Metal Casting: The Complete Engineering Guide to Sand Casting, Investment Casting, Gating Design, Risering, and Solidification Physics

A dense equation-driven engineering deep-dive into expendable-mold metal casting — solidification physics with Chvorinov's rule and the modulus method, the constitutional supercooling criterion, riser design via feed-volume balance and the Niyama criterion, gating system design with Bernoulli, tapered sprues, and choked-area worked examples, green sand and chemically bonded mold systems, the complete process catalog from green sand to investment casting with ISO 8062 tolerance classes, casting alloy grades (FG/SG iron, steel WCB/CF8M, LM6/LM25 aluminum, gunmetal), defect taxonomy with NDT, DFM rules, Indian foundry economics with indicative ₹/kg rates, and a fully worked 8 kg pump-bracket example.

·

Metal Casting: The Complete Engineering Guide to Sand Casting, Investment Casting, Gating Design, Risering, and Solidification Physics

Casting is the oldest manufacturing process and still the largest by tonnage. Humanity cast bronze 5,000 years before it forged steel, and the physics haven't changed: melt the metal, pour it into a cavity, control how it freezes. The 2023 World Casting Census put global production at 113.14 million metric tonnes. China leads, the USA is third — and India is second, producing 15.86 million tonnes in 2024–25 across roughly 5,000 foundries (≈90% MSMEs) worth about US$22 billion. From a ₹55/kg gray-iron manhole cover in Howrah to a single-crystal turbine blade that costs more than gold by weight, casting is one process family.

This guide covers the expendable-mold side of that family — sand casting, shell molding, investment casting, lost foam — and the engineering that makes castings sound: solidification physics, gating, risering, alloys, defects, DFM, and Indian foundry economics, ending with a fully worked example. The permanent-mold, high-pressure route is covered in our Die Casting guide; this is its complementary twin.

1. When Casting Wins

Liquid metal takes the shape of whatever cavity you can build. That single property beats every competing process on three axes:

Casting loses when: tolerances must beat ISO 8062 class DCTG 6 (investment casting is the floor at DCTG 4–6), walls fall below ~0.5 mm, or specific strength is the binding constraint (forgings win there). And the hidden cost is yield — gating and risers return to the melt, so "poured weight" can be 150–200% of "shipped weight."

2. Solidification Physics: The Engine of Every Defect

Everything that goes wrong in a casting — porosity, hot tears, misruns, coarse grain — is written in the solidification sequence. Understand it and the defect chapter reads like a map.

2.1 The Cooling Curve

Pouring at T_{pour} = T_{liquidus} + \Delta T_{superheat} (typically 50–150°C of superheat), the metal passes through three regimes:

  1. Superheat removal — convective and conductive loss to the mold, often 5–20% of the total solidification time for heavy sections.
  2. Nucleation and growth — heterogeneous nucleation on mold walls and inoculants at undercooling \Delta T, then dendritic growth that releases latent heat (L \approx 247 kJ/kg for iron, 398 kJ/kg for aluminium, 205 kJ/kg for copper). The recalescence bump on the cooling curve is the latent heat outpacing mold extraction.
  3. Terminal solidification — the last 10% of liquid freezes in the interdendritic spaces, and this is where shrinkage porosity lives or dies.

2.2 Chvorinov's Rule and the Modulus

Nikolai Chvorinov's 1940 rule remains the workhorse of riser design. Solidification time scales with the square of the volume-to-area ratio:

V/A is the modulus M = V/A, in cm. B is a mold-constant: ~0.9–2.0 min/cm² for steel in green sand, ~1.0–1.6 for grey iron, higher for insulating molds. The squared relationship is brutal: double the section thickness and solidification takes four times longer.

Worked example. A 10 mm cube: M = 1/0.6 = 0.167 cm. A 100×100×10 mm plate: V = 100 cm³, A = 240 cm², M = 0.417 cm. The plate solidifies (0.417/0.167)^2 \approx 6.2\times slower. With B = 1.2 min/cm², the plate takes t_s = 1.2 \times 0.417^2 \approx 0.21 min ≈ 12.5 s — a realistic green-sand number. This is why ribs (high A/V) chill fast and thick bosses (low A/V) stay hot: the modulus is the solidification-time map.

2.3 Constitutional Supercooling: Why Grains Change Shape

A planar solidification front is unstable when the liquid ahead of it is supercooled by rejected solute. The stability criterion:

where G is the thermal gradient (K/m), R the growth rate (m/s), \Delta T_0 the alloy freezing range, and D_L \approx 10^{-9} m²/s the solute diffusivity. Low G/R (the mold wall, early freezing) → constitutional supercooling → equiaxed grains. High G/R (deep in the casting, fed by a riser) → columnar grains growing against the heat flux. Directional solidification for turbine blades is simply this equation engineered: a Bridgman furnace holds a steep unidirectional gradient so the interface stays planar-ish and the entire blade solidifies as one columnar crystal — or, with a helical "pigtail" grain selector, a single crystal.

2.4 Shrinkage: The Numbers That Drive Risering

Metals shrink in three stages, and only two of them need feeding:

Stage · Mechanism · Magnitude

Liquid contraction · Cooling from pour to liquidus · ~1–2% per 100°C of superheat

Solidification shrinkage · Liquid→solid volume change · See table below

Solid contraction · Thermal contraction of the solid · Pattern allowance, not riser-fed

Solidification shrinkage (vol %) — the quantity that must be fed by risers:

Alloy · Shrinkage

Grey cast iron · −1.6 to +1.6 (graphite expansion offsets contraction)

Ductile iron · −2.1 to +2.3

White iron · 4.0–5.5

Carbon steel · 2.5–4.0

Al–Si alloys · 3.5–6.0

Al–Cu alloys · 6.0–8.5

Copper alloys · 4.5–7.7

Magnesium · 3.5–4.2

Grey iron's near-zero shrinkage is the entire reason cast iron risers are small: the graphite precipitation at the eutectic expands ~3% and pushes metal back into the mushy zone. Steel foundries, with 3–4% real shrinkage, spend 30–60% of their poured metal on risers.

Pattern allowance (solid contraction, the patternmaker's rule of thumb): grey iron 10.4 mm/m, steel 20.8 mm/m, aluminium 13 mm/m, brass 15.5 mm/m. A patternmaker's rule is a scale with these corrections baked in.

2.5 Gas: Hydrogen and Sieverts' Law

The biggest dissolved-gas problem in foundries is hydrogen in aluminium. Solubility follows Sieverts' law:

and it collapses at the freezing point: molten aluminium dissolves ~0.69 cm³ H₂ per 100 g, solid aluminium only ~0.036 cm³/100 g (Ransley's classic data). The 20× solubility drop rejects hydrogen at the interface, nucleating spherical gas pores that feed on shrinkage voids. Rotary degassing with argon/nitrogen for 10–20 minutes gets hydrogen below ~0.1–0.15 cm³/100 g, which is the practical threshold for pressure-tight Al castings. Steel foundries fight nitrogen and hydrogen with dry refractories and ladle practice instead.

3. Feeding the Casting: Riser Design

A riser is a reservoir of liquid metal that solidifies after the casting and feeds its shrinkage. Three methods size it, and they should agree within reason.

3.1 Feed-Volume Balance

The riser must supply the casting's shrinkage before it freezes itself:

where S is the total volumetric shrinkage (liquid + solidification, typically 0.05–0.07 for steel poured at 50–100°C superheat) and \eta the riser efficiency — the fraction of riser volume that can actually be fed before the riser itself loses feedability:

Riser type · Efficiency \eta

Open top, uninsulated · 0.14–0.16

Blind riser · 0.30–0.35

Blind + exothermic sleeve + atmospheric vent · up to 0.50

Worked example. A 30 kg steel casting (V_c \approx 3.85 L at 7.8 kg/L), S = 0.06, open riser \eta = 0.15: V_r = 0.06/(0.15-0.06) \times 3.85 = 2.57 L ≈ 20 kg of steel in the riser. Riser metal costs real money, which is why sleeved blind risers pay for themselves immediately: at \eta = 0.45, V_r = 0.06/(0.39) \times 3.85 = 0.59 L — a 4.3 kg riser instead of 20 kg.

3.2 The Modulus Method

Chvorinov inverted: the riser must have a larger modulus than the casting. The classic factors:

with f \approx 1.2 for steel and white iron, 1.1 for aluminium and grey iron (grey iron's graphite expansion makes feeding undemanding), 1.15–1.2 for ductile iron. For a cylinder with H = D, the modulus works out to M = D/6, so a 25 mm plate casting (M = 1.25 cm) in steel needs a riser with M \ge 1.5 cm → D \ge 9 cm — a 90 mm diameter riser, in agreement with the feed-balance method's scale. Insulating sleeves reduce the effective modulus requirement by retarding riser skin growth, which is the mathematical justification for their economics above.

3.3 The Niyama Criterion: Where Microporosity Will Form

Chvorinov and modulus rules size risers; they don't predict where interdendritic microporosity appears. Niyama's criterion does, using the ratio of thermal gradient to cooling-rate square root at the end of freezing:

with G in °C/cm and \dot{T} in °C/min. Steel castings are sound where N_y \gtrsim 1.0 (°C·min¹ᐟ²/cm); below that threshold, interdendritic feeding fails and microporosity nucleates. Every solidification simulation package (MAGMASOFT, ProCAST, SOLIDCast) ships this criterion — the output plot of N_y over the casting geometry is the single most useful defect map in the foundry. Simulation pays for itself on any casting whose pattern costs more than ~₹50,000.

3.4 Feeding Distance and Chills

A riser doesn't feed indefinitely — a solid "skin" chokes the feed path. Steel foundry practice rules of thumb: a top riser feeds ≈ 4.5t in a plate of thickness t, and ≈ 30\sqrt{t} mm in a bar (e.g., 150 mm for a 25 mm bar). External chills (steel plates set in the mold) steepen the gradient and extend feeding distance by roughly 50%; internal chills (steel pins consumed by the melt) do the same from inside. Exothermic sleeves and hot toppings buy another 20–40% of feed distance on the riser side.

4. Gating System Design

The gating system (pouring basin → sprue → runner → ingates) has one job: deliver metal to the cavity fast, full, and quiet. Every rule below serves one of those three words.

4.1 Bernoulli and the Tapered Sprue

Metal falling in a sprue accelerates under gravity:

A straight (untapered) sprue is a mistake: by continuity A_1v_1 = A_2v_2, the accelerating stream must shrink in area, so the metal pulls away from the wall, aspirates air, and oxidizes. The taper that keeps the sprue full:

Worked example. Sprue with basin 50 mm below the metal surface, exit 300 mm below: A_{exit}/A_{top} = \sqrt{50/300} = 0.41. A 40 mm diameter top (1,257 mm²) tapers to a ~26 mm exit (513 mm²). Check any production pattern: sprues are visibly tapered.

4.2 Turbulence: Campbell's Critical Velocity

Oxide films (bifilms) entrained during filling become crack-like defects that no heat treatment removes. John Campbell's work established that for aluminium, surface turbulence begins around a melt velocity of 0.5 m/s — above it, the surface folds over and traps films; the Reynolds-number equivalent is about Re \approx 20{,}000 in a runner. Steel and iron tolerate higher velocities (their oxides dissolve or float more readily), but the design principle survives: bottom-gated, quiet fill, metal front rising smoothly. Ceramic foam filters (10/20/30 ppi) at the sprue base or in the runner both slow the stream and trap inclusions — cheap insurance on every Al casting.

4.3 Gating Ratios: Pressurized vs Unpressurized

The sprue:runner:ingate area ratio decides where the choke sits and whether the system runs full:

System · Typical ratios · Choke location · Character

Pressurized · 1:2:1, 1:4:2, 3:8:3 · Ingates · System runs full, less aspiration; higher velocities at gates — good for iron/steel

Unpressurized · 1:2:2, 1:4:4 · Sprue base · Quieter fill, lower gate velocities — preferred for aluminium

Pressurized systems keep metal against the walls (the "full" requirement); unpressurized systems trade a partly empty runner for gentle cavity entry (the "quiet" requirement). Most Al foundries run unpressurized with a filter; most iron foundries run pressurized.

4.4 Filling Time

Fill too slowly and the metal freezes mid-fill (misruns); too fast and turbulence and mold erosion win. The classic empirical formulae:

Steel (CIS/Russian foundry literature), W in kg, t in seconds:

Grey iron: t = K\sqrt{W} with K \approx 1.4 for thin (≤6 mm) sections rising to ~2.3 for 20–25 mm sections. A 100 kg steel casting: t = (2.4335 - 0.7906)\times 10 \approx 16.4 s.

4.5 Sizing the Choke

Once the filling time is fixed, the choke area follows from Bernoulli with a discharge coefficient:

with W the poured weight (kg), \rho the liquid density (~7,000 kg/m³ steel, ~6,900 cast iron, ~2,400 aluminium), t the filling time (s), C \approx 0.7–0.9 (0.8 typical), and H the effective head (m).

Worked example. A 25 kg steel casting: t = (2.4335 - 0.3953\log_{10}25)\sqrt{25} = (2.4335 - 0.5526)\times 5 = 9.4 s. With H = 0.2 m and C = 0.8:

A 17.5 mm diameter choke — entirely plausible for that pour. This chain (filling time → choke area → ratio → taper) is the complete gating calculation, and it is rarely taught as one sequence; it should be.

5. Mold Materials and Sand Systems

The mold is a heat exchanger with a shape. Its two design variables are permeability (let the gas out) and refractoriness (don't fuse to the metal).

5.1 Green Sand (Clay-Bonded)

Still 60%+ of global tonnage. Typical system for iron: silica sand at AFS grain fineness number (GFN) 55–70, 7–9% active bentonite clay, 3–5% moisture, 4–8% coal dust (for the reducing atmosphere that gives grey iron its clean peel), compactability 35–45%, green compressive strength 0.8–1.6 kg/cm², permeability 100–200 AFS, mold hardness 80–95 on the B scale. The numbers matter because they're all trade-offs: more clay = strength but less permeability; more moisture = better flowability but more gas.

Silica's weakness is the quartz inversion at 573°C, a 0.45% linear expansion that causes veining, rat-tails, and scabs on hot steel and heavy iron work. Fixes: switching to chromite or zircon sand (higher thermal conductivity = faster chilling, near-zero expansion, 4–10× the cost), olivine, or core/mold coatings (zircon/alumina washes).

5.2 Chemically Bonded Systems

System · Binder (typical) · Notes

Furan no-bake · 1.0–1.5% furan resin + acid catalyst · Best dimensional accuracy of sand systems; sulfur pick-up risk on steel at the mold face

Phenolic-urethane (PUNB) · ~1.5% three-part · Fast strip, good for aluminium and iron; amine odor control required

Sodium silicate/CO₂ · 3.5–5% silicate, gassed with CO₂ · Oldest, cheapest; poor shakeout, low sand reclamation

Shell (Croning) · Resin-coated sand, 2.5–3.5% resin, cured on a heated pattern · Thin shells, excellent surface, high-rate production

No-bake and shell molds hold dimension better than green sand (the hardened binder doesn't deform under pour pressure), which is why they own the DCTG 8–10 tolerance band. Shell molding's thin cured shells also shake out to nearly free-flowing sand — one reason it dominates small-parts automotive production.

Cores are the negative space inside castings — made from the same systems but mixed for strength and collapsibility. Core prints (molded recesses) locate them; core wash seals the surface. Hot-strength/collapsibility trade-off: a core must survive the pour but crush rather than hot-tear the casting during contraction.

6. Process Catalog: Capability and Cost

Process · ISO 8062 DCTG · Typical Ra (µm) · Min wall (mm) · Typical max weight · Tooling · Batch sweet spot

Green sand, hand molded · 11–14 · 12.5–25 · 5–6 · 10 t+ · Wood pattern (₹5k–50k) · 1–100

Green sand, machine molded · 9–11 · 6.3–12.5 · 4–5 · ~500 kg · Metal pattern (₹50k–5L) · 100–100k

Chemically bonded (no-bake) · 9–11 · 6.3–12.5 · 4–5 · 100 t+ · Wood/plastic pattern · 1–500

Shell molding · 8–10 · 3.2–6.3 · 2–3 · ~100 kg · Metal pattern + gas-fired machine · 1k–100k

Investment casting · 4–6 · 1.6–3.2 · 0.5–1.5 · ~100 kg (typ <10) · Wax injection die (₹50k–10L+) · 50–10k

Lost foam · 8–10 · 6.3–12.5 · 3–4 · ~500 kg · EPS tooling · 1k–50k

Centrifugal · 9–12 · 6.3–12.5 · 5–8 · Cylindrical parts · — · 10–10k

Die casting (permanent mold) · 5–8 · 1.6–3.2 · 0.8–1.5 · ~50 kg · H13 steel die (₹5L–50L) · 5k–1M

The permanent-mold row belongs to our die casting guide; it's here so you can see the tolerance ladder in one table.

6.1 Investment Casting: The Precision Ceiling

The lost-wax pipeline, step by step, because it is the most common high-value request on any manufacturing platform:

  1. Wax injection — unfilled pattern wax (0.8–1.2% linear shrinkage, compensated in the die) or filled wax (0.3–0.5%) injected at 55–70°C into an aluminium die.
  2. Assembly — patterns are welded onto a central sprue "tree" with wax gates, 10–200 parts per tree.
  3. Shell building — 6–9 alternating coats of colloidal-silica slurry (zircon or fused-silica flour) and refractory stucco, each dried 2–24 h. Total shell thickness 6–10 mm.
  4. Dewaxing — steam autoclave at 150–175°C, 5–6 bar: the wax flashes out in seconds before the shell can crack.
  5. Firing — shell burned out at 1,000–1,100°C; the mold is then poured hot for thin sections.
  6. Pouring — gravity, or vacuum-assisted for superalloys and thin aerofoil sections.
  7. Knockout and cutoff — shell removed by vibration/water-blast; parts cut from the tree, ground, heat-treated, inspected.

Investment casting holds ±0.5% of dimension (DCTG 4–6, ≈ ±0.13 mm at 25 mm), casts 0.5 mm walls in steel, and needs zero or near-zero draft. The cost: wax dies are expensive, and yield is low — a tree is 60–70% gating by weight. It is the default process for turbine blades, medical implants (Co-Cr and Ti), valve bodies, firearms parts, and aerospace brackets. For single-crystal turbine blades the same shell is poured inside a vacuum Bridgman furnace, withdrawn at 100–300 mm/h through a sharp thermal gradient so the solidification front advances as one crystal through the helical grain selector — Section 2.3's criterion made into hardware.

6.2 Lost Foam

The EPS (expanded polystyrene) pattern vaporizes on contact with the melt, leaving a cavity that can include undercuts and internal passages no pattern-pull could produce — and no parting line, so no flash. Best for aluminium and grey iron in the 1k–50k batch range; the cost of the EPS tooling and the need for a dedicated foundry line make it a process you choose, not one you stumble into.

7. Casting Alloys and Grades

7.1 Cast Irons

Grey iron (IS 210) grades are named by minimum tensile strength: FG 150/200/220/260/300/350 MPa. Carbon equivalent predicts structure and fluidity:

Eutectic composition is CE 4.3; production irons sit at 4.2–4.4 for fluidity. FG 200 runs 170–220 HB, machines beautifully, damps vibration — the pump-and-valve workhorse of every Indian foundry cluster.

Ductile (SG) iron (IS 1865) adds a 0.03–0.05% residual magnesium treatment that converts flake graphite to spheroids: SG 400/15 (400 MPa UTS, 15% elongation), SG 500/7, SG 600/3, SG 700/2. SG 500/7 at ₹85–120/kg replaces steel in brackets, hubs, and housings where weldability isn't required — roughly double grey iron's strength at a 30–50% price premium.

7.2 Steel Castings

Carbon and low-alloy steels are specified by ASTM A216 grades: WCB (weldable cast B: UTS 485–655 MPa, YS 250 MPa min, 22% elongation) is the generic engineering grade; WCC the higher-strength variant; LCB/LCC for low-temperature (≤−46°C) service. IS 1030 covers the same space in India. Stainless equivalents via ASTM A351: CF8 (304), CF8M (316), CF3M (316L), plus CA6NM — the 13Cr-4Ni martensitic grade used for hydraulic turbine runners. Steel castings are nearly always heat-treated (normalized, or quenched and tempered) — a not heat-treated steel casting is a defect by omission.

7.3 Aluminium

LM6 (AlSi12): the best-castability alloy — UTS 160–190 MPa, 5–8% elongation, used for thin-walled, leak-tight, marine and electrical work. LM25 (AlSi7Mg ≈ A356) is the heat-treatable structural grade: T6 temper reaches ~280 MPa UTS, ~207 MPa YS at 6% elongation — the material of pump housings, aerospace brackets, and wheels. Grain refinement with Al–Ti–B master alloy and modification of the eutectic with strontium (0.02–0.04%) are standard practice for pressure-tight LM6/LM25.

7.4 Copper Alloys

Gunmetal LG2 (85Cu/5Sn/5Zn/5Pb): UTS 250–310 MPa, the classic valve, pump, and marine fitting alloy — corrosion-resistant and pressure-tight. Phosphor bronzes for wear rings and bushes. Pouring at 1,100–1,200°C with oxidizing fluxes for gas control; zinc loss control matters more than most foundries admit.

8. Casting Defects and Quality Control

Defect taxonomy, with the physics from Sections 2–4 attached:

Defect · Root cause · Prevention

Shrinkage cavity (macro) · Riser too small/cold, feed path choked · Modulus method, sleeved risers, chills

Microporosity · Interdendritic feeding failure · Niyama criterion, simulation, steeper gradients

Gas porosity · Dissolved H₂ (Al), mold gas, cores · Degassing, permeability, core venting

Hot tear · Tensile strain in mushy zone (fraction solid 0.9–1.0) · Wide-freezing-range alloys (Al–Cu), collapsible cores, radii, chills

Cold shut / misrun · Metal froze before streams merged; poor fluidity · Raise superheat, shorten fill time, avoid <3–4 mm sand-cast walls

Inclusions (slag/sand) · Turbulent gating, eroded mold · Filters, pressurized gating, mold washes

Metal penetration / burn-on · Low GFN sand, high ferrostatic pressure · Finer sand, coatings, lower head

Veining, scabs, rat-tails · Silica 573°C inversion expansion · Chromite/zircon sand, coatings, additives

Inspection ladder, cheapest first: visual + dimensional (CMM for DCTG ≤ 8), dye penetrant (surface), magnetic particle (ferrous surface/subsurface), ultrasonic (internal soundness, thickness), radiography (the porosity arbiter — ASTM E446/E505 digital classes), leak testing for pressure-containing parts, and per-heat tensile coupons + hardness for certified grades. Foundry scrap typically runs 3–8%; a foundry that ships castings without an internal RT/UT capability for critical parts is gambling with your project.

9. DFM Rules for Castings

  1. Uniform sections. Keep wall thickness within ±10–15%; transition changes at ≤2:1 ratios with generous radii. Every isolated heavy boss is a hot spot demanding a riser or a chill.
  2. Fillets. Internal corners ≥ T/3 (T = wall thickness); sharp re-entrant corners are where hot tears nucleate.
  3. Draft. 1–3° on sand castings, 0.5–1° on investment castings, more on deep pockets.
  4. Machining allowance. 2–4 mm per machined surface for small/medium sand castings, 6–10 mm on metre-scale work; 0.5–1 mm for investment castings. State it in the drawing — the foundry cannot guess.
  5. Cores. Design cores ≥ 8–10 mm thick, well-printed and vented; avoid long unsupported cantilevers. Every core adds cost — consolidate hollow features.
  6. Parting line. Simplest possible; lettering and bosses on one side of the line; flash tolerance explicit.
  7. Ribs at 0.6–0.8T avoid mass; the modulus (not the CAD dimension) decides soundness.
  8. Consolidation economics. A 12-piece welded assembly becomes a single casting with zero weld distortion, fewer fixtures, and one inspection — the crossover is usually 10–20 units for steel, 100+ for aluminium, assuming the casting passes the tolerance gate (DCTG 9–11 for sand).

If your part currently comes out of bar stock at 90% chip removal, or out of a weldment with 20 weld seams, get a casting quote before optimizing anything else. The tolerance stackup and GD&T guides cover how to spec the result.

10. Indian Foundry Economics (2026)

India's 15.86 Mt of castings (2024–25) come from tight regional clusters: Coimbatore (pumps, motors, valves — arguably the densest foundry cluster in Asia), Rajkot (diesel engine components, automotive castings), Howrah/Kolkata (jobbing grey iron), Belgaum and Kolhapur (automotive and hydraulic), Agra (sanitary castings), Chennai/Hosur (automotive exports), Faridabad (auto components). Nearly 90% are MSMEs, and capability follows the cluster's specialty — a Coimbatore pump-casting foundry is not your Rajkot brake-drum supplier.

Cost structure of a typical grey-iron jobbing casting: metal 40–50%, energy 10–20%, labor 10–15%, sand and consumables 5–8%, overhead 10–15%. Melting energy sets the floor: cold-blast cupola consumes 100–120 kg coke per tonne; medium-frequency induction runs 550–650 kWh/t (≈ ₹5/kg at ₹8.5/kWh industrial tariff). Induction won the quality war — cupola now survives on price for non-critical iron.

Indicative ex-foundry rates, 2026, jobbing quantities (₹/kg, unmachined):

Material / process · ₹/kg

Grey iron, green sand · 55–85

SG iron · 85–120

Carbon steel, sand cast · 90–140

Stainless steel, sand cast · 300–450

Aluminium, sand cast · 250–400

Steel, investment cast (small parts) · 400–900

Machining typically adds 30–100% on top. Lead times: pattern 1–3 weeks, sample casting 2–4 weeks, production 3–6 weeks. Export-grade foundries hold IATF/ISO certifications and deliver with material test certificates; the price gap between a certified and an uncertified foundry can be 30–50% — and so is the scrap-risk gap.

11. Specifying Castings So the Quote Is Meaningful

A casting RFQ that omits tolerance class and testing is an invitation to a ₹45/kg quote that produces a part you can't use. Minimum spec block:

Then let the platform do the matching: on FabFlow, manufacturers quote against your drawing, material, and tolerance class with their verification status visible — the same way you'd qualify a foundry, but without 40 cold calls. Compare with what CNC machining, sheet metal, and metal AM would charge for the same part before committing to the mold.

12. Worked Example: A Grey-Iron Pump Bracket, End to End

Part: mounting bracket for a 15 kW pump set. Envelope 250×180 mm, walls 10 mm, bosses 20 mm, weight 8 kg. Material FG 220 (IS 210). Quantity 100 off. Green sand, machine molded.

1. Pattern. Shrinkage allowance 10.4 mm/m → pattern 0.25% oversize, plus 3 mm machining allowance on the machined faces, plus 1.5° draft. Aluminium pattern + core box: ₹35k–60k, amortized at ₹350–600/part over 100 pieces — trivial against the ₹560 casting cost.

2. Filling. Section 20 mm bosses → K \approx 1.9: t = 1.9\sqrt{8} \approx 5.4 s. Choke (bottom gate, H = 0.15 m, C = 0.8, \rho = 6{,}900):

A 14 mm ingate. Pressurized ratio 1:2:1 → sprue exit 156 mm² (14 mm Ø, tapered from a 22 mm Ø top), runner 312 mm², two ingates of 78 mm² each. Ceramic filter 20 ppi at the sprue base.

3. Riser. Grey iron: modulus factor 1.1. Bosses are the hot spots (M = 1.0 cm for a 20 mm plate section) → side blind riser with M_r = 1.1 cm → D = 6.6 cm cylinder, ~1.6 kg — 20% of casting weight, fed through a neck that freezes after the boss. Graphite expansion does the rest; no exothermics needed on grey iron.

4. Cost. 8 kg at ₹70/kg = ₹560 casting; machining ₹150–250; pattern share ₹400; total ≈ ₹1,110–1,210/part ex-works at 100 off — against ~₹1,800–2,400 for the same bracket machined from a billet, and a 12-seam weldment that would need straightening.

5. Verify. Hardness 170–220 HB on every batch, dimensional layout on the first article, DP on the first three pieces. No RT — it's a pump bracket, not a pressure vessel. That last sentence is quality engineering: test what the application needs.

13. The Decision, In One Paragraph

If your part has internal geometry, weighs more than a kilo, needs 50+ units, and can live with DCTG 8–11 tolerances — sand casting wins on cost almost every time. If it's small, precise, complex, and high-value — investment casting owns the DCTG 4–6 band at a premium. If it needs 0.8 mm walls and ±0.05 mm — you're in die casting or injection molding territory. The casting-specific engineering — modulus, Niyama, feed balance, gating ratios — is the difference between a foundry that guesses and one that engineers; and in a country producing 15.86 million tonnes a year, the guessing foundries are the majority. Spec like you know the difference, and they'll quote like they know you do.

Looking for a foundry? Post your job on FabFlow with your drawing, material grade, and ISO 8062 tolerance class — verified Indian manufacturers quote with their process capabilities on the table.

More FabFlow blog posts