Hmmmmm.. Still doesn't make sense. Let's assume that 5700 lumens are actually utilized from either the MH light source or the LED light source. Just for easier math.
The additive value of all the lumens from the LEDs is equal to the single output of the MH bulb. (63 LEDs) x (113 lumens) = around 5700 lumens. So, if the total amount of lumens out is the same from the MH and the LEDs.. How can the MH be brighter? You can't create light from nothing. The light from the MH bulb is spread over the entire LCD panel. Putting 63 LEDs in "parallel" (spreading them evenly across the LCD) does the same thing. It takes 5700 total lumens and spreads it evenly across the entire LCD; just like the fresnel does with the MH bulb. Something's missing here.
Whether it's from a fresnel (spreading the light) or from an array of lights, the total amount of lumens out from the array / fresnel is the same.
The additive value of all the lumens from the LEDs is equal to the single output of the MH bulb. (63 LEDs) x (113 lumens) = around 5700 lumens. So, if the total amount of lumens out is the same from the MH and the LEDs.. How can the MH be brighter? You can't create light from nothing. The light from the MH bulb is spread over the entire LCD panel. Putting 63 LEDs in "parallel" (spreading them evenly across the LCD) does the same thing. It takes 5700 total lumens and spreads it evenly across the entire LCD; just like the fresnel does with the MH bulb. Something's missing here.
An externally hosted image should be here but it was not working when we last tested it.
Whether it's from a fresnel (spreading the light) or from an array of lights, the total amount of lumens out from the array / fresnel is the same.
Follow me here. If I take a paper and put a hole in it. I get 5700 lumens. Anywhere I put it. If I do the same with your leds, it won't. It will be like 100 lumen. You need to think broader. it isn't a total 5700 lumens getting through. It is at any given point. If you do an led set, it won't be 5700 at any given point. you are adding them up. My metal halide doesn't put one ray of 5700 through. It puts infinity different rays of 5700 lumens through it. Like I said, go take your leds, put them behind the lcd. I bet you anything it will be 100 lumens getting through. Not 5700 like you think. The 5700 come in when those are all squished back into one, going out the lens. I am getting like 10,000-18,000 lumens at the lens hole without the lens in it. Why you ask. It is a culmination of the rays put back into a small area. Then we are dispersing it again out to the way into a very large area.
Here is what you are doing. You are taking the mix of all the light from the leds, and putting into a small circle. If I do the same thing, I am getting much more as well. But you are comparing what I am getting back of the lcd, to what you are getting at the little circle after all the light has been put back together again.
I was just thinking... The only way the MH bulb puts out more light is if you get more light to start (more than the 5700 lumens).
So if you got say, 15,000 lumens from the light source and spread it over a fresnel, THEN maybe you'd get more lumens...
No offense, but I don't trust anyone without a proof. 🙂
Now wait... You were measuring with a LUX meter right? That is something completely different than a lumen. One lux is equal to one lumen per square meter. So:
1 lux = 1 lumen / 1 meter squared
1 lumen = 1 lux * 1 meter squared
1 meter squared = 1521 inches squared
There are 135 square inches on a 15" LCD. That means there are 0.09 square meters on a 15" LCD (135 / 1521). So if you get 7500 LUX all over the 15" LCD screen:
(7500 lux) * (0.09 square meters) = 675 Lumens
Just for math sakes (5700 lux) * (0.09 square meters) = 513 lumens.
Still more than the 113 put out by the LEDs... But definitely not the 7500 lumens per LED that you suggested.
So if you got say, 15,000 lumens from the light source and spread it over a fresnel, THEN maybe you'd get more lumens...
No offense, but I don't trust anyone without a proof. 🙂
Now wait... You were measuring with a LUX meter right? That is something completely different than a lumen. One lux is equal to one lumen per square meter. So:
1 lux = 1 lumen / 1 meter squared
1 lumen = 1 lux * 1 meter squared
1 meter squared = 1521 inches squared
There are 135 square inches on a 15" LCD. That means there are 0.09 square meters on a 15" LCD (135 / 1521). So if you get 7500 LUX all over the 15" LCD screen:
(7500 lux) * (0.09 square meters) = 675 Lumens
Just for math sakes (5700 lux) * (0.09 square meters) = 513 lumens.
Still more than the 113 put out by the LEDs... But definitely not the 7500 lumens per LED that you suggested.
Ok, I'll go along with that. But are we sure that we are getting 113 lumen from the leds through the lcd, or 113 at your ratio 113 x .09 meters
Actually, it's a much better ratio. You are applying 113 lumens to a very small area (whatever the spacing for 63ish LEDs is). With the MH bulb you are pushing 5700ish lumens onto 135 square inches (the 15" LCD) and getting X amount of lux (7500 in your case). With the LED, you taking 113 lumens and pushing it onto a portion of the LCD; something like a 2" x 1" space. So..... Let's do a little math:
1 lux = 1 lumen / 1 meter squared
So 2" x 1" is 2 square inches or 0.0013 square meters (2 / 1521). So you have:
(113 Lumens) / (0.0013) = 86,923 lux (before going through the LCD)
Sounds like a lot, but you're pushing all the light from it into a space 2" x 1".
This is of course assuming you could push all 100% of the light from the LED exactly into a 2" x 1" space; which is only theory. It'll be lower than that in real life.
I hope I don't come across as demeaning. I have no interest in putting people down. I'm interested in discovery, facts, and the truth.
1 lux = 1 lumen / 1 meter squared
So 2" x 1" is 2 square inches or 0.0013 square meters (2 / 1521). So you have:
(113 Lumens) / (0.0013) = 86,923 lux (before going through the LCD)
Sounds like a lot, but you're pushing all the light from it into a space 2" x 1".
This is of course assuming you could push all 100% of the light from the LED exactly into a 2" x 1" space; which is only theory. It'll be lower than that in real life.
I hope I don't come across as demeaning. I have no interest in putting people down. I'm interested in discovery, facts, and the truth.
huh? That made no sense to me. again, it is 5700 lumen with infinity times. It isn't a single ray. And what do you mean you aren't pushing the LEDs through the same area. Of course you are. You have to push as many through the same given area??
I think you are reaching for something rather than conceding. We are talking about what ever you push light wise, to the other side of the lcd, read the lumens at that point. It is still the same, you have to compare apples to apples. The far side of the lcd. That tells you what lumens you have x .09. If you are comparing the apples of led output at the source, than you have to do the same with the small arc of the metal halide. 71mm x 20mm at 95,000 lumens or what ever you are calculating. I am not good at that stuff.
I think you are reaching for something rather than conceding. We are talking about what ever you push light wise, to the other side of the lcd, read the lumens at that point. It is still the same, you have to compare apples to apples. The far side of the lcd. That tells you what lumens you have x .09. If you are comparing the apples of led output at the source, than you have to do the same with the small arc of the metal halide. 71mm x 20mm at 95,000 lumens or what ever you are calculating. I am not good at that stuff.Oh, and bye the way, I am actually having fun posting with you. I have no hard feelings. When you start calling me stupid or something. Then we will have a problem
What I meant was that 1 LED (out of the 63) is being "pushed" through a smaller area than the MH bulb is. 1 LED is being pushed through a 2" x 1" space as opposed to a MH bulb which is being pushed through a 12" x 9" space. The culmination of the 63 LEDs pushes light through the same amount of space (12" x 9"), but each LED only shines on a 2" x 1" patch of the total area. I hope that clears it up.
I see where you are going. But you aren't understanding that for every 2" x 1" square you are talking about, I have 95,000 lux hitting in the same area you have 113. It isn't 95,000 lumen divided by the square size. It is an infinite amount of 95,000's hitting it all over. Like yours is 113 times infinity. The bulb doesn't put 95,000 and divide it into each direction. It puts out that in all directions. Minus obviously directional is higher than outside areas. But none the less, let's say average of 75,000 lumens hit at every 2" x 1" spot.
I don't think you getting 75,000 at your LCD. That's my point. I think you're getting more like 12,000 lumens just before the LCD. If you did get 75,000:
75,000 * 0.08 = 6000 lumens (after LCD with efficiency of 8%)
6000 * 0.90 = 5400 lumens (after field fresnel, assuming split config)
Maybe another 10% loss after projection lens or mirror if you have one, but the point is, you'd have way over 4000 lumens getting to the screen (compared with the 200 - 400 lumens most people are getting and 1200 - 1500 lumens on commercial projectors). Assume that's on a 100" diag screen:
1 lux = 1 lumen / 1 meter squared
lux = 4800 lumens / 3.16 meters squared = 1519.99
So your lux meter would read over 1500 lux on average at the screen. If you get that, please share your secrets. 🙂
Also, it does make a difference how much area you spread the light onto. With a MH bulb, you're spreading it over a 12" x 9" area. With each LED, you are spreading it over a 2" x 1" area. Picture this: If you have a flashlight with an adjustable lens (the one where you twist the front of the flashlight and the light beam changes, like a MAG Light). When you focus the beam to be a spot light, you get a very bright beam that only lights up a small area. If you adjust the focus (turn the head), you get a more distributed beam. It's not nearly as bright, but it covers a whole lot more area.
Keep this concept in mind. The MH is like the distributed beam and the LED would be the spot beam. If the MH and the 1 LED put out the same amount of lumens, the LED would be much brighter because it doesn't have to cover a huge area. But, because the MH is actually brighter, it will be just as bright as the spot LED beam.
This concept is also the reason why you get a brighter picture by having a smaller screen size.
75,000 * 0.08 = 6000 lumens (after LCD with efficiency of 8%)
6000 * 0.90 = 5400 lumens (after field fresnel, assuming split config)
Maybe another 10% loss after projection lens or mirror if you have one, but the point is, you'd have way over 4000 lumens getting to the screen (compared with the 200 - 400 lumens most people are getting and 1200 - 1500 lumens on commercial projectors). Assume that's on a 100" diag screen:
1 lux = 1 lumen / 1 meter squared
lux = 4800 lumens / 3.16 meters squared = 1519.99
So your lux meter would read over 1500 lux on average at the screen. If you get that, please share your secrets. 🙂
Also, it does make a difference how much area you spread the light onto. With a MH bulb, you're spreading it over a 12" x 9" area. With each LED, you are spreading it over a 2" x 1" area. Picture this: If you have a flashlight with an adjustable lens (the one where you twist the front of the flashlight and the light beam changes, like a MAG Light). When you focus the beam to be a spot light, you get a very bright beam that only lights up a small area. If you adjust the focus (turn the head), you get a more distributed beam. It's not nearly as bright, but it covers a whole lot more area.
Keep this concept in mind. The MH is like the distributed beam and the LED would be the spot beam. If the MH and the 1 LED put out the same amount of lumens, the LED would be much brighter because it doesn't have to cover a huge area. But, because the MH is actually brighter, it will be just as bright as the spot LED beam.
This concept is also the reason why you get a brighter picture by having a smaller screen size.
I thought it has been shown that commercial projectors actually only get 500-800 lumens at the screen. I haven't seen any results showing 1200 please put that post up so I can see it. Also, this isn't about commercial projectors, that would put your led theory even further away from what you want to achieve. I am getting 350.06226 lumen, and believe until someone shows me different, that a LED one will give about 10-30 lumen. That is my whole point. If someone can show me I am wrong with actual numbers like I have shown, than be my guest.
I am getting average of 55,888.88 lux in 9 areas before the fresnel lcd, 5029.99 per square meter. After them, average of 6366.66. thus, 572 lumen per square meter after the lcd.
I am getting average of 55,888.88 lux in 9 areas before the fresnel lcd, 5029.99 per square meter. After them, average of 6366.66. thus, 572 lumen per square meter after the lcd.
Here's a 17 inch lcd with 1700 leds. He states he is getting 66 lumens on screen, without the system being tweaked. It's in French so you'll have to translate it.
http://forum.aspect-geek.net/aspect...0w-LED-Allinbox-projet-HDLED-sujet-8342-1.htm
http://forum.aspect-geek.net/aspect...0w-LED-Allinbox-projet-HDLED-sujet-8342-1.htm
Man, I have to learn French. AllInBox is a great site, just can't understand a thing they say. For some reason Babblefish isn't working for me and google only translates 2/3 of the page! Injustice I tell you. 🙂
This guy, I think, was saying he got 1300 lumens from the array of LEDs. He says he got about 66 lumens out of the projector. So if you could get, say, 5700 lumens from the array, you'd get (theoretically) 289 lumens. This guy was also using really weak LEDs (20mA each). The LEDs I was talking about were 1000mA. He used 1710 LEDs (20mA each) that use 50 times less current (less power too). So if you used 50 times fewer LEDs, 34, we're getting close. Of course, the more powerful LEDs (1000mA) are not exactly 50 times brighter. So you will need to use more LEDs... say.. 63 (almost double the theoretical amount here).
I should just go and buy a Luxeon L7 (113 lumens) and a lux meter and see what happens. How much are those Lux meters?
This guy, I think, was saying he got 1300 lumens from the array of LEDs. He says he got about 66 lumens out of the projector. So if you could get, say, 5700 lumens from the array, you'd get (theoretically) 289 lumens. This guy was also using really weak LEDs (20mA each). The LEDs I was talking about were 1000mA. He used 1710 LEDs (20mA each) that use 50 times less current (less power too). So if you used 50 times fewer LEDs, 34, we're getting close. Of course, the more powerful LEDs (1000mA) are not exactly 50 times brighter. So you will need to use more LEDs... say.. 63 (almost double the theoretical amount here).
I should just go and buy a Luxeon L7 (113 lumens) and a lux meter and see what happens. How much are those Lux meters?
I paid $45 shipped off of egay. Well worth it, not only for the projector, but theories about Low-e glass, glass pieces, lcd etc loss. It has told me tons. I was surprised to learn I lose about 20% from the low-e and 90% with the lcd in the way. Very surprising. No wonder we lose so much to the wall. Not the light problem, but the lcd that is sucking our life out.
hey superdave, and guy arguing with him 🙂.
I dont know if you remember me, your name is asdfk or something of that sort on the diypc forums, and I was talking to you as you were finishing up, and I was trying and failing at doing the same.
Ive given up on the MH + ballast because my ballast is loud and my box was huge and heavy. Im going for LED, I dont care what people say I think it can be done well.
The problem with LED lumens is that they are specified with mcd, which dont translate easily to lumens. The best way to translate is to take the angle of projection and use it with the mcd. There is a calculator at linear1's led website, its a nice website for learning about LED's.
Basically, from a lot of the failed attempts I see 30000-40000 mcd LEDs being used, but there are 130000 mcd LEDs out there as well. Granted, they are 10mm, while some of the ones being used currently are 5mm, but I think they still will perform better.
Although Im not familiar with lux, I do know that generally speaking only about 15-20 percent of the light from a point source, even with a reflector, will reach the LCD via the rear fresnel. Thus, the maximum amount of lumens to reach the LCD will be around 4000.
I figured out that with 10mm, 130000mcd, you could get around that number of lumens with around 1000LEDs, which is just about the number I plan on eventually hitting.
I am going to start with half that amount, since they arent super cheap, but I should be able to get 500 for around 150, so it wont be too bad. Depending on how dim those results are I will upgrade and ad more.
Its true that there arent many great results out there, but I think using the full size of a 17" screen and the brightest LEDs out there it just might be possible.
Also the PJ is going into a pitch black basement so it will be ok even if its a bit dim.
I dont know if you remember me, your name is asdfk or something of that sort on the diypc forums, and I was talking to you as you were finishing up, and I was trying and failing at doing the same.
Ive given up on the MH + ballast because my ballast is loud and my box was huge and heavy. Im going for LED, I dont care what people say I think it can be done well.
The problem with LED lumens is that they are specified with mcd, which dont translate easily to lumens. The best way to translate is to take the angle of projection and use it with the mcd. There is a calculator at linear1's led website, its a nice website for learning about LED's.
Basically, from a lot of the failed attempts I see 30000-40000 mcd LEDs being used, but there are 130000 mcd LEDs out there as well. Granted, they are 10mm, while some of the ones being used currently are 5mm, but I think they still will perform better.
Although Im not familiar with lux, I do know that generally speaking only about 15-20 percent of the light from a point source, even with a reflector, will reach the LCD via the rear fresnel. Thus, the maximum amount of lumens to reach the LCD will be around 4000.
I figured out that with 10mm, 130000mcd, you could get around that number of lumens with around 1000LEDs, which is just about the number I plan on eventually hitting.
I am going to start with half that amount, since they arent super cheap, but I should be able to get 500 for around 150, so it wont be too bad. Depending on how dim those results are I will upgrade and ad more.
Its true that there arent many great results out there, but I think using the full size of a 17" screen and the brightest LEDs out there it just might be possible.
Also the PJ is going into a pitch black basement so it will be ok even if its a bit dim.
i'm in a similar situation to some of you...i have an lcd projector (infocus lp240), and i don't want to waste money on electricity and uhp lamps anymore...(read more at http://www.bigscreenforums.com/forum_topic.cfm?which=1263 and at http://www.avforums.com/forums/showthread.php?t=298216 )...both are pretty similar threads started by me, and should both contain any useful (?) links
i've joined your forum, as we may be able to help each other with shared info (though i'm a bit thick and probably know no more than anyone else !)
for me, and my front projector, there are three main issues...
1 - the biggest single issue with one of these pjs is to fool the circuitry that a standard lamp is running without massive energy wasting and hot resistors, this would need to be done with either the help of someone who could rewrite the firmware (very diificult for me/very easy for someone, i just don't know such a person),
or by using a microprocessor or microcontroller directly wired into the control circuit or perhaps by shorting some part of the control circuit...if anyone can help with that, that'd be great !
2 - i don't even know how bright i need the leds to be, i reckon that the polarizers and neutral filters in the pj can be removed/modified/replaced to increase usuable light, i think i need about 200 to 400 lumens at exterior lens, but to do this i'll need anywhere from 1,200 to app 12,000 lumens at lamp
3 - if possible i need to make this lamp only 45mmx45mmx45mm !
using luxeon III emitters (80 lumens each), i think i might be getting somewhere close, as 16 of these will fit into the required space and emit 1200 dimmable lumens
my (a bit vague) ingredients which i am told may be slightly cheaper than i expected here (don't know exactly how much cheaper)...
*the leds have dropped in price*
16no lxhl-pw09 $3.45/£2.06/£2.42inc vat each?????£40
plus dimmable power supply/drivers ??????????????£40
plus ancilliaries/heatsink ??????????????????????£60
total ???????????????????????????????????????????£140
price is something along the lines of an estimated £140/$200us
heres a link to the Luxeon III Emitter - LXHL-PW09, White 5500K, 80 Lumen, Lambertian led http://www.lumiledsfuture.com/produc...cfm?familyId=7
if i can help, or if you can help me, please get in touch
hi stelleg151
i wonder if you could supply a link to your 130000mcd leds...are they the same luxeons ?
thanks, matt
🙂
i've joined your forum, as we may be able to help each other with shared info (though i'm a bit thick and probably know no more than anyone else !)
for me, and my front projector, there are three main issues...
1 - the biggest single issue with one of these pjs is to fool the circuitry that a standard lamp is running without massive energy wasting and hot resistors, this would need to be done with either the help of someone who could rewrite the firmware (very diificult for me/very easy for someone, i just don't know such a person),
or by using a microprocessor or microcontroller directly wired into the control circuit or perhaps by shorting some part of the control circuit...if anyone can help with that, that'd be great !
2 - i don't even know how bright i need the leds to be, i reckon that the polarizers and neutral filters in the pj can be removed/modified/replaced to increase usuable light, i think i need about 200 to 400 lumens at exterior lens, but to do this i'll need anywhere from 1,200 to app 12,000 lumens at lamp
3 - if possible i need to make this lamp only 45mmx45mmx45mm !
using luxeon III emitters (80 lumens each), i think i might be getting somewhere close, as 16 of these will fit into the required space and emit 1200 dimmable lumens
my (a bit vague) ingredients which i am told may be slightly cheaper than i expected here (don't know exactly how much cheaper)...
*the leds have dropped in price*
16no lxhl-pw09 $3.45/£2.06/£2.42inc vat each?????£40
plus dimmable power supply/drivers ??????????????£40
plus ancilliaries/heatsink ??????????????????????£60
total ???????????????????????????????????????????£140
price is something along the lines of an estimated £140/$200us
heres a link to the Luxeon III Emitter - LXHL-PW09, White 5500K, 80 Lumen, Lambertian led http://www.lumiledsfuture.com/produc...cfm?familyId=7
if i can help, or if you can help me, please get in touch
hi stelleg151
i wonder if you could supply a link to your 130000mcd leds...are they the same luxeons ?
thanks, matt
🙂
Attachments
throwit said:
http://cgi.ebay.com/White-LED-Set-o...603505881QQcategoryZ66952QQrdZ1QQcmdZViewItem
Those are the ones that I was looking at, probably are similar to the luxeons, but I dont know if thats what youre looking for since you are not doing a large array. Also the luxeons tend to have heatsinks so they can run at higher currents/brighter.
stelleg151 I remember you!
Yeah, I am that Asdasl on DIYPC. My secrets are getting out!
I think starting small and building bigger is a great way to go. I was suggesting the use of the Luxeon K2. It has a lumen output of 113 at 1000mA (3.42V). So that's a little under 3.5W. I've been trying to find the mcd rating for these LEDs, but the results were varied.
To quote www.superbrightleds.com:
This link is helpful too:
http://led.linear1.org/how-do-i-convert-between-candelas-and-lumens/
Controlling the LEDs is not terribly hard. You can use pulse width modulation (PWM) which involves a chip or just use a potentiometer (variable resistor) to adjust the brightness. There are some calcs involved and some circuit diagrams, but it's not rocket science.
Keeps us updated on your progress!
Yeah, I am that Asdasl on DIYPC. My secrets are getting out!
I think starting small and building bigger is a great way to go. I was suggesting the use of the Luxeon K2. It has a lumen output of 113 at 1000mA (3.42V). So that's a little under 3.5W. I've been trying to find the mcd rating for these LEDs, but the results were varied.
To quote www.superbrightleds.com:
This link is helpful too:
http://led.linear1.org/how-do-i-convert-between-candelas-and-lumens/
Controlling the LEDs is not terribly hard. You can use pulse width modulation (PWM) which involves a chip or just use a potentiometer (variable resistor) to adjust the brightness. There are some calcs involved and some circuit diagrams, but it's not rocket science.
Keeps us updated on your progress!
Yeah, keep us informed. Superdave, just curious (real question) How do you get the leds back into a small point? Do you use only 1 fresnel or what. I was thinking about how you guys intend to use everything that is given off the LED>
- Status
- Not open for further replies.