To get a feeling and gaining knowledge for this group buy, I just finished a 9xXP-L Hi SRK-clone mod that draws 24A (see the what did you mod today thread). It has warm white emitters so the (continuous) output is ‘only’ 4500 lumen (5600lm at start-up), but the power consumption is 90W. It is too hot to hold in (just measured) exactly two minutes (room temperature without air flow). Too hot to hold for a light with a good thermal path in my experience is the correct moment for a stepdown (a safe distance from damage). At the same current, with cool white or neutral XP-L emitters, this light would put out well over 6000 lumen, with added DTP-boards 7000. It feels that the body can handle that, but for a stock light to handle these amps and heat it requires a good design with no mistakes. We are aiming for that of course. And we’re making a 4 emitter light, not 9, so if we really want 6000 lumen OTF that requires very hard driven XP-L leds.
A stable continuous output (room temperature, no airflow) for a host like this appears 6A/2000 lumen to me, with airflow at least double that.
If I were to run a customer BG light, I would want the light to be running above 10000 lumens to make it extremely unique
This group buy is for a 4x parallel battery, 4 led light. You can not get 2500 lumen OTF out of any 3V type led.
Sounds great! Can’t wait to see what specs/price it will have.
I’ve seen the 10K mark mentioned before so surely there will be interest!
I have a UltraFire 7XML SRK clone that puts out more then 3000 lumen stock - guestimate from ceiling bounce with a lux meter. It is 5-8% dimmer then the massive TR-S700 that has been reviewed here at 3500 lumen.
I would love this BLF SRK to go up to more then 5k lumen on turbo and 3k+ on high. More then that would require 6V/12V and a more complicated battery setup which will raise the cost. That would be for a future 12V edition you can start your car with… 
Well, 4x XP-L V6 @3A each will make ~5000 lumens at start. I think I’d be happy enough with that output. It would be ~41.5 watts to the four emitters. Can the host take that? I don’t know. I’ve never owned a SRK. I seriously don’t want it pushed so hard that reliability becomes an issue. And I don’t want any ~6V or ~12V emitters in this light! Maybe we can work out ~6V emitters in V2.
1 xpl nw for me plz
Yes I think maybe you guys are right, the 10k+ lumens obviously needs the series cell setup and so is out of the scope of this.
Maybe a different thread for a totally separate light with series cells and XHP, obviously we would be looking at a $150+ light. Maybe not so much interest on BLF 
I’m still looking forward to this light, shaping up well for the original design plan :sunglasses:
I might re-sub to the thread in 6 months or so and let the wizards do their thing, no way we will have this for Christmas is my guess 
No point getting excited at the ideas just yet, just torturing myself!!!
heheh thanks for the trust
and indeed it will take some time
OK, able to be continuously comfortably held in Brisbane summer night time temperatures (25C).
OP updated with
NEW POLL, community feeler:
Please vote and thanks in advance!
Im 100% with you brother.
go hard or go home 
The answer is simple: direct drive it to as many lumens as we can economically do. Implement thermal control instead of turbo timeout so it remains holdable at any ambient temperature.
I don’t care if it won’t do max output for more than 30-60 seconds before going back to lower output. There are many cases where that’s all I need it for.
+1
Yes fixedit +1
Except that maybe a user configurable turbo timer could prove more stable then thermal control. (maybe could ;))
Wouldn’t know, thermal control is still on my list of things to experiment with. Ideally, it could be user configured to either a timer or a max temperature. One nice thing about the timer is that you can easily override it manually (eg. maybe you’re wearing gloves and can handle the heat for a while longer).
Agree, that's the sort of the direction I would go. Also I kind of like the idea where temp control backs up the turbo timer, so the temp control is the fall-back, set to trip a bit higher than normal. Of course turbo timers have some major faults, cell level for example. With DD FET drivers, fresh cells produce a lot more amps/heat than cells down to 3.6-3.8v for example - I've felt this.
For FET+1 drivers, the optimal most efficient setting I would think would be running the 7135 full out (0.35A or 0.38A), and the FET off. Think I measured this on a 5X modded light and was in the neighborhood of 160-185 lumens. Single LED lights are like in the 130-150 lumens range. It should bump higher the more LED's in parallel, but not drastically, so this result is expected. Most of my mode sets use the 7135 at 100% with the FET off. I call it 10%, but that's a pure estimate based on 1300-1500 is 100% on the 100% FET.
I've found my nice spread out mode sets for a single LED are not so evenly spread on multiple LED lights - the big jump is going from the 7135 to the FET, which of course makes sense. TK's mode sets she's calculated out spread the FET usage more evenly so work better in maintaining the output spread, as well as tint transitions. I preferred utilizing the 7135 exclusively in the lower modes, assuming better efficiency and better tint in the low modes.
This 2 different methods of defining mode sets both have advantages, which is why I include both methods - General Rule: if there's a choice of goin one way or the other for usability, let the user decide 
.
Now doesn’t this sound like heavenly music?
While we’re on the subject, I’d like the high which the turbo or temp timer drops to to be as high as the light can run continuously without toasting anything. Thus if you need max light you simply turn on turbo and forget about it till LV occurs.
Still loose with the rest, just want a useful “lumens bomb” that will impress the peanut gallery :smiling_imp:
Phil