Knife edging, anamorphics and FL. A little advice.

[onryo]

Hi guys,
I have been hanging out on LPF for over a year or so and thought I would drop in here for a little advice and introduce myself. I am working dicking around with a eight diode 445 module I am making.

The ratio of beam aperture and divergence has been driving me nuts all day. Seems to all come down to the diode (M140) properties. The ratio is a constant ... so it seems. The O-Like FL 8mm is quite nice when put though anamorphic prims at Brewster's 56.775 deg (hoped to get a free ride there) x2.3. About a nice 0.7mrad. The down side is the coated BK7 prisms (30 deg) insist on eating 8-10% of my light. Yes they are for coated for 445nm and BK7 glass. Is this normal? Have heard of people only loosing 2-3%. Also wish there was a bigger diameter version of the O-Like lens. They (O-Likes) are clipping tons of light.

My G-2 FL 4mm lenses pass about 18% more light then my O-Likes according to my Ophir. They don't seem to clip my diodes. The FL @ 4mm is not so short that my diodes need to hit them dead center and I don't still don't see that many aberrations. Nice thin beam and ... oh yeah they have ultra crappy divergence! Gee ... go figure.

To keep this short I need to combine eight M140s. Would like a < 0.8mrad. with total aperture of 5x5mm or less and be able to pump out at least 10 watts. Is this even possible? Was thinking the classic 2 rows of 4 diodes being knife edged on prisms/mirrors (depending on price) using Rochester's “405-G-1”. IIIIx(IIII waveplate) into a PBS and then a pair of anamorphic prisms. I am willing to try cylindrical lenses … or heck any setup. This setup is all TEC so heat is not really a problem. Also wondering if the lens the comes with the Nichia NDB7875 would be good for this?

On a side note I noticed that when I chilled one diode to 8 C and drove the other one hard at 1.8A the stripes II -> I I shifted places as they passed though the anamorphics. I knew this would happen but just wanted to see. Hmmm, I wonder if I can calibrate some lines on my wall by lambda and use this as a quick way to test other diodes frequency?

Would you mind sharing what works best for you? Any advice on anamorphics with a suitable magnification? I am always open to any advice.

All the best
Onryo

[planters]

This is an interesting post.

Although you have made significant progress with this work prior to your review here it is remarkable that much of what you have presented has been discussed over the last year or so in various threads on this forum. Your numbers are remarkably similar to the general consensus.

I have built several high count 445 arrays and I strongly recommend staying away from the cylinders for beam expansion because when you are passing more than two beams through these lenses the off axis aberrations are significant. I also recommend sticking with the O-like lens in that the additional loss (my results are more like 6% greater than the Rochester lens) is worth the clipped, lower divergence beam. You might want to consider knife edging AND stacking each 4 diode group to allow greater expansion in the prism. This means 2 horizontally knife edged beams lying above two additional, horizontally knife edged beams. The target should be the largest square array that will fit the available scanner mirrors (assuming that is where you are going).

Now to think outside the box. What about four 9mm 445 diodes, cooled and driven to 2A and 2.5W each? With this you can easily best you goals with something like 4x4mm, 0.5-0.7mrad and 10W.

Even further outside the box the manufacturing variability between the 445 diodes means they have a wavelength range of as much as 7nm from one diode to the next. If you select the order of the diodes left to right as they enter the prism pair then the wavelength difference can allow a tighter packing in the prism than is possible if they were of the same wavelength. In other words, the wavelength dependent deviation can be used to counter the effect of knife edging the beams so that they are actually converging on the prisms. I described this in a thread about a year ago and I am using it in my large projector with a 12 diode array. It's tedious, but it works. If you have gotten this far I think you will see how this might operate.


I would strive for beam quality over power. It is always more impressive and the difference of 10-15% is hardly noticeable.

[onryo]

Thanks for the advice planters!

What about four 9mm 445 diodes, cooled and driven to 2A and 2.5W each? With this you can easily best you goals with something like 4x4mm, 0.5-0.7mrad and 10W.

Sounds perfect! The idea of a 4x4mm beam at 0.6mrad doing 10W+ would be well worth the cost to build!

Looks like Ill be ordering four 9mm diodes from Jordan! Would you recommend the O-Likes even for this diode? I have heard the lens that comes (glued on) with the 9mm is quite nice. Some say better then the G-1? I have never used this diode or lens myself.

To get the above should I just knife edge IIxII(waveplate) into a PBS then pass anamorphic prisms at x1.8 (53 deg BK7) though O-likes to magnify the 4x2mm-ish beam into 4x4mm-ish? Sounds almost too easy =). Did I miss something? Not sure what this beam looks like after the O-Likes but I would think something like the M140s. 4.5x1mm-ish? Maybe I can find some photos on PL.

Ill definitely be looking out for the variations in wave lengths to bend the beams into each other. Might even intentionally try modulating local TECs to get the beam closer. Got a few ideas.

Is it normal to lose 10-12% in the anamorphic prisms. Like I said they are BK7 coated for 445nm. Not sure about the brand. Are the ones Bob or Dave sells better then these?

I am including a spreadsheet I compiled (10 months ago) from some tests Lazeerer did at room temperature. Can change the values around for any diode if that is practical. Maybe his LPM tests a bit low? Interestingly the 9mm voltages seem a bit higher then across the M140. I use to think the 9mm and M140 die were the same except for the little cap. Not so sure now. I know a lot of people that get 3 Watts out of them today.

All the best and thanks for the feedback!

[planters]

Looks like Ill be ordering four 9mm diodes from Jordan! Would you recommend the O-Likes even for this diode? I have heard the lens that comes (glued on) with the 9mm is quite nice. Some say better then the G-1? I have never used this diode or lens myself.

This I don't know. I have not tried this lens, but I have heard good things about it as well. Do you know if it fits the available lens barrels? If it does then I think you should test this. Otherwise, you risk becoming just an assembler of parts and not an experimentalist.


The 9mm beams and the M beams are nearly identical. Yes, knife edge two diodes and do this for each pair and then wave plate one pair into another. Yes, it is that easy. Where you will gain is in the care with which you knife edge. You will loose 10 -12 % with the prism and depending on the quality of your wave plate and cube you will loose anywhere from 5-15% here. These two components can very a lot in quality. I have not seen any wave plates from hobbyist sources that are very good (but thy are cheap). You need true zero order to get your loses down into the low single digits and the cube should not be broad band AR coated, but instead narrow band and centered close to 445nm.

Don't bother with local TEC cooling of the 445s. They have a delta of something like 0.05nm/degree C. Rather, cool them all aggressively as they have close to 0.5-1% increase in efficiency/degree C.

Keep us posted.

[Stiffler]

Hey Planters, why not combine the beams like this: || = so that you end up with #? I know this means correcting the 2 pairs before the cube, but wouldn't this make a more symmetrical beam?

[swamidog]

That's what I did in the 10 x 445 build I worked on a while back.

Hey Planters, why not combine the beams like this: || = so that you end up with #? I know this means correcting the 2 pairs before the cube, but wouldn't this make a more symmetrical beam?

[planters]

One advantage would be the elimination of the wave plate ( the really good ones cost $200). The problem as I see it is that for me knife edging is easier horizontally because by slightly loosening the knife mirror (or prism) mount I can very precisely slide the edge into the beam. Also, I would be concerned with using a prism pair vertically. Normally, you loosen the prism mount a little and slide the prisms around on the base plate ( a tooth pick works well). The way I achieve symmetry with the final product is to beam expand with the prisms the two 4mm high ll lines to a 4mm wide overall width. The beam expansion is in this case as much as possible whatever that ratio might be numerically.

How did that method work out with the 10x build? Any tricks? Any problems?

[swamidog]

We did not do any beam correction.










However, I did a small two diode rig with two sets of prisms and a PBS cube. It works very very well and is currently the blue engine in my projector.



At 40 feet:






Ob note: I wish I had planter's wonka factory...

One advantage would be the elimination of the wave plate ( the really good ones cost $200). The problem as I see it is that for me knife edging is easier horizontally because by slightly loosening the knife mirror (or prism) mount I can very precisely slide the edge into the beam. Also, I would be concerned with using a prism pair vertically. Normally, you loosen the prism mount a little and slide the prisms around on the base plate ( a tooth pick works well). The way I achieve symmetry with the final product is to beam expand with the prisms the two 4mm high ll lines to a 4mm wide overall width. The beam expansion is in this case as much as possible whatever that ratio might be numerically.

How did that method work out with the 10x build? Any tricks? Any problems?

[onryo]

Today has been a lot of fun!

I do have WP but as planter more or less said “you get what you pay for” and I did not pay hundreds of dollars for mine. Swamidogs || = → # seems like an excellent idea. The x 2 mm stripe being on each axis being magnified at 53 deg x 2.0 to give the 4x4. Skipping the WP. Simplicity is beautiful.

Skipping the WP and starting with first one M diode I found the beam to be more like 4.5x1mm-ish with just a O-Like. Hard to tell due to brightness (even with OD7 goggles) and we are talking about tenths of a mm here. I passed two of them knife edged || though the anamorphics at brewster's x2.3. Nice 4.5x4.5mm square. Looked like I had about about 0.8mrad. Is that shabby? I don't have much to compare with. Maybe semi crappy optics? Did not get that 0.6mrad I so badly want. Yes in typical LPF style I did drive the M diodes hard at 1.85A and TEC'd like nuts.

After an hour or two of twisting, pounding, cube flipping, diode twisting and praying at my PBS to stop leaking light, I found a post by planters. Looks like I will be ordering a a narrow band PBS designed for 448nm from JML Optical. Part number CPO21361. Thank you for that post! Thought it was just to press a button and get the order done with. How do I get hold of that thing. 105 USD sounds well worth it!

Unfortunately I only have one pair of anamorphic prisms atm. I'll be ordering two decent pairs of anamorphics in the next few days. This is probably the most subjective question on earth. Can somebody recommend a relativity decent priced pair that are coated for 445nm as planter mentioned? Once I got this all sorted I'll be dropping in four 9mm using Swamidog's II = -> #.

Again I want to think you guys for the input.
All the best!

[planters]

I get my anamorphic prisms from Edmond Optics. They are over priced (I feel) @ $140 but they are coated for 445nm and are equal to the biggest I can find @ 12mm minor dimension. JML has changed their business model over the last year and are going after the custom market, however I understand they are selling off their existing inventory. So, contact them and ask if they have the part you need or if they have a similar cube in slightly different dimensions. An alternative supplier I have been using is Lambda Research Optics. They are just a little more expensive, but the products seem as good.

To get that last little bit of improvement in divergence you will probably find it in your knife edging. Use the 7mm prism based mirrors sold by Lasertack (PL member). These have an exquisitely sharp/smooth edge and allow the two stripes to be packed a little bit tighter. This in turn allows a little more beam expansion.

Ob note: I wish I had planter's wonka factory...

Nice Trilobite. Be very careful with that animating ray.