Browse Prior Art Database

Auto Layer Align/Sort/Stack System Employing a True Best Fit Algorithm To Enhance Electrical Test Yields After Stacking And Lamination

IP.com Disclosure Number: IPCOM000119598D
Original Publication Date: 1991-Feb-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 5 page(s) / 99K

Publishing Venue

IBM

Related People

Celestino, RJ: AUTHOR [+2]

Abstract

This article describes a manufacturing system for automatically sorting, aligning and stacking metallized ceramic sheets used in fabricating multilayer ceramic (MLC) substrates. The disclosed system employs an auto sheet handler with appropriate three-axis positioner, an optical recognition station, and a novel "true best fit algorithm" which is used to calculate the optimum sheet positioning condition necessary to minimize positional errors which would prevent the sheet from meeting stacking specification limits.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Auto Layer Align/Sort/Stack System Employing a True Best Fit Algorithm
To Enhance Electrical Test Yields After Stacking And Lamination

      This article describes a manufacturing system for
automatically sorting, aligning and stacking metallized ceramic
sheets used in fabricating multilayer ceramic (MLC) substrates.  The
disclosed system employs an auto sheet handler with appropriate
three-axis positioner, an optical recognition station, and a novel
"true best fit algorithm" which is used to calculate the optimum
sheet positioning condition necessary to minimize positional errors
which would prevent the sheet from meeting stacking specification
limits.

      Layer stacking alignment requirements for metallized ceramic
sheets become more critical with the more stringent layer-to-layer
alignments, reduced via diameters, linewidths and spacings, etc.,
that are encountered in extensions of the present MLC technology.
The improved system described here employs an optical recognition
system in combination with an appropriate handler to enable
individual sheets to be removed from a tray and to be placed in the
optical recognition station.  This station is atop a closed loop,
three axis x, y, and theta positioner.  Optical recognition of sheet
features is compared to referable data utilizing a "true best fit
algorithm" which will be described later in this article.  The
optimum position to minimize positional error is calculated by means
of this algorithm.  The sheet is rejected if the resulting error
exceeds specification limits.  Acceptable sheets are repositioned to
the calculated optimum position, picked up by a vacuum head, and
transported to a blanking station.  Vacuum hold-down is used to
maintain sheet position during transport.  The metallized portion of
the layer is then blanked into a cavity, completing the process for
the first layer.  The total process, repeated for additional layers,
results in improved electric-test yields, since only stackable layers
are forwarded for lamination into MLC substrates and electrical test.

      The "true best fit algorithm" referred to in the above
description has been developed and implemented in FORTRAN, and, using
linear algebra, the sheet/substrate can be mathematically "moved"
through any translation and rotation, as illustrated in Figs. 1 - 5.
Briefly, error is defined as the magnitude of the vector distance
between a site and its nominal position. ...