I can understand that, the output and beam is really nice, but I would need to know how a really good bin would perform, you may just get the same output for less heat and power.
And for me, I’m not so much into cooler leds, I bought the (cheaper) xpg2 version for modding in the first place 
I was going to mention that if I ever picked up another, it would be for experimental reasons.
The efficiency does indeed drop above 350mA. Each mode above that is less efficient than the one below it, but it at least tries to get the most it can by keeping the 7135 chip on steady while the FET pulses.
Efficiency also drops in moon, making it the least-efficient mode. And it drops in the “low” mode too, though not nearly as much. At a very rough guess, I’d estimate the efficiency of the modes like this (ish):
That is, if I assume that the 7135 chip by itself at full duty cycle is “100%”.
But I haven’t measured a S41 (don’t actually have one), and the numbers change anyway depending on the type of cell used and the voltage. And if the driver has a bleeder resistor, the moon/low/med1 modes get significantly less efficient. It’d probably be like 5, 60, and 90 for those. And all the numbers go down when the emitters are hot.
Anyway, a very rough guess but hopefully it at least illustrates the kind of efficiency curve it has.
I think the point of a multi-emitter 219B is that it gets a great tint with as much output as a single higher-powered Cree emitter. XP-L, for example. It’s the price we pay for high CRI.
The nicest tint I get on any light is actually from a triple 219B with a “moonlight special” 3-amp driver. It’s not exactly a hot rod, but it does give me 800 gorgeous lumens of the finest pure white. It’s what I use when I want to see what color something really is.
I wonder when we’ll have 92+ CRI 219C emitters in 5000K, and how bright they’ll be.
All well and good until you actually hold a 4258 lumen quad 219C X6 in your hand.
Or a 4634 lumen triple XM-L2 U4 1A X6.
Or a 5338 lumen triple XHP-50 SS/Cu X6. Or, with a pair of 18650’s, 7210 lumens.
Point is, sometimes it’s not about efficiency. Of either emitter or driver or cell or thermal transfer. Sometimes it’s just the pure joy of running it.
What TK was referring to above was simply that the A6 firmware, on level 3, is running the single 7135 chip at it’s most efficient output in a tested X6/X5 single emitter light. Most efficient for that particular light. For those that want to consider a long run time, like in camping or emergency use.
Like, Sure, my Corvette (hypothetical) has 600 HP but if I hypermile it I can get 31mpg. OR I can put the petal to the metal and laugh like a demonspawn. 
Edit: By the way, the graph/chart shown above in Sharpie’s post is credit to djozz, he’s the one that worked that up and does enormous amounts of testing. Surprised you left that bit out, Sharpie.
Match did a lot of groundbreaking work as well, clearly the class leader in establishing the importance of copper under the emitter.
Here is just one of his threads on this subject with his accompanying charts.
Edit: Notice the number of reads at the bottom of that OP… over 37,000!
Match was pushing the limits, as was Scaru and a few other’s, like ChicagoX. Comfy started playing with the large 07N02 Mosfet, stripping a Qlite and force feeding the FET onto that board, all kinds of things started cracking open. MattAus drew up board designs, then Wight came in and started doing the same and WarHawk and several others busted it wide open with the MOSFET drivers. TK wrote new code and BAM! We were up and running in entirely new realms.
They’d design boards, I’d cram parts on em and stuff em in lights. Like I’ve said all along, I’m just the hack who cuts all resistance, I don’t design the stuff just make it scream in a flashlight. For me, it’s the build, which is why I’m going on 160 lights now. 2 or 3 close friends used to send me 4 or 5 lights at a time to modify, I don’t even know how many lights I’ve built, well over 300 for sure. Early on it was stacking chips, 12, 16, 20, 24 chips on a Qlite for power. Then the MOSFET’s killed that off almost completely and I learned to rewrite code in AVR Dude, flash the MCU’s and build my own drivers. The rest is history. I stuff these into everything, sometimes with 10mm drivers sometimes 22 or larger.
All along it was on the premise of Match’s charts, what can this take, what can I expect from that, and then the lightbox made it all real so I could see the changes being made, real fun there! I think Justin is the one that really got me started though, about 3 years ago, with his Scratch build contest. I managed to make a flashlight from a live .50BMG round and the door flew off the hinges, started cutting copper on my drill and rotary, no holds barred.
Gotta say, the little quads are great pocket dynamo’s, but really I’ve recently taken it too far… the triple XHP-50 is just too much for a smallish light, there’s so much heat generated even the full copper X6 pill can’t handle it. Not often I find a point that needs to be backed down from, but the SS/Cu X6 with 2 18650’s and 3 XHP-50’s is just excessive. Even more so than running an 18650 on this S41, I’m talking over 130 watts in a pocket sized light with a belt clip! Found the ceiling…
An 18650 in the S41 is capable of de-soldering leads. If the positive lead de-solders and touches ground, say with the light in a backpack and accidentally turned on, well, with luck the springs will collapse and break connection before anything starts arcing and spewing sparks and lithium. Perhaps the 18650 can be controlled when in hand, in use, but what if….
Would this light dissipate heat better if the threads weren’t anodized? I know some people want to be able to loosen the tail for lock out so I can see having anodizing there, but the head end should have been bare, right? Or would there be galvanic stuff going on if the aluminum and copper made contact?
I love how BLF members worked on these drivers and how Manker was willing to run with it and build lights around them. I would hate to see them give up on that. The reason these lights are great is because they listened.
The light shouldn’t get hot in a backpack because it should be off, right? Maybe if the switch was accidentally pressed? Might be a reason to turn off memory or loosen the tailcap if you’re throwing it in a bag.
TK wrote:
Info like this is why I follow you around : - )
I firmly believe that the late Edward Aloysius Murphy jr. is looking down on me…… and he is smiling.
That’s the reason why I always loosen the tailcap of any light that I’m not actually carrying. Not a tiny but, but a full turn.
And remove the batteries from copper and SS lights with a switch like the X5/X6 and the Solarforce.
BTW, the same goes for my Thorfire C8s, with the “powder-like” anodisation. I am under the impression that it is not insulating very well. When I unscrew the tailcap with the switch ON, the light flickers all the way till the cap falls in my hand.
I blame Dale. He built it.
It’s also one of the smallest 18650 lights around, and makes the BLF A6 or Convoy S2 look pretty chunky. I’m still kinda hoping someone will get a manufacturer to make a BLF version of it (CNQG brass 18650), and hopefully also its AA-sized sibling.
I did? That was a triple? REally? The Keeper’s Key? Well, if you say so, I don’t remember that. Glad you like it. 
oooooh. SHINY!
soo.any quad xpl rework?:laughing:)
I put a quad XPL in mine with H17F driver and spring bypasses on both springs.
Very nice. Incredibly bright, especially on a Samsung 30Q 18650.
I was going to try XPG3 until I realized I’d only ordered 3 of them.
Tried the XPG3 and fried one of them, it was pulling just over 18amps on a 30Q ? Now I’m running XPL- HI V3 1C in one and XPL- HI V3 2B in the other, both running the H17F driver.
it mas be hot in seconds right??any beamshots please?…numbers:))
Not really a “Beam” more like a massive wall of day light? Yep she gets hot fast!! 
Hi,
do you have a source for the XP-L Hi V3 2B? I thought these are sold out everywhere.
Regards
Kenjii
Sorry no, I bought a bunch of them last year at IO, But Kaidomain has the V3 1C’s http://www.kaidomain.com/product/details.S024438 The 1C’s are not bad at all in the other (my wife’s) light because of the optic.