If the part has a three-dimensional body — a buckle with a bar and a spring seat, a hook with a thick load-bearing spine, a plate with raised relief — it is a die casting. If the part is flat or formed from sheet, and its job is to bend, spring or clip, it is stamping. Quantity decides the price, but geometry decides the process.
The expensive mistakes are always the same two: casting a part that should have been stamped (paying for a mould the design did not need), and stamping a part that should have been cast (then assembling three stampings to fake a shape a single casting would have made).
The five questions that actually decide it
Most buyer guides start with cost. Cost is the last question, not the first — it depends on the answers to these five.
| Question | Points to die casting | Points to stamping |
|---|---|---|
| Is the shape 3D or flat? | Ribs, bosses, internal cavities, variable wall, thick-to-thin transitions | Uniform sheet thickness, bends, embossments, drawn cups |
| Does it carry load in more than one direction? | Bulk strength, impact, thick sections | Tension, spring return, fatigue, thin flexing parts |
| What is the annual volume? | Few thousand to hundreds of thousands | Tens of thousands upward, where press speed wins |
| What finish is required? | Thick decorative plating, deep relief, painted 3D surfaces | Bright flat surfaces, brushed or polished sheet, light plating |
| Can the part replace an assembly? | Yes — one casting can absorb 3–5 stampings and their fasteners | No — each stamping stays a separate part |
That last row is where die casting quietly wins arguments it loses on unit price. A published industry case converted five welded stampings into one aluminium casting and reported a 35 % total cost reduction at a volume well below 50,000 — the saving came from assembly labour and fixture cost, not from the part price. Treat that number as an example of the mechanism, not a promise about your part.
Cost structure, honestly
Ranges below are from public process-selection guides for small-to-medium hardware. They are reference ranges, not quotations — a 12-gram buckle and a 300-gram bracket do not share a price band.
| Cost item | Zinc die casting | Metal stamping |
|---|---|---|
| Tooling (small hardware) | USD 8,000–35,000 | USD 3,000–20,000 |
| Tooling (broad range, all sizes) | USD 10,000–150,000 | USD 3,000–250,000 |
| Tooling lead time | 4–6 weeks | 3–5 weeks |
| Cycle time per part | 8–20 s (zinc, hot chamber) | 0.3–1.0 s (progressive press) |
| Part cost at 1,000 pcs | USD 0.80–2.50 | USD 0.30–0.80 |
| Part cost at 50,000 pcs | USD 0.25–0.60 | USD 0.08–0.20 |
| Material scrap rate | 3–8 % (runner recycled in-house) | 15–25 % (skeleton waste) |
| Secondary operations | Trimming, deflashing, sometimes machining | Deburring, sometimes welding or assembly |
Two things in that table explain almost every sourcing decision.
Stamping's cost curve falls faster. A progressive die makes a part in under a second; a die casting takes seconds per shot. At high volume that difference compounds into a unit price casting cannot match — but only for parts whose geometry suits sheet metal.
Casting's tooling is not always the more expensive one. For a complex part, the equivalent stamping may need several stations, or several separate dies plus a welding fixture. Below roughly 5,000 pieces, public guides put stamping ahead on total project cost; between 5,000 and 20,000 the geometry decides; above 20,000, casting usually wins for complex 3D shapes because material waste is minimal and multi-cavity tools lift output.
The amortisation trap
Buyers compare unit prices and then get surprised by the tooling invoice. Do the arithmetic on total cost at your real quantity before choosing:
| Line | Die casting | Stamping |
|---|---|---|
| Tooling | USD 12,000 | USD 4,000 |
| Part price | USD 1.60 | USD 0.45 |
| Quantity | 5,000 | 5,000 |
| Tooling per part | USD 2.40 | USD 0.80 |
| Landed cost per part | USD 4.00 | USD 1.25 |
At 5,000 pieces stamping looks unbeatable — until you notice the stamped part cannot have the rib, the boss and the 2 mm thick spine the design needs, and the "stamped version" is actually three pieces plus a welding operation. Run the same table with the assembly cost included and the gap usually closes. Run it at 50,000 and the casting line drops to roughly USD 0.84 while the stamping line drops to about USD 0.53 — still cheaper, still only if the geometry allows it.
One honest caveat about forecasts. A part that is 5,000 units this year may be 40,000 next year, or 2,000. Quote both ends of your range and see where the crossover falls. If it sits inside your range, weigh the cost of being wrong: a stamping die is cheaper to modify, and can be parked and re-run years later; a casting mould is more expensive to revise and ages with use.
Strength: the comparison buyers get wrong
| Property | Zamak 5 (die cast) | Stamped 304 stainless | Stamped cold-rolled steel |
|---|---|---|---|
| Tensile strength | ≈ 330 MPa | ≈ 520 MPa | ≈ 420 MPa |
| Yield strength | ≈ 295 MPa | ≈ 205 MPa | ≈ 350 MPa |
| Elongation | 7 % | 40 % | 15 % |
| Hardness | 95–105 HV | 200–220 HV | 120–180 HV |
| Density | 6.7 g/cm³ | 7.9 g/cm³ | 7.85 g/cm³ |
Read the yield row, not just the tensile row. Cast zinc yields higher than stamped 304 stainless in this comparison — which is why a die-cast buckle feels rigid and does not deform in the hand, while a thin stainless one flexes. But elongation tells the other half of the story: 7 % against 40 %. Under a sudden impact or a bend that must spring back, the casting is the brittle one.
Practical translation:
- Rigidity and shape retention → die casting, because thickness can be placed where the load is.
- Spring, clip, latch tongue, hinge leaf → stamping, because the material must flex a million times without cracking.
- Drop and impact requirements → test both. Cast zinc at 3–4 mm wall absorbs impact well; a 0.8 mm casting with a thin spine does not, and no alloy change fixes a bad wall.
Design rules differ more than the cost does
| Rule | Die casting | Stamping |
|---|---|---|
| Draft angle | 0.5–1° external, 1–2° internal (zinc) | None needed |
| Wall | 0.5 mm minimum, uniform, 1:3 transitions | Fixed by coil thickness |
| Holes | Cast in if parallel to the die opening; side holes need slides | Punched, essentially free |
| Threads | Cast-in threads not recommended; form or machine afterwards | Weld nuts, clinch nuts, or tapped extrusions |
| Edges | As-cast edge, trimmed; may need tumbling | Sheared edge, burr direction matters |
| Tooling change cost | High — a geometry change can mean a new insert or a new mould | Moderate — dies can be reworked, stations added |
The draft requirement is the one that ruins artwork. A relief designed for a flat stamped badge, dropped into a casting, becomes a wall that cannot release from the die. That is a modelling job, not a factory complaint — the relief must be redrawn with 5–10° draft on letter sides and a minimum letter height, which is covered in the logo plate guide.
Surface finish is where the processes separate hardest
Zinc castings take thick copper–nickel–chrome systems, hold deep relief and can be painted or powder coated on complex surfaces. Their weakness is the casting skin: porosity, cold shuts and polishing residue become blisters after plating, usually two to four weeks after shipment.
Stampings offer a bright, uniform, work-hardened surface that polishes beautifully, and their edges are the problem: a sheared edge exposes bare metal, needs deburring, and will not hold a mirror finish without extra polishing. Deep two-sided relief is essentially impossible — the material stretches.
For a bag plate or a belt buckle where the metal is the design, casting wins on shape and loses nothing on finish. For a flat keyring tag with an etched logo, stamping wins on price and finish consistency.
The hybrid answer most programmes end up with
Real hardware is usually both processes in one product:
- die-cast body + stamped spring steel tongue (belt buckle)
- die-cast housing + stamped internal clip (latch)
- die-cast decorative plate + stamped backing plate
- cast hook + stamped gate (swivel snap)
Specifying this deliberately is better than letting a supplier improvise it, because the interface between the two is where tolerance stacks and rattles come from. Put the assembly tolerance on the drawing, and state which part is the datum.
A short decision table
| Part | Process | Why |
|---|---|---|
| Belt buckle body | Die casting | 3D body, thick spine, plating |
| Buckle tongue / spring | Stamping | Must flex repeatedly |
| Zipper pull (sculpted) | Die casting | Shape and hand feel |
| Zipper pull (flat tag) | Stamping | Flat, high volume |
| D-ring, tri-glide, rectangle ring | Stamping (wire or sheet) | Simple formed geometry, high volume |
| Swivel hook body | Die casting | Load-bearing 3D form |
| Cord end, aglet | Die casting or stamping | Depends on relief depth |
| Logo plate with relief | Die casting | Relief and plating |
| Logo plate, etched flat | Stamping | Flat, fine etched detail |
| Ladder lock / side-release buckle | Injection moulding or die casting | Load + snap-fit features |
Next step
Send a drawing, a 3D file, or a photo of the part you want to match to 13928445679@163.com, with your annual quantity and target finish. We will tell you which process we would actually run it on — including when the answer is "this should be stamped, and we will quote it through our supply chain rather than cast it" — plus a quote that separates tooling, part price and finishing.