Flat, Diffuse, Very even, Low Intensity light source

[elmerb]

I am looking for flat light source approximately 1" x 3" to maybe 4" x 4". A flat LED backlight would possibly work. It will only be about an inch away from the object.

There are some special requirement. It needs to be very even in intensity across the area. It also needs very low intensity (20 to 300 mCd !) and be adjustable in intensity between these levels and stable over time. It cannot be pulsed to control intensity. It must generate a minimum of heat. It must be DC powered with no inverter which could generate EMI.

I had thoughts of doing something with an electroluminescent device but these are AC devices (EMI), and have some drift in intensity over time.

Any thoughts on a solution?

 

[Optical Inferno]

Sounds like a custom job that you are looking for. Electroluminescent stuff can be driven by a dc source; I have some EL wire that runs off a 9V battery with an in-line converter. I don't think you can adjust the intensity of EL sources though (i may be mistaken).

An edge lit LED panel may be more what you are looking for. Perhaps someone else may have an idea...

 

[AnAppleSnail]

If I had to build this, I would do the following:

1. 9v battery for power.
2. 3x4 grid of 5mm 'straw hat' LEDs (Not the bullet-shaped ones, but those with wide dispersion), wired 6p2s.
3. Diffusing film
4. Simple LM317 step-down circuit to provide approximately flat regulation (9v in, 5-6v out): See here.

How stable over time must it be? Perceptually even, or do you need precise control? If you can afford slight variation (Very slight at these low current draws) then a 3x3 grid of LEDs (Wired 3s3p) behind a diffuser film, running directly from a resistor on a 9v battery, would do.

 

[elmerb]

Thanks for your thoughts.

The output must be stable over time. It is to be used in testing and grading a photo diode detector array.

The output must be very low level with the ability to adjust it from 1X intensity to 10X intensity and both levels must be stable. This has to be done via current adjustment as pulsing the array with brighter light overdrives the photodiodes.

The linearity of the light intensity vs current must be good. It looks like GaP would be the preferable type for this.

The light intensity across the surface must be uniform.

The color must be between 550 to 600 nm.

I am considering surface mounted green wide dispersion LED's. However, I am concerned about the bin sort of the LED's as to their intensity level. The bin sorts are pretty large.

I had thought of using a wolven fiber panel backlight with one external light to drive it to try to get an even amount of light across the surface. I don't really know how even these sources are.

I am still looking at this...

 

[127.0.0.1]

oled and glass filters

 

[jspeybro]

Hi,

here's an idea:
Regarding wavelength, don't use green leds, these are around 520-540nm. If you want 550-600nm, go for yellow or amber leds: 585-590nm.
To get a nice homogeneous diffuse low intensity illumination, get a piece of endlighten sheet from Evonik. They normally sell the cut to size pieces at a reasonable price. Then place the leds in such a way that they illuminate the side of the sheet. The scattering particles in the material will cause the material to glow in the color of the leds. Placing leds on each side of a square, or maybe even around a circular shape, will probably give you the most homogeneous distribution.
If you want long term stability, you'll need a feedback loop using e.g. a photodiode to monitor the intensity and adjust the current through the leds to keep the photodiode current at a fixed level. I suggest using a data aquisition card that supports analog inputs and outputs.
Since most dimmable led drivers use PWM, you probably won't find a good match for your setup so you'll need to go old school and use a resistor to control the current through the LEDs. If the current is low, you may be able to drive them directly from the data acquisition card. This way, you can adjust the voltage to control the intensity of the LEDs if you placed a well chosen resistor in series with the leds. Using the analog input the read the photodiode current and converting it to a voltage level, you can use this to keep the intensity stable.
If you want to switch between 1x and 10x the intensity, just use 2 analog outputs that connect to the LEDs with a different resistor.
You can find photodiodes with bnc connector at thorlabs. To improve the insensitivity to daylight, you can add a filter that only transmit the led wavelength (plenty available at edmund optics).
All you'll need to do is do some programming with the DAQ card or use labview.

an alternative could be to use an disassembled LCD display. These have a nice even illumination. This is however white light, not yellow.

hope this helps

 

[Optical Inferno]

Check out intematix remote phosphor products too. They create sheets that can be made to emit in that wavelength and are very even. You can create a simple dimming array or 455nm LEDs to excite the phosphor to achieve you intensity desire as well.

 

[jspeybro]

many fluorescent or phosphorescent materials degrade over time when excited for longer periods. if you only switch it on once in a while, this shouldn't be a problem.
The leds itself however will also degrade over time depending on how hard you drive them.

 

[hank]

Microscopists always need this kind of flat even illumination. Here's a new "fly eye lens" method described with pictures:
http://www.nikon.com/products/instr...biological-microscopes/upright/e100/index.htm