We partnered with a global manufacturer to
restore high quality production line

Industry Overview

The semiconductor manufacturing process involves a set of complex, high speed and micro-meter ranged processes along the entire assembly line. It hence becomes increasingly challenging for the manufacturers to identify and improve the processes that lead to production defects.

The problem

Identifying the root-cause for quality defects is a big concern for manufacturers. Even after a through quality testing process, the manufacturer received multiple quality complaints from its customers over the last few months.

Our role

The manufacturer sought our help to restore the production line, identify the bottlenecks and eliminate defective outputs.

Methodology

We used a combination of multi-variate statistical testing and rare-event modelling technique to determine the operating ranges of all machine parameters. We worked with all the stakeholders - Product Heads, Manufacturing Director as well as the Production Engineers to identify the processes associated with all machines in the product line.

Using a hypothesis-based approach and statistical testing, we identified the parameters critical and having a high influence on the production quality.

With the help of this we helped business achieve the following outcomes:

• Dynamic ranges for SPC to ensure smooth production.
• Identify and control the critical parameters to avoid defects.
• Ability to correctly identify bad lots.

How did we enable consumption?

We worked with the engineering team to design a set of experiments with multiple parameter configurations that were tested on sample lines over the course of 1 week to validate the results. Based on the findings, the solution was then rolled out to all the lines.

Our impact

By using a combination of multi-variate statistical testing and rare-event modelling technique to determine the operating ranges of all machine parameters, we achieved 99.99% quality assurance across production line.