How to Vet a Casting Foundry: 12 Questions That Separate Quotes

Vet a metal casting foundry in three passes: confirm process and part fit, observe how the shop controls the proposed production route, then require a sample and first-order release that freeze the production baseline. Compare quotes only after every supplier answers the same technical and commercial scope.

Most supplier searches become unreliable at the same point. Buyers place several attractive websites, brochures and quotations side by side, then treat the lowest total as the strongest offer. The documents may describe different tooling scopes, inspection levels, machining responsibilities and change rights. Price hides them.

Three passes give the question list a job. The first four questions decide whether an audit is worth the time. The next four test whether the proposed control works on the shop floor. The final four turn a sample into a first-order release. An unanswered gate becomes a hold or no-go.

Vet the Buying Task Before You Vet the Foundry

Start with one controlled description of the part and the buying task. Comparability collapses when one quote is based on a sketch, another on an old drawing, and a third on a sample from a different revision. Buyers need a controlled baseline before the supplier search begins.

• Baseline rule: every quote must use the same revision and scope.

For an OEM or ODM casting project, HULK Metal works from at least a 2D drawing or a physical sample, then adds material, quantity and delivery requirements as the review develops. That intake rule matters here because it separates a useful supplier conversation from a price guess. If the foundry can quote without seeing any controlled geometry, ask what assumptions are hidden in the price.

Comparison follows the same rule. Ask every supplier for the process route, included operations, exclusions, inspection documents and tooling ownership using the same line items. A foundry may add useful recommendations, but the final comparison must return to one buyer-controlled scope.

Decision rule: If the process route, material family, size/quantity class or change-control owner is unresolved, do not book an audit. Fix the buying task first. A site visit cannot repair an uncomparable request.

What a Useful First Answer Looks Like

A capable answer names the route and its boundaries. It should say whether the part fits sand casting, investment casting, die casting or another process, which material condition is assumed, which operations sit inside and outside the quote, and which input would overturn the first view. Generic confidence is not enough.

• → Map the answer back to the same RFQ line. The supplier may challenge the requirement, but it should state the change and its consequence.
• ◆ Separate process questions from commercial questions. A capable process route can still have an incomplete quote.
• ○ Record missing information as an open condition, not as a defect hidden inside the unit price.

First Pass: Four Questions That Decide Whether a Visit Is Worthwhile

Triage comes first, not an award decision. It checks whether the foundry normally works in the material, process, size class and quantity range of the project. Relevant capability should come with a route explanation and a named engineering contact.

Confirm the Working Process, Not the Catalogue

Catalogue breadth can hide the route actually used for the quoted part. Ask what most of the foundry's current work looks like. Then ask why the proposed process suits this geometry, alloy and quantity. The explanation should mention at least one tradeoff, such as tooling commitment, pattern cost, surface condition, internal core complexity or machining allowance.

Boundaries strengthen an answer. For example, a supplier reviewing a thin-wall part may ask for a different process, a design change or a stronger inspection method. That boundary is more informative than a broad capability statement because it shows where the quotation stops being reliable.

Put the Part in the Supplier's Real Operating Range

Send the drawing or sample with the material, annual quantity, release batch, critical features and required finish. Ask the supplier to place the part against its normal work. Do not ask only, "Can you make it?" Ask which part characteristic drives the process choice, which operation is subcontracted and what evidence will show that the subcontract scope remains controlled.

• Range test: a useful reply separates routine work from an exception and names the operation that creates the exception.

The response should distinguish routine work from an exception. Unusual size, alloy or quantity can still be feasible, but it deserves a different risk review. If the foundry answers every part with the same confidence, request a more specific route before arranging the audit.

Make Engineering Review Visible Before the Quote

Engineering review is the bridge between the buyer's function requirements and the foundry's process. Ask which person or team reviews the drawing, what issues they normally raise, and how their comments reach the quotation. The buyer needs to see whether manufacturability was considered before the price was fixed.

Written DFM feedback is enough at this stage. It should name the assumptions, the requested clarifications and any design or process change that would alter tooling, inspection or cost. The final article does not have to repeat every DFM point. It must show that the supplier can explain the route it proposes.

Settle Tooling and Change Ownership

Tooling questions are not limited to who pays the first invoice. Establish who owns the pattern or die, where it is stored, how wear is checked, which party can move it, how a drawing revision is controlled and who approves a repair. These details determine what happens when the part changes or the relationship ends.

Ask for the practical change path. If the sample reveals a problem, who decides whether the tool is corrected, replaced or accepted with a deviation? If the buyer changes a machining feature, does the casting tool change, the fixture change, or both? A supplier that can describe those branches is easier to manage than one that answers only with a tooling price.

Second Pass: Four Questions for the Shop-Floor Audit

Control evidence comes before the tour narrative. An audit is useful when it follows one representative part, not when it becomes a tour of equipment logos. Ask to see the pattern or tool condition, the process record, the measurement setup and a recent nonconformance. The questions work on site or by video.

Audit rule: Follow one part and one recent problem. A collection of certificates cannot show whether the shop can hold suspect material, trace a measurement or stop a nonconforming batch.

Table 1. Convert an audit observation into a decision instead of a general impression.

Third Pass: Four Questions That Turn a Sample Into a Production Baseline

Samples earn their place when they answer a defined question and leaves a record that production can repeat. A cosmetic photo may show that a part exists. It does not show the drawing revision, material condition, process route, measured features, packaging basis or quantity basis behind the approval.

• Sample rule: state the approval claim before the part is made.

Define the Sample's Limited Claim

Validation should prove the agreed questions, not everything about the supplier. Casting sample validation can include dimensional checks, chemical analysis, mechanical or hardness testing, surface inspection, machining verification, assembly, pressure testing, coating checks and functional testing where relevant. Inspection depth depends on the part, not on a universal checklist.

Attach the sample to a record. It should identify the drawing revision, material condition, production route, inspection standard and packaging or quantity basis. If a post-sample change affects any of those items, reopen the review before volume release.

Treat Sampling as a Limited Lot Decision

Acceptance sampling selects a sample from a lot and uses the result to decide whether the lot is likely to be acceptable. NIST also notes that the purpose is lot disposition, not a full estimate of every part. That distinction matters when a supplier says a few good castings prove the whole shipment.

Ask what population the sample represents, how the pieces were selected, which characteristics were measured, what rule separated accept from reject, and what happens to a failed lot. For critical features, the agreed plan may require a different level of inspection. The plan should state that choice rather than leave it to habit.

Do not let one clean sample erase lot risk. Sampling provides evidence about a defined decision. It does not replace process control, traceability or disposition of failed parts.

Normalize the Quote Before Comparing the Total

• Quote normalization: compare technical fit and control depth first. Put every supplier on the same line items for tooling, sample, material, casting, machining, finishing, inspection, documents, packing and freight. Ask who owns the tool and what happens if volume changes.

Two prices are not comparable when one includes first article inspection and the other assumes later inspection, or when one supplier machines a sealing face and another leaves it as-cast. A commercial adjustment is not evidence that the technical route changed, and a low total is not proof of a lower total cost.

Table 2. Normalize quotes before ranking price.

Define the Repeat-Order Reopen Rule

A repeat order is not automatically the same order. Tool wear, pattern repair, material supply, subcontracting or a drawing revision can change the risk. Ask which change events force a new sample, a new inspection plan or a new commercial review.

Before shipment, write the repeat-order rule. A buyer should know whether a minor tool repair is accepted internally, whether a material heat change is routine, and which process changes require customer notice. The rule matters most when the first order is designed as a qualification lot rather than a full production release.

Compare Answers Without Letting a Low Price Win

Score suppliers in evidence order. Process fit comes first, shop control second, sample and lot evidence third, commercial scope last. Failure at an earlier gate does not earn a second chance merely because its unit price is lower.

Table 3. A compact scorecard that keeps evidence ahead of price.

Do not score a supplier on how many files it sends. One traceable record can be stronger than a thick unlinked folder. Look for the relationship among the drawing, material, process, measured result, disposition and packed identity. If those links cannot be followed, the buyer still does not know what the evidence covers.

The First Order Is a Controlled Test, Not a Lottery

First-order testing should exercise the production system the supplier described, not a special route assembled for the buyer's visit. It is a controlled step that moves the relationship from claims to repeatable evidence. Buyers need a clear answer for what will be measured, what will be reviewed and what will stop the next release.

First-order stop rules: stop release when the drawing revision is unclear, the tool changed without review, the material identity cannot be linked, a required inspection record is missing, or the packing and shipping quantity cannot be reconciled.

Table 4. What the first order can and cannot prove.

Red Flags and Fair Exclusions

Some supplier differences are legitimate. A prototype may use a temporary pattern. A low-risk part may need a lighter inspection plan. A domestic delivery may not need an export document set. The exclusion becomes acceptable when it is named, agreed and carried into the release record.

• ■ A quote exists, but no one can state the drawing revision or process route it uses.
• ● The supplier can show a certificate but cannot connect it to the part, material heat, inspection result and shipment.
• ★ Tooling ownership, storage or change rights remain unstated after the buyer asks directly.
• ○ A failed check is handled only by rework, with no containment or follow-up rule.
• ▲ The supplier presents sampling as a guarantee for every piece in the lot.

Fair exclusion: If a test is not required, record why and who approved the omission. Silence is not the same as a valid exclusion.

One useful audit question is rarely on a brochure: ask the supplier to explain a recent change without naming the customer. A clear answer shows the chain from request to drawing revision, tool action, sample decision and price effect. A vague answer is a warning even when the final part photo looks correct.

What to Put in the RFQ Before the Audit

RFQ scope lets every supplier respond to the same task. It should include the drawing or sample, material requirement, application context, annual and batch quantity, critical features, inspection expectations, finish, packing and delivery terms. Add the questions that will decide your audit so the supplier knows what evidence will be discussed.

0 of 8 fixed

Questions Buyers Usually Ask Too Late

Send the controlled drawing or sample with the material, application, quantities, critical characteristics, finish, inspection, packing and delivery requirements together. HULK Metal coordinates engineering review, specialized production, inspection, finishing, packing and documentation around that project baseline. Scope that is not yet known should be stated as an open question rather than assumed.

The Vetting Decision

Good vetting does not try to prove that one supplier is perfect. It proves whether the proposed supplier can be controlled well enough for this part at this buying stage. The first four questions protect the buyer from an irrelevant audit. The next four expose the operating evidence behind the claim. The final four turn the sample and quote into a production baseline.

If the evidence stops at a brochure, the decision is not ready. Traceability, explained controls, a normalized quote and a bounded first order give the buyer something stronger than a low price: a repeatable basis for the next decision.

Chapter Index

Source boundary: ISO/TC 176/SC 2 describes ISO 9001 as a generic quality-management-system standard. NIST/SEMATECH e-Handbook sections 6.2.1 and 2.3.1 explain acceptance sampling and calibration limits. HULK Metal project facts come from the Casting knowledge series. Standards and technical references support method boundaries; they do not replace the buyer's drawing, specification and agreed acceptance record.