# Copic split-collet — concept 01

A separate three-part experiment: a holder body with a tapered seat, a removable
four-finger collet, and an end fitting that screws into the body and pushes the
collet into the seat. The taper turns axial motion into inward movement of the
fingers around the pen barrel. This is my interpretation of the supplied drawing,
not a dimensional reconstruction of its original parts. Industrial collets use
the same general taper-closing principle; their published performance does not
apply to this printed design. [Royal Products](https://royalproducts.com/product-line/royal-cnc-collet-chucks/)

Open `review.html` for the illustrated review. `split_collet_concept.blend` contains
three scenes: exploded mechanism, assembled cutaway, and bench parts in print
orientation. The charcoal object is only the measured barrel envelope; its nose
and ink tip are deliberately omitted from the 3D assembly.

The subsequently supplied full PHP31-FIBER manual clarifies the use of a separate
height-setting jig. See [manual findings](manual_findings.md). Its setup sequence
supports an adjustable grip and height datum, while the exact internal taper and
thread geometry remain unspecified.

## What is ready to review

- Actual mating helical geometry, not decorative rings: a custom 2.40 mm pitch
  trapezoidal thread, male diameters 15.20 / 14.00 mm, with 0.25 mm nominal radial
  clearance in the female thread. It is **not a standard M15 thread**.
- Four 0.80 mm slots with 0.40 mm rounded ends, a 12.00 mm slotted length and
  a 2.00 mm closed band. The insert is 14.00 mm long and 13.70 mm maximum OD.
- Relaxed gripping ID 10.25 mm, relieved to 10.60 mm at the closed band. The body
  guide and cap bores are 10.60 mm. These are trial CAD clearances, not measured
  printed dimensions.
- A 9.93° half-angle taper. Ideal taper motion gives 1.14 mm axial travel for a
  9.85 mm barrel and 1.43 mm for a 9.75 mm barrel, starting at nominal seat contact.
  This calculation assumes uniform radial contraction; it is not a flexure or
  clamping-force simulation. Printing error and real finger bending change it.
- Approximately 2.8 full turns of effective thread overlap at the loose datum,
  increasing as the cap advances. The starting cap/body gap is 2.00 mm. The cap
  should grip before its head contacts the body; head contact is not a grip target.
- Minimum nominal radial walls: 1.15 mm at the body thread root, 1.70 mm at the
  cap thread root and 1.025 mm at the collet tip. These are geometry checks, not
  proof of strength. Thin regions cannot accommodate every requested wall loop.

The slotted fingers remain connected by the upper band. Their flexibility comes
from their geometry. Reducing infill is not a substitute for establishing a
repeatable flexure. This first insert has no positive extraction feature: after
unscrewing the cap, the fingers must relax, or the pen and insert must be gently
withdrawn together by the barrel. Taper sticking and fatigue are explicit bench
test items; an extraction feature is a possible later refinement.

## Carriage dimensions retained

| Feature | Dimension |
|---|---:|
| Main insertion body | 13.30 mm OD |
| Local contact collar | 14.50 mm OD |
| Enlarged collar depth below flange | **3.20 mm maximum** |
| Maximum flange / proposed upper head / cap | 18.00 mm OD |
| Flange contact face to old printed nose face | 28.63 mm |
| Intended flange contact face to ink tip | 33.12 mm |

The existing narrow waist and removed intermediate ring stay as in V14. The new
mechanism sits above the flange. Its loose cap top is 30.70 mm above the flange,
**22.70 mm above V14's pen entry**. That extra height and the 18 mm head envelope
must be checked against the moving carriage and nearby machine structure.

The full body in the Blender scenes retains an old nose cone for context. It is
not a finished universal pen seat. The distances from the ink tip to the starts
of the 3.04, 4.03 and 5.71 mm nose sections remain unknown. The 15.70 mm tip-to-full-
barrel length is recorded as supplied/interpreted in the main task. A small outlet
must also allow enough axial adjustment for each pen's nose geometry.

## Bench prototype files

These three STLs are **untested bench prototypes**, supplied at 100% scale in mm:

| File | Bed contact / orientation | Size (mm) | Sliced material |
|---|---|---|---:|
| `bench_body.stl` | Large threaded mouth down | 18 × 18 × 30.70 | 3.38 g |
| `collet_10_25.stl` | Closed band down; finger tips up | 13.70 × 13.70 × 14.00 | 1.00 g |
| `threaded_end_fitting.stl` | Wide cap face down; male thread up | 18 × 18 × 14.70 | 1.87 g |

The bench body has an open bottom and only the mounting region below the flange.
It tests the grip independently of the unresolved nose. It is not a complete
holder for plotting. No full-holder STL or machine print job has been supplied.

The isolated Bambu Studio 02.08.02.61 slices use the copied A1 mini / 0.4 mm nozzle /
Generic PLA profiles, 0.16 mm layers (0.20 first layer), four requested walls,
15% infill, supports off, and a 3 mm outer brim with 0.10 mm gap. These are test
settings, not a new confirmation of your installed material and nozzle.

All three slices succeeded, contain no support toolpaths, and retain 192, 87 and
92 layers respectively. Separately sliced model times are 28m 29s, 12m 53s and
14m 13s; each separate job has additional startup time. Total estimated material
is 6.25 g including those brims. No print has been sent.

Mesh checks found one closed connected component per STL, positive volume and no
non-manifold edges. The steepest downward-facing mesh slopes away from the bed
are approximately 45.00° on the body/cap threads and 1.67° on the collet. No such
face exceeds 46°. This supports a support-free trial; it does not establish actual
print quality, thread fit, or slot clearance. Orientation, tolerances and material
behavior still matter. [Prusa design guidance](https://help.prusa3d.com/article/modeling-with-3d-printing-in-mind_164135)

## First mechanical test

1. Check the unloaded thread pair by hand. Remove brim remnants from mating
   faces and inspect the slots. The cap must run freely before adding a collet.
2. Drop the collet into the seat with the fingers toward the body. Put the cap
   over the pen barrel and insert the pen through the collet into the open bench
   body. Check free sliding before tightening.
3. Hold the barrel and turn the cap in small increments, checking for grip and
   release. Stop at a light, secure grip. This open bench sample has no tip datum;
   support the pen by its barrel during the grip test.
   Record whether the 9.75 and 9.85 mm barrels both grip, whether the cap bottoms
   first, whether the insert releases, and whether repeated cycles damage it.
4. Before a full holder print, confirm upper carriage clearance and finish the
   nose passage. With a finished holder, set the flange-to-ink-tip distance to
   33.12 mm. The full Graphtec manual uses gentle tip contact with a separate
   height-setting jig. A jig for our holder should reference its flange contact
   face and put the datum 33.12 mm below it; that dimension is the user's target,
   not a dimension from the manual. Avoid intentional nib deflection, hold the
   barrel during tightening and remeasure afterward: a pushed collet can drag the
   pen axially. This is not a fixed-length collet mechanism. A matching jig has
   not yet been modeled.

The supported adjustment range, axial holding force, lateral stiffness, creep,
thread durability and repeatable release are not physically established. No
universal-fit claim is made. All work is confined to this experiment; copied
reference hashes were checked and the main live Blender session was not used.