Michael Perry Résumé

Work/01

CALIBRE ZERO

AI-assisted mechanical watch engineering.

CALIBRE ZERO is an original mechanical skeleton watch that I'm developing using AI-assisted workflows, CAD and engineering analysis. It's an ongoing digital project. The design exists as a CAD model and has not been built or physically tested.

Status
Ongoing, digital design
Concept
C2 CHORD
Methods
CAD modeling, AI agent workflows, mechanical analysis, iterative design
Three-quarter view of the watch: round case, squared lugs, crown, and the movement visible through the dial opening

02 / 05Case, lugs and crown, with the movement visible through the dial opening.

CAD renders of the current C2 model. Use the tabs, the arrow keys, or swipe.

01Overview

Designing a mechanical watch movement

A mechanical movement stores energy in a mainspring, passes it through a train of gears, and releases it in small, regular steps through an escapement and a balance. All of those parts have to sit at precise distances from each other, supported by plates and bridges, inside a case small enough to wear.

That makes it a demanding mechanism and packaging problem, which is why I picked it. I also wanted to find out how useful AI tools are for mechanical design when their output has to hold up to engineering checks. CALIBRE ZERO is my way of exploring both.

02Design process

How I'm using AI in the design process

I split the work among specialized AI agents, each responsible for one part of the engineering. I treat what they produce as a proposal. Each proposal is evaluated with geometric and structural checks, and anything that fails goes back for another iteration.

The agents make it faster to explore options and revise the design. The checks, along with my own review, decide what stays in it.

  1. Explore concepts

    I explored several concepts before choosing one to develop further.

  2. Lay out the movement

    Position the barrel, gear train, balance and bridges so the parts mesh, clear each other and are properly supported.

  3. Generate CAD

    Turn the layout into solid geometry for the movement, case, crown and lugs.

  4. Evaluate

    Run geometric and structural evaluations on the model and send problems back for another iteration.

03Concept

The C2 CHORD concept

The concept I chose to develop is called C2 CHORD. Its bridges run in straight lines across the movement, like chords across a circle. That leaves the gear train, jewels and balance visible from both the front and the back of the watch.

Front render of the movement with five numbered markers
  1. 1BalanceUnder its own bridge at nine o'clock, held by two blue screws.
  2. 2Gear trainBrass wheels stepping across the upper half of the movement.
  3. 3BarrelThe largest wheel, next to the crown and winding stem.
  4. 4Chord bridgesStraight, skeletonized bridges with jewels at the pivots.
  5. 5Chapter ringHour and minute markers around the open center.

04Layout and CAD

Movement layout and CAD model

A skeleton movement shows its whole layout, so the layout has to work mechanically and look right. Most of the iterations so far have come from working through these questions:

  • Center distances between meshing wheels, which determine where each pivot can go.
  • Where a bridge can cross and still support the pivots beneath it.
  • Vertical stacking, so wheels, hands and bridges on different levels clear each other.
  • The path from the crown to the barrel for winding.
  • How much material can be removed from a bridge before it is no longer stiff enough.

Each iteration is updated in CAD, checked and revised. The renders on this page come from the current model.

Close view of bridges, jewels and gear teeth in the CAD model
Rear view of the movement through the open case back

05Current progress

Current progress and remaining challenges

The design currently includes the selected concept, a CAD model of the movement, case, crown and lugs, and a movement layout with the gear train, barrel and balance in place.

The main challenges left are mechanical:

  • Confirming that the gear train, escapement and balance would run as modeled.
  • Checking that the skeletonized bridges are stiff enough to support the pivots.
  • Defining tolerances and deciding how the parts would be manufactured.

I'll add analysis results to this page as that work is completed.

06Next steps

Next steps

Next I'm working through those open questions and refining the model. Once the analysis supports the design, I'll decide whether to build a physical prototype.