Publication date: 05/24/2021

Consider a wafer manufacturing process with a target of 4 or fewer defects per wafer. You want to verify that a new process meets that target. How many wafers should you test to detect a difference of 1 that the count of defects per wafer is on target with 90% power at the significance level of 0.05?

1. Select DOE > Design Diagnostics > Sample Size and Power.

2. Click Counts per Unit.

1. Leave Alpha as 0.05 (the chance of failing the test if the new process is as good as the target).

2. Enter 4 as the Baseline Counts per Unit, indicating the target of 4 defects per wafer.

3. Enter 1 as the Difference to detect.

4. Enter a power of 0.9.

This is the chance of detecting a change larger than 1, which is equivalent to 5 or more defects per wafer. In this type of situation, the significance level (alpha) is sometimes called the producer’s risk and the power (beta) is called the consumer’s risk.

5. Click Continue.

Figure 17.14 Counts per Unit Calculator

You must test 38 wafers. The process meets the target if there are fewer than 173 defects.

Note: The 173 defects are the maximum number of defects that can occur in 38 wafers without rejecting the null hypothesis.

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).