Combining diode lasers, a couple of questions

[NobleGas]

I'm putting together an RGB system using DPSS, 445nm and 638nm diodes. To match the visible power of the DPSS, I need to combine 10 445nm diodes (running at ~800mW); I plan to knife edge the diodes in 2 banks and combine the 2 banks using a PBS and a 1/2 wave plate, then correct the final output using an anamorphic prism pair. What I'm wondering is if I might be better to up the number of diodes to 12 so that I can knife edge them in a 3x2 array rather than a 5x1 array. What are your thoughts please?

[swamidog]

I'm putting together an RGB system using DPSS, 445nm and 638nm diodes. To match the visible power of the DPSS, I need to combine 10 445nm diodes (running at ~800mW); I plan to knife edge the diodes in 2 banks and combine the 2 banks using a PBS and a 1/2 wave plate, then correct the final output using an anamorphic prism pair. What I'm wondering is if I might be better to up the number of diodes to 12 so that I can knife edge them in a 3x2 array rather than a 5x1 array. What are your thoughts please?

you could easily drive the lasers harder and reduce the number of diodes you're using.

[NobleGas]

I'm using the A type 445nm diodes (from an XJ-A130), not sure how hard I can drive them for a reasonable lifespan.

[andy_con]

good luck getting 10x 445 diodes corrected using prism pairs ;-)

prism pairs expand the slow axis, so they will also expand any gaps between beams. so put a 0.5mm gap between beams into prism pairs and get a 1mm gap out

[hakzaw1]

you could easily drive the lasers harder and reduce the number of diodes you're using.

OR... sell those and find some M-140 or at LEAST A-140 and that will allow you to just combine a few-- there are lots of threads on safe max power for all kasio diodes(et al) at LPF-- the 130 is the lowest power they ever used---130s sell normally for $20/25 usd each=

hak

[logsquared]

3X2 would be better than 5X1 for the reasons that Andy described above. I would ditch the A series and get two NDB7A75. Waveplate and cube them. Then correct with cylinders. 7-10 Watts way less parts (collimators, mounts, mirrors etc.)

[norty303]

Change them for 4pcs of 2W M462 diodes which should get you 8W of more visible light by knife edging 2 pairs combined with a PBS/waveplate combo and corrected however you like.

[logsquared]

Change them for 4pcs of 2W M462 diodes which should get you 8W of more visible light by knife edging 2 pairs combined with a PBS/waveplate combo and corrected however you like.

I would prefer this over the 445 too.

Why cylinders and not prisms?

The NDB7A75 has 2X worse divergence than the A, M, 9mm, and M262. You would have greater losses correcting with prisms.

[planters]

If you are going to kneif edge a large number of diodes (>6) then the prisms will probably be a better choice because even if you PBS down to 4 and stack them the off axiis aberations and the alignment demands will make the lenses tough to use even if the losses are lower.

I agree with norty. Go to 4, 462nm diodes (or 445 if you prefer), PBS them to 2x2 and stack them vertically then expand with a lens pair as far as the scanners will allow. The smaller number is much easier to deal with and the power will surly exceed 8W and I would expect 10W.

[andy_con]

Which slowly shrinks due to divergence, right?

sure, but you will have a beam that wont fit on the scanner mirrors at aperture