Larry,

I don't believe the power rating would be tied to the failure level.
More than likely it is tide to the temperature rise in a specified
ambient.

If I were working with such a device on the bench, I would monitor temp
so it did not exceed a reasonable rise.

Regards, Ron

I don't recieve e-mail at this address because of spam. E-mail me
at crtsrATmsnDOTcom.