Industrial CT vs Radiography for Casting Inspection

Choose the casting question before choosing two or three dimensions

Industrial CT and radiography can both produce useful evidence about a casting, but the buying decision should begin with the engineering question. Does the customer need an examination against a defined product requirement, investigation of an internal feature, information about the distribution of indications or dimensional evidence? A more elaborate image is not automatically a better answer when the decision has not been specified.

For U.S. foundries, machining suppliers and product engineers, the scope should identify the casting family, material, geometry, production stage and governing customer requirements. Explain whether the item is a development sample, a first article, a production part or an investigation specimen. The same scanner and software may be used in several contexts, but the evidence and acceptance responsibilities can differ substantially.

This guide compares the decision path rather than quoting equipment prices or declaring one technique universally superior. It does not reproduce casting acceptance classes, prescribe exposure settings or authorize a product release. The responsible technical authority determines the applicable examination and evaluation basis. The buyer's task is to make sure the proposed service actually produces the information that basis requires.

Understand what a projection and a reconstructed volume represent

Conventional radiographic images show an attenuation projection through the object along the chosen view. Features at different depths can contribute to the same image region. CT combines multiple projections through a reconstruction process to provide volumetric information. These descriptions explain why the techniques can answer different questions; they do not establish achievable detection capability for a particular casting or configuration.

ASTM E1441's public description addresses CT as a method for reconstructing three-dimensional detail. A buyer interested in the location of an internal feature should still ask how the actual part geometry, material and scan configuration affect the evidence. A rendered volume is the output of an acquisition and reconstruction process, not an unqualified view of every feature inside the part.

For a projection-based proposal, ask which views are required and what information the examiner expects each to provide. For a CT proposal, ask which region is represented, how the reconstruction is reviewed and which limitations remain. Compare those answers against the intended decision rather than comparing a two-dimensional image and a rotating model solely by visual appeal.

Keep examination requirements distinct from product acceptance

A technique document describes aspects of how an examination is performed, while a drawing or product specification may establish what must be evaluated and accepted. Those roles should be explicit in the request. A supplier should not have to invent the casting acceptance basis because the purchase order says only inspect by X-ray. Identify the required documents and revisions before the production examination begins.

ASTM E1742/E1742M's public scope describes radiographic examination requirements and notes the need for a detailed procedure. That is a useful boundary for procurement: a reference to a standard alone is not the complete operating instruction or the entire product acceptance decision. The project must define the applicable technique, extent, evaluation and reporting requirements through its controlled documents.

Ask who may approve an alternative technique or resolve an ambiguous requirement. A CT result should not automatically replace a specified radiographic deliverable without the required agreement. Conversely, an investigation may legitimately need additional CT information beyond the production acceptance record. Keep that supplementary purpose visible so a development study is not mistaken for a formally accepted substitute.

Evaluate geometry, material and the region of interest together

Part dimensions and material affect the practicality of imaging. Thick and thin regions, internal passages and overlapping geometry can create different challenges within one casting. Supply the relevant drawing and identify the region that matters to the decision. A generic demonstration on a convenient specimen may not represent the most difficult area of the actual product.

Ask whether the proposal covers the whole component or a selected region, and how that distinction will appear in the report. A detailed study of one region can be useful while leaving other areas unexamined. The supplier should identify any preparation or positioning assumptions and obtain authorization for changes affecting the part. Do not assume that fitting a casting into a system establishes complete usable coverage.

For a family of related castings, identify geometry and material variants that may affect the approved setup. A change in a wall, insert or configuration should be reviewed against the examination assumptions. The buyer needs a controlled relationship between part revision and inspection setup, especially when a supplier intends to reuse an established recipe for later production lots.

Do not confuse pixel or voxel size with demonstrated detection

A quoted image sampling size can help describe a configuration, but it is not a universal statement that every feature of that size will be detected. Ask the supplier what evidence supports the actual examination objective, considering the complete imaging and review process. The technical reviewer should assess representative conditions rather than accepting a single nominal specification as proof of capability.

NIST's study of CT acquisition parameters investigated image quality and detection behavior using an additive-manufacturing artifact. Its findings should not be transferred as a casting performance number. The relevant general lesson is that configuration and evaluation conditions matter, so a buyer should request evidence tied to the intended product and task.

A useful trial identifies the features of interest, representative parts, review process and criteria for judging the demonstration before scanning begins. Record the configuration and any limitations. If the production service later changes a material setup element or analysis approach, the responsible technical authority should determine what review is needed. A beautiful trial image is not a permanent qualification for every future part.

Keep dimensional metrology separate from flaw evaluation

CT can support dimensional work as well as investigation of internal indications. Those are related but distinct uses. A model that helps locate an internal feature does not automatically establish the uncertainty needed for a dimensional acceptance decision. Likewise, dimensional performance evidence for a system does not, by itself, demonstrate detection of every discontinuity relevant to a casting specification.

NIST's discussion of ASME B89.4.23 concerns performance evaluation for CT dimensional measurements. Use that context to ask the metrology provider how it supports the specific measured characteristics. For flaw evaluation, separately identify the examination basis, relevant capability evidence and interpretation responsibilities. Do not collapse both services into a single unqualified accuracy claim.

If one scan supports both tasks, request separate deliverable definitions. The dimensional report should identify its measurement basis, while the NDT report should identify its examination scope and evaluation basis. Keep both connected to the same part and acquisition record. This allows engineering and quality teams to use the outputs appropriately without assuming one conclusion establishes the other.

Ask how indications are characterized and reported

A report should describe the observed evidence using terminology supported by the examination and assigned evaluation. A visually prominent region in an image does not necessarily establish its manufacturing cause or effect on service performance. Ask who interprets the result and what additional information is needed if the project seeks a metallurgical or engineering explanation beyond the imaging evidence.

For software that segments or measures indicated regions, clarify the analysis settings and review process. A displayed volume or count can depend on the selected interpretation approach. Preserve enough information for another authorized reviewer to understand how the reported result was produced. Avoid presenting a software output as an independent acceptance authority merely because the result appears as a precise number.

The AI-assisted NDT validation guide is relevant if automated classification is proposed. Ask for representative evaluation, handling of uncertain cases and control of changes. An automated system may help prioritize review, but the buyer still needs a clear basis for its use on the actual casting family and the applicable product requirement.

Choose the production stage deliberately

The casting's state at examination matters. An as-cast item, a heat-treated item, a machined component and a repaired casting can present different geometry, accessible regions and customer requirements. Identify the specified stage in the scope and preserve it in the report. An earlier examination should not silently stand in for later-stage evidence if the applicable requirement calls for something different.

Production changes also need a clear communication route. A revised pattern, material change, added insert or authorized repair may affect the examination assumptions. The supplier should know who notifies it and who decides whether the setup or evidence must change. A recurring order is not a reason to ignore a changed product configuration simply because the part number remains familiar.

The manufacturing inspection guide discusses population and traceability questions across product forms. For casting imaging, ask the production team to connect each examined item to the relevant manufacturing record. That connection helps distinguish a problem associated with a particular production change from an apparent difference caused by the imaging process.

Compare sampling and full-population claims carefully

The buyer should define which items are to be examined and why. A development study using selected samples is different from a production order covering every identified part. Do not invent a universal sampling percentage based on a vendor brochure or another company's practice. The customer and responsible quality organization determine the applicable population and any authorized sampling basis.

If samples are selected to investigate a production concern, preserve the selection rationale. A supplier should not generalize the findings to an entire lot without an appropriate basis. Equally, a statement that every part was processed does not establish that every required region produced usable data. Reconcile the population, successful examinations, exceptions and final dispositions separately.

For a production trial, record the complete workflow time, including handling, setup, analysis, review and exception management. A fast acquisition alone may not represent the service delivered to the customer. Compare the actual demonstrated workflow against the required output without inventing throughput guarantees. This makes the result more useful than a headline scan time detached from the rest of the inspection.

Hypothetical example: an internal passage needs clarification

Consider a hypothetical U.S. manufacturer investigating indications near an internal passage in a casting. Its normal production requirement calls for a defined radiographic examination, while engineering wants additional information about the spatial relationship of the observed features. The team separates the required acceptance deliverable from the supplementary investigation instead of assuming that a CT scan automatically replaces the original requirement.

The CT supplier reviews the actual geometry and proposes a representative trial for the region of interest. The parties agree the expected files, interpretation and limitations. The production-quality team retains the original radiographic record, while engineering reviews the CT evidence for its narrower investigation. Any proposed change to the production inspection follows the customer's approval process rather than a marketing claim about three-dimensional visibility.

The resulting package identifies the part, revision, manufacturing stage and relationship between the two examinations. It records what each contributes and what remains uncertain. This hypothetical example does not claim a particular defect or successful customer outcome. It shows why complementary imaging can be valuable without making the two methods contractually or technically interchangeable.

Specify original data, review access and an exit route

Before ordering, define which images, reconstruction files, analysis records and reports the customer receives. Ask whether another authorized reviewer can open the delivered files and understand their relationship to the part. A screenshot or animated video may communicate a finding but can omit information needed for a detailed review. Do not wait until a dispute to discover that the useful data were never included.

The DICONDE feature and native-data handover guide help structure interoperability questions. Where a format or conformance requirement is specified, demonstrate the intended exchange with the actual receiving arrangement. A general compatibility statement does not prove that every required field, annotation or analysis relationship survives the transfer.

Preserve the connection between acquisition, reconstruction, analysis and issued interpretation. If a result is revised, record which stage changed and why. The file-manifest guide provides a practical record structure. The customer should be able to distinguish a new examination from reanalysis of an existing dataset and identify which report is the current issue.

Agree how the delivered dataset will be viewed during a customer review. Confirm that the receiving workstation can access the relevant information and that annotations remain connected to the correct part and analysis. If a proprietary viewer is required, include its access and support terms in the proposal. A file technically delivered but practically unreadable does not provide the same review capability as a demonstrated, usable handover.

Make the buying decision around the required evidence

Compare proposals using the same casting population, product stage, region of interest and deliverable. Ask suppliers to identify assumptions, exclusions, preparation needs and the review included. A CT offer and a radiographic offer may address different questions; the comparison should expose that difference instead of reducing both to a price per part with no technical context.

The equipment RFQ guide is useful when the decision concerns buying or renting a system. For an outsourced examination, the inspection quote planner organizes user-entered service proposals. These are different commercial routes. Neither tool supplies current equipment inventory, market prices or an automatic determination that the proposed examination meets the customer's requirements.

Finish with a controlled scope and a receipt review. The report buyer-review guide and free resources help organize the evidence. Select the approach that answers the actual casting question with supportable capability and usable records. Two-dimensional and three-dimensional outputs are valuable when their roles are explicit and their limitations remain visible.

Compare the evidence you need

Scope preparation checklist

• Identify casting family, material, revision and production stage.
• Separate product acceptance from supplementary investigation.
• Define the region and required evidence before selecting a technique.
• Review representative capability rather than nominal sampling size alone.
• Agree original data, analysis context and report-revision delivery.

Useful planning resources

• Manufacturing scope: Connect product stages and production records.
• Automated evaluation: Review validation and uncertain outputs.
• Data interoperability: Ask specific transfer questions.
• Original files: Define the receiving workflow.
• Data manifest: Trace acquisition through report revisions.
• Equipment buying: Prepare a separate purchase or rental RFQ.
• Report receipt: Check delivered evidence completeness.
• Free worksheets: Organize the casting examination brief.

Sources and editorial scope

Primary sources researched October 31, 2026:

• ASTM E1441 public scope: computed tomography
• ASTM E1742/E1742M public scope: radiographic examination
• NIST: CT acquisition parameters and detection study
• NIST: CT geometry and dimensional performance evaluation

This is original NDT Connect editorial analysis. Worked scenarios are hypothetical. Source references establish only the stated public scope or research context; they do not approve a procedure, set acceptance criteria or certify a provider. Supplier descriptions are attributed and have not been independently performance-tested for your application.