[ 40 ]
PBGA solder ball coplanarity impact evaluation
Keywords
Plastic ball grid array, Reliability,
Solder
Abstract
A study was conducted to exam-
ine the sensitivity of solder joint
integrity for PBGA assembly and
post rework process yields to the
change of the solder ball copla-
narity specification from 0.15mm
to 0.20mm as well as to increase
the level of confidence with
respect to the specification
change. The study considered the
following experimental variables:
solder ball coplanarity; stencil
thickness and aperture design;
whether the assembly also experi-
ences a wave soldering process;
the rework method, i.e. whether
there is dispensing of solder paste
or just a flux. Electrical continuity
tests were performed after PBGA
assembly and rework. Accelerated
thermal cycling and thermal
shock tests were then used to
evaluate the solder joint reliability.
Additionally, standoff heights of
PBGA solder joints were measured
through cross sectioning samples.
Statistical comparisons of the
mean standoff height between
different PBGA coplanarity
groups, where different experi-
ment variables were applied,
allowed for determination of the
variation of BGA solder joint
integrity.
Introduction
The eutectic solder balls of plastic ball grid array (PBGA)
packages tend to collapse to a height of about 0.15-0.20mm
during the reflow soldering process (Lau, 1995a). The
actual final height depends on the solder paste volume,
solder ball coplanarity, package weight, PCB pad size,
reflow temperature profile, etc. Owing to the controlled
collapse characteristics, BGAs which use the eutectic solder
balls require less stringent coplanarity specifications com-
pared with BGAs having high melting temperature solder
balls such as CBGAs.
Indeed, the solder ball coplanarity of BGA packages is a
critical factor impacting not only on the yield of the assem-
bly process but also on the long term solder joint integrity
(Prasad, 1997), especially for the larger packages. JEDEC
has recently changed the coplanarity tolerance specifica-
tions for BGAs with eutectic solder balls on 1.27 and
1.5mm ball pitches from 0.15 to 0.20mm. This change in
the specification by JEDEC has given rise to some concerns
in the industry.
The definition of coplanarity for the arrays of solder balls
used in this study is the distance between the apex of the
highest solder ball and the apex of the lowest solder ball.
The coplanarity requirement specified by JEDEC is depen-
dent on the solder ball melting temperature and pitch size.
For standard PBGA packages, where the solder ball reaches
the liquid state below 220°C, the deviation from coplanarity
allowed by JEDEC is 0.20mm for 1.27mm and 1.50mm
pitch balls, and 0.15mm for 1.00mm pitch balls.
For most package suppliers, it is very difficult to meet the
0.15mm requirement for the larger BGAs (1.27 and 1.5mm
pitch sizes) because these larger packages tend to have
greater warpage (Hwang, 1995). On average, the yield
achieved by the package suppliers can be increased from
60-90 per cent if the 0.20mm instead of 0.15mm coplanarity
requirement is specified for PBGA package production.
With the change by JEDEC of the coplanarity specification
for eutectic BGA solder balls, the pressure for the package
suppliers to meet the 0.15mm coplanarity requirement has
been reduced; however, the impacts associated with this
change on assembly, rework, and solder joint reliability
need to be studied.
This paper describes an experiment designed to analyze
the impact of the PBGA solder ball coplanarity on solder
joint reliability both from a PCB assembly process and from
a post rework point of view. In order to reduce the number
of experimental trials and to minimize the variable interac-
tions, the experiment covered only those variables consid-
ered to have the most significant effect on the PBGA joints
in relation to the coplanarity specification changes. Some
factors such as PCB pad size for the BGA, PCB pad finish,
solder mask defined vs. non-solder mask defined method,
BGA registration, and PCB thickness have been covered in
several studies (Mawer
et al., 1996, Brown and Bromley,
1996) and were therefore not covered in this study. The
experiment included the following activities:
• PBGA coplanarity scan.
• PBGA SMT assembly.
• PBGA wave soldering process.
• PBGA rework.
• X-ray inspection.
• Electrical continuity test.
• Accelerated thermal cycling (ATC) and thermal shock
reliability test.
• Cross section analysis and standoff height measurement.
• Analysis of the mean standoff height.
Three sizes of perimeter-array PBGA packages, with 256,
352 and 388 I/Os, and with controlled solder ball copla-
narity were considered in this study. The 256 I/O was
selected since it was the most popular package type used in
current Adaptec products. The 352 and 388 I/O were
selected because they were the largest BGA packages used
in the company, and they were thought to be more sensitive
to the ball coplanarity during rework. The basic dimensions
of each of these three PBGA device types are listed in Table
I. All of the BGA devices were provided with a daisy chain
layout. A dual daisy chain was used so that a quick continu-
ity test could be conducted during the study. An example of
the daisy chain layout for the PBGA 256 I/O is shown in
Figure 1. All devices were baked at 125°C for 24 hours and
sealed in dry packages prior to usage.
All of the PBGAs were fully tested and then scanned for
coplanarity. They were then separated into three groups
based on their measured coplanarity: those with coplanarity
within 0.15mm, those between 0.15 and 0.18mm, and those
between 0.18 and 0.20mm. This experiment was also
designed to evaluate if the stencil thickness and wave solder
process could impact on the solder ball coplanarity. In
addition, this experiment also verified the effect of different
rework processes including the use of dispensed solder
paste or the use of a paste flux.
Soldering & Surface Mount
Technology
11/3
[1999]
40–45
© MCB University Press
[ISSN 0954-0911]
Tony Huang
Adaptec, Inc., Milpitas, California, USA
Joe Chu
Adaptec, Inc., Milpitas, California, USA
Table I
PBGA devices configuration
Number Pitch Ball array Package size
of I/Os (mm) type (mm
2
)
256 1.27 Perimeter 27
× 27
352 1.27 Perimeter 35 × 35
388 1.27 Perimeter 35 × 35
Figure 1
Daisy chain configuration for PBGA 256