Basically a (1-2S) 26650 host I recently burned out its poor 800mA 20mm driver. Uses a 20mm star over a good heatsinked pill (it was hollow as it came); see here:
Torch is a zoomie with 16mm travel path and 30mm already scratched plastic plano-convex lens.
Parts list:
Non-DTP “XM-L” footprint copper 20mm star board, which I’ll gladly DTP-enable with 800 grit sandpaper (I’ll see…), flathead screwdriver, solder and frying pan.
LD-29 2’8A efficient buck driver.
Glass 30mm plano-convex lens.
I don’t know yet how will I remove the plastic lens. It is held within that donut shaped artistic piece of aluminium. Should I put it in the oven with my next pizza and see if the thing falls off with a nice hammering? LOL!
Well, enough for now.
Sense voltage is 0’07V across a default 0’025Ω SMD shuntsistor, that means it has to dissipate 0’07V × 2’8A = 0’196W max, which seems close to the maximum 0’25W for ¼W resistors so, that shuntsistor at least needs to be ½W to allow some acceptable current derating at high temperatures.
I have no real clue of what he says about those “loop-back” resistors. Just doesn’t makes sense to me to presume they are used as current sensing resistors directly, seems too much of a voltage drop through them.
A MOSFET, mmm, maybe these resistors are there to modulate the source-gate voltage drop, fine tuning the measly source-drain resistance and corresponding current flow.
Hope someone brings additional info on this.
With thorough respect my friend, you’re talking bollocks.
My “heatsink” is just a (much improved) kind of pill.
My host is just MASSIVE compared to an SK98 body: its wider and nearly twice its lenght. UltraFire SK98 XM-L2 at FT
You have impoved total surface area for about 5%, but have lost in thermal conduction several times.
So powerfull LEDs must be used with pills with integrated shelf.
Contact area between your “heatsink” and the pill is to small.
The only way of using such pills with huge hole inside - pressing piece of copper rod in the pill. Cold press fit to maximize contact area.
I’ve spent a little time with these drivers. If you are planning to alter the resistors for more output, they will overheat quickly & reduce power unless they are well sinked.
You are planning to use a DTP board for the LED, aren’t you?
Nice range there. They have the aforementioned driver at FastTech labeled for SST-50s, guess also optimized for higher current, but much more expensive. :-/
A couple of years ago I bought a Nichia NVSL-219B-E there, very nice high CRI emitter.
Maybe you sound a bit dramatic but, I am using thermal paste on the more or less fit inside heatsink piece.
However, I have a blowtorch at hand and, with this kind of acidic soldering flux:
… it should be easy to tin the contact surfaces of the inner piece I cast and, with additional flux on the outer hollow pill and more blowtorch flux, it should be easy to meld both parts into one.
Sounds good! Doesn’t it?
My host uses a ∅29/∅27mm OD/ID aspheric, 11mm thick. I was wondering what kind of improvement (?) would I get from replacing it with an equivalent glass one, like this.
