Reflow Profile Measurement and Control in SMT Production

The reflow profile is the recipe that takes a printed and populated board from room temperature to a molten joint and back again, and it is the last opportunity in the process to influence the quality of the solder. Everything before it has fixed where the alloy is and how much of it there is; the profile decides what happens to that alloy, to the flux and to the components around it. It is also the setting most likely to be inherited from a similar product without being measured, which is why a profile that has never been verified on the actual assembly is a common hidden cause of intermittent defects.

What the Profile Has to Achieve

The profile has four jobs. It has to bring the whole assembly to a temperature at which the flux activates and the oxide is removed, without raising it so quickly that a small component is thermally shocked. It has to give the flux enough time at the activation temperature to do its work. It has to bring every joint above the liquidus temperature of the alloy for long enough to form, and it has to cool the board at a rate that produces a fine grain structure without stressing the components.

The conflict built into those four jobs is that the board is not thermally uniform. A large thermal mass, such as a connector body or a heavy copper plane, heats slowly, while a small component on a thin section heats quickly. A profile that satisfies the heavy area will overheat the light one, and a profile that satisfies the light one will leave the heavy area cold. The measurement exists to reveal that spread rather than to confirm an assumption about it.

Measuring Rather Than Assuming

A profile is measured with thermocouples attached to a sacrificial board of the same design, loaded with the same components and arranged in the same position on the conveyor as the production panels. The attachment matters: a thermocouple held on with a high temperature tape measures the tape as much as the joint, and one placed beside the joint rather than on it reports a temperature that no solder will ever experience.

The standard practice is to measure several points at once. The solder joint on a fine pitch device, the joint on a large thermal mass, the surface of the board in a bare area and the body of a temperature sensitive component are the four useful positions, because together they describe the range the assembly passes through. A profile that is verified at one point only is a profile that is known to work at that point and unknown everywhere else.

The measurement is repeated whenever something that affects the thermal behaviour changes. A different panel array, a heavier copper weight, a change of component type, a change in the conveyor speed or a modification to the oven all move the result, and a profile that was valid for the previous configuration is a starting point rather than an answer.

The Stages and What Each Is For

The preheat stage raises the assembly to the temperature at which the flux begins to work, at a rate slow enough to avoid a thermal shock. The soak stage holds the board in that band long enough for the flux to reduce the oxides and for the temperature to equalise across components of different mass. The time spent here is a compromise between activation and evaporation: too little leaves oxides on the pad, and too much consumes the flux before the alloy melts.

The reflow stage raises the assembly above the liquidus and holds it there for a defined period. The peak temperature has to be high enough to wet the surfaces reliably and low enough to stay inside the rating of every component, and the time above liquidus has to be long enough to form a joint and short enough to limit the growth of intermetallic layers. The cooling stage then brings the board down at a controlled rate, since a rapid quench stresses the components and a very slow one coarsens the joint structure.

Where the Defects Come From

A tombstone on a small passive component is usually a profile problem. One termination reaches the melting point before the other, and the surface tension of the first molten joint lifts the component. The cause can be an unequal thermal path through the pads, an uneven distribution of copper between the two ends or a ramp that is too fast for the assembly to equalise. The correction is in the profile or in the thermal symmetry of the footprint, not in the placement machine.

A cold joint, recognised by a rough and dull surface, points to insufficient peak temperature, insufficient time above liquidus or a pad that was starved of paste. A joint with excessive voiding under a large thermal pad points to flux trapped beneath the device, which is often a soak that was too short or a paste deposit that was too large. Solder beading on the surface around a component frequently follows a ramp that was too fast, causing the paste to spatter before the flux has had time to hold it together.

Keeping the Profile Under Control

The profile is treated as a controlled document attached to the product. It is stored with the product data, the measurement record is kept with the batch and the oven is checked on a schedule so that a drifting heater or a failing thermocouple is found before it produces a batch of marginal joints. Where a product runs on more than one line, the profile is confirmed on each of them, because two ovens that carry the same model number do not necessarily transfer heat in the same way.

For a customer, the practical consequences are straightforward to state. A new product needs its profile measured rather than inherited. A change to the panel, the material or the component mix is a reason to measure again. And a defect that appears suddenly on a product that has been running well is worth checking against the oven before the design is blamed. Our SMT assembly lines keep the profiles with the product record, rapid PCBA prototyping is where a new assembly is profiled, and the measurements are held under quality management beside the PCBA testing results.

Documenting the thermal profile

A measurement that is not recorded is a measurement that will have to be taken again. The thermal profile belongs with the product data, alongside the placement programme and the stencil identification, and it should be readable by someone who was not present when it was taken. The useful record contains the oven identification, the conveyor speed, the zone set points, the measured curve at each thermocouple position, the alloy and paste specification and the date of the measurement.

The record also states what the profile is valid for. A profile measured on a panel with six circuits is not automatically valid for a panel with eight, and one measured with the components of the prototype is not automatically valid for the production bill of materials. Stating the configuration turns the document from a historical note into an instruction, and it is what allows the next engineer to decide whether the existing profile can be used or whether it has to be measured again.

Version control matters as much here as it does for the artwork. When the profile is revised, the previous revision is retired rather than left in circulation, because two profiles for one product on one line is a condition in which neither can be called the process. The measurement record from the batch then references the revision that was used, which makes it possible to establish what changed if a defect appears.

Oven Maintenance and Profile Drift

An oven is a machine with heaters, thermocouples, blowers and a conveyor, and all four drift. A heater that has lost output makes one zone run cooler; a fan that has slowed reduces the heat transfer; a conveyor that runs slightly fast shortens every stage of the profile. None of those produces a dramatic failure. They produce a gradual movement of the profile towards the edge of its window, and the first symptom is an increase in defects that seem to have no common cause.

The protection is a maintenance schedule and a periodic verification of the profile on a production board rather than only at the time of the initial measurement. Where the verification shows a shift, the oven is investigated before the profile is adjusted, since adjusting the set points to compensate for a failing heater is a way of hiding the fault rather than correcting it.

The same logic applies when a product moves between lines. Two ovens of the same model do not necessarily transfer heat identically, because their airflow, their loading and their maintenance history differ. A profile verified on one line is a starting point for the other and is measured again before it is used for production.

FAQ

Why measure at several points on the board? Because the assembly is not thermally uniform, and a profile verified at one point says nothing about the temperature at the others.

Can a profile be copied from a similar product? As a starting point, but it has to be measured on the actual assembly before it is used for production.

What causes a tombstone? One termination melting before the other, which is a thermal balance issue rather than a placement one.