Академический Документы
Профессиональный Документы
Культура Документы
(a)
(b)
(c)
SE
E2
SD
D2
L
D2
C LL
CP L
ZEmax
Y
E2
AEmax
GEmin
X
ADmax
HECB
FCB
HECB
FCB
GDmin
Coverage: 37%
1.35x1.35 mm Squares
@ 1.65 mm Pitch
Coverage: 68%
1.35x1.35mm squares
@ 1.5 mm Pitch
Coverage: 81%
(b)
(c)
(d)
(a)
(b)
(c)
(d)
(a)
3
2.5
2
1.5
1
0.5
0
PLUGGED
PLUGGED ENCROACH ENCROACH
VIA @ 37% VIA @ 67% VIA @ 37% VIA @ 67
PASTE
PASTE
PASTE
% Paste
COVERAGE Coverage Coverage Coverage
48I/O
48I/O
48I/O
48I/O
PLUGGED
PLUGGED ENCROACH ENCROACH
VIA @ 50% VIA @ 81% VIA @ 50% VIA @ 81%
PASTE
PASTE
PASTE
PASTE
Coverage Coverage Coverage Coverage
68I/O
68I/O
68I/O
68I/O
Tg (o C)
Modulus
Cycles
# of
(kg/mm ) Completed Failures
2
1st
Failure
Die Size: The die size inside the package can have a
significant effect on the board level reliability of the
package. To quantify this affect, a number of board level
tests were conducted on 7 x 7mm-48 lead, 10x10mm-68,
and 12x12mm-100 lead packages. The tests were run using
two test conditions (TC1 and TC2) and two board thickness
(0.8 and 1.6mm). Table 2 provides the summary of package,
board, and temperature cycle condition used for these
evaluations.
Figure 10 shows a plot of first failure point Vs Die to
Package size ratio. The data is normalized for the same test
condition and board thickness using the acceleration factor
multipliers for these variables. These acceleration factors
are presented in a later section of this paper. The plot shown
is for first failure based on sample size of 30, as not enough
failures were observed to calculate the mean life for small
die/package ratio. The plot also shows the best fit to this
data as well as the fitted equation.
It is clear that the board level reliability is dependent on this
ratio, not the actual die and package size. The life can be
very low if die/package size ratio is very high, but increases
non-linearly for lower values of die to package ratio.
Table 2: Test matrix for die size effect evaluation.
Package Size
Package Size
Die Size
Board Thickness
Test Condition
10000
Fatigue Life (1st Failure)
3.
8000
6000
y = 341.16x-3.2274
R2 = 0.9886
4000
2000
Mean Life
EMC1
25
125
2650
1846
29
649
978
EMC2
EMC3
7
8
33
35
120
130
2710
2650
4100
5012
29
22
2166
1219
3150
2384
EMC4
35
150
2800
5012
22
2700
3822
EMC5
EMC6
10
11
42
45
135
135
2400
2400
5657
5012
12
12
3747
3578
5320
4708
EMC7
12
49
130
1900
5012
4218
NA
EMC8
14
43
185
1800
5657
24
3684
5090
0
0.30
0.40
0.50
0.60
0.70
0.80
90.0
F=30 / S=0
7mm-48
50.0
99.0
Weibull
F=23 / S=7
Cumulative % Failed
Cumulative % Failed
99.0
10.0
5.0
1.0
100.0
1000.0
Cycles to Failure
10000.0
90.0
W2 RRX - SRM MED
50.0
F=26 / S=4
1.6mm
Thick
Board
0.8mm
Thick
Board
5.0
1.0
1000.0
10000.0
Cycles to Failure
Cumulative % Failed
F=29 / S=1
10.0
100.0
Weibull
90.0
W2 RRX - SRM MED
50.0
Paddle
Not Soldered
F=28 / S=2
Paddle
Soldered
F=28 / S=2
10.0
5.0
1.0
100.0
1000.0
10000.0
Cycles to Failure
1=9.55, 1=1036.80, =0.98
2=10.84, 2=1289.56, =0.88
Plate-up Bump
on Lead
5
32
0.5
1.6
TC1
3785
2
30
3352
NA
5
32
0.5
1.6
TC2
7370
17
30
2610
5960
5
32
0.5
1.6
TC3
5813
0
30
NA
NA
Plate-up Bump
on Die Paddle
Cumulative % Failed
Weibull
90.0
W2 RRX - SRM MED
50.0
Non Bumped
(Std)
F=5 / S=25
W2 RRX - SRM MED
Bumped
F=7 / S=23
10.0
5.0
1.0
1000.0
10000.0
Cycles to Failure
1=10.08, 1=4447.98, =0.97
2=14.00, 2=2255.67, =0.97
(a)
(b)
(c)
(d)