Still may see a diffraction pattern from the aperture edge. 
I applaud your patience as well as your creativity. Thanks for trying and showing us the results. A very interesting idea.
> half a ping pong ball
How about something like a M@g AA reflector, cut down as needed?
They’re fairly soft plastic, nicely silvered.
Thank you scottyhazard, it was fun trying it all.
@hank, comfychair has tried a cut-down mag-reflector and it does make a difference, but how much was not measured. Ideally though, the reflector should be spheric, not parabolic.
Cool idea / experiment, Djozz.
Ah, that used a big M@g C/D reflector; I was thinking about the little AA-size reflectors that would sit much closer to the emitter.
I’d like to see someone make a zoomie with a waiven collar.
I can see how it could be done:
- Mount the LED on a 10mm star and mount the star on a copper pillar.
- Attach the waiven collar to the sides of the sliding bezel that holds the lens. The opening at the top of the collar should be wide enough so that the LED pillar can pass through it.
- When the light is in spot mode, the waiven collar and lens should be at the optimum positions for max throw.
- In flood mode, the bezel would retract back towards the LED. The waiven collar would also retract back around the sides of the pillar and out of the way below the LED.
Would be an interesting design. You could even give it a third beam position by installing a conventional reflector on top of the waiven collar (the opening in the reflector would match the opening on top of the collar). Then you’d have the following:
- Spot mode - lens at focal point of waiven collar and aspheric lens for max throw
- Intermediate mode - lens at focal point of conventional reflector. Waiven collar is out of the way below the LED. It would be fairly floody as it’s still passing through the defocused aspheric lens, but because of the conventional reflector it would have a noticeable hotspot. Beam wouldn’t be that wide due to the LED being quite a distance from the aspheric lens however.
- Flood mode - bezel retracted all the way until it’s almost touching the lens. Both the waiven collar and reflector are now below the LED. Should give a wide even flood with no hotspot.
The dome curvature is diffusing the emerging light (emitter source close to the PCX lens emulates a zoomie creating lots of spill). You could try shaving off that spherical end cap, then refracted light will return to the original incident angle.
A successful Wavien Collar treats the LED as a point source. Each point on your wide LED is near Lambertian, so not all points have a centered return, not even close in most cases (look at a 2D profile of your set up and trace extreme rays). The resulting exit cone of emerging light will be wider than the half angle cone of the silver port.
Also, probably mentioned before, you’ll recycle more photons and collect more light from the wall waste angles if the probability of a reflection is higher. The Wavien collar has a 25% chance of escape (sin 30)^2 and consequently a 75% chance of reflection (1-(sin 30)^2). That higher probability increases the contribution of each bounce in the terms that sum to the gain.