Is there a general consensus on how much current is too much to send through a XML LED? I did my second direct drive conversion on an HD2010 and used thicker wire along with a dedomed XML U3 on aluminum. The light is now running at 5.1 amps. That’s on a TF 5,000mah 26650. Am I going to burn out the led running at this current? Reduced life is fine as long as its more than a few minutes!
Yeah I know copper is better, but I’ve had this lying around for a while and thought I’d finally use it.
Get an infra-red thermometer thingy.
As soon as you turn off the LED after running for 5-10mins measure the temperature.
Compare to max junction temperature rating.
It's just an XM-L. 
Now, I don’t know if those amps are are the emitter or just tailcap that you measured.
But regardless, based on the graph an Xml only increases in lumens up till 4.4 amps directly to the emitter. This test was also done on a big heatsink too. But, through loss in amps through the system you are probably right at the max of an xml(point at which it doesn’t increase in lumens with more current).
I’ve seen many people drive xml’s at 5 amps and they don’t complain it last a few minutes, so I think your good. I know it’s obvious, but don’t leave the light on for very long. And you could measure the temperature after maybe 5 minutes with a infrared temp gun. Those things are pretty cheap, but so are xml’s.
It's all about thermal management.. Here's the other graph by Match showing an XM-L2 on a copper SinkPAD. Lumen output still rising at 6.2Amps.
Excellent graphs! Either way it doesn’t look too far out of line, but with the aluminum mounting it looks like a law of diminishing returns. Any more current and out put is going to nose dive, along with the lifespan of the LED.
When XML2-U3’s become a little more prevalent I will switch it over, along with copper. The XML2-T6 and XML-U3 aren’t very far apart when it comes to output. At least not enough to justify changing it out now.
The beam shot is pretty impressive, but I failed at loading images off my hard drive.
Yeah, those graphs are made by Match. He has a bunch of emitter tests on a ton of crees and even then nichia 219.
I myself am waiting for xml2’s to come down in price.
As PilotPTK pointed out, thermal management is the key.
Testing would indicate that closer to the LED, the thermal pathway is the most critical - as evidenced by the increase in luminous flux in copper star applications.
I would err on the side of caution, being mounted on AL especially. 5A is pushing it, I’d reduce it so about 4.05-4.1A (To get a solid 4A as the driver tends to keep a few ma for itself, as do the springs and wires. 4A on aluminum is going to be thermally equal to about 4.5A on copper, assuming good thermal contact, and adequately sized pill.
I have actually not blown up any led yet just by trying out high currents in a controlled way (it happens just by accident, like shorting something causing direct connection to the battery, then the current indeed can get way too high). What sooner happens is the led floating away because the solder melts, and even then if the led is fixed in place it still lights.