what's inside the several watt laser modules?

[planters]

I mean limits of quality, brightness and power. My projectors tend to push these limits and you might get this impression when you watch the videos. I do not build for cost effectiveness or practicality, but rather to push these limits. You can produce a very nice and powerful module/projector without the extra effort and I think you might enjoy starting out by putting something together and getting a sense for what is involved.

[VDX]

... I'm using 445nm-diodes with 2W and 3.5W slightly above their rated currents (2W-diode with 1.8A, the 3.5W-diode with 3.3A) without TEC's or active cooling, only with good contact to aluminium coolers, which are tightly mounted to the Z-axis of CNC-mills.

The 2W-diodes stays at ambient temp, the housing of the 3.5W-diode with >3A warms up to maybe 35 degC (crystal maybe to 45 degC, what's near to the max. rated temp values) when CW-driving, but I'm using them mostly pulsed, so no need for active cooling with most applications.

Have other 9W-IR-diodes with the same drivers+housing+coolers with 8,5Amps max. current, where the drivers gets hot, but the diode stays at ambient temps too without active cooling.

It's the 'thermal mass' or passive cooling amount in relation to active heating phases, what's defining the limits for me -- but this can change, when switching to higher powered types ...

Viktor

[blue]

Okay, after learning some more and watching planters's videos, I think I have a pretty good idea how the stuff works.
But i still have some questions I couldn't find.

1) Which colimating lens to get? Do they come with the diodes or purchased separately? I've only heard about G2 and 3-element lens sold by DTR. Honestly no idea why anyone would get the 3 element lens right now.

2) What mirrors to use for knife edging? Will first surface mirrors work? Or are they mirrors made specifically for lasers? If yes, does anyone know what is the power loss with using a regular first surface mirror?

In some photos it seems to me like prisms are used instead of mirrors to change the angle of the beam by 90 degrees?



Am I seeing things?

3) And finally, it seems in knife edge setups the diodes are in a 4 diode host where there are screws in the front to attach the lens holder. Are those 4 diode hosts custom machined or are they some kind of standard commercial hosts?
In DTR's site he sells these kind of hosts:



I've seen some people put the colimating lens and diode in that and than put that host itself in a shaft holder, or larger host such as this one:
http://www.ebay.com/itm/Inner-diamet...item3cc65f599f


But does either make sense both for efficient cooling and cost saving if you can just get 1 host/heatsink for few diodes?

[planters]

Now that you are getting into it, I can help with more specific answers.
The collimating lenses you select depend on the wavelength diode you use. As a general rule, if you have the opportunity to use a single aspheric lens this will be a better choice than a multi-element in terms of its lower losses. For red, Dave's (LSP) 2mm aspheric lenses (mounted or unmounted) are a good choice as are the 4mm aspherics from Optima:
http://www.optima-prec.com/molded.htm#NEW%20LENS
There are surely others that would work well.
For blue, the G2 is good, but if you order DTR's 9mm 445nm diode, you can get it with the manufacturer's stock aspheric for a slightly higher price and it is excellent.

I exclusively use dielectric coated prisms as right angle reflectors for knife edging. They are not acting as prisms (the light never goes inside), but they have the finest, sharpest edges which is important because your beam will lie on the raged edge, literally. This fine edge allows you to position the beam closer to the edge and minimizes dead spaces between beams. Chips will degrade the beam and spill power. Lasertak has good prisms:
http://www.ebay.com/sch/*lasertack.c...p2047675.l2562
I did not see them on this E bay store, but he has had them for some time and would be worth contacting.

I would not use these pointer mounts and then in turn mount these in some thermal block/support. This decreases thermal conduction and increases the chances for interfaces to wobble. LSP sells individual brass and copper diode mounts that work well:
http://lasershowparts.com/store/inde...ewCat&catId=84

[blue]

Thanks.
I agree about the pointer mounts. It doesn't seem cost effective if getting a pointer mount then another mount to put the pointer mount in either.
BTW, do you use thermal grease for diode/diode mount, diode mount/base plate?

I understand your explanation of how dielectric coated prisms work and are better. Thanks. However, I might go with regular dielectric mirror to save cost in expense of beam diameter and near field quality for now.
First surface mirrors would also solve the issue of extra gap you mentioned but it seems these mirrors will kill few percent of the beam power each.

Oh and it looks like this item has the mirror prisms? http://www.ebay.com/itm/Spiegelschni...item1e859e055f
I used google translate and not sure if he is offering the complete base plate, diodes mount, mirror holders and prisms for $140 like in the images.

Forgot to say, DTR has a "limited quantity" of "G9" lenses on his site too, not only G2.

[planters]

Yes, those are the prisms. He speaks English. You might message him to see if he will ship the prisms by themselves.
Do not use grease between the diode and the mount. This is too risky. The grease could get on the diode facet. Grease between the mount and the base plate is pretty common and I do this.

[blue]

Gotcha.

About fixing the "far divergence" of the final beam, if I understand correctly cylindrical lenses are better at power loss and around the same cost than prism pairs, but prism pairs are easier to calibrate?

[VDX]

... someone experience with beam-shaping across mirrors instead of lenses?

Should be much better in respect to power losses and 'compressing' of the far axis ...

Viktor

[planters]

There are two important differences between prisms and lenses that are not frequently discussed. Lenses perform best on axis while the prisms perform the same no matter where the beam strikes their face. This means that as soon as you try to expand multiple diode beams with a lens, some of the beams will inevitably lie off axis. The shorter the FL of the lens pair the worse the deterioration of the off axis beams. However, If you do as I do and limit yourself to at most, a pair of PBS'ed diodes (overlayed into a single beam) then the lens provides a significant advantage. The laser coming out of a diode is very asymmetrical, but we always collimate with a radically symmetrical lens. No matter how carefully you adjust the collimator there will be a slightly different spacing that is optimal for each axis. This results in an inherent astigmatism in the output beam. The cylindrical lens pair can be spaced so as to counteract this astigmatism while at the same time performing its beam expansion function.

VDX,
I don't think mirrors would be noticeably superior and may even be worse. A major limitation of lenses is chromatic aberration, but with lasers, at one wavelength, this is not a problem. Good broadband coatings for reflection and transmission have similar losses below 1%. The challenge with mirrors is that they may need to operate off axis to accommodate the beam path on reflection and this will cause distortion.

[VDX]

... for mirrors vs. lenses -- I'm looking on symmetrical focus shapes and smallest focus diameters for best energy denities in material processing, not for the beam or farfield behaviour.

So I've tested with multimode-diodes and -bars with plane and concave mirrors - I've got pretty good focussing of the 445nm-diodes with a single concave mirror ... and a symmetric quadratic square in the farfield, while with the G2 lens the focus shape in the far-field after focussing was a line with maybe 8:1 aspect ratio ...

Viktor