Tip Profile Options

✓

The measurement that fixes the length

The whole protrusion feature measures 10–12 mm along the blade. That settles the overall length and rules one option out immediately: variant D as first drawn runs its taper 15 mm back from the point, which is too long. D12 and E below both come in at 12 mm.

It also confirms the current design was right about this one thing: A's feature is LIP_FLAT 8 + 2 × LIP_RAMP 2 = 12 mm. The length was never the problem — the shape was.

01 — The four profiles, to scale

Tip on the left, root on the right

The horizontal line is the inner rail. The curve above it is how far the protrusion sticks inward, toward the cup. Blue are the two that match the measured 12 mm.

Read left to right as travelling from the point of the blade back toward the mount. A's profile leaves the rail only after 9 mm of nothing, then steps up — that step is what reads as blunt. D12 and E both start rising at the point itself, so the tip stays sharp while the protrusion grows continuously, and both end in a vertical face.

02 — Two-segment vs three-segment

The only difference is whether there is a flat before the hook

D12 — two segments

slope 16.3° all the way, then a square cut

One straight rise from the point to the hook. Simplest shape, sharpest tip, but two mated blades meet on a line, not a face.

E — three segments

slope 23.6° for 8 mm, 4 mm flat, then a square cut

The same square hook, but the last 4 mm before it runs parallel to the rail — so two mated blades touch across a 4 mm face instead of a line. That is the flat the original design had, kept without the blunt step.

03 — The numbers

Variant	Feature length	Rise angle	Flat	Hook angle	Mesh
A — current	12 mm ✓	60.3°	8 mm	60.3°	pass
D — peak at 15	15 mm ✕	13.1°	0	89.7°	pass
D12	12 mm ✓	16.3°	0	89.7°	pass
E	12 mm ✓	23.6°	4 mm	89.7°	pass

Both new variants are watertight, manifold, single-shell. Volumes 25,579 mm³ (D12) and 25,735 mm³ (E), against 29,424 mm³ for the current A — the continuous profiles remove material near the tip, which is where they stop being blunt.

A's rise and return are both 60.3° because its ramps are 2 mm long for 3.5 mm of height. That is the "step" visible in the drawing. Neither of the new ones has a step anywhere except the hook, where one is wanted.

04 — F: the tip hooks inward

Confirmed against the real red gripper — this is the shape

The operator checked the original part on 2026-08-21: the protruding flat runs all the way to the tip, and the tip is that flat's end line. There is no separate "tip" beyond the protrusion. So the correct change is not adding a lump to the inner face — it is translating the whole tip section inward, outer surface and all.

✓

Thickness is preserved exactly — which is what went wrong before

An earlier attempt carried the protrusion to the tip by moving only the inner face. That makes tip thickness = depth + protrusion, so a 1.0 mm tip became 4.5 mm — the blunt trapezoid that was rejected on sight. Moving both faces by the same amount leaves thickness untouched.

d from tip	A thickness	F thickness	inner face shift
0.4 mm	0.997	0.991	−3.469
1.5 mm	2.029	2.029	−3.470
3.0 mm	2.692	2.692	−3.470
6.0 mm	4.802	4.802	−3.470
9.0 mm	6.412	6.413	−3.470
11.5 mm	11.347	11.347	0 — past the hook

A uniform −3.470 mm shift, which is 3.5 × cos 7.5° — the protrusion measured along the rail normal, projected onto x. The tip stays 0.99 mm. Mesh passes: shells 1, watertight, manifold, volume 25,110 mm³.

F — inward hook

whole 12 mm tip section translated 3.5 mm inward

The mating face and the tip are one surface. A square step where the hook rejoins the rail gives the 90° hook face.

One thing left: the mating face is not parallel to the mirror plane

F's mating face follows the inner rail, which sits 5.76° off the blade's own axis. Two mirrored blades therefore meet at an included angle of 11.53° — so a 12 mm face would touch at one end and gap by about 2.45 mm at the other. Face contact would still not be face contact.

Two ways to fix it: rotate the hole pair so the whole blade is mounted tilted, or draw the mating face parallel to the blade axis instead of to the rail. The second is far cleaner — it changes one function and touches nothing else — but which angle is correct depends on how the blade sits on the jaw, and that is the same unknown the HOLE_DX measurement resolves. This is why that page now asks for a third reading, the tip-to-tip distance.

Nothing has been changed in the repo

These two exist only as scratch meshes for looking at. The shipping Blade_v4_final.stl is untouched — still A, still byte-identical to what is committed.

The open question is whether the two blades need to meet across a face or whether a line is enough. The original field-test complaint was that contact area was too narrow and the cup slipped out, which argues for E. If the flat should be longer or shorter than 4 mm, that is one number to change.