I did not write it down too clearly, but if you read the OP carefully you will notice that attempt no3 was done with the same led that untreated produced the control numbers of attempt no2 :-) , so the comparison of #3 was indeed done with the very same led.
Djozz wrote:
The principle of this ring is not just block light, but to reflect light that is going sideways, and would be lost in an aspheric, back into the led die, some of that light will again cause fluorescence of the phosfor layer and so increase the brightness of the die.
I've read this in several places, but it doesn't make sense to me. I would think that the light that is emitted by the phosphor layer has little remaining blue or UV. So I would think the waiven collar or this ingenious ring of yours would be just lighting the surface of the phosphor, but not increasing florescence significantly. This is a hunch. So I could be wrong.
Try aiming one of your flashlights at an emitter. It lights up the emitter, but there isn't much obvious florescence.
Wow Djozz!!! Nice tests!!! 
I guess the best way to get throw is to use a dedomed LED with wavien collar and aspheric.
Keep in mind that to get white light you need Red, Green AND Blue.
The LED die emits blue light. This light passes through the phosphor layer which then emits the red and green. However, the phosphor layer intentionally doesn’t absorb all of the blue, because if it did, you wouldn’t have a white LED.
A wavien collar reflects the outputted light back into the die. Some of this reflected light should cause the phosphor to excite for extra brightness. I suspect another portion of the reflected light will also reflect off the top of the die and out of the emitter.
I agree there is some blue in white light. Maybe the florescence caused by it is drowned out by the red and green. Anyway, the real point of this thread it the technique and results in the OP. Don't want to side track the conversation.
Wow, what a grand project. I don’t think I would be dissappointed in a 20% improvement, especially since the distribution of an led is a bell curve with most of the light already going out the front. Any gains should be an incremental rather than a quantum leap. Maybe doing this in tandem with a partial dedome would be more effective than either one alone.
Interesting idea.
I assume for this kind of partial dedome you’d just use a knife to slice the top of the dome off above the silver coated portion?
I am not a physicist, but I have read a few things about optics and this is what I think would happen: the increase of illuminance by dedoming works because without the dome the difference in refractive index between die and air is so large that internal reflection takes place already in small angles from perpendicular to the surface, and that reflected light adds extra fluorescence to the die and thereby increases the illuminance. By slicing the top off the dome this effect does not take place (or hardly does), so it does not increase throw, while it does make the hotspot smaller.
if thats the case than no need to go this route anymore since there is impossible to have this increased effect and dedoming in the same time and since dedome is said to geave almost double that reamain more reasonable alternative. just curious is that hard to find an appropriate reflector and put it backard?
You need a globular reflector instead of a parabolic one, and that is indeed hard to find (use a decoration ball from the Christmas tree and make a cut-out somehow?) I read somewhere that it is notoriously difficult to center. And I don't know if the patent reaches as far as hobby use?
nice experiment/report
i think the idea is sound and the results are demonstrating that silver isn’t very reflective when deposited on silicone…
I’ve been thinking of sanding some of these in half and polishing the inside
I am not an expert in patent field but I believe that hoby use and experimenting is safe. regarding reflector shape I believe that little imperfections can be tolerated even some deviance toward parabolic can tolerate. I believe too that the biger the reflecting superfice the biger the efect gain will be.
and forgot one more thing, since u can mirror superfices why not to try on 1 half ping pong ball. also some convex lens can be used as secondary and mirrored in half like the dome. some from outside some even from inside. I will try to find a link with some apropriate ones.
www.fasttech.com/products/0/10007135/1583800-replacement-jgf-66-66mm-optical-glass-concave
this for example look like its not flat but curbet from inside. hope its not an ilusion of the photo.
Lots of ideas, I am going to try the xpe2 with silver ring one of these days, I will report :-)
Very impressive experiment and it provided some very good data.
I would like it just to get rid of the square beam. I wonder about driving these hard and what prolonged use would do to them. It would also be interesting to see what if any effect it had on temperature.
Thanks for the experiment, it’s very interesting and informative.
Good point. Did you happen to take any beamshots? You know how we all love those.
I will do some, but as I explained somewhere above, the aperture is at another focal point as the die, so when you focus on the die for maximum throw, you will still see a square hotspot (albeith with a bit darker corners).
Have't read all the replies but honestly a 20% increase in throw is nothing to sniff at...
I was expecting better, so I was disappointed, but you all are right of course, 20% more throw is considerable, and in an aspheric it is completely for free.
Well if you look at lights like the DEFTX they're spending money on wavien collars and seriously complicating their design to get gains that are even less noticeable than yours. They are of course demoing their LEDs. Maybe coat the dome and then SLICE the top off?
beam shots, left zoomed in, focussed on led die, right zoomed in, focussed on aperture just above led die:
I guess not much too throw is lost when focussed on the aperture, if the edge of the aperture is smooth the beam is quite nice, but you can't get around the blue edge with the use of a single aspheric lens, you need an very good achromat to get rid of that