Most foundry evaluations are weak because they focus on the easy paperwork: certifications, price sheets, and polished capability decks. That is check-the-box sourcing, and it is how bad castings get approved.
If you want to know whether a foundry can actually support a production program, you need to get into the engineering controls that determine whether the casting will fill, feed, hold dimension, and survive downstream machining without turning your launch into a containment exercise. The foundries that treat process engineering as a sales accessory are the same foundries where porosity shows up after machining and dimensional drift gets explained away as “normal casting variation.” The foundries that treat engineering discipline as a production control are the ones that keep scrap, rework, and customer escalation under control.
For cast components, three technical areas tell you a lot very quickly: solidification simulation, gating and risering design, and dimensional verification. If those three areas are vague, the rest of the quote package is mostly theater.
1. Solidification Simulation: MAGMAsoft Should Drive Decisions, Not Decorate Presentations
A lot of suppliers say they “use simulation.” That statement means nothing by itself.
The real question is whether the foundry uses a platform like MAGMAsoft to make tooling and process decisions before steel ever hits the mold. Serious simulation work should predict where the process is vulnerable to:
- Shrinkage porosity
- Micro-porosity in heavy sections
- Hot spots and poor feeding behavior
- Misruns and cold shuts
- Air entrapment and turbulence
- Uneven solidification and distortion risk
- Yield loss from inefficient feeding design
Ask: “Show me the MAGMAsoft results for this geometry, what defects were predicted, what changes were made to the gating and risers, and what improved afterward.”
That question matters because simulation is only useful if it changes the process. A colorful image of solidification timing does not protect your program unless the foundry can explain what design decisions came out of it.
What you should expect from a capable foundry:
- A defined simulation workflow before tooling release
- Clear thermal maps showing hot spots and feeding paths
- Iteration history, not just one final screenshot
- Evidence that riser placement, ingate location, and section transitions were adjusted based on results
- Correlation between simulated risk areas and actual tryout or production findings
Weak suppliers use simulation after defects appear, mainly to explain failure. Strong suppliers use it before launch to prevent failure.
The business consequence is straightforward. If the foundry cannot model solidification risk with discipline, then you are paying to discover process limits through scrap, delayed PPAP, machining fallout, and field exposure. That is not engineering. That is trial-and-error with your budget.
2. Gating and Risering Design: This Is Where Foundry Competence Stops Being Abstract
Bad gating and risering design does not always show up in the first conversation. It shows up later as inconsistent yield, localized shrink, inclusions, unstable fill behavior, and arguments about whether the defect is “acceptable.”
A competent foundry should be able to explain, in practical terms, why the metal enters where it enters, how the cavity fills, how turbulence is controlled, and how the part feeds during solidification. If they cannot explain that clearly, they probably do not control it as tightly as they claim.
Your review should cover:
- Ingate location and count
- Runner layout and flow balance
- Riser size, neck design, and placement
- Feeding path into heavy sections
- Chill strategy, where applicable
- Yield impact versus defect prevention
- Expected areas of last solidification
- Controls for oxide generation and mold erosion
Ask: “Where is the last area to solidify, how are you feeding it, and what evidence shows the riser design is adequate under production conditions?”
Ask: “What gating changes would you make if machining later exposes subsurface porosity in this section?”
Those are not academic questions. They expose whether the supplier has an engineering basis for the process or whether the design evolved through habit.
The shops that treat gating and risering as a standard template are the same shops where one part number after another develops repeatable defects in different locations. Whereas the shops that treat gating and feeding as geometry-specific engineering are the ones that improve yield without pushing hidden defect risk downstream.
At ICT, we evaluate foundries on whether they can defend process design with engineering evidence, not just operator experience. Experience matters. But experience without analysis is how old problems keep getting renamed instead of fixed.
3. Dimensional Verification Protocols: If the Casting Cannot Be Measured Reliably, It Cannot Be Managed Reliably
Plenty of foundries will say they “inspect dimensions.” That is not enough. You need to know how, when, against what datums, and what happens when the result drifts.
Dimensional verification for castings is not a clerical exercise. Cast parts move. Datums can be unstable. Machining stock can disappear in one area and pile up in another. If the inspection plan is weak, defects are often discovered after machining, after assembly, or after the parts are already on the water.
A serious dimensional verification protocol should include:
- Ballooned drawing review tied to critical and special characteristics
- Defined datums appropriate for the as-cast condition
- First article dimensional layout
- Fixture strategy for repeatable measurement
- CMM reporting where geometry justifies it
- Attribute checks for high-volume repeat features
- Stock verification on machined-transition areas
- Gage calibration and measurement system discipline
- Reaction plans for out-of-tolerance conditions
Ask: “Show me the first article report, the datum strategy, and the dimensional results on the features most likely to affect machining and assembly.”
Ask: “Which dimensions are checked on every lot, which are audited, and what triggers containment before shipment?”
Those questions separate suppliers who inspect for appearance from suppliers who inspect for function.
For OEM and Tier suppliers, dimensional verification also has to connect to APQP and PPAP discipline. If the foundry cannot produce repeatable dimensional evidence, then your PPAP package gets weak, your receiving controls get heavier, and your plant ends up doing supplier development work it should not have to do.
ICT’s quality model includes dimensional verification before parts ship, supplier oversight, and APQP/PPAP support where required. That matters because global sourcing only works when engineering evidence travels with the part.
What to Require From a Casting Foundry Before You Approve the Program
Before moving forward, require objective proof in these three areas:
- Simulation evidence showing how MAGMAsoft or equivalent analysis was used to reduce shrink, porosity, and fill risk
- Gating and risering rationale tied to the actual geometry, section thickness, and feeding sequence
- Dimensional verification protocol covering first article, in-process checks, final inspection, and reaction plans
If a supplier cannot produce those basics with confidence, you are not reviewing a controlled foundry process. You are reviewing a quoting exercise.
The strategic point is simple: foundry performance is not decided by who talks best in the RFQ phase. It is decided by who can control solidification, feeding, and dimensional stability before failure becomes your problem.
If you are sourcing castings and need a manufacturing partner that evaluates foundries through real engineering controls, ICT can help qualify suppliers, review process risk, and manage the path from quote through production.
Submit an RFQ to ICT or contact our team.
IN Consulting and Trade
Website: inconsultingandtrade.com
Email: mmusleh@inconsultingandtrade.com
Phone: 765 413 4188
LinkedIn: Michael Musleh
Facebook: Indiana Consulting and Trade
Instagram: @inconsultingandtrade_
Twitter/X: @inconsultingand