Well that sucks… So with the U3, the vf is higher so the current ends up being lower but it is supposedly a more efficient LED. Would I get more lumens with the U3 LED (lower amps) or a different LED with the same 4.2A driver (LED with not so high vf)? Say an xm-l2 U2 or what about an XPL??
Last night I also wrecked my LED, I scratched some of the phosphor off and now it has a rather blue tint… All part of learning :bigsmile:
U3 is only a few percent brighter than U2. You’d get more light out of a U2 driven at 4.2A than a U3 driven at ~3A.
That said, I think I heard that whatever they’re doing to U3’s is trickling down to U2’s as well, so watch out for that. If you can ensure that you are purchasing U2 stock from before the U3 was released you’re probably fine.
As I said earlier, using LG D1 cells charged to 4.35v may help a little bit. RMM sells them and he’s got screenshots of HKJ’s 3A discharge compared to an NCR18650B
http://www.mtnelectronics.com/index.php?route=product/product&path=59_88&product_id=191
Unfortunately the difference is even less pronounced at 5A than 3A. See HKJ’s comparator for more info.
If I got those I would need a new charger too wouldn’t I… I might just settle with 3.2ish amps 
Unfortunately yes, and not a cheap one. I’m not saying it’s a great idea - I don’t even think it will be that good a solution to be honest. I’d shoot for the older XM-L2 U2 if you can get your hands on one. You might also get good results with an XP-L, I do not know. Finding an LED with low Vf is the key.
I have a XM-L2 U2 here, but its on aluminium… I put it in place of the U3 and got 4.13A 
what the hell? That much of a difference??
Yes. In order to better understand how this works, please take a look at the XM-L2 graph djozz provided here:
(post #51)
Either graph will do, but the XM-L2 T6 graph is the bottom one. As you can see, every 0.1v is accounting for approximately 0.5A of current. Your linear driver will limit current as intended only when Vf is low enough to give it at least 0.12v of headroom above the Vf at the desired drive current.
Now you just have to learn how to reflow from the aluminum board to the copper one. Yet another place we’ve all wasted a few LEDs.
What tint are you looking for?
Well I guess that’s why I’m getting 4.1A now is 4.1A with a U2 going to be noticeably brighter than a U3 at 3.2A though?
Tint doesn’t really worry me, as long as it isn’t blue like you get with cheap aliexpress lights 
Yeah I’ve already wrecked one, I don’t have any solder paste which doesn’t help!
Meh. Probably not really. You need big differences in output to notice it. But ~3Amps on an XM-L2 U2 is actually sustainable for a little while since it doesn’t start below the minimum input voltage for regulation. ~3Amps on your XM-L2 U3 is with the driver already out of regulation - therefore it will fall continuously. That’s why I’d want the lower Vf part.
Solder paste does help but before I had that it helped to pretin the mcpcb pads and use the same method, heat it up until the led slides into place. Practice on the dead one. The copper mcpcb will take more heat than the aluminum one to reflow.
Well based on that I think I’ll drive a XM-L2 U2 at 4.2A in one C8 and put a U3 in my other (2.8A) C8.
No, that’s just flux. Solder paste is gray. It has has tiny tiny balls of solder mixed with flux.
If you’re in the usa RMM / mtnelectronics.com has solder paste and ships fast. Fasttech also carries it.
There’s rosin(a type of flux), solder, and solder paste. The paste is a mix of flux and ground solder. That link is for flux and is what is used with wire solder.
FT solder paste http://www.fasttech.com/products/0/10003546/1261003-chips-repair-tool-soldering-paste-grease
I use kester but there are plenty of other options that I don’t have experience with.
Right thanks, so that Fasttech stuff will work well for reflowing?
Is solder paste better than solder for reflowing?