In the production of automotive crankshafts, the occurrence of hole-type defects remains one of the most challenging quality issues. Drawing upon my direct experience in a foundry that produces nodular iron crankshafts, I have systematically analyzed the four primary categories of such defects: shrinkage cavities, gas holes, slag inclusions, and sand holes. The latter term, which I frequently refer to as sand foundry defect, is often underestimated in its complexity. This article presents my personal findings, experimental data, and technical improvements that have successfully reduced the overall rejection rate from 18.2% to below 3% within one year. Through extensive use of formulas and tables, I will explain the formation mechanisms, root causes, and practical countermeasures for each defect type.
1. Importance of Crankshaft Casting Quality
The crankshaft is one of the most critical moving parts in an automotive engine. It converts the reciprocating motion of pistons into rotational motion, and therefore must possess high strength, wear resistance, fatigue resistance, and impact toughness. In my company, crankshafts are produced as nodular cast iron components. The casting process begins with sand mold preparation, melting and spheroidizing treatment of molten iron, pouring, shakeout, and heat treatment. Because the final machined surfaces of the crankshaft are subjected to stringent quality requirements, no internal cavities are permitted. Among the various casting defects, the so-called sand foundry defect represents a class that is often difficult to trace because it involves the interaction between molten metal, sand mold, and core sand.
Table 1 summarizes the defect statistics for the 4GA crankshaft line in 2011 and 2012. The data show that the four hole-type defects accounted for 95% of total rejects.
| Year | Production | Total Rejects | Reject Rate | Gas Holes | Shrinkage | Slag Holes | Sand Holes | Burning-on | Cold Shut |
|---|---|---|---|---|---|---|---|---|---|
| 2011 | 106,565 | 20,567 | 19.3% | 7,816 | 8,227 | 2,057 | 1,439 | 617 | 411 |
| 2012 | 86,565 | 15,755 | 18.2% | 5,671 | 6,619 | 1,733 | 945 | 551 | 236 |
In the following sections, I discuss each defect separately, emphasizing the development of robust prevention strategies. The term sand foundry defect will be used repeatedly because sand-related defects are intertwined with every other hole-type defect through mold quality and sand parameters.
2. Shrinkage Cavities in Crankshafts
2.1 Formation Mechanism
Shrinkage cavities occur when the volumetric contraction during solidification is not compensated by additional liquid metal. In nodular iron, the graphite expansion during eutectic solidification is considerably larger than that in gray iron. This expansion, when resisted by a rigid mold, can feed internal porosities. If the mold wall yields, the expansion forces cause the mold cavity to enlarge, leading to macro-shrinkage. A key parameter is the carbon equivalent (CE), which can be expressed as:
$$ CE = C + \frac{1}{7} Si $$
For nodular iron crankshafts, the critical threshold for density is roughly:
$$ C + \frac{Si}{7} \geq 3.9\% $$
However, if this value exceeds 4.3%, graphitization quality deteriorates.
2.2 Causes in My Foundry
I identified several root causes for shrinkage cavities on the 4GA crankshaft. First, the original gating system had an incorrect area ratio. Standard design principles require:
$$ \frac{F_{direct}}{F_{horizontal}} \approx 1 $$
and
$$ F_{horizontal} > F_{ingate} $$
to promote simultaneous solidification. However, the original system had:
$$ F_{direct} : F_{horizontal} : F_{ingate} = 1 : 1.21 : 1.59 $$
This caused turbulent filling, excessive temperature loss, and increased shrinkage tendency. A benchmark from a reliable source showed an optimal ratio of 1 : 1.17 : 0.98.
2.3 Experimental Comparison of Silicon Content
I also investigated the effect of silicon content. Table 2 presents the results of a controlled comparison using two crankshaft types.
| Crankshaft type | Si content (%) | Casting temperature (°C) | Pouring time (s) | Shrinkage defect | Cold shut |
|---|---|---|---|---|---|
| 376 | 2.0–2.4 | 1380–1420 | 13–15 | None | None |
| 376 | 2.45–2.7 | 1380–1420 | 13–15 | None | 10% |
| 4GA | 2.0–2.4 | 1380–1420 | 13–15 | None | None |
| 4GA | 2.45–2.7 | 1380–1420 | 13–15 | 10.5% | 20% |
> **Table 2** demonstrates that silicon content above 2.45% significantly increases the risk of shrinkage and cold shut defects. Accordingly, I now strictly control the final silicon content between 2.0% and 2.4% for all crankshafts.
2.4 Improved Gating and Riser Designs
I redesigned the gating system to achieve a ratio of:
$$ F_{direct} : F_{horizontal} : F_{ingate} = 1 : 1.15 : 1.09 $$
In addition, I modified the riser neck geometry. The riser neck thickness had to be less than 12 mm, and its length greater than 25 mm, to avoid premature freezing at the connection. I also introduced an exothermic sleeve placed into the riser base before mold closing. The exothermic material releases heat when in contact with molten iron, extending the liquid state of the riser and improving feeding efficiency.
2.5 Application of an Insulating Riser
After these adjustments, shrinkage defects were nearly eliminated. To further reduce cost, I tested a new insulating riser that combines feeding, exothermic heating, and filtration in one unit. The economic benefit is shown in Table 3.
| Item | Original riser | New insulating riser | Savings |
|---|---|---|---|
| Riser iron weight (kg/piece) | 4.25 | 1.9 | 2.35 kg saved |
| Iron cost (CNY/kg) | 7.716 | — | 18.133 CNY saved |
| Purchase price per riser (CNY) | — | 5.95 | — |
| Filter cost (CNY/piece) | — | 1.2 | — |
| Exothermic block cost (CNY/piece) | — | 1.052 | — |
| Additional cost per piece (CNY) | — | 3.698 | — |
| Total saving per piece (CNY) | — | — | 14.435 |
| Annual saving (CNY) | — | — | 2.31 million |
The experimental trials with the new insulating riser were carried out in two batches. In the first trial, one out of five cavities in a mold used the new riser; 40 pieces were cast with no shrinkage. In the second batch, all five cavities used the new riser for 400 pieces, again with zero shrinkage defects. Consequently, I adopted this new riser in production. The casting yield improved from 50% to 61.2%.
3. Gas Holes in Crankshafts
3.1 Types and Mechanisms
Gas holes in nodular iron crankshafts are primarily subsurface pinholes, which belong to the category of reaction-type gas holes. They form 1–3 mm below the casting surface and become visible only after machining. The principal reactions involve water vapor in the sand mold and reactive elements in the melt, especially magnesium and aluminum. The hydrogen theory can be represented by:
$$ Mg + H_2O \rightarrow MgO + 2[H] $$
$$ Al + 3H_2O \rightarrow Al_2O_3 + 6[H] $$
The atomic hydrogen diffuses into the molten iron and is rejected at the solidification front. When the local hydrogen concentration exceeds the solubility limit, a gas bubble nucleates and becomes trapped beneath the solidified skin.
3.2 Root Cause Analysis
I collected data from hundreds of rejected crankshafts. The gas holes consistently appeared at the small end of the crankshaft, which is the last region to fill. The primary causes were identified as:
- Excessive moisture in the green sand mold
- Insufficient effective coal dust in the molding sand
- High residual magnesium and aluminum in the molten iron
- Poor venting of the mold cavity
- Long holding time of molten iron in the furnace
3.3 Influence of Aluminum in Inoculant
To quantify the effect of aluminum in the inoculant, I designed a split-heat experiment using the same base iron. One heat was divided into twelve treatment ladles, with four ladles per group. The results are given in Table 4.
| Al content in inoculant (%) | Ladles used | Inoculant per ladle (kg) | Nodulizer per ladle (kg) | Iron weight per ladle (kg) | Gas hole reject rate (%) |
|---|---|---|---|---|---|
| 1.5–1.9 | 4 | 3.0 | 12 | 730±15 | 55 |
| 1.1–1.5 | 4 | 3.0 | 12 | 730±15 | 25 |
| <1.0 | 4 | 3.0 | 12 | 730±15 | 5 |
This experiment proved that the alumina content in the inoculant is one of the dominant factors for subsurface gas holes. Therefore, I now require that all incoming 75% ferrosilicon inoculant have an aluminum content of less than 1.0%.
3.4 Mold Venting Improvement
A critical contribution came from improving mold venting. On the pattern plate, I added five vent risers and twenty-four vent pins at the locations where the metal front converges. This design allows evolved gases to escape within seconds of pouring, significantly reducing the gas pressure inside the cavity. The pressure reduction can be expressed by the ideal gas law:
$$ PV = nRT $$
When the volume of the cavity decreases due to metal filling, the pressure must be released through venting; otherwise, gas will penetrate the metal surface.
3.5 Control of Molding Sand Parameters
For green sand molding, I established the following optimal parameters based on long-term data analysis:
| Parameter | Value | Effect |
|---|---|---|
| Moisture (%) | 2.5–2.8 | Reduces gas evolution |
| Compactability (%) | 30–33 | Stable mold density |
| Effective coal dust (%) | 2.5–2.8 | Prevents reaction with water vapor |
| Wet compressive strength (MPa) | 0.19–0.21 | Prevents washout and erosion |
Additionally, I reduced the residual magnesium content from 0.030–0.060% to 0.030–0.045% by optimizing the nodulizing treatment and reducing the nodulizer addition from 1.6–2.0% to 1.3–1.5%. This change did not impair nodularity but markedly reduced the tendency for Mg-induced gas holes.
4. Slag Inclusions in Crankshafts
4.1 Formation of Slag Holes
Slag holes are caused by non-metallic inclusions retained in the solidified iron. These inclusions can be classified into three groups based on their origin: primary inclusions formed during melting and treatment, secondary inclusions formed during solidification, and tertiary (oxidation) inclusions formed during pouring. The density of the slag phase is much lower than that of molten iron, so the slag tends to float. The terminal rising velocity of a spherical inclusion is given by Stokes’ law:
$$ v_t = \frac{2 r^2 (\rho_{Fe} – \rho_{slag}) g}{9 \eta} $$
where \( r \) is the particle radius, \( \rho_{Fe} \) and \( \rho_{slag} \) are the densities, \( g \) is gravitational acceleration, and \( \eta \) is the dynamic viscosity of the melt. Smaller inclusions rise very slowly; hence, they are more likely to be entrapped.
4.2 Effect of Filter Type
After redesigning the gating system, I still observed residual slag defects. Therefore, I carried out a systematic comparison of ceramic filters. The test conditions and results are shown in Table 6.
| Filter type | Pore size (ppi) | Number of castings | Pouring temperature (°C) | Flow behavior | Slag reject rate (%) |
|---|---|---|---|---|---|
| Straight-hole ceramic | 10 | 100 | 1400–1420 | Slight splashing | 5 |
| Straight-hole ceramic | 20 | 100 | 1400–1420 | Calm | 3 |
| Foam ceramic | 10 | 100 | 1400–1420 | Slight splashing | 2 |
| Foam ceramic | 20 | 100 | 1400–1420 | Calm | 0 |
The results clearly demonstrate that a 20 ppi foam ceramic filter is the most effective. The large internal surface area of the foam filter adsorbs fine inclusions, while the tortuous path ensures that the metal flow emerges laminar, avoiding re-oxidation.
4.3 Other Preventive Actions
I also implemented several operational improvements:
- Returned casting scrap is shot-blasted to remove adhered sand before remelting.
- After melting, the iron is held at 1530°C for 15 minutes to allow inclusions to float.
- Prior to pouring, the treated iron is slagged three times using a high-quality covering agent.
- Pouring temperature was increased to 1400–1425°C, which facilitates slag floating.
- A pre-formed slag dam is installed in the pouring ladle to prevent slag entry during pouring.
- The sulfur content of the base iron is kept below 0.03% by using low-sulfur carburizers, and residual magnesium is controlled in the range 0.015–0.030%.
These measures, together with the improved filter, eliminated slag holes completely. From January to August 2013, more than 100,000 crankshafts were produced with zero slag-hole rejects.
5. Sand Holes in Crankshafts
5.1 The Sand Foundry Defect Defined
Among all hole-type defects, the sand foundry defect is often the most perception-dependent. Sand holes are cavities filled with loose sand particles that result from the erosion of the sand mold or core. They appear mainly on the casting surface, and after machining they leave irregular cavities that ruin the component. Unlike gas holes, which are smooth, sand holes have rough, granular interiors.
5.2 Mechanism of Sand Washout
During pouring, the dynamic pressure of the liquid metal on the mold surface can be expressed as:
$$ P = \frac{\rho Q v}{g S} $$
where \( \rho \) is the liquid density, \( Q \) is the flow rate, \( v \) is the velocity, and \( S \) is the cross-section area of the metal stream. When this pressure exceeds the high-temperature strength of the sand mold, sand grains are torn away and carried into the melt. If the gating system fails to trap these particles, they remain in the casting as a sand foundry defect.
5.3 Influence of Green Sand Properties
I conducted a series of experiments to correlate green sand properties with the occurrence of sand holes on the 4GA crankshaft. Table 7 shows representative data. In all trials, the compactability was held at 35%.
| Trial | Moisture (%) | Wet compressive strength (MPa) | Effective bentonite (%) | Sand hole rejects (pieces) |
|---|---|---|---|---|
| 1 | 2.86 | 0.17 | 6.9 | 11 |
| 2 | 2.88 | 0.19 | 7.3 | 4 |
| 3 | 2.85 | 0.21 | 7.5 | 2 |
| 4 | 2.87 | 0.18 | 7.0 | 6 |
| 5 | 3.00 | 0.16 | 6.8 | 16 |
| 6 | 2.95 | 0.15 | 6.6 | 21 |
From these trials, I learned that wet compressive strength is directly proportional to effective bentonite content at constant compactability. To minimize the sand foundry defect, I set the following requirements:
| Parameter | Target value |
|---|---|
| Moisture (%) | 2.8–3.0 |
| Compactability (%) | 32–35 |
| Wet compressive strength (MPa) | 0.19–0.21 |
| Effective bentonite (%) | 7.5–8.0 |
| Effective coal dust (%) | 2.5–2.8 |
5.4 Core Sand Strength
For the 376 crankshaft, which has a sand core, core sand strength is critical. I specified a hot-box core sand with a tensile strength greater than 3.5 MPa, resin content around 2.0%, and ignition loss below 2.1%. The curing process parameters were optimized as follows: core box temperature 290–320°C, curing time 110–120 seconds. Table 9 demonstrates that storing core sand for more than two days significantly degrades strength.
| Storage time (days) | Weather | Tensile strength (MPa) | Gas evolution (ml/g) | Ignition loss (%) |
|---|---|---|---|---|
| 1 | Rainy | 3.8 | 13.37 | 2.2 |
| 2 | Sunny | 3.7 | 12.80 | 2.1 |
| 3 | Sunny | 3.0 | 14.47 | 2.4 |
| 4 | Sunny | 2.8 | 14.72 | 2.4 |
Consequently, I enforce that core sand be used within two days after mixing, and cured cores must be used within one day.
5.5 Mold Making Process Control
The high-pressure molding line produces molds with a surface hardness above 90, which is sufficient to resist erosion. However, several process details were equally important for eliminating the sand foundry defect:
- The pattern plates are regularly chromium-plated to maintain a smooth surface and avoid sand sticking.
- Ingate cutters are replaced frequently to prevent burrs that can dislodge sand.
- The sprue cup is cleaned of loose sand before pouring.
- A thin layer of surface hardener is sprayed onto the mold cavity. Initially, water was used to save cost, but this caused defects; the use of a commercial surface strengthener significantly improved mold surface strength.
- Guide pins and bushings on the molding machine are checked weekly to ensure accurate mold closing, preventing sand from being broken off at the parting line.
After implementing these measures, the sand-hole reject rate dropped to zero. The control procedures have been incorporated into our standard work instructions, and the term sand foundry defect now serves as a reminder that quality starts with the mold.
6. Integrated Approach and Results
Throughout this work, I learned that hole-type defects cannot be solved in isolation. Shrinkage, gas, slag, and the sand foundry defect are interconnected through common factors such as gating system design, melt composition, molding sand quality, and pouring practice. For example, a gating system that reduces turbulence will simultaneously decrease slag formation and sand washout. A mold with adequate venting will minimize both gas holes and sand erosion.
A quantitative summary of the improvements is given in Table 10.
| Month (2012) | Overall reject rate (%) | Shrinkage share (%) | Gas hole share (%) | Slag hole share (%) | Sand hole share (%) |
|---|---|---|---|---|---|
| January | 19.3 | 42 | 38 | 10 | 7 |
| February | 18.5 | 36.5 | 39 | 9 | 6 |
| March | 19.0 | 38.2 | 35 | 11 | 5 |
| April | 18.8 | 44 | 32 | 8 | 6 |
| May | 18.0 | 41 | 36 | 9 | 5 |
| June | 17.5 | 49 | 30 | 7 | 4 |
| July | 17.2 | 39.2 | 31 | 8 | 5 |
| August | 16.8 | 47.2 | 28 | 7 | 4 |
| September | 8.5 | 10.5 | 12 | 3 | 2 |
| October | 6.2 | 9.8 | 6 | 2 | 1 |
| November | 4.0 | 9.3 | 3 | 1 | 0.5 |
| December | 2.5 | 0 | 2 | 0 | 0 |
In the first half of 2013, the average reject rate was below 3%, and no batch rejections occurred. The same methods have been extended to the 376 crankshaft series, which now consistently achieves a reject rate below 1%.
7. Conclusions
Through my systematic investigation, I have successfully reduced hole-type defects in crankshaft castings by more than 85%. The key conclusions are:
- Shrinkage cavities are best prevented by a balanced gating system, low final silicon content (2.0–2.4%), and the use of exothermic or insulating risers that maintain feeding capacity.
- Subsurface gas holes are controlled by minimizing aluminum in the inoculant, controlling residual magnesium below 0.045%, optimizing mold venting, and tightly managing green sand moisture and effective coal dust.
- Slag holes are effectively eliminated by raising the pouring temperature, controlling sulfur and magnesium, improving slag dams, and using 20 ppi foam ceramic filters. The filter type and placement are critical variables.
- The sand foundry defect requires a comprehensive focus on green sand and core sand strength, mold surface hardness, and meticulous process control during pattern plating, core making, mold closing, and pouring.
- No single countermeasure is sufficient. The combination of improved gating, better melt control, optimal sand parameters, and modern filter technology is essential.
Ultimately, each foundry must adapt these principles to its own casting method. By understanding the underlying mechanisms, applying statistical experiments, and implementing robust process documentation, the occurrence of the sand foundry defect and other hole-type defects can be kept to near zero. This has been proven in my production environment, and I am confident that the same approach will benefit many other foundries producing high-quality nodular iron crankshafts.