Semiconductor Yield Economics: Dies per Wafer, Yield Models & Cost per Chip
Estimate gross dies per wafer, model yield with Poisson, Murphy and negative binomial, cost a good die and a shipped chip after test and packaging, test a die shrink, price speed bins and plan fab utilisation: 8 graded labs on a fictional chip company (illustrative numbers).
The first lesson is free — no account needed.
About this course
A semiconductor yield economics course that costs the chips of one fictional chip company in a spreadsheet, for engineers, product and finance people in the chip industry, analysts and students. You estimate gross dies per 300 mm wafer with the standard approximation and its edge-loss term, turn falling defect density into Poisson yield and the wafer starts for 50,000 good chips, compare the Murphy and negative binomial yield models with a yield curve over die size, compute cost per good die and what five extra yield points are worth, carry a good die through wafer probe, packaging and final test to the cost per shipped chip, test whether a die shrink on a newer, dearer node is cheaper, split good dies into speed bins with NORM.DIST and plan wafer starts and utilisation from the bottleneck tools. Illustrative numbers. The first lab is free.
Lessons
- Wafer Area and Gross Dies per Wafer (with the Edge-Loss Term)Free
- Defect Density and Poisson Yield e^(−A·D0) Pro
- Murphy and Negative-Binomial Yield vs Poisson (with a Yield Curve) Pro
- Cost per Good Die: Wafer Cost ÷ Good Dies Pro
- Test and Packaging Yield: Cost per Shipped Chip Pro
- Die-Size Shrink Scenario: Is the New Node Cheaper per Chip? Pro
- Binning: Speed-Grade Shares and Revenue per Wafer Month Pro
- Fab Utilisation and Wafer Starts per Month Pro