I got to thinking more about the corners being darker and did a little search.

For full text see
http://oemagazine.com/fromTheMagazine/dec02/taminglight.html

Haven't looked into he we DIY'ers can utilize a tandem array. GOing to have to poke around the USPTO

John

I think that's more wild-ass than scientific. The bulb has a spherical-like shape around the filament. It's not perfect, obviously, but it does look like almost all the light that's going to actually fall on the collimating fresnel will pass through the bulb envelope pretty close to the normal.




You're almost right here, but not quite. A corner cube reflector will reflect light back out parallel to the angle it came in at, but it may be displaced a bit. So it's not guaranteed to pass through the point source it came from.

A spherical reflector like the Norpro, however, WILL reflect all incident light back through the point source it came from, provided that point source is positioned correctly at the focus of the reflector.

As I understand it, the fresnels we're using for collimating have a bit of a "fudge factor" to deal with not-quite-point-sources. So if the feature size of the pyramid structures on your corner cube reflector was under say 5mm, then it would prolly work fine.



I did some research on this, and yes, you're right that there are retroreflective surfaces using spherical glass beads. But I think it's a big mistake to think that such an arrangement will be a more efficient reflector than a highly polished silver, chrome, nickel, or stainless steel surface.

If the light is transmitting through glass at any point, then there's loss due to absorption, plus when the light strikes the surface of the bead and enters, some of it will be reflected off in a non-useful direction. Also, a glass bead on its own won't serve as a retroreflector. It needs to be placed on some kind of reflective substrate.


I have to admit that I'm still very interested in seeing some results of these experiments! Let us know how you make out.


Phil.

hey jerseyjohn,

would you mind giving the name of the coating you discovered? and how much is it going to cost?

thanks

Ferday,
The info is in my laptop. Will boot it up tomorrow. It was 30.00 for a can, precharged and would probably coat a few dozen norpros.

Syscrush,

SWAG is a shop term we use for best guess. The corner cube reflector will reflect parallel and offset dependent on the size of the retro. If it is a 2mm retro, the max deviation is 2mm.

The Norpro will not reflect all light back to the source because we have changed it's relative position to the focal point of the norpro. The norpro FP is say 3.375 inches (lets agree on that for a number).

Now, and I have not measured the filament size but lets assume it to be .75 inches. Any light source not in the center will change the bounce and reflect elsewhere. I have not done a ray trace but I am a bit of a solar thermal junkie.
A point source that is 3.00 inches away from a 3.375 sphere will not reflect back from where it came. Only the exact center will do that.

The paint is the source of the back reflector in the spheres.

Where I am going, and again, I have not even built mine yet (waiting for corrected fresnel) is increasing the point light source and trying to use what i know of solar thermal applications to do that. This relies a bit less on reflector shape and more on optics.

Now, let's assume for the moment that the bead coating works. Then nearly 100% of light will go back to the light source regardless of it's relative position to center. Furthermore, additional surfaces could be placed under and above the bulb to reflect more light back to the source, thereby increasing useful reflected light.

I will not be doing any of the bead thing until I get a month or so of use, probably in January. Gotta keep the wife happy and kids entertained and justify my expenses.

I enjoy the exchanges in the forums and appreciate your comments. Please remember, it's just a theory......

John

i'm more interested in the coating as a way to fix error than to boost light output. while the norpro may be close (i hope so as i ordered one..) it is certainly not a perfect sphere, and points on the lamp outside the focal point will indeed be viewing a less than perfect sphere regardless of bowl quality.

this is evidenced by a diffraction pattern of hotspots on my screen (and the fact my current bowl is barely round)

or maybe the norpro is close enough to round to fix my errors, we'll see....it's all fun anyways.

Ferday,

Here is the info on the reflective paint. As I said, It's a theory that I can not test until January or so. If you go for it sooner, be sure to post results.

http://www.usreflector.com/Welcome.html

Clicks
Products---download catalog---reflectors---page 8

Page 8 is a 876k PDF
The paint is in the lower right hand corner $30.00

John

Hey all, thought I'd give my two cents on this one.

I'm confused as to why we need a COMPLETE hemispherical reflector, we don't need a full hemisphere.

The top of the attatched picture shows (not to scale mind you, but that's not TOO important for this demonstration) a hemispherical reflector, the point light fixed at the focal point of the reflector, and the fresnel lens. The yellow shaded portion is the light that is USABLE. The pink shaded portion represents the light that is NOT usable. The image was made by simply ray tracing from the corners of the fresnel lens, though the point light (which is at the 'perfect' focal point of the lens), to the reflector. Since the reflector is (as close as possible) a hemisphere, the light from the point source is reflected right back at itself, I think we can all agree on this part.

Now, any part of the reflector that isn't inside the yellow light cone is reflecting unusable light. The reason the pink shaded portion of the thrown light is unusable is because it is not entering the fresnel lens at the correct angle to be collimated, after bouncing off the sides that is (which in and of itself is going to absorb a great deal of the light energy).

The second scenario presented would be a better way to capture more light from the light source. In this example, the fresnel lens has a shorter focal length, thereby moving it closer to the light source and capturing a bit more light since the angles are greater. I have no idea of this kind of fresnel lens exists, but if it did, it would go a lot further to increase the brightness of our projectors.

Since only the bottom portion of the norpro bowl is the only part that counts, it doesn't matter if the ENTIRE bowl is perfectly hemispherical, only the bottom portion (which as I understand it, isn't, but then I don't have one to check =)

Any thoughts, corrections?

-Jon

we don't need a complete hemisphere, your drawing is correct, the math has been done and i think it was somewhere around 20-30% of the bowl is actually used. some say that the rest of the bowl could control stray reflections however, so they use the whole thing.

my current bowl is cut down substantially, still not as small as it could but pretty small less than half of original size. when my norpro comes i think i will leave it full size to try it out (if it fits in the box), plus cutting a stainless bowl down really kinda sucks.


ps thanks for the info jerseyjohn

The hemisphere is desired because --- Theoretically a sphere reflects back all light back top the focal point if it originated at the focal point of the sphere but our sphere is not perfect, our surface is not perfect, but the entire hemisphere assists in driving light back to the focal point to increase the light intensity on which the fresnel is focused and controls some stray light.

My scheme about the reflective coating is a potential means to increase the amount of light bounced back to it's origination.

IIRC the best guess was about 40% effectiveness on output

John

Excellent, glad to see I'm going in the right direction =) As for the stray reflections, it could be worth it to put two bowls together to form a full circle and cut an opening in the second bowl to allow only the light that is used through. I ran into a picture of a 'standard' projector bulb and it had a configuration like this, though I do think it is too complicated to really try, painting the insides of the box a matte black would probably be best to handle the extraneous light <shrug>

Has anyone done any math on how much actual light from the bulb is being used (not counting what's lost in the glass, lenses or LCD panel, just how much of the light starts out usable?

Thanks again!
-Jon

jerseyjohn,

I understand your want to have all the light reflected back to the source, hence why you would want a hemispherical reflector. My argument is that it doesn't matter if all the light is reflected back to the point source as most of it is unusable anyway. A bowl that has a perfectly round bottom but straight sides would do just as well as a complete hemispherical bowl. I'm just suggesting that we are unnecessarily limiting our options and that finding a lens that can be mounted closer to the light source would be a better way to increase the brightness (or at least as good as finding the 'perfect' reflector).

-Jon

Jon-
I have to agree with you. A small reflector is all that is needed. Extra light just returns more heat to the arc chamber and that can have undesireable results. Only light that is headede straight back through is of any use by the fresnel.

The only other chioce is a parabolic reflector that uses a condenser lens to try and cpapture ALL the light but this adds complexity to an already very simple design. The LL projector woks on brute force not advanced optics. Thats why its cheap and attianable to a lot of people.

Cheers,
Kirk

The amount of light passing through the collector lens versus the light emitted by the bulb even when using a reflector is only 27.76% That is we are using just under 30% of the light emitted from the bulb. This calculation was done for a 15" LCD and a collector fresnel with a focus length of 220 mm which is standard.

another way to get theoretically 100% of the light, is to use an elliptical reflector with the lamp at one focus. all the light is focussed to the second focus, creating a true point source (in theory...). place the collector fresnel 220mm from this point and you have an ideal point source, very bright. good luck finding one the right size for cheap, and the larger an ellipse is the less likely it is truly elliptical.

in reality, no reflector will get 100%, although we could certainly do better than 30%. commercial projectors often use several combined reflectors to direct and collect the light as effeciently as possible.

the only real (although large) benefit to using such an ideal reflector is to lower the wattage of the bulb, the 400W and a basic spherical produce pretty bright images.

Here's a diagram showing why it's only around 30% I assumed perfect mirrors, perfect point source and used a 15" lcd and a collector lens with fl=220 mm.

Modified post

Pickle,

Does this applet confirm or better illustrate what you are saying?
http://webphysics.davidson.edu/Applets/optics4/

I used a mirror with a focal point of 3.4 and a point light source on axis
It gave me and X dimension of 3.5 inches.

Is it then a factor of the angle of the 220mm fresnel (roughly 27 degrees up and down) and working backwards to determine the cone / reflector size that is truly relevant and usable?

John

Jerseyjohn,
Yes that's a helpful applet and you are close. Except the height you want to use is half the diagonal distance of the collector (I'm going to assume it's the same dimensions of the LCD and for this discussion I'll use a 15" panel w=12" and h=9") So what you want to do to find the angle of usable light (this is also going to be the angle describing the useful surface of the reflector) is use atan(7.5/8.66)=41degrees up and down. Let's say you cut down the reflector to the bare minimum here's a drawing showing what light hits the collector depending on which angles you use to calculate the useful reflector area. The rectangle represents the collector lens. The yellow is the light hitting the area which is coming from both the bulb and the reflector. the blue is light that is coming from the bulb but doesn't have a reflector behind it. The first two have cut their reflectors too small and the beam profile is then uneven and will get darker corners. the last one is the minimum you can cut your reflector without cutting into the picture. The area of the circles outside the collector is another place where light is wasted. (see first diagram).[edit] the little reflector diagrams at the bottom are not to scale and the angle is measured from the center of the sphere where the lightbulb center would be[/edit]

The only thing I've seen that would cause a greater amount of light to be captured from the bulb is the parabolic/eliptical reflector. I saw one design that had an eliptical reflector in back and a spherical reflector in front with a hole cut out of it. Of course one can do the math to find out what exact shapes we would want to have but they are nearly impossible to find (small half ellipsoids) and so I would like to find out if there are any substances which can be used by DIYers to paint on a mirror surface. I've never really heard of any and the idea of retro reflective surface wouldn't work for an ellipse geometry since it would still reflect light back to it's source does anyone have a clue about this? I'm thinking it would be hard but doable to make an ellipsoidal shell myself (with some tricky processes I learned from fiberglassing) and then paint the required mirror on or perhaps some sort of metalic flake paint that could be pollished to a mirror shine? Anything short of getting a vapor depo$ited thin film put on it. Anyone have ideas or recomendations?

I saw that drawing, and It looks like a good idea. If you could recreate the dimensions of a reflector like that.. I could get it machined out of aluminum.. if It works i could even make a small run of them for you guys. It would have to be big enough for at least the tubular ushio, if not the normal pulse srike bulb.

lets get reflecting!

-michael

I was working on some geometry today and it wasn't too tricky. The ideal bulb for this would definately be the HQI form factor (the double ended tube). It is very small and most of the rays bouncing off the rear side of the elliptical reflector would "skip" over the bulb entirely on their way to the second focus of the ellipse. My worry is that with the regular EYE/whatever form factor, you get light hitting the elliptical reflector and then on it's way to the second focus it passes through the outer glass of the bulb and gets bent away from the focus. I have no clue how much of a problem this would be. I do know that with a spherical reflector you don't have this problem because the outer bulb is generally circular so the reflected rays don't get bent much by the glass. What do you think? What kind of dimensions do you need for getting it machined? As in could I just give you the major and minor axis dimensions or do you need a function or surface points?

So with an hqi you would have to drill holes in the sides instead of the rear correct? My cousin is a professional machinist, he makes airplane parts.. I just need enough information for him to compute the actual shape. The more the better i think. I think that we should try one for the standard bulb shape.. and one for hqi. I already have a standard lighting setup, and HQI seems to have much less bulb choice. I do understand the advantage however. So lets try everything.

Would there be no need for a precondenser lens with this setup?
I cant seem to find the image of the reflector, do you have anything you can post?

What if we just got high temp resistent reflective paint, and painted the back of the bulb. Would not that bounce all the light back right to the front of the bulb? I dont know much about reflection but it soudns like a good idea to me.

for the thing I'm talking about we would still need a condenser fresnel but not a precondenser lens. Here are some quick drawings explaining the idea.

And...

And, I don't know how to add multiple pictures to a post without them being online??

And finally... I'll work a bit on the geometry I realize that there may be a hot spot potentially, that can be fixed with a little more geometry work. I'll try to have some numbers and such up for you next week...

pickle, Depending on the substrate one could silver leaf a reflector. Available from sign suppliers and art stores. Then polish it to a shine.
toby

I like the way you are headed with this, it might just work. I think that the easiest way to fabricate your ellipsoid reflector might be ceramics. If we could source a highly reflective mirror glaze, you could make a high temp and easily reproducable reflector at a relatively low cost.

Picklejones:

Your right that the current design would cause a hot spot, owing to the sphere's reflected light being spread over a smaller angle than the ellipses. It would be very like your diagram for the reflecter where a small reflector produced a hot spot. The solution is to have the ellipse cut of perpendicular to the line between the focuses and pass through the first focus.

I have drawn a diagram of this.

The angle of the light passing through the hole can be controlled via increasing the distance between the focuses to decrease the angle. If the focues were next to each other than the ellipse would become a sphere and the light would pass out over 180deg. As the focuses part this angle comes down.

The radius also decreases the angle as it decreases, but makeing the reflector very small increases the light lost where the bulb goes through the reflector. In an ideally sized design the reflector lost for the bulb positioning should only cause a dark spot in the circle above and below the rectangle of the screen. I'll post a diagram for that too.

A bit of a poor diagram this, but it shows the effect of the bulb cut outs and how the 'radius' ( sum of distances from the focal points, not really sure what this should be called ) changes the impact of this.

The digrams show too larger ellipse, the right size and too smaller ellipse.

This assumes a vertical bulb that passes through the first focus. Vertical seems better as a smaller ellipse can be used, owing to more wasted light above and below than at the sides.

Thinking about all this a few more thoughts came to mind.

The human eye is good but not perfect. The standard design is roughly 10% brighter in the middle compared to the edges, based on quick maths so don't take it as gospel, but viewers have all said it is unnoticable.

This is not due to any flaw in brains design, it happens to any light. Lights give out the same energy per degree. As you move an object closer to the light it takes a larger angle from the light to cover it and so it received more energy over the same area and so apears brighter. Unless you have a sphere for a screen, which wouldn't look very good I'm sure, the corners of the screen are further away than the center from the light.

The bulb in the light path may cause a darker area in a stripe down the middle, this may or may not be visable, only trying it can really say. I suspect it will depend on how diffused the boundry between the dark and light areas are more than the relative intensities.

If this reflector were built into the bulb, then there would be fewer problems, other than getting very high temp reflectors ;-}>. Does anyone know if bulbs with reflectors within the caseing exist?

MH warehouse lights often use elliptical reflectors. they are poor quality at best and would be likely useless for us. in those designs the bulb lines up with the major (long) axis. this is not an acceptable method for the projectors, we must come in from the minor (short) axis.

the university will build me an elliptical, at a high cost. a machine shop may/may not be able to fab one, it is difficult to get a highly accurate elliptical shell especially considering our space constraints, it would have to be a well designed reflector. there would be no dark spot in the bulb path, but hotspotting would occur without absolutely perfect alignment.

some commercial pj's use the ellipse/sphere combo reflector, seems like a lot of trouble. some guy on the AVS forum tried to make one and gave up.

Moose,
Good idea with the silvering. Could also accomplish a mirror finish on norpro section with jewlers rouge

Left field thought
Apply silver or gold leaf to the bulb.

I have been trying to find a picture of a ellipse / spherical reflector used in commercial PJ's but I keep coming up empty. Anyone got a link or two?

Polishing to mirror finish is not the hard part, it is the shape and size and distance. Can any of your college folks talk to a physics instructor?

John

I can machine a cavity / mold or maybe a core ( like they make fiberglass boats ).
If someone can come up with a process to produce a mirror from the molded or formed part.

jerseyjohn,

never seen a picture of one, just read the schematics on a couple. i used some combo reflectors back in university days for discharge lamps, they tend to be very small. the only pic i can find is a hard to see one (probably cause patent is pending...)

http://www.helicon.de/b/hes/icelptsphr.gif

the design would not be so hard. i have no concept of how to construct one however.

Even though fiberglass is heat resistant, I doubt that the resins are. Are you guys sure that a reflector can be made of fiberglass?

I use Fiberglass and epoxy resin every day (Its my job). I use a few high temp epoxys that are very expensive and still only good to about 350f. I would think a lamp eflector gets a LOT hotter than that. The themal stability is also pretty bad and is prone to warping. I think youy should be barking up another tree.

Kirk

That's what I thought. I still think that the easiest way to make a custom reflector would be with ceramics if we could find a highly reflective mirror glaze.

I still think that a machine shop can make one. At least to the precision that we require.. As I said, my cousin makes airplane parts all by computer, those have to be very precise. Aluminum can get a great finish.. or it can be chromed. If someone could come up with the dimensions for a combo reflector we can probably get them made for cost.

I don't know pricing on polyimide resins but you could get higher temp resistance with them vs epoxy resins. But I think that Deathray has got the best solution with the ceramics. If we had a mold (or a solid plug) via madmike or Haas_man we would just coat it with release agent, stick ceramic clay all over it (it wouldn't have to be even on the outside) dry it, pop the plug out and fire it, then coat the inside with mirror sauce and bam! sounds easy right?

Madmike, I'm going to crunch some numbers for your cousin using a HQI bulb form factor. Do you know where I can get the exact dimensions of that? Including the arc envelope length as well.

400 or 250?

Hey Pickles,

Could you use a hardwood wood mold? If so, I can turn one on a wood lathe if I know the shape. What I could do is make several ellipse templates with say 3.0 inch focal, 3.25 focal and 3.5 focal and send them to you. Perhaps you could mold the three and see what happens.

Will not be able to do for two weeks or so. Gotta hit the road. Need to go the 6 cities in three states in the next 7 days - gotta love sales........

If wood will not work I can try to get a freind to turn one on a metal lathe. Perhaps wood to determine size then go to metal.

John

I think some ceramic research is called for. There are likely shrinkage issues to compensate for or there may be some low shrinkage clay available. Also I'm not sure if we can find a suitable"mirror sauce". It would be best if we could find a glaze that would be suitable so that it could be fused on to the ceramic.

Now it looks like we are getting somewhere. Brain, can you post the specs for both bulbs?

Sorry, gone for a day and.... I wasn' t suggesting making a fiberglass reflector , just the process. Create a plug or core and mold something over it. I was thinking about a ceramic refractory paste. It is used to suspend the heaters in kilns or fix the lining of a furnace.

But we still need a method to apply a reflective surface to it. Thats where the boat method came from. They paint the inside of a mold and then glass over it , when it's removed from the mold, the surface is finished and glossy.

I also am working on a DIY CNC lathe and I am far enough along to create a spherical or eliptical surface up to about 10" in dia.
here is a link to some :
paste

Brain, If this works like I'm hoping it would then it should be around 75-80% efficient and would be best used with a 250 w bulb but since people are willing to test these things out we could try with both 250 and 400W bulbs, do you have dimensions on those? and shouldn't they be about the same dimension wise since they have the same socket setup?

Jerseyjohn, I see nothing wrong with a wooden mold if properly treated though a metal one seems prefered. If we use anything wet to cast with (wet potters clay for example) the wood might warp slightly. I'm not sure how precise this setup would have to be. I use the eye bulb in my current setup and the arc length on that one is around 1 3/4 inches with fairly even picture lighting so I'm thinking the optics aren't terribly sensitive to a true point source. So I think wood wood be fine for initial tests at the least.

The 400 is bigger. I'll get dimensions tonight.

Anyone familiar with poured rubber molds? I was looking this up for molding projecter parts the other day. I am sure the rubber mold is suitable. Sounds like the ceramic stuff someone else was talking about would be pourable onto one of these molds. This stuff takes perfect replica molds of the subject. If it has a mirror polish the mold turn out mirror polished replicas. Sorry I don't have time to give the links right now, but it is easy to find searching for polymer and mold and cast making etc. Looks easy and economical to do at home. Just a matter of having a high temp media to mold and being able to metallic mirror finish it.

What is the current state of the best reflection setup these days?

Still the Norpro at this time.