(Split from this topic. DAZZZLA)

Having the reflector will cerate a second image on the screen. You might be able to see it if you look close. This effect will be less with a reflected image but should still soften the focus some which you might be OK with, it will soften the screen door as well.

The link is for a steady cam which I use to have, there fun.

Lumens are just 9 screens reading averaged and divided by screen size.

Your numbers sound about right for this setup.

Is there a condenser lens?

LCD size? Rear Fresnel FL?


For a room with that much light the PJ is doing good.

Are you saying that having the reflector offset a little will in effect create two arcs and thus create a double image?

Tgreenwood

No, there will be no double image, nor soft focus (at least not so much, that would be visible), nor anything other than better brightness. This is not the way the projection works. Imagine a camera (which is basically a reversed projector), does having two lights in the room make double image on the photo of this room? It's not exactly the same case, but you get the idea. This is the reason we are using projection lens, if we could have absolutely perfect point light source, this lens wouldn't be necessary.

arizona: there is a condensor lens, but cracked, I rotated it until I found a position, that makes those cracks nearly invisible, even on a plain white screen, LCD is 15", 220mm rear fresnel

regards

A double arc creating a double image is what arizona was seeming to say and I was getting clarification before challenging him on it.

It seems to be a common misconception that causes people to think that they have to have a spherical reflector perfectly aligned to get the reflected light going back through the arc for "point source light".

I think a corner cube reflector with a small arc lamp mounted close to the apex would make almost all of the light usable while still keeping it easy to cool.

Tgreenwood

I also think, that "perfect" positioning the arc in the centre of spherical reflector is not optimal, because the arc tube is not very clear, not really transparent, and the reflected light will get partially blocked and dispersed. This is easy to prove with a small 12V halogen globe - moving it at the front of the spherical mirror you will notice, that in the perfect location this bulb will cast shadow. Light on a white sheet of paper, this very exact location is easy to recognize - light will suddenly get dimmer.

Further to the idea of "double picture". In my opinion this is not possible, there is no way to get double LCD image on the screen just by adding another light source (whether real, or a reflection of a single light). Imagine every single pixel of LCD is a light source. Ideally there is only a single, perpendicular light ray, going through the pixel to the projection lens. What happens when you put another light source behind the LCD, close to the first source: there will be another ray, going through that pixel at slightly different angle, and it will hit the projection lens at different angle, but also will be bent and redirected at different angle, and finally will reach the same spot on the screen as the first ray. This is the way the lens work, thats why you get sharp photos of real objects, although everything is lit by enormous amount of light rays, coming from all angles.

And here comes another idea of possibly useful reflector: something on the shape of a pyramid, made of flat mirrors, so looking from the rear fresnel we would see the real arc, and 4 (or more) reflected images of this light. If it all fits within the field of view of collector fresnel and into the projection lens, I guess it could be quite efficient. But I don't have time to try, someone might want to test this idea.

Regards

A halogen lamp is a different animal. It has a filament that will block the reflected light. A HID lamp has no such obstructions so it can pass back threw. The only exception is if the arc vessel is not clear in which case offsetting will help. In fact this is how a lot of old film projectors were set up.

DJ

I’ve tossed that idea around allot. It all boils down to finding a small arc with a physically small outside size that can be placed very close to the apex of the mirrors.

DJ

whats the arc length of the OSRAM 250w w/d bulb?

I don't remember now, I have measured before placing the lamp in my projector, something like 20 or 22mm maybe?

Dazzla: no, a HID lamp is not clear and transparent, you can check yourself, take a welding glasses, turn the lamp on and look at the arc. Put it on a white background and hold something behind the arc, like a piece of metal. Do you see clearly that thing through the arc tube? No, you will see some fuzzy outline of this object. That's why a spherical reflector, when placed according to the theory, is maybe 10% efficient or so (far less than expected), because reflected light is obscured by the arc tube (four layers of curved glass, and some cloudy content inside working lamp). My lamp is mounted on a flexible piece of metal, so I can bend it a little in any direction, and I can obviously see when I get the best screen brightness - when I can see two arc images in the projection lens, and both of these images fit into the lens. That should also work with a standard 400W bulb, but bigger lens (that one you call here "pro" lens). Doesn't do much with a 400W bulb and standard OHP lens, because it's too small, a single arc image hardly fits this lens.

Regards

Interesting. So the actual lamp is placed below (or above) the true centerline and the reflector adjusted so as to create an image just above (or below) that of the actual lamp ? I wish now that I had included an easy height adjustment to my lamp...

I still wonder how far off center you can make this without diminishing returns. Moving the arc away from the fresnels focal point can really make a dim image in a hurry. My other concern is that the image created by the reflector may miss the arc tube, but likely not the outter glass envelope of the lamp as this would really require a large offset.

You wouldn't have any comparable lux measurements between the two setups ?

gs

I don't have any comparative pictures or measurements.

I'm not sure what you mean "above" and "below". My pj is vertical, and I moved the lamp about 1cm maybe 2cm to the side, off the center axis, not moving the reflector, which stays more or less centered. When you move the lamp to any side (not along the axis), its reflection will apear in the reflector, it's easy to see, take the reflector and any small object and check it just in hands.

I guess the rear fresnel has to be a little closer to the lamp, so the whole light setup (both the lamp and its reflection) is within the view cone of this fresnel lens. I haven't measured this distance in my projector, anyway, those cheap plastic lens are not made up to high accuracy, and their focal length vary from piece to piece.

Regards

i'm horiztonal so my above/below is your side/side. at any rate that answers my question... 1-2cm movement off center, with reflector on-center... i'll have to give this a try and make some measurements, but i'll have to get my 250W lamp working first as the 400W arc is just too big for the LL Pro triplet (ends of the arc just fit inside the triplet at the moment)

cheers,
gs

You must have a very dirty lamp. Maybe from overdriving?
Click to view attachment
Click to view attachment
DJ

Edit:
If you mean that you can’t see something behind the arc when it is operating, of coarse you won’t. The arc is the source of the light, the object behind would have to be brighter than the arc for it to appear with any detail.

You have some unusualy clear lamps, perharps new or nearly new ones, even so during operation some stuff evaporates inside the arc chamber and makes lamp more opaque and diffusive.

Check this out, this is one of my 400W lamps, not the one that I was overpowering, it was used for maybe 50 up to 100 hours and I put it into the drawer, because it is a 4000K lamp and I didn't like the colours.



And a pretty new 150W 5200K lamp, used for less than 1 hour:



In both cases you can't really see what's behind the lamp, sure there is some transparency, but due to the curvature of glass and those opaque deposits inside you can't get a good picture of objects behind the lamp, so you can't collect much of the rays bounced from the reflector. Once I tried to measure the gain of using the reflector, but I had no idea of this issue, everything was aligned "as usual" (arc in the centre of sphere). I don't remember exact value, but it was something like 10%, surely less than 20%.
In theory we could expect nearly 100%, so where does that 90% disappear?

Regards

Both of my lamps are old. I got them second hand and used the g12 for experimenting with so it has allot of hours up. The only way I can get the g12 to get a dirty deposit like that is to switch it on for a brief instant.
You might try turning the lamp so that the nipple is not in the path. The nipple will severely distort the light as can be seen in your second pic. The lamp in your first pic, and I’m presuming that this lamp was operating with the nipple up or down, will have its salts solidify on the bottom of the vessel when cold. So taking a photo through this mess will block allot.
These salts probably do affect some light that passes back through but not 90%. After all the arc itself has to pass through them as well, not 4 layers but still 2.
I do realize that the vessel introduces some refraction into the path but if the arc is placed correctly then it can be compensated.

How far away were the lamps from the LCD in your pics?

No matter how you look at it, not having the reflected arc image go through the arc tube should be brighter; there are two layers of glass and the original arc too for the reflected light to go through.

Some more testing here might be a good idea.

With the near round arc that I have with the HMI type lamp when I miss adjusted the reflector I could see it on the screen and in the black screen too.


This might also explain the difference of why some PJ being brighter than others for no good reason


Would you then set the middle of the triplet arc image between the two arc image?

Sim: Well that is a fun site.

I see why they need perfect alighment for their use but do we need perfect alighment? Most of out lamps have 2 shells so the loss will be higher than the single shell lamps they and I use. Most of our arc tubes are long and thin so making two images side by side makes for a more square light pattern. The effect is the same as having a larger arc so all the bad things that happen from having a larger arc will happen but it may be brighter.

The hottest lamp out right now for the stage lighting is this one.

It has three filaments spaced closely.

Click to view attachment

Found this site and thought this may help here. http://www.microscopyu.com/tutorials/java/arclamp/index.html

So there is another idea, rather simple to verify (but I don't have time now): two pieces of a flat mirror, put behind the lamp, as close as possible to the lamp, and positioned at such an angle, that reflections of the arc would be visible as close as possible to the arc, but not covered by the arc tube. Like a shape of letter "V" made of two mirrors, mounted just behind the lamp. A typical mirror will probably crack, but if it's really efficient, it could be glued to some piece of a metal sheet, forming this "V-shaped" support, and a cracked mirror will also be efficient (I'm using badly cracked condensor lens, and it still works).

The main issue is that this big virtual light source should fit the projection lens. A 250W lamp will be almost fine (two images of the arc fit the lens), a 400W lamp is too big for standard OHP lens. And the whole light assembly should be moved closer towards rear fresnel, to keep all of those "virtual arcs" within the view of rear fresnel.

Regards

So what if it doesn't all quite fit in the projection lens? If there is a little spillage, you'll still get a huge increase in light.

BTW, what are "all the bad things that happen from having a larger arc "?

Tgreenwood

A larger arc can cause vignetting, ghosting, chromatic aberration to name a few. If the arc is made artificially large, as in using virtual arcs created by mirrors, the arc now occupies more of a 3D space rather than being closer to 2d so you could expect to see even more vignetting but instead of it just darkening the corners it will discolor them , either blue or red or a combination.

DJ

Edit:
All of this will be related to the size and quality of the triplet and how it is being used.

This is the idea I'm thinking of:



for the moment I don't see the reason it wouldn't work, at least with small lamps, or big ones and big triplet, but I have no time to do real test now. If we could catch all this light in the triplet, I expect lumen output would raise almost 3 times.

Regards

One potential issue with that is that the two mirrored arcs are virtually farther away from the collimator fresnel so the light might not all end up parallel at the LCD.

I like where this is going. A mirror would not be absolutely necessary. Simply two polished aluminum plates. They could be placed virtually touching the glass. Even if they are not mirror polished they may provide enough virtual light on screen to illuminate the whole screen evenly.

Maybe two AL plates with a small hinge connecting them so they can be adjusted to different angles?

Granted his experiment was a slightly different animal but mikyd tried two lamps and if I recall correctly there was some "problem" with "multiple arcs". Real arcs - mind you - not virtual, closely spaced arcs. There have also been some attempts with flat mirrors but again these weren't quite what Pepe is suggesting here. AV did suggest precisely what Pepe is suggesting but that it be done with actual arcs and not virtual arcs.

So I couldn't wait and did a quick experiment. Seems that this idea is not bad, there is some gain in brightness, compared to a spherical reflector (mine is quite a big one, about 13cm diameter, and it's a real reflector, not a big spoon), I have taken pictures for comparison, but I'm tired of uploading files to another server, and since I cannot add attachments here, I'll not post them here.

But it's not practical, because those mirrors got burnt within about 20 minutes. Heat of the lamp burnt the paint on the back and reflective coating, bad smell in the room.

Regards

Start polishing phutton.

Were you using that 150W Topspot lamp when you tested ?

I was planning to put my 150W CDM lamp back in my PJ [maybe] sometime this week (yea, that's a big maybe) to veryify proof of concept and put some numbers to it vs. the pro reflector.

gs

Pepe You may upload your images tho www.imageshack.us and link the thumbnails here. Regarding the mirror - you may use dichroic ones or try phutton's idea with minium.

No, I was using overpowered Osram 250W.

Remember this? It's thin aluminum flashing that I polished with a dremel and Mothers auto aluminum polish sometime back. I had made something of a reflector out of it since then but still in it's testing stage. Gotta get to work now but I'll upload some pics thereafter.

I have a sheet of dead soft mirrored aluminum od about 20"x24" if someone wants a small slice, it's free for the asking.

I knew I shouldn't have opened my mouth!

Give us a pic AV. BTW, still at work

First test results in.

NOTE: Both pictures are under-exposed, to see the light distrobution better.

Reflector centered


Offset roughly 2cm


I would say no noticeable change. (in my case anyway)

LL pro lens and fresnels, ll65K lamp, ll vue eballast, toshiba wuxga, 94% laser mirror from ebay, pro reflector

The little guy was so much smaller then. The metal is a little smaller now I'm using a small piece as an unbreakable mirror for the kids.

http://www.lumenlab.com/forums/index.php?s...st&p=128100

There was also the split elliptical that worked OK perhaps a little brighter than a normal setup. A smaller one might do the same thing or better than a flat plate.

http://www.lumenlab.com/forums/index.php?s...st&p=126599

I would say I can clearly see the second one is brighter. I don't know how much, I guess it's about 10%. This result is better with a smaller lamp.

Regards

Ok, this is what I did with the polished aluminum flashing. I've been reading from some members that they have been trying the aluminum foil around their 400w bulb and have seen some difference. To be honest I was surprised to see that it worked well compared to my 5" ladle but still there were some vignetting issues. BTW, the bulb (Ushio S400dd) with this flashing wrap was used twice @ 3 hours duration each time. The ballast (S51) core was overclocked.

This is what it looked like at the triplet. I found that when I set my background to 'black' on my panel I could actually see the arc a bit clearly without eye protection. Mind you that my projector is a horizontal with my Ushio bulb standing vertically. What I noticed was the arc seemed brighter at the top looking through the triplet. In this case I think I'll be sliding the bulb either down or up to see if it makes any difference.

I did make a pattern of a 'box' like reflector out of cardboard prior to the tube type but after cutting it out of the thin flashing it couldn't hold a much solid shape. I'll have to stop by the hardware store to pick a thicker gauge and try again.

Yes the camera picked up a bit. A lux meter would have been a better measurement device. But to my eye, nothing. I do have a large arc chamber, I will try this again when I get the S400DD like SIMUs lamp. Either way still worth a shot.

The issue may be mute if your arc chamber is too large. If the arc chamber is already too large to fit the triplet then a second virtual image may simply miss the triplet.

However, if you offset the arc slightly and use a virtual image you may be able to get better vignetting. Not necessarily more lumens, but more even lumens on the screen.

Maybe, if you put a piece of paper on the triplet (inside the projector) to see where the arc and virtual arc hit the triplet youi can verify that the virtual arc is actually inside the triplet.

They were side by side, tops and bottoms chopped off, but the centers fit.

Based on my meager experience I'd guess just about anyone could at least get better vignetting if this little trick pans out. My arc image fills the triplet in one dimension but certainly not the other. I can't believe I'm actually even considering trying this! But not to worry . . . I won't get around to actually doing it for at least another two years. Just when I think it's all been done someone comes up with or revives an idea. Check my new signature addition.

Well, I went ahead today and installed the 150W ceramic lamp back in the PJ. (Philips CDM-150/T6/942) I purchased a variac recently and decided to use it tonight to make sure the voltage to the ballast was constant across the readings. Reflector and precondenser did not move at all during these measurements (except for reflector being removed for 'no reflector' test). Precon is a 4.5x6.5, collimator fres is a 330 0.2 pitch, field fres is a 220 0.5 pitch, LL pro triplet.


No Reflector - Variac set at 120.1v

7.4 8.3 6.2
7.4 10.0 6.6
6.1 7.5 5.4

ave lux = 7.2 (x4.55 = 33 lumens)
vignetting = 63%

Click to view attachment


Reflector set to show Single Arc - Variac set at 120.2v

9.5 10.3 7.8
9.8 13.1 8.5
7.7 9.4 6.4

ave lux = 9.2 (x4.55 = 42 lumens)
[27% increase over no reflector]
vignetting = 60%

bench setup
Click to view attachment

setup from triplet - operating
Click to view attachment

Lamp moved to side to show Double Arc - Variac set at 120.1v

10.4 15.4 11.0
11.2 18.4 11.4
9.3 13.7 9.2

ave lux = 12.2 (x4.55 = 56 lumens)
[33% inc over pro reflector setup with single arc image]
[70% inc over no reflector]
vignetting = 54%

bench setup
Click to view attachment

setup from triplet - operating
Click to view attachment

One thing I noticed on screen was that misaligning the lamp to get the two arcs in the triplet caused some darkness at the very corners of the screen. I had to re-tweek the setup to get rid of those, which resulted in the following readings:

10.0 14.1 13.9
10.7 16.8 15.9
8.5 13.0 12.1

ave lux = 12.8 (x4.55 = 58 lumens)
vignetting = 66%

As you can see my adjustments got rid of a generally darker right side of the screen, but I still have some work to do. I think this is proof of concept enought to show that it works rather well for getting a lumen increase over the single arc image. The next thing to try might be the mirror setup... I have plenty of mirror that I can use for a quick test, but it will likely not last long even with a 150W lamp...

Thanks for posting your info Pepe.
gs

Very nice GS. We can say (as pepe said) it doesnt work well with large arcs. Small arcs it looks to be a good choice!

GS: Great job with the LUX readings.

You say you have a 4.5x6 condenser, what is the lamp to condenser distance?

Is it a 10.6" LCD or no LCD?

The screen is 4.55M and you only got 58 lumens from the ceramic with a LCD or 580 raw lumens with no LCD?

I have been waiting for over a year for some one to put a LUX meter on a ceramic setup (with a good condenser lens) but I still don't know exactly what this means.

Your geting 58 lumens with the condenser and a 10.5 LCD?



You have a vertical setup. With the ceramic does the arc move to the top of the arc chamber when mounted horizontal? If so it might be best to set the reflected arc image on the top of the lamp. This will let you have the two images as close to each other as possible.

It does look like the reflector trick gives almost the theoretical maxim for a reflector system. 33 lumens to 58 lumens. That is a 75% gain, Impressive. ( the reflector should reflect 85%)

It may pay to set the two arc images as close together as you can to make sure that you don’t make any ghosting or contrast problems.

EDIT

The refleceted image is slightly smaller too so it could be placed closer too and the readings would be slightly higher

NN, yea probably not with large arc's that's why I changed to the 150W lamp for these tests. I may also try with the 400W HQI lamp too and see...

A/V,
Yea 4.5x6.5 SurplusShed condenser, arc-plano distance is ~25mm (before I tweeked... i'd have to measure cause I did move it a bit)

Yea screen is 4.55m^2 (128" diag 16:9 screen). This is a 15.4" WUXGA Sharp LCD... it's only about 4.5 - 5% transmissive so more like 1200 lumens without an LCD... sorry I was sure i'd published Lux/Lumen numbers for my 150W setup... I just don't remember where

I have a horizontal straight shooter at the moment. The reason for moving it so far to one side was for the arc to completely miss the outter glass envelope. I'm not sure having the arc and arc image touching would do anything to increase the readings, although I could give it a little bit of a try.

I really wanted to see if I could get a 2x150W lamp setup to compete with an overdriven 400W setup... i've got the equipment, I just need to play with a frankenfresnel setup... (oh yea, and I need some time to do this... today's little experiment was easy for me, only about 35-40 minutes to setup and another 10-15 to make the measurements)

I don't want you to do a lot of extra work.

Even with the little extra light from the reflector trick the LUX is fairly low. Has the LCD been stripped? Did you adjust the condenser foward untill you got dark corners and then back up a little?

You still can move the reflector foward some and without the arc in the way you may be able to move it a lot more foward. look free lux

You do need to look at the black screen to see if it gets buggerd up too.

With 86 lumens your almost at what a standard 400 would do with no reflector and condenser. I always thought you could get around 150 lumens fron a 150 watt ceramic.

Adjusting the condenser and lamp is the big pay off in lux. I set the lux meter to the side of the screen and adjust for max LUX, you know the middle will be bright but the outside is what likks you. Then turn up the power. That thing goes to 140V right?

Just curious, Why do you need a frankenfresnel setup? If a double arc image works with one fresnel then shouldn't a double lamp setup also work with one fresnel?

Tgreenwood