The rapid growth of the heavy-duty machine tool industry has placed increasingly stringent requirements on the dimensional accuracy and internal soundness of large structural castings. Among these castings, the rotary table is one of the five major components of a machine tool and is directly responsible for the machining accuracy and service life of the whole equipment. The main challenge in producing rotary table castings is that the guide-way surface and the work-table surface are both critical functional areas, and they are located on opposite sides of the casting. In addition, the wall thickness varies dramatically between thick sections and ordinary walls. In our foundry, the casting process for grey iron rotary table castings has already been mature for many years, but the same geometry in ductile iron castings caused frequent random defects. This contribution summarizes how we systematically prevented and eliminated casting defects in large rotary table castings made of QT600-3 ductile iron castings. The discussion concentrates on the pouring system, the riser design, and the melting practice, all of which had to be optimized together. Special attention is paid to the riser neck geometry because our experiments prove that the large-diameter riser with a small-diameter and sufficiently tall neck is the most reliable solution for feeding ductile iron castings of this type.
A typical case is an 11 m rotary table casting produced from QT600-3 ductile iron castings, with a total mass of about 85 tons and a maximum diameter of 10 500 mm. The structural features that create casting difficulties are as follows. First, the wall thickness difference is very large: the work-table face is 130 mm thick and becomes about 150 mm thick after the machining allowance is added, while the outer guide-way hot spot has a hot-spot circle diameter of about 235 mm, and the ordinary walls are only 40 mm thick. Second, the work-table face and the guide-way face must be entirely free of defects. The guide-way face is a sliding surface that affects the accuracy and lifetime of the machine tool, while the work-table face is visible to the customer and must also have no porosity or shrinkage cavities. Both faces require high strength and hardness with uniform composition. Third, the T-slots on the work-table face are 90 mm deep, which means that the zone within 110 mm below the work-table face must be sound. Fourth, the casting is large and heavy, making gating and risering calculations difficult; the maximum rotary table diameter is 10 500 mm and the mass reaches 85 tons. These constraints demand a robust foundry process that can consistently deliver sound ductile iron castings.
The overall molding practice for this kind of ductile iron castings is based on pit molding with furan resin sand. The casting is oriented with the large planar surface at the bottom of the mold. This orientation is selected for two reasons: it reduces the incidence of surface defects on the critical face, and it makes the pouring of the very large casting easier because the mold is open from the top. Chills are placed in thick sections having a thickness of 60 to 100 mm in the initial stage. However, the chills alone were not able to ensure soundness of the ductile iron castings, as will be explained later. The remainder of the paper details the gating system, riser design and melting process that eventually produced completely healthy castings.
Design of the Pouring System
The primary goal of the pouring system is to deliver liquid iron into the mold cavity uniformly and quiescently, minimizing the generation of oxides and slag. The rotary table casting is essentially a half-disc shape. To achieve the best slag-removing effect, we adopted a semi-circular, profile-following, fully closed gating system with bottom gating. Because of restrictions imposed by the lifting equipment and the foundry pit, the metal was poured at two points separated by the maximum distance on the rotary table. The runner was divided into two segments, and ingates were uniformly distributed along each segment. Computer simulation software was used to verify that the metal front advanced smoothly and filled the cavity without jetting or dead zones.
The gating system is classified as fully closed, meaning that the total cross-sectional area of the ingates is the smallest among the gating elements. This keeps every part of the gating system full of metal during pouring, which greatly reduces the possibility of entraining slag. For the QT600-3 ductile iron castings, ten ceramic tubes of 50 mm diameter were used as ingates. The total effective ingate area can be calculated as follows:
$$
A_{g} = n \times \frac{\pi d_{g}^{2}}{4}
$$
where \( n \) is the number of ingates and \( d_{g} \) is the inner diameter of each ceramic tube. Substituting \( n = 10 \) and \( d_{g} = 50\,\text{mm} \) gives:
$$
A_{g} = 10 \times \frac{\pi \times 50^{2}}{4} \approx 19\,635\,\text{mm}^{2}
$$
This large ingate area allowed the massive mold cavity to be filled quickly. The filling time was estimated using the classical formula:
$$
t_{f} = \frac{G}{\mu \rho A_{g} \sqrt{2gH_{p}}}
$$
In this equation, \( G \) is the total pouring mass of ductile iron castings, \( \mu \) is the flow coefficient, \( \rho \) is the density of liquid iron, \( g \) is gravitational acceleration, and \( H_{p} \) is the average effective pressure head. For bottom gating, the effective head is approximated by:
$$
H_{p} = H_{0} – \frac{c}{2}
$$
where \( H_{0} \) is the metal head above the ingates and \( c \) is the height of the casting cavity above the ingates. The pouring system was balanced so that each of the two segments carried roughly the same amount of metal. Table 1 summarizes the final parameters of the pouring system used for the 11 m rotary table.
| Parameter | Value |
|---|---|
| Number of pouring points | 2 |
| Gating type | Fully closed, bottom gating |
| Runner segments | 2 |
| Number of ingates | 10 |
| Ingate tube diameter | 50 mm |
| Total ingate area | ≈ 19 635 mm² |
| Pouring temperature | 1 330 – 1 360 °C |
| Flow control | Stopper/plug pouring |
The gating ratio of the fully closed system was determined by restricting the total ingate area while maintaining sufficiently large runner areas. The aim was to produce a steady, laminar filling pattern that is particularly important for large ductile iron castings because turbulent flow creates dross inclusions that are difficult to remove later. The two-segment runner also helped to equalize the flow length from each sprue, minimizing the temperature drop between the first and last metal entering the cavity.
Riser Design and Feeding of Ductile Iron Castings
The riser must perform two functions: venting gas and feeding liquid shrinkage. The design of risers for ductile iron castings differs significantly from that for grey iron castings because ductile iron undergoes a graphite expansion during the later stages of solidification. This expansion, if properly harnessed, can provide self-feeding and reduce the size of the external risers. However, if the riser neck is too large or too short, the liquid starved area under the riser will be pulled back into the riser during the expansion phase, causing shrinkage cavities in the casting. Therefore, the riser neck geometry is the key to sound ductile iron castings.
In our early production trials, the riser diameter was chosen empirically as 0.85 times the thickness of the section that required feeding. Thus, for the guide-way section, the riser diameter was roughly 0.85 times the guide-way thickness, and for the maximum hot-spot circle of 230 mm (with chills) a riser of 100 mm diameter was used. The riser neck height was only 20–30 mm and the neck diameter was 30–40 mm. The risers were spaced at intervals of 600–700 mm along the two guide-way zones.
This initial design was based on the old rule that the neck diameter should be about one half of the riser diameter. However, the actual performance on the QT600-3 ductile iron castings was not satisfactory. Although separately cast test bars met the material specification, several shrinkages always appeared under a few risers, and the positions and number varied from casting to casting. In addition, after machining the T-slots, intermittent shrinkage cavities were found at the bottom of the T-slots, usually aligned with the internal cross-shaped ribs. From the radiographs and section examinations, we realized that the hot spots generated at the junction between the ribs and the bottom plate were not sufficiently chilled. The original 60 mm chills on the work-table face and 80 mm chills on the guide-way face were not thick enough to compensate for the heavy local thermal center.
We then increased the chill thickness on the work-table face from 60 to 100 mm, and on the guide-way face from 80 to 110 mm. The simulation results improved, but the actual castings still exhibited occasional porosity under some risers. The reason is that the chill on the cope surface (the upper face of the mold) cannot remain in effective contact with the casting during the whole solidification cycle; once the solidified shell contracts, the chill separates from the casting surface and loses its cooling effect. Hence, merely increasing the chill thickness is not a robust solution for the cope surface of large ductile iron castings.
We therefore decided to completely redesign the risers. Our goal was to create a large reservoir of liquid metal with high fluidity during the initial liquid contraction stage, but to make the riser neck freeze off at exactly the right moment before the graphite expansion stage. For ductile iron castings, the ideal riser neck must be small in diameter and tall in height. A large neck would keep the riser connected to the casting during graphite expansion, causing the expanding casting to push liquid iron back into the riser and leaving a cavity in the casting. A small and tall neck, on the other hand, seals off the riser once the liquid contraction is complete. In our final design, the riser diameter was increased to 1.3 times the diameter of the hot-spot circle. For the guide-way zone with a hot-spot circle diameter of about 230 mm, the riser diameter became about 300 mm. The riser neck diameter was limited to no more than 30 mm, and the neck height was not less than 50 mm. These values were determined by a series of simulation runs and corroborated by production trials.
The solidification modulus of the riser neck can be expressed by the following formula for a cylindrical neck:
$$
M_{n} = \frac{d_{n} h_{n}}{4 h_{n} + 2 d_{n}}
$$
where \( d_{n} \) is the neck diameter and \( h_{n} \) is the neck height. Taking the final dimension \( d_{n} = 30\,\text{mm} \) and \( h_{n} = 50\,\text{mm} \):
$$
M_{n} = \frac{30 \times 50}{4 \times 50 + 2 \times 30} = \frac{1500}{200 + 60} = 5.77\,\text{mm}
$$
For a large riser with diameter \( D_{r} = 300\,\text{mm} \), its modulus is approximately:
$$
M_{r} = \frac{D_{r}}{6} = \frac{300}{6} = 50\,\text{mm}
$$
Thus the neck modulus is much smaller than the riser modulus, guaranteeing that the neck freezes long before the riser center. The casting section modulus around the neck was in the range of 15–25 mm after the application of chills, so the neck modulus of 5.77 mm was less than one third of the local casting modulus, ensuring a positive seal. Of course, the exact values must be adjusted for the actual carbon equivalent and pouring temperature, but the principle remains valid for all sound ductile iron castings.
Table 2 compares the initial and final riser designs used for the rotary table.
| Design parameter | Initial design | Final design |
|---|---|---|
| Riser diameter factor | 0.85 × section thickness | 1.3 × hot-spot circle diameter |
| Riser neck diameter | 30 – 40 mm | ≤ 30 mm |
| Riser neck height | 20 – 30 mm | ≥ 50 mm |
| Riser spacing | 600 – 700 mm | 600 – 700 mm |
| Neck modulus | ~8 – 12 mm | ~5.8 mm |
| Result | Random shrinkage under risers | Sound castings |
It is important to emphasize that the use of large risers is not contradictory to the well-known fact that oversized risers can increase shrinkage defects in ductile iron castings. The correct interpretation is that the riser body must be large enough to feed liquid contraction, but the connection between the riser and the casting must be delicate. A small and tall neck makes the riser self-sealing at the proper moment, so the graphite expansion is fully utilized for self-feeding. This is the real reason why our final design succeeded where the earlier “balanced” risers failed. The improved risers were used in combination with chilled surfaces, and the CAE software confirmed that no shrinkage remained in the riser-neck zone.
Melting and Inoculation of Ductile Iron Castings
The mechanical properties of the QT600-3 ductile iron castings are extremely sensitive to the chemical composition and the treatment process. We used a 20-ton medium-frequency induction furnace to melt synthetic iron. The target chemical composition for the rotary table casting is shown in Table 3.
| Element | Mass fraction / % |
|---|---|
| C | 3.45 – 3.55 |
| Si | 1.8 – 2.1 |
| Mn | 0.5 – 0.6 |
| P | < 0.055 |
| S | < 0.01 |
| Mg | 0.05 – 0.07 |
| Cu | 0.5 – 0.6 |
| Sb | 0.005 – 0.010 |
| Sn | < 0.007 |
The carbon equivalent is calculated using the common equation for ductile iron castings:
$$
\mathrm{CE} = w(\mathrm{C}) + \frac{1}{3} w(\mathrm{Si}) + \frac{1}{3} w(\mathrm{P})
$$
At the target values, \( w(\mathrm{C}) = 3.50\% \), \( w(\mathrm{Si}) = 1.95\% \), and \( w(\mathrm{P}) = 0.04\% \), the carbon equivalent is about:
$$
\mathrm{CE} = 3.50 + \frac{1.95}{3} + \frac{0.04}{3} = 3.50 + 0.65 + 0.013 = 4.163\%
$$
This carbon equivalent is typical for a pearlitic ductile iron grade such as QT600-3. The silicon content was maintained in a narrow range because high silicon increases ferrite and lowers hardness, while too low silicon makes the iron difficult to graphitize. Copper was added as a pearlite stabilizer to achieve the required strength and hardness. Antimony and tin were kept very low to avoid excessive segregation at cell boundaries.
The magnesium treatment was carried out in a covered ladle using a FeSiMg alloy. The residual magnesium was controlled between 0.05% and 0.07%, which proved to be the optimal interval for this section thickness. A residual magnesium above 0.07% produces compacted or chunky graphite and increases the shrinkage tendency; below 0.05% the nodule count decreases and carbides can form. The aim was to obtain a nodularity of at least 85% with a nodule count above 100 per square millimetre.
Inoculation is of paramount importance for heavy-section ductile iron castings. We applied a four-stage inoculation practice: first, an inoculation alloy was placed in the bottom of the ladle; second, a second amount was added to the metal stream during tapping; third, a surface cover was applied immediately after tapping; and fourth, stream inoculation was used during pouring. This multiple inoculation scheme effectively prevents the fading of the inoculant, which is especially dangerous for large ductile iron castings that require a long pouring time. The pouring temperature was strictly controlled at 1 330–1 360 °C. This temperature range was chosen because it ensures good fluidity, but is not so high as to promote excessive liquid contraction or reduce the efficiency of the graphite expansion. At higher pouring temperatures, feeding becomes difficult and the risk of shrinkage porosity under the risers increases.
A useful parameter for controlling the solidification regime of ductile iron castings is the dimensionless feeding coefficient \( f_{f} \), which relates the liquid contraction to the graphite expansion. For a given temperature drop \( \Delta T \) from pouring to the end of liquid contraction, the apparent volumetric change can be written as:
$$
\frac{\Delta V_{liq}}{V_{liq}} = \alpha_{L} \Delta T – \beta_{G} \frac{w(\mathrm{C})}{100}
$$
where \( \alpha_{L} \) is the liquid thermal expansion coefficient, and \( \beta_{G} \) is the specific expansion coefficient of graphite precipitation. By keeping the carbon content high and the pouring temperature moderate, \( \frac{\Delta V_{liq}}{V_{liq}} \) remains small enough to be overcome by the large riser. The small-and-tall riser neck then seals the cavity before the positive expansion takes place. This principle was applied successfully in all later production runs.
Simulation and Production Verification
All process changes were first evaluated with a CAE filling and solidification simulation code. The initial gating and riser design predicted an insufficient feeding capacity, a sunken riser, and large shrinkage pores on the cope surface. After increasing the amount of molding sand around the casting and enlarging the risers, the simulation showed a sound part. However, production of the first few castings still exhibited shrinkages under some risers, with no clear pattern. We therefore suspected that the neck geometry, not the riser volume, was the limiting factor. By modifying the neck to the small-and-tall configuration, the simulation predicted a perfect feeding profile. The subsequent production trial confirmed that every riser-neck region on the rotary table was sound after machining.
Table 4 lists the simulation and production results for the different versions of the process.
| Process version | CAE prediction | Production result |
|---|---|---|
| Initial design (large neck, short neck height) | Riser neck feeding back; shrinkage on cope | Random shrinkage under several risers |
| Enlarged riser + more sand + thicker chills | No visible shrinkage in simulation | Still occasional porosity under risers; T-slot defects |
| Large riser + ≤30 mm diameter, ≥50 mm tall neck | Solid sound; no feeding back | Defect-free after full machining |
The T-slot defects that were originally observed at the bottom of the slots were traced to hot spots generated between the inner cross-shaped ribs and the work-table face. The old chill thickness of 60 mm was not enough to suppress the local thermal center. After the chill thickness was increased to 100 mm on the work-table face and to 110 mm on the guide-way face, the T-slot shrinkages disappeared. The combination of thicker chills and the optimized riser neck finally produced consistently sound ductile iron castings.
Mechanical Properties and Quality Control
The final production castings were evaluated both by non-destructive inspection and by destructive testing on separately cast test bars. The tensile strength of the test bars was between 682 and 820 MPa, the Brinell hardness was between 255 and 275 HB, and the elongation was between 4.62% and 5.50%. These values satisfy the requirements of QT600-3 ductile iron castings. The hardness measured on the casting body was 180–220 HB, which is in the acceptable range for the working surface of a machine-tool rotary table. Table 5 summarizes the mechanical properties achieved.
| Property | Test bar result | Specification |
|---|---|---|
| Tensile strength / MPa | 682 – 820 | ≥ 600 |
| Elongation / % | 4.62 – 5.50 | ≥ 3 |
| Hardness / HB | 255 – 275 (cast-on test bar) | 220 – 270 |
| Body hardness / HB | 180 – 220 | according to drawing |
The microstructure of the ductile iron castings was inspected on the test bars. The typical microstructure consisted of fine pearlite with a nodular graphite structure. The nodularity was above 85%, and the carbide content was negligible. With the optimized riser and chill configuration, final machining of the guide-way face, the work-table face and the T-slots did not reveal any shrinkage cavities, porosity, or non-metallic inclusions. The elimination of defects significantly reduced the time and cost previously spent on weld repair, and also guaranteed the delivery time for the foundry.
Conclusions for Sound Ductile Iron Castings
Through the development described above, we established a robust manufacturing route for large machine-tool rotary table castings in QT600-3 ductile iron castings. The following conclusions can be drawn from our experience.
1. The gating system for large ductile iron castings must be fully closed, bottom-dispersed, and symmetrical to ensure laminar filling and effective slag removal. A two-point pouring system with two runner segments and ten 50-mm ceramic tubes worked successfully for the 85-ton rotary table.
2. The riser design is the decisive factor for the soundness of ductile iron castings. A large riser with a small diameter and a sufficiently tall neck is preferable. For our casting, the riser diameter was set at 1.3 times the hot-spot circle diameter, the riser neck diameter was limited to 30 mm, and the neck height was at least 50 mm. This geometry makes the neck freeze off at the correct moment, using the graphite expansion for self-feeding and preventing liquid iron from being pushed back into the riser.
3. Chills are necessary to suppress hot spots, especially at the intersections between ribs and plates. The chill thickness on the cope side should be sufficient, but because cope chills lose contact during solidification, they cannot replace the proper riser neck design. On the work-table face a chill thickness of 100 mm was required, and on the guide-way face 110 mm was required.
4. Melting practice for ductile iron castings must be tightly controlled. The chemical composition, especially the carbon equivalent and residual magnesium, must be maintained within a narrow window. The four-stage inoculation procedure ensures adequate nodule count and prevents inoculation fading over the long pouring time.
5. The pouring temperature for such heavy ductile iron castings should be maintained at 1 330–1 360 °C. This range gives good fluidity without sacrificing the beneficial effects of graphitic expansion.
6. Modern CAE simulation is a useful tool for designing the gating and risering system, but it cannot replace production verification when subtle neck-freezing phenomena are involved. The combination of simulation and iterative full-scale trials allowed us to develop a defect-free process for ductile iron castings of the rotary-table type.
In conclusion, the prevention and elimination of defects in machine-tool rotary table castings made of ductile iron castings depends mainly on three principles: a properly designed closed gating system that supplies clean liquid iron, a large-diameter riser with a small tall neck that isolates the riser from the casting at the right time, and a rigorous melting and inoculation schedule that stabilizes the graphite expansion and mechanical properties. Following these principles, we now produce large rotary table castings in ductile iron without internal defects, fully meeting the stringent requirements of the high-precision heavy-duty machine tool industry.