OEM Watch Prototyping: From CAD to Sample Approval
OEM watch prototyping is a four-gate process: CAD and design-for-manufacture review, a physical sample built by a route that matches what you need to prove, a structured fit-and-function approval, and a frozen golden sample that production is measured against. A CNC case set typically takes six to ten weeks and costs roughly USD 800 to 3,000, depending on material, finishing and whether the dial, hands and strap are custom. The expensive mistakes are rarely in the drawing. They are in approving a sample that never proved the production process.
What a Prototyping Programme Has to Deliver
A prototype is not a marketing sample. Its job is to turn design intent into evidence: dimensions that can be measured, surfaces that can be finished repeatedly, and an assembly an operator can build rather than a modelmaker. If the sample answers only "does it look right", the programme has not started.
Ask for a document pack with the sample: a dimensioned 2D drawing stamped with the revision, a matching STEP model, a bill of materials with material grades, an assembly drawing showing the gasket and crystal stack-up, a finish specification with roughness values, and a test plan naming each measurement and instrument.
Set three approval tiers. T1 proves form, fit and basic function in a generic finish. T2 proves cosmetic execution, colour, brushing direction and assembly quality. The golden sample is the frozen T2 that both sides sign, and it becomes the reference for first article inspection and every later lot.
The CAD File: What the Factory Needs Before Quoting
Most quoting delays come from incomplete CAD, not factory workload. Send a STEP AP214 solid for the case, bezel, caseback and custom clasp parts, plus a 2D drawing with critical dimensions and geometric tolerances. A mesh file cannot be machined, and a drawing without tolerances cannot be inspected.
The case drawing should state material and temper, case diameter and tolerance, lug width and hole position, case height, crystal seat depth, gasket groove dimensions, caseback thread specification, crown tube torque, and the finish for each face. Mark which surfaces are cosmetic and which functional; a factory that does not know the difference will polish a sealing face that should have stayed turned.
Dial artwork needs vector outlines with colour references, not a JPEG. Specify print method, ink coverage, lume grade, application area and the date window tolerance. If the dial is sunray brushed or stamped, say so, because those processes change the minimum order and tooling cost.
Hands need a drawing each: length to the tip, pipe diameter, height above the dial, counterweight shape and lume mix. A hand set 0.1 mm too low will foul the crystal on a domed case and clear it on a flat one, and the fault only appears at assembly. Where the design reuses a standard movement, send the calibre number so the factory can work from its drawing for the holder and stem.
Design for Manufacture: Tolerances Before Aesthetics
A watch case is a tolerance stack before it is a shape. Critical fits usually sit at plus or minus 0.02 to 0.05 mm, while cosmetic dimensions can carry plus or minus 0.1 mm. Ask the factory which dimensions it will hold and which it will inspect, and get that list in writing before the toolpath is cut.
Water resistance lives in three places: the caseback gasket groove, the crystal seat and the crown tube. Groove depth and width, seat flatness and thread engagement matter more than wall thickness. A groove 0.05 mm too deep will not compress a standard gasket enough, and no amount of torque will fix it.
Internal radii matter as much as external ones. A square internal corner cannot be milled; it needs a radius of at least the cutter radius, typically 0.5 mm or more, or an EDM operation that adds cost and lead time. Undercuts, deep narrow slots and sharp lug transitions are the features that most often turn a two-week case into a five-week one.
Settle material before the prototype, not after. If the production case is 904L stainless steel, prototyping in 316L proves the geometry but not the machining behaviour or the polish response. The two alloys cut and finish differently, and the difference shows up as tool wear, cycle time and surface defects.
Choosing a Prototype Route
There is no single correct route. Choose by what you need to learn, because each one proves a different thing and hides a different risk.
Route | Best for, limits and lead time |
CNC from solid | Geometry, tolerance and finish in real material. Slow per batch; 2 to 4 weeks for a case set |
Resin or SLA print | Form, fit and wear feel only. Cannot be finished, sealed or weighed; 3 to 7 days |
Soft tooling and casting | Cosmetic sampling and small batches. No hard-tooled yield data; 3 to 6 weeks |
Hybrid: CNC case, printed dial | Early brand samples and photography. Dial print quality and lume life unproven; 2 to 3 weeks |
CNC is the default for a first functional sample. It gives you the real material, a real thread and a real gasket groove, so the sample can be pressure tested and worn. It is expensive per unit and slow per batch, but it answers the questions that matter before tooling is committed.
Resin printing is useful for shape and proportion, especially lug geometry and strap integration, but it is not a finishing sample. A printed case cannot be brushed, coated or pressure tested meaningfully, and its weight misleads you about wrist balance.
Soft tooling, usually a silicone mould with a cast polyurethane or low-melt alloy, sits between the two. It suits twenty samples for a sales meeting while the metal case is still being refined. Treat the cast sample as a cosmetic prop, not an engineering reference.
The Sample Timeline, Stage by Stage
A realistic CNC programme runs six to ten weeks from approved drawing to T2. Week one is drawing review and DFM feedback, which usually returns three to ten changes. Weeks two and three are material purchase and machining. Weeks three to five are finishing, where most schedules slip, because brushing, polishing and PVD are batch processes and a small prototype does not get priority on the line.
Weeks five and six are dial, hands and assembly. If the dial needs a new pad print screen or a stamped applique, add two to three weeks. Weeks six to eight are testing and the first review. If T1 produces changes, the T2 loop adds three to five weeks, because re-finishing and re-assembly take as long as the first build.
Plan around two calendar risks: Chinese New Year closes most watch factories for two to three weeks, with reduced capacity either side, and the European summer holiday slows brand-side approvals when the sign-off is yours.
What Sample Approval Should Actually Check
Approval should be a measurement exercise with a signature at the end, not a meeting where everyone agrees the sample looks good. Start with a dimensional report against the drawing, produced on a CMM or profile projector for the case and a toolmaker's microscope for small parts. Compare every critical dimension, not a sample of them.
Then measure what the drawing cannot express. Case mass in grams, because it drives wrist balance. Crown engagement in turns, and the torque needed to seat it. Bezel rotation torque and click count. Crystal height above the bezel, which is cosmetic and durability-critical at once. Lug hole position relative to the case centre line, which decides whether the strap sits straight.
Cosmetic inspection needs a defined light and distance: a brushed surface at 30 cm under a 4,000 K lamp shows grain direction that an office photograph hides. Agree that condition in writing and photograph the approved sample under it. Function checks belong in the same session: winding feel, hand-setting backlash, date change at midnight and midday, quick-set operation, chronograph reset if fitted, and rotor noise. These are cheap to check and expensive to discover at scale.
Fit, Function and Wear Testing
Wear the sample for at least seventy-two hours before approving it. Put it on a wrist matching the target customer, wear it under a cuff, and note where it catches. A case 1 mm too thick is invisible on a bench and obvious under a shirt sleeve.
Test the strap interface with the production strap, not a bench strap. Spring bar length, lug hole depth and strap thickness interact; a 22 mm lug with a 2.5 mm hole and a 4 mm strap will fit on the bench and bind on the wrist. Quick-release bars must clear the caseback.
For function, put the sample on a timegrapher in at least four positions and record rate, amplitude and beat error against the movement supplier's specification. ISO 3159 is the reference for chronometer-grade performance; a standard movement will not meet it without adjustment and, often, a different escapement.
Water Resistance and Finishing on a Prototype
Ratings follow ISO 22810, which defines what 3 ATM, 5 ATM, 10 ATM and 20 ATM mean and how the watch is tested. The rating is static pressure, not diving depth. A 3 ATM watch is splash resistant, 10 ATM suits swimming, and 20 ATM suits snorkelling and water sports. A dive watch needs a screw-down crown, a screwed caseback or a thick O-ring seal, and a test at 125 per cent of rated pressure.
Test the prototype by the method production will use: dry air pressure decay for the case, then a condensation or immersion check for the complete watch. Record pressure, dwell time and machine. A pass with no record is not a pass.
Finishing is where prototype and production diverge most often. PVD and DLC coatings are typically 1 to 3 micrometres thick, applied in a batch chamber, and colour shifts with thickness, substrate and roughness. A prototype coated alone in a development run will not match a production batch coated with twenty other parts. If colour is critical, approve a coupon from a production-intent batch, not the prototype.
Specify brushing and polishing direction per surface; a case brushed one way throughout looks wrong, and the error is invisible in a render.
Common Causes of a Failed Sample Approval
Most failures are predictable. Artwork drift comes first: the dial arrives with a font weight or date window position that differs from the approved file because the printer redrew it. Crystal height is second, where the stack of seat, gasket and crystal puts the glass 0.3 mm proud. Finish direction is third, a specification gap rather than a factory error.
Hand height is fourth. A hand set designed for a flat crystal will foul a domed one, and the collision only appears when the case is closed. Gasket material is fifth: a prototype cut from sheet nitrile seals differently from a moulded FKM gasket, and the difference appears in compression set rather than on day one.
Colour is sixth: anodised aluminium, PVD steel and painted brass will not match a Pantone reference under all lighting, and a match under a studio lamp can fail in daylight. Approve colour under two light sources with a documented Delta E tolerance. Bezel feel is seventh, since click torque and detent depth are set by the spring and machining tolerance. The movement holder is eighth: one made for a single calibre will not accept a replacement without a new drawing.
From Approved Sample to Golden Sample and Production
Once the T2 sample passes, freeze it. Write the revision number, date and signatures on the drawing, seal the sample, and photograph it under the agreed light from six angles. Keep one sealed sample at the factory and one with the brand. From then on, any change needs an engineering change note, a new sample and a new revision, even if the change looks cosmetic.
Before mass production, run a pilot batch of twenty to fifty pieces and inspect it as a production lot. This is where you learn whether the polishing line can repeat the sample finish, whether the dial printer holds colour across a run, and whether assembly hits hand height without the modelmaker. Inspect the pilot with the production AQL plan and record defects by code.
FAQ: OEM Watch Prototyping
How long does an OEM watch prototype take?
A CNC case set typically takes six to ten weeks from approved drawing to T2, including finishing and one review loop. A resin form model can be ready in under a week. If the dial needs new tooling or the case needs a forging die, add four to eight weeks.
How much does a watch prototype cost?
A CNC case set with a basic finish usually lands between USD 800 and 3,000, depending on material, complexity and whether bezel and caseback are separate parts. Dial printing, hands, assembly and testing are usually quoted separately. Treat any figure as a range to confirm with your factory.
Can I 3D print a watch case and call it a prototype?
You can print a case for form and fit, and it is a sensible way to check lug geometry before cutting metal. You cannot print a case and judge finishing, weight, water resistance or thread durability. Use the print to remove shape errors, then cut metal for the sample you approve.
How many prototypes should a brand approve?
Three tiers suit most programmes: T1 for form and fit, T2 for cosmetic and function, and a golden sample for production reference. More tiers add cost without adding information unless the design changes materially between them.
Do I need a new movement holder for a prototype?
Usually yes, if the calibre differs from the factory's standard. The holder, stem and dial feet are calibre-specific, and reusing a holder from another movement is a common cause of a dial that sits off-centre or a stem that binds.
Can a prototype be sold as a finished watch?
It can be sold, but it should not be represented as a production watch. A prototype has no production traceability, may use hand-fitted gaskets and has not been through a pilot run. If you sell samples, label them as samples and keep the records.
CAD Is Cheap, a Bad Sample Is Not
The cheapest part of a watch programme is the drawing; the most expensive is a production run built to the wrong reference. A prototype that is measured, worn and tested costs a few thousand dollars and a couple of months. A production run that repeats a sample error costs the tooling, the components and the season. Treat the sample as a specification, not a souvenir, and make every later decision answer to that frozen reference.
For the manufacturing steps that follow the sample, see our guide to watch case manufacturing from forging and CNC to finishing, the comparison of 904L and 316L stainless steel for watch cases, and the watch assembly and QC checklist. If you are still choosing a partner, start with the OEM watch manufacturer checklist.
Standards referenced: ISO 22810:2010 for water-resistant watches, ISO 3159:2009 for wrist-chronometers, and ISO 764:2020 for magnetic resistant watches.



Comments