During the board verification for a project, I often felt uncertain about whether the results represented the entire project accurately. Sometimes, we gain insight when the first batch is produced and compare the results to a previous board revision, especially when dealing with a large quantity of boards.
An essential data point is the capability, which checks if readings, like voltage, fall within the required range. When plotted on a graph with voltage on the x-axis and the quantity of boards on the y-axis, it should generally form a bell curve.
One of the indicators used is the Cpk, which shows the percentage of boards that fall between the upper and lower limits. Ideally, it should be 1; if it's lower, we have boards that don't meet the expected criteria (in the example below Cpk is 0.9772).
Another indicator is the DPMO, which indicates the number of defects per million boards. Ideally, this number should be zero. This indicator becomes more useful when we have a significant number of boards tested.
---
ref:
