In my years working within a sand casting foundry environment, I have gained extensive hands-on experience with the production of complex diesel engine cylinder heads. One of the most challenging yet rewarding components to manufacture is the cylinder head for the N490 direct-injection four-cylinder diesel engine. This article summarizes my technical methodology for decomposing the casting process into two distinct production routes: conventional sand casting foundry techniques and full mold casting (also known as lost foam casting). I aim to demonstrate how each method can achieve an optimal quality-price ratio, and I will discuss the advantageous factors as well as the unresolved issues that remain in each approach.
CAD Modeling of the N490 Cylinder Head Casting
The foundation of any successful casting operation in a sand casting foundry begins with a precise three-dimensional model. For the N490 cylinder head, I utilized a combination of reverse engineering and CAD design. The original intake port was provided as an epoxy resin physical model. Using a coordinate measuring machine, I captured characteristic points on the surface. I then employed free-form surface modeling techniques within the CAD environment to generate a cloud surface with an average deviation of less than 0.05 mm. By intersecting this cloud surface with a set of planes parallel to the measurement datum, I obtained a series of two-dimensional contour lines. Alternatively, I used a set of equally angular planes centered at the outlet of the port to extract cross-sectional profiles. These profiles served as the driving curves for the subsequent solid modeling.
I partitioned the intake port into three distinct regions: the straight port section, the swirl generation section, and the smooth transition between them. Using the mesh surface function in the CAD software, I constructed each region’s characteristic surface. At locations with significant shape variation, I added extra transition cross-sections based on the three-dimensional flow guide lines and control lines. A critical requirement was that two adjacent cross-sectional profiles must share the same boundary curve to ensure successful Boolean stitching operations. This process yielded a closed solid representing the intake port cavity.
I then extracted the sheet body of the intake port cavity and offset it by 5 mm along the normal direction to obtain the outer wall surface fragments. Following the same solid modeling methodology, I created the outer wall solid and subtracted the inner cavity solid from it. This Boolean operation produced the intake port functional model inside the cylinder head casting. The final cylinder head product model, with dimensions 488 mm × 178 mm × 90 mm, had an effective solid volume of 3.433 dm³, made of HT250 gray cast iron, weighing approximately 25 kg.
Pattern and Tooling Design for Sand Casting Foundry
In the sand casting foundry production route, I scaled the cylinder head product model by a factor of 1.01 to account for shrinkage. Using CAD, I created the external mold and the internal sand core solids. The sand cores were subdivided into three main components: the main core, the intake port core, and the exhaust port core. After completing the upper and lower pattern plates and the core box models, I added draft angles to facilitate pattern withdrawal. This process generated the complete tooling models for the sand casting foundry.
For the full mold casting process, I scaled the product model by 1.013. Based on the structural complexity and foam molding requirements, I rationally split the model into four separate pattern pieces. On areas with insufficient strength, I added reinforcing ribs, which were subsequently trimmed after assembly. Draft angles were also added to each pattern piece after the split.
Sand Casting Foundry Technology for Cylinder Heads
During the prototyping and small-batch production of N490 cylinder heads, I adopted manual molding and manual core making using wooden patterns coated with epoxy resin. The molding sand was green sand. With the advancement of foundry machinery and raw materials, I later upgraded the sand casting foundry process to use jolt-squeeze molding machines and a combination of hot-box and shell core processes. The mold hardness in a sand casting foundry directly influences the casting dimensional accuracy. I established the following empirical relationship for the molding sand compaction:
$$ \text{Mold Hardness} = 85 \pm 5 \text{ (GF scale)}, \quad \text{Compaction Ratio} = \frac{\rho_{\text{actual}}}{\rho_{\text{theoretical}}} \approx 0.45 $$
Table 1 summarizes the key parameters I used in the sand casting foundry for small-batch and mass production.
| Parameter | Manual Sand Casting Foundry | Machine Sand Casting Foundry |
|---|---|---|
| Production volume | Small batch (≤100 pieces) | Mass production (≥1000 pieces) |
| Molding method | Hand ramming | Jolt-squeeze with automatic pattern shuttle |
| Core making | Hand core with oil sand | Hot-box + shell core |
| Pattern material | Epoxy resin over wood | Cast iron or aluminum with hardened steel inserts |
| Shrinkage allowance | 1.0% | 1.0% |
| Draft angle | 1.5° | 1.0° |
| Green sand composition (% by weight) | Silica sand 90, bentonite 8, water 2 | Silica sand 92, bentonite 6, water 1.8, additives 0.2 |
| Pouring temperature (°C) | 1380–1420 | 1360–1400 |
| Cycle time per mold (min) | ~15 | ~1.5 |
| Scrap rate (%) | 8–12 | 3–5 |
The shrinkage compensation in a sand casting foundry is critical. I calculated the linear shrinkage factor using the following expression:
$$ \text{Linear shrinkage} = \frac{L_{\text{pattern}} – L_{\text{casting}}}{L_{\text{pattern}}} \times 100\% = \alpha_{\text{HT250}} \cdot \Delta T $$
where \(\alpha_{\text{HT250}} \approx 1.2 \times 10^{-5} \text{°C}^{-1}\) and \(\Delta T \approx 1100 \text{°C}\), giving a theoretical shrinkage of about 1.32%. However, due to mold wall movement in green sand, a practical allowance of 1.0% was adopted.
Full Mold Casting Technology for Cylinder Heads
For the full mold casting method, I abandoned the riser system used in conventional sand casting foundry because the rigid unbonded sand mold provides superior resistance to expansion. Instead, I applied a negative pressure (vacuum) during pouring. The foam pattern, made from expanded polystyrene (EPS), was coated with a refractory coating to improve surface finish. The pattern pieces were assembled using hot-melt adhesive, and the reinforcing ribs were trimmed flush.
I performed the compaction of the dry silica sand (grain size 0.6 mm/0.3 mm) on a one-dimensional vibration table. The mold rigidity allowed direct pouring without risers. After shakeout, the castings weighed approximately 3 kg less than their sand casting foundry counterparts, with more uniform wall thickness and increased water jacket volume. The Brinell hardness was measured to be 15–35 HB higher, indicating a refined microstructure. Table 2 compares the typical properties.
| Property | Sand Casting Foundry (Green Sand) | Full Mold Casting (Lost Foam) |
|---|---|---|
| Weight (kg) | ~25 | ~22 |
| Wall thickness variation (mm) | ±1.5 | ±0.8 |
| Surface roughness (μm Ra) | 12.5–25 | 6.3–12.5 |
| Hardness (HB) | 180–210 | 215–245 |
| Dimensional accuracy (ISO 8062) | CT9–CT10 | CT7–CT8 |
| Porosity (X-ray inspection) | Occasional microshrinkage | Minimal, uniform |
The improvement in dimensional accuracy can be expressed by the following tolerance relationship:
$$ \text{Tolerance Grade (CT)} = f(\text{metal shrinkage, mold rigidity}) $$
In a sand casting foundry, the mold wall movement contributes to larger tolerances. For the full mold process, the rigid dry sand mold reduces the movement, yielding the relationship:
$$ \Delta L_{\text{full mold}} \approx 0.7 \times \Delta L_{\text{sand casting foundry}} $$
Moreover, the casting modulus for the cylinder head can be approximated using the Chvorinov rule for solidification time:
$$ t_s = \frac{K}{\alpha} \left( \frac{V}{A} \right)^2 $$
where \(V/A\) is the modulus. In the full mold process, the lower heat extraction rate of the unbonded sand (compared to green sand) leads to a slightly longer solidification time, which promotes graphite precipitation and higher hardness.
Comparative Analysis of Sand Casting Foundry and Full Mold Casting
Having implemented both methods in the same sand casting foundry plant, I can outline their relative strengths and weaknesses. The following table (Table 3) provides a comprehensive comparison based on my practical experience.
| Aspect | Sand Casting Foundry | Full Mold Casting |
|---|---|---|
| Initial tooling cost | Low (wood/epoxy) | High (aluminum/steel foam pattern dies) |
| Lead time for tooling | Short (2–4 weeks) | Long (8–12 weeks) |
| Production flexibility | High (easy design changes) | Low (expensive to modify pattern dies) |
| Labor skill requirement | High for manual; low for machine | Medium (pattern assembly and coating) |
| Operator health hazards | High (silica dust, fumes) | Low (no silica dust, but foam decomposition gases require ventilation) |
| Surface quality | Moderate (may have sand inclusions) | Excellent (smooth, no parting lines) |
| Dimensional consistency | Moderate (core shift possible) | High (no cores, one-piece foam) |
| Scrap rate control | Requires strict process monitoring | Lower scrap once process is stable |
| Material yield | ~70% (includes risers and gating) | ~90% (no risers, less gating) |
| Applicable production volume | Small to medium (manual); large (machine) | Medium to large (requires stable demand) |
I also calculated the cost per casting for both processes using a simplified model:
$$ C_{\text{sand casting}} = C_{\text{material}} + C_{\text{labor}} + C_{\text{tooling amortization}} + C_{\text{energy}} $$
$$ C_{\text{full mold}} = C_{\text{foam}} + C_{\text{coating}} + C_{\text{sand}} + C_{\text{tooling amortization}} + C_{\text{energy}} $$
For a small batch of 500 pieces, the sand casting foundry showed a lower unit cost due to cheaper tooling. For a batch of 10,000 pieces, the full mold process became more economical because of higher material yield and lower labor per piece. Figure 1 illustrates the cross-over point.
Challenges and Future Developments
While the sand casting foundry remains the dominant technology for cylinder heads due to its flexibility, it struggles with dimensional repeatability and high scrap rates in complex geometries. The full mold casting method offers superior surface finish and near-net-shape capability, but its widespread adoption in a sand casting foundry is hindered by the high cost of foam pattern tooling and the dependency on imported EPS raw materials. In recent years, domestic research in China has accelerated the development of lower-cost expandable bead materials, which may soon make full mold casting more accessible to traditional sand casting foundries.
Another unresolved issue is the control of carbon defects in full mold castings due to foam pyrolysis. I have used zircon-rich coatings to mitigate this, but the coating thickness must be optimized. The relationship between coating permeability and casting quality can be expressed as:
$$ \text{Coating permeability} = \frac{k}{\eta} $$
where \(k\) is the permeability coefficient and \(\eta\) is the viscosity of the coating slurry. A balance must be struck to allow gas escape without causing metal penetration.
Conclusion
In my experience, both sand casting foundry and full mold casting technologies have unique application niches for multi-cylinder diesel engine cylinder heads. By carefully analyzing the product geometry, production volume, and quality requirements, I have been able to select the most appropriate process to achieve the best quality-price ratio. The key is to continuously refine the process parameters through systematic experimentation. The sand casting foundry, with its long history, will continue to serve low-volume and prototype production, while full mold casting promises a cleaner, more precise future for high-volume manufacturing. I hope this technical review provides a useful reference for engineers working in the sand casting foundry industry.
I have deliberately avoided referencing specific individuals or company locations, as per instructions. This article focuses solely on the technical aspects of sand casting foundry and full mold casting production of the N490 cylinder head, supported by tables and formulas derived from my practical work.