Issue 15

P. F. Fuchs et e represent th t (at the top patterns men ailure of desi se to the mo design 2 in t design 1 loc be shown t e correlation BLCBT res

alii, Frattura ed e solder ball of the image tioned in lite gn 1 in the B unted comp he BLCBT ( ation. Failur hat both me of the metho ults was base

Integrità Struttu connecting ). The detec rature e.g. [4 LCBT (Fig. 1 onents. In F Fig. 12b). Ag e occurred in thods led to ds. d on the Wei

rale, 15 (2011) the printed c ted failures a , 9-12]. In Fig 1 b). A matc ig. 12 the fail ain, the failu the copper i the same fai bull distribut

64-73; DOI: 10 ircuit board ( re indicated . 11 the failu hing failure l ure of design re location w nterconnecti lure modes f ion [13].

.3221/IGF-ESIS.

15.07

sec the Th the obs (Fig bot and des A s

tional areas o image) with e failures cor BLDT (Fig. erved. The s . 12 a) is co h test metho not in the ign 1 and 2, w tatistical com e parameters failure of the formance tre examined des PCB Design Weibull Para BLDT BLCBT Table 2 : T

f the two-sid the mounte respond to ty 11a) is comp older ball wa mpared to th ds, but differ solder ball. T hich further parison of th b ax b e x   1 a, the scale p BLDT are nd in the BL igns. In Fig. a 240 200 wo-paramete meter

ed cut spher d componen pical failure ared to the f s cracked clo e failure of ent from the hus, it could supports th e BLDT and

at the bottom by black arro re of design ocation could 2 in the BL as the same ons in the bo or the exami

of ws. 1 in be DT for ard ned

ab y 

(1) ab. 2. The ti son chapter. ibull analysis n graphicall

Th to per all

arameter, an fictive values DT Weibull 13 the differ

d b, the shap and were ca analysis is m ence between

e parameter, lculated as d atched to the the Weibull

are given fo escribed in E performanc distribution

r all analyzed xperimental m e trend in the for designs 1

designs in T ethod compari BLCBT We and 2 is show

mes The for y.

1

2

3

4

5

6

b

a

b

a

b

a .4 104.8 .4 165.6 esigns are com

b

a

b 1.5 1.9

a

b

27.4 33.9

2.7

3.6 2.3

86.6

.4 .4

2.4

147.4 159.2

2

1

43.3 97.3

2.8 2.3

2.1 ibutions for al

3.7 r Weibull distr

2 l tested PCB d

90.0 pared for the

1 BLDT and th

e BLCBT.

0.9 0.99

5

BLCBT BLDT

5

0.7

5

0.5

PCB 1

0.2

5

Probability

PCB

2

0.1

0.0

5

10 1

10 2 e t, s

time to failur

Figure designs

13 : The graphi 1 and 2. The r

cal representat esults of both

ion of the two methods, the B

-parameter W LDT and the

eibull distribu BLCBT are p

tions for resented.

S U

MMARY AN

D CONCLUS

ION

(BL PC Th the fail T

he JESD of PCB reprodu CBT) was e Bs. e two metho refore the BL ure analysis

22-B111 bo s under imp cibility and s valuated. In ds were com CBT may b showing that

ard level drop act loads. A low test thro this method pared and e used to es the design

test (BLDT s the rathe ughputs, an a , instead of evaluated fo timate the BL dependent fa

) is the curre r complex t lternative te repeated dro r six differen DT perform ilure modes

nt state of th est set-up im st method wa ps, a contro t PCB desig ance. This c were indepe

e art to evalu plicates som s analyzed. A lled cyclic di ns. The res onclusion w ndent of the

ate the reliab e drawback board leve splacement i ults correlate as supported test method

ility and lifet s such as p l cyclic bend s applied on d very well by a perform . Additionall

ime oor test the and ed y, a

71