Eighteen months ago, a Texas contractor called me about a shipment of excavator bucket teeth that was falling apart. Fourteen of sixty teeth had visible cracks by week six. The machine sat idle at a Midland job site, burning roughly $800 an hour in downtime plus the cost of re-mobilizing a three-man crew. When he pulled the material certs, the “ASTM A128 Grade B-3” manganese steel he’d paid for didn’t match the spectro analysis. Carbon was high, manganese was low, and there was no evidence the castings had ever been water-quenched.

The supplier? A trading company in eastern China that listed itself as a heavy equipment castings manufacturer, quoted 18% under the nearest real factory, and answered every question in flawless English within forty minutes — at 2 a.m. Beijing time.

This article is the vetting process I’ve used for over two decades of buying castings for construction and mining equipment. It takes about fifteen minutes to read, and it can save you six figures. Here’s what to check before you send a single dollar.

A “castings manufacturer” label proves nothing

Spend an afternoon searching “heavy equipment castings manufacturer” and you’ll see why this article exists. Hundreds of websites, dozens of logos, and nearly all of them claim to be “a professional foundry with 20 years of experience.” A meaningful share of those sites are trading companies — offices with laptops and good English that forward your inquiry to a real factory somewhere else and add a margin on top.

Trading companies aren’t automatically bad. A good agent handles inspection, logistics, and communication, and that service costs money either way. The problem is when a trader pretends to be the factory. Because when a casting fails, the trader’s answer is “let me ask the factory” — and you suddenly realize you have zero visibility into how your parts were actually melted, poured, and heat treated. You bought the story, not the steel.

You can verify who you’re talking to in under 30 minutes. Ask for a live video walkthrough of the production floor — not a recorded tour, live, where they show you the melting control panel, the sand system, and the heat-treatment record sheets. A trader can’t do this because they don’t have a shop floor. Get their business license and check the registered scope: does it actually cover casting production, or only import and export? Then ask what their melting capacity is, in tonnes per day, and which alloys they pour. A real foundry answers instantly with specifics. A trader paraphrases a brochure.

Another common trick: the “factory tour” video. Traders send a polished, YouTube-style walkthrough of a foundry floor — often shot at a completely different company, sometimes bought from stock footage. Reverse-image search the still frames. I’ve found the same “factory” video used by three different “manufacturers” in three different provinces. If the video looks too cinematic, it probably is.

One more tell: ask how they control manganese when they pour A128 manganese steel. A foundry that actually melts this alloy will talk about spectro verification at the furnace, tap temperature, and why manganese content drifts if you’re not careful. A trader will say they’ll check with the engineer. You want a direct line to the engineer, not a polite messenger.

The price you pay isn’t the price you pay

Let’s run the numbers on that Texas contractor. He paid $21 per bucket tooth from the trader. A direct factory quote for the same spec would have run roughly $26 to $30. On sixty teeth, that’s a difference of $300 to $540 — pocket change next to what the failure actually cost. The excavator idled for the better part of a week before replacement teeth arrived. At $800 an hour, forty hours of downtime is $32,000. Add freight, labor for swapping the broken teeth, and the near-loss of the client on that Midland job.

Buyers compare the wrong numbers. Unit price is the number on the invoice. Landed cost includes freight, inspection, and financing. But failure cost is the number that actually matters, and it can be fifty to a hundred times the unit-price difference. A bucket tooth that costs $21 and survives 40% of its expected life is the most expensive tooth you’ll ever buy.

There are quieter hidden costs too. Rework when hard spots in a casting destroy machine-shop tooling. Back charges when your component fails an OEM’s spec. Product liability — in earthmoving and agriculture, a broken cast part can injure or kill someone. Cheap castings aren’t a bargain. They’re a liability with a shipping label.

Inspection is another line item buyers forget. Third-party inspection of a casting shipment typically runs $300 to $800 per day depending on the port and the inspector. Smart buyers build that cost into the deal and make payment conditional on the inspection report. If a supplier pushes back hard on third-party inspection, that’s a signal the parts won’t pass.

Know your alloys before you open negotiations

“Heavy equipment castings” covers everything from a 30-kg bucket tooth to a 14-ton counterweight. Those parts do not share the same metallurgy, and the foundry that’s excellent at one can be terrible at the other. Before you source anything, know which family your part belongs to:

• Bucket teeth and wear parts — austenitic manganese steel, typically ASTM A128 Grades B-3 or C. These are water-quenched from around 1,050°C (1,900°F), and their toughness comes from work-hardening in service. Heat treatment is where cheap foundries cut corners, which is exactly what happened to that Texas contractor’s teeth.
• Structural castings — linkages, boom components, suspension housings. Usually ASTM A148 Grade 80-50 or 90-60 low-alloy steel, or ASTM A27 Grade 65-35 carbon steel. These need consistent chemistry and weldability, not just raw strength.
• Counterweights — heavy, simple, often plain carbon steel or ductile iron per ASTM A536 65-45-12. They’re cheap to cast and expensive to ship. For these, weight and chemistry matter more than exotic metallurgy.
• Gear and axle housings — ductile iron or cast steel, often requiring pressure testing, tight machining tolerances, and careful control of internal soundness.

Heat treatment deserves its own paragraph because it’s the most commonly faked step in the whole process. Manganese steel has to be fully austenitized, then quenched fast enough to keep carbon in solution. Skip the quench and the casting looks right on the surface but wears out at a fraction of the expected rate. Ductile iron housings need the right annealing or normalizing cycle to hit machinable hardness. Put heat treatment in the contract: furnace chart records due with the shipment, no exceptions.

The point isn’t to give you a metallurgy exam. It’s to help you spot a mismatch before you waste a month of lead time. If you’re sourcing manganese steel bucket teeth, the foundry needs a melt line that can handle manganese steel and real water-quenching capability. A shop that only pours gray iron counterweights has no business quoting your teeth. If it quotes anyway, quote someone else.

And ask for raw spectro reports, not just a certificate that says “ASTM A128 Grade B-3.” The certificate is a piece of paper. The spectro data is the actual chemistry from the heat. If a supplier won’t share it before the order, what makes you think you’ll get it after?

Our sourcing guides go deeper on alloy-specific requirements — including the exact chemistry limits you should put in your RFQ.

Where heavy equipment castings actually come from

Global foundry output runs about 100 million tonnes a year (World Foundry Organization data). China produces roughly half of it — on the order of 50 million tonnes in recent years. India follows at roughly 10%, and the United States, Germany, and Japan are each in the low single digits. For heavy-equipment-scale steel castings in volume, China still dominates, and that’s not changing soon.

If you’re serious about finding a real heavy equipment castings manufacturer, the region matters more than the country. Shandong province alone has over 1,200 registered foundries. Jiangsu — Wuxi, Changzhou, Xuzhou — is the cluster around construction equipment OEMs like Sany and XCMG. Liaoning and Hebei are strong for large steel castings because they sit close to steel mills and ports. The cluster effect matters: foundries sit near scrap yards, ferroalloy producers, refractory makers, and machine shops. That density is why Wuxi can turn a drawing into a finished, machined casting in weeks. It’s also why a “foundry” in a city with no industrial base is almost certainly an office with a website.

You can browse verified foundries in these regions — facilities we’ve already audited for melting capacity, heat-treatment capability, and NDT coverage.

Geography also drives lead time. A real casting-plus-machining program for a typical heavy equipment component runs 45 to 90 days. Add ocean freight and you’re at 60 to 110 days. You want one supply chain in one region, not a trader stitching together three foundries in three provinces. Every handoff is a risk you can’t inspect.

The 7-point vetting checklist

Run these seven checks on every candidate. If you’re short on time, checks 1, 3, and 5 catch the most failures.

1. Structure: factory or trader?

Business license, live video walkthrough, furnace specs. You already know the drill from the first section. Make it pass/fail — if they can’t show you a melting furnace on a live call, they’re not the manufacturer.

2. Certifications that actually matter

ISO 9001 is table stakes; nearly every export foundry has it, so don’t be impressed. Check that the scope covers steel castings for construction machinery, that the certificate is current, and that the last external audit produced no major findings. Ask to see the audit summary. A foundry that won’t share it usually has a reason. If they hold IATF 16949 — the automotive-grade quality standard — that’s a much stronger signal, because it means they survived far tougher customer audits.

3. Melting and alloy control

Ask to see melt logs. A proper melt log records furnace number, tap temperature, heat-treatment times, spectro verification, heat number, and pour date for every heat. Ask them to pull this week’s logs on the spot. A foundry with real process control hands you a binder or a screen in minutes. Then ask about raw materials: certified charge material, or whatever the scrap yard drops off? Your alloy starts in the charge, and if the charge is junk, the chemistry will be too. Also ask how they mark castings with heat numbers. If a finished part can’t be traced back to its heat, you’ll never be able to trace a failure to its cause.

4. NDT: who tests, and how much?

You need magnetic particle testing (MT) for surface cracks and ultrasonic testing (UT) for internal voids. Thick sections may require radiographic testing (RT). The real question isn’t “do you do NDT?” — it’s who runs the equipment, what certification they hold, and what percentage of castings get tested. ASNT Level II or ISO 9712 Level 2 is the industry baseline. For safety-critical parts, you want 100% testing, not sampling. If NDT is outsourced to a lab that visits twice a week, you’re not getting real coverage. And ask about acceptance criteria: ASTM E186 is the standard radiographic reference for steel castings. If a foundry quotes a looser internal standard, push back or walk.

5. Heat treatment: where the truth lives

Most mystery failures in heavy equipment castings trace back to bad heat treatment. Water-quenching manganese steel at the wrong temperature destroys its ability to work-harden. So ask directly: what furnaces do you run, do you have continuous temperature recording, and what’s your quenching setup for a 500-kg component? A foundry that can’t document soak times and temperatures is gambling with your parts.

6. Machining and finishing: in-house or outsourced?

A sand casting isn’t a finished part. Heavy equipment components need CNC machining, drilling, shot blasting, and painting or coating. Ask who machines and where. In-house machining signals real vertical integration. Outsourced machining is common, but it adds coordination risk — and a casting that travels to a different shop for machining is harder to trace when something goes wrong. For pressure-tested housings and tight-tolerance mating surfaces, you want to see the machine shop, not just a video. Shot blasting has a quality dimension too: the wrong media or excessive pressure can stress a surface and cause failures later. Ask what media they use and at what pressure.

7. References, track record, and defect data

Ask for existing customers in your industry — OEMs, not vague claims of “we export to the US.” Ask for their defective parts per million (PPM) over the last twelve months, and ask to speak with one customer in a similar time zone. A foundry with genuine export history answers without hesitation. “All customers are confidential” is a red flag, not a policy.

What you should expect to pay

Here are realistic export ranges for 2025-2026, ex-works:

• Carbon steel structural castings (ASTM A27/A148), simple geometry, machined and painted: $2.50–4.50 per kg
• Manganese steel wear parts (ASTM A128), heat treated: $4.50–8.00 per kg
• Ductile iron gear or axle housings, fully machined: $5–12 per kg depending on complexity and tolerances
• Tooling: $3,000–30,000 depending on part size and core count
• Typical MOQ: 100–500 pieces for smaller castings; 50 pieces can work for large parts

A quote that lands 25% below the bottom of these ranges means someone is cutting a corner — usually in chemistry, heat treatment, or inspection. The smart way to spend money is on a foundry that can sit with you for an hour and explain how it controls the alloy.

Red flags vs green flags

FAQ

Can I visit the foundry before ordering?

If you can, do. If you can’t — and many buyers can’t justify the trip for a first order — a live video walkthrough plus a third-party inspection by SGS, Bureau Veritas, or a trusted local inspector covers most of the value. For high-volume or safety-critical programs, send someone in person. It’s the cheapest insurance you’ll ever buy.

What’s the minimum order quantity?

MOQs range from about 50 pieces for large castings to 500 or more for small wear parts. But don’t negotiate MOQ before you’ve passed the vetting checklist. The lowest MOQ in the world is expensive if the foundry can’t confirm chemistry and heat treatment.

How do I verify quality when parts arrive?

Check the material certificate against the heat number stamped on the castings. Sample hardness with a Brinell or Rockwell tester, run spectro analysis if you have lab access, and verify critical dimensions. For safety-critical parts, pay for third-party inspection before shipment — it’s far cheaper than finding defects after arrival.

What’s a realistic lead time?

45 to 90 days for casting plus machining, plus 15 to 35 days of ocean freight. Add a 30% buffer on your first order. If a supplier quotes 20 days for a complex casting, run in the other direction.

Should I work with an agent or go direct to the foundry?

A good agent adds value on inspection, logistics, and communication. The best setup is direct access to the foundry engineer plus an independent third-party inspector. If you go through an agent, demand at least one direct conversation with the factory’s technical team before production starts.

How do I compare quotes fairly?

Normalize every quote to delivered cost per part, same Incoterm, same inspection scope, same heat-treatment documentation. A quote that excludes NDT and heat-treatment certification isn’t cheaper. It’s incomplete — and you’ll pay for the difference somewhere, usually in disappointment.

Next steps

Run the 7-point checklist before you send an RFQ, not after. The Texas contractor from the beginning of this article would have caught his trader on check #1 — a live video walkthrough would have shown no melting furnace, no sand line, and a meeting room with a projector.

When you have two or three real foundries on your shortlist, ask each for three things: raw spectro data from their most recent comparable heat, one customer reference in your industry, and a line-item quote that separates melting, heat treatment, NDT, machining, and packaging. Any supplier that can’t produce all three in seven days isn’t a manufacturer. It’s a middleman — and you already know how that story ends.