Thanks for the detailed explanation and advice.
How important is the choice between smooth vs orange peel reflectors?
For ceiling-bouncing it does not matter. But as you may also want to use such a light as a general balanced light, I would pick an orange peel.
2 Thanks
Not important. I find that orange peel reflectors have a tendency to produce lopsided beams (the extent of OP being radially asymmetric across the reflector) with an accentuated yellow-green corona. I’d stick with a smooth reflector for the extra throw and nicer beam.
If you ever want a more floody, soft, smooth and homogeneous beam, smooth reflector+frosted tape is a much better solution than OP reflector IMO.
I have hands made S21A with XHP50.3 (4000K) and SMO reflector.
And it is quite good for lighting a room with light reflected from the ceiling.
Please explain: there was smoke with 3*B35AM, but 4 is ok? Another vote for this. I have B35AM in M21F and I love it. Exceptionally throwy for a Nichia LED with beautiful tint, perfectly set halfway between 219b and 519a.
With the quad version each emitter will receive 25% less power than the triple, so should be less likely to smoke. Also, the quad will have the emitters spaced further apart.
3 Thanks
The current is divided.
5A/3 = 1.7A per LED
5A/4 = 1.25A per LED
Less Amps = Less Heat
4 Thanks
Ooops got it. Thanks and sorry guys lol. Same current/power divided by 4 instead of 3.
I was not thinking in the right direction. 
1 Thank
I would take one each in M21H, M21F, and M21E. 
This would upgrade my small Convoy collection to more lights with side-switch and on-board charging.
I measured ~1100 lm using lumen tube and 260 m throw in M21F with single B35AM. I can hardly wait for the quad version.
I don’t know if all B35AM’s are the same but I could vouch for the 4500k batch you have. Very beautiful tint in between 219b and 519a. Definitely time for a multiple LED version.
Hi Simon!
Never asked you anything 
I vote for a XHP 70.3 and/or XHP 70.2 light with the new 46950 battery.
Boost and reverse charging too, if possible.
That would be a killer for camping/power outage/emergency.
All the best!
8 Thanks
I agree. I want a single led only. I don’t need insane peak output, I have other lights for that. I just want insane runtime. If we imagine this light has 30,000mah and is similarly as efficient as the, say, wurkkos ts22 (5000mah and 5+ hours at 600 lumen). This light would have ~30 hours at 600 lumens.
3 Thanks
Yes, one sole led is ok.
I’m already saving money for this, Simon =)
Why I like to have more lights with B35AM. Numbers below from Simon (my amateur test shows even more negative Duv). Check out the 99 R9:
Nichia B35AM 4500k
CCT 4454k
Ra 98
R9 99
R12 77
Duv –0.00503
This LED (and Nichia E21a) “looks” on white walling very much like the legendary 219b, but less “rosy tinted”. If you think 219b’s look is “too much,” this one should be perfect. The single LED in my M21F w/ B35AM tested with Texas Ace lumen tube gives ~1100 lm. So 4 of them could be quite yummy lol.
4 Thanks
Hi Simon my Convoy M25F w/ B35AM was bought post the smoke incident, so the current had been throttled back to 2.4A, and even then output still excellent for single Nichia 9080 LED and the light runs cool, actually coolest of all my Convoy lights. It acts as if it’s just “cruising,” easily.
I am not technical and don’t understand, if 2.4A is ok in my M21F, why does the triple LED set up smokes with 1.7A at each LED? Am I missing something? Thanks.
Looking at the MCPCB, could be the close proximity of the emitters, fragility of the B35AM, lack of thermal pad, and just not wicking away the heat quick enough leading to premature destruction. In a quad emitter setup with spacing, perhaps it can prevent centralizing the heat in the center.
5 Thanks
I also think the proximity of the emitters and lack of thermal pad is main factor in the smoking issue. Also, being aluminum instead of copper could have something to do with it. Though the quad B35AM D4SV2 mcpcb YBF650 has made up is also aluminum, and I have no smoking issues giving each emitter ~1.8A.
2 Thanks
I’m glad to hear it’s survived that without causing trouble, it was originally intended for only 1A per emitter in the boost-driver D4Sv2.
That means it should also work fine with the higher-power LumeX1, very exciting!
I believe that due to the lack of a DTP in the emitter, the material of the core isn’t as important.
I think that the main factors are the area of the traces/pours and the thermal resistance of the dielectric, as it needs to be a large area with a low-resistance dielectric to function well, and ideally a heavy-gauge copper layer, but it seems like 1oz is fine for lower current use.
1 Thank
Thanks guys for the explanation. I thought the smoke is from too much current frying the LED, like too much current blowing a fuse, but instead it is thermal destruction from having too many LEDs without adequate heat management?
Yeah, B35AM are rated for 1.8A continuous or 2.4A pulsed, but you need adequate cooling.
1 Thank
