reflow instructions19397719 Reflow Manual

Research International Solder Reflow Technology Handbook

Section 1: Solder Reflow Basics

Introduction Electronics assembly is the general term for joining electrical components to printed circuit boards (PCB). Surface mount technology (SMT) is the electronics assembly process where electrical component leads are joined to the PCB via individual pad connections located on the board surface. The basic SMT process consists of the following steps: 1. Solder paste is applied to the PCB using

a screen printer. Stencils designed with holes over individual pads control the solder application to the board.

2. Electronic components are positioned on the PCB using placement equipment (pick-and-place machines, chip shooters, etc.). Component leads are placed in direct contact with the solder-pasted pads.

3. The solder paste is heated until liquidus (reflowed) then cooled until the solder hardens and creates permanent interconnection between the component leads and the PCB. This process is performed in a SOLDER REFLOW OVEN.

4. After reflow, the assembled circuit board can be cleaned, tested or assembled into a final product.

High-volume SMT lines use automated equipment to perform these steps. These lines can typically produce a completed circuit board in less than 20 seconds, with placement machines that can position upwards of 40,000 components per hour on the PCB. However, once the solder paste is

applied and the components are placed on the board, the only way to create a functioning circuit board is an effective solder reflow process.

A high-volume SMT assembly line. Reflow Know-How: Prevent and Resolve SMT Process Problems The reflow oven is the key to the soldering process. Properly working ovens should be invisible to line operators. Yet if process problems occur during board production, the reflow oven is often the first place manufacturing engineers look to find answers. Knowing the reflow process and ways the oven effects soldering results is critical to consistent SMT production. The purpose of the Reflow Technology Handbook is to explain the reflow process in a way that answers basic soldering questions. The handbook also helps qualify the performance factors you should consider to find a reflow oven to best suit your process needs.

Research International Page 2 Solder Reflow Technology Handbook

Solder Alloys The vast majority of electronic interconnections are accomplished with tin-lead alloy solders.

Tin Pure tin is a soft, shiny metal that is most often found in ore form rather than metallic form. It is easily shaped and molded without breaking. The essence of the solder process is the ability of molten tin to dissolve nearly any other metal. Copper is one of those metals, and copper is used extensively in the manufacture of printed circuit boards. In the next subsection we will discuss the role of compounds formed when surface copper is dissolved by tin.

Lead Lead, when exposed to air, has a dull gray appearance. Like tin, lead is an easy metal to work with because it is both soft and flexible. Lead does very little to aid the bonding of metal to metal during soldering. However, when lead is combined with tin at a ratio of 63% tin to 37% lead, the melting point of the resultant alloy becomes lower than that of either pure tin or pure lead. The abbreviation for tin in chemistry is Sn and for lead is Pb, so this alloy is usually written Sn63Pb37. This type of alloy is called a eutectic composition. Pure tin melts at 232C (449F) and pure lead melts at 327C (621F). The melting temperature of Sn63Pb37, however, is 183C (361F). At this temperature, the alloy goes from a completely solid to a completely liquid state

without going through a pasty stage. This is called the eutectic temperature. When eutectic tin/lead is heated to a temperature higher than 183C it is said to be in its liquidus stage. The relatively low melting temperature has made eutectic tin/lead the solder alloy of choice for printed circuit board assembly because circuit boards made of FR4 substrate or similar materials can be safely processed at this temperature. In addition, the low melting temperature means that the equipment that solders the boards does not have to operate at higher temperatures. Finally, a benefit most managers and accountants appreciate is the low cost of tin/lead compared with other alloys. Intermetallic Compounds As mentioned above, molten tin dissolves most metals. During the solder process, the primary metals in the solder form compounds with the metals in component leads and circuit board pads at boundaries between the solder and the pad or lead. (Figure 1-1).

Sn63\Pb37 solder

Copper pad

One to two micron

intermetallic layer: Cu3Sn and Cu6Sn5

Figure 1-1. The intermetallic layer at the boundary of the solder and copper surfaces.


These intermetallic compounds form a boundary layer that is extremely strong when it is no thicker than one to two microns. At this thickness the natural brittleness of the intermetallic layer is not a problem. As the intermetallic layer thickens, it becomes more susceptible to cracking. The cracking is caused by expansion and contraction of the circuit board substrate as it heats and cools during normal operation of the electrical device in which it operates. The thickness of the intermetallic layer is controlled primarily by controlling the liquidus time of the solder. Most solder manufacturers recommend a liquidus time between 45 and 60 seconds. (See Section 3 Profiling for additional information.) Other Solder Compounds Some special applications require alloys other than eutectic tin/lead solder. A common application for other solder compounds is double sided surface mount assembly. To solder on both sides of the board requires bottom side components be soldered with an alloy that has a melting temperature higher than eutectic tin/lead. Then, the board is flipped over and top side components are soldered with an alloy that has a melting temperature lower than that of eutectic tin/lead. The higher melting temperature allows the solder used for the bottom side components to stay solid during the top side reflow, reducing the risk of intermetallic growth.

Bismuth and Low Temperature Soldering The addition of bismuth to tin/lead solder reduces the melting temperature significantly without adding serious solderability problems. A typical tin/bismuth/lead alloy is Sn43/Bi14/Pb43. This alloy has a pasty stage from 143C-

163C (289F-325F). The pasty stage means the solder is no longer completely solid at 143C, but not yet completely liquid until it reaches 163C. The lower melting temperature of this and similar alloys means that the flux used in the solder paste must also be different from that used in eutectic tin/lead. This may change the shape of the thermal profile. Consult your solder paste supplier for the recommended thermal profile for the particular low temperature solder you use. Another matter to consider when you use low temperature solder is that the melting temperature of the solder is affected by any metals dissolved from leads or pads that were plated to prevent oxidation. Usually these contaminant metals raise the melting point of the solder. If this becomes a problem in your process, it might be necessary to strip the plating just prior to the low temperature soldering operation.

Silver Some printed circuit boards have components with silver plated leads. As noted in the section on tin above, molten tin dissolves most metals, including the silver in silver plated leads. One way to reduce the amount of plated silver that dissolves in the molten tin/lead is to add a small quantity of silver to the solder itself. The alloy Sn62/Pb36/Ag02 (62% tin, 36% lead, 2% silver) is good for this application.

Another popular silver alloy Sn96/Ag04 melts at 221C (430F) and is sometimes used for double sided reflow.


Lead Free Solder Among the most discussed topics in the late-1990s is the effort to remove lead from electronic products. Concerns about ground water lead contamination are forcing the debate. Many governments are demanding that Tin/Lead alloys be phased out and replaced with lead-free solders. Solder paste manufacturers are working to create new alloys to meet what is likely to become a new regulatory requirement for SMT assemblers. According to Phil Zarrow of ITM Consulting, the most promising lead-free alloys are in the Tin-Silver family with variations incorporating relatively low amounts of copper of bismuth. (Circuits Assembly August 1999). These alloys cost about the same as Sn63/Pb37, but the 221C melting temperature means that electronic components must be exposed to reflow temperatures as high as 240C for effective soldering. Another reflow alloy being considered is Sn91.8/Ag3.4/Bi4.8. Adding bismuth results in a melting range between 208C and 215C (Phil Zarrow: Circuits Assembly August 1999). Lead-free alloys are being evaluated to verify their performance to standard Tin-Lead compounds. Anyone involved in SMT production should be prepared to convert their lines to lead-free soldering by the mid-2000s.

Flux Oxygen in the air combines with metal on circuit board pads and surface mount component leads and forms oxides. This oxidation blocks the molecular attraction

between the solder and the metal at the surface of the pad or lead. Flux removes oxides from pad and lead surfaces, as well as from the surfaces of the solder particles themselves. Molten solder then can flow evenly over the surface of pads and leads and form intermetallic bonds with them.

Surface Tension and Wetting Surface tension is the attraction that molecules at the surface of a drop of liquid have for each other. If that attraction is greater than the attraction for the material which the liquid touches, the liquid will not spread, but will remain in drop form. Surface tension explains why a drop of rainwater on a waxed car stays in drop form. The wax surface is intended to repel water and keep it from bonding to the cars surface. The earths gravity works against surface tension and tries to flatten the raindrop into an oval. If the attraction of the molecules of rainwater for each other is greater than the pull of gravity and the attraction of the cars surface, the drop will remain somewhat round and self contained. If gravity is stronger and the wax is worn down, the drop will disperse. Wetting is the word used to describe the extent that solder flows over the surfaces to be bonded. Poor wetting is the result of solder particles in the paste bonding to themselves to form a sphere, so the edges of a poorly wetted solder joint are rounded rather than flat. Wetting is measured by the angle made where the edge of a layer of solder paste meets the pad on a circuit board. Poor


wetting leaves a thick edge and a large angle (Figure 1-2).

What Flux Cannot Do Even the strongest flux does not remove thick layers of oxidation. At most, flux removes the oxide a few molecules thick from a metallic surface.

Flux Chemistry The basis of flux is usually a solid that has been dissolved by a solvent. The most commonly used solid for many years in PCB assembly was rosin. Rosin is derived from pine trees, and is an inert substance, which does not conduct electricity at room temperature. Rosin becomes liquid between 125C and 130C (257F -266F). A substance known as water white flux (abbreviated w/w) is pure rosin dissolved in isopropyl alcohol. Water white is mild and able to remove only the thinnest layers of oxide during soldering.

To make rosin based flux more aggressive when removing oxides, activator substances are added. Many different substances are used as activators, but it is not as important to know exactly what the activator is, as to know:

at what temperature the activator begins to work, because this affects the thermal profile in the reflow oven, and

how corrosive the activator is, which determines whether the board needs to be cleaned and what cleaning solution is necessary. Your solder paste supplier can give you this information.

RMA and RA are the most common activator fluxes. RMA is mildly activated rosin based flux, and is much stronger in removing oxidation than water white. RA is activated rosin based flux, and is more aggressive than RMA.

Flux Residue and Cleaning Water white and RMA fluxes leave a residue after soldering that is not corrosive or electrically conductive at levels high enough to affect the function or life of a circuit board. Residue from these fluxes has often been removed by washing, however, for aesthetic reasons, or because it is non-conductive it interferes with bed of nails testing devices. The residue from RA fluxes can be more corrosive, and usually needs to be cleaned from the board. RA fluxes require two types of cleaning:

1. A nonpolar solvent to remove the rosin, as well as any oils or waxes that may have contaminated the board during handling. Nonpolar solvents such as chlorofluorocarbons

Poor wetting: >90 angle to pad Strong Attraction

Weakened Attraction

Acceptable wetting:


(CFCs) were used for years, but environmental concerns eliminated them from consideration. Semi-aqueous cleaning systems use a terpene as the nonpolar solvent. The terpene is then removed by a surfactant, which makes the terpene soluble in water.

2. A polar solvent, which is usually very pure water, removes the activator, as well as salts and other water-soluble contaminants that may have gotten onto the circuit board during handling.

A substance called a saponifier can be used to clean PCBs also. One way to understand and remember what a saponifier does is to know that the root word is sapo, which is Latin for soap. A saponifier acts like soap in reacting with substances that are not water soluble (oils, for example) so that they become water soluble and can be washed off.

Water Soluble Flux Water soluble fluxes usually are based on the oxidation removing qualities of an organic acid such as citric or glutamic acid. Organic acid fluxes are more aggressive than RMA, so they provide good cleaning, but they are usually corrosive as well. Because of the corrosiveness of these fluxes, a thorough water wash following solder reflow is necessary. Boards stored without washing for more than an hour after exiting the reflow oven may have serious oxidation buildup.

No Clean Fluxes Millions of printed circuit boards are manufactured that are never cleaned following solder reflow. This is especially true in consumer product electronic

assemblies, primarily to save money. Inert, non-corrosive flux residues are left on the boards and do not cause problems. By virtue of the fact that the boards are not cleaned, it can be said the fluxes used are no clean fluxes. However, in recent years the term no clean has come to refer to a type of flux specifically designed so flux residues are minimized. No clean fluxes generally have the same aggressiveness as RMA fluxes. Lower residue is achieved in no clean fluxes by using a lower solids content than in other fluxes. Solids content refers to the ratio of solvent thinner to solid component in the flux. A typical no clean flux now has less than 15% solids, compared with 50% in other types of flux. Since the solids in a flux are the basis for its effectiveness, reducing the solids content tightens the process window for no clean fluxes. Flux density control thus is more important than with other fluxes. No clean fluxes can be either rosin or resin based. The most noticeable difference is that rosin based no clean flux produces a thin layer of sticky residue that can be removed if desired. Resin based no clean flux produces a thin residue that is hard and not sticky, but cannot be removed by water or other solvents. This can occasionally be a consideration if rework is necessary.

Nitrogen gas inert atmosphere solder reflow is often helpful when no clean flux is being used, especially when it is used with fine pitch surface mount devices. Nitrogen gas eliminates oxygen from the heat chamber of the reflow oven so that oxidation is minimized. (See Section 4 Inert Atmosphere Soldering for additional information.)


Section 2: Printed Circuit Board Fabrication & Solderability

Good Boards = Results Board fabrication is one aspect of the electronics production industry that SMT assembly engineers often know little about. It cannot be stressed enough that poorly fabricated boards WILL cause solderability problems during assembly. This section covers PCB production basic steps as it effects the assembly of populated circuit boards. In addition, this section will describe how to apply this knowledge to prevent board solderability problems BEFORE the boards ever get to the assembly line.

Soldered boards exiting a reflow oven. Pad Oxidation and Board Storage Circuit board fabricators prefer to produce large numbers of boards in a single run. This reduces setup time and offers the benefits of economy of scale when ordering

laminate, copper foil, prepreg, and other raw materials of printed circuit boards. These purchasing and manufacturing issues cause them to offer lower per board costs to assemblers if the assembler buys in large lots. However, even in the era of just-in-time manufacturing, many companies keep boards on the shelf too long. If the assembler goes for the low per board cost and does not populate and solder the boards in a reasonably short time, oxidation buildup on pads, even if they have been plated, can cause solderability problems. Often reworking boards that exhibit poor solderability is more expensive than the savings gained by buying larger lots of cheaper PCBs. Therefore, it is important for the processing engineer in a surface mount assembly line to work with purchasing and management to determine the safe shelf life of bare boards. Solderability tests can determine the safe time limit. Ensuring a good raw material of boards is one of the most important parts of high yield surface mount assembly. What makes board oxidation particularly critical is that it dooms the process before solder even touches the PCB.


PCB Manufacturing

Substrate Manufacturing Substrate is manufactured from thin sheets of a dielectric material bonded to a sheet of electrically conductive material. Figure 2-1 illustrates the first three steps of circuit board fabrication. FR4 is the most common substrate used in printed circuit boards. Epoxy resin is used to bond fiberglass to copper foil in the creation of FR4. A fire retardant is added so the substrate can be safely soldered in later processes. Some other substrates are: polyimide/fiberglass, which can sustain

higher temperatures and is much harder than FR4

FR2-fire retardant coated phenolic/paper, which is cheap and used mainly in low cost consumer electronics

Flex circuits, usually polyimide, in some cases polyester, used in automotive and other applications where space and weight are at a premium. New technology allows adhesiveless material for even thinner flex circuits.

Problems that occur during substrate production usually have little impact on solderability during SMT assembly. The Subtractive Process After the substrate has been made, drilling machines bore holes of different diameters in the exact locations on the board. These holes are called vias. These vias are where circuits electrical connections are created between different layers of the board. Next, an image of the circuit pattern is transferred to the copper foil on the surface of the board with either a UV photoresist film or an ink screening process. In the UV process, another step is required to remove the resist material from areas where the circuit will be. Incomplete removal of this resist material can cause solderability problems later. Resist residue on the copper will not allow solder to bond to the pad. Then 0.0000150 - 0.000020 inches of copper is chemically deposited (called electroless plating) in the drilled holes. This plating provides a base on which more copper can be electrically plated. (See Figure 2-2.) Then, a 0.0010 - 0.0020 inch layer of electroplated copper is added to the

1-Substrate: copper foil bonded to fiberglass

2-Drilling and deburring

3-Transfer of circuit pattern to copper foil on substrate

Figure 2-1. The first three steps in circuit board fabrication.

4-Electroless copper plating in drilled holes

5-Add plating resist; electroplate copper

6-Electroplate tin/lead over copper plating to protect from next etching

Figure 2-2. The next three steps in producing a PCB.


chemically deposited layer. Tin/lead plating---not solder, because the metals are still discrete---covers the copper to protect it from oxidation and the subsequent etching solution step. Now it is necessary to etch off the plating resist which again exposes the copper foil on the board surface that remains from the very first step of the process. Then the tin/lead can be etched off the circuit traces and pads.

Figure 2-3. Final steps of PCB fabrication The Protective Process When the electroplated tin/lead (Figure 2-2) has been removed, the board is electrically complete. Solder mask is applied to seal out contaminants. Applying Solder Mask; Hot Air Surface Leveling (HASL) or Organic Coating Solder mask can be dry film or liquid. Both are photoimageable and require a very clean environment. After mask is applied over the entire board surface, UV-blocking artwork is placed over the board. The artwork blocks UV radiation from those places where solder mask must be removed so that the pads can be exposed for component placement during assembly.

UV radiation is then applied to the board. The mask that is protected by the UV-blocking artwork remains soft. The mask not protected by the artwork is exposed to UV, which begins polymerization of the mask. Polymerization is the linking of the polymer molecules so the mask becomes hard enough to protect the board surface from scratches and contaminants.

The mask that was protected by UV radiation is still soft, and is removed by high pressure water spray in a machine called a developer. After this, the board is heat cured to harden the mask to its final form.

Then the pads must be coated for a last time to prevent oxidation. For about a decade this usually was done with a hot air leveler, which dips the board into a solder pot for 3-10 seconds.

As the board comes out of the solder, air knives blast it on both sides, leveling the solder on the pads.

This, however, can cause problems during assembly:

Bad air knives leave an uneven pad surface so components cannot lay flat.

The solder in the hot air leveler is contaminated by a tiny amount of copper from each board. The solder left on the pads thus becomes less solderable.

Too much time in the solder pot creates a thick layer of intermetallic alloy, which weakens joints made in assembly.

To overcome these issues, some board fabricators are applying an organic compound that protects the pads, but

7-Etch off resist material; then etch tin/lead from circuits & pads

8-Apply solder mask to protect board, and legend ink with part no. & other information

9-Apply solder or organic coating to pads (HASL, fuser, or chemical bath)

xxx-xxxxxx-x

xxxxxx-x xxxxxx-x

xxx-xxxxxx-x

Figure 2-3. The finishing steps PCB fabrication.


without the thermal and mechanical stress that happens during hot air leveling.

Solderability Testing PCB fabrication can be a source of assembly problems, particularly with regards to oxidation and misapplication of solder mask. Engineers and machine operators who understand fabrication learn to make careful inspection of the raw board as the first step in troubleshooting SMT solderability problems. Several solderability test methods are commonly practiced. Regardless of the test method you use, the process engineer should evaluate the tests level of complexity, the time required and cost compared to the negative impact of the solderability problem. Here are some quick and easy test approaches that have proven useful for basic troubleshooting.

A Simple and Fast Test for Determining Pad Solderability on Bare Boards If solderability is in doubt, a simple test is to stencil solder paste on a bare board and then reflow it without components. If the joints look good, but there is still doubt, have a testing lab do an X-ray to determine

copper content of the solder or other potential causes of poor solderability. If the pads exhibit proper adhesion, then check the electronic components themselves. The following tests measure solderability of both component leads. The Dip and Look Test for Component Leads The dip and look test is similar in its ease and simplicity to the pad test. It requires a few pieces of equipment and some work space, but can be done quickly without interrupting production.

If you have established that solderability of the pads is good, take a board that has been through the screen printer, but has no components on it, to use for this test. Heat the board on a hot plate, or by whatever means are available that can be controlled.

With a thermocouple attached to one of the leads, use a tweezer to place the surface mount component on its pad. Some kapton tape placed over the component to hold it in place helps secure the part.

Heat the assembly through the thermal cycle (See Section 3 Reflow Profiling) appropriate for the solder paste, using the thermocouple as a guide. Let the assembly cool, remove flux residue if necessary, and then inspect the lead for wetting.

Notes:


Wetting Balance: A More Involved Component Lead Test Wetting balance is good test because its accuracy relies on instruments rather than the judgment of the person doing the test. The drawback is that those instruments are often not readily available and they require training for proper use. In addition, this test requires a comparison against a known specification established on a component with proven solderable leads. To perform the wetting balance test, attach a lead with a clamp to the arm of a force

sensor. A solder pot with molten solder is raised so that the lead goes at least 1/8 (3mm) into the solder. Leave the lead in the solder for 5 seconds. Remove the lead. The force pushing against the lead should be reduced as the flux activates and cleans the surface of the lead if the flux is working and the lead is not too corroded for it to work. This reduction of force will be recorded on the force sensor, and is compared with force results from the lead on the good solderability component.

Notes:


Section 3: Reflow Profiling

What is a Temperature or Thermal Profile? High quality, low defect soldering requires identifying the optimum temperature profile for reflowing the solder paste. Achieving SMT process consistency means repeating this profile over and over. Every solder joint on every board needs to be heated similarly if the desired soldering results are to be accomplished. From the solders point of view, it does not matter what the heat source to the solder joint is. What does matter is that the heat is applied to the solder joint in a controlled manner. The heating and cooling rise rates must be compatible with the solder paste and components. The amount of time that the assembly is exposed to certain temperatures must be defined and maintained. In other words, the solder reflow profile must first be defined and then maintained.

The Typical Profile The reflow profile is defined by the relationship of temperature versus time during heating. A typical profile consists of three heating slopes (the time vs temperature relationship or rate of temperature rise) defined by Figure 3-1. This three step profiling approach has been commonly used since the early days of SMT. Each solder paste defines the heating slopes and time and temperature limits within each slope. It is best to consult your solder paste supplier to determine the exact heating condition required for the paste you are using. For the purpose of discussion, we will use the traditional three step profile, which is typical of RMA pastes. The three step heating profile slopes are called preheat, dryout, and reflow. Preheat In the preheat section, the goal is to fully preheat the entire SMT assembly to temperatures between 100C and 150C. The most critical parameter in the preheat section is to control the rate of rise to between 1-4C/second. The main concern is minimizing thermal shock on the components of the assembly. For example, multilayer ceramic chip capacitors can be vulnerable to cracking if heated too fast. In addition, rapid heating can cause the solder paste to spatter.

Time (Sec.)

Preheat Dryout Reflow Cooling

50

100

150

183

200

25

75

125

175

225

T( C)o

WETTINGTIME

SOLDER LIQUIDOUS

RATEOF

RISE

PEAK TEMPERATURE

Figure 3-1. A typical thermal profile.


Dryout The second heating section, referred to as the dryout, soak, or preflow zone, is used primarily to ensure that the solder paste is fully dried before hitting reflow temperatures. It is characterized by a consistent temperature (often between 150C - 170C) for an extensive (60-120 second) time period. The dryout portion of the profile acts as a flux activation zone for RMA solder pastes. Dryout provides thermal stabilization of large and small components to ensure uniform heating as the SMT assembly enters the reflow zone. Convection ovens have reduced the need for the thermal stabilization, as the entire profile tends to be uniform (referred to as Delta T defined as the temperature difference between the warmest and coldest component lead on the board). Reflow The reflow section of the profile elevates the solder paste to a temperature greater than the its melting point. For Sn63/Pb37 eutectic solder, the melting temperature is 183C. This temperature must be exceeded by approximately 20C to ensure quality reflow for every solder joint lead. The amount of time the solder joint is above the melting point is referred to as the wetting time or timeover. The wetting time is 30 to 60 seconds for most pastes. If the wetting time is excessive, intermetallic layer may form in the joint, which result in brittle solder joints. Excessively slow cooldown while the paste is liquidous can also cause the solder joint to consist of a larger grain structure, resulting in a potentially weaker solder joint. Common cooling rates are controlled between 1-2C/second. Many reflow ovens

have water-cooled or refrigerated cooling sections so the timeover and cooling rates can be precisely controlled. What are the Profile Control Limits? The profile is not a simple line graph between time and temperature, but rather a band or process window as defined by upper and lower control limits. Heating within the limits will result in high quality solder joints providing all other aspects of the process are in control. The size of the reflow profile band is defined by the range of temperature deviation (thermal tolerance) that can occur during while yielding high quality solder joints. In Figure 3-2, the profile band size is indicated as 25C. The actual limits of the band will vary depending upon solder paste, component type, and circuit board material.

Time (Sec.)

Preheat Dryout Reflow Cooling

50

100

150

183200

25

75

125

175

225

T( C)o

PROFILE PROCESS WINDOW

GOOD

COLDSOLDER

THERMALSHOCK INCOMPLETE

DRYING

EXCESSIVELIQUIDOUS

TIME

BURNT BOARDS

Figure 3-2. A graphical representation of a 25C reflow profile process band.


It should be noted that the band width is defined by the temperature and time axis. If the product temperature profile is not maintained within the control limits, a defect will occur. Figure 3-3 indicates some

of the production problems from improper profiling and possible causes. One should note that these problems may also be caused by non-reflow processes as well.

Problem Possible Cause (Profile Related) Cracked chip capacitors Excessive rise rate in the preheat zone Solder balls Incomplete drying before reflow (dryout section too cool and too short a duration) Excessive drying temperature (fluxes skin over) Improper gas atmosphere (nitrogen versus air) Cold solder joints Insufficient time over reflow temperature Solder not wetting to leads Excessive drying time causing fluxes to deteriorate Excessive reflow temperature/time causing oxidation Solder not wet on pad Lead is heating faster than board (too much airflow) Component/board burning Excessive reflow temperature

Figure 3-3. Basic Solder Reflow Troubleshooting.

Notes:


Defining the Profile Control Limits The profile control limits need to be defined for the given SMT assembly. To define these limits, one should be aware of not only the solder paste requirements, but also any specific requirements of components or the circuit board material. The process band width is defined as the total deviation in temperature that can occur and yield reliable results. To determine the profile control limits, thermocouples are attached to a sample board populated with components. The Thermocouple Attachment Process NOTE: This section contains information from KIC Application Note #00001, Rev: 98-05. The entire document is found at www.kicthemal.com. It is strongly recommended to attach TCs with high-temperature solder for best results. An alloy like Indium #228 (88Pb/10Sn/2Ag) with a 267C solidus temperature is a good solder for TC attachment. TC Locations TC location is usually determined by identifying critical or sensitive components and attaching the TCs to the appropriate pads. Additionally, it is best to place the TCs so that you receive temperature readings from the hottest and coldest points on the board. The measurements received from these three to twelve TCs (depending on the size of the board) will allow you to determine whether the thermal profile is consistent throughout the product, and also to measure the heating and cooling profiles of heat sensitive components. In placing

your TCs, it is important to keep in mind that areas that are densely populated or have larger components will take longer to heat up and also will hold heat longer. It is also important to place TCs at the edges of the board and in any areas with small or non-existent components that will heat up faster than the rest of the board.

Thermocouple placement for determining profile control limits. (Photo courtesy of KIC Thermal Profiling.)

Surface Preparation Thoroughly clean your selected TC locations, taking care to remove any residual low temperature solder and other contaminants that might prevent complete high temperature solder wetting. Use alcohol or a suitable solvent to scrub the attachment surface. Then place the solder wick on the surface and press the soldering iron into the wick, heating both the wick and the surface to be cleaned. Use as much wick as necessary to remove all free flowing solder droplets. When solder has been completely removed, re-scrub the attachment surface to remove any remaining contaminants. Attachment Do not attempt to attach the TC by melting solder into the spot and then shoving in the bead. Place the TC bead on the attachment


surface and heat both evenly, then touch the solder to the heated TC bead and let the heat from the TC bead melt the solder. This method gives you superior wetting and a stronger solder attachment to the pad or lead. Your temperature reading will come from the first point of contact between the two wires leading from the TC. To insure accurate readings, it is critical to carefully separate the two wires all the way up to the TC bead after soldering. High temperature solder is an efficient heat conductor, so if a tiny bit gets between the TC bead and the lead or pad, you will still get an accurate reading. However, too much solder at the measurement point will increase the heat capacity of the TC and cause your peak temperature measurement to read low. Oven Setup Once the thermocouples are placed on the test board, the reflow oven is set up to create a three step thermal profile on the board. Common oven parameters that are adjusted to create the thermal profile are:

Zone setpoint temperatures Conveyor speed Fan speed Refrigerated cooling rates (if

applicable) Profiling then becomes essentially an iterative process comparing the test board temperature results to the desired profile. If the results differ, oven parameters are adjusted and the test is rerun.

Each of the curved lines is the temperature as sensed by a thermocouple attached to a product as it passes through the oven. (Graph courtesy of KIC Thermal Profiling.)

How Does Oven Performance Relate to the Reflow Profile Control Band? Once the three step profile is defined and the oven parameters are determined, the reflow oven must be able to produce consistent results within the reflow profile band. The user should be aware of all variables that influence the band. To achieve high quality, low defect soldering results, the sum of all the non-uniformities must fall within the defined reflow profile band. These variables may be defined as product related and oven related. Product Related Variables By far the most critical product related non-uniformity is that created by large mass differences on the product. Since it is easier to heat an area that has no components as


compared with an area with large components, temperature differentials will exist on the product. Oven Related Variables Consistent thermal repeatability under product load (oven temperature changes due to mass variation) and edge-to-center and front-to-back product heating uniformity are the two most common factors that effect the reflow process. The following sections describe these three main causes of non-uniform heating in the SMT reflow process: board mass differential (product related variables), conveyor and heater edge effects and product loading (oven related variables). Mass Differential The amount of temperature rise of a product as it travels through an oven heating zone depends upon a number of variables. The temperature rise of an object subjected to heat is determined by the following equation: DT = Q x A x t / (M x Cp ) WHERE: DT = Product Temperature Rise (C) Q = Heat Absorbed (W/cm2) A = Exposed Area (cm2) t = Heating Dwell Time (seconds) M = Mass of Object (kg) Cp = Specific Heat (W-sec/kg-C)

The mass (M) times the specific heat (Cp ) is often referred to as thermal mass. Greater thermal mass requires more heat to achieve a given temperature rise. In Figure 3-4, the area populated with the PLCCs has considerably more thermal mass per unit area than the area with the discrete components.

LOW MASS DIFFERENTIALHIGH MASS DIFFERENTIAL

The high mass area will be more difficult to heat. The non-uniformity caused by thermal mass differential may be measured by thermocoupling a lead on the largest and smallest device, and running a profile. (As described previously.)

Figure 3-4. Comparison of high mass versus low mass differential boards.

Notes:


Oven Uniformity A second main cause of non-uniform product heating is the oven heating pattern. This can be caused by edge effects on the heaters (less heat at the end) or by heat sinking of the conveyor system. The edge effect may be caused by insufficient airflow around the edges of the machine, or by non-uniformities that are inherent in the heaters. The heat sinking of the conveyor is really a mass effect, similar to the one described above. The oven uniformity may be measured by thermocoupling a bare board over a matrix of points, and running a profile. The matrix should include points on the board edges (front, back, left, and right), and the center of he board. Oven Repeatability Repeatability refers to the ovens capability to repeat a given profile. Repeatability is affected by machine loading (number of boards in the oven). Figure 3-5 shows a typical product loading condition. The loading factor is defined in the figure. The higher the value of the loading factor, the more difficult it is for any oven to give

repeatable results. Ovens will generally specify a maximum loading factor (ranging from 0.5 to 0.9). The loading factor becomes an important consideration when sizing an oven. (See Section 7 Choosing Your Reflow Oven.) The non-uniformity caused by repeatability can be measured by running a profile on a test board, then loading the oven down. Periodically, run the test board within the oven load and compare with the unloaded profile.

Figure 3-5: Load factor calculation combines board length and the space between the boards.

Board Length + Space

Board Length Space

Load Factor = Board Length

Notes:


Section 4: Nitrogen Inert Atmosphere Reflow

Reflow soldering in a nitrogen atmosphere is a common process consideration in SMT assembly. The issue is not the ability to reflow in nitrogen, but rather the ability to reflow in the absence of oxygen. Heating solder in the presence of oxygen will create oxides, which are generally non-solderable surfaces. Oxidation is effected by temperature, surface area, flux content, metal content and condition, and oxygen levels around the solder joint. Lets take a look at the variables that effect oxidation and discuss their impact. Temperature Oxidation is directly related to temperature. Higher temperatures mean faster rates of oxidation. Thus, reflowing at the lowest possible temperature reduces the amount of oxidation created during the reflow process. Convection technology helps (as compared with IR) in that heating uniformity is better and thus the maximum temperature on the product may be reduced. Surface Area More surface area means more area available for oxidation. As the pitch between leads becomes finer and finer, the tin/lead solder particles in the paste must become smaller to obtain good paste print definition.

As solder particles become smaller, the ratio of surface area to volume of the particles increases. This can be shown by: Ratio = (p * D2 ) / (1/6 * p * D3) Ratio = 6 / D Where: D = Particle Diameter Thus, as solder particle size decreases, the surface area to volume ratio increases.

SMALLER PARTICLES HAVE GREATER SURFACE AREAPER UNIT VOLUME THAN

LARGE PARTICLES

With more available surface area in the solder paste, there is a greater tendency towards oxidation. Flux Content In the early days of reflow soldering, solder paste contained aggressive enough fluxes to easily remove oxides. Often CFC solvent cleaning was used to remove any remaining residue from the board. As CFCs were phased out, so has the most effective means of cleaning rosin based

Figure 4-1. Surface area of solder particles.


pastes from the board. Today, water-soluble Organic Acid (OA) fluxes are used, and the cleaning process has changed to a water wash system. Many companies have eliminated cleaning completely and have gone to a no-clean process. The term no-clean is misleading, as it only implies that water wash cleaning has been eliminated. Some people have changed nothing in their process, except they no longer clean the product. This is satisfactory providing the desired functional life of the end product is not high, which may be the case with many consumer products. After several years of operation and thermal cycling, the residues that were left on the board can cause short circuits, or simply corrode the circuit away. If the desired long term reliability of the product is important, then this is not satisfactory. The amount of residues left on the product will determine its long-term reliability. The bottom line is that higher residue pastes are more solderable in air. Although the line is moving, the low residue pastes (


Oxygen Levels None of the variables in Figure 4-2 by itself is enough reason to use a pure nitrogen environment in the reflow process. However, if several of these combine (such as a fine pitch, low residue paste), then an atmosphere void of oxygen may be required in order to accomplish satisfactory reflow. This requires an oven that can run at low levels of oxygen. The oxygen level is measured in parts per million, or ppm. Air atmosphere is 21% oxygen, which is 210,000 ppm. Most ovens have the option to operate at less than 100 ppm oxygen environment.

The purity level required (if any) is dictated by the process variables discussed previously. There is no magic number of purity that is required for a given board. Solder pastes are continuously changing in the quest towards air reflowable pastes that leave no residue. While the perfect paste does not exist, solder pastes have continually improved to the point where one should experimentally determine whether nitrogen has any significant impact on your process. If it does not, do not use nitrogen!

Notes:


Section 5: Reflow Oven Heat Transfer

The Three Heat Transfer Modes in Reflow Soldering There are three different heating modes involved with most SMT reflow processes: conduction, convection, and infrared radiation (IR). All three of these heating modes occur naturally in our daily lives. Perhaps the easiest way to understand each of these heating modes is through example.

Conduction Conduction heat transfer occurs when two solid masses of different temperatures are in contact with each other. A good example is when a pan is placed on an electric burner. Most of the heat is transferred to the pan by the contact between the pan and the burner. Conduction also occurs within the same mass if a temperature differential exists within the mass. Conduction can both help and hinder the SMT reflow process. Conduction helps in

uniform product heating, as heat will conduct from a hot spot to a cold spot in the product. Thus, if a product is difficult to reflow, often a good solution is to reduce the oven heater setpoint temperatures and conveyor speed. This allows more time for the conductive flow to occur and the product will heat more uniformly. Conduction hinders the process if an edge conveyor in contact with the board is cooler or hotter than the product. Conduction heat transfer can result in a hot or cool spot along the edge of the product, preventing uniform solder joints on the outer fringes of the PCB. Infrared Infrared Radiation (IR) occurs when two bodies of different temperatures are in sight of each other. The best example of IR is the heating of the earth by the sun. Dull, rough surfaces absorb the suns rays better than shiny, smooth surfaces. An object in the direct sun light will become hotter than if in the shade.

Conduction

Hot Spot

Cold Spot Heat

Heat Flows From Hot to Cold To Help Equalize Temperatures

Figure 5-1. Example of conduction heat transfer.

Figure 5-2. Example of infrared radiation.

Radiation

Sun

Radiant Energy From The Sun

Heats The Earth


IR in SMT applications works similarly. Fluxes, plastic components, and epoxy glass laminate absorb IR very well. Shiny, reflowed solder will reflect the IR energy away. Solder joints around small packages (such as chip resistors, capacitors, and SOICs) are in sight of the IR energy and heat very well. Solder joints around larger devices (such as PLCCs) are shaded, and do not heat as well. Convection Convection heat transfer occurs when a fluid (such as air, nitrogen, or water) passes over an object (such as an SMT assembly). A cool breeze on a hot sunny day provides convective cooling. Hot air from a hair dryer provides convective heating. Convection heating or cooling requires contact of the flow with the solid part. Only the layer of the flow that is in contact with the part is actually transferring heat. Convection may be classified as natural or forced. Natural convection occurs when no flow is being forced over the object. Temperature differences between the object and the environment create the convective heat transfer. Perhaps the best example is the chimney effect, where a strong convective current rises away from the hot embers to the cooler outdoors.

Forced convection requires an external force that pushes or pulls the flow over the object. A common house fan is a good example.

Typically, forced convection heating or cooling rates are higher than natural con-vection rates. Most reflow ovens today use forced convection as the primary heat transfer mode.

Research International won the 1998 SMT Magazine Vision Award for advances in Soldering Equipment.

Convection

Hot Air

Flow

Cold Object Is Heated By Convection

From The Hot Air Flow

Figure 5-3. Example of convection heat transfer.

Notes:


Heat Transfer Equations Conduction In order to understand what parameters are important in conduction heat transfer, one can consider the variables in the general conduction heat transfer equation. Q = K * A * ( T1 - T2 ) / DX Where: Q = Conduction heat transfer (W ) A = Cross sectional area (cm2 ) K = Thermal conductivity (W/cm -C) T1 = Temperature at point 1 (C) T2 = Temperature at point 2 (C) DX = Thickness of material between points 1 and 2 (cm) The thermal conductivity (K) is a measure of how well the object conducts heat. Insulators (such as epoxy glass) have very low thermal conductivity and permit very low amounts of conductive heat transfer. Conductors (such as copper) have a very high thermal conductivity and permit high amounts of conductive heat transfer. Thus, an SMT assembly with high amounts of copper will heat more uniformly than one without much copper. Conduction through the circuit board during the reflow process improves heating uniformity, as the heat conducts from the hot spots to the cold spots.

t1

t2

x

Surface 1 Surface 2

k = Thermal Conductivity

D

Conductive heat transfer is proportional to the cross sectional area (A) between point 1 and point 2. In SMT reflow, a thicker board will provide greater populated uniformity than a thin board since the cross sectional area is greater. Conductive heat transfer is also dependent upon the distance between point 1 and point 2 (DX). The greater the distance, the less the heat transfer. Thus, a thicker board has a better chance of maintaining a higher temperature difference between the top and bottom side of the board than a thin board. IR - Infrared Infrared (IR) heat transfer occurs when two objects at different temperatures are in sight of each other. The heat is transferred by electromagnetic waves of 0.78 to 1000 micron wavelengths. All objects emit some level of infrared energy. The quantity of infrared energy emitted and the wavelength of the emission are both dependent upon the absolute temperature of the object. As the source temperature increases, the heat transfer output increases exponentially to the fourth power. Increasing the source temperature results in shorter wavelengths. Decreasing the source temperature results in longer wavelengths. In order to understand what parameters are important in infrared heating, one can consider the general equation for heat transfer between the heat source and the object being heated. The purpose is not to memorize the equation, but rather to point out the significance of what is involved.

Figure 5-4. Conduction heat transfer model.


The general equation for infrared heat transfer is:

Q/A = Fv * es * at * s * ( Ts4 - Tt4 ) Where: Q/A = Infrared heat transfer (W/cm2) Fv = Geometric view factor (0 - 1) es = Emmissivity of the source (0 - 1) at = Absorptivity of the target (0 - 1) s = Stefan-Boltzmann Constant (5.67 x 10-12 W/cm2/K4) Ts = Source temperature (K) Tt = Target temperature (K)

The geometric view factor Fv is the fraction of energy that leaves the source that hits the target. In SMT reflow, the oven chamber designs yield very high view factors in the range of 0.90 to 0.95. An important aspect of the view factor comes in product design. If two very large components are in close proximity to each other, the view factor to a solder joint between them is decreased, which makes it more difficult to reflow. The source emissivity (es ) and the target absorptivity factors (at ) are in the range of 0.90 to 0.95 for most SMT applications. Solder paste is an excellent absorber of infrared energy. Shiny gold components may be difficult to heat, as they tend to be reflective. Most often, however, the board material, solder paste, and the components all absorb quite well.

Control of infrared is generally done by controlling the source temperature. Providing that the overall source emission can be regulated, IR energy provides high levels of repeatability. Convection Convection heat transfer occurs when a fluid at a given temperature contacts a solid mass at a different temperature. If the fluid is hotter than the mass, the mass will be heated. If the fluid is cooler than the mass, the mass will be cooled. Perhaps the easiest way to understand convection is to look at the convection equation and note the significance of each variable. The following is the general equation for convection heating or cooling: Q/A = H * (Ta - Tt ) Where: Q/A = Convection heat transfer per unit area (W/cm2) H = Convective film coefficient (W/cm2-C) Ta = Fluid temperature (C) Tt = Target temperature (C) The amount of heat transfer can be modified by either changing the convective coefficient (H) or the temperature difference (Ta - Tt ). Increasing the fluid temperature will increase the temperature difference, and thus the amount of heat transfer will increase proportionally. The convective coefficient (H) can have a significant impact on the heat transfer. The value of H is related to a number of variables, including velocity of the flow and angle of attack of the flow. Increasing the flow velocity will increase the value of H. For example, when one is outside in the wind, a low velocity cool breeze on a warm day will feel quite nice. However, a strong wind on that same day may feel very cold.

IR Source

This Area is Shaded by Adjacent ComponentsResulting in a Lower View Factor Figure 5-5. IR view factor.


That is because the higher velocity wind has a higher convective coefficient, and thus removes heat more effectively from your skin. In SMT reflow applications, increasing the velocity of the flow also improves the amount of convection heat transfer. There are practical limits to this, as too high velocity may cause components to shift around.

The flow direction also has a significant impact on H. Convection heat transfer relies on contact between the flow and the object. Parallel flow can result in a stagnation or boundary layer in which the heat transfer is diminished. Perpendicular flow can break up formation of a boundary layer and enhance the convective heat transfer. Perpendicular flow thus provides higher heat transfer rates than parallel flow.

Heat Transfer Mode Interaction and their Effect on SMT Reflow In SMT reflow applications, there is no such thing as pure convection or pure IR. The only way to achieve a 100% convection system would be to have gas flowing over the product with no chamber walls inside the reflow oven. Obviously the reflow oven is a mechanical device with walls that absorb or emit heat, so 100% convection is a physical impossbility. To have a 100% IR reflow system, the chamber would need to be operated in a vacuum (no air fluid heat transfer). Most reflow processes currently occur in some process gas whether air, nitrogen or vapor phase techniques. However, reflow oven heat transfer designs utilize a greater convection component, while others utilize IR as the primary heat source. For the sake of simplicity, this text will refer to a convection dominate heat transfer machine as a convection machine, and an IR dominated machine as an IR machine. In SMT reflow applications (with the exception of vapor phase), convection, conduction, and IR all play a role in the heating process. Referring to a system as 100% convection or 100% IR is an error in terminology. An oven may be dominated by convection or IR, but the other heat transfer is always present. IR dominant systems usually range in the convection/IR ratio of 60/40 to 40/60. Convection dominant systems have ratios in the range of 70/30 to 90/10.

Parallel Flow

Figure 5-6. Fluid flow patterns.

Air Velocity

(V) Boundary Layer Thickness At Surface, V=0

Perpendicular Flow Air Velocity (V) Surface

Agitation Prevents Boundary

Layer Buildup


Section 6: Reflow Oven Convection Methods

Research International DeltaFlo 10LN Low Nitrogen Convection Oven

Convection Oven Technologies While convection is a relatively simple technology, there are several technical aspects to consider. This section discusses some of the key differences including:

Forced convection versus natural convection

Turbulent flow versus laminar flow

Fresh gas input versus recirculated gas input

Defined convection flow path versus uncontrolled flow path

Methods of forcing convection gas flow Natural convection occurs when the air (or nitrogen) surrounding an object is at a different temperature than the object. The density differences between the hot and cool air causes a physical motion to occur, which assists in heat transfer. Forced convection

occurs when an independent device (such as a fan) forces the air over the product. Typically the convective coefficient is greater with forced convection than with natural convection, which results in improved heating of the product. Early infrared/convection ovens relied on natural convection to help improve heating uniformity. Forced convection ovens usually have better heating uniformity than the infrared/convection ovens. Forced convection can be either laminar or turbulent. Laminar flow is characterized by very stable flow patterns. Convection heat transfer requires contact between the air and the object. (See Figure 6-1.) Thus, a stable gas jet allows more air to come in contact with the object. Because of this contact, laminar forced convection can have heating efficiencies of greater than 30 percent. Turbulent flow is characterized by vortices, eddy currents, and high heating rates. Thermal efficiencies are typically only 10 to 15 percent.


Defined vs Non-Defined Reflow Oven Flow Patterns There are two methods of directing the flow within convection ovens. The defined flow path oven (Figure 6-2) moves the gas from zone to zone. It does not allow the same gas to become trapped in a given zone, which could become excessively contaminated with fluxes. These ovens will typically stay cleaner by preventing flux buildup.

The non-defined flow path ovens (Figure 6-3) do not direct the flow from zone to zone. Gas eventually reaches the exhaust and leaves the oven. These ovens work satisfactorily thermally, however they usually have oven flux cleaning issues associated with high-volume PCB production.

Laminar Flow

Turbulent Flow

Air Jets Remain Stable. Most of the Air Flow Contacts the Object High Efficiencies (25 - 35%) Medium Heat Transfer Rates

Air Velocity (V)

Air Jets are Unstable. Most of the Air Flow Does Not Contact Object Low Heating Efficiencies (5 - 15%) High Heat Transfer Rates

Figure 6-1. Forced convection flow patterns.

Figure 6-3. Example of a non-defined flow pattern reflow process gas is recirculated within each zone. Flux is not removed from the zone and builds-up in the heater cavity.

Fresh Air Input Exhaust Exhaust

3 5 6 7 4 1 2

Figure 6-2. Example of a defined flow pattern process gas is recirculated to the previous zone. Flux is removed via exhaust collectors. Fresh process gas keeps oven clean (free of flux residue). Also see Figure 5-4.

LAMINAR FLOW CONVECTION TO BOARD

PREHEAT DRYOUT ZONES REFLOW COOLING SECTION

FRESH GAS IN REFLOW ZONES


Recirculated Gas: Too Much Recirculation Means Flux Cleaning Air introduced into a heating zone may either be fresh or recirculated. Fresh gas input requires heating the air from ambient to the zone temperature. In comparision, recirculated gas input requires only heating the air a small amount, since the gas is already hot . Thus, an advantage of recirculated flow is a lower power requirement to heat the zone.

Recirculated flow has the disadvantage of picking up contamination each time the air is re-used. No-clean fluxes prevent solids from building up on the PCB. These solids end up in the reflow process gas and will collect in the heater cavity if the oven is not designed properly. The result is a dirty process atmosphere or machine downtime due to the need for oven flux cleaning.

31 2 4 5 6

How is the reflow process gasrecirculated to the previous zones?

Highest pressure hereforces recirculated gas to

be constantly purged.

Fresh input gas into every bottom zone insures clean processatmosphere and versatile bottom/top temperature control.

Heater configuration continuouslycleans oven convection flow jets.

Figure 6-4. Positive pressure in the reflow zone purges flux laden gas to the exhausts. This design feature creates the defined flow pattern within the reflow oven.

Notes:


Ways to Generate Forced Convection in Reflow Ovens There are three methods of generating forced convection in reflow ovens fans, compressors and flow amplifiers. Fans Fans are an inexpensive and reliable method of moving high volumes of air. Fans typically have low pressure generating capability, so exit velocities may be limited. However for most board profiles this should not be a problem. Most convection reflow ovens use fan forced convection.

Convection fan in a preheat zone. Note that the fan motor is outside the heater chamber for improved fan reliability.

Compressed Gas Sources Compressed input gas sources (such as compressed air or nitrogen) are also used to create forced convection. The pressure must be regulated in order to achieve constant flow. Compressed sources have the advantages of providing a controlled atmosphere, which is free of particulate contaminants. If compressed air is used, the humidity can be controlled using an inline dryer. Flow velocity is easily controlled with flow

meters. Compressed sources have high pressures, allowing high velocities if desired. An air compressor or compressed nitrogen supply is required. Compressed gas reflow is used when the application requires an extremely accurate thermal profile with filtered process gas. Flow Amplifiers The third common method used to create convective flow is the flow amplifier. The flow amplifier uses a compressed source to induce flow. In a flow amplifier, compressed gas flows through the inlet into an annular chamber. The gas is then throttled through a restriction at high velocity. The compressed gas flow towards the outlet adhering to the outside wall. A low pressure area is created in the center, which induces flow.

The advantage of a flow amplifier is that flow volumes are similar to a fan without any of the fan life issues. However, a major drawback to the flow amplifier is that the flow amplifiers tend to clog with fluxes. The ring nozzle causes the gas to expand and cool. The cooldown of the gases causes flux to precipitate out of the flow. There is also a thermal efficiency loss with the gas cooldown.

Figure 6-5. Schematic of a flow amplifier.

Inlet Gas(Compressed)

EntrainedAir

ExitFlow

Figure 5-1. Forced convection flow patterns.

Flow is restricted with small ring nozzle causing cold spot.

Compressed gas flow through annulus


Section 7: Choosing Your Reflow Oven

What Size Oven do I Need? The basic requirement of any reflow oven is to be able to consistently reflow the product at the required speed to meet production needs. Selecting a reflow oven requires that the profile process band has been defined and the oven can successfully heat within the band. Production throughput rates can be met providing the oven has been sized correctly.

Size The Oven Properly The oven must be sized to handle the production rate of the pick-and-place equipment. The calculation is shown below: Line speed (minimum) =

Boards per minute x Length per board Load Factor

(See Figure 3-5 for description of Load Factor.) Operating the oven conveyor speed greater than the line speed ensures that no product jams will occur on the onload area of the reflow oven.


In order for the oven to be properly sized, the process speed defined below must be greater than the minimum calculated line speed. This will ensure that bottlenecks will not occur at the reflow oven. The process speed should be determined through testing or by the following equation: Process speed =

Oven chamber heated length Process dwell time

Example of Oven Sizing Calculation An SMT assembler wants to produce 8-inch boards at a rate of 180 per hour. The solder paste manufacturer recommends a 4 minute, three step profile. How long an oven do I need to process boards at this throughput? Boards per minute = 3 (180/hour) Length per board = 8 inches Load Factor = 90% (2-inch space between boards) Process Dwell Time = 4 minutes Calculate Line Speed: (3 boards/min) x (8 inches/board) 0.9 Line speed = 26.67 or 27 inches/minute Therefore, the reflow oven must have a process speed of at least 27 inches per minute. Determine oven chamber heated length with process speed equation: 27 in/min = Oven chamber heated length 4 minutes Oven heated length = 108 inches (9 feet)

To meet process requirements of this example, the reflow oven needs at least 9 feet of heated length so that an 8-inch board can be processed at 27 inches/min conveyor speed for a 4-minute profile. Note that the overall length of the oven will exceed 9 feet including the cooling section and conveyor loading sections. The calculation is for HEATED LENGTH NOT OVERALL OVEN LENGTH needed.

Comparing Reflow Ovens When selecting reflow equipment, it is important to take into consideration all aspects that may affect the process and production. Try to avoid becoming too focussed on one issue, such as populated uniformity, as it may not lead you to the optimum decision. The recommended steps in the selection process are described in the following paragraphs. Thermal Performance Oven thermal performance is defined by the sum of the temperature non-uniformities caused by mass differential, oven uniformity, and oven repeatability under production loading conditions. A set of tests must be conducted with the desired product on the oven in question.


Throughput Oven throughput (in boards per unit time) is defined by the following equation: Throughput =

Conveyor Speed x Load Factor Length per board Note that in the previous example, the throughput in boards per minute was a given. Process Gas If a nitrogen atmosphere is required for solderability or no-clean processing, one needs to consider the purity required and the gas consumption. Be careful how the purity is measured, such that it is not measured at the gas inlet. Measure purity in an area representative of what the board will see on the conveyor, not the nitrogen inlet in the oven. Purity in reflow is the most critical area since solder is liquidus. Conveyor Design One needs to define whether a mesh conveyor or an edge-type conveyor is required. Typically edge conveyors are used for in-line processing or if double-sided assembly is required. The most important issue with edge conveyors is the ability for the conveyor to remain straight to avoid dropping boards. Find out how parallel the edge conveyors will be during heating conditions avoid simply taking measurements in a cold oven only.

Research Internationals patented tensioned rail stays parallel over any temperature range. Software The best technically designed oven in the world may be no better than the worst if the software is difficult to operate. When comparing software, keep the user in mind. Consider file access and edit security. Ask questions about networking machines to a host.

Research International StarWin 3.0 Software


Reliability/Serviceability/ Documentation Once a machine is placed into service, it is critical that it remains operational. Evaluate the product design from a structural as well as a functional standpoint. Is the oven designed to withstand the rigors of production? Review the product manuals, drawings, and spare parts lists. Find out where the nearest servicing center is located. Is training available? Maintenance Downtime Maintenance downtime is perhaps the most overlooked aspect of equipment selection. The key to meeting throughput requirements is a continuously operating reflow oven. Flux cleaning downtime and frequent scheduled maintenance will reduce the number of hours available to solder boards. Compare maintenance requirements carefully before selecting an oven.


Section 8: Oven Price vs Cost of Ownership

When multiple oven vendors pass the previous technical selection criteria listed, then a proper cost evaluation should be used to compare ovens. This is especially important when comparing ovens that utilize different heating technologies, as initial price and operating costs can be significantly different. Often, the initial cost of the equipment is the only cost considered when comparing equipment. It is certainly important when companies have tight budgets, BUT IT IS OFTEN FAR LESS SIGNIFICANT THAT THE OPERATING COSTS. To make an informed comparison, all costs should be analyzed with a common basis based on an annual cost of ownership approach. As an example, compare two ovens with a vast difference in sell price (Figure 8-2).

The expected oven life is 5 years, with a discount (interest) rate of 10%. Assume the oven is fully depreciated at the end of 5 years with no residual value (worth $0 the worst case scenario). The cost of ownership should reflect all the operational and opportunity costs of running the oven. Direct costs like electrical power usage and process gas requirements (such as nitrogen) are included. Indirect or lost opportunity costs related to lost production caused by maintenance downtime. Flux cleaning and chain lubrication are examples of machine downtime that will reduce board production. Figure 8-1 gives a real world example of these costs. When included in the Cost of Ownership model (Figure 8-1), it shows that when the impact of major cost of ownership factors is analyzed, the cheaper oven actually costs TENS OF THOUSANDS OF DOLLARS MORE OVER JUST ONE YEAR OF OPERATION! The next 2 pages describe how for a typical high-volume SMT line:

Oven purchase price differences of 20-25% are INSIGNIFICANT when

compared to increased cost of ownership

Advanced ovens can payback 20-25% price premiums vs basic ovens IN THE FIRST 6 WEEKS of operation, based on reduced maintenance and downtime (opportunity) costs, even with a 10% cost of money factored in

Advanced ovens improve profitability by NEARLY $170,000 PER YEAR based on the Real World example described by Figure 6-1 and Figure 6-2

Reflow oven downtime is the MOST SIGNIFICANT cost factor for high-volume machine evaluation


Real World Case Study: Flux Cleaning Creates Regularly Scheduled Bottlenecks A large contract manufacturer operates four SMT assembly lines in a plant that runs approximately 22 hours per day except for changeover and maintenance. The maintenance technician describes the reflow oven flux cleaning maintenance procedure:

Every three weeks, two technicians disassemble each reflow oven for flux cleaning. Reusable c ondensers collect flux in the cooling zones, and air jets require cleaning for the flux not collected by the heat exchanger. Newly cleaned flux condensers replace soiled units in the oven so that the latter s fins can be degreased off-line. The cleaning process requires 10 hours/month for EACH REFLOW OVEN.

Based on its cost structure, the plants per line annual maintenance expense for reflow oven is:

Labor: 2 people @ $40/hour (burdened) * (10 hours/month oven cleaning) * (12 months/year) = $9600/year Flux Condenser Defluxing Cost: $100/month covering degreasing solvent, using a commercially available degreaser (not degreased by hand scrubbing) = $1200/year Direct Maintenance Cost per Line = $10,800 excluding the cost of repl acement flux condensers LOST BOARD THROUGHPUT MEANS LOST PROFIT The cost is in lost production, however, is many times greater. For continuous board production, reflow oven maintenance creates regularly scheduled bottlenecks, i.e. the capacity of the plant is equal to the capacity of the bottlenecks wherein each hour of reflow oven flux cleaning is a production hour lost for the entire plant (Goldratt, The Goal, p. 157). For example, a typical production line might include:

Board Sell Price: $100 Board Cycle Time: 0.3 minutes (180 boards per hour) Board Bill of Material Cost (PCB and components): $90 Labor Cost per board (direct and overhead): $1

SOURCE: Lasky, Baldwin, Throughput: The Critical Cost Variable in DCA Assembly , SMTA National Symposium, November 1996, p. 1.

Applying 10 maintenance hours per month imposes the following reductions in board production throughput (and profitability) per assembly line: Board Production Rate * Maintenance Downtime = Lost Throughput per Line (180 boards/hour) * (10 hours/month * 12 months/year) = 21,600 boards/year lost per line Lost Line Throughput * (Board Sell Price - Board Cost [Material & Labor]) = Lost Line Profit (21,600 boards lost/year) * ($100 - 91) = $194,400 ANNUAL LOST PROFIT PER LINE

Figure 8-1. Maintenance cost scenario.


Reflow Oven Cost of Ownership Model Assumptions (Figure 6-1) Number of assembly hours/year (2 shifts)

4160 4160

Board Cycle Time 180 boards/hour 180 boards/hour Oven A (Advanced) Oven B (Average) Purchase Price $95,000 $75,000

Capital Recovery Factor 10% Interest for 5 Years (A/P, 10, 5)

0.2638 0.2638

Annual Oven Cost $ 25,061 $ 19,785

Direct Costs Nitrogen Gas Consumption (cubic meters/hour) Nitrogen Cost

30

$0.19/cubic meter

54

$0.19/cubic meter Annual Nitrogen Cost $ 23,712 $ 42,682 Power Consumption (KVA) Electrical Power Cost

43 $0.16 KVA/hour

38 $0.16 KVA/hour

Annual Power Cost $ 28,442 $ 25,292 Direct Maintenance Costs (See Figure 6-1 for calculations)

Flux Cleaning Maintenance (hours/month)

1 8

General Maintenance (hours/month)

2 2

Annual Direct Maint. Costs (2 people @ $40 incl. Overhead)

$ 1,440 $ 4,800

ANNUAL DIRECT COST OF OWNERSHIP (Oven + Nitrogen + Power + Maintenance)

$ 78,655

$ 92,559

Opportunity Costs (See Figure 6-1 for calculations)

Board Profitability ($100 board sell price @ 9% Gross Margin)

$ 9/board $ 9/board

Annual Production Lost ($9/board x 180 bds/hour x Maint. Hours)

$ 38,880 $ 194,400 (From Figure 6-1)

TOTAL ANNUAL COST OF OWNERSHIP (Direct cost + Opportunity Cost)

$ 117,535

$ 286,959

Figure 8-2. Advanced reflow ovens cost more, but improve profitability by more than $847,000 over the five-year life of the oven.


Section 9: Vertical Reflow Multitasking PCB Production

Vertical reflow is designed for SMT assemblers who: Need to fit more high-volume lines into

an existing manufacturing plant Need to process jobs of varying board

types and profiles Need to convert SMT lines

IMMEDIATELY to a new board type Current Reflow Approach A typical conveyor oven reflows a board by moving it laterally through a series of horizontal temperature zones to create the desired thermal profile. Proper reflow

requires a smooth mechanical motion, many non-fluctuating temperature zones, exacting timing for each zone, and a significant cooling area. Modular vertical stationary reflow (MVSR) oven operation is quite different. How Vertical Reflow Works MVSR systems like TOWERTM by Research International use individual chambers, vertically stacked, to perform the entire thermal profile without moving the board. These chambers are self-contained and computer controlled mini-ovens and each provides the necessary heating and cooling required for all reflow operations. Profiles are generated by controlling the power of a single heater instead of moving the board through a series of heated zones at different temperatures.

Each one of these chambers performs the entire thermal profile.

Circuit boards entering the TOWERTM are placed on an elevator and are moved to an available chamber for reflow processing. Those boards that have already been


processed are removed from the chambers by the same elevator and then passed on to the in-line conveyor.

The elevator feeds boards into each chamber. Boards are in motion only when they are in a cool (solder is non-liquidus) state.

Since the board is stationary during reflow , energy can be focused to high component mass areas (like shielded connectors) and shielded from low mass areas (populated by capacitors, for example). In this way, each component on the board is brought to reflow temperature in a single step and temperature gradients are minimized for even the most difficult boards. These boards can potentially be reflowed entirely instead of having subsequent wave or hand soldering processes post-reflow. For typical surface mount boards, standard convection profiles are used without focusing heat to high-mass board regions. TOWERTM is installed in a standard production line just like an in-line conveyorized reflow oven. Mechanical and electrical interfaces to periphery equipment are the same for both MVSR and conveyor ovens. A comparison of the conventional reflow versus MVSR is shown in Figure 9-1.

Traditional Method

TOWERTM/MVSR Approach

Board passes through multi-zone oven to reflow solder and cool board

Convection minimizes Delta T caused by mass differences

Set point temperature, air flow, and belt speed are typical profile variables

User inputs zone setpoint temperatures to create recipe

Desired profile verified through trial and error Underboard support runs through one plane Large components on underside not possible due

to conveyor vibration Inert atmospheres created by purging oxygen

Performance determined by initial capital

investment (buy large and nitrogen version) In-line oven operating costs incurred regardless of

boards produced

Board is stationary during reflow and cooling cycles Convection heat input optimized for different regions of

board (optional) Set point temperature and user defined time duration

controls profile User determines time/temperature profile and oven

automatically adjusts to create recipe Oven adjusts heat input automatically for test board Stationary board allows multi-point support Underside components do not vibrate since board is still Inert atmospheres can be created by purging oxygen OR

creating a vacuum Buy the capabilities you need today and expand/retrofit

as customers needs change Operating costs (nitrogen use and power consumption)

proportional to board production levels

Figure 9-1. Traditional (in-line) reflow vs vertical reflow.


Vertical Oven Throughput Board production rates are no longer tied to the heated length of the reflow oven, but rather to the number of chambers utilized for the specific production run. More chambers mean higher production volume. For example, a typical 4-minute board cycle consisting of a 3-minute heating profile, 30 seconds of in-chamber cooling (liquidus timeover) and 30 seconds of elevator transport/secondary board cooling means an individual chamber produces 15 boards per hour (assuming the chamber is loaded with only one board per cycle). If six chambers are operational for this production run, the production rate is 6 chambers times 15 board/hour/chamber = 90 boards per hour. If higher volume is needed, activating 8 chambers produces 120 boards per hour. Compared to current reflow ovens, 120 boards/hour is the rate of a 10 foot heated length in-line oven, running at 40 inches (102 cm) per minute conveyor speed for an

18-inch (46 cm) long board, at a 90% load factor (2-inch [5 cm] space between boards). Besides a smaller footprint for equivalent assembly volume, the impact of this MVSR throughput on board production is the ability to multi-task PCB assembly.

Multitasking Board Production

Unlike dual-lane conveyors for in-line ovens which can only process 2 types of thermally similar boards, MVSR ovens can reflow thermally unique board types in each chamber. Since each chamber is like a lane of a dual-lane conveyor, MVSR provides the equivalent of 8 lanes each with an independent profile and board size (length). Multiple pick-and-place lines dedicated to each board type can be routed to a centralized MVSR oven (Figure 9-2) instead of having separate reflow ovens throughout the plant.

Chamber 5

Chamber 6

Chamber 7

Chamber 8

Chamber 4

Chamber 3

Chamber 1

Chamber 2

TOWER Reflow Oven

Dedicatedto

Board ABoard

A

BoardC

BoardB

StencilPrinter Placement



Production FlexibilityMultiple Lines

Dedicatedto

Board B

Dedicatedto

Board C

LOW VOLUME LINECOMBINATIONS

BOX BUILD ASSY

MAINTENANCEWORK-AROUNDS

Figure 9-2. Multiple SMT production lines can be routed into a single TOWERTM.


Section 10: Semiconductor Reflow

Most SMT components use wire leads, which extend from the edge of the package. A new generation of chips relies on solder ball leads instead of wire leads. Common names for these semiconductor chips are ball grid arrays (BGAs) and chip scale packages (CSPs). Reflow processes are used to attach solder balls to BGAs. Semiconductor reflow has much the same issues as SMT reflow. The approach for profiling, determining inert atmosphere purity requirements, cleaning post-reflow flux residues, etc. are essentially common to both types of products.

Solder Ball Attachment Process Solder ball attachment consists of three primary processes. Solder ball placement Nitrogen convection reflow Aqueous cleaning Solder Ball Placement The solder ball placement process typically begins by screen printing a layer of solder paste onto the chip substrate. The paste provides adhesion sites for the next step - placing preformed solder spheres on the

chip. These solder balls provide the package-to-PCB connection in a CSP. Common ball placement processing techniques are gravity transfer and vacuum transfer. Solder ball inspection is often the final step in the placement process. Most automated ball placement machines accomplish all of these steps in a single platform.

A solder ball attachment line: (L-R) Placement machine, reflow oven, cleaner.

Nitrogen Convection Reflow CSP ball attach reflow is usually performed in a nitrogen atmospheres with purity levels from 10-25pp

Documents

reflow instructions19397719 Reflow Manual