The Five-Watch Journey

A long-arc plan: each watch deliberately stretches a different skill so that by Watch 5 the full toolkit is in hand. Each build's spec lives in builds/watch_NNN/spec.md; this file is the map.

Philosophy

• Movement assembly is not the goal — movements are cheap and excellent from the factory. The craft is in case, dial, hands, crystal, finishing, and sourcing judgment.
• Every build teaches specific skills. The complexity escalates deliberately across five watches.
• These watches are worn. They are not shelf pieces. Office-capable, daily-wear quality is the standard.
• Wrist: 8 inches circumference. Comfortable case range up to 44–45mm. Sweet spot 39–42mm.
• Budget: not a constraint. Don't pay for brand names or marketing. Pay for excellent components at honest prices.

Prior experience: one kit build completed (Watch-Supply.co.uk West Coast Mayfair, NH38A, parts swap only). No prior independent sourcing or scratch assembly.

Skill progression — what each build adds

The skill ladder is one-way. Don't attempt Watch 3 before Watch 1 ships. Each build re-teaches what the previous build taught, plus one new thing. This is deliberate — the "one new thing" rule prevents a multi-failure cascade where you don't know which mistake caused the issue.

Tool acquisition timeline

Tools acquired for one build belong to all subsequent builds. Log them in tools/bench_inventory.md so future planning can assume them.

Before Watch 1 — bench bring-up (the big spend):

• AmScope SM-4TZ-144A stereo microscope with boom stand and ring light — order immediately, longest lead time of any bench item
• Bergeon 6899 screwdriver set (0.6 – 2.0mm)
• Dumont #3 and #5 anti-magnetic tweezers
• Crystal press with 26–32mm die range (model TBD — must match Watch 1 crystal OD once seat ID is confirmed)
• Horotec hand press set with plastic-tipped dies
• Bergeon 4040 movement holder (NH3x)
• Bergeon 7033 Rodico (replace annually)

Before Watch 2 — regulation comes online:

• Weishi 1000 timegrapher (deferred from Watch 1 because the Lucius regulated NH38A is no longer in spec — Esslinger NH38A is unregulated, so for Watch 1 either accept the factory −20/+40 spec or send it to a regulator. Acquiring the timegrapher before Watch 2 means by Build 3 self-regulation is the default.)
• Finer tweezers — Dumont #5 (already have) plus Dumont #7 for hairspring work if needed
• Anti-static mat / soft work pad (sunburst dials need a soft surface during dial-foot trim)

Before Watch 3 — inspection upgrade:

• Loupe-grade dedicated lighting for the microscope (the AmScope ring light is fine for assembly; an external goose-neck adjustable LED makes movement finishing visible at angle)
• A second movement holder sized for the new movement family if it's not NH3x (Miyota 82S5 has a different footprint)

Before Watch 4 — custom-work workflow:

• No new bench tools, but acquire: a calibrated USB caliper for measuring real-world tolerances; a dial print art-spec template (Adobe Illustrator file) ready to send to commissioning artists
• Optional: small bead-blasting cabinet for case finishing experiments — only if Watch 4 calls for it

Before Watch 5 — complications:

• Chronograph-specific oils (Moebius 9010, 9020, 9415 for chronograph applications — different than 3-hand)
• GMT bezel-insert seating dies if Watch 5 turns out to be GMT and the case requires an insert press

Calendar — realistic per-build duration

Each build has three serial phases: sourcing, bench, wear-test. They cannot be parallelized — sourcing must complete before bench, and bench must complete before the wear-test can declare a watch "done."

Total expected calendar from Watch 1 first parts ordered → Watch 5 shipped: ~14–21 months. This is realistic for someone who isn't full-time on this. Compress only if (a) a build's sourcing is actually faster than estimated, or (b) you decide an off-ramp (see below).

The wear-test is non-negotiable. A watch isn't done because it ticks. It's done because it survives 14 consecutive days on the wrist without losing time outside spec, without crystal-seal failure, without a hand falling, without a crown popping. A build that fails its wear-test goes back to bench — it doesn't ship.

Risk register — by build

Watch 1

• Highest-likelihood failure: dial foot misalignment after trim (snipping the wrong set of feet; trimming both sets accidentally; misjudging the cut angle). Dial is a $29 part — recoverable, but the rebuy is a 4-week lead time.
• Second highest: crystal seat tolerance mismatch — Tandorio crystal seat ID hasn't been confirmed yet; the email is drafted, not sent. Ordering crystal before this is confirmed risks pressing-in a crystal that won't seat or that cracks under pressure.
• Third highest: hand setting — first hand-setting on a real movement is the highest-error operation in the whole build. Plastic-tipped Horotec press is the mitigation.
• Mitigation strategy: trim feet under microscope with the dial taped to the soft mat, dry-fit before pressing, hand-press with the gentlest die that contacts the tube cleanly.

Watch 2

• Highest-likelihood failure: fingerprint or scratch on a sunburst dial. Sunburst dials show everything — even glove residue.
• Second highest: applied indices coming loose — applied indices are glued; bumping them during dial-foot work is fatal.
• Mitigation: dial work happens under finger cots, on the soft mat, with the dial loaded into the case as the very last step before crystal press. Index check under microscope before crystal press.

Watch 3

• Highest-likelihood failure: open-heart cutout fouls the dial-side rotor or escape wheel — i.e., the cutout was sized for the wrong movement variant. Verify cutout dimensions on the dial before the dial ships.
• Second highest: poor movement finishing makes the open heart underwhelming — the design depends on the movement looking good. Source the movement before sourcing the dial cutout.
• Mitigation: photograph the movement under angled light before approving the dial cutout; reject the movement if the rotor or bridges look unfinished.

Watch 4

• Highest-likelihood failure: custom dial print quality — colors don't match the proof, or the print shifts on the dial blank. There's no second dial coming for a month.
• Second highest: custom CNC case dimensions don't match standard movement seat or crystal seat. CNC shops don't always understand watch-internal tolerances.
• Mitigation: order a printer's proof / sample dial first (often free); for CNC, send the shop a finished sample case as the dimensional reference, not just numeric specs.

Watch 5

• Highest-likelihood failure: chronograph wheel-train re-oiling — chronograph movements have specific oil points that shouldn't be touched without the right oil set. Wrong oil = wrong viscosity = pusher feels weird.
• Second highest: GMT bezel insert mis-seats — bezel inserts have a specific orientation and pressure. Wrong seat = the GMT hand reads wrong even when the movement is fine.
• Mitigation: for chronograph, regulate without touching the chronograph oils unless the pre-purchase service paperwork says they were freshly done; for GMT, dry-fit the bezel insert with a press at minimum pressure first.

Off-ramps — when to slow down or simplify

You do not have to do all five builds in order. The goal is a craft, not a checklist. Reasonable off-ramps:

• After Watch 1: if independent sourcing felt overwhelming, the right move is Watch 2 but with a shorter sourcing list (re-use Watch 1's case supplier, etc.) before tackling Watch 2's dress-watch precision demands.
• After Watch 2: if precision was the wall, repeat Watch 2's profile with a different movement before moving to open-heart territory.
• Before Watch 4: if commissioning custom feels like a leap, do a "Watch 3.5" — a non-stretch build between Watch 3 and Watch 4 that re-uses a lot of Watch 3's sourcing decisions, just to compound bench experience before the commissioning step.
• Before Watch 5: if complications feel like a bridge too far, consider a more refined three-hand build — premium movement, hand-finished case, exhibition back — instead. The skill that Watch 5 teaches (research and problem-solving beyond three-hand) is valuable, but it's not the only valuable skill.

The five-watch frame is a roadmap, not a contract. The goal is excellent watches and accumulated craft.

Cross-cutting themes

• Each build adds at least one entry to sourcing/compatibility_notes.md. Every supplier interaction yields a line of compatibility intelligence — every dial foot mismatch, every crystal seat tolerance, every drop-shipping lead time surprise.
• Each build leaves sourcing/parts_database.md richer than it found it (including parts that didn't make the cut — "rejected with reasoning" entries are often more valuable than "good" entries).
• Tools acquired for one build belong to all subsequent builds — log them in tools/bench_inventory.md so future planning can assume them.
• Each build appends a ## [YYYY-MM-DD] research | … or decision line to LOG.md for every non-trivial sourcing pass or judgment call.
• After each build ships, builds/watch_NNN/notes.md gets a "lessons" section. Promote anything that generalizes to the cross-build files.

Quick links per build