Selective Soldering Design Guide
Selective soldering solders one joint at a time with a small fountain of molten alloy. It exists because a modern board mixes fine pitch surface mount parts with a handful of through hole connectors, and the wave process that would once have soldered those connectors would also damage everything else on the board. The process is slow and precise, and the layout has to give the nozzle room to work.
What Selective Soldering Is
A selective machine pumps solder through a nozzle that is a few millimetres across and moves the nozzle under the board, or moves the board over the nozzle. The joint is wetted in a fraction of a second and the nozzle moves on.
The rest of the board is protected by the fact that the solder is applied locally, so the surface mount parts on the same side are never immersed. That is the whole advantage of the process over a wave.
The alloy is the same tin silver copper that a reflow process uses, and the nozzle holds a small volume of it at temperature throughout the shift. The composition drifts as the alloy is used, so the pot is analysed periodically and topped up with fresh metal.
The machine usually includes a fluxing station and a preheat station, and the three operations are programmed as one cycle. The cycle time per joint sets the cost, and a board with many through hole pins is expensive to run.
The process is most often used for connectors, for press fit hardware that has to be reinforced, and for parts that cannot survive a reflow or a wave because of their temperature limit.
Nozzle Types and Wave Shape
The nozzle comes in a round version for a single pin and in a rectangular version for a row of pins. The rectangular nozzle solders a whole connector in one dip, which is much faster than a pin by pin cycle.
The shape of the wave at the nozzle is critical. A nozzle with a flat top gives a stable contact area, while one that is worn or damaged gives a wave that is too tall and floods the board or too short and fails to reach the joint.
The nozzle diameter has to be larger than the pad and the surrounding mask, because the solder has to surround the pin and form a fillet on the solder side. A nozzle that is too small leaves a joint that is filled on one side only.
The nozzle also has to fit between neighbouring components. A row of connectors placed close together leaves no room for a rectangular nozzle, and the machine has to fall back to a round nozzle and a much longer cycle.
Layout for Nozzle Access
The layout has to leave a path for the nozzle to reach every joint from the solder side. Components on the solder side are the usual obstruction, so the space under a selected joint should be kept clear.
The clearance is measured from the centre of the pin to the nearest obstruction, and it has to include the nozzle body and not only the wave. A nozzle is a physical object with a holder, and the holder is often what collides first.
The shape of the board also matters, because a long thin board can flex as the nozzle rises to meet it. A support pin or a small pallet is used under the board so that the joint is presented to the nozzle at a consistent height.
A common rule is to keep a clear area of several millimetres around a selective joint, and to group the selected pins together so that one rectangular nozzle can serve them all. Grouping also reduces the cycle time.
Where the pins cannot be grouped, the layout should place the selected joints on a grid so that the machine can plan an efficient path. A random arrangement of selected pins is slow to program and slow to run.
Thermal Mass and Preheat
Selective soldering needs more preheat than a wave process, because the heat has to come from the nozzle in the fraction of a second that it is in contact. A board that is cold will not wet.
The thermal mass of the joint is the variable that the layout can influence. A pad connected directly to a plane draws heat away from the joint, so a thermal relief is used on the plane, and the relief is a design feature that makes the process possible.
The bottom side of the board is usually preheated with an infrared lamp or with a hot air stream, and the temperature is set below the melting point of the alloy. The preheat also drives off moisture, which reduces the spatter from the flux.
A board with a large copper area on the solder side needs a longer preheat and a longer contact time, and the designer should flag such areas so that the process engineer knows about them before the first article.
Flux and Residue
The flux has to be applied precisely, because it is not washed off the whole board afterwards. A no clean flux leaves residue that must be harmless, while a water soluble flux has to be washed and the board must allow the water to drain.
The flux is applied by spray or by a drop jet that follows the selected joints. The amount has to be enough to activate the surfaces and small enough to leave no pooling, since a puddle of flux under a connector is a residue problem that appears only after the product has been in the field.
The flux choice also interacts with the surface finish. A board with a HASL finish presents a rough surface that holds more flux, while an ENIG finish is smooth and easier to clean.
The residue question is one of the main reasons to choose selective soldering over a wave, because only the selected areas see the flux. The advantage is lost if the flux is applied over the whole board.
Design Review
The review should list the joints that will be selectively soldered and check each one for nozzle access, thermal relief and spacing. A joint that cannot be reached is a layout error that only appears when the machine is programmed.
The second item is the grouping, since a row of pins that can be soldered with one rectangular nozzle is far cheaper than the same pins soldered one at a time.
The third item is the flux and the cleaning requirement, which has to be settled before the first article. The solder defects that selective soldering produces are mostly incomplete fill and residue, and both are influenced by the layout.
The last item is the first article, where the fill is checked on a sectioned sample and the residue is checked under a microscope. The quality check on a selectively soldered board concentrates on those two things.
FAQ
Is selective soldering slower than a wave? Much slower per joint, but it avoids the pallet and the damage to surface mount parts. The comparison should be made on the whole assembly cost rather than on the joint alone.
What is the main layout rule? Keep a clear path to each selected joint from the solder side. A nozzle has a body and a holder, and both need room.
Can selective soldering be used for a fine pitch connector? It can, but a rectangular nozzle is needed for a row and the clearance between the pins has to allow it. Above a certain density the process becomes impractical.
