It’ll look like a double-barrel shotgun. Sharp!
I’m interested in anything that comes from here 
I’m interested.
I wouldn’t use royal blue LEDs! There is sufficient evidence that they will damage your eyes.
They pretend to be brighter in specs (higher lumen values) but visible output is similar to standard blue.
Instead of using multiple 7135 for each color LED you might consider a single D882 transistor per LED, I’m using them in my RGBW D4 EDC since January 2018 without any problems. D882 are slower than 7135 so I recommend reducing the PWM rate for near moonlight levels.
I prefer a double head light (not double-barrel but one head at each end), similar to my RGBW Q8 but sized as a D4. One powerful quad head with TIR for a couple of thousands lumens and one head with 4 color LEDs behind a TIR. This would require a special tube which leads the positive terminal of the cell to the rear head (hint: FW3A).
I find double-barrel heads really ugly and I fear that I’m not alone in that. And I do not see why the white led should need a larger reflector than the coloured ones, a EDC-type beam is even preferable for a light like this in my opinion.
Okay, some more thoughts and rough ideas...
Multicolor flashlight with reflector
Let's just imagine we had a quintuple reflector as in Acebeam's X65 flashlight. Now, the X65 is a heavy duty flashlight with a large reflector. So, please take this as reference for the reflector design only. Getting there entire reflector smaller the configuration could be:
center emitter: Cree XP-L HD V6 3D (4.885K) or for even more throw Cree XP-L HI V3 3D (4.885K) or - more cost effective - Luminus SST-20 4.000K 95CRI / SST-20 5.000K 70CRI
adjacent emitters: Cree XP-E2 red/green/blue/amber or just three colors and UV (e.g. LG LEUVA33W70RL00 365nm or Seoul Viosys Z5 365nm).
the adjacent reflectors would at least need to have a structured surface (orange peel) to reduce artifacts. However, I'm not sure if the shifted angle (not being in the center of the host) would cause an uneven beam shape.
Multicolor flashlight with LEDIL ANNA-40-7-X (where "X" defines clear spot, medium spot or frosty wide beam)
source: https://www.ledil.com/search/?families%5B%5D=Anna-40&q=
Ideal would be an optic where the center LED part is clear for a narrow spot beam of white light and the adjacent "rim" part would be frosty wide spot for a diffused colorful beam without artifacts. A compromise could be ANNA-40-7-M with light diffusion.
center emitter: Cree XP-L HD V6 3D (4.885K) or for even more throw Cree XP-L HI V3 3D (4.885K) or - more cost effective - Luminus SST-20 4.000K 95CRI / SST-20 5.000K 70CRI
adjacent emitters: 2x Cree XP-E2 red/green/blue each.
In any case, the biggest challenge however is the driver. The most intriguing questions are:
- Is the Dr. Jones (or any other) driver capable to fulfill are requirements in terms of hardware and software (UI)?
- How difficult will it be for Sofirn to source enough drivers from third party manufacturers?
- Alternative: How complex and difficult would it be for Sofirn to develop a comparable driver on their own?
- How much costs will be generated in developing/sourcing the driver and will these costs be reasonable with regard to the targeted sales price?
- How much effort in terms of manpower will be necessary both by BLF members and by Sofirn's engineers? I deliberately pose this question because we may overstrain things too far here. The development of the SP10S Andúril NiMH/Li-Ion boost driver is stuck for some weeks for reasons unknown. I assume that Sofirn's engineering team is kept busy with a lot of other topics right now and even if there was room for additional developing our driver experts still have no direct communication channel to the manufacturer's engineering department.
For sofar as I have come to know Sofirn, they are pretty good in developing driver hardware at the moment: for the SP36 they managed to combine the Q8 Narsil driver with USB charging hardware, including extra indicator leds in the switch. All without any BLF involvement. This tells me that they are pretty capable in that department. They are not good in designing good user interfaces though, every UI that they developed sofar have flaws. It is because they are not end users so they have to guess what their customers want and are not very good in that. In contrast: Hank from the Emisars is a flashlight enthousiastic himself and he actually carries a new design around for a while for testing.
So I think that Sofirn can make pretty sophisticated designs, but will need a lot of input for the user interface, and they are stubborn enough to have an opinion of their own about it which you will have to overcome or meet in the middle.
Yep, most of these look like ugly Frankenstein this would biass the sales probably. And it’s crucial to gurantee good sales since the volume will not be huge I believe.
How about three reflectors then? :laughing:
Anyway I do hope to have a bright and throwy red LED - anti-insect, no-blue, preserving night vission goodness…
In terms of blue&UV - I think there should be some warnings stated (inprinted on the body?). Also those regarding melanine, prostate cancer and eyes damage. I think these hazards are also covered by world standards (EN/ISO/UL/ASTM etc.)
Left picture is a beamshot of a single color LED in my RGBW D4 with TIR.
Right picture is a beamshot of a single color LED in my RGBW Q8 with quadruple reflector.
Which one do you like more (if you are not a Mickey Mouse fan)?
Yeah, headshot Mickey looks kind of bizzarre. Looks like one should take into account either separate reflectors or TIR optics. Seems the TIR is cheaper, easier and nicer, but it reduces the output a little.
Btw., when using a TIR, the white LED does not have to sit in the center, you will have the same perfect beam with any of the TIR lenses. You don’t even need a TIR with centered LED.
With the proposed 7-LED TIR I would rather prefer 3 or 4 powerful white LEDs and 1 LED for each extra color.
Jeez... :O :FACEPALM:
Okay, let's drop the whole idea about a multireflector light with independently running emitters. I think we should stick to the idea of a TIR optic.
I agree. As long as the output of one LED suffices it should be fine. I assume with the right (over)current applied one XP-E2 should emit up to 150-250 LED lumens depending on wavelength. Red XP-E2s are known to be more sensitive to overcurrent and heat than blue XP-E2s. I could also imagine the LEDIL ANNA-40-7 optic with the Osram S2WP in the center and some powerful white LEDs around it.
Not meaning to disagree with you Djozz but I find it very hard to get substantial feedback from Sofirn so far. I sent them quite some suggestions for improvements about hardware and user interfaces in the past, both here on BLF and also directly to the factory in Shenzhen. As usual, patience is the key but without proper mutual communication in place the odds to run a project like this are very remote.
I agree, there is a lot to be done right in a succesful multicolour flashlight. Sofirn not only must have the capability to build it but also the will to do it in close cooperation with BLF. My experience is that cooperation does not start smooth, only when after some prototyping they have the confidence that it will be a good product it becomes much easier get the last bits done well (see the lantern project).
I was wondering, is there any difference in the emitted light between a coloured LED, and a normal white LED with a coloured filter, like the ones available for some Acebeam flashlights ?
There is. Significant. White light is a mix of all colors. A color filter filters all wavelenghts except for the own color (so all other wavelengts are “wasted”). Monochrome led puts all it’s power to emit just one color (although it usually contains also some neigbour wavelenghts too).
I’m not an expert but if you search this forum you can find more information including those really PRO.
There is a catch though: white leds are more efficient than coloured leds, some colours are so inefficient (orange, lime) that filtered white light may come close in efficiency to dedicated colour leds. But a third option for colours are phosfor-converted colour leds (I know green, lime, amber and red exists), just like white leds the underlaying led is a highly efficient blue led but with a phosfor that converts in into a narrow(-ish) wavelength band instead of a wide white band. For lime and orange they are the most efficient leds.
That thing is dang short to have screw in chokes… but I kind’a like it anyway. Well, Except for the grip design. Ouchy! 
Interested.
Surely this would need Toykeeper’s input multicolour lightening storm anyone?
interested
I agree with TIR’s.
The 7 emitter concept (Fireflies E07 style) has some appeal, although whether or not it makes sense depends on the driver.
Dr. Jones’ RGBW driver design for example, only provided 700mA to each channel, because it had the same 2x7135 on that channel as the color channels. There’s no reason to have more than 1 white LED for that, and even any of the color LED’s under discussion can handle it.
Flashy Mike went a bit further and made his white channel FET + 1. In that case, it would be possible to make proper use of multiple white LED’s, although even a single XP-L or LH351D should be good for 1500+ lumens.
Question for Flashy Mike: with the driver and firmware you created, are you able to ramp each color channel individually to create different mixes? It looks to me like what you have already created achieves the basics of what a lot of us are looking for.
Personally, I’m still attached to my idea of supporting RBGA for a little more variety (and again, because the E17A phosphor converted color LED’s should mix to a better white than standard color LED’s), but it sounds like I’m in the minority on that, so I’ll go along with what the majority wants.
I used the DrJones driver with 4x7135 per channel (stacked) and a FET on the white channel, so the concept is versatile.