What Makes Lock Parts a Distinct Finishing Case

Lock parts are small, often under a hundred grams, and run in enormous variety — bolts, cylinders, cases, roses, strike plates, and keys, frequently sharing a line. Variety is the first design constraint.

The second is the edge split. Outside edges are cosmetic and want a soft, consistent break; inside edges are functional and want to be burr-free without being over-radiused. A chamfer that flatters the eye can wreck a keyway.

Third, materials mix: zinc die-cast bodies, brass cylinders and bolts, steel strikes. Each alloy answers a different tool and force, so one path rarely fits all.

The Defect Catalogue

The table below lists the failure modes that recur across lock-hardware cells. Read it left to right: what you see, why it happens, what fixes it.

Two rows deserve emphasis. The bolt-stick and key-drag rows are functional failures — they escape cosmetic inspection and reach the customer as a lock that does not lock. Both trace to path exits and force, not to machine capability.

The zinc-smear row rewards attention: smeared alloy hides porosity from later inspection, converting a visible reject into an invisible one. Sharp tools are a quality control, not just a productivity item.

Edge Strategy: Cosmetic Outside, Functional Inside

Split every part drawing into edge classes before programming. Outside corners and show perimeters take the cosmetic budget: a consistent radius, achieved with brushes and light passes.

Functional edges — keyways, slots, pin channels, bolt tails — take the functional budget: burr-free with minimal radius change. These edges want small tools, controlled exits, and verification by feeler or function test, not by eye.

Program the exits deliberately. A tool that lifts straight out of a slot drags a burr with it; a lead-out that sweeps along the wall leaves a clean edge. Three extra seconds of path prevents the worst field failures in this industry.

Stability Monitoring Points

Running cells stay stable by watching five points. Each is cheap to monitor and each catches a catalogue row before it becomes a batch.

• Force log, slot passes. Rising force means tool loading; falling means wear. Alarm on either trend.
• Edge-break gauge, first part each nest. A five-second comparator check per nest catches offset drift instantly.
• Function test, bolt-in-case, hourly. One bolt cycled ten times beats a hundred cosmetic checks for finding functional burrs.
• Tray integrity, per shift. Worn nests cause half the cosmetic defects in the catalogue; inspect the trays, not just the parts.
• Tool hours versus parts. Parts-based tool life beats hours-based when mix varies — a steel strike hour wears tools ten zinc-part hours’ worth.

Put the five monitors on one page at the cell. Operators who can see all five states at a glance run tighter processes than any dashboard can compel.

Best Practices From Running Cells

The habits below separate plants that run years without field failures from plants that firefight monthly. They cost little and compound.

Segregate alloys by station, not by schedule. Zinc dust and steel sparks in one extraction path is a maintenance and safety problem. Dedicated stations for steel strikes pay for themselves in filter life alone.

Keep two nests per high-runner. When one wears past limits, swap in the fresh one and send the worn to re-lining — never run a worn nest “one more shift”. The catalogue’s rattling-marks row is always a worn nest at the end of its shift.

Verify with function, not just appearance. The hourly bolt cycle test catches what gloss meters miss. Lock plants that skip it eventually learn the lesson from customers.

Program change discipline. Lock-part programs multiply across SKUs; keep revisions stamped and backed up. A program reverted to last month’s version is a subtle batch defect generator.

For the setup and cell-building side, robotic cell setup for castings remains the best walk-through, and the same discipline applies at lock-part scale.

Consumable Discipline on Small Parts

Small-part finishing burns tools in deceptive ways. A brush that looks fine can be thirty percent worn, and thirty percent wear moves edge breaks visibly on parts this small.

Anchor tool life to parts counts per alloy, reviewed monthly. Zinc counts and steel counts belong in separate ledgers; averaging them is how offsets end up wrong for both.

Compound for lock show faces earns its own stock discipline. Brass and zinc share compounds more easily than either shares with steel, and cross-contaminated compound produces the streak defects the catalogue lists.

Keep dressing in the schedule. A dressed wheel at moderate life outperforms an undressed wheel at fresh count, and dressing costs one minute against hours of rework. Maintenance scheduling for uptime applies directly.

When to Escalate Backwards to the Casting

Some catalogue symptoms are finishing problems wearing casting costumes. Recurring porosity on a case face, or a parting line that always flashes thick, will consume finishing effort forever.

Track defects back to die position. A flash zone that needs a heavy pass this month needed a heavier one last month; the die is wearing and the fix belongs in the tool room.

Feed the defect log to the casting team monthly. Plants that close this loop watch finishing costs fall quarter over quarter; plants that treat casting and finishing as separate kingdoms pay for the same flash twice.

Changeover Efficiency for High Variety

Lock plants run dozens of SKUs through one cell, and changeover minutes multiply by frequency into whole shifts. A cell that changes over in twelve minutes loses two percent of the day; at forty minutes it loses seven — the difference is a shift per week.

Nest design is the lever. Group parts by edge-class profile rather than by SKU number, and one nest serves five to eight parts with program variation only. The catalogue’s chamfer-variance row disappears when the nest keys the orientation positively.

Stage the next run during the current one. Trays loaded, program queued, tools pre-staged: the swap becomes load, dock, select, verify-first-article — and the verification is the only step that deserves new minutes.

Track changeover times as a metric with a target. What gets measured gets shortened; plants that log every changeover find their averages falling for months without any capital spend at all.

Automated Feeding for Small Parts

Lock parts are small enough that feeding, not finishing, often becomes the labour bottleneck. Vibratory bowls feed bolts and strikes reliably; cylinders and roses want pallet trays with vision locate, because their geometry defeats bowl orientation.

Choose feeding by part damage tolerance. A bowl that tumbles strike plates leaves edge nicks the finish pass then polishes into permanent witness marks — trays cost more and protect the part you are paid to perfect.

Synchronous feed beats max-rate feed on mixed lines. A tray conveyor presenting parts on the cell’s own cycle absorbs mix changes without re-tuning, and the buffer between feed and finish is where scheduling flexibility actually lives.

Keep one manual load position even after automating. The odd rush order, the prototype batch, and the part that jams the feeder all route through it, and it costs a bench and a fixture plate.

Instrument the feeder like any other station. A jam counter, a rate trend, and a part-present log turn “the feeder is acting up again” into a pattern you can fix on a schedule — usually a worn rail liner or a spring past its cycle count.

Feeder maintenance deserves the same ledger discipline as the spindle and the nests. Log every intervention, review the pattern monthly, and the feed system earns the reliability the finishing cell already has.

Lock hardware finishing succeeds on discipline at small scale: edge classes split early, exits programmed deliberately, five monitors visible, function tested hourly, and defects traced back to the die when they belong there. The catalogue above is the map — the fixes are proven, and the choice to apply them is the variable.

ROI estimates assume stable volume and the cited yield gains; model your own numbers before investing.