Changing the focus of the LED might have some part to play in it, but it’s pretty well documented that reducing the apparent size of the LED die increases throw and reduces the size of the hotspot by varying degrees. Other than being warmer and smaller the hotspot should be just as uniform as before, though this is dependent on the reflector.
Im really impressed with the polished dome results, Thanks to MattSPL!
If only Cree released a factory dedomed XML, now that would be sweet. Affordable throwers for everyone!
What a fun mod!
After reading the OP, I realized this is a task I could probably pull off, but didn’t want to make an attempt on any of the XM-L lights I currently have because I like them too much to ruin an emitter and have to wait for a replacement.
Then I remembered I still have a generic 5-mode XM-L T6 drop-in that came with my first P60 - an UF wf-502b from amazon. This drop-in noticeably under-performs compared to the MF UF 3-mode T6, and I suspect it’s because it is under-driven. Either way, seemed like a good candidate for my first intrusive mod in the form of de-doming that LED.
I do not have a razor blade, but I do have a very small Kershaw pocket knife with a VERY thin blade. I sharpened it up a bit, and quite easily sliced off about 50% of the silicone dome. It was not a perfectly straight cut, and seemed to leave some artifacts and ridges, etc. But I tried it out anyway, re-assembled the drop-in and put it in my 501a with a 18350.
The hot spot, while not particularly pretty or even, was noticeably smaller and seemed hotter too. Compared to a MF UF 3-mode T6 in my L2M with the same TFF 18350, it now was ALMOST as bright as the higher-driven drop-in. NICE. However, like I said, the spot was kind of ugly, uneven, and even had some small rings around it.
Of course, given that this was yesterday, now after revisiting this thread and seeing several more posts and ideas thrown about, I decided to make further attempts to improve the beam on this emitter. I pulled out my dremel, and used the polishing point as pictured in one of the posts above, and carefully touched the remaining “dome” to the spinning head until it was flattened and smoothed out evenly. It left the surface frosted, so I switched the polishing head out for a soft felt disc, and went back at it… polished the silicone surface out til it appeared to be perfectly clear.
I reassembled the light and eyeballed the beam on white wall again, and yes - it has improved drastically. It is, as reported by others, a slightly warmer white than the unaltered UF 3-mode T6, but damn if it isn’t somehow, at least to the naked eye, matching the output, even under-driven.
Last night before today’s polishing session, I played around with this light outside, and couldn’t tell for sure how much better the throw was, but now I’m pretty sure it’ll be more obvious. Again - I had only removed about 50% of the silicone originally, and I suspect there’s still a good 30% or so left after polishing. There is definitely less spill, and the spot is definitely hotter and a bit smaller. So I’m happy with the mod thus far - but does anyone think a 30% is still too much? Could I improve this even more? Should I even bother? 
At this point, I’m really tempted to de-dome one of my 3 MF UF 3-modes. With the results on this generic, I’ll bet the same on the hotter drop-in would be killer.
Yeah, i wonder why they dont? I suppose one poke of a finger and the LED could be ruined without a dome.
I’m contemplating having another polish of the XM-L to see if i can get it really shiny, but don’t want to end up going too far and ruining it :~
No it’s not. If that’s the case, then a reduction lens can make a supreme thrower, rather than larger reflector.
It makes no sense that it would be warmer in aggregate. It’s the exact same light coming out of the emitter and passed the phosphor layer. The lens itself isn’t tinted. Perhaps the warmer light coming out at more extreme angles that was directed to the spill by the lens is now making it into the hotspot.
Removing the dome reduces appearant die size, and so increases lumen concentration at the focal point of the reflector. The brighter you can get a point of light at the focal point of a reflector determines the throw from what I know. (the furthest throwing led flashlights in the world are de-domed)
As for the light being warmer after de-doming, I am no expert here, but it seems to be a general trend when de-doming most high power LED’s, including cree and luminus. I still don’t know the mechanism as to why exactly the light yellows after dedoming, but it is a well documented occurance. It may be possible that warm light from the spill is now making its way into the hotpot through the reflector, but that wouldnt explain why the warm tinting still occurs in de-domed lights that are fitted with asphericals. The dome is actually more complicated than it seems, I remember reading somewhere that it’s actually composed of multiple layers of material that have a different refractive index at the top of the dome compared with the bottom, in order to help light transition from the emitter to air and not be reflected back and lost.
JaffoAZ,
Well you could always test it out and report to us if theres any improvement when taking off more material .
If you plan on doing so, I highly suggest that you take beamshots of the before and after, which is something
that I neglected to do and im kicking myself for it.
DrJones (who’s posted in this thread) figured differently a while back via this phenomenon: Etendue - Wikipedia. It turns out the angle of the light is important, too.
Has anyone ever measured color with respect to angle of emission? Unless it’s magic, the frequency of the light is conserved.
it’s not the exact same light coming out of the emitter, that’s the point. The dome not only provides a wider, more even scattering of light, it also reflects some of that light back onto the phosphor via internal refraction within the dome. The phosphor is then excited to a slightly higher, aggregate, degree, which increases the wavelength coming out i.e., a shift towards the blue end of the visible spectrum. No dome = no extra excitation = shift towards the red end of the visible light spectrum.
It may be worth reading up on this stuff more before slamming peoples observations with your gut feelings
(same applies to the smaller surface = smaller hotspot, but I don’t know enough about that to explain it properly)
No, the dome concentrates the light, not a “wider, more even scattering”. I also doubt there’s any internal refraction given its spherical shape and the fact that the emitter is not a reflective surface.
Thanks HAL, I agree. I was actually just thinking about getting some shots after realizing I have no good way of comparing what I’ve done so far.
I’ve not yet taken any beam shots before. So we shall see. Figured I’d do comparison shots between my altered emitter and an unaltered MF UF 3-mode, and then comparison shots of the altered beam spots before and after further layer removal. Hopefully, I don’t destroy the emitter.
now I remember why I usually stop following any threads you contribute to. I really have neither the time nor the inclination to spend smacking you round the head with the appropriate technical documents to overcome the insurmountable burden of proof you seem to require.
My apologies to all the great contributors to this thread and I hope that this in now way diminishes their efforts. I’ll now unsubscribe to avoid feeding the flames further.
I would highly recommend reading these “technical documents” yourself to gain some understanding of how lenses work at a basic level.
agenthex,
I’m not quite sure what you’re getting at here in terms of angle of emission? De-doming increases the angle of emission allowing more light to hit the reflector I agree. The size of the actual die also matters, for example an xml in a 20mm reflector cannot throw further than a xr-e in said reflector no matter how well designed it is. the dome magnifies the die making it behave as a larger die would. An example of this is the mc-e vs p7. The p7 always tended to have an edge on throw in identical mods involving both reflectors because the p7 has a larger radius dome and so the die has a lower apparent die size than the mc-e. Maybe there is some other reason for apparent die size having this effect, but practically it seems to work out pretty well.
And about the warm shift, I never said that the sum total of the LED’s output is affected tint wise. All I said was that the hotspot becomes slightly warm, which is true in just about every successful de-dome case I’ve read about.
I never talk in absolutes, these are just my opinions, im here to learn like everyone else.
HAL, I find it interesting, based on what you’ve said above about MORE light hitting the reflector after de-doming, that I am seeing less spill from mine. I was under the impression that the 120 degree dispersion of light from an XM-L was supposed to decrease to closer to more of the 90 degree angle of an XR-E after de-doming, and we get a smaller hotter spot as well, which seems evident in my new beam pattern. Wouldn’t more light hitting the reflector cause more spill?
Thanks again!
Yes, learning rocks. I love this stuff.
It’s not so much the size of the die that matters as the luminous output per unit area. The xml cannot throw further because its brightness is spread over a larger area.
The dome also focuses the light to go forward; a narrower angle of emission. IOW, helps make the resultant spot more focused since it was more focused to begin with, see link above.
Or it’s possible that simple parabolic reflector designs tend to favor the output pattern of a simple flat die, and the p7 output is closer to this. I don’t know, but at least that makes sense.
I agree there’s something odd going on, evident from plain observation. That’s why we should investigate why it does instead of taking lesser explanations at face value.
Do you have a link to the same effect in asphericals? They tend to be cooler tint for the same emitter, which would at least confirm that warmer light comes out further towards the sides.
The light hitting the reflector goes forward and forms the hotspot.
All flat surfaces emit with the same magnitude w/re: to angle: a Cosine curve assuming equidistant from emitter (more or less true at a distance).
The xp-g2 throws further with the same die as the xpg, and it has a narrower emission angle… discuss
Okay - I understand all flat surfaces will emit at the same angle. But from what I understand in this thread is that because the XM-L emitter and dome toss out light at the much wider 120 degrees, the reason we de-dome is to reduce that dispersion angle and focus more through the front. We lose lumens, but gain spot intensity. If light hitting the reflector is what forms the hot spot, how does adding an aspheric lens and blackening out the reflector give me a hot spot with no spill?
Maybe I’m confused. I thought I was kinda getting it before. And when I did my little de-doming job, I got less spill, and a smaller, tighter hot spot that appears to compete with a high-driven emitter in terms of brightness, and especially in throw. That’s why I de-domed. So if I’m wrong about my above assessment, I guess I’m maybe still confused about why it works. But I am glad it works.
An aspheric (or any lens in front) is doing the concentration instead of the reflector. You black out the reflector because the “focused” light it’s projecting out is messed up by the lens (which assumes a point source more or less, not collimated light).