Hex fit spectra

Updated finished models: All three Bondhus fans, native FreeCAD, print STEP and 100% infill modifier projects. The earlier revision below is preserved as history.

Two sizes, five lettered fits per size, in fan print orientations.

Fit coupon v6 · native geometry checked; slicing and physical selection pending.

HX04 v2 three-tier fan: historical geometry studied on this page. The owner socket-fit trial failed; later straight-socket revisions have their own evidence.
HX04 v2 three-tier fan: historical geometry studied on this page. The owner socket-fit trial failed; later straight-socket revisions have their own evidence.

Final straight-socket models

Your complete fan and the three Bondhus single-set racks — rendered native models, STEP, FreeCAD and print meshes. Owner selected 3C / 5C: +0.38 mm total across-flats clearance. New sockets have at least 40 mm straight engagement, separate 3 mm / 1.5 mm lead-ins, square stops and vents. Other sizes use this allowance provisionally; full-rack print and installation remain unqualified.

Extended engineering studies

Open the interactive motion and FEA lab · Full numerical report

All 26 tool paths, 25 initial load/fixture cases, exaggerated computed flexure, stress maps, full-rack loading/unloading and locking-hoop pinch. The deeper review finds front-host collision on the old assumed swing and shows why mouth forces do not represent loads at the handles. These are prototype studies, not a printed load rating.

Sources, native CAD, display fields and review ZIP · Corrected unscaled v2 brace project—reslice with compensation once

Fit spectrum v6: pick one letter per size

Two requested test sizes for 3 and 5 mm hexes, each with five slots A–E, engraved with size and letter. C uses the latest requested clearance: 0.38 mm total across flats. Each square 3 mm floor has a 1 mm through vent.

Download the 3/5 mm spectra on one plate
3 mm coupon STL · 5 mm coupon STL
FreeCAD · STEP · CAD, adjustable source and checks ZIP

LetterTotal AF clearance3 mm bore AF5 mm bore AF
A0.20 mm3.20 mm5.20 mm
B0.30 mm3.30 mm5.30 mm
C0.38 mm3.38 mm5.38 mm
D0.46 mm3.46 mm5.46 mm
E0.56 mm3.56 mm5.56 mm

Print at the fan angles

Keep the supplied flat bases on the bed; do not auto-orient these models. The bores follow the HX04 fan’s full print transform: 3 mm rises 25.10° above the bed; 5 mm rises 38.53°. Their six-flat directions retain the revised flat-to-flat handle clocking. The coupons have built-in feet to preserve these orientations.

Shaft guides are 40 mm for 3 and 5 mm, plus a separate 0.8 mm entrance. fan_orientation_v6.json records both shaft and handle vectors, the original fan layout hash and print transform. To change an angle, adjust that frame and rebuild; FreeCAD dimension properties are metadata.

Pick the letters after printing

Use the same printer, filament and calibrated compensation as the rack. Inspect the inclined bore roofs, vent paths and supports in your actual slicer. Starting at E, try each slot down toward A and select the smallest clearance that seats fully to the floor and slides out comfortably, without forcing or binding. Compare rocking and rotational play. Report 3 mm = C, 5 mm = C. The loose 3E bore permits full rotation geometrically; do not choose it when a rotational stop is required.

Native solids, nominal insertion, vent openings, retained stop shoulders and mesh/STEP/FreeCAD exports pass the geometry checks. The supported 3/5 mm Orca job was submitted through the Bambu bridge on physical AMS slot 3, then paused before layer 1 on a filament-feed error. The owner subsequently completed the fit trial and selected 3C and 5C. No G-code is supplied.

The v5 spectra used two-tier HX05 angles and are unsuitable for evaluating the fan’s print-angle fit. Earlier coupons: vented v4 · snug v3. The archived full rack below retains its earlier seats.

Archived HX04 v2 rack and original coupon

Historical geometry and computational evidence. The owner’s socket-fit trial failed; this full rack has not yet adopted the new guide geometry.

What changed in v2

  • A peg in every hole the plate covers: hook pegs every inch across the top row, locking hoops across the bottom row, and gravity-load pegs in the 72 holes between (9 columns × 10 rows). The reviewed peg profiles are unchanged; the back is now their 5.4 mm wall throughout. Hook the top row in, then the intended final motion swings the bottom toward the board. Whole-rack clearance and physical hoop recovery remain unqualified; see the installation audit.
  • The solid-infill helpers are braces: a solid web through each row of sockets, a short bulkhead at every socket and a disc round every peg root, instead of padded sleeves.

Original socket design (superseded)

  • Every seat is at least four times the key size. Long hex keys and the smallest T-handles get deeper seats so they sit lower.
  • Each L-key’s short arm points straight downhill, so a key dropped in at any turn swings to rest. T-handles have no preferred turn, so T3 and up sit in hex-keyed bores clocked to their grip turn.
  • The original design used a short neck, a widened chamber and a sloped floor. Owner testing found inadequate centering and rotational restraint; that seating approach is superseded by the straight-guide trial above.
  • To take a tool, straighten it, lift it out of its seat, then pull it forward. Every route clears the other tools and a 3″-deep shelf 2″ above.

236 mm wide, on the nine-cell peg pattern (columns at ±101.6 mm) · 178 mm from the board to the furthest tool · 236 mm tall in the print, on its left end.

ToolSizeSeatTurn held by
Hex L-key1.5 mm12 mmgravity
Hex L-key2 mm15 mmgravity
Hex L-key2.5 mm17 mmgravity
Hex L-key3 mm20 mmgravity
Hex L-key4 mm26 mmgravity
Hex L-key5 mm33 mmgravity
Hex L-key6 mm40 mmgravity
Hex L-key8 mm50 mmgravity
Hex L-key10 mm61 mmgravity
Torx L-keyT1012 mmgravity
Torx L-keyT1513 mmgravity
Torx L-keyT2015 mmgravity
Torx L-keyT2518 mmgravity
Torx L-keyT2720 mmgravity
Torx L-keyT3022 mmgravity
Torx L-keyT4027 mmgravity
Torx L-keyT4531 mmgravity
Torx L-keyT5035 mmgravity
T-handle2 mm71 mmset by eye
T-handle2.5 mm32 mmset by eye
T-handle3 mm12 mmhex-keyed, ±19°
T-handle4 mm16 mmhex-keyed, ±12.3°
T-handle5 mm20 mmhex-keyed, ±9.3°
T-handle6 mm24 mmhex-keyed, ±7.5°
T-handle8 mm32 mmhex-keyed, ±5.4°
T-handle10 mm40 mmhex-keyed, ±4.2°

Printing

  1. Original v2 coupon (history) (coupon STL, plain 20%, about 113 g / 5h 26m 9s). It is cut from the final model and prints the same way. The end slice hangs on the board from the top of the left peg column and holds Torx T10/T15, hex 1.5/2 and T-handles T2, T2.5 and T3; drop the tools in, knock the board, pull them out. The PEGS strip is the rest of that column down to the locking hoop, so you can feel the gravity-load pegs and the hoop snap. Blocks T5 C and T5 F key the T5 shaft to its corners and to its flats: the one that holds the bar at the right turn tells you how your handles are made. This archived full part assumes corners; the owner has since confirmed flat-to-flat handles.
  2. Archived full part (older seats): open the Bambu Studio / OrcaSlicer project with solid-infill zones. It is the body plus brace zones at 100% infill (socket row webs, socket bulkheads, peg roots), about 636 g / 1d 6h 26m 33s. A lighter alternative is the print-oriented STL at plain 20% infill, about 554 g / 1d 1h 39m 23s. Print it on its left end as oriented, with tree supports on the build plate only (they land under the peg hooks behind the mounting face).
  3. ASA: use an enclosure and a 5 mm brim. Then paint the debossed size labels.

Native CAD, source and checks (ZIP) · Review evidence

v1 (two peg columns, padded solid sleeves): print STL · v1 CAD bundle