Choosing the wrong mold type leaves you choosing between scrap and profit. This guide breaks down the five main casting mold types (sand, investment, die, permanent metal, ceramic shell), showing exactly when each becomes economical by production volume.
Table of Contents
The mold you choose determines your part's final tolerances, surface finish, and production lead time. Yet most manufacturing buyers choose mold types for metal casting on cost alone - ignoring the technical fit, lead time, and supplier verification that separate working parts from scrap bins.
This guide breaks down mold types by production volume requirements, shows you exactly when each mold type is appropriate, and explains how to vet mold makers without guessing. We've analyzed mold sourcing across 800,000+ factories so you don't have to.
Looking to get started? Submit your casting requirements to Haizol and receive quotes from verified foundries within 24 hours.
A casting mold is the negative form that shapes molten metal into a finished or near-finished part. The mold cavity defines the part's final geometry, tolerance, and surface quality - get the mold wrong, and you're pouring money into scrap.
Metal casting has been used for 6,000+ years, but modern molds have become engineered systems. The mold's material, porosity, thermal properties, and surface preparation all influence whether your casting comes out dimensional or warped.
This is why "getting a quote" for a mold means understanding not just price, but what you're paying for: how the mold handles heat, how fast it cools the metal, and how many times it'll survive reuse. For a deeper dive into how castings become finished cast metal parts, read our guide to metal casting processes.
Manufacturing buyers underestimate mold selection because it feels technical. It's not. It's a volume and budget decision dressed up in metallurgy.
Metal casting molds split into two families: expendable (used once, then discarded) and non-expendable (reused across multiple production runs).
Expendable molds include sand, investment (ceramic shell), and plaster - each burns away or falls apart after the casting cools. Non-expendable molds include permanent metal molds and die casting molds - they're pressed steel or ductile iron, designed to survive hundreds or thousands of production cycles.
Your production volume almost entirely determines which family fits your budget.
Here's the quick picture:
|
Mold Type |
Material |
Reusable? |
Lead Time |
Tolerances |
Surface Finish |
Best For |
|
Sand |
Silica sand + clay binder |
No (expendable) |
2-4 weeks |
±0.5-1 mm |
Fair (rough, Ra 6.3-25 μm) |
Large parts, low volume (< 10K) |
|
Investment |
Ceramic shell |
No (expendable) |
3-6 weeks |
±0.1-0.2 mm |
Excellent (Ra 1.6 μm) |
Complex geometries, tight tolerances (100-50K) |
|
Permanent Metal |
Cast iron or steel |
Yes |
4-8 weeks |
±0.1-0.3 mm |
Good (Ra 3.2-6.3 μm) |
Medium volume repeatable runs (1K-50K) |
|
Die Casting |
Hardened steel |
Yes (100K+ cycles) |
6-12 weeks |
±0.05 mm |
Excellent (Ra 1.6-3.2 μm) |
High-volume precision (100K+) |
|
Ceramic Shell |
Ceramic coatings |
No (expendable) |
3-5 weeks |
±0.1 mm |
Excellent (Ra 1.6 μm) |
Aerospace, medical, jewelry |
Sand casting molds are the oldest and most forgiving casting method. You pack silica sand mixed with a clay binder around a pattern (wooden or 3D printed), remove the pattern, and you've got your mold cavity. Simple.
What you get: Tolerances of ±0.5-1 mm are typical (ISO 8062-3 CT11-CT14 grade). Surface finish is rough (Ra 6.3-25 μm), meaning you'll almost always need secondary machining. Venting and gate design matter enormously - bad venting creates porosity defects; bad gating creates cold shuts and misruns.
When to choose sand: Your part is large, your geometry is simple, and your volume is under 10,000 pieces. Sand handles thick-walled castings beautifully. It's also the only economical choice if your part is huge (aerospace forgings, heavy machinery frames).
Red flags: If a supplier quotes sand molds without discussing vent placement or gating, they're guessing. Ask them how they'll prevent shrinkage porosity - the answer tells you their metallurgy experience.
Investment casting, also called lost-wax casting, uses a ceramic shell mold. You 3D print or machine a wax pattern, coat it in ceramic layers, burn the wax out, and pour molten metal into the hollow ceramic shell. The ceramic is "sacrificed" - hence "expendable."
Why you'd pick this: Investment casting delivers ±0.1-0.2 mm tolerances (ISO 8062-3 CT5-CT9 grade) and Ra 1.6 μm finishes in a single pour. Thin-walled complex parts that would cost a fortune to machine? Investment casting handles them economically.
Common defects: Ceramic shell cracking (if thermal cycling is too fast), swelling (if the shell reheats unevenly), and wax bleed-through (if wax residue cakes on the shell before dewaxing). Preventing these requires controlled furnace ramps and careful shell strength management.
When to choose investment: You need tight tolerances, thin walls, or complex internal passages. Medical implants, aerospace turbine blades, jewelry - investment casting wins. Volume should be 100-50,000 pieces; below 100, you're dealing with higher per-piece overhead; above 50,000, die casting becomes more practical at scale.
Die casting forces molten metal under high pressure (1,500-15,000 psi) into a hardened steel mold. The metal cools fast, the mold opens, and your part ejects. Die casting molds survive 100,000+ cycles because they're engineered for repeated shock and thermal cycling.
What you get: ±0.05 mm tolerances on suitable features (ISO 8062-3 CT1-CT4 grade). Surface finish Ra 1.6-3.2 μm. Repeatability across 100,000 cycles with minimal variation. You're paying for consistency.
When to choose die casting: Volume is 100,000+ pieces, part geometry is medium complexity, and you're casting non-ferrous metals (aluminum, zinc, magnesium). Die casting is the only economical path at scale.
Catch: Die casting molds are expensive and inflexible. If your design changes mid-production, you're scrapping the mold. Get design-for-manufacturability (DFM) right before committing.
Ceramic shell molds are investment casting's sophisticated cousin. Instead of a single ceramic shell, you coat a wax or plastic pattern in dozens of ceramic layers, each precisely controlled for thickness and porosity. The result is superior shell strength and surface finish. Often used for aerospace and medical casting where failure isn't acceptable.
Why choose it: Ceramic shells survive thinner patterns and more complex internal geometry than standard investment shells. Aerospace parts with tight tolerances and zero tolerance for defects. Medical devices where every microgram of material counts.
When to avoid: Skip ceramic shells if your part geometry or tolerance requirements can be met with sand or investment casting alone. Don't over-specify - choose ceramic shells only when the technical requirements demand the added precision.
A permanent mold is a steel or cast iron block with a cavity, used repeatedly with gravity-feed casting (metal poured in by gravity, not forced by hydraulic pressure like die casting). The mold is preheated, metal is poured, it cools, the mold opens. Repeat. Because the mold survives 1,000-50,000 cycles, the per-unit overhead is favorable for mid-volume production.
When to choose it: You're casting non-ferrous metals (aluminum, bronze), volume is 1,000-50,000 pieces, and you need better tolerances than sand but don't have the volume to justify die casting. Permanent molds are the "goldilocks" option.
Trade-off: Surface finish (Ra 3.2-6.3 μm, per ISO 8062-3 CT9-CT12) is decent but not investment-casting excellent. You'll likely need some machining afterward.
3D printing has opened two paths to molds: printing wax patterns (which you then use for investment casting), and printing sand molds directly (no pattern needed - the 3D-printed sand IS the mold).
When to choose it: You're prototyping, your geometry is complex, and you can't wait 6 weeks for a traditional mold. 3D printing lets you iterate quickly. For production (100+ pieces), traditional molds become cheaper.
Catch: 3D-printed molds have lower durability and surface finish than traditionally cast molds. Sand printed at low resolution produces grainy surface finishes. Use 3D printed molds to validate design and then switch to traditional molds for production.
Here's the framework every manufacturing buyer should follow:
|
Part Geometry |
Tolerance Required |
Production Volume |
Budget |
Mold Recommendation |
|
Simple, large |
±1 mm |
< 5,000 |
Low |
Sand |
|
Complex, thin walls |
±0.2 mm |
100-10,000 |
Medium |
Investment |
|
Simple, medium |
±0.5 mm |
1,000-50,000 |
Medium |
Permanent Metal |
|
Precision, high volume |
±0.05 mm |
> 100,000 |
Medium-High |
Die Casting |
|
Aerospace/Medical complexity |
±0.1 mm |
100-50,000 |
High |
Ceramic Shell |
|
Situation |
Recommended Mold Type |
Why |
|
Large parts, simple geometry, < 5,000 pieces |
Sand |
Fastest lead time, lowest tooling complexity |
|
Complex thin-walled parts, 100-10,000 pieces |
Investment |
Tight tolerances (±0.1-0.2 mm), excellent surface finish |
|
Medium-complexity parts, 1,000-50,000 pieces |
Permanent Metal |
Reusable mold spreads investment across multiple batches |
|
Precision high-volume, > 100,000 pieces |
Die Casting |
Repeatable ±0.05 mm tolerances, fastest cycle time at scale |
|
Aerospace/medical/high-reliability parts, any volume |
Ceramic Shell |
Superior precision and zero-defect capability; use when failure isn't acceptable |
Every casting process carries specific defect risks that stem from metal flow behaviour, thermal gradients, and mold material properties. Understanding these failure modes before production starts is what separates buyers who get first-article approval from those who cycle through expensive rework iterations. The defects below are the most common across sand, investment, and die casting processes, along with the prevention strategies verified foundries apply.
Sand casting defects:
Investment casting defects:
Die casting defects:
How mold makers prevent defects: Ask suppliers about their casting simulation software. Modern foundries use casting simulation to predict porosity, shrinkage, and flow patterns before cutting a mold. If they're not simulating, they're guessing.
Most manufacturing buyers have never sourced a mold. They call local machine shops, get vague quotes, and hope for the best. Don't do that.
Red flags in foundries:
Questions to ask every mold maker:
How to vet multiple suppliers: This is where most European buyers fail. They get one quote, assume it's fair, and award the job. Get at least 3 quotes. Compare not just price, but:
Haizol difference: Instead of cold-calling foundries individually, submit your casting requirements on Haizol's platform and get structured quotes from 8+ verified foundries within 24 hours. Each quote includes the supplier's equipment list, certifications, capacity, and past work samples - side by side, so you compare apples to apples. Review the top die-casting companies
Different industries place fundamentally different demands on casting quality, certification scope, and allowable defect rates, which means the right mold type for one sector may be entirely unsuitable for another. Selecting the correct casting process for your industry vertical before shortlisting suppliers saves significant time and prevents you from evaluating foundries that cannot meet your regulatory or tolerance requirements.
Automotive (OEM & Tier 1):
Medical Devices:
Aerospace & Defense:
Industrial Machinery & Heavy Equipment:
Yes. Use a temporary 3D-printed mold or a short-run aluminum mold (cheaper than steel). Validate your design, then switch to a production mold. This costs 20% more upfront but avoids the catastrophe of discovering a design flaw in your steel die mold.
You need a second mold. Many foundries keep molds on file, so you can order a second cavity without waiting weeks. Plan for this in your cost estimates if you expect volume to grow.
Sand molds: one use (expendable). Investment molds: one use (expendable). Permanent metal molds: 1,000-50,000 cycles before wearing. Die casting molds: 100,000-1,000,000 cycles. After that, cavity walls erode and tolerances drift. Refresh molds by polishing or cavity relining.
Not always. Haizol's platform supports three NDA levels: no NDA (for non-sensitive parts), standard platform NDA, or custom buyer NDA. Choose based on your IP sensitivity. This protects your drawings from the moment they're uploaded.
Yes. A test casting is cheap insurance - use it to validate your mold before you're committed to 10,000 pieces. Most foundries recommend test castings to catch design flaws and validate mold performance before full production.
Sometimes. If the mold is yours (paid for outright), you can transfer it. If the supplier owns it, you'll need to get a new mold or negotiate ownership transfer. Always clarify mold ownership in your purchase agreement.
Sand castings typically need post-machining to reduce surface roughness. Investment castings may need light finishing. Die castings usually require minimal finishing because they're designed to eject near-net-shape. Discuss secondary operation requirements with your mold maker during design review.
Mold selection comes down to three decisions: define your production volume, match the mold type to your tolerance requirement, and verify your supplier holds the certifications your industry requires. Most buyers skip that last step, choose on price alone, and end up with a foundry that cuts corners on gating design or skips simulation entirely.
Haizol connects you with verified foundries that have the equipment, certifications, and process capability documented before you commit. Create a free buyer account to get started, or submit your part drawing directly and receive structured quotes from multiple verified foundries within 24 hours, each including equipment lists, certifications, and capacity data so you are comparing real capability, not just price.
Join Haizol for free - Asia’s leading custom manufacturing marketplace. Connect with over 800,000 suppliers and get multiple quotes with one request.