Building my own parabolic reflector--technical questions.

[Gomer]

I have some 19watt 6500k screw ins that I really like, but the lighting isn't as even as I want.

...so, I have decided to build a parabolic reflector. I will be using some 1/64" thick mirror finish aluminum (looks just like my AH reflectors)

Does anyone have any recommendations for parabolic coefficient(y=ax^2 + b) "a" with respect to optimal lighting?

I understand that the all light entering a parabola will cross the focus point on the first reflection. it would make sense to me that I sould have the axial center of the light at this point. Is this correct, or should I offset in a particular direction?

Also, I will be using a smooth parabola rather than faceted like the AH reflector. Any reason to go faceted other than easier (I have the tools to fold the angles).

Thanks!

[Guttboy]

Gomer is that algebra????? LOL....honestly I have no Idea but where did you get your aluminum?

Mike

[Gomer]

www.mcmaster.com
They have a warehouse an hour or so away. Ordered yesterday around 5pm. Shipped UPS ground and I got it ~11am today

and I got 1/32 (typo)

search polished aluminum. I have the 24x24" sheet. I am using the extra elsewhere

[GulfCoastAquarian]

I've never seen a design guide for using an actual parabolic reflector. Probably because as you suggested, it is going to be nearly impossible to replicate the curvature accurately, unless it is a molded plastic part. I've done a few parabolic molded reflectors for a point source light on a vacuum cleaner, which is pretty straightforward.
But for a fluorescent tube, a faceted, angular reflector is your best bet. The most ideal design is a spider reflector that will help prevent restrike directly above the point source.
But there really isn't any voodoo science to it. Lay it out on paper, or even better - on a Modeling package like Autocad or Pro/E. Angle of incidence = angle of refraction.

[m.lemay]

Why don't you just model them after your AH reflectors. I agree with Sam that a parabola will have quite a bit of restrike on the back side because your point source ( the light) is not a point but more like a donut. The AH reflectors address that problem by having an angle to divert the rays away from the backside of the bulb.

Marcel

[Guttboy]

http://w1.864.telia.com/~u86438141/a...g_lighting.htm

Check out this article for some information.....its pretty good.

Mike

[Gomer]

Thanks, I did make a modified "spider" reflector. Restrike was minimized fairly well for the light dimensions that I used (2.5" outer diameter spiral compact).

Design Info:

Total of 10 facets.
All internal angles are 157°
The lengths from the outside to the center are:
1.76", 1.19", 1.01", 0.76", 1.15"
The light's axis is 2" from the ground plane.

Hope this isn't too confusing. after designing, it only took ~1/2 hour to make 2 reflectors.

I'll post pics when I get the camera back (boss is borrowing it)

[Gomer]

ok, here are the pics. As you can see, when modeling as a point source radiation, there is very very little restrike on a NON point source light.

[GulfCoastAquarian]

Looks good! When you take a picture of your tank, try taking one with the reflector, and one without, with the exact same shutter speed, f-stop, etc. to compare light output in the tank. I'd be curious to see!

[Gomer]

opps...sorry, I had an image doubled. That is fixed now.

I'll try and take both pictures. I'll just need to get a different bulb socket to mount on the outside instead of centrally.

In retrospect, I should have increased the 0.76" and created a steaper internal peak for slightly less back strike on the central facet.

[GulfCoastAquarian]

Hindsight is always 20/20, and you could tweak this thing for weeks and never reach a truly optimal design. With a good specular grade aluminum, you could even calculated double reflection and tweak the design for light that is reflected off the Al and reflected again.
I drove myself nuts before settling on the idea that I was only increasing efficiencies by a percentage or two over a design as simple as two bends (a top and two angled sides).

[Gomer]

(PLEASE correct me if I am wrong...I am just trying to justify ALL the time I put into the multi facets over the 4 facets ...well..the few hours I didn't want to do "work" at work)

The problem with the design that you posted is restrike still.

If we look at the long side, you still have ~82° of restrike (40.8x2)
On the short side, you get ~17.5° (8.74x2) of restrike. There is also another ~20° (~10°x2) of restrike off the double bounce.
This nets a loss of ~120° or ~33% due to restrike.

my restrike drawings show my minimum clearance before restrike. Unless I messed up, I don't get any other restrikes even along the path to the next facet (assuming point source light and non-point source backstrike light). This even acounts for second reflections.
The design I have only has 26.6 (13.3x2), 1.84 (0.92x2), 2.8 (1.4*2), 4.72 (2.36x2) degrees of restrike.
This nets a loss of 36° or 10%

[DLeDeaux]

I'm just curious what kind of equipment you would use to bend the metal to such specific angles.

[Gomer]

I am a grad student. We have a machine shop in the basement.

[Gomer]

Here are some pictures. Be sure to see my reflector comparison thread to see how it worked

(sorry for the blurr)
Just showing the reflectivity of the polished aluminum


here is my prototype reflector. The one I am using are a little longer (made with scrap)