If the rear fresnel only 'processes' light coming from the bulb as a point source, would it make sense to use a globe reflector that was larger than 1/2 a sphere? If you trace rays back from the corner of the fresnel to the target point, you'll notice that there is plenty more room for more reflector surface in front of the bulb, that won't interfear with the light path to the fresnel.

My solution would be to buy 2 half sphere reflectors, and cut out a rectangle that's proportionate to the fresnel. Inverse square law would say the closer the light is to the target, the brighter it will be. If anbody could find a 100mm FL fresnel, that might make for quite an increase. But for now, I'm planing on using the 220mm.

There's only 1 question I have abot my proposed design. Are there any disadvantages in terms of reflection for using a smaller reflector diameter. Would there be any issues with the reflector being just large enough to house the bulb? Would there be a better reflective response if the reflector was larger? Is there a sweet spot for reflector diameter?

EDIT: I found a local machine shop who said the could make something like this with a near mirro polished ceneter, out of aluminum sheet for around $40

The short answer is no.

Although there is more room for the reflector, that won't get more light on your fresnel. The only surface of your reflector that does you any good is the area that reflect light straigh back through the bulb and to your fresnel. The rest of it just heats up the bulb and shortens its life.

There may end up being some miniscule benefit, but IMO, it's just not worth the costs.

Hey Mad

Check it out in Raytrace.exe. I read in another post you just downloaded it. Since I spent alot of time with raytrace I can recommend a walkthru / tutorial so you can see what you'd get. Also might help as up to speed on raytrace

1. fire it up, "go"

2. = reflector =
2a. Create->Element->Surface. You should get a "Start" cursor
2b. Left-click 1/3 the way down the middle of the workspace. This is the bottom lip of your reflector. You should get an "End" cursor
2c. Right-click to get context menu. Choose "Circular Arc". Still "End" cursor
2d. Left-click 1/3 the way up the middle of the workspace. This is the top lip of your reflector. Try to get the line vertical. You should get a "Tangent" cursor
2e. Move the line so it points to the lower right at a ~45 degree angle. This is the angle of the reflector's lower lip as you drew in your bmp. Left click (length of line doesn't matter). You should get the reflector.
2f. If you don't like it, right-click and choose "Undo segment" then back to step 2c. If you like it, right-click choose "Finished"

3. = screen =
3a. Create->Element->Screen. This is your fresnel+LCD. Left click the upper left corner of the workspace
3b. left-click the lower left so the screen covers the left side of the workspace.
3c. right-click "Finished". At this point the workspace should look alot like your bmp

4. = lamp =
4a. Create->Source->Point. Choose 3 rays. Don't check "Drag on create". Cursor should say "Center"
4b. Left-click between the screen and the reflector (not inside the reflector, yet). Cursor should change to "Aperature 1"
4c. Left-click in the upper right corner of the workspace, and just to the left. Cursor: "Aperature 2". (Again this is a tangent)
4d. Left-click in the upper right corner of the workspace, and just below. Down and to the right of 4c. You should get a point source with three evenly spaced rays. This is your "lamp"
4e. If you have the patience: select one of the rays, and Modify->Ray->Green->Ok. Select another, modify it to blue. Leave the third red

5. THE INTERESTING PART
5a. Left click to select the lamp. It should show little red boxes
5b. Left click it again. Popup: "Source Drag Options". You want to "Drag all rays".
5c. Drag the lamp into the reflector.

Watch what happens to the rays.

Notes
- Raytrace only bounces a ray 10 times, then stops it. 90% ^ 10 = 35%
- You can change the radius of the sphere by left-clicking it, then dragging the tangent. The lips will stay in the same place. To get the smaller sphere you could
a. select the sphere
b. Modify->Scale->Scale by dragging
c. base: right click "Center snap", left click the sphere
d. point: left-click+drag near the sphere.


Edit: i guess this was a long answer

I don't know what I'd do without ya supra guy.

Thanks for the quick guide charlie.

Played around with it last night just so i could get use to the gui. Any way to construct lenses based on given measurements, diameter, widths, and focal lengths? Or do you need to use a more advanced optical planning program to do so?

I did make a region that looked like a convex/concave lense, but I can only guess what the actual focal length of the thing I created. Any way to display a measured grid system...show some kind of ruler so that things could be built to scale, yadda yadda yadda?

The lens button on the left lets you specify radus of curvature, height for a spherical lens. Not focal length. Help->Tutorial has a chromatic aberration tutorial that gives you an adjustable doublet. I was eventually able to make an adjustable triplet; let me know if you want details. I was in over my head at that point :-)

There's also Options->Grid->Snap To Grid.

Can you post your files???

I wish I could -- unfortunately the demo version doesn't allow saving!

I'm not at my computer with raytrace on it so I don't know the licensing fee...I'd be interrested in buying in on a 'group' license of some optical softwaer if the price is right. The more tools provided to DIYers, the better the quality will be for the rest of us.

The may be some good deals out there for educational licenses...some of us are actually in school.

I wonder how difficult it would be to design something like raytrace. I've been teaching myself opengl...if I can dig up formulas for the linear transformations performed by lenses, maybe I could throw something together that meets our basic needs.

I don't want to loose the developer any sales, but he allows you to print the order and fax it and enables the saving files for something like 14 days before it realizes that you haven't gotten your unlock code yet. During this period you can save and load files fine.

Hrrm... A site license... Interesting idea - though probably not legally applicable to all LumenLab members, it seems like a small enough enterprise that one might be negotiable.

I really like that idea. I could see paying a separate fee or something for access to "site licensed" sw.
Worth noting that the demo version of Raytrace (v2.20) has an annoying bug or two, at least on my laptop. They seem to occur after extended use. Maybe the full version doesn't have the bugs.

I think the demo version is the full version...untill you put in the key...like winzip.

Initially, the program was VERY unstable on my machine - crashed all the time.

Lately, I can get it running for long periods of time without problem, but every once in a while when I'm messing with one too many rays, it crashes.

I would be fine with it crashing and me closing it, but it bluescreens my whole win2k pro machine.

I didn't get bluescreens, phew. but did get 100% cpu usage (according to task mangler) and learned to stay away from more than 10 rays per source since the program seemed to start freaking out once it tried to draw too many rays and gave the "Out of rayspace" message.

Their website ( http://members.ozemail.com.au/~imesoft/ ) does say
"Every attempt is made to address genuine bug reports and to issue updated files as necessary".

My PC can handle up to 100 (max for the program)rays , smoothly, untill you introduce more than 3 surfaces, then is tells me to cut back on rays.

Going to run this one when I get home...Don't know how that front projecting lense handles light coming in at wierd angles. May have to introduce some sort of prism to get light bending the right way.

You'll have big gaps in between the three beams that are hitting the condensor. Generally, setups that seperate the light and then recombine it suffer from that.

Lighting setups either suffer from a nonuniform beam(shadows on the screen, that sort of thing), low light output, not being adequately focused as a point source, or light distribution gradients.

The LL design is focused enough to use our triplet, has about 10% of light hitting the fresnel, suffers from a gradient wherein somewhere between 50%-75% dimming at edges, and has zero obstructions to the light path.

Pretty damn hard.

You'd be much better off experimenting with POVRay, a free ray-tracing tool. Have a look at povray.org


Take it easy,
Phil.