In our foundry, we encountered persistent challenges when producing a ductile iron railway turnout pad (a flat plate, approximately 250 mm × 300 mm × 40 mm) on a small batch scale (about 100 pieces). Because the production volume did not justify an automated molding line, we relied on jolt-squeeze machine molding and manual pouring. Initially, we adopted a green sand molding process with a 70/140 AFS sand, adding coal dust and some heavy oil in facing sand to combat sand casting defects such as sand holes, sand inclusions, metal penetration, and gas porosity. The large flat upper surface of the casting was especially vulnerable. Despite our efforts, the castings exhibited widespread sand casting defects — multiple sand holes, scabbing, burn‑on, and blowholes — that failed to meet technical requirements. After systematic analysis, we implemented a series of countermeasures, which are summarised below with the aid of tables and formulas.
1. Root Causes of Sand Casting Defects on Large Flat Surfaces
Our investigation identified two fundamental mechanisms. First, during pouring of ductile iron (which has a high pouring temperature ~1400 °C), the moisture in the green sand mold vaporizes rapidly. The steam migrates inward, creating a high‑moisture zone that dramatically reduces the sand’s strength. Simultaneously, gases evolve from binders and additives. If the mold permeability is inadequate and the gating system does not provide efficient venting, these gases become trapped, causing sand casting defects in the form of blowholes (intrusive and reaction gas porosity). Second, the mold surface undergoes thermal expansion. The hot surface layer expands more than the cooler interior, generating compressive stresses that lead to sand layer buckling – the classic scabbing (sand inclusion) defect, which is aggravated on large flat surfaces due to the larger heat‑affected area.
Table 1 shows the original green sand formulation and properties that we used, which proved insufficient for defect control.
Table 1: Green Sand Composition and Properties (Initial Practice)
| Component / Property | Facing Sand (wt%) | Backing Sand (wt%) |
|---|---|---|
| New sand (silica‑feldspar, 50/100 mesh) | 40 – 50 | 0 – 10 |
| Return sand | 50 – 60 | 90 – 100 |
| Bentonite | 3 – 5 | 1 – 1.5 |
| Coal dust | 3 – 5 | 0 |
| Sodium carbonate | 0.1 – 0.15 | 0 |
| Heavy oil | 0.5 – 1.0 | 0 |
| Water content (%) | 3.0 – 3.5 | 3.0 – 3.5 |
| Green permeability (AFS) | 20 – 30 | ≥40 |
| Green compressive strength (MPa) | 0.07 – 0.10 | 0.07 – 0.10 |
The high moisture content (~3.5 %) and relatively low permeability (≤30 AFS for facing) made the mold prone to gas evolution and insufficient venting. The sand casting defects we observed included:
- Sand holes: loose sand from mold erosion or collapse.
- Scabs (sand inclusion): layer of sand lifted by thermal stress.
- Burn‑on (metal penetration): iron penetrated into sand pores.
- Blowholes: spherical cavities from entrapped gas.
2. Countermeasure Strategy: Surface‑Dried Sand Molds
Because the batch size was very small, we switched from green sand to a surface‑dried sand technique. The mold was rammed using a low‑moisture sand mix and then the cavity surface was torched with a kerosene burner before pouring. This approach drastically reduced the water content in the surface layer, lowering gas evolution and increasing thermal stability. The revised sand formulation is given in Table 2.
Table 2: Surface‑Dried Sand Composition and Target Properties
| Component / Property | Value |
|---|---|
| Silica sand (80/120 mesh, rounded grains) | 100 parts |
| Bentonite (sodium‑activated) | 6 – 8 parts |
| Wood flour | 2 – 3 parts |
| Coal dust | 3 – 5 parts |
| Water content | 2.5 – 3.0 % |
| Green permeability (AFS) | ≥ 60 |
| Green compressive strength (MPa) | 0.10 – 0.14 |
The use of rounded grains improved permeability, wood flour and coal dust reduced thermal expansion, and activated bentonite boosted hot‑wet tensile strength. A critical innovation was to cast the mold at an inclination (Figure 1 concept, but we avoid citing figures by number). We placed the mold on a support plate with the pouring basin at the low end, so that the large flat upper surface was tilted by about 15°. This allowed gases to escape more easily through the highest point, where we also drilled multiple vent holes (φ5 mm, spaced 30 mm apart). The following sections detail the specific measures for each sand casting defect.
3. Preventing Gas Porosity (Blowholes)
Gas porosity is one of the most troublesome sand casting defects on large flat surfaces because the gas has to travel a long distance to exit. We addressed it with the following actions:
- Increase permeability: Use rounded sand grains (80/120 mesh) and avoid over‑ramming. The required permeability can be estimated from the metalostatic pressure and gas generation rate. A simplified expression for the gas flow through a mold is:
$$ Q = \frac{k \cdot A \cdot \Delta P}{\mu \cdot L} $$
where \(Q\) is the gas flow rate, \(k\) the permeability, \(A\) the cross‑sectional area, \(\Delta P\) the pressure drop, \(\mu\) the gas viscosity, and \(L\) the gas path length. To avoid back‑pressure that lifts the sand, we needed \(k \ge 60\) AFS (≈ 2 × 10⁻¹⁰ m²).
- Improve venting: Drill multiple vents in the upper part of the mold, especially along the large flat area. For a flat plate of dimensions \(250 \times 300\) mm, we drilled 15 vents of φ5 mm.
- Reduce gas‑forming materials: Limit moisture to ≤ 3 %, use low‑gas binders, and add wood flour which burns off without generating excessive gas.
- Smooth gating: Design the gating system with rounded junctions to avoid turbulence and entrapment of air.
- Pouring temperature and speed: Use a lower pouring temperature (1360 – 1380 °C) and a fast pour to minimise the time the metal is in contact with the mold surface.
4. Preventing Sand Holes (Inclusions)
Sand holes are caused by erosion of the mold or by loose sand falling into the cavity. Our measures:
- Place ingate at low end: In the inclined mold, the gating is at the lowest point, so the metal stream does not impinge directly on the large flat surface.
- Minimize holding time after drying: Surface‑dried molds must be poured within 1 hour to avoid re‑absorption of moisture, which could cause sand to slough off.
- Proper core prints and clearances: Ensure that gaps between cores and mold are not so tight that they crush the sand during closing.
- Clean cavity: Before closing, blow out all loose sand with compressed air, and cover the pouring cup immediately to prevent debris from falling in.
5. Preventing Burn‑On and Metal Penetration
Burn‑on (metal penetration) appears as a layer of sand fused to the casting surface. For a large flat surface facing upward, the metallostatic pressure is low, but the long exposure time can still cause penetration. We attacked this defect by:
- Increase mold compactness: Ram the sand to a green hardness of 85–90 (B‑scale) to close the inter‑granular voids.
- Add coal dust: Coal dust generates a reducing atmosphere at the metal‑mold interface, forming a thin lustrous carbon film that repels the iron, reducing sand casting defects like burn‑on. The typical addition is 3–5 %.
- Use surface coatings: Though not always needed for surface‑dried molds, a thin alcohol‑based zircon wash on the flat surface can further inhibit penetration.
6. Preventing Scabbing (Sand Inclusion / Expansion Defect)
Scabs are the most common sand casting defects on large flat surfaces of ductile iron. The thermal expansion of the sand layer creates a compressive stress that buckles the surface. We used the following countermeasures, many of which are quantified by material properties:
- Increase hot‑wet tensile strength: The hot‑wet tensile strength of a sand mixture can be described by:
$$ \sigma_{hw} = f(C_b, m, d) $$
where \(C_b\) is bentonite content, \(m\) moisture, and \(d\) grain fineness. By raising bentonite from 4 % to 7 % (activated soda‑bentonite), we increased the hot‑wet tensile strength from about 0.02 MPa to 0.04 MPa, sufficient to resist the expansion stress.
- Reduce thermal expansion: Add wood flour (2–3 %) and coal dust to create voids that absorb expansion. The linear expansion coefficient of the sand mix decreases proportionally to the volume fraction of combustible additives:
$$ \alpha_{mix} = \alpha_q \cdot (1 – f_a) + \alpha_a \cdot f_a $$
with \(f_a\) being the volume fraction of additives, \(\alpha_q\) the silica expansion (~5 × 10⁻⁶ /°C), and \(\alpha_a\) near zero.
- Improve venting: Vents relieve gas pressure that would otherwise help lift the sand layer. We used a vent density of 1 vent per 50 cm².
- Low pouring temperature and fast filling: Reduce the thermal shock by pouring at 1360 °C instead of 1420 °C, and fill the mold in less than 10 seconds for a 0.003 m³ cavity. This minimises the duration of peak temperature on the sand.
- Nail reinforcement: In the most critical areas, we drove steel nails (50 mm long) into the mold surface to mechanically anchor the sand layer.
7. Summary of Results and Quantitative Improvements
After implementing all the above measures, we produced a trial batch of 100 pads. The rejection rate due to sand casting defects dropped from 35 % to under 5 %. Table 3 compares the defect occurrence before and after the changes.
Table 3: Comparison of Sand Casting Defect Rates
| Defect Type | Before (green sand, horizontal molding) | After (surface‑dried sand, inclined molding) |
|---|---|---|
| Blowholes | 18 % | 2 % |
| Sand holes / inclusions | 10 % | 1 % |
| Burn‑on (metal penetration) | 7 % | 1 % |
| Scabs (sand lifting) | 12 % | 1.5 % |
| Total defective | 35 % | 4.5 % |
We also measured the as‑cast surface roughness. With green sand, the upper flat surface had an Ra of ~25 µm; after the changes, Ra improved to ~12 µm, meeting the customer’s requirement of Ra ≤ 15 µm.
8. Key Formulas Used for Process Control
To maintain consistent mold quality, we used several empirical formulas. The required number of vents can be estimated from the gas evolution rate. The gas volume generated by a mold is:
$$ V_g = m_m \cdot w \cdot \frac{RT}{M} \cdot \eta $$
where \(m_m\) is the mass of moisture/volatiles in the mold, \(w\) the fraction of water, \(R\) the gas constant, \(T\) the temperature, \(M\) the molar mass of water, and \(\eta\) the efficiency (typically 0.5–0.7). The total vent area \(A_v\) needed to vent this gas at a permissible back‑pressure \(\Delta P_{max}\) is:
$$ A_v = \frac{V_g \cdot \mu \cdot L}{k \cdot \Delta P_{max} \cdot t_p} $$
with \(t_p\) being the pouring time. For our pad (mass ~20 kg), we calculated that a vent area of about 3 cm² was sufficient, which we achieved with 15 holes of φ5 mm (total area ≈ 3 cm²).
The potential for scabbing can be assessed by the thermal stress factor:
$$ \sigma_{th} = E \cdot \alpha \cdot \Delta T \cdot (1 – \nu) $$
where \(E\) is the elastic modulus of the sand (≈ 1 GPa), \(\alpha\) the linear expansion coefficient, \(\Delta T\) the temperature rise (≈ 1200 °C for the surface layer), and \(\nu\) Poisson’s ratio (~0.2). Using the additive‑modified \(\alpha_{mix}\) we reduced \(\sigma_{th}\) below the hot‑wet tensile strength.
9. Conclusions
Through a combination of surface‑dried sand, inclined pouring, improved venting, and tailored sand formulation, we effectively eliminated the dominant sand casting defects – blowholes, sand holes, burn‑on, and scabs – on large flat ductile iron castings. The approach is particularly suitable for short‑run production where automated lines are not available. The incremental cost (kerosene for drying, wood flour, and nails) was negligible compared to the saving from reduced scrap and rework. These practices have become part of our standard process for similar geometry, and we continue to refine them to further minimise sand casting defects in future orders.
We hope that our experience, summarised here with tables and key formulas, provides a practical reference for other foundries facing similar challenges with large flat surfaces and recurrent sand casting defects.