Best material to diffuse light

[drodrig]

Hi everyone,

We are building a setup to distribute the light from LEDs to optical fibers. We have 3 LEDs (blue, red and green). They are connected to a plexiglass light mixer (with a in between LCD as a filter). On the other end we put optic fibers.

The plexiglass has a cross section of 18x18mm2 and it's 150mm long.

The problem we have is the light is not homogeneous on the fibers end so some fibers get more light than others.

The light mixer should be as short as possible.

Does anybody have recommendation for a better material for the mixer?

Attached a sketch showing the idea

thanks

http://files.engineering.com/getfil...a-486e-bfab-593565cf3c34&file=Light_Mixer.png

 

[IRstuff]

You could try a holographic diffuser?

http://www.luminitco.com/

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[MacGyverS2000]

A holographic light diffuser would do the job nicely. But if you don't care about efficiency as much, separate the plexi rod into two pieces and frost the interface close to the LEDs... instant/cheap diffuser.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[drodrig]

IRstuff,
thanks for the recommendation of this company. I have contacted them for a sample

MacGyverS2000,
sorry, it might be a language problem from my side. What do you mean with frost the interface? to de-polish? roughen the end face?

thanks

 

[MacGyverS2000]

A bit of sandpaper... say, 240 grit.

Looks like IR beat me to the punch on the holographic diffuser.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[drodrig]

ok, nice.

your idea is also good, we can try it quickly

thanks!

 

[MacGyverS2000]

In case my post wasn't clear, it should be two separate plexi rods between the LEDs and fiber. Rod #1 should be polished at the LED entrance end and frosted at the exit end. The second rod can just be polished on both ends. For extra mixing, consider a slight concave on the frosted end, but then we're getting back into a true optics design.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

There are some caveats with using a surface diffuser vs. volume diffuser, namely, uniformity.

Surface diffusers tend to be somewhat Lambertian, so any standoff distance can drastically reduce the transmission efficiency. Some holographic diffusers are designed for specific exit cone angles,, which can greatly improve the overall efficiency.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[drodrig]

MacGyverS200,
nice explanation
how would you connect the rods? optical grease? optical-transparent glue? just pressing one against the other?

IRstuff,
Are the holographic diffuser always surfaces? Our setup will be placed in a non very well controlled area so I fell like a solid diffuser will be more stable

 

[MacGyverS2000]

Any sort of optical glue/grease would defeat the purpose of the frosting.

Is this for a prototype or a production unit? I would definitely do things differently for each... what are you trying to create?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[drodrig]

Now we are prototyping, doing tests but looking for a solution for a final setup (to be used for some years)

We have some 500 crystals to be cured (from radiation) and calibrated (photosensors) with light. We have 3 LEDS (blue, green and red)and 4 fibers (200microns core diameter) per crystal.

We need to bring light from the LEDs to the crystals and we need to have the same light on each crystal.

For that we use this light mixer (here again the sketch)

http://files.engineering.com/downlo...a-486e-bfab-593565cf3c34&file=Light_Mixer.png

right now we have 15cm of plexiglass but the homogeneity is not good.

cheers,

 

[MacGyverS2000]

In such a case, I might consider putting the fibers right at the LED chip interface (let's say 4 per LED). Place each of the 12 fibers into an array bundle that mimics an LCD/TV screen (RGB stripes). If you need further mixing, run the output of the fibers into a frosted plexi rod, then back into fiber... I doubt you could see any pinpoints of a specific color at that point.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[drodrig]

But I have over 2000 fibers, this solution would require big amount of LEDs. If I make bundels of 200-800 fibers with less than 10 mixers (30 LEDs) we get the work done

The problem is finding the proper material/length for the homogeneous light at the end of the rod

I have found this:

http://www.edmundoptics.co.uk/optic...ryid=231&_ga=1.258154121.221542912.1453910329

what do you think? looks more promising than plexiglass

cheers,

 

[MacGyverS2000]

Cripes, if you can afford the cost of those homogenizing pipes, what are you talking to us little guys for? 76 pounds for a 1" long pipe?!

I'm imagining something (relatively) simple... a cast piece of plastic in a triangular form with a dimple on each "point" to hold a few fibers each. The spacing of the triangle is the same as the spacing of the three LEDs in the cluster. The fibers are immediately pulled together into a bundle, or you could use a much smaller triangle that mixes a fiber from each color at each point, essentially homogenizing the colors there.

Of course, I was also assuming you were calibrating each crystal one at a time... if you have a need to do all 500, you'll need to figure out a way to efficiently route the fibers. But then again, you would have to do that anyway, regardless of what solution you used if it involved fiber.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

if the fiber bundle is a critical part, then doing a heated draw to reduce the bundle diameter would be on way to reduce the requirements on the diffuser.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[drodrig]

Hi

We will have like 5 or 6 of these diffusers, so the cost is assumable.

About the fibers side; we also have to study it, but the first step is checking the homogeneity (for that we use the same fiber to be able to compare values from different points)

cheers,

 

[IRstuff]

That may be an imposition of an unnecessary requirement. Once you reach a certain level that's generally tolerable from a transmission perspective, you can theoretically calibrate out the remaining errors.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[MikeHalloran]

A former employer bought a light splitter assembly, in quantity, from Welch-Allyn of Skaneateles Falls, NY.

At its input end, it had a single port, ~6mm dia, that was aligned with a light source.
At the other end, ~50mm away, it had two similar smaller ports, spaced apart ~25mm.
In between was a forked bundle of optical fibers, 'randomized', and potted.

We never found out exactly how they randomized the fibers, but the output was sufficiently uniform for our purposes.


I see that Welch-Allyn has been acquired by Hill-Rom.
I do not know how that affects things; they were extraordinary partners for us.





Mike Halloran
Pembroke Pines, FL, USA