The human brain excels at auto white balance. Differences that are obvious in side-by-side comparison are often undetectable in isolation.
Good to know that, thanks. I have a 219b and a couple of 219c lights. I enjoy their respective CCT, but I’d been wanting something more efficient and it sounds like the 519a fills that spot well. As for dome removal, I saw someone post about removing one in a triplet, leaving the other 2 intact and it balanced out the CCT nicely.
That is true, to a degree. I do have a couple of older lights that have either cold or greenish/yellow tints and it’s painfully obvious even without a compare. There is some zone of comparison as you start to get to a smaller delta on CCT that starts to become a wash. Still, I think once you get to know a particular emitter CCT and have a strong appreciation for it, you’re more apt to notice in usage.
When my eyes adjust to low lighting, my tint perception shifts to yellow-green quite a bit, so I need a significantly rosy emitter to balance it out. I think it's the Purkinje effect.
not a binning issue… tint shift is caused by flip chip design, it can be equalized with Tir optics, they blend the beam.
dedoming 519a does eliminate the tint shift in the corona of the hotspot.
At the cost of a smaller hotspot with irregular margins, and 13% lower output. Also lowers the CCT by about 1000K and lowers the DUV by about –0.0030…
sw45k DUV is about –0.0100
519a dedomed DUV is about –0.0040
as far as efficiency of 519a, depends what you compare it to:
examples in a V11r:
5000K LH351d 280 lumens
4500K 519a 260 lumens, is 7% less than LH351d
dedomed 4500K 519a 230 lumens, (becomes 3400K), and is 12% less than domed
Personally not a fan of dedomed 519a, imo it damages the quality of the beam. I do use 519a w dome ON, in V11r, because its large dome is more compatible with that reflector. dedomed 519a in V11r makes a shadow ring in the spill worse… similar to sw45k, but the sw45k is more pink and has a cleaner hotspot perimeter.
look at the beam profile w dome (the wall photo bottom right):
.
and without dome:
.
compare to the beam profile of sw45k:
.
still a shadow in the spill, but not as bad as dedomed 519a
the 519a dedomed is 18% brighter than sw45k. But I dont use Max output, so brightness is not a motivating factor, for me. Or if brightness is traded for runtime, 18% more battery life is not motivating, for me, since I use rechargeable batteries, I have unlimited runtime.
imo, efficiency is a red herring, a hold over from the days when people used disposable batteries and did runtime tests to get some idea how long their disposable battery might last.
I choose the LED that gives me the best beam profile, with a Tint I like.
I do not care for the 519a domed 4500K, too much “green” for my preferences. otoh, I like the 519a domed 5700K, its my goto for daylight white.
No. Flip chip design has nothing to do with tint shift. The flip chip is responsible only for the fact that the conductors do not fit to the outside of the crystal.
Tint shift is caused by a lens on the crystal - dome - that collects a lot of light from the ends of the crystal.
Domeless leds have no tint shift.
ok, got any links to support your opinion?
see this post:
and look at this XP-G3:
.
fwiw, the intergrated DUV is –0.0014, but that is the average across the beam
Shure, did you read what I wrote?
Dome on the led is the reason of tint shift.
Domeless leds don’t have tint shift - example any xpl-hi, xhp35-hi, xhp50.3 hi and so on.
And if you properly dedome even xhp50.2 with ‘slice and dice’ you can get rid of tint shift. Read about it.
Some of hicri leds with dome have not big tint shift but domeless leds doesn’t have any.
There are flat LEDs with significant tint shift like E21A (high CCT) and Optisolis, in both case it’s because the phosphor area is larger than the LED chip and the rays going trough that extra thickness of phosphors on the sides have a different color than the rest.
Thank you, totally agree. And, the E21a is a flip chip.
My impression is that flip chips have phosphor on the sides that emit a different color, so in a reflector, we get a corona that is a different color than the center.
here is a light that I modded to 519a (also a flip chip design). I saw green in the corona (more than the photo shows):
.
so I intalled a pebbled Tir and now the tint is the same across the beam:
.
Here is an sw45k (not a flip chip), in another Ti3, I like it, no need for a blending optic:
.
Having a larger area of phosphor is independent form the construction being flip chip or not, A dedomed or sliced 519A or LH351D barely have any tint shift even though they display a noticeable amount of it with the dome on, B35A is flat flip-chip and also have litle tint shift (except artefacts due to the quad dies) conversely Optisolis isn’t a flip chip construction yet displays large amount of tint shift.
Another example are the XHP 1st, 2nd and 3rd gen :
1st : lateral chip with bond wire, no large phosphor pour, domed, reasonable tint shift
2nd : flip-chip, phosphor pour on the whole package, domed, extreme tint shift
3rd : flip-chip,no large phosphor pour, domed, reasonable tint shift
3rd HI : flip-chip,no large phosphor pour, flat, lower tint shift (still some though)
The size of the dome seems to have an impact, though I’m not sure about it, small dome LEDs seems to have lower tint shift (SST-20, 219B)
Thank you very much for all the detailed info.
It does seem to make sense that a larger dome would spread the tint shift over a larger area of separation, to make the tint shift more obvious, even w non flip chips like LH351d. I do see a green corona:
.
it is also possible that the hotspot does not look green simply because of the overexposure, and in fact the whole beam is green 
Interesting… so while dedoming can mitigate tint shift, it won’t necessarily get rid of it altogether. In the end, if the degree isn’t much then it becomes a wash in practical use.
XP-G3 has been an LED of choice for RovyVon and it took them a long while to start branching out to Nichia and others. It was the ‘leash’ that kept me from buying certain models. I didn’t like the CCT rendition… and the tint shift is likely the culprit. Noticeable on close-range usage with simple & smooth surfaces.
So with flat flip-chip style emitters, is tint shift less noticeable with lower CCT than higher CCT? I’m a bit confused, because I saw this:
Anyway, “flip chip” is a nomenclature that was unfamiliar to me. Fascinating,, the premise of the chip flipped inside the LED resulting in 20% greater efficiency and double the lifespan! What LED’s are made with flip-chip design aside from E17, E21, 519a and B35AM (E21x4)?
Any without bond wires I think, LH351D, 219C, new XP-L, new XM-L2, XHP gen2/3 and many others.
I didn’t realize the flip-chip design has been around that long, going back 4 years? I thought the whole 219 series was before that; so 219c has it as well. But in terms of efficiency, isn’t a 519a is still appreciably more efficient than 219c?
Very interesting question! I also notice more tint shift on my 4500K 519As compared to other CCTs. My conjecture is that this particular color temp is right between cool white and warm white, and tint shift easily puts the spill into the cool category, and the hotspot into the warm category, which makes the difference seem very apparent. 4000-4500K is also kind of an awkward CCT range that tends to make things look greener than they should, I'm not sure why.
In my experience the amount of tint shift does not vary monotonically with CCT, positively or negatively. I will say, however, that at extremely low CCTs (3000K and below), the same amount of tint shift becomes less noticeable than at, say, 4500K.
I believe there is no correlation OR causation between flip chip design and amount of tint shift. The dominant factors are generally 1) whether the LED is domed and 2) whether the phosphor pour extends past the underlying blue die. An external factor is the choice of optic/reflector--shallow reflectors like S2+ tend to bring out tint shift, while TIRs generally do a good job of smoothing out tint variations.
The 519A is much more efficient than 219C. On paper you don't see much gain in lumens, but keep in mind that the 519A is 95+ CRI and 80+ R9, while the 219C is not! Gaining R9 (deep red) at the expense of green makes the lumen count go down quite a bit, but makes the light so much nicer.
not sure if you are asking about
Tint Shift Across the Beam, or
Tint Shift at Higher Output.
dont undestand what kind of Tint Shift you mean at Lower CCT?
imho, efficiency is the wrong question.
I have not seen any numbers… what percent do you consider meaningful?
Bear in mind the 219c has lower CRI R9.
The 519a is not more efficient than LH351d… I just like the higher R9 and lower DUV of 519a… efficiency is not a priority to me, I care about spectrum quality, not lumen quantity.
Efficiency favors greener tint and Lower CRI… neither of which I want.
It’s more efficient yeah, the die is larger and the thermal resistance lower, lower temperature means higher efficiency, the glass + silicone dome construction supposedly increases light extraction and there could be other improvements as well
Just to be super precise and pedantic: what you just described (presumably measured in lumens/watt) is actually efficacy, not efficiency. This appears to be a very common misuse of terminology within the flashlight world.
One nice way to distinguish them: efficacy measures how well a source produces visible light, and would thus have lumens (which is computed from the luminosity function of the human eye) in the numerator. Efficiency is always a unitless quantity: it measures the proportion of electrical power that has been converted into radiant power.
Given that the 519A still produces more lumens than the 219C despite having a R9580 spectrum (which is less efficacious than a 9050 spectrum), it is reasonable to conclude that the 519A is more efficient.
I can’t remember if it was Cree that introduced flip chip for emitters or not, but they were making them at least as far back as 2017 in their XD series, and in those they were able to mitigate wasted light and tint shift a little by putting tiny reflectors around the die, attempting to emulate a LES with one point instead of five (edges of the die). If memory serves the increase in output was around 15% or so, just in rescuing those lost wayward photons.
Flip chip as a concept goes much further back, in computer chips and maybe even transistors…certainly was happening in the 80s but probably years before that.
Thanks. Especially helpful to learn about how tint shift isn’t necessarily progressive by CCT, that it can be more prominent in certain ranges and depending upon domed vs. dedomed. Yes, I’d seen an ambiguous mention about efficiency, and I’m getting a better idea of it now. Of course, output efficiency has to be qualified because it’s conditional on duration and comparative CCT, given equivalent host/driver/amperage. “Luminous Efficacy” helps explain a lot too
Thanks. Larger die and comparatively lower thermal resistance sounds like a good prescription for better efficiency.
Tint shift across the beam or tint shift manifesting at different levels of output. Perhaps both are worth considering. I had read someone seeing tint shift manifesting in a dedomed 519a emitter rated 4000k~4500k, but not seeing it in 5000k~5500k. They didn’t qualify it by brightness setting, but that is of course an important consideration.
Wrong? Didn’t seem to be taken wrong by others here. Does luminous efficacy sound more appropriate? I don’t think efficiency is wrong generally speaking, but it should be qualified. The presumption is, given same power level being driven, does one LED deliver more lumens or nearly equal lumens but with less thermal buildup (and thus able to sustain the output longer before needing to step down)? From what I’m seeing in responses, it sounds like the 519a will achieve same lumens output as a 219c while drawing less on the battery and producing less thermal buildup, given the same host.