The point of a DIY projector is to build one that is as high of quality as a professional projector without spending as much money - mainly on the expensive lamp replacements. It seems to me that we've all done a really good job at finally figuring out good LCD's, lens triplets, fresnels, etc. BUT our light sources, in my opinion, still have a lot of room for improvement.
I would like to start a collaborative attept to figure out if there is any possibility of using LED's as a light source. They would be very beneficial because of their very low power consumption and heat disapation. I know there has been discussion about this in other threads, but I wanted to make one thread that was dedicated to this for sake of organization so that we could actually get somewhere with this.
...
I once tried to cluster 10-20 16,000 mcd LEDs into a couple of IC sockets and put them behind my LCD panel to see if it would illuminate that part of the image enough for me to consider using a full array of them. My results were terrible - bad coloration and very dim image. Recently, I've realized that I didn't have the LED's very well oriented, nor was I monitoring my voltage to see if I was getting optimal ouput. I will try carefully soldering these 20 LED's onto a circuit board as densly as possible and try again. I'll post my results here as well as in our forums at my website. Hopefully other people will join in as well.
Thanks for reading!
I would like to start a collaborative attept to figure out if there is any possibility of using LED's as a light source. They would be very beneficial because of their very low power consumption and heat disapation. I know there has been discussion about this in other threads, but I wanted to make one thread that was dedicated to this for sake of organization so that we could actually get somewhere with this.
...
I once tried to cluster 10-20 16,000 mcd LEDs into a couple of IC sockets and put them behind my LCD panel to see if it would illuminate that part of the image enough for me to consider using a full array of them. My results were terrible - bad coloration and very dim image. Recently, I've realized that I didn't have the LED's very well oriented, nor was I monitoring my voltage to see if I was getting optimal ouput. I will try carefully soldering these 20 LED's onto a circuit board as densly as possible and try again. I'll post my results here as well as in our forums at my website. Hopefully other people will join in as well.
Thanks for reading!
Here's as much as I know:
In order to be able to concentrate all available light through the lense, a point source is required.
Experiments with LEDs have therefore been limited to projectors using only one LED, or maybe one per colour. Lumileds have made a proto that produces a good image about the size of a sheet of paper.
I'll need to hone my optics skills before I can say more.
Cheers,
Bruno
In order to be able to concentrate all available light through the lense, a point source is required.
Experiments with LEDs have therefore been limited to projectors using only one LED, or maybe one per colour. Lumileds have made a proto that produces a good image about the size of a sheet of paper.
I'll need to hone my optics skills before I can say more.
Cheers,
Bruno
Allan, with all due respect, i think your projector systems are the most behind when it comes to light enines, i havnt seen you use a proper spherical reflector yet or even front surface mirror, so how are you even thinking in using leds? if you cant get a mh to work right, then you seriously wont get leds to work.
There are manny ways to get low wattage metal halides to be as bright as the larger wattage ones, right now im looking at ways to crack the cdm-t lamp on making it brighter then the 250w, and as it is, ive gotten it a tad brighter, a new reflector that ive just orderd will make it brighter ( still spherical not parabollic btw).
Now imagine that new reflector on a 250w, it will come close to a 400w, as it is ive gotten the 250w brighter then a 400w before, but heating issues on condensers was the problem, ( i did manage to crack that btw, but overheating the bulb is a big no no). Ive also learnt about over heating bulbs, and by all means ive soon learnt that soup laddels are the worst thing for overheating a bulb and risking an explosion. Its all in the reflector allan for a crisp clear sharp and bright image in a safe manner, it starts right back at the source.
I read somwhere the otherday about leds, aparently the light from a led is magnified 100 000 times to get the brightness it is, now with that magnifacation and by adding other lenses, u aint gonna get that led much brighter at all. Large singular leds are the way to go and they are currently in research, the only way we could use one of those these days is if we had a lcd with micro lens tech, and in a normal pro projector being as a retrofit, and even then it still will be dim. Leds also run very hot to acheive the brightness we require.
I realy dont see any point to this led stuff cos it aint gonna happen any time soon, atleast not with us, metal halides are the way to go for price, heat, and lifespan, atleast in the diy community. If your metal halide setups are not good enough for you, then i sujest doing abit more hands on research on the matter cos if and when you got them set up right, you cant beat em.
No pun intended on this post allan and no personal hard feelings, but thats just my 2cents.
Trev
Can we talk about LED's in this thread?
I do use a spherical reflector also in addition to my condenser lens - don't know what more you want from me (though this one is pretty dirty and could use a lot of polishing):
...
Anyways, I was hoping that we could try experimenting with LED's. I don't know what you meant by the light being magnified 100,000 times ... I'm going to go solder some to a circuit board and see how bright they are. I think that their alignment is the biggest issue.
I do use a spherical reflector also in addition to my condenser lens - don't know what more you want from me (though this one is pretty dirty and could use a lot of polishing):
An externally hosted image should be here but it was not working when we last tested it.
...
Anyways, I was hoping that we could try experimenting with LED's. I don't know what you meant by the light being magnified 100,000 times ... I'm going to go solder some to a circuit board and see how bright they are. I think that their alignment is the biggest issue.
Allan thats not a spherical reflector, its what its called, a soup laddel, its the wrong shape for starters and it, if it has a focal is all over the place. Its peice of metal thats round and just scatters light lol. Spherical reflectors are precision made, a soup laddel is just pressed with no optical surface coating at all, it makes a big diff.
As i stated in my last post, no pun intended, kind of hoping you would learn somthing and maybe try the real gear.
Trev
array
is it possible to focus a big LED array to a point source so it can then illuminate an LCD evenly? there would be some losses, but that would give the advantage of being able to build a very large array
is it possible to focus a big LED array to a point source so it can then illuminate an LCD evenly? there would be some losses, but that would give the advantage of being able to build a very large array
You can illuminate an LCD evenly but that's not the point. All photons should be heading straight for the lense. Which is why you start from a point source, which then allows you to collimate the light beam as desired.
There's no way of making 2 physical light sources into 1 virtual source (unless they occupy different parts of the spectrum, or unless you accept losses so great the whole exercise becomes moot). By extension, a light source of substantial size cannot be reduced to a virtual source of smaller dimensions.
If I remember right the term is Brightness. You can't increase the brightness of a source.
There's no way of making 2 physical light sources into 1 virtual source (unless they occupy different parts of the spectrum, or unless you accept losses so great the whole exercise becomes moot). By extension, a light source of substantial size cannot be reduced to a virtual source of smaller dimensions.
If I remember right the term is Brightness. You can't increase the brightness of a source.
Jezz, all so negitive. Granted, I don't think it would work, but let the man try! Also wanted to say that if you can get surface mount small white LED's, this could work. The light is not being shot trough the large plastic lens of a conventional led and you could put these very close together. You could mabye make a 'super bright' backlight from these for a 15 or 17 inch lcd and use the $5 100' tv idea to get a watchable picture in a dark room(who with a diy proj watches in a well lit room anyway?). it would be a little pricey, but the 50 to 100 THOUSAND hr life would be nice. The proj would also be smaller b/c your light engine no longer needs a f/l of 14 inches.
just throwin' out an idea,
Don't get discouraged be the nay-sayers Alan, good luck, although I don't think you'll need it-you got skill
Brad
just throwin' out an idea,
Don't get discouraged be the nay-sayers Alan, good luck, although I don't think you'll need it-you got skill
Brad
golftdibrad said:
Phew! I'm glad there are other positive thinkers out there somewhere - I was starting to worry. Thanks for that post, Brad. I agree that it most likely won't work, but I've heard quite a few people talk about it in terms of theory, so I thought maybe we just ought to put it to the test. It would be excellent if we could use this as a way to reduce the size of our projectors!
That is a cool toy, but WAY too expensive - they are making a killing on it. I pay about $1 per LED at 16,000 mcd per LED. What they are using is probably about the same quality/brightness, but I'm sure they can get a better price than I can. Also, assuming those LED's are wired in series to get a total current draw of ~110v, there is very little extra electronics that would need to go into that. I think you could make your own using a round circuit board, 20 LED's, a resistor or two, and a rectifier diode for about $30-40. Also, they mention that the lamp has to be positionable since the beam is so narrow, but I'll bet you could use the lens from a cheap outdoor flood lamp to diffuse the output.Tiarb said:
Back to using LED as a projector light source. A couple of you have spoke about how it won't work because you can't get a bright point source out of it ... but what if you don't need a point source anymore? The reason we use a point source with a fresnel lens is so that we can get as much of the initial light to shine onto the back of our projection lens. But what if we arranged a bunch of LEDs immediately behind the LCD - no fresnel at all - and aligned them so that each one shines all of its light onto the projection lens? The LED's that I have are 10 degree beam, and they shine a very small pattern on my wall even from 5 feet away, which I think makes them ideal for this experiment. Alignment is, of course, the hardest part in that case because you'd have about 600 LEDs that had to be individually aligned. Sheesh ... this will be a trick!
I guess people are used to thinking how they are used to thinking...
I always thought you meant, shine the LEDs right through the panel from straight behind, not focused, collimated, whatever.
With 10 degrees from a few hundred sources, what happens to the f-o-c-u-s of the overall light source? Sure it's a small pattern, but it's not a ( . ) is it? A fresnel between the diodes and panel still eh?
Flatness should not be much of a problem...use a thick, flat PCB to mount them LEDs on.
on a side note, I have been thinking about doing this on a much dimmer scale for a tracing panel/ light table that currently uses a single fluorescent bulb.
I always thought you meant, shine the LEDs right through the panel from straight behind, not focused, collimated, whatever.
With 10 degrees from a few hundred sources, what happens to the f-o-c-u-s of the overall light source? Sure it's a small pattern, but it's not a ( . ) is it? A fresnel between the diodes and panel still eh?
Flatness should not be much of a problem...use a thick, flat PCB to mount them LEDs on.
on a side note, I have been thinking about doing this on a much dimmer scale for a tracing panel/ light table that currently uses a single fluorescent bulb.
Not to get off topic, but for your light table you might also considering using a few cold cathode tubes in the way that LCD panels use for their backlight. I'm sure that if you could make the same thing - less the LCD - that you'd have a very nice bright tracing table that you could even run off of batteries if you wanted to.
Back to LED's though, I didn't quite follow what you meant about the alignment. I know that light coming from different angles sounds like a terrible idea at first, but that's exactly what were doing when we used non-split fresnel lenses under out LCD panels - the LCD and projection lens don't know the difference, so I don't think it will matter ... true to the rest of this thread though: there's only one way to find out for sure ...
Back to LED's though, I didn't quite follow what you meant about the alignment. I know that light coming from different angles sounds like a terrible idea at first, but that's exactly what were doing when we used non-split fresnel lenses under out LCD panels - the LCD and projection lens don't know the difference, so I don't think it will matter ... true to the rest of this thread though: there's only one way to find out for sure ...
I was thinking at first about putting the LEDs behind the panel without any sort of a lens, and their various mis-focusing of 10deg each might give you a color smudge instead of a good focused image....
CCLs as a light source: That was my first idea, but there are 2 problems with using CCLs:
The power supply - I don't have one and am not interested in making a 1500V SMPS with no prior experience and the first penny to buy a (not cheap, I've checked) supply would be an infinite increase in the cost of the device...so far it's all free. LEDs cost less than $1 each, by far.
It's free because the sheets came from various backlight assemblies that were rejected from Samsung's LCD facility next door, and God only knows the condition of their lamps even if I did have a supply for them. I don't feel like hanging around to wait for the trash man at 5 in the morning every wednesday to see if they will drop a functional lamp; I picked up these parts when I came in to work over a period of months...and I don't want to buy any CCLs.
OTOH if you want a tracing table, and have maybe 3, 17" to 19" LCDs that you are turning into projectors, the diffusers, fresnels and various hard sheets make a DANDY when you experiment with them...and the guys next door have LCD monitors at all their computers the size of which would make a geek fall over backwards...
CCLs as a light source: That was my first idea, but there are 2 problems with using CCLs:
The power supply - I don't have one and am not interested in making a 1500V SMPS with no prior experience and the first penny to buy a (not cheap, I've checked) supply would be an infinite increase in the cost of the device...so far it's all free. LEDs cost less than $1 each, by far.
It's free because the sheets came from various backlight assemblies that were rejected from Samsung's LCD facility next door, and God only knows the condition of their lamps even if I did have a supply for them. I don't feel like hanging around to wait for the trash man at 5 in the morning every wednesday to see if they will drop a functional lamp; I picked up these parts when I came in to work over a period of months...and I don't want to buy any CCLs.
OTOH if you want a tracing table, and have maybe 3, 17" to 19" LCDs that you are turning into projectors, the diffusers, fresnels and various hard sheets make a DANDY when you experiment with them...and the guys next door have LCD monitors at all their computers the size of which would make a geek fall over backwards...
If you're into REAL DIY, how about making your own electroluminescent panel?
I remember an article in (I think) practical Electronics, must have been 30+ years ago, before LCDs were commercially available, about making your own completely from scratch. Dealt with tin-plating glass, etc. and also covered using it for e/l construction.
Don't know if anyone but the author ever tried it, though!
I remember an article in (I think) practical Electronics, must have been 30+ years ago, before LCDs were commercially available, about making your own completely from scratch. Dealt with tin-plating glass, etc. and also covered using it for e/l construction.
Don't know if anyone but the author ever tried it, though!
There's an article on www.howstuffworks.com about how to make your own, but it's only really good for making an on-off panel, not a panel with multiple colors, let alone multiple pixels. There are a lot of timing circuits that you'd have to make, and a bunch of multiplexors also that you'd need in order to make an LCD panel of even very low resolution. Would be cool though.
Interesting article, though not exactly practical instructions 
I wasn't really suggesting building a whole display (I agree that it would be a 'challenge!).
What I had in mind was a plain e/l panel for the lightbox application, or possibly for projector use - though I've never seena panel with anything near enough brightness for that.
I wasn't really suggesting building a whole display (I agree that it would be a 'challenge!).
What I had in mind was a plain e/l panel for the lightbox application, or possibly for projector use - though I've never seena panel with anything near enough brightness for that.
The instructions seemed straitforward as I recall, but didn't leave me with the sense that I could actually complete the project with the instructions given. By the way, what exactly does e/l mean?
E/L = electroluminescent (ever seen the backlighting on H&H amp control panels, for example, or certain types of illuminated 'exit' sign?).
In comparison, the article I read was very practical (in the sense of giving full and detailed instructions), but I recall warnings about avoiding inhaling highly toxic vapours evolved during the process. Probably wouldn't be allowed to publish it nowadays!
In comparison, the article I read was very practical (in the sense of giving full and detailed instructions), but I recall warnings about avoiding inhaling highly toxic vapours evolved during the process. Probably wouldn't be allowed to publish it nowadays!
In my best Homer voice: "Hmmmm ... vapors ..." 
So what you were trying to say is that we could try to build a bright backlight, basically? The problem with that, as with current backlights, is that the light diffuses everywhere and we want it to all go straight to the projection lens. Hopefully today I can experiment with my LED idea. I have about 19 LED's soldered to a small PCB (about 2"x2"). The light from those LED's shines onto a small region about 3" in diameter from 4-5 feet away, so it should be near ideal for trying in this experiment. I'll be attempting to project a 2"x2" part of my LCD onto the wall using the LED's. If it seems to have some potential, I'll post the images I get.
So what you were trying to say is that we could try to build a bright backlight, basically? The problem with that, as with current backlights, is that the light diffuses everywhere and we want it to all go straight to the projection lens. Hopefully today I can experiment with my LED idea. I have about 19 LED's soldered to a small PCB (about 2"x2"). The light from those LED's shines onto a small region about 3" in diameter from 4-5 feet away, so it should be near ideal for trying in this experiment. I'll be attempting to project a 2"x2" part of my LCD onto the wall using the LED's. If it seems to have some potential, I'll post the images I get.
- Status
- Not open for further replies.