I’ve been playing around with lens shifting and I seem to have confirmed the theory that the arc size adversely effects the image focus. I’ve got a 150W 6mm arc MH and a 400W 25mm arc MH. When the 25mm arc is used the outer edges of the image are very hard, if not impossible, to focus but when I switch to the 6mm arc every thing becomes crisp. It could be a result of the entire arc image not entering the triplet but I can’t see how that would effect the image focus. So everything points towards the actual size of the arc image causing the crappy results. Ie the 6mm arc uses less area of the triplet lens than the 25mm arc so less aberration. Any thoughts?

DJ

Wow this new interface is cool...

Dazzla... ever experiment with an iris for the arc?

Not on the actual arc but I have on the triplet.

edit I tried stopping down the triplet when I used the 25mm arc but it had little effect.

edit2 the new editors cool but have a go at the quick edit function.

Well... your thoughts? Any issues there? Do you think that there would be a significant brightness loss?

If you mean stopping down the lamp then yes I think it would effect brightness too much. Just to qualify what I’m doing, the results I’m seeing are from pushing my lenses to there limits I don’t think there would be the same thing observed when used in the normal way.
Dam I was only going to post this thread and go to bed.

DJ

hmmmmm, after having done the enhancements (LCD)

this is exactly what i was working on,,

the glowies sent me there,,, but along the way worked out why glowies were happening, hmmmm

used a commercial lamp,,, wow what a difference,,

crisp is right.. I was salvaging an iris from an old piece of equipment,, durable metal for heat reasons,,,

just for that exact reason

what i think is happening are multi collimated sources, because of edge ray effects

After getting out of shower, thinking clearer.

It would seem to me that and iris would only work well if it was very close to the arc. Otherwise the outer edges of the fresnel will see less arc than the center.


[edit] Aussies need sleep... sorry to keep you up. I gotta go to work.

So let's see if I'm gettig the gist of this.

Small arc, sharper image. Larger arc, blurred image.

An iris at the lens does not help.

This would seem to discount your theory about the amount of the triplet used, since if there's an iris at the triplet, the amount used is restricted that way.

I would expect that this is due to less ideally collimated light from the larger source.

Question: did you experiment with unsplit fresnels? With split fresnels, this would be more noticable, as the field fresnel would magnify the changes in collimation after the image. Before the image, it would matter less to the lens.

Please could you make a drawing and how large is your LCD?

Thanks Dave

I was reading a physics book at the Library a few weeks ago and came across a section dealing with point light sources. It was explaining that you could shift a point source(considering the lamp remains in one spot)by simply using an apature or iris. It even had a diagram explaining it. The diagram made it look as though there would be no loss of light, or should i say the iris would be as bright as the original source. It DID NOT mention the arc length of the bulbe though.

Point light sources are different animals, and are one of the things that appear in theeoretical physics, but don't really exist in real life. You can do many things with a point light source that you can't do with a real arc lamp.

Correct me if I'm wrong but according to my calcs the 150W light source should be brighter than the 400 watt lightsource in your projection due to the size of the arclength. A 6mm arclength distributing 150watts has a power density of about 25watts/mm. A 25mm arclength distributing 400 watts has a power density of 16 watts/mm. If the fresnal only focuses 1 mm in length then the 150 watt bulb should be about 50% brighter than the 400 watt bulb.

Was that your perception, also?

Edit: What is the brand name of your 150W MH bulb?

No, the 150W/6mm lamp should not be brighter.

The fresnel does not limit what light it accepts, it merely bends or focuses the light that's there. Any light that can be bent/focused into the triplet gets projected and adds to the brightness. Light that cannot won't be. Light that originates from a point that is too far from the lens' principal axis to allow the lens to bend it into the triplet doesn't make it. Anything that is close enough does.

Given the magnification factor, and diameter of the standard triplet, 25mm is pushing it, but still possible.

So what really matters, still is lumens.

Osram's MH literature indicates that the majority of the light comes from the two ends of the arc (I have heard that its two ends for AC driven bults and one end for DC bulbs). Depending the the triplet used there may be a substantial loss from even moderate arc lengths (S400DD is 27mm some of the venture bulbs are as high as 44mm). The question is how much of the ends is making it through 1) standard triplet 2) PRO triplet. I would quess that everything makes it through the PRO but a substantial portion will not make it through the Standard, probably not enough to make the 150W brighter on a standard triplet but maybe enough to make the compention close. With the PRO the higher wattage would win out. My previous posts came to the conclusion that we really wanted to keep arcs below 20mm to get best performance through a standard triplet.

Where does the light come from on an S400DD lamp?


#823 - S400DD projected - from Parabolic Tests

Great thread. The precision of the light source angle is what makes the image. A brighter bigger source will wash it out.

Try this experiment: Put on a white T shirt, put a scenic picture on your monitor - full contrast - medium brightness, grab an objective lens, Turn out the lights. From the far end of the room to the screen project onto your tshirt. walk closer to the monitor and focus again. There will be a point where the light is getting brighter but the image is fading. when you are close you cant make an image.

When you are far away you are only getting a very narrow angle of light (narrow cones) the image is sharp but not much light comes from the screen to the lens. As you get closer the lens collects more light but its in fatter cones. The image is not as sharp. Eventually it washes out all together.

Id love to have a 6mm 150w!

Try the experiment with the TV, its easier. Block out light from the sides.

I’ve had a bit more time to think about what is happening, still not 100% though. I’ll try to post more tomorrow, have to go to bed.
The 6mm arc MH is a GE Arcsteam, I got it very cheep $10 only good for testing as its colour and lumens are off but it has got a reasonable out put compare to the 400w, maybe due to its arc size?

DJ

Dazzla,

visually, are both outputs comparable. Is it possible to perform a luxmeter test to make certain?

Supra: I didn’t actually use an iris I simply truncated parts of the triplet but I see what you’re saying though

Phutton: I can’t say for sure as I don’t have a meter handy. The lamp is only 4000K so it is a little yellow this makes the comparison hard as well. What I meant is that considering it is only 150W its image is reasonable but I would still use a 400w.

Me2: you’re on the right track, I’ll explain shortly

Pagercam : I tend to agree with what osram said. The arc image appears to have more light at the ends.


Here’s what I think is happening. A good quality lens will be able to resolve an image of the LCD using most of its surface area. A poor quality lens will only be able to resolve the LCD image using the centre of its surfaces. Because of lens shifting and stretching the fresnel FL the majority of the light of the arc image is forming at the top of the triplet. This part of the lens is where most of the aberrations are. When I switch to the 6mm arc the amount of the outer edge of the lens being used is less so the aberrations are less. To put it another way, using the outer edge of the triplet is like using a crappy triplet. The large arc image means that more of the crappy part of the triplet is required to resolve the LCD, the 6mm arc uses less surface area so less aberration.

Another observation is that by moving the arc I can change the focus on the screen. Some times the centre is in focus then a slight movement of the arc and it becomes blurry. This indicates that different parts of the lens surface are not equal. By that I mean not all of the lens should be used if possible. So a smaller arc will use less of the surface so a better image will result. I hope that makes sense.

What makes this annoying is that I originally tested lens shifting using the 6mm arc and a 17” LCD at 16:9 and 50 % lens shift and it worked beautifully. So I set up this test using the 25mm arc and it failed. !@#$$%. I really need to find a better triplet. Or a fresnel with a focal of 400mm, this would reduce the arc image so less lens area would be used. (Sigh) .

DJ

so according to your thinking, the new pro triplet ,being larger, should be much better at lens shifting since the same arc image will be more thowards the center of the lens(or does the new triplet also increase the arc image size so no gain there?)

Dazz,

What fresnels and triplet are you using now?

I did a quick calc a while back and I don’t think there will be any gain as the magnification will be greater on the pro fresnel. But it all depends on the quality of the triplet.





220mm 330mm 3dlens.com the triplet is a 3M 342mm FL

I did do a split set-up a fair while back but I liked the fresnels behind the LCD.

DJ

Standard
330/220= 1.5 magnification
1.5* 25mm=37.5mm arc image
80mm/37.5mm=2.1

Pro
600/220=2.7 magnification
2.7*25mm=68.2mm arc size
100mm/68.2mm=1.4

Going by these quick and nasty calcs the standard is better but it all depends on the pro triplet. Just thought of another test for those who have one of the defective pro triplets. Try using a smaller arc and see if the blurry corners improve.

DJ

Couple of corrections the standard triplet has an outside dimension of 80 mm I think the clear glass portion is closer to 65mm, the PRO fresnel is 650mm according to Brain, so both numbers will grow a bit, conclusion still stands thou.

same coment about standar triplet diameter and pro fresnell focal as above.

and new coment, . I don´t know why did you choose effective diameter/arc sice ratio.... If you had chosen arc size/effective diameter you would have the % of the lens used by the arc.

% standar (64mm diam) = 37.5/64 = 58% used.
% pro ========= 73.8/100= 73% used.

I can say from personal experience that the arc image on my focus mechanism is much larger than 37.5mm in size, probably due to the scattering of light going through the LCD.

Makes a lot of sense that the outer edge is poorer quality. Olny aspherics would not have this problem.

I'm not sure about the second part? Are you saying that you can get a smaller arc into a smaller point and therfore it is in the better part of the lens? Move the lens so the quaility point is at the edge. Is it still lousy? probably.

If the problem with a large arc is only that it uses the edges put a cover around the back of the lens to block light from those edges. The remaining image should be darker but good acording to that thinking. Is that what happens? Im guessing no. This is a second issue.

With a large arc the light through a point on the lcd opens into a larger cone. The lens tries to recombine light cones into a focused point on the screen. If the cone diameter is beyond the lenses ability you cannot focus. Or more precisely the center of the bell curve of that cone focuses but the edges are just defused light. This is what the spot tests in the lens design software try to mimmic.

Using a double fresnel setup cuts the side angle in half too.

Could you explain what you mean by that a bit better?

Pagercam, I realised that the triplet was smaller just after I posted it. Then my computer crashed and I was in sleep mode so I went to bed. I did say it was a quick and nasty calc though. Same answer to you rox I was tied but I only wanted to show the difference, didn’t matter if it was a ratio or percentage. Some how though I new Rox was going to pic this.

Me2 the problem that I’m faced with is due to lens shifting but the theory should be the same for normal use just not at the extremes I’m seeing. The most intense part of the arc image is at the top of the triplet with both 6mm and 25mm. Just as an example spherical aberration is where the outer rays are bent to focus closer to the lens than the centre if you could limit the light to just the outer edge then you could still form an image. Or another simplistic way to look at it is, the outer edge forms one image and the centre forms another when the two are forming images at the same time the image is poor. Block ether of them and you should have a good image, they won’t be equal in quality of coarse. This is what I believe is happening in my set-up, by using a small arc I’m confining the light to a small section of the lens so the differences are minimised its just unfortunate that this spot is at the edge of the lens.

I’ve got another test I can try to see if the theory is true. By using the 6mm arc I should be able to get a bit more lens shift.

DJ

Just finished the test. I was able to get greater than 50% shift! But the entire image was not visible at the same time because not all of the light was entering the triplet. What I did was adjust the triplet so that the centre of the image was focused. The corners were blurred, as you would expect from such an extreme angle, but by moving the lamp to various positions I was able to get sharp focus at any part of the screen image. At no time did I adjust the triplet to get it focused again This suggest that the image quality is a product of where the light hits the triplet. I’m think that it might be viable if I get at fresnel more closely matched to the triplet or a smaller arc that is 400W then the triplet would play less of a role in the final image quality.

DJ

edge of field rays,, plus divergent collimated light effect....caused by arc size ??

dazz read what I have done other thread,,,heheheheheh

To Paladin: The Frankenfresnel test.
The link is here: http://www.lumenlab.com/forums/index.php?showtopic=8219

Basically you use two rear Fresnels and two lamps angle the Fresnels to cover the seem line It works good but not perfect. It does cut the angle from the farthest point of the Fresnel by half this might sharpen up the focus some or if two smaller and shorter arc lamp were used then the focus could be even better

There is definitely something to be said for using an iris to increase image quality. Every camera lens has a specific aperture setting that is optimal for that lens. I just wanted to make sure we realize this is not just a theory, it is common knowledge in photography. Unfortunately, while film and shutter speed makes up for the change, the projection screen will not.

The optimum resolution out of a lens is thought to be around 2 to 2.5 stops down from full aperture as a rough guideline.

This apparently due to the reasons you have mentioned: aberrations in the lens. Stopping down samples a higher quality portion of the lens. I would also think it would cut down on some reflections off the inner walls of the lens mount.

But if you stop down beyond the optimal aperture, the resolution gets worse. This is apparently due to diffraction at the edges of too small an opening.

I know it is a given (and has probably already been said), but a smaller arc will provide the optimal light path for the panel as well: normal incidence.

Hope that helps (and isn't just stating what is known ),
Mark.

Wouldn't this increase your arc image size, if you are using a 330mm fresnel and 342mm FL triplet?

My first PJ has a severe pincushioning and focus problem in the corners - using an OHP varifocal lens that was designed for a chamfered 11" square "stage glass". Obviously since a 15" LCD is 12" wide, it might explain the imaging problem.

But...I wonder if moving the lens further away from the field fresnel and then refocusing using the varifocal, will improve my image quality for the better?

Or am I moving in the wrong direction, or a horse chasing its tail?...

I think, haven’t re-read my post , I wanted a 400mm FL fresnel to alleviate the distortion the short fresnel was causing.

DJ

I read some where in these forums that most, if not all, OHP varifocals do not have a large enough field of vision for a 15 inch lcd. I wonder if measuring the opening on your varifocal and comparing it to the dimensions of the standard triplet in the LL store would tell us anything. Also, I had an OHP with a varifocal. The focus ring faced the glass and fresnels. If the varifocal was turned around so the focus ring was facing the screen, the image became unevenly lit and distorted. I don't know if all varifocals display this characteristic.

thanks for the reply S.T. - I believe it is smaller than the LL triplet. That's why I gave up trying to improve it - a year ago...gave it to my brother's kids for Xmas - they don't mind the flaws ...I may have even tried moving the lens further away. Can't remember the results...

almost forgot...I discovered I could eliminate the pincushioning by distorting the field fresnel - applying pressure with my finger.

This would take a complex adjustable mounting system, but may be worth it if I overhaul the PJ. Or I could just get a larger varifocal. I think they are available at the Euro DIY site.

Moving my lens further away might also help...

I presume this was a split design.

DJ

boy, I tell ya ... I didn't get it at first, but after some practical observation this afternoon, I understand the arc issue better.

I took a look at the light hitting my triplet after adjusting my reflector and my arc to be more perfectly centered, and now I can tell that I have some alignment problems. But when I put a piece of white paper against my triplet, I could see for myself how big the arc length was, and how it was overlapping the edges at the top and bottom of the triplet.

LOL So now I understand that my bulb has an arc bigger than the width of my thumb, and its color temp is 4000k, not 6500k, and it's as big as a football. That Powerball is looking pretty good right now. hehe

Yes