IPC-7095C-2013.pdf - 第112页

this technology is the inability to view interior rows with the same level of quality and clarity . It is sometimes pos- sible to focus on interior solder joints but not at the same level of detail as the exterior rows. …

have to be nearer 3%, and this is before any other errors in

X-ray measurement are taken into account (from grayscale

sensitivity variation, geometric magniﬁcation effects, etc.)

and lowers the action level still further.

7.3.5 Scanning Acoustic Microscopy Scanning acoustic

microscopy (SAM), also called scanning acoustic tomogra-

phy (SAT), is a nondestructive failure analysis tool. It uses

sound waves to scan the internal layers of an assembly. It

is commonly used in the semiconductor packaging ﬁeld to

detect delamination or voids located inside an electronic

assembly. It can locate delamination or voids inside a BGA

package. It can also locate similar anomalies in underﬁll

after the BGA has been attached to the substrate.

The resolution of delamination or void detection depends

on the acoustic frequency used for analysis. Resolution

increases with increasing frequency. A 230 MHz transducer

can bring detection resolution down to around a 25 µm gap.

A single point observation is called a SAM, a line scan is

called a b-SAM, and an area scan is called a c-SAM. The

c-SAM image in Figure 7-33 shows the location of voids

in underﬁll in a ﬂip chip assembly. The sample being

inspected needs to be in water during the SAM analysis.

Voids or delaminations that are open to water ingress can-

not be detected with this method.

7.3.6 BGA Standoff Measurement Feeler gauges pro-

vide a nondestructive method of determining the approxi-

mate ﬁnished standoff of the BGA after reﬂow. Feeler

gauges can be used at each corner after reﬂow and the

combined results can be used to determine an average

standoff measurement. This method is not as accurate as

cross-sectioning, but it is much less expensive and nonde-

structive provided that the operator does not attempt to

force a feeler gauge under the device. Doing so could result

in breaking solder joints. It does require adequate space

around the BGA for the feeler gauges to be inserted. The

standoff height of a BGA can give some indication that the

solder balls reﬂowed completely and uniformly. The stand-

off of a typical PBGA with a 0.75 mm ball is approxi-

mately 0.60 mm prior to reﬂow and drops to 0.45 mm

(including solder paste) after reﬂow. Other BGA packages

have their own stand-off height characteristics based on the

ball size, alloy used, and whether the BGA contains a metal

heat spreader. Since each package has its own stand-off

characteristics, users should develop a part assembly pro-

ﬁle in order to make the use of feeler gauges applicable.

7.3.7 Optical Inspection Endoscopy is an optical inspec-

tion method that permits visual inspection of tiny objects in

a small, conﬁned area. This technology has been adapted

and applied to BGA solder joint inspection. BGA solder

joints can be inspected and analyzed for a variety of criti-

cal factors such as:

• Overall Solder Joint Quality – evidence of proper wetting

• Solder Joint Shape – evidence of proper reﬂow

• Solder Joint Surface Texture – smooth vs. irregular

• Overall Solder Joint Appearance - ﬂux residue, etc.

• Solder Joint Defects - solder shorts, opens, cold solder

• Missing balls

This technology is best suited for inspecting exterior row

BGA solder joints as shown in Figure 7-34. A limitation of

IPC-7095c-7-33

Figure 7-33 Scanning Acoustic Microscopy

Reflected

Pulse

Fluid

Input

Pulse

Sample

- Application

Zone

T - Transducer

- Ultrasound

January 2013 IPC-7095C

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---

this technology is the inability to view interior rows with

the same level of quality and clarity. It is sometimes pos-

sible to focus on interior solder joints but not at the same

level of detail as the exterior rows. It is usually not possible

to see paste on the second or third row of balls. The anal-

ogy would be the inability to see trees in the center of a

forest from the outside.

Lens design is a distinguishing feature of this technology.

The highly advanced lenses are able to focus and redirect

an image 90° using a mirror or prism. A high-resolution

CCD camera and monitor are used to capture and display

the image. Magniﬁcation, depending on working distance,

ranges from 50x to 200x (see Figures 7-35 and 7-36).

Lighting is a critical factor. The image quality will be poor

if the light source does not properly illuminate the solder

joint being inspected. Front lighting permits frontal inspec-

tion of a solder joint while back lighting is useful for

detecting solder shorts and other obstructions. Back light-

ing also displays the solder joint outline which makes it

easy to view the overall shape.

A robust positioning system that provides adequate support

and protection for the lens and CCD camera is essential. It

must eliminate motion due to shock and vibration and it

must be adjustable through the desired range of motion.

Analytical software is also desirable. In addition to display-

ing a real time image of the solder joint it is useful to have

features such as image capture and measurement. Some

systems provide reference photographs of acceptable and

unacceptable solder joints. These images can be compared

simultaneously with the image being evaluated, which

reduces subjectivity during inspection.

7.3.7.1 Protocols for BGA Assembly Reliability and

mechanical strength of SMT joints are very important,

especially in high reliability applications. Voids and other

defects may hurt thermal cycle reliability. However, there

is temperature cycling data that exhibit voids, improving

the test results.

Voids in solder joints are common and unavoidable at low

levels. Voids in BGAs are very difficult to characterize

quantitatively, nondestructively. X-ray is the only practical

solution, but it’s tricky!

Void quality protocols must apply to all assembly situa-

tions. Rework, even when ‘‘required,’’ can make matters

worse. In-process and lot-acceptance protocols must be

developed. Out-sourcing/extended supply-chains pose

additional challenges.

Several early studies document the inevitable occurrence of

voids. Testing included some void characterization of con-

ventional SMT joints, as well as an objective to link abnor-

mal failures to abnormal causes, (e.g., extreme voids).

Figure 7-34 Endoscope Example

Figure 7-35 Lead-Free 1.27 mm Pitch BGA Reﬂowed in

Nitrogen and Washed Between SMT Passes

Figure 7-36 Lead-Free BGA Reﬂowed in Air and Washed

Between SMT Passes

IPC-7095C January 2013

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---

Most SMT joints have voids (under foot, plus inside/toe/

heel ﬁllets). The majority are between 5 to 20 area percent

voids with a small amount at 0% or around 25%. Typical

BGA solder joints average ~10 area percent, with some

averages as low as 5% and others in the 15% range. Within

the BGA the joints had different area percent voids ranging

from 0% to 35%.

Early-fail SMT joints do not have more voids than later-fail

joints. This is based on hundreds of comparisons of sym-

metrical parts: the ‘‘ﬁrst-fail’’ joint does not have a larger

area percent void than the ‘‘last-fail’’ joint. Strong data

shows that variations in normal levels of voids are not

linked to thermal cycle life.

Methods (x-ray and fracture plane visuals) are approxi-

mately equivalent but both are very limited, difficult, or

subjective. One analysis compared void content of several

styles of early-fail BGAs vs. void content of equivalent

later-fail BGAs. No correlation was found. Early and late

failures had the same void content.

Note that all this BGA data involved ‘‘all-ball nets.’’ No

data exists to statistically link ‘‘ﬁrst-fail joint’’ of a BGA to

its life. Therefore there is no numerical correlation of life

vs. void percentage. No data was taken that would be good

enough to establish trends within the 5-10 accuracy that

might be useful.

At this time, there is no new data that would establish a

numerical link (voids vs. life) for any/all of the component

types, under any/all conditions. Some information suggests

that normal voids variations are:

• For in-process control, use >35% area percent as the

threshold with a >50% void diameter threshold

• For reject/rework, use >45% area percent as the criteria

with a >65% void diameter threshold

7.3.7.2 Cracked Peripheral Interconnect Determina-

tions

A torque-limited screwdriver can be used to identify

cracked peripheral interconnects (solder joints). Applying a

small force between the component and substrate will sepa-

rate fractured surfaces, as shown in Figure 7-37. This tech-

nique can identify open connections in a nondestructive

manner, and can determine if an open is due to a lifted

land, interfacial structure or bulk solder structure. This

technique does not work on some substrates, typically

those of low thickness laminates which produce a more

ﬂexible component substrate (interposer).

7.3.8 Destructive Analysis Methods If nondestructive

methods fail to identify the cause of an anomaly, it may be

necessary to use destructive methods to isolate the problem

area. Such techniques will render the analyzed assembly

unusable. Once the cause of the failure has been identiﬁed

the information can be used to implement corrective

actions to eliminate the problem.

7.3.8.1 Cross-Sectioning Cross-sectioning is a destruc-

tive analysis method which looks at a section of the com-

ponents, substrate and solder joints after cutting it apart.

The ﬁrst step in cross-sectioning is to identify or make a

best guess regarding the area that needs to be examined. If

more than one area is suspect then it needs to be deter-

mined whether those areas can be accessed sequentially on

the same component. If not, then the areas will need to be

prioritized according to the possibility of ﬁnding the prob-

lem or more than one component will need to be analyzed.

Next, if the problem area is a part of a larger assembly, it

may need to be isolated into a small more manageable por-

tion by cutting it out of the larger assembly. Care should be

taken to ensure that the evidence is not altered or destroyed

during the cutting process.

IPC-7095c-7-37

Figure 7-37 Engineering Crack Evaluation Technique

Side View

End View

BGA

Torque Screwdriver set at ~ 0.02 Nm

~ 0.03 mm minimum gap

January 2013 IPC-7095C

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

--`,,```,,,,````-`-`,,`,,`,`,,`---