Green sand is the cheaper mold for small and medium ferrous castings made in volume; no-bake is the better process for large castings, short runs, and heavily cored parts, and above about 200 lb it also holds tighter dimensions. Below roughly 50 lb the two processes hold similar tolerances, so the choice there comes down to volume and price alone. Between those weights the foundry’s process control matters more than its binder, which is why a part near the line should be quoted both ways.
What Is Green Sand Casting?
Green sand casting molds the part in silica sand bonded with bentonite clay and water — roughly 5–10% clay and 2–5% water — compacted around the pattern and poured while the mold is still moist. Nothing is baked or chemically set, so the mold is ready to close the moment the pattern is drawn.
The bond comes from compaction, which is why the process lives on molding machines. Jolt-squeeze and high-pressure lines pack the sand hard enough to resist mold-wall movement during pouring, and harder molds hold tighter dimensions. The pattern takes that squeeze on every cycle, so production tooling is metal or a pattern mounted on a matchplate.
After shakeout the sand goes straight back into the system. It is cooled, re-tempered with water and clay, and reused continuously, with a fraction of burnt sand pulled out each turn and replaced with new. That closed loop, with no reclaimer in it, is what makes green sand the cheapest mold per part at volume.
What Is No-Bake Sand Casting?
No-bake casting bonds the sand with a liquid resin and a catalyst mixed in seconds before the sand reaches the pattern; the mix hardens by chemical reaction at room temperature, with no baking and no squeeze.
Because the mold cures into a rigid block before it leaves the pattern, it does not depend on compaction for strength. Three things follow: the mold resists erosion and metal penetration, so it carries heavy sections and deep cores cleanly; the pattern sees no ramming, so sealed wooden tooling lasts; and each mold ties up its pattern until it cures, which caps throughput. Very large floor and pit molds for steel are almost always no-bake for these reasons.
Spent no-bake sand cannot simply be re-tempered — the burnt resin shell has to come off each grain, and the reclaimer that does it is capital the foundry’s price has to carry.
No-Bake vs. Green Sand Casting
The differences that hold up are mold rigidity, pattern requirements, and throughput; the accuracy and finish gaps most comparison charts show are smaller and more conditional than they look, and the tooling-cost row is usually printed backwards.
| Factor | No-Bake Sand Casting | Green Sand Casting |
|---|---|---|
| Binder | Liquid resin (furan, phenolic urethane, alkaline phenolic) + catalyst, roughly 1–2% of sand weight | Bentonite clay + water |
| Mold curing | Chemical cure at room temperature on the pattern, minutes to an hour or more | None; bond set by compaction, mold poured moist |
| Mold strength | High and rigid; resists erosion and metal penetration | Moderate; rises with ramming pressure |
| Dimensional accuracy | Tighter above 200 lb; no advantage below 50 lb; added scatter across the parting line | Similar to no-bake on small castings; looser as weight rises |
| Surface finish | Good; takes fine facing sand and washes easily | Fair; tracks sand grain fineness, not the binder |
| Large castings | Excellent; the default for floor and pit molds | Possible on the floor, but dimensions harder to hold |
| Complex shapes | Good | Fair |
| Production rate | Slower: about 20 molds/hr on an automated loop | Faster: 65 molds/hr on an automatic machine, 300–400 on high-pressure lines |
| Tooling cost | Lower: about 80% of a green sand pattern; sealed wood is fine | Higher: metal or mounted patterns to survive machine molding |
| Sand cost | Higher (resin and catalyst on every mold) | Lower (clay and water, recirculated) |
| Sand reuse | Mechanical or thermal reclamation; over 90% recovery is routine | Continuous in-system reuse with make-up sand |
| Typical production volume | Low to medium | Medium to very high |
| Common metals | Steel, ductile iron, gray iron; the usual route for large steel | Gray iron, ductile iron, steel |
The accuracy row is the one buyers over-weight. SFSA’s dimensional-capability data on production steel castings puts the two processes at similar tolerances below 50 lb, with no-bake pulling ahead only above about 200 lb, and it finds a larger spread between foundries running the same process than between the two processes. On a large part with a tight CT grade, no-bake is the safer choice; on a small one, judge the foundry, not the binder.
No-bake also carries a penalty the chart omits: about 1 mm (0.040 in.) of added six-sigma scatter on any dimension that crosses the parting line, where green sand shows none worth measuring. The cured cope and drag are rigid blocks closed on pins, so any pattern or pin misalignment lands whole on the joint, whereas a green sand matchplate makes both halves from one plate. Keep close-tolerance features inside one mold half and the penalty never applies.
Tooling cost runs the other way from what most charts print. A no-bake pattern costs about 80% of the equivalent green sand pattern and builds in around 12 weeks against 18, because sealed wood is enough: it survives production runs that would wear a green sand pattern out of tolerance.
A pattern built for one process cannot simply move to the other. Low-alloy steel needs about 1.60% shrink allowance in green sand but 2.39% in no-bake. Switching processes on an existing pattern means re-checking every toleranced dimension, not just re-quoting.
Which Process Should You Choose?
Choose green sand for a casting under a few hundred pounds that will run in enough volume to pay for a metal pattern and keep a molding machine busy; choose no-bake for anything large, heavily cored, or short-run, and for any casting above about 200 lb whose tolerances are tight.
Green sand wins on price when the machine is running. A gray iron counterweight or a ductile iron bearing housing that repeats every month belongs on a matchplate in green sand: the mold costs almost nothing in materials, the sand recirculates without a reclaimer, and 65 molds an hour from one automatic machine amortizes the metal pattern quickly. Below the 200-lb line the tolerances are no worse than no-bake would give.
No-bake wins when the mold has to do work a machine cannot. A carbon steel valve body with a deep core, a gearbox housing with multiple cored passages, or a pattern that comes off the shelf a few times a year all favor a rigid cured mold: it holds heavy sections without erosion, its wooden pattern costs less and arrives sooner, and above roughly 200 lb it holds tighter dimensions. The sand costs more per mold, but at low volume the tooling saving usually outweighs it.
The volume line between the two is judgment, not a table value; it moves with part size and with what the foundry already has on the floor. For a part near the line, quote it both ways and check two things on the drawing — whether any tight feature crosses the parting line, and whether an existing pattern’s shrink allowance was cut for the process now being quoted.
Conclusion
Size and volume settle this choice. Below 50 lb the two processes hold similar tolerances and green sand is cheaper at volume; above 200 lb no-bake holds tighter dimensions, tolerates wooden tooling, and carries the heavy sections, with parting-line scatter as its one dimensional weakness — orient tight features into one half and it is managed. In between, the foundry’s process control matters more than its binder, so quote both and judge the foundry.