Investment cast fittings are poured into a ceramic shell built around a wax pattern. Machined fittings are cut from bar or from a forging on a CNC lathe or mill. Both can deliver the same part number at the same pressure rating, and still land on completely different cost curves.
Below about 100 pieces, machine. Above roughly 500 pieces of a complex shape, cast. Between the two the wax die and the chip volume decide, and either route has to be qualified by NDT and hydrostatic test.
Key takeaways
- Investment casting forms near-net shapes from a wax pattern: injected pattern, ceramic shell, dewax, pour, knockout — complex elbows, tees and valve bodies come out nearly finished.
- Machining from bar or forging holds the tight tolerances — CNC typically ±0.025 mm class — and the fine surface; investment casting runs ±0.1–0.5 mm and an as-cast surface around Ra 1.6–6.3 µm.
- Cost breaks at volume: investment casting tooling commonly pays back in the 100–500 piece range; below that, machining usually wins total cost.
- Cast pressure parts carry porosity risk — qualifying them means NDT and hydrostatic testing to the applicable valve/fitting test standards, with traceable heat numbers.
- Grades differ by process: cast stainless is CF8/CF8M (ASTM A351, the cast analogs of 304/316); free-machining 303 is a bar-stock grade, not a standard casting alloy.
Which one do I need?
Two questions settle the process: how many pieces a year, and how much of the shape would have to come out of solid bar. Grade, tolerance and test requirements follow from those two.
| Investment cast | Machined from bar or forging | What decides it | |
|---|---|---|---|
| Shape cost | Complex internal geometry forms in the wax | Every internal corner is cutting time, paid on every part | How many chips the part would remove from solid stock |
| Linear tolerance | ±0.1–0.5 mm, ISO 8062 CT4–CT6 territory, size-dependent | ±0.025 mm class as CNC routine | Which dimensions are actually sealing or gauging surfaces |
| As-delivered surface | Ra 1.6–6.3 µm | Ra 0.4–1.6 µm class | Where the seal lands |
| Tooling | A wax die, commonly low thousands of dollars | None | Whether the die amortises over the run |
| Volume band | Pays back in the hundreds of pieces; above roughly 500 the per-part economics flip | Wins below about 100 pieces, and on any lot with no die budget | Annual quantity |
| Grade route | CF8 and CF8M per ASTM A351 cover most industrial demand | 303, 316L, 2205, 625/718, Monel and C-276 from bar | Alloy and run size |
| Pressure integrity | Qualified, not assumed: NDT plus hydrostatic test | Wrought stock; cast strength can sit 10–30% below it | Whether the part is pressure-retaining |
The crossover is not a single number. A part that would hog out half a bar billet pays the die back earlier than a simple straight coupling does, because the die cost is spread over the run and the chips are not.
How is each one made?
Investment casting builds the part backwards from its finish. Machining starts from the other end: solid bar or a forging, with the shape cut out of full-section stock.
Investment casting, in order:
- Inject wax into a metal die to form the pattern.
- Assemble the wax patterns into a tree.
- Dip and stucco the tree in ceramic slurry until a hard shell forms.
- Dewax in a steam autoclave or by flash-fire, leaving the cavity.
- Pour molten stainless, knock off the shell, cut the parts off the tree, then machine the sealing faces.
Machining from bar, in order:
- Cut bar, or start from a forging.
- Turn or mill the shape, one chip at a time.
- Machine the sealing faces last, to the tolerance they actually need.
What comes out of the mould is near-net-shape: a 90° elbow or a tee with its internal geometry already formed, needing only finishing and the machining of sealing faces.
Where does casting win?
Complexity is free in wax and expensive in chips. Elbows with swept internal radii, tees, valve bodies, caps and Y-patterns come out of the mould nearly complete, where machining would hog out half a bar billet.
Tooling is real but modest. Investment casting dies commonly run in the low thousands of dollars, against die casting’s far higher tooling and volume thresholds, and break-even lands in the hundreds of pieces.
Above roughly 500 pieces the per-part economics flip decisively. Below about 100, machining almost always wins on total cost. The cast fittings range is built on this logic.
Where does machining win?
Tolerances first. CNC work holds ±0.025 mm class on critical features as a routine matter. Investment casting runs a ±0.1–0.5 mm linear band, ISO 8062 CT4–CT6 territory and size-dependent.
Surface follows the same split: as-machined Ra 0.4–1.6 µm class against as-cast 1.6–6.3 µm. The difference matters where a seal lands.
No tooling is the second edge. Prototypes and small lots skip the die entirely, free-machining grades like 303 cut fast from bar, and a drawing change is a program edit rather than a new wax die. Non-standard and low-volume specials live here, which is the custom fittings territory.
Material access is the quieter advantage. The casting ecosystem standardises on a handful of grades, CF8 and CF8M covering most industrial demand. Duplex and nickel alloys are castable, but under separate specifications with tighter melt chemistry and heat treatment, and they carry small-heat premiums that make them disproportionately expensive at low quantity.
| Grade family | As a casting | As bar | What decides it |
|---|---|---|---|
| 300-series stainless | CF8 or CF8M per ASTM A351 | 303 free-machining, 316L | Whether the design depends on free-machining bar stock or on a standard cast grade |
| Duplex | Castable under separate specifications, with small-heat premiums | 2205 | Run size, because the cast route carries the small-heat premium |
| Nickel alloys | Castable under separate specifications, tighter melt chemistry | Inconel 625 and 718, Monel, Hastelloy C-276 | Run size, because the same alloys are routine on distributor shelves as bar |
The same alloys sit on distributor shelves as bar. For a 20-piece run in a speciality alloy the bar route is often the only practical route.
Schedule is the other edge. With no die to build, prototypes and small lots move in days, and simple parts commonly run 3–5 days standard, while casting mould setup alone runs weeks. One caveat: speciality alloys machine slowly, so the bar route buys access and schedule, not cheap chips.
What goes wrong with a cast pressure part?
Cast metal can hide voids that machining cannot see. Shrinkage and gas porosity stay inside the section until a sealing face is cut or a hydrostatic test finds them.
Process references put cast strength potentially 10–30% below wrought stock of the same nominal grade. That is why pressure-retaining castings are qualified rather than assumed.
Radiography or ultrasonic examination covers internal defects; dye-penetrant or magnetic-particle covers surfaces. Hydrostatic shell and seat testing runs to MSS SP-61, API 598 or ISO 5208, and MSS SP-25 marking carries the heat number that makes the heat traceable.
| Symptom | Most likely cause | What to change |
|---|---|---|
| Weep or leak at the shell on hydrostatic test | Shrinkage or gas porosity inside the section | Radiograph or ultrasonically examine that heat before accepting it |
| Porosity appears only after the sealing face is machined | A void opened up by stock removal | Examine the casting before machining critical faces, or move the part to bar stock |
| Dimensions drift across a cast batch | The ±0.1–0.5 mm linear band, size-dependent | Hold the band on non-critical faces and machine the sealing faces |
| A seal leaks although the face is flat | As-cast Ra 1.6–6.3 µm where the seal lands | Machine that face to the Ra 0.4–1.6 µm class |
| The drawing says 316 and the certificate says CF8M | A wrought designation written for a cast part | Restate the material line as CF8M per ASTM A351 |
| Machining cost runs away on a complex body | Swept internal radii hogged out of solid bar | Cast the shape instead, or start from a forging |
| The heat number cannot be traced on the part | Marking not carried to MSS SP-25 | Require MSS SP-25 marking with the heat number |
This is the same proof-versus-burst discipline covered in proof pressure vs burst pressure, and the trail in EN 10204 3.1 certificates is what makes a cast heat auditable.
What goes on the RFQ?
The RFQ is where a process decision becomes an order. Name the process you expect, and give the factory the numbers that justify it.
- Drawing plus annual quantity. The process decision rides on volume.
- Critical dimensions with tolerances, marking which faces are sealing or gauging surfaces.
- Grade by process:
CF8Mfor a cast 316-equivalent,316L barfor machined. Name the form, not only the alloy. - NDT level: radiographic or ultrasonic for internal defects, dye-penetrant or magnetic-particle for surfaces.
- Hydrostatic test: shell and seat, to MSS SP-61, API 598 or ISO 5208.
- Certificate type, EN 10204 3.1 where the heat has to be auditable.
- Marking to MSS SP-25, with the heat number on the part.
- Surface callouts on wetted or sealing surfaces, given as Ra.
Between 100 and 500 pieces, put both routes on the same drawing and let the chip volume argue its own case.
FAQ
When does investment casting beat machining on cost? Above roughly 100–500 pieces, where the die cost spreads thin; below about 100, machining almost always wins total cost. The more chips a part would hog out, the earlier casting pays.
Which process holds tighter tolerances? Machining. CNC work holds ±0.025 mm class routinely, while investment casting runs ±0.1–0.5 mm. Specify which dimensions actually need the tight band, because demanding it everywhere is how RFQs get expensive.
Is a cast fitting weaker than a machined one? Potentially. Porosity and cast grain structure can put cast strength 10–30% below wrought stock, which is why pressure castings are qualified by NDT and hydrostatic test. A tested CF8M casting is a sound pressure part.
What do CF8 and CF8M mean? The ASTM A351 cast designations: CF8 is the cast analog of 304, CF8M of 316. If the drawing says 316 and the part is cast, the material line is CF8M, same corrosion family on a different specification route.
Can I get 303 in an investment casting? Not as a standard option. 303 is a free-machining wrought bar grade, sulfur-added for machinability, while standard cast stainless grades are CF8 and CF8M per ASTM A351. A design that assumed 303 needs a grade conversation first.



