Issue 15

P. F

. Fuchs et alii, F

rattura ed Integ

rità Strutturale,

15 (2011) 64-7

3; DOI: 10.3221

/IGF-ESIS.15.07

For par

the BLDT ameter. Anal

, the numbe ogous, for th

r of drops u e BLCBT th

ntil the first e number of

failure, det cycles until fa

ected by the ilure was me

event detec asured.

tor, was tak

en as evalua

tion

up of the boar

Figure 6

: The test set

d level cyclic b

end test.

To eac me

compare the h other and thod compar

BLDT with a Pearson’s ison is shown

the BLCBT r factor [7] w in Fig. 7.

the method as calculate

average resul d to rate the

ts for the dif linear corre

ferent PCB b lation. A sch

uild-ups wer ematic repre

e plotted aga sentation of

inst the

Fig

ure 7 : Schema

tic representat

ion of the com

parison of BL

DT and BLC

BT.

Fur ana mic Fin (e.g eva Th to-

types introdu cted failure cur despite th tion in reliab llow for a co ted for all BL duced by div

a failure pat ed with a l e BLDT anal is of a statis eter to the ti build-ups in

thermore, to lysis was pe roscope. It w ally, as it is a . [8, 9]), We luation. erefore, a fict failure in the

check the c rformed. M as tested if t dvantageous ibull-paramet itious time-to BLCBT. A c

orrelation of icrosections he same failu to use the W ers were det -failure had onversion fa

the failures of the expe re modes oc eibull distribu ermined to a to be calcula ctor was intro

ced by the d locations w e different lo ility and life mparison o DT results to iding the av

ifferent testi ere prepared ading condit data analysis f the method have a corr erage time-to

ng methods, and analyz ions. , similar to th s on the bas elating param -failure of all

tern ight ysis tical me- the

68