DIY Video Projector Part II

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
diyAudio Member
Joined 2003
Ok, I think I see where I got confused , mutiplying the candela by 12.57 seems to be a way to convert to lumens when your light source radiates sphericaly or something (or at least I think thats what they meant, still not sure really), where as leds only cast their light at a particular "viewing angle".

Playing with leds and thier brightness can be tricky, i personally dont fool with leds, Yet! but may do further down the line so realy i have no say on the matter, but what i did know is that 250lumens will never come from a 20000cmd led.

Parabolic reflected lamps are measured like leds aswell, anything with a beam wont be measured in lumens on a manufacturers site, why? not sure, but its to do with the beam angle and there is a way to convert it back.

Regarding light loss, there is a way we can tackle this without a great loss of light, but a few calculations need to be kept in mind, here is the thread where i posted the idea.

http://www.diyaudio.com/forums/showthread.php?postid=459580#post459580


Trev:)
 
My 2c:

A narrower beam may give a lower lumen output per LED but will probably produce less light loss (to the surroundings before the LCD) and give a much crisper picture.

I have yet to see anyone here do any decent pulsing with these white LED's, traditional LED's should take a couple of Amps if it's for a very short period of time giving a perceived increase in brightness. It would be great to find out how much current the cheap white LED's can take and if it made their colour better or worse.

verbose mustafa's calculations look spot on, great work with the dos program by the way. :)
 
White LEDs

Personally, I love the new super-bright white LEDs, but there are a few issues that are being ignored here.

1) LEDs are less efficient than MH lamps. That power consumed, but not converted to light, has to go somewhere. Where it goes is... heat! So an LED projector light source will actually get hotter than a same-lumen MH lamp light source. (The heat WILL be spread out more, but you will need more airflow to get it out of the box.) It will also cost you more to power it.

2) This is the killer: All the very bright whites I have seen use a narrow-band blue LED to stimulate a narrow-band yellow phosphor. When we see blue + yellow, the blue stimulates our eye's blue receptors, while the monochromatic yellow stimulates both our red and green receptors. So it looks "white" to us. But it does not look like an even mixture or red, green, and blue to the red, green, and blue filters on the LCD panel! The blue will get through fine, but reds & greens will be very dark. This is why LED projection tests go through the lenses just fine, but look horrible when you pass it through an LCD.

3) There are true RGB LEDs that are used for devices that need to create any color on demand, like paint matchers. But those LEDs are much more expensive and much too dim. The other option is to make your own array of R, G, & B LEDs. Then you would get each color you need, but you would also introduce lots of color artifacts in the image.
 
70000:1 idea

They get "70000:1 contrast ratio" by using a backlight made of individual pixel LEDs? But that leads to a more interesting question:

Once you have a full-sized panel of RGB LEDs,why would you put an LCD in front of it? Why not use the LEDs as the display device directly. You could run them at 10% of the power needed to get the same amount of light through LCD filters, etc.
 
Re: 70000:1 idea

Guy Grotke said:
They get "70000:1 contrast ratio" by using a backlight made of individual pixel LEDs? But that leads to a more interesting question:

Once you have a full-sized panel of RGB LEDs,why would you put an LCD in front of it? Why not use the LEDs as the display device directly. You could run them at 10% of the power needed to get the same amount of light through LCD filters, etc.


Size?
Speed?
Soldering?
 
70000:1 idea

>Size?
>Speed?
>Soldering?

My point was that the 70000:1 contrast ratio idea doesn't make much sense. If you have individual pixel backlights, why not just use those as the display device? I don't see what putting an LCD in front of them gets you.

Size: If you have individual pixel LED backlights, then that has to be exactly the same size as the LCD panel. (In fact, they would have to very precisely aligned.)

Speed: LEDs are thousands of times faster than LCD pixels.

Soldering: An LED array like this would have to be fabricated in a semiconductor processing facility, rather like where they make LCD panels. No soldering would be involved, just lots of photo-etching and vapor depositing using very nasty chemicals.

AND it doesn't help us at all, since it would result in a panel with a fixed backlight. Even if you could seperate the backlight from the LCD, the LCD would still be 400:1 max.
 
LED

Hello all again , i have something which you lot will find very intresting. I have still been followig the project through , and a new LED device has just come into a local electronics store claming this new LED gives more light that a Halogen bulb and is priced at £8 each. Here are the details.

• Very high brightness
• Very high flux per LED
• Very long operating life (to 100k hours)
• Cool beam, safe to the touch
• Instant light (less than 100ns)
• Fully dimmable
• No UV Superior ESD protection
• More energy efficient than incandescent and most halogen lamps

Luxeon star hexagonal modules can be connected and powered in a close packed hexagonal configuration for tight spacing and increased light output per unit area. They are available in white, green, cyan, blue, royal blue, red and amber. Typical applications include passenger reading lights, portable lighting, orientation/emergency lighting, mini accent lighting, decorative lighting, fiber optic alternatives and other appliances. This range of products have a batwing (low dome) radiation pattern and slots in the aluminium-core PCB for M3 or #4 mounting screws. All codes have a viewing angle of 110°

if anyone has any views i be pleased to hear them :)
 
or there is this one , data sheets are avalible

Luxeon V Emitter Lambertian


• Highest flux per LED family in the world
• Very long operating life (up 100k hours)
• Available in white, green, cyan, blue, royal blue
• Lambertian radiation pattern
• More energy efficient than incandescent and most halogen lamps
• Low voltage
• DC operation
• Cool beam, safe to touch Instant light (less than 100nS)
• Fully dimmable
• No UV Superior ESD protection

Luxeon III is a revolutionary, energy efficient and ultra compact new light source, combining the lifetime and reliability advantages of LEDs with the brightness of conventional lighting.
Luxeon III is rated for up to 1000mA operation, delivering increased lumens per package.
Luxeon emitters give you total design freedom and unmatched brightness, creating a new world of light.
 
diffuse light sources

I suppose the key to using a diffuse light source would be obtaining a sheet of plastic formed as an array of thousands of PCV lenses. Each lense would "see" the entire light source and converge all the photons it gets from any angle, into a single beam perpendicular to the sheet. This would give you a set of parallel beams that you could send to a field fresnel, either above or below the LCD panel.

That way you could make the most of a diffuse light or even multiple light sources like LEDs or fluorescents. Or even a mixture of light types to give you a better color mix.

Has anybody heard of such a lens array material?
 
Has anybody heard of such a lens array material?
It's a nice thought but practically impossible to implement as you'd need to know the divergence of the type of LED you were using, and this is likely to vary slightly between each unit. It would therefore probably have to be custom made and I can't help but think that this reaches way beyond a DIY realm.

A few things to cheer LED lovers up though:
* New reflectors are expected to be introduced into LED's in 2005 whos efficiency is claimed to be 99%. Hopefully this will mean a wider variety of divergence angles as well (source).
* This along with much brightness progress where they are currently acheiving 75 lumens per watt in the lab (source) should give us something serious to play with (I think I'm correct in saying MH is between 70 and 90 lumens per watt). The latter could be a year or two away yet though...
* I still maintain that pulsing LED's produces a much greater perceived brightness than constant current, the danger here is that it could negatively affect the colour temperature of the LED's further towards blue.
 
lenses for rent for $0

Has anybody noticed that you can "borrow" various long focal length lenses for quick experiments by purchasing reading glasses from a local drug store, WalMart, etc. that lets you return things for a refund? The glasses are labelled 1.0 through about 3.0. These are not magnifications, but rather Diopters. (The same as eyeglass prescriptions.)

There is a very simple relationship: D = 1/focal length in meters

So here is a handy table for the standard reading glasses:

D focal length
1.00 = 1000 mm
1.25 = 800 mm
1.50 = 667 mm
1.75 = 571 mm
2.00 = 500 mm
2.25 = 444 mm
2.50 = 400 mm
2.75 = 364 mm
3.00 = 333 mm

When you buy them, make sure you ask if you can return them if they "don't work". (give you a headache, make your eyes hurt, etc.) They aren't very big, and they have a funny shape, but they should do for a quick check before you order a "real" lens.
 
Maybe we are taking the wrong approach here...

Maybe we are taking the wrong approach here with LED projectors in general. Everyone keeps talking about how we can make led projectors comprable with MH projectors, but right now it just doesnt look like they really can be, at least without spending far more time/effort/money than its worth. So why not use them for a projetor with a much smaller screen, that is a great deal closer to the source? Do we really NEED an image that is 10 feet wide? Doesnt the brightness increase exponentially as you get the source closer the the image or something? I would be pleanty satisfied with an image that is say the same size as a 70 inch big screen TV. The longer life, small profile, and ability to turn it off without having to worry about the bulb breaking would make it worth considering. I dont know about the colour problems though, if as some people have suggested, a white led wouldnt react the same way with an lcd, then perhaps it is merely a question of adjusting the lcd's colour settings, and accepting a hit on the over all brightness in order to compensate.

I digress though. The major dificulty for me with diy projectors, regaurdlesss of the light source, has been tracking down a reasonably priced lcd. Considering how much more noticable poor resolution is in a larger image, any way you slice it, you need an lcd that is going to do at the very least 1024x768, and if you want to justify not simply going out and picking up a commercialy produced projector, that lcd really cant cost any more than $200 - $300 at the most.

Im curious though, why dont people go with more of an opaque projector like set up, and just stick a run of the mill computer lcd/crt in it? Does the image look to washed out or something? I know people on this forrum seem to think that fresnels are the devil or something, but in combination with other lenses used to ajust the focal length, would they be so bad?

-Orochi
 
I think there's several approaches been taken by various people to LED backlights, personally I'd love a small projector using something like a PS LCD screen, however using a more standard 14-15" TFT is just as feasible (and the latter is easier to fit more LED's behind). However I'd still want the projected image size to scale up to similar to my current MH projector and, as you say, to get a sensible brightness using LED's this would require a lot of time/effort/money.

Having not really played with LED's behind an LCD enough I have no idea how the colour is affected however because white LED's tend to emit very narrow sets of wavelengths (which the eye interprets to be a broad spectrum white) it may be the case that we're talking about huge light losses, especially through the red and green filters. You are also right in saying that this can be evened out by adjusting the LCD's colour settings however the concern is that this will be a significant loss making it even harder to get a sensible brightness out of a projector.

As for your digression, I use the screen out of a standard Compaq TFT500 computer monitor in my MH setup, it is by no means the best monitor available but does allow 1024x768, with enough colours and a good rise and fall time. The problems with this are:
* that it's a 15" panel and as a result the actual projector box is massive (especially in my small living room).
* That it's a bit awkward to take apart the case of the monitor and refit the panel into a custom design. For this reason people sometimes go with a panel specially designed for projection, especially if it is to be sat on top of an OHP. If you want to use a computer TFT monitor then check out the good panel / bad panel thread to see which are suitable for using.

I'm not sure how you have the impression that people don't like fresnels, every design on here that I've ever seen uses them and IMHO they are paramount to a DIY design. The only problems people have had with them as far as I'm aware is when they have them to close to the focal length of the projection lens at which point their rings can become visible on the projected image. This is easily fixed by moving them up to 1/4" away from the fl.

Hope this clears up a few things,

Steve
 
alternatives to MH

The brightness of the screen goes down at the square of the magnification, just by spreading out the same amount of light over a wider area.

One of the reasons I am working on a 15" TFT LCD monitor projector, is that I want to experiment with different forms of lighting. (Also: $200 new from pcclub.com, 300:1 contrast ratio, 30 ms response, 1024 by 768) You can also get a 14" TFT LCD monitor for about $150, but they all seem to have a 50 ms response time.

These monitors are backlit by a single CCFL tube, and have about 10 layers of light spreaders and diffusers the light has to get through before it even reaches the LCD. So I suspect we don't need to use a blinding light source with so much LCD area. I have a 250W MH bulb coming in, but I also want to try a solid array of fluorescent tubes right behind the LCD with a field fresnel above the LCD. Another idea I want to try is a fresnel-free design using a reflector to focus light on the objective lens through the LCD.

I do not plan to try white LEDs, since they are less efficient than MH and have a horrible spectrum. People only think LEDs are ideal because a single LED doesn't generate much heat. But if you put enough white LEDs in your box to match the lumens of a MH bulb, it would generate twice as much heat! (And VERY little red or green in your image.)
 
Anyone got any circuit / part diagrams for these DIY projector units?

I'm just tinkering and hunting on the web for information on projectors. At present I work with AV equipment, but don't really have the additional cash flow to purchase expensive equipment for myself, so I make do with an InFocus projector I bought a few years back and a Hitachi I bought on eBay for next to nothing because the seller had NO idea of it's value.

Problem is that I'm sick of shelling out cash every 400 / 500hrs and would probably have been better off buying a 60" plasma by now!

I've read through a few (dipped more like) threads on DIY LCD projectors and frankly, I'm impressed. I read somewhere that white LEDs last 90k hrs. In the time it's taken for these threads to be created, the cost of purchasing a reasonably sized LCD TV has dropped by vast amounts here in the UK.

So, these threads appear to have either moved on, closed down or ground to a halt in Dec 04. Where did everyone manage to get with these projects? Anyone actually put together a decent prototype and create circuit diagrams?

Did anyone have any screenshots worth posting?

curious...
 
You are quite right, it has gone very quiet around here just lately!

By now I'd imagine there's enough information in these forums for anybody to be able to create their own DIY projector and I'm guessing a lot of people are following pre-made build instructions (e.g. the ones available on http://www.lumenlab.com/ ) rather than experimenting.

You'll find plenty of good pictures around the forum and I'm sure you'll see for yourself that people get stunning results, there's nothing quite like having a 70" screen in your living room!

Most designs use Metal Halide bulbs, I think they're approx. 20k hour lifetime although don't quote me on that one, they're also available for about £20 in the UK if you look in the right places. Last time I looked LED's were still much more expensive than this and less efficient.

In terms of you shelling out for replacement bulbs, you can almost certaintly build yourself a DIY projector for the cost of the next replacement bulb!

If you're contemplating building one then the best advice I can give is read read and read.
 
Sounds advice

No doubt good advice. I spent the morning contacting bulb suppliers in the UK, Europe and even in Australia to discover that a replacement bulb (700hrs) for the Hitachi is £300.00 plus P&P (inc 17.5% vat).

Considering I just picked up the Hitachi for £50.00 BIN on eBay (laughs out loud - cheaper than my last set of AV leads) I'd rather not shell out such a silly amount of cash. I knew it was going to be expensive, but prices aren't coming down fast. Don't the manufacturers realise that more people would go for these units instead of plasma if the bulbs cost 50% less and lower. They could do it, the cost of production is measured in £1's.

My InFocus projector has roughly 150hrs left on the bulb, which will not last long and it's started to dim already.

How practical are re-conditioned bulbs? Has anyone on these forums tried? The bulb from the Hitachi looks to me to have died through shock (over heating because of poor venting probably) rather than over use.

As you can see I am new to this side of things as frankly I like to park my ar*e and view films, not spend weeks tweaking. That said, I really want to have a good set-up going.

Cheers.

P.S. I'll look those schematics up as I have built equipment in the past (mostly studio) and wouldn't mind trying my hand at it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.