← taylorinalaska.com

the real Fischer Easy Skin connector on a ski
the real part
the finished v28 CAD model of the receiver
v28 — finished CAD

Rebuilding this ski-binding clip as editable CAD, from photos alone. The black receiver was the hard half — it took 28 revisions.

Reading a Ski-Skin Connector

Rebuilding a Fischer Easy Skin connector in Fusion 360 — and the one habit that would have saved 25 of the 28 tries.
Thanks for the inspiration, Matthew! 🙏

A climbing skin is the grippy strip you stick under a backcountry ski to walk uphill; a small plastic connector clips it to the ski tip. I tried to rebuild that connector as clean, editable CAD — working only from a rough 3D-print file and photos. I drove an AI agent that runs Fusion 360; I supplied the eyes and the corrections.

Two halves

A white insert bonded to the skin, and a black receiver set into the ski. The insert was quick — I imported the print file as an exact solid in a few tries. The receiver was the problem: 28 versions, nearly all of them me misreading how it locks.

top-down photo of the connector showing the receiver
The receiver I was chasing — the black part cradling the insert's wide end.

Where it went wrong

Every wrong turn was a sentence read on the wrong axis. I heard "enclosed" and built a roof — it meant enclosed in the top-down view, open on top. I assumed the lock was a click; it's a tapered ramp, held by the tension of the skin's glue. Matthew's circled photos, not my reasoning, fixed each one.

annotated photo circling the locking ramp annotated photo circling the engaging ramp
One circle was worth a hundred of my words.

How it evolved

■ red a wrong turn  ·  ■ amber getting warm  ·  ■ green right idea. Oldest to newest; full notes in the render gallery.

The receiver — v3 → v18 (the long road)
1rec v3
v3
carved blob, wrong feature
2rec v4
v4
a slide-through tube — wrong
3rec v5
v5
cleat over the top — wrong
4rec v6
v6
wrong end seated
5rec v7
v7
correct end at last
6rec v9
v9
no more roof
7rec v12
v12
first real cradle
8rec v13
v13
cradle reaches the ramp
9rec v15
v15
enclosed — but a wrong roof
10rec v16
v16
open cradle
11rec v17
v17
mates the full ramp
12rec v18
v18
caps the ramp (oversized)
The receiver — v26 → v28 (locked, trimmed, rounded)
13rec v26
v26
open cradle + side hooks — locked
14rec v27
v27
on top only at hooks + tip
15rec v28
v28
rounded to match — the keeper
The insert — v1 → v6 (mesh to exact solid)
1ins v1
v1
no barbs yet
2ins v2
v2
barbs + OEM colors
3ins v3
v3
9 ribs
4ins v4
v4
barb cut silently failed
5ins v5
v5
barb finally visible
6ins v6
v6
exact solid — the keeper

The lesson

The two photos that finally cracked it — the receiver by itself, and a clean top-down — were Google Images results the whole time. I spent over twenty versions inferring the black part's shape from the white one it clips into, when a photo of the part itself settled it in three.

When you're rebuilding a real object, get a clear photo of the exact part — from a few angles — before you model anything. The reference is the truth; your reasoning about a part you can't see is not.

The CAD program was never the bottleneck. Modeling on a guess was, while the answer sat one image search away.