SMT Soldering Handbook surface mount technology 2nd phần 5 potx

Parameters of the solderbath and the wave The level of molten solder in the machine should at all times be kept strictly at theheight recommended by the maker.. On the exit side of this

demand a solder with a different melting point or maybe a certain percentage ofsilver, the user will be well advised never to change the specification of hissolder The upheaval which would be caused by changing from the standardtin–lead solder to a lead-free one (Section 3.2.3) explains the general reluctance

of the industry to adopt a lead-free technology, unless forced to do so

4 The standard wavesoldering temperature of 250 °C/480 °F plus or minus a fewdegrees is, like the conveyor angle, the result of over four decades of practicalwavesoldering experience Without a compelling need, it is advisable not todepart from it

4.7.2 Choosing and monitoring operating parameters

Condition of the flux

Given that the choice offlux is settled, the contents of the fluxer should at all timesmatch the density and/or the acid value which is specified in the vendor’s data sheet.Section 4.2.2 discusses in detail how this requirement can be met, by automaticequipment if required It is worth restating at this point that the success ofwavesoldering depends critically on the consistent quality of theflux, and that thisconstancy is assured more easily with sprayfluxers than with foamfluxers

Amount of flux per unit of board area

This parameter also affects the soldering success, though to a lesser degree than thedensity and the activity of theflux Too much flux means more solvent in the fluxcover and, unless the preheater is adjusted accordingly, a risk of boiling andsolder-prill formation as the board passes through the solderwave If boards have to

be cleaned after soldering, too muchflux reduces the cleaning efficiency Too littleflux, uneven fluxcover or, worse, unfluxed patches inevitably cause soldering faults,such as bridges, icicles, solder adhering to the board and open joints, especially withlow-solids fluxes With these, the margin of error is much narrower than withhigh-solidsfluxes

The thickness of theflux cover can be controlled to some extent with the varioustypes of sprayfluxer, but foamfluxers permit very little, if any, control over thisparameter At the time of writing (1997), there is no equipment on the market forautomatically monitoring the thickness of theflux cover A frequent visual check ofthe overall appearance of the soldered boards is the best method of ensuring thestability of this important factor Automatic video surveillance of the output of asoldering line should be capable of giving warning of a malfunction of thefluxingunit

Intensity of preheating

Insufficient preheat leaves too much solvent in the fluxcover, which is thereforemore liable to be washed off in the solderwave, leading to bridging or open joints.This factor is particularly critical with double waves, where a substantial portion of

theflux cover must survive the passage through the first, turbulent wave Moreover,

if the board is too cool, the solder may not rise through all plated holes and form therequired solder meniscus on the upper board surface

Too sharp a preheat can cause trouble with rosin-basedfluxes: overbaking such aflux will cause the rosin to polymerize This reduces its mobility, so that it mayobstruct the solder in tinning all solderpads or in rising to the top surface of theboard It will certainly make cleaning less efficient

By contrast, fluxes with a low solids content and very little rosin, and theso-called ‘no-clean’ fluxes (Sections 3.5 and 8.1) which are mostly rosin-free,require more intense preheating to ensure that theflux coating is not washed off inthe double solderwave With thesefluxes, most vendors suggest that the underside

of the board should have a temperature of 120 °C/250 °F on emerging from thepreheating stage

The methods of controlling the intensity of preheating are dealt with in Section4.2.3

Parameters of the solderbath and the wave

The level of molten solder in the machine should at all times be kept strictly at theheight recommended by the maker Many machines arefitted with an automaticsolder feeder, which maintains the correct solder level Failing an automatic levelcontrol, the solder level must be regularly checked at intervals depending on theusage of the machine, and if necessary topped up Unlessfitted by the maker, it isadvisable to install a simple solder-level sensor, which gives an audible or visiblewarning as soon as the solder level drops below the maker’s danger mark

If the solder level drops too low, dross and flux-residues which float on thesolderbath can be sucked into the inlet of the solderpump Once in the solderstream, they tend to deposit on the solder conduits and the pump impeller Thesedeposits interfere with the steady running of the solderwave, as will be discussedbelow Particles of dross andflux which reach the wave nozzle emerge in the wave

as small, but conspicuous, black spots, which pop up in the wavecrest andfinish up

on the surface of the solderjoints Such dross orflux inclusions do not necessarilythreaten the function or reliability of the affected joints, but they are a legitimatecause of rejection by quality control or by the customer

The temperature of the solder is one of the most basic wavesoldering parameters.The general suitability of 250 °C/480 °F for most wavesoldering tasks has beenmentioned already Close adherence to this value is less critical than is oftenassumed, an accuracy of ±2–3 °C/4–6 °F being quite sufficient It is much moreimportant to guard against a slow, unnoticed upward or downward drift of thesolder temperature away from its set value The temperature readout on the controlpanel of the machine, together with its warning signals, may be misleading: software

or functional errors are not unknown The safest way to guard against this danger is

to check the actual solder temperature halfway through every working shift bychecking it with a reliable, preferably occasionally re-calibrated, handheld tempera-ture measuring instrument, with its sensor placed in the solderwave about 5–

10 mm/0.25–0.5 in below the crest

Figure 4.30 Checking the wave height : Conveyor angle

The height of a wavecrest is directly linked to the speed of the solder pump,which with most good machines has a slip-free, tachometrically controlled drivewhich is protected against variations in the supply voltage The waveheight and itsconsistency across the whole width of the wave can be checked very simply bysliding a piece of plain FR4, with gradations marked on it, across the length of thewavenozzle while the pump is running (Figure 4.30) It is advisable to carry out thissimple check at the beginning of every shift Some computer-controlled machinesarefitted with a sensor-operated surveillance of the height and integrity of theirsolderwave(s)

The depth of immersion of a board into the crest of the solderwave is normallyequivalent to the thickness of the board It is therefore important that the underside

of the board is strictly parallel to the line of the wavecrest to well within thismeasure This is easily checked by letting a piece of plain FR4 without copperlamination, as wide as the largest board, run across the wave and stop briefly over thewavecrest Theflattened wavecrest will be clearly visible through the translucentFR4 If the board is parallel to the wavecrest, the width of the band formed by theflattened wave will be the same across the whole breadth of the testboard (Figure4.31)

If the board is not parallel to the wavecrest, the whole conveyor must be tiltedsideways until a parallel position is achieved Provisions for carrying out thisadjustment are, or should be, a feature of every wavesoldering machine As analternative to the FR4 board, many machine vendors can supply a plate of heat-resistant borosilicate glass which carries a pattern of parallel lines to make it easy tocheck the width of the wavecrest across the plate To make sure that the glass platedoes not crack during this manoeuvre, it is advisable to pass it over thefluxer and thepreheater before arresting it over the wave With FR4, this is not necessary.Uneven or rough running of the solderwave, such as fluttering of the

Figure 4.31 Checking depth of immersion and horizontal alignment between wavecrest and board

waveheight, can be a sign that deposits of dross orflux residues have formed on thepump impeller or the solder ducts, often as a consequence of an unduly low solderlevel in the machine (see above) With all jetwaves, even quite small accretions ofdross orflux residue in the exit slot of the wavenozzle can ruin the smooth profile ofthe solderjet The wavecrest becomes ragged One big dross particle can depress it

by quite large amounts, naturally leading to serious soldering defects Regularcleaning of the nozzle aperture is an important requirement with all jetwaves It isbest carried out by drawing a scraping tool, made from soft steel or PTFE, along thewhole length of the nozzle aperture Most vendors can supply suitable implementsfor the purpose A scraper made from aluminium, brass, copper or hardened steelshould on no account be used

With all double-wave machines, the second wave is of the ‘asymmetrical’ type(see Section 4.4.4) On the exit side of this kind of wave, the board lifts off from ahorizontal pool of solder, whose surface moves in the same direction and ideally atthe same speed as the board on its conveyor for reasons which have been explainedalready (see Figure 4.14) The match between these two speeds can be checkedsimply byfloating a small steel ball, for instance from a ball bearing, on the soldersurface and comparing its movement with that of the board conveyor

Conveyor speed

The conveyor speed is a critical wavesoldering parameter On the one hand, theheat received by a board is inversely proportional to the speed at which it travelsthrough the preheating unit at a given setting of the heater panels On the otherhand, the maximum practicable soldering speed of a wave machine is governednot only by the ability of the solderwave to get the necessary amount of heat intothe board within the time available for this, but also by the complexity of itspattern and the density of its population of components Furthermore, multilayerboards with high heat capacity must travel more slowly than simple single-layerboards Boards with closely set SMDs andfine-pitch multilead components musttravel over the wave more slowly to give the solder a chance to flow into thenarrow gaps between neighbouring components, and to drain away from thefinepattern of leads

With most soldering machines, the set and the actual conveyor speed aredisplayed on the control panel Nevertheless, it should be part of the machine-minder’s task to check the actual against the displayed conveyor speed of themachine once every day, with the aid of a stopwatch and a simple marker, travel-ling on the chain conveyor over a measured distance marked on the conveyorrail This very simple test can save hours of expensive rework of boards, should theconveyor speed have drifted from its set value, or should the machine controlhave started to malfunction

Computer-controlled soldering machines

The large number of interlinked operational parameters makes wavesoldering anatural subject for computer control, which generally has two tasks

Thefirst task is the monitoring and stabilizing of all parameters, which will havebeen established as optimal and stored in the program It is worth saying again thatthis automatic pilot does not relieve the operating personnel from periodicallyverifying that the machine does in fact run correctly The functions which must bewatched are those which are difficult if not impossible to monitor by sensors, such asthe behaviour of the foamwave, the spraypattern of the fluxer and the correctbehaviour of the solderwave

The second task relates to parameters which can be adjusted to suit a given type ofcircuit board The main parameter here is the conveyor speed, which can be raisedwith boards of simple pattern and modest thermal requirements, or which may needlowering for complex or multilayer boards Linked to this is the intensity of preheat.Low-temperature emitters respond slowly to a change in heater current (see Section4.3.2), and this factor must be considered in the program Alternatively, theinclusion of one or more high-temperature emitters in the preheating unit willpermit a much faster response to the commands of the computer Some types ofmachine allow for a choice between foamfluxing and sprayfluxing to suit differenttypes of board

With computer-controlled soldering lines, each board normally carries a barcodewhich calls up the correct parameter as it enters the machine Nevertheless, it is not

at all advisable to run a soldering line with a random mix of different types of board.

It is advisable to gather them in as large batches as possible

The following strategy for starting up a new machine, or changing to a new type ofboard, has proved its worth in practice:

1 Check that the conveyor angle is near 7°, and that the conveyor is laterallyhorizontal Place a plain piece of FR4, of the same thickness and size as theboards to be soldered, in a board carrier or into the chain conveyor and move itforward into thefluxer With a foamfluxer, adjust the air pressure so that thewave can hold the required height with a good margin With a sprayfluxer, setthe width of the spraypattern to suit the width of the board

2 Move the board forward to the solderwave With double-wave machines, setthe primary wave, whether it is of the turbulent type or a jetwave, as high as ispossible without causing the solder to push through apertures in the board andflood the top surface With a jetwave of the type where the solder flows in thedirection of the travelling board, make sure that the baffle at the trailing edge ofthe boardfixture is high enough to prevent the solder from flooding the top ofthe board as it leaves the wave Single jetwaves of the counterflow type need asafety baffle at the leading edge of the board Again, set the wave as high as ispossible withoutflooding the board

The secondary wave is always of the asymmetrical laminar type The board ismoved forward so that its leading edge is just in front of the wavecrest Adjustthe waveheight so that the crest comes approximately level with the top surface

of the board This means that thicker multilayer boards dip deeper into thewave, spend more time in contact with it, and in consequence receive moreheat

Having done this, move the board forward into the wave and check whetherthe board is parallel with the wavecrest, as shown in Figure 4.31 If youfind thatthe board is not parallel with the wavecrest, do not try to adjust the setting ofthe whole solderbath or of the wavenozzle, but tilt the conveyor, as has beendescribed already

3 Next, set the conveyor speed at half the value which the vendor gives as itsmaximum speed, unless operational requirements make it necessary to workfaster than that It is worth remembering that it is a fact of life in engineeringthat the failure-rate or fault-rate of any given equipment or process begins torise exponentially as it is driven at a rate or speed approaching its designedmaximum (Compare the number of pit stops during an Indianapolis race withthe service requirements of a family car.)

Run a board of the pattern which is to be soldered on the machine, withoutcomponents, through thefluxer and the preheating stage and check its tem-perature on leaving the latter (Section 4.3.4) If it gets too hot, reduce thesetting of the heaters If on the other hand the heater, even at its maximumsetting, does not get the underside of a heavy multilayer board hot enough,fit a

top reflector to the preheater if none is provided If the board is still too cool,reduce the distance between the heaters and the conveyor Only if all else fails(and in this case the design of the machine must be at fault) lower the transportspeed or, better, modify the heating stage yourself.

4 Having balanced the setting of the heating stage and the conveyor speed againsteach other, proceed to solder about ten fully assembled boards, and check thesoldering quality of each carefully If this is satisfactory, enter the set of workingparameters into the machine computer or your production control manual Iffaults persist in an erratic pattern, check the stability of thefluxer and wavesetting; as a last resort, lower the conveyor speed and reduce the setting of theheating stage accordingly If faults persist systematically with one or more givencomponents, check their solderability or the suitability of the layout For details

of the systematic analysis, interpretation and elimination of soldering faults anddefects, see Chapter 9

Daily

Clean the wavenozzle at the end of the shift, and if necessary also at the mid-shiftbreak Turn on the solderpump and check whether the wavecrest is level and stable

If you are not satisfied, switch off the pump and scrape the inner walls of the solder

conduit with an annealed hacksaw blade An annealed hacksaw blade will not snap

and constitute a potential danger, nor will it damage the conduit Restart thesolderpump and skim dross andflux residues, which may now be flushed throughthe wavenozzle, from the solderbath This maintenance is especially important withjet nozzles

At the end of the day, clean splashes of solder andflux from the top of themachine and the rims of the solderbath With endless-chain conveyors, check thecondition and functioning of the automatic chain cleaners With board carriages,remove excessiveflux buildup from the holding jaws

This schedule can be greatly relaxed with soldering units which work under anitrogen atmosphere (Section 4.5.2)

Monthly

Lift the wavenozzle assembly and the pump impeller from the solderbath andremove all adhering dross andflux residue With foamfluxers, renew the air filterand clean the foaming stone With sprayfluxers, clean spraynozzles, if any, andremove buildup of driedflux, if any Renew flux in fluxer, unless in the case of afoamfluxer the scheduled flux renewal has taken place earlier Clean the exhaustsystem of thefluxing unit Clean and, if necessary, renew air filters

Annually

Carry out a complete overhaul of the soldering machine Check all board carriages,

if any, for correct setting and alignment

Figure 4.32 Taking a solder sample

This schedule applies for soldering lines which are in constant use With machineswhich are used only sporadically the schedule will, of course, be stretched accord-ingly

4.7.5 Check-analysis of the solderbath

Depending on the utilization of the machine, the solderbath should be checked atleast once a year for its tin content and its impurity levels Most solder vendors areable to carry out this analysis for their customers Unless your own organizationmaintains a central analytical laboratory, and sometimes even then, it is better andquicker to employ the services of an outside specialist For the interpretation of theanalytical report, and the measures to be taken if it is unsatisfactory, see Section3.3.3

An analytical laboratory requires a sample of solder weighing 100–200 g/3–6 oz,

in the form of a small ingot In order for this sample to be meaningful andrepresentative of the contents of the solderbath, the following sampling procedureshould be followed The wave is switched on and kept running for 1–2 minutes.The sample is then taken from the over-run of the wave with a small stainless steelladle, which must be absolutely dry This is best assured by preheating it in ablowflame The sample is then poured into a simple mould fabricated from heavysteel or stainless steel sheet, as sketched in Figure 4.32 This mould too must beabsolutely dry, but it ought not to be too hot because the sample should solidifyreasonably quickly

4.7.6 Dealing with dross

The nature of dross and the manner of its formation are discussed in Section 4.4.5.The layer of dross which forms on the surface of the solderbath (unless the machine

is run under nitrogen) must be removed periodically to prevent it from beingsucked into the pump inlet Skimming the dross twice daily is sufficient for this

purpose (Section 4.4.5) The simplest and best method is to gather the dross intoone corner of the bath surface with a simple stainless steel implement, and then to liftthe lump of dross out of the bath with aflat stainless steel spatula such as is obtainable

in any hardware shop Tilting the spatula after lifting the dross allows most of theclean solder trapped in it to drain back into the bath The rest is then put into a steelcontainer which is provided with a lid Most solder vendors are prepared to takeback solder dross from their customers once a sufficient quantity has accumulated,and will credit them for a portion of the clean solder contained in it Depending onthe circumstances, it may be advisable to demand an analysis of the returned drossfor its metal content from the solder vendor Compact, fully enclosed electricallyheated melting pots for the in-house recovery of solder from dross are offered bysome equipment vendors

Skimming the dross from the solderbath twice a day should be enough Frequentskimming in order to make the solderbath look attractive, and maybe to impressvisitors or the management, is not only unnecessary but also increases the amount ofdross which forms on the machine An existing layer of dross helps to protect thebath from further oxidation

4.7.7 Hygiene and safety

Lead and its toxic nature

Solder contains about 40% lead, and lead is toxic However, if treated and handledwith common sense, there need be no danger to any person working with solder inany of its many forms, such as solderwire, solder ingots, molten solder or solderpaste, provided a few basic facts are recognized

Lead can be absorbed into the human body only through the digestive system,while skin contact is harmless Put crudely, the basic rule is therefore ‘Do not eatlead, in any of its forms.’ In practice, this means strict observation of a number ofsimple rules

Don’t smoke, eat or consume drinks on the job Having handled solder or dross,wash hands thoroughly before smoking, eating or drinking The reasons for theserules are obvious: handling a cigarette or food with solder-contaminatedfingerscarries the danger of ingesting lead-containing solder Even small amounts matter,because lead is a cumulative poison which is not excreted by the normal bodilyfunctions Quite apart from that, soft drinks should not under any circumstances beconsumed near any part of an electronic assembly line Fruit juices and fruit sugarform reaction products on metallic surfaces which severely affect solderability, andwhich are difficult to remove Aerosol, formed for example by a fizzy soft drink, can

be fatal for the solderability of a circuit board

An often neglected danger point is the habit of chewingfingernails The spacesunder the fingernails are notorious collectors of dirt and dust, picked up fromeverything that is being handled or touched (as any forensic scientist knows).Habitual nailbiters should therefore on no account be given jobs which involve thehandling of solder in any of its forms

Dross must be handled with caution and common sense: it contains a proportion

of powdery lead oxide, which is more dangerous than metallic lead because it isabsorbed more readily into the digestive system Hence the rule of placing drossskimmings into a metal container which isfitted with a lid Dross must be handledgently, so that it does not form a cloud of dust It is a sensible precaution to issue adust mask to all operators who have to handle dross in larger quantities On theother hand, there is no reason for wearing a dust mask when skimming dross from asoldering machine, because in this form the oxide is trapped within the bulk of themetal and its adheringflux residues.

Handling molten solder

Molten solder is quite hot and must be treated with respect The main danger whenhandling it arises from the fact that it will spit and splatter violently when it comes incontact with a wet or even slightly damp surface This spitting is caused by theexplosive evaporation of any surface moisture trapped under the molten metal.Hence the strict rule, already mentioned, that every implement which comes intocontact with molten solder must be meticulously dried by preheating By contrast,small amounts of liquid spilled onto the surface of molten solder will hiss awayquietly without spitting

Drops of molten solder on the skin can be painful, and cause small but relativelyharmless local burns To stop the pain quickly, touch a cold metal surface, or runcold water onto the burn Never apply oil or grease, which will only make mattersworse Application of a small amount of burn-ointment, which normally containspicric acid, stops the pain, promotes quick healing and prevents blistering It isuseful to keep a tube or tin of it handy near any machine or bench where moltensolder is handled

On the other hand, even a minute drop of molten solder which reaches the eyecan fatally damage sight It is therefore important to issue all operators who have tohandle molten solder, for example when taking a sample of solder from thesolderwave or when emptying a solderpot, with safety goggles There is, however,

no need to wear goggles when removing the safety screen to watch a board passingover the solderwave, or when skimming a solderpot (provided the skimming tool isdry)

Wearing protective gloves is a wise precaution when sampling the solder orcleaning the wavenozzle When handling larger amounts of molten solder, such aswhen emptying a solderpot, it is advisable to wear an apron or protective clothing.Solderdrops clinging to clothing are easily removed by touching them with a smallsoldering iron set at a low soldering temperature, provided the material is entirely

of naturalfibre such as wool, cotton or linen With synthetic fibres, this methodwould not work, and scraping or plucking the solder off is the only way

When faced with the task of handling larger amounts of molten solder, it is best toplan one’s strategy in advance: decide what you want to do, and how best to do it,before you start Have all implements and receptacles dry and ready in their properplaces Do not hurry, and move slowly and with deliberation

4.8 The role of adhesives in wavesoldering

SMDs must be anchored to the board before they are wavesoldered because theyhave no leadwires or legs with which to hang on to the substrate Adhesive jointshave been found to provide the best answer to the problem Their mechanicalproperties are adequate for the task, and they can be broken without undue force ifnecessary

4.8.1 Demands on the adhesive and the glued joint

The glued joint must be strong enough to hold the component securely to the boardduring any handling operation which may precede the soldering process, forexample the insertion of wired components with a ‘mixed’ board, and above allduring the wavesoldering procedure itself These mechanical loads are only modest,

at most of a magnitude of a few newtons It is important, however, that the jointdoes not distort or disintegrate under the influence of the flux solvents during thepreheating stage, and especially during the passage through one or two solderwaves.Should the removal of a glued and soldered component become necessarybecause on inspection it has been found to be faulty or wrongly placed (Section10.2), the joint should be capable of being broken without undue force andconsequent damage to the substrate (Section 4.8.5) Finally, during the life of theassembly, the glued joints should not give off or leak any substance, particularly notone of an ionic nature, which could lower the surface resistance of the board orinterfere with the function of the assembly

4.8.2 Storage and handling behaviour of adhesives

Adhesives for SMDs are of the reactive, single-component epoxy or acrylic type.Solvent-containing adhesives and two-component reactive adhesives, which re-quire mixing before use, are unsuitable for industrial SMD wavesoldering.Single-component adhesives are a mixture of two ingredients, a polymer-resinand a hardener, which are capable of reacting with one another, forming a rigidstructure of crosslinked molecules This reaction requires a trigger to set it off,which may be a rise in temperature, or exposure to light in the visible or the UVrange, or both these triggering agents, acting simultaneously or in sequence

A good SMD adhesive must satisfy a number of specific requirements:

1 During storage, resin and hardener should not of course react with one another.With some adhesives, this may require storage in a refrigerator at about

5 °C/40 °F, to ensure a storage life of up to one year, which is what theindustrial user expects With many modern adhesives, refrigeration is no longernecessary, and storage times of up to one year at room temperature (say

25 °C/78 °F) are not unusual

2 A drop of adhesive, as dispensed onto the board, may have to bridge a gap between0.01 mm/0.4 mil and 0.3 mm/12 mil in height (the standoff height of thecomponent)while, depending on the geometry of the layout, its base may have to

Figure 4.33 Demands on the adhesive spot

fit into a very narrow gap between two footprints (about 1 mm/40 mil with amicromelf) (Figure 4.33) As a general rule, the dot as put down on the boardshould be about 0.05 mm/2 mil higher than the standoffheightofthecomponentwhich it has to hold down This requires the dispensed adhesive to retain its shapewithout sagging or ‘slumping’ Any sideways spread of the dispensed drop wouldnot only lower its height, so that it might fail to contact and hold the component,but it could also spread over the adjacent solderpads, totally and possiblyirreparably ruining their solderability, and thus the whole circuit board The type

of behaviour in which semiliquid substance retains its shape is called ‘thixotropy’.For similar reasons, the solder pastes which are used in reflowsoldering must alsoexhibit thixotropy, which is discussed fully in Section 5.2.1

Apart from a sideways slump of the adhesive drop, it would be equally fatalshould one of the more mobile constituents of the adhesive leak out sidewaysfrom the drop and contaminate an adjacent solderpad Finally, the adhesivemust separate neatly from the dispensing nozzle or placement pin, withoutforming a tail or thread which might tip over and fall on a solderpad

Theflow behaviour of an adhesive is necessarily temperature dependent,making it more mobile at highter temperatures Most manufacturers havesucceeded in reducing this temperature dependence to a minimum However,since a very precise dosage of the dispensed adhesive drop is of the essence,especially with very small melfs and chips, the dispensing ampoules on someplacement systems are heated to a standard temperature

As a rule, the adhesive does not sit directly on the FR4 of the board, but onthe solder resist This places certain demands on the adhesion and the surfaceproperties of the latter which are discussed in Section 6.1

Very often one or more conductors will pass between the solder pads of a

component It is important that these conductor tracks are not covered with alayer of solder, as might be the case with boards made by a ‘subtractive’ process.

If they are, the solder will melt underneath the solder resist as the board passesthrough the solderwave Because the solder resist starts to crinkle as the solder

on which it sits melts, the result is called the ‘orange peel effect’ An SMD glued

to the solder resist loses its safe anchorage when the solder underneath the resistmelts, so that it is in danger of being washed off in the wave For this reason,boards for wavesoldering SMDs should preferably be of the ‘solder mask overbare copper’ or ‘SMOBC’ type (see Section 6.1)

3 After an SMD has been placed on its adhesive dot, it must stick to it stronglyenough to prevent it from shifting its position or falling off, while the board ishandled between the placement of the components and the curing of the gluedjoints This holding power of the uncured adhesive is called ‘green strength’.Furthermore, an adhesive must be able to maintain its thixotropic behaviourand its green strength for at least 24 hours between being taken from itscontainer, or discharged from its dispenser, and its being hardened or curedprior to soldering (open time)

4 Last, but not least, the adhesive should have a distinctive and conspicuous,perhaps luminous, colour, so that missing or misplaced dots are easily spotted.Orange or bright red seem to be the preferred shades

4.8.3 Applying the adhesive

The precision of both the placement coordinates and the size of every individual dot

of adhesive are critically important, especially with small melfs and chips: a placed dot, or one which is too large and becomes squeezed out during placement,

mis-is liable to cover a solderpad and make it unsolderable Removing cured adhesivefrom a pad surface is one of the most costly and hazardous operations in correctivesoldering (Section 10.1.1) On the other hand, a dot of insufficient height may notconnect with the component it is supposed to hold

In most situations adhesive dots have to be of varying height, for reasonsexplained above There are several alternative methods to achieve this:

Sequential application of single dots

Dispensing the adhesives from the nozzle of a cartridge or ampoule is widelypractised Most vendors offer adhesives in air-pressure operated ampoules, whichcan discharge the content in a controllable manner For manual placement, thepressure impulse in the hand-held ampoule is controlled by the operator through afootpedal or a press-button The operation is simple, and misplaced adhesive can bewiped off, with solvents supplied by most vendors

Dispensing adhesive from ampoules can be mechanized in two ways:

1 For putting down dots of adhesive onto boards before the components areplaced, automatic equipment, which is capable of being programmed, is on the

market With these machines, the dispensing ampoule is mounted on an xy

Figure 4.34 Dispensing gantry

movable gantry (Figure 4.34) The precision and repeatability of dosage andplacement are of a high enough order to meet the requirements of adhesiveapplication to modern, closely populated boards

The distance between the nozzle tip and the board can be varied andprogrammed, not only to suit the size of individual dots, but also to enable twodots to be placed on top of one another (piggy-back) to cater for exceptionallyhigh standoffs A ‘suck-back’ at the end of a delivery impulse prevents theformation of a dangerous string of glue when the nozzle is lifted off Thedispensing program can be derived from the software of the board layout Thespeed is limited by the various manoeuvres which have to be performedbetween discharges, such as starting and stopping the dispensing head, adjusting

its position in the vertical z axis, actuating the displacement mechanism, and

executing the ‘suck-back’ Within these limitations, vendors claim maximumachievable dispensing rates of up to 17 000 dots/hour, i.e over 4 dots/second.Another technique for putting down individual, metered amounts of adhes-ive is derived from ink-jet printing The duration of an individual discharge is asshort as 0.001 seconds, which makes it possible to use ‘on-the-fly’ dispensing,where the dispensing head does not stop moving Thus, dispensing speeds of 20dots/sec, i.e 72 000 dots/hour, are possible Jet dispensing units can be integ-rated in-line with high-speed pick-and-place equipment The major adhesivevendors are able to supply adhesives suitable for jet-dispensing

2 Many automatic sequential pick-and-place machines are or can befitted with

an adhesive dispensing station, which puts down measured dots of adhesive andprecedes the component placement station The dispensing details are the same

as with the dispensing gantry With simultaneous pick-and-place machines, theadhesive is often placed directly on the underside of each component in thetime interval between pickup and placement (see Section 7.3.2)

Stencilprinting the adhesive

Some major vendors offer equipment and adhesives, which by using speciallydeveloped plastic stencils, and by controlling the thixotropic behaviour of the

Figure 4.35 Hardening curve of an epoxy adhesive

adhesive, permit the print-down of adhesive dots of varying and controllable height

on to a board in one single printing operation By choosing the right aperturediameter, dots varying in height from 0.125 mm/5 mil to 1 mm/40 mil can beprinted with a 0.3 mm/12 mil thick stencil onto a given board in one pass Also, dots

do not necessarily have to be round, but may be given other outlines if required.The advantages are obvious: apart from speeding up the process, parameters of theadhesive such as a tendency to form strings are no longer critical Vendors of theprinting equipment claim printdown rates of up to 14 4000 dots per hour

4.8.4 Curing the adhesive joint

After all components have been placed on the ‘green’ adhesive, the joints must becured Curing transforms the adhesive from a viscous mass into afirm, solid body.This is achieved by triggering a reaction between the two constituents of theadhesive, which crosslinks the individual mobile polymer molecules of the resinconstituent into a coherent, semicrystalline mass It is important that the joint doesnot shrink or crack during curing

With most epoxy and acrylic adhesives, heat provides the required trigger As isnatural with any chemical reaction, the higher the temperature, the shorter the time

in which the crosslinking reaction proceeds throughout the joint (Figure 4.35).With a modern adhesive, all joints on a board are sufficiently hard for wavesolderingafter two to three minutes at a curing temperature of 120 °C/250 °F

Curing can be carried out in an infrared oven, similar to or identical with areflowsoldering oven (Section 5.4.4) It must be remembered, however, that with

Figure 4.36 Location of adhesive spots for ultraviolet hardening

the exception of small melfs and chips, the adhesive joint is shielded from theradiation by the sometimes large and thick component For adhesive curing,therefore, an efficient convection oven is best There is no advantage in carrying outthis curing operation in a nitrogen atmosphere

As an alternative to heating, or possibly in addition, exposure of the adhesive toultraviolet light is also effective in promoting the curing process This mechanismwill only work if the adhesive, or at least some of it, is ‘visible’ to the light source andnot covered by the component (Figure 4.36) Simultaneous exposure to heat andlight can shorten the curing time by up to 50 per cent

A type of adhesive is available which contains a light-sensitive initiator This hasthe effect of adding a ‘hairtrigger’ mechanism to the crosslinking reaction: exposure

of the dots of adhesive to a dose of visible light for one-half to four minutes(depending on the intensity of the light source), before the components are placed,firms up the adhesive and increases its green strength without starting the crosslink-ing proper Once the components are placed, which must be done within 30minutes after the light exposure, full curing of the joints can be completed in a muchshorter time or at a much lower temperature than is the case with conventionaladhesives

4.8.5 The glass transition temperature

No assembly process is entirely free from faults, and occasionally glued and dered components must be removed from a board However, while soldered jointscan be unsoldered, a glued joint cannot be ‘unglued’ but must be broken Theso-called ‘glass transition’ mechanism of crosslinked polymers makes it possible tobreak cured joints without damaging the board Above a certain temperature,

wavesol-which is called the glass-transition temperature (T%) the crosslinked bonds between

neighbouring molecules begin to open and the molecules start to regain theirmobility For a cured joint this means that the adhesive loses its rigidity and begins tobehave like a highly viscous substance Depending on the type of adhesive and on its

curing history, its T% lies between 35°C/95°F and 80°C/175°F This means that after heating it to its T%, the joint can be separated without much force In

desoldering practice (Section 10.2) this is done by twisting the SMD The resulting

shearing force will readily break the joint Pulling the joint apart would require amuch greater force than twisting, and is liable to damage any conductor trackswhich may pass underneath the glued joint.

References

1 Brit Pat 798 701, 1956, Fry’s Metals, Improvements Relating to Soldering

Components to Printed Circuits.

2 Brit Pat 639 178, 1943, Eisler and Strong, Manufacture of Electric Circuits and

Circuit Components.

3 Kirby, P L and Pagan, I D (1987) The Origin of Surface Mounting, Proc.

Europ Microelectronic Conference, Bournemouth, UK.

4 Klein Wassink, R J (1989) Soldering in Electronics, 2nd ed., Electrochemical

Publications, Ayr, p 489

5 Smernos, S and Strauss, R (1988) Low Temperature Soldering Electronic

Communications, pp 148–151.

6 Klein Wassink, R J (1989) loc cit., pp 498–500

7 Oates, W A., Todd, D D (1962) Kinetics of the Reduction of Oxides J.

Austral Inst Met., 7, pp 109–114.

8 Leibfried, W (1979) Soldering without Flux German Min for Res & Technol.

(BMFT ), Rep T., pp 79–164 (in German).

9 Keeler, R (1990) New Fluxless Soldering Process El Pack & Prod., 30, 10,

p 15

10 Albrecht, J., Scheel, W., John, W., Wittrich, H., Grasmann, K H and

Liedke, V (1992) Fluxless Wavesoldering DVS Report 141, Duesseldorf,

Germany, pp 90–99 (in German)

11 Hendry, M (1995) Is there a future for nitrogen? El Manuf and Test, 1995,

pp 11–12

12 Ehrlich, M et al (1994) Investigations into the Reliability of and Inert-Wavesoldered Capacitors Proc Techn Conf Nepcon West.

Atmospheric-13 German Patent DE 195 19 188 A1, 24.5.95 (Scheel, Ring, Hafner & Leicht)

14 Anon (1997) Wavesoldering in a Vapourphase Protective Atmosphere

Prod-uctronic 5/6, p 6 (in German).

5 Reflowsoldering

5.1 The reflow concept

As has been said in Section 3.1, the making of a good soldered joint needs the rightamount of solder, flux and heat, in the right place, and at the right time Withwavesoldering as with handsoldering, theflux always comes first, and solder andheat together come afterwards With all reflowsoldering methods, the heat alwayscomes last

To begin with, solder andflux are placed on one or both joint surfaces, eithertogether in the form of a solder paste or separately,first the solder in the form of ametallic coating and then theflux at a later stage Subsequently the joints are puttogether The important point is that all this happens at room temperature though,with some procedures, the solder may have been predeposited on one or both jointsurfaces by a hot-tinning method

With all reflow strategies, the assembled joints are finally heated to a temperaturehigh enough to melt the solder, and for long enough to let it tin the joint surfacesandfill all the joint gaps As soon as this has been achieved, heating is discontinuedand the solder is allowed to solidify, the faster the better

5.1.1 SMDs and reflowsoldering

Reflowsoldering is a much older process than wavesoldering, going far back intoantiquity; under the name of ‘sweatsoldering’ it is used in plumbing to this day Withthe advent of hybrid technology in the early 1960s sweatsoldering was recognized asthe logical way of joining SMDs, which were specifically developed for hybrids, tothe metallic conductor pattern of the ceramic substrate Rosin-based soldercreamswere already in existence, though not as yet screen-printable, and the assembledhybrid circuits were mostly soldered on simple hotplates Professional reflowsolder-ing equipment and printable solder pastes became commercially available by the earlyseventies, when SMDs had begun to be used on conventional circuit boards.SMDs and reflowsoldering are ideal partners With wavesoldering SMDs, themolten solder needs help tofind its way to the joints (Figure 4.14) With reflowsol-dering, both solder andflux are already in place before the joints are heated by one