Can You Weld Cast Iron? What to Assume When You Don't Know

Yes, cast iron can be welded, and grey and ductile iron castings are repaired by arc welding every working day using nickel filler metals. Whether your casting should be welded is a separate question with a separate answer. Cast iron names a family of iron alloys carrying roughly 2–4% carbon, about ten times the carbon of most steels, according to TWI's guidance on the subject, and weldability inside that family runs from routine down to nil: TWI describes white iron as "generally considered as unweldable." When the grade, the service duty and the true extent of the crack are all unknown, the safest assumption is that your part is not a fusion-welding candidate until you prove it is. Braze welding and cold mechanical stitching, which put far less heat and no dilution into the casting, are what you fall back to in the meantime.

Why one cast-iron welding recipe fails

The advice that dominates search results is one ritual: preheat hot, run nickel rod, peen every bead, bury the part in sand. It works often enough on grey iron to survive as folklore. The document that governs this work refuses to write it that way.

AWS D11.2/D11.2M:2025, Guide for Welding Iron Castings, second edition, approved by ANSI on 17 September 2025, opens its scope with a sentence worth reading slowly: "The term cast iron encompasses a family of ferrous alloys with a variety of metallurgical, mechanical, and physical properties. The chemical composition and welding requirements of the metals vary within the family."

I spent years translating appliance safety manuals, where the distance between should be tested monthly and must be tested monthly decides whether a sentence is advice or an obligation. D11.2 says plainly which register it uses: "Although this guide is not written with mandatory requirements, mandatory language, such as the use of 'shall', will be found in those portions of the document where failure to follow the instructions or procedures could produce inferior, misleading or unsafe results."

The hedging reflects metallurgy. The guide's micrographs show five carbon structures under the single word iron: flake graphite in grey iron, iron carbide in white iron, temper carbon nodules in malleable iron, spherical nodules in ductile iron, and the mixed form of compacted graphite iron. A procedure tuned to flakes has no claim on carbide.

Which cast iron are you actually holding?

You can narrow it without a laboratory. Grey iron breaks with a dull, faintly sooty fracture face, because you are looking at exposed graphite flakes. White iron fractures bright and crystalline. Ductile iron behaves closer to steel under a file and bends a little before it lets go; grey iron will not. Part identity helps as much as any test: blocks, heads, manifolds and machine housings are usually grey; crankshafts, hubs, steering knuckles and gearbox internals are frequently ductile iron to ASTM A536; small forged-looking fittings are often malleable iron to ASTM A220.

None of that is chemical analysis, and the honest description of it is guessing with better odds.

The number that would settle it is carbon equivalent, conventionally calculated as CE = %C + (%Si + %P) ÷ 3, with 4.3% marking the eutectic composition. D11.2 devotes a whole figure to the "Relation of No-Crack Temperature to C.E. #2 for Several Grades of Iron Castings": the temperature below which a grade cracks when welded is read off carbon equivalent, not off appearance. Computing it needs a spectrometer reading or the foundry certificate. Lacking both, assume high carbon equivalent, low ductility and high restraint, and let every later decision inherit that.

The repairability check that comes before the electrode

Grade is one of five inputs, and electrode choice sits downstream of all of them.

Contamination comes first because it is the one you can fix. A casting that ran in oil for thirty years holds oil inside its graphite network, and welding heat drives it out into the weld pool as porosity. Degrease, then bake the part above boiling point for several hours; a wash never reaches the graphite.

Restraint decides how much shrinkage stress the base metal must absorb. A crack across a free-standing bracket arm can move as it cools; the same crack in a closed engine block cannot. Section thickness sets thermal behaviour, thick sections cooling unevenly while thin ones distort. Consequence of failure sets your tolerance for all of it: a cracked vise body drops a workpiece, a cracked steering knuckle does something else entirely.

Then map the crack. Dye penetrant, cleaned and re-applied, usually shows it running past the visible line, and stop-drilling each true end arrests propagation while you work. Assume it is longer than it looks until the penetrant says otherwise.

Welding, braze welding or metal stitching?

The real comparison is not MIG against stick. It is fusing the casting, sticking something to it, or bolting it back together cold.

| Method | Heat into the casting | What carries the load | Machinable after | Best case | Where it fails | |---|---|---|---|---|---| | Arc welding, nickel filler | Highest; local melting plus preheat to 1,200 °F | Fused weld metal, diluted with base iron | Yes with ENi-CI; carbide tooling with ENiFe-CI | Known grade, clean metal, low restraint | Unknown grade, oil-soaked iron, white iron | | Braze welding, bronze filler | Moderate; filler melts at 1,595–1,630 °F, base metal never melts | Adhesion at the interface, plus joint geometry | Yes, readily | Thin or intricate grey iron; parts you cannot preheat evenly | Loads above the bronze's strength; heat above its melting range | | Metal stitching | None | Interlocking pins and locks in tapped holes | Yes | Large or immovable castings; engine blocks | Tight access; thin walls without thread depth | | Replacement | None | The new part | n/a | Safety-critical and pressure-retaining parts | Obsolete castings with no supply |

Braze welding earns its place because the filler melts below the casting. AWS RBCuZn-C low-fuming bronze melts at roughly 1,595–1,630 °F, several hundred degrees under grey iron, so the base metal is heated to red and wetted rather than melted. TWI notes the consequence: minimal impact on the base metal, because the filler adheres instead of diluting. Aufhauser specifies a preheat of 320–400 °C (610–750 °F), well under what fusion welding the same part demands. RBCuZn-D nickel silver is the stronger of the two common braze fillers.

Metal stitching is the option most home mechanics have never seen, though it dates to the late 1930s and the Texas oil fields. Lock-N-Stitch, founded in Turlock, California in 1990 and now part of Wärtsilä, refined it with threaded pins whose thread form draws the crack closed instead of wedging it open, overlapped along the crack and cross-tied by locks set at 90 degrees. No heat enters the part, so no distortion, no heat-affected zone and no new residual stress.

How to control heat if you do weld it

Lincoln Electric frames the choice as two incompatible routes: weld with preheat and plenty of it, or keep the casting cool. Once you select a method, stick with it.

The hot route means preheating the whole casting, typically to 500–1,200 °F, never above 1,400 °F, which enters the critical temperature range. Sodel's procedure narrows the band to a 1,000–1,200 °F preheat with interpass temperature held above 600 °F and below 1,250 °F, and permits beads of 3 to 5 inches (75–125 mm).

The cool route holds the part near 100 °F, and Lincoln offers a usable field test: never heat the casting so hot you cannot rest a bare hand on it. Beads get short. Lincoln calls for roughly 1-inch segments, Sodel specifies 3/4 to 2 inches (20–50 mm), Aufhauser says 50 mm at a time. Across three manufacturers' procedures maximum bead length runs from 3/4 inch to 5 inches, and which end applies depends on the route. There is no universal inch rule.

Preheating only around the crack is the trap that catches careful people. Lincoln's crack-repair guidance is specific: preheat the entire casting, not just the area to be welded, and hold that from start to finish. A hot patch inside a cold brittle casting manufactures the exact thermal gradient preheating exists to remove.

Filler is chosen on dilution and ductility, never on which machine sits in the corner of your shop. AWS A5.15 sets ENi-CI at 85% nickel minimum, sold as Nickel 99 and by Lincoln as Tech-Rod 99; its deposits stay soft and machinable even on single-pass, high-dilution repairs. ENiFe-CI is specified at 45.0–60.0% nickel with iron as the balance, sold as Nickel 55, reaching about 84 ksi tensile with better crack resistance for multi-pass work and ductile iron, at the cost of carbide tooling afterwards. Sodel adds a step worth copying: butter the groove faces with pure nickel first, then fill with ferronickel, so the highest-dilution metal sits in the softest deposit.

What peening and slow cooling actually do

Peening compresses hot weld metal so it shrinks into less tension. TWI attaches a condition generic advice drops: peening should only be used with relatively ductile weld metal, meaning welds made with nickel consumables. Striking a hard steel-electrode deposit adds stress to brittle metal. Sodel specifies a rounded tool while the bead is above 1,000 °F and warns against pointed tools, whose impressions become crack starters.

Slow cooling is where "bury it in sand" needs a number attached. Sodel states that cooldown after hot welding should never exceed 30–55 °F (15–30 °C) per hour. Take a casting from a 1,100 °F interpass down to a handleable 200 °F and that rate implies 16 to 30 hours under insulation. The shop rule of 24 hours buried sits inside that arithmetic, which is the strongest thing you can say for it. Dry sand, a welding blanket or vermiculite slow cooling by insulating; none measures the rate. What you can control is the negative: no water, no compressed air, no draught, and never a hot casting on cold concrete.

How to tell whether the repair will hold

A bead that holds coolant is not evidence of a structural repair. Porosity under the surface, residual stress locked in during cooling and cracking at the edge of the heat-affected zone announce themselves later, when the part returns to load or thermal cycling.

Inspect after the casting reaches ambient, never while it is warm, because heat-affected-zone cracks can appear hours into cooling. Dye penetrant across the weld and 25 mm either side, or magnetic particle if you have it, shows what visual inspection misses. D11.2 carries a normative annex classifying welds in iron castings by quality level, with acceptance criteria for those tests and surveillance recommendations for repairs.

Machinability doubles as a diagnostic. If an ENi-CI deposit will not take a file or a cutter, dilution was high and hard martensite formed in the weld metal or beside it. Heat control failed, whatever the surface looks like.

The castings to leave alone

White iron, identified by its bright crystalline fracture, is treated as unweldable by TWI, and no home procedure changes that. Safety-critical parts belong to the same category by a different route: braking, steering, lifting and pressure-retaining components carry consequences no unqualified repair should be asked to hold. D11.2's procedure qualification apparatus exists for parts of that kind.

Engine blocks cracked through into a water jacket are the classic case for stitching rather than welding: the restraint is total and the casting cannot be evenly preheated on a bench. A manifold cycles from ambient to over 1,200 °F every time the engine runs, so a repair that passes inspection on Saturday is re-tested every morning after, and a warped flange face will leak whatever the crack repair does.

The casting most worth walking away from is the one whose grade you cannot establish and whose failure you cannot afford. Two of the four routes above, brazing and stitching, hold that part without asking you to name its grade.

Frequently asked questions

What happens when cast iron is welded?

The base metal melts and carbon from the casting dissolves into the weld pool. Rapid cooling traps that carbon as brittle martensite and iron carbide in the weld and heat-affected zone, which then crack under shrinkage stress. Preheat, short beads, peening and slow cooling all exist to fight that.

What is the best welding method for cast iron?

There is no single best method. For known grey or ductile iron with good access, stick welding with a nickel electrode (ENi-CI or ENiFe-CI) plus full-casting preheat gives a machinable, load-bearing repair. For unknown grades or parts you cannot preheat evenly, braze welding or metal stitching is safer.

How can broken cast iron be repaired?

Four routes exist: arc welding with nickel filler, braze welding with bronze filler melting at 1,595–1,630 °F, cold metal stitching with interlocking threaded pins, and replacement. Choose by casting grade, contamination, restraint and consequence of failure. Stitching and brazing put far less heat into the part.

Can cast iron be welded to steel?

Yes. Nickel-based fillers are specified for this joint because they tolerate the carbon picked up from the cast iron side while staying ductile. ENi-CI and ENiFe-CI electrodes are both classified for welding grey iron castings to mild or stainless steel. Preheat the cast iron side and keep dilution low.

Can cast iron be MIG welded?

Yes, using nickel-based wire such as ERNiFe-CI rather than ordinary ER70S-6 steel wire, which absorbs carbon and cracks. AWS D11.2 lists gas metal arc welding variables for nickel-base electrodes. Expect a specialist shielding gas mix and 1 to 2 lb spools against steel wire's 33 lb.

When is metal stitching better than welding?

When heat is the problem. Stitching adds none, so no distortion, no heat-affected zone and no new residual stress. It wins on large or immovable castings, engine blocks and heads, unknown grades, oil-contaminated iron, and any part you cannot preheat evenly.