A QUALIFICATION VEHICLE IS
DESIGNED THAT INCLUDES STUDIES OF SUCH PROCESS PARAMETERS
AS SOLDER-PASTE PRINTING, COMPONENT PICK-AND-PLACE,
AND SOLDER REFLOW.
B y means of a number of experiments
on pad-design identification, pick-and-place machine
evaluation and assembly process investigation, understandings
for 0201 assembly can be summarized as follows.1,2
Two types of pad size (non-soldermaskdefined)
are recommended in previous work.
The pick-and-place machines evaluated are able
to place 0201-size components side-by-side, as close
as 0.004" and at acceptable defect levels.
Placement height vs. solder paste location is
critical to achieve low defect rates. Thus, any
placement machine having the highest ability to
control placement height will be recommended.
For reflow, low soak time and low peak temperatures
will help reduce the number of components skewing.
Based on the experience, a Qualification
Vehicle (QV) is designed and more experiments performed
to further understand the 0201 assembly process. These
experiments include assembly with solder paste and
components from various vendors.
Qualification
Vehicle Design
As
shown in Figure 1, the 0201 QV is a double-sided
panel with mirror images.
The outside dimension is 5 x 7" with 0.030"
thickness. Four boards with a cell phone form
factor are designed into this panel.
For 0201s, two types of pad design (Pad U
and Pad H) also are used; boards A
and B are designed with Pad U,
boards C and D with Pad H.
In addition, Boards A and C will test 0201-to-0201
spacings (0.006, 0.008 and 0.010"), while
Boards B and D will test the spacing between
0201 and other components.
The same spacings (0.006, 0.008 and 0.010")
will be tested for between 0201 and 0402,
while between 0201 and chip scale and S08
packages, an 0.008" spacing is designed.
Figure
1.
Finally, with a 0.005"-thick
stencil, the Area Ratio (AR) is 0.60 for Pad U and 0.74
for Pad H. (AR is the ratio of the aperture opening
area to that of the aperture wall.)
On the top side of the QV there are total 5,728 components,
including 5,092 locations for 0201s, 624 locations for
0402s, eight locations for S08s and four locations for
CSPs.
The 0201 resistors from two vendors and 0201 capacitors
from two suppliers are distributed equally in the pick-and-place
program.
Experimental
Work
No-clean
solder pastes from three vendors are selected for
this experiment.
All are particle Size 3 with a mesh size of 25 to
45 µm and with metal contents from 90 to 90.25
percent (by weight).
An electroform stencil of 0.005" thickness and 1:1
ratio (between the aperture size and pad size) is
used.
A standard assembly line consists of a solder paste
printer, a turret pick-and-place machine, and a nine-zone
reflow oven.
During the experiment, solder
paste height, area and volume for each board are measured
by using an automatic measurement machine, and the
process control (Cpk) for solder paste volume also
is calculated.
Manual inspection is performed after printing, pick-and-place,
and reflow. At least 35 panels are assembled for each
solder paste, yielding a total of more than 555,000
0201s and 68,000 0402s in the experiment.
Results
and Discussion
Post-printing
Data Analysis. As can be seen from Table 1, for all types of
solder paste, on both Pad U and Pad H, the average height
is about 0.0055" with 0.005"-thick stencil, and the
average solder paste volume transfer ratio is 65 to
70 percent for Pad U and 84 to 89 percent for Pad H.
The higher the AR, the higher the Cp of solder paste
volume.
During visual inspection, no
bridging or missing solder paste is observed in any
of the printed boards in this experiment, indicating
that the printing process is under control during
the entire experiment.
No misalignment above 20 percent of pad width is observed
in either X or Y direction.
Table 1. Post-printing
Data Analysis.
Post
Pick-and-Place Inspection. Post
Pick-and-Place Inspection reveals that missed
components is the primary defect.
A typical appearance of a location with a missing
component is presented in Figure 2, showing that
the component had been placed onto the solder
paste but did not remain in place.
Figure2.
Figure
3. Defects
charted on Pad U using solder paste from vendor
A exhibits five types of defects with missing
component and tombstoning being the primary faults.
Figure
3.
Post-reflow
Analysis on Pad U. Post-reflow Analysis on Pad U. Figure 3 shows
that when using solder paste from vendor A, five types
of defects are observed, such as:
component missing
billboarding
bridging
tombstoning
and off-pad
with component missing and tombstoning
being the main defects.
Compared with the results from
solder paste vendors B and C, more defects of missing
components and tombstoning are observed with solder
paste A.
(In a previous solder paste study, the experiment
data had revealed that a higher tack force was observed
with solder paste B than with solder paste A. The
tack test procedure was based on IPCTM-650 and the
tack force was measured up to eight hours.)
Accordingly, the reason for missing components may
be due to the relatively low tackiness for solder
paste A than that for paste B.
Figure 3 also indicates that there were many more
missing component defects from locations with 0201
capacitors than with 0201 resistors, probably due
to the height and weight difference between these
two parts.
As
shown in Figure 4, the solder paste height before
placing a component is about 0.0055".
The average height is about 0.012" for a 0201 capacitor
and about 0.009 to 0.010" for a 0201 resistor. The
average weight is about 0.00028 g for a 0201 capacitor
and about 0.00014 g for a 0201 resistor.
During
the pick-and-place process, when a 0201 resistor is
placed into the solder paste about two-thirds of its
body will sink into the solder paste while only about
half of the 0201 capacitor body will settle into the
material.
With air blow or board movement, a 0201 capacitor
would be more likely to fall out of its correct location
in the solder paste, causing the defect of missing
component.
With respect to the bridging defect, more than 98
percent come from locations with 0.006" spacing with
solder paste from vendor B, while 100 percent of bridging
defects come from locations with 0.006" spacing with
solder paste from A and C, most likely due to more
solder paste being deposited when printing with solder
paste B.
More defects concerning 0201 capacitors (missing)
are observed with components from vendor C than from
vendor A.
Post-reflow
Analysis on Pad H. Post-reflow Analysis
on Pad H ( Figure 5) shows a similar trend to
Pad U.
More defects of missing
components are observed with solder paste from
vendor A, and most from 0201 capacitors rather
than 0201 resistors (especially from component
vendor C).
More defects of bridging
were found with solder paste from vendor B than
the other vendors.
Figure
5.
Post-reflow Analysis on Pad U vs. Pad H. Post-reflow Analysis
on Pad U vs. Pad H. For all type of defects, fewer
are observed on Pad H as compared with Pad U.
Reason: the pad size (and therefore the aperture size)
for Pad H is larger than that of Pad U.
Solder
Paste Comparison. As
shown in Table 2, regardless of component types,
the results from assembly with solder paste B
exhibit the lowest defect level, with solder paste
A having the highest defect level, probably due
to the different tackiness. Based on previous
experience, after a printer idles for more than
five minutes, solder paste will exhibit a release
issue. Solution: it will need to print a couple
of boards before achieving a consistent print.
Solder paste C has less of that problem than displayed
by solder paste B.
Table
2.
Component Type. As shown in Table
3, for different component types, 0201 resistors display
fewer defects than 0201 capacitors.
The latter from component vendor C have twice as many
defects as those from component vendor A.
To understand the relationship
between the defect level and 0201 capacitors from
different vendors, another experiment is performed:
Using a different assembly machine, five types of
capacitors from four 0201 component vendors (Comp
A, C, E and F) are tested.
Two types are tested from Comp C, labeled as "Comp
C(1)" and "Comp C(2)." The only difference is that
Comp C(2) has a smaller spacing between it and the
pocket. Twenty-three 0201 test panels are used in
this study.
Five types of capacitors are equally distributed in
panels A and C and a total of 82,800 capacitors are
assembled, with 16,560 components for each type of
capacitor.
Table 3. Defects
per Million Oppertunities for types of 0201
Components.
The pickup errors
for each type are presented in Table 4 and Figure
6.
As can be seen, there are mainly three types
of defects:
mispick,
component
pickup in vertical position
andcamera-recognition
error.
No component drop
is detected for any type of component tested.
Results: Comp A shows the best pick-up. The
pocket-size difference between Comp C(1) and
C(2) shows some improvement on the pick-up process.
However, the improvement is not very significant.
Figure
6.
Table 4. Pickup
Errors for Various 0201 Components.
All boards are inspected post-reflow
and five types of defects are observed
(Table 5 and Figure 7).
The defect level from component F and
C(1) is lower than 100 DPMOs.
Figure 7.
Table 5.
Number of Defects Post-reflow (By
Component Type).
The
final comparison (Table 6) is made by
combining the results from post pick-up
and post-reflow.
Table
6.
Conclusion
In
this work, a QV is designed for 0201 packages including
more than 5,000 locations for 0201s and 1,400 locations
for 0402 components, with the pad-to-pad spacing for
0201 components ranging from 0.010, 0.008 to 0.006".
No-clean solder pastes and 0201
components from several different vendors are evaluated.
Detailed studies are carried
out on important process parameters, including solder
paste printing, component pick-and-place and reflow.
Based on this study, 0201 capacitors
show a higher defect level than that of 0201 resistors,
and solder pastes with higher tackiness exhibit a
lower defect level for 0201 assembly.
References
M. Wang, D. Shangguan, D. Geiger, F. Mattsson,
S. Yi, "PCB Design Optimization of 0201 Packages
for Assembly Processes," IMAPS 2002, Dallas.
M. Wang, D. Geiger, K. Nakajima, D. Shangguang,
C. C. Ho, S. Yi, "Investigation of Printing Issues
and Stencil Design for 0201 Package," SMTA 2001,
Chicago.
ACKNOWLEDGEMENT:
The authors recognize the assistance of their colleagues
at Flextronics worldwide as well as engineers and
managers from several suppliers for their help and
support during the course of this project. This article
was originally published at SMTA International.
Mei Wang, Dr. Dongkai Shangguan, M. T. Ong, Fredrik Mattsson, David Geiger, and SammyYi may be contacted at Flextronics, 2090 Fortune Dr., San Jose, CA 95131; (408) 576-7000; Web site: WWW.flextronics.com
David
Clegg, Rebecca Cole, John Franka, Doug Mitchell, Dave
Wontor,
Motorola Semiconductor Products Sector,
Austin, Texas
