Maybe a crazy idea....

[tgreenwood]

A fellow holography enthusiast (he's not into projection like me) has been using an lcd panel to make multi-channel hologram portraits with a red laser pointer.

To get to my point quickly, he mentioned that the lcd transmits a lot more than just 10 or 12 percent of the laser light. He estimates that about 55 percent of the laser light gets through the lcd.

The theory is that the diode lasers used in laser pointers are linearly polarized, and if you line up the polarization of the laser pointer with the polarization of the lcd (they are polarized, you know) tons more light gets through than ordinarily would.

I know that red laser pointers and green laser pointers are available, does anyone know if blue laser diode pointers are available yet?

Wouldn't it be incredible to have a DIY laser projector?

I've heard somewhere about blue laser diodes being used, but I can't remember where. Anybody know?


T.

 

[Megaman]

I'm not sure about the polarization of the laser light, but the fact that the lcd area are divided into parts with red, green and blue color filters will result in no more than 1/3 of the light getting through (in average that is, assuming that the light point is much larger than the individual (sub)pixels).

Anyway, that is not the main issue here.
There certainly exists green and blue lasers but the green ones cost a bit and the blue ones cost a smaller fortune.
I was actually looking around for lasers last week in thought of building a projector using lasers, some optics and some electronics (no lcd). Here is a link to a guy who have done it and can project a small text, if you are interested. http://cscott.net/Projects/FabClass/3/

The only blue laser I found on ebay costed around $2000.

The main problem with lasers however is its light output. The standard laserpointers sends out only a few milliwatts at most. It might be fine for a small point to be visible, but spreading that light out to a larger area like a projector image it will be wery faint.

As I see it, there is no realistic way to use lasers for projectors within the next few years.

 

[tgreenwood]

Oooohhhh, shot down immediately.

Oh well, I'm going to play around with the idea anyway starting with LEDs (cheap) and work my way up to lasers.

T.

 

[Me2!]

The laser is a great idea. It would be very efficient use of the light generated. You still need to pump enough light to make a real projection as megaman stated. That takes a lot of energy and time. You would have intermittent laser blasts while it is pumping.

Powerful lasers would be more expensive that a commercial projector.

 

[Me2!]

I was thinking about this again and im wondering if 1 wavelength of RGB each would combine to white? Do you need a broad spectrum of each colour? If so lasers wouldnt work. If not, what shade of green etc.?

My reason for asking isnt just curiosity, im thinking that pulsed lasers would really be no different than the filtering colour wheel of DLP. In both cases you get bursts of each colour. Lasers may stay out of the home budget but considering the expense of cinema lamps would they be interested in this laser approach?

PS, im a lousy 2 finger typer, but colour is the correct English spelling!

 

[Guy Grotke]

human color perception

The color "sensors" in the human eye are molecules that absorb light over a wide band. You can look at a white screen lit by any color light that stimulates only the green sensor, and you will perceive it as pure green. If you put another color that stimulates only the blue sensor, then you will perceive it as pure blue. It doesn't matter how wide the frequency band of each of those light sources are.

But our sensors all have wide bell-curve responses distributed around the "red", "green", and "blue" lines on the spectrum. So many colors will stimulate more than one of the sensors. For example, a cyan color will stimulate both the blue and green sensors equally. That "color" can consist of a narrow band of light at the point on the spectrum where the blue and green curves overlap, or it can consist of a "pure blue" and a "pure green". Looks just the same to us!

Likewise, "white" just means that all three color sensors are equally stimulated. Could be from R + G + B, or it could be from a single line of blue plus a single line of yellow (red + green sensors). Or it could be a single line of red plus a single line of cyan (green + blue sensors). A good example of this is most of the high-intensity white LEDs: They are really blue LEDs with a yellow phosphor.

On pulsing: The sensor molecules average their response over time, so a display that gets updated over 24 times per second looks just the same as a display with the same average amount of light per pixel. I don't think the duty cycle matters much. Just the average. Otherwise movies and laser light shows would not work.

 

[tgreenwood]

A lesson... never give up

Well, nuts, I should have hung in there with the laser idea.

Check this out... Tiny laser projector .

These guys are freaking incredible.

I can see a monochrome projection as being practical, because it worked well with the virtual keyboard, http://www.pocketpccanada.com/pages/Reviews/Virtual Keyboard/virtualkeyboard.asp , but achieving full color is going to be hard, just from alignment problems. My guess it that they'll use a prism system like the one used in commercial 3-LCD projectors to combine the three color images.

And no moving parts!

P.S. A lot of the laser pointers available are CW (continuous wave), no pulsing.

 

[cbm5]

I've been thinking about an idea along similar lines. The LCD can only pass certain wavelengths, so why not give it only those wavelengths to start with?

There's two ways to do this. One, is to have some kind of mixed-gas halide lamp that has a spectrum with big spikes in the red, green, and blue areas. I don't know if this kind of lamp exists, or is possible.

Another way involves three monochromatic sources, some kind of powerful neon bulbs maybe, or single-color halides if they exist. The lamps would be lined up with condensor lenses in between, so the focus of the previous lamp passes through the arc of the next one. This would provide a virtual three-color source from the final lamp. A single reflector at the end of the lamp array would work for the light passing back through the condensers. I'd guess you could get away with individual lamps of 50 or less watts, and still end up brighter than a white-light projector because most of the light would actually be used instead of filtered.

 

[Ben_Tech]

How about these lamps?

Check the 'technical bulletins', but they appear to come in 250, 400 and 1000 watt versions and the color temp for the 250 and 400 watt lamps seems to be 6500K. What caught my eye is:

The exclusive Iwasaki chemistry in the Hortilux - Blue provides ratios of red, blue and green spectral energy that.......

There may be nothing to it, but worth a look. It would be difficult to fit optics around a large globe-shaped lamp.

 

[Guy Grotke]

more efficient projectors

Even if you had a lamp that only produced narrow red, green, and blue peaks, it would not help: When you sent that light through a color LCD, each individual sub-pixel color filter would absorb the other two colors. So you would lose 2/3 of your light right there. Same problem with combining different color lamps.

You CAN benefit from having individual red, green, and blue lamps by sending the light from each one through a simple black and white LCD. (Or you could just use a white lamp with a dichroic beam splitter to seperate the R, G, and B light.) No filters means no filtering loss. The downside is that you have to recombine the three color images. Simple way to do that is to use three identical lenses to recombine them on the screen (like CRT projectors). You also need some custom circuitry to drive each LCD with its own color's luminance, but that is pretty easy with component or VGA input.

But try finding a small black and white LCD with specs as good as a dirt-cheap 15" LCD color monitor! And don't forget, you will need three of them! I think maybe the cheapest way to get all of this stuff is by buying a used LCD projector on eBay.

 

[tgreenwood]

I love lasers

I have to admit that I'm really excited about using lasers for projection.

I've done some research on the tiny laser projector mentioned above and I think I've figured out pretty much how they are doing it.

The full color one will use a conventional imaging system consisting of 3 microdisplay LCDs, one for red, one for blue and one for green, with a image combiner prism for the output, just like commercial 3-lcd projectors made now.

The really cool part is how the lasers are going to be used. Since diode laser light is linearly polarized, there will be little or no light loss through the polarization layer of the LCD. Regular lamps lose at least 50% of the light right there.

The next neat part is the use of diffraction optics. Instead of losing light at the lcd by light being blocked by the parts between the pixels, the laser light will be diffracted using computer generated pattern producing diffractive optical elements (DOEs).
These have been around for a while to generate patterns with laser light. I just ordered one from Thorlabs for $12.

The whole point of all these big words is that the optics used with the lasers can produce an 1024x768 array of laser light dots at whatever spacing you want, without losing overall brightness. The light isn't being blocked to produce the dots, it's being "steered". The light will go directly through the pixel holes with no loss.

Follow me so far?

Now for the "light engine". I think they'll use three diode lasers, one red, one green (both readily available) and one blue or blue-green. The only way I know of currently getting a blue/blue-green diode laser right now is from one of those blu-ray style dvd players/burners. Too expensive at the moment for a parts scrounger like me, but someone is bound to get a screwed-up aiming or tracking mechanism in one eventually, and if the laser still lights up it's MINE.

Another cool thing is that the projector will have infinite focus, a perfectly focused image at any size/distance without lenses. I imagine that they'll use a "zoom style" lens on it simply to adjust the size of the image with no need to refocus.

What really puts the icing on the cake is that the whole thing can probably be run with 8 AA batteries (or less).

I am soooo tempted to look for a cheap old LCD projector on Ebay (most of them use the three lcd microdisplays) and at least play with the red and green LCDs, but since I am currently unemployed there is no chance that I'll get to indulge myself.

Did I mention that I love lasers?

 

[cbm5]

I think the main issue with homebuilt laser projectors so far has been projected intensity. I've seen a few designs that used a modulated laser pointer and a faceted spinning-mirror scanning arrangement. The physics of the matter dictate that if you beam a laser at a wall, it will make a dot of a given brightness.

If you somehow split the beam into two dots, discounting factors like optical loss, the maximum brightness of each dot is 1/2 the original dot. For a 1024x768 display, we are dividing the laser into 786,432 dots. Again, not counting losses from optics and offtime while scanning the laser over areas where a display is not desired.

So we are dividing the laser dot brightness by at least 786,432 times.

Another, more convincing way to look at it: a typical laser diode is about 3 to 5 mW and some are even less. Lasers are rated by output power, and there are about 680 lumens per watt. I think you're beginning to see where I'm going with this. 0.005 x 680 is 3.4 lumens output, and that doesn't account for the reduced visibility of monochromatic light, which would take the apparent brightness down at least a third.

In order to begin to approach the brightness of the low-end commercial projectors, you're looking for lasers of 2 to 5 Watts each. Only recently have diode lasers this powerful been available, and they aren't cheap. And try finding blue.

 

[superdaveumo]

That's very true if you split the beam. But, if you take a beam with a width equivilant to the width of the pixel (the sum of the RGB components) and move the beam to each pixel, not split but move, then you would get the full brightness of the laser; not one that is split 768,000 times. This is what the mirrored experiment that has been posted did. By angling the mirrors in a different way, each mirror moved the laser dot to a different location.

Now, his experiment was for 32 x 8 I think. It's a whole different story to do even 640x480. If you did it the way he did it, you'd need 480 mirrors and a lot of patience to align them.

I think a more efficient way to approach this is to find something that will bend the laser beam like a CRT bends an electron steam.

Ooo! Just thought of another idea.. If you had a circular piece of glass that spins. The laser beam is shot through the glass. As the glass spins, it is shaped in a way that bends the laser beam down in a linear fashion. This would eliminate the need for 480 or however many mirrors. You'd just have to keep track of where the one laser beam is.

 

[cbm5]

You're getting dangerously close to pseudoscience here, certainly the eye is somewhat of a peak detector and a pair of switched dots would look a little brighter than a pair of beamsplit dots. But that effect will hardly matter with a 1:768,000 duty cycle. No matter where the laser is at full brightness, it isn't at 767,999 other locations.

Just look at the facts of it: you're trying to spread the light of a laser diode over a surface to make an apparent large display. The total output power of the laser is 3 lumens, compared to 1500 lumens for a typical projector.

The idea is like trying to cook a turkey by "scanning" a pocket blowtorch all over the turkey instead of putting it in an oven. The local temperature of the blowtorch may be much higher than the oven's temperature, but the blowtorch doesn't have enough total output to cook the turkey.

As I said before, it's not an impossible project as long as you can find red, green, and blue lasers in the multiwatt range, instead of milliwatt laser pointers.

 

[superdaveumo]

I wouldn't call it pseudoscience...... You'd need a laser that when shown through an LCD, the resulting pixel was as bright as a normal bulb. The power of the laser is still questionable. I haven't read of anyone trying this so we can't be sure how much light actually gets through.

The only requirement of the laser is that it needs to achieve full brightness at 1/768000th of a second, but that's for only 1 frame per second. Of course if you had a beam that was wider, you could cut down on the amount of scans you'd need to do.

 

[cbm5]

The concept of pulse width modulation is lost on you people.

 

[pinkmouse]

About 15 years ago I was at a product launch for Ford Motors. They used lasers to project full colour artwork over the top of video. To get brightness comparable to the Barco 5000 they were using for projection, they were using three 10W dye tunable lasers, each with a 32Amp three phase supply.

 

[superdaveumo]

PWM is a great idea! Love PWM. Works really well. It would work perfectly for control of brightness levels on a laser.

I think the Ford laser show is a little (lot) more than what we are trying to do here. Think about it. If you take a laser pointer and point it at the wall you see a dot. If you move the pointer to another place on the wall, you see the same dot at the same brightness just in another location. Picture this movement REALLY fast and in a grid-like manner.

 

[pinkmouse]

Sorry, but it's exactly the same. The 3D graphics had to be as bright as the video image, and were driven from scanning motors, exactly as you would in a video projector. If you're going to all the trouble of using a laser, you might as well scan it like video and modulate the beam, rather than passing it though a lossy LCD screen.

 

[cbm5]

PWM is a great idea! Love PWM. Works really well. It would work perfectly for control of brightness levels on a laser.

I think the Ford laser show is a little (lot) more than what we are trying to do here. Think about it. If you take a laser pointer and point it at the wall you see a dot. If you move the pointer to another place on the wall, you see the same dot at the same brightness just in another location. Picture this movement REALLY fast and in a grid-like manner.

No, you don't understand...is what I was trying to say. It doesn't matter if you do it REALLY fast, the duty cycle is what matters. Strobing the laser across 768,000 individual points is the same as a 1/768,000 PWM duty cycle. The brightness of the laser will be the same wherever it's pointed, but the OTHER places are dead black at the same time.

Try this experiment: get in your car with a brave passenger carrying a stopwatch. On a long stretch of road, floor the accelerator for one second out of every ten. See how fast you go. Then try the same experiment, only floor the accelerator five seconds out of every ten. Then try ten seconds solid. Notice that the power of the impulse is always at maximum, but at 1:10 you go much slower than at 10:10.

If your laser idea worked, then you could also reach the same speed in the same amount of time, regardless of whether the accelerator were floored for one or ten seconds.