Keywords: sand casting defects; cylinder head; gas porosity; core venting; casting scrap reduction
In our foundry, the 226B cylinder head has been one of the most demanding castings we produce. The model has been manufactured since 1999, when our company entered a joint venture with a German engine partner and introduced a complete machining line, molding line, core-making line, and pattern equipment for this engine family. Over time, annual output reached 15,000 units, and we are now preparing to increase this to 25,000 units. The cylinder head is a complex gray iron casting with thin internal walls and a strict pressure-tightness requirement. Any porosity or entrapped gas in the head can lead to leakage during pressure testing and to premature failure in service. For this reason, I have spent many years on the shop floor studying how sand casting defects form in this component. The purpose of this article is to explain how we analyzed and solved the most severe of these sand casting defects.
Production Process and Material Specification
The 226B cylinder head is cast in a copper-alloyed gray iron designated GG30Cu. This grade is selected because it combines good machinability with the strength and pressure tightness required for a diesel engine cylinder head. The cast wall can be as thin as 4 mm, especially in the water-jacket region, so the iron must have excellent fluidity and a stable graphitic structure. The target chemical composition is summarized in Table 1.
| Element | Target Range (wt%) |
|---|---|
| Carbon, C | 3.35–3.45 |
| Silicon, Si | 1.90–2.10 |
| Manganese, Mn | 0.70–0.90 |
| Sulfur, S | ≤0.10 |
| Phosphorus, P | ≤0.10 |
| Copper, Cu | 0.80–1.00 |
| Chromium, Cr | 0.20–0.30 |
| Molybdenum, Mo | 0.30–0.40 |
For a gray iron of this type, the carbon equivalent is a useful first-order measure of the alloy behavior. I frequently use the following expression for carbon equivalent:
$$CE = w(\mathrm{C}) + \frac{1}{3}\left[w(\mathrm{Si}) + w(\mathrm{P})\right]$$
With our target composition, the carbon equivalent is roughly 4.0 to 4.2%. This places the alloy close to the eutectic composition, which is beneficial for feeding thin sections and for avoiding shrinkage-type sand casting defects. The copper, chromium, and molybdenum additions are expensive, and their prices have risen sharply in recent years. As a result, every scrap casting caused not only a loss in output but also a substantial loss in alloy cost. This economic pressure made the reduction of sand casting defects a very clear priority for our production team.
| Process Parameter | Value |
|---|---|
| Material grade | GG30Cu |
| Molding process | Green sand, jolt-squeeze molding line |
| Castings per mold | 4 |
| Cores per group of two castings | 9 |
| Large water-jacket core | Cold-box process |
| Other cores | Hot-box process |
| Gating system | Open, single-side middle injection |
| Pouring temperature | 1400–1420°C |
| Core treatment | Assembled, dipped in core wash, oven dried |
| Venting | Vent sticks on core prints and in mold cavity |
In the original arrangement, the cores were grouped in pairs. Each group contained one large water-jacket core, upper and lower water-jacket cores, intake port cores, exhaust port cores, and core prints. The large water-jacket core was made by the cold-box process, while all other cores were made by the hot-box process. After assembly, the whole core package was immersed in core wash and then passed through a drying oven. The assembled core package was placed on storage racks and later lowered into the lower mold cavity on the molding line. The upper mold was a green sand mold, and vent sticks were added at the core prints and in the cavity to allow mold gas to escape.
In principle, this process seemed straightforward. In practice, it produced a large number of sand casting defects. The scrap rate was around 10%, which is far too high for a high-volume cylinder head line. More importantly, most of the scrap was caused by the same recurrent defect: gas porosity. It took us a long time to understand why the defect was so persistent and why it always appeared in the same location.
The Dominant Sand Casting Defects in the 226B Cylinder Head
When we started the detailed scrap analysis, we separated all sand casting defects into categories. The results were striking. Gas holes accounted for about 90% of all scrap. Sand inclusions, internal fins, broken cores, and core shift accounted for the remaining 10%. Table 2 shows the defect categories and the main contributing factors that we identified before making any major process changes.
| Defect Category | Share of Scrap | Primary Contributing Factors |
|---|---|---|
| Gas holes / blowholes | ~90% | Blocked vent paths, thin wall sections, high binder gas evolution |
| Sand inclusions | ~4% | Crushed green sand during mold closing, loose sand in core assembly |
| Internal fins / metal penetration | ~3% | Bonding paste failure in core joints, high metal temperature near ingate |
| Broken cores | ~2% | Buoyancy force on large cores, insufficient support |
| Mismatch and surface defects | ~1% | Worn tooling, large parting-line clearances |
The gas holes were not randomly distributed. They appeared mainly at the bolt holes on the side of the cylinder head that has no process hole. After knockout, and after the vent sticks were removed, we could see smooth, rounded cavities. In many castings, the holes were clustered together. This location was the farthest point from the only open vent path. The core gas generated during pouring had to travel a long distance, through a thin section, and against the pressure of the liquid metal. Because that region solidified quickly, the gas bubbles were trapped before they could reach the vent.
The second group of sand casting defects was related to the interaction between the green sand mold and the core package. During mold closing, the upper mold sometimes touched the top of the assembled core package before the mold was fully closed. This crushed the green sand and created loose sand on the cope surface. The loose sand was then washed into the casting by the iron stream, producing sand inclusions. In a cylinder head, sand inclusions are especially dangerous because they can block a water passage or create a crack initiation site under service loading.
The third group of sand casting defects was internal fins in the water jacket. The upper and lower water-jacket cores were joined with a conventional sodium silicate paste. Near the ingate, the binder was exposed to temperatures high enough to melt and decompose. When the sodium silicate softened, the joint opened slightly, and liquid iron entered the gap. The result was a fin-like projection inside the water jacket. This fin reduced the cooling-water flow area and degraded the thermal performance of the cylinder head.
The fourth group was broken cores. During pouring, the liquid iron pushed upward against the large water-jacket core. If the support provided by the core prints was not sufficient, the core bent or even broke. A broken core can produce a casting with a completely wrong internal geometry. Even a small crack in the core can cause core shift and a wall-thickness deviation. These defects are particularly difficult to detect, because the external shape may look perfect while the internal water jacket is useless.
Root-Cause Review of Sand Casting Defects
I have learned that sand casting defects are rarely random events. They are almost always the result of an imbalance between gas generation, venting capacity, solidification time, and mechanical support. In the case of the 226B cylinder head, all four factors were unfavorable.
Gas Generation and Venting
Every resin-bonded core evolves gas when it comes into contact with molten metal. The gas consists mainly of water vapor, carbon monoxide, hydrogen, nitrogen, and light hydrocarbons. The amount of gas produced depends on the binder content, the core-baking conditions, the moisture content, and the pouring temperature. The pressure inside the core can be estimated from the ideal gas law:
$$P_g V_g = n_g R T$$
where Pg is the gas pressure in the core, Vg is the free volume available in the core and its vent passages, ng is the number of moles of gas released, R is the universal gas constant, and T is the absolute temperature. If the free volume is too small, or if the vent is blocked, the pressure rises quickly and forces gas into the molten metal.
An important observation is the thermal expansion of gas. Gas generated at room temperature will expand very significantly when heated to pouring temperature. At 25°C, the absolute temperature is about 298 K. At 1420°C, it is 1693 K. The volume ratio is:
$$\frac{V_2}{V_1} = \frac{T_2}{T_1} = \frac{1693}{298} \approx 5.68$$
This means that a given amount of gas occupies nearly six times more volume at pouring temperature than at room temperature. The venting system has to cope with this enormous expansion. In the 226B cylinder head, the large water-jacket core was almost completely surrounded by liquid iron. The gas could only escape through a small number of core prints. On the side without a process hole, the escape path was long and narrow. The result was a high gas pressure that produced gas holes.
Solidification Time and Thin Walls
The cylinder head has internal walls only 4 mm thick. These thin sections freeze very quickly. The solidification time of a casting section can be estimated with Chvorinov’s rule:
$$t_s = C_m \left( \frac{V}{A} \right)^2$$
where ts is the local solidification time, Cm is a mold constant, V is the volume of the section, and A is the cooling surface area. For a flat plate of thickness t, the modulus is approximately t/2. A 4 mm wall therefore has a modulus of 2 mm. A thicker section with a modulus of 4 mm has four times longer solidification time because solidification time is proportional to the square of the modulus. During that very short time window, the gas bubbles have to escape from the metal. If they do not escape before the solidification front reaches them, they remain in the casting.
This combination is the main physical reason for the gas hole problem in thin-wall cylinder heads. The thin wall simultaneously increases the gas pressure by restricting the vent path and decreases the available escape time. The solution therefore had to do more than simply improve the vent size. It had to make the top surface of the casting form through a stable core that could carry the venting channels directly to the dangerous region.
Buoyancy and Core Support
The large water-jacket core experiences an upward force during pouring. This force can be written as:
$$F_b = \rho_{\mathrm{iron}} g V_{\mathrm{core}} – m_{\mathrm{core}} g$$
where ρiron is the density of liquid iron, g is gravitational acceleration, Vcore is the submerged volume of the core, and mcore is the mass of the core. When the submerged core volume is about 0.8 L, the upward force is roughly 45 N after subtracting the core’s own weight. This is not a huge force, but the core is slender and has a long unsupported length. The resulting bending moment can crack the core or lift it from its seat. A lifted core changes the wall thickness and can allow iron to flow around the core, creating flash and sand casting defects.
We also considered the metal pressure acting on the core prints. The local metal pressure at a given height is:
$$p = \rho_{\mathrm{iron}} g h$$
where h is the vertical height of the liquid metal above the core print. A higher pouring basin or a higher sprue creates more pressure, which can help reduce gas expansion but can also increase metal penetration into core joints. Therefore, every solution had to balance these competing effects.
First Major Solution: Adding a Top Cover Core
The most important change we made was to add a new top cover core to the core package. This one change addressed several sand casting defects at the same time. In the old design, the top surface of the cylinder head was formed by the green sand cope. The green sand could be crushed during mold closing, and it did not provide a reliable venting path. In the new design, the top surface is formed by a separate core, made on the same core-shooting machine with two cavities per cycle.
The top cover core is coated separately and located by core prints on the large water-jacket core. It can be assembled in advance and stored on a dedicated rack. We were careful to prevent loose sand from falling through the cover-core vent holes into the assembled core package. The upper mold also has a small riser added at the position where the gas holes used to form. This riser moves the thermal hot spot upward, so that the last liquid metal remains in the riser rather than in the casting. Any small gas bubble that is late to form can rise into the riser instead of becoming trapped at the bolt hole.
| Feature | Before Improvement | After Improvement |
|---|---|---|
| Cylinder head top surface | Green sand cope | Dedicated top cover core |
| Venting path to bolt-hole side | Long, narrow, unreliable | Short, stable, repeatable |
| Riser at former gas-hole location | None | Small riser added |
| Risk of crushed sand during closing | High | Very low |
| Surface finish of top face | Irregular | Consistent and clean |
The top cover core was a turning point. It reduced the gas-hole defect rate dramatically because it created a continuous, unobstructed venting channel to the region where gas had previously been trapped. It also eliminated most of the sand inclusions caused by crushed green sand, because the top surface was no longer part of the green sand mold. The core is rigid and stable, so there is no loose sand to be washed into the metal stream. The surface quality of the top face improved, and the cleaning and grinding work was reduced.
From a venting point of view, the core vent area is now more than sufficient. The venting capacity can be approximated by:
$$Q_v = C_d A_v \sqrt{\frac{2(P_g – P_m)}{\rho_g}}$$
where Qv is the volumetric flow rate through the vent, Cd is the discharge coefficient, Av is the total vent cross-sectional area, Pg is the gas pressure in the core, Pm is the pressure in the surrounding metal, and ρg is the gas density. By adding the top cover core, we increased the effective Av and reduced the distance over which the pressure drop must occur. The result was a much lower gas pressure at the critical side of the casting.
Second Major Solution: Developing a Specialized Core Paste
The second group of sand casting defects was internal fins in the water jacket caused by failure of the core bonding paste. The original paste was based on sodium silicate. It was inexpensive and easy to use, but it had a serious weakness. Near the ingate, the temperature of the liquid iron exceeded the decomposition temperature of the sodium silicate. The binder melted, the joint opened, and iron penetrated into the gap between the upper and lower water-jacket cores.
We decided to make our own specialized refractory paste. The development was guided by the thermal and mechanical conditions in the casting process. The paste has to withstand high temperature without decomposing, fill the gap between the two cores completely, and provide enough shear strength to prevent the cores from moving during assembly and pouring. The shear strength of a bonded core joint can be expressed as:
$$\tau = \frac{F_{\mathrm{applied}}}{A_{\mathrm{joint}}}$$
where Fapplied is the force acting on the joint and Ajoint is the bonded area. Because the mating surfaces of the upper and lower water-jacket cores are small, even a small force can produce high local shear stress. Therefore, the paste must penetrate into the surface pores of the cores and form a strong mechanical key.
| Property | Previous Sodium Silicate Paste | New Refractory Paste |
|---|---|---|
| Room-temperature shear strength | Moderate | Higher and more stable |
| High-temperature resistance | Softens near ingate | Remains stable above pouring temperature |
| Gap-filling behavior | Poor on thin edges | Excellent, less core distortion |
| Gas evolution | Moderate | Low |
| Risk of internal fin | High | Significantly reduced |
After introducing the new paste, the internal fin defects decreased sharply. The water-jacket cores remained aligned during assembly and pouring, and the inner surfaces of the cylinder head were much cleaner. This also improved the cooling performance of the finished cylinder head. The extra cost of the paste was small compared with the reduction in scrap and the savings in cleaning labor.
Third Major Solution: Strict Process Control
The third solution was not a new tool or a new material. It was a set of strict operational rules for the molding and core-assembly stations. I have often observed that the best tooling still fails when the process is not executed consistently. In the 226B cylinder head, many gas holes were caused by blocked vents. A vent may look open from the outside, but at the mold-metal interface it can be closed by a thin layer of core wash, loose sand, or seal paste.
Our team introduced a standard check before every mold closing. The operator must pierce every vent hole with a wire to ensure that it is fully open. The top cover core is inspected for any sand debris before it is placed on the water-jacket core. In addition, seal paste is applied to the top of the cover core in a controlled way. The paste is kept away from the vent holes. This prevents the paste from being squeezed into the vent when the mold closes. A blocked vent, no matter how small, can create enough back-pressure to produce gas holes.
This may seem like a simple detail, but it had a large effect. During the summer months, when humidity and core-moisture levels are high, sand casting defects usually increase. In the past, the scrap rate rose sharply in hot and humid weather. After the top cover core and the new operational rules were introduced, the scrap rate stayed stable even during the most difficult part of the summer. That was the first real evidence that the process had become robust.
Measured Reduction of Sand Casting Defects
After implementing these changes, we followed the scrap data for several months. The results are summarized in Table 5. The total scrap rate dropped from about 10% to 4–5%. More importantly, the dominant gas-hole defect was almost eliminated. The bolt-hole gas porosity that had caused 90% of all rejections became a rare event.
| Defect Type | Before Improvement | After Improvement |
|---|---|---|
| Gas holes / blowholes | ~90% of total scrap | Almost eliminated in the critical bolt-hole zone |
| Sand inclusions | Frequent | Significantly reduced |
| Internal fins in water jacket | Common | Rare |
| Broken cores / core shift | Occasional | No recurrence |
| Total scrap rate | ~10% | Stable at 4–5% |
| Cleaning and grinding workload | Heavy | Notably lighter |
| Surface quality | Variable | Consistent |
The economic benefit can be described with a simple cost model. Let Q be the monthly production quantity, U be the cost of a finished good cylinder head, r0 be the scrap rate before the improvement, and r1 be the scrap rate after the improvement. The monthly value of the saved castings is:
$$S = Q \cdot U \cdot (r_0 – r_1)$$
The added costs include the cost of making the top cover core and the cost of the new refractory paste. If Ccover is the added per-casting core cost and Cpaste is the added per-casting paste cost, then the total added cost is:
$$E = Q \cdot \left(C_{\mathrm{cover}} + C_{\mathrm{paste}}\right)$$
The net economic benefit is therefore:
$$B = S – E = Q \cdot \left[U \cdot (r_0 – r_1) – \left(C_{\mathrm{cover}} + C_{\mathrm{paste}}\right)\right]$$
When these numbers are put into the formula with our actual production volume and cost data, the value of the saved castings is much larger than the added process cost. The improvement not only reduced the scrap rate but also increased the effective capacity of the molding line, because we no longer waste melting capacity, molding capacity, and machining capacity on castings that will be scrapped.
Practical Lessons for Sand Casting Defects
The experience with the 226B cylinder head changed the way I think about sand casting defects in complex core packages. The first lesson is that sand casting defects are often connected to each other. Gas holes, sand inclusions, and internal fins all shared a common root: the top of the casting was not controlled well enough by the core package. A single change, the addition of the top cover core, attacked several different sand casting defects at the same time. This is why system-level thinking is so important.
The second lesson is that venting must be designed from the physics of gas evolution. More cores mean more gas. Thin walls mean less time for the gas to escape. The thermal expansion of gas by a factor of nearly six must be considered in every venting calculation. A venting system that is sufficient at room temperature may be completely insufficient at pouring temperature.
The third lesson is that the core binder or paste must match the thermal conditions at each location. A material that works perfectly in one part of the mold can fail near the ingate. The local temperature, the hydrostatic pressure, and the solidification time all determine whether a core joint remains closed or opens to allow metal penetration.
The fourth lesson is that process discipline is essential. Even the best venting system can be defeated by a blocked vent. The operator’s habit of piercing every vent wire and keeping seal paste away from the vent holes became a routine that guaranteed the theoretical design was realized in practice.
Conclusion
In summary, the 226B cylinder head taught us a practical method for solving sand casting defects. The main defect was gas porosity, caused by a combination of high core-gas generation, inadequate venting, thin wall sections, and the absence of a controlled top surface. We solved the problem by adding a top cover core, developing a high-temperature refractory paste for the water-jacket core joints, and enforcing strict venting procedures. The scrap rate decreased from about 10% to 4–5%, and the quality remained stable through hot, humid weather. The increase in process cost was small compared with the reduction in scrap loss.
The same approach can be applied to other complex castings. Start by classifying the sand casting defects and their locations. Then examine the physical mechanisms: gas generation, venting capacity, solidification time, buoyancy, and thermal stability of the core materials. Finally, choose the smallest set of changes that attack the root causes. In our case, one top cover core solved a problem that had resisted many smaller adjustments. It proved that a well-designed casting process is not merely a collection of separate operations. It is a system in which every core, every vent, every paste, and every operator action contributes to the final quality of the cylinder head.
I hope this account helps other foundry engineers who are struggling with similar sand casting defects. The solution is not always a new molding machine or an expensive alloy. Sometimes it is simply a smarter core package, a better paste, and a more disciplined operator. Those are the elements that allowed us to turn a difficult casting into a stable and profitable production part.