Treatise on Cree xr-e rings.

[qwertyydude]

It has come to my attention that it seems there's a misunderstanding on this site as to the cause and solution to the infamous Cree xr-e rings. Some say it is simply a matter of mis-matched optics to the xr-e. I say it is a fundamental flaw in the design of the led. I in no way hate the xr-e, and in fact appreciate the output and efficiency, but I will admit it has flaws that are difficult to overcome without losing either efficiency, throw or a smooth spill.

The consensus here is that the metal ring around the led causes the rings. The question is how it does so. A prevailing theory is that a reflector simply needs to cover just the dome of the led without exposing the metal ring and this will solve it. My experience is that it does not. The problem I see is that the metal ring causes reflections inside the dome of the led that project outward. The way most lights solve this is to use orange peel reflectors, I view this as a band-aid. Orange peel does hide the rings but it also defocuses the hotspot, which may or may not be seen as an advantage, but it also reduces throw and decreases OTF lumens due to increased incidental light hitting the flashlight lens, more orange peel=more losses. I view orange peel therefore as a crutch for the xr-e and also challenge anyone to find a light with an xr-e with a SMO reflector that doesn't have cree rings. I think that will be a good challenge as I've yet to run into one.

Any smooth parabolic optic will simply project whatever is in its focus and as you can see, the xr-e it emits bands of light and dark light coming out the sides of the xr-e. This will project as rings. The advantage is that most of the light is projected outward so aspherics work very well with the xr-e since little light besides the rings is projected out the sides of the led.

The xp-g and xp-e led on the other hand project a well focused hotspot AND smooth spill, two previously opposed features that flashlight makers had to pick one or the other. The reason being that they emit a smooth lambertian light pattern which is easy for any parabolic optic to focus and not project any artifacts. These emit more light out the side and are not good choices for aspherics due to light wastage out the side of the led.

 

[DM51]

Interesting topic! Moving it to LED.

 

[selfbuilt]

Nice pics. :kewlpics:

Shows the difference of bare emitters quite nicely.

 

[Black Rose]

I grabbed my Fenix L0D Q4 to see if a covered metal ring eliminates the Cree rings (the reflector completely covers the metal ring).

The rings are still there, though not as pronounced as with my P60 drop-ins where the opening of the reflector surrounds the ring instead of covering it.

That first pic really shows what the ring does to the light pattern.
Almost looks like light is coming from under the ring as well as the dome.

 

[kaichu dento]

The consensus here is that the metal ring around the led causes the rings. The question is how it does so. A prevailing theory is that a reflector simply needs to cover just the dome of the led without exposing the metal ring and this will solve it. My experience is that it does not. The problem I see is that the metal ring causes reflections inside the dome of the led that project outward. The way most lights solve this is to use orange peel reflectors, I view this as a band-aid. Orange peel does hide the rings but it also defocuses the hotspot, which may or may not be seen as an advantage, but it also reduces throw and decreases OTF lumens due to increased incidental light hitting the flashlight lens, more orange peel=more losses. I view orange peel therefore as a crutch for the xr-e and also challenge anyone to find a light with an xr-e with a SMO reflector that doesn't have cree rings. I think that will be a good challenge as I've yet to run into one.

Maybe the orange peel is a crutch for the XR-E, but the fact remains that by matching the right reflector to the emitter you'll have no rings. In your next pictures get a Draco or Nitecore EZ series and you'll see. You may value throw over other types of beam pattern but some of us prefer a floodier beam.

 

[spencer]

Very nice and informative. Thank you.

 

[qwertyydude]

http://www.candlepowerforums.com/vb/showthread.php?t=233122

I agree not all of us want throwers, me included. In fact that is why the xp-g is leaps and bound ahead of the xr-e is it's spectacularly smooth flood, but not having to sacrifice a smooth spill to get throw.

Here is a review of the Nitecore EZ-AA, if you look at the -2 stop underexposed you'll see a dark donut around the corona there, they may not be as obvious as a P60 but if they are visible to the camera they're usually more obvious in real life. Perhaps this reflector wasn't properly "matched" to the led but if the xr-e is so sensitive to reflector variances then it's still not an ideal emitter when it comes to beam quality especially when I have several cheap xp-g modules that provide smoother beams whether flood or throw. But I won't apologize for nitpicking the dark donut if people are willing to buy Malkoff modules at a premium because they're annoyed with the slightly darker and cooler centers of the hotspot on most led's which is hardly even visible even when photographed.

To me the xp-g's versatility and smooth beam trumps the xr-e's ability to throw better. Yes I said it the xp-g floods better and I appreciate that feature because it floods smoothly with SMO reflectors which increases OTF lumens. I don't just want throwers but the xp-g floods better and throws almost as well as the xr-e not to mention is more electrically efficient and due to not needing orange peel will be more OTF efficient. So that in my opinion means the xp-g is the better option for smooth floods and mid to long distance throw. This relegate the xr-e only to extreme throw which is obviously not what the Nitecore EZ was meant for. So now that the xp-g is here and floods better there is no need to even defend xr-e's for flood.

All in all my challenge still stands unanswered, that you can not find an SMO xr-e without rings. Until then no matter how matched the reflector is you can not get rid of rings without the orange peel.

 

[kaichu dento]

I'm not sure what you're trying to prove because with the exception of mule type lights we don't tend to use emitters without the benefit of a reflector and it's how the light performs as a package that really matters. It doesn't matter to me at all how the light is made into a functional tool, I only care that it works.

I doubt there are many people who prefer the XR-E over the XP-G, and I'm most assuredly an XP-G and P4 fan, but the XR-E does fine in a number of lights and is still popular enough that it's unlikely to go out of production any time soon.

I just went in the other room and after once again examining the beam patterns from my Draco and EZAAw found nothing to be unhappy about. It's very possible that I would like both lights better if they had XP-G's in place, but I don't really care to think about it because the reflector matching makes them excel at what I want from them.

My XR-E Draco is one of the nicest 24/7/365 lights ever and my XR-E EZAAw is one of the slimmest and lightest throwy lights that can be carried EDC.

 

[uplite]

I doubt there are many people who prefer the XR-E over the XP-G

The XR-E beats the XP-G in terms of surface brightness, and therefore throw. That's why the farthest-throwing flashlights on this forum still use the XR-E.

I own two XR-E dive lights. They both have ringy beams. I haven't disassembled them to investigate why, but it makes perfect sense that this comes from the metal ring around the primary optic.

Why is the ring there? I would guess it helped lighting manufacturers to center the LED before they made their own jigs. And perhaps it helped with heatsinking the less-efficient low-bin XR-E's early in their manufacturing lifetime.

-Jeff

 

[kaichu dento]

The XR-E beats the XP-G in terms of surface brightness, and therefore throw. That's why the farthest-throwing flashlights on this forum still use the XR-E.

I own two XR-E dive lights. They both have ringy beams. I haven't disassembled them to investigate why, but it makes perfect sense that this comes from the metal ring around the primary optic.

Not being a throw fan I haven't followed the reasoning but was certainly aware that XR-E's were still highly favored by those in search of more throw.

Again the lights that you have which show the rings in their beams have mismatched reflectors and if the rings bothered you enough you could eliminate them by talking to others about what reflector would be a better match.

I understand that the rings are coming from the metal ring, but with the right reflector, that problem becomes negated.

The Draco, which I keep bringing up, has it's XR-E matched up to a reflector that McGizmo helped Jonathan pick for it and even as tiny as it is, nothing but beautiful flood with a very XP-G-like hotspot, that is wide and blends smoothly with the rest of the beam.

 

[saabluster]

The consensus here is that the metal ring around the led causes the rings. The question is how it does so. A prevailing theory is that a reflector simply needs to cover just the dome of the led without exposing the metal ring and this will solve it. My experience is that it does not. The problem I see is that the metal ring causes reflections inside the dome of the led that project outward.

The "prevailing theory" is as you have noted clearly wrong. This is due to people misunderstanding what is meant when it is said the metal ring is causing the artifacts. They think it is coming from the outer surface not the inner part that is immersed in the optical gel. The inside of that metal ring looks just like a tiny reflector and aids in light extraction. The ring itself is necessary as with the optical system used there must be something to mount the glass dome on.

 

[qwertyydude]

I don't think the ring is entirely necessary though because as far as surface brightness is concerned the R2 xr-e is where it tops out but the xp-e using the same die manages an R3 bin. I think with the xr-e Cree was intending it to be more for home lighting than flashlights so they wanted most of the light in a narrower cone so they wouldn't need fancy reflectors to make xr-e downlights. A pure dome won't block as much light but a lot will go out the sides and that would be inefficient in your run of the mill nonreflectored led downlight

 

[Th232]

I'd agree with that, pages 6 and 7 of Cree's luminaire design guide go into that exact issue, even down to comparing the efficiency and distribution curves for a CFL fixture vs an XR-E.

 

[MikeAusC]

Two weeks ago I used the worklights that Police Rescue & Bomb Disposal Squad have - they're fitted with XRE's, don't use an OP reflector - but there's not a trace of rings !

They use 24 XRE's with just a flat white reflector

http://www.peli.com/en/lights/9430-remote-area-lighting-system

 

[SemiMan]

I believe the XP-E has higher surface brightness than the XRE. The XP-E advantage for throw only comes in if one is using aspheric optics due to the narrower cone of light from the XR-E.

About XRE being in R2 and XPE in R3, note that they do not use the same die. The die is different. If you don't believe me, go to Cree's output simulation tool and compare the lumens output of the XRE and XPE in the same bin at different currents. The XPE has much less current droop.

Semiman

 

[qwertyydude]

Semiman I belive you have your led's confused. The xr-e is a known good thrower when coupled with aspherics specifically because they concentrate more light out in a narrower cone. And as for current droop they both seem to have about a 170% light output at 700 ma so it's still a draw. The big problem is that also the xr-e data sheet stops at Q5 so the data might be off because the xr-e's data sheet is based still using the ez1000 die vs the xp-e which is a newer led using the ez900 die and R3 is listed on the data sheet. So conclusion can not be concretely drawn using different data sets unless the data sheet includes the newer xr-e R2 info.

 

[SemiMan]

Semiman I belive you have your led's confused. The xr-e is a known good thrower when coupled with aspherics specifically because they concentrate more light out in a narrower cone. And as for current droop they both seem to have about a 170% light output at 700 ma so it's still a draw. The big problem is that also the xr-e data sheet stops at Q5 so the data might be off because the xr-e's data sheet is based still using the ez1000 die vs the xp-e which is a newer led using the ez900 die and R3 is listed on the data sheet. So conclusion can not be concretely drawn using different data sets unless the data sheet includes the newer xr-e R2 info.

Doh.... I started off not confused ... ... The XPE has higher surface brightness.... and then I meant to type XRE for aspheric but my mind and my fingers lost touch with each other.

I think we are both confused as you typed "I don't think the ring is entirely necessary though because as far as surface brightness is concerned the R2 xr-e is where it tops out but the xp-e using the same die manages an R3 bin." You just clarified that they use a different die which was a point I was trying to make.

I typically see about 7-10% more at 700mA on XPE in an integrating sphere on metal core versus XRE in the same bin. Cree's product characterization tool has a smaller but still noticeable difference. The additional difference in my results is likely binning/measurement differences either at my end or their end. I have a reasonably large sample size, but it does not take much to throw it off.

Semiman

 

[Th232]

I think we are both confused as you typed "I don't think the ring is entirely necessary though because as far as surface brightness is concerned the R2 xr-e is where it tops out but the xp-e using the same die manages an R3 bin." You just clarified that they use a different die which was a point I was trying to make.

Just a small bit here, some XR-Es also have the new EZ900 die, here's one thread with some results, and what I think was the original thread about it.

 

[qwertyydude]

It's possible too there are more internal losses in the xr-e making the same die achieve only R2 binning in the xr-e whereas it makes R3 in the xp-e. Any time you have reflectors and lenses in an optical system it means there are going light losses. The xr-e has two things against it's light output ability, the ring which could block some light and is not a high polish reflector, then you have the silicone/glass barrier making for diffractive and internal reflection losses. The only barrier to light output in the xp-e is a single material silicone dome so theoretically there would be less losses.

 

[saabluster]

I believe the XP-E has higher surface brightness than the XRE. The XP-E advantage for throw only comes in if one is using aspheric optics due to the narrower cone of light from the XR-E.

About XRE being in R2 and XPE in R3, note that they do not use the same die. The die is different. If you don't believe me, go to Cree's output simulation tool and compare the lumens output of the XRE and XPE in the same bin at different currents. The XPE has much less current droop.

Semiman

I believe the XR-E has higher surface brightness than the XP-E. That is the reason it works better in an aspheric setup not the fact that it has a narrower cone of light. Lensing cannot increase surface brightness.

The XR-E and XP-E do in fact use the same die. I have also found that simulation tool to be faulty on more than one occasion. Take it with a huge grain of salt.

It's possible too there are more internal losses in the xr-e making the same die achieve only R2 binning in the xr-e whereas it makes R3 in the xp-e. Any time you have reflectors and lenses in an optical system it means there are going light losses. The xr-e has two things against it's light output ability, the ring which could block some light and is not a high polish reflector,

This is exactly the conclusion I have come to. The same die perform better in the XR-E package due to the better thermal transfer but since it sits down in the package there are significant losses to the ring. That's why it performs better for a lens but the spec-sheet shows less total output.

then you have the silicone/glass barrier making for diffractive and internal reflection losses. The only barrier to light output in the xp-e is a single material silicone dome so theoretically there would be less losses.

This bit is unclear as we do not have enough info to reach a proper conclusion. The approach in the XR-E seems like a better optical foundation as far as preserving the brightness of the die/light extraction(even if only in one direction). The glass is of higher clarity and they use three different materials of varying indices to prevent a large gap in RI which would cause internal reflection. I believe the XP-E uses only two.