|The Asscon VP2000 in-line vapor-phase soldering system.
Vapor phase reflow systems have had a curious history. They were the first reflow systems, back in the 1960s, before surface-mount technology and infrared reflow were introduced. When infrared and convection infrared reflow ovens came along, vapor phase systems faded away, but not quite entirely, because they had yield advantages and an ability to reflow difficult panels when compared to infrared and convection systems. Finally, five years ago, when Pb-free solder became mandatory in many applications, vapor phase systems blossomed again.
The basic functioning of a vapor phase system is straightforward: a fluid is selected that vaporizes at a specific temperature — a temperature 5 to 10 degrees above the temperature at which reflow will occur for a given solder. The fluid is heated to its vapor state, and the board is introduced. As the vapor rises, the fluid droplets condense onto the board, much as water vapor condenses onto a glass holding an ice-cold liquid. Condensation raises the temperature of the board to the temperature of the fluid droplets, and the solder reflows at the required temperature.
Vapor phase systems tend to cost more initially than reflow ovens, but also tend to have a lower cost of ownership over time. Allen Duck, president of A-Tek, a Colorado firm that distributes vapor phase systems for the German manufacturer Asscon, explains that a 10-zone in-line convection reflow system might cost anywhere from $35,000 to $160,000. The average is probably around $100,000. An in-line vapor phase system with equivalent productivity might cost $175,000. But vapor phase systems use less than half the power, are more flexible in terms of changeover, and operate in an inert environment without the need for any third-party material such as nitrogen.
Vapor phase systems are commonly used in applications where high yield and high long-term reliability are important. They are widely used by makers of high-end servers, medical electronics equipment, and military/aerospace electronics. They are much less common at companies that are assembling lower-cost consumer items that are expected to have short life spans. Convection is adequate and well established, for such applications; the electronics being assembled are not extreme in value or complexity and they often have a limited life expectancy, so there is no need to change the reflow process. High value, extremes of mass or life expectancy and Pb-free present different scenarios and in these situations vapor phase is most appropriate.
There is no separate pre-heat zone in an Asscon vapor phase system. The boards move directly into the chamber, at the bottom of which is the fluid that will be heated to create the vapor. The chamber itself is not pressurized; it remains at atmospheric pressure, but as a consequence of displacement contains only the fluid's vapor — no atmospheric air, no oxygen, and no inert gas such as nitrogen. The vapor itself is inert.
|Inside the VP2000.
The ramp-up rate therefore depends only on one thing: the amount of vapor available for condensation onto the boards. The amount of vapor created is controlled by managing the external heat source attached to the tank. Low-mass high-response heaters clamped to the outside wall of the tank allow for precise thermal control and vapor production. Condensation of the vapor gradually raises the temperature of the boards. The differing masses of the components, their different colors and the juxtaposition of tall components to short components are unimportant because no part of the board can exceed the temperature of the vapor. There are no hot spots and there are no defects caused by uneven heating. The condensed vapor will get under array packages by capillary action, and so provide a direct thermal link to solder connections ensuring equal heating of every single solder connection on the circuit board. The result: there are no cold or brittle joints, often found under the package body in convection reflow of Pb-free solders.
The board itself is held at a single elevation. Allen Duck explains, "That elevation is the point at which condensation occurs because you have a temperature mismatch between the vapor and the circuit board.
"As vapor rises, it reaches the circuit board and it condenses. When that circuit board reaches the same temperature as the vapor, you have a temperature stability situation and condensation ceases." Since no more condensation is occurring, vapor now begins to rise above the level of the circuit board. Above the circuit boards, the vapor contacts temperature sensors that detect the arrival of the vapor. "As soon as they see that temperature rise, they know that the circuit board has hit peak temperature and liquidus," Allen Duck notes. "And at that point there are software parameters that start to count down the period of liquidus that you desire. At the end of that clock, then the circuit board is drawn back up into a cooling zone, and cooling takes place." The system as a whole provides a new level of precision in managing reflow parameters and results. The overall simplicity of the process means that vapor phase systems are easy to profile. An infrared reflow system might require multiple runs to work out the right pre-heat time, soak and reflow time. In a vapor phase system ramp-up depends on the quantity of fluid that is vaporized. "Somebody who is familiar with vapor phase could literally look at a circuit board, look at the paste manufacturer's profile recommendation and say, OK, I need 2.1°C per sec ramp rate, 65 seconds above liquidus. They could literally profile that machine just from experience — it's that predictable. If you're looking at a board that weighs 1kg and another that's 2kg, you have a ratio of 2x the mass, you're going to be looking at a ratio of temperature input, which is not quite linear, but it's not far off. And then you can program the machine accordingly."
Extreme Size Boards
One area in which vapor phase systems have an unbeatable advantage is in reflowing boards of extreme sizes. How do you reflow a board that is 3 feet long and 2 feet wide (0.91 x 0.609m)? A maker of convection reflow ovens once tried to design a system that would handle these giant boards. You can imagine the complexities involved in trying to achieve anything like thermal stability across such a large area populated by hundreds of components of different sizes and heights and colors. "It was an exercise in futility," Duck observes. "You simply cannot get a convection process to operate across a planar surface that is that big without impacting the way in which airflow conducts itself. And you create any number of eddy currents, any number of pressure points and cold/hot spots." As a result of the close temperature control provided by the vapor, the process of reflow can actually be observed as it occurs. Asscon systems have windows in the chambers. The transition to liquidus commences at the edge of a board and proceeds to the center in very symmetrical fashion. On a board that measures 18 x 18-in. (0.457 x 0.457mm), the whole process might take half a second, Duck says. On a truly large board — 3 x 2 ft — the reflow process might take a second to a second and a half.
A solder that has been reflowed in a vapor phase system may also possess long-term resistance to the formation of tin whiskers. Tin whiskers were rare in the days of SnPb solder. The introduction of Pb-free solders raised the fear that tin whisker failures might become widespread.
Keith Howell, technical director of the solder manufacturer Nihon Superior, suspects that part of the reason for the relative lack of tin whisker-related failures is the short life of many of the products involved. Tin whiskers take years to appear in Pb-free solders. "With SnPb," he explains, "you didn't get tin whiskers because the alloy will relax itself, whereas Pb-free (solder) doesn't have that ability. That's why you can get the stresses built up and you can get the whiskers to grow."
A tin whisker may begin with corrosion from flux residue in a joint. Corrosion puts compressive stress on the metal, as a study by Nihon Superior has shown. If the metal is SnPb, it simply creeps to relieve the stress. Pb-free solder, however, is more likely to sprout a tin whisker. "One of the ways for metal to get out of the way is by atoms migrating to the surface and piling on top of each other to form a low-stress extension of the crystal, which is a whisker," Howell notes.
The oxygen-free environment of a vapor phase system should make it harder for corrosion to occur. But absolute proof is hard to come by. Tin whiskers take a very long time to form and even then are notoriously hard to study. Keith Howell's conclusion from the available evidence: a halogen-free flux used in the inert atmosphere of a vapor phase system will make it very hard for tin whiskers to form.
Contact: A-Tek LLC, 5213 Longs Peak Road, Suite G, Berthoud, CO 80513 970-532-5100 fax: 970-532-5101 Web: http://www.atekllc.com