Sourcing OEM Aluminum Casting from China: What the Capability Page Doesn't Tell You

Every OEM casting manufacturer's website looks broadly the same. Cold chamber and hot chamber. A380, A383, ADC12. Tolerances to ±0.05 mm. ISO 9001. Fast turnaround. The list of capabilities is real — it describes what the factory can do on a good day with a straightforward part. What it doesn't describe is what happens when your part is the one that tests the edge of those numbers.
That gap is where sourcing decisions actually get made. So here's the version built around what to verify, not just what to read.
OEM casting is a supply relationship, not just a process. You own the design, you own the tooling you paid for, and you're accountable to your customer for every dimensional and functional requirement. The factory's job is to execute your specification reliably across every production lot.
That accountability chain matters because it shapes what you need in writing before the first shot is ever made. Custom Die Casting Services sold on a project basis and OEM casting services sold as a long-term supply program require different conversations upfront. A one-off casting can tolerate ambiguity that a recurring supply program cannot. If you're planning repeat orders, the first conversation needs to cover more than unit price.
Most capability pages list four or five aluminum alloys and two or three zinc alloys. Choosing between them is less about the list and more about two questions you bring to the table.
First: what does the part actually need? A housing that sits in ambient air and takes a painted finish doesn't need the same alloy as a hydraulic valve body that sees 200-bar internal pressure and a corrosive fluid. A380 covers the housing well. A413 — higher silicon, better fluidity, lower porosity tendency — is the right choice for pressure applications. A383 (ADC12 in Japanese industrial standard) fills thin walls better than A380 and is standard in electronics enclosures where 1.5 mm walls are routine.
Second: does the supplier actually run that alloy regularly, or does it appear on the capability page as a theoretical option? A shop that runs A380 every day and A360 twice a year will have process data for one and educated guesses for the other. Ask which alloys they poured in the last three months and request a melt certificate from a recent lot. An EN 10204 3.1 certificate ties the chemistry to a specific heat. If they can't produce one, that tells you something.
Zinc casting sits alongside aluminum on most OEM capability pages, and the choice between them is less obvious than it looks. Zinc — Zamak 3 in most cases — runs hot chamber, cycles three to four times faster than cold-chamber aluminum, and the tooling lasts five times as long. For small, intricate aluminium casting parts where weight isn't a constraint, zinc is often the better answer on total program cost even though the raw material costs more per kilogram.
Tooling is where OEM programs go wrong most predictably, and it almost always traces back to something that wasn't put in writing before the tool was built.

The number most buyers focus on is tool build time. It's the largest single block — typically six to eight weeks for a medium-complexity die — and it's the one that feels fixed. What's actually variable, and where most programs lose time, is the approval and revision phase after T1 samples arrive.
T1 is the first article off the production tool. It comes with a dimensional report, a visual inspection, and if you've specified it, a functional test. If the dimensional report shows a characteristic out of tolerance, someone has to decide whether to rework the tool, adjust the process, or accept a deviation. That decision loop, between your engineering team and the supplier's, is where two weeks becomes six.
Two things compress this. First, a DFM report before steel is cut. A proper DFM review catches the wall section that will shrink, the boss that will pull a sink mark, the thread location that needs a side action. Changes in CAD cost nothing. Changes to a hardened die are expensive and slow.
Second, a clear definition of what T1 must demonstrate before it passes. Write down the critical-to-function dimensions, the surface finish requirement with a reference standard, and if the part is pressure-bearing, the leak test rate in mbar·L/s. "Pressure tight" is not a specification; it's an intention, and it will be read generously.
ISO 9001 certification means a documented quality management system exists. It doesn't mean the system produces parts that meet your drawing. The two are related but not the same.
What's more useful for an OEM program is process capability data on the specific characteristics that matter to your application. Ask for Cpk values on critical dimensions from a recent production lot — not the initial capability study run at T1 approval, but from ongoing production. A Cpk of 1.33 or above on a critical bore diameter from month-six production means something. A Cpk of 1.67 from the qualification run alone doesn't.
For aluminium casting products going into regulated industries — automotive, medical device, aerospace — ask about IATF 16949 or AS9100 certification specifically. ISO 9001 is not a substitute. The documentation requirements, control plan depth, and reaction plan discipline are different categories.
Lot traceability is worth confirming separately. A part number on the carton is not lot traceability. A heat code, a tool cavity number, a date of pour, and a link to the melt certificate — that's lot traceability. If you ever need to isolate a suspect batch, you need that chain.

Before the purchase order, twelve items need resolution:
The packaging item sounds minor. It isn't, for cast metal parts. Aluminium casting parts shipped in bulk cartons without foam or tray separation arrive with ding marks on machined faces. The parts meet every drawing requirement; they also look like they've been kicked around a warehouse. If your customer inspects on cosmetic as well as dimensional, the packaging spec belongs in the purchase order.
Go back to those capability claims with a different lens. "Cold chamber and hot chamber" means they run both, which matters if you're comparing aluminum and zinc options. "Tolerances to ±0.05 mm as-cast" means that's the tightest they can hold without machining, and only on features within a single die half — features crossing the parting line inherit die-close variation. "Fast turnaround" usually means fast once the tool is built and approved, which is not where the time goes.
None of this makes the capability page dishonest. It makes it incomplete. The job of the sourcing conversation is to fill in the rest: which alloy, which tolerance matters, who owns the tool, what T1 must demonstrate, how you'll verify ongoing production, and what happens when something's wrong.
A manufacturer who answers those questions with specific numbers and documented commitments is describing a program. One who answers with adjectives is describing a brochure.
The supplier who tells you T1 samples failed two characteristics and here's what they're changing is the supplier running a program. The one who tells you everything passed and ships anyway is the one you find out about later.

