From 3D Print to Casting: What Changes
Printing the pattern is the easy half. Shrinkage, ash content and gate position decide whether the casting comes out right. A short guide for prototypes.
10/18/20264 min read
A customer sent us twenty printed brackets last spring. They had been printed at final dimension, in a castable resin, and the request was simple: cast exactly these, twenty pieces.
The castings came out 2% small. Every dimension, uniformly. Nobody had done anything wrong at the printer, and nothing was wrong at the foundry — the metal simply shrank on solidification, which is what metal does.
That is the whole subject of this article. Printing a pattern is the easy half.
Scale the model before you print
Every casting shrinks. The pattern has to be made oversized by roughly that amount, or the finished part will be undersized by it.
Carbon and low alloy steel — Linear shrinkage: 1.8–2.0%; Scale a 100 mm dimension to: 101.8–102.0 mm
304 / 316 stainless — Linear shrinkage: 1.8–2.2%; Scale a 100 mm dimension to: 101.8–102.2 mm
17-4PH — Linear shrinkage: 2.0–2.2%; Scale a 100 mm dimension to: 102.0–102.2 mm
Duplex stainless — Linear shrinkage: 2.0–2.2%; Scale a 100 mm dimension to: 102.0–102.2 mm
Aluminium A356 — Linear shrinkage: 1.2–1.4%; Scale a 100 mm dimension to: 101.2–101.4 mm
Copper alloys — Linear shrinkage: 1.3–1.6%; Scale a 100 mm dimension to: 101.3–101.6 mm
Inconel 718 — Linear shrinkage: 2.0–2.4%; Scale a 100 mm dimension to: 102.0–102.4 mm
Two caveats that catch people out. Shrinkage is not uniform — a part with a heavy boss and a thin flange will not shrink equally in both places, and no scale factor fixes that. And the scale factor applies to the pattern, not to the finished part: if your printed model is already the final geometry, you have to scale it up before printing, not shrink the casting afterwards.
If you would rather not deal with it, send the STEP file at final dimension and tell us the alloy. Applying the shrink is a normal part of setting up a job.
Resin is not wax
Standard SLA resin burns out badly. It leaves ash inside the shell, and ash shows up later as gas porosity or surface defects on the casting — usually in the places you cannot machine.
You need a castable wax-filled resin, or better, a printed wax. The burnout schedule also has to change: slower ramp, longer hold, more air. A shell built around printed resin and fired on a wax schedule will crack or leave residue.
We have seen both. The residue version is worse, because the castings look fine until they are machined.
Surface finish transfers, roughly
The layer lines on a printed pattern do transfer to the casting surface, softened by the ceramic coat but still visible under a light. Two things reduce them: print orientation, so the lines run along the face nobody looks at, and a slightly thicker prime coat on the printed pattern.
For a visual part, sanding the pattern before dipping is worth the hour.
Gates still have to be designed
A printed pattern gets glued onto a wax sprue like any other pattern, and the joint is a weak point. Printed patterns are heavier and stiffer than wax, so the joint carries more load when the cluster is dipped and handled.
Place the gate where the metal enters a thick section, same as always. Printing does not change the rules of feeding. It only changes how the pattern was made.
If you are printing one pattern to test a design, print two gate variants. The second cluster costs very little and answers the question that a single sample cannot.
When printing is the right call
One-offs and spares. A single replacement part, no tooling, casting in two weeks.
Design iterations. Print, cast, measure, change the model, repeat. Three iterations is cheaper than one aluminium tool that turns out to be wrong.
Bridge production. Ship the first fifty from printed patterns while the die is being cut.
When it is not
Above roughly a hundred pieces, printed patterns stop making sense. The per-piece pattern cost does not fall, and a machined aluminium die amortises quickly. Compare the two quotes at the quantity you actually expect over two years, not the quantity in the first purchase order.
Also, if the part has a tolerance that needs machining anyway, the surface quality argument for printing weakens — you are going to cut it true regardless.
What to send
The STEP file, the alloy, and the quantity. Tell us whether the model is at final dimension or already scaled. If the part has thin walls, read how thin a wall can be before you commit to the geometry — printed patterns are more forgiving than dies, so a wall that casts from a print may not cast from tooling later.
For wax grades and pattern materials in normal production, see wax materials for precision casting. For what we can run in house, the services page lists the processes and alloys.
Questions we get on this
Can you cast directly from my printed part without a model file? Yes, if it is a castable wax or wax-filled resin and it is undamaged. Send the physical part and tell us the alloy; we will measure it and apply the shrink.
How many pieces before tooling is cheaper? Usually somewhere between 50 and 150, depending on part size and how many features need machining. Below that, printing wins on total cost even though the unit price looks higher.Related reading
Quality
Precision manufacturing for diverse industrial applications.
Services
Products
+8615912702921