The thermal behaviour of electronic components during soldering

Transcription

1 Philips ech. Rev. 38, , 1978/19, No. 4/5 135 The hermal behaviour of elecronic componens during soldering R. J. Klein Wassink The hermal behaviour of small componens like resisors and capaciors during sof soldering is unlikely o reveal many new aspecs of a ruly scienific naure. Even so, as can be seen from he aricle below, a heoreical reamen of he subjec can be of considerable pracical value. The large-scale auomaion of soldering in modern mass producion is of course only efficien if i is reliable and if he heaing applied does no noiceably impair he performance of he componens. A rough esimae shows ha he number of componens sof-soldered per year in he enire indusrial world is in he region of Soldering and mass producion Mos soldered joins were formerly made by hand, including hose used in he elecronics indusry. Owing o he enormous growh of' producion needs, his manual work has now almos enirely been superseded by highly mechanized and auomaed mehods. Nowadays hundreds of componens like resisors and capaciors can be soldered simulaneously by a single auomaic operaion, for ëxample o a circui wired on a prined board. Recenly a grea deal of aenion has also been paid o soldering in hybrid circuis made by he hick-film echnology n, This aricle deals only wih soldering on prined boards. Among he principal mehods used dip soldering should be menioned firs. n his mehod he erminal leads projecing from he componens are simulaneously immersed in a saionary solder bah in such a way ha he underside of he board jus ouches he solder. The second mehod is called wave soldering; here he solder is pumped up like a spou from a founain, bringing i ino conac wih he leads on he underside of he prined board (fig. ). Boh mehods are used for sof soldering, for which he emperaure a he poin of conac mus remain above abou 200 oe for a few seconds. To speed up he operaion - and for oher reasons as well - he pars o be soldered are someimes preheaed. This poin will be discussed again a he end of he aricle. The soldering mehod used should be compleely reliable, i.e. all he soldered joins required mus be good ones. The desired reliabiliy can be achieved by only using componens whose characerisics corre- r R. J. Klein Wassink, wih Philips Phille and Mealware Works, is seconded o he Philips Cenre for Technology (CFT), Eindhoven. Q J. Fig. 1. The principle of dip soldering (a) and wave soldering (b). PR prined board wih elecronic componens, such as capaciors and resisors. Each componen has wo wire leads, which have o be soldered o conducing racks on he underside of he board. SB solder bah. n dip soldering all he leads are simulaneously immersed in he solder for a few seconds by a verical movemen of PR. n wave soldering he leads are soldered consecuively. The solder is forced upwards by a pump. The prined board ravels horizonally. spond o he condiions and daa ha are ypical of he soldering mehod. mporan in his respec are he choice of maerials, he solderabiliy of he various meals, he choice of solder and flux, he dimensions of he componens and heir leads, and also he dimensions of he prined boards. There are wo main ypes of prined board: single-sided - wih conducing racks on he underside - and double-sided - wih racks on boh sides. n he double-sided ype he [1) W. Funk, Philips ech. Rev. 35, 144, 1975.

2 136 R. J. KLEN WASSNK Philips ech. Rev. 38, No. 4/5 soldering akes place in meallized holes ha run hrough he full hickness of he board. This is ofen he preferred mehod for professional equipmen. This aricle deals wih he hermal aspec of he reliabiliy of he soldering process. This involves wo requiremens, which a firs sigh seem diamerically opposed. The firs is ha he leads of he componens should be raised o a emperaure high enough for hem o be weed by he solder. The second requiremen is ha he emperaure o which he capacior or resisor rises during soldering should no be so high as o affec is operaing value. Each componen has is ypical soldering range, i.e. a range of values for he solderbah emperaure and for he soldering ime (immersion ime) wihin which boh of hese requiremens can be saisfied. There mus be a good overlap beween he soldering ranges of componens of differen ypes. The simples means of ensuring his is o adap he lengh of he leads. n he nex secion an elecrical nework will be discussed ha can give a fairly good simulaion of he hermal behaviour of a componen during soldering. The mehod of calculaion based on his nework allows he required solder-bah emperaures and soldering imes o be found easily from a graph. Conversely, he mehod can also be used o good advanage for designing new componens ha have o be suiable for a paricular soldering mehod. The general validiy of he model will hen be examined. Finally, he model is applied o wo cases and a number of pracical maers are discussed, such as he use of copper wire or iron wire for he leads and he quesion of preheaing. emperaure is everywhere he same. To calculae he behaviour 'of he poenial as a funcion of ime i is sufficien in he case of fig. 2 o solve a simple firsorder differenial equaion. The model used here is herefore generally classified as a 'firs-order model'. Solving he differenial equaion gives Tw(), he emperaure of he lead a he heigh of he underside of he prined board (afer dipping), and also Tb(), he emperaure of he body, boh as a funcion of ime. These wo emperaures mus be known for an assessmen of he success of he soldering process. The momen a which dipping sars is aken as he ime = O.The soluion found for he emperaure Tb{) is: Tb{) = Tpre + {- exp(-/rc)}x [ TSO- Tamb ] X 1 + {Rs + Rw)/Rb - (T pre - Tamb). The resisance R is given by: l/r = /Rb + liers + Rw), which is he expression for a parallel arrangemen of he hermal resisances Rb and Re + Rs: The significance of he yarious emperaure symbols in eq. (1) can be found in fig. 2. The soluion for he emperaure Tw{) is Rs Tw{) = TSO- {TSO- Tb{)}. (2) Rs + Rw hus appears ha he emperaure of he lead, which is beween he (high) solder-bah emperaure and he body emperaure of he componen, follows from a () The model and he empéraure-ime diagram The elecrical nework ha serves as he simples analogue for calculaing he hermal behaviour of a componen is shown in fig. 2 [21. The componen consiss of a body and a lead, which is insered hrough an opening in a single-sided prined board. (The hermal behaviour in he case where he lead is insered hrough a meallized hole in a double-sided board requires a more complicaed elecrical analogue, which will be referred o briefly in he secion dealing wih he validiy of he model.) Dipping in he solder bah corresponds o closing he baery swich in he model. This has he effec of applying a sep-funcion volage o he nework. The resulan increase in he poenial a he various poins in he nework corresponds o he rise in emperaure a he corresponding poins in he componen. The value of he capaciance in he ne-. work corresponds o he hea capaciy of he body. For simpliciy i is assumed ha no emperaure gradiens occur in he body of he componen, so ha is Rb e l B J ~ff).--= J - ORw l Tomb ~ PR- r---- ~ ur) P!sol ~Rs - SB J-o""'"0- ~ E Sw Fig. 2. An elecrical nework (lef) for simulaing he hermal behaviour of an elecronic componen (righ), consising of a body (B) wih a lead (P) during soldering on a single-sided prined board (PR). SB solder bah. RB inerface resisance beween solder bah and lead. Rw hermal resisance of he lead. Rb inerface resisance beween body (B) and air. C hea capaciy of B. soldering disance ('free' lengh) of he lead. S crosssecional area of he lead. Tw() emperaure of he lead as a funcion of ime. Tb() body emperaure as a funcion of ime. T.O solder-bah emperaure. Tamb ambien emperaure. Closing he swich Sw corresponds o he momen of immersion in he solder bah a = O. f here is preheaing, he body emperaure (and also he lead emperaure) is equal o Tpre a = O.

3 Philips ech. Rev. 38, No. 4/5 SOLDERNG ELECTRONC COMPONENTS 137 T T we T dom T.ini W simple 'poenial division' in fig. 2. This division involves only he hermal resisances Rs and Rw, and is independen of ime. The emperaures calculaed from equaions (1) and (2) are ploed in fig. 3 for a componen whose characerisic daa are lised in Table J. The ime in fig. 3 is expressed in unis of RC, he characerisic ime consan of he componen. When he damage emperaure Tdam of a componen and he weing emperaure Twe of is lead are also known he corresponding ime duraions dam and we can be derived from fig. 3. The soldering ime mus be chosen beween he limis represened by hese quaniies. Fig. 4 shows a emperaure-ime diagram derived from fig. 3 by giving values o he coordinaes such ha he curves for Tb() and Tw() are sraigh lines. Since a dimensionless quaniy (T*) is ploed insead ofhe emperaure, he diagram also gives a conversion line from which he emperaure T corresponding o T* can be read, where T* is equal o (T - Tpre)/(Tsol - Tpre). The linearized diagram only applies of course o componens whose characerisic daa (emperaures and hermal resisances) correspond o hose given in Table. Saring wih he damage emperaure Tdam of 145 oe and he weing emperaure Twe of 183 oe, i is firs necessary o find he corresponding dimensionless emperaures on he verical scale. Wih hese and he wo emperaure lines he limis Tdam and Twe for [2] From H. J. Ver beek, nvs-ne-. 40, 35, T'sol 3 4 -ire Fig, 3. A calculaed emperaure-ime diagram for he soldering of an elecronic componen (fig. 2). The numerical values used are given in Table. The emperaures Tb(!) of he body and Tw() of he lead are boh ploed as a funcion of he ime ha has elapsed since he momen a which he lead is firs dipped ino he solder bah. Tpre empéraure of body and lead a he ime = 0, as a resul of preheaing. Tamb emperaure ofambien air. Tso! emperaure of solder bah. Twe weing emperaure of he lead. Tdam emperaure a which he componen is damaged. n he model he hea capaciy of he lead is negleced so ha a he same ime as immersion begins is emperaure becomes Twini. RC is he characerisic ime consan of he componen (see eq. ()). The limi values we and Jdam for he soldering ime follow from he values of Twe and Tdam applicable o he componen. he dipping duraion are hen found on he dimensionless ime axis. Each value beween hese limis, afer muliplicaion by he ime consan RC, gives an appropriae soldering ime for he componen. The conversion of fig. 3 ino a graph conaining only sraigh lines is a simple maer. From eq. () i can be direcly deduced ha {Tb() - Tpre} is proporional o { - exp( - irc)}. The curves in fig. 3 can herefore be ransformed ino sraigh lines by aking he scale disances along he irc axis from { - exp( - irc)}. n he case, frequenly found in pracice, where Rb» R«+ Rw, he proporionaliy facor is equal o (TSOl - T pre ), so ha T* hen goes from 0 o 1 when goes from O.o 00. f (Rs + Rw) is no negligibly small compared wih Rb, hen T* a T = 00 is somewha less han 1. A similar approach applies for Tw(). The values of Tb max and Twini can be calculaed from he equaions given in Table. The main hermal quaniies in he model Hea capaciy and hermal resisances The hea capaciy C of a componen is equal o Vee, where V is he volume of he body, e is is densiy and c is specific hea capaciy. The produc çc and he hermal-conduciviy coefficien A for various maerials are given in Table J. can be seen ha he produc ec has abou he same value for various maerials, bu he values of A differ considerably. The hermal resisance Rw of he lead can be calculaed from he relaion for he conducion resisance, which saes ha Rw is equal o AS where is he soldering disance (or 'free' lengh) ofhe lead and S he area of is cross-secion (fig. 2). The conducion resis- Table. The values of he hermal quaniies applicable o he componen used by way of example as he basis of fig. 3 and fig. 4. Tbmax maximum emperaure ha he body of he componenwould have afei'an infiniely long ime. Twini emperaure of he lead immediaely afer immersion in he solder. The values of hese emperaures were calculaed from he equaions given in he Table, which were derived from equaions (1) and (2). The oher quaniies are defined in he capion o fig. 2. Temperaure (DC) Thermal resisance (ocjw) Tpre 50 Rw 15 Tso! 250 Rs 15 Tamb 20 Rb 315 Tb rnax 230,._ Tb max 1 + ( + Rs/Rw)RwTamb/RbTso! Tso! 1 + ( + Rs/Rw)Rw/Rb Twini Tdam 145 Twini Tso! 1 + RsTpre/RwTso! + s.»; Twe 183 The euecic emperaure of in-lead

4 138 R. J. KLEN WASSNK Philips ech. Rev. 38, No. 4/5 ance of copper, aluminium and iron leads is calculaed per millimere of lengh, and wih he widely used diameer of 0.6 mm i is equal o 10, 16 and 50 CjW respecively. The hermal resisances Rs and Rb are inerface resisances, R; beween he solder and he immersed par of he lead and Rb beween he body and he ambien air. The values are calculaed from he general formula ljha, where h is he appropriae hea-ransfer coefficien and A he inerface area. There are uncerainies abou he hermal resisance Rs, and hese will be menioned in he secion on he validiy of he model. sol T* we Table l. Produc of he densiy e and be specific hea capaciy c, wih he hermal-conduciviy coefficien À of some meals, alloys and maerials for prined boards. Maerial ron Copper Aluminium Fernico Brass Plasics Paperboard Glass Solder (euecic in-lead) -T(=f/RC) abou À (W/m 0c) abou pre 100 Fig. 4. The emperaure-ime diagram from fig. 3 wih he coordinae axes ransformed o make he curves sraigh lines. H conversion line for ransforming he emperaure T(OC) ino a dimensionless quaniy T*. The dimensionless ime axis is a he op; he uni of ime is he RC consan of he componen. The emperaure curves for he body and for he lead are he line Tb() and he line Tw{); hese lines are found from he calculaed Tbmax and Twini. Taking Tdam = 145 C and Twe = 183 C as given daa for he componen and he solder, he ime limis Tdam and Twe for a properly soldered join can be read from he graph. Wih an inerface area of 10 mms Rs can be assumed o have he order of magniude of 10 CjW, wih h aken as 10 kwjm 20 C. The value given for h is reasonably accurae for wave soldering. For dip soldering a lower value mus be used, abou 6 kw /m2 C; he hermal resisance R; hen has a value beween 15 and 20 C/W. This means ha in dip soldering he quaniy Rs/ Rw is raher higher, and his is refleced in he longer weing ime required. The increase in Rs affecs he emperaure of he body mainly hrough he increase in he ime consan RC. f he free lengh ofhe leads is no oo small, however, he body emperaure afer wave soldering is found o be only a few per cen higher han afer dip soldering (during he same ime). Measuremens and calculaions agree compleelyon his poin. The inerface area for he body is abou 100 mm-. The hea-ransfer coefficien h for he hea ransfer from he body o he air is abou 10 Wjm2 "C, The hermal resisance Rb hus amouns o abou 10 3 CjW, a hundred imes higher han R«, so ha i is ofen reasonable o allow for an infinie hermal resisance beween he body and he air. The damage emperaure n he discussion ofhe model several references have already been made o he damage emperaure Tdam. n fac Tdam is no a fixed quaniy, nor is i fixed for each ype of componen. The damage emperaure can bes be regarded as he emperaure a which cerain characerisics of he componen change by more han a permissible percenage. The applicaion deermines he magniude of his percenage. The percenage change ha acually occurs depends no only on he emperaure bu also on he ime during which he emperaure prevails. We shall no go furher ino his here; lile informaion on he maer is available, for one reason because damage sudies are usually limied o a es a one sandard combinaion of ime and emperaure, e.g. 10 seconds and 260 oe.

5 Philips ech. Rev. 38, No. 4/5 SOLDERNG ELECTRONC COMPONENTS 139 The validiy of he model The model described conains a number of simplificaions ha require some explanaion. is assumed, for insance, ha he hea capaciy of he lead is negligibly small. This is of course admissible when he hea capaciy of he lead is very much lower han ha of he body. n general his is in fac he case, alhough here are excepions, such as small ceramic capaciors. Because he hea capaciy of he lead differs from zero, he iniial emperaure Twini ofhe lead is no produced insananeously on dipping. There is a delay of he order of 0.1 s. This means ha he body emperaure sars o rise raher less seeply han indicaed by eq. (1). A couneracing effec here is ha of he hea capaciy of he solder iself round he end of he lead. The presence of he prined board does no significanly affec he hermal behaviour, alhough in heory he board should increase he hermal ineria. When single-sided boards are dipped he underside wih he conducing racks comes ino conac wih he surface of he solder bah immediaely. Calculaions show ha he emperaure a he underside, which originally had a value of say 20 C, can hen exceed he weing emperaure in a fracion of a second; see fig. 5. This ime is so shor ha here is no significan effec on he hermal behaviour, and herefore he model is enirely accepable in his respec. For double-sided boards, where he solder peneraes over he whole hickness of he board, he model is sill valid, hough here are some reservaions. n his siuaion addiional hea is ransferred by he meallized wall of he hole in he prined board, and he rising solder also brings is own hea wih i. The elecrical nework ha could be used as an analogue conains a larger number of resisors and capaciors han he one in fig. 2. The behaviour of he exended nework has been calculaed from compuer programs, bu he resuls for he hermal reliabiliy were almos he same. is herefore preferable o make he calculaions wih he original model. Since he upper side of he prined board now has o be weed as well, he posiion for calculaing he emperaure of he lead mus be aken a he upper side, and no he underside. This means ha he 'poenial division' mus be modified when eq. (2) is used: he hea resisance Rw in eq. (2) has o be reduced by he correcion erm ij/j..s, where zl is he hickness of he board. The dashed curve in fig. 5 represens he calculaed ime needed o raise he upper side of a board o he weing emperaure, if he presence of he holes is no aken ino accoun. n fac he weing emperaure will be reached much more quickly, because he meallized wall of he holes also conducs hea. The body also heas up more quickly han he model suggess. The final facor limiing he validiy of he model is he variaion in he hermal resisance Rs. n he expression /ha for his resisance neiher A nor hare consan. The coefficien h varies because of he marked variaion of emperaure a he lead/solder inerface; in addiion he hea ransfer is affeced by he layers of oxide and flux iniially presen, and also by he solven ha evaporaes from he flux during soldering. The surface area A is no consan owing o he 'climbing' of he solder level; immediaely afer dipping he meniscus of he solder is curved downwards; a he end, for complee weing, he meniscus is curved upwards (fig. 6). Wih a lead of 0.6 mm diameer he 10s / / , ;'..",..;""", " 001 i_ ' mm --,1 Fig. 5. The solid curve represens he calculaed ime required for he underside of prined boards o reach a emperaure of 183 DC afer conac wih a solder bah a 240 DC, as a funcion of he board hickness,1. The underside of he prined board, which is a ambien emperaure (20 C), comes ino conac wih he solder bah a he ime = O. The dashed curve gives he corresponding resuls for he upper side of he prined board. (The hea-ransfer coefficien is aken as 10 kw 1m 2 -c. The presence of holes in he boards is no aken ino accoun.) -- 06mm PR Fig. 6. Solder 'climbing' up a lead during soldering. mmediaely afer immersion of he lead (P) in he solder bah (SB) he surface meniscus becomes convex (mini). When he solder meniscus becomes concave (mwe) he lead is fully weed. PR prined board. For clariy he diameer of he hole in he prined board is shown disproporionaely large.

6 140 R. J. KLEN WASSNK Philips ech. Rev. 38, No. 4J5 difference in solder level can be as much as 0.9 mm. The effec ofhis is o reduce he free lengh of he lead, wih full weing, by abou mm (maximum), which means ha he body is heaed up raher more rapidly han he model predics. Table ll. Daa for calculaing he hermal behaviour of an elecrolyic capacior (Cl) and of a microcapacior (C2). Come RwJl Rs Rb Tdam Twe Tamb = ponen (WC) (ocjw mm) (ocjw) CCJW) (0C) = Tpre Cl [*) 20 C [*) 20 Model calculaions on capaciors As an example he model in fig. 2 was used o calculae he hermal behaviour of wo very differen capaciors. The daa used are lised in Table ll; he free lengh of he leads was varied from a few mm o 10 mm. This parameer, which has a considerable effec on he hermal behaviour, usually provides he simples means of adaping various componens o he soldering. process used. The resuls of he calculaions are summarized in fig. 7 and fig. 8. Each curve consiss in principle of wo branches, one more or less horizonal and one more or less verical. Of he upper four curves in fig. 8 he verical branch falls compleely or parly in he emperaure region below 200 oe. n he region on he lef of he verical branch he solder-bah emperaure is oo ow, so ha no weing can occur; in he region above he horizonal branch he soldering ime is oo long, causing damage o he componen. The region beween he wo branches has he correc combinaions of solder-bah emperaure and soldering ime, so ha he weing occurs wihou damage o he componen. Since he branches are roughly horizonal and verical, wo pracical rules for soldering may be formulaed: - if he leads are no weed in a soldering process, i is preferable o increase he emperaure of he solder bah; - if he body of a componen is damaged in a soldering process, i is preferable o shoren he soldering ime. n boh cases he oher possible change (longer soldering ime or lower solder-bah emperaure) would have o be relaively much greaer o obain a properly soldered join. A sudy was also made of he exen o which he maerialof he leads affecs he hermal ineria. The maerials compared were iron and copper (fig. 9). The weing ime for iron leads is abou a sixh of ha for copper leads, under oherwise idenical condiions. The emperaure of a body wih an iron lead is significanly lower han ha ofhe same body wih a copper lead, even a he ime corresponding o he weing of a copper lead. should be emphasized here ha he beer soldering performance of iron leads is no he only facor ha deermines he choice of maerial, bu he maer will no be discussed furher here. [*J euecic emperaure of solder (m-lead) r 10s 5 -T so1 Fig. 7. Calculaed hermal-behaviour profiles of he elecrolyic capacior Cl in dip soldering (see also Table 1l). The soldering ime is ploed verically, he solder-bah emperaure Tsol horizonally. Beside he 'horizonal' branch of each profile he free lengh l is given for he lead o which he curve relaes (see fig. 2). Complee weing of he lead wihou damage o he capacior occurs only for combinaions of and Tsolon he righ of he 'verical' branch and below he 'horizonal' branch of he appropriae profile. To he lef of a verical branch he weing is insufficien, and above a horizonal branch he capacior is damaged. f 15s,,, Fig. 8. Calculaed hermal-behaviour profiles for he microcapacior C2 in dip soldering (see also Table ll). A fuller explanaion is given in he capion o fig. 7. The 'verical' branches of some of he profiles lie o he lef of he verical axis (Tsol = 200 C) for his capacior and are herefore no shown. f in addiion o a capacior C2 a capacior Cl also has o be soldered o a paricular board (fig. 7) an accepable soldering ime and solder-bah emperaure can easily be found for boh capaciors by varying he lengh.

7 Philips ech. Rev. 38, No. 4/5 SOLDERNG ELECTRONC COMPONENTS 141 Preheaing We shall now reurn o he quesion of preheaing, which has been he subjec of some debae. is obvious ha preheaing will make he momen of weing earlier. Our calculaions, which closely resemble hose made wih he examples given in he previous secion, indicae ha he ime inerval beween he momen of weing and he momen a which he body reaches he damage emperaure is independen of he preheaing. Thus, he momen a which damage begins is brough forward o he same exen, so ha preheaing does no offer any greaer margin, When he condiions of he soldering process are such ha he lead is immediaely weed on immersion, Temp =O 20~ ~20 Fe Cu =15 Temp 75 Temp Fe Fig. 9. Comparison of he behaviour of iron leads (lef in he phoographs) and copper leads (righ). a) A = 0 boh leads are dipped ino he solder. n boh cases he meniscus is curved downwards (see fig. 6). b) Afer 1 second he iron lead is compleely weed (meniscus curved upwards). The meniscus for he copper lead is sill curved downwards. c) Afer 6 seconds he copper lead is also compleely weed. The hisograms give he corresponding lead emperaures Twand body emperaures Tb in "C, Plead. B body.

8 142 SOLDERNG ELECTRONC COMPONENTS Philips ech. Rev. 38, No. 4/5 he preheaing only makes he damage sar earlier. Under hese condiions he preheaing herefore shorens he available soldering ime. nee Thermal resisance ec/w) Rw = 25 RB = 10 Rb = 1000 R Ri 34' ~ = Rb(RB + Rw) Rb+RB+Rw Temperaure ec) TBol = 240 Twe = 183 Tomb = 20 Hea capaciy C OPEN CHOCE n fig. 10 a number of calculaed (dimensionless) soldering imes are ploed as a funcion of he preheaing emperaure. The characerisic daa for he case seleced are indicaed in he figure. The lowes curve is he 'weing line', he oher curves are he 'damage lines'. The damage emperaure is indicaed beside each of hese curves. The curves in fig. 10 give a furher indicaion ha he ime inerval beween weing and damage is consan, a leas a preheaing emperaures lower han abou 50 oe, since he curves in ha range are parallel. Finally, we should noe ha preheaing may be applied for reasons oher han hose considered here, in paricular o give he flux he desired viscosiy. ---T pre Fig. 10. The effec of preheaing on soldering. The calculaed (dimensionless) soldering ime is ploed verically, he preheaing emperaure Tpro horizonally. The preheaing emperaure is he emperaure o which he body of he elecronic componen and is lead are raised before soldering sars. W weing line; his gives he minimum ime necessary for weing he lead o be soldered (fig. 2). The hree oher curves are 'damage lines'; he value beside hem is he emperaure a which damage occurs. The accepable soldering imes lie beween he damage line and he weing line W. A preheaing emperaures below abou 50 oe he ime beween he occurrence of weing and damage is roughly consan, a higher emperaures i is shorer. The values used for he calculaion (fig. 2) are shown in he figure. Summary. The aricle describes he calculaed hermal behaviour of elecronic componens, such as capaciors and resisors, during sof soldering o prined-circui boards. The componens consis of a body, which mus remain below a defined damage emperaure, and a meal lead o be soldered, which mus a leas reach he weing emperaure. Each componen has a ypical range of solder-bah emperaures and soldering imes in which boh condiions are fulfilled. For differen componens hese ranges can be made o overlap sufficienly by varying he lengh of he lead. Conversely, he resuls of he calculaions can be a help in designing new componens o mee he soldering condiions. The hermal behaviour is simulaed by an elecrical nework ha saisfies a firs-order differenial equaion. A dimensionless emperaure-ime diagram derived from he nework simplifies is applicaion o various componens. The condiions deermining he validiy ofhe model and he main hermal quaniies are discussed. The prined boards have lile effec on he hermal behaviour. The weing ime for iron leads is abou a sixh of h a for copper leads. n cases where weing does no occur 'immediaely', preheaing has no effec on he ime beween weing and damage. Wave soldering gives body emperaures ha are a few per cen higher han in dip soldering, wih slighly shorer weing imes.