Cree XR-E at 240 lumens?

[The_Doctor]

I just light my room with mine. Not sure if they'll collimate or focus to burn tape. Maybe if the lens was carefully taken off and a larger lens used instead, they might. I don't know the emitter size, there's a curious lack of any statement on that in all of the Cree and Luxeon PDF's I looked at. I really looked hard too, there is no statement that I can find. They're definitely small though. I tested a pair of lithium CR123A rechargeables today (very rewarding, first battery type I ever had that charges about as fast as it will discharge at a useful safe rate, meaning two sets, no waiting). WHile they were at their weakest on an LED torch, I looked closely at the chip through a camera lens I use as a magnifier. The chip is square, has three emitting areas under the phosphor coating, and the bonding wires were visible attached to two adjacent corners. I don't know how much of the magification was due to its own lens, but as the bonding wires can carry an amp, the total size must be too big for good collimation unless you use a lens a few inches wide.

[steve-o]

Hmmm.. interesting..

[PNjunction]

What was that Planck reference?

That's just me potentially being an idiot... seemed like the thing to say at the time. More seriously though, it was a semi-weak reference to the chromaticity index. IIRC, it has something to do with some Planckian formula.

You had also asked about buying the Rebel. If this was a couple months ago, I'd probably say to hold off... but now, Lumileds have released their 100 (@ 350mA) lumen piece, and is available through either Cutter Electronics or Future... I forget which. In singles, I think they're $9 US.

Here's where I have to do some more homework, but the data states they can withstand a 175 degree C junction temp. If underdriven, I wonder if they can be used without heatsinking. I'd love to stick one of these in my natural HA Mag Solitaire, running a 10440 Li-ion.

BTW, a 4 degree optic for the Cree?... about to start Googling.

[The_Doctor]

BTW, a 4 degree optic for the Cree?... about to start Googling.

Walking the Planck? Nm...

The 4 degree optic is a Ledil make thing, I think, called a Rocket. I haven't tried one yet. Ideally I want the smaller reflector types like Fenix use. Narrower, and no deeper than the Rocket.

The idea I had for a Rebel LED was a small light for a Psion Organiser II's LCD screen, fitted internally. I doubt it would work though, the clear plastic in there is in no shape to act as a decent waveguide.

[PNjunction]

Walking the Planck? Nm...

Arrrrrrr, me hardy... that was a good one.

Ah, yes... Ledil. Not being an optics guy, purely because I always thought they're ugly lookin' in a torch (I like the aesthetics of bright, shiny reflectors), I haven't kept tabs on much of the offerings out there.

The Rocket however, looks quite interesting, although looking at the datasheet brought a bit of disappointment. I really could've used one in a mod host, but it can be no bigger than 24mm dia. The Rocket is 26mm. I'm not sure if the outer edges can be ground down without ruining the optical properties, and I don't have a proper means of doing so anyway. Shame... for a moment there, I had idea for another offbeat, yet effective torch mod.

Out of curiosity, I'd like to look into how the Optic's 90% efficiency translates to torch usage... meaning, is it more efficient than a reflector and lens in a flashlight? To make a long story very short, a while back on CPF there was a long running inquiry/debate on determining an accurate ratio between bulb lumens (amount of light produced by an emitter or filament) and torch lumens (what actually comes out a flashlight's front end). Eventually, a figure of 65% of bulb lumens was determined to be an effective rule of thumb, given a high quality reflector coating and using an ultra-clear glass lens.

How does this compare to an optic? Will an optic truly throw 90% of a Cree's bulb lumens forward and out of the torch, or are there other losses to consider? Time to go info hunting.

[The_Doctor]

The Rocket could be ground down a bit, I think you might get a vague ring-shaped scatter at worst, and not much light lost. Tedious work to do it though, and figuring out how to hold it while working on it won't be easy either. Probably by hand would be best, rotating the rim gently against a disk sander with a very fine grit disk on it. If you wanted more than 1 mm off the rim it might significantly affect the beam as the curve changes the angle, but that first mm looks ok as it takes back from a sharp edge and will eat less than 1 mm from the reflector surface. It's actually a loose 26 mm too, actual diameter I just measured is 25.75 mm. They're very cheap, so worth experimenting on.

The efficiency might not be easy to gauge. While it has total internal reflection, the light must first be coupled into it, then out of the front. Neither surface is AR coated. On the other hand, a metal reflector and window offers more chance of losses, as you still have to pass two surfaces AND you lose in the reflector, and in the Rocket, you lose next to nothing in the reflector as it is based on total internal reflection (there's no metallisation there); that's the basis of fibres and they can carry light for hundreds of miles without losing much.

An uncoated acrylic lens can pass about 90% of a laser diode's output, >95% if coated, so the Rocket might be able to pass 90% out.

I wanted to work a pirate joke in there somewhere but I'm not on form today, and nothing about that text gave opportunity. Next time, maybe... Maybe this will do for the time being:
A slight shift in wind direction causes the mainsail to flap, and the Mate says to the Captain,
"The mainsail's not happy."
"Yes it is," the Bosun says; "It's luffing every minute of it!"

[wes]

even the timbers are groaning at that one!

[steve-o]

.................................................. ..............lol

[The_Doctor]

even the timbers are groaning at that one!

As well they might.

Steve-o, the operative word is 'Shanghaied', or 'Shanghied', I forget which. Don't ask me why though, I have no idea... Maybe someone from Shanghai can tell me?

Stanwax is sending me some extremely thin single sided FR4. 0.2 mm thin. The copper will spread the heat, and even though FR4 is a poor heatsink, it is so thin that with the copper spreader it should pass enough through to a copper or aluminium block on the other side, to effectively keep LED's cool. Pnjunction, I'll send a bit on to you to try if it looks like it works well. If it really DOES work well, it might find favour with many here, for making easy secure and reworkable connection in tight places where heat must be removed.

[The_Doctor]

I got the LED's and the thin PCB today. Thankyou Stanwax and PNjunction. I'll test the PCB first, as that's needed before I try other things.

It's VERY thin, 0.15 mm including its single 0.02 mm copper layer. I might consider electroplating a thicker layer on it to form a heat spreader, but as I'm using it for small parts it might be better to use pure indium as solder, and tin the copper with that as it's a very good conductor of heat. The natural filletting of the indium will be an ideal heat spreader too.

Superglue doesn't seem right for gluing it to aluminium, as it peels off too easily. I'm currently trying a piece glued with a thin smear of two part epoxy (araldite) with cure speeded by heating to 80C once the PCB is held against the aluminium under firm pressure.

So long as this stuff sticks ok to bare metal, sanded and cleaned with acetone before the epoxy goes on, it will be an ideal way to heatsink small and complex structures without the expense of a metallised ceramic plate. It will never conduct heat as well as that, but it looks like a neat way to handle things for which no other cheap or easy answer exists.

Edit: It's so thin that some printers might handle it directly, if they can handle stiff paper or transparencies. That can allow complete optical and electronic layouts to be printed in ONE stage right to PCB for etching if the prining ink can stand it. That would allow a good (and very cheap) compromise between mass production and one-off hand building, which might be what is most needed for people building lasers.