Our company, a major manufacturer of automotive powertrain components, has long faced the challenge of accelerating product development cycles for complex castings such as engine cylinder blocks and cylinder heads. Traditional sand casting methods rely on expensive and time-consuming metal tooling, which typically requires 3 to 6 months for fabrication. In response to rapidly evolving market demands, we adopted 3D printing casting technology to produce sand cores and molds directly from CAD data, thereby eliminating the need for physical patterns. This paper presents our comprehensive experience in applying 3D printing casting to a gray iron cylinder block, covering the technological principles, process validation, and key results.
The core advantage of 3D printing casting lies in its ability to fabricate geometrically complex sand molds layer by layer without any tooling. This approach drastically reduces lead times—from design to first prototype castings—down to 15–20 days. Moreover, it enables iterative design optimization with minimal cost, as modifications are made only in the digital model. In the following sections, we detail two common 3D printing casting methods, then describe our full-scale application on a 4-cylinder gasoline engine block, including material verification, process simulation, printing parameters, assembly, pouring, and inspection.
Principles of 3D Printing Casting Technologies
Two primary 3D printing casting technologies are widely used in the foundry industry: Selective Laser Sintering (SLS) and Three-Dimensional Printing (3DP). Both methods build sand molds layer by layer, but they differ in the binding mechanism and the type of sand used. We evaluated both and selected 3DP for our production due to its higher throughput and compatibility with our existing core-making materials.
Selective Laser Sintering (SLS)
In SLS, a thin layer of resin-coated sand (shell sand) is spread over a build platform. A CO₂ laser beam selectively scans the cross-section of the part, sintering the resin binder at the exposed areas. After one layer is fused, the platform lowers by a layer thickness (typically 0.2–0.5 mm), and a new layer of sand is applied. The process repeats until the entire sand core is formed. The unbound sand remains as support and is later removed. The resulting sand core has strength comparable to that produced by the hot-box process (typically 1.2–1.5 MPa). However, SLS machines have relatively slow build rates because the laser must trace every contour, making them less suitable for large-volume core production.
Three-Dimensional Printing (3DP)
3DP, also known as binder jetting, uses a print head that travels over a bed of sand (mixed with a catalyst) and selectively deposits a liquid binder (usually a furan or phenolic resin) according to the 2D slice data. After each layer, a fresh layer of sand is spread, and the binder is applied again. The chemical reaction between the binder and catalyst hardens the sand locally. Once all layers are printed, the loose sand is evacuated, leaving the finished core. 3DP machines can be much faster than SLS because the print head covers the entire layer width simultaneously. For our cylinder block, we used a Chinese-made 3DP printer with a build volume of 800 mm × 500 mm × 450 mm, capable of printing one full set of sand cores for a cylinder block in about 36 hours.
Verification of Sand Properties for 3D Printing Casting
Before committing to production, we needed to ensure that the printed sand cores met the same mechanical and thermal requirements as conventionally produced cores. We printed standard “8-shaped” tensile test specimens using the 3DP process and measured key properties: gas evolution, permeability, tensile strength, bending strength, and loss on ignition. Table 1 summarizes the results compared with our internal production requirements.
| Property | Production Requirement | Measured Value |
|---|---|---|
| Gas evolution at 1000 °C (mL/g) | <16 | 13.5 |
| Tensile strength (MPa) | >1.2 | 1.65 |
| Bending strength (MPa) | >2.1 | 2.8 |
| Loss on ignition (%) | 1.3–2.0 | 1.42 |
All measured values satisfied the requirements. Particularly, the tensile strength of 1.65 MPa provided a safety margin against core breakage during handling and pouring. The gas evolution of 13.5 mL/g was below the 16 mL/g limit, which is crucial to avoid gas defects in the casting. Since we used a full-core assembly technique (no green sand mold), the permeability requirement was less stringent, but the printed cores still exhibited adequate venting capacity.
Application to a Cylinder Block Casting
Cylinder Block Geometry and Material Specification
The selected component is a 4-cylinder gasoline engine cylinder block with overall dimensions of 382 mm × 356 mm × 256 mm and a minimum wall thickness of 4.5 mm. The material is HT250 gray iron, requiring a minimum tensile strength of 250 MPa, hardness between 180 and 220 HB on the top deck and cylinder bores, and a hardness variation within the same casting of no more than 20 HB. Dimensional tolerance follows CT9 per GB/T 1804.
The complete mold assembly consists of seven separate sand cores: left and right side cores, front and rear end cores, top cover core, oil gallery core, and water jacket core. These are printed individually, then dip-coated, dried, and assembled manually with bolted connections.
Process Design and Simulation
The casting weighs 50 kg, and the total poured metal mass is about 70 kg. We designed a bottom-gated, open runner system with a cross-sectional area ratio of:
$$ \sum F_{\text{straight}} : \sum F_{\text{horizontal}} : \sum F_{\text{ingate}} = 1 : 1.5 : 2 $$
The pouring time was controlled between 10 and 15 seconds. All gating channels were directly formed as cavities in the 3D-printed sand cores, eliminating the need for separate core boxes. Based on our experience, we applied a shrinkage allowance of 0.3 mm on non-dipped surfaces and 0.6 mm on dip-coated surfaces to compensate for the coating thickness (approximately 0.3 mm) and the loose sand layer (0.3 mm).
Before printing, we performed casting simulation using commercial software to optimize the filling and solidification sequence. The simulation predicted minimal shrinkage porosity and good feeding through the risers located on the top deck. The final design incorporated a group of small risers (riser cluster) to ensure directional solidification.
3D Printing Parameters
For the cylinder block cores, we used a domestic 3DP machine with the following parameters:
- Sand grain size: 140–200 mesh (70–140 mesh for the test specimens)
- Resin content: 1.8%–2.0% by weight
- Layer thickness: 0.2–0.5 mm
- Repeat positioning accuracy: 0.2–0.5 mm
- Sand core dimensional precision: 0.1–0.2 mm
- Sand core tensile strength: 0.8–1.5 MPa
- Gas evolution: ≤12 mL/g (measured at 13.5 mL/g for test specimens)
The entire set of cores for one cylinder block was printed in 36 hours. After printing, the cores were cleaned of residual loose sand using compressed air.
Coating and Core Assembly
Each sand core was immersed in a water-based zirconia coating and then dried in a continuous oven at 180 °C for 60 minutes. The residual moisture content after drying was checked and found to be less than 0.76%, which is within the acceptable range for gray iron casting. Manual assembly was performed using threaded rods and nuts to clamp the cores together. Special care was taken with small bosses and protrusions to avoid breakage due to their fragility.
Molding, Pouring, and Cleaning
Since the cores could not be placed on the automated molding line, we used a simple sand box filled with green sand to back up the core assembly, ensuring uniform support and preventing mold shift during pouring. After pouring with HT250 iron at 1420 °C, the casting was allowed to cool for 8 hours before shakeout. A total of six castings were poured; three were fully intact, while the other three suffered from core damage during handling or pouring.
Inspection Results
We conducted mechanical testing and dimensional inspection on the three acceptable castings. Table 2 shows the tensile strength measured from separately cast test bars (same ladle). Table 3 summarizes the Brinell hardness readings on the top deck and cylinder bore.
| Sample | Tensile Strength (MPa) |
|---|---|
| #1 | 265 |
| #2 | 270 |
| #3 | 269 |
| Average | 268 |
| Location | Point 1 (HB) | Point 2 (HB) | Point 3 (HB) | Range | Average |
|---|---|---|---|---|---|
| Top deck | 201 | 202 | 200 | 5 | 201 |
| Cylinder bore | 189 | 191 | 195 | 6 | 191 |
All values met the specification. The average tensile strength of 268 MPa was 7.2% above the minimum requirement, confirming the soundness of the casting process.
One casting was selected for full dimensional inspection. Table 4 presents the measured cylinder bore diameters at the four cylinders. All readings fell within the tolerance of ±1.0 mm.
| Cylinder # | Required (mm) | Actual (mm) |
|---|---|---|
| 1 (nearest front) | 65 ± 1 | 65.69 |
| 2 | 65 ± 1 | 65.81 |
| 3 | 65 ± 1 | 65.80 |
| 4 | 65 ± 1 | 65.86 |
Table 5 lists the positional accuracy of the four cylinder axes relative to the casting datums. The allowable deviation was ±0.5 mm, and all actual values were within limits, demonstrating that the 3D-printed sand cores maintained dimensional stability during pouring and that the core assembly alignment was precise.
| Cylinder # | Required (mm) | Actual (mm) |
|---|---|---|
| 1 | 115.7 ± 0.5 | 115.02 |
| 2 | 32.7 ± 0.5 | 32.688 |
| 3 | –50.3 ± 0.5 | –50.05 |
| 4 | –133.3 ± 0.5 | –133.3 |
The excellent dimensional conformance validated our approach of accounting for coating thickness and shrinkage allowance. The small deviations also indicate that the sand cores underwent negligible distortion under the ferrostatic pressure and thermal loads during solidification.
Machining and Engine Testing
All three acceptable castings were machined to final blueprint specifications and assembled into complete engines. The engines were then subjected to a cold test (motoring test) on a dynamometer bench. No abnormal vibration, leak, or dimensional interference was observed. The test results confirmed that the 3D printing casting process produced cylinder blocks with properties equivalent to those made by traditional tooling.
Conclusion
Our successful application of 3D printing casting technology to a production-intent gray iron cylinder block demonstrates that complex sand cores can be manufactured directly from CAD data without any hard tooling. The entire cycle—from process design to delivery of the first castings—was completed in 22 days, including 5 days for simulation and optimization. In contrast, a conventional approach would have required 3–4 months for pattern fabrication.
The key findings are:
- 3D-printed sand cores using the 3DP binder jetting process exhibit sufficient strength (1.65 MPa) and low gas evolution (13.5 mL/g) for successful gray iron casting.
- Full-core assembly with bolted connections yields dimensional accuracy within CT9 tolerance; all 4 cylinder bores and axis positions were within specification.
- Mechanical properties—tensile strength (268 MPa) and hardness (201 HB deck, 191 HB bore)—meet HT250 requirements.
- The elimination of pattern tooling reduces lead time by 80–90%, enabling rapid iteration for new engine designs.
We believe that 3D printing casting will become an indispensable tool for prototype and low-volume production of complex iron and steel castings. Future work will focus on optimizing the print speed and surface finish, as well as extending the method to larger components such as cylinder heads and transmission housings.