Multiple transistor heat dissipation.

[Boscoe]

If, for example, I have one transistor with a jc of 0.8C/W and a c-heatsink of 0.5C/W then and then a sink to air resistance of 0.5C/W I can dissipate 50W without exceeding a junction temp of 100C. If I now have two of the same transistor but now each dissipating 50W each is there an effective increase in junction to heatsink resistance or less now there are two paths for the heat to go? In other words is it more efficient to dissipate the same heat through multiple devices than one?

Thanks
Boscoe

[Andrew Eckhardt]

It's only an efficiency improvement in terms of maybe required heatsink size, but yes, adding parallel transistors decreases the total junction to ambient thermal resistance. You can run through the calculations to show how. Two transistors instead of one doubles die area and halves junction to case thremal resistance. So with a perfect sink it would take twice the power to raise the two junctions to the same temperature as one. Of course a real sink of fixed size is going to get hotter, so there's a diminishing return to be calculated.

[sreten]

Hi,

It is more efficient to to use two devices more than one as you effectively
half the resistance of the thermal paths. In fact for the same power two
devices need a smaller heatsink than one for the same junction temperature.

In reality keep the heatsink the same size (as one device) and up power say
30% (for two devices) to get the best out of the heatsinks and 2 devices.

rgds, sreten.

[Tajzmaj]

And something like 40-60W per device is safe and nice value.....
Best regards, Taj