Improving process
yield by utilizing smart SPC rules
Oded Tal
MAX International
Engineering Group
1
Tel: (201) 295-2953
Fax: (201) 295-3255
Email: oded_tal@maxieg.com
Keywords: SPC, Quality,
Yield Improvement, Operational Controls
Abstract
The purpose of Statistical
Process Control is to create dynamic measures that track our manufacturing, and
assures that our process remains stable. Typically, there is a high cost associated
with implementing a full-blown SPC mechanism both in system purchasing cost and
in management overhead. The article will give an overview of SPC, from original
Shewhart charts to more recent SPC schemes like Western electric rules. We will
demonstrate that most SPC tools relay on assumptions that are not valid in most
compound semiconductors manufacturing fabs. However we will show that by
creating good operational process control, and scrutinizing SPC measures it is
possible to achieve excellent process control by just staying with a basic set
of control charts, such as X Bar and R plotted for specific products,
operations, and tools. Furthermore we will show the relationship between
reaction times to discrepancy events and overall yield losses. Remembering that
our target is to reduce discrepancies and not to manufacture charts is the
clear advocate for an operational approach to controls.