No, that would not be realistic. The stock L6 at about 3200 lumen just barely is sustainable. The temperature gets way up beyond safe levels, but it does not destroy itself. For an L6 sized light, about 2,000 to 2,500 lumen might be sustainable without burning anyone.
Iām just gonna ignore all these āwide rangingā ideas and focus on a modern version of the L6/S70s.
I agree. The 26650 is like the C cell Maglite. I am used to the D size Maglite as itās not too big. I wish there were good 32650 cells with the same energy density as top 18650 cells. The 32650 is 3 times the physical size of an 18650 so imagine it having 9,000mah to 10,000mah. That would be sweet, but unfortunately 26650 is the best choice we have today. The newer Shockli 5500 and KeepPower 6000mah seem to be the same cell as both measure about 5750mah.
Also keep in mind the L6/S70s are still using the older xhp70 and OP reflectors. A modern version should use the 70.2 and SMO. You get a nice even hotspot and more throw. I use this combo daily on my L6 and it works great. You do get a slightly yellow corona, but itās not that big a deal.
The real question is the driver. Should it be a FET (like so many folks currently use, myself included) or a Buck (like in the stock L6)?
I think Sofirn could definitely build a FET driver as we have seen on their other lights. It would probably pull 17A or so (with good double springs or bypasses).
Maybe a 9A to 10A driver like the K70 was going to use would be best? You lose maybe 1700 lumen, but itās a lot less heat and gives more runtime on turbo before the thermal protection can kick in.
If Sofirn could somehow use a UI like NarsilM or Anduril we could set up the top of the ramp to be about 9A to 10A and then turbo could be full 17A to 18A. This makes sense to me, but Sofirn probably wants to design their own driver and UI. Whatever the case, I hope it has one click on/off on the side switch. Please donāt be one click on and then a āpress and holdā for off.
I have no problem with a tail switch. My L6 does 17A and itās been going strong for a year or so. I tried to measure any losses in the tail switch by measuring lumens with it and with a piece of wire. I got the same output.
A flat tail cap is okay as well, I suppose. Just as long as we can twist it to do a mechanical lock out.
Would you guys want Sofirn to use a similar UI to their C8F which has multiple āgroupsā you can choose from?
The MF03 is not time sensitive so it had to be put aside temporarily to get the MF04 and MF04S to market (to compete with Lumintop). They will start up the MF03 project later on.
For right now I think Sofirn wants to do a single emitter light.
Later on they can do just like they did with the C8 to C8F. They can machine the head with a taller shelf, run 3 emitters and make a triple reflector. We know the 3ā head on the L6/S70s lights can support this. Just look at the Haikelite MT09R. It has a 3ā head and it can do 20,000 lumen! You wonāt get the same power from 2 x 26650, but I could still see it easily doing 15k lumen. (A pair of 30T cells, if they fit, might get 20k lumen)
I really like the idea of a MT09R head using a 2 x 26650 battery tube. No one has done this. The old BTU Super Shocker was a bit bigger in diameter and only had a tail switch. I think Sofirn could do it, but later on as it would cost quite a bit more money.
It would depend on the heatsinking solution though. Copper heatpipes for more efficient transfer of heat away from the LED, and more fins/surface area to dissipate that heat would theoretically mean a higher stable output than a solid chunk of copper with a small number of thick fins.
I wonder about using a design like the NoFan CR-95C CPU heatsink as a flashlight head. Flip the orientation of the CPU block, add gaps for the electrical, mount the MCPCB, LED, and reflector on the inside. Then add structural support with horizontal aluminum fins and aluminum posts. Waterproof or economical? Not in the slightest. The cooling potential should be excellent, though. Some napkin math says a single XHP70.2 cooled in such a manner (assuming some airflow and quality direct thermal path MCPCB) could theoretically sustain more than Calebās mentioned 5k lumens.
More practically, I doubt copper heatpipes would be economical, and thin aluminum fins flatly wouldnāt be sturdy enough. Maybe a vertical orientation of medium thickness, more densely packed fins, curved to provide better airflow when the flashlight is held parallel to the ground, attached to a copper heatsink would be an effective next step? It would be a unique look, at least.
Has the Sofirn K6 project been abandoned or left in stand-by? :zipper_mouth_face:
The light that basically looks like the Q8?
They have been talking about it, but not by that name. Sofirn has access to the Thorfire Q8 parts and have been selling replacement drivers, switches and mcpcbās. They want to do their own version of it. Maybe even use battery carriers and double up on the length. You can check it out here.
BTW, I see they do currently make a Skyray King clone called the SF34. I donāt think it shares any parts with the Q8, though.
Yup, Jason, that was the light! I somehow missed that post, by that times I was away from the country and saw the forum less frequently! :zipper_mouth_face:
Thanks for pointing it out, gonna read that thread :+1:
Assuming 3x70.2 LEDs, wouldnāt each LED be operating in a more efficient range thereby generating less heat?
Iām not sure why your assuming 3 emitters since Sofirn wants 1. Anyway, I will use the L6 as an example as Iām more familiar with it. When it comes to heat generation, itās directly connected to how many watts of power we are generating. The stock L6 pulls about 4.7A at the tail cap and that is basically too much to be sustainable and not burn you. Letās say the voltage sag of two good batteries gives 8v. So 4.7 x 8 = 38 watts.
A single xhp70.2 would need to draw about 9A to generate 5,000 lumen. Letās make the voltage sag a bit more and say 7.8v x 9A = 70 watts. Thatās way too hot.
Now to get 5,000 lumen with 3 x 70.2 we would need about 6.6A. Letās say the voltage sag is 7.9v x 6.6 = 52 watts. It can not sustain that. It will have to activate thermal protection and drop down in power.
I would estimate that we need about 30 watts (about 3.7A) to be both sustainable and to not burn your hand. 3.7A with a single emitter is about 2250 lumen. 3.7A with a triple emitter setup is about 2800 lumen.
So even though a triple 70.2 setup is more efficient, there is no way it can sustain 52 watts.
Currently there are only a couple of lights that can sustain 5,000 lumen. One of them is the Acebeam X65. You can get an idea of its size and heat sinking fins here.
It is really big! Much bigger than the light Sofirn wants to build.
Stick with the original K70 theme with slightly larger head 1 emitter and a bit more power and keep the price budget.
Yep, well said āedsgāā¦ :+1: . If weāre not careful this could get all BLFād up. :person_facepalming: . 
Yes, they do not want to invent a totally new design. We should keep it along the L6/S70s theme.
Does anyone have the new SP32A v2.0 light? Do you like the ramping mode on it? I hear itās pretty good compared to v1.0.
What do you guys think of using the C8F UI?
I think combining the 4 modes of the C8F plus the ramping on the other light would give a ton a choices for users. (Assuming they can combine all 5 modes)
Sofirn says they do like the idea of one click on/off (instead of press and hold for off), so I hope they can do that. Do you guys agree?
Iām a huge fan of NarsilM personally. Q8, BLF GT, TA modded MT09Rā¦
Iām not familiar with the C8F UI, but Iād like ramping to 50%, double click to turbo.
I agree. :+1: ā¦ The 21700 / 20700 may or may not be the futureā¦ that remains to be see as far as we are concerned in the ātorchā world.
But for now & the foreseeable future in the world of lightsā¦. the 18650 is far more popular than the 26650 which is far more popular than the 21700. That is neither good or badā¦ it is just how it isā¦
- So if it is to be quad batteriesā¦ keep it 18650.
- If it is to be triple batteriesā¦ā¦ keep it 18650.
- If it is to be two batteries in seriesā¦ make it 26650.
And please SOFIRNā¦ā¦
ā¦ā¦.Keep up what seems to be your tradition of offering āTrue Neutral Whiteā emitters instead of āUrine Yellow or Harsh Blueā.
It seems you have many offerings in the 5000K - 5700K rangeā¦ā¦ āTrue Neutral Whiteā that adds little tint one way or the other.
THANK YOU & KEEP UP THE GOOD WORK!!! :+1:
Sofrin agreed to keep the K70 rolling the K70 is a modded S70/L6. Its NOT a sodacan light, Its NOT a triple XHP70 monster. Slight changed are okay but i dont think this thread is intended as a whole new build?
The K70 was meant to be a functional simple versatile light at a decent price. 50USD was a similiar price to what the Q8 cost and the K70 would of had similiar output or even more output then the Q8 stock.
My S70/L6 see some use while my MT03 is just covered in dust because that style does not suit what is needed when outdoors.
You could even mount a S70 on a rifle if you wanted to because it has the skinnier battery tube and the beam profile is well suited to actual real world use.
Plus running 2 26650s in series is as much W/H as running 4 18650s that is why this style works.
Looks like you jumped in the deep end Jason 
lol.
!
!
I like the 2x26650 platform. Along with the L6 size head. (Althought 3 XHP70ās would be awesome!
Iām always about max output, so is a single XHP70.2 feasible?
Well said everydaysurvivalgear. :+1: ā¦ Sounds good also. Two 26650ās in series is pretty darn nice too. 
From what I understand, Sofirn would like to use their own driver design and UI instead of a whole new driver to run NarsilM. (Not that they are ruling that out)
Using their own driver and UI helps keep costs down.
They would like to use their new v2.0 ramping mode (maybe even a more refined version), plus 1 or more regular mode groups. Iāll let you know if the give more details on the UI and itās modes. More than likely it will be a mix of the UI on the SP32A v2.0 and the C8F. I listed both UIās below.
Here is the ramping mode instructions from their SP32A v2.0. See what you think. Looks pretty good to me. Especially if they tweak it to make it smoother like Narsil.
Group 2: Ramping (takes about 4 seconds)
When the light is in Group 1 (distinct brightness levels), press&hold switch for 3 seconds, it switches to Ramping, the default brightness is xxx lumen (Medium Mode).
When the light is in Group 2 (ramping), Fast click to turn on or turn off (Nice!)
Press&hold the side switch to ramp up and down. It changes direction when ramping again within 1.5 sec; same direction when ramping again after 1.5 sec.
The light flashes two times when it reaches Eco or High.
Double clicks to activate Turbo, one click goes back to the last brightness.
Triple clicks to Strobe, one click goes back to the last brightness.
When the light is OFF, press&hold for 3 seconds, it flashes 2 times and then switches back to Group 1.