I think there is a misconception in regards to heat. I think you will find that if the efficiencies stated in previous posts are correct a MH lamp will run cooler than a cluster of led’s, having the same lumens outputting. The fact is you can’t change the laws of physics. Nearly all the electrical energy that is applied to an ark or filament is transformed into light energy and heat energy, this is where the efficiency is derived, the more energy that is converted to lumens, the less energy will be converted to heat. So heat and light always equal the energy applied, nothing more and nothing less. If an led cluster is 50Lm/W and a MH is 80Lm/W then there is at least 30Lm/W worth of energy in the led cluster not being converted to light so it must be converted to heat instead.
Until the led becomes more efficient than the MH, the MH will win on heat and on Lumens. Just remember that because you can't feal much heat from an led, that doesn't mean its not there. multiply the watts used by the led, so that its in the same range as a MH and you will feel it then.
DJ
Would it be feasable to mould a lightguide from, say, polyester resin, perhaps embedding the LED matix in the process?
Hey Ace I was looking at your LED light guide and it looked like a fiber optic cable. If you are going to use this idea it will work bacause fiber optic can transfer light to one piont with little or no intensity lost, but LED will not be a great light source. Unless you use a lot of them.
rmccoll said:
heya buddy thats exactly what it is except we are making ours out of square optical plexi, if we use round cable we have a problem, when they are all placed together ( say 400) we will have gaps and those gaps inbetween the cables will be dark spots on the lcd so we use square. Each led is asigned to its own guide so there will be alot of them lol, intensity losses will be very minimal even if the eye can notice at such a short distance, it all depends on how polished you get the plexi and the material used, put it this way it will be far brighter then having the same light from that led going through air, infact that brings me to another point, glue the led in with plexi resin, this will even be less loss of light.
Plexi resin = solvent and plexi shavings, stair till desolved and till its a paste, and there you have plexi resin, Drying time is anything from 24 - 72hrs and the strength is the same as the original plexi. USE CLEAN PLEXI SHAVINGS! lol
Trev🙂
dnsey said:
You could but it wont work right, the guide has to be no wider then the led, ( or you loose intensity ), if we make a big solid block you will have the leds in the middle making hot spots in the center of the lcd, you will also loose light and the light wont be guided correctly.
Trev🙂
Ive got one of those here actually, i should take a pic of it for ya, they are very bright condensers and have very high power.
Trev🙂
Trev🙂
I was looking at various lighting and just found this about car makers going to LEDs for headlamps. I am sure the price for the bright ones will have to be coming down some if the do go into production.
http://www.canadiandriver.com/news/030429-4.htm
http://www.canadiandriver.com/news/030429-4.htm
Have any of you guys checked out this Lamina light engine thing?
Have any of you guys checked out this Lamina light engine thing?
http://www.laminaceramics.com/
They claim to have a "7 cavity light engine" that will produce 1000 lumens. I checked at www.mouser.com, and it sells for about $40. Whats even more interesting, is that they sell a 39 cavity light engine for $160. Im a little weary though, Im curious as to how exactly they come to this lumen rating, and whether or not it is truely comprable to a metal halide lamp at that rating.
-Orochi
Have any of you guys checked out this Lamina light engine thing?
http://www.laminaceramics.com/
They claim to have a "7 cavity light engine" that will produce 1000 lumens. I checked at www.mouser.com, and it sells for about $40. Whats even more interesting, is that they sell a 39 cavity light engine for $160. Im a little weary though, Im curious as to how exactly they come to this lumen rating, and whether or not it is truely comprable to a metal halide lamp at that rating.
-Orochi
Ok, I think I am now officialy disillusioned.. =/
Well, I dug arround a little deeper at the lamina site, and it looks like their claim of 1000 lumens for the 7 cavity part, isnt even close to correct. I dont know what exactly they where trying to pull by putting the sentance "up to 1000 lumens" right under the 2000 series 7 cavity part, but when I downloaded it's product spec sheet, it said it had a rating of 100 for luminous flux. Maybe Im getting myself confused again though, so it might still be a good idea for someone else to take a look at their site.
-Orochi
Well, I dug arround a little deeper at the lamina site, and it looks like their claim of 1000 lumens for the 7 cavity part, isnt even close to correct. I dont know what exactly they where trying to pull by putting the sentance "up to 1000 lumens" right under the 2000 series 7 cavity part, but when I downloaded it's product spec sheet, it said it had a rating of 100 for luminous flux. Maybe Im getting myself confused again though, so it might still be a good idea for someone else to take a look at their site.
-Orochi
have you guys (hope it wasnt mentioned earlier) Super Bright White LEDS in the Dichroic Halogen reflector??
might help for that colour problem....
might help for that colour problem....
Your LEDS may not last as long as you think
A lot of discussion is going on about 5mm white LEDS for your light engine. While the cost may be higher than MH, there is an expecation of very long life to compensate.
Most 5mm white LEDS DO NOT have a long life, at least not putting out full brightness and nice white light. The combination of the phosphor and epoxy encapsulation actually results in a significant reduction in light output at 10K hours, possibly being down as much as 50%.
About the only white LEDs that guarentee long life white output are LuxeonI and LuxeonIII LEDS.
Many vendors talk about 100K life, but do they actually show a light output versus time curve, and do they show it at the same current as they do for their maximum brightness?
Another thing to consider when you are putting all those LEDS together is that the color temperature of the LEDS can be signicantly different from LED to LED. To that end, make sure you are mixing the light of numerous LEDS in order to get consistency. The color temperature is likely lot dependant, so watch how you mount them in groups. You could end up with big differences from one part of your display to another.
p.s. You may also want to wait a little... big things on the horizon.
A lot of discussion is going on about 5mm white LEDS for your light engine. While the cost may be higher than MH, there is an expecation of very long life to compensate.
Most 5mm white LEDS DO NOT have a long life, at least not putting out full brightness and nice white light. The combination of the phosphor and epoxy encapsulation actually results in a significant reduction in light output at 10K hours, possibly being down as much as 50%.
About the only white LEDs that guarentee long life white output are LuxeonI and LuxeonIII LEDS.
Many vendors talk about 100K life, but do they actually show a light output versus time curve, and do they show it at the same current as they do for their maximum brightness?
Another thing to consider when you are putting all those LEDS together is that the color temperature of the LEDS can be signicantly different from LED to LED. To that end, make sure you are mixing the light of numerous LEDS in order to get consistency. The color temperature is likely lot dependant, so watch how you mount them in groups. You could end up with big differences from one part of your display to another.
p.s. You may also want to wait a little... big things on the horizon.
If it was me I would look at trying to use surface mount LEDs. The heat dissipation factor of the LED is limited by the clear plastic around the diode. Then I would try to place a heat sink on the back side of the PCB holding the LEDs with lots of little holes in the board with copper in them to transfer the heat to the heatsink.
I would then take a single sheet of plexi and mill lens into it in front of each LED.
The previous post about efficiency is correct but, an LED gives off light at only one wavelength. A MH light gives off light at a spectrum almost equal to the sun, if so designed. I use them over my aquariums for live coral. (another time consuming hobby) What this means is that the amount of light per watt appears to be greater to the eye. White LEDs only became possible when the blue LED was invented 12-14 years ago. A white LED only gives off light at the 3 basic wavelengths.
Where am I going with this... SM LEDs are very small, it may be possible to install enough of them to get the light output you're looking for. 2-3 SM LEDS can be mounted in the same space as 1 T13/4 device. With a heatsink attached the LEDs can be run at a higher level. It's true that the half life of a white LED is in the 10K hour range, maybe. But I replace my MH lamps on my aquariums once a year because the half life of the bulbs are only 2K hours.
Don't give up trying. You don't know what you might stumble across.
I would then take a single sheet of plexi and mill lens into it in front of each LED.
The previous post about efficiency is correct but, an LED gives off light at only one wavelength. A MH light gives off light at a spectrum almost equal to the sun, if so designed. I use them over my aquariums for live coral. (another time consuming hobby) What this means is that the amount of light per watt appears to be greater to the eye. White LEDs only became possible when the blue LED was invented 12-14 years ago. A white LED only gives off light at the 3 basic wavelengths.
Where am I going with this... SM LEDs are very small, it may be possible to install enough of them to get the light output you're looking for. 2-3 SM LEDS can be mounted in the same space as 1 T13/4 device. With a heatsink attached the LEDs can be run at a higher level. It's true that the half life of a white LED is in the 10K hour range, maybe. But I replace my MH lamps on my aquariums once a year because the half life of the bulbs are only 2K hours.
Don't give up trying. You don't know what you might stumble across.
Same issue with SMT LEDS for the most part. The encapsulation method results in drastically reduced output over time. The other problem with the 5mm and most SMT LEDS is the lumens per W is quite low compared to the truly high powered LEDS out there.
The best white LEDs that are actually shipping are on the order of 30 lumens\watt. There is nothing better shipping in volumn.
Most white LEDS do not emit at the three primary colors. They emit Blue + a yellowish color that is derived from a phosphor. The holy grail of white LED development is not considered to be RGB led, but UV leds with a phosphor. The reason for this is that you will be able to achieve a much higher consistency level w.r.t. the white point or perceived color of the white LED.
Alvaius
The best white LEDs that are actually shipping are on the order of 30 lumens\watt. There is nothing better shipping in volumn.
Most white LEDS do not emit at the three primary colors. They emit Blue + a yellowish color that is derived from a phosphor. The holy grail of white LED development is not considered to be RGB led, but UV leds with a phosphor. The reason for this is that you will be able to achieve a much higher consistency level w.r.t. the white point or perceived color of the white LED.
Alvaius
You guys are saying that the main reason to pay more for LEDs is the amount of time they last. 10,000 hours in my opinion, is more than enough. I would be willing to say many diy projectors don’t make it past 1,000 hours before being revamped or improved. I'm not even sure if anyone on this forum has ever blown a metal halide bulb from normal use since their life is so long, lol 😀. LEDs will last long enough. And if you could build a light guide device that made is possible to hot swap dim or burned out LEDs, all the better. If one LED was brighter then another, you could dim it down with a resistor. At this point the main reason for going LED is bragging rights lol. They are really not at all economical and for me makes it fun to attempt. One of these days I'll start my prototype and let you guys know how it goes.
At 10,000 hours, some of them could be down 60%, and certainly the color could have dramatically shifted. You need to expect that even when you first turn them on, that if they are not all from the same batch, that there will be large differences in both color and brightness. Just something to keep in mind when are building your unit.
I think you will find the viability of a white LED light source for projectors will increase greatly in the next 6 months as new products come to market.
Alvaius
I think you will find the viability of a white LED light source for projectors will increase greatly in the next 6 months as new products come to market.
Alvaius
Watch out for blue LED + phosphor!
You may have a problem using most of the super bright "white" LEDs, since they use a blue LED to stimulate a yellow phosphor layer. If the right balance of blue leaks through, we perceive it as white. But it really only has two spectral lines: Blue and yellow. The blue line will get through the LCD's blue filters fine, but the yellow line may fall outside the bands of the LCD's red and green filters! The result would be a very blue image with dark reds and greens.
There are white LEDs that incorporate red, green, and blue emitters, but these are more expensive and rare (and a lot dimmer). Maybe the next round of technology (UV LEDs with a multi-color white phosphor) will work.
This problem points out an interesting phenomenon: Humans see a single color based on how much it stimulates our seperate red, green, and blue receptors. (The bands overlap a bit.) We see a light as yellow if it contains a narrow red band and a narrow green band. We also see a light as the very same shade of yellow if it has a single band of the proper frequency (true yellow color). If you apply a filter that cuts off just the green band, the single-band light source will still look yellow, but the two-band light source will turn red! You can see this effect by looking at different objects through blue-blocking sunglasses.
You may have a problem using most of the super bright "white" LEDs, since they use a blue LED to stimulate a yellow phosphor layer. If the right balance of blue leaks through, we perceive it as white. But it really only has two spectral lines: Blue and yellow. The blue line will get through the LCD's blue filters fine, but the yellow line may fall outside the bands of the LCD's red and green filters! The result would be a very blue image with dark reds and greens.
There are white LEDs that incorporate red, green, and blue emitters, but these are more expensive and rare (and a lot dimmer). Maybe the next round of technology (UV LEDs with a multi-color white phosphor) will work.
This problem points out an interesting phenomenon: Humans see a single color based on how much it stimulates our seperate red, green, and blue receptors. (The bands overlap a bit.) We see a light as yellow if it contains a narrow red band and a narrow green band. We also see a light as the very same shade of yellow if it has a single band of the proper frequency (true yellow color). If you apply a filter that cuts off just the green band, the single-band light source will still look yellow, but the two-band light source will turn red! You can see this effect by looking at different objects through blue-blocking sunglasses.
Sorry for not going back in the thread to check...
I did an experiment at work involving 15,000 mcd red LED's in a 5mm package.
Because of thier fast turn on time, we found out we could increase the brightness substantialy by pulsing them with a higher voltage for very small amounts of time at about 100Hz. At that frequency, you can't tell its actually being pulsed and you get the benfit of the increased brightness. It actually 'seemed' brighter when the pulses where smaller (small enough so the oscilloscope couldn't measure them.. its not a great scope though).
You can even use them at their rated voltage for less current and extend the life of the LED's.
All you have to do is keep the pulse short enough so the average current isn't greater than the 30ma most can take.
The whole thing was based on a 555 timer and a transister (A FET would definately be better in this case)
I just though I'd mention it because I have some experience using it first hand.
I did an experiment at work involving 15,000 mcd red LED's in a 5mm package.
Because of thier fast turn on time, we found out we could increase the brightness substantialy by pulsing them with a higher voltage for very small amounts of time at about 100Hz. At that frequency, you can't tell its actually being pulsed and you get the benfit of the increased brightness. It actually 'seemed' brighter when the pulses where smaller (small enough so the oscilloscope couldn't measure them.. its not a great scope though).
You can even use them at their rated voltage for less current and extend the life of the LED's.
All you have to do is keep the pulse short enough so the average current isn't greater than the 30ma most can take.
The whole thing was based on a 555 timer and a transister (A FET would definately be better in this case)
I just though I'd mention it because I have some experience using it first hand.
- Status
- Not open for further replies.