Die Casting, Sand Casting or CNC? The Volume Where Each One Wins
beats machining and where it doesn't — real crossover volumes, buy-to-fly ratios, hybrid cast-and-machine routes, aluminum vs zinc selection, and how to sanity-check a process recommendation before committing to tooling.Meta description (crawler-facing, not customer-facing): Where aluminum alloy casting
Ask a Die Casting Shop whether you should die cast a part and you already know the answer. Ask a machine shop the same question and you get the opposite one. Both are being honest. Both are answering from inside their own cost structure.
The useful question isn't which process is better. It's at what annual volume each one stops being the cheapest, and those numbers are calculable.
The two crossover points
Every process trades tooling cost against per-piece cost. CNC needs no tooling but bills machine hours forever. Sand Casting needs a modest pattern. Die casting needs an expensive steel die and then produces parts for very little.
Plot total program cost against volume and you get three lines with different slopes, crossing at two points.
For a mid-size aluminum housing, the first crossing usually lands somewhere between 300 and 800 pieces a year. Below that, machining from billet wins outright — no tooling, no lead time, and you can revise the design on Tuesday.
The second crossing typically sits between 3,000 and 10,000 annually, where a die pays for itself against sand casting. Wide range, because it moves with part complexity and how much finish machining survives either route.
Two caveats before anyone quotes those numbers back at me. They shift with part size — a large structural part pushes both crossings right because tooling scales faster than machine time. And they assume you actually hit your forecast. A program that projects 8,000 and delivers 2,000 has bought an expensive die.
Buy-to-fly, and why machining a housing hurts
Here's the number that decides most casting-versus-machining arguments, and hardly anyone calculates it.
Buy-to-fly is the ratio of starting material weight to finished part weight. Machine a 400-gram enclosure from a solid billet and you might start with 3.2 kilograms. That's a ratio of 8:1, meaning 87% of the aluminum you purchased becomes chips.
Chips have scrap value, but you bought that metal at billet price and you're selling it back at scrap price. The spread is real money at volume, and you paid machine hours to create it.
A casting arrives at roughly 1.2:1. The comparison isn't close once volume climbs.
Where machining still wins on the same part: prototype quantities, designs still moving, and geometries that are genuinely just a block with features in it. Buy-to-fly on a simple bracket might be 2:1, and at that ratio the argument evaporates.
Nobody actually picks one process
This is the part the "casting vs machining" framing gets wrong. Almost every real Aluminum Casting program is a hybrid.
You cast the shape and machine the features that need precision. Sealing faces, bearing bores, threaded holes, anything with a position tolerance tighter than about ±0.13 mm. Everything else stays as-cast.
Which turns the design question into a boundary question: how much of this part must be machined, and can I move that boundary?
Two moves worth knowing. First, consolidate machined features onto one face or one axis so the part needs a single setup rather than three — each additional setup adds fixturing, handling and a stack-up error. Second, cast bosses slightly proud where they'll be faced, so a light skim cleans them up rather than a heavy cut fighting the as-cast skin.
I've seen a housing drop 40% of its machining cost purely by rotating two threaded bosses onto the same plane. The casting didn't change. The fixture went from three operations to one.
Aluminum or zinc
Different question, and it gets decided too casually.
Aluminum runs cold chamber because molten aluminum at around 690 °C would chemically attack the submerged steel gooseneck of a hot chamber machine. Zinc melts near 420 °C, gentle enough to run hot chamber, and that single difference cascades through everything.
Zinc cycles in 3 to 8 seconds against 20 to 90 for aluminum. Zinc tools last four to five times longer because the thermal shock is milder. Zinc holds thinner walls — down to 0.6 mm against about 1.5 mm — and tighter as-cast tolerances. It also plates beautifully, which aluminum does not.
The catch is density. At 6.6 g/cm³ zinc is roughly two and a half times heavier than aluminum. On a small connector shell nobody cares. On a handheld tool housing it decides the product.
Rough rule: aluminum when weight, thermal conductivity or corrosion resistance matters. Zinc casting when you need very thin walls, fine detail, long tool life or a real plated finish, and the mass is acceptable.
Sand casting still has a place
It gets dismissed as the crude option, which is wrong in two specific situations.
Large parts. Die casting is bounded by available clamp tonnage, and above roughly a square metre of projected area you're looking at a very short list of machines globally. Sand has no such ceiling.
Heat-treatable properties. Conventional die castings blister above 500 °C because entrapped gas expands, so full T6 is off the table. Sand and permanent mold in A356 take T6 properly, which matters when you need the mechanical properties rather than just the shape.
Sand also wins on low volume with an evolving design, since a pattern is cheap to modify and a hardened steel die is not.
Sanity-checking a recommendation
When a supplier proposes a route, these questions tend to reveal whether they modeled your job or defaulted to their equipment.
What volume does this recommendation assume, and where's the crossover if we're wrong by half? A shop that has run the numbers answers immediately.
How many machining setups after casting, and can any be eliminated by moving a feature? Silence here means nobody looked.
What's the buy-to-fly if we machined it instead? Even a rough figure shows they've considered the alternative rather than dismissing it.
What tonnage will it run on, and what happens if that press is busy? Capacity is a real constraint that rarely appears in quotes.
Does the price include trimming, deburring, machining and finishing, or just the raw casting? This single question explains most inexplicable price gaps between quotes.
Where it lands in practice
Electronics and energy work runs die cast housings and heat sinks where thermal conductivity and volume both favour aluminum. Fluid power takes valve bodies and pump housings, cast for shape and machined for sealing faces. Automotive pulls structural nodes and motor housings at tonnages that barely existed a decade ago. Plumbing runs cast metal pipe fittings and flanges, and most any pipe fittings manufacturer will keep aluminum, brass and zinc cells running in parallel with separate melt records.
Hand tools are the interesting case, because that's where aluminum and zinc genuinely compete on the same drawing and the decision comes down to grams.
Industrial machinery mixes castings with custom metal fabrication as a matter of course, and this is where the framing question returns one level up. Casting and structural fabrication aren't alternatives. A cast node replaces six welded plates and holds its geometry without a weld fixture, but the frame it bolts into still gets cut, formed and welded. A supplier offering casting services and custom fabrication together can put that split wherever total cost lands lowest — which is rarely where either specialist alone would have drawn it.
FAQs
At what volume does die casting beat CNC machining? Usually somewhere between 300 and 800 pieces annually for a mid-size aluminum housing. Complex parts with high buy-to-fly cross earlier; simple bracket-like geometries cross much later or never.
When is sand casting the better choice? Large parts beyond practical clamp tonnage, low volumes where a steel die can't amortise, designs still changing, and anywhere you need genuine T6 heat treatment — which conventional die castings can't take without blistering.
What is buy-to-fly and why does it matter? Starting material weight divided by finished part weight. Machining a housing from billet often runs 6:1 to 10:1, meaning most of your purchased aluminum becomes chips. Castings sit near 1.2:1. It's usually the decisive number in the casting-versus-machining comparison.
Do die cast parts still need machining? Almost always, for sealing faces, bearing bores and threaded holes. As-cast holds roughly ±0.13 mm, so anything tighter gets machined. The goal is minimising setups rather than eliminating machining entirely.
Why can't aluminum run on hot chamber machines? Molten aluminum near 690 °C attacks the submerged steel gooseneck and dissolves it. Zinc melts around 420 °C, mild enough for the submerged assembly to survive, which is why zinc runs hot chamber and cycles far faster.
Should I choose aluminum or zinc? Aluminum for weight, thermal performance and corrosion resistance. Zinc for walls below 1.5 mm, fine detail, tool life four to five times longer, and genuine plated finishes — provided the extra density is acceptable in the application.
How accurate are those crossover volumes? Directional, not absolute. They shift with part size, complexity, alloy and how much machining survives each route. Ask your supplier to model both routes at your actual forecast rather than relying on a generic figure.
Can I start with machining and move to casting later? Yes, and it's often the sensible sequence. Machine early units while the design settles, then cut a die once the geometry is frozen and volume justifies it. Design for casting from the start though — draft, radii, uniform walls — or the transition means a redesign.
The client I mentioned the machining-cost figure for had been quoted both routes and assumed the cheaper per-piece number was the answer. It wasn't, quite. At his real volume of 1,400 a year, sand casting with two machined faces beat both the billet route and the die he'd been offered.
He cut the die eighteen months later, when volume tripled. That was the right time. Not before.