Junction Temperature: The Number That Actually Limits a Design
Every semiconductor has a maximum safe junction temperature, and exceeding it — even briefly — risks permanent damage or a dramatic reduction in service life. Junction temperature isn't measured directly in most designs; it's calculated by adding up thermal resistances along the heat path from the junction to ambient air, each stage adding its own temperature rise proportional to the power being dissipated.
The calculation chains together junction-to-case resistance (intrinsic to the component's packaging), case-to-sink resistance (dependent on the thermal interface material and mounting quality), and sink-to-ambient resistance (dependent on the heatsink's size and airflow) — each stage's temperature rise is simply power multiplied by that stage's thermal resistance, and the stages add in series like resistors in an electrical circuit.
This is exactly why a heatsink that looks adequate on paper can still let a component overheat — a poor thermal interface (a dry or improperly applied thermal paste layer) can dominate the total resistance even when the heatsink itself is well-designed.