Skip to content
Ch 07Before you start

Before you start Chapter 07

Before you start · Chapter 07 — A/B belts and tensioning

Cut, route and clamp the two CoreXY belts, set a provisional tension, and finish the X carriage with the inductive probe. This makes the gantry a working XY motion system — everything after it (toolhead, wiring, bring-up) assumes the belts are on.

What you're building in this chapter. Two belt loops and the printed carriage that clamps them. In CoreXY neither motor owns an axis: the rear-right motor drives the A belt, the rear-left drives the B belt, and the toolhead moves in X when both turn the same way and in Y when they turn opposite ways. Each belt is one long loop that starts and ends at the X carriage, and on a Voron the two loops sit at different heights and never cross — each stays in its own horizontal plane the whole way round. Where a belt has to be driven it wraps a toothed motor pulley teeth-first; everywhere it is merely turned, it runs smooth-side-on over a plain bearing stack — except at the XY joints, where each belt meets one toothed 20T idler and one plain F695 pair: A rides the left joint's upper plain stack and the right joint's upper toothed idler; B rides the left joint's lower toothed idler and the right joint's lower plain stack (p.105, p.135, p.136, p.138). Whether a face is teeth-on is fixed by which joint and which height — never twist a belt to change it. The A belt is the upper of the two planes, B the lower. Both belts start in the left carriage half and leave it heading left, so A's first turn is on the left joint's plain stack and B's first turn is on the left joint's toothed idler. Both loops are the same length by design, so two belts cut identical arrive at identical tension — and no amount of tensioner adjustment can undo a length error, which is why they are cut against each other and their protruding tails compared at the carriage. The chapter finishes by fitting the inductive probe that later levels the gantry and maps the bed.

Revali reading a caliper
Belts that sound alike are not alike. Measure both. My ears are excellent. I still measure.

Time: 2.5–4.0 h hands-on, first build (survey §5.1 P07 / §7.2 Ch 07).

Sessions: 5 × ~30 min (first-build estimate; each Pause: line carries its own segment minutes — note that segment 3, threading both belts, cannot be broken and needs ~80 min in one sitting — there is a soft stop after Step 07.16 if you must).

Prerequisites:

  • Ch 04 (A/B drives and front idlers assembled, motor pulley set screws threadlocked), Ch 05 (gantry, XY joints, X carriage on the rail, Ti backers), Ch 06 Part A (gantry installed on the Z joints, Z belts clamped).
  • Print batches B02 (accent: [a]_tensioner_left/right, [a]_cable_cover), B03 (front idlers), B04 (XY joints + X carriage + probe_retainer_bracket).
  • The gantry must move freely by hand over its full travel with no belts on. Fix binding now, not after belting.

Tools

  • Ball-end hex 2.5 mm, for every screw in this chapter
  • Needle-nose pliers and tweezers (belt threading; the manual calls these out on p.132)
  • Flush cutters or sharp scissors for belt; steel rule or tape ≥ 2.2 m; 150 mm steel rule for the tension span; marker
  • Phone with a spectrum app: Spectroid (Android), Sound Spectrum Analysis (iOS), or Gates Carbon Drive (both — use the "motorcycle" option) (Voron tuning docs)
  • Temperature-controlled soldering iron + M3 brass tip, only if the heat-set inserts on p.129 were not done in the Ch 00 insert pass

Printed parts

Parts arrive in the bins named below (see the bin map in print/README.md#bins); check the bin label's list before you start.

Looks like STL Bin Qty Colour
Gantry/X_Axis/X_Carriage/x_frame_V2TR_MGN12_left 07-X 1 Black
Gantry/X_Axis/X_Carriage/x_frame_V2TR_MGN12_right 07-X 1 Black
Gantry/X_Axis/X_Carriage/probe_retainer_bracket 07-X 1 Black
Gantry/Front_Idlers/[a]_tensioner_left 04-B 1 Blue — fitted in Ch 04, adjusted here
Gantry/Front_Idlers/[a]_tensioner_right 04-A 1 Blue — fitted in Ch 04, adjusted here
Gantry/AB_Drive_Units/[a]_cable_cover 07-X 1 Blue

Hardware (chapter totals)

Fastener / part Qty Where
Gates 2GT open belt, 6 mm wide 2 lengths, cut equal A and B belts (p.125, p.131 — see Step 07.8)
M3×8 SHCS 4 belt clamps, 2 per X carriage half (p.131, p.140)
M3×12 SHCS 2 1 per X carriage half, in the plain hole beside the belt slot, head 3 mm off the plastic (p.130)
M3×30 SHCS 4 2 carriage bolts (p.140) + 2 probe retainer bracket (p.143)
M3 nut 2 right X carriage half, capture the M3×30 carriage bolts (p.129)
M3 heat-set insert 4 1 per X carriage half, in the bottom hole + 2 in probe_retainer_bracket (p.129); the M3×12 stop screws do not use them
Inductive probe, Omron TL-Q5MC 1 X carriage (p.143)
Fibreglass tape (LDO-supplied) as needed probe insulation, front + sides only (LDO note p.143)

Read first

  • The tension you set here is provisional. Gantry squaring (Ch 06b) begins by fully releasing A/B belt tension, and it needs a running printer (G28, QUAD_GANTRY_LEVEL, SET_STEPPER_ENABLE), so it happens inside Ch 13. Set enough tension to home and QGL. A/B tension is set three times on purpose: Ch 07 (provisional — enough to home and QGL), Ch 06b Step 06b.15 (provisional again, after squaring, cold and open) and Ch 14 Step 14.4 (final — panels on, cold, door open, immediately before the closed-chamber soak and hot Z-joint tighten of 14.6). Z belts: 06b.3 (so QGL converges for squaring), set for the last time at 14.5, same cold session. Ch 14 owns the final number; tensioning to final now just gets undone — that is survey §5.2 W1, a ~1.5 h rework.
  • Equal length is the whole trick. Both belt paths are the same length by design, so if the two belts are cut identical, equal tension follows (p.125). Cut each belt against the other, never to a tape measure alone: fold-cut both at once (Step 07.8), or do as p.125 recommends: route one, remove it, and cut the second to match it.
  • There is no sourced deflection number. The official manual gives no tension figure at all — only the equal-length rule. The one sourced number in this build's document set is 110 Hz over a 150 mm span (Voron secondary tuning). voronldo.com's 80–100 Hz is discarded (survey §4.3). Do not substitute a "push it 2 mm with your finger" test — nothing here calibrates it.
  • Nothing in this chapter is torqued. Belt clamps and carriage bolts go on lightly, the belts get pulled tight by hand, then the carriage bolts are fully tightened (p.140–141). The manual specifies no torque values; don't invent one.
  • Three LDO deviations land in this page range: the X-carriage variant check (p.129–130), probe insulation (p.143), and p.145 skipped entirely.
  • The inductive probe is built once, here. Insulation, fitting, the 6 mm height and the cable route are Steps 07.34–07.37 (manual p.143–144). The tape goes on here and nowhere else: Ch 10 Step 10.30 only confirms it. The lead is never cut: it arrives pre-terminated for the toolhead board, and Ch 08 Step 08.50 plugs it in. Ch 08 otherwise only confirms the decision and bags the Klicky set. The probe_retainer_bracket and the fibreglass tape are counted in this chapter's totals, not Ch 08's.

Sources for this chapter:

Video coverage (Steve Builds, pre-release LDO kit — differs from Rev D+): Part 4 @1:38:20 (+6m), Part 5 @1:52:44 (+8m), Part 6 @0:01:10 (+18m), Part 8 @1:43:45 (+6m)

helperHelper steps: 9 · 07.5, 07.23, 07.27, 07.29, 07.30, 07.31, 07.32, 07.36, 07.37

← PreviousCh 07 — chapter overview

Next →Step 07.1 — Confirm the gate before you start