Step 14.6 — Closed-chamber soak, settle QGL, lock the Z joints hot, re-verify¶
Step 6 of 36 · Chapter 14 — Calibration and tuning
(no image — see text)
Parts:
- the four M5×40 Z joint SHCS · already on the bench
- 4 mm ball-end hex key · tool
- 150 mm machinist square · tool
Do:
SET_IDLE_TIMEOUT TIMEOUT=99999,G28, note the Z height where a ball-end 4 mm reaches all four Z-joint heads.- Shut the panels and door, bed to 110 °C, hotend 150 °C, and soak 1½–2 hours. Then work the table hot.
Check
chamber_temp was plateaued for at least 20 minutes before the QGL runs, and each run's correction got smaller rather than bouncing around.
What you're looking at
Aluminium grows as it warms, so a gantry squared cold in Ch 06b is not square at chamber temperature. A long soak brings the frame to its working size, and the four Z joint M5×40 bolts get their first and only full tighten hot.
| Order | Hot, with the chamber closed | Accept |
|---|---|---|
| 1 | PROBE_ACCURACY at bed centre, then QUAD_GANTRY_LEVEL three to five times in a row |
Each run converges and its correction shrinks run over run rather than bouncing around |
| 2 | Jog Z back to the height you noted, open the door, and repeat Ch 06b's machinist-square check at both front corners, gantry forward and back | No light under the square at either front corner, hot |
| 3 | Fully tighten the four M5×40 Z joint bolts while the machine is still hot, rear pair first, door open as briefly as you can | First and only full torque on these four bolts |
| 4 | Refit any side panel you took off, its M3×12 clip screws back in, door shut. Still hot: G28, one more QUAD_GANTRY_LEVEL, then PROBE_ACCURACY. Only then heaters off and SET_IDLE_TIMEOUT TIMEOUT=1800 |
QGL converging in ≤3 retries within retry_tolerance: 0.0075, PROBE_ACCURACY σ < 0.003 mm with no trend, idle timeout back at the config's 1800 |
Caution
Light showing under the square hot when it did not cold means the squaring did not survive thermal expansion or the belt change: leave the Z joints light, go back to Ch 06b Step 06b.14, and return here. Do not tighten the Z joints cold; cold-tightening throws away the point of the soak and shows up as first-layer inconsistency.
Caution
If the QGL runs need all 5 retries where Ch 13 converged in 2, one of the six belts is off. Fix it at 14.4/14.5 before tightening anything: the tighten freezes whatever geometry the belts are holding.
Tip
bed to your print temperature, 110 °C for ASA. If a rear Z joint is out of reach, take that side panel off and refit it before the last QGL.
Good stopping point · ~15 min since the last one
since the last pause (plus the 1½–2 h soak) — all six belts at final tension and even, the machine soaked closed, QGL settled over 3–5 runs, the four Z joint bolts locked hot, PROBE_ACCURACY re-verified hot afterwards and the idle timeout set back to 1800. The whole belt-and-geometry item is closed; heaters can go off. The next item is the extruder.
Source: Voron docs § V2 Gantry Squaring, steps 14–17 · Voron startup wizard § Quad gantry level · Klipper docs § QUAD_GANTRY_LEVEL · Klipper docs § SET_IDLE_TIMEOUT · leviathan-printer-rev-d-sbv2.cfg · Ch 06b Part B · Video: Part 9 @2:49:16