How the alloy is chosen before the pattern is cut, what changes when the same grade is written in five national standards, and which substitutions fail in service.
Choosing a casting alloy, and reading a grade you were not trained on
Most material arguments at the quoting stage are not about metallurgy. They are about naming. The drawing says ZG10Cr18Ni12Mo2, the purchase order says CF8M, the inspector’s sheet says 1.4408, and somebody has to confirm those three describe one part. This page is how we work through that, and how we pick an alloy when the drawing states a duty instead of a grade.
One rule comes first. The nearest equivalent is not the same grade. GB, ASTM, EN and JIS each write their own composition window, their own test bar position and their own acceptance rules. In general valve and pump work the swap usually holds. In pressure vessels, marine class or anything with a notified body, it does not, and the difference is almost always in the test rather than the chemistry.
Start from what the part touches, not from the grade name
A grade name is shorthand for a set of properties. Work backwards from the duty and the shortlist writes itself. The order of questions below is the order we ask them in, and it is also the order in which a wrong answer gets expensive.
| Duty | First choice | Second look | What rules it out |
|---|---|---|---|
| Potable water, food contact | CF3M | CF8M below 60 °C | residual lead and sulphur from scrap in the melt; free-machining grades are out |
| Seawater, brackish water, chlorides | CD3MN (2205) | CD3MWCuN (2507), C95800 bronze | CF8M pits in stagnant chloride above roughly 60 °C |
| Dilute sulphuric and phosphoric acid | CW-12MW | CD4MCuN | the 304 and 316 families are limited to dilute, aerated acid |
| Steam and hydrogen service, 400–550 °C | WC6, WC9 | CF8C where the part must be stainless | carbon steel creeps; CF8M forms sigma phase above about 425 °C |
| Abrasion with impact | CA15, CA40 | CA6NM where it must also weld | austenitic grades work-harden and gall instead of resisting wear |
| Cryogenic, to −196 °C | CF3M, CF8 | A356 for light parts | martensitic and ferritic grades go brittle |
| Structural, painted, cheapest per kilo | WCB, ZG230-450 | QT450-10 where damping matters | corrosion allowance and weight both get worse |
| Furnace fixtures above 900 °C | HK40 | CU5MCuC in carburising atmospheres | any grade with free carbon oxidises and scales at the section edge |
Where a duty sits between two rows, the deciding factor is usually temperature, not chemistry. Chloride pitting potential falls as the metal gets hotter, and a grade that lasts ten years in a cold line can fail in a heat exchanger loop in eighteen months.
What actually changes when you cross standards
Six things move when a grade is re-written in another system. None of them show up on a one-line description, and all of them have cost somebody a rejected lot.
- A window, not a number. CF8M allows carbon to 0.08 % and molybdenum from 2.00 to 3.00 %. EN 1.4408 caps carbon at 0.07 % and holds molybdenum between 2.0 and 2.5 %. A heat at 0.075 C and 2.9 Mo is legally CF8M and not legally 1.4408. That is a real heat, not a hypothetical one.
- Where the test bar came from. ASTM takes tensile from a separately cast keel bar; EN 10213 uses a separately cast test block of different dimensions; GB has its own. Two heats that both pass can report numbers several percent apart, and neither supplier is cheating. Ask which bar, cast when, and from which ladle.
- Impact testing. LCB, LCC and CA6NM are frequently ordered with Charpy values at −46 °C. A WCB certificate carrying room-temperature tensile only does not cover that requirement, and the gap is discovered at the witness point, not at the quote.
- Heat treatment condition. “CA15” on its own means nothing in service. It needs austenitising and tempering to a stated hardness, and as-cast CA15 is not the material the designer calculated. If the condition is missing from the drawing, put it in the purchase order.
- Non-destructive testing level. “PT inspected” is not a specification. Name the standard, the acceptance level, the surface coverage and who signs the report. Two foundries quoting dye penetrant can be quoting two different jobs at a 30 % price difference.
- Ferrite content. ASTM A351 sets no ferrite number for CF8M. Plenty of purchase specifications add one anyway, most often 5 to 15 %. It changes weldability and it changes the machining feel, so it is worth agreeing before the pattern is cut rather than after the first castings are inspected.
Five swaps that look safe and are not
These are the substitutions we see most often on incoming drawings. Each one is reasonable on paper and each one has a failure mode.
1. CF8M to CF3M, to fix a corrosion problem
The carbon limit drops from 0.08 to 0.03 % and the point of that is weld heat-affected zones on thick sections. What it does not fix is chloride pitting, which is the failure most people are actually looking at. ASTM quotes the same minimum tensile and yield for both grades, so the wall does not get stronger, and the low-carbon grade is slower to fill and gassier in the mould.
2. WCB to CF8M, because stainless sounds safer
Thermal expansion nearly doubles, from about 12 to about 17 µm/m·K. A bolted joint designed around a carbon steel body grows differently and the gasket starts weeping at temperature. The stainless body also cannot be welded straight to carbon steel pipe without a transition piece, which nobody budgets for.
3. CA15 to CA6NM, for a part that has to be welded
CA15 is 12 % chromium martensitic, hard and cheap, and it cracks in the weld heat-affected zone unless you preheat to 200–300 °C and temper straight afterwards. CA6NM carries about 4 % nickel, welds without that ritual and keeps its toughness, which is why hydro runners are made from it. It is also softer, so it is not a wear upgrade.
4. CF8M to CD3MN duplex, for chloride stress corrosion
The move is correct for stress corrosion cracking and wrong for everything else people assume comes with it. Duplex must not run above roughly 300 °C, where 475 °C embrittlement starts to matter, it is not the answer for concentrated sulphuric acid, and the casting has to be solution annealed and water quenched. A slow cool leaves sigma phase and the impact values fall away without any visible change in the part.
5. Wrought 316 plate to a CF8M casting
On a certificate the chemistry matches closely. In the part it does not: the casting carries 5 to 15 % ferrite, has coarser grains, lower elongation, and most design codes apply a casting factor to it. A welding procedure qualified on plate does not automatically cover the cast body, and the welder needs a different WPS before the first arc.
What the alloy costs you in the foundry
Metal price is the visible number. The invisible one is what the alloy does to yield and to the wax die, and it is usually larger. The allowances below are ours, on our dies and our shell system.
| Family | Die allowance | Fluidity | Hot tear risk | Where it bites |
|---|---|---|---|---|
| Carbon and low alloy steel | 1.6–1.8 % | good | low | cheapest metal, worst corrosion; machining stock has to be generous |
| Martensitic 12 % Cr | 1.8–2.0 % | fair | high on thin webs | cracks in the shell where sections change quickly |
| Austenitic stainless, CF8 / CF8M | 2.0–2.4 % | fair to good | medium | gas porosity on heavy sections; ferrite has to be held, not just measured |
| Duplex and super duplex | 2.0–2.2 % | poor | high | quench has to be fast, and the gating eats yield |
| Nickel base | 2.4–2.8 % | poor | medium | dedicated crucible and melt practice; metal cost dominates the quote |
| Cobalt base | 2.2–2.4 % | poor | medium | hard to machine and hard to weld, so scrap is expensive twice |
| Copper base | 1.5–1.8 % | good | low | gas pickup and dross; watch the pour temperature |
| Aluminium | 1.2–1.4 % | good | low | oxide film and thin walls are easy; feeding is the constraint |
Another foundry will quote a different allowance because their wax, their shell and their pour temperature are different. If two quotes differ by one percent on linear shrinkage, this table is usually why. The dimensional side of the same question sits on the capability data sheet, next to the tolerances it has to live with.
When the drawing is silent
Five items decide the price of a casting and are missing from a surprising number of drawings. We ask for all five before quoting, because guessing wrong on any of them turns into a rework loop.
- Heat treatment condition. Write it on the drawing, not in an email. It changes hardness, it changes machining, and on martensitic grades it changes whether the part is usable at all.
- Test bar location. Separately cast, integrally cast, or cut from the part. Three different numbers, three different prices.
- NDT level and coverage. Standard, acceptance level, percentage of surface, and who witnesses it.
- Weld repair policy. Allowed with approval, allowed freely, or not at all. This decides whether a 2 % internal indication becomes scrap.
- Machining stock. Two millimetres a face on a cast surface is normal. Half a millimetre on a warped flange is a repair, and the repair is charged to somebody.
Questions that come up at the quoting desk
Is CF8M the same as 316?
Close enough to substitute for a bracket, not for a pressure boundary. CF8M allows up to 0.08 % carbon and normally carries 5 to 15 % ferrite; 316 plate carries no ferrite by design. Same corrosion family, different weld procedure, lower elongation, and most design codes add a casting factor. The alloy finder lists both, with the numbers, if you need to compare them side by side.
Can you cast 2205 duplex?
Yes, as CD3MN. It is not a drop-in for CF8M. Gating and risering are different, the quench has to be fast, and we specify impact testing because a slow cool destroys toughness without changing how the part looks. Below 300 °C in chloride service it is the right answer; above that it is the wrong one.
Which grade for a seawater valve body?
CD3MN for most sizes. C95800 aluminium bronze when the body is large and the water is moving. CF8M only with cathodic protection or a temperature ceiling around 60 °C, and even then we would ask about stagnation.
Is GB ZG230-450 the same as WCB?
Same strength class, different rules. They line up on tensile and yield, but the carbon and manganese windows differ and WCB orders often carry Charpy requirements the GB grade does not mention. For a domestic pump casing either is workable. For an export order the certificate has to match the standard named on the purchase order.
How small a run will you do in a nickel alloy?
One heat, if the part justifies it. Nickel alloys need a clean crucible and a separate melt practice, so the economics are per heat rather than per piece. Below about 50 kg poured, the setup stops being the problem and the metal price takes over.
Where to go next
- Capability data sheet — 27 grades across five standards, as-cast tolerances and minimum wall thickness.
- Casting tools — Alloy finder by duty or by grade, piece weight and price, melt additions from a bath analysis.
- Defect reference — 46 defect types, each tied to the stage that causes it, because half of them start with the alloy choice.
- Articles and case notes — Longer write-ups: shell problems, gating, shrinkage, the jobs that went wrong first.
- Send a drawing — If you have a part number and a quantity, this is faster than reading further.
Quality
Precision manufacturing for diverse industrial applications.
Services
Products
+8615912702921
