Thanks a lot for sharing your thoughts about Sofirn and the often long and sometimes even tedious way to convince flashlight manufacturers. Regardless of any specific brand I will not give up convincing them to invest some extra efforts in driver development, well-chosen emitters and comprehensible and sophisticated user interfaces and not only following the war on lumens. Sofirn and WildTrail are two of those brands who strive to become better by listening to our suggestions but as you said, it sometimes takes more than a just a good advice to guide them into the right direction.
Concerning this very project:
I told Jacky to please avoid using FET drivers with a tactical/tacticool flashlight like the one we're talking about here. FET is nice for posing high lumen numbers with ugly-tintshift emitters like XHP50.2 3V 5.000K (3B). Instead, I would like him to use a boost driver with a 6V SST-70 but that might increase both efforts and costs for development at least by factor 2. So, this option has been ruled out, at least for this project. That's why we have to stick with 3V emitters. In a tactical flashlight I expect something more reliable than FET, something with constant output, preferably with the option to use 2x CR123A for emergencies or the always imminent zombie apocalypse. If I was lost deep down in a cavern and my life was at stake, I wouldn't want to rely on something that comes with unstable output und unpredictable runtimes. I don't know how much extra efforts it will take to use a suitable buck driver that will work both with 1x 18650 and 2x CR123A. It cannot be that difficult as Sofirn can also offer a cheap flashlight with a good buck driver, see SP31 V2.0. I also like the UI the SP31 V2.0 uses even though I still see lots of room for improvement.
Still, it's totally uncertain where we will end up with the driver as Jacky is still looking for semi-professional / professional support in driver development. Maybe we don't have to reinvent the wheel and can use something that is tried and tested.
UI-wise there's the question if we need Andúril (preferably stripped to the essential functions incl. thermal calibration/configuration). If the answer is yes, it will take additional time and testing including the involvement of ToyKeeper. Eventually, it could be worth it as this has some potential for selling a whole new generation of (tactical) dual switch flashlights, regardless of any specific brand. I try to think ahead both for BLF joint-collaborations and the future of flashlights and their manufacturers. If the answer is no, we could do a shortcut and use something proven which should be easy to implement.
Last but not least, there's the decision about the emitter. Osram's White Flats seem to be out of the question (too much throw, too coldwhite) and require very exact focussing (see Fireflies E01 issue with Osram LED). If we aim for about 1.400 - 2.000lm there are several options:
3535:
SST-20: cheapest option but highest risk of tintshifts between green (low current) and blue (high current), small die size could result in a too small hot spot
XP-L HI / HD: might be a safe option to use
XP-L2: most efficient in terms of lumens but looks awful in a reflector but might be useful underneath a TIR lens that eliminates the "rainbow-effect" of a flip-chip design
LH351D: probably less efficient and with a high risk of green tintshift, HighCRI version will not reach the target lumen range
5050:
XM-L3: not yet available, pending further testing
SST-40: might be a safe option to use and works fine with a CC driver, 4.800K tint was reported to look well on beamshots
XHP50.2 3V: only available in 6.500K and 5.000K while the latter one is subject to heavy tintshifts in green and yellow