How is a google search supposed to change the fact that the light coming out the central angles of the LED is NOT what forms the hotspot? Did you even read the thread?
Believe what you want to believe. Yes, I’ve read this thread. Did you? Look at the pictures that djozz posted. The picture on the right tells me that the outer part of the reflector gives the hotspot while the inner part gives the corona.
If you do a Google search you’ll see that there are thread after thread on CPF where people are backing up the claims of better throw from the XR-E with higher lux numbers. You don’t have to believe it. Heck, until about two months ago, I thought that an XP-E R4 would out throw an XR-E R2 despite everyone telling me otherwise. Since then I’ve seen the light.
So according to this theory how is the xr-e a better thrower given it focuses more light on the spill and less light on the reflector?
Also, if you look at the experiment, the “corona” is just a less magnified version of the hotspot. This becomes more obvious if you try the experiment with a smaller hole (basically approx a pinhole camera).
XP-E’s aren’t necessarily ez900 size dies. In any case, dedomed emitters seem to throw better, and those have no lens to focus anything.
Correct me if I’m wrong but I think in custom builds XR-E’s seem to handle more current all else being equal. I’m talking about overdriving here.
Also, many lights built for throw with XR-E are using an aspheric and the XP-E is usually used in reflectors so it that scenario a XR-E would throw further (all else being equal) just because aspherics of similar size throw better than reflectors.
My overall impression was that the XP-E was more efficient at lower currents. I also don’t see a R4 for the XP-E in the Cree datasheets. R3 is as high as I see.
The Ez900 die is significantly smaller than ez100 size die so it’s not even comparing the same thing.
19% difference in surface area
The XP-E is also available in ez900 vs 1000 isn’t it?
To make it easier to compare:
Die ratio:
73/92
Ans = 0.793478261
R2/R4 ratio:
251/286
Ans = 0.877622378
Cree doesn’t seem to differentiate even in their own lit. The only xp-e teardown I can find is of ez1000. I would assume new bins debut with larger die if that’s an option since it’s easier to make brighter given a bin (ie more high output ones out of a wafer).
I also remember reading that XP-Es are EZ1000, but that information might be outdated.
PCC: The corona is formed by the inner part of the reflector, while the spot is formed by the whole reflector. The corona is less bright, because only a part of the reflector contributes to it.
BTW, the DEFT EDC uses an XP-C…
I thought I was going to have to say that, until I read this far.
That is true. And also approximately true for any continuous shaped mirror or lens that throws a small spot. What has a varying focal length, and therefore a mixed magnification, is a Fresnel lens. Each segment is a section of a different continuous lens or mirror and therefore has a different focal length. So, instead of giving a clear image of the emitter a Fresnel lens gives a mixed magnification.
So there are two reasons why Fresnel lenses are not much used for telescopes. One, based on geometric optics, is that it has mixed magnification. The other, based on wave optics, is that the aperture that determines the diffraction limit of the resolution is the segment size, not the device size.
Saying that a parabolic mirror has ONE focal length is... an approximation only valid for paraxial optics, which is what is taught in school and even in many basic optics lectures. It's acceptable if the focal length is (much) bigger than the reflector diameter (and even for 'recoil' reflector flashlights). But the typical flashlight reflector is way beyond the scope of that approximation.
The outer parts of the reflector have a significantly bigger focal length than the inner parts.
Explanation: Each patch of the reflector projects an image into infinity, thus the LED must be in each patch's focus, thus the focal length of each patch is the distance to the LED - which varies a lot along the reflector. That also implies that those patches project images of the LED in different sizes; inner areas project bigger images, outer project smaller images (longer focal length).
Experiment: Take a flashlight with a big, but not that deep reflector, and some cardboard. Make a little hole (~2mm) into the cardboard, place it directly in front of the flashlight and move the hole around. Watch the image created by the part of the reflector under the hole. Areas near the center project a bigger image (implying a shorter FL), those near the edge project a smaller image (implying a longer FL).
Dr. Jones, I like your explanations.
What is really needed is lux at 1 meter and 350 ma for various emitters, using 26 mm/34 mm 41 mm smo relectors. Lux (and diameter at 1 meter) of hotspot and typical corona, at varying reflector depths. Would not this, over confusing theory, better clarify and allow us to better choose a reflector?
A parabola is the locus of points with equal sum of their distances from a point and a line, so all rays emerging from the focal point travel equal distance to a line perpendicular to the axis and so equal distance to a far away point in front of the reflector. So it is only the size of the source that needs to be small to get a perfectly focused image. The reflector diameter can be arbitrarily large, as long as it is small compared to the distance to the target. I would describe a system that generates a perfect image as having a single focal length.
Good text book recitation!
I have problem with the work “only”. In short, an xml can out throw an xpg (at same current), if the reflector is well chosen. My Fenix TK35 xml slightly out throws my Fenix xp-g at one watt or two watts. The TK35 has a reflector 38mm by 30 deep (or close to that), while the xp-g uses a typical 26 mm (both smooth reflectors and both drivers are comparable 92% efficient buck drivers.)
Not much inherent meaning here.
Anyway…
Obviously, a 35 KM wide reflector would be like a bare emitter, if the depth is not right: most light would miss the reflector and end up in the corona. Same with 3 meter reflector. So, I think depth-probably at some sweet ration to width-must have to do with throw.
I just installed a 30 mm by 20 deep that throws horribly (xml), while my build last week 28 mm by 24-28 mm (more or less, deeper anyway) throws okay. It ain’t in the Fenix TK 35 league, though.
There should be tangible numbers that come with these reflectors that would tell me: hey my 150 lumen xml will throw 4k candela (8 degree hotspot) and have a 100 degree 100 candela corona.
Short of this, I/we need tangible ratio, rule of thumb. Though, dealing with parabolas, it may not be a linear multiplier ratio.
That is from college calculus. I don’t know when I first understood the relevance to optics.
More later.
This is awesome, but a little bewildering with only one read. One guy put that it is mind blowing; which really means it did not blow his mind, but rather went over his head. I need to go use my lights that I made for work now, no time to study this and break it down. It would also be far easier if several examples were done. Examples teach faster and more completely than explanations (which always are in danger of relying on logic jumps and vague definitions).
I am wondering what rules of thumb from this post that we can extract. We already knew, led intensity matters (efficiency over surface area), and we know xml is less than xp-g, which I believe is less than xp-c, etc. From one read, it looked to me that surface area matters (so a deep reflector should add area). I am not sure how the small diameter is measured (is this led opening or some other point). If it is the led opening, I don’t see the relevance. I didn’t glean too much else which would offer any rule that we could use. Also, I am not sure the klux that they are measuring. I bet it is the lux at focal point of a few centimeters in front of the led optic. (I forget the implied current.) It simply is not the lux at one meter.
The inner and outer radius part is strange because it doesn't directly account for the height of the revolved parabola. This must be an approximation that's saying the surface area of a revolved parabola is the same as area of the big opening minus the area of the small opening. Btw, I read the small area as being the outer diameter of the flat area, not just the LED hole, although in some reflectors that is the same thing. Unfortunately that approximation isn't even close to correct according to example 2.
Perhaps it's not literally the surface area that matters. We know that if it's a parabola, then every point of light emanating from its vertex will be parallel after it hits the surface of the parabola. Given that, approximating as a circle could be sufficient. I'm still bugged by depth though. A deeper reflector will throw further, all else equal. So maybe the formula above is only for x^2 parabolas.
I'm probably very wrong. I took two semesters to barely get through calc I, and calc II was a disaster.
Unfortunately I didn't link or quote the other thread he mentioned. This is one of the very few times I wish the other forum was working. That probably explains why depth isn't directly factored into this formula.