I was just referring to the drop off above 4A, that seems low compared to Djozz’ tests. Poor thermal conductivity between the die and heat pad might explain that and not be surprising from a knock off emitter.
Ah, I got you now, 
Yeah, the thermal resistance in these is most assuredly much higher then real Cree emitters. In fact I would say that is by far the largest difference. If the thermal performance was better I bet it could match the XP-G2 that is should be competing against.
The reflow should be good, same way I do all of mine although anything i possible. If I feel like it I might run this one up till it pops and replace it with another and then test the new one.
What heat sink are you attaching it to for the tests?
One of my many CPU heatsinks from the parts box. It is rated for a 90W CPU and I have ran it quite a bit higher then that back in the day.
It keeps the temps down so I can easily hold my fingers on it, I have not measured the exact temps though.
Thanks, that definitely puts the onus on the die.
Even with the drop off, up until then it has a lower Vf and slightly higher lumen output than the XPG-2, at least when well sinked. The question is whether the higher thermal resistance will nullify that in the less perfect sinking of a host where heat does build more quickly.
I’ve got a couple of cheap “diving” lights that had magnetic switches and Latticebright XML clones. The drivers went out, so I put a regular buck driver into one of them, since the battery tube was 2S and the emitters were wired parallel. Several weeks later, I forgot that I’d taken the buck driver out and replaced it with a blank contact board, because the buck driver was acting weird. I put a couple 18650s in it and turned it on. To my surprise, the LEDs were fine that way. They lit up as normal, not even turning angry blue like the Real Cree’s do when you over power them. I played with it like that for a few minutes with no apparent adverse affects. Has anyone else ever seen a 3V Latticebright XML running on 8.4V and surviving?
Very strange that it survived at 8.4V. I have killed a few even with a single high drain cell actually, although they were in badly heat sinked hosts, still death was only a matter of seconds after turn on.
I think one of the largest issues with the LB LED’s is quality control. The quality varies massively from LED to LED. This can be clearly seen with tint variations it makes sense it would also apply to output as well.
i have some decent lb xm-l knockoffs.
i suspect they can make good stuff but like any other mfr the more expensive bins do not go in cheap lights.
was very surprised to get 3 matching pure white 5000k ish lb in a cheap light.