Posting before first coffee of day , doh!
From experience do not rely on the emitter slug to be any particular
connection....
What should have said is that the emitters are not consistently and reliably
connected to their slug.
My thanks to R Lewis for correcting this before any LSs were smoked.
The emitter slug is silver plated and solderable, perhaps not reccomended
but possible.
Interesting thought about detaching the chip and remounting it on a sapphire
substrate, just wait for commercial diamond production to ramp up, then can
really get into overdriving.
Yep.
The sapphire substrate isn't very good, it's about a tenth of the
thermal conductivity of copper. (this is countered by the fact that it's
a better shape than the long bar of the emitter)
However, copper isn't transparent, and I'm using a reflector to bounce
light back past the LED.
For a rather silly upgrade to a minimag solitare.
I'd really like to use diamond, but it's only really a bit better at
these sorts of power densities.
Sapphire does 40W/m/K, diamond 2000W/m/K.
The chip is around 2mm square.
80mw/2mm/K.
So the cube behind the chip adds around 12C, and the four in contact
with it around 3C, and the the rest maybe another couple.
So, call it 17C rise for a watt, for a 4mm thick infinite disk.
Diamond would be interesting to try to overdrive it, but in this case,
sapphire is just fine.