How to create super bright laser beams

[dave]

Its not like you even need to remove it?

Undo the lid screws, remove lid, tweek the alignment, refit lid, replace screws

[planters]

OK, back in the States and back on topic.

Before I go into some details, let me start with...it works! I will post a video to demonstrate this, but the prism technique does what it is supposed to do, without any surprises. The losses are about 8%/prism with a wave plate to rotate the polarization and about 15% per without it. With the P73 diodes and their substantial divergence, especially with the 2mm aspherical collimators, it is necessary to knife edge them so that the slow axis is vertical. The stripes need to be vertical. I learned this the hard way by starting the other way; the divergence screws up the dispersion and soon prevents the beams from fitting on the 25mm prisms.

The set up is pretty simple (NOW it is after spending hours wrestling with positioning the optics). After carefully knife edging the collimated beams to a target 12M away, I overlapped the beams so that they looked like a single stripe on the 12M target. Given that they were each about 2mm wide at the knives, the convergence is about 200urad, so essentially parallel. The cooler is switched on and allowed to stabilize and this takes about 10 minuets to fully stabilize. The first prism is inserted and rotated to bring the knife edge beams to a convergence 300mm away. At this convergence point the second prism is inserted and rotated to reestablish the overlap at the 12M target. That's it. The three beams are now as wide as a single beam within the projector and there is no discernible deterioration at 12M.

The exciting thing is that this works despite the really poor performance of the cooling system. The gradient bar is awful! My coldest diode is @ -10C. Diode #2 runs @ +9C and the warmest @ +15C. The warmest diode also has a little manufactured red shift @ .25nm as well. So, with an average delta of 13C the system works well. With some proper cooling, 5-6 diodes might be combined. I will have to set up with greater precision to be sure of the measurements, but there does not seem to be any modulation induced beam positioning shift at 12M. It is at the least small enough not to be obvious.

[CDBEAM]

WOW....Simply put....Another " Grail Moment " !!! The pioneer's on this site, whom have NO vested interest in one approach vs. another.... pump out a new technique to step us all into a previous " Underscovered Country " Bravo !!!

The best driving force continues to be the audacity to forge ahead, and not let conventional wisdom / prior art be a lead anchor !!! Study it, Think it, Design it, and Build it !! And sometimes, that spirit just pushes our hobby a little forward !!! Passion propels progress !! Exciting times indeed. Thank you for cutting a path !! Look forward to the Show & Tell Video !!!

[logsquared]

300mm prism to prism. That's fantastic!!

So the knifed beams before the first prism are ~6mm and overlapped after the last prism ~2mm? And the FA divergence is the same as before the prisms?

Losses are lower than I expected also. SWEET!

[logsquared]

I have another question....

Did you turn off the power to the "gradient beam" to see what happens to the laser beams/ spots?

[planters]

I'll answer your last question first. I've had the power on and off, on and off, again and again. The beams align and then spread out each time to the same stabilized relationship. It takes about 10 minutes to stabilize at full cooling power, but by reversing the polarity to the TEC, I can get the diodes to stabilize within 4 min at room temperature.

So the knifed beams before the first prism are ~6mm and overlapped after the last prism ~2mm?

This is not quite right. There is an anamorphic property to the second prism and so the beam coming out is also close to 6mm, but the divergence is 1/3 as great as well. You will see this when you compare the width of the far field stripe to the waste beams that reflect off the second prism and travel the same distance to the target screen. Most of the anamorphic effect happens in the second prism which also has a steeper angle for operation than the first prism (haven't figured this out yet). Where you best see the unmistakable effect of the convergence is when you look at the three, distinct stripes just before the second prism and before the cooling begins. These stripes continue, through the second prism, to slowly converge to the distant target. As the diodes cool the three stripes merge into a single stripe just before the second prism and then...remain converged all the way to the distant target.

[logsquared]

Very cool. Only thing that seems strange is the differential between the first pair is 19deg and 6deg on the last pair. Seems like that would cause a problem. have you checked them with your spectrometer to see if the wavelength diff is the same?

So are we needing a temp controller to move forward?

[planters]

This differential is a little curious to me as well, but there are some mitigating factors. My measurements could be a little off, but I am assuming this could go either way. The fact is that the delta lambda/delta T seems to be non-linear with the biggest red shift/degree at the higher temps and that the warmest diode already has a 0.25nm manufactured red shift (I measured this). Also, the spacing between the stripes is not equal.

We absolutely need a temperature controller to move forward. My set up is not sloppy and I did my best to design a reasonable gradient bar. I think this approach is just not going to improve a lot. The long stabilization time is inconvenient. My poor control of the temperature intervals probably means that a controller does not have to nail a precise temperature interval, but rather it needs to be adjustable and then hold a fixed temperature. If each TEC/diode could hold a particular setting then the resulting system will probably work well.

Oh, and I made a miscalculation. The losses /prism are 12% with the wave plate and 18% without. Still, not too bad.

[logsquared]

I have limited time over the next few weeks but could probably work on something in my spare time. Others have mentioned working on one. Not sure if anyone has committed to actually doing it?

My idea is to build a simple prototype to use for this project specifically. It would have 4 bi-directional channels, thermistor sensors, and a simple LCD display. Not sure how I would implement input at this point. It would use the picaxe 18m carrier board for simplicity and a sub-board with the driver chips, caps, terminals etc. lemme know what you think.

[planters]

I think 4 channels is a good number. Do you mean you are not sure if you could use say a typical 10k thermistor input? Or, do you mean the multi-amp input to the TECs? I think it should allow at least 15 V and 8A per channel. The coldest diode is where almost all the power will go. I wouldn't be surprised if the quad drew 10 A with 7A to the coldest diode. I also like the idea of a controller board and higher power, modular daughter boards for each bi directional channel.