making the new NUBM44 450nm diode handheld?

[Alaskan]

Even with mine having a damaged window I was getting 5 watts at 5 amps, I expect 6 or more watts when de-canned.

 

[RedCowboy]

Even with mine having a damaged window I was getting 5 watts at 5 amps, I expect 6 or more watts when de-canned.

I was shocked when my NUBM44 popped it's window in the first week of testing. I reproduced the result with a M-140 but I had to run it for 5 minutes plus in a flashlight host and get it pretty darn warm.... ok ok HOT! After busting the window out of that m-140 it still works fine....for now, but the window failed in it too.
I think the problem is going to be heat with these diodes and I bet they are more than fan cooled in the projectors. We don't have a spec sheet but my guess is 3 amps max on paper.
Mine was running at 5 amps and it seems that was too much. I'm going to guess that 4 amps it the safe limit for these with the window intact and no TEC because past a point any more power is just making more heat. Also as these diodes warm up the output drops and more of the power is converted to heat, especially when overdriven. Planters did a very controlled test with TEC and 4 amps is about it. If I was going to go higher then duty cycle will have to be very short unless these diodes prove to be able to last without the can/window. When I get mine back from DTR I will test it and see how long it lasts without the can, but I am going to cut it from 5 amps to 4.5.

 

[Benm]

The diode is outputting part of that 25 watts as light, so it isn't ending up as heat dissipation. The way you worded part of your post seemed to indicate the heat sink would need to handle that much dissipation.

Strictly speaking you are correct: what is emitted as light does not have to be dealt with by the heatsink. But since this amount is actually quite small (25% at best) i usually advise to ignore that fact and design the thermal system to handle the full input power to the diode has heat.

But in all honesty you should over-dimension heatsinks a bit usually: their K/W performance is often rated in optimum orientation and such, which may be feasible for a device sitting on a bench, but not a portable unit. It actually does matter if the fins on a heatsink are mounted vertically or horizontally for example, but this cannot be predicted for a torch-style device in real world application.