Cold Solder Joints and Poor Wetting
Two Different Defects
A cold joint and a poorly wetted joint are often described as one problem, and they are not. A cold joint has not reached a temperature sufficient to melt the alloy properly, so the solder is grainy, dull and mechanically weak. A poorly wetted joint has reached temperature but the molten solder did not bond to the surface, so the solder sits as a blob with a high contact angle and the joint looks convex. Both produce a high resistance connection and both fail in service, but they have different causes and they need different corrections.
What Insufficient Heat Does
When the joint does not reach the liquidus temperature the solder does not flow, and the flux does not activate fully. The result is a joint that may look continuous but that has not formed a metallurgical bond, with a resistance that rises with temperature and vibration. The causes are usually the profile, the thermal mass of the joint or the board’s heat path. A large ground plane draws heat away from the joint, a thick copper layer spreads it, and a component with a large thermal mass absorbs it. A reflow profile that was developed on a small board and applied to a heavy one will produce cold joints at the heavy features.
What Poor Wetting Looks Like
A wetting failure is a surface condition. The molten solder is unable to reduce the oxide on the pad or the termination, so it beads instead of spreading, and the contact angle stays high. The joint may have the correct volume and the correct shape on the surface, but it is not attached. The causes include an oxidised or contaminated surface, a pad finish that has degraded, a flux that is too weak or too old, and a plating that is not properly solderable. In the extreme, the solder de-wets and pulls back from the surface, leaving a thin layer with a bare patch beside it.
Thermal Causes in Detail
The thermal causes are worth separating because each has a different cure. An inadequate peak will leave the joint below the liquidus everywhere; a peak that is adequate but too brief will not let the alloy coalesce; a soak that is too short will leave the heavier features behind the lighter ones; and a cooling that is too slow will allow the intermetallic layer to grow and embrittle the joint. On a wave soldered board, a joint that does not reach temperature is usually a preheat or a dwell problem, and on a hand soldered joint it is usually a tip temperature, a tip size or a dwell problem. The remedy follows the cause, so the first question should be which one it is.
Material and Surface Causes
The material side includes the pad finish, the component termination, the flux and the storage. A finish that has oxidised in storage, a component whose termination has a contamination, a paste that is past its working life and a flux that cannot cope with the surface all produce wetting failures. The board’s own cleanliness matters, since a residue from fabrication or from handling prevents contact. Where the failure is concentrated on one component or one area of the board, the cause is likely to be the material or the surface; where it is spread, the process is more likely.
Detection and Its Limits
A cold joint is visible as a dull, grainy surface, and it is the one joint defect that optical inspection can catch well. A wetting failure is harder, because a joint that looks convex may still be adequate and one that looks flat may be de-wetted, and the judgement depends on the acceptance criteria and the viewing angle. X-ray shows the internal structure but not the surface bond, which is why a wetting problem can pass an X-ray and fail a mechanical test. Where the risk is high, a pull or shear test on a sample is the measurement that settles the question.
Correcting the Process
The correction should follow the diagnosis. For a thermal cause, measure the profile at the joint that failed rather than at a convenient point, and adjust the peak, the soak or the ramp. For a surface cause, verify the storage and the handling of the boards and the components, check the paste’s age and condition, and confirm that the flux matches the surface finish. Where the cause is a design issue such as an inadequate thermal relief, the layout is the answer. And where the defect keeps returning on the same component, the process is being adjusted for a symptom rather than a cause, which is the moment to look at the component and the footprint.
Acceptance Criteria and Judgement
The criteria for a solder joint should be written so that two inspectors reach the same conclusion. A cold joint is generally rejectable, but a joint that simply looks dull after a lead free reflow is not, because lead free alloys produce a matte surface as a matter of course. Confusing the two is a common source of unnecessary rework. The criteria should distinguish between the appearance of the alloy and the evidence of a missing bond, and where the judgement is difficult, a mechanical test on a sample provides the evidence that appearance cannot.
FAQ
What is a cold solder joint? A joint that has not reached a temperature sufficient to melt and coalesce the alloy, leaving a grainy, weak connection.
What is a wetting failure? A joint that reached temperature but did not bond to the surface, so the solder beads with a high contact angle.
Can optical inspection find both? It finds cold joints well. Wetting failures are harder, because appearance depends on the viewing angle and the criteria.
Why do heavy copper areas fail? The copper conducts heat away, so the joint reaches a lower peak and a shorter time above liquidus.
How should the correction start? By identifying whether the cause is thermal or surface, then measuring the profile at the joint that failed.
Conclusion
Cold joints and wetting failures both produce unreliable connections, and they are cured by different means, so diagnose before adjusting. Measure at the joint that failed. Reflow control belongs to SMT PCB assembly, the assembly result sits in PCB assembly, and the acceptance criteria are part of quality management. Process development for a new product happens during prototype PCB assembly in 2026.
