Originally Posted by bountyhunter View Post
A TO-92 standing up in free air has a theta J-A of about 180C/W. I guarantee it.
It's nice that figure gets a mention on the datasheet. Why was the reason for omitting it or have I missed something?
basically, because the person who wrote it was an idiot who did not understand what theta J-C means, which is thermal resistance junction-to-case. There is no defined heat conducting surface on the case to attach to a heatsink.
Originally Posted by bountyhunter View Post
You can also solder it down to a PCB with a 2 square inch copper area under it and it will drop the thermal resistance from about 65C/W to maybe 30C/W.
Does that also help with the TO-92 package?
No and No. How do you solder a plastic case to a PC board? And if you do glue the plastic down to copper or aluminum, the thermal rsistance of plastic is so high that the effect of the heatsink is negligible.
Originally Posted by bountyhunter View Post
Usually, they cycle in and out of thermal shutdown a bit then establish an equilibrium where the output voltage (and load current) is reduced to a point that the power dissipation keeps the die temp right at the shutdown threshold.
What's the likelihood of that damaging the IC?
Zero, for a properly made part because the thermal shutdown will save it. If it dies, it means the shutdown was not working and that happens fairly often
Originally Posted by bountyhunter View Post
Usually, they cycle in and out of thermal shutdown a bit then establish an equilibrium where the output voltage (and load current) is reduced to a point that the power dissipation keeps the die temp right at the shutdown threshold.
Does that happen when the IC is short circuited? Not in my experience but I could be wrong.
Yes, in most cases with small packages, the power dissipation in short circuit is enough to cause the thermal limiter to cycle on and off. In many cases, the package and die get thermally saturated and establish a balanced equilibrium where the current limiter is holding power dissipation steady right near thermal shutdown. The cycling behavior may stop, but it's not always obvious whether the curr limit or temp limiter is driving.