Looking to build a DIY IR laser (or maybe LED) illuminator ... 980-1064nm

[LazyBeam]

So I've been gone awhile, but had built a few "lower" power lasers in 405nm and a 445nm in the past. I'm planning on getting into night vision and want to build an illuminator for it. I haven't decided on a method yet (LED vs LD), but i do have experience with home built lasers and I have no experience with LEDs... so I'm leaning towards LDs right now. I've been out of the game awhile and I'm having trouble knowing the best places to source these things so I figure I'd ask the forum here. Fair warning: here's a handful of questions. I don't know anything about building a IR laser and I know even less about building an LED flashlight. After a couple days of research, here's the questions I've accumulated.

Requirements:
Powered by 2 CR123A Batteries, optionally rail mountable.
I'm probably going to use this Surefire M600 Clone as a host.
Wavelength(s) over 950nm ... no emmisions under 950nm.
Round or square beam profile. Both axis near 1:1 divergence.
Continuous Duty Cycle - from fresh batteries to dead.

Option 1, Laser Diode:
I'm thinking a 980nm diode. They're reasonably priced and common. I'm thinking I want a diode power in the 100-200mW range. I plan on machining the head of the laser to be twist/adjustable; from collimated up to 275mrad or so. Alot of commercial products have illuminator divergences of 5-16 degrees. One thing I'm not sure about is the fast & slow axis to get a nice profile. Most of the 980nm diodes I'm seeing have about a 5:1 difference between the fast & slow divergences.
Anyone know where to find a 100-200+ mW 950-1064nm diodes that are more symetric? If I can't find a better diode, I'm assuming I need a cylindrical lens to increase the slow axis divergence to match the fast axis... where can I find an IR coated cylindrical lens and which model? What focal length/type? Do I just glue it to the diode heatsink between the diode and the collimator lens? For the collimating lens, again, where can I find one with 980nm coatings? Ashperic or 3-element lens? Any recommended drivers suitable for 4.0-6.4V input with about 1.5-2.1V/400mA output? I don't think either of my Lava Flexdrives can do this.

Option 2, LED:
So I don't know much about building an LED light. But I'm also assuming that because I'm STARTING with an LED flashlight host, swapping out the emitter and driver light be pretty simple. LED emmiters have a much broader spectrum than laser diodes so I'm thinking if I go the LE route that I want a 1000nm, 1020nm, or 1050nm diode in the 3-5+ watt range. Staying over 1000nm peak wavelength will keep almost all of the power OVER 950nm but I think even then I'd still want a bezel window that is also longpass filter that cuts on around 950-960nm to ensure there's no emission leakage over 950nm. What would be a good Diode, star heatsink, and single-function driver for this application. Any ideas where I could find 950nm range longpass filter?

 

[lasersbee]

LEDs will be cheaper and a lot safer... IMO

Jerry

 

[Cyparagon]

LEDs are cheaper, more robust, easier to drive, easier to heat sink, easier to mount, are ostensibly available in higher powers, and have a much more even illumination.

I've got a flashlight that I swapped in a 3W 980nm LED. (They're only a few dollars, and it's an intriguing device, so why not.) I can only tell it's on in pitch black room if it's pointed right at your face, and only if you're looking for it. Try it without the filter first. A 980nm laser will have some (spontaneous) emission in the NIR range too, so if it ends up being a problem, it's going to end up being a problem for both.

They're cheap enough, just buy a few and experiment. That's half the fun!

 

[LazyBeam]

That's what I was worried about with the 9XXnm LEDs... wavelengths below 940/950nm. So I think I'm definitely going with a long pass filter and if I can use a 1050nm emitter that will put more of the power curve over the cut filter cut frequency.

Is there a particular place to pick up suitable drivers for 4-6.2V input and whatever voltage a 5W IR LED needs? Or do I just use the driver that's coming in the white light I'm getting? The light I got originally set up for white 3W CREE R5 and 2X CR123A.

Also, what about a suitable TIR optic... any rules of thumb for selecting a TIR optic? I think the bezel diameter is 25mm... I'm not sure until it shows up in a couple weeks. Are TIR lenses even suitable for 900-1100nm or do they need special AR coatings?

Here's some 5W 1000-1050nm diodes I found ($24-57 ea):
https://www.aliexpress.com/item/32815069642.html
https://www.aliexpress.com/item/32842202891.html
https://www.aliexpress.com/item/33046937554.html
https://www.ebay.com/itm/5W-45mil-I...s-690nm-770nm-810nm-980nm-1050nm/223525031791

 

[Cyparagon]

Why TIR instead of standard reflector? Plastic and glass are both transparent at NIR as far as I know.

Why the 940nm hard cutoff? The linewidth of a 980nm LED isn't that wide, and 980 looks to be a lot cheaper than 1050.

 

[LSRFAQ]

Most cameras do not see anything longer then 930-940 very well. Unless your using NVGs, long wavelengths have poor response on anything with a silicon sensor. I'd avoid 1050 nm or longer, period. I frequenty work with 1064 and now 1080, and Silicon barely sees it.

Steve

see:

silicon responsivity - Bing

Intelligent search from Bing makes it easier to quickly find what you’re looking for and rewards you.

www.bing.com

 

[LazyBeam]

Why TIR instead of standard reflector? Plastic and glass are both transparent at NIR as far as I know.

Why the 940nm hard cutoff? The linewidth of a 980nm LED isn't that wide, and 980 looks to be a lot cheaper than 1050.

TIRs generally have a tighter beam throw on a same-size basis. I'm looking for 15-20 degrees out the front. It will come with some kind of reflector that I can try out just to see if it'd be good, but I'd like it to reach out over 100-150+ meters. I want to stay over 950nm (ideally even over 1000nm) and filter out any shorter wavelength because Gen 3 image intensifier tubes are basically blind to wavelengths over 950nm. Even the newer European 4G tubes start to fall off around the 1000nm mark.

 

[LazyBeam]

Most cameras do not see anything longer then 930-940 very well. Unless your using NVGs, long wavelengths have poor response on anything with a silicon sensor. I'd avoid 1050 nm or longer, period. I frequenty work with 1064 and now 1080, and Silicon barely sees it.

Steve

see:

silicon responsivity - Bing

Intelligent search from Bing makes it easier to quickly find what you’re looking for and rewards you.

www.bing.com

black silicon strongly absorbs light from 400-2500nm. For example, the Sionyx Aurora NV handheld camera has been shown to see wavelengths exceeding 1150nm. 1000-1050nm should be no problem for it. Moreover, it's monocle sized and can mount to a helmet much like a traditional NVG unit. Except it's 1/5th the price of a gen 3 NVG unit, sees in color, and can detect wavelengths the NVGs cannot. Resolution is a bit lower, but the other strengths and price make it an interesting project.

 

[LSRFAQ]

I have image converter tubes with long response cathodes, but that's just me...

Steve