IPC-7095C-2013.pdf - 第109页

the X-ray source, utilizing the peripheral portion of the cone of radiation to produce an oblique view with the high- est achievable magniﬁcation. Figures 7-29 and 7-30 illustrate the beneﬁts of oblique viewing at high m…

detecting PCB assembly defects that cause changes in the

solder joint proﬁle. These types of assembly defects

include but are not limited to: solder shorts, solder opens,

insufficient solder joints, missing devices, skewed devices,

and solder voids.

7.3.3.1 Transmission X-Ray Technology In transmis-

sion X-ray, the X-ray source and detector can be ﬁxed or

movable in a variety of motions for variations in magniﬁ-

cation and angle of view. Generally, all features in the ver-

tical ‘‘line of sight’’ are viewed concurrently without dis-

tinguishing depth.

Differences in material thickness or density will result in

different transmitted X-ray attenuation at the detector

resulting in brighter or darker intensities within the image

display. For a single material type, such as eutectic solder,

the attenuation of the X-ray photons received at the detec-

tor is proportional to the material thickness. A gray scale

image is created which can be interpreted to determine

whether or not solder joints are acceptable. Figure 7-26

shows a transmission inspection illustration.

7.3.3.2 Oblique Viewing Inspections with Transmission

X-Ray Technology

There are two basic methods of

oblique viewing of objects with transmission X-ray sys-

tems. One method consists of tilting the sample in order to

get the oblique angle as illustrated in Figure 7-27. Although

this method enables the oblique view, the technique may

not allow for the highest achievable level of magniﬁcation

to be realized.

Another method of oblique viewing utilizes a wide angle

transmission X-ray source as illustrated in Figure 7-28. In

this method, the detector rotates around the center axis of

Figure 7-25 Laminographic Cross-Section Image (3D)

Figure 7-26 Transmission Example

IPC-7095c-7-27

Figure 7-27 Oblique Viewing Board Tilt

Oblique Viewing with Board Tilt

Detector

Board with

FBGA

Tilted View

High magnification is

required, but loss of

magnification occurs

through object tilt.

(Due to longer source

object distance.)

IPC-7095c-7-28

Figure 7-28 Oblique Viewing Detector Tilt

Oblique View without

Compromising Available Magnification

Detector

Board with

FBGA

Oblique View

High magnification is

required and maintained by

rotating the detector

through the cone of

radiation while keeping it

perpendicular to the center

of the x-ray source.

(Due to shorter source to

object distance.)

IPC-7095C January 2013

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

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

the X-ray source, utilizing the peripheral portion of the

cone of radiation to produce an oblique view with the high-

est achievable magniﬁcation.

Figures 7-29 and 7-30 illustrate the beneﬁts of oblique

viewing at high magniﬁcation with actual images of a

FBGA comparing a top down transmission view with an

oblique view.

The images in Figure 7-29 may be slightly different in tex-

ture and darkness for a lead-free application. The tube

intensity and/or power may need to be slightly reduced in

order to avoid over exposure.

7.3.3.3 Cross-Sectional X-Ray Technologies On double

sided boards, some subset of the solder joints will be inac-

cessible to the transmission X-ray technique due to over-

lap, whereas the cross-section technique will have greater

test access. Because the transmission X-ray technique cap-

tures information from the entire solder volume and the

cross-section technique captures speciﬁc slice information,

these techniques have both unique and common capabili-

ties to detect some types of solder defects. For more

detailed information, contact the manufacturers of AXI sys-

tems who can provide information about the capabilities of

their techniques and systems.

Both tomosynthesis, Figure 7-31, and laminography, Figure

7-32, are radiographic techniques that provide image

‘‘slices’’ of the device that can be viewed. With laminogra-

phy, the X-ray source and the X-ray image plane are

moved in a coordinated way with respect to the electronic

device being inspected. A clear image of only one layer or

slice of the device appears; all other layers in the image

plane are blurred out. Tomosynthesis collects several trans-

mission X-ray images from different angles during the

inspection and combines the digital data of those images

mathematically.

This enables virtual slices to be created at the desired plane

for analysis.

7.3.4 Analysis of the X-Ray Image An understanding of

the construction of the particular BGA device and of the

reﬂow process will aid in the interpretation and analysis of

the X-ray image of the ball bonds. The concurrent factors

that should be considered for X-ray image analysis might

include:

• Determine if the balls are collapsible (eutectic) or noncol-

lapsible (noneutectic).

• Deﬁne if noncollapsible balls have been placed in the

corners to retain coplanarity.

• Was reﬂow temperature maintained sufficiently to permit

full alignment and collapse?

• Does BGA package appear to have physically deformed

in some way during reﬂow?

These factors will add further insight into the interpretation

and analysis of the X-ray image.

7.3.4.1 Field of View In determining the criteria for

inspections it is also important to determine how much can

be seen of the BGA at any one time. As the pitch of the

BGA gets smaller, the ball size is also reduced. Several

magniﬁcation levels can be applied to the evaluation. Table

7-5 provides the different pitch and ball size characteristics

applicable to any ﬁeld of view. The magniﬁcation range

varies; however, should be between 30X and 50X. Depend-

ing on the ball size, the ﬁeld of view can be determined by

simply dividing the number 15 by the ball size. Thus, 15

divided by 0.75 (for the 0.75 mm ball) results in being able

to see 20 balls of that size for evaluation. As the ball size

decreases, a greater number of ball images are viewed for

quality evaluations.

The display of the ﬁeld of view from a real-time x-ray sys-

tem depends on the number of pixels available from the

x-ray detector and how this is displayed on the operator

screen. This can vary from 640 x 480 pixels to 1600 x 1200

pixels, or more, in commercially available x-ray systems.

As one of the quality measurements that is used in this

document is voiding, unless a minimum number of pixels

is suggested for the diameter of the solder balls when under

Figure 7-29 Top Down View of FBGA Solder Joints

FBGA

Solder Joints

Top Down

View

Figure 7-30 Oblique View of FBGA Solder Joints

Three solder joints are clearly open (No. 2,3,4).

The void (No.1) is located near the component pad.

The solder joint above right (No.6) is in contact to the

board but with insufficient wetting (no meniscus).

FBGA

Solder Joints

Oblique View

January 2013 IPC-7095C

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

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

inspection, then the precision of any subsequent measure-

ments will be severely compromised. For example, assume

a4x4array of 0.75 mm diameter balls (pitch 1.5 mm) and

a7x7array of 0.30 mm diameter balls (pitch 0.5 mm) are

shown fully on screen (similar to that shown in the table).

On an X-ray system with a 640 x 480 pixel detector

(assuming 1:1 presentation on the display), then for the

shortest screen axis, each type of BGA ball will be 68 and

44 pixels in diameter respectively. With an error of linear

measurement of 2 pixels (one at either edge), for example,

each BGA has an error of 2/68 and 2/44, or 3% and 4.5%

respectively. Should these values then be used to calculate

the area of the balls for the crucial void percentage calcu-

lation, then the error in the area will be 9% and 20%

respectively. In contrast, the values for an x-ray system

with a 1600 x 1200 detector would be 171 and 109 pixels

with a linear error of 1.2% and 1.8% and an area error of

1.5% and 3.25% respectively. For Class 3 products, a

maximum of 4% voiding by area is proposed. In the worst

case above (but as devices continue to shrink then the

lower spec X-ray systems will progressively worsen) the

minimum action level through system measurement will

Figure 7-31 Tomosynthesis

IPC-7095c-7-32

Figure 7-32 Scanned Beam X-Ray Laminography

Rotating X-ray Beam

Rotating X-ray

Detector

Result

Focal Plane

Table 7-5 Field of View for Inspection

Nominal Ball Diameter (mm) Pitch (mm)

Magniﬁcation

Recommendation Field of View

0.75 1.50, 1.27

30X - 50X

15/0.75 = 20 Balls

0.60 1.00 15/0.60 = 25 Balls

0.50 1.00, 0.80 15/0.50 = 30 Balls

0.45 1.00, 0.80, 0.75 15/0.45 = 34 Balls

0.40 0.80, 0.75, 0.65 15/0.40 = 38 Balls

0.30 0.80, 0.75, 0.65, 0.50 15/0.30 = 50 Balls

IPC-7095C January 2013

Provided by IHS under license with IPC

Not for Resale

No reproduction or networking permitted without license from IHS

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