Talking about a P2 binned XHP50.2 is not realistic because as far as we are concerned it doesn’t exist if we can’t get our hands on it.
We are talking about same bin leds or what? Just compare same bin - H4
I mean no one in their right mind would buy the lowest bin XHP70.2, if you even could, which I don't think you can.
Basically the comparison is comparing the lowest bin XHP70.2 against the highest bin XHP50.2, and concluding the XHP50.2 is more efficient. It's not realistic because no one would make that choice. They would look at the highest bin available for each, then compare.
We know the CREE specs are unrealistic in several respects:
- they list bins not available
- in a few case they don't list bins that are available
- they limit amps, which is fine for a manufacturer, but we at BLF don't follow those limits. I'm stating the obvious here with the BLF proliferation of FET based drivers, or at least an available FET output channel. We've been doing this for many years now.
To me, whether two different LED types/models are in the same bin is meaningless. In this case they are binned at different currents (1050mA 12V for XHP70.2, 700mA 12V for XHP50.2) so it doesn’t necessarily make any sense to use bins as a common reference across different LED types.
Datasheets, like everything else, have to be taken for what's good and what's not good or relevant. CREE doesn't cater to flashlight enthusiasts, so we can only get what they choose to give us, or the resellers can make available through their sources.
Here we go again with the “efficiency” thing? :GLASSES:
Luminous efficiency/efficacy is biased to a presumed human eye/mind sensitivity curve. Since the relative weight of the different frequencies isn't thus equal, it is clear that unless two different emitters are to have exactly the same spectral power distribution curves any actual efficiency (defined as actual or meaningful radiant flux emission versus input power) comparison between them is, in my opinion, wrong. This is the reason emitters whose emission curves are more in tune with those of blackbody radiators or the sun usually have lower efficiency.
But 
well, where's that Cree emitter without tint shift and all high and balanced R9 (namely) to R12 values?
So yes I like the beams of Osram (or other low-CRI) throwers, but comparatively colors and contrast look quite dull.
Take it with a grain of salt. Data sheets are for reference only and don’t give you the full picture of the led output performance, just potential. It’s when folks from the form test them thoroughly with real-world use (at higher than advised drive currents/voltage, in reflectors, under lenses, behind optics, etc) that we get true performance. I have been pleasantly surprised by some emitters that had so-so numbers on data sheets, but when run at higher currents really came alive and had great performance, performing way ab0ve what ai thought was possible (like the xp-g3, Xhp50.2, xp-l2, Samsung 351D, Luxeon V) and others that performed less so.
There is no H4 in XHP70.2 datasheet. The ds doesn’t list anything below J4.
And even if it did, what would be the point of comparing poor bin of one LED with a good bin of another?
To get fair results compare the best bins you can get.
Is it too radical for such a small flashlight to install such high-power LED lamp beads?
It’s completely ridiculous, but that’s what makes it fun.
[/quote]
Is it too radical for such a small flashlight to install such high-power LED lamp beads?
[/quote]
I was curious how it will turn out and it was ok. I was bored at home :)
I did solder 3.5mm thick copper on the bottom of the pill for extra heatsinkining and added L4P 12A CC driver. Results are impressive. My guess is around 4000lm. Enough for wow efect. Even on low/mid it’s bright as hell ut having it in 4x18 host with e-switch and Anduril clone will be the best combo ever.
M21C is an M21A variant? Are the mirrors the same for the SST40?
Also, am I correct that the C has better cooling than the A, or there is no such difference?
I ask these because the 12 group models are very tempting to buy.
Convoy 4X18A is interesting also. How does it compare to the M3/L6 in beam pattern?
Oh yes, that would be interesting!
I would love the 4x18A with SBT90.2
I haven’t considered it. There is no suitable driver for the time being,
The heat dissipation effect of M21C is better than that of M21A,
The 4x18a spot effect is between m3 and l6,
There you go:
I also have M21C review with SST40
Thanks.
Youtube even shows that I have seen the M21C review video 
Great review Zozz! Thanks 
Maybe a little off topic but I wanted to know if Simon’s ePacket free shipping is going to go away on July 1st. USPS is supposedly going to hike up their processing fees for Chinese vendors.