optimal lens

I have been lurking for some time now and looking through older threads for information on lenses and have come to the conclusion that starting a new thread might just prompt the kind of discussion I'm looking for regarding a "silver bullet" collimating lens solution. Forgive me if I don't make proper use of vernacular in my descriptions. In what I have seen of lens comparisons, one would opt for a "3 element" lens or "LFL" lens for somewhat improved beam quality at the cost of power, or a "G-2" or "G-1" type lens for increased optical power at the cost of beam quality (in general, but especially for multimode/wide emitter LDs).

My understanding of the basic principles responsible for the advantages/disadvantages of lens types (please correct me if I have missed something) is that the G type lenses have shorter focal length and therefore get placed closer to the emitter, as well as having fewer interfaces which leads to no clipping of the divergent incident beam and low losses respectively, but because the resulting beam is quite narrow, the divergence is high. The 3 element lens has a longer focal length and several interfaces resulting in its placement further from the emitter which results in some beam clipping and higher transmission losses, but a wider resultant beam with lower divergence. Now, I have recently purchased a 9mm 445 diode in module from DTR and it looks like with no lens, the fast axis (which comes from the shorter dimension of the emitter) diverges quickly enough as to be clipped by the female threaded barrel of the module. A small lens deep within the barrel that collimates the beam before it is wide enough to be clipped could then actually result in more optical power output than the module without a lens (e.g. G2 glass)

The wheels have been turning in my head, and I propose the following: What if one were to truncate the length of the module extending "down-beam" from the diode window such that it could no longer interfere with the divergent beam. Then thread the outside of some portion of the upper end of the module (again, to avoid putting any opaque components in the way of the beam) which could be used to attach a lens holder with a single element, long focal length (on the order of say 12-20 mm) aspheric lens with the end result of... wait for it... A wide beam with low divergence, and high optical power, sans beam expander to boot. I'm picturing a good 3W, 1.5 cm wide beam with sub mrad divergence on both axis from a 9mm 445. I realize that threading the outside of the top portion of the module would make for less heat sink contact area (although the lens holder could be a heat sink as well), and this might not be a plug-n-play alternative to the standard aixiz setup, but would fit right in to most hosts as long as the module axial position can be adjusted forward relative to the heat sink. I recently moved and have several home projects that need to be finished before I'm allowed to pull the lathe out and start messing around with my own hobbies, but I wanted to throw the idea out to see what LPF members might see as problems with this setup, or issues that I have neglected as an optics novice.

Well, that linked thread doesn't seem to be quite along the same lines as my question. Or at least I was unable to find an answer there. I'm not worried about corrective optics at this point, but maybe the astigmatism, rather than the divergence, is the most prominent factor inhibiting perceived beam quality. I have found a supplier of relatively inexpensive (for good sized aspherics) lenses and ordered both AR coated and uncoated 18mm EFL lenses with numerical aperture of 0.7. I do not currently have a LPM (although I'm in the market for an Ophir head if anybody has a spare) to compare the power throughput, but I should be able to do some spotshot comparisons in a couple weeks when the lenses arrive.

Not to dissuade your exploration into optics, but I'm not sure what the benefit would be over just using a single larger lens with a longer focal length to collimate the light into a larger beam. The only reason beam expanders are used is because the light is already collimated. Here you can just use the uncollimated output of the laser diode and pass it into a larger lens with a larger focal length, resulting in a fatter beam with lower divergence.

If you'd like to try this yourself, go to DTR's shop and get one of his optics kits from a projector. Many of the lenses are wide-band AR coated. You could even use the lens from the knife-edge array which is made for those blue lasers. Not a bad way to go if you want a fat beam.

That's exactly the idea Badger. Instead of using the very short FL lenses that seem to be common practice here, I am going to try using one with 18mm fl. That will be the only lens I am using. The problem is finding low cost, large, aspheric lenses with a large enough numerical aperture to accommodate that widely divergent raw beam. The projector optics kits look like they have one lens that could be suitable, and as leftovers the price is certainly right. Thanks for that tipoff. That will give me some things to toy with.