I forgot to add, it looks like the standard ANSI C78.377-2008 is old; the current revision is ANSI C78.377-2017. Interestingly, within the current standard, ANSI does “define” recommended light as being as low as 2200K as seen within this .pdf. You can also read on 16 of 51 on this .pdf from Design Lights Consortium that ANSI defines 2200K, and that the 2015 standard is out of date.
That looks pretty official, updated to 2200K to 6500K. Again, maybe doesn't mean much but at least they are defining tolerances, so for manufacturers to claim compliance, there's some well defined spec's to adhere to.
I think we should update the flashlight wiki. It looks like they went through several revisions, and that 2008 standard is showing its age. Looks like they added 2200K and 2500K in C78.377-2015. You can read on page. 4 of this .pdf from DOE.
I used to have edit/update access to the wiki - haven't tried it though in years.
Yes. As I was saying, the concept of “white light” varies quite a bit with personal preference and perception. There are standards which set guidelines for the most common preferences, but they are not the Absolute Authority On All Things White, so we shouldn’t get too hung up on them. It’s just a recommendation… for someone who was asking for a recommendation.
It’s like saying a zero or slightly negative duv is a safe bet, but a positive duv is not… because most of the population prefers lighting which is on or slightly below the blackbody line. So we generally try to get “A” or “D” (below BBL) tints instead of “B” or “C” (above BBL), since it results in fewer people being unhappy.
That doesn’t mean the less-popular flavors shouldn’t be available… just that they’re an uncommon preference.
In general, LEDs in the neighborhood of ~2200K are warm enough that they’re outside the mainstream concept of “white light”. It’s often described as being yellow or orange like fire, instead of being white like the sun, so it typically falls into a different category. For example:
- Cree doesn’t offer 2200K for most of its LEDs, presumably because the demand isn’t high enough to bother. In product lines where it is offered, like XP-E2, that range is segmented off into its own specialty product class, separate from its mainstream white offerings.
- Nichia categorizes its LEDs into “White”, “Warm White”, or “Candle Color”. The candle color is specified as 2580K to 1870K.
Some people really like it… especially for electric candles and late-night bathroom runs. It’s a great option for that sort of thing. But it’s not typically considered useful for general-purpose white lighting. So it’s more of a niche product for specialty items and people with uncommon taste.
In case it provides context, a while back, we tried to measure people’s lighting preferences. There was a poll where people could rank their favorite color temperatures and LEDs, to help figure out which LEDs to offer, in which CCTs, and how much of each type to produce. Here are the results, graphed as a heat map:
For anyone interested in a below 2700K light source, I purchased this light bulb after SKV89 tested it:
Waveform Lighting Lux24™ Circadian LED Bulb - 2400K 95 CRI E26 A19 LED Filament Bulb
According to SKV89’s test it is roughly 2350K, below BBL, CRI more than 95, R9 more than 80. I like it a lot. Colors look perfectly natural to me at night. If you like it, you will probably like E21A 2200K.
Absolutely.
And if you spend a lot of time at a certain color temperature outside and go fast inside at a different color temperature things can look awful for several seconds.
Flashlights are really not natural light, bad CRI, variable CCT from hotspot to spill (CREE is the master of wide CCT transition from hotspot to spill, depends on the reflector and optic of course), trained expectations, involuntary expectations, all also play a role.
Even side by side comparisons of 2 different tints cause the brain to look for color opposite (red-green, orange-blue, yellow-purple), if you have something warm that dominates with red the cool one will have the tendency side by side to look more green than it is.
If anyone studied serious works on how brain perceives color by whatever other color are around, you know things can be very tricky, not only that but not everyone has the same color depth perception as the other one, it sounds harsh but it is something unaccepted in society, the same how hearing depth varies between people and it is not really accepted.
I think that’s where the confusion might have happened. I can see limited polls being used to determine preference, but I certainly would be wary of limited polls or old recommendations to determine what should be considered the scientific definition of “white light.” This report from DOE showed people preferred light with a CCT below 3000K for outdoor general lighting as a counterpoint to the poll you mentioned.
You might want to double check that…XP-G3 in 2200K. or the CMT 2890. which shows 2200K is being developed for high powered COB platforms. The manufacturer might leave a small gap between 2700K and 2200K on the chromacity charts, where there is a gap in their product line, but you can see 2200K listed with the rest of the white light emitters.
Markets take time to mature, and super low CCT high efficacy LEDs are still being developed. But to that end. in the world of 2200K.
“As customer requests for 2200K CCT increase, Evluma releases new warm color temperature option for OmniMax an LED retrofit for decorative streetlights… We began building 2200K CCT OmniMax as a custom order item,” stated Evluma’s Marketing Director Cathleen Shattuck. “We’ve seen an increase in demand for a wider range of warm CCT decorative streetlights. We are happy to be releasing 2200K as a standard option.” The 2200K model falls in between the 2000K and 2700K models offering greater CCT refinement for architectural lighting designers and municipal lighting specifiers.”
Hi everyone, I’m new to this forum and placed my first order with Hank. I was wondering what the turnaround time to shipment might be? I ordered a Ti/Cu KR4 E21A 4500K over the weekend. Are the lights usually ready to go or are they hand built to order? TIA for your help!
Yesterday I received my KR1 from intl-outdoor. I asked Hank for a custom led, a 4000K 95CRI SST-20 that was not listed as an option, and he was so kind to make it for me. My plan was to slice the led to bring the hotspot to around 3500K and close to the BBL, and that is what I managed to do this afternoon.
The results are:
Hotspot tint at 350mA: 3510K, 98CRI, R9=91, duv=+0.0006
Hotspot tint at max : 3660K, 95CRI, R9=78, duv=–0.0033
Output at max (30Q battery, 30 seconds) 610 lumen
Throw at max (30Q battery, 30 seconds) 58 kcd
The output is very stable btw, I’m not sure what driver is used but I assume the 5A current-regulated driver that is also used for the 1mm2 White Flat? Can’t measure current easily myself in this light.
The beam is very good, with the focus perfect. Before slicing the hotspot had some slight artifacts (the focusing of the KR1 is not entirely correct for a domed led), that were completely gone after slicing.
Thanks Hank for getting the custom led done! I love the result!
Wow I cannot believe someone else also asked for the SST20 in their KR1! I even got the same body color. Twinsies!
I’m surprised the tint is still duv=+0.0006 with sliced dome. I thought Hank is using the FA3 tint. The FA3 tint should be about that duv without slice. Did you measure tint before the slice?
The previous tests were integrated or in TIR and still above BBL. SMO reflector will be worse than that even with respect to Duv so I wouldnt plan on it.
SKV89, at 350ma though? Never seen an SST20 thats close to BBL @ 350ma.
–0.0033 at 5A is nice. Got to start slicing some SST20s myself, i love the throw, hate the green.
Care you share technique? Ive been quite happy with my sanding skills on the LH351D recently. Turns a meh LED into something unique and quite amazing. I understand its far from prefered by most though.
Ive got an ROT66 i would love to slice the 12 SST20s, seems like quite the job! Get some great throw and nice tint!
On my Fireflies E07 with FA3 tint, I measure –0.0002 at 4018K at about 30 lumens
Fair enough. Maybe i havent seen the FA3. My ROT66 is only a few months old, grrr.
Ive had some single emitter light from Sofirn that were pretty good, ive actually got 7-8 SST20 4000k emitters on order from them, fingers crossed they are FA3.
A TIR keeps the duv lower than a reflector. In my experience you must start with a very rosy led to keep the tint under the BBL when measured in the hotspot of a reflector. And the purple AR coating likely does not help either, works as a +green filter.
I did measure the tint before slicing but forgot to save it as a file, CCT was 3910K and the duv in the hotspot of the KR1 was clearly above the BBL, like +0.0025-ish
Nice to hear I’m not the only one 
. A throwy high CRI light in (close to) EDC size was missing in my collection. It makes walking in the dark outside in nature following a trail so much more pleasant than a cold light.
I use a series of 4 steel washers going from thick to thin and slice the dome bit by bit with a (oldschool Wilkinson) razorblade. The washers fit between the solder blobs of a 20mm board but for a 16mm board I need to remove all the solder to fit the washers flat on the board. Brass washers would actually be better for keeping the blade sharp. The thinnest washer was sanded to exactly the right thickness to get the slice right above the phosfor layer. The led must be reflowed flat on the board and the angle that you keep the blade at during the slice makes a difference too. I actually made the washer set for the LH351D (I have several lights with shaved LH351D leds that have wonderful tint and CRI) but it was almost correct for the SST-20 too (had to keep the blade flatter than with the LH351D during the last slice to clear the bond wires of the SST-20). I find that the remaining surface need not be perfect for a great result, it is not critical and I do not bother sanding it anymore.
It would be quite the challenge to slice 12 of them without chopping any bond wires. Requires closely monitoring the process!
Great minds think alike. I was actually surprised at how compact the KR1 is. I was worried it might be too much larger than the D4V2. Well actually it is larger but it’s manageable for me.
Hello
My KR is super flashlight, I love it very much.
Only one question. Is it normal that flashlight starts working when tail cap is srew last half of mm.
Is this normal?
Thank you
Kind regards
Yes that is normal.
Because of the tube in tube system.