Aspheric collimators for red and blue laser diodes

[lasertack]

@ taggalucci

here are the pics.

[andythemechanic]

Ok, so:

Beam diameter at aperture with O-Like lens: 4,5 x 2mm
Beam diameter at aperture with aspheric lens: 4 x 1mm
...
Divergence with O-Like lens: 1,4 x 0,2mrad
Divergence with aspheric lens: 2,2 x 0,4mrad

Just a quick comment to the numbers: From the physics behind collimation this has to be axactly like that. These Aspheric lenses have roughly half the focal length of the O-Like so they should have half the beam size and double the divergence. Given the very different axes of the blue or red multimode diodes, it is anyway better to use short focal length and expand the high divergence axis alone later, or even directly start with a collimation with two cylindrical lenses. One thing thatdirectly jumps into the eye is, that although twice the focal length the o-like lens has not double the beam size in the fast direction. This is simply because it has a too small aperture for the fast axis and clips the beam at the outer edge of the lens what means power losses.

[planters]

lasertack,

Thanks for the link. As always,interesting. Earlier, you posted divergence data that I found confusing. The aspheric lens produced a longer stripe than the O-like at 5M and this is not surprising as it produces a longer stripe in the near field and the ratios are proportional. However, the aspheric produces a significantly narrower stripe in the near field AND a narrower stripe in the far field, yet you specify a higher divergence for this lens in that axis. Was there some error or revision in these numbers? Otherwise this seams like a significant advantage for this lens. As an aside I had always assumed the O-like to be an aspheric. Was I wrong and it is either simply spherical or multi-element? If there is such a distinction, or if the conic of the Rochester is better suited to this diode then these results do not necessarily violate any physical laws. I can only hope. Look forward to your feedback.

[lasertack]

@planters

The divergence of the aspheric lenses is larger, because of the larger NA and shorter FL. They produce a smaller beam at aperture but a higher divergence. By using prisms you´ll get nearly the same or better beam as with O-Like´s but at much higher output power.
O-Like´s consist of 3 spherical lenses. Because of the larger NA the RPO lenses catch much more light (260mW more than O-Like´s).

[planters]

Andy,

Forgot to add above. Your comment about the markedly different axis ratios of the emitters in multi-mode lasers favoring shorter FL primary collimators followed by correction optics is intriguing. If the FL of the collimator multiplied by the power of the telescope were a constant then it would seem that the trade off would be based on issues such as the ease of alignment, spacing constraints etc and the beam quality would not change. But, if you are correct, then the quest for better beams should seek the shortest FL collimators that have sufficient NA to accommodate the emitter. That seems to be the trend on the G71 thread where the consensus is to use a 2mm FL lens supplied by Dave @ LSPs. Thorlabs sells a 1.45mm FL asphere. What do you think about that?

[planters]

Lasertack,

Look at post# 8. At 5M the O-like is 3mm wide and the aspheric is 2mm wide. This seems like LOWER divergence for the aspheric unless you accidentally reversed these numbers.

[lasertack]

@planters

Compare the numbers to the numbers at aperture.

O-Like: 4,5x2mm - after 5m -> 11x3mm what means 1,4 x 0,2mrad

Aspheric lens: 4x1mm -after 5m -> 15x2mm what means 2,2 x 0,4mrad

Do you recognize the difference?

From 2mm to 3mm -> 50% larger (O-Like)
From 1mm to 2mm -> 100% larger (aspheric)

[planters]

So you stand by your numbers. Good. Then maybe the old devil of the definition of divergence is rising again. Look at the angle of the marginal ray. If I were to pass only the beam from the O-like lens through a cylindrical telescope and reduce its beam width to 1mm in the near field, what would you expect the spot size at 5M to be? What would you call its divergence? Similarly, double the aspheric to 2mm; what spot size @5M and what divergence?

What matters is what is the spot size going to be at some downrange distance typical of projectors, say 10 to 20 M, given the constraint of placing the near field beams in some arrangement that fits on a 5x5mm scanner. I believe this lens may be a substantial improvement over the O-like.

[lasertack]

@planters

I cannot tell you the divergence just by assuming 1mm beam after the telescope. What is the definition of "near field", what are the specs of the telescope, how did you collimate your diodes. Before you answer "far field", what does it mean? 10m, 20m or infinity collimated with a interferometer?

[andythemechanic]

Andy,

Forgot to add above. Your comment about the markedly different axis ratios of the emitters in multi-mode lasers favoring shorter FL primary collimators followed by correction optics is intriguing. If the FL of the collimator multiplied by the power of the telescope were a constant then it would seem that the trade off would be based on issues such as the ease of alignment, spacing constraints etc and the beam quality would not change. But, if you are correct, then the quest for better beams should seek the shortest FL collimators that have sufficient NA to accommodate the emitter. That seems to be the trend on the G71 thread where the consensus is to use a 2mm FL lens supplied by Dave @ LSPs. Thorlabs sells a 1.45mm FL asphere. What do you think about that?

The beam quality in general does not change if you don't clip the aperture. The product of beam with in a beam waist and far field divergence is a constant, thats the basic physics behind collimation. So yes at the end you can choose you lens combination for the easiest alignment or whatever other constraint you have provided that the "effective" focal length stays the same. E.g. a f=2mm lens with a 4x telescope is eqivalent to a f=8mm lens (provided there is no clipping of course). Regarding the very short focal length: Alignment tolerances are proportinal to the aperture size of the lens which is in turn proportional to the focal length (given constant NA). That is why for the 2mm lens it seems to be necessary to have a special lens centering mechanism.
It would make sense to use small focal length plus a telescope in the slow axis because at some ratio you can easily start to stack beams in the fast axis. E.g. if you use af=1mm lens the beam in the fast axis is only 1.5mm and ~0.2mm in the slow axis. Now you can expand 20x and stack 3x1.5 which gives you a 4.5mmx4mm beam with ~1.4mrad. This would probably be hard to align. For collimation of the fast axis however focal length down to 200µm are common.

Andreas