A Really Big Laser?

[planters]

Has anyone thought about this?
http://www.photonics.com/Article.asp...ures&AID=57248
http://www.osram-os.com/osram_os/en/...tors/index.jsp
http://www.okotech.com/microlens-arrays

The module is rated at 50W and if actively cooled could probably exceed this by a lot seeing as Osram has specified each diode @ 3.2W.
These are almost certainly polarized, so doubling in a big PBS should be a snap.

By placing the micro-lens array a little farther from the diode plane than the FL, the individual beams could be focused into fibers. Alternatively, a 2-D array of prisms could do this the conventional way and probably produce a better far field.

This would be a big beam (20 diodes). But, 100++ watts?

[krazer]

Seems like the packing density is pretty terrible on them, so it would require some intensive knife-edging to get the intensity to be higher than a decent LED (phatlights are about 100lumens/mm^2 for blue, this thing would be 8lumens/mm^2, admittedly with a much lower divergence if you put a collimating lens on each diode but still nothing compared to the 2,500,000 lumens/mm^2 at the surface of the diodes).

That said, as long as you are clever in how you layout the knife edging optics there is no reason you shouldn't be able to get identical optical performance as you get with individual diodes, and presumably you could make it slightly more compact with that package. But at the end of the day, it basically the same as just pulling the whole diode block out of a projector and starting from there. I am holding out to see when they release 19 emitter bars, since there is already very established technology for combining them into nice tight fibers (50w into 100um 0.22NA fiber is common, double if you allow for polarization multiplexing, and even more with wavelength multiplexing). If you could start with a 100w module into a 0.22na you are only a single lens away from a 10mm perfectly round beam with 1 milliradian divergence.

[planters]

You are right about the packing density, but that only means that the lens array has to be fabricated with the same pitch and if too dense, the knife edging will be too difficult. I'm not suggesting that the entire array be focused like an LED, hence the micro-lens array. The specified, 3.2W/diode is also really high.
Nevertheless, I agree that the fiber approach is better and that is the thrust of the author of the article (he works for Powertechnology, duh). However, I fear the cost of the fiber module and it doesn't exist yet, but then I don't know the cost of the array either.

[DTR]

I have not seen the Osram block in the wild but we can get this 8 diode block. Rated about 30W and a member over on LPF has tested the wavelength which is over 460nm.
https://drive.google.com/file/d/0B_6...ew?usp=sharing

I have thought about sticking PBS cubes right on the top of each diode to combine the output.

[VDX]

... look at the 'hyper-diode-laser' here: http://www.ringfocus.de/en_page/prod...ringflaser.htm

They did some really complex optics+adjusting for a pretty poor focussing quality ... could be much better (and simpler) with fibercoupling :-/

When testing my 6x 455nm-diodes head, I've got an overall focus diameter of 0.3mm in 60mm distance from the diodes, but shouldn't be a problem to reduce it down to 0.2mm with lower distance and even more diodes.

So coupling some ten diodes into a fibre with 0.2mm shouldn't be a problem ... a bit more fine-tuning for a 0.1mm-fibre ...

Viktor

[planters]

DTR,
The neat thing about that module is that the lenses are already in place and the leads suggest that the polarizations are aligned as well. I know I could mill a plate with 8 holes that step up along the array so that the right angle cubes effectively knife edge the beams.

Something I don't understand well is the fiber splicing technique. I have some IR diode modules that include several independent diodes, each fiber coupled and then spliced to a single fiber. Lets say the final fiber is 100um in diameter. Does that mean that if there are four combined diodes that each needs to launch into a 50um fiber?

What about the launch conditions? If 10 diodes were angled, free space, into the end of a 100um, multimode, fiber (20 mm FL, given VDX's example), would the output contain the convergence angles of each of these diodes or does it homogenize. In other words, would the NA be the inverse of the focal ratio of the individual diode or the inverse of the convergence angle of the array?

Given krazer's example of the 19 emitter bars it would seem that the only real difference with an individual diode/collimator/focuser and fiber launch alignment is the compact nature of the bars. Putting aside multiplexing and polarization, each element of a bar still has to be coupled into a fiber. Even if glued directly to the output facet, the laser can't be coupled any more intensely than what a transfer lens would accomplish. It would just be more compact, robust and hopefully done for you. I'm not saying that it won't work, it does. I think Dilas makes something like this in red ($$$$?). Is there a system that could allow individual diodes to be fiber launched rather than knife edged and retain the far field performance?

[DTR]

Here is a thread on a project using this over on LPF.
http://laserpointerforums.com/f61/ea...ml#post1356206

[planters]

locked out of link

[DTR]

Sorry did not realize it was in the Vets section over there. Anything to do with N!tch!a over there is in the Vets due to past problems with them.

[dsli_jon]

Heya Sire -

I'm sure The P'fessor will be along soon to comment in this / correct my mild-mannered understandings, but... Meantime..

..If 10 diodes were angled, free space, into the end of a 100um, multimode, fiber..would the output contain the convergence angles of each of these diodes or does it homogenize...

No, that's the whole 'point' of using fiber, *to* 'homogenize'.. The output ends-up based on the properties of the fiber / diameter, not what you 'pour into the funnel', that's the beauty of it. However, 'efficiency' with diode-coupling largely comes-down to the achievable spot-size, at the fiber-input, vs the size of the fiber..

ie: In your example above - Aside from it being a nightmare, since you'd need to have like, practically indie-gimballed super-finely-positionable mounts to 'target' the end of One fiber - (which would be really difficult, because you need to have some 'meat' around them for heatsinking) - The 'farther apart they're spaced', the harder getting them all-into the same spot, at an 'efficient angle' (so they all go INto the fiber-end, vs glancing off, etc..) - is going to be.. Aside from all that muck, especially with a 'multimode' chip, there's only so-far down that 'spot size' is going to image.. And, 100µm is 'tough-enough' with an Ion-quality beam, so.. You're more-correct in going the direction, here..

...it would seem that the only real difference with an individual diode/collimator/focuser and fiber launch alignment is the compact nature...

..IS 'the Real difference'.. Coupling 1 fiber to 1 diode-face (whether in a bar or array or can, etc) is - as I understand - The-most efficient coupling, possible.. Yes, much more expensive, but.. Much-better results. And, 'splicing', simply takes all those indie-diode 'inputs', and - literally - fuses them into One, bigger fiber.. (glass is nicely cooperative like that..

Now, this is not to say that, 'it's just not possible to 'gang bang' a bunch of closely-stacked diodes into the end of a fiber', BUT.. because of that 'ultimate spot-size', it becomes more a question of, IF the Fiber-diameter (..that you CAN efficiently cram all n-Diodes into) will be 'ok', for Your-output needs.. ie: You might, say, be able to get most / all of n-Diodes into a 200µm, or 400µm.. But not a 100µm. But - given the finite beam-size you can glean out of a 200 or 400 (vs that 100µm) - Will that 'be ok' for your App?.. Or, are you simply 'better off', just indie-collimating / prism-edging / combining.. For 'graphics shows', well, we know the answer to that.. But.. if 'Rrraw Powuh' is what you're after, IF you can 'live with that bigger beam' (ie: 10mm) then it all becomes more attractive..

ie: Waaay-back, when the '445 C@$i0 block' first dropped, I did-try to 'cram all the raw output into a fiber'..

..And - it "worked", but.. the 600µm fiber didn't exactly yield the 'best beam'.. Of course, the main issue there, was largely that the 'raw block diode-spacing' was way-too far apart / no adjustability, but.. Again, only so-small that mission-critical 'spot size' would be, with those diodes / lenses, etc, So.. that was Fail for a 'quick and dirty 20W blue-engine'.. :/ ..BUT..

..I will-say, that there's a *good reason* we put the '4 small tapped-holes' around the Aperture of the 'BlüFlame' chassis-design.. http://i.ebayimg.com/00/s/NDYxWDYxNQ...U6mOT/$_12.JPG ..Even now, with 9mm 445s, we 'extrapolate' better than 45W out the aperture.. with the *3W* spec Blues? - x 20-up - in as tight of a-core as these pups feature?? Pff.. Monster-Power potential, Tiny Little Body... Will report-back when we get a chance to test out a 'fiber-coupler' for that pup..

..Is there a system that could allow individual diodes to be fiber launched...

Yep, and they are Beauty-full, but $$$$... Coherent makes the 'Cube', and.. there's another Co I found that made really nice indie 1-up and 2-up modules, but.. I can't find my literature on them, right now.. I'll keep lookin, later.. Must have named it something wakky..

..If you 'already know all this', and I've somehow 'missed yer point', please forgive.. Been nearly-floored all week sick with some ungodly Mongolian flu or whatever it is.. +_+ and am just now starting to recouperate brain-cells, so.. Maybe I'm off base with what you were asking..

..Anyhoo, I'm sure the P'fessor will correct me where I'm smoking banana peel, here..

.02
j

PS

..we can get this 8 diode block.

Wow, what a sweet little cookie!! And... Hah, gotta love that 'last bullet point' on the spec-sheet.. "High safety structure for prevention of removing LDs" ..Lol..