Casting defects · analysis

Cavity-wise rejection analysis

Castings from one mould share the same metal, the same sand and the same moment of pouring. When one impression rejects far more than the others, the difference is usually the impression itself.

By Versatile Equipments Pvt Ltd · equipment manufacturer & calibration provider · updated 2026-09-07

Why the cavity is a fair comparison

Most analysis of rejections compares things that are not really comparable: this month against last, this shift against that, one pattern against another. Too much changes between them for the comparison to be clean.

A multi-cavity pattern is different. Every impression is filled from the same ladle, in the same sand, within seconds of the others, on the same machine and by the same operator. Almost everything is held constant. So when one impression rejects consistently more than its neighbours, the metal and the sand are unlikely explanations, and the impression itself becomes the reasonable suspect.

Recording it without slowing the floor

The analysis only becomes possible if the cavity number is captured with the defect. On most patterns the impression number is cast into the part already, so the inspector is reading a figure that is in front of them rather than working anything out.

• One casting carries one cavity number. An inspector picks up a single casting, so there is one impression to record, not several.
• Keep it optional. Where the number is not legible, allow it to be skipped or marked as unreadable rather than guessed. A guessed cavity number is worse than a blank one.
• Consistency beats completeness. Even partial capture will show a strongly misbehaving impression, provided the entries that are made are honest.

Reading the pattern

Compare each cavity against the average for its own pattern, not against other patterns. On an eight-impression pattern you would expect the defects to spread roughly evenly; what you are looking for is one cavity sitting well above that spread over a reasonable number of moulds.

An illustrative month. Seven impressions behave alike and the eighth does not, which no total by pattern or by shift would ever have shown. Give it enough castings before drawing a conclusion - on a low-volume pattern a single week may be too thin to separate a real difference from ordinary variation.

What one bad cavity usually means

• Wear or damage to the impression. A worn, dented or badly repaired cavity gives poorer surfaces and dimensional drift.
• Venting. An impression that vents less freely than its neighbours tends to collect gas defects such as blowholes.
• Feeding and gating imbalance. A cavity fed last or through a longer path can show shrinkage, misruns or cold shuts while the rest fill cleanly.
• Local mould hardness. Uneven ramming or squeeze pressure over one impression can bring penetration or swelling in that position.
• Core seating. A slack or worn core print in one position gives shift and wall thickness variation there.

Acting on it

The value of the finding is that it points at one physical object. The impression can be inspected on the pattern plate, its vents opened, its print checked, its gating reviewed against the others, or the plate scheduled for repair. That is a far smaller and cheaper action than adjusting the sand system or the pouring practice for the whole line, which is what a pattern-level total might otherwise have prompted.

Afterwards, the same cavity comparison is how you find out whether the repair held. Recording the cavity as part of the defect register makes that check routine rather than a special exercise.

Frequently asked questions