Microsoft Word - numero 15 articolo 7 P. F 64 C pr P.F Poly Z. Ins AB ele dif thi fail dra the sup Th at the KE IN imp eva et a B1 PC Dr the imp Th circ lev T F. Fuchs et alii, F yclic be rinted c F.Fuchs lymer Compete Major stitute of Polym BSTRACT. T ectronic prod fferent tests is test, the b lure driver is awbacks, an e results of pporting the he advantage lower comp e local stress EYWORDS. P NTRODUCTIO he perf increasin product pact loads ca aluate the dro al. reviewed t 11 was defin CBs are repeat rawbacks of th e poor reprod pact perform hus, in this w cuit board, du vel cyclic bend T Frattura ed Integ end test circuit b ence Center L mer Product E The reliabilit ducts indust were develo boards are d s a flexural o alternative both tests w e suitability o es of the alte puting time. ses. Printed Circu ON formance of ng market fo ts are especial an result in se op performan the developm ed under the ted until failu he standardiz ducibility, the mance. work an altern ue to input a d test (BLCB grità Strutturale, ts for the boards u Leoben GmbH, Engineering, Joh ty of printe try. In orde oped. The cu dropped und oscillation o test method were compa of the board ernative test In future an uit Board; Re printed circu or handheld e lly prone to b erious damag nce and to dev ment of the m e Joint Electro ure is detected zed BLDT ar e main criticis native test to acceleration c T), applying 15 (2011) 64-7 e reliab under dy H, Roseggerstras hannes Kepler ed circuit bo er to predict urrent indus der defined of the board d was evalua ared. A very d level cyclic method wer nalysis, test eliability; Dr uit boards (P electronic pr be dropped, i ge of the PCB velop reliable methods in the onic Device E d. The rather re discussed in sms are the s determine th created from a repeated an 73; DOI: 10.3221 ility eva ynamic sse 12, 8700 r University, A oards under t the perform try-wide sta conditions due to the i ated in this s y good corr bend test fo re shorter te simulations rop Test; Cy PCBs) under roducts. Beca i.e. exposed t B interconnec e designs, diff eir work [1]. A Engineering r complex tes n literature e. slow test thro he drop perfo dropping, is nd well define /IGF-ESIS.15.07 aluation loads Leoben, Aust Altenbergerstra r dynamic l mance of th ndard testin until a failu impact even study. A boa relation betw or the determ esting times, will be used yclic Bend T r dynamic lo ause of their to an impact ctions and th ferent board A standardize Council (JED st set-up and .g. by Wong oughput and ormance was the primary d ed flexure, wa n of tria asse 69, 4040 loads is a k he boards in ng method is ure in the bo nt. As this te ard level cyc ween the m mination of better adap d to generat est; Fatigue. oads is of sig size and app load during t hus in the ma level test met ed board leve DEC) [2]. In load applicat et. al [1, 3] an the lack of r s performed a driver for PC as used to rep 0 Linz, Austri key issue in n their appli s a board lev oard is dete st method h clic bend tes ethods coul the drop tes tability and te Wöhler cu gnificant inte plication, han their useful se alfunction of thods were in el drop test (B this test defin tion affect the nd Luan et. a relevance to t and evaluated CB failure. Th produce the l ria n the handh ication lifeti vel drop test cted. The m has a numbe st was used ld be observ st performan test simulati urves related erest due to ndheld electro ervice life. Th f the product. ntroduced. W BLDT) JESD ned drops of e test robustn al [4]. Apart f the product d d. Flexing of herefore, a bo loading situat held ime, t. In main er of and ved, nce. ions d to the onic hese . To Wong D22- f the ness. from drop f the oard tion. http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it Th reli In the Th und def MA dif it c we ma abb Pre reg the thic Th def On to t BO P B he simplified iable results. this study, co e suitability of he test metho der different flection ampli ATERIALS AN CBs with The aim The ana ffered regardin can in some c re considered aterials used i breviations M epreg layer (g garding the gl e layers marke cknesses, is sh he PCB geome fined board g nly two comp the PCB man OARD LEVEL efore the deforma Except f P B and accelerat onventional B f cyclic bend d applied wa t loading con itudes were u AND SPECIM h different lay was to simpl alyzed specim ng the interm cases be adva d in the test in the differe M1, M2, M3 a glass fibre rei lass fibre mat ed with a. A hown in Fig. Table 1: An o Figur etry was chos geometry is sh ponents instea nufactures exp L DROP TEST e developmen ation in a stan for some ada ted load appl BLDT and BL tests as an alt as based on s nditions. The used. MENS yer build-ups lify the comp mens were eig mediate epoxy antageous to plan. Additi ent layers (C and M4 repr inforced) and t used for rein schematic re 1. All copper Design Nr. 1 2 3 4 5 6 overview of the re 1: A schemat sen following hown in Fig. ad of 15 were perience, wer T ANALYSIS nt of an altern ndard BLDT. apted parame P. F. Fuchs et lication shou LCBT were p ternative test similar ideas t e specimens s, expected to parison of the ght layer PC y layers which use neat epo ionally, differ ore, IL1, IL2 resent differen d R stands fo nforcement. T epresentation r layers had a Core M1 a M1 a M1 a M2 M2 M1 b e chosen mater tic representati g the JEDEC . 2. The dark e used to sim re chosen. Th native test me The drawn c eters, the anal alii, Frattura ed uld result in a performed for method. to the high sp were loaded o differ signifi testing meth CBs and six d h are glass-fib oxy resin. Thu rent epoxy r 2 and IL3) o nt manufactu or an unreinf The layers m of the layer thickness of IL1 Mater M1 PP 1a M1 PP 1a M1 PP 1a M2 PP1 M2 PP1 M1 PP1 b rials for the sp ion of the laye standard, but ker squares at mplify the test he component ethod, it was conclusions sh lyzed BLDT IL3 (40 m) IL2 (40 m) IL1 (60 m) Core (200 m) IL1(60 m) IL2 (40 m) IL3 (40 m) d Integrità Struttu a more robus r a set of PCB peed cyclic b d under thre icantly regard hods by cover different laye bre reinforced us, variations esin types w of the tested ures and epox forced and ne arked with b assignments f 18 m. IL2 rial M1 PP 2a M1 PP 2a M1 PP 2a M2 PP2 M2 PP2 M1 PP2 b ecified layers i er assignments t was adapted t the board ce t. The two m ts were daisy necessary to hould help de was perform urale, 15 (2011) st test metho Bs. The resul bend test (HS ee-point bend ding their BLD ring a wide ra er build-ups w d in most case in the reinfo were analyzed PCB design xy resin type eat epoxy res are the low d in the build- IL3 M3 R M1 PP 2a M4 R M3 R M2 PP2 M3 R n the six PCB in the PCB bu d according to entre represe ost failure-pr chained acco analyze the a esigning a rea med according ) 64-73; DOI: 10 od, providing lts were comp SCBT) [5, 6], ding and sig DT performa ange of BLDT were tested. es. Only for t orcement of t . A detailed s is provided es respectively sin layer. PP1 dielectric con -up, including designs tested uild-up. o customer re nt the moun rone compon ording to the acting loads an asonable test. g to the JEDE 0.3221/IGF-ESIS. faster and m pared to eval but was real gnificantly hig ance, were tes T performanc These build- the surface la the surface la summary of d in Tab. 1. y, PP indicat 1 and PP2 di nstant version g the epoxy l d. equirements. ted compone nents, with reg standards. nd resulting P EC standard. 15.07 65 more luate lized gher sted. ces. -ups ayers ayers f the The tes a iffer ns of layer The ents. gard PCB . An http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it P. F 66 alte we a p Dr res dis For set tab In set Th def Fur Th Fro the sin BO F F. Fuchs et alii, F ernative boar re used. The pulse duration rops were rep istance great continuities a r the BLDT a -up is shown ble on the stri Figure 2: Fig. 4a), start in a dampe homas Kerstin flection could rthermore, st he observed lo om the analys e bending-os usoidal cyclic OARD LEVEL or the lo Prairie, U used to aF Frattura ed Integ rd geometry, drop height a n of 1 ms. peated until fir ter than 1000 a special even analysis a hig n. The board i ke surface wa A schematic p ting from the d bending-os ng, Institute d be generated train gauges w ongitudinal st sis results it c cillation. Kn c bending load L CYCLIC BE oad generatio US) was used apply a sinuso grità Strutturale, described in and strike sur rst failure wa 0 ohms lastin nt detector (25 gh-speed cam is fixed with as filmed with presentation of Figure 3: Th e top, three co scillation, wh for Didactic d (Fig. 4b). T were placed at train values at could be con nowing this, d is applied. END TEST DE on an electro d. This machi oidal displace 12 12 15 (2011) 64-7 Materials and rface were ad s detected. Fa ng for 1 mic 56STD, Analy era (Fastcam four screws a h the high-spe f the used PCB he test set-up o onsecutive hi hich can be r cs and Physi The initial osci t the PCB sur t the first defl ncluded that t it seemed re EVELOPMEN o-dynamic te ine has the a ement amplitu 100 12 mm mm 100 12 mm mm 73; DOI: 10.3221 d specimens cha djusted to ach ailure was def crosecond or ysis Tech, Wa SA1.1, Photr at its corners eed camera. B geometry. Th of the JESD22- igh-speed cam recorded by ics, Universit illation freque rface centre t lection were 2 the main defo easonable to NT sting machin advantage of ude. 0 mm m 0 mm m /IGF-ESIS.15.07 apter, and a d ieve a specifi fined as an el r longer. To akefield, US) ron Inc., San on the drop he darker squar -B111 board le mera images s a camera. U ty of Essen, ency was dete to measure th 2 x 10-3. ormation mec expect corr ne (Bose Elec a very precis 44 different boa ed G level of lectrical disco measure the was used. Diego, US) w table. The te res represent th evel drop test. show the imp sing image a DE), an osc ermined as 28 he maximum chanism and relating resul ctroForce 34 e force and d 48 mm48 mm ard fixation (l f 1500 g peak ontinuity in th e very short was used. In est, i.e. the im he mounted co pact of the PC analysis softw cillogram of 80 Hz. occurring lon consequently ts from a B 450, Bose Co displacement longer stando k acceleration he daisy chain lasting elect Fig. 3 the BL mpact of the d omponents. CB. The boar ware (Viana 3 the PCB ce ngitudinal stra y failure drive BLCBT, wher orporation, E control and offs) and ns of rical LDT drop rd is 3.64, entre ains. er is re a Eden was http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it For (Ø mm sch It n onl sys wa Th BL 4m we def Du BL Hz A p EX set T r the board fi 5 mm) which m), clamping hematic repre Fig Figur needs to be c ly be applied stem. Robustn s accepted th he test param LDT was abo mm with a me re chosen. Th flection of -1m ue to machine LDT (280 Hz) z. picture of the XPERIMENT he BLD statistics used. Th -up another f T fixation a thre h have a dist the board at esentation of t gure 4: The res consecutive re 5: A schema onsidered tha in one direc ness and repr hat in contrast meters were d out 5 mm and ean deflection he maximum mm was defin e limits, it wa ). To remain e final test set TAL METHOD DT was perfo s. For the BL hree boards f five test repet ee-point bend ance of 90 m t half length. the fixation is (a) sults of the JE high-speed cam the impact. I tic representati at the board i ction. A set-u roducibility o t to the actual derived from d the maximu n of 3mm (BL m fibre strains ned to ensure as not possib at the desired t-up is shown D COMPARIS rmed with ni LCBT, as ind for all design titions were p P. F. Fuchs et ding fixture w mm to each ot The clampin s shown in Fi SD22-B111 bo mera pictures In b) the accor ion of the thre is not clampe up with clamp f the test wou l BLDT altern the BLDT um fibre stra LCBT Set-Up s were 1.7 x 1 e that no lift-o ble to perform d optimum m n in Fig. 6. Th SON ine samples o dicated in Boa s and both se planned to app alii, Frattura ed was constructe ther. The load ng force was a ig. 5. oard level drop show the deve rding amplitud ee-point bendin ed at the posit ps at the sup uld have been nating loads c analysis. The ain was about p 1) and 3mm 10-3 and 2.2 x off of the boa m the tests at machine displa he event detec of every PCB ard level cyclic et-ups were t ply statistics a d Integrità Struttu ed, where the d is applied b applied over p test high-spee elopment of th de oscilloagram ng fixture used tion of the su port rollers w n affected, co could not be e maximum t 2 x 10-3. Fo m with a mean x 10-3, that is ard from the t the oscillati acement cont ctor was used B build-up to bend test develo tested to deci according to 0 0 -3 -2 -1 0 1 2 3 a m p lit u d e s , m m urale, 15 (2011) board is plac by a fin, whic spring plung (b) ed camera anal e PCB deflecti m is presented. d in the board l upport rollers would have r ounteracting p applied. peak-to-peak or the BLCBT n deflection o in the range support rolle on frequency trol precision d in the same w obtain repre lopment chapte ide which per the BLDT ev .01 0.02 time ) 64-73; DOI: 10 ced centred o ch consists of gers to have a ) lysis. In a) cho ion right after level cyclic ben s, implicating esulted in an previous effo k oscillation a T peak-to-pe of 2.5mm (B of the BLDT ers could occu y of the impa n the frequenc way as for th esentative resu er, two differ rforms better valuation. 0.03 0.0 t, s 0.3221/IGF-ESIS. on support ro f two rollers ( a defined load osen nd test. that the load n overconstrai orts. Therefor amplitude in ak amplitude BLCBT Set-U T. The minim ur in the tests acted PCB in cy was kept a e BLDT. ults and to ap rent set-ups w r. For the cho 04 0.05 15.07 67 ollers (Ø 5 d. A d can ined re, it the es of Up 2) mum s. n the at 25 pply were osen http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it P. F 68 For par To eac me Fur ana mic Fin (e.g eva Th to- F. Fuchs et alii, F r the BLDT rameter. Anal compare the ch other and ethod compar rthermore, to alysis was pe croscope. It w nally, as it is a g. [8, 9]), We aluation. herefore, a fict failure in the Frattura ed Integ T, the numbe logous, for th e BLDT with a Pearson’s rison is shown Fig o check the c erformed. M was tested if t advantageous eibull-paramet titious time-to BLCBT. A c grità Strutturale, er of drops u he BLCBT th Figure 6 h the BLCBT r factor [7] w n in Fig. 7. gure 7: Schema correlation of Microsections the same failu to use the W ters were det o-failure had conversion fa 15 (2011) 64-7 until the first e number of 6: The test set- the method was calculate atic representat f the failures of the expe ure modes oc Weibull distribu termined to a to be calcula actor was intro 73; DOI: 10.3221 t failure, det cycles until fa up of the boar average resul ed to rate the tion of the com types introdu ected failure cur despite th ution in reliab allow for a co ated for all BL oduced by div /IGF-ESIS.15.07 tected by the ailure was me rd level cyclic b lts for the dif e linear corre mparison of BL uced by the d locations w he different lo bility and life omparison o LDT results to viding the av e event detec easured. bend test. fferent PCB b elation. A sch LDT and BLC different testi were prepared oading condit data analysis f the method o have a corr verage time-to ctor, was tak build-ups wer hematic repre CBT. ing methods, d and analyz tions. , similar to th ds on the bas relating param o-failure of all ken as evalua re plotted aga esentation of a failure pat zed with a l he BLDT anal sis of a statis meter to the ti l build-ups in ation ainst f the ttern light lysis tical ime- n the http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it BL fict RE rein bui wa effe pro ver Fig BL Set app per adv Fiv BL T LCBT through titious times-t ESULTS AND he resul average perform nforcement in ilt with glass s used for the fect, regarding onounced, ap rsion of desig gure 8: Results LCBT with tw t-Up 1 and B plication of th rcentage, the vantage regar ve more spec LDT results w T h the averag to-failure bas D DISCUSSIO lts of the perf of nine m mances for the n the surface fibre reinfor e IL3 layers, g the reinfor pparently due gn 1, resulting s of the BLDT Figure 9: B wo different s BLCBT Set-U he lower defle differences ding the mea cimens were were compare ge number of sed on the nu ON formed BLDT measurements e six tested bu layers. While rced IL3 layer resulting in a rcement of th e to the differ g in a BLDT p T performed on BLCBT results set-ups were Up 2 are pres ection amplit between the asurement sca tested with s ed with the B P. F. Fuchs et f drops of al mber of drop T are present respectively uild-ups. The e design 1, ha rs, performed a worse BLDT he IL3 layers rent material performance n six PCB desig showing the am the standar performed o sented in Fig tude, resulted tested design atter, the faste set-up 1 to b BLCBT result alii, Frattura ed ll PCB build ps until failure ted in Fig. 8 i y. The resul e large gap be aving unreinf d worst of all T performanc s, could be o l manufacture in the range o gns. Additiona mplitude depe rd deviations a on three speci g. 9 in terms in significant ns were very er set-up 1 wa be able to per s of set-up 1 d Integrità Struttu -ups in the B e of the BLD in terms of d lts show the etween design forced outer l l designs. In ce compared observed as es. Design 6 of design 3. al to the averag endence. Addit re indicated. imens for eve of cycles unt t longer times y similar for as chosen for rform a statis . The numbe urale, 15 (2011) BLDT. This DT. drops until fai e desired w ns 1 and 2 res layers (IL3), p design 3 a di to design 1. for designs 1 was a low d ge values the st tional to the av ery design. T til failure. As s until failure the set-ups. r further tests stic evaluatio er of cycles un ) 64-73; DOI: 10 factor was u ilure. The pill wide range o sults from the performed ve ifferent neat For designs 4 1 and 2. The dielectric cons tandard deviati verage values The average re s expected, se . Nevertheles As set-up 2 . on of the dat ntil failure in 0.3221/IGF-ESIS. used to calcu lars represent of board BL e influence of ery well, desig epoxy resin 4 and 5 the s e effect was stant epoxy r ions are indicat esults of BLC et-up 2, with ss, if expresse did not have ta. In Fig. 10 the BLCBT 15.07 69 ulate t the LDT f the gn 2, type ame less resin ted. CBT h the ed in e an the was http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it P. F 70 plo Pea fro loa can Mo spe fail F. Fuchs et alii, F otted over th arson’s r fact om the linear ading conditio n be expected oreover, to ve ecimens. The lure mode. In Frattura ed Integ e number of tor, determin correlation, w on dependen d to be reflect Figure Figure Figure erify the acco e analysis was n Fig. 11 an grità Strutturale, f drops until ned as 0.87, which can be t material be ted in the BLC 10: Compariso Additional to (a) 11: Compariso design (a) 12: Compariso design ordance betw s focused on d Fig. 12 the 15 (2011) 64-7 failure in the supported th e attributed t haviour on th CBT. on of the BLDT o the average v on of a) the fai n 1 in the BLCB on of a) the fai n 2 in the BLCB ween BLCBT n designs 1 an e light micro 73; DOI: 10.3221 e BLDT. Th he method hy to the existing he other han T and BLCBT values the stand ilure of design BT, analyzed w ilure of design BT, analyzed w T and BLDT nd 2, expect oscopy image /IGF-ESIS.15.07 he plot is rev ypothesis. Ho g data scatter nd, only signi T set-up 1 resul dard deviations n 1 in the BLDT with light micro n 2 in the BLDT with light micro a failure patt ed to exhibit es of identifie ealing a clear owever, beca r on the one ficant differe lts for six PCB s are indicated. (b) T and b) the fa oscopy. (b) T and b) the fa oscopy. tern analysis t the most p ed failure loc r method cor ause of existi hand and on ences in the B B designs. . ) ailure of ) ailure of was perform ronounced d cations are sh rrelation and ing discrepan n the unregar BLDT behav med on the te differences in hown. The c d the ncies rded viour ested n the cross http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it sec the Th the obs (Fig bot and des A s Th to per all SU (BL PC Th the fail T ctional areas o e image) with he failures cor e BLDT (Fig. served. The s g. 12 a) is co th test metho d not in the sign 1 and 2, w statistical com aby  he parameters failure of the rformance tre examined des PCB Design Weibull Para BLDT BLCBT Table 2: T UMMARY AN he JESD of PCB reprodu LCBT) was e CBs. he two metho erefore the BL lure analysis T of the two-sid h the mounte rrespond to ty 11a) is comp solder ball wa ompared to th ods, but differ solder ball. T which further mparison of th baxb ebx 1 a, the scale p e BLDT are end in the BL signs. In Fig. n ameter a 240 200 Two-paramete Figure designs D CONCLUS D22-B111 bo Bs under imp ucibility and s evaluated. In ods were com LCBT may b showing that ded cut spher d componen ypical failure pared to the f as cracked clo he failure of rent from the Thus, it could r supports th he BLDT and parameter, an fictive values LDT Weibull 13 the differ 1 a b 0.4 2.4 0.4 3.7 er Weibull distr 13: The graphi 1 and 2. The r SION ard level drop pact loads. A low test thro this method mpared and be used to es t the design 0.0 0.1 0.2 0.5 0.7 0.9 0.99 P ro b a b ili ty P. F. Fuchs et re represent th nt (at the top patterns men failure of desi ose to the mo design 2 in t e design 1 loc d be shown t e correlation d BLCBT res nd b, the shap s and were ca analysis is m ence between 2 a b 27.4 2.7 33.9 2.1 ributions for al ical representat esults of both p test (BLDT As the rathe oughputs, an a d, instead of evaluated fo timate the BL dependent fa 10 1 05 1 25 5 75 95 95 BLCBT BLDT alii, Frattura ed he solder ball of the image ntioned in lite ign 1 in the B ounted comp the BLCBT ( cation. Failur that both me of the metho ults was base pe parameter, alculated as d matched to the n the Weibull 3 a b 147.4 2 159.2 2 ll tested PCB d tion of the two methods, the B T) is the curre r complex t alternative te repeated dro r six differen LDT perform ailure modes time to failur PCB 1 T d Integrità Struttu l connecting e). The detec erature e.g. [4 BLCBT (Fig. 1 ponents. In F (Fig. 12b). Ag re occurred in thods led to ods. ed on the Wei , are given fo described in E e performanc l distribution 4 b a 2.4 104.8 2.4 165.6 designs are com o-parameter W BLDT and the ent state of th test set-up im st method wa ops, a contro nt PCB desig mance. This c were indepe 10 2 re t, s PCB urale, 15 (2011) the printed c cted failures a , 9-12]. In Fig 11 b). A matc Fig. 12 the fail gain, the failu n the copper i the same fai ibull distribut or all analyzed Experimental m ce trend in the for designs 1 b a 3.6 86.6 2.3 90.0 mpared for the Weibull distribu e BLCBT are p he art to evalu mplicates som as analyzed. A lled cyclic di gns. The res conclusion w endent of the B 2 ) 64-73; DOI: 10 circuit board ( are indicated g. 11 the failu ching failure l lure of design ure location w interconnecti ilure modes f tion [13]. d designs in T method compari e BLCBT We and 2 is show 5 b 6 1.5 1 0 1.9 1 e BLDT and th utions for presented. uate the reliab me drawback A board leve isplacement i ults correlate was supported e test method 0.3221/IGF-ESIS. (at the bottom by black arro ure of design location could n 2 in the BL was the same ions in the bo for the exami (1) Tab. 2. The ti rison chapter. eibull analysis wn graphicall 6 a b 143.3 2.8 197.3 2.3 he BLCBT. bility and lifet ks such as p l cyclic bend is applied on ed very well d by a perform d. Additionall 15.07 71 m of ows. 1 in d be LDT e for oard ined imes The s for ly. time poor test the and med ly, a http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it P. F 72 stat goo Fin     In Th sim Th fail pre AC and the RE [1] [2] [3] [4] [5] [6] [7] [8] [9] [10 [11 [12 [13 T F. Fuchs et alii, F tistical evalua od approach nally, the follo The BLCBT times depen BLDT. In differences i The BLCBT by default. I testing spac impact ener The BLCBT displacemen The BLCBT Simulations challenging acceleration future analy herefore, BLC mulations are p hus, having th lure stress ov ediction of th CKNOWLEDG he resea Austria) Technol d the Austria e State Gover EFERENCES E. H. Wong JEDEC Sta (2003). E. H. Wong 139. J. Luan, T. Y S. K. W. Sh technology E. H. Wong Leoni, Y. –S J. L. Rodger X. Qu, Z. C G. Heaslip, 0] D. Y. R. Ch 1] P. L. Tu, Y. 2] W. H. Zhu Component 3] O. Beucher T Frattura ed Integ ation based o of the two m owing advant T testing time nds on the bo our case the is an approxim T can be perf In contrast th e. Furthermo rgy, an adjustm T can be perf nt and frequen T can be si of the BLD [14, 16]. The n signal is app ysis it is plan CBT will be p planned. he local failure ver the numb he PCB fatigu GEMENTS arch work o within the logy with con Technologie rnments of St S g, S. K. W. Se andard JESD2 g, S. K. W. Se Y. Tee, E. Pe hea, E. H. W conference, ( g, S. K. W. Se S. Lai, C. –L. rs, W. A. Nic Chen, B. Qi, T J. M. Punch, hong, H. J. To . C. Chan, K. u, J. H. L. Pa ts and Techn r, Wahrschein grità Strutturale, on a two-para methods. ages of the B e is significan oard perform e BLCBT w mate break o formed consi he BLDT und ore, in the BL ment to prede formed on b ncy. imulated at DT are difficu e simplest wa plied directly t nned to gene performed at e stresses in d er of cycles. T e behaviour. f this paper framework ntributions by e & Systemtec tyria and Upp eah, V. P. W. 22-B111, Boa eah, W. D. va ek, C. T. Lim, Wong, Y.-W. (2006) 1003– eah, C. S. Selv YEH, J. Ele cewander, The T. Lee, J. Wan , Electronic a oh, B. K. Lim C. Hung, J. K ang, X. R. Zh nology Confer nlichkeitsrech 15 (2011) 64-7 ameter Weibu BLCBT over t ntly shorter th mances. The b as about two f 5 seconds b idering differ der temperatu LDT only the efined amplit asically every lower compu ult, as the lo ay to simulate to the board, erate Wöhler different def dependence o This resulting was perform of the Kplu y the Institut chnik Aktieng per Austria. Shim, Micro ard Level Dro an Driel, J. F. , Z. Zhong, M Mai, R . Raj 8 lvanayagam, R ectron. Mater e American S ng, Microelec and Photonic m, P. T. H. Lo K. L. Lai, Mic hang, E. Poh rence, (2008) nung und Sta 73; DOI: 10.3221 ull distributio the BLDT co han the BLD better a boar o times faste between each ent influence ure influence i e initial ampli tudes is difficu y dynamic tes uting times oad applicatio e the BLDT i which is still r curves [22] flection levels of the PCB be g Wöhler cur med at the P us-program o te of Material gesellschaft. T electron Relia op Test Meth . J. M. Caers, Microelectron oo, C. T. Lim R. Rajoo, W. r., 38 (2009) 8 Statistician, 42 ctron Reliab., Packaging, 3 ow, In: Electr croelectron R h, Y. F. Sun, 1667. atistik mit Ma /IGF-ESIS.15.07 on was perfor uld be identif DT testing tim d performs, er than the drop in the B e parameters is costly, as th itude can be c ult. The frequ sting machine than the BL on and the d is the ‘Input- computation ] to evaluate s. To determi ending amplit rves should h Polymer Com of the Austr l Science and The PCCL is ab., 48 (2008) hod of Comp N. Owens, Y n Reliab., 47 (2 m, In: Procee D. Driel, J. F 884. 2(1) (1988) 59 47 (2007) 21 (2003) 125. onics Packagi Reliab., 41 (20 A. Y. S. Sun atlab. Springer rmed. The st fied: me (The facto the faster is BLDT. The BLDT.) (e.g. tempera he temperatur controlled. A uency in BLD e which is ca LDT using t definition of -G method’ [ nally expensiv e the fatigue ine the acting tudes, it woul help to impro mpetence Cen rian Ministry d Testing of P s funded by t ) 1747. ponents for H Y. –S. Lai, Mi 2007) 450. edings of 56t F. J. M. Caers 9. 97. ging Technolo 001) 287. n, C. K. Wan r, Berlin Heid tatistical resul r between th the BLCBT reason for ature, frequen re has to be a As this is done DT can hardly apable of app the Finite E the boundar [13, 17, 18, 2 ve. behaviour o g local stresse ld be possible ove the under nter Leoben y of Traffic, Plastics, Univ the Austrian G Handheld Elec icroelectron R th electronic s, X. J. Zhao, ogy Conferen ng, H. B. Ta delberg, (2005 lts supported e method tes compared to the testing t ncy or amplitu applied on a l e by adapting y be varied at plying the des Element Meth ry conditions 0, 21]), where of PCB desi es, finite elem e to plot the l rstanding and GmbH (PC Innovation versity of Leo Government ctronic Produ Reliab., 49 (20 components N. Owens, L ce, (2005) 26 an, In: Electro 5). d the sting o the time ude) large g the all. sired hod. s are e an igns. ment local d the CCL, and oben and ucts, 009) and L. C. 2. onic http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it [14 [15 [16 [17 [18 [19 [20 [21 [22 4] M. I. Sakri, 5] Y. Wang, K 6] D. Yu, J. B Reliab, Arti 7] T. Y. Tee, J 8] T. Y. Tee, J 1094. 9] J. E. Luan, Technologi 0] C. –Y. Cho (2008)1149. 1] J. Bai, F. Qi 2] D. Radaj, M S. Saravanan K. H. Low, H. . Kwak, S. Pa icle in Press (2 . Luan, E. Pe J. Luan, E. P T. Y. Tee, E es, 29 (2006) ou, T. –Y. H . in, T. An, In: M. Vormwald, , P. V. Mohan . L. J. Pang, K ark, J. Lee, D 2010). ek, C. T. Lim, Pek, C. T. Li E. Pek, C. T 449. Hung, S. –Y. Internationa , Ermüdungs P. F. Fuchs et nram, Int. J. C K. H. Hoon, F Dynamic resp , Z. Zhong, E im, Z. Zhong T. Lim, Z. Zh Yang, M. –C al Conference festigkeit Gru alii, Frattura ed Comput. Met F. X. Che, Y. ponses of PCB EuroSimE200 g, In: Electro hong, J. Zho C. Yew, W. – e on Electroni undlagen für d Integrità Struttu thods Eng. Sc S. Yong, Mic B under prod 04, (2004) 133 onic Compon ou, EEE Tran –K. Yang, K ics Packaging Ingenieure, S urale, 15 (2011) ci. Mech., 9 (2 croelectron R duct-level free 3. nents and Te nsactions on K. –N. Chiang g Technology Springer, Berl ) 64-73; DOI: 10 2008) 138. Reliab., 46 (20 e drop impac echnology Co n Component g, Microelec y, (2007) 187. lin Heidelberg 0.3221/IGF-ESIS. 006) 558. ct, Microlelect onference, (20 ts and Packag ctron Reliab., g (2007). 15.07 73 tron 004) ging , 48: http://dx.medra.org/10.3221/IGF-ESIS.15.07&auth=true http://www.gruppofrattura.it