How to create super bright laser beams

[logsquared]

If the diodes are too close in wavelength you may see injection locking of any two adjacent diodes. Nothing a quarter waveplate fragment won't cure.

Steve

That could be a pain. is that caused by stray light from one diode "seeding" another?

[planters]

Steve,
I don't think there will be much feedback from the optics after the combining prism. I would be surprised if it was more than a fraction of 1 percent especially if any flat optics in colimated beams are tipped slightly.

I am not sure what the line width of the SM diodes is. But if the dispersion is around 20-25urad per nm, then even a 5nm line width should only increase the size by 30% or so.

Yes, and that's what I would have thought. So,that's why I got nervous when you described an increase in the far field spot of four fold.

I think the number of same diodes able to be combined will have to due with the variation in the sample we get

This might cost a bit, but I would plan to order a handful and measure the wavelength of each. Then, order them along the brass bar or mount them on the TEC's (still a work in progress) according to their wavelength, so that the temperature variation works WITH the manufacturing variability, not against it or independent of it.

I don't think temp regulation will be as critical as one might think at first. Especially with around 4 diodes. Slight variations will not cause the spots to become completely misaligned.

I agree with you here as well. Where you might get into trouble is by trying to create a huge amount of temperature induced deviation to accommodate badly knife edged beams. This reduction of temperature induced deviation will reduce the smear as well even if it is a small effect.

If the anamorphic effect needs to be countered this can be achieved with a cylinder pair of low dispersion glass (what you normally use anyway)

I will test the P73's as well because they are so powerful. (1,2,3,4) -PBS- (1,2,3,4) into a SINGLE STRIPE, expanded and spatial filtered would put 10W into a far field beam half the size of a stock Laserwave DPSS beam. If more powerful single modes become available, so much the better.

I do agree that this approach will probably work and if so, it has legs. It will be available to increase the brightness of red laser beams for a long time to come. This is not limited to one particular diode that may go extinct.

[planters]

I've ordered the diodes.

I've worked out a preliminary design for the mount. The thermal conductivity of the metal needs to be pretty low so that a practical cross sectional area with enough material to thread into and with enough bulk stiffness, can be used. Mg bronze or Al bronze look good with about 1/2 -1/3 the conductivity of brass and 1/4 - 1/6 that of the Al base plate that this module will undoubtedly be sitting on. The rectangular bar with a cross section of 1cm thick by 3cm wide and 15cm long will conduct 12.6 W with a delta temp of 70C between its two ends. With 5 diodes each contributing a little less than 1 W along this bar, the thermal load should be around 17W. Each diode will see a 14C step interval.

I will mount each end on the surface of a two stage TEC device, but with one of the two "coolers" flipped so that it will act to heat the one end. Obviously, the flip can be electrical as opposed to actual. This means the thickness of the two TEC's will be the same and this will make the clamping much easier. The 15cm length will allow 30mm for each of the diode mounts to be fitted. The two coolers I have selected should easily permit a 100C differential along the bar.

http://www.customthermoelectric.com/...Q_spec_sht.pdf

If a greater temperature differential between the diodes is required I can push the TEC's a little harder or drop the diode count to 4. The manufacturing variability can help, but its significance is uncertain. Also, one issue with the manufacturing variability is diode replacement. If, say diode #3 burns out, I may not have a close replacement for that particular wavelength and I don't then want to play musical chairs with the others.

The big question is what is the bandwidth of these diodes?

Maybe I was wrong and a three stage cooler with a powerful single stage added under the above cooler on one end only and the now warm base plate as the "heater" on the other would work better.

http://www.customthermoelectric.com/...W_spec_sht.pdf

[planters]

I've elected to go to stainless steel for the "thermal bar". I hate machining SS, it's naaaaasty, but is has an even lower thermal conductivity, it's stiff, cheep and the only machining is the small threaded holes for the diode mount hold downs and one large clearance hole on one end only to allow the assembly to be bolted to the base plate and press down on the TEC on the opposite end. I will limit the cooling to just the two stage module I listed above. This should be enough and keeps the set up simpler and the power requirements more practical.

Last night, I took a look at an old 650nm laser pointer that I had in a drawer. I'm pretty sure it's single mode and I got a 1nm line width. This is near the base of the curve and not up at the 50% level as I assume I want almost all of the spot envelope to shift over its neighbor. I then tested a He Ne and got about 1/2 nm line width. I suspect that I'm at the limit of resolution of my spectrograph. If so, then I'm hopeful that with a line width of less than 1 nm for the Oclaro 170 mW single mode diodes then a delta of 15C plus whatever I get from the manufacturer will be plenty.

I also looked at one of my old videos and the P73's look to have no more than a net 2 nm line width and so I believe it may be possible to overlap at least three and maybe 4 of these as well.

One powerful tool to make this whole project work as well as open the door for larger numbers of overlapped beams is the arm length to the prism. As the arm lengthens, given a fixed knife edge spacing between the beams, the number of diodes that can be overlapped rises almost without limit. I'm trying to work out a practical method to fold this path that is stable, easy and allows a significant lengthening within a limited space. I'm not sure of the axis of polarization of the diodes, but I may introduce a single wave plate in the converging beams as only one axis of polarization into the prism has almost zero loss.

Another advantage to this technique is the absolute elimination of knife edge spacings between beams. Anyone who has knife edged multiple diodes knows that no matter how careful you are there is a point where you loose power or introduce a black gap between neighboring beams and this inevitably increases the near field beam dimensions beyond what they should be "on paper" and is also visible to the observer in certain circumstances. It is not really attractive.

For the set up, you will bring the thermally modified, knife edged beams to a single overlapped spot at the approximate center of the prism's intended position, much as you do with a PBS. Then the far field spots will be spaced out along a line. You insert the prism into the beam and rotate it to converge all the far field spots as close as possible. Then, you fine tune the overlap with the individual knife edge prisms. The microscopic movement this causes in the prism will be proportional to the ratio of the distance from the knife edges to the prism and the kneif edge prisms and the far field spots. Piece of cake.

[tribble]

Wait, why are you trying to minimize thermal conductivity of the bar providing the thermal gradient, or did I mis-read somewhere? If the thermal pumping at the ends of the bar should dominate, I would think you would want at least medium thermal conductivity to sink the heat generated by the diodes.

[planters]

No, the thermal gradient from one end of the bar to the other is related to the conductivity vs the heat flow and given a finite amount of watts that any cooler can produce, then the maximum thermal gradient occurs when the conductivity is lowest. I looked at the relative conductivity of practical metals allowing enough materiel to be screwed into and stiff enough for the mounts not to move much while say focusing. The gradient for the cooler I looked at needs to be allowing somewhere around 10 W or hopefully less to be conducted to get about 60C from one end to the other. Copper, for example, would require 150W for the same cross section and length.

The diodes each add about 1/2 W at each of their contact points along the bar and so end to end heat flow will dominate.

[logsquared]

I'm trying to work out a practical method to fold this path that is stable, easy and allows a significant lengthening within a limited space.

I think "zig-zag" between two quality dicroic mirrors may work.

For the set up, you will bring the thermally modified, knife edged beams to a single overlapped spot at the approximate center of the prism's intended position, much as you do with a PBS. Then the far field spots will be spaced out along a line. You insert the prism into the beam and rotate it to converge all the far field spots as close as possible. Then, you fine tune the overlap with the individual knife edge prisms. The microscopic movement this causes in the prism will be proportional to the ratio of the distance from the knife edges to the prism and the kneif edge prisms and the far field spots.

That is exactly how I envisioned it.

The knife edge mounts will need to be 3 axis I'm thinking

Piece of cake.

We'll see

[planters]

I think the zig zag would be the most compact, however the adjustment of either mirror will have a multiplicative, not additive alignment effect. What about the knife edged beams, as I envision two sets facing away from each other and exiting opposite sides of the box, sent toward the converging prisms and rather hitting a retro mirror? Then, the beams head back toward the center of the close end of the box, hit another mirror and once again head toward the converging prisms, but this time entering them and so on to the PBS, dichro etc? A length of 500mm would be pretty easy to fit and for 1mm beams would only require 2mrad of correction in the prism.

The knife edge mounts will need to be 3 axis I'm thinking

I was planning to use standard Flex mounts with X/Y tip and tilt. You would slide them along their base slot to control the intercept position. This is how I would usually do it. Do you see a special issue?

Piece of cake? I really think so. I don't see any deal breakers other than a possible broad line width. DTR is fast and I expect I will be able to test these in a couple of days. The thermal box is done and later tonight I will finish the Al mounting block for the "thermal bridge". I have the TEC and the stainless bar was added to a McMaster Carr order. I ordered a few more 7 x7 mm red prisms from Lasertack (already sent) and I have all the other optics and mounts in stock.

The reason I'm so excited is that the potential is great. For 2-3 diodes this should be pretty easy to implement with a few mrad of convergence and a delta of 20C between the diodes. This woud be equal to or more powerful than PBSing. With care, even more diodes could be overlapped. The beam parameters might be unachievable by any other means. What are the near field beam dimensions and divergence of an OPSL?

[logsquared]

I was planning to use standard Flex mounts with X/Y tip and tilt. You would slide them along their base slot to control the intercept position. This is how I would usually do it. Do you see a special issue?

The lateral position of the knifed beams is the "fine tuning" adjustment of where the far field spots will overlap. Actually its the angle it enters the prism. But that is dependent on the position. I envision fiddling with aligning a loose mount might be a PITA.

[logsquared]

Any idea how long it will take to reach equilibrium of the thermal gradient in the bar?