Accidentally this week and last week I did some non-standard thermal simulations with through-hole pins by imitating hot solder solder wave underneath the PCB. Of course a mere thermal simualtion can only look at heat spreading in detail but cannot cover metallurgical or fabrication aspects. Nevertheless, the results gave a correct agreement of positions of the problematic pins (and those without problems) compared to the solder wave result. Among other effects it showed that local copper around a solder pin can help heat to flow toward the pin rather than flowing away from it. But all that depends on details of the layout in the bottom layer and maybe in other layers as well, the mass of the component, how the flux is dispensed and many more influencing parameters. A numerical investigation can help to reduce the number of prototypes.
Why is temperature of printed boards so important?
Because there are limiting temperatures for operation:
- Secure maximum operational temperature of components
- Glass transition temperature of the board material
- Stability of solder joints
Why should temperature be known as good as possible?
Because too many items depend on temperature:
- Material properties, e.g. electric resistance of traces and thus voltage feed or impedance for transmission of signals
- Performance of components
- Derating of power for hot ambients needs knowledge of component temperature
Why don't simplifying spread sheet calculations work?
Because the local thermal conductivity (the artwork) controls the temperature of a component. It is impossible to include copper patterns, drills and influence from other components in a simple formula, a data sheet or a simple network. Better do a holistic simulation using dedicated software.