I've been looking at the diy projecter thread and saw somthing about leds been an alternative light source. But in the end a cluster of these was still to dull. The post mentioned them been about 10 lumen to 17 lumen (or something like that).
I was researching leds and found a manufactuer that sells 5 watt leds that give off aprox 120 lm each!
does "120" lm mean 120 lumens?
And if so, wouldn't a cluster of these be perfect for lcd projectors.
The site is: www.luxeon.com
and its the 5 watt emitter
I was researching leds and found a manufactuer that sells 5 watt leds that give off aprox 120 lm each!
does "120" lm mean 120 lumens?
And if so, wouldn't a cluster of these be perfect for lcd projectors.
The site is: www.luxeon.com
and its the 5 watt emitter
Try searching the net for Luxeon Star-O. They seem to be very powerful. They sell for $12 to $16 U.S. each.
Vince
PS, when considering lumens, it should be ANSI lumens. If it's not listed as ANSI lumens, they are not following any standard.
V
Vince
PS, when considering lumens, it should be ANSI lumens. If it's not listed as ANSI lumens, they are not following any standard.
V
Marc5193
Leds would be perfect. My OHP has a light rated at 50 hours use costs me NZ $25 each time for a new one. Makes you feel you have to ration large screen use to conserve the bulb.
Leds have low heat, You can turn them on and off anytime you want without waiting for the bulb to cool, they are robust not easily damaged and last a very long time. 5-10 years
My OHP is rated at 2000 lumens. Not very bright but ok for me at home and gives a good picture. If your leds give 100 lumens you only need 20 -30 assuming ANSI lumens which would be not costly considering the lifetime.
Thing is you cant just stick then behind your panel. You have to pass the led light through your panel on to a projection lens and because projection lenses are usually small the light has to converge (get smaller) down to the lens. Thats the same as using a reular OHP but the problem with led's is they are multiple units not a point source like a metal halide light. That means you have a problem of how to design a reflector so all you led light goes through the panel.
If you had a 100 or so led it would be easy just crowd them all on one circuit slightly larger than the lcd panel and angle the led's so they are converging through the LCD panel to the projection lens. But with only 20-30 leds you would have gaps between the leds which would give uneven light.
It depends a lot on the angle of light projection the leds have.
If the light exits the led and fans out in a large angle you have to attempt to collect all that light otherwise isome will be wasted, and you cannot afford to waste any. If the light on the newer type led exits in a tight angle with small spread you have a much better chance of collecting all the light, or as much as possible.
If the newer led have a small exit angle and high lumen output you could be getting close to a very useful light source.
So what you need is some serious research on what led gives the highest possible lumen output with the smallest beam spread for an economic price. With the pace of progress on led I would not be surprised if you found some good ones next week perhaps or if not wont be very far away.
If you do find a good led let me know and I will try and help with a means of getting all the light going in the right direction. Or someone else will.
Best of luck
Leds would be perfect. My OHP has a light rated at 50 hours use costs me NZ $25 each time for a new one. Makes you feel you have to ration large screen use to conserve the bulb.
Leds have low heat, You can turn them on and off anytime you want without waiting for the bulb to cool, they are robust not easily damaged and last a very long time. 5-10 years
My OHP is rated at 2000 lumens. Not very bright but ok for me at home and gives a good picture. If your leds give 100 lumens you only need 20 -30 assuming ANSI lumens which would be not costly considering the lifetime.
Thing is you cant just stick then behind your panel. You have to pass the led light through your panel on to a projection lens and because projection lenses are usually small the light has to converge (get smaller) down to the lens. Thats the same as using a reular OHP but the problem with led's is they are multiple units not a point source like a metal halide light. That means you have a problem of how to design a reflector so all you led light goes through the panel.
If you had a 100 or so led it would be easy just crowd them all on one circuit slightly larger than the lcd panel and angle the led's so they are converging through the LCD panel to the projection lens. But with only 20-30 leds you would have gaps between the leds which would give uneven light.
It depends a lot on the angle of light projection the leds have.
If the light exits the led and fans out in a large angle you have to attempt to collect all that light otherwise isome will be wasted, and you cannot afford to waste any. If the light on the newer type led exits in a tight angle with small spread you have a much better chance of collecting all the light, or as much as possible.
If the newer led have a small exit angle and high lumen output you could be getting close to a very useful light source.
So what you need is some serious research on what led gives the highest possible lumen output with the smallest beam spread for an economic price. With the pace of progress on led I would not be surprised if you found some good ones next week perhaps or if not wont be very far away.
If you do find a good led let me know and I will try and help with a means of getting all the light going in the right direction. Or someone else will.
Best of luck
Hi Guys,
I have been thinking about this for a bit now. My father gave me some advise months ago on how to converge light. Use back to back convex lenses. I believe this would work best w/ a small LCD screen- 1.8"-2.5" or smaller. Something like this....
< )( [] )
Light src.> convex lenses> LCD> Projection Lens
I think frensels were making the light source bigger, not gathering then redistributing it.
The drawing above, from convex lenses to the projection lens should be enclosed, so that light does not escape.
With LEDs, you could have the components even closer together.
Vince
ps. you could even do this...
< )( )( [] )
V
I have been thinking about this for a bit now. My father gave me some advise months ago on how to converge light. Use back to back convex lenses. I believe this would work best w/ a small LCD screen- 1.8"-2.5" or smaller. Something like this....
< )( [] )
Light src.> convex lenses> LCD> Projection Lens
I think frensels were making the light source bigger, not gathering then redistributing it.
The drawing above, from convex lenses to the projection lens should be enclosed, so that light does not escape.
With LEDs, you could have the components even closer together.
Vince
ps. you could even do this...
< )( )( [] )
V
vdi_nenna,
that's the principle of slide projectors. Fresnels do the same what normal lenses do. They are only divided into small zones, which represent parts of the normal lens curvature.
Richard,
looking at multiple LEDs for illumination system, there could be a way, if we could use a multi condensor system.
If each LED behind the LCD panel is arranged, that it's center ray hits the objective lens, and each LED has it's own condensor.
This means, LEDs must sit on concave surface. Since every LED has a flood angle, this angle has to be reduced to near parallel beam, which is not larger than objective's entrance pupil. To avoid gaps between LED beams, the condensors should sit side by side, so that paralaxial rays slghtly 'touch' each other.
Of course, this can only become actual, if we see real bright LEDS, and we can sing the song:
BLINDED BY THE LED...!
Greetings
xblocker
that's the principle of slide projectors. Fresnels do the same what normal lenses do. They are only divided into small zones, which represent parts of the normal lens curvature.
Richard,
looking at multiple LEDs for illumination system, there could be a way, if we could use a multi condensor system.
If each LED behind the LCD panel is arranged, that it's center ray hits the objective lens, and each LED has it's own condensor.
This means, LEDs must sit on concave surface. Since every LED has a flood angle, this angle has to be reduced to near parallel beam, which is not larger than objective's entrance pupil. To avoid gaps between LED beams, the condensors should sit side by side, so that paralaxial rays slghtly 'touch' each other.
Of course, this can only become actual, if we see real bright LEDS, and we can sing the song:
BLINDED BY THE LED...!
Greetings
xblocker
Xblocker
Yeh your right. Its the same as the MLA micro lens array the big boys are using in their 3 LCD projectors. Take all the light, straighten it up with a lens so its going right through the LCD panel in parallel beams rather than at all sorts of angles.
With our big panels what do you do after the light has gone through the panel parallel. The light bundle is much too big to use a glass projection lens. Have to use a very high quality fresnel.??
At the moment cost of LED and lens would still be a problem but if someone bought one or two of these superbright led it would be possible to experiment on a small scale with just a small section of the LCD panel.
Yeh your right. Its the same as the MLA micro lens array the big boys are using in their 3 LCD projectors. Take all the light, straighten it up with a lens so its going right through the LCD panel in parallel beams rather than at all sorts of angles.
With our big panels what do you do after the light has gone through the panel parallel. The light bundle is much too big to use a glass projection lens. Have to use a very high quality fresnel.??
At the moment cost of LED and lens would still be a problem but if someone bought one or two of these superbright led it would be possible to experiment on a small scale with just a small section of the LCD panel.
One of the problems with LED apart from cost is because each one is small you do not have one large powerful light source as with a Metal Halide or similar bulb but a number of point source illuminators.
How do you get them all to provide one beam of light that has good uniformity and enough power to go through an LCD panel and give a bright image on the wall without wasting illumination.
There may be several possibilities.
(1) Instead of using a small number of superbright & costly led it may be possible to use a larger number of "bright" led to achieve similar lumen output but at less cost. Each led would need its own focusing lens so you could have for a 10 inch LCD panel with several hundred led aimed through the panel and converging to a projection lens. Cost of led+lens may work out less.
(2) Joining multiple light sources into one beam is a common problem in optics but is not all that easy for DIY people. Usual method is to aim one or more lasers or led at the entry point to a fibre optic. Fibre has an acceptance angle so if the light sources are within the acceptance angle the fibre will accept it. There are losses at the entry point (and the exit point) .
It is possible to have air wave guides which minimise losses. These can be constructed from first surface mirrors. The better the mirrors the less losses but construction details are very exacting.
It is also possible to inject light into water contained in a tube. The tube can have multiple fingers joining together to give a composite light source. Teflon AF tubing is a suitable material. Not expensive in short lenghts. Waveguides are a possibility to join many "bright" led or a smaller number of "super" bright led into one beam. Led have a much larger source size than lasers so the combining problems increase
(3) Led's can be pulsed. It is common practice to pulse relatively weak radiators with very high bursts of energy for a very short period of time to give very large increases in brightness.
The flashlight for a camera is an example. If a superbright Led could be pulsed above the flicker rate for tv, each led could appear many times as bright as for steady drive conditions.
The dissapation of the led must still be within its ratings. Led are solid state devics and have no inertia from a filament so could be pulsed by simple electronic circuits at hundreds of pulses per second which gives the impression to the eye as being very much brighter than unpulsed. Persistence of vision allows you to still see a very bright source even though its no longer illuminating. Like this LLLLLLL. On for a short time, off for a longer time.
Pulsing can provide a substantial increase in apparent illumination. Many devices that on DC run at 100ma maximum can be pulsed at several amps with the correct mark space ratio.
How do you get them all to provide one beam of light that has good uniformity and enough power to go through an LCD panel and give a bright image on the wall without wasting illumination.
There may be several possibilities.
(1) Instead of using a small number of superbright & costly led it may be possible to use a larger number of "bright" led to achieve similar lumen output but at less cost. Each led would need its own focusing lens so you could have for a 10 inch LCD panel with several hundred led aimed through the panel and converging to a projection lens. Cost of led+lens may work out less.
(2) Joining multiple light sources into one beam is a common problem in optics but is not all that easy for DIY people. Usual method is to aim one or more lasers or led at the entry point to a fibre optic. Fibre has an acceptance angle so if the light sources are within the acceptance angle the fibre will accept it. There are losses at the entry point (and the exit point) .
It is possible to have air wave guides which minimise losses. These can be constructed from first surface mirrors. The better the mirrors the less losses but construction details are very exacting.
It is also possible to inject light into water contained in a tube. The tube can have multiple fingers joining together to give a composite light source. Teflon AF tubing is a suitable material. Not expensive in short lenghts. Waveguides are a possibility to join many "bright" led or a smaller number of "super" bright led into one beam. Led have a much larger source size than lasers so the combining problems increase
(3) Led's can be pulsed. It is common practice to pulse relatively weak radiators with very high bursts of energy for a very short period of time to give very large increases in brightness.
The flashlight for a camera is an example. If a superbright Led could be pulsed above the flicker rate for tv, each led could appear many times as bright as for steady drive conditions.
The dissapation of the led must still be within its ratings. Led are solid state devics and have no inertia from a filament so could be pulsed by simple electronic circuits at hundreds of pulses per second which gives the impression to the eye as being very much brighter than unpulsed. Persistence of vision allows you to still see a very bright source even though its no longer illuminating. Like this LLLLLLL. On for a short time, off for a longer time.
Pulsing can provide a substantial increase in apparent illumination. Many devices that on DC run at 100ma maximum can be pulsed at several amps with the correct mark space ratio.
LED's always seemed to have a pretty narrow beam angle, hence making them more like little lasers than little light bulbs -- so i dont think the "multiple point source" problem had with the LOA lights will be as much of a factor here as it was before... if you were to cluster them close together behind a larger LCD panel (maybe 10.4" ?), then use a page magnifyer after the panel to converge it down to the projection lens then you might have something.... this is assuming the LEDs put out fairly parrallel beams................but then again, i really dont know too much about what i am talking about.
my Luxeon's should arrive today
I've been playing with the Agilent super-brite RGB leds for a couple years, but the specs on the Luxeon's caught my attention -- the back order was about a month -- but my distributor says they should arrive today.
I've been playing with the Agilent super-brite RGB leds for a couple years, but the specs on the Luxeon's caught my attention -- the back order was about a month -- but my distributor says they should arrive today.
http://www.bit-tech.net/review/82/
the luxeon LEDs do indeed get hot...hot enough to need a heatsink on the back of them. The previous link reviews these LEDs and has pictures of the red and blue models in action. Unfortunately they don't have a white one, but I'd imagine the intensity would not differ incredibly much between them.
the luxeon LEDs do indeed get hot...hot enough to need a heatsink on the back of them. The previous link reviews these LEDs and has pictures of the red and blue models in action. Unfortunately they don't have a white one, but I'd imagine the intensity would not differ incredibly much between them.
the STARs are not yet available in white, but the EMITTER's are:
http://www.lumileds.com/pdfs/DS25.PDF
my application requires RGB
http://www.lumileds.com/pdfs/DS25.PDF
my application requires RGB
Optics considerations
Hi Guys,
Found this info about secondary optics considerations for LED lights. Might be of use...
http://www.lumileds.com/pdfs/an1149-5.pdf
KD
Hi Guys,
Found this info about secondary optics considerations for LED lights. Might be of use...
http://www.lumileds.com/pdfs/an1149-5.pdf
KD
For Luxeon's, it is likely that you would want to use Red, Green and Blue ones. If you use white, you have no guarentee that the phosphor used for the white matches up with the filters in your panel and the efficiency may not be great. You will probably find that you need an awfully large number of them, and they will not be cheap. However, there is something to be said for a light source that is only down 30% after 50,000 hours, IF you keep them cool! You also need to keep the temperature stable or the wavelenght, hence color temperature of your light will change. One nice thing about them will be that there is no IR in the light, hence your panel does not heat up. Realistically, cold cathode flourescents offer much higher efficiencies today and the bulb life if pretty high, i.e. 40,000 hours at 50% brightness.
Alvaius
Alvaius
- Status
- Not open for further replies.