Watch Assembly and QC Checklist for OEM Brands
An OEM watch assembly and QC checklist has to fix four things before production: the approved golden sample, the assembly sequence and its environment, the measurement gates with their sampling plans, and the records that prove each watch passed. In practice that means written tolerances for hand height and dial alignment, a defined caseback gasket and closing torque, rate testing on a timegrapher in at least four positions, water-resistance testing to ISO 22810 at the rated pressure, cosmetic inspection to an AQL plan under ISO 2859-1, and a batch record tying movement, case and gasket lots to the results. Without those four blocks, "QC passed" is an opinion rather than evidence.
What a Watch Assembly and QC Checklist Must Contain
Separate the document into two parts. The process checklist tells the factory what to do at each station: cleaning, dial mounting, hand setting, casing, sealing, testing and final cosmetic. The acceptance checklist tells both sides what counts as a pass, with a number, a method and a sample size. A checklist without numbers is a reminder, not a specification.
The gates below are typical rather than mandatory; agree sample sizes with your factory and inspection provider before production.
Gate | What is checked, how, and what is recorded |
Incoming components | Case dimensions, dial print, hand finish, gasket lot and crystal checked with callipers, pin gauges and 3× visual; AQL per lot, recorded by lot number. |
Pre-casing movement | Rate, amplitude, beat error and hand height on a timegrapher and height gauge; 100% or AQL II, with calibre serial and position results. |
Post-casing function | Winding, setting, date change and crown feel checked by hand; 100%, with operator sign-off. |
Water resistance | Dry pressure test at the rated pressure plus a condensation check in a warm bath; 100% at 10 ATM and above, AQL for 3–5 ATM, with pressure, duration and machine ID. |
Final cosmetic | Case, dial, hands, crystal, bracelet and lume at 30–40 cm under defined lux; AQL II, major 0.65 / minor 1.5, with defect codes counted. |
Pre-shipment | Carton audit, serial match and accessories from a random pull; agreed plan, audit report and photos. |
Fix the Golden Sample Before Assembly Starts
The golden sample is the reference every inspector compares against, so it has to be approved before the first production watch is assembled, not after. Seal it, photograph it under the same lighting the inspection bench will use, and keep one signed copy at the factory and one with the brand. Record the finish direction on brushed surfaces, the lume colour and texture, the dial print weight, the date disc font and the crown logo orientation.
The golden sample also settles arguments that a drawing cannot. Two brushed cases can meet the same Ra value and still look different because the grain runs at a different angle. A photograph under uncontrolled light will not resolve that; a physical sample will. Any change after approval needs a written engineering change note, a new sample and a revised revision number on the checklist. Otherwise the factory builds to last year's sample while the brand inspects against this year's expectation.
The Assembly Sequence, Stage by Stage
Assembly normally runs case preparation, dial mounting, hand setting, dust removal, movement casing, stem and crown fitting, caseback closing, function testing, pressure testing, final cosmetic inspection and strap fitting. Crystal and bezel pressing can happen before or after dial mounting depending on the case construction, but the sequence must be fixed in the process sheet so every line builds the same watch.
Case preparation is where sealing is won or lost. Clean the gasket seat, check the thread for burrs, confirm the caseback and case thread engage by hand before torque, and verify that the crystal seat is free of polish residue. Assembly cannot correct a case that arrived out of tolerance, which is why forging, CNC and finishing need their own inspection gates upstream. Dial mounting comes next: dial feet must sit square in the movement ring, the dial washer must not distort the dial, and the date window must align with the date disc at both midnight and midday. A dial that is 0.2 mm off is obvious on a clean dial and almost invisible on a textured one, which is why alignment is checked against the printed index, not the case edge.
Hand setting follows. Fit the hour hand, check it clears the dial and indices, then the minute hand, then stake the seconds hand flat. Blow the assembly clean before casing and cap it immediately; an open movement on a bench is a dust trap.
Dust Control, Dial Protection and Handling
Dust is the most common cosmetic reject in watch assembly and the easiest to prevent. Dial and hand fitting should happen under laminar flow or in a room of ISO 8 or better, with ionising blowers at the bench, no cardboard packaging in the assembly area, finger cots changed at every shift, and capped trays for work in progress. Compressed air must be dry and filtered; a contaminated line puts oil mist straight onto the dial.
Write the acceptance limit rather than leaving it to the inspector's mood: no visible particle under 3× magnification at 30 cm, with zero tolerance inside the date window and on the crystal underside. If a particle is found after casing, the watch is reworked, not wiped through the crown opening. Rework means reopening, replacing a compressed gasket, re-cleaning and re-testing water resistance. That cost belongs in the quotation, not a surprise debit note.
Hand-Setting Tolerances and Dial Alignment
Hands must be parallel to the dial, clear of each other and clear of the indices and crystal at every point of rotation. The exact height comes from the calibre and case drawing; what the checklist must state is where to measure and what the pass limit is. A typical production limit is 0.10–0.30 mm clearance between the highest hand and the crystal underside, measured at the tallest point, with no contact allowed at any position. Treat that as a starting range and confirm it against the actual case and crystal geometry.
Also specify seconds-hand staking height, no tool marks on the hand tops, and the date-change window: most contracts accept the date starting to move within about 30 minutes of midnight and completing within 60, though the calibre datasheet governs. Test at midnight and midday.
Caseback Sealing, Gaskets and Crown Function
The gasket is a wearing component, not a permanent seal. Specify the material, hardness and lot, inspect for moulding flash and cuts, and never stretch a gasket over a thread. Whether the case is 904L or 316L stainless steel, the sealing seat and thread quality decide the result. NBR, EPDM and silicone are the common choices, typically in the 50–70 Shore A range; the correct grade depends on skin contact, temperature and whether the watch will see chlorine. A light film of silicone grease helps seating, but excess grease attracts dust and can migrate onto the movement.
Closing torque must be written down and checked with a torque driver, typically 1.2–2.0 N·m for screw-down casebacks, with the exact figure set by the case design and gasket compression. Too little leaks; too much shears the gasket or strips the thread. For press-fit casebacks, specify press force and check caseback height after closing.
Crown function is tested before and after casing: winding must be smooth, the stem must engage setting and hacking positions cleanly, and the crown must screw down fully where applicable. Stem length is cut to the calibre, deburred and secured with the specified thread-locking compound. A stem that is 0.1 mm long loads the crown and produces the "tight crown" complaint six months later.
Movement Regulation and Rate Testing
Rate testing is meaningless unless the timegrapher lift angle matches the calibre and the watch is fully wound and settled. The calibre's own datasheet, and our Japanese automatic movement briefing checklist, should set the target before the first watch is tested. Test after 20–30 minutes of running, then again after 24 hours, and record each position rather than a single average. A common OEM acceptance window for a standard automatic is −10 to +20 seconds per day averaged across positions, with beat error at or below 0.5 ms and amplitude in the 250–310° range at full wind. Those are typical figures, not universal ones.
Chronometer claims need certification, not a tightened internal limit: ISO 3159 defines the testing behind the title. If the brief says COSC, the movement must come with the relevant certificate and number; a watch regulated to −4/+6 seconds a day on the bench is not a certified chronometer. For quartz, test voltage, current draw and that the seconds hand lands on the indices.
Water Resistance Testing After Assembly
Every watch rated 10 ATM or above should be pressure tested individually, and lower ratings should be tested to an agreed AQL sample at minimum. Use a dry overpressure test first so a failure does not put water inside the case, then a condensation test on a warmed watch against a cold plate. Any mist under the crystal is a fail, even if the pressure gauge held.
ISO 22810 sets the baseline for water-resistant watches, while ISO 6425 adds the diver-specific requirements, including a 100 m minimum and additional tests for legibility, shock and salt water. Our guide to water resistance testing sets out the test sequence in more detail. Testing must happen after the caseback is closed and before the strap is fitted, and the watch must be re-tested after any reopening, including a battery change or a rework. Record the pressure, hold time, machine identification and operator. A pressure test certificate without those fields cannot be traced to a specific watch.
Cosmetic Inspection and AQL Sampling
Cosmetic inspection needs a defect classification before it needs a sampling plan. Write down what counts as critical, major and minor, with photographs: a scratch on the crystal is major, a faint brush mark on the caseback is minor, a sharp edge on the case flank is critical because it is a safety issue. Then apply ISO 2859-1 at general inspection level II, typically AQL 0 for critical, 0.65 for major and 1.5 for minor. For high-volume programmes, tighter major limits are normal.
Inspect at a defined distance and light level, usually 30–40 cm under 500–1,000 lux, with 3× magnification for the dial and hands and 10× for edges and lume. Check bracelet articulation link by link, clasp closing force, end-link fit to the case, and lume evenness under UV.
QC Records, Batch Traceability and CAPA
A QC record should identify the watch, not just the batch: movement serial, case serial, dial and hand lot, gasket and crystal lots, line, operator, test equipment and calibration due date, and the result of every gate. Keep records for at least three to five years, longer with a long warranty, where a claim can be answered within a day.
Corrective action needs the same discipline. A failure without root cause becomes a repeat failure. Require containment, a root-cause statement, a corrective action with an owner and date, and an effectiveness check on the next lot. Put that obligation in the purchase order, with the right to audit the line and reject a batch that fails the agreed plan.
Pre-Shipment Audit: What the Buyer Checks
Do not audit the samples the factory prepared for you. Pull cartons at random from the finished goods area, open them in front of the factory, and match serial numbers to the QC record. Re-test a small number of watches for water resistance and rate, inspect the accessories, manual, warranty card and packaging, and photograph the carton labels and pallet count. If the factory cannot produce the matching record, the audit stops there.
A pre-shipment audit is the last cheap moment to catch a systematic error, such as a lot assembled with the wrong gasket or a dial print that drifted after sample approval. It does not replace in-process gates; it confirms they were real. Budget a third-party inspection on the first two shipments, then reduce frequency as the process stabilises.
FAQ: Watch Assembly and QC for OEM Brands
How many watches should be water-resistance tested?
Test every watch rated 10 ATM or above, and sample lower ratings to an agreed AQL plan. For a 3 ATM watch, a common start is a sample from each lot plus 100% testing on any lot that had a caseback rework. Agree the plan before production, not after.
What AQL should an OEM brand specify?
A workable starting point is critical 0, major 0.65 and minor 1.5 at general inspection level II under ISO 2859-1. Brands selling at higher price points often tighten major to 0.4 or 0.25. The number matters less than applying it consistently and recording the defects that were counted.
Who should write the QC checklist, the factory or the brand?
The brand owns the acceptance criteria; the factory owns the process detail. The brand writes the gates, tolerances and defect definitions, the factory adds station instructions and fixtures, and both sign. A checklist written only by the factory describes what the line already does.
How long should assembly and QC records be kept?
Keep them for at least three to five years, and for the full warranty period plus one year where the brand offers extended cover. The record must be retrievable by serial number, because that is the only way to answer a warranty claim without inspecting the returned watch first.
What is a realistic rate tolerance for an OEM automatic?
For a standard automatic, −10 to +20 seconds per day averaged across positions is a common production window, with beat error at or below 0.5 ms and amplitude around 250–310° at full wind. A chronometer claim needs a certificate; a tight bench average alone does not make one.
Can QC catch problems after the watches are cased?
Only some of them. Dust, hand contact and sealing faults are visible or testable after casing, but a pinched gasket may pass a short pressure test and fail in service. That is why in-process gates, 100% pressure testing on higher ratings and a batch record matter more than a final visual check.
A Checklist Is a Contract, Not a Form
The value of the checklist is the agreement it forces before tooling. Once the golden sample, gates, sample sizes and records are written down, a disagreement about a scratched dial or a leaking caseback becomes a conversation about a specification rather than blame. That is the difference between a factory that builds to your standard and one that builds to its own.



Comments