Heat Sink Calculator

Heat leaves a semiconductor through a series circuit of thermal resistances: θ_jc inside the package from junction to case, θ_cs across the interface material, and θ_sa from the sink to the air. Each one turns watts into kelvin of rise, so T_j = T_amb + P·(θ_jc + θ_cs + θ_sa). Rearranged, the sink a design needs is (T_j,max − T_amb) ÷ P − θ_jc − θ_cs. The page also shows the share of the rise each stage takes, which is where you find out that a better sink will not help.

Running a device at its absolute maximum junction temperature is not a design, it is a limit: lifetime roughly halves for every ten kelvin, so the margin you set takes a percentage off the allowed rise before anything else is worked out. When several devices share one sink, each one’s θ_jc and θ_cs act on its own dissipation, but the whole load flows through θ_sa — which is why sharing raises every junction, and why the hottest device, not the average, sets the limit.

Steady state and one-dimensional: transient pulses, spreading resistance in the sink base, radiation, altitude and recirculated hot air are not modelled. All three resistances are your inputs — θ_jc from the device data sheet, θ_cs from the pad or grease at the mounting pressure you will actually achieve, and θ_sa from the sink maker’s curve at the airflow you will actually have, since natural-convection figures collapse if the sink is boxed in or mounted the wrong way up. A thermocouple on the real assembly is the only proof. Nothing leaves the browser; the same four anonymous usage counts as the rest of the site apply.