This might be of interest,
A patent design for projectors to use up to four lamps rays together to provide a single light source using direct view 15 inch LCD displays like we are using.

The design talks about using 1mm arc commercial lamps (expensive!) however Elken is currently testing a DIY 1mm arc lamp.

The Pdf also talks about a/g removal and methods of increasing contrast.


http://www.freepatentsonline.com/6924849.html

Once your on that page.. view the detailed patent design and see the design drawings
by clicking on:

"View PDF Images" (just below "Publication Date:") it will open the pdf within the browser, you then can save a copy to your machine for viewing later.


Never really delved into looking at patents, Me2's talk of patenting his ideas got me looking.
Found another patent that talks about using LED's as an auxillary light source together with the main light source to compensate for the lack of RED in our lamps and WHITE LED's so that the Projector can be watched instantly when power is applied !

Sounds very interesting. Unfortunately I can't read the PDF file. Downloads at 5K. Hmm....

PLJack

Sorry ...it appears that you cant link directly to the PDF, apologies i should have checked first.

I will edit the first thread...so now just goto the html link

http://www.freepatentsonline.com/6924849.html then on that page click on the link that says

"View PDF Images" (just below "Publication Date:") it will open the pdf within the browser, you then can save a copy to your machine for viewing later.

Cheers

There is still no PDF to save at your corrected link.

That's strange, the pdf link worked fine for me.

On the HTML page, the pdf link is there, and works. Wait, let me check something...........
Well, that's interesting....... I just checked and the pdf link works in IE but doesn't work in Mozilla Firefox.

Tgreenwood

Seems that the website is a tad flaky..for me it works in both Firefox and after a couple of refreshes in IE (after clearing browser cache.)


I tried uploading the pdf but error says something about not having the right permissions..?!


SIM, PLJack and anyone having trouble getting this to work PM ur email and I will send it ..no worries

Cheers

Why dont you guys try USPTO website. I alreadly checked the patent its pretty cool.

6924849

Johnzo,

Thanks for the alternate link...for all if you use Johnzo's link you will have to install a TIFF
plugin to your browser to view the detail and the design drawings. I used:

http://www.alternatiff.com

To view the the orgiinal patent (includes all the drawings) click "Images"

Here's the patent zipped up.
Click to view attachment

Thanks Paladin !

No problem, mods have a higher upload limit.

Croma,

That is an outstanding find!! Aside from the multiple lamp design, which appears to be difficult to implement, there are other things
mentioned I've never seen anywhere else. Such as purposely sending light through the LCD panel at a small angle to maximize
contrast and use of a polarization rotator to improve corner brightness.

I am pinning this thread.

What a PITA PDF! Good grief.
For those of you having as hard a time as I did, I've attached an image of the juicy bit.
Looks very interesting.

Patent No: US6,924,849 B1


BTW, never drag a PDF into Photoshop. The conversion nearly took down my PC.

That PDF makes for excellent reading,, and it has confirmed several points that I have suspected for a long time..

Uhg... my head hurts from reading that but ugly PDF. Anyway... here are some interesting things I pulled away from it. For those of you who don't want to read, these are my cliff's notes =P

The first part of the paper discusses using multiple projectors for a bigger display, and how the projectors can be constructed to make them easily "stackable". Not much of an application for us... we're sticking with 1 projector.

Figure 10 shows an LCD sandwiched between two fresnels "having centers that are offset to direct light rays through the LCD panel at an optimal ray angle for achieving a peak contrast ratio from the LCD panel". The angel shown in the picture is 5% off of normal. I think this could be achieved with a very slight split design lens shift. I've never heard that LCDs look better having light comming into them at a slight angle. Latter in the paper they say that the exact angle needed is particular to the LCD construction, and that it is determined by emperical measurement.

Has anyone ever observed this? If this were true this would mean that unsplit designs would have better contrast near the edges. I suppose it is possible this is true... it just hasn't been noticed because our lighting is rarely even. One thing they don't mention is the distortion created by a split design from the fresnels or how they plan to get around that. They also talk about why having small light sources close together is better... again to control the angle that the light hits the LCD to improve contrast.

The method explained for removing layers from the LCD sounds much more envasive than what is currently being done at LL. In addition to what LL members have removed, they remove a "Wide Viewing Angle" film (which I haven't heard of) that they say reduces contrast in order to increase viewing angle. They take pretty much every layer apart. From what I gather, the ONLY thing used from the original LCD is the TFT and liquid crystal layer, and the polarizing layers (although they say they are removed... and unless they can do this without damaging they'd have to use new ones). A "Wide band rotator" layer is added... more on that in a sec. Also... they mount the LCD "backwards" so that the "black mask" in the liquid crystal layer blocks "unintentional photoconducting". I assume they mean it keeps unwanted light from passing through =P

Apparently a "Wide band rotator" is designed to turn the horizontally polarized light from the light source, to diagonally polarized light for the first polarizer, then after it passes through the 2nd polarizer it rotates it back to be horizontally polarized again. Apparently most projectors diagonally polarize their light. This appears to mostly be to help their multi-screen technology. The distortions they say it can cause appears to be most visible when trying to match the edges of multiple diagonally polarized projectors. I haven't heard of anyone complaing about diagonal polarization here... so I don't think we need to worry about "Wide band rotators"

The light source is suggested to be 4 120W Philips UHP100 bulbs. These are made for rear projection TVs I think... they're expensive. However, the paper states that any lamps having 1mm to 7mm arc gaps are suitable. Through the use of parabolic reflectors and a "pinwheel mirror" the paper claims that the light is very collimated. (+-3 degrees). The pinwheel mirror is 4 mirrors, each with 1 corner touching the rest, and resting at a 45 degree angle. This can effectively place the 4 light source's focal points very close to each other. The goal is NOT to have each light light a different portion of the screen, but rather for all 4 light sources to light the entire screen. Another benifit to the pinwheel is that the "light bundle" is now in a generally rectangular shape now.

After this the light is shot through two "lenslet arrays". I wasn't familiar with this term but I found a good paper on it:

http://www.zemax.com/kb/articles/91/1/Fly&...rojector-Optics

Note this paper says a 11x9 array is optimal.

If we can find a cheap source of lenslet arrays... (which I wouldn't think hard) we should certainly be able to greatly improve our "ROX".

I’ve seen this mentioned before but on another forum. You can see this affect if you look at a LCD monitor from below the normal view height. The screen gets a lot dimmer at the top. I think this was a big problem with allot of older LCDs.



The Wide Viewing Angle film they mention is I believe Anti-Glare.
There has been some testing done by DIYers for the best direction for light to travel though the LCD. From what I’ve read and from my own experience you would be hard pressed to see any difference. Photoconduction refers to the transistors used in the LCD produce electrical current when light falls onto them so the mask is there to shield them from light. I remember on another forum someone talked to the designer of the SGI LCD and designer said that it wouldn’t make much difference which way the light travelled. Again I haven’t observed any ill effects, although it could be possible.



I think they have actually removed both polarisers and placed new ones, horizontally and vertically orientated, into the rotators for fine adjustments. This they say helps with multiple stacked projectors.


DJ

Guys I know that PDF is quite detailed but Ive found some more of potentially of use patent designs:

Instant-on projector
http://www.freepatentsonline.com/6988806.html


Single-path color video projection systems employing reflective liquid crystal display devices
http://www.freepatentsonline.com/6839095.htm


Achieving color balance in image projection systems by injecting compensating light in a controlled amount
http://www.freepatentsonline.com/6698892.html


Illumination systems employing corrective optics for use in reduced
etendue color video projection systems
http://www.freepatentsonline.com/7055967.html


Cheers

Well... I think this is related to the "Wide Viewing Angle" film now that I think of it. If it remains on the panel, you wouldn't see this effect be nearly as pronounced. What you are talking about sounds exactly like what the Wide Viewing Angle film is suppose to prevent. This must be how manufactuers have improved LCD viewing angle.



No I don't think this is the case. They claim the WVA film is in the Liquid Crystal layer. They claim the Anti-glare film is on the polarizer. They speak of the two as completely seperate.



Ya I hadn't heard this either... maybe this was a problem with the LCDs this inventor was using?



I agree this is probably not something of much interest for us (unless someone wants to use multiple projectors). My understanding however was that LCDs normally diagonally polarize light, each side in the opposite diagonal direction. The WBRs were designed to produce horizontally polarized light. However I did re-read what they said and they did mention having to replace one of the polarizers. I don't think it's polarizing direction is any different though... obviously this would not work with the original polarizer if they did. They say:

"Light propagating through substrates 96[TFT] and 98[Liquid Crystal Layer] is typically diagonally polarized. WBR 118 receives horizontally polarized light from the light source 70 and delivers through polarizer 120 diagonally polarized light to substrate 98. Likewise, WBR 114 receives diagonally polarized light from polarizer 102 and delivers horizontally polarized light to substrate 116. "

Anyway great find on the patent. It wouldn't surprise me if some here had already thought of some of these ideas, but seeing it in a patent really gives the idea weight as something viable. Way to go =P

Alright. I donloaded and read that PDF. It IS butt ugly. Fascinating read though. Im thinking the angle of the light might have something to do with the transistor alignment. I think I heard elken speak about this before in difference in the twist paterns etc. I suppose its possible that certain layout patterns might allow more light through at certain angles, but you have to wonder if it damages the clarity in any way.
The light engine is interesting. They say it is designed to cover the whole LCD from each light source, but it looks like it was adressing 1/4 of the panel. Has anyone done a raytrace for this?

I cant tlook at that drawing without thinking of a VW air cooled engine.

No this isn't the case. The reason for using small arc lamps with the pinwheel reflector is to effectively place these light sources so close together, relative to the huge size of the LCD screen, that they act like a single source. The diffusing lenslet arrays further evens out the light from the different sources.

I know thats not the case but if you look at the design, it SEEMS ( Im stressing the word SEEMS here...)to look like it would cover a 1/4 of the panel. Im just wondering about the accuracy of the drawings. I was kinda hoping someone would raytrace this one to see how it behaves.

Any takers?

Never mind I did a really rough simlation and it actually works out pretty nice. I was surprised at the spread across the panel. VERY interesting. Makes the design array I was working on look quite a bit less efficient.

Mirror 172 seems too short to capture all light coming down from output lenslet 170. And the diameter of lenses 174 and 176 seems
too small to capture all light from mirror 172 if it were larger. I think the lens diameters should match the size of lenslet 170.

I’m with SJ on the ¼ screen illumination. I drew up a rough raytrace in Raytrace last night but Raytrace doesn’t have a fisheye lens so I simulated one. It was way too tedious to see exactly what was going on so I read up a bit more about integrators. From what I could gather the fisheye lens adjusts for the arc length so that the ray angles from either ends of the arc seem to be coming from the centre of the arc.
It seems to me that it is basically four collimated beams reflected from four individual mirrors angled so that each beam illuminates ¼ of the total. The pair of fisheye lenses (integrator) then smooths the individual beams as well as smooths the transition between them to form one seamless total. The advantage of using four individual beams as opposed to just one is that the light evenness that a fisheye lens adds to a one beam set up can be further enhanced by splitting it into four.
Some of the ideas in the pdf seam a bit vague. One thing I noticed is that the integrator doesn’t seem to have all three components, two fisheye lenses and another plain lens, so maybe there is something I’m missing.

DJ

edit:
Hang on. Maybe the third component of the integrator is the ashperical lens 174. Now I’ll have to draw another trace to see if it overlays all four beams.

Seems like the actual drawing and the theory dont quite mesh together. His description definitely says they are 4 sections of collimated light moving into a "lenslet array" ( Im thinking thats a mild fresnel?)
So Im not sure how thats supposed to cover the whole LCD. But he is talking about a rectangle of light thats only 5 inches across. Page 10, second column, line 15.

Still its got my noodle goin...

For those of you confused about the lenslet array like I was when I first read the patent, I highly suggest the zemax article I posted. Page 4 of the article in particular. It looks like a fantastic way to distribute the light.

Those still puzzled by how the 4 light sources can light the whole screen, I would think of it in this manner. Lets say I have 2 flashlights. I tape them together and then shine them on a book right in front of the two flashlights. There are two circles right? Now lets say I shine it on a brick wall 30 feet away. Now it is more or less 1 circle right? The same principle is what the patent is trying to achieve by using very small arc lamps, placed very close to each other with the use of mirrors. Since the arc lengths are 1mm, and the LCD is many times that size, the light sources behave more or less like one by the time they hit the LCD.

My understanding is that the collimating lens and relay lens are in place to keep the light path narrow by refocusing the light. I don't think we need this. I think we could achieve the same evenness results by having a parabolic reflector on a light, then 2 lenslets, then our rear fresnel. Basically, the lenslets are replacing the job of the precondensor lens most people have been using.

Im also having trouble understanding the original pdf...it does say that up to 4 lamps can be used. So it seems that the patent is like a jigsaw which can use different pieces.

I read sling_blades last post but Im still not sure if a lenslet light ends up covering all the LCD Panel or do we have 4 lenslets 1 for each quarter...would a raytrace give an answer ?

BTW Sling_Blade Nice link on flyseye lens !, I wonder how much the lens arrays are

Thanks. I've been trying to find some cheap ones and haven't been successful. So far I've seen them for 80-180. The patent is a bit confusing because it is talking about using multiple projectors stacked in a "jigsaw" like method as you mention to make a bigger screen. That is why they talked about the difficulty of getting edges to match. Then they talk about just 1 of the projectors that could be used in that "stack" of projectors. Which could have up to 4 bulbs.

Ive just been looking for flyseye lens but have had zilch luck... unusual for google

So what is the answer as to how it illuminates one panel ?

Hi guys! First post here, and might I say I enjoy the DIY attitudes here.

Some things I see from this diagram:

1) Inefficiency? It looks like the reflector is arranged to provide parallel light toward the first mirror(s). If this is so, then (at least from the diagram) there is a size misfit between the edges of the reflector and the pathway to the first mirror. This could be an artifact of the poor drawing, however, and easily overcome in adjustment and selection.

2) The off-angled position of the lamps might lead to some better cooling options, which would be necessary with the higher power lamps and the multiples of them.

3) I agree that the front side mirror placement is probably only representative, rather than accurate.

4) If the fly eye lens in fact evenly mixes the light from 1 - 4 lamps brought together by the pinwheel mirror, then why is the mirror necessary at all? Light from the right side lamp cannot illuminate the left side of the LCD panel prior in the absence of the fly eye (at least from the drawing as presented. It would seem 4 lamps in close proximity at a distance from the fly eye would have the same effect without all the complications.

5) Taken to a logical extreme, imagine a torus of light surrounding an inverted cone pointed at the LCD panel. That would be some serious lumens.

weldonjb

Welcome to the club !!

Well I havnt found the answer exactly how this patent design illuminates the whole panel but I have found the company who submitted it is still alive.

Interestingly their main focus is LCD/DLP Rear Projection Displays for business:

http://www.clarityvisual.com

From looking at the company it seems safe to say that these guys know what they are talking about.
I wonder if they are using this patent or parts of it in any of their current model designs !?

After re reading this patent I realised I had missed what this patent really was ... The patent apppears designed for a small depth rear projection 67" display.

For the simple people like me:
It is a folded path design. Clarity Visual systems who's patent this is are into BIG displays for industry/commerce hence the reference in the design for a frame to stack these 67" displays !! to any number.

Further simple explanation:

The light source (engine) 70 (fig 12) projects light---> the LCD Panel Assembly 80 (fig 6)
then onto a fold mirror thru projection lens 86 onto a large fold mirror before it hits the rear projection screen (67")

I had missed the fact that the light engine (fig6) 70 has a condenser lens 174 and aspheric lens 176 which I assume spread the rays from the integrator (flyseye lens) to fit the LCD panel

The patent link below is referenced in this patent it shows the folded light path for the rear projection display used by clarity visual systems...not sure if it transposes exactly as per this projector design but would be similar.

http://www.freepatentsonline.com/6273570.html

I tried to explain the way I understand it with my flashlight example. I'm not sure I understand why you don't think it would work as they says it does. Nothing we use is a single point light source. For the LL projectors, even though we are using one light, we are using a light produced by a "line" of light comming from the fillament. You could say that this line is like having a bunch of point light sources along that line. Since this setup obviously works fine for the projectors we have built so far, I don't think it is much of a stretch to see how using a bunch of smaller arc length bulbs grouped together with a mirror would behave all that different.

You're absolutely right about the "stacking" not meaning stacking the projectors but the entire rear projection unit. Now that I look at it that way the drawings make a lot more sense =P

This has my brain working hard, I tell ya. One difference I see is the blending of the light. For a single filament bulb, while the light is not evenly spread throughout the sphere centered on the middle of the filament (length of the filament, shape of the bulb, thickness of the glass, the mogul, etc.) in the cone that reaches the panel, there is only a VERY slight more light along the axis of the filament.

For the patent arrangement, the center along the axis of the point of the pinwheel mirror seems like it would either be significantly darker or lighter, depending on the placement of the mirror. But, you can bet I am going to experiment with some cheaper bulbs and maybe a penlight laser.

I wonder where I could scrounge up a 45 degree chrome cone ... hmmm.

I found a possible source for the lenslets:

http://www.fresneltech.com/pdf/FresnelLenses.pdf

I'm going to send them an email and see what they have to say.

Holy Moly that is technical. Going to have to put my physicist hat back on to work through that.

Physics/Journalism Baylor 1983 Go Bears!

Sling_Blade

Good find !.....looking forward to hearing how much the lenslets cost.
I tried searching again but still failed to find anything other than manufacturers of lenslets.

Perhaps they could provide a lens array that would fill the LCD panel without the need for relay or aspheric lenses.

I m tossing up wether to contact the patent holders Clarity, as they obviously have them..perhaps they sell spares ?!

Maybe we could use 4 x 70 watt small arc lamps in a similar configuration to the patent design ?

What kind of mirrors are they using for the pinwheel mirror setup that could stand the heat of the lamps ?

One question I have is the aspherical and relay lens in the light engine, what could be/would be the metrics of the lenses to give the correct spread across the 15" panel..from the the last lens array.

I guess we are talking math here..not my strong point.

Paladin maybe you have the answer to the lens aspheric and relay lenses size questions ...!?

Where can you purchase the aspheric/relay lens from and how much would they cost..and do we need both of them....and...... no that s enuff for now !

Perhaps if you had a 5" inch LCD panel u might not need these lenses......the homogenised rectangular rays from the lenslet's would fill the LCD panel.... ?

putting two point five cents worth here,,

to a degree,, the KISS principle needs not to be forgotten,,, for all the builds and experiments I have done, it always reverts back to basics,, which with the tools and items available give us the best results,,

no poo pooing it,, just how much money do we throw at a PJ, before it becomes uneconomical...

with what we are doing one lamp is the way,,, far less complication,,,
introducing extras multiplies our issues by an order of four or more

I completely agree with working at the lamp end,, in all of this,, but,, one reflector,, one lamp,, one pre lens,, and or other enhancements, then fressie,, thats where the answer lies...

my tests with a 1mm arc prove that the fressies are not upto the challange,,,

I have found this is a fixed factor, born by going down to a 10mm arc already,,, I had to change my fressie setup, to get rid of onions heheheh shrek slips in.....

Elken,

I agree KISS is best.. Im not spending money ...yet

talk is cheap !

Perhaps though a lenslet is not to expensive to try with one lamp..then maybe 2 lamps later if it proves worthy.

sorry i was putting that up before uniformed go running off on a tangent

I have read some of the info, you have triggered by the thread,,, and have some ideas myself,,, but arhhhhhhhhhhhhhhhhhhhhhhhhhhh

mind bending stuff... having brain meltdown.... too many tests to do.... and i will be back to 1mm arc,,, as she was a sight to behold standing at screen looking back at trippy,,, a quantum singularity she was mmmmm

elken
btw .. what u are doing at the moment is huge...small arc, small wattage great results...the idea of using the small arc has been confirmed as the way to go from what is found in the patent design...how about a lenslet array for the 1mm..

I agree,, but the fressies limit it,,

ohhhh oops remiss of me,,, Thanx,,, gotta push the envelope,,,

Geez I cant keep.....I started to read the fresneltech's pdf but only just got to the page that has prices for the arrays...lol

but which one's ..............

Ya the prices for the arrays are pretty cheap... about $55. I'm with elken on the multiple lamp idea. I'm sticking with 1 lamp for now. The aspherical/relay lens are NOT necessary as far as I can tell. They're purpose is not discussed in any length, and I conclude that they are simply a way of refocusing the light back to a more narrow beam so that their rear projection box can maintain a small size.

If I were to go with the multiple lamp idea, I'd start out with 2 lamps first and see if that worked out well. The same idea would apply to two lamps as it would to 4, and it would be cheaper.

Unless I'm totally mistaken, the lenslet array can replace the precondensor many people are using. I'm just going for better distributed light for the moment... not more light.

Sling

$55 not as dear as could be, even more so if your idea of not needing the aspheric and relay lens is correct...are you thinking of using one lenslet or 2 ?

I had thought in the patented design the aspherical lens and relay lens are to spread the rays from 5" rectangle of light coming from the the last lens array to cover the whole 15" inch panel (via the fresnel) however my mind is open to be convinced otherwise....anyone care to add ?

Or to say it in another way can you describe how the spread of light would cover the panel size if coming from the lenslet array in your idea ?

Cheers


A bit more about Lenslets/Flyseye/Integrator lens/Lens Arrays:

An animation of how adding Integrator Lens Technology magically makes your projected image
uniformly bright !

http://www.3lcd.com/eg/pup_4_e.html

Heres some working drawings of a commercial projector : "The Integrator Lens"
http://www.3lcd.com/structure_close_up.html

Perhaps if we think through the problem in reverse? I you had collimated light coming down a chamber, how would you break it out into 2 equal parts of light condensed into small points with minimal attenuation? If you can do that, maybe you could think of how to reverse it back to a light engine?

Condensing lense, beam splitter? Polarizer? Alien artifact/technology?

could the lenslet in the way it works..possibly help the issues with the 1mm and the fres..