IPC-7095C-2013.pdf - 第120页
any occurrence of exceeding the characteristics for void size shown in T able 7-7, requires 100% evaluation of the total lot. The appropriate corrective action may vary based on class of product and customer requirements…
For in-process control, use >35% area as the threshold with
a >50% void diameter threshold. For reject-rework, use
>45% area as the criteria with a >65% void diameter
threshold.
Use of X-ray requires some caution regarding radiation
overexposure on vulnerable materials or components, as
the radiation dose from different x-ray systems can vary
widely. The implications of radiation dose to radio-
sensitive components may need to be discussed with your
component supplier and the dose rates that will be achieved
under typical x-ray inspection should be discussed with
your x-ray equipment supplier.
Although much description data is provided on identifica-
tion of voids, there is not always correlation to significant
factors related to thermal cycle life of the joint. In addition,
no data is currently available on rework vs. thermal cycle
life.
Normal area percent voids is 15 ± 10% of the projected
area.
Note: Voids at ~30% warrant process-control attention.
Voids >50% warrant rework (see Section 7.5).
Considering a 0.20 mm diameter void as an example, Table
7-7 lists the void percentages for different ball sizes. The
percentage void diameter detectable becomes larger as sol-
der joint size decreases; that is, 27% on a 0.75 mm joint
inflates to 67% on a 0.30 mm solder joint.
Defect determination is made by the product’s reliability
requirements. As an example, if the maximum allowable
void size is 31% of the solder ball diameter, the equivalent
of the void area is 9%. This can be either one void, or the
summation of many voids. Some of the newer X-ray equip-
ment use algorithms that are able to summarize the void
areas. The current algorithms for X-ray tomography do not
perform the summation of the voids.
For a single void, X-ray tomography can identify a defect
caused by a void that is greater than the pre-determined
size.
Example: If the solder ball size = 0.75 mm and the maxi-
mum allowable void size = 30% of the ball diameter, the
maximum void size at the center of the ball would be cal-
culated as follows:
30% of 0.75 mm
(0.75 mm)(0.3) = 0.225 mm maximum void diameter
When the void is not in the center of the ball and near the
land of either the board or the component, the cross-
sectional diameter of the ball will be reduced as well as the
maximum allowable size for a void.
Example: If the ball diameter at the land is approximately
equal to the land and the land size is 75% of the ball size
(25% reduction), then:
75% of 0.75 mm = 0.56 mm ball diameter at the land
30% of 0.56 mm = maximum void diameter
(0.56)(0.3) = 0.17 mm maximum void diameter
at the land
7.6.4 Sampling Plans for Void Evaluation Because of
the industry concern regarding voids, this standard attempts
to define criteria for void baseline goals and process con-
trol techniques. Everyone agrees that it makes no sense to
throw away good product, or rework product where a void
is identified without some indication as to the complexity
and the impact on reliability of that condition.
Void occurrence criteria are not based on 100% inspection,
but are accomplished through the use of sampling plans.
The sampling plan conditions are identical to those shown
in IPC-6012 and are repeated in Table 7-8. It should be
noted that the table is based on C=0. What this means is
that, as within all IPC standards, when a sample is selected,
IPC-7095c-7-45
Figure 7-45 Examples of Suggested Void Protocols
025" land 06" void
24% DIAMETER
6% AREA
025" land 013" void
52% DIAMETER
27% AREA
.020" X .030"
.010" VOID
50% DIAMETER
10% AREA
20 X 40 + 10,5,5,5,5,5,5
24% AREA
20 X 40 + 10,10,5,5,3
26% AREA
.020" X .030"
.014" VOID
70% DIAMETER
19% AREA
.020" X .030"
.016" VOID
80% DIAMETER
25% AREA
.020" X .030"
.018" VOID
90% DIAMETER
31% AREA
025" land 015" void
60% DIAMETER
36% AREA
025" land 010" void
50% DIAMETER
16% AREA
25 + 6,6,6,3,3,3,3,3
24% AREA
25 + 10,10,6
38% AREA
25 + 10,6,6,6
37% AREA
25 + 6,6,6,4,3,3,3,3
25% AREA
Table 7-7 Ball-to-Void Size Image -
Comparison for Various Ball Diameters
Solder Ball
Diameter/
X-Ray Image
Void 0.20 mm Diameter
% Void Diameter % Void Area
0.85 mm 24% 6%
0.75 mm 27% 7%
0.65 mm 31% 9%
0.55 mm 36% 13%
0.45 mm 44% 20%
0.40 mm 50% 25%
0.30 mm 67% 44%
January 2013 IPC-7095C
105
Copyright IPC-Association Connecting Electronics Industries
Provided by IHS under license with IPC
Not for Resale
No reproduction or networking permitted without license from IHS
--`,,```,,,,````-`-`,,`,,`,`,,`---
any occurrence of exceeding the characteristics for void
size shown in Table 7-7, requires 100% evaluation of the
total lot.
The appropriate corrective action may vary based on class
of product and customer requirements. The ultimate action
is to remove and replace the affected component; however,
those solutions must be carefully evaluated as the product
should have been designed to permit repair procedures that
include reevaluation.
Once a void protocol has been established as to the percent
attachment permitted, random samples should be taken
from the production units and evaluated as to compliance
with the accepted conditions identified in the protocol. The
decision as to how many samples to take depends on the
number of units being produced. A consideration should
also be given to the number of BGAs that are part of the
particular assembly.
The index value of 2.5 is a good representation that can
provide an insight into the capabilities of the process
regarding BGA assembly and the occurrence of voids. The
Level A 2.5 index can be used for evaluating commercial
product used in consumer goods and computer applica-
tions. It is also appropriate for telecommunication equip-
ment. For a Class 2 equipment where high performance
and extended life is required, and for which uninterrupted
service is desired but is not critical, and certain cosmetic
imperfections are allowed an index indicator of 1.5 should
be used.
Class 3 is for High Reliability Electronic Products and
includes equipment for commercial and military products
where continued performance or performance on demand is
critical. Equipment downtime cannot be tolerated, and
must function when required such as for life support items,
or critical weapons systems. Printed board assemblies that
contain BGAs in this class are suitable for applications
where high levels of assurance are required and service is
essential; thus, the evaluation requires a sampling of 1.0
index value.
The purpose of the index value is to define the possibility
that, although the samples reviewed all meet the defined
void protocol, there is still a statistical possibility of the
percentage shown by the index that some of the product
may not meet the conditions established. A 1.0 index for
level C requires that 13 samples be examined from a pro-
duction lot of 125 assemblies. Even if the samples all meet
the protocol, there is still a possibility that 1.0% of the lot
size do not. This relates to less than 2 assemblies and
would be a satisfactory risk in many applications.
7.7 Process Control for Void Reduction
7.7.1 Process Parameter Impact on Void Formation
In
order to establish process control for BGA assembly it is
important to understand what parameters can impact void-
ing frequency and size in board assembly process. The
voids found in solder joints after assembly are usually
called process voids, which are also known as macrovoids.
The process voids are normally caused by the evolution of
volatile ingredients of fluxes and solder pastes. However, if
big voids in solder joints are found after the SMT process,
they may not be generated during the SMT process only.
The ball attach process for attaching solder balls to the
BGA package substrates can also generate voids in the sol-
der balls. It is therefore necessary to investigate the amount
of voids within the solder balls of BGA components. This
voiding level within the balls of BGA components can be
very high. The JEDEC guideline, JESD 217, suggests the
Pre-SMT BGA voids percentage should be less than 15%.
High levels of voids within the BGA solder balls may
Table 7-8 C=0 Sampling Plan (Sample Size for Specific Index Value*)
Lot Size
Class 1 Class 2 Class 3
2.5* 4.0* 6.5* 1.5* 2.5* 4.0* 0.10* 1.0* 2.5* 4.0*
1-8 5 3 2 ** 5 3 ** ** 5 3
9-15 532853**1353
16-25 533853**1353
26-50 555855**1355
51-90 765876**1376
91-150 11 7 6 12 11 7 125 13 11 7
151-280 13 10 7 19 13 10 125 20 13 10
281-500 16 11 9 21 16 11 125 29 16 11
501-1200 19 15 11 27 19 15 125 34 19 15
1201-3200 23 18 13 35 23 18 125 42 23 18
3201-10,000 29 22 15 38 29 22 192 50 29 22
10,001-35,000 35 29 15 46 35 29 294 60 35 29
*Index Value is associated to the A.Q.L. value. If a particular product is determined to be ‘‘critical’’ by the user and a smaller index value is required, the user
shall designate the requirement in the procurement document and should state the ‘‘critical’’ requirement on the master drawing.
**Denotes inspect entire lot.
IPC-7095C January 2013
106
Copyright IPC-Association Connecting Electronics Industries
Provided by IHS under license with IPC
Not for Resale
No reproduction or networking permitted without license from IHS
--`,,```,,,,````-`-`,,`,,`,`,,`---
result in higher levels of voids in these solder balls after the
SMT reflow soldering process.
Reflow time/temperature profile, solder paste formulation,
solder paste amount, component and PCB contamination/
oxidation and multiple reflow are the parameters believed
to impact the voiding level in solder joints after assembly.
However, these parameters may impact BGA voids by dif-
ferent levels. The details of each parameter’s impact to sol-
der joint voiding are discussed as follows.
7.7.1.1 Solder Paste Formulation Flux used to assemble
BGA components tends to generate lower voiding in solder
joints than solder paste. The reason is the oxide of the sol-
der powder in solder paste reacts with the solder-paste-flux
and generates outgassing in reflow process, which leads to
more voiding in BGA solder joints.
BGA solder joints assembled with water soluble solder
paste tend to produce higher voiding than with no-clean
solder paste. This is because the flux formula in water
soluble and that in no-clean paste is different. The water
soluble solder-paste-flux is more aggressive than the
no-clean solder-paste-flux, and reacts faster and more com-
pletely with the oxides on the solder powder. This reaction
leads to more outgassing in the reflow soldering process.
For no-clean lead-free paste, the voiding level in solder
joints with different types/manufacturers of no-clean paste
does not make a big difference. Even though the pastes
vary in flux solvent and metal content, the voiding level is
not significant in solder joints (<1%).
Solder powder type in solder paste also impacts the void-
ing level after board assembly. The finer the solder powder,
the more the surface area to volume ratio of the solder
powder particles and hence the higher their oxide content.
There is more outgassing which is caused by more oxide
reacting with flux in the reflow soldering process.
7.7.1.2 Solder Paste Volume Addition of solder paste
applied to lands on PCBs may result in more voiding in the
solder joints. This is because the more paste applied, the
more flux solvent is present, which leads to more outgas-
sing from reacting with solder powder oxide and solder
ball oxide. Solder paste volume will become more essential
to voiding generation since ball size is getting smaller, and
the ratio of paste volume to solder ball volume is getting
larger.
7.7.1.3 Component/PCB/Solder Paste Contamination/
Oxidation
Solder ball and land surface contamination
may lead to potential process voids generation in the SMT
process. The contamination may present a nonsolderable
surface on the PCB lands which may keep macrovoids
attached to the lands and may generate volatile gases at
soldering temperatures if the contamination decomposes.
Solder powder in the solder paste may oxidize if the paste
is exposed to air or humidity conditions. In the reflow pro-
cess, flux reacts with the oxidation layer of solder powder
and generates volatile gases, which form voids in solder
joints. Finer sized powder has more surface area and leads
to a potential of more outgassing in the SMT process.
BGA solder balls exposed to elevated temperature, humid-
ity condition or ambient temperature long enough develop
a layer of oxide crust formed outside of the solder balls.
This oxide crust can prevent voids from escaping from the
molten ball during reflow soldering.
Compared to solder ball oxidation, the oxide of the PCB
land surface does not impact voiding level solder joints
significantly. This may be because compared to solder
powder surface area, the land surface is much smaller and
flux may not react with oxide on the land surface com-
pletely.
7.7.1.4 Reflow Profile Reflow peak temperature and
time above liquidus (TAL) are two key parameters consid-
ered to impact voiding in reflow profile. Some manufactur-
ers believe that increasing peak temperature and prolonging
time above liquidus will let the voids in solder joints
expand and grow. However, the test results show increas-
ing peak temperature and prolonging time above liquidus
only bring limited impact to voiding level in solder joints.
For example, increasing peak temperature from 235°C to
250°C and TAL from 35 sec to 85 sec, the mean value of
voiding level only increased less than 1%.
There are two types of conventional oven profile, ramp
soak spike (RSS) and ramp to spike (RTS) profiles. The
RSS profile is recommended to control voids in solder
joints by some paste suppliers, because the soak stage may
allow the flux to function well to get rid of voiding. How-
ever, the RSS profile and RTS profile do not have a big
difference in voiding level. For example, if the RSS and
RTS with the same peak temperature and the same time
above liquidus (TAL) are used to assemble the BGAs,
using the same solder paste, the voiding level will be very
similar. The mean value of %voiding in solder joints is
(<1%).
7.7.1.5 Reflow Atmosphere (Nitrogen or Oxygen)
Atmospheres with lower oxygen content produce less
amount of oxidation of the solder ball and PCB surfaces
during the reflow soldering process. Less oxides will result
in less process voids since reduction of oxides creates
water vapor which is one source of macrovoids and lack of
an oxide crust on the other surface of the balls when mol-
ten; this oxide crust can prevent voids from escaping from
the molten ball during reflow soldering.
7.7.1.6 Multiple Reflow Cycles Sometimes, the same
BGAs will go through multiple reflow cycles due to being
located on double sided boards or due to rework. Voiding
level of BGA solder joints after multiple reflow cycles is
January 2013 IPC-7095C
107
Copyright IPC-Association Connecting Electronics Industries
Provided by IHS under license with IPC
Not for Resale
No reproduction or networking permitted without license from IHS
--`,,```,,,,````-`-`,,`,,`,`,,`---