emits it. ff_print
exposes seven status fields, not nine, and never emits TOOLS=. Package sizes
reconciled against the real artifacts (62MB built, ~93MB stock, ~370MB of CI
assets, ~60s not 90s). testing.md's sample output block matched no header the
code prints, so it is replaced with the real ones; its pytest row listed two
of six suites.

F -- stale claims. _FF_PREFLIGHT is not "missing". SDCARD_SET_CHANNEL is not
still issued (764af2a). The app and Moonraker do not coexist; the app is gone.
TOOLCHANGE_STATUS has nothing to compare against since the JSON reader was
removed. The JSON is re-imported every boot until a nozzle is saved, not once.
The 23-check render harness was deleted in 91604c2. PURGE now wipes by
default. 50-print-lifecycle.md gained the section on how the sequences are
actually entered now -- [ff_print], not the slicer. The dead "profiles"
concept is gone from three files, one of which printed it on every replica run.

G -- overclaims. The README no longer says a stock reflash puts the printer
back "exactly as it was": it says what is restored and what survives inert,
which is what test-recovery actually asserts. The random ssh password is
conditional on ROOT_PW_HASH being unset. verify.sh no longer warns "ssh will
not start" on every build for a dropbear we deliberately do not ship -- the
check now reads the way round it means. case-install.sh's read-only block says
it asserts nothing. An uncalibrated tool contributes z_adjust, not zero, in
three operator-facing places. offset_base does not cancel out in X/Y. The
safe-moves file's Z50 promise is scoped to its own lines, with the implicit
prepare's G1 Z10 spelled out.

48 passed, 5 skipped.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-08-24 19:48:43 +03:00

5.2 KiB

Raw Blame History

Toolchange: sequences, sensors, mounted-tool state

Address-verified against the live binary (CommMgr methods). Ported in payload/klipper/extras/ff_toolchange.py; divergences are deliberate and listed at the end.

Physical model

4 docks on the right, X ≈ 297 (250 and 280 are the staging and approach waypoints, x_safe / x_approach), per-dock X/Y from extruder.json (x/y_check_pos..3), station base x/y/z_station_pos.
Lock motor on the carriage grabs/releases the head. Its endstop is registered on manual_stepper gear_stepper; MOTOR_GRAB/MOTOR_RELEASE are STOP_ON_ENDSTOP-style moves that respond "endstop triggered / not triggered".
Sensors (gcode_button): extruder_pos1..4 = head present in dock N; extruder_grab(1..4) = something locked on the carriage; servo_min/max = lock mechanism end positions.
Invariant checked constantly: a head is EITHER in its dock OR on the carriage (E0127-134, E0139-146 cover every disagreement).

Grab — `doGrabExtruderLatest` @0x7a8190

Home gate (abort if not homed, E0160).
Dock-must-be-OCCUPIED precheck: 20 × 50 ms polls of checkInLocation(tool); failure → E0127+tool, no motion.
SET_VELOCITY_LIMIT ACCEL=8000; SET_GCODE_OFFSET X=0 Y=0 MOVE=1 MOVE_SPEED=100.
Approach, once, before the retry loop: G1 X250 F<fast> (x_safe), G1 Y<dockY>, G1 X280 (x_approach).
Then up to 3 attempts. Each begins by re-engaging the dock — G1 X<dockX> F<slow> (+grabOffset correction from test.json) — which is what undoes the previous attempt's back-off. Poll the grab sensor for up to 1 s; on success MOTOR_GRAB, G1 X<dockX-20> F4800 (the app's literal pullback feed, not the calibrated one), MOTOR_GRAB2, and leave the loop. On failure back off to G1 X280 and wait. G1 X250 F1500 is the retreat after the loop, not part of an attempt.
Verify: dock sensor released AND grab sensor pressed → else E0051+tool (ERR_GRAB_VERIFY_BASE; the app's own code here was not recovered).
Apply per-tool G-code offsets (see 40-offsets.md).
SET_VELOCITY_LIMIT ACCEL=20000.

Feedrates: test.json grabSpeed/grabSpeedSlow × 60, fallbacks 24000/6000.

Release — `doReleaseExtruderLatest` @0x7aa394

Not symmetric with grab; the details matter:

Home gate.
Dock-must-be-EMPTY precheck, 20 × 50 ms; failure → E0131+tool, no motion.
ACCEL=8000; clear X/Y offsets. No ACTIVATE_EXTRUDER anywhere on release.
Approach once: G1 X250 F<fast>, G1 Y<dockY>. No X280 stage (grab-only).
Up to 3 attempts, each: G1 X<dockX-10> (no F, modal feed) → G1 X<dockX> F<slow> (slow fallback here is 5400, not 6000) → wait up to 20 × 50 ms for the dock sensor to read "seated" → only then MOTOR_RELEASE → supervise the lock endstop; "ended untriggered" → re-issue MOTOR_RELEASE, max 3 sends per attempt.
Success only: retreat G1 X250 F4800. On failure the carriage stays at the dock.
Verify: in dock AND grab sensor clear → else E0144; all-attempts-exhausted → E0135+tool.
Unconditional tail (success AND failure): MOTOR_STOP, ACCEL=20000 (hence try/finally in the port).

In-process advantage: the app greps response text for "endstop (not) triggered" and polls 40 × 50 ms; a klippy extra just catches printer.command_error from MOTOR_RELEASE synchronously — a strictly stronger signal.

Mid-print change (`changeExtruderChannel`)

Save state (SAVE_GCODE_STATE NAME=PAUSE_state), manage temps, release old → grab new, prime G1 E15 F240 / retract with part-fan ooze control, RESTORE_GCODE_STATE MOVE=1, restore M220.

Which tool is mounted

The app persists an imperative index (now_extruder in extruder.json), reset to −1 by homing, and never derives it from sensors — because its grab-sensor read (getGrabSensorStatus @0x76f294) is a plain OR of four bytes: it knows only that something is held, never which.

ff_toolchange.py derives identity instead: extruder_pos1..4 are per-tool dock switches, so the mounted tool is the one whose dock is empty (while the grab sensor is pressed). Exactly-one-empty + held → that tool; all-occupied + held, or several-empty + held → abort (faulty switch / ambiguous — refusing beats guessing, since guessing wrong would drop the carried tool into another tool's dock). Nothing is persisted; the value cannot go stale. now_extruder is never written (see 40-offsets.md on why writing the JSON is forbidden), and it is no longer read either: the JSON reader was removed, so TOOLCHANGE_STATUS reports only sensors and Klipper config and has nothing to compare against.

Deliberate divergences in the port

No 4th unsupervised MOTOR_RELEASE (the app fires one before checking its own failure counter — energising the lock with nobody watching).
No M119 refresh polls — Klipper's gcode_button state is event-driven; the 20 × 50 ms gating semantics are kept as wall-clock deadlines.
now_extruder not written; identity is sensor-derived.
Modal-state restore added: the app leaks G90 + last feedrate + accel into whatever runs next; the port snapshots and restores absolute/relative mode, feedrate, and the pre-sequence accel limit.

5.2 KiB Raw Blame History Unescape Escape