How to create super bright laser beams

[planters]

That is the real competitive alternative to the metal bar thing. I think the only place I would differ is that in both cases I would use a single housing because this allows a lot of savings in work, cost and space especially as you move to higher diode counts. But, the individually controlled diode approach is attractive and its how I had first envisioned this set up. This approach requires a TEC, probably a two to three stage module to accommodate the coldest diodes and a controller for each diode. You could probably avoid the computer controlled window interface part (fun, but extravagant) as long as the driver was closed loop. It would be nice to adjust each diode to produce an even wavelength interval and this would undoubtedly require the freedom to adjust for an uneven temperature interval. The stabilization time should be much quicker as the controller would max out the TEC's that have to get down further.

One issue might be cost. Let's say you had 8 diodes, 4 in each set and each set to be PBSed. That requires 8 TECs and 8 controllers. What would be a good source for each and how much would it cost?

Thinking about it did bring up another question.... When we are thinking of light entering the prism we are generally (me anyway) thinking of the text book drawing of a white line going in and the colors coming out. In fact the diode beam is not perfectly collimated. Different Fl collimators may have any effect on the way the beam is refracted due to the spread of the beam when it interacts with the prism faces???

I looked at this and it is interesting. With a SF-11, 60 degree apex prism, a 2mrad correction of two beams that are 4nm apart, the angle of incidence needs to be approximately 51.6 degrees. This angle is only 1 degree from where total internal reflection occurs. If the incident beam is diverging at 1mrad then the prism causes the divergence (obviously, in one axis only) to be amplified 4x and similarly if converging by the same amount then the convergence also amplifies by 4 X. This is not chaotic, its not a blurring or smearing as the outer rays angle faster than the inner rays and so this acts as a focusing effect. This can be totally eliminated with the correction lenses that would be used in any case to correct any anamorphic effect or to size the beam to match the other lasers or the scanner.

[planters]

How do you determine where on the conductivity spectrum you want to be? Somewhere between copper and, say, Delrin, there must be a sweet spot. You could probably find a lower conductivity solution, but at some point the diodes are effectively isolated.

I agree. Based on my calculations, I think the stainless steel is not bad. If you move from this point up to say bronze and the conductivity increases x 2-3 the heat flow increases by the same amount and the achievable differential decreases. The curves that are on the link I provided show how much. You can adjust the cross sectional area and this will effect flow, but also there is the questio0n of stiffness to keep in mind. The lower the bar's mass the quicker equilibrium is approached.

I am still drawn to the idea of individual loops as I discussed above if the bar gets too touchy. Maybe see how it goes with the bar I have coming and get a different bar if the result seems to show where to go.

[planters]

Logsquared,

You brought up two good points I'm looking at. Firstly, the injection current red shifts the diodes. I had originally discounted this because if you drive all the diodes at the same current the "package" should remain pretty tight. I think this red shift is not going to change for the individual diodes, but the deviation in the prism as opposed to the dispersion may be a problem. The whole beam may shift relative to the other colors as you modulate the red power. Secondly, your idea of a telescope to compress the path length is a good one. The alignment and losses in four lenses as opposed to a couple of folds between mirrors could be a better alternative.

On the first, I am going to PM Steve and run an idea past him before I go on and on about something that may be silly.

[logsquared]

The whole beam may shift relative to the other colors as you modulate the red power.

This is clearly evident in my test set-up with my SM RGB. From threshold to full power the red beam deviates aprox. .5mrad. The G and B do as well. I am not sure if I am seeing the effect of localized heating or current induced shift?

Secondly, your idea of a telescope to compress the path length is a good one.

Not sure if its feasible or not? And actually, I think it would be reverse of what I previously said. PCV should go before prism and PCX after. Its too complex for me to rationalize without drawing a ray diagram and doing math. That's no fun


I really look forward to seeing your experimental set-up. Even if its not able to perfectly overlap the FF spots. It should show show us where we need to go from there. And also allow for experimentation with telescopes and such.

[kecked]

The telescope is what I was trying to get across by modifying the prism surface.

[planters]

An independent telescope would have higher losses, but it would be stock and independently adjustable.

This is clearly evident in my test set-up with my SM RGB. From threshold to full power the red beam deviates aprox. .5mrad. The G and B do as well. I am not sure if I am seeing the effect of localized heating or current induced shift?

I think it is likely if we really want to nail quality we will find the modulation drift of all colors is an issue that has been over looked. The DPSS lasers (the good ones) include pointing stability as a spec and the blues are frequently "corrected" with an anamorphic prism as well. I am going to review this with steve so that if the pointing drifts more than trivially, it can be prevented

[flecom]

That requires 8 TECs and 8 controllers. What would be a good source for each and how much would it cost?

nothing a micro controller like an arduino couldn't do fairly easily

[kecked]

The picaxe runs 20 MHz and has three hardware pwm, adc 10 bit, and plenty ports running off basic. Also costs 2.00 and the program softwRe is free. You can add a display and set point pot as well. Three chips and your done with 9 lines. Same setup I use to run rgb leDs just instead of a pot for control you use a thermocouple.

[planters]

I was thinking that the slickest system would be a central control like a single screen that presented all the temps and from which you would adjust the temps. The Picaxe, Arduino or another Pic or Atmel controller would all have to talk to the current or voltage control unit (s). The typical thermistor/pot/current output devices that we use to regulate diode temps in projectors are pretty much there, but 8 or more and the lack of a slick display that avoids independently measuring each diode is nice. Where I'm stuck is the unit that switches a few A at say up to 12 V under the control of these central devices.

[logsquared]

Where I'm stuck is the unit that switches a few A at say up to 12 V under the control of these central devices.

All that is needed is a Fet with a resistor on its gate pin driven by the controller. Supper easy! Especially if using one of the tab isolated packages.