Re: VK2AUJ's µBITX project

Jerry Gaffke

That math may look silly.
We found that the silicon was just a couple degrees C hotter than the tab.

But the math becomes important on something like the IRF510.
Many Bitx40's are getting around 20 Watts out by raising the PA-PWR voltage to 24 volts.
Assuming 50% efficiency, that's another 20 Watts going up as heat in the single IRF510.
The IRF510 has nearly the same numbers as the LM7805:  150 C max,  junction to case of 3.5 C per Watt.
With 20 Watts dissipation, the die (junction) is  20*3.5 = 70 C hotter than the tab o the TO220 case.
And the tab is hotter than the heatsink, perhaps by quite a bit if the tab is electrically insulated.
So figure the die might be up to 100 C hotter than the heatsink.
The IRF510 heatsink should never exceed 50 C to keep the die below 150 C.


On Tue, May 1, 2018 at 04:21 pm, Jerry Gaffke wrote:

I measure the current from my 12v supply into the LM7805 on the Raduino at 60 ma.
So the LM7805 is dissipating (12v - 5v) * 0.060amps = 0.420 Watts
And we can calculate the junction temperature to be 4 * 0.420 = 1.68  degrees C higher than the TO220 tab.
The junction temperature is just a couple degrees C above that of the metal tab you see on the outside of the package.

Join to automatically receive all group messages.