16mm star for IR (infra red) illuminator. Help wanted.

Maybe a layer of Kapton tape? Frown That might be thinner and offer more protection than a thin layer of epoxy?

Probably not a bad idea - lay down a bit of kapton tape, then a dollop of AA, then mash the LED down hard on top of it.

Before I start my apologize for late reply,

Thanks on help guys...

Well i need single 1W IR 850 16 mm base or complete drop-in. As you all pointed out there are plenty of 3W and 5W drop ins but not 1W...

keltex78 it seems that i can not find 1W 16 mm IR 850's already mounted on 16 mm stars on DX? I was searching for half an hour and nothing? Your link for this is pointing on 16mm stars(thanks for this I have it in option).

gmarsh thanks on very nice idea... Very nice explanation how to mount it without star(led pcb)... So you recommend Osram SFH4715s (850nm)? Would you be so kind and tell us what is maximum drive current for this 1W IR's? I have few AMC 7135 0,7A drivers... Will they suit for them or for those cheaper ones from DX?

Helios I would also stick to 850 because they seem to be brighter than 940, although plenty of people claims that 940 is less visible(or completely invisible) for game observation.

Why I want 1W 850 IR? It works best with aspheric

Why I want brightest possible? It will throw further...

Any of your experience with this type of modding especially max drive current(most suitable drivers) for this babies would be much appreciated.

Noctigons and Sinkpads would probably give the best from this IR's but they don't have such pcb's.

Artic alumina(i have it at home, and used it to made few drop ins in vinhnguyen style) is great idea but 16mm star mounted IR that lays pressed on artic silver 5(or better thermal paste) seems like most logical choice for me at this moment. In fact I can use AA on 16 mm led PCB if I will not find 16 mm IR mounted ones :)

After you apply artic alumina there is no returning back. It is concrete solid. Imho even more solid than solder.

Any good and reliable source for this Osram SFH4715s (850nm) ones?

I was just searching for what DX had for IR emitters. To me, it looks like your only option (at least from DX) would be to buy a cheaper white emitter mounted on a 16mm board remove the white emitter to solder your own IR emitter on the same board. I'm trying something similar with a UV emitter to make my own drop-in.

I understand what you are saying about needing the 1W (single emitter) unit to use with an aspheric... I'm curious to know how well that will work!

Friend you don't need to be curious because it works excellent! Especially with bigger aspherics. But it all depend whether you need long range illumination or not... For close encounters P60 reflector light + 5W IR would be the best.

Could someone kindly tell us what is maximum drive current for this 1W IR 850's and recommend suitable drivers?

But I think experiment will be necessary for brightest IR configuration possible, and since most people does not have night vision device and they are not modding IR's I understand lack of data on this type of modding...

The maximum drive current for that chinese 1W LED is about 0.7 amps, assuming a 1.4-1.5V forward voltage. A dual AMC7135 driver is your best choice for it.

For the maximum IR light output possible, I recommend SFH4715S - make sure you order it with the S, as it’s the dual series die version with twice the light output. The LED is rated at 1A forward current, a 3*AMC7135 driver will output 1.05A which the LED will certainly handle. A 4*7135 driver (1.4A) should work reliably if the LED is adequately cooled, on a good XP-G MCPCB you shouldn’t have any thermal problems.

The SFH4715S can handle twice the current (one doubling in light output) as the chinese LED, has two dies versus one (doubling the light output again) and I imagine the LED is at least twice as efficient. So I imagine you’ll get 8-10x as much light out of the SFH4715S versus your 1W chinese LED.

Excellent and superb advice! 8-10× more light is so fascinating that you hooked me on...If I will manage to made at least twice as bright over the regular 1W IR 850 I will have to repay you somehow :)


I just got to find that SFH4715S somewhere. If someone find vendor with international shipping please post here.

This will give you a bunch of electronics distributors stocking the part, pick one in your part of the world.

It seems that I can not order any SFH4715s.


I will be stuck with China ones...

Anyone tried to de-dome IR's?

I don't think the Chinese IR emitter have domes like the Cree emitters. I think it's just a protective plastic bubble...

mouser has the 4715s and they also have the berquist stars for them. The 4715s will not go on an xpe/g sinkpad because the 4715s and 4725s have the thermal path and anode connected. I blew up a few before i figured it out.

I have not found the need to dedome the osram IR emitters. They put out enough light to illuminate way beyond sensible shooting ranges when used in IR illuminators for nightvision scopes.

i’m actually starting a project a bit similar to yours, i’m building ir light for my camera, which is old sony 8mm 0lux cam.
it has 2 small 850nm leds build in, but i also build array from 13 0,2W 10mm leds. night and day difference.

now i’m building 10w ir light, i’ll use 2 5w ledengin 850hm leds. and 38mm 14* lenses, it’ll run on 4 18650 cells.

so far i got most parts and as time allows start the build, will make a new thread than.

very cool project. I will keep an eye on your progress for sure. The Vf is very high on that emitter. I will have to look around for a boost driver to use in a dual 18650 light. this emitter might be just the ticket in a slight defocused aspheric gun light application.

At what point does one have to start being concerned about IR safety ? a single oslon is rated at 800 mW/sr of radiation. When used to create a flood illuminator i guess its not going to be too dangerous but I imagine i would not walk in front of it if it was focused with a reflector or a lens … what kind of precautions do you guys take? there seems to be very little in way of guidelines on the net.

by the way, why are reflectors way less common in IR builds than aspherics or lenses ? Its the other way around with flashlights.

Only reason I can think of is the oslon emitters half angle which favors aspherics, but is that all or is there some other reason ? are the losses for reflectors/aspherics somehow different with IR and visible light ?

I bought a oslon ir-emitter at rs-online together with stuff from work. Another one I bought from mouser, they ship worldwide to anyone but have high shipping costs (but it is there in three days :-) )

Guys I will just say that you can not buy China 16 mm 1W 850 IR's mounted on 16 mm pcb's anywhere...

They only sell 20 mm ones...

4715s really looks good but Even if I order 10 pcs and pay more than 100$ for them I can bet that I will ruin at least 2 of them...

I seen that Djozz made one 4715s on xpg sinkpad so I guess it would work.

I have XPG noctigons and now some guys says 4715s can bee mounted and some says it can't.

4715s is to expensive for experiment.

@ fred…safety is always a concern. never look directly into a light with an IR emitter.never.
I use a camera that is sensitive to the 850nm and 940nm wavelength to
“see” the beam. They work well in reflector lights. They yield a beam pattern
similiar to an xpe. The aspherics are popular because most applications for the
emitters are in illuminators for weapons mounted nightvision equipment. With
aspherics the user has the ability to adjust the beam to meet the needs of the
situation. Plus the extra distance you get out of a tightly focused beam.

@lum iac…the only way a 4715s can be used on a SinkPad or NOCTIGON is to electrically isolate the star from the negative pathway of the flashlight. As I said above the anode and the thermal slug are connected on the 4715s and 4725s. screwing the star to the light housing short circuits the emitter and poof. I have used a thicker layer of thermal epoxy to attach the star and it works but defeats the whole purpose of mounting the emitter on a Sinkpad. The berquist stars are the best for oslon IR emitter at this time, they can also be filed or turned down to 16mm.

I glued the sinkpad with ir-oslon on the pill with arctic alumina adhesive. The layer can be really thin, it is a perfect isolator, so heatsinking is still very good. You can measure continuety with a DMM while glueing to check what is going on

You can see 850nm/940nm as a faint purple flash on a cellphone camera, you don’t need a specialized camera. I use my cellphone camera for testing TV-B-Gone’s after I build them.

Putting an Oslon on a sinkpad/noctigon doesn’t really defeat the point - I’m pretty sure the the delta-C/watt across the few square centimeters of thermal epoxy under a star should be less than the delta-C/watt across the usual MCPCB dielectric in a the tiny bit of area under the LED.

Thanks guys, plenty of good advices in this thread now.

Well I did make some orders:

1W IR 850 from DX

http://dx.com/p/1w-850nm-infrared-ir-led-emitter-5-piece-pack-145716

1W IR with different optics(dome) from ali express:

http://www.aliexpress.com/item/Free-shipping-10pcs-Infrared-transmitting-tube-1w-high-power-infrared-LED-lamp-bead-850-nm-45mil/692588556.html

I also ordered above mentioned 16 mm LED PCB's for 1-5W leds from DX and Ali express ( I want they come as fast as possible so if one of them fails probably other won't :) )



But I am still concerned about suitable driver for China 1W IR 850

I have few 4×AMC7135 (i don't have 2×AMC7135 as I mentioned above)


Main question for you guys is do I really have suitable driver for this or I have to order something else?
If I don't have please recommend some other drivers.Maybe more suitable than this one?

I mean we mentioned that max drive current for cheap China 1W ir is 0,7A. I am planing to use it on 1×18650 flashlight and this 4×amc7135(I will unsolder 2 amc's) will heat much more than maybe other type of driver?

Well I can slap things together but I am not driver expert. Need your opinion guys with link if any.