Sand casting explained: what it suits and what it costs you
The oldest casting process still in daily use, and still the right answer for one off parts, very large parts and alloys no die will take.
Sand casting makes a mould out of bonded sand, pours metal into it, then breaks the mould to get the part out. Every casting needs a new mould. That sounds wasteful until you price the alternative: a steel die costs thousands before you see a single part, and a sand mould costs the sand, the pattern and the labour.
That is the whole commercial case. Sand casting suits one off parts, small batches, very large parts, and alloys that no die casting process will take. Cast iron and bronze go into sand all day long. Neither will go into a high pressure die.
What it costs you is precision. Expect around plus or minus 0.8mm on as cast dimensions, and a surface in the Ra 6.3 to 25 micron range depending on the sand and the pattern. If your drawing needs a sealing face, a bearing bore or a tight bolt pattern, you are not getting it straight from the mould. You leave machining stock and cut it afterwards.
How a sand mould is made and destroyed
A pattern is made first, usually in wood, resin or aluminium, slightly oversized to allow for the metal shrinking as it cools. The pattern is packed in bonded sand inside a two part box, then withdrawn, leaving a cavity the shape of the part. Cores, also in sand, sit in the cavity to form internal passages. Metal is poured under gravity, left to solidify, and the mould is broken apart.
Two binder systems cover most work. Green sand uses clay and water, suits higher volumes, and is cheaper per mould. Chemically bonded sand, often resin, holds a cleaner edge and a better surface, and costs more.
The pattern survives. The mould does not. That split is what makes the economics work the way they do.
One off economics: why no permanent tooling changes the maths
Permanent tooling is the dominant cost in most casting decisions. A high pressure die is a block of hardened steel, cut and polished and tried out before it produces anything sellable. It only makes sense once production volume spreads that cost across enough parts.
Sand casting moves the money. The pattern is a fraction of a die’s cost, and a wooden pattern for a simple part can be made in days. One part is a reasonable order. Five is reasonable. So is a replacement for an obsolete pump housing where the drawing is a photocopy and the original foundry closed in 1988.
That makes sand casting the default for spares, prototypes of parts destined for a die later, and anything with a lifetime demand measured in dozens. Most of the sand casting we place falls into exactly that band.
The process with no practical size limit
Ask a die caster for a part weighing 400kg and the conversation ends. Ask a sand foundry and they ask about handling and lead time.
Sand moulds are built up rather than machined, so size is limited by the foundry’s melting capacity, crane and floor space rather than by the process. Engine blocks, pump casings, machine bases, valve bodies, counterweights: these live in sand because nothing else takes them at a sensible price.
Thickness follows the same logic. A sand casting can carry a 30mm section without trouble, and heavy sections solidify slowly enough to feed properly if the foundry has designed the risers correctly. Pressure die castings want thin walls and get into difficulty above about 5mm.
Alloys: iron and bronze, where a die cannot follow
High pressure die casting runs aluminium, zinc and magnesium. The reason is melting point: steel dies survive molten aluminium at around 700C, and they do not survive molten iron at 1,400C.
Sand does not care. The mould is consumed anyway, so pouring temperature stops being a tooling problem. Grey iron, ductile iron, bronze, brass, aluminium and steel all go into sand.
If your part is specified in cast iron, sand casting or shell moulding is the realistic route. If it is bronze, the same applies. Those materials are chosen for a reason, usually damping, wear or corrosion behaviour, and substituting an aluminium alloy to suit a different process is a design decision, not a sourcing one. Talk it through before you commit. Our materials guidance covers what we see specified most often.
Tolerance and finish: what you actually get as cast
ISO 8062 sets out casting tolerance grades, and sand castings sit several grades coarser than investment or die castings. In practical terms:
| Process | Tooling cost | Sensible minimum order | As cast tolerance | As cast surface |
|---|---|---|---|---|
| Sand casting | Pattern only, low | 1 part | Around plus or minus 0.8mm | Ra 6.3 to 25 |
| Investment casting | Wax tool, moderate | 50 to 100 parts | Around plus or minus 0.1 to 0.3mm | Ra 1.6 to 3.2 |
| Gravity die casting | Die, high | 500 to 1,000 parts | Around plus or minus 0.3mm | Ra 3.2 to 6.3 |
| High pressure die casting | Die, highest | 10,000 parts and up | Around plus or minus 0.1 to 0.2mm | Ra 0.8 to 3.2 |
Figures are typical rather than guaranteed, and if the finish column is what rules sand out for you, investment casting is the next process to price. Section thickness, part size, alloy and the foundry’s own practice all move them.
Two more things change as cast dimensions. Draft, because the pattern has to come out of the sand, so vertical faces are tapered by one to three degrees. Parting lines, because the mould is made in two halves and a small mismatch is normal.
Machining stock: how much to leave and where
Leave stock on every surface that has a tolerance tighter than the casting can hold. Around 3mm on faces that will be machined is a sensible starting point for a medium sized part, more on large castings where distortion is a factor.
What matters more than the number is deciding early which faces are the datums. Pick them, mark them on the drawing, and make sure the foundry and the machine shop are working from the same ones. Castings that fail at inspection usually fail because the machinist picked a different reference face from the one the pattern maker assumed.
Most of our cast parts go straight from the foundry into CNC machining before inspection, which is also the point where the as cast datum decision shows up as either a non event or a scrap report. Our design tips cover draft, radii and section transitions in more detail.
When to move to a die: the volume that justifies permanent tooling
The switch is not about part size or complexity. It is about how many you will buy, over how long.
- Count the lifetime demand, not the first order. A thousand a year for five years is a die conversation. A thousand once is not.
- Price both routes properly. Ask for the sand price per part at your real quantity, then the die price including tooling. Divide the tooling cost by the lifetime quantity and add it to the die unit price.
- Check the alloy will transfer. Iron and bronze stop the conversation. Aluminium continues it.
- Check the wall thickness. Below about 5mm throughout, high pressure die casting is worth pricing. Thicker or heat treated, look at gravity die casting instead.
- Decide how settled the design is. Changing a pattern is cheap. Changing a hardened die is not.
- Add the machining you will save. Tighter as cast tolerances remove operations, and that saving is often larger than the difference in casting price.
The practical version. Sand casting wins when the quantity is small, the part is large, or the alloy is iron or bronze. It costs you roughly plus or minus 0.8mm and a rough surface, so leave machining stock on anything that has to fit. Once lifetime demand runs into the thousands and the alloy is aluminium, price a die against it.
When sand casting is not what you need
If you need twenty standard flanges or a common bracket that already exists as a catalogue item, buy them from a stockist. A casting is slower and dearer than a part somebody already makes by the thousand.
If the part is small, thin walled and needed in high volume, sand casting will quote badly against a die and the finish will cost you in machining.
If the geometry is simple enough to cut from bar or plate in a few operations, CNC machining or laser cutting and fabrication will usually beat a casting at low volumes with no pattern cost at all.
Casting earns its place when the shape is genuinely three dimensional, when the alloy demands it, or when removing that much material from solid would be absurd.
Common questions
How accurate is sand casting?
Plan on around plus or minus 0.8mm on as cast dimensions for a medium sized part, with more variation on large castings and across the parting line. ISO 8062 grades are the formal way to specify it. Anything tighter gets machined.
What surface finish does sand casting give?
Typically Ra 6.3 to 25 microns. Chemically bonded sand sits at the better end, green sand at the rougher end. Shot blasting improves the appearance but does not change the dimensions much.
Can I get one part sand cast?
Yes. One part is a normal order, because the only tooling is the pattern. It is the main reason obsolete spares get remade this way.
Is sand casting cheaper than machining from solid?
At one or two parts, usually not. Past about ten, and on anything with internal passages or a lot of material to remove, casting starts winning quickly.
What alloys can be sand cast?
Grey and ductile iron, bronze and brass, aluminium alloys and steel. The mould is destroyed each time, so high pouring temperatures are not the obstacle they are in die casting.
Do I need to supply a pattern?
No. Send the drawing or the 3D model and the pattern gets made to suit the chosen foundry’s practice, including shrinkage and draft. If you already own a pattern, say so, because it changes the price.
Have a part you think belongs in sand?
Send the drawing or the model, the alloy and the quantity you expect to buy over the life of the part, and we will tell you which casting process it belongs in and place it with a foundry we have already qualified. If a stockist is the better answer, we will say that too. Request a quote.