Step 1. Lower/mill the dielectric by the sides of the emitter's thermal pad for a nice window of raw copper surface where a nice deal of the reflector's base ring will land.
Step 2. Put some solder paste over the uncovered window surface, and on top of it a 0.1mm copper sheet. We are to allow the round reflector base to land on top of it, shape the copper sheet shim accordingly. Carefully reflow it at a low temp.
Step 3. Centering rings get in the way? Sand them until the reflector's butt lands onto the copper sheet (not tried yet). Add some thermal paste.
Mmmkay, had to write it in a flash. Sounds good? Maybe or maybe not worth the improvement.
High quality reflectors made from glass or electroformed metal won’t conduct heat very well. It would only work with the cheap reflectors milled out of solid aluminum.
Having too much pressure which is important for heat transfer will distort or damage the reflector so it needs to be appropriately tightened and also use thermal paste.
The contact area between the pcb and reflector needs to be as large as possible so a 32mm or larger mcpcb with a large base on the reflector is important.
Finally, heat will make the aluminum expand so the light needs to be cooled very well already to prevent the reflector from being distorted by heat expansion.
Barkuti, I have thought about this too, but only for torches with big MCPCBs, large solid reflectors. Specifically Q8 and SRK.
Thermal gap filler pad was my thought. Cho-Therm specifically, I have access to offcuts.
But the reflector also needs to have good thermal contact to the head if it is to work steady-state.
In the case of the Q8 or SRK the reflector is a big chunk of Al, so could extend high power modes for a short while due to extra thermal mass, but when I did the sums I didn’t see much point.
Now if the reflector could also be bonded in good thermal contact with the head, it would be quite different. I could have stuffed Cho-Therm in to do that, even considered potting it together.
I agree with Enderman that thin shell-like reflectors would contribute little to thermal management, but the big chunky ones I mentioned could be a useful extra thermal path. If it is worth it (the materials I mention are not inexpensive, nor easy to apply).
Dunno if it ever did any good, but when I’d have a reflector pressing right up against the pcb, maybe a centering ring but not a spacer or insulator, I’d just “wet” the lip of the reflector with some AS5 to fill the gaps and make better contact.
Not a perfect dtp connection to it, but if every little bit helps…
Some companies already did this to get some extra cooling. PflexPro is one of them. I had the same idea too, but I use the entire “dummy” reflector to cool the light (collimation done by the small TIR optic inside it): Dimmable Cafe Light Prototype
And it works really well. In my case its because my reflector exposed to the open air directly.
Several things to consider….
Most cheap reflector don’t have good efficiency and in higher output flashlights it gets hot just by absorbing the photons out of the LED. If you think this is hard to believe then first try to:
- insulate the base of the reflector.
- Run the flashlight in full power.
Measure the temperature difference between the reflector and the MCPCB.
If the reflector is cooler than the MCPCB then it is worth the hassles. Otherwise thermal coupling the reflector could adds more heat to the MCPCB. But for most under powered flashlight I bet the reflector should be cooler than the MCPCB.
Couple observations:
It’s not dtp, infact there’s a ceramic insulating layer!
Did I mention it’s ceramic, they must really trust heatsinking to the reflector!
