3D Printed Sand Casting: What To Check Before You Commit

• 1
  Key takeaways
• 2
  3D Printed Sand Mold vs Machined Pattern
• 3
  How 3D Printed Sand Casting Actually Works
• 4
  Which Parts Suit 3D Printed Sand Casting
• 5
  Tolerance, Draft and Machining Stock
• 6
  What Actually Drives The Cost
• 7
  How To Judge A Supplier
• 8
  Seven Steps From Drawing To Casting

A printed sand mold starts as a CAD solid of the mold cavity, not of the part. The printer deposits silica sand mixed with a binder, layer by layer, and cures each layer where the mold or core should be. Layer thickness usually falls between 0.2 mm and 0.4 mm. After the build, loose sand is removed and the mold is ready for pouring.

The important difference from a conventional foundry is what is missing. There is no pattern to cut, no core box, and no draft machined into a block of aluminum. That removes the longest item on the schedule and it removes the point where a design change becomes expensive.

The part is still a casting. Metal is poured into a cavity, it shrinks as it cools, and it needs draft to release. Printing the mold only changes how the cavity is made. All the foundry rules that apply to green sand or no-bake molds still apply here.

• 1
  Mold is printed, part is pouredTwo separate processes with separate tolerances.
• 2
  Layer lines transfer partly to the surfaceExpect a rougher as-cast skin than a machined pattern gives.
• 3
  Binder choice sets strength and gas behaviorFuran and phenolic binders behave differently at the pour.

The strongest fit is a part with internal channels, undercuts or a geometry that would need several cores in a conventional mold. A printed core can be one continuous piece, so a water jacket or a curved internal passage no longer needs to be split and glued.

Prototype and bridge volume is the second strong fit. When the design is still moving, a printed mold lets you pour a real part in the actual alloy without committing to a pattern. If the next revision changes a wall or a boss, you print again.

Thin walls are the limit to watch. Sections below roughly 3 mm are hard to fill consistently in sand, and they cool fast enough to create cold shuts. If the drawing calls for a 2 mm wall, cast it thicker and machine it back.

Very large parts are a different problem. Print bed size, mold handling weight and pour volume all cap the practical envelope. Ask the supplier for their maximum mold box before you assume your part fits.

Do not read a casting drawing the way you read a machined drawing. As-cast dimensions on a printed sand mold typically hold around ±0.5 mm on small features, and the variation grows with part length. Shrinkage adds another variable because it depends on alloy and section thickness.

The practical answer is to leave stock. Critical bores, sealing faces, bearing seats and any datum used for inspection should be cast oversize and finished by CNC. On a typical aluminum part, 1.0 mm to 2.0 mm per face is enough for a clean-up cut.

Draft is not optional. A vertical wall with zero draft will tear sand on withdrawal, and the tear lands exactly where you need a clean surface. One to two degrees is a common starting point; deeper pockets need more.

Tell the supplier which faces are machined and which stay as cast. That single note changes the mold layout, the gating and the price. Without it, the quote is a guess.

• 1
  Cast oversize, machine to sizeKeeps tolerances on the machine, not on the mold.
• 2
  Allow 1–2° draft on vertical wallsMore on deep pockets and rough sand.
• 3
  State the shrinkage allowance in the RFQAluminum, iron and stainless shrink differently.

Print time is the largest single line item. It scales with mold volume, not with part complexity alone, so a tall mold box with a small part inside wastes money. Ask the supplier to orient the mold to minimize build height.

Binder and sand consumption follow the same logic. A mold that is mostly sand around a small cavity costs nearly as much as a mold with a large cavity. Consolidating several parts into one printed mold plate is a common way to spread that cost.

Post-processing is often underestimated. Shakeout, cut-off, gate grinding, heat treat and machining all add time. A quote that lists only the mold and the pour is incomplete, and the difference shows up as a schedule slip rather than a line item.

Volume is the last driver. At one to twenty parts, printing usually wins. As quantity climbs, the per-part print cost stays flat while a machined pattern amortizes, and the crossover commonly lands in the low hundreds.

Ask what they print and what they outsource. A supplier who prints molds but sends every pour to a third-party foundry adds a handoff, and handoffs are where schedules and quality records get lost.

Ask for the casting tolerance they will actually put on the drawing, not the tolerance of the printer. Those are different numbers and the gap between them is where disputes start.

Check the finishing side. A casting is rarely the shipped part. If the supplier can also machine, heat treat and inspect, you avoid shipping a rough casting between vendors and losing the datum reference each time. GreatLight holds ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022, and runs 127 high-precision CNC machines across three plants, with 16 simultaneous 5-axis machining centers for the finishing cut.

Finally, ask how they handle your files. Uploads should be treated as confidential, and an NDA should be available before you send a drawing.