LEAD FREE WAVE SOLDERING:

PROCESS OPTIMIZATION FOR SIMPLE TO HIGHLY COMPLEX BOARDS

Denis Barbini
Vitronics Soltec, Inc.
Stratham, NH, USA
[email protected]

Paul Wang
Microsoft, Inc.
Mountain View, CA, USA
[email protected]

Peter Biocca
ITW Kester, Inc.
Allen, TX, USA
[email protected]

Quyen Chu
Jabil, Inc.
San Jose, CA, USA
[email protected]

Tim Dick
Cisco Systems, Inc.
San Jose, CA, USA
[email protected]

Denis Jean
Plexus, Inc.
Neenah, WI, USA
[email protected]

Stuart Longgood
Delphi Corporation
Kokomo, IN, USA
[email protected]

Chrys Shea
Cookson Electronics, Inc.
Jersey City, NJ, USA
[email protected]

Keith Sweatman
Nihon Superior, Inc.
Maroochy River, Australia
[email protected]

Mike Yuen
Foxconn, Inc.
Houston, TX, USA
[email protected]
ABSTRACT assembly chapter, the technology forecast for wave
Today’s wave soldering processes solder electronic soldering is described and is listed in Table 1.
assemblies within the large eutectic tin lead processing
window. The issues surrounding the conversion to lead free
assembly are multiple and varied. Logistics, cost, material
selection, and equipment choices/options are some of these
challenges that require planning and organization.
However, at the core of lead free assembly is soldering. The
iNEMI lead free wave soldering team embarked on a multi-
tiered project that focuses specifically on two aspects:
identifying the impact of critical parameters on the
development of a reliable, robust lead free wave soldering
process as well as determining the process that achieves the
optimized soldered joint.

The critical aspects of the lead free wave soldering process

can be broken down into several categories which include
component reliability, board reliability, flux type and
specifications, alloy and alloy composition, its melting Table 1. The 2007 iNEMI Technology Forecast for Wave
behavior, wave equipment capacity and power Soldering.4
requirements, profile attributes and flexibility.
This updated table provides increased resolution on alloy
This collaborative effort has investigated process parameter composition, component pitch, contact time, and
and material impact on the soldering process and joint atmosphere compared to the 2004 iNEMI Technology
formation. Addressed in this lead free implementation study Forecast.
were inspection criteria, failure definition and measurement,
followed by root cause analysis and ultimately optimized The technology forecast continues to provide consistent data
process confirmation. The result of this investigation was to on the nature of the wave soldering landscape. The issues
lay the foundation for a broader effort to characterize the today continue to be led by the enduring changes in
reliability of through-hole joints on a test vehicle materials and their maturity combined with a significant
specifically designed to test the norms and practices used in effort to minimize cost. Alloys containing silver adversely
tin lead wave soldering and develop new standards for lead affect cost thus pressuring assembly houses to find
free wave soldering. alternative cheaper materials. Another pressure is the
complexity of board technology. Today’s high end board
Overall, this investigation looks to bring an understanding designs are quickly migrating to increased layer count,
of how critical wave soldering parameters correlate and thicknesses, and complexity, thus exacerbating the
provide the reader with the information necessary to make challenges already associated with developing and
educated decisions in selecting materials. This study will maintaining a robust and reliable wave soldering process.
also provide insight into the process issues that one will This results in an overall situation where materials are
encounter so that a rational implementation strategy for a pushed to the limits of their respective specifications in
reliable and robust lead free wave soldering process can be terms of exposure to elevated temperature for extended
achieved. times. As the lead free implementation progresses, various
questions have arisen. This research takes an in-depth look
Key words: Lead Free, Pb Free, wave soldering, process, at the challenges encountered in developing a lead free wave
solder, alloy, flux. soldering process based on the specific products as well as
on specific materials.
INTRODUCTION
The 2004 iNEMI roadmap identified a gap in industry For these and other similar reasons, the iNEMI Lead Free
information and research relating to lead free wave Wave Soldering Project focused on three critical areas:
soldering processes and reliability of these lead free 1. Materials Selection
assemblies. Since this time, only a limited number of 2. Process Optimization
comprehensive investigations have taken place.1,2,3 As a 3. Solder Joint Performance
result, the Lead Free Wave Soldering project was developed In order to achieve these goals the project participants
in order to characterize and quantify the impact caused by developed a two-phase approach.
the transition to lead free solder on the wave process itself
as well as the impact on the performance and reliability of The companies supporting this project during the execution
lead free solder joints. The 2007 iNEMI roadmap was of Phase I are shown in Figure 1.
recently made available to member companies and will be
released to the public in March 2007. Within the board
 with timely and statistically backed understanding of lead
free wave soldered joints.

In order to achieve this goal, the team designed and

fabricated a test vehicle that aims to understand future
assembly requirements and consequently develop new
standards and best practices for lead free wave soldering
assembly. This board’s call name is “GTLO”, Get The
Lead Out! As shown in Figure 2.
Figure 1. iNEMI Lead Free Wave Soldering Project
Companies Supporting Phase I.

The first phase of the project focused on characterizing

process-related challenges and optimization of a lead free
wave soldering process for various factors including: fluxes,
alloys, and board thicknesses. Characterizing the window
of opportunity for various materials specifically designed
for lead free assembly and its impact on wave soldering
process is based on the quantification and analysis of
specific defects. The research performed in Phase I
accomplished the aforementioned goal by investigating
various levels of two factors:
• First, selection of a broad variety of materials allows for
the determination of specific interactions. Alloys,
fluxes, board laminate and finishes, component types
and metallurgy, board design including thickness,
thermal tie design, finish, and component orientation
exert individual restrictions on the wave soldering
process that results in an overall window of
opportunity.
• The second factor is the flexibility allowed in
developing the robust soldering process. Development
of the wave process requires selection and control of
various parameters including flux amount, atmosphere,
preheat temperature, contact time, alloy temperature, Figure 2. The GTLO Board
and wave configuration.
It is with this board and component mix that the team will
The findings provide insight into the optimization of the derive solutions and specifications to the challenges posed
wave soldering process for given material combinations. in the technology forecast.
Confirmation of the data analysis was achieved by soldering
boards utilizing the optimized parameter settings for the EXPERIMENTAL
respective material combinations. The focus of this latter In the design of this experiment a variety of materials were
effort was to provide a data driven solution for the used. Criteria for the selection of these materials was based
optimized wave soldering process which is an essential part on industry assembly norms, gaps in information or industry
of a robust and reproducible lead free assembly process. need, and known reliability of specific materials not
The scope of the project included accomplishing this goal considered in the experiment. Wave soldering equipment
for three different lead free alloys as well as for tin lead on use and setup was determined based on common industry
three different board thicknesses. practices, basic configuration and operation.

The goal of the iNEMI lead free wave project is to Materials

ultimately characterize solder joint performance. Phase I Fluxes
provides the optimized settings that result in IPC class 3 Soldering fluxes are known to exert a significant influence
acceptable through-hole fill for specific material on the resulting solder joint. Without performing a
combinations. Phase II of the project focuses on comprehensive flux investigation, three fluxes from three
standardizing the lead free wave assembly process based on flux composition categories were selected.
the Phase I process development and optimization so that 1. No Clean Alcohol Based Flux- 4.6% solids content
only solder joint performance will be evaluated. The intent 2. No Clean Water Based Flux
of Phase II is to provide the electronics assembly industry 3. Organic Acid Water Soluble
Alloys The color code is followed in the schematic per through-
The various, available lead free alloys on the market today hole. There are three locations for 120 pin PCI and twelve
consist of tin in excess of 95%. As a result, the melting 20 pin dual row berg stick connectors. This was done in
behavior of the alloys will be dominated by the melting order to assess the impact ground plane connection to the
point of tin, 232°C. In this experiment three lead free alloys barrel has on through-hole penetration. The green arrow
were selected based on project interest at the time the indicates the direction of the board as it travels in the wave
experiment was executed. Tin lead was also included to soldering machine.
provide a benchmark to the majority of today’s production
lines. Board finish was divided into two categories:
1. SAC 305 1. Cu OSP
2. Sn100C 2. Alloy Specific HASL
3. SACx For each of the three lead free alloys and tin lead, the
4. SnPb respective board finish was applied. ie, In the experiments
where SAC 305 alloy was used, the board finish was also
Board SAC 305 HASL. The OSP applied was Entek Plus HT.
The test vehicle used in Phase I was a Cookson Electronics
designed test board referred to as the “Skate” board. The The board thicknesses utilized in this experiment were 62
fabrication included the use of Polyclad 370 HR laminate mils, 94mils, and 135 mils.
with a 175°C Tg and a high Td.
The layout of the Skate test vehicle used in this experiment Components
included two types of through-hole components, four banks PCI Connector
of passives, two banks of SOT-23’s oriented in two One of the through-hole components selected was the 120
directions, and three QFP’s with varying lead pitch. A pin PCI connector with a lead free finish. This component
bottom view is illustrated in Figure 3. was only available in one lead length. All three PCI
connector locations were populated.
No connection Shorter spokes on WW Longer spokes on WW Direct to ground

Berg Stick
The 20 pin dual row Berg Stick connector came in a lead
free finish as well as a tin lead finish. The bottom side lead
protrusion for this component was kept constant over the
three board thicknesses by mechanical manipulation of the
pins within the plastic housing. Seven out of the twelve
locations were populated. Only one tin lead finished
component was populated per board. The remaining six
were lead free finish.

QFP
Three QFP sizes were used in this board design. Pad layout
included solder thieves with varying design based on lead
count and pitch.
ƒ 64 lead – 0.5mm pitch
ƒ 80 lead – 0.65mm pitch
ƒ 64 lead – 0.8 mm pitch

SOT
A total of twelve SOT-23’s were placed on each board. The
twelve were divided into two groups of six. One group was
parallel to the direction of board travel while the other group
Figure 3. A Bottom Side View of the Skate Test Vehicle was perpendicular to the direction of board travel.

Provided at the top of the board is a color code of the four Passive Resistors
copper connection types designed into the test vehicle. The A total of four banks of passives were assembled onto each
four connection types are: board. Two sizes were used including 0805 and 0603. As
1. No connection with the SOT configuration, the two banks of 0805’s and
2. Short Spokes on Wagon Wheel 0603’s were designed parallel and perpendicular to the
3. Long Spokes on Wagon Wheel direction of board travel.
4. Direct to Ground
Equipment Variable Level 1 Level 2 Level 3 Comments
A Vitronics Soltec Delta 6622 wave soldering machine was Atmosphere Nitrogen Air N/A
used to assemble the boards for Phase I. The wave
Belt speed 3 (91) 4.5 (137) 6 (183) .062”/.094”
soldering machine was configured with a dual head spray ft/min (cm/min) 2 (61) 3.5 (107) 5 (152) .135”
fluxer which was operated with a pump system and
Preheat Temp 90°C 110°C 130°C (alcohol flux)
delivered each flux with specifically designed nozzles. The 100°C 115°C 130°C (VOC free flux)
preheat technology and configuration for the three zones of
Flux Quantity low med high
preheat consisted of forced convection modules in the first
Flux Type Water Alcohol OA
two preheating zones and a calrod module in the last
Chip Wave on off on
preheating zone. The wave configuration consisted of a
Solder Temperature 255°C 265°C 275°C
chip wave and main wave. The chip wave is designed to
Board Thickness 0.062 (1.6) 0.094 (2.4) 0.135 (3.5) Inches (mm)
deliver a turbulent flow of solder and the main wave is
designed to deliver a laminar flow of solder. The nitrogen Table 2. Definition of Parameters and Materials Used for
inerting system works on the blanketing concept and was the Inner Array.
operating at the following N2 flow settings of 30,50, and 80
l/min. A Taguchi L18 orthogonal array, listed in Table 3, was
required for this set of materials and process parameters.
Parameter settings for each subsystem were controlled as The L18 was executed for each individual alloy and board
required by individual runs in the design of experiments. finish resulting in a total of eight separate but comparable
Further, all systems, including transport, were calibrated to sets of data. Three replicates were assembled for each run.
deliver directly comparable outputs. ie, The flux amount The run order was determined by minimizing the time
delivered was calibrated by flux type and conveyor speed so required between runs to complete the various parameter
that a low amount of flux delivered to a board traveling at changeover.
slow or fast speed was identical regardless whether the
water based no-clean flux or the alcohol based flux was The specific values of preheat temperature and flux quantity
utilized. This was accomplished by measuring the mass, used in each run are based on the flux manufacturer’s
volume, and through-hole penetration of flux delivered to specification. Low and high settings are near the respective
the board at all potential experimental conditions. min and max of the recommended values and the medium
setting is midway between to the two.
All alloys were contained in individual solder pots that were
switched out as needed to complete the DOE. This was The criteria used to determine the optimized process is
done in order to eliminate any possible cross contamination based on the characterization of the through-hole
issues. penetration as per IPC 610D Class 3 or 75% hole-fill. The
target was also consistent for desired hole of 100%. Hole-
Design of Experiment fill was determined by 5DX laminographic analysis. Other
The goals of Phase I include ascertaining the relationship or defects such as shorts, solder balling, skips, and also solder
influence of various parameters on the formation of defined shrinkage were visually characterized but not included in
defects; derivation and confirmation of an optimized process determining the optimized process.
for each lead free alloy and board thickness. The team
Actual Run DOE Speed Preheat Flux Chip Wave Solder Board
selected the Taguchi methodology to achieve these goals. Order Standard Atmospher Ft/min Temp Quantity Temp (°C) Thickness
During the development of the experiment, a number of 1
Order
5
e
N2
(cm/min)
4.5 (138) med med
Flux Type
OA on 255 0.062
process parameters and materials were of interest, based on 2 1 N2 3 (92) low low Water on 255 0.062
3 16 Air 6 (183) low high Alcohol on 255 0.094
the perspective of various industry segments. Ultimately a 4 12 Air 2 (61) high med Alcohol on 255 0.135
consensus was reached with eight parameters and materials 5
6
14
9
Air
N2
3.5 (107)
6 (183)
med
high
high
low
Water
OA
off
off
255
255
0.135
0.094
placed into the inner array at three levels. This is listed in 7 10 Air 3 (92) low high OA off 265 0.062
8 2 N2 3 (92) med med Alcohol off 265 0.094
Table 2. The four alloys and two board finishes were placed 9 15 Air 4.5 (138) high low Alcohol on 265 0.062
in the outer array. 10
11
7
6
N2
N2
5 (152)
4.5 (138)
low
high
med
high
Water
Water
on
on
265
265
0.135
0.094
12 17 Air 5 (152) med low OA on 265 0.135
13 13 Air 4.5 (138) low med OA on 275 0.094
14 4 N2 3.5 (107) low low Alcohol off 275 0.135
15 8 N2 6 (183) med high Alcohol on 275 0.062
16 11 Air 3 (92) med low Water on 275 0.094
17 3 N2 2 (61) high high OA on 275 0.135
18 18 Air 6 (183) high med Water off 275 0.062

Table 3. The Taguchi L18 Experiment Listed by Actual

Run Number.

Profiles
Characterization of the temperature profile and contact time
was accomplished by using the ECD Mole© profiler and the
ECD WaveRider©. Similar to the flux calibration, contact
time and preheat temperature were adjusted accordingly to
ensure that the three levels of preheat and contact time were
identical for each board thickness. Figure 4 illustrates a
typical profile used to quantify topside preheat temperature
and peak temperature on the Skate test vehicle.
Defect Defect
Analysis No defect Defect
S3-25% S3-25% S3-25%
S1-50%
Various forms of analysis were employed in Phase I. The S1-50%
S4-75%
reporting in this research is based on visual characterization
of shorts and skips as well as x-ray analysis of hole-fill.

Defect Defect Defect Defect

All Slices S2-0% S3-25% S4-75%
S1-50% Void
S4-75%

Figure 6. Data Reporting Scenarios for Through-Hole

Penetration Based on 5DX Analysis.

Cross Sectioning
Select samples were cross sectioned to perform various
analyses including hole-fill, barrel integrity, board integrity,
and intermetallic formation. The procedure used to make
these cross sections has been documented previously.5

Figure 4. Typical Lead Free Wave Profile. RESULTS AND DISCUSSION

The execution of Phase I included the fabrication of the test
5 DX X-Ray Analysis vehicle, applying the correct finish onto the defined board
Hole-fill was determined by using a programmed Agilent thicknesses, procurement of the components, surface mount
5DX x-ray instrument. The data collection of barrel hole- assembly and glue cure, flux procurement, and equipment
fill was based on quarter percentiles as illustrated in Figure setup. Once this was in place, a total of 324 boards were
5. The types of defects observed are illustrated in Figure 6. assembled per the Taguchi L18 orthogonal array experiment.
For the purposes of optimizing the process, only defects at The boards were visually inspected on site, followed by
the 75% hole-fill or the top quartile were processed. further inspection by professionally trained inspectors at an
However, increased sensitivity to process variations was EMS, and then sent for the characterization of through-hole
provided by defects occurring at the various levels. Theis solder penetration by 5DX. As explained in the
measurement was made on both the Berg Stick and PCI Experimental section, the criteria employed to determine the
connectors. output response was hole-fill greater than 75%. In Table 4,
the results of x-ray analysis are provided for SAC 305 on
CuOSP and HASL finished boards. The same procedure
was performed on the SACx and Sn100C runs.
Actual Run DOE Speed Preheat Flux Chip Wave Solder Board SAC 305
Order Standard Atmospher Ft/min Temp Quantity Temp (°C) Thickness HASL OSP
Order e (cm/min) Flux Type
1 5 N2 4.5 (138) med med OA on 255 0.062 0.0 0.0
2 1 N2 3 (92) low low Water on 255 0.062 2.0 2.7
3 16 Air 6 (183) low high Alcohol on 255 0.094 41.0 44.0
4 12 Air 2 (61) high med Alcohol on 255 0.135 32.7 200.0
5 14 Air 3.5 (107) med high Water off 255 0.135 399.0 394.3
6 9 N2 6 (183) high low OA off 255 0.094 48.0 44.3
7 10 Air 3 (92) low high OA off 265 0.062 0.0 0.0
8 2 N2 3 (92) med med Alcohol off 265 0.094 3.0 0.0
9 15 Air 4.5 (138) high low Alcohol on 265 0.062 0.0 0.0
10 7 N2 5 (152) low med Water on 265 0.135 358.3 350.3
11 6 N2 4.5 (138) high high Water on 265 0.094 0.3 0.3
12 17 Air 5 (152) med low OA on 265 0.135 56.3 47.3
13 13 Air 4.5 (138) low med OA on 275 0.094 6.0 25.0
14 4 N2 3.5 (107) low low Alcohol off 275 0.135 56.7 80.3
15 8 N2 6 (183) med high Alcohol on 275 0.062 2.3 3.0
Figure 5. 5DX Analysis of Through-Hole Penetration 16 11 Air 3 (92) med low Water on 275 0.094 1.7 37.0
17 3 N2 2 (61) high high OA on 275 0.135 58.7 55.0
Based on 25% Increments from Bottomside. 18 18 Air 6 (183) high med Water off 275 0.062 1.0 1.0

Table 4. Through-Hole Penetration Defect Reporting for

SAC 305 on CuOSP and SAC305 HASL.

The defect data provided is the average number of defects

per board as measured by 5DX. It is interesting to note that
there are three runs which resulted in no defects while two
runs resulted in hundreds of defects. The remaining runs are
distributed between 1 defect and 80.3 defects. This
provided the team with the distribution of data points that to an extent compared to results on the CuOSP finished
allows for a statistically significant determination of both boards. The derived optimized process for best through-
process interaction and optimization. Figures 7 and 8 chart hole penetration for SAC 305 on SAC 305 HASL is:
the influence of the eight inner array parameters on through- Atmosphere: nitrogen
hole solder penetration for CuOSP and SAC305 HASL Transport Speed: 3ft/min or 6ft/min
finished boards, respectively. A lower y-axis value Flux type: OA
represents fewer through-hole defects. Flux amount: medium
Preheat temperature: 90°C
Response: Mean Count
Chip Wave: on
Average Number of Holes per Board w/<75% Fill
SAC305
Pot Temperature: 265°C
200
180 305OSP Board Thickness: 62 mil
160
140
120
The intent of Phase I was to develop and confirm an
100 optimized process for three lead free alloys and on varying
80
60 board thickness and board finish. Furthermore, the team
40 regarded as a priority, developing a process that most
20
0 resembled the majority of typical assembly methods utilized
F lu x M e d
F lu x L o w

C h ip O ff
C h ip O n
F lu x
C ool
S lo w

W a rm

0 .1 3 5
0 .0 6 2
0 .0 9 4
Fast

H ot

1868
OA3 3 5 5
Med

255
265
275
330
A ir

by project members. Based on this, it was not possible to

N2

Alcohol
Water

select one board thickness or OA flux. The team elected to

confirm the statistical model by soldering all board
Figure 7. Response Chart for the Eight Inner Array thicknesses with the no clean alcohol based flux.
Parameters Influence on Through-Hole Solder Penetration
for SAC305 on CuOSP . Table 5 lists the derived optimized processes for the three
lead free alloys. The only difference listed is the required
The behavioral analysis for the eight parameters on through- topside preheat temperature of SAC305 versus both Sn100C
hole solder penetration for CuOSP finished boards is and SACx. The former requires a temperature of 130°C
characterized by both linear or first order interactions except versus 110°C for the two latter alloys.
for flux amount which peaks at the medium amount. Based
on this response chart, the optimized process is readily Alloy Atmosphere Speed
Preheat Flux
Flux Type
Solder Board
Temperature Amount Temperature Thickness
identifiable. The derived optimized process for best
(ft/min) (°C) (°C) (mil)
through-hole penetration for SAC 305 on CuOSP is: SAC 305 N2 3 130 Med Alcohol 265 62, 93, 135
Atmosphere: nitrogen SACx N2 3 110 Med Alcohol 265 62, 93, 135
Transport Speed: 3ft/min SN100C N2 3 110 Med Alcohol 265 62, 93, 135

Flux type: OA Table 5. Derived Optimized Process for the Three Lead
Flux amount: low Free Alloys.
Preheat temperature: 130°C
Chip Wave: On Confirmation of the derived optimized process is required to
Pot Temperature: 275°C ensure that the design and results of the L18 Taguchi
Board Thickness: 62 mil experiment are valid. Upon finalization of the optimized
process, confirmation runs were performed on the original
Response: Mean Count lot of boards. Due to the amount of time that passed
Average Number of Holes per Board w/<75% Fill
SAC305 between the executions of the original experiment to the
60
confirmation runs, the stored boards had absorbed a
50
305HASL significant amount of moisture that resulted in board
40 outgassing during the wave process and thus created
30 numerous defects. A second batch of confirmation runs
20 were then performed on boards that were baked at 125°C for
10
a minimum of 24 hours in a high purity pure nitrogen
atmosphere. Only HASL boards were utilized in the
0
confirmation run since the OSP finish was compromised.
Alcohol
Flux Low
Flux Med
Flux High
Slow

Hot
N2

Med

Warm
Cool

Chip Off
Chip On
Fast

1868

0.062
0.094
0.135
330

OA3355

255
265
275
Air

Water

Completion of the confirmation runs for each lead free alloy

Figure 8. Response Chart for the Eight Inner Array was successful after taking the proper precautions and PWB
Parameters Influence on Through-Hole Solder Penetration conditioning. Through-hole solder penetration was then
for SAC305 on SAC305 HASL. measured using the same technique and criteria as in the
analysis described for the L18 experiment.
The same analysis was performed for SAC305 on SAC305
HASL finished boards. As shown in Figure 8, the
optimized process for best through-hole penetration varies
Table 6 provides the defect results for through-hole solder
penetration as a function of alloy and board thickness as
measured on all Berg Stick locations.

Board
SAC305 SACx Sn100C
Thickness
62 mil 0 0 0
93 mil 0 0 0
135 mil 2 7 3
Table 6. Through-Hole Solder Penetration Defect Count
for HASL Boards Assembled at the Optimized Settings.

The derived optimized wave soldering process for the three

alloys were confirmed based on defect free soldering for the
62 mil and 93 mil thick boards. The 135 mil thick boards Figure 9A. PCI Lead Protrusion on the 62 mil Thick Card.
had minimal instances of less than 75% hole-fill for each of
the alloys. Overall, it is possible to conclude that the L18
Taguchi experiment was successful in providing the data
needed to develop an optimized process for a variety of
materials including flux type, alloy type, and board
thickness. The information gathered in Phase I will be used
in the development of a fixed soldering process for Phase II
in which the focus is not process optimization but solder
joint performance.

GENERAL OBSERVATIONS AND TRENDS

Over the course of Phase I, a number of challenges had to be
addressed and solved in order to meet the stated goals. This
section provides some insight into various industry
questions, challenges, and wave soldering assembly defect Figure 9B. PCI Lead Protrusion on the 93 mil Thick Card.
trends.

Impact of Component Design on Through-Hole Solder

Penetration
Although the Skate test vehicle only had two types of
through-hole connectors, sourcing the PCI connector with
varying lead lengths was not possible. As a result, the PCI
connector lead protrusion changed as a function of board
thickness to the extent that no lead protrusion was observed
on the 135 mil thick card. Figures 9A-C illustrate the
challenges encountered with lead protrusion on the PCI
connector. A direct correlation was observed between
through-hole solder penetration and lead protrusion. An
inverse relationship exists wherein as the lead protrusion
decreases, defects increase. Figure 9C. PCI Lead Protrusion on the 135mil Thick Card.

Impact of Board Construction on Through-Hole

Penetration
The Skate test vehicle was designed with specific barrel tie-
ins consisting of varying amounts of copper as defined in
Figure 3. This investigation collected and analyzed the
impact of copper connection to through-hole solder
penetration. Figures 10A and 10B illustrate the impact of a
marginal wave soldering process on through-hole solder
penetration. Figure 10A provides a topside view of the PCI
connector through-holes which is visible to the inspector
while Figure 10B provides the cross sectional view which
allows one to measure actual hole fill. Moreover, the
influence of copper tie-ins to the barrel is observed in one of
the L18 runs. In these images the barrel construction is
described from left to right as short spoke on wagon wheel,
no connection, long spoke on wagon wheel, and direct
connection.

Figure 11A. A Five Pin Short on the 0.5mm QFP

Figure 10A. Visual Inspection of Specific Through-Hole

Vias.

Figure 11B. Closeup of Multi-lead Bridging on a QFP.

Figure 10A. Cross sectional view of Specific Through-Hole

Vias.

In the cross sectional view it is possible to determine that

the long spoke wagon wheel (third from the left) and direct
connection (fourth from the left) have less then 75%
through-hole penetration. This behavior was observed Table 7. Relationship of Bridging to QFP Lead Pitch and
regardless of alloy or board finish and is mainly dependent Alloy.
on two factors: the wave soldering process and board
complexity/construction. The analysis of the bridging occurrence clearly provides the
challenge one has in assembling QFP’s with pitches less
Impact of QFP Pitch on Bridging than 0.65mm. All alloys were characterized by greater than
Based on the visual inspection data collected after the 1000 leads with bridges on the 0.5mm pitch QFP.
assembly of the 324 boards, it was possible to determine Comparison between alloys is less definitive since it is
some empirical trends on QFP bridging. Figures 11A and observed that each alloy performed equal to or better than
11B illustrate two instances of shorts occurring on a QFP. the other two for different QFP’s.

Impact of Board Layout on Skips

The Skate test vehicle component layout included surface
mount components aligned parallel and perpendicular to the
direction of the board over the wave. In addition, the DOE
inner array design focused on the utilization of the turbulent
flow chip wave. Analysis of the visual data collected on
skips in the area of SOT-23’s, as shown in Figure 12, formation. All types of solder joints were produced ranging
provides insight into which parameter(s) are most critical in from highly oxidized, not solder-able surfaces as illustrated
minimizing skips. As listed in Table 8, the single most in Figure 13 to the optimum solder joint in Figure 14.
important parameter is the use of the chip wave. The
turbulent solder flow penetrates the areas of the component
lead/pad that are biased against proper wetting. Thus
allowing for a proper joint to form.

Figure 13. Poor Wettability Resulting in Unacceptable

Joints.
Figure 12. Board Layout as Per IPC and Contrary to IPC
Design Guidelines.

Figure 14. Perfectly Formed Joints.

Table 8. The Influence of Process Parameters on
Minimizing Skips.
Addressed in this study are inspection criteria, failure
definition and measurement, followed by root cause analysis
CONCLUSION
and ultimately optimized process confirmation. The result
This research provides insight into several of the key
of this investigation was to lay the foundation of a broader
questions and challenges observed in today’s lead free wave
effort to characterize the performance of through-hole solder
soldering process. The iNEMI lead free wave soldering
joints on a test vehicle specifically designed for testing the
team embarked on a multi-tiered project that focuses
norms of tin lead wave soldering.
specifically on two aspects: Identifying the impact of
critical parameters on the development of a reliable, robust
Overall, this investigation looks to bring an understanding
lead free wave soldering process and determining the
of how critical wave soldering parameters influence the
process that achieves the optimized soldered joint.
various outputs. It also attempts to provide the reader with
the information necessary to make educated decisions in
It is clear that in order to achieve the optimized lead free
selecting materials and controlling various process
wave soldering process one must identify component
parameters in order to execute a rational implementation
features, board finish and thickness, flux type and
strategy for a reliable and robust lead free wave soldering
specifications, alloy and alloy composition, its melting
process.
behavior, wave equipment capacity and power
requirements, profile attributes and flexibility.
ACKNOWLEDGEMENTS
The authors and project members would like to gratefully
This collaborative effort investigated process parameter and
thank the following people for support during the execution
material impact on the soldering process and joint
of Phase I and its analysis: John Norton, Norm Faucher,
Gerjan Diepstraten, and Bruce Quigley for build support,
Ursula Marquez de Tino for the many cross sections, Sam
Greenfield for performing all of the 5DX analysis. The
authors would also like to thank the management of each
participating company for their support in this endeavor.
Without their continued support, this project would not have
achieved its goals.

REFERENCES
1. Gleason, J., Reynolds, C. “iNEMI Advanced Lead Free
Assembly and Rework Project”, April 2005.
2. Hilman, D. et.al. “JCAA/JGPP Lead Free Solder
Project” SMTAI, 2005.
3. Holder, H. et. al. “Reliability of Partially-filled SAC305
Through Hole Joints” 2006 APEX Proceedings.
4. Excerpt from the Board Assembly Chapter of the 2007
iNEMI Roadmap.
5. Marquez de Tino, U. “Procedure for the Preparation of
Cross Sections” Vitronics Soltec information document.