Known Good Die Selection Tradeoffs: A Cost Model

Abstract

A Monte Carlo simulation model was developed for examining the cost of Known Good Die (KGD) relative to packaged parts, with a focus on identifying the dominant cost drivers. The cost of KGD and packaged devices that go through test and burn-in are studied by varying these key drivers and examining their sensitivities. The model is capable of manipulating IC fab parameters (e.g., defect density, sort coverage, etc.), as well as packaging, test, and burn-in costs. In determining relative costs, an important consideration is the package type to which KGD is being compared (plastic vs. ceramic). The key cost drivers are number of KGD carrier uses, complexity of the IC, cost of the carrier, number of burn-in hours, and KGD assembly/disassembly time.