Doc #93 — Foundry Work and Metal Casting

---

RECOMMENDED ACTIONS

First month (Phase 1)

1. Identify all operating foundries in NZ — include dedicated foundries, in-house casting operations within engineering firms, and any educational institutions with foundry equipment. Classify as critical-capability facilities (Doc #1).
2. Classify experienced foundry workers as critical-skills personnel. Prevent redeployment.
3. Inventory all foundry consumables nationally: induction furnace crucibles, refractory materials, moulding sand stocks, bentonite clay, ferrosilicon, and metal charge stocks.
4. Secure foundry engineering references — locate and preserve copies of standard foundry texts (Principles of Metal Casting by Heine, Loper & Rosenthal; ASM Metals Handbook Volume 15: Casting; American Foundry Society publications). These references are essential for training and quality control.

First 3 months (Phase 1–2)

5. Begin knowledge capture from experienced NZ foundry workers — document techniques, film operations, pair experienced founders with learners.
6. Assess NZ sand deposits for foundry suitability — collect samples from known deposits (Waikato, Bay of Plenty, Northland), conduct standard foundry sand tests (grain size, clay content, permeability, strength).
7. Assess NZ clay deposits for bentonite and moulding binder suitability.
8. Inventory national scrap stocks of cast iron, bronze/brass, and aluminium. Classify by alloy where possible.
9. Begin design work for a prototype NZ-built cupola furnace — target a small unit (600 mm internal diameter) for trial operation at an existing foundry. Use Glenbrook plate steel for the shell, NZ fireclay for the lining.

First year (Phase 2)

10. Build and commission the prototype cupola at an existing NZ foundry location. Conduct trial melts using NZ charcoal and locally available cast iron scrap. Evaluate iron quality (chemistry, fluidity, casting soundness).
11. Establish a standard pattern library for high-demand cast parts — pipe fittings (DN 25 through DN 150), common bearing housings, pump impellers for widely used pump types.
12. Begin foundry training programs at Te Pūkenga institutions with existing or restorable foundry facilities. Curriculum to cover: moulding, pattern-making, furnace operation, gating design, metal pouring, quality control.
13. Trial NZ-produced moulding sand and clay binder in production casting. Compare results to imported materials. Document any adjustments needed.
14. Establish at least one functioning foundry in each main NZ region — either by expanding existing operations or by equipping new facilities. Priority: Canterbury (if existing foundry can expand) and any region lacking current capability.

Phase 2–3 (Years 1–5)

15. Expand cast product range from basic pipe fittings and bearing shells to gear blanks, pump housings, valve bodies, and machine tool components.
16. Develop lost-wax casting capability for precision parts — start with plaster-mould variant using NZ gypsum, progress to ceramic shell as materials become available.
17. Begin casting machine tool components (beds, headstocks, tailstocks) in coordination with machine tool bootstrap program (Doc #91, Section 9).
18. Develop NZ-produced refractory crucibles for induction furnace use — extend the life of the induction furnace fleet by ensuring crucible supply.
19. If trade with Australia develops, prioritise import of ferrosilicon (for iron chemistry control), copper and tin (for bronze casting), and high-quality refractory materials.

Phase 4+ (Years 7+)

20. Engine block and cylinder head casting — develop the capability once foundry skill, pattern-making, and core-making are mature.
21. Steel casting capability — for highly stressed components, using induction furnace melting.
22. Machine tool castings at scale — supporting the machine tool bootstrap (Doc #91, Section 9) with cast iron beds and structural components produced to machining specifications.

---

ECONOMIC JUSTIFICATION

The question: Is establishing new foundry capacity worth the investment of labour and resources?

Person-years to establish a basic foundry: Approximately 2–4 person-years of effort to build the physical plant, fabricate the equipment, source materials, and train the initial workforce (assuming at least one experienced founder is available to lead the effort).³

The alternative — not having a foundry: Without local casting capability, every broken pump housing, cracked valve body, or failed bearing housing requires either:

• Scavenging a replacement from another piece of equipment (depleting the scavenging pool)
• Attempting to fabricate a substitute by welding and machining from solid stock (possible for some shapes, impractical or impossible for complex hollow parts)
• Accepting the loss of the equipment that needed the part

For a region with agricultural machinery, a water system, and any industrial equipment, the steady failure of cast components makes a local foundry economically worthwhile within 1–3 years of establishment, considering that each critical casting produced locally replaces a component that cannot be obtained any other way.⁴

Breakeven: If a regional foundry produces 30–100 functional castings per year (a modest-to-moderate output for a small operation with one melting furnace), and each casting keeps a piece of essential equipment in service that would otherwise be lost, the foundry pays for itself within 1–3 years. The value of each casting is measured not in the cost of the metal but in the value of the equipment it keeps running.

---

1. WHY CASTING MATTERS FOR RECOVERY

1.1 What casting does that machining cannot

Machining removes metal from a solid block to achieve a desired shape. This works well for shafts, bushings, and parts with simple external geometry. But consider a gate valve body for a hydro station (Doc #65): it is a hollow object with internal passages, flanged connections, and a seat for the gate — a shape that cannot practically be machined from solid stock. It must be cast.

Similarly:

• Pipe fittings (tees, elbows, reducers, flanges): NZ imports these by the thousands. Every plumbing, water, and industrial system depends on them. They are cast parts.
• Pump housings and impellers: The volute shape of a centrifugal pump cannot be economically machined — it is cast, then critical surfaces are finish-machined (Doc #91).
• Engine blocks and cylinder heads: The complex water jacket passages and oil galleries of an internal combustion engine are produced by casting with internal cores. No other process achieves this geometry.
• Gear blanks: Large gears start as cast blanks, with teeth cut by machining. The blank itself — especially for large gear sizes — is too expensive or impossible to machine from solid bar stock.
• Machine tool beds and bases: Traditionally cast from grey iron for its vibration-damping properties (Doc #91, Section 9). Steel weldments are a substitute, but with measurably worse vibration damping — grey iron’s graphite flake structure absorbs 20–25 times more vibrational energy than steel, which is why precision machine tools historically use cast iron beds.⁵
• Bollards, cleats, and heavy marine hardware: For sailing vessel construction (Doc #138), cast bronze fittings are stronger and more corrosion-resistant than fabricated alternatives.

1.2 Casting and machining are complementary

Most castings require some machining to reach final dimensions. A cast valve body needs its sealing faces machined flat. A cast gear blank needs its bore turned true and its teeth cut on a mill. A cast pump housing needs its shaft bore and seal faces finished to precise tolerances.

This means the foundry and the machine shop work in sequence: the foundry produces the near-net-shape part, and the machine shop finishes the critical surfaces. Neither capability is useful without the other for most functional parts. The interaction also runs the other way — machine shops produce the patterns, core boxes, and moulding tools that the foundry needs.

1.3 Products NZ will need to cast

A rough priority list based on recovery demand:

Immediate demand (Phase 2):

• Pipe fittings: tees, elbows, flanges, valves for water and industrial systems
• Bearing housings: replacements for failed housings on agricultural and industrial machinery
• Brake drums and discs: for vehicle fleet maintenance (Doc #33)
• Pump components: impellers, volute casings, wear plates
• Bollards and cleats: marine hardware for port operations and new vessel construction

Medium-term demand (Phase 3–4):

• Engine blocks and heads: for rebuilding internal combustion engines as spares deplete
• Gear blanks: for gearbox repair and new gear production
• Machine tool castings: beds, headstocks, tailstocks for the machine tool bootstrap (Doc #91, Section 9)
• Stove and furnace components: fireboxes, grates, doors for domestic and industrial heating
• Rail fittings: chairs, fishplates, points components if rail system is maintained

Ongoing:

• General replacement parts for any machinery where the original was a casting

---

2. SAND CASTING: THE CORE PROCESS

2.1 Overview

Sand casting is the oldest and most widely used casting process. The principle is straightforward: pack sand around a pattern (a replica of the desired part), remove the pattern to leave a cavity in the sand, pour molten metal into the cavity, let it solidify, break away the sand, and clean up the casting. The sand can be reused.

The simplicity of the principle conceals significant skill requirements. Producing a sound casting — one that is free of internal porosity, shrinkage cavities, sand inclusions, and cracks — requires understanding of mould design, gating and risering (how metal flows into the mould and how shrinkage is fed), sand properties, metal chemistry, pouring temperature, and solidification behaviour. A poorly made casting looks fine on the outside but fails in service because of internal defects that are invisible without inspection. This is not a process where trial-and-error is cheap — each failed casting wastes metal, sand preparation time, and furnace energy.

2.2 The pattern

The pattern is the shape that forms the mould cavity. For sand casting, patterns are traditionally made from wood — typically a close-grained hardwood or, more commonly in modern practice, medium-density fibreboard (MDF) or engineered board. NZ has abundant timber suitable for patternmaking: kauri (existing stocks — no longer harvested), rimu, beech, and radiata pine (with appropriate selection for grain stability).⁶

Key patternmaking requirements:

• Draft angles: Pattern surfaces must be tapered slightly (typically 1–3 degrees) so the pattern can be withdrawn from the sand without damaging the mould. A straight-sided pattern would tear the sand as it is removed.
• Shrinkage allowance: Metal shrinks as it solidifies and cools. The pattern must be oversized to compensate. The shrinkage allowance varies by alloy and casting geometry: approximately 0.8–1.2% for grey cast iron, 1.0–1.6% for steel, 1.2–1.8% for bronze, 1.0–1.5% for aluminium — exact values depend on section thickness and mould rigidity.⁷ Patternmakers use special “shrink rules” — rulers that are deliberately oversized by the shrinkage percentage.
• Machining allowance: Where surfaces will be finish-machined after casting, extra material must be added to the pattern — typically 3–6 mm per surface, depending on the casting size and the precision required.
• Core prints: Where the casting has internal cavities (holes, passages), the pattern includes extensions called core prints that form recesses in the mould to support sand cores. Core design is one of the most skill-intensive aspects of patternmaking.
• Split patterns: Most patterns are split along the parting line (the plane where the two halves of the mould meet), with alignment pins to ensure the halves register correctly.

The patternmaker’s skill: Patternmaking is a distinct trade, historically separate from both moulding and machining. A skilled patternmaker must understand the entire casting process — how metal flows, where shrinkage occurs, how moulds are made — to build a pattern that produces a sound casting. In NZ, dedicated patternmakers are rare; in many foundries, the moulder or a machinist makes the patterns. Under recovery conditions, patternmaking skill becomes a training priority because every new cast part starts with a pattern.

Patterns can also be made from metal (for high-volume production — the pattern lasts indefinitely) or from plastic and resin (for short runs). Under recovery conditions, wood patterns are the practical default because NZ has the materials and the woodworking skills.

2.3 Moulding sand

The mould is made from sand — specifically, a mixture of sand and a binder that holds the sand grains together after compaction around the pattern.

Green sand (the standard): The most common moulding sand is “green sand” — a mixture of silica sand (SiO₂), bentonite clay (as binder, typically 5–10% by weight), water (2–4%), and sometimes additives like coal dust or sea coal (to improve surface finish on iron castings by creating a reducing atmosphere at the mould-metal interface).⁸

NZ sand sources: NZ has extensive deposits of silica sand suitable for foundry use. Known deposits include sands in the Waikato, Bay of Plenty, and Northland regions.⁹ Not all sand is suitable for foundry work — the grains must be of the right size distribution (AFS grain fineness number, typically 40–80 for general work), reasonably round (angular grains produce weaker moulds), and low in contaminants that could cause casting defects. The specific suitability of NZ sand deposits for foundry use is established by the existing NZ foundry industry, which has been using NZ-sourced sand for decades.

Bentonite clay: The clay binder in green sand is typically sodium or calcium bentonite. NZ has bentonite deposits, though the quality and accessibility for foundry use would need assessment.¹⁰ Small quantities of bentonite can also be imported from Australia via trade if it develops. Alternative clay binders — including some NZ clays not classified as bentonite — may work for lower-quality moulding sand, though with reduced mould strength and surface finish.

Chemical-bonded sand: Modern foundries increasingly use chemically bonded sand systems — sand mixed with resin binders (furan, phenolic, or alkyd resins) that are cured by chemical reaction or heat. These produce stronger, more dimensionally accurate moulds than green sand. However, the resin binders are imported petrochemical products. As resin stocks deplete, NZ foundries will need to revert to green sand for most work — which is functional but represents a real capability reduction: dimensional tolerances widen from ±0.3–0.5 mm (resin-bonded) to ±1–3 mm (green sand), surface finish is rougher requiring more machining, and complex cores are harder to produce without chemical binders. This is manageable for the majority of recovery castings, but means some precision parts will require more post-casting machining.¹¹

Sand reclamation: Used moulding sand can be recycled. Green sand is reconstituted by adding fresh clay and water. Chemically bonded sand requires thermal reclamation (heating to burn off spent resin) for full reuse, though it can be blended with new sand at reduced proportions. Sand recycling is standard foundry practice and reduces the demand for new sand.

2.4 Moulding

Moulding is the process of packing sand around the pattern to create the mould cavity. For most sand casting, this involves:

1. Flask preparation: A flask is a rectangular frame (typically cast iron or steel) in two halves — the cope (top) and drag (bottom). The pattern is placed on a flat surface (the moulding board), the drag half of the flask is positioned around it, and sand is packed in. The drag is then turned over, the cope half is placed on top, the upper portion of the pattern is positioned, and more sand is packed in.

2. Ramming: Sand is compacted around the pattern by hand ramming (using a hand rammer — a weighted tool) or by machine (jolt-squeeze or sand-slinger machines). Proper compaction is critical — too loose and the mould collapses when metal is poured; too dense and gases cannot escape, causing porosity.

3. Pattern withdrawal: The pattern is carefully removed from the sand, leaving the cavity. This is the moment of truth — if the sand is not properly bonded, or if the pattern lacks adequate draft, the mould face crumbles. Repair of minor damage is possible with hand tools, but major collapse means starting over.

4. Core setting: If the casting has internal cavities, preformed sand cores are placed in the mould cavity. Cores are made separately in core boxes (patterns for the cores), typically from sand bonded with oil, resin, or — historically — linseed oil and flour mixtures that harden when baked.¹²

5. Closing: The cope is placed on the drag, and the mould is clamped or weighted to resist the metallostatic pressure of the molten metal (which can be substantial — liquid iron is dense).

6. Gating and risering: The mould includes channels (gates and runners) through which metal enters the cavity, and risers (reservoirs of metal) that feed the casting as it solidifies and shrinks. Gating design is one of the most important skills in founding — incorrect gating causes misruns (incomplete filling), turbulence (which entrains sand and oxides), and shrinkage defects. There are principles and rules of thumb, but much of the skill is learned through experience.¹³

2.5 Pouring

Molten metal is transferred from the furnace to the mould in a ladle — a refractory-lined container with a pouring lip. Pouring temperature, pour rate, and pouring technique all affect casting quality.

Pouring temperature: Must be high enough that the metal fills the mould completely before solidifying, but not so high that it causes excessive shrinkage, hot tearing, or sand burn-on. For grey cast iron, typical pouring temperature is 1,300–1,400°C; for bronze, 1,050–1,200°C; for aluminium alloys, 680–750°C.¹⁴ A pyrometer (temperature-measuring instrument) is valuable for consistent results, though experienced founders can judge temperature approximately by colour — bright yellow-white for iron, dull red-orange for bronze.

Pour rate: Too slow and the metal solidifies before filling the mould (misrun). Too fast and turbulence entrains air and sand, creating porosity and inclusions. The correct rate depends on the casting size, gating design, and metal fluidity.

2.6 Shakeout and finishing

After the metal has solidified (minutes to hours depending on casting size), the mould is broken apart — “shaken out” — and the casting is removed. The sand falls away and is returned to the sand system for reconditioning. The casting then requires:

• Fettling: Removing the gates, risers, and flash (thin fins of metal at parting lines) using grinders, chisels, and cutoff wheels.
• Cleaning: Shot blasting or wire brushing to remove adhering sand.
• Heat treatment: Some castings require annealing (to relieve internal stresses), normalising (to refine grain structure), or hardening. Grey iron castings are often used as-cast; steel castings typically require heat treatment.
• Inspection: Visual inspection, dimensional checking, and — for critical parts — pressure testing (for valve bodies and pipe fittings) or non-destructive testing (see Section 10).
• Machining: Finish-machining of critical surfaces (Doc #91).

---

3. METALS FOR CASTING

3.1 Grey cast iron

Grey cast iron is the workhorse of the foundry. It is the easiest ferrous material to cast (it flows well, has low shrinkage, and is forgiving of mould imperfections), it machines readily, and its graphite flake structure provides excellent vibration damping and self-lubricating properties. Most of the castings listed in Section 1.3 — pipe fittings, valve bodies, machine tool beds, engine blocks, brake drums — are traditionally made from grey iron.

Composition: 2.5–4.0% carbon, 1.0–3.0% silicon, with smaller amounts of manganese, sulfur, and phosphorus. The high carbon content (compared to steel at 0.1–0.6% carbon) is what makes iron “cast” — it lowers the melting point to approximately 1,150–1,200°C and improves fluidity.¹⁵

NZ source material: Grey iron is made by melting a charge of:

• Scrap cast iron: NZ has a large stock of existing cast iron in engine blocks, brake drums, machine bases, pipe fittings, stove parts, and industrial equipment. This is the primary raw material for the foundry. Scrap cast iron is identifiable by its characteristic fracture (grey, granular) and by the fact that it is brittle — it breaks rather than bends.¹⁶
• Steel scrap: Can be blended with iron scrap to adjust chemistry, though pure steel scrap has lower carbon and silicon than needed for grey iron.
• Pig iron: Not produced in NZ. Glenbrook (Doc #89) produces steel, not pig iron. However, the reduced iron product from Glenbrook’s kilns is high in carbon and could potentially serve as a pig iron substitute if diverted before the steelmaking (refining) stage. This would require coordination with NZ Steel and is not current practice.
• Ferrosilicon: An alloying addition used to control graphite structure. NZ does not produce ferrosilicon. Existing stocks will deplete. Silicon can in principle be produced from quartz sand in an electric arc furnace (a major project), but this is a Phase 4+ capability at best.

Honest assessment: NZ can produce grey iron castings from scrap for the foreseeable future — the scrap stock is large relative to likely demand. The limitation is alloying control: without ferrosilicon and other additives, the metallurgical quality of castings will be less consistent. Scrap-based iron of variable composition produces castings of variable properties. This is manageable for many applications (pipe fittings, stove parts, simple machine components) but problematic for critical parts (engine blocks, high-pressure valve bodies) where material properties must meet specification.

3.2 Ductile (nodular) iron

Ductile iron is cast iron treated with magnesium (or cerium) to change the graphite structure from flakes to spheroids (nodules), dramatically improving strength and ductility. Ductile iron approaches steel in mechanical properties while retaining the castability of grey iron. Modern applications include crankshafts, suspension components, pipe, and heavy-duty gears.¹⁷

NZ feasibility: Ductile iron production requires magnesium treatment of the liquid iron — typically adding a magnesium-ferrosilicon alloy to the ladle just before pouring. NZ does not produce magnesium metal or magnesium-ferrosilicon alloys. Without imports, ductile iron production is not feasible. This is a significant limitation — many modern castings that replaced older steel forgings are ductile iron, and NZ cannot replicate them without the treatment alloy. The fallback is either grey iron (with roughly one-third to one-half the tensile strength of ductile iron — typically 150–250 MPa vs. 400–700 MPa — and essentially zero ductility, meaning it fractures rather than deforms under overload) or steel castings (stronger than ductile iron but significantly harder to produce — see Section 3.5).

3.3 Bronze and brass

Bronze (copper-tin alloy) and brass (copper-zinc alloy) are the traditional casting alloys for bearings, marine fittings, valve components, and any application requiring corrosion resistance. Bronze bearings (cast and machined to size) are the local substitute for ball and roller bearings that NZ cannot produce (Doc #91).

Common casting alloys:

• Gunmetal (leaded tin bronze): ~88% copper, 10% tin, 2% zinc. Excellent for valve bodies, pump components, and general marine fittings. The standard foundry bronze.
• Phosphor bronze: Copper-tin with phosphorus addition. Excellent for bearings and wear-resistant applications.
• Manganese bronze: Higher-strength alloy used for propellers and heavy marine hardware.
• Brass: Copper-zinc alloys. Easier to cast than bronze (lower melting point) but lower strength and corrosion resistance.

NZ copper and tin supply: This is the constraint. NZ has minor copper production — small deposits exist, particularly as byproducts of gold mining in the Coromandel and West Coast.¹⁸ This is not enough for significant bronze casting. NZ has no tin production at all.

NZ’s bronze casting capability therefore depends on:

1. Existing stocks: Scrap bronze and brass in the NZ economy — old valves, plumbing fittings, propellers, electrical components, decorative brassware. The total stock is unknown but probably modest relative to long-term demand.
2. Trade with Australia: Australia has copper mines and smelters. If maritime trade develops (Doc #138), copper and tin are exactly the kind of high-value, low-volume goods that sail trade handles well.
3. Careful allocation: Bronze should be reserved for applications where its properties are essential (bearings, marine fittings, valve seats) and not wasted on applications where iron or aluminium would serve.

Melting point: Bronze alloys melt at approximately 900–1,050°C — significantly lower than iron, making them easier to melt in smaller furnaces. Bronze can be melted in a simple crucible furnace fuelled by charcoal, coke, oil, or electricity.

3.4 Aluminium alloys

Aluminium is light, corrosion-resistant, and easy to cast — its low melting point (approximately 660°C for pure aluminium, 570–630°C for common casting alloys) makes it the easiest metal to work with in a small foundry. The Gingery approach to bootstrapping a machine shop (Doc #91, Section 9) uses aluminium castings specifically because they are the most accessible to a beginner.

NZ aluminium supply: NZ’s only aluminium source is the Tiwai Point smelter near Bluff, which produces aluminium from imported Australian alumina. If the smelter is operating at the time of the event and has alumina stocks, NZ has aluminium for some period. Once alumina stocks are exhausted, primary aluminium production ceases.¹⁹

However, NZ has a large accumulated stock of aluminium in the economy — in vehicles (engine blocks, cylinder heads, transmission cases, wheels), in buildings (window frames, cladding), in cookware, and in industrial equipment. This scrap stock provides a long-term supply for casting, as aluminium can be remelted repeatedly with minimal loss.

Common casting alloys: The most widely used aluminium casting alloy is A356 (or its variants) — an aluminium-silicon alloy (~7% Si) that casts well and has good mechanical properties. When casting from scrap, the alloy composition will be uncertain and variable — this is manageable for non-critical parts but means properties cannot be guaranteed to specification.

3.5 Steel casting

Steel castings are stronger than iron castings and are used for highly stressed components — gear blanks, structural nodes, heavy machinery parts. However, steel is significantly harder to cast than iron:

• Melting point is higher (~1,500°C vs. ~1,200°C for iron)
• Fluidity is lower — steel does not flow into thin sections as readily
• Shrinkage is greater (~2% vs. ~1% for iron) — risering must be more generous
• Steel castings almost always require heat treatment
• Moulding sands must withstand higher temperatures

Steel castings are within NZ’s capability if the furnace can reach the required temperature (electric induction furnaces can; cupola furnaces cannot, as they produce iron, not steel). Steel casting should be reserved for parts that genuinely require steel’s strength and ductility — for most applications, grey iron or bronze is adequate and far easier to produce.

---

4. MELTING EQUIPMENT

4.1 Electric induction furnace

The induction furnace is the standard modern foundry melting unit and NZ’s primary advantage for foundry work. It uses an alternating current through a water-cooled copper coil to generate a powerful electromagnetic field that induces eddy currents in the metal charge, heating it to melting temperature. No fuel, no electrodes, no combustion atmosphere — only electricity, a refractory crucible, and cooling water.

Advantages for NZ:

• Powered by NZ’s renewable grid — no imported fuel required
• Produces clean metal (no combustion gases contaminating the melt)
• Temperature is precisely controllable
• Can melt any metal (iron, steel, bronze, aluminium) by changing the crucible and charge
• Rapid melting (a medium-frequency induction furnace melts a charge in 30–60 minutes typically)
• Existing units in NZ foundries

Capacity range: Induction furnaces range from small bench-top units (a few kilograms — used in jewellery and dental work) to large industrial units (several tonnes per heat). NZ’s existing foundry induction furnaces are likely in the 250 kg to 2 tonne range for medium foundries.²⁰ Larger units exist at major operations.

Dependencies:

• Electricity: A medium-frequency induction furnace melting 500 kg of iron draws approximately 300–500 kW. This is a significant but not extraordinary load — comparable to a large workshop’s total connected load. Grid power under baseline conditions is adequate.
• Crucible (refractory lining): The crucible is a refractory vessel that holds the molten metal inside the coil. It is consumed over time — crucible life depends on the metal being melted, the temperature cycles, and the refractory quality. Crucibles are typically made from silica-based or alumina-based refractory materials. NZ has silica and fireclay suitable for basic crucible production, though imported high-alumina and magnesia crucibles perform better. Crucible replacement is a running cost, not a show-stopper — NZ can produce adequate crucibles from domestic materials.²¹
• Copper coil: The induction coil is made from copper tubing. Copper is limited in NZ (Section 3.3). If a coil fails, replacing it requires copper and the skill to wind and braze a new coil. This is feasible but copper availability is the constraint.
• Power electronics: Modern induction furnaces use power electronics (thyristors or IGBTs) to convert grid-frequency power to the medium-frequency (150–3,000 Hz) that efficient induction heating requires. These electronic components are imported and have a finite life. When they fail, the furnace is inoperable unless the electronics can be repaired or a replacement unit is available. This is the same controller-failure problem that affects CNC machines (Doc #91, Section 6). Older furnaces using mains-frequency (50 Hz) induction are simpler — they use less electronics — but are less efficient and are suitable mainly for larger furnaces.
• Cooling water: The coil and power electronics require water cooling. A recirculating water system with a cooling tower is standard. Water is not a constraint in NZ.

Failure modes and timeline: The power electronics are the weak link. Thyristors and capacitor banks have finite service lives — thyristor modules typically last 15–30 years under normal industrial duty, and capacitor banks 10–20 years, though both depend heavily on thermal management and utilisation intensity.²² With careful maintenance and reduced utilisation, existing NZ induction furnaces might remain operational for 10–25 years. Replacement electronics would need to come from trade or, eventually, from NZ-produced components (a long-term prospect). The coil itself is robust if cooling is maintained.

4.2 Cupola furnace

The cupola is a simple vertical shaft furnace — essentially a large metal cylinder lined with refractory, charged from the top with alternating layers of iron, coke (or charcoal), and limestone flux, and blown with air from tuyeres (nozzles) near the base. The coke burns in the air blast, melting the iron, which collects in a well at the bottom and is tapped into a ladle for pouring. The cupola has been the standard iron foundry melting unit since the 18th century and produced the vast majority of the world’s iron castings until the mid-20th century.²³

Advantages:

• Structurally simple — can be built from a steel shell lined with fireclay brick. NZ has the materials and fabrication capability (Doc #91, Doc #89).
• High production rate — a medium cupola (600–900 mm internal diameter) can melt 1–5 tonnes of iron per hour.
• Fuel is coke or charcoal, both producible in NZ. Coke from NZ coal (Waikato sub-bituminous or West Coast bituminous); charcoal from NZ timber (Doc #99, if written, or general charcoal production — well-established process using hardwood or dense softwood).²⁴
• Does not depend on electricity or electronics.
• Produces hot metal continuously during a “campaign” (a melting run lasting hours to a full working day).

Limitations:

• Produces cast iron only, not steel (the carbon content increases as the iron passes through the coke bed). Not suitable for bronze or aluminium.
• Metal chemistry control is less precise than with an induction furnace — the coke bed and air blast affect carbon, silicon, and sulfur levels in ways that are manageable but harder to control finely.
• Requires good-quality coke or charcoal. NZ coal is mostly sub-bituminous, which produces lower-quality coke than bituminous coal — this affects cupola performance. West Coast bituminous coal produces better coke but is further from likely foundry locations. Charcoal is an effective alternative that NZ can produce abundantly, though charcoal-fuelled cupolas operate somewhat differently from coke-fuelled ones (lower carbon pickup, different melting rate).²⁵
• Air blast is required — traditionally from a hand- or motor-driven blower. An electric blower is convenient; a belt-driven blower from a steam or water engine is the historical alternative.
• Emissions: cupola operation produces significant smoke and fumes. This is a workplace health issue that requires ventilation and, ideally, outdoor or well-ventilated siting.

NZ application: The cupola is the appropriate melting unit for a new foundry established in a region without existing induction furnace capability, or as a backup when induction furnace electronics fail. Building a cupola is a straightforward engineering project — the shell can be fabricated from steel plate (Doc #89), the refractory lining from NZ fireclay brick, the tuyeres from steel pipe, and the blower from available motor and fan components. A functional cupola could be built in 4–8 weeks by a competent fabrication team, assuming steel plate, fireclay brick, and blower components are available. The dependency chain is: steel plate (Doc #89) → cutting and rolling to cylindrical shell → fireclay brick (from NZ clay deposits, requiring mining, forming, and kiln firing — itself a multi-week process if brick stocks are not available) → steel pipe for tuyeres → blower (electric motor-driven, or belt-driven from available power source). If fireclay brick must be produced rather than sourced from existing stocks, add 2–4 weeks for brick production.²⁶

4.3 Crucible furnace

For small-scale and non-ferrous work (bronze, brass, aluminium), the crucible furnace is the simplest option. A crucible — a refractory container — is placed inside a furnace (a simple firebrick enclosure) and heated externally by combustion of coke, charcoal, oil, gas, or by electrical resistance elements. The charge is placed in the crucible, melts, and is poured directly from the crucible or transferred to a ladle.

Capacity: Small — typically 10–100 kg per heat for a manually handled crucible. This is adequate for individual castings of bearing shells, valve components, small fittings, and similar parts.

NZ materials: Silicon carbide crucibles (the standard modern type) are imported. Clay-graphite crucibles can potentially be produced from NZ clay and imported graphite (limited). Alternatively, steel crucibles lined with a refractory wash can serve for aluminium melting (aluminium does not attack steel at its lower melting temperature). For bronze, a clay-bonded crucible or a commercially produced crucible from existing NZ stocks would be used.

4.4 Melting equipment summary

Furnace type	Metals	Capacity	Fuel/power	NZ buildable?	Best application
Induction (medium-frequency)	All	50 kg–5 tonnes	Electricity	No (electronics imported) — existing units only	Primary foundry melting where grid power available
Induction (mains-frequency)	All	500 kg–10 tonnes	Electricity	Partially — simpler electronics	Large foundry melting; longer-term option
Cupola	Cast iron	1–5+ tonnes/hour	Coke or charcoal + air	Yes — fully NZ-buildable	Iron foundry production; electricity-independent
Crucible (fuel-fired)	Bronze, brass, aluminium	10–100 kg	Coke, charcoal, oil, gas	Yes — fully NZ-buildable	Small-scale non-ferrous casting
Crucible (electric resistance)	Bronze, brass, aluminium	10–100 kg	Electricity	Partially — elements imported	Small foundry non-ferrous work

---

5. NZ’S EXISTING FOUNDRY CAPABILITY

5.1 Known foundries

NZ’s foundry sector is small and has contracted over recent decades as imported castings — primarily from China, India, and Southeast Asia — undercut domestic production on price. The remaining NZ foundries serve applications where local production still makes economic sense: urgent repairs, custom one-off work, and niche products.

Foundries known to be operating or recently operating in NZ include:²⁷

• Bradken (formerly Aimex/Christchurch): Industrial castings including mining and materials-handling components. One of the larger NZ casting operations.
• De Beer Engineering / NZ Foundry Group (Auckland area): General engineering foundry work, including iron and non-ferrous castings.
• Harris Pumps (Auckland): Pump manufacturer with in-house casting capability for pump components.
• Metcast Group (Hamilton area): Iron and steel casting for industrial applications.
• Various small non-ferrous foundries: Several small operations in Auckland, Christchurch, and elsewhere producing aluminium and bronze castings for marine, engineering, and art applications.

Honest assessment: The exact current status, capability, and capacity of NZ foundries requires verification through the national census (Doc #8). The sector has changed — some operations listed in older directories may have closed, and some casting work may have moved to in-house operations within larger engineering firms that are not primarily identified as foundries. The census should specifically ask engineering firms whether they have any melting and casting capability, even if it is not their primary business.

5.2 Related capability

Beyond dedicated foundries, NZ has casting-adjacent capability in:

• Jewellery and art casting: Small-scale lost-wax casting operations in several NZ cities. The technical process is identical to engineering lost-wax casting (Section 8), though at smaller scale and with finer tolerances than most engineering applications require.
• Dental laboratories: Some NZ dental labs cast precision metal prosthetics using lost-wax investment casting — directly relevant skills.
• NZ Steel Glenbrook (Doc #89): Not a foundry in the traditional sense, but has substantial metallurgical knowledge, refractory experience, and furnace operation expertise that is transferable.
• Polytechnic workshops: Some Te Pūkenga institutions include foundry practice in their engineering trades programs, though this has been declining as the industry contracted.

5.3 Workforce

The NZ foundry workforce is small — probably fewer than 200 people directly employed in casting operations nationally, though this figure requires census verification.²⁸ As with machine shop workers (Doc #91), the experienced foundry workers tend to be older. The skill set — moulding, pattern-making, furnace operation, gating design, metallurgical quality control — takes years to develop and is not taught widely in current NZ training programmes.

Knowledge capture from experienced NZ foundry workers is a Phase 1–2 priority, following the same model as for machinists (Doc #91, Section 5): identify experienced practitioners, pair them with learners, document techniques, and ensure the knowledge is transferred before it is lost to retirement or death.

---

6. ESTABLISHING NEW FOUNDRY CAPABILITY

6.1 Why existing capacity is not enough

NZ’s existing foundries serve a small domestic market for custom and repair castings. Under recovery conditions, demand for castings will increase dramatically as imported parts deplete and NZ must produce replacement components locally. The existing foundries cannot meet this demand without expansion, and additional foundries will be needed — particularly in regions (the South Island, rural North Island) where no foundry capability currently exists.

6.2 What a basic foundry requires

A functional sand-casting foundry requires:

Physical plant:

• A building or covered area large enough for moulding, melting, and pouring operations. A building of 200–500 m² is adequate for a small-to-medium operation.²⁹ Existing industrial buildings, farm buildings, or large sheds can be adapted.
• A concrete floor (for safety — molten metal spills on a wooden floor cause fires). If no concrete floor exists, a rammed earth or gravel floor is acceptable if kept dry.
• Adequate ventilation — molten metal produces fumes, and sand moulding generates dust. Natural ventilation in a large shed may suffice; forced ventilation (fans) is better.
• An overhead lifting capability — a chain hoist or small crane, because ladles of molten metal are too heavy to carry safely by hand for all but the smallest pours. A 1-tonne chain block on a beam serves a small foundry.

Equipment:

• Melting furnace (induction, cupola, or crucible — Section 4)
• Ladles: refractory-lined steel containers for carrying molten metal. Several sizes, from handheld (10–20 kg capacity for small pours) to crane-handled (100+ kg). Ladles can be fabricated from steel plate and lined with refractory cement — NZ-producible.
• Moulding flasks: steel or iron frames in matched pairs. Can be fabricated locally.
• Sand preparation equipment: a muller (machine for mixing sand, clay, and water — essentially a heavy mixer) or, for small operations, a flat mixing surface and hand tools. A simple muller can be built from available components.
• Pattern-making tools: woodworking hand tools, which NZ has in abundance.
• Moulding tools: hand rammers, trowels, slicks, sprue pins, vent rods. Simple hand tools, most of which can be made by a blacksmith (Doc #92) or machinist (Doc #91).
• Basic pyrometer or thermocouple for temperature measurement (existing instruments — not NZ-producible, but existing stocks should be adequate for foundry allocation).
• Personal protective equipment: face shields, heat-resistant gloves, aprons, safety boots. Leather-based PPE can be produced in NZ.

Materials:

• Moulding sand (NZ-sourced — Section 2.3)
• Bentonite clay (NZ-sourced or imported)
• Refractory material for furnace/crucible lining (NZ fireclay, dolomite — Section 5 of Doc #89)
• Metal charge stock (scrap iron, bronze, aluminium — Section 3)
• Pattern material (NZ timber — Section 2.2)
• Fuel for cupola or crucible furnace (coke or charcoal — NZ-producible)

People: This is the real constraint. A foundry requires:

• At least one experienced founder who understands mould design, gating, metal behaviour, and furnace operation
• Moulders — people who can prepare and ram moulds. This is a learnable skill but takes months of practice to do consistently well.
• A patternmaker (or a skilled woodworker who can learn patternmaking)
• A furnace operator
• Labourers for sand preparation, shakeout, and fettling

A small foundry can operate with 4–8 people.³⁰ Scaling up requires more moulders and support staff, not more specialists — one experienced founder can supervise multiple moulding stations.

6.3 Recommended geographic distribution

Under recovery conditions, foundry capability should exist in at least the following regions:

• Auckland/Waikato: Existing foundry base. Expand capacity. Proximity to NZ Steel (Doc #89) for steel supply and to the largest concentration of machinery requiring cast replacement parts.
• Christchurch/Canterbury: Existing foundry base (Bradken and others). Critical for South Island agricultural and industrial equipment maintenance.
• Wellington/Lower North Island: New capability likely needed. High concentration of port and marine infrastructure requiring cast fittings.
• Otago/Southland: New capability needed unless an existing operation can expand. Critical for South Island agricultural district.
• Regional North Island (Bay of Plenty, Taranaki, Hawke’s Bay): At least one foundry per major agricultural region, even if small-scale (crucible furnace for non-ferrous work plus cupola for iron).

---

7. CUPOLA FURNACE: DESIGN AND OPERATION

The cupola is the most important new foundry capability to develop because it is fully NZ-buildable, does not depend on electricity or electronics, and can produce iron castings at useful scale. This section provides enough detail for an engineering team to build and operate one.

7.1 Construction

A small-to-medium cupola for a regional foundry:

Shell: A steel cylinder approximately 600–900 mm internal diameter and 3–5 m tall, fabricated from 8–12 mm steel plate (from Glenbrook, Doc #89). The shell is set on legs or a support structure at a height that allows a ladle to be positioned under the tap hole. A hinged bottom door allows the remaining charge and slag to be dumped at the end of a campaign.

Refractory lining: The interior is lined with fireclay brick or rammed fireclay refractory, approximately 75–100 mm thick in the melting zone. NZ fireclay is suitable for this application — cupola operating temperatures (typically 1,400–1,600°C at the tuyere level) are within the range of basic fireclay refractories. Lining life is variable — weeks to months of campaign operation depending on the refractory quality and operating severity. Relining is straightforward and uses NZ materials.³¹

Tuyeres: Typically 4–8 openings (steel pipes, 50–75 mm diameter) equally spaced around the circumference at approximately 200–300 mm above the bottom. Through these, the air blast enters the furnace, burning the coke and melting the iron.

Air supply: A centrifugal blower delivering approximately 100–200 m³/min of air per square metre of cupola cross-section at a pressure of 5–15 kPa. An electric motor-driven blower is the standard modern arrangement. For an electricity-independent installation, a belt-driven blower from a steam engine, water wheel, or wood gas engine serves the same purpose — this was the historical practice.

Tap hole and slag hole: Openings at the base for tapping molten iron (lower) and slag (higher). The iron tap hole is plugged with clay between taps. The slag hole allows the lighter slag to overflow continuously or periodically during operation.

Charging door: An opening in the upper portion of the shell, typically 2–3 m above the tuyeres, through which the charge (iron, coke/charcoal, limestone) is loaded. Charging is done in alternating layers: a layer of coke, a layer of iron scrap, a layer of coke, and so on, with limestone flux added periodically.

7.2 Operation

Bed coke: Before charging iron, a bed of coke is built up in the cupola from the tuyere level to approximately 1–1.5 m above the tuyeres. This bed is ignited (typically by lighting wood kindling at the bottom and allowing the coke bed to light progressively, or by using a gas burner through a tuyere). The bed must be fully incandescent before charging begins.

Charging: Iron scrap and coke are charged in alternating layers from the top. The ratio of iron to coke (the “charge ratio”) determines the melting rate and iron temperature. A typical charge ratio is 8:1 to 10:1 (iron to coke by weight). Limestone is added at approximately 2–4% of the iron charge weight as flux.³²

Melting: Once the iron charges descend through the coke bed and melt, liquid iron collects in the well below the tuyeres. Iron is tapped periodically (every 5–15 minutes for continuous operation) by opening the tap hole with a pointed bar and allowing the iron to flow into a ladle. When sufficient iron has been tapped, the hole is plugged with clay.

Campaign length: A cupola campaign (continuous melting run) typically lasts 4–12 hours. At the end, the air blast is shut off, the bottom door is opened, and the remaining contents are dumped onto a sand bed.

Typical output: A 750 mm internal diameter cupola operating with charcoal or coke can produce approximately 1–3 tonnes of molten iron per hour, depending on the coke ratio, blast rate, and iron charge characteristics. A single campaign might produce 5–20 tonnes of iron — enough for a large number of castings.³³

7.3 Metal chemistry control

Cupola iron chemistry is influenced by the charge composition, the coke quality, and the operating parameters. Key variables:

• Carbon: Iron absorbs carbon from the coke bed, typically producing metal with 3.0–3.5% carbon. This is generally suitable for grey iron castings. If lower carbon is desired (for stronger iron), the charge can include more steel scrap; if higher carbon is desired, more pig iron or cast iron scrap is used.
• Silicon: Silicon is lost during cupola melting (it oxidises in the blast). To maintain the silicon content needed for grey iron (1.5–2.5%), ferrosilicon additions may be needed in the ladle. Without ferrosilicon (an imported alloy), silicon can be partially maintained by using high-silicon iron scrap and by operating at higher temperatures. This is a genuine limitation — low-silicon cupola iron tends to produce hard, brittle “white iron” that is difficult to machine. Managing silicon without ferrosilicon requires careful scrap selection and operator skill.³⁴
• Sulfur: Coke contains sulfur, which transfers to the iron and makes it brittle and difficult to cast. Charcoal contains very little sulfur — this is a significant advantage of charcoal-fuelled cupola operation for NZ. The limestone flux also helps remove sulfur into the slag.

---

8. LOST-WAX (INVESTMENT) CASTING

8.1 When sand casting is not precise enough

Sand casting produces parts to a tolerance of approximately ±1–3 mm, with surface roughness that requires machining for functional surfaces. For most industrial castings — pipe fittings, valve bodies, pump housings — this is adequate because the critical surfaces are machined after casting.

Some parts require greater precision or finer surface finish directly from the casting, or have geometry that is too complex for sand moulding (thin walls, fine detail, undercuts that cannot be withdrawn from a sand mould). Lost-wax casting (also called investment casting) addresses these applications.

8.2 The process

1. Wax pattern: A precise wax replica of the desired part is made, either by carving/machining wax directly or (for production quantities) by injecting molten wax into a metal die. Under recovery conditions, hand-made wax patterns are the practical method. Beeswax, blended with pine resin for hardness, is a NZ-producible pattern material.³⁵ Paraffin wax from petroleum stocks also works while stocks last.
2. Assembly: Multiple wax patterns are attached to a wax “tree” (a central sprue with patterns branching off), allowing multiple parts to be cast in a single mould.
3. Investment: The wax tree is coated with a ceramic slurry — typically fine silica flour mixed with a liquid binder (colloidal silica or ethyl silicate). Multiple coats are applied, with each coat dipped and stuccoed with progressively coarser refractory grain. The result is a ceramic shell, typically 5–10 mm thick, encasing the wax patterns.
4. Dewaxing: The shell is heated (typically in an oven at 100–200°C) to melt and drain the wax. The wax runs out, leaving a precise cavity in the ceramic shell. The wax can be recovered and reused.
5. Firing: The shell is heated to approximately 900–1,100°C to sinter the ceramic and burn out any remaining wax residue.
6. Pouring: Molten metal is poured into the hot shell while it is still at elevated temperature (pouring into a hot mould improves metal flow into thin sections).
7. Shakeout: After solidification, the ceramic shell is broken away (it is designed to be brittle after casting) and the castings are cut from the tree.

8.3 NZ applications and feasibility

Lost-wax casting in NZ would serve:

• Turbine and pump impellers: Complex blade geometry that sand casting handles poorly.
• Precision valve components: Valve trim (plugs, seats, cages) where dimensional accuracy reduces the machining required.
• Small hardware: Marine fittings, chain links, structural connectors.
• Dental and surgical instruments: If NZ’s medical equipment maintenance requires it.

Feasibility: Lost-wax casting is feasible in NZ using local materials, but the ceramic shell materials are the constraint. Silica flour can be produced from NZ silica sand by grinding. The liquid binder (colloidal silica) is an industrial chemical that NZ does not currently produce — it can be made from sodium silicate (water glass) by acid treatment, but this requires chemical processing capability. Ethyl silicate is a more complex chemical. Plaster of Paris (gypsum-based) can serve as an alternative investment material for lower-temperature alloys (aluminium, bronze), producing what is sometimes called “plaster mould casting” — less precise than ceramic shell but simpler to prepare and within NZ material capability.³⁶

Feasibility rating for lost-wax casting specifically: [B–C] — the basic process is feasible with NZ materials (particularly for plaster-mould variants), but achieving the precision and consistency of pre-war commercial investment casting requires ceramic shell materials that NZ does not currently produce.

---

9. CAST PRODUCTS: SPECIFICATIONS AND STANDARDS

9.1 Pipe fittings

One of the highest-volume casting demands under recovery conditions. Every water system, industrial plant, and building plumbing system uses cast pipe fittings — tees, elbows, flanges, reducers, valves, couplings.

Material: Grey cast iron for pressurised water mains and industrial piping (traditional). Bronze for valve trim and corrosion-critical applications. Aluminium for low-pressure, lightweight applications.

Pattern library: A standard set of pipe fitting patterns in common sizes (DN 25, 40, 50, 80, 100, 150 — matching NZ’s predominantly metric pipe sizes) should be produced as a regional foundry priority. Once patterns exist, any number of fittings can be cast from them. Patterns should be made in metal (cast aluminium is ideal) for durability, after initial wooden prototypes are proven.³⁷

Quality requirement: Pressure-containing fittings must be sound — free of porosity that would cause leaks under pressure. Every pressure fitting should be hydrostatically tested before installation (filled with water and pressurised to 1.5x the rated working pressure). This is standard practice and does not require sophisticated equipment — a hand pump, pressure gauge, and blanking flanges suffice.

9.2 Bearing shells (plain bearings)

Bronze bearing shells are the local replacement for anti-friction bearings (ball and roller bearings) that NZ cannot produce (Doc #91, Section 4.4). A cast bronze bushing, finish-machined to size and fitted with an oil groove, is a functional bearing for shaft journals, linkage pins, and many rotating applications where ball bearings were used as a convenience.

Material: Leaded tin bronze or phosphor bronze — the traditional bearing alloys. Lead improves the bearing’s ability to conform to shaft imperfections and provides emergency lubrication if the oil film breaks down.

Process: Cast a slightly oversized blank in sand or in a simple metal mould (permanent mould casting — pouring metal into a reusable iron or steel mould). Machine the bore to the required diameter on a lathe (Doc #91). Drill oil holes and cut oil distribution grooves. Press into the housing.

Performance gap: Plain bronze bearings have higher friction, require more lubrication, generate more heat, and have shorter life than ball bearings under the same loads and speeds. For many applications (slow-speed agricultural equipment, structural pivots, non-critical rotating assemblies), this gap is acceptable. For high-speed or high-load applications (machine tool spindles, electric motor shafts), plain bearings work but require careful design and lubrication management.³⁸

9.3 Gear blanks

Gears transmit power in virtually every piece of machinery. When a gear tooth breaks or wears beyond use, the gear must be replaced. Under recovery conditions, replacement gears cannot be imported.

Process: Cast a gear blank (a disc or cylinder of the right diameter and width) from grey iron or steel. Machine the bore and faces on a lathe. Cut the teeth on a milling machine using a dividing head and gear cutter (Doc #91). Heat-treat if required.

The gear cutter problem: Gear tooth profiles are involute curves cut with specialised cutters. NZ has stocks of gear cutters (held by machine shops, polytechnics, and tooling distributors), but these are imported items that will deplete. Eventually, NZ must be able to produce its own gear cutters — a challenging but feasible precision-machining and heat-treatment task.

9.4 Engine blocks and cylinder heads

Rebuilding internal combustion engines requires casting new blocks and heads when the originals are cracked, warped, or worn beyond machining limits. This is the most demanding sand-casting application — engine castings are complex shapes with internal water jackets (formed by cores), thin walls, and tight dimensional requirements.

Feasibility: Engine block casting is within the capability of a well-equipped foundry with experienced personnel, but it is not a beginner project. The pattern is complex (requiring multiple cores that must be precisely positioned), the gating must be carefully designed to avoid porosity in critical areas, and the iron chemistry must be well-controlled to produce the correct hardness and machinability. NZ’s existing automotive engine reconditioning shops (Doc #91) have the dimensional knowledge — they know what the finished engine should measure. The foundry provides the raw casting; the engine shop finishes it.

Realistic timeline: Engine block casting is a Phase 3–4 capability — not because the process is unknown, but because it requires a level of foundry skill, equipment quality, and metallurgical control that takes time to develop. Simpler castings (pipe fittings, bearing housings, gear blanks) should be mastered first.

---

10. QUALITY CONTROL AND DEFECT PREVENTION

10.1 Common casting defects

Casting quality is the difference between a part that works and one that fails in service — potentially catastrophically, as in a valve body that cracks under pressure or a bearing housing that fractures under load. Common defects:

• Porosity: Gas bubbles trapped in the solidified metal. Caused by dissolved gas in the metal (particularly hydrogen in aluminium), moisture in the mould sand, or inadequate venting. Reduces strength and causes leaks in pressure-containing parts.
• Shrinkage cavities: Voids formed where the last metal to solidify has no liquid feed to compensate for solidification shrinkage. Caused by inadequate risering — the risers are too small, too cold, or incorrectly placed. The most common serious defect.
• Sand inclusions: Loose sand particles trapped in the metal. Caused by erosion of the mould surface during pouring (turbulent flow), or by loose sand falling into the mould before closing. Weakens the casting and causes machining problems (hard sand particles damage cutting tools).
• Hot tears: Cracks formed during solidification when the casting is constrained by the mould and cannot shrink freely. Caused by poor mould design (mould is too rigid), wrong alloy composition, or incorrect gating.
• Misrun: Metal solidifies before completely filling the mould cavity. Caused by pouring temperature too low, metal fluidity too low, or gating too restrictive. The result is an incomplete casting.
• Cold shuts: Lines or seams where two metal flows met but did not fuse properly. Related to misrun — the metal was too cold.

10.2 Prevention

Most defects are preventable through:

• Proper gating and risering design: Following established principles for runner size, gate placement, and riser sizing. Published foundry engineering references (Heine, Loper & Rosenthal, Principles of Metal Casting; American Foundry Society publications) provide design rules and methods.³⁹ These references should be secured as part of the NZ knowledge preservation effort.
• Sand control: Maintaining the correct sand properties — moisture content, clay content, grain size, permeability (the ability of the sand to pass gas). Testing sand properties with standard tests (clay content by wash, moisture by drying, permeability by airflow) should be routine practice.
• Metal temperature control: Pouring at the correct temperature for the alloy and the casting section thickness. A pyrometer or thermocouple is essential equipment.
• Clean metal practice: Removing slag and dross from the ladle before pouring. Using proper ladle techniques (quiet pouring, avoiding turbulence in the pouring stream).
• Mould preparation: Ensuring the mould cavity is clean, properly vented (thin vent holes poked through the sand to allow gas to escape), and the cores are correctly placed and supported.

10.3 Inspection

Visual inspection: The first check — look for obvious surface defects, incomplete filling, sand adherence, and cracks. Many defects are visible on the surface.

Dimensional inspection: Measure critical dimensions with calipers and gauges. Compare to drawing or to the original part being replicated.

Pressure testing: For pressure-containing castings (valves, fittings, pump housings), hydrostatic testing is essential. Block all openings, fill with water, pressurise with a hand pump to 1.5x rated pressure, and hold for a defined period (typically 5–10 minutes). Any leak indicates porosity or cracking.

Hammer testing: A crude but informative test — a sound casting rings when struck with a light hammer. A casting with internal defects (cracks, voids) produces a dull thud. Experienced foundry workers use this instinctively.

Non-destructive testing: Magnetic particle inspection (for ferrous castings — reveals surface and near-surface cracks), dye penetrant testing (reveals surface cracks in any material), and ultrasonic testing (reveals internal defects) are all used in modern foundry quality control. NZ has NDT capability in engineering inspection firms, and the consumables (magnetic particles, dye penetrant fluid) have finite stocks but long shelf life. For critical castings (pressure vessels, structural parts in safety-critical applications), NDT should be employed.⁴⁰

Destructive testing: Periodically sacrificing a casting from a production run for sectioning and examination provides direct evidence of internal quality. Section the casting, polish and etch a face, and examine under a magnifying glass or low-power microscope. This reveals porosity, shrinkage, graphite structure (in iron), and grain size — information that helps the founder adjust the process.

---

11. HISTORICAL NZ FOUNDRIES

NZ had a more substantial foundry industry in earlier decades. Notable historical operations include:

• A & G Price, Thames: Established 1868, produced locomotives, marine engines, and heavy engineering castings. One of NZ’s largest foundries for over a century. The expertise developed at Thames over generations is part of NZ’s industrial heritage — some of this knowledge may still be accessible through retired workers and their families in the Thames-Coromandel area.⁴¹
• Dunedin foundries: Dunedin was NZ’s early industrial centre, with multiple foundries in the 19th and early 20th centuries serving the gold mining, agricultural, and marine industries.
• Various railway workshops: NZ Railways workshops (Hutt, Addington, Hillside) operated foundries for locomotive and rolling stock casting production. The Hillside workshops in Dunedin had significant casting capability.⁴²

The closure or contraction of these operations means the physical plant is largely gone, but the institutional memory may be partially recoverable — retired foundry workers, apprenticeship records, technical libraries, and engineering heritage collections.

---

CRITICAL UNCERTAINTIES

Uncertainty	Impact if Wrong	Resolution Method
Number and capability of operating NZ foundries	Cannot plan expansion if baseline is unknown	National census (Doc #8) — foundry as specific category
NZ moulding sand deposit quality and accessibility	If NZ sand is unsuitable, moulding becomes the constraint	Geological assessment and test casting using NZ sand
NZ bentonite clay availability for sand bonding	Without adequate clay binder, mould quality is poor	Geological survey; test NZ clays for foundry properties
Scrap iron, bronze, and aluminium stocks in NZ	Determines how long casting can continue from existing materials	National scrap inventory (part of census, Doc #8)
Induction furnace remaining service life	Electronics failure stops electric melting	Condition assessment of existing foundry induction furnaces
Ferrosilicon stocks in NZ	Without ferrosilicon, cupola iron quality degrades (white iron tendency)	Inventory existing stocks; assess NZ silicon production feasibility
Foundry workforce age and skill level	If fewer experienced founders than assumed, expansion is slower	Census skills survey
Charcoal quality from NZ timbers for cupola use	Cupola performance depends on fuel quality	Testing — produce charcoal from available NZ species and trial in cupola operation

---

CROSS-REFERENCES

• Doc #1 — National Emergency Stockpile Strategy (foundry consumable requisition, critical-skills personnel)
• Doc #8 — National Skills and Asset Census (foundry workforce, equipment, and scrap inventory)
• Doc #56 — Wood Gasification (gasifier construction requires cast components; charcoal production for cupola fuel)
• Doc #65 — Hydroelectric Station Maintenance (valve bodies, gate components — major casting demand)
• Doc #74 — Pastoral Farming (agricultural equipment repair generates casting demand)
• Doc #89 — NZ Steel Glenbrook (steel supply for foundry construction; plate for cupola shells; coordination on refractory development)
• Doc #91 — Machine Shop Operations (complementary capability — castings require finish machining; machine shops need foundry for pattern tooling and machine tool castings)
• Doc #92 — Blacksmithing and Forge Work (complementary metalworking; blacksmiths produce foundry tools)
• Doc #94 — Welding Consumable Fabrication (welding for foundry equipment construction and casting repair)
• Doc #97 — Cement Production (refractory cement for furnace linings)
• Doc #105 — Wire Drawing (wire drawing dies may require cast components)
• Doc #113 — Sulfuric Acid (industrial chemistry for some refractory and investment casting materials)
• Doc #138 — Sailing Vessel Design (marine hardware — cast bronze fittings; trade route for copper/tin imports)
• Doc #157 — Accelerated Trade Training (foundry as a training priority trade)
• Doc #160 — Heritage Skills Preservation and Transmission (patternmaking, hand moulding as heritage trade skills)

---

---

---

1. NZ foundry industry: The NZ foundry sector is small and not comprehensively documented in publicly available sources. The Maintenance, Engineering and Manufacturing Association (MEMA) and the Employers and Manufacturers Association (EMA) represent some NZ foundry operators. The Australasian Foundry Institute (AFI) is the regional professional body for the foundry industry. Specific NZ foundry company information is based on business directory listings, industry publications, and the Australian & New Zealand Foundry Directory. The current operational status of individual foundries requires direct verification.↩︎

2. NZ silica sand deposits suitable for foundry use are documented in GNS Science (Institute of Geological and Nuclear Sciences) mineral resource publications. See: GNS Science, NZ Mineral Occurrence Database. https://www.gns.cri.nz/ — Specific suitability for foundry moulding requires assessment of grain size distribution, grain shape, and contaminant levels, which varies by deposit.↩︎

3. Foundry establishment effort: The estimate of 2–4 person-years assumes adaptation of an existing industrial building, fabrication of basic equipment from available materials, and training a small initial workforce. A purpose-built facility on a greenfield site would require more effort. The estimate does not include the furnace itself (which is a separate project — Section 7 for cupola construction).↩︎

4. Foundry breakeven estimate: The 1–3 year breakeven range is an estimate based on the establishment cost of 2–4 person-years (note 23) weighed against the value of locally produced castings for equipment maintenance. The lower end assumes high demand for replacement castings (agricultural region with aging machinery); the upper end assumes lower demand or a longer ramp-up to consistent production quality. The estimate is illustrative rather than precise — actual breakeven depends on regional equipment stocks, failure rates, and the availability of alternative repair methods (welding, fabrication from solid stock).↩︎

5. Vibration damping of grey iron vs. steel: Grey cast iron’s damping capacity is approximately 20–25 times that of steel, due to the graphite flakes acting as internal friction surfaces. This property is the primary reason machine tool builders traditionally specify cast iron for precision machine bases. See: ASM International, “ASM Metals Handbook Volume 1: Properties and Selection — Irons, Steels, and High-Performance Alloys”; Bickford, J.H., vibration damping data in machine tool design references.↩︎

6. NZ timber for patternmaking: Radiata pine is abundant and workable but has open grain that requires sealing. Native timbers (kauri, rimu) have finer grain and better dimensional stability but are less available. Medium-density fibreboard (MDF), while not a NZ raw material (MDF is manufactured in NZ from radiata pine fibre), is widely used in NZ foundry patternmaking and stocks exist. See general patternmaking references; NZ timber properties in: NZ Wood, timber species guides. https://www.nzwood.co.nz/↩︎

7. Shrinkage allowances for casting alloys are standard foundry engineering data. Values cited are typical; exact allowances depend on the specific alloy, casting geometry, and mould rigidity. See: Heine, R.W., Loper, C.R., and Rosenthal, P.C., “Principles of Metal Casting,” McGraw-Hill, various editions; ASM International, “ASM Metals Handbook Volume 15: Casting.”↩︎

8. Green sand composition and properties: Standard foundry practice. See: American Foundry Society (AFS) publications; Heine, Loper & Rosenthal (note 4). The coal dust addition (sea coal) is typically 2–5% by weight of the sand mix and is used specifically for iron casting — it creates a reducing gas atmosphere that prevents iron oxide formation on the casting surface.↩︎

9. NZ silica sand deposits: Known deposits include Parengarenga Harbour (Northland — high-purity silica sand, currently exported for glass manufacture), deposits in the Waikato region, and various Bay of Plenty locations. The suitability of each deposit for foundry use depends on specific grain characteristics that must be tested. See: GNS Science mineral resource data; Crown Minerals NZ. https://www.nzpam.govt.nz/↩︎

10. NZ bentonite deposits: Some bentonite deposits are known in NZ, though production has been limited. The quality and quantity of NZ bentonite for foundry use requires geological assessment. Bentonite is also imported from Australia for various industrial applications (drilling mud, wine fining, cat litter). See: GNS Science mineral occurrence data.↩︎

11. Chemical binder sand systems (furan, phenolic urethane, alkyd) are petrochemical products manufactured outside NZ. See: American Foundry Society, “Mold & Core Test Handbook”; Brown, J.R., “Foseco Ferrous Foundryman’s Handbook,” Butterworth-Heinemann. Reversion to green sand is a straightforward process change but requires more skill in mould making and accepts somewhat lower dimensional accuracy.↩︎

12. Traditional core binders (linseed oil and flour, baked to harden) are described in older foundry texts. See: Walton, C.F. and Opar, T.J., “Iron Castings Handbook,” Iron Castings Society, 1981; various historical foundry practice references. Linseed oil is producible in NZ from flax seed (Linum usitatissimum, not to be confused with NZ native flax/harakeke).↩︎

13. Gating and risering design: The science and art of designing the runner system and risers for a casting is extensively documented. Key references include: Heine, Loper & Rosenthal (note 4); Beeley, P., “Foundry Technology,” Butterworth-Heinemann; Chvorinov’s rule for solidification time calculation. Modern approaches use computer simulation, but the principles can be applied using hand calculation and experience.↩︎

14. Pouring temperatures are standard foundry data. Values cited are typical ranges; exact pouring temperature depends on the specific alloy, casting section thickness, and mould material. See: ASM Metals Handbook Volume 15 (note 4); AFS publications.↩︎

15. Grey cast iron composition and properties: Standard metallurgical data. The iron-carbon phase diagram shows that carbon content above approximately 2.0% produces “cast iron” (as distinct from steel) with dramatically different properties — lower melting point, better fluidity, and the formation of graphite flakes that give grey iron its characteristic appearance and properties. See: any introductory metallurgy text; ASM International, “ASM Metals Handbook Volume 1: Properties and Selection.”↩︎

16. Identifying cast iron in scrap: Grey cast iron fractures with a grey, granular appearance (visible graphite flakes). It is brittle — it breaks rather than bends when struck. It is magnetic. It sparks with short, dull red sparks on a grinding wheel (distinct from steel’s longer, brighter sparks). These identification methods are standard scrap-sorting practice.↩︎

17. Ductile (nodular, spheroidal graphite) iron: Developed in the late 1940s, ductile iron has largely replaced steel castings and forgings for many structural applications due to its combination of castability and mechanical properties. The magnesium treatment that converts graphite flakes to nodules was a transformative metallurgical development. See: Ductile Iron Society publications; ASM Metals Handbook.↩︎

18. NZ copper deposits: Small copper occurrences are known in NZ, primarily in the Coromandel Peninsula, West Coast, and Southland. These have been intermittently mined on a small scale but NZ has never been a significant copper producer. See: GNS Science; Crown Minerals NZ.↩︎

19. Tiwai Point aluminium smelter: See Doc #89, footnote 43. The smelter produces aluminium from imported alumina. Alumina stocks at the time of the event determine the duration of primary aluminium production.↩︎

20. Foundry induction furnace sizes: Medium-frequency coreless induction furnaces for foundry use are manufactured by companies including Inductotherm, ABP Induction, and others. Typical NZ foundry installations use furnaces in the 250 kg to 2,000 kg (per heat) range. Exact specifications of NZ-installed furnaces require direct survey.↩︎

21. Refractory crucibles for induction furnaces: Crucibles are consumable items with a finite campaign life (typically 50–200 heats for silica-based crucibles used with iron, longer for alumina-based crucibles). NZ’s silica sand and fireclay can produce basic crucibles, though their service life may be shorter than imported high-performance crucibles. See: Inductotherm crucible selection guides; refractory engineering references.↩︎

22. Power electronics service life: Thyristor module life depends on junction temperature cycling, typically rated for 15–30 years in industrial applications under managed thermal conditions. Power capacitors (used in the resonant tank circuit of medium-frequency induction furnaces) degrade with thermal cycling and voltage stress, with typical service lives of 10–20 years. Both figures assume operation within rated conditions; overloading, poor cooling, or power quality issues reduce life significantly. See: Inductotherm and ABP Induction technical service documentation; general power electronics reliability literature.↩︎

23. Cupola furnace history and operation: The cupola has been the primary foundry melting unit for iron since its development in the 18th century. See: Kirk, E., “Cupola Furnace — A Practical Treatise,” Baird, various editions (historical but informative); AFS, “Cupola Handbook,” American Foundry Society; Heine, Loper & Rosenthal (note 4).↩︎

24. Charcoal for cupola operation: Charcoal was the original cupola fuel before coke became dominant in the 19th century. Charcoal-fuelled cupolas produce iron with lower sulfur content (an advantage) but lower carbon pickup and somewhat different thermal characteristics. Charcoal production from NZ timber is well-established in principle — see general charcoal production references. Hardwood charcoal is preferred for metallurgical use due to its density and compressive strength; NZ native hardwoods (beech, pohutukawa, rata) are better suited for cupola charcoal than radiata pine, which produces lower-density charcoal that is more prone to crushing under furnace charge weight. Radiata pine charcoal can serve but requires larger piece sizes and adjusted charge ratios.↩︎

25. Charcoal vs. coke in cupola operation: Charcoal has lower sulfur (~0.01%) compared to coke (0.5–1.0%+), producing cleaner iron. However, charcoal has lower compressive strength and can crush under the weight of the furnace charge, affecting gas flow. Using larger charcoal pieces and adjusting the charge ratio helps manage this. See: historical foundry practice literature; Kirk (note 19).↩︎

26. Cupola construction timeline: The 4–8 week estimate assumes a fabrication team with oxy-acetylene or electric arc welding capability, steel plate available from stock or from NZ Steel (Doc #89), and fireclay brick available from existing stocks or NZ clay deposit. The fireclay brick production timeline (2–4 weeks if produced from raw clay) is based on the time required for clay mining, brick forming, drying, and kiln firing — a well-understood process with NZ materials. See: general refractory engineering references; NZ clay deposit data from GNS Science.↩︎

27. NZ foundry industry: The NZ foundry sector is small and not comprehensively documented in publicly available sources. The Maintenance, Engineering and Manufacturing Association (MEMA) and the Employers and Manufacturers Association (EMA) represent some NZ foundry operators. The Australasian Foundry Institute (AFI) is the regional professional body for the foundry industry. Specific NZ foundry company information is based on business directory listings, industry publications, and the Australian & New Zealand Foundry Directory. The current operational status of individual foundries requires direct verification.↩︎

28. NZ foundry workforce size: This estimate is based on the small number of known operating foundries and typical staffing levels. The actual number could be higher if in-house casting operations within engineering firms are included, or lower if some listed foundries have further contracted. Census verification is required.↩︎

29. Foundry building size: The 200–500 m² range is based on typical small-to-medium jobbing foundry layouts that accommodate a sand preparation area, moulding floor, furnace and pouring area, fettling area, and pattern storage. Larger operations (particularly those with cupola furnaces) typically require 500–1,000+ m². See: Brown, J.R., “Foseco Ferrous Foundryman’s Handbook,” Butterworth-Heinemann — foundry layout guidance; AFS publications on foundry design.↩︎

30. Foundry staffing: A small jobbing foundry producing sand castings typically operates with 4–8 people: 1 furnace operator, 2–4 moulders, 1 patternmaker (often part-time or shared), and 1–2 labourers for sand preparation, shakeout, and fettling. This estimate is based on historical small foundry staffing patterns and modern NZ small foundry practice. See: AFS publications; Brown (note 40).↩︎

31. Cupola refractory lining: Fireclay brick is the traditional lining material for small cupolas. Lining life varies from several campaigns (weeks of actual operation) to many months, depending on the refractory quality and the severity of operating conditions (temperature, slag chemistry). Relining involves removing the spent lining and laying new brick — a straightforward but labour-intensive process. See: AFS Cupola Handbook (note 19).↩︎

32. Cupola charge ratios and flux additions: Standard cupola operating practice. The 8:1 to 10:1 iron-to-coke ratio is typical for coke-fuelled cupolas; charcoal-fuelled cupolas may require different ratios (typically more charcoal, as charcoal has lower energy density than coke). Limestone flux at 2–4% of iron charge is standard for sulfur removal. See: AFS Cupola Handbook; Heine, Loper & Rosenthal (note 4).↩︎

33. Cupola output rates: A cupola’s melting rate depends on its diameter, the blast rate, and the coke ratio. Published data for a 750 mm internal diameter cupola suggest melting rates of approximately 1–3 tonnes per hour. These figures are from standard cupola engineering references and should be validated by trial operation with NZ charcoal, as charcoal-fuelled operation differs from the coke-fuelled data most commonly published. See: AFS Cupola Handbook (note 19).↩︎

34. Silicon control in cupola iron: Silicon is oxidised during cupola melting, typically losing 10–30% of the charged silicon content. Without ferrosilicon ladle additions, the resulting iron may have insufficient silicon for grey iron solidification, producing “white iron” — hard, brittle, and unmachineable. Managing this by selecting high-silicon scrap and optimising furnace conditions is possible but requires skill and experience. See: Heine, Loper & Rosenthal (note 4); AFS publications on cupola metallurgy.↩︎

35. Beeswax for lost-wax casting: Beeswax has been used for lost-wax (investment) casting for thousands of years. NZ has a domestic apiculture industry and beeswax is a byproduct of honey production. Beeswax is typically blended with rosin (pine resin) or paraffin to adjust its working properties — harder wax holds shape better during mould making. See: general investment casting references; Horton, R.A., “Investment Casting,” ASM International.↩︎

36. Plaster mould casting: Gypsum plaster (plaster of Paris) mixed with silica sand or other filite produces moulds suitable for aluminium and copper alloy casting. Not suitable for iron (the plaster decomposes above approximately 1,200°C). Gypsum is mined in NZ, making this a fully NZ-producible process for non-ferrous castings. See: ASM Metals Handbook Volume 15 (note 4).↩︎

37. Metal patterns for production foundry work: For parts that will be cast repeatedly (pipe fittings, standard bearing housings), the initial wooden pattern should be used to cast a metal pattern — typically in aluminium — which is then used for all subsequent moulding. Metal patterns are more dimensionally stable, more durable, and produce more consistent moulds than wooden patterns. See: standard patternmaking references.↩︎

38. Plain bearing vs. rolling element bearing performance: Plain (journal) bearings have higher friction coefficients (0.01–0.10 depending on lubrication conditions, compared to 0.001–0.005 for ball bearings), require continuous lubrication, and have lower speed limits. However, they are simpler, quieter, have high load capacity relative to size, and can be produced locally from cast bronze — which makes them the appropriate bearing technology for an import-constrained economy. See: standard bearing engineering references; Shigley, J.E. and Mischke, C.R., “Mechanical Engineering Design,” McGraw-Hill.↩︎

39. Foundry engineering references: The essential texts for NZ foundry operations are: Heine, R.W., Loper, C.R., and Rosenthal, P.C., “Principles of Metal Casting,” McGraw-Hill (the standard academic text); ASM International, “ASM Metals Handbook Volume 15: Casting” (comprehensive reference); American Foundry Society publications (practical handbooks); Brown, J.R., “Foseco Ferrous Foundryman’s Handbook” and “Foseco Non-Ferrous Foundryman’s Handbook,” Butterworth-Heinemann (practical industrial references). Securing multiple copies of these references for NZ foundry operations is a knowledge-preservation priority.↩︎

40. Non-destructive testing (NDT) for castings: Magnetic particle inspection (MPI), dye penetrant testing (DPT), ultrasonic testing (UT), and radiographic testing (RT) are standard NDT methods for casting inspection. NZ has NDT inspection companies and qualified inspectors. NDT consumables (magnetic particles, penetrant fluids, coupling gels) are imported with finite stocks. See: American Society for Nondestructive Testing (ASNT) publications; NZ NDT inspection industry.↩︎

41. A & G Price, Thames: One of NZ’s most significant historical engineering and foundry operations. Established 1868, the company produced railway locomotives, marine engines, gold dredge components, and general engineering castings. It operated continuously for well over a century, serving as a major centre of NZ foundry and heavy engineering skill. See: NZ Heritage / Heritage New Zealand records; Thames historical society publications; Hearn, T.J., various publications on NZ engineering history.↩︎

42. NZ Railways workshops: The NZ Railways Department operated major engineering workshops at Hutt (Lower Hutt), Addington (Christchurch), and Hillside (Dunedin), among others. These workshops included foundry operations for casting locomotive and rolling stock components. The Hillside workshops in Dunedin were particularly notable for their casting capability. Most of these operations have been closed or dramatically reduced. See: NZ Railway and Locomotive Society publications; Churchman, G. and Hurst, T., “The Railways of New Zealand: A Journey Through History.”↩︎