A surface defect does not automatically identify faulty sand
To investigate metal penetration and burn-on, first describe the defect before shot blasting, then examine material, mould preparation, coating and pouring conditions. A casting with adhering sand, a crust or a rough surface does not immediately establish that its sand supplier should be changed: surface evidence needs to remain visible to guide diagnosis.
The Ministry of Industry’s regional profile identifies automotive activities in Rabat-Salé-Kénitra 1. This supports a qualification dossier for industrial-component projects, without asserting that any particular plant experiences the defects described here. The objective is an investigation applicable to actual manufacturing, with its alloys and moulding system.
The American Foundry Society’s training programme distinguishes defects involving sand expansion, adhesion and strength 2. A useful method must retain these separate families. Several mechanisms can coexist on one component, so workshop vocabulary must be connected to verifiable observations.
1. Describe the anomaly before choosing a correction
Write a description that remains valid without knowing the cause: location, extent, relief, adhesion, repetition and stage of appearance. “Rough area on the lower face, present before cleaning” is a better starting point than “sand too coarse”. The second expression already converts a hypothesis into a conclusion.
Photograph the entire component and then the detail with a dimensional reference. Identify its position in the mould and proximity to a gate, core or thick section. Retain useful portions when the laboratory procedure requires a section. The file should permit comparison of defective and acceptable components with the same geometry.
Also define the industrial criterion: surface finish, dimensions after machining, cleaning access or cavity function. A cleanable superficial anomaly and inaccessible adhesion within an internal passage are different problems. Classify them before comparing the economics of corrective measures.
2. Mechanical penetration: examine metal passage
Penetration describes liquid metal entering mould or core pores. Foseco identifies metal pressure, stream impact and open structure among factors to investigate 3. The question therefore concerns a prepared interface rather than mean grain diameter alone.
First compare affected locations. Concentration beneath a substantial metal height directs investigation differently from concentration opposite a gate. This map does not demonstrate the cause by itself; it helps choose trials separating mould structure from filling conditions.
Distinguish the grading of loose sand from porosity in the prepared mould. Teams using the same batch can produce different interfaces. If defects follow a compaction area or preparation method, changing the entire grading without examining that stage may simply move the problem.
3. Burn-on: investigate chemical interaction
Workshop descriptions of “burnt sand” can refer to several observations. Nakae and Matsuda’s cast-iron study examines burn-on as chemical penetration, including wettability and interaction between silica and an FeO-containing phase 4. This explains why a crust should not be attributed only to excessive pore opening.
The research abstract does not provide a recipe applicable to every alloy. For an actual component, ask the laboratory whether examination can identify a metal network between grains, a reaction phase or both. Supplement surface observation when the decision requires these mechanisms to be distinguished.
Retain alloy composition and available data on adhering material. Commercial-sand chemistry helps characterize feedstock, but does not replace analysis of the post-pouring interface. Overall SiO₂ content cannot describe the entire reaction layer or the conditions producing it.
4. Distinguish veins from generalized roughness
Veins involve metal entering cracks in the core or mould surface. Foseco connects this family with thermal stresses associated with silica expansion 5. A thin metal rib can therefore require a different investigation from a surface covered with adhering grains.
Examine the defect pattern: line, network, plate or diffuse area. Relate that pattern to core geometry and surroundings. Photographs after grinding can remove the most useful evidence. Preserve at least one observation before that operation when the programme permits.
Do not assume that all defects near a crack have one origin. Local rupture can also modify metal flow or release material. Document what is observed, then test the sand-binder combination and thermal loading corresponding to actual manufacture.
5. Separate sand inclusion from penetration
An inclusion involves material introduced into the component; penetration involves metal entering the mould structure. Foseco notes that inclusions can arise from dislodged sand and remain hidden until machining 6. The word “sand” in both descriptions therefore does not constitute a common diagnosis.
Compare surface and section where sectioning is justified. A particle or particle group inside the metal is interpreted differently from an external crust. Location may also direct examination of mould cleaning, fragile edges and gating.
Keep components inspected before and after machining separate. A more extensive inspection campaign can increase detected anomalies without manufacturing having deteriorated. Assess improvement with comparable inspection methods and the same control stage.
6. Read the complete grading
Grading decisions must address the distribution, rather than only a fineness index or commercial range. Retain sieve fractions, method and sample condition. An identical average can describe different mixtures; an unrepresentative portion can conceal segregation.
Examine received material separately from sand actually used in the mould or core. Transport, mixing or preparation can change the available fraction. Identify the sampling stage so that investigation does not compare the wrong materials.
If a finer variant is studied, also monitor permeability and preparation. Assess correction across the whole system, using the foundry’s selected tests. Improved surface appearance is insufficient if another property required for the component becomes inadequate.
7. Examine compaction at the relevant location
An overall mixture measurement does not demonstrate uniformity across every mould surface. Identify deep areas, junctions, edges and parts prepared differently. An acceptable average can coexist with an inadequately prepared detail.
Compare results at defined positions and comparable waiting times. Where a local measurement method is used, preserve its protocol, geometry and application range. Avoid turning an isolated reading into a value characterizing the whole mould.
For cores, examine preparation and handling before closure. An edge damaged after manufacture can leave an area different from the specimen originally inspected. Observations should follow the actual chain: preparation, extraction, storage, coating and assembly where those stages exist.
8. Qualify coating as an operation
The presence of coating does not demonstrate that a consistent barrier is available during pouring. Describe the product appropriate to the system, its preparation and application. Foseco emphasizes selection and application thickness in defect prevention 6.
During investigation, inspect accessible areas and difficult details. Distinguish missing coverage, accumulation and deterioration after application. A layer observed before assembly may be damaged during fitting; a photograph at one stage does not document the entire operation.
Follow drying according to product and process procedures. Do not simultaneously change formulation, dilution, method and waiting time when the trial seeks to understand a defect. Identify one modification and preserve the other parameters needed for interpretation.
9. Connect defects with pouring history
Retain conditions actually available: alloy, metal preparation, measured temperature, delay and filling geometry. A theoretical setpoint without measurement cannot explain the component. Also identify the measurement location and timing so that series can be compared.
A defect concentrated in a thick section warrants examination of local thermal exposure. A defect facing a gate warrants examination of filling. These remain hypotheses to confront with observations and an appropriate trial, rather than immediate proof of overheating or poor design.
The foundry must validate pouring modifications against filling and quality requirements. Reducing temperature or modifying a gate for one defect can change other outcomes. Follow the targeted defect together with the component’s necessary characteristics.
10. Build a matrix of possible causes
Organize hypotheses by mechanism: entry into an open structure, chemical interaction, cracking, dislodgement or coating application. For each, associate expected evidence and an observation that could contradict it. This prevents one explanation from becoming impossible to challenge.
For example, if the defect is attributed to changed sand, ask whether acceptable components were made from the same batch under the same other conditions. If so, examine preparation or geometry differences. Otherwise, the material hypothesis remains open, but controlled comparison becomes necessary.
Keep the matrix short and decision-oriented. Twenty causes without an examination order do not support action. Begin with discriminating observations and achievable checks, then use specialized analysis when initial results fail to distinguish mechanisms.
11. Example: one batch, two locations
Imagine a component with adhesion only on its lower face while another face made with the same batch is acceptable. This is a teaching example, rather than a Minerallum observation. It proves neither a pressure defect nor a coating defect; it shows that globally blaming the batch would be premature.
The first trial compares mould surfaces corresponding to both areas: preparation, coating and geometry. The second confronts this map with filling history. Sectioning can then be requested if metal intrusion must be distinguished from a reaction layer. Each stage answers a specific question.
If a sand variant is eventually tested, preserve the component, binder system and relevant conditions. Improvement after several simultaneous changes may validate the new combination, but does not demonstrate that grading was the sole cause.
12. Choose a diagnostic section that preserves the original interface
A diagnostic section should retain metal, adhering material and a neighbouring acceptable area together. If the sample is taken only from cleaned metal, the interface being investigated has disappeared. Agree the cutting path, preservation method and faces to examine with the laboratory before preparing the fragment. Sampling must answer a hypothesis, rather than merely provide a small piece that is easy to send.
Cross-sectional observation can examine metal continuity between grains and the position of an intermediate layer. Elemental analysis can then help compare that layer with sand and alloy, according to the available method. An isolated silicon signal cannot name a mineral or demonstrate a particular reaction. The conclusion must connect observed form, composition and limitations of the analytical equipment.
Where possible, prepare an acceptable reference from the same geometrical location. Differences between component locations can arise from thermal history or preparation. Differences between two components become interpretable only when the origin of each fragment is preserved. Mark the mould face and direction towards metal on the accompanying sampling drawing.
Prior cleaning also requires an explicit decision. A washed, blasted or ground surface no longer represents precisely its shakeout condition. Where several preparation methods are needed, retain separate portions for each. The laboratory can then distinguish an initial observation from an effect introduced during sample preparation.
Finally, one section may explain a local mechanism without establishing its frequency throughout the series. Compare its result with the defect map and component count. Investigation becomes more precise when detailed analysis answers the mechanism question and series inspection separately answers the question of repetition.
13. Measure correction without concealing the defect
Before testing, define component count, inspected areas and classification method. Retain acceptable results alongside anomalies. Presenting only the best component cannot establish whether the correction is consistent.
Separate the defect at mould removal from the result after cleaning. More aggressive blasting can improve appearance while increasing finishing time or modifying dimensions. Distinguish prevention from additional capacity to remove the defect.
For internal areas, examine actual function and accessibility. An externally satisfactory component may retain material in a passage. The criterion must reflect customer requirements and available inspection, rather than a commercial photograph of the surface.
14. Convert diagnosis into a material requirement
Explain what was established: defect family, tested combination and validation scope. If sand contributes to correction, specify the grading and characteristics actually useful, together with their control methods. Avoid requesting “better quality” without a measurable criterion.
Purchasing teams can then distinguish material modification from process modification. A precise enquiry gives the supplier a verifiable objective and lets the foundry control receipt. It does not transfer all responsibility for an interface built through several operations to the material supplier.
Our silica-sand page complements the material dossier. For penetration or burn-on investigation, sand is assessed within the prepared mould and resulting component. Useful outcomes are a better identified cause and a combination whose correction remains observable in production.