Investment Casting vs Sand Casting: 7 Differences

Investment casting and sand casting are separated by one decision: what the mould is made of. Sand casting packs bonded sand around a rigid pattern that has to be withdrawn, so the mould needs a parting line and a draft angle. Investment casting coats a wax pattern in ceramic slurry, then melts the wax out, so the mould is one piece with no parting line at all. Everything else — tolerance, finish, minimum wall, tooling cost, lead time — follows from that.

The measured gap is wide. Investment casting normally holds CT4–CT6 dimensional tolerance and Ra 0.4–3.2 μm surface roughness. Conventional sand casting lands around CT9–CT11 with a finish several grades coarser. Investment casting costs more per piece; sand casting costs more per finished part once machining is counted.

This page compares the two on the seven points buyers actually argue about, using the tolerance, roughness and cost tables published in Chinese foundry handbooks.

Sand casting vs investment casting: the seven differences at a glance

• Dimensional tolerance — Investment casting: CT4–CT6 (CT3 achievable); Sand casting: about CT9–CT11

• Surface roughness — Investment casting: Ra 0.4–3.2 μm; Sand casting: coarse, needs machining to clean up

• Minimum wall thickness — Investment casting: 0.5 mm; Sand casting: thick walls only

• Parting line and flash — Investment casting: none; Sand casting: parting line plus draft angle

• Tooling — Investment casting: wax die, 100,000+ shots in steel, high cost; Sand casting: pattern is cheap and fast

• Part size — Investment casting: 1 g to ~1,000 kg, up to 2.2 m; Sand casting: small to very large

• Cost structure — Investment casting: high fixed, low machining; Sand casting: low fixed, high machining

Why the mould decides everything

The two process routes are structurally different, and the handbook flow charts make it obvious.

Sand casting runs: mix sand and binder → mould (cope) → set cores → close the mould → pour → shake out → clean. The pattern stays in the box until the sand is compacted, then it is pulled out, which is why the cavity must be tapered.

Investment casting runs: inject wax into a die → assemble a pattern tree → dip in slurry → stucco with sand → dry → repeat until the shell is thick enough → dewax → burn out the shell → pour → knock out and cut off.

Practical Investment Casting Technology reduces the difference to three process characteristics:

1. A fusible pattern means no opening the mould. The wax is melted out, so the shell is a single continuous cavity. No parting plane means no flash and no burr on the casting.

2. A liquid slurry shell copies the pattern better than sand can. The handbook states the shell made from liquid coating reproduces the wax pattern more accurately and more smoothly than a sand mould does.

3. Pouring into a hot shell. The shell is at temperature when metal enters, so the metal stays fluid longer and can fill thin, complex sections that would freeze off in a cold sand mould.

Those three points are the entire commercial case for investment casting. Everything below is the arithmetic of them.

1. Dimensional tolerance: the CT grade gap

GB/T 6414 sorts casting tolerances into 16 grades, CT1 to CT16, with CT1 the tightest. The tolerance band is normally split symmetrically, half plus and half minus. Some published values, in mm, total band:

• 25–40 mm feature — CT4: 0.32; CT5: 0.46; CT6: 0.64; CT7: 0.90; CT8: 1.3; CT9: 1.8

• 100–160 mm feature — CT4: 0.44; CT6: 0.88; CT8: 1.8; CT9: 2.5

A 30 mm dimension cast to CT6 is therefore 30 ± 0.32 mm. The same dimension cast to CT9 is 30 ± 0.90 mm — nearly three times the spread.

Investment casting is classed as precision casting and normally reaches CT4–CT6, with CT3 possible on high-precision work under controlled conditions. The standard also allows tighter non-linear tolerances: form tolerances such as straightness, roundness, parallelism, perpendicularity and concentricity are achievable, not just linear size.

For sand casting, the handbooks do not quote one number because the family is broad, but they do give a calibration point: lost foam casting reaches CT7–CT9 and is described as one to two grades better than conventional sand casting. Read backwards, that puts conventional sand casting at roughly CT9–CT11 — which is consistent with the relative grading table in the next section.

Practical consequence: if a drawing carries ±0.1 mm on a 25 mm dimension, that is a 0.2 mm band, tighter than CT4. Neither process hits it as-cast. Investment casting gets close enough that a light finishing cut holds the tolerance; sand casting needs real machining allowance.

2. Surface roughness: Ra 0.4 versus "machine it off"

Investment casting roughness depends heavily on which shell system is used. The handbook table gives two bands:

• Wax-based pattern, water-glass shell — as-cast Ra 12.5–6.3 μm; after finishing Ra 3.2 μm

• Resin-based pattern, ethyl-silicate or silica-sol shell — as-cast Ra 6.3–3.2 μm; after finishing Ra 1.6–0.8 μm

The second row is the process most export buyers mean when they say "investment casting" — silica-sol shell castings rather than water-glass work. The first row is the low-cost route still common in volume production, and its as-cast finish overlaps the sand casting range — a point worth remembering when two quotations differ by 40%.

Roughness also degrades with part size and weight, because a bigger, heavier section holds more superheat and reacts longer with the shell:

• Max dimension 50 mm, wall under 3 mm, under 0.2 kg — Ra 1.6–3.2 μm

• 150 mm, wall 3–5 mm, 0.2–1.0 kg — Ra 3.2–6.3 μm

• 250 mm, wall 5–10 mm, 1.0–5.0 kg — Ra 6.3–12.5 μm

• 500 mm, wall 10–20 mm, 5.0–15.0 kg — Ra 12.5 μm

• Over 500 mm, wall over 20 mm, over 15 kg — Ra 12.5 μm

So the headline Ra 1.6 is a small-part number. On a 10 kg casting the same process delivers Ra 12.5, and the finishing argument weakens.

Sand casting is graded 6 out of 6 on surface roughness in the relative table below, i.e. the worst of the four processes compared. Its surface is described in the same handbook as visually poor but easy to machine — the sand mould simply cannot reproduce fine detail, and the claim is not that it can.

3. The relative grading table, four processes at once

This is the single most quotable table in the source material. It scores four casting processes from 1 (best) to 6 (worst) on four criteria:

• Dimensional tolerance — Sand casting: 6; Investment casting: 2; Permanent mould: 4; Die casting: 1

• Surface roughness — Sand casting: 6; Investment casting: 1; Permanent mould: 4; Die casting: 1–3

• Porosity level — Sand casting: 1; Investment casting: 4; Permanent mould: 3; Die casting: 3–5

• Cost per casting — Sand casting: 3; Investment casting: 6; Permanent mould: 3; Die casting: 1

Two results in that table are uncomfortable for anyone selling investment casting, and they are worth stating plainly.

On porosity, investment casting scores 4 — worse than sand casting's 1. A sand mould vents through the sand; a ceramic shell does not. Combined with the slow cooling the handbooks list as a limitation, this is why investment castings in critical service are often hot isostatically pressed (HIP), and why the process is normally followed by heat treatment to develop properties.

On unit cost, investment casting scores 6 — the worst of the four. That is the honest starting position: an investment casting costs more to produce than a sand casting of the same weight. The case for it is never unit price, it is total cost.

4. Minimum wall, holes, and part consolidation

Investment casting's operating envelope, from the handbooks:

• Minimum wall thickness — 0.5 mm

• Minimum hole diameter — below 1 mm

• Casting weight — about 1 g to nearly 1,000 kg

• Maximum envelope — up to roughly 2.2 m

The reason thin walls work is the hot-shell pour: metal stays liquid long enough to reach the end of the section. In a sand mould the metal front freezes earlier, which is why sand castings are described as thick-walled.

Part consolidation is the underrated saving. The handbook documents a titanium investment casting 1.3 m in diameter and 182 kg that replaced an assembly of 88 separate parts, cutting weight by 55% while raising both strength and dimensional accuracy. Sand casting can also consolidate, but the parting line and core complexity limit how far, and every joint that is removed is machining and inspection that disappears.

5. Size and weight ceiling

This is where sand casting wins outright. The handbook's process comparison places sand casting across small, medium and large castings with thick walls, while investment casting is bounded by shell strength and burnout furnace capacity. Beyond a few hundred kilograms the ceramic shell becomes difficult to handle and the economics collapse.

A rough rule used in quoting: under 50 kg, investment casting is competitive on total cost; over 50 kg, sand casting usually wins unless the geometry is impossible any other way. The 1,000 kg figure is the extreme edge of the process, not a normal order.

6. Tooling and the break-even volume

The handbook classifies wax dies by life and cost:

• Machined steel die — service life over 100,000 shots, high manufacturing cost; used for high-volume, tight-tolerance parts

• Low-melting-point alloy die — life of several thousand shots or more, lower cost; used for a few thousand pieces and for cavities that are hard to machine

• Plaster die — low accuracy, low cost, short lead time; used for single parts, prototypes and trial production

For development work the pattern can come from a rapid-prototype part instead of a die at all, which removes most of the entry cost for a first article — prototype investment castings are quoted that way.

Sand casting's tooling is a pattern and a core box, cheap and quick by comparison, but the handbook notes the trade-off explicitly: low equipment and mould cost, higher machining cost.

Crossovers, in practice:

• 1–50 parts — sand casting or investment casting from a printed pattern; investment casting wins if the part needs no machining

• 50–5,000 parts — investment casting, low-melting alloy die

• 5,000–100,000+ parts — investment casting with a steel die, or die casting if the alloy allows

• Very large parts at any volume — sand casting

The economics table in the source material measures the payoff differently: as a share of machining allowance removed, investment casting ranks highest at about 90%, against 50–60% for sand casting. That single number explains why a part with a 15-minute machining cycle can be cheaper overall as a more expensive casting.

7. Alloys, and the heat treatment you must budget for

Investment casting is not restricted by alloy. The handbook lists carbon structural steel, stainless steel, alloy steel, cast iron, aluminium, copper alloys, cast superalloys, magnesium, titanium and precious metals. Alloys that are hard to forge, weld or machine are often better suited to investment casting than to anything else — which is why the process dominates turbine blades and orthopaedic implants.

Sand casting covers iron, steel, aluminium and copper alloys, and handles the very large work investment casting cannot.

The cost item buyers forget: investment castings cool slowly in the ceramic shell, so mechanical properties usually need a heat treatment step after knockout. That is a line item in the quotation and a week in the lead time — heat-resistant and stainless castings are normally quoted with it included. Sand casting's cooling is faster but its section thickness limits the designs it can serve.

For alloy-by-alloy selection logic, see the investment casting alloys guide.

The trap: "investment casting" is not one specification

This is the part that decides most sourcing arguments, and it is documented in the handbook's purchasing chapter.

Chinese investment casting splits into two production routes:

• Wax-based pattern, water-glass shell, air melting — dimensional tolerance only CT7–CT9, surface roughness Ra 3.2–12.5 μm. This route is aimed at carbon and low-alloy steel parts with moderate quality requirements, and the handbook states directly that it does not reach international investment casting standards.

• High-quality pattern, high-pressure wax injection, silica-sol or ethyl-silicate shell, steam dewaxing, fast air or vacuum melting — tolerance CT4–CT6, roughness Ra 0.4–3.2 μm. This is the internationally comparable route.

Both are legally "investment casting". One delivers sand-casting-grade tolerance at an investment-casting price. If a quotation looks 40% cheaper than the others, the first question is not the price, it is which shell system and which wax injection method the shop is running.

Ask three questions before comparing quotes:

1. Which shell system — water glass, ethyl silicate, or silica sol?

2. What CT grade do you guarantee on this feature size, in writing?

3. What Ra do you guarantee as-cast, and what is the sampling method?

For how tolerances are specified and inspected, see the investment casting tolerances guide. For the step-by-step shell and pouring sequence, see the investment casting process page.

When to choose which

Choose sand casting when:

• The part is large — roughly over 50 kg, or over 500 mm with heavy sections

• Walls are thick and the design is simple

• Tolerances of CT9–CT11 are acceptable and machining is planned anyway

• Volume is low and tooling must stay cheap

• The alloy is a common cast iron or carbon steel grade

Choose investment casting when:

• Walls are thin — down to 0.5 mm — or holes are under a few millimetres

• The part should come off the mould without machining, or with a single finishing pass

• Geometry is complex enough that a parting line would ruin it

• Several parts can be consolidated into one casting

• The alloy is difficult to machine — stainless, heat-resistant, cobalt- or nickel-based

• Volume justifies a die, or the part is small enough that the die cost amortises quickly

Choose neither alone when the tolerance is tighter than CT4 as-cast: specify investment casting plus a finishing operation, and put the datum scheme on the drawing rather than leaving it to the shop.

FAQ

Is investment casting more expensive than sand casting?

Per casting, yes — it is graded the most expensive of the four processes compared in the handbook. Per finished part, it often is not, because it removes up to about 90% of the machining allowance against 50–60% for sand casting, and it eliminates flash, parting-line cleanup and assembly of welded subcomponents. Compare landed cost after machining, not casting price.

Which process is more accurate, investment casting or sand casting?

Investment casting, by roughly three to five CT grades. Investment casting holds CT4–CT6 as standard and CT3 under special control; conventional sand casting sits around CT9–CT11. On a 30 mm dimension that is ±0.32 mm versus ±0.90 mm.

Can sand casting achieve Ra 1.6 μm?

Not as-cast. Sand moulds cannot reproduce that level of detail; the surface is coarse and is normally machined where finish matters. Ra 1.6 μm as-cast requires a resin pattern and a silica-sol or ethyl-silicate shell in investment casting, and even then it applies to small parts — the same process gives Ra 12.5 μm on a 10 kg casting.

What is the minimum order quantity for investment casting?

There is no process minimum — the constraint is tooling. With a plaster die or a printed pattern, a handful of parts is viable. With a machined steel die rated over 100,000 shots, the die cost has to be absorbed, so the practical minimum is usually a few hundred to a few thousand pieces depending on part value.

Do investment castings need heat treatment?

Usually yes for structural work. Cooling in a ceramic shell is slow, so the handbook lists the need for heat treatment to develop properties as one of the process limitations. Budget it as a separate operation and a separate lead-time item.

Why do two investment casting quotes for the same drawing differ so much?

Most often because they are two different processes. A water-glass shell shop quotes CT7–CT9 and Ra 3.2–12.5 μm; a silica-sol shop quotes CT4–CT6 and Ra 0.4–3.2 μm. Ask which route the quote is based on before negotiating the number.

About the data in this article Process routes, the three process characteristics, the CT tolerance values (GB/T 6414), the surface roughness tables, the four-process relative grading table, the wax die classification and the two Chinese production routes are taken from Practical Investment Casting Technology (《实用熔模铸造技术》, Jiang Buju), Investment Casting Handbook (《熔模铸造手册》, China Foundry Association) sections 1.2, 1.4 and 3.4, and Precision Casting Technology and Quality Inspection (《2005精密铸造实用新技术与质量检测实物全书》) tables 1-2-1, 1-2-8, 1-2-10, 1-4-6 and 1-4-7. The CT9–CT11 figure for conventional sand casting is derived from the handbook statement that lost foam reaches CT7–CT9, one to two grades better than conventional sand casting, and is flagged as a derivation rather than a quoted value. Figures are typical reference data for specification work, not a substitute for a drawing-specific agreement with your foundry.

Comparing processes for a specific part? CireCraft is written by investment casting practitioners from XINYOUYUAN, a precision casting plant in Dongguan, China, casting stainless steel, carbon steel, alloy steel and heat-resistant alloys in the 0.01–50 kg range. Send the drawing and the annual volume to castinmetal.com and we will tell you whether investment casting or sand casting is the cheaper route for it — including when the answer is sand casting.

Updated: October 2026

Contact

Questions about our casting process? Reach out anytime.

Email

Phone

info@castinmetal.com

+1-555-123-4567

© 2025. All rights reserved.