1 / 1
Remix of Hex2000's Ball Clock - Minutes 4 Balance Bar — image 1

Remix of Hex2000's Ball Clock - Minutes 4 Balance Bar

Order a print of this model

We calculate the final price from your selected file. You review and accept it before payment.

Saved on this device.

Rebalanced top minutes bar for Hex2000's Ball Clock

Rating
0 / 5
Likes
1
Downloads
21
Makes
0
Views
73
Reviews
0

Updated at source:

Model information

License
Creative Commons — Attribution — Noncommercial — Share Alike

Tested with 10mm steel ball bearings (~4g each, 20-pack from Amazon). The printed bar weighs approximately 11g in PLA. With the original pocket position (~87mm from pivot), the counterweight overpowers the 5-ball side and the bar won't tip. Moving the pocket to ~62mm from pivot restores proper balance — the bar holds steady with 4 balls and tips reliably on the 5th.

If your bar isn't tipping correctly, this is likely your fix. Your results may vary slightly depending on ball weight and print density.

How this remix was made

This modification was designed and executed entirely by Claude (Anthropic's AI assistant) running in Claude Code (CLI), with no manual CAD editing. Here's the process:

Mesh Analysis

Using Python with trimesh and numpy, the original STL was loaded and analyzed programmatically. The bounding box, vertex positions, and face normals were examined to map every feature on the bar — the pivot slit, the counterweight pocket, and the multi-ball track.

The counterweight pocket was profiled at multiple Z heights to confirm it was a perfect spherical depression (radius 5mm, ~10mm opening at the top surface).

Physics Calculation

A torque balance analysis was performed using the measured values:

  • Ball weight: 4.05g each (81g per 20-pack of 10mm steel bearings from Amazon)

  • Bar weight: 11g (PLA)

  • Bar center of mass: calculated from mesh volume, ~5.6mm left of pivot

  • Original counterweight distance: ~87mm from pivot slit

The first 4 balls on the 5-ball side are roughly symmetric around the pivot (net torque ≈ 0). Only the 5th ball at ~23mm from pivot creates the tipping force. With the counterweight at 87mm, its torque far exceeds what 5 balls can produce, confirming the design imbalance. Physical testing confirmed that moving the counterweight ~1 inch closer to the pivot corrected the behavior.

Mesh Modification

The pocket was moved using boolean operations via manifold3d (a geometry processing library):

  1. A cylinder was unioned with the bar to fill the original pocket

  2. The result was intersected with the bar's bounding box to keep the top surface flat

  3. A sphere (R=5mm) was subtracted at the new position to cut the relocated pocket

The non-manifold edge in the original STL (Error 3 in most slicers) was also repaired in the process.

Tools Used
  • Claude Code (Opus 4.6) — AI assistant driving the entire workflow

  • Python 3 with trimesh, numpy, scipy, networkx, manifold3d

  • No GUI CAD software was used at any point