Small-Batch Aluminum Die Casting: When It Makes Sense and How to Run It Without Overpaying
The question that nobody puts in a RFQ but everybody is really asking: is my volume high enough to justify a die?
It's the right question. Opening a die for fifty parts a year and then watching the tooling cost sit in the denominator of every unit-cost calculation for three years is a real outcome, not a hypothetical. So is the alternative — machining from billet at quantities that make per-part cost four times what casting would produce. Getting the answer right the first time matters more than most buyers realise when they send out that initial enquiry.
The arithmetic of opening a die
A die is a fixed cost. Once it's built, you pay the same tooling charge whether you make ten parts or ten thousand. What changes is how many parts that fixed cost is spread across.
A simple single-cavity die for an aluminium bracket — nothing exotic, no side actions, moderate complexity — might cost between 18,000 depending on size and the number of machined features in the cavity. Add the trim die for gate removal and you might be at 25,000 all in. That number sits in the numerator of your tooling-amortised unit cost, divided by however many parts you make each year.
At fifty parts a year, that tooling amortisation per part is brutal. At two hundred, it starts to look reasonable. Past five hundred, the curve flattens and the story changes: casting's low variable cost — fast cycle, minimal material waste, near-net shape — starts winning against machining even when tooling is factored in. The crossover point isn't the same for every part, but two hundred to five hundred annual pieces is the range where the conversation gets serious for most standard-complexity aluminium Casting Parts.
The number most buyers miss is the variable cost asymmetry. Machining from billet at low volumes looks cheap because there's no tooling. But aluminium billet costs roughly three to five times more per kilogram than die casting feedstock, and on a complex part you might buy six kilograms to produce one kilogram of finished part. That scrap goes back as recycling value, not purchase price. At quantities above a few hundred pieces, the material efficiency of casting — buy-to-fly ratios of 1.2:1 rather than 6:1 — changes the total program economics substantially.
What part weight tells you about machine selection
Not every die casting inquiry needs the same machine. The clamp tonnage required to keep the die from flashing scales with the projected area of the part, and projected area broadly tracks with weight for a given material and section thickness.
The four weight brackets above are rough guides, not rigid rules. A flat thin part can have high projected area and need more tonnage than its weight suggests. A compact dense part — a thick-walled hydraulic body, for instance — can be heavier than its projected area implies. But as a first filter for matching enquiries to machine availability, weight is a useful proxy.
For small-batch programs, machine availability matters more than it does for volume work. A foundry running twenty cells can absorb a 150-part short run without disrupting its anchor programs. A shop with three machines and two anchor customers cannot. When you're placing short-run custom casting work, ask directly whether they run mixed schedules or whether small batches get scheduled as fill-in work between larger runs. Fill-in scheduling means your parts ship when the machine is free, not when you need them.
Alloy selection for short runs — where it gets complicated
Standard volume programs run a fixed alloy on a dedicated cell. The melt is set, the shot parameters are dialled in, and the operator knows the part. Short-run mixed programs don't have that luxury. If a shop is running A380 on the morning shift, A383 on the afternoon shift, and a customer asks for A360 in a short-run job, melt management discipline matters a lot more than it would in a volume program.
The alloy question for a short-run program isn't just which alloy is right — it's whether the foundry can demonstrate clean melt practice across a changeover. Request an EN 10204 3.1 melt certificate from a short run they've completed in the last ninety days.
Tooling strategies that make short runs viable
Process control on short runs — what's different
Volume programs generate enough parts per shift to run statistical process control (SPC) in real time. Short-run programs produce fifty or a hundred parts total — there isn't enough data for traditional SPC to function as intended.
What fills the gap is tighter first-article discipline and more thorough pre-production setup verification. Before the first shot on a short run, the shot profile — slow shot velocity, fast shot velocity, switch point, intensification pressure — should be logged and compared against the approved profile from previous runs of the same part.