Search This Blog

Friday, 2 August 2013

The Catches: Device Thermal Resistance

I been designing with power devices (BJT, Mosfet) for years – but I will always remember when I first started looking at their thermal resistance – by not reading through those fine prints,  I made gravely mistake of picking the wrong device for my prototype.

Take an example from the figure above – the device might be promoted to have less than 25degC per watt of junction-ambient thermal resistance (with fine print indicates test condition) – but without heatsink or large enough copper area, the thermal resistance easily go up to 40degC/Watt - as shown in the graph when copper area is less than 1 inches square.

Power dissipation specification at times shows best number - but only when mounted on large copper area - not practical for actual circuit – often heatsink is needed for low thermal resistance.

But it was the standard way for the industrial to compare using this method - too fresh and naive and some said silly to take it literally :)