De-doming is an interesting thing… But there I wondered for some time - or is something even more effective?
I was inspired by the car LED bulbs:
light is reflected to the sides
Maybe the same can be done with CREE LED’s?
here is a standard XM-L:
(the rays drew on one side only, but I hope that you understood the idea)
here is a de-domed XM-L:
And here is My idea - We might drill a hole right through the middle of dome:
This could reflect light to the sides - to the reflector, and we can get more concentrated beam (more throw)…
interesting idea, but im not sure how reflective the dome material would be if not entirely smooth, or at the right angle like the auto bulbs you pictured.
i thought of a different idea once of mounting 4 XMLs “sideways” ( facing directly at the sides of a large reflector ( in a similar manner that a HID discharge tube directs its light horizonally,) but mount the emitters ditectly to a square heat-pipe protruding from the center of a reflector as a heat sink.
I think polishing the drilled/tapered dome part will be very difficult. Without polished surface, this idea may not work, but simply scatter the light even more.
I doubt if the idea will work at all. The principle of dedoming a LED is not about allowing more lights to scatter sideways, if this is the case then why dedomed LED would work as well in an aspherical setup?
Dedoming LED is to increase luminance at the die by reflecting the photons back to the die, to “reuse” them, hence you get higher luminance, or surface brightness as we used to call. Check out DrJones’s thread here .
The cut would have to be polished, obviously. But it would be easier to cool than with the led facing backward. I don’t think it increases the luminance like de-doming does. Not much light would be reflected back to the phosphor.
dedoming reduces apparent die size and scatters more light sideways.
In aspheric lenses it works very well because the apparent die size the lens "sees" is reduced so much. Hence you get smaller projected image of the die when in full throw.
It works well in reflectors because more light get directed to the reflector surface and becomes part of the hotspot.
I think for something like OP wants you have to do the exact opposite. remove everything except a small cone of reflecting material in the middle that has a size that exactly reflects all the light that would otherwise become part of the spill on to the reflector. However it cannot be perfect because the focal point of the reflector shoul still be on the surface plane of the dedomed emitter and the focal point will not have enough depth to include the reflected light perfectly into the hotspot.
Yeah the lens “sees” smaller die size, but if only die size is appeared smaller, how would the throw would improve so much from there if the luminance (surface brightness) of the die never changed?
My question is how simply a smaller die size if without changes of luminance would improve throw?
when dedoming we loose overall output, measured as lumens in a sphere.
So surface brightness is less after dedome.
but before dedome lens or reflector "see" much bigger area, maybe double the area, then dedome and loose maybe 20 % output, but the remaining 80% is seen as coming from only 50% of the original area so surface brightness is seen as 0,8÷0,5= 1,6 times brighter. Then allow for better focusing on smaller emitting plane and you get even more throw.
I am not entirely sure how to 3xplain that. But wit an aspheric lens you will often see that the tint with bare led is slightly less blue than when you put lens in front of the led. It seems that the optics push the light to the cool side. Maybe the dome also does this. Cannot remember right terms for this effect but it is described in optics litterature.
Why dedoming improve throw by large margin, and why is the tint shift, are explained here .
Actually to be frank, I used to understand the dedoming mechanism as what you do now, but I think Dr. Jones explanation makes more sense. On a side note it is very difficult to find some recognized sources or journals which directly explain on the dedoming effect of the LED. Well, who else on earth except flashaholics like us would attempt to dedome a LED? It takes someone who has deep knowledge in every other aspect of optics to interpret the dedome mechanism.
I was reading some old CPF posts on aspherics and the image projection is partially a function of the lens design. I was surprised to read that a larger aspheric lens diameter could project a smaller image if the lens design allows. I’ve now got three XML zoomies and saw this seeming (happy, for me) dilemma. The curvature and distance from source where it is focused have something to do with this.
I hope somebody who know opitics can give a short tutorial on ashperic behavior. Most of the stuff I’ve found wasn’t grounded with definitions and really most of what I found was for eye contacts and light entering instead of projecting.
The light is re-radiated isotropically by the phosphor, and light that comes out close to the plane of the led has a longer path through the phosphor. So a reflector reflects the yellowest light and misses some of the bluest to the spill.
I see a lot of colors with my Fresnel lens, which reminds me that aspherics must also have chromatic aberation and that is part of why there is a yellow ring around the spot. The other part is that the light that goes through the side of the lens de-focuses farther as with a mirror.
I learn new things everyday, this is why BLF is an excellent community.
I don’t know the reason behind, but I only know that given a lens diameter as fixed variable, the thicker of the lens (thicker protruding) the shorter of the focal length of the lens would be; with the shorter focal length more lights are captured by the lens hence more OTF lumens, and the projected die size appeared larger as well.
If comparing two lens both with identical diameter but different focal lengths, they will still produce the same candela (throw), given that the light source, lens materials & quality… etc are the same. The only differences between them are the OTF lumens and projected die size. They both have the same light intensity unless you change the other factors.
The idea isn't new (link), but it won't increase throw, neither does that funny SP6 setup, which just pushes it's more flux into a wider beam.
The one thing that increases throw when dedoming is the total inner reflection sending light back to the emitter and increasing it's luminance. The 'emitting more light to the sides' explanation for the increased throw is wrong.
About aspheric lenses: Lenses have a focal length f and a diameter D, both can vary widely independently, but there's a technical limit: The ratio between f and D, i.e. f/D (also called F/#), can not get below about 0.7 with a single lens. The f/D ratio is determined by the lens shape/curvature: The more curved, the smaller the f/D ratio.
That also means: The bigger the lens, the longer the smallest possible focal length.
The LED sits in the focus of the lens, and that distance determines the image size: The longer the focal length, the smaller the image. Thus bigger lenses must have a smaller image than smaller lenses (assuming both have the same shape i.e. f/D ratio).
