Lost Foam Casting of Machine-Tool Beds: Pattern, Gating and Process Design

A machine-tool bed carries the guideways, the headstock and the tailstock, so its casting has to combine high static stiffness, good vibration damping and enough dimensional stability to hold guideway geometry after machining. Those requirements collide with the geometry: beds are long, thin-walled, ribbed and often heavier than the rest of the machine combined. Shrinkage porosity and distortion are the two failure modes that matter, and both are decided long before the first ladle is poured — in how the mould is built, how the metal enters it, and how the sections solidify in sequence.

Lost foam casting addresses the mould side of that problem by replacing the bonded sand mould and cores with a polymer pattern that is vaporised by the incoming metal. The pattern is a faithful copy of the part, embedded in loose dry sand, so the cavity that forms is defined by the foam rather than by pattern withdrawal or core assembly. For a ribbed bed with cored oil passages and varying wall thickness, this removes a whole class of tolerance stack-ups.

Lost foam casting mould: polymer pattern embedded in loose sand inside a flask

What the evaporative pattern changes in the mould

Four consequences follow directly from the absence of a parting plane and cores, and each maps onto a bed-specific benefit.

  • No core assembly, no core-related mismatch. In a conventional mould, wall thickness and bore positions depend on how well cores are located and how much they shift during ramming. With a one-piece foam pattern there is no core to shift, so the dimensional error contributed by core-to-mould assembly disappears and the as-cast surface is smoother.
  • No draft and no parting-line flash. Ribs, bosses and internal passages can be placed where the stiffness analysis wants them rather than where a pattern can be withdrawn. Complex, thin-walled geometry becomes a pattern-making problem rather than a moulding problem.
  • Uniform wall thickness. Loose sand flows into every recess without ramming pressure, so a 15 mm wall can be held over a large face. Uniform walls mean more even solidification and fewer hot spots.
  • Fewer process steps and less labour. No core boxes, no core baking, no shakeout of a bonded mould; the sand is dry and reusable, which lowers tooling investment and cycle time.

The trade-off is that the pattern is consumed by every casting, and its decomposition products must escape through the refractory coating and the sand bed. Foam density therefore becomes a process variable rather than a purchasing detail, and the coating and the sand’s permeability become part of the filling design.

Casting envelope and process sequence

The reported production case is a large bed-type component with a maximum pattern envelope of 2400 mm × 1280 mm × 700 mm, a 15 mm wall, a poured mass of 1900 kg and a height of 630 mm. Filling is completed in about 55 s, at a pouring temperature of 1350 °C. The metal is delivered through a branched (multi-ingate) system rather than a single choke, so that flow, pressure and temperature fields stay even across the length of the casting.

Parameter Value
Casting mass (poured) 1900 kg
Nominal wall thickness 15 mm
Casting height 630 mm
Maximum pattern envelope 2400 × 1280 × 700 mm
Pattern material / density Expanded polystyrene, 20 kg/m³
Static head Hp 600 mm
Mean head at ingate hp 142 mm (14.2 cm)
Pouring time 55 s
Pouring temperature 1350 °C
Phosphorus limit ≤ 0.08 %
Dimensional distortion after casting about 5 mm

Gating: from static head to sprue area

For a system discharging through several orifices of different resistance, the driving head at the ingate is not the full static head of the sprue. It is reduced by the losses in the sprue and runner, expressed through the ratios of effective cross-sections:

$$h_p=\frac{K_2^{2}\,H_p}{1+K_1^{2}+K_2^{2}}$$

where Hp is the static head above the ingate, hp the mean head actually available at the ingate, and K1, K2 the ratios of effective cross-sectional area between successive elements of the system. With a static head of 600 mm, the calculation returns an ingate head of 142 mm — roughly a quarter of the static head, which is why a large-head, small-orifice assumption would badly overestimate the fill rate.

The choke area then follows from the mass to be poured, the density of liquid iron, the pouring time and the discharge coefficient:

$$S=\frac{G}{\rho\,\tau\,\mu\sqrt{2g\,h_p}}$$

with G = 1900 kg, τ = 55 s, μ = 0.5 and hp = 14.2 cm. Resistance coefficients of 0.4 for the ingate and 0.5 for the runner and sprue were used in the layout, and the sections were distributed in the ratio Ssprue : Srunner : Singate = 1 : 2 : 2, giving 36.5 cm², 73 cm² and 73 cm² respectively.

Element Resistance coefficient Area ratio Calculated area As built Number
Sprue 0.5 1 36.5 cm² 60 × 60 mm (36 cm²) 1
Runner 0.5 2 73 cm² 60 × 60 mm (36 cm² each) 2
Ingate 0.4 2 73 cm² 10 × 30 mm (3 cm² each) 25

The built sections reproduce the design areas closely: 25 ingates of 3 cm² give 75 cm² against the calculated 73 cm², and two 60 mm square runners give 72 cm². The hydraulic consistency can be checked independently. A poured mass of 1900 kg in 55 s corresponds to about 34.5 kg/s, or roughly 4.9 × 10⁻³ m³/s of liquid iron at 7000 kg/m³. Passing that flow through 73 cm² of ingate area implies a mean ingate velocity near 0.67 m/s, which matches the Bernoulli estimate μ√(2g hp) = 0.4 × √(2 × 9.81 × 0.142) ≈ 0.67 m/s. The low velocity is deliberate: it keeps the metal front from breaking up the foam too violently and reduces the risk of entraining decomposition gas and oxide film into the advancing front.

Distributing the ingates along the bed, rather than feeding from one end, shortens the distance the front must travel before it reaches the far flange and keeps the thermal field more uniform over the 630 mm height. That uniformity is what allows the sections to solidify in a controlled order afterwards.

Temperature, chemistry and feeding

Pouring temperature is the most direct lever on shrinkage. Too high a superheat enlarges the liquid contraction that must later be fed and makes shrinkage cavities harder to eliminate; too low a temperature risks cold shuts in 15 mm sections and incomplete vaporisation of the foam. The reported practice is 1350 °C, a moderate superheat for a 1900 kg iron casting.

Phosphorus is held to 0.08 % maximum. Above that level the casting skin becomes brittle and shrinkage porosity tends to enlarge, both of which are unacceptable in a bed that will be machined and scraped on its guideway faces. Carbon is controlled at the same time, and the evaporative pattern is the reason: the polystyrene decomposing against the metal front releases carbon-bearing species that can be absorbed into the surface layer, so the pattern density, the coating and the melt chemistry together determine whether the surface carbon stays inside specification.

Feeding depends on gate and chill placement. If chills and gates are not positioned for directional solidification, the metal cannot freeze progressively from the far end back towards the feeders, and the ingates — whose number and connection geometry determine how much feed metal they can pass — will not compensate the sections they serve. Gate and chill layout is therefore treated as part of the solidification design, not as a finishing detail.

Pattern fabrication and dimensional control

Patterns are built by bonding polystyrene foam slab and machining the assembly under numerical control. A foam density of 20 kg/m³ is the practical optimum for this class of work: lighter foam decomposes into a larger volume of gas for a given section, heavier foam machines less cleanly and leaves more residue. Tool shank diameter has to be kept modest, since an oversized cutter tears the foam surface and the tear is reproduced faithfully in the casting.

Length is the logistical constraint. Bed patterns run to about 8 m, and the foam is fragile enough that patterns are shipped in sections and assembled at the foundry rather than moved in one piece. After pouring and cleaning, castings are checked for chemical and physical properties as well as for the control dimensions, and distortion is held to roughly 5 mm — a figure that determines the machining allowance left on the guideway faces. Because the foam pattern eliminates draft and parting-line mismatch, that 5 mm represents most of the dimensional error budget, and assembly of the machined bed is correspondingly simplified.

Production economics

The economics of the route come from four places: the pattern replaces core boxes and a bonded mould, which cuts tooling investment and lead time; machining allowance and distortion are small enough to shorten finishing; design freedom allows ribs and passages that would be expensive or impossible to core conventionally; and the dry sand is recovered. Reported output of beds by this method is currently around 300 t and rising, which is the usual sign that the per-piece pattern cost has been amortised over a stable order book rather than absorbed by one-off work.

Where the route stays sensitive

Lost foam is not unconditionally superior to conventional moulding; it shifts the difficulty from moulding to pattern and gas management. Its accuracy advantage is real only while the foam stays dimensionally stable during storage and coating drying, and its surface quality advantage is real only while the coating permeability matches the quantity of decomposition gas the poured section must vent. Thin sections poured at low superheat and thick sections poured with excessive superheat both defeat the process — the first by incomplete vaporisation, the second by enlarging the shrinkage volume that the feeders must handle. The 15 mm wall, 1350 °C and 55 s reported here describe one operating point that satisfies all of these constraints simultaneously, and that balance, rather than any single parameter, is what makes the process repeatable at production scale.

Scroll to Top