An appliance manufacturer was rejecting more than 7% of a large ABS front housing for sink marks. Fixing the rib design, bosses and gate, rather than the process, cut rejects to 1.1% and took four seconds off the cycle.
Key numbers
Every morning the quality inspector at a Melbourne kitchen appliance manufacturer laid the previous night's rejects on a bench under a daylight lamp. Most days there were dozens of large ABS front housings, each with the same faint shadows on the show face: one behind every internal rib, one behind every screw boss. The setters had tried higher hold pressure and longer hold times. The sink got slightly shallower, flash appeared at the parting line and the reject rate stayed stubbornly above 7%.
The situation
The housing was the most visible part of the product. Customers judged quality from across a kitchen, so the cosmetic standard was strict, and any visible sink was a reject. At 7.4%, the scrap was costing roughly $86,000 a year, and the inspection effort was tying up a person on every shift.
What the data showed
The first step was to separate rejects by cause rather than treat "cosmetic rejects" as one problem.
Sink marks made up more than two thirds of all rejects. Measuring the part told the rest of the story: the internal ribs were 85% of the nominal wall thickness, far above the 50–60% guideline for ABS, and several screw bosses were solid. A gate freeze study showed the gate was freezing before the thick areas could be packed, so extra hold time was doing almost nothing.
In other words, the process was not the problem. The geometry was creating thick spots that no setting could pack out.
What changed
- Ribs reduced to 55% of wall thickness with a small tool insert change.
- Screw bosses cored out to remove solid sections, with gussets for support.
- Gate enlarged by 0.4 mm to extend pack time into the thicker areas.
- Process reset using a gate freeze study, then documented as a standard window.
The rib changes required adding steel to the cavity inserts, the expensive direction, so they were planned carefully and done in a single tool modification.
Results
Rejects fell to 1.1% within three weeks. Cycle time dropped by four seconds, from 38 to 34, because the thinner sections cooled faster. Annual scrap cost fell from about $86,000 to $14,000, a saving of roughly $72,000 a year, and the extra inspector was redeployed.
Lessons for your operation
- Classify rejects by cause before trying to fix them. "Cosmetic" hides several different problems.
- When sink sits opposite ribs and bosses, measure them. The answer is often in the CAD model, not the machine.
- A gate freeze study tells you in an hour whether more hold time can help at all.
- Fixing geometry usually saves cycle time as well as scrap, because thick sections are slow to cool.
The same principles became the foundation of our design guidance on wall thickness, ribs and sink marks, and they remain the cheapest way to prevent this kind of problem: fix it in design, before steel is cut.
