Thought this would be a good topic to discuss, I know lots of people are considering ceiling or floor mounting their projectors, and are worried about keystoning and correction of it.


Here's some info I found on calculating the amount of keystoning:

Combating Keystoning
The percentage change in image width (or height, for horizontal distortions) necessary to prevent keystoning can be calculated as follows:



where
angle A is the full beam angle (not the half angle as is sometimes quoted!) of the lantern or projector
and
angle B is the deviation of the projector from horizontal (-ve if pointing up, +ve if pointing down)



This equation can also be used to calculate horizontal distortions. Note that in any case the projector or lantern is assumed to be on the centreline of the projection surface in the direction which is not being distorted.

You can read about the derivation of this equation here.

In deriving this, it has been assumed that the distance between the projector and the screen at the centre of the beam is the same as the distance between the projector and the screen at the edge of the beam.
For small beam angles (upto about 25°) this assumption will not present a problem.
However for wide angle projectors, the equation will not compensate adequately for spherical distortions.

... or in english, this isn't the full answer, but at least it'll get us thinking

*This information was taken from http://business.virgin.net/tom.baldwin/keystoning.html

I am definitely going to have to adjust for keystone as I will be ceiling mounting. I am going to try to mount as low as possible without being unobtrusive. I really hope I don't have to calculations as it has been a long time since any math class
Has anyone built a projector with keystone correction? Should I be able to get the very same results as a straight aligned projector if I set up the keystone correction correctly?
Here is what I gather from the design guide:
I will clip my condensor fresnel to the glass panel. I mount the LCD 10mm away from the condenser fresnel. The collector fresnel is mounted 1/2" away from the LCD. I am guessing that although the collector fresnel is tilted it should still be on center. I am thinking that a center pivot point for the tilting fresnel with a tightening/locking mechanism would work well as I might occasionally move the projector. Does this sound like a good plan? So basically the the only measurement changing is moving the collector fresnel out a half inch, therefore moving my triplet forward 1/2".
I really would prefer to get as many things built as possible before the lenses arrive.

It's a good plan, but leave yourself some room for experimentation. The collector fresnel is now in front of your LCD panel. You may need more room than 1/2" to allow enough tilt to compensate for the keystone. Also keep in mind that when either fresnel gets too close to the panel, lines will show up in your projection. This includes the tilted part of the fresnel.

Man, I think I need to go re take math. Brainchild how about making us a form to fill out like you did with the FL and lens. Think you can do it? That would be soooooooo helpful.

How good is the keystone correction on the 1515 or 1512 LCD monitor? Does it have any?

Thanks for your help.

PS any news on the group sale we talked about?

Could I just tilt the lcd insted of the fresnel? then both fresnels would remain behind the lcd. Mike

How about tilting the screen to match the angle of the projetor?

You can't tilt the LCD to correct for keystone. All that would happen if you tried is that part of your LCD would be out of focus. The LCD needs to remain flat because the projection lens is focused on it.

Brainchild,
So can you or some one make a form to figure keystoning that we can type numbers into and get a figure out?

Lamps, I haven't done a lot of research along these lines but I'll look into it. I'll try to get back to you tonight on the order thing, very busy, please forgive.

Brainchild,
Sounds good, yes I'm sure you have a lot going on with the update and other stuff. Let me know.

I've written a program that accepts the two angles and outputs the percentage width change. Its DOS-based and should be used in conjunction with the schematic above. Is anyone interested?

twilite: send me the source code, and I'll rewrite it in php... then it can be put up on the website

Wouldn't it just be easier to manualy find the angle by adjusting the lens in the enclosure by repositioning it? I am guessing you will have to do exactly that anyway... just as you have to with the other lenses to find the sweet spot.
I wanted to share some experiences with you guys.
I put together an experimental projector using unkown focal length used scrathced optics yesterday. My first experience at all with a projector. It was not all that difficult to find the correct focus etc, but small things made a huge difference in the projected image. Following the guide "to a T" even with correct lenses I would imagine you will still need to tweak. I intiially couldn't get the 400W MH bulb's gas pocket to line up center with the optics. I had the outer 1/3rd portion lined up thinking that was OK. Not at all OK! It resulted in a broad light rainbow effect over the screen. I don't know if my bulb is slighlty bigger than the guides, but I could not align the center mounting just as in the guide. I had to tweak it by having it pointed at a 40 degree angle towards the fresnel so it is aligned with the gas pocket the long way at an angle (maybe it even captures a little more light that way?). That made all the difference. Even with the low qulity optics I was able to get a somewhat uniform white square.
Be carefull with your LCD! I tried to bend the ribbon to LCD connectors just a little and lost 1/6th of the screen! Then i walked across the floor and touched the to of it and ZAP, I lost video completely! Hopefully I can exchange it!

Jo, can you tell me which LCD you used? Thanks.

He used the Circuit City liquid video (1512). These panels are fragile guys, be very careful with them. Also it's best not to strip them until you're ready to use it. This will minimize the chance of anything happening to your panel in the interim. Grounding yourself is also very important, static discharge will waste a panel.

Yeah, as with any electronics and/or computer work, it's a good idea to use a 'grounding strip' I've been fortunate to never have a problem without grounding myself, but I also am sure to always touch something metal (like the screw on a wall plate) before I grab up the piece of electronics I'm working with..

Most any computer store or radio shack should have a grounding strip, otherwise a simple DIY way to make one would be a stripped wire. Basically you wrap a stripped wire around your wrist, like a bracelet, and attach the other end to something grounded in the house, again the screw on a wall plate would work good for this. or if you have one, put an aligator clip on that other loose end, and clip it onto something, a lamp, computer case, etc.. anything metal and grounded, just make sure you actually get a decent connection. don't count on static discharge being able to jump through a painted surface. It's possible, but not probable.

Here's one at CompUSA: Grounding Strip @ CompUSA

BTW: Thanks for the code twilite.. I'm going to look into this equation a bit more and see if I can't work out something that will spit out how much the front fresnel needs to adjusted to correct a keystoned image, when you enter in a few variables about the projected image and relative location of the projector.

I think that would be a bit more handy than just figuring out how much keystoning there is.

I'll keep everyone up to date on my progress on this, but I intend to work on it tonight at school.

BTW: Rorschach brings up an interesting point.. you could tilt the screen (projection surface, not the LCD) a bit to compensate, however I think this would only really be a likely option in perhaps two cases.. if it was floor mounted, this probably would work fine, since undoubtedly the screen would be much taller than you anyway, so pulling the top side out a bit to match the angle of the projector with the screen could work.. OR, if it were floor or ceiling mounted with only minor keystoning.. I'm fairly sure that you don't want to have to stand up to watch anything with your projector

It just occurred to me on my way to school that this may very well be pointless to calculate out.

Seems to me the answer to this is simple.

When the projector is at a 90 degree angle to the screen (straight on centered vertically and horizontally) the front fresnel would be at a 90 degree angle.

Suppose you ceiling mount it. Now this is way too extreme of an angle, but just for simplicity sake, assume it's at a 45 degree angle. Obviously the keystoning would be insanely bad.

What seems to make sense in my head is that the front fresnel must be parallel to the screen. So in the case of it being at a 45 degree angle, the fresnel would also be at a 45 degree angle with the top edge leaning toward to LCD.

this is obviously much too extreme of an angle and probably wouldn't work at all. However, the basic theory should work. So, assume that the projector was at a 15 degree angle vertically, and off center a bit to the right (as you look at the screen)

to correct the vertical is easy, same as before, just adjust the fresnel so it's vertically parallel to the screen. But, uh oh! now it's keystoned horizontally too, what now? same idea for the horizontal.. the result would be that the closest corner would be the top right (looking from the lights point of view) then the top left, then lower right, with the lower left being the furthest.

sounds weird, but in simpler terms, as long as the fresnel is parallel to the screen (Projection screen, not LCD) keystoning should be history..

You will probably have to play with the distance from the panel to the fresnel, but the basic principle should apply. I haven't tested this, but it makes sense to me logically. But then, I'm finding more and more often that logic very well may just be a crock people use to justify a 'best guess' answer

Just what I was thinking Oog. Very good points Brainchild. Leave that LCD alone until you are ready for it.

I went with what a guy had suggested to Brainchild on diyaudio forums to use the frame/enclosure for the LCD. With the 1512 it was easy to just make a cut with a dremel just below the face of white plastic frame. The metal frame snaps to the white plastic frame and whalla, I nice solid LCD with protected ribbon cables that you can be cofident hanling and easily mount. If I would have followed the guides' measurements I would have been able to fit the enclosed LCD in there perfectly. Oh well, I will just make a little groove in one side of the box for it.
Will post pics of this with some test shots using my crappy optics tonight!

Hey Jo,

I would love to see some pics of how the frame stays on the LCD panel. Would be very helpful.

OK, what I meant was tilting the screen the image is projected on to, not the lcd panel (screen?)

Maybe we need an online terminology dictionary? That could help avoid any misunderstandings?

Sure you could tilt the screen, would make for a rather odd viewing angle though. Of course, if you're rather odd this works out.

That would be me!

If it was only a few degrees it should't be too noticeable, might be ok if you had stadium seating to go along with it though.

wouldnt it be possible to make the focusing frenel adjustable wth in the enclosure for keystoning. maybe have the focusing half of the encolsur on a movable mount of some sort. Any one get me??

hey oog, it sounds correct that the front fresnel needs to be parrallel to the screen. So maybe it would be posible to put that fresnel in a small frame, pivot it from the exact centre, and put an exactly aligned weight on the bottom. that way the lense would always stay vertical. I guess you would need some small stoppers to stop the lense moving further than what produces a useable result. anybody follow me, lol.

Yes that's how auto keystone correction works.

Dougy, yeah, that sounds good.. probably fairly easy too.. once you ascertain the limits to how far it can rotate, just pop a small peg in there..

Yeah thats what i was talking about!!

This forum is awesome!

I was dwelling with the keystoning issue and try to find a way to make it work without having to readjust it all the times, and then I thought that my swivel frame needed some sort of control to move the fresnel... first I thought about a motorized solution using the mecanism (modified) from a broken Computer CD tray, then I thought to let mother nature land me a hand (gravity) and use a weight (Hey dougy same wavelenght! )...that should work best to auto keystone the projector everytime is moved from one place to another, but I think that in an advanced projector design (future?) the modified CD tray idea may still be valid to fine tune keystoning moving the fresnel exactly to the position that gives best results.

Oooohhh! Nice idea, having a 'hanging' fresnel so gravity automatically adjusts if for you.

I am curious, when the fresnel is at an angle to the lcd, does it affect the focus of the projected image? Or cause any distortion?

Also brainchild mentioned auto-keystoning, in what I assume was the manufactured projectors. I would like to learn more about how my InFocus X1 manages to project a square image while positioned at the bottom of the screen, and while at a 90 angle to the screen.

NICE thinking,I like the theory,but I think it needs some testing.

Definitely needs testing, that's why we're here

Tony, the focus should not be affected, may affect slightly the contrast because of the limits of the panels viewing angle, make sure you get at least 35 vertical viewing angle (the more the better for keystoning). The image will of course be distorted that's what keystoning is about, it'll compensate the difference between position of the projector being placed off center (ceiling or floor mounts) correcting the angle of projection so that the image will not result be showing the bottom corners being closer thatn the the ones at the top (or viceversa in case of floor mount) and looking as a trapezioid shape. If you are talking about other types of distortions then, no that should not occur. Again keystoning needs to be tested more thoroughly once we all have out projector put together

I know most CRT's have trapezoid adjustment, but does anyone know of an LCD monitor with this feature? To me that would be the best solution for countering the keystone effect.

I searched for quite a while to see if I could find one that did have trapezoid adjustment, and still haven't seen any.

Same here - They may not exist as lcd's by their very nature aren't prone to the magnetic interference that would necessitate such an adjustment.

How is keystoning accounted for in a commercial projector? Optically - or does the smallish lcd panel have trapezoid adjustments?

Optically, weighted lens.

Peculiar thing...
When I split my used crappy fresnel lenses last week and started playing with placement I noticed I got a much better picture than with them combined. When the fresnels were combined the bottom of my image had problems i was not able to solve no matter what tweaking or repositioning I did. The center bottom was the worst with it lightening and getting better going out in a half circle pattern . The condition was like looking at negative film with black's, but color wasn't really affected. Skin tones would look richer and the image overall darker in that area. generally an unbalanced image that anoyed the heck out of me. I tweaked and tweaked to no avail. If I took the wall with the Triplet lens out and looked at the LCD I saw the same thing as projected on my screen, the effect everyone knows all to well if you look at an LCD at a bad angle. I was guessing that it must be my crappy lenses, and maybe that has something to do with it...
Finally, I decided it was time to give the keystone correction and fresnel splitting a shot so I could get the projector mounted on the ceiling. After splitting the fresnels I got a brighter and perfectly brightness balanced image.
My hypothesis:
The dual fresnel sends light out at an angle from it. You get the same distortion when looking at an LCD from a similar angle. Splitting the fresnel, placing one on either side of the LCD send light through straight.
I really did try adjusting everything to correct this before I split the fresnels.

Has anyone experienced anything like this? Does anyone know what I am talking about?

When do the good lenses arrive?

Hey Joe, The lenses have been shipped from China and should arrive early next week at which time we will ship immediately.

Yes you are experiencing a contrast variation due to the angle of the light leaving the collector fresnel lens. Some panels show this and others don't. I've only recently become aware of this problem and I have a solution for us. I will be ordering custom manufactured lenses next go round that have a higher groove count and that are actually larger than our panels. This higher quality lens will be suitable for use in front of our panels with a minimum influence on the projection. Since this is a custom manufactured product the turnaround is longer and the price is higher, but still reasonable (<$40 for both lenses). Normally these lenses would cost $200. This is a great advance for us and is an option since we have established a trading relationship with the supplier.

Thanks Brainchild,

Advancements left and right.

Having found a better bulb and lens after I had already purchased is just fine for me as I have several interested in this projector so I will just seel or give this other stuff to them. I am sure others in my situation would have no problem figuring out a solution to get rid of unwanted stuff.

Any chance you could permenantly post I buy/sell/trade section for projector parts. I personally want to sell the Kworld box I tried out and maybe some glass.

Speaking of advancements. I found when ordering my glass that there are several types with different levels of clarity. There were about 5 types at the place I went to. From the standard with blue/green edge to totally clear (clear edge). As you might guess, the higher the clarity, the higher the price. Depends on how far you want to go with this I guess. I actually used 1/8th inch non-tempered glass just fine for a couple weeks of testing. My heat actually reached near 120F for a moment. Ever heard of non-tempered breaking or cracking on anyone? This would save $$ as tempered throws a $25 extra (as opposed to $3) anywhere around me.

If your non-tempered glass were to break, and your fresnels and monitor melt, would the $25 seem like a savings then?

Yes I tried a piece of standard glass and it broke. It isn't the heat, it's the temperature difference from the incoming air. My untempered glass lasted about 2 weeks.

Also I started the 'trading post' forum a few days back, sell away

http://www.lumenlab.com/forums/index.php?showforum=10

Same pass as the other forums.

I don't suppose a mirror ( or series of them) AFTER the projection lens would be able to perform keystoning would it (they)?


Second thought path. What happens if you tilt the projection lens? Lose focus on part of the image? Barrelling? Etc. etc. etc.?

Tilting the mirror would cause part of the image to be out of focus. So would tilting the projection lens.

Anybody considered adjustable keystoning? Is the keystoning so delicately tuned that you couldn't just put the lens on a pivot and attach an external knob so that you can just twist it around until it looks right? How significant a keystoning effect does a slight nudge of the lens tilt have on the output image?

That sounds swell. How would you attach the knob to the fresnel lense? Are there knob/pivot kits out there that would be applicable?

Another question about seperating the fresnels and placing one in front of the LCD to adjust for keystoning - will there be a problem with that fresnel warping as it is not clipped to a rigid piece of glass? Or would you place another piece of glass there to mount it to - to prevent it from warping?

I have keystoning set with a permenantly mounted fresnel on a peice of glass. Seems to work pretty well. It doesn't seem to be that touchy as in having perfect center alignment. I don't have a screen built yet to cover my entire image, but results look promising. It seems hard to get the entire image pefectly in focus, however very slight. The larger the angle the worse the results. I have my second fresnel mounted at a 15 degree angle with my box actually at a 13-14 degree. Looks pretty good. Strangely, left center and right center of the fresnel circles gets projected in the image with the circle direction opposite of what you would expect... maybe I am getting a chunk of fresnel #1 backwards in my image. Either way it is very slight. It is no more noticeable than the scrrendoor effect from my liquidvision 1512, in other words, don't worry about it. As i am still using old crappy fresnels too small for my LCD I am definitely going to wait until I have the quality fresnels before I make any final conclusions.

I am no artist. But I have tried to draw up a little diagram, where you mount the fresnel in a slotted wooden frame (see A). I assume the fresnel will need to be pivoted from the center? The previous post seems to confirm that.

Anyway, use a wooden dowel rod to create a pivot axis for the frame (blue arrow), connecting it to the box/case.

Use a threaded bolt for the adjustment point (red arrow). Cut a slot large enough in the side of the box/case to allow adequate adjustment fore and aft. Use a very large washer to prevent light leakage. Use a wingnut to clamp 'er into position.

I doubt you will need an adjuster on each side, but you will naturally need the pivot point on both sides.

Great idea, Brainchild please add this to the v3 guide.

If it works. But I am quite flattered you like my idea, Lamps. (hey, they didn't provide an 'embarassed' icon, it's the best I could do on short notice.)

Of course all this 'frame' is going to make the insides of the enclosure need to be enlarged.

Actually, I think my project will be 'modular.' Every section will be framed for easy tweaking and protection. One module I definitely will drop in and out is the heat shield module. And the reflector module.

Also, I am quite nervous about a naked LCD, so I want it framed for protection.