Watch Case Manufacturing: From Forging and CNC to Finishing

Most OEM stainless steel watch cases are produced by closed-die forging, CNC machining from solid billet, or stamping and deep drawing, then finished by tumbling, brushing, polishing and optional PVD or DLC coating. Forging suits medium-to-high volumes and sculpted shapes; CNC suits low volumes, prototypes and fast iteration; stamping suits very high volumes of simpler cases. The right route is decided by volume, geometry, tooling budget and surface finish, not by which process sounds more premium. This guide covers how each route works, where cost sits, the tolerances that decide assembly, and what to put in the brief.

What the case has to achieve

Before comparing processes, fix the requirements. A case must hold the movement, seal against water and dust, carry the crystal and caseback, accept the crown and stem, provide lugs at the right width, and survive daily knocks. Radii, chamfers, brushing direction and polish quality are part of the product, not decoration added afterwards.

The brief should state the movement calibre and holder, case diameter and thickness, lug width and spring-bar position, crystal type and thickness, caseback type, crown type, water-resistance rating, finish on every surface, material grade and expected annual volume. Without those inputs a factory will quote a generic case, and the quotations will not be comparable. The pre-production checklist in our guide to choosing an OEM watch manufacturer covers the commercial side; this article covers the process side.

Forging: near-net shape and grain flow

Closed-die forging heats a billet of 316L (UNS S31603, 1.4404), the grade described in Aperam's 316L product data, to roughly 1,150–1,250°C, places it in a die cavity and presses or hammers it into a near-net shape. The steel flows along the die contour, so the grain structure follows the part rather than being cut through. That is the metallurgical argument for forging: better toughness and fatigue resistance than a part machined from bar, because grain flow stays continuous around corners and lugs. For a watch case the practical benefit is a stronger lug and a more reliable thread boss, not a dramatic weight saving.

Forging is not a finishing process. The shell still needs machining for the crystal seat, caseback thread, lug holes, crown tube and sealing surfaces, then full finishing. Its advantage is starting close to final shape, so less material becomes swarf. Its disadvantage is tooling: a die set is specific to one design, takes weeks to make and is expensive to change.

Ask for the die life and the refurbishment plan. A die that has run its rated strokes wears at the radii and parting line, and the flash line can migrate into a visible surface without maintenance. A credible forging supplier will state the die material, the number of strokes between refurbishments and how it controls billet temperature, because cold spots cause laps and cracks that only appear after machining.

CNC machining: freedom at the cost of cycle time

CNC machining starts with a block, slab or bar and removes material until the case emerges. Three-axis milling handles the outline, lug slots and caseback recess; five-axis machining or a turn-mill centre handles angled surfaces, crown guards and complex bezels in fewer set-ups. The route is the default for prototypes, small batches and designs that are still changing, because there is no die to cut and the program can be revised in hours.

The price is paid in cycle time and material. A medium-complexity steel case can spend 45–120 minutes across roughing, semi-finishing, finishing and drilling; complex cases take longer. Much of the billet leaves as swarf, and thin walls move as residual stress is released. Good machinists rough, stress-relieve, then finish rather than holding final tolerance in one pass.

Stainless steel is not aluminium. 316L work-hardens, conducts heat poorly and tends to built-up edge, so Sandvik Coromant's guidance on stainless machining matters: sharp edges, adequate coolant, controlled chip load and a program that does not rub the surface. A workshop that cuts aluminium well will not automatically hold ±0.02 mm on a thin 316L wall.

CNC also makes economic sense as a hybrid. Many factories forge or stamp the middle, then CNC the critical features and the bezel. That combination buys the material efficiency of forming and the precision of machining where it is needed.

Stamping and deep drawing: the high-volume route

Stamping and deep drawing form a flat disc or sheet into a cup or shell with a press and progressive dies, then trim, pierce and machine the critical details. The route is fast and repeatable at high volume, and unit cost falls once tooling is amortised. It is common for casebacks, bezels, inner rings and simpler case bodies, and it is the natural choice for a mature product that will run for years.

The limits are geometric and financial. Deep drawing cannot produce the sharp internal corners, undercuts and thick-to-thin transitions of a sculpted sports case without extra operations, and every design change can mean a die change. Draw marks, thinning at the radius and springback are process realities that must be designed around. Tooling is cheaper than a forging die but far from free, and the minimum economic run is usually in the thousands, not the hundreds.

For a first collection, stamping is rarely the right starting point. It becomes attractive when a best-selling shape is stable, forecast volume is firm and the brand wants to take cost out of a proven design.

Comparing the routes

Route	Best volume	Tooling	Material waste	Typical tolerance	Best for
Closed-die forging	Medium to high	High, design-specific die set	Low to medium	±0.05 mm general, ±0.02 mm machined features	Sculpted cases, strong lugs, long programmes
CNC from solid	Low to medium	Low, program only	High, often over half the billet	±0.02 mm critical, ±0.05 mm general	Prototypes, small batches, complex geometry
Stamping and deep drawing	High	Medium, progressive dies	Low	±0.05 mm formed, ±0.02 mm machined	Simple shells, casebacks, mature high-volume designs
Hybrid forge or stamp plus CNC	Medium to high	Medium to high	Low to medium	±0.02 mm on machined features	Most OEM sports and dress cases

The tolerances are typical starting points, not guarantees; they depend on part size, wall thickness, machine condition, fixturing and inspection method. A supplier promising ±0.01 mm everywhere without discussing the feature and gauge is promising a number, not a capability.

Tooling, MOQ and the cost curve

Tooling is the hinge on which the route decision turns. A forging die or a progressive stamping die is a fixed cost that must be recovered over the order, so the quoted unit price is only meaningful alongside the tooling charge, the amortisation assumption and the volume band. Ask for price breaks at your forecast volumes and what happens to the tooling if you move supplier. Ownership of tooling should be agreed in writing, with a condition report at handover.

The minimum order quantity follows the route. CNC shops can start at tens of pieces because the only set-up is the program and fixtures. Forging and stamping need enough volume to justify the die, so the economic minimum is usually in the hundreds to thousands depending on complexity and how the factory loads its presses. A factory that quotes a very low MOQ on a forged case is either absorbing tooling into the unit price or planning to machine from solid and call it forging; both are worth clarifying.

There is also a learning curve. The first articles from a new die or programme carry the highest risk of dimensional drift, porosity at a corner or a finishing mismatch. Budget for a pre-production sample, a pilot run and a first-article inspection before the full order. A pilot costs far less than a container of cases that will not accept the crystal.

Tolerances that decide whether the watch assembles

The tolerances that matter most are not the cosmetic ones. They are the interfaces. The crystal seat must be flat, concentric and at the right depth, or the crystal will not seal or will sit proud. The caseback thread must engage smoothly over its full length, with the right pitch diameter and no burr at the start. The crown tube bore must be concentric with the stem, or the crown will bind and the seal will wear. The lug holes must be parallel and at the correct spring-bar span, or the bracelet will not fit.

A practical drawing will call out each critical feature with its tolerance, datum and inspection method. Diameter and roundness of the crystal seat, perpendicularity of the caseback face, position of lug holes, thread gauge and surface roughness are the usual candidates. General cosmetic surfaces can carry a looser tolerance and a finish specification instead.

Ask for the first-article inspection report with actual measurements, not a tick sheet. A CMM report on the case, a thread gauge result and a roughness trace on the brushed and polished surfaces tell you whether the process is in control. If the factory cannot produce those, the tolerance conversation is theoretical.

Finishing: from tumbling to PVD and DLC

Finishing creates most of the perceived quality and holds most of the labour. The sequence starts with deburring and tumbling, then brushing or polishing, then cleaning and, if specified, coating. Brushing uses abrasive belts or wheels for a directional satin surface; polishing uses compounds and mops to remove scratches and bring up lustre. Masking steps keep brushed and polished zones distinct.

Surface roughness is the measurable part of the specification. A typical brushed finish sits around Ra 0.4–1.0 µm, while a mirror polish is well below that, often under 0.1 µm. Agree the target on a physical sample: Ra alone does not capture the direction and uniformity of brush lines the eye reads as quality.

Coatings come after finishing and cleaning. PVD is a vacuum deposition process that lays down a thin ceramic or metallic layer, typically 1–3 µm, in colours from steel and gold tone to black. DLC, a form of carbon coating, is typically 1–4 µm and is chosen for its hardness and low friction rather than for colour alone. Both depend on substrate preparation: a coating over a poorly cleaned or poorly polished surface will show the defect and may fail adhesion. If the case is titanium or ceramic rather than steel, the finishing route changes; our guides to titanium watch cases and ceramic cases and bezels cover those material-specific sequences.

Inspection and quality control

Inspection should be built around the interfaces and the finish, not only the final visual. Incoming material needs a mill certificate and, for critical programmes, positive material identification to confirm the grade. In-process checks catch drift before a batch is finished: dimensions after machining, thread gauges after tapping, roughness after brushing and before coating. Final inspection adds water-resistance testing, a visual under controlled light and a function check with the movement fitted.

Water resistance is tested to a defined standard, not to a feeling. ISO 22810:2010 sets the requirements and test methods for water-resistant watches, and ISO 6425:2018 covers divers' watches. The test method, sample size and acceptance criteria should be written into the order. A pressure test on one sample from a batch is evidence about that sample, not proof that every case is sealed.

For steel, a salt-spray test to a recognised method is a useful screen for finish and passivation quality, but not a universal pass mark. Agree duration and acceptance criteria with the brand and finish supplier; brushed, polished and coated surfaces behave differently. The same applies to PVD and DLC adhesion tests: put the method and acceptance level in the specification.

Writing the case brief

A good case brief is a drawing plus a schedule. The drawing defines geometry, datums and tolerances. The schedule defines material and grade — our 904L versus 316L comparison explains where the higher-alloy grade earns its premium — plus finish by surface, coating, water-resistance rating and test standard, crystal and caseback interface, crown type, expected annual volume, tooling ownership and the approval process. It should also state what is not acceptable: visible tool marks, polish over a brushed zone, a thread that catches, a mismatch between case and bracelet finish.

Send the same brief to every supplier. If one quotes a different process, ask why. A forged quote against a CNC quote is not like-for-like unless finish, tolerance, inspection and tooling terms match; the cheapest quote is often the one with the least inspection.

Finally, keep a golden sample. A signed, finished case at the agreed finish is the reference for every later batch and the fastest way to settle a dispute about what brushed or black means.

Frequently asked questions

Is forging better than CNC for a watch case?

Not inherently. Forging gives a near-net shape with continuous grain flow and less machining, suiting medium-to-high volumes and sculpted designs. CNC gives freedom and fast changes at low volume but wastes more material and takes longer per part. Match the route to volume, geometry and tooling budget.

Can a forged case be machined to tight tolerances?

Yes. Forging produces the shell; machining produces the interfaces. Crystal seats, threads, lug holes and sealing surfaces are cut after forging, so critical tolerances come from machining and inspection, not the die.

What tolerance should I specify on a watch case?

Specify tolerances by feature, not one blanket number. Critical interfaces such as the crystal seat, caseback thread and lug holes may need ±0.02 mm or tighter; cosmetic surfaces can be looser. State the datum and inspection method so the supplier measures what you will.

How many cases do I need for forging to make sense?

It depends on die cost and factory loading, but forging usually needs hundreds to thousands of pieces to amortise tooling. Below that, CNC from solid is normally cheaper. Ask for tooling separately to see the break-even volume.

Does PVD or DLC change the case dimensions?

A coating is thin, typically 1–4 µm, but it lands on every exposed surface and can affect a precision fit already at its limit. Tell the coater which surfaces are fits, mask them, and inspect the coated part.

How do I check the finish before approving mass production?

Approve a golden sample under the lighting the product will be sold in, and record roughness and gloss readings with it. Check the first production batch against it and repeat at each batch. Photographs are no substitute for the sample.