Cast iron or welded steel for a machine base

The question comes up on every new machine programme and it is usually argued on the wrong axis. People debate which material is better. The useful question is which one is better at your production volume, for your accuracy class, on your schedule — because the answer genuinely flips depending on those three. Material is one of six decisions in our guide to machine frame design.

What each material is actually good at

Grey cast iron earns its place in machine tools through damping. The graphite flakes in the matrix dissipate vibration energy as internal friction, and the effect is large — grey iron damps roughly an order of magnitude better than a comparable steel section. In a structure where the limit on surface finish is chatter rather than deflection, that matters more than anything else on the datasheet.

Steel wins on specific stiffness. Young’s modulus for structural steel is about 200 GPa against roughly 100–130 GPa for grey iron, so per kilogram of material in the right place, steel is roughly twice as stiff. A properly boxed and ribbed weldment can be lighter than the iron casting it replaces and still deflect less under the same load.

So the honest summary: iron damps, steel is stiff. Every real decision is about which of those you are short of.

Where damping stops being the deciding factor

The damping argument is strongest for machines whose limit is chatter in interrupted or heavy cuts — lathes, boring machines, older-generation milling machines with long unsupported spans. It weakens considerably for:

• Machines with high natural frequencies. A stiff, well-boxed weldment pushes its first mode above the excitation frequencies the process actually produces. Damping matters most when you have a mode sitting where the cutting forces live; move the mode and the problem moves with it. This is covered in more depth in natural frequency and chatter in machine structures.
• Machines where joints provide the damping. Damping in a weldment concentrates at the welds and at bolted interfaces. A structure with well-designed joints and a bolted, preloaded interface can approach cast iron in practice even though the bulk material does not.
• Filled weldments. Sand, concrete or epoxy-aggregate fill inside the box sections adds damping to a steel structure cheaply. It is common on gantry crossrails and column bases and it closes much of the gap. The full comparison against a purpose-cast polymer structure is in mineral casting versus welded steel.

The volume argument, which usually decides it

Cast iron carries a pattern. For a machine base that means a wooden or resin pattern, core boxes, and a foundry willing to run it — typically several thousand dollars and six to ten weeks before the first casting exists. After that, each additional casting is cheap and identical.

A weldment carries no tooling cost. Cutting, fitting and welding start the week the drawing is released. Each additional unit costs roughly what the first one did, because labour does not amortise the way a pattern does.

The crossover is real and reasonably predictable:

Annual volume	Usual answer
1–5	Weldment, almost always
5–25	Weldment unless damping is the binding constraint
25–100	Genuinely contested; run both quotes
100+	Casting, unless the part is too large to cast economically

Size shifts the crossover upward. Above about 3 metres in the longest dimension, foundries capable of the pour thin out quickly, freight for a single heavy casting gets expensive, and the weldment stays competitive at volumes where a smaller part would have gone to iron long ago.

Lead time and design freedom

A weldment can be redesigned between units. If the first machine reveals that the column wants another 20 mm of wall or a rib in a different place, the change costs a drawing revision and a week. The same change on a casting costs a pattern modification and another foundry cycle.

For a machine still finding its final form — a prototype, a pilot batch, a design in its first two years — that flexibility is usually worth more than the damping difference. Several builders run welded frames through the first 20 or 50 machines and switch to iron only once the geometry stops moving. That is a sensible sequence, not a compromise.

Where the weldment route can go wrong

Two failure modes account for most of the disappointment with welded machine bases, and both are process problems rather than material problems.

Residual stress left in the structure. Welding locks stress into the part. Machine a face without relieving it and the structure relaxes over the following months. The geometry you inspected and accepted quietly walks away. This is the single most common reason a welded base underperforms an iron one in service, and it is entirely avoidable — see stress relieving welded machine bases.

Distortion during fabrication. A poorly sequenced weld on a long frame can pull it out of plane by several millimetres, and the machining allowance you left will not cover it. Sequence, fixturing and restraint decide this; welding distortion control on large weldments sets out the levers.

A shop that handles both of these competently produces welded bases that hold geometry for decades. A shop that skips them produces bases that pass FAI and fail in year two. The difference is not visible in the quote.

A decision sequence that works

1. Identify the binding constraint. Is the machine limited by chatter, by static deflection, by thermal growth, or by cost? Only the first favours iron strongly.
2. Count the units over the next three years, honestly, including the ones marketing is promising.
3. Check the size against the foundry map. If the part exceeds roughly 3 m or 5 t, the list of foundries that can pour it — and ship it — shortens fast.
4. Ask whether the design is frozen. If it is not, the weldment buys you the right to change your mind.
5. If iron still leads on damping but loses on volume, price a filled weldment before conceding. Fill is cheap and it moves the damping number a long way.

Most machine builders working in the 1–50 unit range land on a welded, stress-relieved, finish-machined structure — and land there for schedule and tooling-cost reasons rather than because the metallurgy demanded it. That is a perfectly sound engineering answer, provided the weldment is specified and processed properly.

If you have a base design that could go either way and want the fabricated route priced properly before you commit to a pattern, send the drawing through — the review comes back within 24 business hours and includes whether the geometry is sensible to weld at all. The full capability envelope sets out the sizes the network handles.