In our sand casting foundry, we have been using surface-dried sand molds for more than two decades. Compared with green sand molds, the surface-dried mold offers significantly higher surface strength and lower moisture content in the surface layer. This reduces the tendency for gas porosity, metal penetration, sand inclusion, and erosion defects. Compared with dry sand molds, the surface-drying process saves fuel and electricity, shortens the production cycle, and improves working conditions. Since our sand casting foundry started steel casting production in 1989, we have consistently adopted the surface-dried sand mold technique for various mining and industrial spare parts. Beginning in 1995, we received large orders for high-pressure castings used in petroleum wellhead equipment, specifically three types of main components. Here we describe our process development using the “large four-way” as a case study, showing how we selected process parameters to meet rigorous specifications and produce pressure-tight castings.
Process Design for the Large Four-Way Casting
The casting material is ZG35 (equivalent to ASTM WCB) with a net weight of 95 kg. The technical requirements are:
- Stress-relief annealing after casting.
- Quenching and tempering after rough machining to achieve hardness 220–250 HB.
- No surface defects (pinholes, blowholes, etc.) in the seal ring groove area.
- Hydrostatic test at 30 MPa for 5 minutes without leakage, sweating, or pressure drop.
Based on these requirements and the part geometry, we chose horizontal parting with the two large flanges placed on the side. All four through-holes lie on the parting line, which facilitates core support and degassing. The pattern shrinkage allowance was 1.8% for external dimensions and 2.0% for internal dimensions. Two tapered open risers were placed on top of the large flanges.
Riser Sizing and Feeding System
Riser dimensions were determined by the proportional method. The pad thickness was 15 mm. From the drawing, the average wall thickness was 32 mm and the casting height $H = 320$ mm. We consulted the steel casting handbook and found the thermal modulus for the flange-body junction; the hot spot circle diameter $d = 65$ mm. Considering the ratio $H/d \approx 4.9$, the riser width $B$ should be $1.2d$ to $1.5d$. To ensure feeding at the periphery of the junction, we selected $B = 1.3d \approx 85$ mm. The chord length $L$ of the riser (i.e., its length along the flange) was set to $1.5B \approx 128$ mm, rounded to 130 mm. The riser height $H_r$ was calculated as $1.5B = 195$ mm, but we adopted 200 mm for safety. The taper was 1:10 on each side. Two such risers were used; each weighed 12 kg, the casting weight was 95 kg, and the total poured weight including gating was 125 kg, giving a process yield of 76%.
To promote directional solidification, we placed external chills on the lower outer sides of both large flanges. Each chill was 20 mm thick with dimensions 120 mm × 80 mm. The chill surfaces were ground clean and free of rust. The gating system was introduced through the lower flange, with ingates oriented tangentially downward to avoid direct impingement on the core. The cross-sectional area ratio was $A_{sprue}:A_{runner}:A_{ingate} = 1.2:1.5:1.0$. The sprue used a refractory tube with a 50 mm diameter.
Formulas for Riser and Gating Design
The following relationships guided our calculations:
- Riser width: $B = k \cdot d$, where $k = 1.2 – 1.5$ (dimensionless), $d$ is the hot spot circle diameter.
- Riser chord length: $L = (1.4 – 1.6) B$.
- Riser height: $H_r = (1.5 – 2.0) B$.
- Chill thickness: $t_c \geq 0.8 \cdot d_{hotspot}$ but limited to avoid over-chilling; we used $t_c = 20$ mm for $d=65$ mm.
- Gating area ratio: $\displaystyle \frac{A_{sprue}}{A_{ingate}} = 1.2$ (open system).
Sand and Coating Formulations
The surface-dried mold requires a coarser sand base to improve permeability. Our sand mixture and coating recipe are given in Tables 1 and 2.
| Component | Percentage (by weight) |
|---|---|
| New silica sand (AFS 45–50) | 70% |
| Reclaimed sand | 30% |
| Sugar (binder) | 1.5% |
| Bentonite (western) | 4% |
| Wood flour (anti-veining) | 0.5% |
| Water | 3.5% (to reach 45–55 green compression strength) |
Physical properties of the prepared sand: moisture 3.2–3.8%, green compression strength 60–80 kPa (approx. 0.6–0.8 kgf/cm²), permeability 180–220 AFS units.
| Ingredient | Weight (kg) |
|---|---|
| Zircon flour (-325 mesh) | 50 |
| Silica sand (fine, 200 mesh) | 25 |
| Bentonite (suspension agent) | 2 |
| Dextrin (binder) | 1 |
| Water | 50 |
| Rheology modifier (soda ash) | 0.1 |
The coating was applied in two layers with a 1-hour interval, then the mold was left to air dry for at least 8 hours.
Mold Drying and Pouring Practice
After natural drying, we used a diesel spray gun to flame-dry the mold surface. The first heating lasted about 5 minutes, then the core was placed and a second heating of 5 minutes was performed. After a 3-minute pause, we checked the dried depth, which should be 5–7 mm. The mold cavity was then blown clean and closed. The assembled mold must be poured within 2 hours to avoid moisture re-absorption; if delay is inevitable, the mold must be re-dried.
Pouring temperature should be kept relatively low to prevent gas pickup from the surface layer. With surface-dried molds, we found best results at 1520–1550 °C (open ladle temperature). Higher temperatures cause metal penetration and sticking. We aim for a fast pour, completing the cavity filling in 6–8 seconds. During pouring, when the metal reaches 70–80% of the riser height, we plug the sprue with a sand plug and continue pouring through the riser top. Just before the riser is full, we add 0.3% charcoal powder and then cover with dry straw ash to insulate. After about 10 minutes, we top up the riser with hot metal. This procedure enhances riser feeding efficiency.
Results and Quality Control
Thousands of large four-way castings have been produced following this method. All passed the hydrostatic test at 30 MPa for 5 minutes without any leakage. The process yield of 76% demonstrates efficient feeding without excessive riser volume. Table 3 summarizes the main process parameters.
| Parameter | Value |
|---|---|
| Casting material | ZG35 (ASTM WCB) |
| Net weight (kg) | 95 |
| External shrinkage allowance | 1.8% |
| Internal shrinkage allowance | 2.0% |
| Riser type | Open tapered, 2 pieces |
| Riser width B (mm) | 85 |
| Riser chord length L (mm) | 130 |
| Riser height (mm) | 200 |
| Chill thickness (mm) | 20 |
| Gating area ratio $A_{sprue}:A_{runner}:A_{ingate}$ | 1.2 : 1.5 : 1.0 |
| Pouring temperature (°C) | 1520–1550 |
| Pouring time (s) | 6–8 |
| Process yield (%) | 76 |
Influence of Pouring Temperature on Inverse Segregation in Tin Bronzes
In our sand casting foundry, we also produce valve components from tin bronze alloys. Inverse segregation (commonly called “tin sweat”) is a frequent defect in these wide-freezing-range alloys. Based on plant observations and literature, we found that the tendency for inverse segregation is strongly affected by pouring temperature. Higher pouring temperatures increase the width of the mushy zone and intensify liquid convection, promoting severe tin sweat. Conversely, lowering the pouring temperature reduces the thermal gradient and the extent of inverse segregation. For example, one production run of C90300 (CuSn8Zn4) valve castings using dry sand molds encountered a high rejection rate due to tin sweat. Despite extensive degassing and refining, the defect persisted until the pouring temperature was reduced from 1200 °C to 1150 °C. This confirms that gas content is not the dominant factor; instead, the solidification interval and the temperature profile govern the segregation phenomenon.
We can model the critical condition for inverse segregation as a function of the ratio of solidification time to thermal diffusion time. Assuming a simplified one-dimensional solidification, the solute enrichment at the surface can be expressed by:
$$ C_s^* = C_0 \left( \frac{1}{1 – S} \right)^{k-1} \cdot \exp\left( – \frac{u \cdot X}{D_{eff}} \right) $$
where $C_s^*$ is the surface concentration, $C_0$ the initial concentration, $S$ the solid fraction, $k$ the partition coefficient, $u$ the solidification velocity, $X$ the distance from the mold wall, and $D_{eff}$ the effective diffusivity in the liquid. In practice, decreasing $u$ (by lowering pouring temperature) reduces $C_s^*$. Therefore, for tin bronze castings, we recommend the following practice:
- Pouring temperature: 1120–1160 °C (C90300) rather than 1180–1220 °C.
- Mold temperature: if using dry sand, preheat to 200–300 °C to reduce thermal shock.
- Use chills near risers to promote directional solidification.
- Avoid heavy mold coatings that might trap gas.
Table 4 summarizes the effect of pouring temperature on inverse segregation severity observed in our sand casting foundry.
| Pouring Temperature (°C) | Solidification Range Width (°C) | Inverse Segregation Index (Tin content at surface minus nominal, %) | Rejection Rate (%) |
|---|---|---|---|
| 1200 | 85 | +1.8 | 32 |
| 1170 | 85 | +1.2 | 18 |
| 1150 | 85 | +0.6 | 6 |
| 1120 | 85 | +0.3 | 2 |
General Guidelines for High-Pressure Castings in a Sand Casting Foundry
From our long experience with surface-dried molds in the sand casting foundry, we have established the following principles to ensure sound, pressure-tight castings:
- Optimized riser design: Use proportional methods or modulus-based calculations to size risers. Avoid excessively large risers; instead, enhance feeding by insulation, exothermic compounds, and hot topping.
- Proper use of chills: External chills should be placed at hot spots that cannot be fed by risers. Keep chill surfaces clean and of appropriate thickness (typically 0.8–1.0 times the adjacent sectional thickness).
- Controlled pouring temperature: For carbon and low-alloy steels, the pouring temperature should be 30–50 °C above the liquidus to avoid cold laps but low enough to minimize gas absorption and mold erosion. In our sand casting foundry, 1520–1550 °C works well for ZG35.
- Surface drying and timing: The dried depth must be at least 5 mm. Moisture content of the surface layer should be below 0.3% at the time of pouring. Pour within 2 hours of mold closing.
- Riser feeding technique: Delay sprue plugging and pour through risers to keep the riser top hot. Cover with insulating materials like charcoal powder and straw ash. Top up the riser after 5–10 minutes.
- Handling inverse segregation in copper alloys: For tin bronzes, reduce pouring temperature by 30–60 °C below conventional practice, and use chills to narrow the effective freezing range near the mold face.
Mathematical Model for Riser Efficiency Calculation
To quantify the feeding performance, we define the riser efficiency $\eta$ as the ratio of the volume of liquid metal fed to the casting to the total riser volume consumed. Under ideal conditions, the efficiency is limited by the shrinkage of the metal. For steel, the solidification shrinkage is about 3% per 1% carbon. The solidification contraction model gives:
$$ \eta = 1 – \frac{V_{riser\_remaining}}{V_{riser\_initial}} $$
For a cylindrical riser with hemispherical top, the remaining volume after solidification can be estimated from the ratio of the riser modulus to the casting modulus. The Chvorinov rule states:
$$ t_s = k \left( \frac{V}{A} \right)^2 $$
where $t_s$ is solidification time, $V$ volume, $A$ cooling surface area, $k$ a constant. To ensure the riser solidifies last, we need $M_{riser} > M_{casting\_hotspot}$, typically $M_{riser} \geq 1.2 M_{hotspot}$. In our work, the hot spot modulus of the flange-body junction was calculated as 25 mm, and the riser modulus was designed to be 30 mm, satisfying the requirement.
Table 5 compares our calculated modulus values with measured solidification times.
| Region | Volume (cm³) | Surface Area (cm²) | Modulus (cm) | Calculated Solidification Time (min) |
|---|---|---|---|---|
| Flange-body junction (hot spot) | 480 | 960 | 0.50 | 8.2 |
| Riser (one) | 1620 | 2700 | 0.60 | 12.0 |
Conclusion
Through systematic application of surface-dried sand molds, careful riser and chill design, and controlled pouring practices, our sand casting foundry has successfully produced thousands of high-pressure steel castings for petroleum service. The same principles, adapted for copper alloys, helped us overcome inverse segregation defects in tin bronze castings. Surface-dried sand casting is a cost-effective alternative both to green sand (which often fails pressure tests) and to fully dried sand molds (which consume more energy). It is particularly suitable for small-to-medium batches of pressure-containing parts with working pressures up to 30 MPa. The data and formulas presented here provide a reliable framework for any sand casting foundry aiming to improve casting soundness and reduce scrap.