AN1231 Motorola / Freescale Semiconductor, AN1231 Datasheet - Page 17

• Manufacturer Part Number: AN1231
• Description: Plastic Ball Grid Array (PBGA)
• Manufacturer: Motorola / Freescale Semiconductor

can be readily observed. It should be noted that fracture
length measurements taken from perpendicular PBGA
cross-sections can be erroneous, depending on the row
being sectioned, due to the fact that fractures generally prop-
agate in a direction radially from the device center.
tude of fractured joints is dye penetrant analysis. Dye pene-
trant analysis can be used to visualize an overall distribution
of fractures on all joints. For this procedure to work the best,
the mounted device should be cleaned with a solvent prior to
the application of a dye penetrant. This cleaning removes
any flux residues, soldermask and other particles which are
mobile during thermal cycling and may inhibit the flow of the
dye into the fracture. There are many dyes available for the
specific purpose of penetrating fractures (such as Ardrox
Tracer-Tech), however, a machinist’s layout dye made by
ITW and called Dykem (Steel Red is typically the most vis-
ible) has shown excellent results at Motorola and elsewhere.
After cleaning, a dropper is used to repeatedly flush the dye
underneath the mounted device. The excess dye is then al-
lowed to drain and the remainder is dried. This drying is ac-
celerated by baking at 100 C for 10 to 30 minutes depending
on the amount of dye under and around the device. Placing
the assembly in a vacuum chamber prior to baking may in-
crease dye penetration into small fracture openings. Follow-
ing removal from the bake oven and cooling, the PBGA is
mechanically removed. It can be pried off with a screwdriver
or similar, which may damage joints on the outer one or two
rows, or the board repeatedly flexed until the device “pops”
MOTOROLA FAST SRAM
A very useful technique in analyzing devices with a multi-
95
90
80
70
60
50
40
30
20
10
5
2
1
1000
W/rr
361 Pin PBGA Cycled from 0 to 100 C at Two Cycles per Hour
Figure 14. Weibull Failure Distribution by Device Row for a
NUMBER OF THERMAL CYCLES
10000
off. After removal, the board and device can be readily in-
spected. Fracture surfaces which are dyed were obviously
present prior to device removal and presumably caused by
the accelerated testing. An example with a fracture approxi-
mately one quarter of the way through the joint is presented
in Figure 15.
Penetrant Following 1885 –40 to 125 C Thermal Cycles
Figure 15. Micrograph of a 361 Pin PBGA Solder Joint
SYMBOL
Partial Fracture Surface After Application of Dye
DIE PER–
IMETER
(Magnification approximately 120X)
MIDDLE
OUTER
INNER
ROW
11468
7958
11640
13958
10.3
13.0
10.3
8.1
100000
AN1231
17