Two Ways to Get a Metal Part, With Opposite Economics

Machining starts with a solid block and removes everything that is not the part. Investment casting starts with nothing and adds metal in the shape you want. That difference drives every downstream decision: cost curves, achievable geometry, surface finish, alloy availability, and how badly you get hurt if the design changes after tooling. The question is never which process is better in the abstract, but which fits this geometry at this volume with this tolerance requirement, and how often the answer is to use both on the same part.

What Investment Casting Actually Is

Lost-wax casting builds a wax replica in an aluminum die, assembles several patterns onto a sprue to form a tree, dips the tree repeatedly in ceramic slurry until a hard shell builds up, melts the wax out, fires the shell, pours molten metal in, then breaks the ceramic away. Gate removal, grinding, heat treatment, and machining of critical features follow.

The consequences of that sequence matter. Because the mold is a single-use ceramic shell, there is no parting line dictated by die halves, so undercuts, internal passages, and organic shapes are achievable. Because the shell is fine ceramic, surface finish lands around 125 to 250 microinches Ra, far better than sand casting. And because you need a wax die, there is real tooling cost and lead time before the first part exists.

Where Casting Wins

  • Complex organic geometry. Turbine blades, impellers, implants, and brackets with flowing ribs are cheap to cast and brutally expensive to machine, since a five-axis machine must reach every surface. If the part looks topology-optimized, casting is probably the answer.
  • Material removal ratio. When a machined part starts as a 40 lb billet and finishes at 6 lb, you pay for 34 lb of scrap plus the hours to remove it. Near net shape eliminates both.
  • Hard-to-machine alloys. Stainless, tool steels, cobalt-chrome, and nickel superalloys chew tooling. Casting them costs the same as casting anything else.
  • Consolidating an assembly. A weldment of six machined pieces can often be cast as one, removing fixtures, welds, and inspection.

Where Machining Wins

  • Low volume. Under roughly 100 to 250 parts, tooling amortization kills casting outright, and a machined prototype ships in days where first castings take weeks.
  • Tight tolerances. Casting holds about plus or minus 0.005 in per inch, typically no better than plus or minus 0.010 in on a mid-size part, against plus or minus 0.001 in for machining. Bearing fits, sealing faces, and threaded holes need machining regardless.
  • Design still moving. Changing CAD and re-running a program costs hours; changing a wax die costs weeks and thousands of dollars.
  • Prismatic geometry. Flat faces, bores, and pockets are what mills are built for; the cost drivers are in CNC machining cost.

Numbers to Plan Around

Wax die tooling for a small to medium cast part typically runs $3,000 to $15,000, with complex multi-core tools reaching $30,000 or more, an order of magnitude below die cast tooling and a large part of the appeal. First articles usually take 6 to 10 weeks, production lead times 4 to 8.

Per-part pricing depends on alloy and weight, but the useful model is that casting cost per part falls steeply and then flattens above a few hundred pieces while machining cost per part is flat everywhere. For a moderately complex steel part the crossover often lands between 200 and 800 units per year. For a simple bracket it may never cross; for an impeller it crosses at 50.

Minimum practical wall thickness is around 0.06 in (1.5 mm) in steel, and part sizes from a few ounces to 50 lb are routine. Alloy selection is broad: carbon and alloy steels, 300 and 400 series stainless, aluminum, bronze, cobalt-chrome, and nickel alloys, a wider menu than die casting offers. If the real question is which metal, start with aluminum vs steel.

Against the neighbors: aluminum die casting is cheaper per part at high volume but limited to aluminum, zinc, and magnesium, with far higher tooling cost and draft requirements, while metal 3D printing wins below about 50 parts and for geometry no casting can produce, then loses badly on cost per part. The framework for picking by quantity is in choosing a manufacturing process by volume.

The Answer Is Usually Both

Production metal parts are rarely purely cast or purely machined. The standard approach is to cast near net shape and machine only the precision features: bearing bores, mating flanges, O-ring grooves, tapped holes. That buys casting's material economics with machining's tolerance where it counts, typically removing 70 to 90 percent of the machining time versus cutting from billet.

It requires drawing discipline. The casting drawing and the machined drawing are separate documents with different datums, and the casting must carry enough stock, usually 0.03 to 0.06 in on machined faces, plus locating features a fixture can grab. Datum strategy and stock allowance are exactly what GD&T exists to communicate, and getting it wrong produces castings nobody can fixture.

Quality Issues to Watch

Castings carry failure modes machining does not. Porosity from trapped gas or shrinkage weakens sections and appears when a machined face opens a subsurface void. Inclusions from entrained shell material create hard spots that break cutters. Dimensional drift means a first article can be in tolerance while the hundredth walks out.

Countermeasures: specify radiographic or dye penetrant inspection on critical parts, require documented pour and heat-treat records, and treat the first production lot as a real qualification event under a proper first article inspection. Ask the foundry for its scrap rate on parts of similar complexity; a straight answer tells you a lot.

Run the Crossover Before You Commit

Projects House models both routes on your geometry: casting tooling and piece price against machining time and material, including the secondary operations casting still needs. Send a STEP file and annual volume through our contact form.