There was a thread here a couple of weeks ago that has disappeared. A vendor was offering XML2 U3 and some members were sent samples. Tom e was one I think. But where has the thread gone?
Marc.
I think he was banned for spamming? Not so sure…
The seller had a fake label.
Ok, but the problem with that is if anyone goes to his website and then tries to do any research, they won’t find out he was a scammer. 
Marc.
bestflashlights is the member's name I think. They're still around, but threads and posts keep getting deleted. Iirc the LED's were received and tested, but the results were too close to confirm if it was a U2 or U3.
Tom E’s testing showed his U3’s to have performance comperable to a T6
Oh, crap! Spent this morning finishing the testing.. Just spent 10 mins searching for the thread, and you guys are right -- it's totally gone, and so are other threads be posted.
HHmmmmmmm.......
Let me summarize the test procedure and results:
- used a standard, stock Small Sun T08 host to test with because of it's stable low amps (1.95A), 2 parallel cells, amps can be measured by simply jumper'ing across the switch, easy to work with
- since the 2 samples came on standard 20mm aluminum MCPCB's, I stayed with that for all comparative testing, accept the last XM-L2 U2 tested on a Noctigon
- measured battery voltage and amps on each run to ensure consistency
Results (lumens 1st number is at start, 2nd number is at 30 secs):
| Emitter | Source | Lumens (bare in host) | Lumens w/reflector |
| XM-L U3 1A (16mm star) | Kaidomain | 258-241 | --- |
| XM-L2 T6 0D | CNQualityGoods | 265-258 | --- |
| XM-L2 U2 1A (#1) | Kaidomain | 255-252 | 673-656 |
| XM-L2 U2 1A (#2) | Kaidomain | 255-248 | 673-660 |
| XM-L2 U3 ?? (#1) - maybe 1B? | willbuying.com | 262-255 | 680-663 |
| XM-L2 U3 ?? (#2) - maybe 1C? | willbuying.com | 248-241 | 656-642 |
| XM-L2 U2 1C (#1) | IlluminationSupply | 272-269 | 717-704 |
| XM-L2 U2 1C (#2) | IlluminationSupply | 272-269 | 711-694 |
| XM-L2 U2 1A /20mm Noctigon | Intl-Outdoor | 286-282 | 765-758 |
Notes:
- the lumens reads low for the "bare in host" -- probably because the inside of the head is black and side radiating beams are absorbed, however, results are pretty consistent with the reflector #'s
- Sorry, didn't measure the XML U3 or the XM-L2 T6 0D with the reflector
Conclusions:
- hard to believe the supposed XM-L2 U3's are in fact U3's. They are consistently lower than known good U2's (from IS and IOS). I'm thinking they are T6's.
- another scary result is the poor performance of the KD U2 1A's - I don't feel they are true U2's based on these test results. They are probably T6's as well
- I didn't expect the IOS U2 1A to do as well as it did. Didn't think a copper star would make much a difference at 2 amps based on Match's and relic's test results. So, what could be going on there? Is a true 1A that much brighter than a 1C? Maybe a better reflow? Didn't check but maybe the LED sits a bit higher getting better reflector #'s but that doesn't explain the better bare #'s.
Ok, I'm really like'n Calvin/Craig and Hank's offering on emitters, as suspected before this....
Thanks Tom! Really appreciate your time testing this. What I don’t appreciate are threads being removed when someone is trying to provide this kind of info. Sorry sb, but maybe we need to look at a way to keep a thread alive without the spammer rather than wiping the lot.
Marc.
At least we now know he is a spammer and a liar!
Marc.
Nice work Tom E as usual. Thanks for going to all the effort. I agree 200% that you usually get what you pay for and the supplier counts for everything. I’ve had XP-G leds That blew supposedly XP-G2 leds out of the water. Buyer beware.
are not xp-g2 leds a different pattern on the emitter?
yep, but you can get good high output xp-g’s and crap low output xp-g2’s
Thanks Tom, I didn't get to see your original post before it was deleted.
Agreed! Ban, lock and edit is better so we can see what happened previously.
Thanks for the tests Tom. I think this is the only way one can tell if the LED’s are rated correctly, and the more samples the better. 
Do you have any plans to test these LED’s: http://www.aliexpress.com/store/group/CREE-XML2-LED/330416_251473169.html ?
Thanks again.
Well, really don't want to get into the emitter testing biz, too much of a PIA in time and expense, though think it's beneficial to everyone -- I vote for relic38! 
Actually looked at this site and it's all T2-T6's, no U2's, though prices look real good but didn't check out shipping. Big risk to take with an unknown seller to buy qty from. Also it's a PIA to buy them on stars you will be throwing away, at least for me, because copper direct path is the only way to go -- alum is a total waste now. I'll probably never be using aluminum again, ever.
Yeah Hank offering all on noctigon is awesome.
I'm only using aluminum when I CAN'T use copper. Like the pill on my Xeno e03 that only fits a 14mm star, or the Trustfire Z2 that probably takes an even smaller star.
I suppose I shouldn't say that I can't because I could get a custom copper Sinkpad made if I wanted a thousand stars. Anyone want 999 14mm copper Sinkpads?
Well, guys have been cutting down the 16's to 14 - that's what they were designed for, though the thickness may be an issue, but they can be sanded down fairly easily. I sand down all the SinkPAD 16's until the indent is totally gone, but never cut the diameter of one down.
Yeah, it does take a lot of time and money. We should have someone like HKJ to test emitters as well. 
lol what
(XML2 T6 all-copper direct drive module for 3x NiMh AA Minimag LED)