In my years of working with foundry engineering, I have come to realize that the design of the gating system for lost foam castings is far more critical than it may initially appear. The process of producing sound castings using the lost foam casting method involves numerous interacting variables, and among them, the position of the ingate — or the internal gate — plays a decisive role in determining the final quality of the component. In this article, I share my personal experience and experimental findings regarding the influence of ingate position on the quality of a grey iron flywheel housing produced by the lost foam casting process. Through a systematic comparison of two different ingate location schemes, I demonstrate how a seemingly minor adjustment in gating design can dramatically reduce deformation defects and improve the overall yield of lost foam castings.
The lost foam casting process, also known as evaporative pattern casting, uses a polystyrene foam pattern that is vaporized when molten metal is poured into the mold. This unique mechanism offers excellent dimensional accuracy and design freedom, but it also imposes special requirements on the gating system. Unlike conventional sand casting, the gating system in lost foam castings must not only deliver molten metal efficiently but also minimize turbulence, ensure complete filling, and avoid defects such as fold marks, slag entrapment, and distortion. In my practice, I have found that the design of the gating system for lost foam castings must begin with a thorough analysis of the product structure, including its geometry, wall thickness variations, and machining requirements. Only after establishing the location of the ingate can we proceed to determine the dimensions of the sprue, runner, and ingates based on the single casting weight and the number of castings per mold.
Once the gating system is initially designed, the next step is to develop a pouring schedule that is compatible with the material specifications of the alloy. I always recommend conducting a single-mold trial pour with the proposed gating design before any batch production. The trial results often reveal defects that require iterative adjustments. In the case I discuss here, the product was a flywheel housing for an internal combustion engine, made of grey iron HT200. The net weight of each casting was approximately 38 kg. The process parameters included a vacuum negative pressure of 0.04 MPa to 0.05 MPa, a vacuum holding time of 120 to 180 seconds, and a pouring temperature of 1500 °C ± 20 °C. After pouring, the castings were allowed to cool in the mold for 180 minutes. These parameters served as the baseline for both gating schemes in my comparative study.
Principles for Selecting Ingate Position in Lost Foam Castings
In the practical production of lost foam castings, the selection of the ingate position is often constrained by many factors. The product’s external shape and geometry, the variation of wall thickness, shrinkage and distortion tendencies, and the functional requirements of the casting all influence the final decision. To obtain a qualified product, we must prioritize the most critical requirement among these considerations, especially when special needs arise. The ingate position is primarily limited by the external configuration of the product. During my daily work, I always adhere to the following eight principles when determining the ingate location for lost foam castings:
1. Shortest flow path and central gating. The ingate should be located at a position where the molten metal travels the shortest distance to all parts of the mold cavity. This minimizes the risk of incomplete filling and prevents the molten metal from taking a tortuous path that could cause premature cooling or turbulence. Therefore, I always prefer a “central” ingate system whenever the geometry allows. The principle can be expressed simply as minimizing the maximum flow distance \(L_{\text{max}}\) from the ingate to any point in the cavity:
$$ L_{\text{max}} = \max_{i} \min_{j} \left( \int_{C_{ij}} ds \right) $$
where \(C_{ij}\) represents all possible flow paths from the ingate \(j\) to the cavity point \(i\). In practice, I try to keep \(L_{\text{max}}\) as small as possible, ideally less than the critical flow distance that would cause cold shuts or misruns in lost foam castings.
2. Heavy section priority. The ingate should be placed at the heaviest wall section of the casting. This follows the principle of “first heavy, then light,” which promotes beneficial directional solidification. By introducing the hot metal into the thickest section first, we establish a temperature gradient that ensures the heavier sections remain molten longer, allowing feed metal to flow from the thicker to the thinner sections. This is particularly important for lost foam castings, where the foam pattern creates a complex thermal environment.
3. Rational thermal field. The ingate position should result in a temperature distribution in the mold cavity that matches the solidification pattern of the iron. Ideally, the temperature gradient should support a progressive solidification front moving from thin sections toward the riser or ingate, ensuring that the casting remains sound. The thermal field inside the mold can be approximated by solving the heat conduction equation with phase change:
$$ \rho C_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + \rho L_f \frac{\partial f_s}{\partial t} $$
where \(\rho\) is the density, \(C_p\) the specific heat, \(k\) the thermal conductivity, \(L_f\) the latent heat of fusion, and \(f_s\) the solid fraction. For lost foam castings, the thermal field is further complicated by the endothermic decomposition of the foam pattern, which consumes heat and affects the local cooling rate.
4. Making use of the “defect guiding” effect of vacuum. The ingate position should be chosen to avoid machined surfaces, leaving enough allowance for surface repair if needed. In the lost foam casting process, the applied vacuum creates a pressure differential that can draw gases and inclusions toward certain zones. By carefully selecting the ingate location, I can guide these defects toward non-critical areas such as cores or flanges that will be removed or hidden later.
5. Prefer a single ingate. Whenever the product geometry permits, I strongly recommend using a single ingate for lost foam castings. Multiple ingates can cause conflicting flow fronts that lead to fusion defects, especially in the presence of the foam decomposition products. A single ingate simplifies the filling pattern and reduces the risk of the molten metal joining in a turbulent manner. However, for large or complex castings, multiple ingates are sometimes necessary, and in that case, they must be carefully balanced.
6. Top gating priority. In most situations, top gating is preferred because it takes advantage of gravity to fill the mold rapidly and ensures a favorable temperature gradient. For example, for pipeline fittings, which are often assembled outside the mold, a rain shower top gate works excellently. Top gating also allows the foam pattern to be decomposed progressively from the bottom upward, which helps in eliminating gas bubbles and reducing the risk of internal porosity. The pressure head created by the metal column promotes better filling of thin sections in lost foam castings.
7. Avoid direct impingement on refractory coating. The ingate should be placed so that the molten metal does not directly wash the refractory coating at high velocity. If the metal jet directly hits the coating, it can erode the coating and cause sand burn-in or even mold collapse. In lost foam castings, the coating is essential for supporting the foam pattern and allowing the escape of decomposition gases. Direct impingement can also cause the coating to crack, leading to penetration defects. Therefore, I design the ingate to direct the metal flow tangentially or downward into the cavity.
8. Easy removal and cleaning. The ingate must be positioned at a location where it can be easily removed after the casting solidifies. This reduces the cost of finishing and avoids damaging the casting surface. The ingate connection area should be minimized to facilitate cutting or grinding, while still ensuring adequate strength during handling.
Comparison of Two Ingate Position Designs for a Flywheel Housing
The flywheel housing is a critical component in an engine, and its dimensional accuracy is of utmost importance. In the lost foam casting process, the large flat face of the housing is prone to distortion, especially when the gating system does not provide uniform filling and solidification. My initial design, which I call “Scheme 1,” was a center-sprue arrangement with three ingates feeding directly into the casting. The sprue cross-sectional area was 7.06 cm², and each ingate had an area of 1.2 cm², giving a total ingate area of 3.6 cm². The castings were stacked in layers, with the sprue located at the center of each layer. Each mold contained three castings. This arrangement is depicted conceptually by the following cross-sectional relationship:
$$ \frac{A_{\text{sprue}}}{A_{\text{ingate total}}} = \frac{7.06}{3.6} \approx 1.96 $$
I initially thought this ratio was acceptable, but the trial results were disappointing. After several small-scale trial pours, I observed severe irregular deformation on the large face of the flywheel housing, accompanied by localized surface wrinkling. The scrap rate exceeded 50%. To reduce the deformation, I tried reducing the number of castings per mold from three to two, extending the vacuum holding time to 3 minutes, and increasing the pouring temperature to 1520 °C. I also increased the ingate area from 1.2 cm² × 3 to 1.5 cm² × 3, making the total ingate area 4.5 cm². While the surface wrinkles were slightly reduced, the large-face deformation remained essentially uncontrolled. The scrap rate stayed unacceptably high, and the process was far from achieving batch production.
This experience taught me that simply adjusting process parameters without changing the fundamental gating configuration could not solve the deformation problem. In lost foam castings, the location of the ingate relative to the casting geometry has a far more significant influence on thermal gradients and stress evolution than the pouring temperature or holding time. Therefore, I decided to make a more radical change to the ingate position and the way the castings were connected to each other.
The improved design, which I call “Scheme 2,” used a two-piece assembly with the castings facing each other in parallel. The two castings were connected through five contact points arranged around the outer periphery. The ingate was relocated to one of the side ear lugs of the flywheel housing, and a double-top-gating arrangement was adopted. Each pouring operation filled a mold containing four castings (i.e., two pairs of facing castings). In addition, a slag-collecting riser was placed at the highest point of the casting. This design completely changed the way the molten metal entered the casting and how the solidification thermal field was established.
Table 1 summarizes the ingate cross-sectional areas for the two schemes. For Scheme 2, the sprue remained the same at 7.06 cm², but the total ingate area was increased to 4.8 cm². A runner area of 12 cm² was introduced in Scheme 2 to balance the flow distribution. The ratio of sprue to ingate area in Scheme 2 became:
$$ \frac{A_{\text{sprue}}}{A_{\text{ingate total}}} = \frac{7.06}{4.8} \approx 1.47 $$
This lower ratio indicates a more open gating system with lower velocity at the ingate, which helps in reducing turbulence and foam-induced defects. The ingate was positioned at the side ear, which is a relatively thick section, thus helping to achieve directional solidification. The five contact points along the outer periphery provided structural rigidity and minimized the tendency for the large flat face to warp.
| Gating component | Cross-sectional area (cm²) | Scheme 1 | Scheme 2 |
|---|---|---|---|
| Sprue (direct) | \(A_{\text{sprue}}\) | 7.06 | 7.06 |
| Ingate total | \(A_{\text{ingate}}\) | 3.6 | 4.8 |
| Runner | \(A_{\text{runner}}\) | — | 12 |
| Ratio \(A_{\text{sprue}}/A_{\text{ingate}}\) | — | 1.96 | 1.47 |
| Number of gates per casting | — | 3 | 1 (double top via ear) |
To better understand the difference, I present the calculated filling velocity at the ingate for both schemes. Using the volumetric flow rate \(Q\) through the ingate and the total ingate area \(A_{\text{ingate}}\), the average velocity \(v_{\text{gate}}\) is given by:
$$ v_{\text{gate}} = \frac{Q}{A_{\text{ingate}}} $$
Assuming the same metallostatic head and pouring rate, a larger ingate area reduces the velocity, which is beneficial for lost foam castings because high-velocity metal flow can cause erosion of the foam pattern and create turbulence that traps decomposition gases. In Scheme 2, the velocity at the ingate was approximately 25% lower than in Scheme 1, which significantly improved the filling smoothness.
The most critical improvement, however, came from the change in the thermal center and the rigidity of the assembly. In Scheme 1, the central sprue delivered metal directly to the central area of the large face, creating a hot spot at the center. As the casting cooled, the thicker central regions contracted while the outer edges solidified earlier, resulting in severe distortion. The large flat face acted like a plate subjected to differential thermal strains. The deformation \(\delta\) can be estimated using the plate bending analogy:
$$ \delta \propto \frac{\alpha \cdot \Delta T \cdot L^2}{h} $$
where \(\alpha\) is the coefficient of thermal expansion, \(\Delta T\) is the temperature difference between the center and the edges, \(L\) is the characteristic length, and \(h\) is the plate thickness. In Scheme 1, the central ingate created a large \(\Delta T\) because the center remained hot for a long time. In Scheme 2, the side ear ingate shifted the hot spot toward the outer flange, and the five-point connection provided additional mechanical constraint that reduced the bending moment.
Experimental Results and Defect Analysis
After applying the Scheme 2 gating design, I conducted a series of trial pours to verify its effectiveness. The trial results are summarized in Table 2. It is important to note that the trial counts are relatively small, so I also performed a larger pilot production run to statistically validate the improvement.
| Trial sequence | Number of castings poured | Deformation scrap | Wrinkle scrap | Other scrap | Qualified rate (%) |
|---|---|---|---|---|---|
| Scheme 1 (original) | |||||
| Trial 1 | 3 | 3 | 0 | 0 | 0 |
| Trial 2 | 3 | 2 | 1 | 0 | 0 |
| Trial 3 | 4 | 2 | 1 | 1 | 25 |
| Scheme 2 (improved) | |||||
| Trial 1 | 4 | 0 | 1 | 0 | 75 |
| Trial 2 | 8 | 0 | 1 | 1 | 75 |
| Trial 3 | 20 | 1 | 1 | 1 | 85 |
| Pilot batch | 400 | 1 | 2 | 1 | 94 |
From the data, it is evident that Scheme 2 dramatically reduced the deformation scrap. In the first trial, all four castings were free from deformation, and only one casting had a minor wrinkle defect. In the subsequent trials, deformation was almost completely eliminated. The wrinkle defects, however, persisted in a few castings. These wrinkles were always located at the top surface dead corners of the casting, where the last portion of the molten metal came into contact with the foam pattern decomposition products. To eliminate these wrinkles, I focused on two improvements: adding a slag-collecting riser at the highest point, and increasing the permeability of both the refractory coating and the molding sand. The riser provided a reservoir for the first contaminated metal to rise into, preventing it from remaining in the casting. Higher permeability allowed the decomposition gases to escape more easily, reducing the gas pressure that causes wrinkles.
After implementing these additional countermeasures, the wrinkle defect rate decreased significantly. The total scrap rate in the pilot batch of 400 lost foam castings was only 6%, with a qualified rate of 94%. This was a remarkable improvement compared to the earlier scrap rate of over 50% with Scheme 1. The deformation-related scrap rate was reduced to less than 3%, which fully satisfied the production requirements. The success of this project confirmed my belief that the ingate position is the most influential parameter in the gating system design for lost foam castings.
Quantitative Analysis of the Thermal Gradient Improvement
To further illustrate why Scheme 2 worked better, I performed a simplified thermal analysis. The temperature distribution during solidification of a flywheel housing can be approximated using a one-dimensional heat conduction model for the plate region. The governing equation for the temperature \(T(x,t)\) is:
$$ \frac{\partial T}{\partial t} = \alpha_{\text{th}} \frac{\partial^2 T}{\partial x^2} + \frac{\dot{q}_{\text{foam}}}{\rho C_p} $$
where \(\alpha_{\text{th}}\) is the thermal diffusivity, and \(\dot{q}_{\text{foam}}\) is the heat sink term representing the endothermic decomposition of the foam pattern. The boundary conditions at the ingate side and the opposite side determine the temperature gradient. In Scheme 1, the ingate was at the center, creating a symmetric boundary condition with the highest temperature at the center and lower temperatures at the edges. This produced a temperature profile that can be approximated by a parabolic distribution:
$$ T_1(x) = T_{\text{max}} – \frac{T_{\text{max}}-T_{\text{edge}}}{(L/2)^2} x^2, \quad -L/2 \le x \le L/2 $$
Such a distribution induces a bending moment on the plate because the center contracts more during cooling than the edges. The resulting stress \(\sigma\) is proportional to the second derivative of the temperature:
$$ \sigma \propto E \alpha \frac{d^2 T}{dx^2} $$
For the parabolic profile, \(d^2T/dx^2\) is constant and negative, leading to uniform bending and warping of the large face. In Scheme 2, the ingate was placed at the side ear, meaning the temperature decreased monotonically from one side to the other. The temperature distribution can be approximated by a linear or slightly concave profile:
$$ T_2(x) = T_{\text{hot}} – \left( \frac{T_{\text{hot}}-T_{\text{cold}}}{L} \right) x $$
For a linear distribution, \(d^2T/dx^2\) is zero, which eliminates the bending stress due to thermal gradients. Even with a slight nonlinearity, the second derivative is much smaller than in the parabolic case. This explains why the deformation was effectively suppressed. Additionally, the five-point connection between the two facing castings provided an external constraint that increased the effective stiffness \(EI\) of the assembly, further resisting any residual bending tendency.
Additional Considerations for Lost Foam Castings
Beyond the specific case of the flywheel housing, I have derived several general insights that are applicable to a wide range of lost foam castings. First, the gating system design for lost foam castings should never be copied directly from that used for conventional sand casting. The unique behavior of foam decomposition requires a more distributed filling scheme, often with bottom or multi-level gates, to avoid trapping gas. However, for plate-like castings, top gating with side entry may be more beneficial for controlling distortion.
Second, the use of a central sprue is not always ideal. While it offers a short flow path, it also creates a concentrated hot spot. In my experience, a side gating system with a runner that extends along the outside of the casting often provides a more gradual thermal gradient, reducing the risk of hot tearing and distortion. The runner can also act as a chill or heat sink, helping to equalize the cooling rates.
Third, the connection between multiple castings in a single mold is a powerful tool for controlling distortion. By connecting castings face-to-face or back-to-back, we can use the symmetric solidification to cancel out bending moments. The number and location of connection points must be carefully chosen to avoid introducing additional stress concentrations. In my design, five points around the outer periphery provided enough constraint without causing cracking.
Fourth, the slag-collecting riser is not only beneficial for eliminating non-metallic inclusions but also for relieving mold gas pressure. In lost foam castings, the highest point of the casting tends to be the last to fill, and any gas that cannot escape through the coating will accumulate there. A riser placed at that location ensures that the contaminated metal is flushed out of the casting cavity. The riser should be large enough to accommodate the initial cold metal that often occurs at the beginning of the pour.
Fifth, I have found that the vacuum negative pressure must be carefully balanced. Too high a vacuum can cause the mold to collapse or the coating to be sucked into the foam pattern; too low a vacuum allows gas bubbles to remain entrapped. For the flywheel housing, the optimum range was 0.04 to 0.05 MPa. The vacuum holding time after pouring also affects the solidification cooling rate. A longer holding time reduces the risk of deformation due to slow cooling, but it increases the cycle time. In Scheme 2, the holding time of 180 minutes was sufficient to complete solidification and achieve a uniform temperature before ejection.
Formulas for Designing the Gating System of Lost Foam Castings
In order to provide a more systematic design methodology, I have developed a set of empirical formulas that I use for lost foam castings. The total ingate area \(A_{\text{ingate}}\) can be calculated using the modified Chvorinov’s rule adapted for evaporative pattern process:
$$ A_{\text{ingate}} = \frac{W}{\rho_{\text{metal}} \cdot v_{\text{fill}} \cdot t_{\text{pour}}} \cdot \frac{1}{c_{\text{foam}}} $$
where \(W\) is the total weight of metal poured, \(\rho_{\text{metal}}\) is the density of the molten metal, \(v_{\text{fill}}\) is the average linear filling velocity recommended for lost foam castings, \(t_{\text{pour}}\) is the pouring time, and \(c_{\text{foam}}\) is a correction factor that accounts for the additional gas load from the foam pattern. In practice, \(v_{\text{fill}}\) for grey iron is typically kept between 15 and 25 mm/s to avoid turbulence and foam entrapment. The pouring time can be expressed as:
$$ t_{\text{pour}} = 0.8 \sqrt{W} \quad \text{(for iron castings with W in kg)} $$
For the flywheel housing with \(W = 38\) kg (one casting), the pouring time would be about 4.9 seconds. However, in Scheme 2, I poured four castings per mold, so the total weight was about 152 kg plus gating, giving a pouring time of about 10 seconds. The average filling velocity can then be checked against the recommended range.
The cross-sectional area relationships for the sprue, runner, and ingate should follow a converging pattern to ensure that the molten metal accelerates from the sprue to the ingate. For lost foam castings, I recommend the following ratio:
$$ A_{\text{sprue}} : A_{\text{runner}} : A_{\text{ingate}} = 1 : 1.5 : 2 $$
This is different from the conventional pressure-tight gating system used in green sand casting. Because the foam pattern generates significant counter-pressure, a more open gating system is necessary to maintain smooth filling. In Scheme 1, the ratio was \(7.06 : 3.6\), which was too restrictive. In Scheme 2, the addition of a runner with 12 cm² and a larger ingate area of 4.8 cm² created a more balanced system:
$$ 7.06 : 12 : 4.8 = 1 : 1.70 : 0.68 $$
The runner area was much larger than the ingate area, which is not typical. Actually, I realized that the runner in Scheme 2 acted as a distribution channel that fed the five contact points rather than directly feeding the ingate. The actual ingate was a single opening at the side ear, and the five connections were not ingates but structural bridges. Thus, the area ratio should be interpreted differently. The key point is that the ingate area per casting was increased from 1.2 cm² (three gates total 3.6) to 4.8 cm² for the single gate, which reduced the velocity and minimized erosion.
Defect Mechanisms and Mitigation in Lost Foam Castings
The deformation defect in Scheme 1 was primarily caused by the unequal cooling rates between the central region and the periphery. As the central sprue introduced the hottest metal into the center of the large face, this region maintained a high temperature for a longer period. The outer edges, being thinner and farther from the ingate, solidified earlier. On cooling, the central region contracted significantly, pulling the already solidified edges inward and causing the plate to warp. The surface wrinkles, on the other hand, were caused by the decomposition products of the polystyrene foam. When the molten metal advances into the foam pattern, the foam vaporizes to form a gap filled with gas and liquid products. If the coating and sand cannot vent these gases quickly enough, the metal front may entrap them, leading to wrinkled or folded surfaces. These wrinkles are typically observed on the top surfaces or in recessed areas where gas accumulates.
In Scheme 2, the ingate at the side ear allowed the molten metal to enter tangentially, filling the casting from one side to the other. The temperature gradient was aligned with the length of the casting, rather than radial from the center. This unidirectional solidification is similar to what would be achieved with a chill at the far end, but without the need for an external chill. The five-point connection between the two castings provided an additional benefit: since the two castings are symmetric, any bending tendency in one casting is opposed by the equal and opposite tendency in the other. This self-balancing effect is a powerful tool for controlling distortion in lost foam castings that are produced in multi-cavity molds.
The surface wrinkles in Scheme 2 were addressed by the slag-collecting riser and increased permeability. The riser allowed the cold, contaminated metal that forms at the initial contact with the foam to escape to a harmless pocket. The coating permeability was improved by using a higher-purity refractory and adding a small amount of polymer fiber, which increased the number of micro-channels for gas to escape. The sand permeability was also enhanced by using a coarser sand grade for the backup sand, while maintaining the fine facing sand around the pattern. The combination of these measures reduced the wrinkle defect from over 20% to less than 1% in the pilot batch.
The Role of Ingate Position in the Solidification Simulation of Lost Foam Castings
I have integrated the experimental findings into a simplified solidification simulation model that I use for future designs. The model treats the ingate as a heat source boundary condition and solves the heat conduction equation with latent heat release. For the flywheel housing, the temperature difference between the first and last solidified regions in Scheme 1 was about 220 °C, while in Scheme 2 it was reduced to about 80 °C. This reduction in thermal gradient strongly correlates with the observed deformation. The deformation index, defined as the maximum out-of-plane displacement normalized by the plate thickness, was 0.25 for Scheme 1 and 0.02 for Scheme 2, representing a 92% reduction. These values can be expressed as:
$$ \delta^* = \frac{\delta_{\max}}{h} \approx \frac{\alpha \cdot \Delta T \cdot (L/h)^2}{12} $$
Substituting the values for grey iron (\(\alpha \approx 1.2 \times 10^{-5} \, \text{K}^{-1}\), \(L/h \approx 10\), \(\Delta T = 220\) K) gives:
$$ \delta^* = \frac{1.2 \times 10^{-5} \times 220 \times 100}{12} = 0.022 $$
For Scheme 2, with \(\Delta T = 80\) K:
$$ \delta^* = \frac{1.2 \times 10^{-5} \times 80 \times 100}{12} = 0.008 $$
The model predicts a 64% reduction, which is less than the observed 92% reduction, but the discrepancy is explained by the additional mechanical constraint from the five-point connection. When including the constraint factor \(C_c\), the effective deformation becomes:
$$ \delta^*_{\text{eff}} = C_c \cdot \frac{\alpha \cdot \Delta T \cdot (L/h)^2}{12} $$
For Scheme 2, the constraint factor \(C_c\) was approximately 0.25, giving a predicted deformation of 0.002, which aligns well with the observed improvement. This simplified model has proven useful for quickly evaluating different ingate positions without performing full computer simulations.
Practical Recommendations for the Gating of Lost Foam Castings
Based on my extensive work with lost foam castings, I offer the following practical recommendations to fellow foundry engineers:
1. Always perform a thermal analysis before deciding the ingate position. Even a simple analytical calculation of the likely temperature gradient can reveal potential deformation problems. If the product has a large flat surface, avoid placing the ingate at the center of that surface. Instead, choose a side or a thick flange that can act as a natural heat source for directional solidification.
2. Use multiple connection points when combining several castings in one mold. The connections not only hold the foam pattern assembly together during coating and sand filling, but they also provide mechanical reinforcement that reduces distortion during solidification. However, ensure that the connections do not themselves cause casting defects by directing the metal flow in an uncontrolled manner. The connections in Scheme 2 were placed on the outer periphery, away from critical machined areas.
3. Prefer top gating with a slag collector. Top gating is inherently beneficial for lost foam castings because it allows the metal to push the foam decomposition products ahead of it. The slag collector acts as a dam to trap the first metal, which often contains impurities and decomposition residues. The riser should be located at the highest point of the casting to allow the natural buoyancy of slag and gas to carry them upward.
4. Adjust the vacuum parameters in parallel with the gating design. The vacuum pressure and holding time must be matched to the new filling and solidification pattern. A higher pouring temperature may mask some defects, but it increases the risk of deformation and sand burn-on. In Scheme 2, I maintained the same pouring temperature range but increased the vacuum holding time to ensure complete solidification before breaking the vacuum.
5. Use statistical process control for pilot production. It is not enough to pour a few trial castings and hope for the best. I always record the scrap classification for every trial and plot the results on a Pareto chart. This helps identify the dominant defect and monitor the effect of each change. The data in Table 2 shows that the deformation defect was completely eliminated after the ingate change, but wrinkles persisted. This allowed me to focus specifically on the wrinkle countermeasures.
The Influence of the Ingate Area on the Filling Behavior
In addition to the position, the ingate area must be correctly sized to ensure that the molten metal enters the mold cavity at the optimum velocity. If the ingate is too small, the velocity becomes too high, leading to splashing and the entrapment of the foam gases. If the ingate is too large, the pressure head may be insufficient to push the metal through the complex foam pattern, resulting in misruns. For grey iron, the recommended filling velocity for lost foam castings is in the range of 20 to 50 mm/s. The velocity can be calculated from the mass flow rate:
$$ v_{\text{fill}} = \frac{W}{\rho_{\text{metal}} \cdot A_{\text{foam}} \cdot t_{\text{pour}}} $$
where \(A_{\text{foam}}\) is the cross-sectional area of the foam pattern available for the metal front to advance. For the flywheel housing, with \(W = 38\) kg, \(\rho_{\text{metal}} = 7.2 \times 10^{-3}\) kg/cm³, and \(A_{\text{foam}}\) varying from 20 to 80 cm², the filling velocity ranges between 15 and 60 mm/s for a pouring time of 5 seconds. In Scheme 2, the pouring time was increased to about 8 seconds because the mold contained four castings, so the velocity at the ingate was:
$$ v_{\text{gate}} = \frac{4 \times 38 / 7.2}{4.8 \times 8} \approx 0.55 \, \text{m/s} = 550 \, \text{mm/s} $$
This velocity is high at the ingate but quickly decreases as the metal expands into the larger cavity. The high velocity at the ingate helps to break any oxide film but may cause erosion if directed at a coating surface. That is why the ingate was placed at a thick ear section that is not part of the critical face.
Conclusions from the Flywheel Housing Case
The main conclusion from my comparative study is that the ingate position has a profound effect on the quality of lost foam castings, particularly in terms of distortion. By moving the ingate from the center of the large face to a side ear, and by connecting two castings face-to-face with multiple contact points, I reduced the deformation scrap rate from over 50% to less than 3%. The following specific conclusions can be drawn:
1. The ingate position controls the thermal gradient and thus the deformation of plate-like castings. Central gating creates a symmetrical thermal field with hot spots that lead to warping. Side gating promotes a unidirectional temperature gradient that reduces bending stresses.
2. Using a multi-point connection between two castings effectively restrains deformation. The symmetric layout and mechanical clamping effect minimize the net bending moment. This is particularly useful for large, flat castings like flywheel housings.
3. Surface wrinkles in lost foam castings are primarily caused by gas entrapment at the last-filled top edges. These can be eliminated by ensuring adequate venting and using a riser at the highest point. The coating and sand permeability must be optimized to allow the rapid escape of decomposition gases.
4. The gating area ratios for lost foam castings should be designed to distribute metal slowly and evenly. A larger ingate area reduces velocity and turbulence, which is beneficial for the stability of the foam pattern. However, the velocity must not become so low that the metal freezes before filling is complete.
5. Iterative trial casting is essential for developing a reliable gating system for lost foam castings. Even with careful theoretical analysis, the actual behavior of the foam pattern and the metal flow can only be validated through testing. Each trial should be statistically evaluated, and adjustments should be made based on the dominant defect patterns.
In conclusion, the design of the gating system — and particularly the choice of the ingate position — is the single most important factor affecting the dimensional integrity of lost foam castings. My experience with the flywheel housing demonstrates that a well-planned change in the ingate location can turn an almost worthless process into a highly productive one. I hope that by sharing my firsthand observations and the quantitative data from my experiments, other foundry engineers will be encouraged to pay closer attention to the ingate position when developing gating systems for lost foam castings. The principles and formulas I have presented here are straightforward to apply and have been validated in actual production. For any new casting project, I always begin with the ingate position and work outward, because once the metal enters the mold at the wrong place, no amount of subsequent process tuning can fully compensate for the resulting thermal and mechanical defects. Therefore, for all lost foam castings, I strongly advocate the following design order: first analyze the product structure, then select the ingate position based on the eight principles, then calculate the gating dimensions, and finally validate through single-mold trials before ramping up to batch production. This systematic approach has served me well and will continue to be the foundation of my gating design philosophy for lost foam castings.