A month or so ago I purchased a Cree P3 LED and a few optics to go with it. Only now have I got around to playing with them. I initially was just testing out an idea of using a tapered light pipe to direct the light into an old commercial 3 LCD projector, a separate R G and B LED is what I had planed but that’s another story. What I just tried tonight was a crude test (very crude) holding the LED behind a busted 7” LCD and one Fresnel. Surprisingly what I found was I was able to get the central spot of the LED to fit into a spot on the wall the same size as a triplet. I flipped the whole thing so that I was shinning it onto my ceiling and I could see the spot with a full lights on. What this suggests to me is that these LEDs could make a very good projector, as long as everything doesn’t go astray when I put a triplet in place.
I also moved the LCD away from the wall about 600mm, this time with no fresnels at all, and I could get a spot about the size of a the pro triplet. A projector made like this would be interesting. It would require each LED to be aimed at the triplet separately so individual heat sinks would be required. If it worked there would be no losses from the fresnels to eat into the light output.

I think I’ve caught the LED bug. or maybe I’m just sleepy.


DJ

That's interesting for sure. If a person just used a regular magnifying glass to project the light through maybe it would intensify the amount of light. But then again it would be just a different type of fresnel.
Go Dazzla, it's peaked my curiosity!

DAZZ
More than interesting ! how many Cree's have you got to try ?

Do these Crees have a reflector or lens builtin ? or are you just shining them at the LCD bare

Good luck Dazz ..LED's do make a lot of sense if we can get them to work optimally

I’ve only the one at the moment. It’s mounted on a 20mm star and using a narrow optic(lenflector? ). With the LCD and fresnel stationary and set back from the wall about at the fresnel’s FL I could move the LED at any position on the LCD and the spot on the wall would stay roughly in the same position.
I had little setback, some bugger made a dodgy connection using duct tape to hold the wires into a wall wart plug and now the transformer mysteriously doesn’t work any more . I’ll have to go hunting in the depths of my draws to find another.

The P3 was a typo, it’s actually a Q2

DJ

OK if I want to pursue this further then I need some help finding a supply of cheap LEDs and optics.
This is the optic I’m looking for.

DJ

Now I see what you mean when you said lenflector !

First Ive seen of such a thing...Ive not dabbled into LED's yet so Im gonna be unhelpful for this one..Im sure others have more knowledge of where to source

There’re actually quite common, other brands that is. And the beam pattern looks nothing like the one in the PDF, I don’t know where they got that from. The pattern is very square rather than round which makes it much easier to blend in an array.

DJ

I’m actually toying with the idea of modifying the optic or a cheaper generic version. The beam angle is stated as 8° but my measurements show it to be closer to 12°. At this angle the resulting beam spot will slightly overfill the triplet. I can get the spot size to shrink by using another lens in front of the optic. If the front surface of the optic was to be slightly convex then it would act in the same way as the separate lens. But I’m not sure if I want to sacrifice this optic just yet.

DJ

Ahhh . . . never mind. Too easy.

Was thinking the same thing, Durachko. *pass*

I resurrected my 17” projector to test out the led. In short it works. I removed the rear fresnel and positioned the LED at various positions on the LCD. They all seem to pass the majority of the beam through the triplet. Now some crude LUX measurements. The distance between the LED and the LCD will determine the lumens. My screen is about 2600mm wide and with the LED about 300mm from the LCD I get about 25LUX. The total screen area is 5.4M2 so that works out to be about 135 ANSI lumens. I moved the LED to about 80mm from the LCD and it went up to 85LUX. That's 460ANSI lumens!!! Now to the catch. With the LED at this distance it would require about 220 to cover the 17”LCD .
I still need to re-measure the LED current, I think I was stooged. If I remember correctly it was only drawing 350ma.

DJ

Well whatever you decide to do that pretty impressive figures so far ..is this using the lenflector & cree on its own ?


Im assuming you have a spot ( how big) on your screen

Whats the lumens on a P3 Cree ?

How about doing this on a smaller panel ?

It was with the Cree and lensflector.


The screen is 2600mm wide and the LED illuminated a spot about 360mm in diameter on the screen when it was at 300mm from the LCD. And a spot size 200mm in diameter on screen when is 80mm from the LCD.


It’s stated as 87.4 to 93.9 Lumen @ 350ma (It’s a Q2) I should be able to at least double its output if I run it at 700ma.



The size of the LCD is not a real issue. For a bigger LCD to use the same number of LEDs only means that the LEDs just get moved back further from the LCD. The biggest hurdle is getting the light to pass trough the triplet.

That means with 220 of these suckers, it theoretically should be around 920 ANSI @ 660Watts, not into the commercial efficiencies but bright enough. With some polarization recycling it could be increased further. The only problem is that the LEDs and optics are not cheap. Unless we could find a Chinese supplier.

92 ansi with leds could already be a tempting proposition for some...
Curious about colour rendering though - not to mention other minor issues such as getting uniform illumination

Uniformity can be improved by increasing the distance between the LEDs and LCD and adding mirrors to the sides. Polarisation recycling can also be used to further increase uniformity.
The colour temp I’m not sure of.

DJ

Just remembered I have a Q2 powered flashlight (+luxmeter), might just have to take the back off the pj later on

You’ll probably need one of these as well. Or something with a simular beam angle.

You could be right. I thought this light threw a fairly well collimated beam as is but the luxmeter tells a different story.
From 270cm away, the max is 85 lux, and the spot is 42cm across (half the lux). So thats a 9° beam, but less than 12lumens. Not quite the 90% efficiency of the lensflector

Dazz, any ideas if this would be suitable?
http://www.dealextreme.com/details.dx/sku.12437
they also do this one, cheaper, but even less info http://www.dealextreme.com/details.dx/sku.4572
oh and tons of info on CPF, http://www.google.com/search?q=cree+optics...powerforums.com just need to find a 'minute' to go through it

Only one way to find out

Yeah I know... also involves waiting three weeks for delivery

This one looks like the closest match to yours:
http://kaidomain.com/ProductDetails.aspx?ProductId=1606 - it's even fairly affordable at the current rate of exhange

That's the cheapest I've seen it so far.

Hi,

New to the forums and just posted my first build that should be done soon. You got me thinking though. What if you just did a LED surface or hole through mount array, roughly the size of the LCD. For sake of argument, say you had a 15" display and put a 15" array offset a few inches and used that as the light source? This would drastically reduce the size of the box since the rear light source/focal length could be virtually eliminated!! LED light source won't be as hot and hence no or little fans, for cooling, no noise! Not to mention saving power and money for expensive replacement bulbs.

Hi DAZZZLA !
could you show some pictures of your test?

Hello Dazla,

I think this way of using LEd is the futur of DIY projector. Ye could use a lot of LED on a 15" panel and get a bright and even image, with less heat.

Do you have any development news?

Thank.
MichelSpasCie

Seeing DAZZZLA's great results (in theory at least), I thought that this might be appropriate for this thread Researchers develop higher-efficiency LEDs

http://www.engadget.com/2009/01/14/researc...fficiency-leds/

some design, not mine, not so bright

ehhh i'd love to get one of those leds to run a ramp up test on it........


could provide some extremely accurate info helping to overdrive it. but just a nice voltage increase should do the trick quite well...

any progress lately/at all? and if you'd like i could stop by my microE teacher's office and ask if he knows of any sources for these cheap. if anyone knows off the top of their head it'll be him.

posted a link the other day on some very high powered LED's that can produce 4000 lumens white light to be used in commercial projection, however they are very expensive at $120 USD

http://www.lumenlab.com/forums/index.php?s...st&p=302012

here's the direct link to phlatlight LED

http://www.luminus.com/content1463