Achieving Low-Risk, Volume Manufacturing on Leading Edge Technology Probecards - Definition and Methodology for Implementing a Process Control System - IEEE SW Test Workshop

Sep 11, 2026

Achieving low-risk volume manufacturing on leading-edge technology probecards by defining a practical process control system (PCS) and outlining a method to implement it. The material explains why wafer-test challenges and shrinking process windows demand robust, no-luck monitoring, using control charts as the core caution mechanism. It details how to create control charts (performance parameters, baseline stability, control limits, and documented response procedures) and emphasizes complementary PCS components such as document control, process change control, training management, and recurring reviews. Examples show learning over time, improved chart signals, and corrective actions during breakdowns in control.

LanguageEnglish

CategoryPresentations

FormatPDF

Pages13

Reading Time5 min

AccessFree to download

Motivation for PCS

The Stoplight

How to Create a Control Chart

Define performance parameters

Study variation and set control limits

Run baseline process and establish response procedures

What is a Process Control System?

Control charts and supporting PCS components

Review and process change governance

Examples of PCS in action

June 12 to 15, 2011 San Diego, CA Achieving Low‐Risk, Volume Manufacturing on Leading Edge Technology Probecards; Definition and Methodology for Implementing a Process Control System Douglas A. Sottoway Anil Kaza Intel STTD Wafer Test The Title • Made the title so it sounds very important. Why? Because it is. But what is it? • Here is the title to put in your mind... – What is a process control system, why do I need it? 2 Motivation for PCS • Wafer test technology challenges continuing – Shrinking pitch, increasing array size/probe count, current carrying requirements pitch, increasing array carrying requirements • Shrinking process window at test – Performing test efficiently and producing correct results is very difficult l h h i ll ! – No onger have huge margins to allow test process to vary! • A robust process control system is now a matter of SURVIVAL! – In my opinion, if you’re not doing this already, you need to start now – Within spec monitoring not enough – 7 years go we needed PCS in sort • Every couple months the yield department would find something wrong with tool at sort • Control charts were put in place to detect issues, this was better but still too many issues • Simple PCS charts not enough, need fab solution, entire system for good process control June 12 to 15, 2011 IEEE SW Test Workshop 3 The Stoplight • In the days of horses and buggies, the stoplight had only red/green – In 1920 William Potts realized the advent of automobiles required a better method, i.e ., a yellow caution signal • Specifications (red/green only) not enough to ensure a robust output and expected results • Now need yellow, proceed with caution (control charts) – Something bad could happen unless good decision‐making utilized • Control charts are the core of PCS bad could happen unless good decision 4 How to Create a Control Chart 1. Define Performance Parameters USL – Contact resistance LSL – Find Critical Process Parameters (CPPs) • probe length mean 2. Study variation and set control limits – Control limits should be practical 5 How to Create a Control Chart 3. Run baseline process – Engineer identify special cause and fix – Find out if process stable and capable 4. Response procedures documented and trained USL UCL Target LCL LSL – Empower workforce with correct tools Of Note: • Expect many signals during initial volume – Engineering spends time responding to and learning about the signals – Root cause fixes are put in place and response procedures are documented and trained • The process gets healthier, control charts are improved • The amount of learning and process improvement made simply from responding to good control charts is amazing, and I’ve been doing this for 21 years 6 What is a Process Control System? • Control charts provide the caution light but... – There are other critical components that the organization must have to make control charts fully effective – Any weakness in the other PCS components will eventually lead to process health issues If any piece is missing, predictable output becomes a matter of luck when i d ti ht process windows are tight Control Charts Document Control Process Change Control Training Management Review Example: During Development Worn Part Signal Found USL UCL ~~ ~~ 2 sigma Signal Found • Control chart: f b l h – Mean of probe length – Limits set ~ 3 sigma k li i i Target 2 sigma LCL LSL s 1.3 • Learning: • Cp to spec limits – Concern with worn part – Implement quarterly PM – Process step is stable and capable 8 Example: 1 Year Later • Production team sees f l ( ) out o controls (OOCs ) – Huge increase in volume Target LCL LSL • 2 tools to 5 tools – Conclude 3 added tools run a little different – Still within spec – Increase lower control limit 9 Example: Breakdo

FAQs

Why is a process control system (PCS) necessary for leading-edge wafer test volume manufacturing?

Because wafer-test technology faces shrinking pitch, higher probe counts, changing current requirements, and tighter process windows where producing correct, efficient results becomes very difficult. The material frames robust PCS as a survival need beyond simple within-spec monitoring.

How are control charts created in the PCS methodology?

First define performance parameters and critical process parameters (CPPs), then study variation to set practical control limits. Next run a baseline process to confirm stability and capability, and document response procedures and train the workforce to act on signals.

What other components are required besides control charts for PCS effectiveness?

The presentation stresses that control charts only provide the caution light. To prevent predictable process health issues, the organization must also manage document control, process change control, training management, and regular reviews of control limit changes.

Sep 11, 2026

Uploaded by Nguyễn Văn Học