Focus lasers with bending mirrors ?

[badger1666]

Has anyone ever had any luck focusing a farfield beam
By bending mirrors to compress the beam
Could the wide farfield stripe on a p73 be focused to say 30mm at 80m instead of 30mm at 25m which i normally get.
Just wondering.

[andy_con]

how would you bend a mirror without snapping it?

[The_Doctor]

how would you bend a mirror without snapping it?

Glass is elastic! More bothersome, it is a liquid... With time it will deform under stress so even if this works, it won't stay working. Got to wonder if in the short term this might beat anamorphic prisms for correcting a narrow-stripe multimode diode though..

[planters]

Sure, this can work.

What you are describing is a manually adjustable adaptive optics mirror. There are a lot of variables such as mirror thickness and shape (round vs square) and single point or an array of points. But, there is no free lunch. If the diverging beam of the p73 was just an astigmatic stripe then the right lens FL positioned in the expanding beam would correct this and a beam limited only by diffraction would result. Sadly this is not the case. The multimode output of the P73 is pretty awful as lasers go and the ray paths that end up in the periphery of the far field stripe originate across the area of the beam. Maybe if you analyzed the beam with a wavefront sensor and used this to design a complex high order lens that corrected the high order aberrations this would make a significant difference, but well...

Of course, if the beam off your curved mirror was large enough then the divergence could be reduced to achieve the 30mm @ 80M.

[logsquared]

I happen to be experimenting with this as we speak (type). Theoretically the curved mirror should have less aberrations. I few years ago I built flexible mirrors when the cas diodes first came out. This worked well until prisms were available for a reasonable price.

The problem with using a single mirror or lens for correction is as planters said "there is no free lunch". So, the curved element has to be pretty far from the diode to work. 300mm is about good for a .9mRad/ 5mm beam with the 71 diode.

[badger1666]

Hey eric are you saying the beam need to be fairly large to do this nearfield ?
I know i can focus the beam some distance from the laser, as i found out stressing the bounce mirror i was using
10 m away from laser ,but near the laser it did not work, if anything the beam looked a little larger

[The_Doctor]

Is there no benefit to a longer mirror, held obliquely across the beam so the 'impact' area is a very elongated ellipse? That ought to get you a lot closer to the source, and more freedom in mirror thicknesses, etc. Seems to me that you could put one edge right after the lens that way to get very compact layout.

Edit:
Judging by the lengthening silence, I'll take that as a 'yes' on whether there is a benefit. And that at least one person is currently exploring it to see if it continues to justify initial impressions...

[planters]

I do not think this is worth the effort.

If you had a lot of time and wanted to experiment to see if you could iteratively tweak the aberrations this might be interesting but too bulky for any type of projector. The way I would do this is to use a diverging lens to enlarge the beam to as large as I could re-collimate ie the size of the second positive lens. Send this 100mm...150mm diameter beam to a decent first surface mirror that I attached a 5x5 to 6x6 array of push /pull screws to the back of (glue the flange nuts to the back surface) and mount the screws in an aluminum plate tapped for this array. Re-collimate this beam at a more useful diameter.

The reason you need to enlarge the beam is to allow the individual actuators to effect different areas of the beam. These mult-mode beams are really like a cluster of "kinda parallel" individual laser beams not unlike a high count knife edge array. And similarly, the way to correct the far field of such an array is to address the individual beamlets.

[The_Doctor]

It's always worth the effort. That said I didn't follow up because there are all kinds of curves, catenary, parabolic, etc, that are similar until you need the fine detail, and the devil always shows up in that kind of detail. Unless you can carefully control the bend radius everywhere on that mirror, the results aren't likely to be worth it. In the sense of final results, that is. But the effort of proving how far it will go wouldn't be wasted.

If I did do it I wouldn't expand the beam though. That;s because you then have to deal with aberrations of curvature in both axes. Granted, bending in one tends to straighten the other in a plane bent to cylinder, but even so... As far as getting benefits from divergence at all, why do that if you can expand the beam's occupancy on the mirror plane neatly in one axis just by having a very shallow incident angle?

I still wouldn't do it though, it's sort of elegant in notion but ugly in practise. Can't separate pointing accuracy from divergence adjustment, for one thing...

[The_Doctor]

These mult-mode beams are really like a cluster of "kinda parallel" individual laser beams not unlike a high count knife edge array. And similarly, the way to correct the far field of such an array is to address the individual beamlets.

Ah, well you saw that other thread where I asked about that. This is my point, is there really a great deal of incentive to use multiple single mode diodes if it is easier by far to use one lens and two prisms on a multimode (whose other axis behaves like single mode diodes)? After years, there is still no consensus. And if it were purely dowqn to the physics there should be, these are issues as old as Newton, so what's missing? Why is it still a debate? I know that laser diodes are new(ish) but old enough to resolve this, I'd hoped, anyway..