(Those emitters only has quality, when they are dedomed and excess phosphor around LES is removed, but I am not aware of any manufacturer willing to do that.)
Would be interesting, but 3×21700 could qualify for much more powerful designs as well. 40W could be good for a 3 × SST-20 95CRI thrower, or any other flashlight design with at least 18-20 × E21A.
For which optic this cadavre wold be adopted?
Additional one flash FL? No any 3000+lm can be efficiently dissipated in near soda can size. May be 5000lm in night cool weather.
If the optic would be floody,that even with 15k lm the real distance would be not amazing. And if optic would be throwy, it wouldbe uncomfortable in close distances.
The future is on FL with chageable light distribtion (not lens zoomer )))), with such light ability
I think you are correct. From what I read, the XHP50.2 and 70.2 have very high efficiency for lumen / watt, and the CRI is high. However it often has green tints.
Maybe the best idea is using new B35AM from Nichia but I have not seen it before in any flashlight.
Not really. CRI is not high for any of XHPxx.2 . Yes, they are efficient, but even the highest CRI class usually gets measured with R9 under 50 (I remember seeing it 39 for XP-G3, which has mostly identical phosphor technology).
Still, I would accept that for some cases, but without the changes I mentioned (which was implemented already by a few members here, most notably ZozzV6), tint shifts are so huge across the beam, that it shall be called CreeRainbowPuke™.
Just an observation. I picked up the Astrolux FT02S recently and it’s max output is with the Samsung 30T. That’s 3000 mah where as a triple 21700 could do the same with 3 Samsung 50G. So instead of triple capacity you now have 5 times the capacity with the same brightness.
No. Yuji are special. This one gets 60 lm/W. I don’t remember the exact numbers and don’t care to calculate again but even heavily underdriven it stays very inefficient.
F.e. Nichia NFCWL060B-V4 offers the same CRI with 135 lm/W. Even warm Optisolis is 110+ lm/W.
ADDED:
Nichia NVCWL024Z-V1 has 11 mm LES, so 21% larger area.
It makes almost 14000 lm with ~78 lm/W. CRI 95min as well.
At the rated current (1.2A) it gives 4980 lm with 117 lm/W.
My crude chart read gives it 146 lm/W and 2340 lm at 500 mA. Might be significantly incorrect but the max efficiency is much more than 117 lm/W.
Note that both Yuji and Nichia do their ratings when the LED is cool.
Since Simon over at Convoy Flashlights came out with an affordable triple 21700. Since he’s an active member of this forms. Since he listens to the wants and desires of his customers. I’m going to declare the BLF triple 21700 done. I just ordered the triple SST40 5000K. Thank you Simon for building affordable quality lights.
Make the body an empty can at 52mm or so. Keep the same contact ring on the head/driver. All parallel, so who really cares how they’re situated?
A cheap “speedloader” that’s got a cross or trefoil cross-section in the middle to keep the 4 18650s or 3 21700s upright, and a bottom-with-springs that would have the springs make contact with the cells on one end and with the body on the other end.
One for 3×21700 (shorter springs) and one for 4×18650 (longer springs).
Nb: I don’t mean a full cell-holder or carriage that’d make the assembly any wider, but just a disc-with-springs at the bottom and central spacer/form to just keep the cells upright. Kinda like a big-ass thumbtack but with cross/trefoil pointy-part.
Then you can have the same light take both configs, 3 or 4 cells.
Eg, if I’m remembering right, the SD01 with its 1pc body, has the cloverleaf-shaped insert-with-springs just drop in and snap-fit to stay, vs the Q8 where the cap unscrews and the PCB-with-springs is hard-mounted inside. The former is so much simpler and cheaper, and works just as well if not better (loses both threads and O-rings, as well as another potential for water-ingress).