To significantly step up from the BLF Q8’s sustained brightness of approx. 1500lm you’ll need to spend a lot more money. Think Olight Marauda 2 or Acebeam X80-GT which can sustain over 3000lm.
The higher the max. output, the higher the sustainable brightness, because LEDs are more efficient at lower current.
So probably a Trustfire TR-J20 modded with XHP50.2’s would be a good choice
I am kinda disappointed that noone has made a flashlight with high-efficient LEDs like the Samsung LM301B/H yet - sure, you need a lot of them, but 220Lm/W in you hands would be nice.
For sustainable brightness there is only two parameters involved: led efficiency and flashlight outer surface area. All other parameters are not in play.
By asking for a soda can light you have already more or less fixed the surface area (are there soda cans with extreme finning?, that could stretch it a bit).
The option leftover is led efficiency. Q8’s with XPL’s are running the led at their max, so with very much reduced efficiency (but good for throw). High 90+ CRI leds have down to half the efficiency as cool white 70 CRI. So those big cool XHP50’s (lousy throw and ugly beam) are the way to go.
I’ve just never needed to use 3k lumens for more than 5 minutes. Turbo is over 10,000 lumens and I’m pretty sure it can sustain 3k indefinitely. For a 3V XHP 50.2 emitter, 2A will get you close to 1000 lumens. Therefore, under 8A for all 4 emitters should be quite sustainable for a soda can size light.
One (small) gap in the market is a super sustainable small form factor light.
I’d be interested in a 4x XHP70.2 with an efficient boost/buck driver and with the emitters under driven. The emitters should be driven only within the 200 lumen/watt efficiency range. A sustainable 2000 lumens at 10W for a single 21700 cell light sounds nice.
It would even nicer if someone could design an optic to create a nice beam from 4x XHP70.2. Although, it would require a pretty unconventional shape.
I would guess your next step in progression would be a soda can size thrower. Have you considered a Convoy 4X18A Cree XHP70.2 ? HERE
.
Or at Neals Gadgets HERE
.
I have the 90.2 version, but the run time is short, the 70.2 version is better. If you get the 4K version the CRI should be higher and the price is cheaper.
.
I have the Convoy 3×21A and its a marvelous light, Tremendous power and sustained output. Takes standard 21700 batteries, not that proprietary nonsense. At $65, its a no brainer.
I did not include driver efficiency as parameter because even the FET drivers without 7135-bank are in most modes sufficiently efficient, the difference with nice current regulated buck-drivers is often not more than 10, but agreed that is a factor too. In the worst case design the FET in fully open mode drives the leds right at their max (I do not consider this a well-designed flashlight) and is then PWMed to 30 to reach max sustainable lumens.
This is very interesting to me in that efficiency by definition should be a critical factor to this. However, the least efficient LED light that I have can sustain about 1k lumen for hours on end and not even get very warm to the touch. Roughly 76 lumen per Watt. However, that is not a flashlight — it is an under counter light. No idea how many LEDs it has but certainly at least dozens. These LEDs are laid out over about 62 X 600 mm. Don’t think there is a way to do anything like this with a flashlight. Must be just too expensive to use more efficient LEDs in the under counter lights. Should be closer to 6 watt rather than 13 but apparently, getting the efficiency out of counter lights would be just too expensive to justify.